history of the 1997 economic census · history of the 1997 economic census issued july 2000...

History of the 1997Economic Census

U.S. Department of CommerceEconomics and Statistics AdministrationU.S. CENSUS BUREAU

Issued July 2000

POL/00-HEC

This history was prepared by Michael A.Hovland and Jason G. Gauthier, underthe direction of William F. Micarelli,Chief, History Staff.

The following individuals providedvaluable information and/or reviewedthe manuscript: Irma F. Harahush,Lillyana Najafzadeh, Valerie A. Strang,Ewen M. Wilson, and Sheldon Ziman,Company Statistics Division; Shirin A.Ahmed, Larry A. Blum, Bernard J.Fitzpatrick, Kathleen V. Friedman,Paul S. Hanczaryk, James E. Kristoff,Robert A. Marske, Douglas J. Miller,Jeannette K. Mon, Dennis Shoemaker,Mark E. Wallace, and Paul T. Zeisset,Economic Planning and CoordinationDivision; Pamela Coan and Mark I.Kronisch, Economic Statistical Methodsand Programming Division; Patricia A.

Angus and Robert W. Marx, GeographyDivision; William G. Bostic, Jr.,Tamara A. Cole, and Judy Dodds,Manufacturing and Construction StatisticsDivision; Carol A. Ambler, Ruth AnnBramblett, John L. Fowler, andKimberly P. Moore, Services SectorStatistics Division.

Kim D. Ottenstein, Bernadette J.Gayle, Crystal M. Pate, Theodora S.Forgione, and Laurene V. Qualls of theAdministrative and Customer ServicesDivision, Walter C. Odom, Chief,provided publications and printingmanagement, graphics design, composi-tion, and editorial review for print andelectronic media. General direction andproduction management were provided byMichael G. Garland, Assistant Chief, andGary J. Lauffer, Chief, PublicationsServices Branch.

DedicationJohn P. Govoni

March 6, 1943 - January 21, 2000

The 1997 Economic Census was the last to benefit from the leadership ofJohn P. Govoni, Chief, Economic Planning and Coordination Division.

The history of this census is just a small part of the history of John’s tremendouscontribution to the Economic Programs Directorate, to the U.S. Census Bureau,

and to the people the world over whose lives he touched throughout

a long and inspired career of public service.

ACKNOWLEDGMENTS

History of the 1997Economic Census

Issued July 2000

POL/00-HEC

U.S. Department Of CommerceWilliam M. Daley,

Secretary

Robert L. Mallet,Deputy Secretary

Economicsand Statistics

AdministrationRobert J. Shapiro,Under Secretary for

Economic Affairs

U.S. Census BureauKenneth Prewitt,

Director

Nancy A. Potok,Principal Associate Directorand Chief Financial Officer

Gerald W. Gates,Chief, Policy Office

William F. Micarelli,Chief, History Staff

ECONOMICS

AND STATISTICS

ADMINISTRATION

Economicsand StatisticsAdministration

Robert J. Shapiro,Under Secretaryfor Economic Affairs

U.S. CENSUS BUREAU

Kenneth Prewitt,Director

William G. Barron,Deputy Director

CHAPTERS

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12. The Players and the Plans . . . . . . . . . . . . . . . . . . . . . . . . . 73. Introduction of the North American Industry Classification System

(NAICS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134. Questionnaire Design and Development . . . . . . . . . . . . . . . . . 255. Promoting Census and Product Awareness . . . . . . . . . . . . . . . . 356. Geographic Area Coding . . . . . . . . . . . . . . . . . . . . . . . . . 477. Preparations for Census Mailout . . . . . . . . . . . . . . . . . . . . . 578. Data Collection and Processing . . . . . . . . . . . . . . . . . . . . . . 719. The 1997 Economic Census in the Outlying Areas . . . . . . . . . . . . 8710. Transportation Surveys . . . . . . . . . . . . . . . . . . . . . . . . . . 9511. Economy-Wide Programs . . . . . . . . . . . . . . . . . . . . . . . . . 10512. Publication Program. . . . . . . . . . . . . . . . . . . . . . . . . . . . 11713. Completing the Census Bureau’s Mission Through Data Integration . . 127

APPENDIXES

A. Provisions of Title 13, United States Code, Relating to the 1997Economic Census. . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1

B. Historical Development of the Economic Census of the United States . B-1C. Principal Advisory Committees on the 1997 Economic Census . . . . . C-1D. Consultation on the Census Questionnaires . . . . . . . . . . . . . . . D-1

CONTENTS

Contents iiiHistory−1997 Economic Census

Chapter 1. INTRODUCTION

What Is the Economic Census? . . . . . . . . . . . . . . . . . . . . . . . . . 2

Legal Authority . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

The Evolution of the Economic Census. . . . . . . . . . . . . . . . . . . . . 2

What’s New About the 1997 Economic Census? . . . . . . . . . . . . . . . . 3The North American Industry Classification System (NAICS) . . . . . . . . 3Expanded Role of Electronic Technology. . . . . . . . . . . . . . . . . . . 3

Major Users and Uses of Economic Census Data. . . . . . . . . . . . . . . . 4

Data Collection and Basic Content . . . . . . . . . . . . . . . . . . . . . . . 4

How the Data Were Classified . . . . . . . . . . . . . . . . . . . . . . . . . 5

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

CONTENTS

Chapter 1 1History−1997 Economic Census

Introduction

WHAT IS THE ECONOMIC CENSUS?

The economic census can be defined as the systematicmeasurement of the following:

• Almost all economic activity in the U.S. (During the1997 Economic Census, approximately 97 percent ofthe Gross Domestic Product (GDP) was covered. Onlyagriculture [1.6 percent of GDP] and a few other activi-ties [1.2 percent of GDP] did not fall under the purviewof the 1997 Economic Census.)

• Major business activity in Puerto Rico, the Virgin Islandsof the United States, Guam, and the Commonwealth ofthe Northern Mariana Islands.

• Business activity within the scope of special programs,including—the Survey of Minority-Owned Businesses,Survey of Women-Owned Businesses, Business Expendi-tures Survey, Vehicle Inventory and Use Survey, andCommodity Flow Survey.

The economic census is the major economic statisticalprogram of the United States. It constitutes the chiefsource of data about the structure and functioning of theNation’s economy, and provides the foundation and frame-work for a host of other statistical endeavors by publicand private sector alike.

LEGAL AUTHORITY

The 1997 Economic Census was carried out by the U.S.Census Bureau, an agency of the U.S. Department ofCommerce, under the authority granted by Title 13, UnitedStates Code—Census. (Appendix A shows Title 13, UnitedStates Code.) Title 13 directs the Census Bureau toconduct economic censuses every 5 years, for yearsending in ‘‘2’’ and ‘‘7.’’ The title’s language also makesresponse mandatory by business firms or individualsqueried by the census, establishes penalties for noncom-pliance, and requires that the Census Bureau maintain theconfidentiality of the information provided by individualrespondents.

THE EVOLUTION OF THE ECONOMIC CENSUS

The economic census has come a long way since 1810,when a few questions on manufacturing activity wereadded to the population census schedule.

The U.S. Constitution requires that a census of popula-tion be carried out every 10 years to apportion representa-tion of each state in the U.S. House of Representatives.Even as delegates to the Constitutional Conventiondiscussed various provisions James Madison—principalauthor of the Constitution—urged that the census be usedfor more than a head count. Nothing came of his initialrecommendations until 1810, after he had become Presi-dent. In that year, the census collected data not only onthe population, but also some data on manufacturingestablishments. Questions were asked about—

• The kind, quantity, and value of goods manufactured.

• The number of manufacturing establishments.

• The number of machines of various kinds used inselected types of manufacturing establishments.

From this modest beginning, the economic censusevolved into a major institution involving—

• The successive expansion in scope of business activitymeasured.

• An accompanying increase in the systematization ofdata collection.

• Advances in the systematization of data processing.

• Increasing standardization of formats for data presenta-tion and dissemination.

• Increasing realization of the potential for integratingthe data sets to disseminate to the data user, not justraw data, but value-added, integrated information—particularly via the Internet.

The few manufacturing questions asked in 1810burgeoned, by 1997, into a separate economic census,using 479 questionnaire variations to collect data from 3.7million companies representing over 5 million businessestablishments in 1,056 industry classifications. (SeeChapter 4, ‘‘Questionnaire Design and Development,’’ formore information on the census questionnaires.) Inaddition, through the use of administrative records, theCensus Bureau compiled data on 14 million businesseswithout paid employees (‘‘mom and pop’’ businesses) andon 1.5 million small employers.

2 Chapter 1 History—1997 Economic Census

U.S. Census Bureau, 1997 Economic Census

WHAT’S NEW ABOUT THE 1997 ECONOMIC CENSUS?

The North American Industry ClassificationSystem (NAICS)

The 1997 Economic Census collected, tabulated, anddisseminated the bulk of the data in terms of the newNorth American Industry Classification System (NAICS).(Chapter 3, Introduction of the North American IndustryClassification System (NAICS) describes the new system indetail). The NAICS was developed after extended negotia-tions among the United States, Canada, and Mexico. Thisinternational effort was undertaken to create an integratedindustrial classification system for use throughout NorthAmerica, providing comparable economic statistics for allthree national economies. The 1997 Economic Census wasthe first major statistical activity conducted by one of thethree signatory nations to collect and publish data interms of the new system.

In addition to providing comparable statistics, theNAICS was designed to articulate a consistent conceptualsystem for classification. Since industrial statistics requirethat information on outputs and inputs be used together,it was agreed that the NAICS measures of economic activ-ity would be ‘‘production-based.’’ Carving the economicactivity into 20 sectors, instead of the 10 sectors used bythe Standard Industrial Classification (SIC) system, theNAICS provided a more detailed and accurate measure-ment of industrial activity in the United States. This levelof detail within NAICS was particularly true for industriesproviding services. For example, gone was the old ‘‘Ser-vices’’ sector, as such, and in its place was a host of farmore specific industrial classifications for the servicesthat— today—account for 34 million jobs. The SIC, devel-oped in the late 1920s, was appropriate for measuring amanufacturing-based economy. The NAICS is more suitedto a services and information-oriented economy.

The NAICS divides economic activity into these sectors:Agriculture, Forestry, Fishing, and Hunting; Mining; Utili-ties; Construction; Manufacturing; Wholesale Trade; RetailTrade; Transportation and Warehousing; Information;Finance and Insurance; Real Estate and Rental and Leasing;Professional, Scientific, and Technical Services; Manage-ment of Companies and Enterprises; Administrative andSupport, Waste Management and Remediation Services;Educational Services; Health Care and Social Assistance;Arts, Entertainment, and Recreation; Accommodation andFood Services; Other Services [except Public Administra-tion]; and Public Administration.

Figure 1-1.Employment by NAICS Sector

Sector 1997 employment

Not covered . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1,500,000Governments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18,000,000Agriculture. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,000,000Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547,590Utilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 711,053Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5,732,691Manufacturing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16,999,584Wholesale Trade . . . . . . . . . . . . . . . . . . . . . . . . . . 5,820,391Retail Trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14,115,559Transportation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,927,060Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3,221,988Finance and Insurance. . . . . . . . . . . . . . . . . . . . . 5,767,840Real Estate and Rental . . . . . . . . . . . . . . . . . . . . 1,761,179Professional/Scientific/Technical . . . . . . . . . . . . . 5,416,055Management of Companies . . . . . . . . . . . . . . . . 2,656,196Administrative Waste Mgmt. . . . . . . . . . . . . . . . . 7,439,208Educational Services . . . . . . . . . . . . . . . . . . . . . . 332,490Health Care and Social Asst. . . . . . . . . . . . . . . . 13,616,915Art/Entertainment/Recreation . . . . . . . . . . . . . . . 1,571,886Accommodation/Foodservices . . . . . . . . . . . . . . 9,496,136Other Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3,317,947

Figure 1-2, illustrates the basic structure of the NAICScoding. The first five digits are common to industrial activ-ity in the three participating nations. The NAICS, however,provides for a sixth digit in the code. This digit is for eco-nomic activity classifications that may vary from countryto country.

Figure 1-2.NAICS Coding Structure

NAICS Code Description

Sector 51 InformationSubsector 513 Broadcasting and telecommunicationsIndustry group 5133 TelecommunicationsIndustry 51332 Wireless telecommunications carriers, except

satelliteU.S. Industry 513321 Paging

Expanded Role of Electronic Technology

Other 1997 Economic Census innovations include asignificant expansion of the use of electronic technologyin data collection, processing, and dissemination. A fewexamples include—

• The DocuPrint technology used to print thecensus questionnaires. DocuPrint made it possible toprint simultaneously all forms—including question-naires, cover letters and inserts—that would be mailedto a single company. This eliminated the need for acomplex clerical assembly line, used in all previouscensuses, to assemble the mailing packages.

Chapter 1 3History—1997 Economic Census


• Computer Self-Administered Questionnaires(CSAQ). This newly developed technology, providingcompanies the capability of sending in their censusreports electronically—was extended well beyond 1992Economic Census capabilities. The Census Bureauauthorized a private company to develop a diskette-based questionnaire that would run on a personalcomputer.

• Standardization, across economic sectors, of thetable formats. Increased consistency in data presenta-tion greatly assisted data users in making nearly instan-taneous comparisons across databases.

• The expanded use of the Internet. The Internet wasused, among other purposes, to promote awareness ofthe census at the data collection stage and to dissemi-nate data products over the ensuing years of the censuscycle.

• The development of new products, including newsoftware, for data users. In response to data userrequests for timely data with easy navigational capabili-ties, the Census Bureau augmented its electronic publi-cation system and products. The majority of data fromthe 1997 Economic Census were released throughimproved CD-ROMs and on the Internet, with most dataavailable free of charge.

As part of the increased use of electronic technologyin the 1997 Economic Census, the issuance of paperproducts was reduced. The Census Bureau bridged thetransition from hard copy to electronic media by providingusers with the capability to access page images via por-table document format (pdf) files. Users with the freeAdobe Acrobat reader could print high quality pages forall key reports that—for earlier economic censuses—wereissued in print.

For previous censuses, the production processes lead-ing up to the printed reports delayed data release, some-times, for months. Releasing data electronically gave usersfar more immediate access—often on the day of release.

MAJOR USERS AND USES OF ECONOMIC CENSUSDATA

Federal Reserve Board Chairman, Alan Greenspan,called the economic census ‘‘indispensable to understand-ing the American economy.’’ Economic census data arestudied and recombined into a vast array of patterns toyield essential information for government, business,industry, and the general public. The following were majorusers of economic census data:

• The Federal Government uses the data for compositemeasures such as the Gross Domestic Product (GDP),input-output measures, production and price indexes,and for other statistical series that measure short-termchanges in economic conditions.

• Policymaking agencies of the Federal Government usethe data for, among other things, monitoring economicactivity.

• State and local governments use the data to assessbusiness activities and tax bases within their jurisdic-tions and to develop programs to attract newbusinesses.

• Business firms compare their own internal statisticswith published figures for their industry or geographicarea. Then, they compute market shares and evaluatetheir own growth and product lines relative to how, andwhat, other businesses are doing.

• Manufacturers and distributors that sell their productsto (or through) other businesses use the data to studythose markets to plan sales territories; locate plants,warehouses, and stores; and make economic or salesforecasts.

• Small businesses use the data about their markets ortheir industry to enhance their business opportunitypresentations to bank officers or venture capitalists, asthey seek funds for expansion.

• In colleges and universities, the data are used to teachmarketing and business management.

• In research units of all types, researchers apply the datain their study of long-term economic trends.

• Trade associations, business journals, and the dailypress widely disseminate facts derived from the eco-nomic census. Business magazines use census statisticsas technical background for articles.

Census statistics are available in most major public andcollege libraries designated as government depositories.Information services specialists in the Census Bureau’sregional offices (located in 12 large cities) and the staffs ofthe Census Bureau’s State Data Center and Business andIndustry Data Center programs throughout all 50 states,the District of Columbia, Puerto Rico, the Virgin Islands ofthe United States, Guam, and the Commonwealth of theNorthern Mariana Islands have economic census reports.Finally, the Department of Commerce’s International TradeAdministration staff answers questions about the availabil-ity and uses of economic census data.

DATA COLLECTION AND BASIC CONTENT

For the 1997 Economic Census, questionnaires weremailed out in December 1997 to about 3.7 million compa-nies to collect data on 5 million business locations. TheCensus Bureau requested that responses be mailed tothe National Processing Center ((NPC)—the NPC was theData Preparation Division until the Census Bureau’sreorganization effective May, 1998) in Jeffersonville, IN,by February 12, 1998.



Responding to the census can be particularly burden-some on small businesses. To save very small businessesthe expense and effort of filling out census question-naires, the Census Bureau used administrative recordsprovided by the Internal Revenue Service (IRS) and theSocial Security Administration (SSA) to obtain basic busi-ness information—such as payroll, total receipts, and thelike—for approximately 1.5 million small employer and 14million nonemployer businesses. (The sharing of recordsby these agencies was strictly one-way; IRS and SSA pro-vided their records to the Census Bureau, but the CensusBureau did not share its respondents’ data with any otheragency.)

Economic census data are collected at the level of the‘‘establishment.’’ An establishment is defined as a businessor industrial unit at a single-physical location that pro-duces or distributes goods or that performs services. Anexample would be a factory or a store. Each ‘‘establish-ment’’ is owned by a ‘‘company.’’ A company may own justone establishment (a ‘‘single-establishment’’ company) orit may own many establishments spread across differentgeographic or political jurisdictions, or across industrytypes. These are ‘‘multiestablishment’’ companies. A set ofcompanies under common ownership is an ‘‘enterprise.’’

Economic census questionnaires are mailed to companyheadquarters. One mailing package to a company head-quarters might contain scores or even hundreds of ques-tionnaires for all the company’s many establishments. Thequestionnaires requested basic information about eachestablishment, including—

• Kind-of-business activity (manufacturing, wholesale,retail, etc.).

• Physical location.

• Form of ownership.

• Dollar volume of business in 1997.

• Number of employees.

• Dollar amount of the payroll.

In addition to these core items, the questionnairesasked industry-specific questions (for, respectively, retail,wholesale, manufactures, and the like). For example, theCensus Bureau gave high priority to establishing productlines because of the need to compare domestic outputwith imports and exports.

HOW THE DATA WERE CLASSIFIED

After collecting the data, the Census Bureau classifiedeach data item into a category that made them manage-able for data users, namely—

• ‘‘Kind-of-business activity’’ translated into industrial cat-egories, also called ‘‘economic sectors,’’ such as manu-facturing, retail trade, and the like. For the 1997 Eco-nomic Census, this meant that the activity was

classified into one of the 20 NAICS-defined economicsectors. The classification of industrial data formed thestarting point for the Census Bureau’s Industrial Reportseries.

• ‘‘Physical location’’ translated into geographic catego-ries, and at the Census Bureau’s NPC the data were sub-classified into the following levels of geography:

• Nation.

• State.

• Metropolitan area.

• County.

• City or place.

• ZIP Code.

The classification of geographic data led, ultimately,to the Census Bureau’s Geographic Area Report series;however, not all data items were published for eachgeographical level. The most detailed 1997 EconomicCensus statistics were published only for the UnitedStates as a whole. Key data were published for states,metropolitan areas, counties, and places with 2,500 ormore inhabitants. Selected data also were provided atthe five-digit ZIP Code level. The smaller the geographicarea, the less detailed were the data released for thatarea. The Census Bureau follows this practice to makesure that the identity of the individual businesses thatprovided the data cannot be inferred.1

• ‘‘Form of ownership’’ translated into categories oforganizational structure, namely, single- or multiestab-lishment companies or enterprises.

SUMMARY

Each time economic data have been collected in acensus—from 1810 to 1997—there has been a progres-sion toward—

• Expansion in scope or content of business activitycovered.

• Increasing systematization in data collection, process-ing, presentation, or dissemination.

1For example, if there were just one auto dealership in a par-ticular ZIP Code, it would violate the confidentiality of theanswers provided by the dealership to publish payroll or salesreceipts data on ‘‘Auto dealers by ZIP Code.’’ In order to keep indi-vidual business identity confidential, in this example auto dealer-ship data would need to be published for a larger geographicentity than for a ZIP Code. Perhaps countywide or even statewidedata would be the only statistic that could protect the identity ofthe auto dealership. At the opposite end of the spectrum, ofcourse, would be the publication of sales receipts for auto dealer-ships nationwide. Since no one can infer a single dealership fromdata covering an entire Nation, national level data can be pub-lished in great detail.



• Further integration of data sets to provide the user withgreater value-added.

The 1997 Economic Census represents a substantialleap forward on most of these dimensions. Where appro-priate, these advances will be emphasized throughout theHistory of the 1997 Economic Census. One fact is certain—

1997 was the year for a state-of-the-art U.S. EconomicCensus to close out both the century and the millennium.2

2In November 1997, the Government Information QuarterlyJournal published a special issue on the 1997 Economic Census.It can be found at www.census.gov/econ97 under ‘‘GovernmentInformation Quarterly: Symposium on the 1997 Economic Cen-sus.’’ The Journal contains six articles with extensive informationon how this census was planned and conducted.



Chapter 2. THE PLAYERS AND THE PLANS

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

The Economic Directorate Reorganization . . . . . . . . . . . . . . . . . . . 8

Consultation on the Census . . . . . . . . . . . . . . . . . . . . . . . . . . 9The Census Advisory Committee of Professional Associations . . . . . . . 9Survey of 1992 Economic Census Data Users . . . . . . . . . . . . . . . . 9

The Economic Census Planning Teams . . . . . . . . . . . . . . . . . . . . . 9Organization and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . 9‘‘Must Do’’ Lists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Planning Team Recommendations . . . . . . . . . . . . . . . . . . . . . . 10General information . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Products and data user needs . . . . . . . . . . . . . . . . . . . . . . . 10SSEL improvements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10Data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Multiestablishment processing improvements . . . . . . . . . . . . . . 11Post-collection activities . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Budget Concerns and Contingency Planning. . . . . . . . . . . . . . . . . . 11

CONTENTS


The Players and The Plans

INTRODUCTION

While an economic census is conducted every 5 years,preparations for each enumeration start well before theprevious one is finished. Typically, preliminary planningfor an upcoming census begins at about the time theprincipal data collection effort for the preceding censushas been completed, and may continue (due to changes inperceived data publication requirements, available fundingor even technological advances) well into the operationalphase of the census.

In planning and conducting the economic census, theU.S. Census Bureau must weigh the data needs of theFederal government and other public and private datausers against the ability and willingness of private compa-nies to provide the information requested. Every censusplanning effort involves extensive review of the datacontent of the questionnaires to make certain that onlythe most urgently required information is collected, whileresponse burden is kept to a minimum. (Chapter 4, ‘‘Ques-tionnaire Design and Development,’’ treats this subject.)

Costs are another major consideration. This wasparticularly true for the 1997 Economic Census: thecontinuing effort by the Administration and the Congressto reduce Federal spending had a significant impact on theCensus Bureau’s activities. This led the 1997 EconomicCensus planners, at one point, to prepare several alterna-tive census plans based on the possible reduced appro-priation levels. Continued funding uncertainty, from fiscalyear to fiscal year, at times led to delays in finalizingplans.

A significant new factor that had to be considered inplanning the 1997 census was the decision by the Officeof Management and Budget (OMB) to develop and adopt anew industrial classification system—the North AmericanIndustry Classification System (NAICS)1—to replace theStandard Industrial Classification (SIC) code. Since classifi-cation of economic activity is at the heart of the statisticaltabulations derived from the census, the Census Bureauneeded to integrate the new classification system at everylevel of the census, from mailout to publication.

THE ECONOMIC DIRECTORATE REORGANIZATION

In June 1994, the Census Bureau reorganized theEconomic Directorate to facilitate conducting both the1997 Economic Census and the agency’s current economic

programs. The Census Bureau established the EconomicPlanning and Coordination Division (EPCD) to provide cen-tralized planning. The EPCD was charged with (1) formu-lating and developing comprehensive plans and programsfor the Economic Directorate, as well as (2) coordinatingthe planning efforts and related activities for conductingthe census. These activities included mailout, data cap-ture, editing, tabulation, and the like.

The reorganization created a second new division—theEconomic Statistical Methods and Programming Division(ESMPD)—to support and improve economic census andsurvey programs. This division was responsible for over-seeing joint development and implementation of statisticalprogramming and survey methods. In addition, the ESMPDensured the consistent development and application ofstatistical methodologies, survey operations, automatedprocesses, and computer programs and systems through-out the Census Bureau’s economic statistical activities.

Also in 1994, the Census Bureau merged the Construc-tion Statistics and Industry Divisions to establish a newManufacturing and Construction Division (MCD). TheBusiness Division transferred its financial statistics func-tions and staff to the Agriculture Division (which becamethe Agriculture and Financial Statistics Division [AGFS])and Business Division was renamed Services Division(SVSD).

After the transfer of the agriculture census (and part ofthe personnel of the AGFS) to the Department of Agricul-ture,2 the remaining staff was reorganized once again. InJune 1998, the AGFS was renamed the Company StatisticsDivision (CSD). At the same time, the SVSD was renamedthe Service Sector Statistics Division (SSSD).

MCD and SSSD retained responsibility within the1997 census for developing plans for data content,questionnaire design, and publication of statistical datafor business enterprises falling within their respectiveareas. These activities were coordinated by the EPCD. TheCSD was primarily responsible for the 1997 Surveys ofMinority- and Women-Owned Business Enterprises, and the1997 Business Expenditures Survey—both components ofthe 1997 Economic Census industry-wide programs—as

1Chapter 3 describes the North American Industry Classifica-tion System (NAICS) in detail.

2Funding and responsibility for the Census of Agriculture wastransferred from the Census Bureau, Department of Commerce,to the Department of Agriculture (USDA), National AgriculturalStatistics Service (NASS) effective October 1, 1996. EffectiveFebruary 2, 1997, NASS also received the transferred programpositions and staff from the Census Bureau.



well as the Quarterly Financial Reports Survey and othercurrent financial statistics reports.

The Census Bureau’s Data Preparation Division (DPD),located in Jeffersonville, IN, was renamed the NationalProcessing Center (NPC) in a reorganization effectiveMay 1, 1998. The NPC was responsible for much of themailout, and handled receipt and check-in of responses tothe questionnaires, initial processing of questionnaires,and data entry. (Chapter 7, ‘‘Preparations for Census Mail-out,’’ treats these issues.) Computer editing, tabulation,disclosure analysis, and final preparations for data publi-cation were carried out by EPCD, ESMPD, and staffs of thesubject divisions at the Census Bureau’s headquarters inSuitland, MD. (Chapter 8, ‘‘Data Collection and Process-ing,’’ provides more information on these activities.)

CONSULTATION ON THE CENSUS

The Census Advisory Committee ofProfessional Associations

In planning any census, the Census Bureau seeks theadvice of a variety of experts, both within and outside theFederal government. The primary sources of advice for theCensus Bureau, from the nongovernment sector, are itspublic advisory committees. Prior to April 1994, theCensus Bureau had nine public advisory committees3 thatprovided recommendations and advice to the Director onvarious Census Bureau programs. Four of these commit-tees, one each made up of representatives of the AmericanStatistical Association (ASA), the American Economic Asso-ciation (AEA), the American Marketing Association (AMA),and farm and agribusiness organizations, were the mostactively involved in advising the Director on the economicand agriculture censuses.

Presidential Executive Order 12838 (signed in February1993 and implemented the following year) reduced thetotal number of advisory committees and reorganizedthose that remained. The order created a combinedadvisory committee drawing its membership from fourprofessional associations that previously had eachprovided a separate committee.

The new committee of professional associationsconsisted of representatives of the ASA, AEA, and AMA,together with the Population Association of America (PAA).Participating organizations submitted nominations ofrepresentatives for membership on the committee to theDirector of the Census Bureau. Terms of appointment were

3 years and could be renewed, although members gener-ally were limited to a maximum of 6 years continuousservice. No member received pay, other than per diem andtravel expenses, for attending meetings. The new commit-tee made recommendations on the scope, content, meth-odology, tabulation plans, and data product plans for the1997 Economic Census (as well as for the Census Bureau’scurrent survey programs and other facets of the agency’swork). The committee usually met twice a year. FromApril 1, 1994, when the new organization was imple-mented, through September 31, 2000, when the bulk ofthe 1997 Economic Census activities were completed, theCommittee met 14 times. (For more information on theadvisory committees, including a list of members duringthe census period, see Appendix C, ‘‘Principal AdvisoryCommittees on the 1997 Economic Census.’’)

Survey of 1992 Economic Census Data Users

In mid-July 1995, the Census Bureau surveyed 5,147users of economic and agriculture census data. Includedwere subscribers to census publications through theGovernment Printing Office (GPO), the economic programarea’s list of associations maintained by the (then-named)DPD, and the Census Bureau’s Customer Services’ list ofsubscribers to the 1992 Economic Census CD-ROMs.The survey also extended to State Data Center lead agen-cies, Census Bureau Information Centers, Business andIndustry Data Centers, and Census Bureau RegionalOffices, together with a 1-in-4 sample of State Data Centeraffiliates and a 1-in-10 sample of Federal DepositoryLibraries.

The survey was designed to identify users of economiccensus data, understand their product needs and uses,and obtain feedback to help the Census Bureau plan futureproducts and/or adjust services provided.

THE ECONOMIC CENSUS PLANNING TEAMS

Organization and Objectives

After the June 1994 reorganization of the EconomicDirectorate, the planning effort was expanded toward amore comprehensive redesign and reengineering of theeconomic census and its associated processes and prod-ucts. In September 1994, the Directorate establishedseveral cross-divisional planning teams, each assigned toevaluate and recommend improvements for a specifiedcensus program area or activity. These Economic PlanningTeams (EPTs), 15 in all, were to identify problems encoun-tered in the 1992 census and to recommend solutions.Their overall objectives were to—

• Improve the Census Bureau’s Standard Statistical Estab-lishment List (SSEL), the Census Bureau’s Business Regis-ter, which served as the primary source of businessestablishment addresses for the census mailing list andalso as the sampling frame for the current economicsurvey programs.

3A ‘‘public advisory committee’’ is any group of persons—notcomposed wholly of Federal employees or officials—organizedunder Government authority for the purpose of obtaining advice,recommendations, or other assistance. The Census Bureau’s advi-sory committees operated under the provisions of the FederalAdvisory Committee Act of 1973, which stipulatted that the Direc-tor of OMB prescribe administrative guidelines and managementcontrols.



• Improve the dissemination, as well as the usefulness of1997 Economic Census data products.

• Improve the tools available to the Census Bureau’sanalysts and programmers.

• Introduce a generalized processing system to eliminateredundancy among the subject areas.

Each team was given an objective and was asked toestimate the resources (in terms of the number of‘‘full-time employee’’ years) each proposal would require.The EPTs focused on identifying changes required forexisting processing systems that could be implemented byJanuary 1998—i.e., by the time respondents would startreturning their 1997 census questionnaires.

‘‘Must Do’’ Lists

During the latter half of 1994, representatives of EPCDand ESMPD met with subject division staff (CSD, MCD, andSSSD) to develop the ‘‘must do’’ lists. These were the listsof improvements or changes that were considered criticalto the success of the 1997 Economic Census. The EPTs, inturn, continued to work on the lists through January 1995.Ultimately, the lists from the subject divisions and theplanning teams amounted to dozens of proposals, rangingfrom narrowly specific suggestions to major changes inCensus Bureau procedures.

EPCD established an evaluation group composed of theDivision Chief and Assistant Division Chiefs to review anddetermine the value of each item, as well as the likelihoodthat the Economic Directorate could implement thechanges by January 1998. In considering each item, theevaluation group assumed that—

• Available resources would be limited.

• The Census Bureau would maintain the timing and qual-ity of its key current survey programs (i.e., the censuswould not have absolute priority over survey needs).

• Several census-related activities had an overriding‘‘must’’ priority. These included implementing the NAICSand correcting the systems programming problemsidentified in the 1992 census.

• Data infrastructure improvements would be necessary.

• The Census Bureau would improve the product offeringsfrom the 1997 Economic Census compared to the 1992census.

• Response burden must be reduced.

EPCD staff conducted its evaluations of the lists duringthe last quarter of 1994 through January 1995. In Febru-ary and March 1995, EPCD met with EPT leaders to discussevaluation results and to specify the direction of eachteam’s activities thereafter.

Planning Team Recommendations

General information. May 15, 1995 was the target datefor establishing the Economic Directorate’s processinggoals for the 1997 census. The principal recommendationsmade by the planning teams and implemented for the1997 Economic Census fell into five functional areas—(1) products and data user needs, (2) SSEL improvements,(3) data collection, (4) multiestablishment processingimprovements, and (5) post-collection activities— outlinedbelow.

Products and data user needs. The Census Bureauused the results of the July 1995 survey of 1992 censusdata users to develop plans for the 1997 Economic Censusdata products. The agency wanted to make data availablein a more timely manner, make them easier for its custom-ers to use, and increase the perception of the CensusBureau’s economic products as a unified whole rather thanas a disjointed set of discrete products.

The planning teams recommended that the censuspublication program focus on electronic data productsas its initial and primary dissemination media. In addition,the teams recommended that the Census Bureau imple-ment a print (and FAX) on demand capability for dataotherwise available only in electronic files. The teams sug-gested that data presentation be improved by standardiz-ing file formats across subject areas to the greatest degreepossible, and that data files be as comprehensive as pos-sible to preclude the need to use multiple files. Also, theEPTs recommended that the Census Bureau issue, by early1999, an Advance Report—at the trade sector level andusing national data on both the NAICS and SIC bases.

Using these recommendations, the ensuing productimplementation phase concentrated on fulfilling userrequirements, including the provision of—

• Timely and relevant data, particularly electronic datafiles.

• Electronic data products that ease the transition from‘‘hard’’ copy to new electronic media.

• User capability to quickly find and manipulate data formultiple sectors and census/survey years.

• Easy access to free and ‘‘for fee’’ data sets, togetherwith easy and secure electronic product purchasingprocedures.

SSEL improvements. Improvement initiatives for theCensus Bureau’s Business Register—the SSEL—includedediting, imputation, and multiestablishment processing.A primary objective of the planning teams was to improvedata infrastructure. This also would mean substantial costsavings, as well as improvement in both the timeliness ofvarious census activities and in the data publication pro-gram. The specific recommendations cited a strong needfor more effective editing and imputation of employment



and payroll data. Previously the Census Bureau’s approachhad been to spread this work across several processingsteps belonging to Business Register (that is, SSEL), activi-ties, the County Business Patterns program, and thesubject divisions’ census programs.

The teams also made suggestions that would havesubstantial short-term benefits while the Census Bureaucarried out the major redesign of the SSEL (scheduled forcompletion by 2002). These recommendations focusedupon quality assurance measures; user documentation;linkages for company organizations and longitudinal con-tinuity; improved interactive query tools; expanded cover-age of nonemployers, railroads, and municipal utilities;and the assignment of Congressional District geographiccodes.

In addition, the teams wanted to strengthen basicSSEL concepts to provide a basis for the comprehensiveBusiness Register redesign, particularly with regard toenterprise definition and sub-enterprise organizationalstructures and methods for maintaining them.

Data collection. The teams recommended using theCensus Bureau’s DocuPrint equipment for electronic print-ing of questionnaires and other materials for the 1997Economic Census multiestablishment mailout, remails,and follow-ups.4 Ultimately, this recommendation wasexpanded to include census mailing packages for single-establishment companies in the manufacturing andmineral industries.

To reduce response burden, the planning teamssuggested expanding electronic reporting. Electronicmedia were already in place, such as the electronic datainterchange, and computerized self-administered ques-tionnaires (CSAQ). The Census Bureau had previously usedthese methods, mainly for the Company OrganizationSurvey (COS) and the Annual Survey of Manufactures(ASM). The 1992 census had used electronic reporting ona limited basis; about 80 companies responded electroni-cally, mostly on magnetic tape. EPCD staff estimated thatnearly 600 respondents might elect to report their dataelectronically in the 1997 census. Ultimately, CSAQs wereextended to companies in retail; finance, insurance, andreal estate; and other sectors.5 In addition, the teamsrecommended that data could be accepted on magnetictape provided by respondents.

Multiestablishment processing improvements. Mul-tiestablishment processing had been a chronic bottleneckfor previous economic censuses. Completeness andcoverage edits were essential for maintaining up-to-datemultiestablishment coverage, but they complicated theworkflow and delayed data release. The EPTs recom-mended streamlining multiestablishment workflow by—

• Consolidating or eliminating steps in the processingflow.

• Focusing completeness and coverage edits on criticalproblems.

• Improving control and tracking systems.

• Imputing for multiestablishments earlier in the process-ing flow.

Post-collection activities. The teams’ principal recom-mendations on data processing and other post-collectionactivities involved the development and adoption of ageneralized edit system, as well as improvements to theCensus Bureau’s analytical capabilities. Time constraintsand the complexities of developing such a system for theentire census originally led planners to consider that itwas unlikely that one could be designed and fully imple-mented for the 1997—census the custom coding edits forall economic sectors would need to be revised as part ofthe implementation of the NAICS. Hence, the GeneralizedEdit and Imputation System—better known as ‘‘PlainVanilla’’—was developed. (Chapter 8, ‘‘Data Collection andProcessing,’’ elaborates on ‘‘Plain Vanilla.’’) It replaced thecomplex edits for the 1996 Annual Survey of Manufac-tures (ASM) and for the 1997 census.

The EPCD established an independent cross-divisionalplanning team in May 1995 to develop ‘‘Plain Vanilla.’’ Keyto the team’s effectiveness was the release of its membersfrom day-to-day operational responsibilities to enablethem to devote time to developing the edit system asquickly as possible.

To improve the agency’s analytical capabilities, theteams recommended that EPCD and ESMPD develop toolsfor analysts that would (1) enable them to access data setsacross programs, (2) use graphic data analysis to supple-ment conventional reviews, and (3) provide a commonmenu for these items across both the census and the cur-rent economic surveys. This would require a significantplanning and design effort. A new planning team wasorganized, composed of staff from the EPCD, ESMPD, andthe subject divisions. The team evaluated existing soft-ware systems and developed plans for acquiring andinstalling new highspeed, high-capacity personal comput-ers and the appropriate software.6

BUDGET CONCERNS AND CONTINGENCY PLANNINGFrom the latter part of 1993 into 1995, planning for

the 1997 Economic Census was based on the assumptionthat both the census and its funding would be comparableto that of the 1992 census. However, an increasedemphasis on funding reductions for all Federal agencieseventually weakened this assumption.

4This subject is covered in detail in Chapter 7, ‘‘Preparationsfor Census Mailout.’’

5Chapter 8, ‘‘Data Collection and Processing,’’ discusses elec-tronic reporting.

6The processing infrastructure upgrade is discussed inChapter 7, ‘‘Preparations for Census Mailout.’’



Pressures to hold down costs were constant, butbecame most severe during debate on the fiscal year (FY)1998 Federal budget. This was, in large part, due to pres-sure for a balanced budget agreement and the determina-tion of the political leadership of both the Administrationand Congress to reduce, and eventually eliminate, thebudget deficit. Congress considered significant ‘‘across-the-board’’ budget cuts for the Census Bureau and for the1997 Economic Census. These reductions would havecome at just the point in the census budget cycle whenfunding normally would have been sharply increased inorder to support final preparations for the economic cen-sus mailout, as well as for a major proportion of the datacollection phase of the census.7

Facing the possibility of large cuts in the requestedbudget for FY 1998, the Census Bureau’s executive staffdecided that, rather than undermine the statistical cred-ibility of the entire census, potential reductions would beaccommodated by eliminating discrete programs withinthe overall census. The principal casualty of the budgetreductions was the Enterprise Statistics Program, whichwas dropped from the 1997 Economic Census.

The Administration requested $63.7 million for the1997 Economic Census in its FY 1998 budget. However,based upon Congressional views with respect to theAdministration’s budget request, the Census Bureau pre-pared contingency plans for various alternatives: (1) fullfunding at the requested level ($63.7 million), (2) a $45million budget, and (3) a $35 million budget.

Fortunately, Administration and Congressional negotia-tors were able to reach an agreement that provided fund-ing at the $63.7 million level.

The expenditures for the 1997 Economic Census, byfiscal year, were as follows:

Fiscal year Expenditure

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . $220,273,000

1995 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1,550,0001996 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26,386,0001997 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29,681,0001998 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62,129,0001999 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54,083,0002000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46,444,0007Previous budget constraints had already led the Census

Bureau to drop test plans for the 1997 Economic Census.



Chapter 3. INTRODUCTION OF THE NORTH AMERICAN INDUSTRYCLASSIFICATION SYSTEM (NAICS)

The Standard Industrial Classification Code (SIC) . . . . . . . . . . . . . . . 14The Development of the SIC Code . . . . . . . . . . . . . . . . . . . . . . 14Limitations of the SIC Code. . . . . . . . . . . . . . . . . . . . . . . . . . 14

The Economic Classification Planning Committee (ECPC) . . . . . . . . . . . 15

Joint Planning of the NAICS . . . . . . . . . . . . . . . . . . . . . . . . . . . 15The Need for Comparative Data . . . . . . . . . . . . . . . . . . . . . . . 15Design Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Industry Structure Agreements . . . . . . . . . . . . . . . . . . . . . . . . 16Schedule for Implementing NAICS . . . . . . . . . . . . . . . . . . . . . . 17

SIC and NAICS Compared . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Revisions of SIC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17New Sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17The NAICS Numbering System . . . . . . . . . . . . . . . . . . . . . . . . 19

Implementing NAICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Implementation Procedures . . . . . . . . . . . . . . . . . . . . . . . . . 19The 1997 Refile Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20The 1996 Company Organization Survey . . . . . . . . . . . . . . . . . . 20‘‘Birth’’ Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21Changes to the 1997 Census Questionnaires . . . . . . . . . . . . . . . . 211997 Economic Census Classification Questionnaire Mailout. . . . . . . . 21Processing System Modifications. . . . . . . . . . . . . . . . . . . . . . . 21

Comparability of the Data and Maintenance of Time Series . . . . . . . . . . 22Scope of Data Comparability . . . . . . . . . . . . . . . . . . . . . . . . . 22Maintenance of Time Series . . . . . . . . . . . . . . . . . . . . . . . . . 22

NAICS and economic census time series . . . . . . . . . . . . . . . . . 22Publication plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22The reclassification of 1992 Economic Census data . . . . . . . . . . . 23

CONTENTS


Introduction of theNorth American Industry Classification System (NAICS)

THE STANDARD INDUSTRIALCLASSIFICATION CODE (SIC)

The Development of the SIC Code

The economic census is the principal source of detailedinformation on economic activity in the United States, butis not, and never was, the only source of that information,even within the Federal Government. One of the majorproblems with multiple sources of information is that theymay not, indeed probably do not, share the same conceptsand definitions of the information they collect and dis-seminate. Standardization in collecting and publishingeconomic statistics has, consequently, been a primary goalfor the Federal Government. To this end, in 1939 the Cen-tral Statistical Board of the Bureau of the Budget (BOB—thepredecessor of today’s Office of Management and Budget[OMB]) established an Interdepartmental Committee onIndustrial Classification, together with a Technical Sub-committee on Industrial Classification, and directed themto develop a standardized list of industries that could beused by all agencies or organizations and could facilitatecomparisons of industrial data produced by differentagencies. By 1940 the committees had produced separatedraft lists and descriptions of industries and alphabeticalindexes for manufacturing and nonmanufacturing indus-tries. After review, the BOB published the first StandardIndustrial Classification (SIC) Manual for manufacturingindustries in 1941. In the following year, the BOBpublished a similar manual for the nonmanufacturingindustries. The dual classification systems continued inuse until 1954, when they were combined into a singleManual, which was periodically revised and reissuedthrough 1987. The 1987 revision was used for the 1992Economic Census.

Limitations of the SIC Code

The SIC Code system was developed during a timewhen the bulk of the American economy was involvedwith manufacturing. After World War II, the Americaneconomy not only began a prolonged period of expansion,but became more and more complex. New industriesappeared in the manufacturing sector, as did entirely newareas of commercial economic activity. The OMB and con-cerned agencies (including the Census Bureau, the Bureauof Economic Analysis [BEA], and the Bureau of Labor Statis-tics [BLS]) updated the SIC Code approximately every 10years to try to compensate for the dislocations within the

economic statistical system caused by the developmentand growth of these new industries and activities.Nevertheless, the basic structure became increasinglyobsolescent.

The problem was exacerbated by the explosion of newindustries that appeared in the 1970s. Collecting andtabulating economic statistics became increasinglydifficult when more and more economically active andimportant establishments could not be defined using theSIC Code. Even before the 1992 Economic Census, anoma-lies within the SIC Code had become painfully obvious; forexample—

• While technology and the economy moved rapidly toimplement and expand electronic information dissemi-nation and telecommunication, the SIC placed printpublishing activities in the Manufacturing Division,online services in the Services Division, and communica-tions (including radio and television broadcasting, cabletelevision services, paging services, and satellite masterantenna systems services) in the Transportation,Communications, Electric, Gas, and Sanitary ServicesDivision.

• The SIC industry 7389, Business Services, Not ElsewhereClassified, included everything from bronzing babyshoes, to telemarketing, to sound recording studios.

• Satellite communications systems were not recognizedby the SIC Code.

• At a time when the Nation’s medical care system was amajor focal point of public and economic interest, the1987 SIC lacked any classification that separately identi-fied health maintenance organizations (HMOs).

The SIC Code’s bias toward manufacturing industriescaused considerable dissatisfaction as well. The systemdid not reflect the accelerating trend within the economytoward growth in the nonmanufacturing sectors, particu-larly the various services industries. By the late 1980s,there was increased interest within the Federal Govern-ment and throughout the U.S. economy in improving theability of economists and statisticians to assess the impactof international trade and industrial practices on thedomestic economy. In addition, the growing need to relateU.S. output to that of Canada and Mexico under the pro-posed North American Free Trade Area (NAFTA) agreementrequired a major overhaul of the SIC.



THE ECONOMIC CLASSIFICATION PLANNINGCOMMITTEE (ECPC)

In November 1991, the Census Bureau sponsored theInternational Conference on the Classification of EconomicActivities, in Williamsburg, VA (the ‘‘Williamsburg Confer-ence’’), which focused on the need to reexamine the entireindustry classification system. The conference involvedparticipants from Federal and state agencies, academia,and the international statistical community, includingrepresentatives of Statistics Canada. Acting on recommen-dations from the Conference, the OMB, in March 1993,established the Economic Classification Policy Committee(ECPC) to consider a thorough overhaul of the classifica-tion system in place. The ECPC was chaired by the BEAand consisted of representatives from the BEA, the CensusBureau, and the BLS.

The Committee was charged with carrying out a ‘‘freshslate’’ examination of economic classifications for statisti-cal purposes, including the following:

1. Identifying the essential statistical uses of economicclassifications.

2. Identifying and developing, if needed, economic con-cepts, new structures, and statistical methodologiesthat address the essential statistical uses of economicclassifications.

3. Developing classification systems based on conceptsdeveloped under No. 2 above.

4. Planning the implementation of the new classificationsystem.

5. Ensuring an opportunity for public participation in thereview and planning process.

Between March 1993 and March 1994, the ECPCcompiled and released six papers addressing specificissues involved in the development of a new industrialclassification system—

1. ‘‘Conceptual Issues,’’ reviewed the purpose of aneconomic classification system, the need for a consis-tent conceptual framework for such a system, whatgrouped or classified data were needed, which classifi-cation approach or approaches should be adopted indesigning a new system, and the classification unit tobe used.

2. ‘‘Aggregation Structures and Hierarchies,’’described classification hierarchies and reviewed thepossibility of using multiple hierarchies and whetherany new classification system should have a flexibleaggregation structure.

3. ‘‘Collectibility of Data,’’ reported on the codingpractices and requirements of the U.S. statisticalagencies that would be the most frequent users of the

new classification system, together with the ability ofbusinesses to supply data, and the kinds of datarequired for reclassification under a new system.

4. ‘‘Criteria for Determining Industries,’’ reviewedthe criteria of economic significance, specialization,and coverage ratios in determining industry classifica-tion and considered alternative criteria.

5. ‘‘The Impact of Classification Revision on TimeSeries,’’ considered the need for historical continuity,the problems involved with time series continuity, theimpact of previous classification code revisions, andthe implications of a fundamental SIC revision.

6. ‘‘Services Classifications,’’ reviewed questionsrelating specifically to the services sector of theeconomy, including the boundary between ‘‘goods’’and ‘‘services,’’ the impact of production-orientedand/or supply-oriented classification for services,examples of market-oriented classification, questionsabout specific services industries (e.g., nonprofit orga-nizations), past approaches to classifying services,and the services classifications in the restructuredclassification system.

The ECPC asked for comments on each paper frominterested users and reviewed these comments as part ofits study.

In addition to the studies described in the issuespapers, the ECPC evaluated the U.S. and Canadian SICstructures to determine the conformance of the respectivesystems to economic concepts. To this end, the Commit-tee undertook a study of the ‘‘heterogeneity index’’ pro-posed for U.S. industries to decide whether establish-ments in the old SIC industries met the conditionsrequired for the production-oriented classification systemdescribed in ‘‘Conceptual Issues,’’ Paper No. 1.1

JOINT PLANNING OF THE NAICS

The Need for Comparative Data

The SIC Code system was not compatible with industryclassification systems used in other countries, and ana-lysts had difficulties comparing industrial characteristics,trends, and developments internationally. The adventof the North American Free Trade Agreement (NAFTA)spurred consideration of a new system that would bemore comparable with those of the other NAFTAsignatories.

1The results of the joint U.S. Canadian SIC study were releasedin two reports, one covering the U.S. and the other the CanadianSIC Codes. The U.S. report was ECPC Report No. 1, ‘‘EconomicConcepts Incorporated in the Standard Industrial ClassificationIndustries in the United States.’’ The Canadian SIC was covered in‘‘The Conceptual Basis of the Standard Industrial Classification,’’published by Statistics Canada. ECPC Report No. 2, ‘‘The Hetero-geneity Index: A Quantitative Tool to Support Industrial Classifica-tion’’ evaluated U.S. industries’ new ‘‘index of heterogeneity.’’



Statistics Canada, a participant in the Williamsburgconference, also planned to revise its 1980 SIC. Shortlyafter the establishment of the ECPC, the Canadians joinedthe discussions on restructuring the respective nationalclassification systems and creating a joint system, fol-lowed thereafter by Mexico’s Instituto Nacional de Estadis-tica, Geografia e Information (INEGI). As a result of thesetalks, the OMB announced in the July 26, 1994, FederalRegister, that the statistical agencies of the North Ameri-can countries had agreed to develop a North AmericanIndustry Classification System (NAICS) to introduce com-mon industry classifications for all three countries.

Design Principles

Planning an integrated industry classification systemrequired agreement on a basic conceptual framework.Statistics Canada, INEGI, and the ECPC agreed to developand put in place a single system based on the followingmajor principles:

• Using industrial statistics requires that information onoutputs and inputs be used together. The NAICS shouldconform to a production-oriented economicconcept—i.e., industries that use like processes to pro-duce goods or services should be grouped together.

• The respective national statistical systems should com-pile product data to meet the need for market-orientedor demand-based groupings of economic information.Each country agreed to improve commodity classifica-tion systems, including the Harmonized System of theCustoms Cooperation Council and the United Nationsprovisional Central Product Classification system.

• The long-term goal of the NAICS should be to imple-ment a production-oriented framework for economicclassification. The framework would be used in 1997and subsequently in reviewing changes to the existinglist of industries.

• Any new system should include developing classifica-tions for (1) new and emerging industries, (2) servicesindustries in general, and (3) industries engaged in theproduction of advanced technologies.2

• Time series continuity should be maintained to theextent possible, but changes in the economy and evolv-ing user needs must be taken into account. Proposalsrelating to all parts of the classification will be consid-ered, so long as they are supported by a rationale andby factual information that furthers the long-term goalsof NAICS.

• The participating countries’ statistical agencies agreedto conduct a detailed review of their current industrydefinitions to determine where differences existedbetween the three systems, and to move toward fullcommonality and the implementation of production-oriented reasoning into the new classification system.

• The participants agreed to try to develop a NAICScompatible at the two-digit level to the current Interna-tional Standard Industrial Classification of All EconomicActivities (ISIC, Revision 3) of the United Nations.

• When changes were required to classifications to attaininternational comparability, those changes would bemade in the direction of the country or countries whoseexisting classification definitions most closely corre-sponded to the production-oriented concept.

The ECPC established seven subcommittees—agricul-ture; mining and manufacturing; construction; distributionnetworks; finance, insurance, and real estate; businessand personal services; and health, social assistance, andpublic administration—composed of economists, statisti-cians, and classification specialists representing 19Federal agencies that used the SIC in their work. The ECPCalso established a U.S. Coordinating Committee to coordi-nate the work of the subcommittees with that of INEGI andStatistics Canada.

Industry Structure Agreements

The ECPC subcommittees and representatives of INEGIand Statistics Canada developed the structure of theproposed NAICS in a series of meetings. Proposals for indi-vidual industries from all three countries were consideredfor acceptance if the recommended industry was based onthe production-oriented concept required for the generalsystem. As the subcommittees considered and agreed ongroups of subsectors of NAICS, the ECPC published theindustries proposed for inclusion in those subsectors.Between May and December 1995, a total of 31 separateagreements were developed and approved by the partici-pating statistical offices. Each agreement described thesector and the industries included in the sector, outlinedthe limitations and constraints of the classification, relatedthe NAICS industries to the United Nations’ ISIC (Revision3), and summarized changes to the respective nationalclassification systems.

Agreement 18, for example, covered NAICS sector 51,‘‘Information.’’ The ‘‘Information’’ sector drew many ofits industries from SIC divisions D, ‘‘Manufactures’’; E,‘‘Transportation, Communications, Electric, Gas, andSanitary Services’’; and I, ‘‘Services.’’ Some 20 newindustries were created in the NAICS Information sector,including—

• One new five-digit industry in the Publishing Industriessubsector (511) for database and directory publishers.

2‘‘Advanced technologies’’ included, but were not limited to,electronic components, telecommunications equipment, comput-ers and associated equipment, computer software, medical equip-ment, and advanced materials.



• A NAICS industry for software publishers was createdfrom part of the SIC industry 7372, ‘‘PrepackagedSoftware.’’

• Eleven five- and six-digit industries were created in the‘‘Motion Picture and Sound Recording Industries’’subsector (512).

• Eleven five- and six-digit industries were created in the‘‘Broadcasting and Telecommunications’’ subsector(513).

The individual agreements were approved by therespective national authorities in five groups, the first (ofthree agreements) on May 19 and the last (for a singlesector—finance and insurance) on December 13, 1995. Asthe agreements were approved, the OMB published theproposed industries for those sectors and subsectors andasked for public comment in the Federal Register. Fivenotices were published between May 1995 and November1996, with final comments requested by December 20,1996. (The OMB received approximately 400 comments onthe proposed new classification system, most involvingsuggestions for additional industry classifications, or titlechanges for industries covered by NAICS.)

The ECPC submitted its final proposals for NAICS to theOMB based on the proposals published in the last (Novem-ber 5, 1996) Federal Register notice. In general, the OMBaccepted ECPC’s recommendations, with a number ofrelatively minor exceptions. These exceptions typicallyinvolved renaming or renumbering specific industries. Inthe manufacturing sector, however, NAICS subsector 337,‘‘Furniture and Related Product Manufacturing,’’ wasrevised to emphasize furniture manufacture by type offurniture, rather than by material (i.e., wood, metal, andother materials).

Schedule for Implementing NAICS

Having designed a new industry classification system,the three countries faced the task of implementing thatsystem. Delay would only exacerbate the problems inher-ent in the original system, so the participants determinedthat NAICS should be applied as soon as possible.

The Census Bureau agreed to implement the newsystem beginning with the 1997 Economic Census, whichwas scheduled to begin collecting data less than a yearafter the final approval of NAICS.3 The Census Bureauplanned to begin publishing advance general statisticsdata using the NAICS system early in 1999. Industry,geographic area, and subject series reports, using NAICSclassifications only, would be published beginning in the

second quarter of 1999. A final—report with bridgingtables to enable users to relate NAICS statistics to datapublished using the SIC Code—was scheduled for publica-tion in March 2000.

Several components of the 1997 Economic Census werenot to be converted to NAICS; these were the economiccensus in the outlying areas (Puerto Rico, Guam, the VirginIslands of the United States), the Commodity Flow Survey,and the Vehicle Inventory and Use Survey. The CensusBureau’s current economic surveys—programs e.g., the1998 Annual Survey of Manufactures (scheduled topublish using NAICS in March 2000), monthly wholesaleand retail trade surveys (both in 2001), 1999 AnnualCapital Expenditures Survey (Feb. 2001), Quarterly Finan-cial Reports (March 2001, for the fourth quarter of 2000),etc. would begin using NAICS classifications after the1997 census.

SIC AND NAICS COMPARED

Revisions of SIC

The SIC Code had been revised periodically since itsinception over 50 years ago. The most recent changes tothe SIC were made in 1987, when the system was modi-fied to identify new high-tech industries, particularly thoseinvolved with computers and other electronic equipment.These revisions, while updating the detail of the classifica-tion system left the core characteristic of the SIC Code—itsclassification of industries using a mix of production- andmarket-based (i.e., commodities or services that havesimilarities in use, that belong together, or are usedtogether for some purpose) categories—largely unaltered.The 1997 ‘‘revision’’ changed all this in an integratedeffort to create a unified classification system for theNorth American free trade area based on groupingtogether industries that used similar productionprocesses.

New Sectors

The new system that was created (in the 31 industrystructure agreements) grouped industries into 20 broadsectors, doubling the number of major groupings com-pared to the SIC Code system.4 Many of the ‘‘new’’ sectorswere recognizable as similar or modified versions of theold SIC divisions; for example, NAICS Agriculture (sector11), Mining (sector 21), and Construction (sector 23)sectors were virtually identical in coverage to the SICAgriculture (division A), Mineral Industries (division B), andConstruction (division C) divisions respectively.

There were, however, significant changes in otherareas. Under NAICS, the old Finance, Insurance, and RealEstate Division (H) was split into separate Finance andInsurance (52), and Real estate and Rental and Leasing

3The BEA and BLS also were scheduled to implement NAICS asquickly as possible; the earliest published data for BEA would bethe 1997 Foreign Direct Investment Benchmark Survey (to be col-lected 1999), whole BLS would begin releasing NAICS-based datain 2002 (for the 2001 Employment and Wages Report).

4The SIC Code classified industries in 10 divisions (A-J), withan eleventh division (K) used for unclassifiable establishments.



(53) Sectors; Wholesale Trade (43) and Retail Trade (44-45)were redefined. The most extensive changes involved theSIC Services Division I, which was broken up under NAICSand furnished industries to eight new sectors, includingfour (Education Services; Health Care and Social Assis-tance; Arts, Entertainment, and Recreation; and Other

Services, Except Public Administration) drawn entirelyfrom the former Services Division.

The NAICS sectors and their corresponding SIC divi-sions are as follows:

Table 3-1.NAICS Sectors and Their Corresponding SIC Divisions

NAICScode NAICS sector

SICcode SIC divisions making the largest contribution

11 Agriculture, Forestry and Hunting and Fishing (The Census of A Agriculture, Forestry and FishingAgriculture is conducted by the Department of Agriculture.)

21 Mining B Mineral Industries

22 Utilities E Transportation, Communications, and Public Utilities

23 Construction C Construction Industries

31-33 Manufacturing D Manufacturing

42 Wholesale Trade F Wholesale Trade

44-45 Retail Trade G Retail Trade

48-49 Transportation and Warehousing (The economic census excludes E Transportation, Communications, and Public Utilitiesthe U.S. Postal Service, large certificated passenger airtransportation, and all rail transportation.)

51 Information D ManufacturingE Transportation, Communications, and Public UtilitiesI Service industries

52 Finance and Insurance (The economic census excludes funds H Finance, Insurance, and Real Estateand trusts.)

53 Real Estate and Rental and Leasing H Finance, Insurance, and Real EstateI Service Industries

54 Professional, Scientific, and Technical Services (The economic I Service Industriescensus excludes landscape architecture and veterinaryservices.)

55 Management of Companies and Enterprises Parts of all divisions

56 Administrative Support; Waste Management and Remediation I Service IndustriesServices (The economic census excludes landscapingservices.) E Transportation, Communications, and Public Utilities

61 Educational Services (Census excludes elementary and I Service Industriessecondary schools, colleges, and professional schools.)

62 Health Care and Social Assistance I Service Industries

71 Arts, Entertainment, and Recreation I Service Industries

72 Accommodation and Food Services G Retail TradeI Service Industries

81 Other Services Except Public Administration (Census excludes I Service Industriespet care; labor, political, and religious organizations; and privatehouseholds.)

93 Public administration (Separate census of governments does not J Public Administrationpresent data according to the NAICS or SIC systems.)



The NAICS not only doubled the number of sectors/divisions, it also increased the number of defined indus-tries in the U.S. economy from 1,004 under the SIC to1,170. The new system defined over 350 new industriesnot previously identified separately (including 250 newservice industries), and revised the scope of over 330more. Under the SIC, 45 percent of industries defined hadbeen manufacturing activities. In NAICS, 473 of the totalof 1,170 defined industries (40.3 percent) still representedmanufacturing.

The NAICS Numbering System

The SIC Code used a four-digit numbering system inwhich the first two digits identified the major group, thethird digit the industry group, and the fourth the specificindustry. The Census Bureau had developed its ownsupplement to the SIC in the form of five-digit productclass codes, and seven-digit product codes, to enable itto apply SIC Codes to specific products. It also had createdsix-digit ‘‘kind-of-business’’ codes in the services (or non-goods producing) sectors.

NAICS planners decided that the first two characters ofthe new system should designate the sector (in NAICS, theterm ‘‘sector’’ replaced ‘‘division’’ as used in the 1987 SIC).They also thought that modern economies were too com-plex to be described adequately if the first character ofthe coding system restricted the number of sectors to 9 or10 as in the SIC. Therefore, NAICS used a six-digit code inwhich the first two digits designated the sector, the thirdthe subsector, the fourth the industry group, and the fifthNAICS industry. The five-digit industry level is the mostdetailed level at which comparable data will be availablefor all three participating countries’ economies. A sixthdigit identifies specific national industries—under NAICS,the United States has 659 such industries, i.e., those withsix-digit classifications unique to the United States. Thestructure may be illustrated as follows:

NAICSlevel

NAICScode Description

Sector 72 Accommodations and foodservices

Subsector 722 Food service and drinking places

Industry group 7222 Limited service eating places

Industry 72221 Limited service eating places

U.S. industry 722211 Limited service restaurants722212 Cafeterias722213 Snack and nonalcoholic beverage bars

The Census Bureau added additional digits for stillmore detailed classification in some industries. Forexample, for Sector 72, the six-digit ‘‘national’’ industry722213, ‘‘Snack and Nonalcoholic Beverage Bars’’ wasfurther broken down into three seven-digit industries:

7222131, ‘‘Ice Cream and Soft Serve Shops’’; 7222132,‘‘Frozen Yogurt Shops’’; 7222133 ‘‘Donut Shops.’’

IMPLEMENTING NAICS

Implementation Procedures

The adoption of NAICS for the 1997 Economic Censusrequired the Census Bureau to introduce about 300 newindustry classifications and to revise about the same num-ber of existing SIC industries. The extent of these changesmeant that implementing the new system required farmore complex procedures than had been necessary forprevious revisions of the classification system. Theseprocedures included the following:

• Developing a bridge code system using the 1987 SIC.The Census Bureau assigned two extra digits for eachpart of the 1987 SIC that had been reclassified to a dif-ferent NAICS sector or industry. This allowed previouscensus data collected and tabulated using the SIC to becompared with the 1997 census tabulations done usingNAICS.

• Conducting a ‘‘refile survey’’ in December 1996. Thisinvolved a mailout and follow-up to approximately215,000 business establishments on the StandardStatistical Establishment List (SSEL) where the SIC Codefor the establishment would be reclassified under adifferent sector in the NAICS. The SICs involved werethose activities that were reclassified into multipleNAICS codes and establishments that had to be classi-fied into a correct NAICS code. The results of the 1997Refile Survey were used to determine which economiccensus questionnaire the establishment would receive.

• Using the 1996 Company Organization Survey—anannual survey used to update the SSEL’s address files onmultiestablishment companies—to request classificationinformation from approximately 51,000 multiestablish-ment companies.

• Modifying the 1997 Economic Census questionnaires toenable the Census Bureau to accurately classify estab-lishments that were transferred to different industrieswithin the same sectors or to industries in new sectors,as a result of the adoption of the NAICS.

• Mailing classification questionnaires in November 1997to approximately 1.6 million small, single-establishmentcompanies that were classified in SICs affected by thechange over to the NAICS.

Once the businesses included in the economic censushad been reclassified, the Census Bureau’s data-processingsystems had to be able to identify, edit, tabulate, and cor-rect the individual data records, as well as the tabulationsand cross-tabulations made from those records. Thisrequired modifying the Census Bureau’s data-processing



procedures to produce the tables needed for comparisonsof 1997 data tabulations with those of previous censusesand for publishing data using NAICS classifications andhierarchy.

The 1997 Refile Survey

The 1997 Refile Survey covered establishments on theSSEL for which economic activities were reclassified in anew industrial sector under NAICS. The survey wasdesigned to classify establishments in the correct NAICScode and so ensure that the correct questionnaire wasmailed in the census. The Census Bureau mailed the sur-vey questionnaires in December 1996 so that the informa-tion could be collected and the affected establishmentsidentified prior to the census mailout late in 1997.

The Census Bureau’s Economic Planning and Coordina-tion Division (EPCD) identified the 80 SIC Codes that weremoving to new sectors because of the change to NAICS.The Census Bureau established payroll cutoffs for eachgroup of establishments (most did not have a minimumpayroll required, but 23 had minimums ranging from$15,000 to $225,000 to use in selecting the sample. Thefinal sample consisted of 215,829 establishments.

The Census Bureau designed a special classificationquestionnaire, Form NC-9922, ‘‘1997 Economic CensusClassification Report,’’ to use in the survey, in 17variations—Forms NC-9922(V01) through (V17)—for speci-fied groups of industries. The questionnaires were singlesheets of paper, with a cover letter on the front side and adata-collection section on the reverse, asking the respon-dent to identify the establishment’s principal economicactivity. This provided the information for the CensusBureau to use to assign the establishment to an industryclassification.5 Check boxes were used to minimizeresponse burden.

The questionnaires and other materials were printedat the Census Bureau’s Data Preparation Division(DPD—renamed the National Processing Center [NPC] inthe agency’s reorganization effective May 1, 1998) inJeffersonville, IN, whose staff also assembled the mailingpackages. Each mailout package included the appropriateversion of Form NC-9922, 1997 Economic Census Classifi-cation Report, instructions on completing it, and a returnenvelope.

The DPD conducted a single mail follow-up on March10, 1997, remailing survey packages (with the outgoingenvelopes overprinted ‘‘Second Request’’) to 47,819addresses still nonrespondent at that time. The surveyquestionnaires were returned to the Jeffersonville office,

where DPD staff processed the responses, sorting, screen-ing, and keying the data from the NC-9922 to a computer-ized data file. (This file was later used to update the SSELprior to the 1997 Economic Census classification question-naire mailout in November 1997.)

Processing returns from the 1997 Refile Survey contin-ued into the summer of 1997 to enable the data obtainedto be used for preparing the 1997 census mail lists.Establishments that did not respond to the survey by theend of July were added to the 1997 census classificationquestionnaire mailing list as establishments with doubtfulclassifications.

The 1996 Company Organization Survey

Further reclassification information was collected aspart of the 1996 Company Organization Survey (COS).The Census Bureau uses the annual COS to update theSSEL’s address files on multiestablishment companies.The survey involves sending Form NC-9901, ‘‘Report ofOrganization,’’ questionnaires to all multiestablishmentcompanies with 50 or more employees, and to a sample ofsmaller multiestablishment companies. As part of the1996 COS, the Census Bureau requested informationneeded to classify the individual establishments of multi-establishment companies in SIC Codes significantlyaffected by the adoption of NAICS. This effort particularlyinvolved those establishments whose reclassificationcrossed ‘‘division/sector’’ boundaries (e.g., from SICDivision D, ‘‘Manufacturing,’’ to NAICS Sector 51,‘‘Information’’).

The Census Bureau reviewed the 1996 COS sample(about 53,000 addresses in all) of smaller companies,retaining those with SICs impacted by the change over toNAICS and dropping the remainder (about 20,000 firms)from the sample. It then added approximately 18,000more small multiestablishment companies from affectedSICs, for a total sample mailing of 51,000. The COS mail-ing package contained not only the Form NC-9901, butalso any 1 or more of 11 specialized inserts requestingclassification information for establishments in industriesmost affected by NAICS.

Each of the inserts listed a set of industry descriptions,together with the bridge classification code for eachindustry described. Form NC-9901-IG, ‘‘Personal Services,’’for example, listed six-digit industry bridge codes for 10‘‘personal care and related services’’ and for 12 ‘‘otherservices to individuals.’’6 The form listed the bridgeclassification codes covering the respondents’ generalarea of economic activity, and asked them to enter theappropriate code in the ‘‘Major activity’’ section of thequestionnaire.

5Form 9922(V04), for example, asked whether 50 percent ormore of sales were to businesses that would resell merchandiseor to household consumers, what primary format the establish-ment used to sell in 1996, how the establishment attracted newcustomers (e.g., location and store attractiveness, advertising tothe general public, etc.), and what was the principal line of mer-chandise in 1996.

6The various services listed were drawn from the 1987 SIC,with the corresponding NAICS Code shown for each. In severalcases, a single six-digit NAICS Code covered several industries;e.g., under personal care and related services; massage salons;tanning salons; saunas, steam baths, and Turkish baths; and tat-too parlors were all classified as NAICS industry 729998.



The Census Bureau mailed the COS packages inDecember 1996, requesting response in 30 days. Threemail follow-ups were carried out between the beginning ofFebruary and the end of May 1997. Responses werereceived from approximately 44,000 companies.

‘‘Birth’’ Classification

The Census Bureau used administrative records fromvarious Federal agencies (primarily the Internal RevenueService [IRS], the Social Security Administration [SSA], andthe BLS) to update the SSEL.7 The Census Bureau incorpo-rated data identifying ‘‘births’’ during the first three quar-ters of 1997 into its census mail list preparation, and usedthe individual records to classify as many new companiesas possible. Those companies for which no informationwas available, or that could not be classified using admin-istrative records, were referred for the November 1997classification mailout and were sent the Form NC-9923‘‘General Classification’’ questionnaire. When sufficientinformation existed to classify the company by economicsector or industry group, but not at the industry level, thecompanies also were added to the November mailout, butwere sent either a Form NC-9926 (manufactures and min-ing companies) or the appropriate ‘‘services’’ sectors (i.e.,virtually every other economic sector except manufactur-ing and mineral industries) classification questionnaire.

Data on fourth-quarter births were compiled by theCensus Bureau from monthly transfers of IRS and SSAadministrative records. The Census Bureau prepared anadditional mailing list of new businesses and conducted asecond classification mailing in February 1998 for thosefourth quarter births it was unable to classify from theadministrative records.

Changes to the 1997 Census Questionnaires

The technical committee that had periodically reviewedthe SIC following its adoption in the 1940s, rarely movedmany industries from one SIC division to another. Theadoption of NAICS involved doubling the number of eco-nomic divisions, called ‘‘sectors’’ under NAICS. This exten-sive reorganization involved the creation of scores of newindustries and required not only establishing entirely newproduct and industry codes, but moving significant num-bers of industries between sectors and into the new sec-tors. The designs of over 450 1997 Economic Censusquestionnaires were modified to accurately categorizeestablishments.

The changes made to the questionnaires specifically toimplement NAICS generally fell into two categories—

• Additional products, product codes, and activities werelisted where necessary to provide for the movement ofan activity from an SIC division or industry to a newNAICS sector or industry.

• Special inquiries were added to the questionnaires toaid in identifying specific types of economic activitiesnot previously identified under SIC. Any additional prod-ucts or activities were added to one of the followingitems on the questionnaires:

• Kind of business or selling characteristics.

• Method of selling.

• Merchandise lines (retail trade).

• Commodity lines (wholesale trade).

• Sources of receipts or revenues.

• Type of construction (construction industries).

• Consumption of selected materials (manufactures andmineral industries).

• Products and services of this establishing during 1997.

The ‘‘special inquiries’’ questions frequently were incor-porated into other items on the questionnaires. That is,the response to items on the questionnaires enabled ana-lysts to confirm the classification of the respondent estab-lishment. (For more information on changes to the ques-tionnaires, and for examples of changes made specificallyto accommodate NAICS, see Chapter 4, ‘‘QuestionnaireDevelopment.’’)

1997 Economic Census ClassificationQuestionnaire Mailout

Once the 1997 Economic Census mail list had beencompiled from the SSEL in the fall of 1997 (see Chapter 7,‘‘Preparation for Census Mailout,’’ for details), the CensusBureau sampled the list for small companies or establish-ments from the SICs most affected by the introduction ofNAICS. In November 1997, the Census Bureau conducted amailout to those establishments considered most likely torequire reclassification once NAICS was adopted, sendingclassification questionnaires to approximately 1.6 millionaddresses. (By comparison, the 1992 Economic Censusmailing had included 886,120 classification forms of allkinds.) Addressees were asked to complete and return thequestionnaires within 30 days. This mailing was timed toenable the DPD, which carried out the mailing and towhich all the census mail questionnaires were to bereturned, time to process at least part of the responsesbefore the main census mailout. (For further details of thecensus classification mailing, see Chapter 8, ‘‘DataCollection and Processing.’’)

Processing System Modifications

The problems created for the Census Bureau’s 1997Economic Census data-processing systems by the adop-tion of NAICS involved the systems’ capacity to accountfor the new classifications in both the microdata (i.e., pro-cessing individual census records from the questionnaires)

7These records were supplied by the agencies monthly andquarterly on a flow basis. The IRS provided the Census Bureauwith its complete Business Master File (BMF) each May.



and the macrodata (the census data file) areas. Changesfor microdata processing involved the complex edit ofrecords and the interactive editing systems that enablededit clerks and/or subject-matter specialists to reviewand edit individual records. The complex edit had to bemodified by:

• Adding classification routines to bridge NAICS and theSIC Code systems.

• Modifying code validation procedures to account for thenew code ranges and lengths.

• Rewriting the classification programs to enable the sys-tem to properly classify establishments that were newto the specific economic sector involved (i.e., industriespreviously part of a different division under SIC).

The Census Bureau used interactive systems to keydata from questionnaires to the economic census datafiles and also to review and edit the individual datarecords created when the data were keyed from thequestionnaires. The adoption of NAICS required that theinteractive systems originally designed for use with theSIC Code be modified to account for the new bridge codesallowing linkages between the NAICS and SIC Code.Similar modifications had to be made to the economiccensus tabulation programs.

COMPARABILITY OF THE DATA ANDMAINTENANCE OF TIME SERIES

Scope of Data Comparability

Data users will be able to derive NAICS data for roughlytwo-thirds of the four-digit SIC Code industries, eitherbecause the industries have not been changed (other thanthe numerical code), or because any new industries aresubdivisions of an old industry. For example, the SICindustry 4953, ‘‘Refuse Systems,’’ was subdivided underNAICS into eight new industries. These eight differentiatedamong collection and disposal, separating hazardouswaste, materials recovery, landfills, and incinerators.However, many other industries were significantly alteredby the adoption of NAICS—544 SIC industries either werenot continued or could not be identified under the newsystem. In turn, 710 industries were created, or had theirclassifications significantly revised by NAICS. To assistdata users, OMB published tables showing the full NAICShierarchy, as well as tables showing NAICS industriesmatched to the equivalent 1987 SIC Code, and vice versa.

Maintenance of Time Series

NAICS and economic census time series. Economicdata are generally most useful in time series, providingcomparable data over a significant period of time. Theadoption of a new classification system breaks time series,since the data collected using the new definitions and

codes will not be identical to that produced by the old sys-tem. In the best of all statistical worlds, the need to enabledata users to familiarize themselves with the new systemwould require the application of the new classifications toexisting data before (or at least contemporaneously with)the collection and publication of new data using the newclassifications. In the case of the census, it would be desir-able to reclassify the data from the 1992 Economic Censusprior to the publication of 1997 statistics. However,for accurate classification, NAICS requires more and/ordifferent information about the operations of individualestablishments than did the SIC, so that while some of the1992 data will have direct counterparts in the 1997 statis-tics, much of the statistical data produced will have nosuch immediate connection. The Census Bureau recog-nized that users need to maintain critical time series datafor comparability, and the agency addressed this problemin two ways— through its plans for data publication and ina research program to determine the feasibility of reclassi-fying 1992 census data using NAICS.

Publication plans. The sheer volume of data that wouldbe generated by the census, as well as time and costrestraints, meant that any reasonable response to theproblem of complete tables ‘‘bridging’’ between the oldand new classification systems necessarily would beextremely difficult and costly. Ideally, a complete parallelpublication program, showing all the data classified underboth systems, would have been done. Since that wasimpossible if the data were to be published on a timelybasis and within any possible budget, the Census Bureauhad to develop a more limited plan.

The 1997 Economic Census data were published inthree phases—an Advance Report, detailed data, andbridge reports. Early in 1999, the agency released theAdvance Report, covering all sectors—at the nationallevel—included in the 1997 Economic Census. This reportshowed data down to the three-digit industry group level,using NAICS. Preliminary ‘‘comparative statistics’’—that is,tables presenting 1997 and 1992 data on the 1987 SICbasis also were shown at the two-digit levels for theUnited States. (A separate Comparative Statistics reportwas released later in the publishing cycle, with two-,three-, and four-digit SIC summaries that superceded thedata in the Advance Report.)

Detailed data from the Industry Series and GeographicArea Series reports were released throughout 1999, cover-ing industry and summary statistics for various NAICS andgeographic area levels. In early 2000, a final ComparativeStatistics Report was issued, covering 1997 and 1992basic data items at the SIC trade area and selected two-,three-, and four-digit levels for all establishments for theUnited States and states.

At approximately the same time, the Census Bureaureleased ‘‘bridge tables’’ for the United States that included1997 statistics showing the distribution of NAICS-based



industries at the more detailed level among SIC-basedindustries at the four-digit level, as well as 1997 statisticsshowing the distribution of SIC-based industries at thefour-digit level among NAICS-based industries at the moredetailed level.

All other data from the 1997 census were publishedusing only NAICS. The first industry-specific numberspublished in 1999 did not enable comparisons to be madeto earlier censuses for any industries that changed in defi-nition. Only limited historical data appeared for industriesthat were unmodified by the change to NAICS, or thatcould easily and cleanly (i.e., without fragmentation of theSIC Coded industry) be connected to NAICS. As in previouscensuses following changes in the SIC, no comparabilitytables were available below the state level—i.e., substatelevel statistics were published using NAICS only.

The reclassification of 1992 Economic Censusdata. The disruption of data time series occasioned bythe adoption of NAICS caused considerable concernamong many economic census data users. The Census

Bureau’s Professional Associations’ Advisory Committeeurged that the agency try to alleviate this situation to thegreatest extent possible. While it proved impossible tocompletely reclassify the 1992 Economic Census data, theCensus Bureau’s Center for Economic Studies (CES) under-took a study of the possibility of reclassifying at least partof the 1992 statistical data to conform to NAICS.

CES staff found that, by making extensive use ofmodeling, some degree of reclassification could be accom-plished, although the resultant data would have to beused with considerable caution. Accordingly, the CES staffcarried out the reclassification of the 1992 manufacturing,retail, and wholesale data, completing the process bymid-1999. The services industries, which had undergoneperhaps the most extensive changes in the adoption ofthe new system (e.g., the creation of eight new majorservices economic sectors under NAICS), was consideredthe most difficult sector to attempt to reclassify, and theCES was still working on that reclassification at the time ofthis writing.



Chapter 4. QUESTIONNAIRE DESIGN AND DEVELOPMENT

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Design Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Consultation on the Questionnaires . . . . . . . . . . . . . . . . . . . . . . 27General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Retail Trade; Wholesale Trade; Service Industries; Finance, Insurance, andReal Estate; Transportation, and Utilities . . . . . . . . . . . . . . . . . . 27Manufacturing, Construction, and Mining . . . . . . . . . . . . . . . . . . 27

Design and Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27Retail Trade; Wholesale Trade; Service Industries; Finance, Insurance, andReal Estate; Transportation, and Utilities . . . . . . . . . . . . . . . . . . 27Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28Manufacturing and Mining . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Changes to the Questionnaires . . . . . . . . . . . . . . . . . . . . . . . . . 29Retail Trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30Wholesale Trade. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31Mining. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Service Industries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32Finance, Insurance, and Real Estate . . . . . . . . . . . . . . . . . . . . . 33Transportation and Utilities Industries . . . . . . . . . . . . . . . . . . . . 33

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

CONTENTS


Questionnaire Design and Development

INTRODUCTION

Questionnaire design and development began inJanuary 1994. The majority of the work on the designand content of the census questionnaires was done bythe staffs of the Census Bureau’s economic subjectdivisions—Manufacturing and Construction Division(MCD), Company Statistics Division (CSD), and ServiceSector Statistics Division (SSSD). These divisions wereassisted by Economic Planning and Coordination Division(EPCD) and Administrative and Customer Services Division(ACSD). Subject specialists, together with forms designspecialists and EPCD staff, made many decisions: size andformat of the questionnaires, questionnaire identification(i.e., form numbers and titles), wording and sequencing ofstandard questions, file copy requirements, design tech-niques, color schemes for the paper and ink, shading, typeof processing equipment to be used, and microfilmingrequirements.

The Census Bureau developed separate questionnairesfor each of the economic sectors covered by the census,and, within each sector, developed questionnaire varia-tions for the different industries. Table 4-1 shows thenumbers of questionnaires used for the major economicareas covered by the 1997 Economic Census.

Table 4-1.1997 Economic Census Questionnaires

Total 479*Manufacturing 253Mining 12Construction 11Retail Trade and Accommodations andFoodservices

44

Wholesale Trade 48Service Industries 54Finance, Insurance, and Real Estate 24Utilities (utilities, transportation, andcommunications)

8

Vehicle Inventory and Use Survey 4Business Owners and Self-Employed Persons 2Business Expenditures Survey 9Economic Census of Outlying Areas 6**General Schedules 4

*The Economic Directorate also was responsible for coordinatingthe design and printing of the 48 kinds of envelopes used in thecensus, and the report forms for the 1997 Census of Agriculture (43questionnaires and 37 envelopes).

**Includes Spanish and English questionnaires for Puerto Rico andthe Virgin Islands of the United States.

While the Census Bureau designed the questionnaires,the Office of Management and Budget (OMB) exercisedfinal authority for approving or denying approval for alleconomic census forms. The OMB was required, by law,to review the questionnaires and to determine that theinformation requested—

• Was necessary.

• Was not available from other government or privatesector sources.

• Was available from the records customarily maintainedby businesses or organizations being enumerated, orthat reasonable estimates of the requested informationcould be made by respondents.

The OMB also verified conformance to the reportingburden control program for Federal agencies, the objectiveof which is to keep the response burden of the businesscommunity to a certain minimum number of hours.

Prior to its submission of the draft questionnaires forOMB approval, the Census Bureau was required to publisha notice in the Federal Register of its intention to carryout the economic census. This was to inform interestedparties of their opportunity to comment on the question-naire content. For the 1997 Economic Census, the CensusBureau made draft versions of the questionnaires availableat its economic census Web site, www.census.gov/econ97.

DESIGN CONSIDERATIONS

The chief goal of questionnaire design was to obtaininformation, while imposing the least possible respondentburden. This task was complicated by the requirementthat the 1997 census collect, tabulate, and publish datausing the new North American Industry ClassificationSystem (NAICS). (The NAICS is described in detail in Chap-ter 3, ‘‘Introduction of the North American Industry Classi-fication System (NAICS).’’) Replacing the Standard Indus-trial Classification (SIC) Code in use in the U.S. since the1930s, the NAICS doubled the classification of sectors atthe two-digit level from 10 to 20, and added over 300 newindustry classifications to capture business activity in theU.S. economy.

The adoption of the NAICS, only a year before themailout of the census questionnaires, imposed a very tightschedule for designing, approving, and printing thequestionnaires. Given the additional requirements forclassification information, the Census Bureau decided thatminimal changes should be made to the overall content



and format of the questionnaires. To avoid complicationsand possible delays in the design and approval of the vari-ous economic census forms, the Census Bureau simplifiedand standardized the designs to the greatest extent pos-sible.

CONSULTATION ON THE QUESTIONNAIRES

General Information

The Census Bureau asked for comments and sugges-tions on questionnaire content and design from a widevariety of professional, industry, and business associa-tions.1 Since this process began before the final proposalsfor the adoption of the NAICS were published in theFederal Register (November 1996), the Census Bureau’sconsultations with the various associations were based onSIC industry divisions. In the discussion that follows,therefore, the consultations are described in terms of SICcategories.

Mailing packages sent to the associations eachconsisted of a cover letter explaining the request forcomments, a draft copy of the questionnaire for therespective industry, a response sheet that provided bothspace for comments as well as a check box for respon-dents who had no comments, and a return envelope.

Census Bureau staff also consulted various Federalagencies (e.g., the Federal Reserve Board, the Bureau ofLabor Statistics, the Office of Trade and Economic Analy-sis, and the Bureau of Economic Analysis) that use eco-nomic census data as benchmarks for their own statisticalwork, or as primary data for their own programs. Inaddition, the Census Bureau sought comments on ques-tionnaire content from its advisory committees.

Retail Trade; Wholesale Trade; Service Industries;Finance, Insurance, and Real Estate;Transportation and Utilities

In January and February of 1995, the Census Bureaucanvassed national professional and trade associationsin services; retail and wholesale trades; transportation;and finance, insurance, and real estate sectors. Packageswere mailed to 558 business associations, Federal agen-cies, and other organizations with response requested byMarch 31, 1995. The Census Bureau followed up selectednonrespondents by mail and telephone, receiving 160replies by the end of August. Census Bureau staff carefullyreviewed the responses, and also met with selected publicand private organizations for detailed discussions of theirrecommendations.

Manufacturing, Construction, andMining

The Census Bureau mailed requests for comments on aflow basis beginning in the spring of 1995, as drafts ofthe proposed questionnaires became available. Requests

went to over 700 manufacturing, construction, and miningbusiness and trade associations, and to various Federalagencies that use data from these sectors. The packagesfor the manufacturing, construction, and mining differedfrom those used in the round of consultations for theother economic sectors in that they included descriptionsof both proposed changes to the questionnaires and publi-cation formats for the 1997 census. Many of the organiza-tions on the mailing list received multiple packages—over2,000 were mailed in all—because they conducted busi-ness across several industrial classifications.

By the end of August 1996, the Census Bureau hadreceived comments on over 960 of the packages sent forreview. The MCD staff reviewed the responses and metwith selected respondents to discuss their recommenda-tions before finalizing the questionnaire content.

DESIGN AND CONTENT

General Information

The enormous variety of activities within the economyand within the major economic sectors precluded conven-tional standardization in questionnaire design andcontent.2 However, the following basic data items wererequested of all respondents to the 1997 EconomicCensus:

• Employer identification number (EIN).

• Physical location of the economic activity.


• Payroll.

• Value of sales, receipts, work done, or equivalent.

The general instructions and certification (name,address, telephone number, etc., of the person completingthe questionnaire) were similar for all questionnaires.Within each major economic sector, the Census Bureaucustomized the questionnaires to the greatest extentpossible for ease of response, but the data requestedvaried enormously from sector to sector. For example,product data for manufacturing establishments wasnecessarily different from product data for financial andinsurance businesses. The sector-specific data contentof the questionnaires for each major economic sectorcovered in the 1997 Economic Census is outlined below.

Retail Trade; Wholesale Trade; Services Industries;Finance, Insurance, and Real Estate;Transportation and Utilities

The questionnaires for the retail and wholesale trades;the service industries; and for finance, insurance, and realestate, asked for legal form of organization (i.e., individual

1The names of these organizations are contained in AppendixD, ‘‘Consultation on the Census Questionnaires.’’

2For example, 48 separate questionnaires ultimately were usedto enumerate wholesale trades.



proprietorship, partnership, cooperative association,government, corporation, other), and operational status(in operation, temporarily or seasonally inactive, ceasedoperation). The wholesale trade questionnaires includedadditional items on type of operation, operating expenses,and inventories, while the service industries’ question-naires asked for Federal tax status and operatingexpenses for tax-exempt organizations. The transporta-tion and utilities industries’ questionnaires included anitem on operational status, and requested data on sourcesof revenue. Many of the questionnaires included specialinquiries—e.g., sales by merchandise, commodity, orreceipts line; sales by class of customer; and employmentby occupation.

Construction

Eleven questionnaires were used for enumerating con-struction firms. These forms included relatively detailedinquiries on costs, as follows:

• Employers’ cost of fringe benefits (voluntary and legallyrequired).

• Costs of (1) work subcontracted out; (2) materials,components, and supplies; (3) selected power and fuels;(4) selected purchased services; (5) rental machinery,equipment, and buildings.

• Value of business done.

• Estimated percentage of dollar value of constructionwork done for various kinds of business.

• Estimated percentage of dollar value of constructionwork done for various types of construction and howmuch was new construction, additions, alterations, orreconstruction; or maintenance and repair work.

• Percentage breakdowns of government-owned andprivately-owned construction, of construction worksubcontracted from other contractors, and for thelocation of construction work by state.

• Capital expenditures, depreciable assets, andinventories.

Manufacturing and Mining

The Census Bureau used 35 short- and 218 long-formquestionnaires for the manufacturing sector. The shortquestionnaires were two pages each, with a standardizedfirst page that requested—

• Employer identification number.


• Employment in 1997.

• Annual payroll.

• Value of products exported.

• Total capital expenditures and gross value of depre-ciable assets (excluding land).

• Total rental payments (buildings, machinery, andequipment, including land).

• Cost of materials for 1997.

• Value of inventories.

• Legal form of organization.

• Operational status.

The second page of the short forms (identified by formnumbers based on the industry-specific NAICS Code)asked for value of products shipped and other receipts (bytype of product), as well as total value of shipments andother receipts.

The first four pages of all 218 long questionnaires werestandardized (three pages of questions and one of instruc-tions) and were virtually identical to the general question-naire used in the Annual Survey of Manufactures (ASM).The remainder of each questionnaire was designed forenumerating the specific industries. In fact, the 1997 ASMquestionnaire, Form MA-1000, was used as the first fourpages of the manufacturing questionnaire for companieson the ASM sample list. A variant of the ASM questionnaireformed the first four pages of the questionnaires used forthe rest of the census of manufacturing long-questionnairemailing list.

The long questionnaires included the followingstandardized items:

• Employer’s cost for fringe benefits.

• First-quarter payroll before deductions.

• Plant hours worked by production workers.

• Total shipments and other receipts, value of productsexported, and shipments to other domestic plants.

• Cost of materials and contract work.

• Establishment inventories at end of year. Depreciableassets, capital expenditures, and retirements (of debt).

• Rental payments.

• Selected purchased services (MA-1000 only).

• Quantity of electricity used (kilowatt hours) (MA-1000only).

• Total depreciation charges (MA-1000 only).

• Method of valuation for inventories not subject to last-in/first-out (LIFO) costing (MA 1000 only).

• Ownership, control, and location.



The individual industry long questionnaires askedfor information on consumption of selected materials,products and services provided by the respondent estab-lishment, total shipments and receipts, and also includedspecial inquiries for particular industries (e.g., selectedproducts that were produced and used to manufactureother products by the responding establishment).

For mining, the census gathered data using 2 short- and10 long-form questionnaires. All of these included thestandardized questions, as well as many of the itemsappearing on the MA-1000 questionnaire. The short-formquestionnaires (Forms 1371 [Oil and Gas Field Operations]and 1471 [Stone, Sand, and Gravel]) requested data on—

• Capital expenditures and gross value of assets at end ofyear.

• Total rental payments and cost of supplies.


• Operational status.

• Products and services provided in 1997.

• Total shipments and other receipts.

The long-form questionnaires asked for all the data onthe short versions, and the following additional informa-tion:

• Establishment inventories.

• Debt retirement.

• Total depreciation.

• Current account expenditures for mineral properties.

• Detailed cost of supplies and contract work. Ownership,control, and location of operations.

• Detailed cost of supplies used, minerals for preparation,and purchased machinery installed in 1997.

• Detailed fuels used.

• Detailed rental payments.

• Type of operations.

• Selected purchased services.

CHANGES TO THE QUESTIONNAIRES

The 1997 Economic Census questionnaires were basedon those used in the 1992 enumeration, and were updatedas necessary. The specialized nature of the questionnairesfor each economic sector meant that most of the contentmodifications for the 1997 census applied only to a singlequestionnaire, or to a small group of questions covering aparticular economic sector or group of sectors. By con-trast, the general design changes focused primarily onidentification of the questionnaires and their handlingprior to data processing. For example, the most prominent

change in the overall design of the questionnaires was thetitle ‘‘1997 Economic Census’’ in place of the specific cen-sus designations previously used (e.g., ‘‘1992 Census ofManufactures,’’ ‘‘1992 Census of Retail Trade,’’ etc). Thischange applied to questionnaires for all trade areascovered by the census except mineral industries and (partof) transportation, which retained their individual censusdesignations (e.g., ‘‘1997 Census of Mineral Industries,’’‘‘1997 Census of Transportation, Vehicle Inventory andUse Survey’’). The trade area identifiers appeared in theindividual form numbers; that is, an alphabetic prefixidentified the general economic area that would processthe forms:

Prefix Economic area

AU AuxiliariesCC ConstructionFI Finance and insurance, and real estate

and rental and leasing industriesMC ManufacturingMI MiningRT Retail trade and accommodations and

foodservicesSV Service industriesTC Vehicle inventory and use surveyUT Transportation and utilitiesWH Distributive (wholesale) trades

The Census Bureau also added a barcode to the rightmargin of each page of the questionnaire. This served asan alignment aid for microfilming equipment. The censusdue date was changed from February 15, in 1993 toFebruary 12, in 1998.

The adoption of the NAICS was a substantial factorleading to changes in the content of individual question-naires. These changes were primarily of two kinds—

• Additional products shipped and materials consumed orkinds of business activities were listed on the question-naires where necessary to provide for the movement ofan activity from an SIC sector or industry to a newNAICS sector or industry. It also provided for compara-bility between data from previous censuses—publishedusing the SIC—and 1997 data collected and publishedunder the NAICS.

• Special inquiries were added to the questionnaires toaid in classifying specific types of economic activitiesnot previously covered.

Depending on the specific sector, additional or ‘‘new’’products shipped and product or activities codes wereadded to one of the following items on the questionnaires;(1) kind of business or selling characteristics, (2) methodof selling, (3) merchandise lines (retail trades), (4) com-modity lines (wholesale trade), (5) sources of receipts or



revenue, (6) type of construction (construction), or (7) con-sumption of selected materials (manufacturing and min-ing). The ‘‘special inquiries’’ questions frequently appearedin separate items at the end of the questionnaire, but alsocould be incorporated in other items. In the latter case,responses could be used by analysts to confirm the classi-fication of the establishment without asking an explicitquestion.

The changes made to the questionnaires are summa-rized below, by industry sector or related sectors, togetherwith some illustrative examples of specific additions,deletions, or changes to the questionnaires.

Retail Trade

The retail trade sector expanded under the NAICS toinclude establishments primarily engaged in providingaccommodations (e.g., motels) and food services (e.g.,restaurants). The changes made to the 48 retail tradequestionnaires generally fell into three broad categories—

• Deletion of special inquiries not critical for kind-of-business classification.

• Changes to the merchandise lines (item 10) of the ques-tionnaires, made at the request of the industry to collectmore useful data.

• Changes made to facilitate kind-of-business classifica-tion under the NAICS.

A considerable variety of special inquiries were addedto the retail trade questionnaires, primarily to providedetailed data on specific characteristics of limited sectorsof the retail sector.

Retail trade is a relatively volatile sector of theeconomy, in terms of the importance of specific kinds ofmerchandise at any given time. There were considerablechanges made to Item 10, ‘‘Merchandise Lines,’’ on manyquestionnaires. For example, on Form 5901, ‘‘Health andPersonal Care Stores,’’ lines of merchandise considered nolonger relevant to the establishments covered—e.g., majorhousehold appliances, furniture and sleeping equipment,floor coverings, automotive fuels, and paint and relatedpreservatives and supplies—were deleted, and new linesrequesting data on hearing aids and seasonal decorationswere added.

The adoption of the NAICS also had an impact on theretail questionnaires, including the addition of two newquestionnaires—Forms RT-7001, ‘‘Traveler Accommoda-tions,’’ and RT7002, ‘‘Recreational and Other Accommoda-tions.’’ These covered industries reclassified from servicesto retail trade. Other changes made to facilitate kind-of-business classification under the NAICS required modifica-tion of existing forms, such as combining items on paint,glass, wallpaper, and hardware dealers from two 1992census questionnaires into Form 5201, ‘‘Building Materials

and Supplies Dealers.’’ Also, a single 1992 questionnaireon contract feeding and vending operators was split intoForms RT-5802, ‘‘Contract Feeders,’’ and RT5803, ‘‘VendingOperators.’’

Wholesale Trade

The bulk of the changes to individual questionnaires inthe wholesale trade sector involved either Item 11, ‘‘Kindof Business,’’ or Item 13 (13a on some forms), ‘‘Commod-ity Lines.’’ However, there also were some general contentchanges to nearly all questionnaires for wholesale estab-lishments, as follows:

• An additional option was added to Item 6, ‘‘Employ-ment,’’ section b, ‘‘List the above employees by theemployee’s primary function’’—‘‘Sales support (includingoffice and clerical, warehousing, customer service,maintenance employees, and drivers).’’

• The Census Bureau added Item 11b, ‘‘Selling Character-istics,’’ asking (1) ‘‘In what format did this establishmentprimarily sell in 1997?’’ and (2) ‘‘How did this establish-ment primarily attract new customers in 1997?’’ to alllong-form questionnaires (except Form WH-5170,‘‘Petroleum and Petroleum Products’’).

• In Item 10, ‘‘Sales by Class of Customer,’’ choices ‘‘b’’through ‘‘f’’ were reworded to accommodate changesrequired by the adoption of the NAICS (e.g., Item 10b,originally read ‘‘Restaurants, food services, and contractfeeding,’’ and was changed to ‘‘Restaurants, hotels, foodservices, and contract feeding,’’ while Item 10d, origi-nally ‘‘Other wholesale establishments’’ was reworded‘‘Other wholesale establishments for resale.’’

Most other changes made to individual questionnairesstemmed from the NAICS classification requirements, orresulted from the aforementioned consultations. Forexample, under the NAICS, all fluid-power equipment wasclassified under industrial machinery and equipment.Consequently, the Census Bureau added a line to Item 11,‘‘Kind of Business,’’ on ‘‘Hydraulic and pneumatic valves,cylinders, rotary actuators, connectors, and assemblies,filters, accessories, and parts,’’ to Form WH-5084, ‘‘Indus-trial Machinery and Equipment.’’ This, in turn, led tochanges in Item 13, ‘‘Commodity Lines,’’ of the same form,adding lines for hydraulic and pneumatic valves, cylindersand rotary actuators, and connectors.

Consultation with the industry led to considerablerewording of the detail lines in Item 13 of Form WH-5060,‘‘Electrical Goods,’’ where 10 new lines—on power anddistribution transformers; switchgear and switchboardapparatus; motors and generators; conduits, raceways,and fittings; electric bulbs; interior wiring; exterior wiring;lighting fixtures; industrial controls; and other electricalapparatus and equipment—were added to the item.



Six of the 48 wholesale questionnaires were intendedfor small companies.3 Aside from the prefix (i.e., ‘‘WH’’)change and the request for county in the item on location,the principal changes made to the short-form question-naires was the adoption of the items on ‘‘Sales By Class ofCustomer’’ (Item 9 on most questionnaires) and ‘‘SellingCharacteristics’’ (Part b of Item 11, ‘‘Kind of Business’’) onthe long forms.

Manufacturing

The adoption of the NAICS that had the greatest impacton content changes for the manufacturing sector. Thesechanges required additions and elaborations of productlines on most of the manufacturing questionnaires. Forexample, Form 2023, ‘‘Dried Food, Pickles, and Macaroni,’’added ‘‘Vegetable flour’’ (product code 20343 38) to Item18. Form MC-2207, ‘‘Spinning and Throwing of Yarn andThread,’’ expanded the 1992 equivalent product line for‘‘rewound and novelty yarns’’ (1992 product code 2282221 7) into two new product lines: ‘‘Rewound yarns, otherthan wool, ...’’ (product code 22822 22 5) and ‘‘Noveltyyarns, other than wool, ...’’ (product code 22822 23 3).

For the 1997 census, the Census Bureau made severalchanges to the standardized portion of the manufacturingquestionnaire by eliminating the following items thatappeared on the 1992 questionnaire:

• Separate data for voluntary and legally required fringebenefits. This item was dropped in order to reduceresponse burden.

• Separate data for new and used assets and expendituresand detail for new machinery expenditures. These datawere collected in the Census Bureau’s Annual CapitalExpenditures Survey (ACES).

• Request for data on methods of inventory valuation notsubject to last-in, first-out (LIFO) costing (eliminatedfrom the MA-1000 only).

The industry-specific items of the manufacturing ques-tionnaires (pp. 5-11 for long-form questionnaires, p. 2for the short-form questionnaires) covered the particularcharacteristics of the individual establishments’ opera-tions. Changes in these items were made based on (1)expected growth or decline in importance of a material,product, or special inquiry item, (2) revisions requiredto promote comparability between the Census Bureau’sclassification system and the international Harmonized

Commodity Description and Coding System (HS)4, and (3)revisions needed to implement the change over from SICto NAICS. For example, Form MC 2004, ‘‘Dairy Products,’’Item 18, ‘‘Products and Services of This Establishment in1997,’’ was modified to add new product lines on ‘‘LowfatNatural Cheese, Except Cottage Cheese’’ (shipped in pack-ages of less than three pounds, and three pounds ormore). The Census Bureau added the new product linesbecause it expected the total value of these productsshipped in 1997 to exceed $50 million. A product linecould be deleted if the value to be shipped was expectedto fall below $25 million.

Construction

For the 1992 census, the Census Bureau used 22questionnaires to cover construction—11 long and 11short. Both the long and short questionnaires were 4pages, with virtually identical layout and design for pages1, 2, and 4. Page 3 on each questionnaire was customizedand contained items requesting information on kind ofbusiness and type of construction.

For the 1997 Economic Census, MCD reevaluated the1992 questionnaire designs and decided that there werefew, if any, advantages to using short questionnaires. Theshort questionnaires requested information from a sampleof small construction firms and, for 1997, the CensusBureau planned to reduce such sampling.

There were limited changes to the overall design andformat of the long questionnaires, as follows:

• On page 2, an additional instruction (to include receiptsor billings derived from contract construction, design/build contracts, and construction management) wasadded to two of the questionnaires (Forms 1508 and1608).

• Page 3 of each questionnaire, which contained Items 13,‘‘Kind of Business,’’ and 14, ‘‘Type of Construction,’’ wascustomized for each questionnaire. The Census Bureaudeleted requests for detail on new and used depreciableassets and detail on buildings, structures, and machin-ery from Item 18, ‘‘Assets, Capital Expenditures, andDepreciation in 1997.’’

• On page 4, Form 1508 included new items asking fordata on housing starts and nonresidential design/buildwork and Form 1608 asked for information on nonresi-dential design/build work.

3Payroll cutoffs for small wholesale establishments varied bykind of business, but small establishments typically includedfirms with one to three employees. These companies representedabout 10 percent of total annual wholesale sales.

4The U.S. adopted the Harmonized Commodity Descriptionand Coding system (HS) in 1989 for classifying commodities ininternational trade. Under HS, both exported and imported com-modities carried the same codes to the six-digit level. Thisallowed closer comparisons than under previous systems (theprocess was assisted by adjustments made to the SIC Code). TheHS was intended to serve as a universally accepted productnomenclature and could be used for classifying goods for cus-toms programs, as well as for collecting data on exports andimports.



Other relatively minor wording changes were made,primarily to clarify the data requested. Since write-inresponses to the kind-of-business and type-of-constructionquestions would require time-consuming manual codingby analysts during processing, the MCD tried to reducethe need for write-in response by anticipating activitiesand adding ready-made codes.

As with the questionnaires for other sectors, construc-tion industry forms underwent certain changes to accom-modate adoption of the NAICS. Reclassified activities andtheir attendant codes were added to the appropriate ques-tionnaires. For example, stucco contractors were part ofSIC Industry 1771, ‘‘Concrete Work.’’ Under the NAICS,stucco construction was transferred to NAICS Industry23542, ‘‘Drywall, Plastering, Acoustical and InsulationConstruction.’’

Mining

The 1997 Economic Census used 12 questionnaires toenumerate mining. Content changes to the questionnairesparalleled those made for manufacturing, deleting—

• Breakdown between legally required and voluntaryfringe benefits paid to employees.

• Detailed breakdowns of (1) beginning and endingassets, (2) retirements, and (3) depreciation/depletionchanges between buildings, machinery, mineralexploration/ development, and land and rights. Detailedbreakdown of capital expenditures between buildingsand machinery.

• Detailed breakdown of current account expenditures formineral properties between mineral exploration/development, and mineral land and rights.

Industry-specific changes also were made to the ques-tionnaires, primarily to Item 17, ‘‘Supplies Used, Ores forTreatment, and Purchased Machinery Installed During1997,’’ and 18B, ‘‘Products and Services of This Establish-ment in 1997.’’ For example, on Form MI-1001, lines inItem 17C for ‘‘Iron ore concentrate’’ and ‘‘Bauxite’’ werecombined with ‘‘Other ores and concentrates’’ for 1997.Separately, they were no longer as significant to the indus-try as was the case in 1992. Other changes to Item 18Bstemmed from the transition to the NAICS. On Form 1303,‘‘Oil and Gas Field Services,’’ a product line on ‘‘Geophysi-cal surveying for oil and gas’’ was added to facilitate clas-sifying establishments in the NAICS. Similarly, Form 1403,‘‘Chemical and Fertilizer Minerals and Nonmetallic Miner-als,’’ split ‘‘Bituminous limestone and bituminous sand-stone’’ from the product line for ‘‘Native asphalt andbitumens’’ to attain comparability within the NAICS. Onseveral questionnaires, ‘‘Lubricating oils and greases’’ werecombined with ‘‘All other supplies’’ because separate datacould not be published without disclosure problems.

Service Industries

For the 1997 census, the former SIC Service Industriesdivision was expanded and reorganized into eight NAICSservice sectors. The new NAICS sectors, derived largely,but not exclusively, from Service Industries were as fol-lows: Information; Professional, Scientific, and TechnicalServices; Administrative Support, Waste Management andRemediation Services; Educational Services; Health Careand Social Assistance; Arts, Entertainment, and Recreation;Other Services (Except Public Administration); andManagement of Companies and Enterprises (the latterdrew elements from all the SIC divisions).

The items asking for the EIN and physical location onthe 54 services questionnaires were changed to conformto those used on the rest of the census questionnaires.Specific lines or items were dropped from various ques-tionnaires due to poor response or to budget limitations(e.g., for Form SV-7903, ‘‘Gambling, Amusement andGambling,’’ lines asking for data on docking and storagecapacity, as well as on bowling, were deleted from Item 9,‘‘Sources of Revenue’’). On taxable and nontaxablebusinesses questionnaires (e.g., Form 8704, ‘‘Managementand Consulting Services’’ (taxable) and Form SV 8004,‘‘Hospitals’’ (nontaxable), instructions were added to Item4, ‘‘Legal Form of Organization,’’ to help companies selectthe appropriate check box and to resolve problems withdata on limited liability partnerships.

The adoption of the NAICS also led to the creation ofseveral new questionnaires, including long-form question-naires Forms 7303, ‘‘Specialized Design Services,’’ and7309, ‘‘Miscellaneous Professional Services,’’ and fiveshort-form questionnaires, to cover newly identified indus-tries. In addition, communications industries (former SICMajor Group 48, covering telecommunications, includingtelephone and telegraph, radio and television broadcast-ing, cable and other pay television services, and communi-cations services not elsewhere classified) was transferredfrom Transportation, Communications, Electric, Gas, andSanitary Services to the NAICS ‘‘Telecommunications(consolidated)’’—one of the new service classifications.

Most changes to the service questionnaires addedinquiries to Items 8, ‘‘Kind of Business,’’ and 9, ‘‘Sources ofReceipts.’’ These changes either improved the level ofdetail data available from the census or facilitated industryclassification. For example, Form 7302, ‘‘Administrativeand Other Support Services,’’ combined inquiries fromthree 1992 questionnaires into eight new lines in Item 8,‘‘Kind of Business’’—(1) collection services, (2) adjustmentbureau, (3) secretarial services, (4) court reporting, (5) allother support services, (6) all other administrative ser-vices, (7) telemarketing bureau and soliciting services, and(8) recovery and repossession services—while deletingtwo lines (blueprinting and photocopying, and duplicatingservices) to facilitate classification under the NAICS.

Form 7301, ‘‘Advertising and Related Services,’’ hadseven new lines added to Item 9—(1) display advertising



services, (2) direct mail, (3) distribution or delivery ofadvertising materials, (4) fees for soliciting advertising formedia, (5) commissions from buying advertising space ortime and reselling, (6) sales of ‘‘specialty advertising’’products, and (7) sales of merchandise except specialadvertising products—to provide support for industrycoding and improved industry detail.

Finance, Insurance, and Real Estate

Changes made to the 24 questionnaires for these indus-tries reflected (1) improvements from experience gained inearlier census coverage, (2) detail needed for classificationunder the NAICS, and (3) items deleted because of poorresponse in previous censuses or because of budget limi-tations. Based on previous experience, the Census Bureauadded ‘‘Subchapter ‘S’ corporations’’ to Item 7, ‘‘LegalForm of Organization’’ on all questionnaires to helprespondents select the correct check box. The 1992 FormCB-6000, ‘‘Depository Institutions,’’ was split into threequestionnaires for the 1997 census—FI-6001, ‘‘Banks andDepository Credit Intermediaries, Except Credit Unions’’;FI-6002, ‘‘Credit Unions’’; and FI-6300, ‘‘Foreign Banks andInternational Trade Finance’’—because the amount ofdetail required to cover all these industries on a singlequestionnaire had confused the respondents. Items onquestionnaires directed to insurance agencies and to realestate agents and managers were reworded to improvedatacapture on commissions paid to independent agents.

Changes made to accommodate the NAICS requiredadding, deleting, or rewording lines in Item 8, ‘‘Kind ofBusiness or Activity,’’ and Item 9, ‘‘Sources of Revenue.’’For example, on Form FI-6102, ‘‘Activities Related to CreditIntermediation,’’ the Census Bureau added a kind-of-business check box and a revenue line for loan servicing.This activity is recognized in the NAICS industry 52239,‘‘Other Activities Related to Credit Intermediation.’’ Threenew questionnaires were added to cover rental and leas-ing of automotive and other machinery and equipmentwere FI-6504, ‘‘Automotive Equipment’’; FI-6505, ‘‘Rentaland Leasing of Machinery and Equipment’’; and FI6506,‘‘Miscellaneous Rental and Leasing.’’

Finally, certain inquiries originally intended to assist inassigning industry classification codes were dropped dueto poor response or to budget limitations (e.g., from the1992 Form FI6501, ‘‘Real Estate Operators and Lessors,’’an item asking about managed or leased properties).

Transportation and Utilities Industries

The biggest change within this group of questionnairesresulted from the transfer of communications industriesout of the sector. Also, in previous censuses, transporta-tion, communications, and utilities industries had coveredtelecommunications firms, but for the 1997 EconomicCensus, this subsector was transferred into one of theservice industries.

For the 1992 census, the design of the transportation,communications, and utilities industries questionnairesdiffered from those for other economic sectors in theplacement of the kind-of-business and source-of-revenueitems. By contrast, in the 1997 census—for the transporta-tion and utilities industries—this organization waschanged to conform with that of other questionnaires. Theeffect was to standardize overall design and to facilitateprocessing of all census materials. As with other question-naires, minor wording changes were made and the ques-tion on physical location of the establishment waschanged to request the county.

Most of the more substantial changes in the transporta-tion questionnaires were made to individual lines (i.e.,sources) listed in Item 9, ‘‘Sources of Revenue,’’ either tofacilitate classification or to improve comparability underthe NAICS. For example, on Form UT-4200, ‘‘Trucking andWarehousing,’’ the Census Bureau—

• Added a line asking for revenue from ‘‘Process, physicaldistribution, and logistics consulting.’’

• Deleted the line on ‘‘waste collection—operation ofincinerator, landfill, transfer station, or other disposalsite.’’

• Modified the wording on several lines to conform to theNAICS language (e.g., ‘‘sanitary services’’ was changedto ‘‘waste collection,’’ and ‘‘...and disposal’’ was deletedfrom the line originally reading ‘‘hazardous waste collec-tion and disposal’’).

The NAICS classification requirements also affected thecontent of the utilities questionnaires. On Form 4900,‘‘Electric, Gas, and Water Utilities,’’ the Census Bureauadded a line to Item 9, ‘‘Sources of Revenue,’’ for ‘‘electricpower transmission,’’ and deleted lines on ‘‘sanitaryservices’’ and ‘‘sales of scraps and waste materials.’’

SUMMARY

Questionnaire design and development for the 1997Economic Census was a carefully considered and lengthyprocess. From the consultations, to notifications in theFederal Register, and the incorporation of the NAICS, theCensus Bureau created questionnaires gathering the mostspecific data ever reported on the Nation’s businessactivity.



Chapter 5. PROMOTING CENSUS AND PRODUCT AWARENESS

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37Two Moments of Truth: 1954 and 1997 . . . . . . . . . . . . . . . . . . . 37NAICS and the 1997 Promotional Campaign. . . . . . . . . . . . . . . . . 37

Lessons Derived From the 1992 Economic Census . . . . . . . . . . . . . . 38Unify and Clarify the Economic Census . . . . . . . . . . . . . . . . . . . 38Integrated Response and Product Marketing . . . . . . . . . . . . . . . . 38Objectives of response marketing . . . . . . . . . . . . . . . . . . . . . 38Objectives of product and services marketing . . . . . . . . . . . . . . 39

Remind Respondents That Response is Mandatory by Law . . . . . . . . . 39Improve the Ease/Reduce the Costs of Business Response to the Census . 39Target Large Companies . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Promotional Roles Played by Various Census Bureau Divisions . . . . . . . . 40Economic Planning and Coordination Division (EPCD). . . . . . . . . . . . 40Lisboa Associates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

Public Information Office (PIO) . . . . . . . . . . . . . . . . . . . . . . . . 40Congressional Affairs Office (CAO). . . . . . . . . . . . . . . . . . . . . . 41The Marketing Services Office (MSO). . . . . . . . . . . . . . . . . . . . . 41Customer Liaison Office (CLO) . . . . . . . . . . . . . . . . . . . . . . . . 41

Large Company Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41Groundwork for Identifying and Contacting Large Companies . . . . . . . 41Economic Statistical Methods and Programming Division (ESMPD) . . . . 41Governments Division (GOV) . . . . . . . . . . . . . . . . . . . . . . . . 41Subject divisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41National Processing Center (NPC) . . . . . . . . . . . . . . . . . . . . . 42

The Account Manager/Company Contact Program . . . . . . . . . . . . . 42Advance information . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42SSEL search and query tool. . . . . . . . . . . . . . . . . . . . . . . . . 42Electronic company notepad . . . . . . . . . . . . . . . . . . . . . . . . 42

Account Manager Calling System . . . . . . . . . . . . . . . . . . . . . . 43Account manager’s calling cycle . . . . . . . . . . . . . . . . . . . . . . 43

The Mail Campaign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43Original mailout letters. . . . . . . . . . . . . . . . . . . . . . . . . . . 43Successive letters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Special Role of the Internet . . . . . . . . . . . . . . . . . . . . . . . . . . . 43Writer’s Tool Kit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Product Marketing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Target Markets for Product Marketing . . . . . . . . . . . . . . . . . . . . . 44Congress and the Administration . . . . . . . . . . . . . . . . . . . . . . 44Corporations, Small Businesses, and Communications Sector TradeAssociations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44Federal Agencies, State, County, and Local Governments. . . . . . . . . . 44

CONTENTS


Target Markets for Product Marketing—Con.SDCs/BIDCs, Depository Libraries, Small Business DevelopmentCenters (SBDCs), Chambers of Commerce, Manufactures ExtensionPlants (MEPs), Agricultural Extension Offices, Universities and OtherResearchers and Educators . . . . . . . . . . . . . . . . . . . . . . . . . 44Data retailers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

Marketing Tactics to Promote Data Products . . . . . . . . . . . . . . . . . . 44


Promoting Census and Product Awareness

‘‘Every census—whether of people or of business establishments—must persuade those from whom they would col-lect the data to provide the information. No data response, no census results.’’

John P. Govoni, ChiefEconomic Planning and Coordination DivisionU.S. Census Bureau

INTRODUCTION

In promoting the economic census, the U.S. CensusBureau must consider how to increase the willingness ofbusinesses to participate. Even though response isrequired by law, many businesses fail to participate. Vari-ous causes have been identified, including the cumulativeeffects of the paperwork burden, changing attitudestoward government, and diminishing administrativeresources because of business downsizing.The economic census is absolutely critical to support-

ing ongoing survey programs, to maintaining the StandardStatistical Establishment List (SSEL), which serves as theCensus Bureau’s Business Register, and to providing thebenchmark data upon which many other national eco-nomic measures are predicated—including the GrossDomestic Product (GDP) estimates. Therefore, the CensusBureau undertakes a wide array of activities to guaranteethat this most important of all business data collectioneffort gains notice and obtains responses from thebusiness sectors. Those activities are the subject of thischapter.

Two Moments of Truth: 1954 and 1997

Each economic census occurs within unique historicalcircumstances that condition the Census Bureau’sapproach to promoting awareness. The 1954 EconomicCensus, for example, received tremendous advance cover-age. Scheduled to take place in 1953, funds for the censuswere never appropriated. Instead, an Intensive Review—Committee headed by Dr. Ralph J. Watkins—was commis-sioned in October 1953 by the Eisenhower Administrationto study if the economic census was necessary. Business,financial, professional, and governmental groups wererepresented on the committee which, in February 1954,issued its unreserved support for the economic census—

We are in the midst of great technological revolutionwhich is accelerating change, hastening obsolescence,creating new industries and transforming old ones,

remaking the industrial map of the country, and bring-ing within the range of the feasible great heights ofproduction, productivity, and well-being. . . The needfor the benchmark statistics provided by the CensusBureau is greater today than ever before, and promisesto grow in intensity.— Watkins Commission Report,1954.Within months, the economic census was reinstated,

and the 1954 Economic Census had taken its special placein the history books. No promotional campaign has quitetopped that one, but to promote awareness of the 1997Economic Census, a booklet was produced, The EconomicCensus—Two Moments of Truth: 1954 and 1997. Thebooklet drew parallels between the1954 and the 1997economic censuses. An original print run of 10,000 book-lets in September 1997 was quickly followed by anotherprint run of 14,000 in June 1998: the item was highlypopular in the 1997 promotional effort. (See Appendix B,‘‘Historical Development of the Economic Census of theUnited States,’’ for more information on the history of theeconomic censuses.)

NAICS and the 1997 Promotional Campaign

Like the 1954 Economic Census, the 1997 EconomicCensus found itself at a special juncture in history—particularly with reference to the creating of new indus-tries and the transforming of old ones. The 1997 censuswas to be the first to collect data in terms of the newNorth American Industry Classification System (NAICS)which replaced the Standard Industrial Classification (SIC)system. The SIC had been in use, and revised severaltimes, since the 1930s. The new system identified severalhundred new industries that had not been recognizedseparately in earlier censuses, and it revised the classifica-tion of many hundreds more. Therefore, in addition totraditional activities, the promotional campaign for the1997 took into consideration that—

• Respondents needed to know that their answers wouldestablish the baseline for a new coding system.



• Data users needed to know that 1997 Economic Censusproducts would have many advantages not known tothem before, and that using these products wouldpresent unfamiliar challenges.

Special brochures, question and answer sheets, andother promotional materials were formulated to explainthe NAICS and its effects on both those preparing thecensus forms as well as data users.The introduction of the NAICS gave additional impetus

to the integration of two objectives for the awarenesscampaign:

• Promoting awareness of the economic census duringthe data collection effort.

• Promoting awareness of products and services that theCensus Bureau would provide after data were tabulatedand released.

LESSONS DERIVED FROMTHE 1992 ECONOMIC CENSUS

Following the 1992 census, the Census Bureau formedseveral committees to evaluate the census and to recom-mend improvements for the 1997 enumeration. Thus, the1992 Economic Census Multiunit Response ImprovementCommittee came into being and issued a report. In 1995,this report was used as a departure point by the 1997Economic Census Response Improvement Team which,in turn, made additional recommendations of its own.In effect, these recommendations centered around thefollowing principles:

• Clarify—in the minds of respondents—that theeconomic census is a unified event.

• Coordinate and integrate response and productmarketing.

• Remind respondents of the mandatory nature of theirresponse.

• Improve the ease—that is, reduce the costs—ofcompanies in responding to the census.

• Target large companies.

Unify and Clarify the Economic Census

Prior to 1997, the economic census lacked a unified,recognizable identity. As a whole, it was routinely referredto as ‘‘the economic censuses.’’ It was perceived and con-ducted in terms of a series of discrete events, namely, thecensus of retail trade, the census of manufactures, and thelike. Therefore, a business executive might read about‘‘the economic census’’ in the media, but fail to connect itto the data he may have referenced from a census ofmanufactures report of the last census, or to the census ofservices industries questionnaire that just arrived in the

mail. Each of these censuses had equal status, and theeconomic census was characterized as a unified endeavoronly administratively, via budget submissions.Lack of a clear identity for the economic census pre-

sented a singular task to those charged with promotingresponse, given that no business would receive a form bythat name. To correct this situation, the ‘‘1997 EconomicCensus’’ was adopted as the formal name for the datacollection effort. This title appeared above—and in largerprint than—references to sector-specific titles on thecensus forms, instruction sheets, cover letters, and dataproducts. Keeping the sector titles in a subordinate statuswhile publicizing the larger effort enhanced the presenta-tion of the economic census as an integral whole, ratherthan as a collection of separate economic measures.More importantly, this approach raised the profile of theeconomic census as the major measurement event of theNation’s economy.

Integrated Response and Product Marketing

Response marketing—i.e., making respondents awarethat the census was being taken and that, by law, theymust fill out the census forms and return them to the Cen-sus Bureau by the February 12, 1998 due date—was a keystep in successfully conducting the 1997 Economic Cen-sus. Even though the Census Bureau had long observed a‘‘disconnect’’ between the user and the provider of censusinformation, the agency knew these were two sides of thesame coin. Therefore, in its promotion efforts, it wasimportant to link collection with dissemination. Awarenessof benefits would motivate some respondents and alsowould help legitimize the Census Bureau’s requests formedia attention.Moreover, the Census Bureau was intent upon showing

appreciation to ‘‘intermediaries’’—i.e., the Chambers ofCommerce, associations, news media, and others thatcarried its promotional messages to potentialrespondents—by informing those intermediaries about theresults of the census. Since results are always in terms ofproducts, the close synchronization of census and productpromotion was reinforced.To coordinate the collection and dissemination func-

tions, staff members at all levels of Census Bureaudivisions participating in this census were informed andinvolved in the 1997 awareness campaign. It was impera-tive that the themes developed in marketing the censusresponse were carried into the marketing of economicproducts and services, providing a focused, consistentmessage. In order to accomplish this, the data productportfolio was developed early on in the census planningphase. The publicizing of the mix of data products andservices was anticipated from the outset of respondentoutreach efforts.

Objectives of response marketing. The chief goal ofresponse marketing was to raise awareness that data col-lection was taking place and that response was important.



Both those preparing the forms (such as accountants) andthose who might influence the preparers (such as com-pany CEOs, the media, Chambers of Commerce and otherorganizations) needed to be targeted with this message.The Census Bureau expected that increased awarenesswould help to increase the overall level of response tothe census, and to accelerate the rate of response—particularly by very large companies. Also, responsemarketing was intended to set the stage for product andservices marketing.

Objectives of product and services marketing. Thechief goals of product and services marketing were toincrease awareness of the Census Bureau’s products andservices and to increase their use by data users. The Cen-sus Bureau wished to cultivate satisfied users who could,in turn, serve as advocates, spokespersons, and role mod-els for the 1997 Economic Census. This would facilitatefuture data collection activity. The Census Bureau does,after all, conduct a population census every 10 years, aneconomic census every 5 years, and hundreds of eco-nomic and demographic surveys every year.

Remind Respondents That Response IsMandatory by Law

Taking a cue from the 1992 Economic Census—in whichit was demonstrated that placing the mandatory messageon the mailout envelope1 was perhaps the most effectiveelement of 1992 response improvement efforts—planningfor 1997 expanded on this principle as follows:

• Included the message—‘‘Your response is requiredby law’’—on all outgoing form packages, multiestablish-ment companies, as well as single-establishmentcompanies and on follow-ups and remails.

• Remailed to reinforce the message. In 1992, the CensusBureau only remailed questionnaires to single-establishments (3 remails); in 1997, questionnaireremails also were sent to small and medium multiestab-lishments companies (1 remail). For 1997, businessesthat had been granted time extensions were remindedof the reporting requirement; in 1992, businesses withextensions were excluded from reminder mailings.

Improve the Ease/Reduce the Costs ofBusiness Response to the Census

Responding to the census takes time. This is the casewhether in-house staff of large companies, or accountantshired by small companies, complete the questionnaires.How might the Census Bureau make this response lessburdensome? An important lesson in this regard from the1992 census was applied to the Census Bureau’s focus onthe response rate for the 1997 census. During the 1992

census, the Census Bureau inadvertently exacerbatedunwillingness to respond to the census. That is, twoeconomic surveys were conducted for the first time—theAnnual Capital Expenditure Survey (ACES) and the Com-modity Flow Survey (CFS)—and responses to these surveyswere to be compiled by businesses already answering the1992 Economic Census questionnaires. The paperworkburden was substantial. The Census Bureau took care notto repeat this error in 1997—the CFS was conductedthroughout the year, previous to the mailout of the 1997census questionnaires, so that this survey was not in the‘‘field’’ at the same time as the 1997 Economic Census.The ACES was done as well, but the survey actually low-ered respondent burden by allowing the Census Bureau todiscontinue other surveys of business investment.Further lessons from the 1992 census were applied for

the 1997 census to ease the way for businesses torespond in a timely, cost-effective manner. To this end, theCensus Bureau—

• Made electronic reporting easier. For 1997, the CensusBureau greatly expanded its outreach to the businesssector in this regard, inviting approximately 600 compa-nies to report electronically—using either the Windows-based Computerized Self-Administered Questionnaire(CSAQ) or magnetic computer tape. (See the ‘‘AccountManager/Company Contact Program’’ below for moreinformation on electronic reporting. Chapter 8, ‘‘DataCollection and Processing,’’ also discusses electronicreporting.)

• Projected greater accessibility. To answer business ques-tions, the Census Bureau prominently included an 800number on the questionnaires and provided its Web siteaddress—www.census.gov/econ97—on all census coverletters. The web page included an ‘‘Ask Dr. Census’’feature.

• Provided sample questionnaires on the Internet.

• Uploaded Portable Document Format (pdf) question-naires to the Internet, so that respondents could pre-view the 1997 Economic Census questions specific totheir industry. This was done in March 1997.

Target Large Companies

Multiestablishment companies classified as ‘‘large’’ (‘‘L’’)companies constituted only 7 percent of all establish-ments to which forms and classification cards were mailedin the census, but these companies represented over 30percent of the economic activity of the country. (The clas-sification mailout is discussed in Chapter 8, ‘‘Data Collec-tion and Processing.’’) Their responses were essential toensure data accuracy—particularly because of the need toobtain correct NAICS Codes. Thus, a significant part of thecensus awareness program was devoted—as in previouseconomic censuses—to ‘‘large company response.’’ (Forthe 1997 Economic Census, the final response rate formultiestablishment ‘‘L’’ companies was 88.9 percent.)

1For the 1992 census, the message appeared only on enve-lopes for mailout to single-establishment companies.



PROMOTIONAL ROLES PLAYED BY VARIOUSCENSUS BUREAU DIVISIONS

The following Census divisions and other organizationscoordinated their actions to promote awareness of thiscensus.

Economic Planning and Coordination Division(EPCD)

Census and product awareness activities were centeredwithin the Economic Planning Staff (EPS) of EPCD. Thisstaff coordinated with, and provided input and guidanceto, the outside contractor advertising agency—LisboaAssociates, hired in the Fall of 1996. An outside contractoralso had been used in the 1992 Economic Census.

Lisboa Associates. The advertising agency was respon-sible for implementing most aspects of the communica-tions contract, under oversight by the ContractingOfficer’s Technical Representative. Lisboa Associates’functions included the following:

• Research: Interviewing Census Bureau staff. Conductingfocus groups with potential respondents (accountants)and intermediaries (news media, Chambers of Com-merce, and trade associations) and interpreting theresults.

• Strategy: Recommending approaches based uponresearch findings.

• Creativity: Developing artwork and editorial copy. Thisartwork was to convey a unified census, and in thatregard, distinguished itself from the graphics of previ-ous economic censuses.

• Development and distribution of media kits: Facts forMedia contained press releases, sample editorials, factsheets, a NAICS brochure, clip art, response dates, infor-mation on available assistance, answers to typicallyasked questions about the economic census, examplesof how the data are used, findings from the 1992 cen-sus, and Public Service Announcements (PSAs). In all,10,000 media kits were distributed to general businesspublications (1,021), industry associations (2,112),newspapers (905), regional business publications (823),Chambers of Commerce (3,027), and DataCenter/Business and Industry Data Center affiliates(2,189).

• Distribution of over 50,000 Accountants’ Guide to the1997 Economic Census. These booklets were sent to

independent accountants who were members of theNational Accounting Association and the American Soci-ety of Certified Public Accountants. The 1987 EconomicCensus Advertising & Response Behavior Study showedthat about six percent of small businesses used outsideaccountants to complete their census forms.

• Duplication/distribution: Producing print materials andtwo radio PSAs. In three waves, print materials in infor-mation kits were ‘‘direct-mailed’’ to media, businessgroups, and intermediaries. The press releases, editori-als, and PSAs were tailored to reflect the census time-table, as follows:

• Wave 1 Message: ‘‘The questionnaires are coming.’’(for use September to December 1997).

• Wave 2 Message: ‘‘The questionnaires are here. ’’ (foruse January to February 1998).

• Wave 3 Message: ‘‘The questionnaires responses arepast due.’’ (for use March 1998).

• Internet: Creation of digital versions of editorial and adcopy, radio programs, and other materials. Developmentof web pages organizing the material to be posted onthe Census Bureau Internet site.

• Analysis and evaluation: Providing a newspaper clippingservice, maintaining statistics on clips, response cards,and the like.

• Fulfillment: Sending profiles to media requesting them(in response to offers contained in response cards), andsending messages to media who provided coverage ordonated—space thanking them for their help.

After the above activities had been implemented, theCensus Bureau extended the contract of Lisboa Associatesto handle promotion of 1997 census products (see lastsection of this chapter).In addition to coordinating with the ad agency, EPS

handled most media interface and interviews. For use inits own direct contacts with associations, trade press, andrespondents, EPS also distributed copies of promotionalmaterials in quantity.

Public Information Office (PIO)

The PIO was responsible for—

• Working with the contractor as part of the creativeteam, specifying contract details, developing andevaluating components of the package, and integratingcontract output with other publicity activity.

• Developing press releases and other publicity materialsrelated to contract elements.

• Developing video and radio materials, and distributingthem through existing channels, including networkfeeds.



• Engineering cooperative efforts, such as the NationalAssociation of Broadcaster satellite transmission onFebruary 12, 1998.

• Cultivating and/or assisting in interviews betweenCensus Bureau staff and government officials, businessleaders or data users. These taped interviews wereuseful as elements of the media kit, in radio and TVmaterials, and in brochures to make appeals to‘‘stakeholders.’’

Congressional Affairs Office (CAO)

The CAO engaged in—

• Assisting in development of strategies for communicat-ing with the Congress about the 1997 EconomicCensus.

• Helping design a version of the media kit for congres-sional offices.

• Setting up Congressional staff briefings on theeconomic census.

The Marketing Services Office (MSO)

The MSO participated by providing guidance withrespect to product marketing activities.

Customer Liaison Office (CLO)

The CLO had jurisdiction over—

• Assisting in development of strategies to communicatewith the business community through State DataCenter/Business and Industry Data Center (SDC/BIDC)units.

• Helping to develop mailing lists of SDC/BIDC affiliates.

• Setting up briefings for SDC/BIDCs—telling them of theCensus Bureau’s plans.

The promotional roles of Economic Statistical Methodsand Programming Division (ESMPD), Governments Division(GOV), various subject matter divisions, and the NationalProcessing Center (NPC)2 in Jeffersonville, IN, are dis-cussed in the next section.

LARGE COMPANY PROGRAM

If the goal of the Census Bureau was to publish resultsof the 1997 Economic Census earlier than ever before,then the corresponding mission of the Large CompanyProgram was to promote a prompt turnaround time for thecensus questionnaires. As mentioned, nearly one-third of1997 Economic Census data were furnished by seven per-cent of the Nation’s companies. However, the issue wasnot simply one of ‘‘large company response’’—large firms

do respond, eventually. But, responses from some largefirms come in so late that they significantly diminish theutility of census results. The question addressed by the1997 Large Company Program, therefore, was ‘‘How canthe Census Bureau motivate large companies to respondearlier than ever before—while maintaining the excellentrates of response (89 percent) attained in the 1992 Eco-nomic Census?’’Survey staff from subject divisions served as ‘‘account

managers’’ to specific companies with whom they formeda direct ‘‘contact.’’ The Account Manager Program will bedetailed below, but first, it is necessary to describe thework of the many census divisions that established thefoundation for the program.

Groundwork for Identifying and ContactingLarge Companies

Economic Statistical Methods and ProgrammingDivision (ESMPD). ESMPD assembled a list of potential‘‘L’’ and ‘‘M’’ (medium) companies.3 The Large CompanyResponse improvement team set the criteria: namely, thetop 1,000 companies—by establishment and bypayroll—in each SIC division were to be chosen. Thisresulted in a list of about 10,000 companies—

• At mailout there were 1,000 ‘‘Ls’’ and 9,000 ‘‘ Ms.’’

• After mailout, for purposes of follow-up and Manage-ment Information System reporting, the Census Bureauredefined ‘‘M’’ companies to include any company with1,000 or more employees.

ESMPD also produced several mailing materials, includ-ing the Contact Exchange Card and cover letters, both ofwhich are referred to below.

Governments Division (GOV). For the 1997 census,the ESMPD did not have the resources to develop theaccount manager Intranet application. Therefore, GOVdeveloped the application, and account managers used itto record their contacts with ‘‘L’’ companies. Since therewas no direct interface between the intranet applicationand the SSEL, data were downloaded and sent to ESMPD tobe applied to the SSEL. This took place in July 1997, inorder to apply the account manager and Contact ExchangeCard names/addresses to the SSEL before mailout.

Subject divisions. The subject divisions participated inearly planning and fact-finding stages of the census. Theyexamined the list of potential ‘‘L’’ companies assembled byESMPD and then applied criteria for selecting companiesfor the Large Company Program. The criteria included

2Until a reorganization effective on May 1, 1998, the NPC wascalled Data Preparation Division (DPD).

3The designations ‘‘L’’ and ‘‘M’’ do not refer to statistical datacollection, tabulation, and publication categories. ‘‘L’’ and ‘‘M’’ aresimply ways the Census Bureau identifies companies needing suc-cessive levels of follow-up (see also footnote 4 in Chapter 8, DataCollection and Processing).



importance of the company in the sector, as well asresponse history of that company. Subject division staffalso helped evaluate alternative strategies.The largest involvement of subject division staff was

through serving as account managers in the Large Com-pany Program.4 Selected division staff were so designated,and each staffer was assigned 5 to 25 companies. Thesubject divisions and the number of companies assignedto each were as follows—

• MCD—400 companies

• SSSD—500 companies

• EPCD—136 companies

National Processing Center (NPC). While ‘‘M’’ compa-nies were not assigned individual account managers, aContact Exchange Card sent to these companies listed aphone number from either MCD or SSSD that the companycould call. In addition, NPC staff conducted several tele-phone follow-ups from June through August 1999.

The Account Manager/Company Contact Program

Advance information. In November 1996, the NPCmailed advance materials to ‘‘L’’ and ‘‘M’’ companies. Eachpacket contained—

• A customized cover letter (including lists of forms thecompany would receive).

• A Contact Exchange Card consisting of two compo-nents: (1) a ‘‘rolodex’’ card with either an account man-ager’s name, phone number, and e-mail address or a listof census sectors and corresponding phone numbersfor assistance, and (2) a blank form that requestedaddress correction, name of contact person at thecompany, and contact person’s supervisor.

• An electronic reporting flyer inviting about 600companies to report electronically. Ultimately, 671companies—representing 317,593 establishments—availed themselves of the electronic reporting option.

• An Advance Information booklet (featuring a list ofcensus contacts, a census Question and Answer sheet, aguide for obtaining census questionnaires from theInternet, and two sample forms).

One of the key tasks for an account manager was toidentify to whom to address correspondence, phone calls,questionnaires and the like throughout the duration of theprocess until completed census forms were received. The‘‘addressee’’ was the ‘‘company contact,’’ and was

requested on the Contact Exchange Card, as well as intelephone calls from the account manager. This name andaddress also would be used to mail forms to the company.

For 1997, initial calls by account managers occurredearlier than at the equivalent juncture in the 1992 census.The purpose of the calls was to—

• Verify receipt of advance materials, obtain the identityof the company contact, and verify the mailing addressfor sending census forms before the census mail filewas frozen in July 1997.

• Raise awareness of the coming census forms and theneed for the company to begin preparations.

• Offer assistance throughout the census-taking interval.

One of the guiding principles of the AccountManager/Company Contact Programs was that onlyaccount managers should be able to grant time extensionsor negotiate reporting arrangements. To do this, accountmanagers had more information than ever before abouttheir ‘‘L’’ companies—including a company profile andcopies of all advance information sent to companies. Inaddition, account managers had access not only to Inter-net resources (including e-mail, and access to Security andExchange Commission filings and company lookups) butalso to several new tools such as the SSEL search andquery tool and the electronic company notepad.

SSEL search and query tool. This feature displayedcompanies and their establishments based upon size,industry, form, extension date, and other factors. Withsearch and query, it was easier for account managers to—

• Isolate reporting problems—such as a single subsidiaryof a company in which most other units had reported.

• Identify delinquent companies significant in a particulargeographic area or industry.

Electronic company notepad. In 1992, the capacity ofaccount managers to act effectively was limited by a pre-dominantly paper-based record-keeping system. Sharing ofinformation was difficult or impossible. A business mightcontact any of several dozen persons located in threeseparate buildings at Census Bureau headquarters in Suit-land, MD, or any of hundreds of analysts and telephoneoperators in Jeffersonville, IN. Thus, one individual mighthave worked out an arrangement with a company—andrecorded it on a paper document that no one else mayhave seen—only to have someone else grant the companyan extension or reach a different agreement. By develop-ing the electronic company notepad linked to the SSEL, theCensus Bureau enabled staffers to leave a record of whathad been done, and why, so that all arrangements wouldbe visible and up-to-date. The electronic company notepadwas used, predominantly, by NPC staff.

4Account managers were assigned only to ‘‘L’’ companies, notto ‘‘M’’ companies. However, ‘‘M’’ companies were included inother aspects of the Large Company Program. Fifty mining com-panies were selected as ‘‘Ls.’’ but were converted to ‘‘Ms’’ when itappeared that the mining census might be canceled.



Account Manager Calling System

The GOV Division developed and maintained theaccount manager calling system for the documentation ofaccount manager contacts with companies. Using the sys-tem, it was possible to record to whom the account man-ager talked, when, why, and with what results. The con-tact name and address was used to update the mailingaddress for census forms, particularly if the company didnot return a Contact Exchange Card.

Account manager’s calling cycle. Figure 5-1 shows theaccount manager’s calling cycle i.e., the intervals whenaccount managers were to contact companies and obtaininformation or offer assistance.

The Mail Campaign

Original mailout letters. The essence of the mail cam-paign was to emphasize the existence of the census, thedue date, and the legal requirement to respond, as well asto incrementally escalate the messages contained in each

successive letter. For example, the original mailoutletters— each of which were signed by the Director of theCensus Bureau—

• Highlighted the due date of February 12, 1998.

• Emphasized the importance of the data.

• Cited applicable sections of Title 13, United States Code.

• Indicated that response was mandatory and would bekept confidential.

• Provided the toll-free telephone number. Provided theWeb site address.

• Contained frequently asked questions, with answers onthe back of the letter.

• Quoted Alan Greenspan, Chairman of the FederalReserve Board, in endorsing the unique function of theeconomic census in gathering benchmark economicdata for the Nation.

Successive letters. Successive letters ‘‘upped the ante.’’There were 17 variations of follow-up letters, dependingupon whether the company was in the ASM or not, was asingle-establishment company or a multiestablishmentcompany, or had replied with only a partial response tothe census questionnaire. These letters included the fol-lowing additions to the original message:

• ‘‘OVERDUE NOTICE’’ printed in bold caps at the top ofthe page.

• ‘‘Please complete and return your census forms within10 days’’ or ‘‘immediately.’’

• Stronger assurances of confidentiality.

• ‘‘You may be subject to penalties for not returning thisform’’ or ‘‘You are subject to. . . ’’

• Reminder notice, if there was a pending extension.

SPECIAL ROLE OF THE INTERNET

For the 1997 Economic Census, the Internet played agreater role in both response and product marketing thanin any previous economic census. The economic censusweb page (www.census.gov/econ97) invited businesses to‘‘Ask Dr. Census.’’ E-mail queries addressed to ‘‘Ask Dr.Census’’ peaked at about 300 per month during Januarythrough March 1998, then slowed to 200 per month untilphase-out in September 1998. In addition, a directory ofquestionnaires in portable document format (pdf) receivedthousands of ‘‘hits’’ between December 1997, and June1998, when data were being collected. In addition, a pagewas established to describe and provide an example ofhow to use the Computerized Self-Administered Question-naire (CSAQ).

Figure 5-1.Account Manager’s Calling Cycle

• Verify receipt of advance information booklet.• Encourage return of exchange card.• Introduce yourself as the Account Manager and offer assistance in preparing for the census.

Follow-up mailing of advance information to companies not returningcards or receiving calls.

Calls to companies not satisfied in Round 1, even if they returned cards.Complete by March 31.

• If we received the card, acknowledge it.• If we have not received card, alert company that material is coming, encourage return of contact exchange card.• Introduce yourself as the Account Manager and offer assistance in preparing for the census.

Re-establish contact to obtain name/address changes prior to censusmailout. Complete by July 15.

• Discuss special requirements such as split mail or electronic reporting.• Discuss ways to improve quality and timeliness of reporting.• Remind company that forms will be mail December 1997.• Offer assistance in preparing for the census.

Final mail file preparation

Forms mailed

• Verify that company received its census forms. Remind that forms are due February 12, 1997.• If necessary offer extension not to exceed May 1, 1997.• Offer assistance in reporting for the 1997 Economic Census.

Forms due

Companies that respond

• Thank the company for their response

Companies that do not respond

• Remind company that forms were due February 12, 1998.• Get an estimate of when forms will be returned.• Offer assistance in reporting for the 1997 Economic Census.

Reminder mailings to delinquents and unexpired extensions


• Remind company that forms were due Feb. 12, extensions have expired, forms are due now.• Obtain an estimate of when forms will be returned.• If the company contact is uncooperative, appeal to the executive on the contact exchange card.• Offer assistance in developing estimates.


• Remind company that forms were due Feb 12, extensions have expired, forms are due now.• Obtain an estimate of when forms will be returned.• If the company contact is uncooperative, appeal to the executive on the exchange card.• Offer assistance in developing estimates.

Account Managers Calling Cycle

Call No. Purpose of Call Timing of Call

Dec 96 - Jan 97

mid-Feb 97

Feb - Mar 97

June 97

Sep 97

mid - Dec 97

Jan 98

Feb 12, 1997

After check-in

Mar 98

April - June 98

June 98

As required

1

**

2

3

**

**

4

**

5

6

**

7

8



Writer’s Tool Kit

The economic census Internet page also was used tohelp reporters prepare stories about the census. The‘‘Writer’s Tool Kit’’ was established to provide an electronicversion of public service announcements (PSAs) and othermedia kit contacts, also listing quotes from AlanGreenspan and other prominent economists and businessleaders.

Product Marketing

In promoting products and services from the 1997census, the Census Bureau was responding to data users’wishes for flexible products that would keep up withchanging needs. Moreover, the Census Bureau fulfilled animplied promise—made during the response marketingphase—by providing census results to the media, businessorganizations, and intermediaries who helped promotecensus response.Therefore, the 1997 product mix included fewer paper

products and more electronic products. In addition, therewere fewer ‘‘fixed’’ format data tables and more access todata bases. This enabled users to extract, download, andwork with the data in their own formats. While Censusproducts and services resulting from the 1997 EconomicCensus will be discussed in detail in Chapter 12, Publica-tion Program, suffice to say here that the Internet washighly instrumental in product marketing. It provided—

• Database files.

• ‘‘Screen shots’’ enabling customers to ‘‘test drive’’selected data products, before deciding to buy them.These included downloadable summary data files andCD-ROMs, and accompanying software.

TARGET MARKETS FOR PRODUCT MARKETING

Congress and the Administration

This constituency relies upon census information tobuild profiles, evaluate the need for legislative changes,determine the performance of programs, judge the effectof legislative and program changes on constituency, andestimate the costs of legislation. Users looked at both theprimary data and the value added products, such as GDP.

Corporations, Small Businesses, andCommunications Sector Trade Associations

These users employ census information to prepare theirmarketing plans, growth plans, estimates, trends and fore-casts, news reports, articles, feature stories and books,research projects. They use economic data to establishnew associations and to build membership of existingones, as well as to create strong positions for proposedlegislation, debates, and other issues. Included in this cat-egory are the questionnaire respondents.

Federal Agencies, State, County, and LocalGovernments

Public agencies use census data to prepare legislationand to plan development and programs. This group alsouses the data as a reference point for current statisticssuch as the GDP, the Producer Price Index, and the like.

SDCs/BIDCs, Depository Libraries, Small BusinessDevelopment Centers (SBDCs), Chambers ofCommerce, Manufactures Extension Plants (MEPs),Agricultural Extension Offices, Universities andOther Researchers and Educators

This array of users relies upon economic census data tomaintain a snapshot of the economy for various levels ofgeography, and for a given time frame. They typicallyassist the Census Bureau in distributing data locally, andin educating the public about myriad data uses.

Data Retailers

Data retailers acquire census data and repackage it forsale to their own customers. They need high-quality data,as well as metadata (descriptive and formatting material interms of which data are presented), to assure proper useof the information.

MARKETING TACTICS TO PROMOTEDATA PRODUCTS

The Census Bureau understood that any good promo-tional campaign consists of a multifaceted approach toannouncing its products to the user community. Theagency pursued the following strategies:

• Extended the contract with Lisboa Associates topromote the data products resulting from the 1997Economic Census.

• Produced the Preview of the 1997 Economic Censussummarizing facts about the census, such as itspurpose, classifications, data items, products, and thedata release schedule.

• Selected ‘‘New Data for Today’s Economy’’ as the newtheme.

• Assembled media kits for products and sent them to theoriginal media kit mailing list.

• Used drop-in advertising in trade and professionalmagazines to reach nonusers. After development,the ads were placed on the Internet for even furtherdissemination.

• Disseminated news releases (press releases, statestories, and feature stories) to business organizations,trade associations, state and Federal governmentagencies, universities, newspapers and magazines—allannouncing the availability of data.



• Mailed information kits directly to key data user organi-zations (journalists, Members of Congress, marketers,and the like).

• Targeted key media (through PIO) at the time of releaseof key reports.

• Conducted user conferences, coordinating them withCensus 2000 conferences. This included schedulingconferences to coincide with the availability of localarea data products.

• Highlighted major data results in multimedia(audio/video) presentations.

• Made presentations and staffed booths at national andstate exhibits and at trade shows, complete with anarray of census publications. These were augmented byspecialty items (pens, ‘‘post-its’’), as well as by variousbrochures and data summaries.

• Produced the Economic Census Slide Rule, a data sum-mary device showing 1997 summary results by sector(manufacturing, services, etc.) on both a SIC and aNAICS basis.

Figure 5-2. Hungry for Business Data?

Representing the essence of the effort to integrate census and product awareness, this post card was sent torespondents—both to thank them for their response and to let them know that the data were now available for their use.



Chapter 6. GEOGRAPHIC AREA CODING

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49The Geographic Interface Team (GIT) . . . . . . . . . . . . . . . . . . . . 49

Geographic Code Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 49General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Divisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50United States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Nationwide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Foreign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Offshore Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

Counties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51The 50 states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51District of Columbia . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Montana. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Maryland, Missouri, Nevada, and Virginia . . . . . . . . . . . . . . . . . 51Hawaii. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Puerto Rico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Guam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Northern Mariana Islands . . . . . . . . . . . . . . . . . . . . . . . . . 51Virgin Islands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Undistributed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

Places . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Incorporated place . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Census designated place (CDP) . . . . . . . . . . . . . . . . . . . . . . 51Consolidated city . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52Balance of consolidated city . . . . . . . . . . . . . . . . . . . . . . . . 52Special Economic Urban Areas (SEUA) . . . . . . . . . . . . . . . . . . . 52Balance of metropolitan area. . . . . . . . . . . . . . . . . . . . . . . . 52Balance of county. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Metropolitan Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53Congressional Districts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53Census Tract/BNA and Census Blocks . . . . . . . . . . . . . . . . . . . . 53

Geographic Recode and Reference Files . . . . . . . . . . . . . . . . . . . . 53General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53Census to Economic Recode File . . . . . . . . . . . . . . . . . . . . . . . 531992 Economic to Current Census Recode File . . . . . . . . . . . . . . . 531997 Economic to Current Census Recode File . . . . . . . . . . . . . . . 53Clustering Program Module . . . . . . . . . . . . . . . . . . . . . . . . . 54County Code Change File. . . . . . . . . . . . . . . . . . . . . . . . . . . 54County Alpha Recode File . . . . . . . . . . . . . . . . . . . . . . . . . . 54Geographic Stub File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54Subject Review City Reference File (CRF) . . . . . . . . . . . . . . . . . . 54

CONTENTS


Geographic Recode and Reference Files−Con.Geographic Notes File . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Geographic Reference Manual Table Files . . . . . . . . . . . . . . . . . . 55ZIP Code Legality File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

Geocoding Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Freestanding Geocoding System . . . . . . . . . . . . . . . . . . . . . . . 55City Reference File (CRF) . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Address Reference File . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Legality Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

Validating and Updating Incoming Data . . . . . . . . . . . . . . . . . . . . 56

48 Chapter 6 History−1997 Economic Census

Geographic Area Coding

INTRODUCTION

General Information

To publish tables of economic activity by geographicareas, the U.S. Census Bureau geocoded−that is, assignedgeographic codes−to each establishment record for whichdata were collected for the 1997 Economic Census. Aftergeocoding, the data from the establishment records weretabulated by geographic area. These geographic areaswere legally or administratively established, and statisti-cally defined, geographic entities.

The Geography Division (GEO) developed the geocodingprocess used in the 1997 Economic Census processing. Tobe geocoded, each establishment in the 1997 EconomicCensus needed to have an address, or supplementalinformation, that provided the physical location for thatestablishment.

Prior to mailout, the selected address for each estab-lishment record in the 1997 Economic Census control filewas processed through the geocoding system and a state,county, and place code was assigned to each record. Thisprocess set the initial geocodes in the control file for the1997 Economic Census cycle. Subsequent geocoding ofthe physical location response information and addresschanges from the responses to the economic censusquestionnaires were used to update the geocodes in thecontrol file as follows—

• For single-unit establishments, the physical locationaddress was selected for geocoding. If no physical loca-tion address was available, the 1992 Economic Censusmailing address was selected.

• For multiunit establishments, only one address field wasmaintained on the control file. The address was definedas both the physical location and the mail address, andwas used for geocoding.

• For the economic data collected by the Census Bureaufrom administrative records, only one address was avail-able on the record for geocoding and no additionalphysical location information was received.

Other geographic products, described separately in thischapter, were provided by GEO to the economic censusdivisions for use in tabulation, analysis, and publication ofthe 1997 Economic Census data.

The Geographic Interface Team (GIT)

In the fall of 1994, the Census Bureau established the1997 Economic Census Geographic Interface Team (GIT).Their mission was to plan the geocoding operation for

both the 1997 Economic and Agricultural Censuses.1

The team consisted of representatives from the followingCensus Bureau divisions or offices—

• Agriculture and Financial Statistics Division (renamedCompany Statistics Division [CSD] in April 1998).

• Services Division (renamed Service Sector StatisticsDivision [SSSD] in April 1998).

• Manufacturing and Construction Division (MCD).

• Economic Planning and Coordination Division (EPCD).

• Economic Statistical Methods and Programming Division(ESMPD).

• Geography Division (GEO).

• Computer Assisted Survey Research Office (CASRO).

The team met periodically until the summer of 1998,by which time the basic geographic plans for the 1997Economic Census had been finalized. It was decided thatthe geographic code structure for the 1997 EconomicCensus would be similar to that used for the 1992Economic Census.

The major questions addressed by the GIT includedthe overall geographic code structure—i.e., how the geo-graphic boundaries were to be set and what sourcesshould be used—and the designation of respectiveresponsibilities of the various Census Bureau divisions.In particular, GEO was responsible for developing a free-standing geocoding production system that would includethe geocoding software and all reference files needed tomeet the geocoding processing requirements of the 1997Economic Census. GEO supplied the geocoding productionsystem to ESMPD, who, in turn, installed the system on itsown computer platforms for production work.

GEOGRAPHIC CODE STRUCTURE

General Information

For the 1997 Economic Census, the Census Bureau rec-ognized regions, divisions, states, counties, consolidated

1While responsibility for the 1997 Census of Agriculture wastransferred to the U.S. Department of Agriculture’s NationalAgricultural Statistics Service (NASS) in 1996, the planning stagewas underway well before that. Thus, the GIT played a major rolein preparing for the census of agriculture, as well as for theeconomic census. Note 3 in chapter 2, The Players and the Plans,describes the transfer of the census of agriculture in more detail.



cities, places, and metropolitan areas. The 1997 EconomicCensus included the United States, Puerto Rico, Guam, theNorthern Mariana Islands, and the Virgin Islands of theUnited States.

The boundaries honored for geographic entities for the1997 Economic Census were those in effect on January 1,1997, except for MAs, for which June 30, 1997 bound-aries, and the 105th Congressional Districts, which tookeffect in January 1997, were used. Changes occurring afterthese dates were not recognized for the 1997 EconomicCensus. Boundaries for states, counties, and selectedincorporated places were reported to the Census Bureau inresponse to the annual Boundary and Annexation Survey.Where the code structure includes a population size crite-rion, the population was based on the 1990 DecennialCensus, tabulation counts, corrected counts from CountQuestion Review, updated counts based on special censuscounts, and counts as updated by the Population Divisionto reflect major annexations.

Regions

The Census Bureau recognized four regions within the50 states and the District of Columbia: Northeast, Mid-west, South and West. Each region was subdivided intodivisions and was identified by a set of one-digit codes.Puerto Rico and the outlying areas were not part of anyregion, they were assigned a region code of 9.

Divisions

The Census Bureau recognized nine divisions withinthe 50 states and the District of Columbia, and within acensus geographic region: East North Central, East SouthCentral, Middle Atlantic, Mountain, New England, Pacific,South Atlantic, West North Central, and West South Cen-tral. Divisions were identified by a set of one-digit codesthat were shown as the first digit of the component cen-sus state codes. Puerto Rico and the outlying areas werenot part of any division, they were assigned a divisioncode of 0.

States

The Census Bureau recognized the following areas asstates or as statistically equivalent entities of states—

• The 50 states of the United States.

• District of Columbia.

• Puerto Rico.

• Guam.

• Northern Mariana Islands.

• Virgin Islands of the United States.

• Nationwide.

• Foreign.

• Offshore Areas.

The state codes used for the economic census were thetwo-digit Federal Information Processing Standard (FIPS)2

state codes. The FIPS state codes were assigned in alpha-betic sequence for the 50 states and the District of Colum-bia. The FIPS state codes for Puerto Rico and the outlyingareas were assigned in a separate alphabetic sequence fol-lowing the states.

For the special situations of United States, nationwide,foreign, and offshore areas, state equivalent codes wereused as follows—

United States. A United States record was in the stubfile, with a state equivalent code of 99 in both the FIPSCode field and the census code field. This code was pro-vided in the stub file for publication purposes only. It wasnever assigned directly by the geographic coding system.

Nationwide. If the establishment could not be assignedan acceptable state code and was not categorized as for-eign, the geocoding system assigned a nationwide stateequivalent code of 97. Code 97 was used in both the FIPSstate and the census state code fields. The county andplace fields on this record were filled with zeros.

Foreign. Foreign establishments were assigned a foreignstate equivalent code of 98 in both the FIPS state codefield and the census state code field. The county and placefields on this record were filled with zeros. To be assigneda foreign state code, an establishment record had to meetone of the following pairs of conditions—

State abbreviation with ZIP Code

ZZ 00000YY 00000AE 09000-09899AA 34000-34099AP 96200-96699

Note: Ships at sea, previously recognized as a separate categoryassigned with 00 as the FIPS state and census state codes, were nowassigned a foreign state code of 98.

Offshore Areas. The mineral industries geographic codestructure provided for either a two-digit FIPS state code ora Census Bureau pseudo-state code, and a three-digitpseudo-county code. The codes provided for offshoreareas were—

2The FIPS Code is the standard geographic coding schemeused by all Federal agencies. The FIPS codes are unique identifica-tion numbers for geographic entities. They are assigned by theNational Institute of Standards and Technology.



Offshore areaState

abbre-viation

FIPSstatecode

Censusstatecode

FIPScounty

code

Atlantic . . . . . . . . . . AT 81 29 886Northern Gulfof Mexico . . . . . . . NG 82 69 886

Pacific . . . . . . . . . . . PC 83 89 886

These codes were not assigned directly by the geo-graphic coding system. Addresses identified as ‘‘offshore’’were assigned a state code for one of the 50 states, Dis-trict of Columbia, Puerto Rico, outlying areas, or nation-wide.

MCD assigned the offshore area codes, where appli-cable. This enabled the tabulation of oil and gas establish-ment statistics to a state, an offshore area within a state,and offshore areas not associated with a particular state,rather than the usual state, county, or place level. Itreflected a classification of oil and gas operations basedon where their ‘‘bottoming out’’ actually occurred. Opera-tions that bottomed out on land or inland waters wereclassified as ‘‘onshore,’’ if within a coastal state; if withina noncoastal state, as ‘‘statewide.’’ Operations thatbottomed out seaward of the coastal line were classifiedas ‘‘offshore.’’

Counties

The Census Bureau recognized the following areas ascounties or as statistically equivalent entities of countiesfor the economic census—

The 50 states. (except Alaska and Louisiana) Countieswere recognized.

• Alaska: Boroughs and census areas were recognized.

• Louisiana: Parishes were recognized.

District of Columbia. The District of Columbia wasrecognized as both a statistical equivalent of a county anda state. The county and state are coextensive.

Montana. The part of Yellowstone National Park locatedin Montana was not included in any county and was recog-nized as a statistically equivalent entity (in addition to thecounties).

Maryland, Missouri, Nevada, and Virginia. Indepen-dent cities (in addition to the counties). Independentcities were independent of any county and thus constituteprimary divisions of their states. Independent cities aredifferent from coextensive cities and counties because nocounty existed for an independent city. In order to givecomplete coverage to the state at the county level, countyequivalent records represented the independent cities.Data for the independent city were identical to the data forits county equivalent.

Hawaii. Kalawao County, which had no functioning localgovernment, was included with Maui County for statisticalpurposes.

Puerto Rico. Municipios were recognized.

Guam. No county or statistical equivalent entities wererecognized.

Northern Mariana Islands. No county or statisticallyequivalent entities were recognized.

Virgin Islands. No county or statistically equivalententities were recognized.

Undistributed. Code 000 was used if the county codewas not known or not applicable, as in nationwide, for-eign, offshore and undistributed addresses. Code 000appears in the stub file with the name ‘‘undistributed.’’

Counties were identified by a three-digit FIPS countycode. The FIPS county codes were assigned in alphabeticsequence with the names of all counties and countyequivalents (excluding independent cities) within eachstate, usually beginning with code 001. County codes forindependent cities were assigned as a separate alphabeticsequence beginning with code 510.

Places

The economic census recognized the following areas asplaces or as statistically equivalent entities of places—

Incorporated place. Incorporated places with a popula-tion of 2,500 or more were recognized for economic cen-sus data tabulations. New places reported to the CensusBureau and legally in effect on January 1, 1997 wererecognized if they had a population of 2,500 or more.

Three incorporated places with a population of lessthan 2,500 were recognized as places for previouseconomic census tabulations because of their economicimportance. The Census Bureau continued to recognizethese places for the 1997 Economic Census. They were:Industry, California; Vernon, California; and Bedford Park,Illinois.

Coextensive cities and counties occur where two gov-ernmental units, the city and the county or county equiva-lent, both have the same boundaries. In these cases, boththe city and the county were recognized. Data for the citywere identical to the data for the county.

Census designated place (CDP). CDPs were not recog-nized as places for the economic census, except in Hawaii.For Hawaii, which does not have incorporated municipali-ties recognized for decennial census purposes, placeshave been delineated for statistical purposes by the stategovernment in cooperation with the Census Bureau. Inagreement with the state government, the Census Bureaudoes not recognize the City of Honolulu, coextensive withthe county of Honolulu, as a place for decennial census



purposes. Those CDPs in Hawaii with a population of2,500 or more (from the 1990 Decennial Census) wererecognized as place equivalents for the economic census.

Maui County, Hawaii consisted of four islands. At therequest of the state of Hawaii, the Census Bureau recog-nized the ‘‘Island of Lanai’’ and the ‘‘Island of Molokai’’ asplaces for previous economic census tabulations, and con-tinued to recognize them as places for the 1997 EconomicCensus. Kalawao County, a nonfunctioning governmentalunit, was included in the ‘‘Island of Molokai’’ place.

Consolidated city. Some consolidated governmentsinclude separately incorporated municipalities. TheCensus Bureau referred to such entities as consolidatedcities. Each consolidated city had a one-character alpha-betic census code. See the discussion of ‘‘balance ofconsolidated city’’ below for a list and description of theconsolidated cities.

Balance of consolidated city. Some of the consolidatedgovernments included separately incorporated municipali-ties. The Census Bureau referred to such entities as ‘‘con-solidated cities.’’ For the economic census, the CensusBureau recognized the portion (balance) of a consolidatedcity that excluded the incorporated place(s) that are recog-nized separately for the economic census. If a municipalitywithin a consolidated city had a population of 2,500 ormore, the Census Bureau treated it as a separate place forthe economic census and omitted it from the balance ofthe consolidated city. If it had fewer than 2,500 inhabit-ants, it was included as part of the balance of the consoli-dated city. The balance of a consolidated city may refer toa whole consolidated city if none of the included placescontain a population of 2,500 or more. The consolidatedcities and their components were—

• Milford, Connecticut consolidated city is coextensivewith Milford town. It included Woodmont, which wasnot populous enough for separate tabulation. Therefore,‘‘Milford’’ referred to the whole consolidated city.

• Jacksonville, Florida consolidated city was coexten-sive with Duval County. It included Atlantic Beach, Jack-sonville Beach, and Neptune Beach, which were tabu-lated separately. ‘‘Jacksonville (balance),’’ which was aplace equivalent, includes Baldwin, which was not popu-lous enough for separate tabulation.

• Athens-Clarke County, Georgia consolidated city wascoextensive with Clarke County. It included Wintervilleand Bogart (part in Clarke County), which were notpopulous enough for separate tabulation. Therefore,‘‘Athens-Clarke County’’ referred to the whole consoli-dated city.

• Augusta-Richmond County, Georgia (new for 1997Economic Census) consolidated city was coextensivewith Richmond County. It included Hephzibah, whichwas tabulated separately. ‘‘Augusta-Richmond County

(balance),’’ which was a place equivalent, includedBlythe (part in Richmond County), which was not popu-lous enough for separate tabulation.

• Columbus, Georgia consolidated city was coextensivewith Muscogee County. It included Bibb City, which wasnot populous enough for separate tabulation. Therefore,‘‘Columbus’’ referred to the whole consolidated city.

• Indianapolis, Indiana consolidated city included all ofMarion County except Beech Grove, Lawrence, South-port, and Speedway, which were tabulated separately.The consolidated city included Cumberland (part inMarion County), which also was tabulated separately.‘‘Indianapolis (balance),’’ which was a place equivalent,included Clermont, Crows Nest, Homecroft, MeridianHills, North Crows Nest, Rocky Ripple, Spring Hill,Warren Park, Williams Creek, and Wynnedale, whichwere not populous enough for separate tabulation.

• Butte-Silver Bow, Montana consolidated city wascoextensive with Silver Bow County. It included Walker-ville, which was not populous enough for separatetabulation. Therefore, ‘‘Butte-Silver Bow’’ referred to thewhole consolidated city.

• Nashville-Davidson, Tennessee consolidated citywas coextensive with Davidson County. It included BelleMeade, Forest Hills, Goodlettsville (part in DavidsonCounty), and Oak Hill, which were tabulated separately.‘‘Nashville-Davidson (balance),’’ which was a placeequivalent, includes Berry Hill, Lakewood, and Ridgetop(part in Davidson County), which were not populousenough for separate tabulation.

Special Economic Urban Areas (SEUA). SEUAs weretownships in Michigan, Minnesota, New Jersey, and Penn-sylvania; and towns in Connecticut, Maine, Massachusetts,New Hampshire, New York, Rhode Island, Vermont, andWisconsin with a population of 10,000 or more. If a townor township that qualified as a SEUA contained an incorpo-rated place that was recognized separately for economiccensus tabulations, the Census Bureau recognized the bal-ance of the SEUA as a place equivalent, if the balance hada population of 10,000 or more.

Balance of metropolitan area. In the six New Englandstates (Connecticut, Maine, Massachusetts, New Hamp-shire, Rhode Island, and Vermont) ‘‘Balance of metropoli-tan area (MA)’’ was used for cities and towns located in anMA but not qualifying for economic census tabulations.This was needed to maintain the ability to tabulate datafor MA’s in the New England states because MAS aredefined in terms of cities and towns rather than countiesin these states.

Balance of county. ‘‘Balance of county’’ included thoseareas not recognized as places and place equivalentsunder the above definitions.



The 1997 Economic Census used the four-digit decen-nial census place codes for internal processing. Theseplace codes were assigned in alphabetic sequence withinstate or state equivalent. Additional place equivalentcodes were assigned in the sequence as needed to accom-modate economic census needs. A place code of 9999designated ‘‘Balance of county.’’ The first digit of the‘‘Balance of MA’’ census place code is 9, and the last threedigits were the same as the first three digits of the metro-politan statistical area (MSA) or primary metropolitanstatistical area (PMSA) code.

Metropolitan Areas

The Census Bureau recognized the three types of met-ropolitan areas (MAs) established by the Statistical PolicyOffice, Office of Information and Regulatory Affairs, Officeof Management and Budget. The three types of MAs were:Metropolitan Statistical Areas (MSAs), Consolidated Metro-politan Statistical Areas (CMSAs), and Primary MetropolitanStatistical Areas (PMSAs). MSAs and CMSAs were equiva-lent levels of geography; PMSAs were subdivisions ofCMSAs. GEO recognized the June 30, 1997 boundaries forthe MAs. The portion of Sullivan city in Crawford County,Missouri was legally part of the St. Louis, MO-IL MSA. How-ever, it was not included in the 1997 Economic Censuscode structure because it did not adhere to the require-ment that whole counties be used to define the MAs(except in New England).

Each MA had a four-digit FIPS Code that was assigned inalphabetic sequence. The code was 9999 for nonmetro-politan areas. In addition, each CMSA had a two-digit FIPSCode also assigned in alphabetic sequence.

Congressional Districts

The Census Bureau originally planned to publish eco-nomic census data for the congressional districts of the105th Congress of the United States. Budget reductionsresulted in the cancellation of this program for the 1997Economic Census.

Census Tract/BNA and Census Blocks

The geocoding system assigned census tract/BNA andblock numbers to as many economic establishments aspossible during the geocoding process.3 These censustract/BNA and block numbers were used as a linkage tothe higher level geographic codes required for economiccensus data publications. The census tract/BNA and blockcodes, however, were not posted to the economic censuscontrol file, nor were data tabulated for these areas.

GEOGRAPHIC RECODE AND REFERENCE FILES

General Information

In addition to the freestanding geocoding system, GEOcreated other products to support economic census pro-cessing and data tabulation. A two-thirds reduction in theGEO budget allocations for the 1997 Economic Censusfive-year budget cycle resulted in significant reductions inthe level of services, as well as elimination of selectedproducts provided by GEO.

Products eliminated were: initial geocoding processingand monitoring by GEO; the Metropolitan Area Appendixesfor the geographic publications; publication-ready tablesfor, and the printing of, the Geographic Reference Manual;the County Place Code Change File; the Duplicate NamesFile; Establishment Cluster Files; and clerical geocoding.The level of services were greatly reduced for the follow-ing products: manual reviews and updates to enhance theCity Reference File (CRF) beyond the capabilities of theautomated file creation process, and the Publication StubFile. The CRF will be treated in detail below.

Census to Economic Recode File

The purpose of the census to economic recode file wasto provide a source of geographic codes for use in con-verting the current (1997) decennial census based geo-codes to the economic census tabulation geocodes. Thisfile was created uniquely for use in the economic censusgeocoding process.

1992 Economic to Current Census Recode File

The purpose of the 1992 economic to current censusrecode file was to provide a source of geographic codesfor use in converting the historic (1992 Economic Census)tabulation geocodes on the establishment records to theirequivalent current (1997) decennial census based geo-codes. The historic geocodes were used as one of the tiebreakers in the geocoding process. This recode file wasused only during the initial geocoding of the control file.The historic geocodes on the establishment records in thecontrol file were defined as the 1992 Economic Censustabulation geocodes. The 1992 economic to current cen-sus recode file was one of two historic recede files createduniquely for use in the economic census geocoding pro-cess.

1997 Economic to Current Census Recode File

The purpose of the 1997 economic to current censusrecode file was to provide a source of geographic codesfor use in converting the historic (1997) economic tabula-tion geocodes on the establishment records to theirequivalent current (1997) census based geocodes. The his-toric geocodes were used as one of the tie breakers in thegeocoding process. This recode file was used during thegeocoding of the physical location responses from the

3Census tracts and BNAs were subdivisions of a county orcounty equivalent. Census blocks were small areas bounded onall sides by visible (e.g., streets, streams, or railroad tracks) orinvisible features (e.g., place or congressional district boundaries,property lines, or short, imaginary extensions of streets or roads).



economic census questionnaires, and all subsequenteconomic census geocoding. The historic geocodes on theestablishment records for physical location and subse-quent geocoding were defined as the previously assigned1997 economic tabulation geocodes. The 1997 economicto current census recede file was one of two historicrecode files created uniquely for use in the economiccensus geocoding process.

Clustering Program Module

The purpose of the establishment cluster review pro-cess was to identify multiple establishments reporting tothe same address. This situation often occurs when theaddress reported for an establishment was the mailaddress, not the physical location address. Once identi-fied, the establishment clusters were eligible for reviewand address correction where applicable.

An establishment cluster was defined by SSSD as agroup of establishment records with a common addressand industry code. Address types requested by SSSD forconsideration in the clustering program module were:structure number and street name, building name, andpost office boxes. The criteria for industry codes consid-ered in the clustering program module was limited tothose requested by SSSD.

The GEO clustering program module selected a subsetof control file records that later became part of the inputfor the traditional clustering process. The GEO clusteringprogram module was incorporated into the initial controlfile geocoding only. Control file records that met thecriteria defined by address type and industry code wereidentified and written to a separate file, along with theirstandardized address, for further establishment clusterprocessing by ESMPD. The traditional automated cluster-ing was programmed and executed by ESMPD for the1997 Economic Census.

County Code Change File

The purpose of the county code change file was toconvert the 1992 county codes to the 1997 county codestructure. The GEO provided a one-to-one recode in thisfile. Only counties with code changes were included in thecounty code change file. For whole county changes, thefile contained one county record with its historic state andcounty codes (1992) and its current (1997) state andcounty codes. When one 1992 county split into two coun-ties for 1997, the file contained a record for each 1992qualifying economic census place. Prior to creating ques-tionnaire mailing labels, ESMPD used the county codechange file to recode the 1992 county codes assigned toestablishment records.

County Alpha Recode File

The purpose of the county alpha recode file was to con-vert the FIPS county code to a six-character abbreviationof the county name, which was printed on the question-naire mailing label. The ESMPD performed the county

alpha recode operation after the county code changes hadbeen made. The county alpha recode file contained onerecord for each 1997 county or county equivalent.

Geographic Stub File

The purpose of the geographic stub file was to providethe staffs of the subject-matter divisions with the ability toinsert geographic entity names in the stubs of the publica-tion tables and other data products. The ‘‘stub’’ is the listof geographic entities in the left column or in the box-head.

This file contained all the information necessary to con-vert the geographic codes associated with establishmentrecords to the publication geographic entity names forpublications and other data products. The stub file hadone record for each of the following—

• Metropolitan area (Metropolitan StatisticalAreas/Consolidated Metropolitan StatisticalAreas/Primary Metropolitan Statistical Areas).

• Region.

• Division.

• State or state equivalent (included the 50 states, theDistrict of Columbia, Puerto Rico, Guam, the NorthernMariana Islands, the Virgin Islands of the United States,ships at sea, three offshore areas for mineral industries,foreign establishments, and nationwide establishments),plus one record representing the United States as awhole, in the same format as a state record, with a statecode of ‘‘99.’’

• County or county equivalent, plus a record in eachstate or state equivalent for ‘‘undistributed,’’ used forindustries with operations not designated to specificcounties, and a county record for each offshore area.

• Consolidated city.

• Place and place part within county (qualifying economiccensus places only).

For consolidated cities that contained separately incor-porated places, this file had a record for each consolidatedcity, each separately incorporated place that qualified foreconomic census recognition, and each consolidated citybalance. Each place record included the place description,which distinguished between various types of places andthe ways they related to metropolitan and urban areas.

Subject Review City Reference File (CRF)

The purpose of the subject review CRF was used by theeconomic census subject matter divisions to facilitate theirreview of the geographic coding and tabulation results.The subject review city reference file was an extract of thecity reference file used for general geocoding. It contained



all post office name, state abbreviation, and ZIP Code com-binations from the city reference file, collapsed to the eco-nomic census tabulation geocode level.

Geographic Notes File

The purpose of the geographic notes file was to iden-tify and document geographic changes that occurredbetween January 1, 1992 and January 1, 1997, as well asany changes that occurred prior to January 1, 1992, butwere reported too late for inclusion in the 1992 EconomicCensus. The geographic notes file was used to prepareappendixes for the relevant census report publicationproducts.

The types of changes documented in the geographicnotes file were—

• Economic census-recognized places that were in morethan one county and the names of the counties in whicheach multicounty place was located.

• Special economic urban area balances and economiccensus recognized incorporated places excluded fromthem.

• Differences between 1992 and 1997 Economic Censusrecognized geography and the reasons for the changes(i.e., new incorporations, dis-incorporations, annex-ations, detachments, name changes, consolidations,mergers, and county changes).

• Anomalous situations that involved economic censusgeography.

Geographic Reference Manual Table Files

The purpose of the geographic reference manual tablefiles was to provide an easily accessible reference sourcefor the components, names, and geocodes of geographicareas. The tables encompassed all of the geographic areasthat qualified for economic census tabulations.

The geographic reference manual tables were madeavailable as a series of 12 files, one for each table, in thefollowing sequence—

• States and Statistically Equivalent Areas.

• Metropolitan Statistical Areas, Consolidated Metropoli-tan Statistical Areas, and Primary Metropolitan StatisticalAreas.

• Metropolitan Statistical Areas, Consolidated Metropoli-tan Statistical Areas, and Primary Metropolitan StatisticalAreas by State.

• Counties.

• Consolidated Cities.

• Places.

• Places by County.

• Special Economic Urban Areas.

• Places Located in More than One County.

• Census of Mineral Industries: Offshore Areas.

• Places in 1992 Economic Census, Not in 1997 EconomicCensus.

• Places in 1997 Economic Census, Not in 1992 EconomicCensus.

ZIP Code Legality File

The purpose of the ZIP Code legality file was to providea source of all legal ZIP Codes within each state or stateequivalent. GEO provided ESMPD with a file extracted fromthe U.S. Postal Service’s City State File. The ZIP Code tostate relationships were extracted from the City State Fileand supplemented with additional relationships from theCRF.

GEOCODING PROCESS

Freestanding Geocoding System

For the economic census the GEO developed a free-standing geocoding production system that encompassedthe geocoding software and all reference files necessaryto meet the geocoding processing requirements of theeconomic census. After extensive testing, the economiccensus geocoding production system was installed on theESMPD computer platform in November 1997. The installa-tion included a comprehensive production test thatensured all components were performing as required onthe ESMPD platform.

The GEO general purpose geocoding system was usedas a basis for the economic census geocoding system.Modifications were made to meet the specific require-ments of the economic census. The Census to EconomicRecede File, the 1992 Economic to Current Census RecodeFile, and the 1997 Economic to Current Census RecodeFile were incorporated into the general purpose geocodingsystem to provide conversions between decennial censusbased geocodes and economic census tabulation geo-codes.

For the economic census, place level geocodes wererequired for all establishment addresses, including thosewithout a structure number and street name type address.To meet this requirement, both detail geocoding andheader geocoding were needed. The process referred to asdetail geocoding matches the structure number and streetname part of an address to the Address Reference File toobtain geographic codes. The process referred to asheader geocoding matches the last line components of anaddress (post office name, state abbreviation, and ZIPCode) to the CRF to obtain geographic codes. For the eco-nomic census, both detail and header coding were linkedto ensure all establishments addresses received the mostaccurate codes based on the address information pre-sented to the geocoding system. The Census to Economic



Recode File converted the decennial census based geo-codes from the general purpose geocoding programs tothe economic census tabulation geocodes.

City Reference File (CRF)

The CRF contained records that link the post officename, state abbreviation, and five-digit ZIP Code (referredto as the last line of an address) to its associated decen-nial census state, county, MCD, and place codes. Each CRFrecord contained a post office name, state abbreviation,and five-digit ZIP Code. In addition to the official postoffice names and alternate names, the CRF also containedvariant spellings for post office names and postal neigh-borhood names, together with supplementary countyname information for matching county names when acounty name response is given. The GEO general purposegeocoding system matched to a CRF with decennial cen-sus based geography. Modifications made uniquely for theeconomic census geocoding system subsequently con-verted the decennial census based geocodes to economiccensus tabulation geocodes.

Address Reference File

The Address Reference File (ARF) was an extract of theTopologically Integrated Geographic Encoding and Refer-encing (TIGER) Files, reformatted and reorganized to aconvenient matching structure. The ARF contained streetaddress information from TIGER linked to its associatedstate, county, tract, and block codes. The components ofthe street address were house number, street name, andZIP Code. The geocodes and linkages associated with eachARF record could be associated to all of the higher levelgeographic codes. The GEO general purpose geocodingsystem matched to an ARF with decennial census basedgeography. Modifications made uniquely for the economiccensus geocoding system subsequently converted thedecennial census based geocodes to economic censustabulation geocodes.

Legality Files

The legality files were maintained by GEO for generalgeocoding applications. The purpose of the legality fileswas to provide a source of geographic codes, names, and

other attributes for editing legal geographic codes in otherfiles, for assigning attributes based on geographic codes,and as a source of information for the geographic stubsfor publications. The Census to Economic Recode File, the1992 Economic to Current Census Recode File, and the1997 Economic to Current Census Recode File were addedto this group of files as part of the modifications made tothe GEO general purpose geocoding system uniquely forthe economic census.

VALIDATING AND UPDATING INCOMING DATA

Responses to the 1997 Economic Census were receivedby mail or telephone at the National Processing Center inJeffersonville, Indiana, where respondent records under-went data entry processing.

In turn, ESMPD staff, each night, processed addressinformation received from respondents in order to updatethe geocodes on the Standard Statistical Establishment List(SSEL). The SSEL functioned as the Census Bureau’s regis-ter of all business addresses. This updating continuedthroughout the 1997 Economic Census processingoperations.

The geocoding system referred for EPCD review anyestablishment that it was unable to fully code, or that hadbeen assigned a low confidence code (e.g., only a statecode assigned with no corresponding county or placecode). EPCD staff, often in collaboration with GEO staff,checked problem records against maps and other sourcesof geographic information and made corrections to therecords.

Subject-matter staff also participated in identifyingproblems. During their review of geographic tabulations,analysts found some individual establishments, or groupsof establishments, that appeared to have questionablecodes. GEO staff researched the geocodes to determinecorrect tabulation geography, and returned the referrals tosubject matter staff for retabulation.

In sum, the goals established by the (Geographic Inter-face Team (GIT) in 1994 had excellent results for the 1997Economic Census. The freestanding geocoding productionsystem was successful in meeting the requirements of thiscensus, and set the pattern for future economic censuses.



Chapter 7. PREPARATIONS FOR CENSUS MAIL OUT

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

Upgrading the Processing Infrastructure . . . . . . . . . . . . . . . . . . . . 59

Compiling the Census Mail List. . . . . . . . . . . . . . . . . . . . . . . . . 60The Standard Statistical Establishment List (SSEL) . . . . . . . . . . . . . . 60Background information . . . . . . . . . . . . . . . . . . . . . . . . . . 60SSEL content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60Sources for the SSEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60Using the SSEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

Split Mail and Electronic Reporter Identification . . . . . . . . . . . . . . . 61Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61Background information . . . . . . . . . . . . . . . . . . . . . . . . . . 61Construction industries . . . . . . . . . . . . . . . . . . . . . . . . . . 61Manufacturing industries. . . . . . . . . . . . . . . . . . . . . . . . . . 62Retail trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62Wholesale trade. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62Service industries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62Financial, insurance, and real estate industries . . . . . . . . . . . . . . 62Utilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

Geographic Coding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

Administrative Records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63Background Information . . . . . . . . . . . . . . . . . . . . . . . . . . . 63Obtaining the Records . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63IRS records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63SSA records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64BLS records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

Using Administrative Records in the Economic Census . . . . . . . . . . . 64

Assigning Census File Numbers . . . . . . . . . . . . . . . . . . . . . . . . 64

Printing and Assembling the Mailout Packages . . . . . . . . . . . . . . . . 64Cover Letters, Instruction Sheets, and Other Mailing Materials . . . . . . . 64General information . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64Cover letters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Inventory sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Printing the questionnaires and other mailing materials . . . . . . . . . 65

Using DocuPrint for the 1997 Economic Census . . . . . . . . . . . . . . 65Background information . . . . . . . . . . . . . . . . . . . . . . . . . . 65DocuPrint production. . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

Address Labels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66Printing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

CONTENTS


Printing and Assembling the Mailout Packages—Con.Quality Assurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67Contract printing/package assembly . . . . . . . . . . . . . . . . . . . 67DocuPrint printing and package assembly . . . . . . . . . . . . . . . . 67Label printing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

Pre-Census Company Contacts . . . . . . . . . . . . . . . . . . . . . . . . . 68Information Mailings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68The Account Manager Program. . . . . . . . . . . . . . . . . . . . . . . . 69


Preparations for Census Mailout

INTRODUCTION

Planning, development of questionnaire content, and—new for the 1997 Economic Census—the development andimplementation of the North American Industry Classifica-tion System (NAICS) were long-term activities that merged,eventually, into the more particular preparations for theactual conduct of the enumeration. Census preparationsincluded (1) upgrading the Economic Directorate’s elec-tronic processing infrastructure, (2) compiling the censusmail list, (3) acquiring and using administrative records,(4) assigning Census File Numbers (CFNs) to provideidentification for each company, (5) printing and assem-bling mailing packages of questionnaires, instructions,and cover letters for over four million companies, and(6) establishing and maintaining pre-census contactsbetween U.S. Census Bureau staff and the largest compa-nies covered by the 1997 Economic Census.

UPGRADING THE PROCESSING INFRASTRUCTURE

During the 2 years prior to the 1997 Economic Census,the Economic Directorate dramatically improved itsdata-processing infrastructure’s capabilities. This wasaccomplished by upgrading existing computer systems,replacing obsolete hardware, and installing new softwareapplications. The upgrade reduced the Directorate’ssupport costs, increased processing efficiency, and createda user friendly network.

The upgrade—

• Replaced the Local Area Network (LAN) servers withmore powerful machines.

• Installed faster Virtual Memory System (VMS).

• Replaced Virtual Access Extension (VAX) 7600 processorequipment with Alpha processors.

• Installed Common Use Interface (CUI) applications.

• Expanded use of the operating system and its associ-ated hardware to replace the Unysis computer

The LAN upgrade included both the staff work stationsand computer servers. At Census Bureau headquarters inSuitland, MD, the Economic Directorate acquired nearly600 new personal computers with Pentium chips, whilethe National Processing Center (NPC)1 office in Jefferson-ville, IN received 110. Adding Pentium technology

increased the speed of the individual workstations by 60percent (from 166 megahertz [MHZ] to 266 MHZ each) andallowed more rapid communication with the mainframecomputers.

Eight new servers increased the total to 46. All localnetworks had identical applications installed to maintainconsistency among work stations. New software wasadded to eliminate the need for paper copy of documentsand files in favor of electronic transmission to individualwork stations and to the public. This promoted informa-tion sharing, reduced mailing demands, and diminishedstorage and filing needs. It also resulted in controlledaccess to information and it provided a search function ofall electronically stored material.

At the Census Bureau’s headquarters, the upgradeeliminated one VMS station, but the new technologyincreased input/output speed and total storage capacity tothe remaining systems. The development cluster2 upgradeincreased processing speed over 700 percent, from 77million lines of information per minute to 551. Memoryincreased 600 percent, from 640 megabytes to fourgigabytes, and disc space nearly tripled (50 to 141gigabytes). The production cluster counted three fewersystems, yet increased processing speed from 154 to4,085 million lines of information per minute. In turn,memory was increased from 2.6 to 12 gigabytes, and discspace grew from 325 to 865 gigabytes. The NPC used 10fewer VMS stations, but increased processing speed bytwo million lines per second and expanded disc space by14 gigabytes. Memory was reduced from 672 to 352megabytes.

The infrastructure upgrade also addressed the real andpredicted difficulties of the aging Virtual Address Exten-sion 7600 clusters (support for this equipment wasbecoming difficult, since the system was gradually beingretired for more open architecture systems). The CensusBureau replaced all but one of the 7600 processors withAlpha 4100 and 8400 processors by the spring of 1998.The Alpha processors, which incorporated 64-byte proces-sor technology, outperformed the capabilities of a largernumber of the Virtual Address Extension 7600 systemsthat they replaced.

The final aspect of the technology upgrade was thereplacement of older and slower UNIX systems with new

1Until a reorganization effective May 1, 1998, this unit wascalled Data Preparation Division (DPD).

2A cluster is a group of disc sectors, or units, into which theoperating system divides the disc. This allows the operatingsystem and disc drive to identify where information is stored.



equipment. Advances in UNIX technology required fewermachines to handle the required workload. The newsystem featured an open architecture that enabled mostcomputer languages to be converted into a language thatthe new—system and subsequently all systems on thelocal network—could use.

As a result of the UNIX upgrade, the Economic Director-ate discontinued using the Unysis computer. Unysis wasincapable of conversion to an open architecture system:the networks could ‘‘understand’’ Unysis communications,but Unysis could not understand the networks. The deacti-vation of Unysis required the transfer of its originalCOBOL/FORTRAN processing programs to UNIX. There wasinsufficient time for the complete conversion of Unysisprograms to UNIX for use in the 1997 census. Therefore,the Census Bureau concentrated on the most vital process-ing programs, and delayed completing the changeover forall programs until the 2002 Economic Census.

COMPILING THE CENSUS MAIL LIST

The Standard Statistical Establishment List (SSEL)

Background information. The SSEL is a centralized,multipurpose, computerized data file of U.S. businessestablishments and companies. Prior to 1974, Federalagencies collecting and tabulating economic data eachcompiled and maintained their own business registers.Since each agency designed programs for its own pur-poses, there were serious comparability problems, signifi-cant duplication of effort, and increased costs for thoseusing economic data across agencies. In 1968, the Officeof Management and Budget (OMB) addressed the problemby designating the Census Bureau to develop, establish,and operate a general, multipurpose business directory.While the new directory would be used as the samplingframe for the entire Federal statistical establishment, theCensus Bureau was never permitted to share the listingitself with any other Federal agency. Funding for theproject began with Fiscal Year (FY) 1972 and the newdirectory—the SSEL—became operational for data year1974.

The SSEL includes the Nation’s employer and nonem-ployer business establishments (except private house-holds and governments) and organizational units ofmultiestablishment businesses. ‘‘Businesses’’ are legal oradministrative entities assigned an employer identificationnumber (EIN) by the Internal Revenue Service (IRS), and‘‘units’’ include divisions, subsidiaries, companies, andother affiliated organizations. For the 1997 EconomicCensus, the SSEL included over 19 million businesses—180,000 multiestablishment companies (with approxi-mately 1.5 million affiliated establishments), 5 millionsingle-establishment companies, and nearly 14 millionnonemployer businesses.

SSEL content. The SSEL was designed to (1) identify themost useful definition of business establishment units for

statistical analysis, (2) build bridges between these statis-tical units and legal entities for which administrativerecords were available, and (3) facilitate more efficient useof administrative records for statistical purposes. The SSELsystems design was based on a variable word-lengthrecord that allowed additional information to be addedwhenever necessary. The establishment was the basicbuilding block of the list, and, for economic censuspurposes, was the smallest unit for which key data oneconomic activity (e.g., employment and payroll) wereavailable.

For each organizational unit, the SSEL contained—

• Employer identification number (EIN).

• SSEL identification number.

• Company name, mailing address, and physical location.

• Industry classification.

• Geographic code.


• Operating data (i.e., receipts, payroll, and othereconomic variables).

By matching the EINs assigned to each establishmentwith employees to the SSEL identification numbers, theCensus Bureau could link and identify affiliations for par-ent companies, subsidiary firms, and their establishmentsthrough all phases of economic activity.

Sources for the SSEL. The SSEL database was main-tained as two cross-classified lists, one each for single-and multiestablishment companies. The list informationfor single-establishment companies was updated continu-ously, using both payroll tax records and receipts datafrom tax records provided by the IRS. Information on mul-tiestablishment companies was updated annually basedon information from the Census Bureau’s current eco-nomic surveys, such as the Current Industrial Reports-(CIRs), the Company Organization Survey (COS—seeChapter3, ‘‘Introduction of the North American IndustryClassification System (NAICS),’’ for more information onthis survey), and the like.

The operating data were initially derived from informa-tion collected in the COS for multiestablishmentcompanies, and from the IRS records for first-quarteremployment, quarterly payroll, as well as for annual pay-roll. The size codes covered payroll, sales, and receipts atthe establishment, legal-entity, and enterprise levels, andused a set of standard size ranges for each characteristicfor all industries. The SSEL system retains administrativedata for establishments for the current and immediatelyprevious years.

Using the SSEL. For the economic census, the SSEL func-tioned as a universe file to control questionnaire mailingand data processing. The Census Bureau used the SSEL



source files for a final pre-census update of the SSEL in theSummer of 1997. These source files included administra-tive data from the Social Security Administration (SSA), IRS,Bureau of Labor Statistics (BLS), as well as the results ofthe 1996 COS. Of particular importance in the updatewere the mailing addresses for individual establishmentsand the industry classifications.

During mail-list compilation, the Census Bureau usedthe SSEL as the principal source for mailing addresses forindividual establishments. For multiestablishment compa-nies, the Census Bureau used the SSEL to—

• Select all active companies (excluding farms) and allactive establishments.

• Identify consolidated companies/establishments.3

• Assign report form (questionnaire type) numbers for thefollowing:

• Out-of-scope or suspected out-of-scope establish-ments (Form 9901).

• Finance, insurance, and real estate, and communica-tions establishments (consolidated questionnaires).

• Auxiliary establishments (Form 9202).

• Other establishments based on program divisioninformation/NAICS/SIC.

For single-establishment companies, the Census Bureauused the SSEL to—

• Select establishments with 1997 payroll.

• Split the single-establishment company universe, usingNAICS and program division information, to assignappropriate questionnaire.

• Sample by trade area.

During data processing, NPC staff keyed informationfrom census questionnaires into a computer system thatinterfaced with the SSEL and individual records werematched to the SSEL to confirm coverage. During qualityassurance processing, the SSEL file enabled the CensusBureau to determine whether all the establishments of amultiestablishment company had been enumerated.

Split Mail and Electronic Reporter Identification

The multiestablishment mailout system also wasdesigned to make it easy to identify two other contingen-cies: ‘‘split mail companies’’ and those companies thatwanted to respond electronically. This was done using aparameter-driven identification system that enabled an

analyst to identify which parts of a company needed ques-tionnaires mailed separately to another location within thecompany (i.e., ‘‘split mail’’) and/or allowed companies thatwished to respond electronically also to receive the properelectronic reporting medium in the company package. Ini-tially, about 600 companies were mailed the electronicreporting media. Seventy one more companies requestedthe media during data collection. (Chapter 8, ’’Data Collec-tion and Processing,‘‘ discusses electronic reporting.)

Sampling

Background information. The 1997 Economic Censuscovered over 14 million establishments. The smallest ofthese—based on payroll cutoffs set for each four-digit SICor five-digit NAICS industry—were covered using adminis-trative records provided to the Census Bureau by the IRSand SSA. These establishments were not required to com-plete census questionnaires. The cutoffs established foreach industry were set so that the establishmentsexempted from the requirement to complete census formsaccounted for only about three percent of the total valueof shipments for that industry. Establishments that wouldotherwise have been accounted for using administrativerecords—but for which the Census Bureau lacked thenecessary information to assign a six-digit NAICS industrycode—were sent classification forms in the census mail-out.

These latter totaled approximately 1.6 million. Thepreparation of the mail list, and the arrangements madefor collecting detailed information on the long question-naires, varied by trade area. All multiestablishmentcompanies in a given sector or trade area were included inthe census and were enumerated using the long question-naires. The largest single-establishment companies—based on payroll cutoffs that varied by trade area—wereincluded and were sent long questionnaires. These largesingle-unit companies, together with the multiestablish-ment companies, were subjected to intense follow-up.

For smaller single-establishment companies in all tradeareas except transportation and auxiliaries,4 the CensusBureau used stratified samples to select establishments tobe included in the census or to determine whether to sendthem long or short questionnaires. Selection for thesamples was based on formulae designed to enable theCensus Bureau to compile reliable estimates for industriesat specified geographic levels, while keeping overallrespondent burden to a minimum.

Construction industries. The Census Bureau selected aprobability sample of 131,000 of the approximately625,000 establishments in the NAICS construction sector

3Consolidated companies/establishments were multiestablish-ment companies that, because of their relatively small size,and/or the kind of economic operations of their various individualestablishments, were sent a consolidated questionnaire request-ing summary data on all their establishments, rather than indi-vidual questionnaires for each establishment.

4Transportation industries were covered in two surveys—the1997 Truck Inventory and Use Survey and the Commodity FlowSurvey—while auxiliary establishments were part of multiestab-lishment companies. For more information on these surveys, seeChapter 10, ‘‘Transportation Surveys.’’



to participate in the 1997 census. The sample included all12,000 identified multiestablishment construction compa-nies, plus a stratified random sample of single-establishment companies. The single-establishmentsample was stratified by state and industry, then by pay-roll size. Within each state/industry level, the CensusBureau established five payroll intervals. The cutoff inter-val for each industry was set so that the sample drawnwould include approximately 50 percent of the total pay-roll attributable to single establishments within that indus-try. Single-establishment companies in the largest payrollsize category were identified as ‘‘certainty’’ cases. Thecomplete sample consisted of 12,000 multiestablishmentcompanies, 51,000 single-establishment ‘‘certainty’’ cases,and 68,000 other single-establishment companies. Thedata collected were used to generate national- and state-level estimates for each of the 27 NAICS constructionsector industries.

Manufacturing industries. The 1997 Economic Censusincluded approximately 380,000 manufacturing establish-ments. Nearly 83,000 single-establishment companies forwhich the Census Bureau lacked sufficient information toassign a six-digit NAICS industry code were sent theNC-9922, ‘‘Classification Form,’’ in the census mailout. Allother single- and multiestablishment manufacturing com-panies were split into three categories—Annual Survey ofManufacturers (ASM) establishments, large and medium(based on payroll; the cut-offs varied by industry) non-ASMestablishments, and small non-ASM establishments. Therewere approximately 61,000 ASM cases, about 21,000 ofwhich were single-establishment companies. The mediumand large non-ASM establishments—about 167,000 in all—were enumerated using the more than 200 manufacturingindustries census long questionnaires. Approximately29,000 small non-ASM establishments were sent one ofthe short census questionnaires.

Retail trade. The 1997 Economic Census initial mailoutplans called for inclusion of all multiestablishment retailfirms (approximately 530,000). An additional 320,000single-establishment companies for which the CensusBureau had insufficient classification information would besent the classification form. The remaining small single-establishment retail trade companies were sampled at arate of 13.8 percent, creating a sample mail file of121,309 firms. These were sent one of 37 retail trade longquestionnaires.

Wholesale trade. The initial census mailout included allmultiestablishment wholesale firms, and all wholesalesingle-establishment companies with employees approxi-mately 458,600. (Nonemployer wholesalers were consid-ered statistically insignificant.) Some 266,000 large- andmedium-size wholesalers—those with a specified numberof employees or level of payroll—were sent one of 42product-specific long questionnaires. The remaining

61,000 smaller wholesalers were enumerated using one ofsix short questionnaires asking for summary data on prod-ucts, materials, and key variables.

Service industries.5 The Census Bureau identifiedapproximately 330,000 multiestablishment firms in theservice industries for the initial census mailout. An addi-tional 662,263 companies, for which the Census Bureaulacked adequate information to classify at the six-digitlevel, were selected for inclusion in the classification ques-tionnaire mailout. Large and medium-size single-establishment service—firms some 651,145 companies,identified based on payroll and/or volume of sales cutoffsestablished for each five-digit industry—were identifiedfor the mailout and were sent the appropriate long formquestionnaire. The Census Bureau drew an approximate3.1 percent stratified sample of the remaining 1.55 millionsmall single-establishment firms within the service sector.This yielded a sample of 48,654 addresses, which weremailed one of the 41 services industries long question-naires.

Financial, insurance, and real estate industries.Mailout plans for the Financial, Insurance, and Real Estatesector called for mailing questionnaires or classificationforms to approximately 157,000 multiestablishment firms,and to just over 246,000 large- and medium-sized single-establishment companies. A 2.4-percent sample of smallsingle-establishments added 8,384 companies to themailing list to receive the long questionnaires.

Utilities. All utilities industries multiestablishment com-panies were sent the long questionnaires in the initial cen-sus mailout. An additional 32,000 single-establishmentcompanies were selected to receive classification ques-tionnaires, while 66,500 large- and medium-size single-establishment firms were sent long form questionnaires.The remaining small single-establishment utility compa-nies were sampled at a 5.9 percent rate, yielding 6,615additional companies that were sent questionnaires.

Geographic Coding

Conducting a mailout/mailback enumeration, and pro-cessing the responses, required place-level geographiccodes for all establishment addresses. The CensusBureau’s Geography Division developed a freestandinggeocoding production system that incorporated the geoc-oding software and all necessary reference files to meetthe geocoding processing requirements of the 1997 Eco-nomic Census. The production system was installed on

5Part or all of the following NAICS sectors were in the old SICservice industries—51, Information; 54, Professional, Scientific,and Technical Services; 56, Administrative Support; Waste Man-agement and Remediation Services; 61, Educational Services; 62,Health Care and Social Assistance; 71, Arts, Entertainment, andRecreation; 72, Accommodation and Foodservices; and 81, OtherServices, Except Public Administration



the Economic Statistical Methods and Programming Divi-sion’s (ESMPD’s) computer platform, and ESMPD used thecoding system to apply geographic area codes to theaddress records on the Census Bureau’s 1997 mail list.(For more information on geocoding see Chapter 6,‘‘Geographic Area Coding.’’)

ADMINISTRATIVE RECORDS

Background Information

The census first used administrative records to collectdata in the 1890 enumeration, when special enumeratorsvisited real estate recorders’ office records to obtain dataon individual and corporate debt. However, the CensusBureau did not try large-scale use of administrativerecords in the economic censuses until 1954, whenselected data items (employment, payroll, and sales) forsmall retail ($2,500 or more in sales) and services ($1,000or more in sales) nonemployers were tabulated from IRStax returns. Thereafter, administrative records obtainedfrom Federal agencies were a significant data source forthe Census Bureau’s economic statistical programs.

Access to already existing administrative records andtheir use in the 1997 Economic Census—

• Reduced the cost of data collection for the CensusBureau.

• Reduced company response burden, particularly forsmaller companies.

• Provided company and establishment names andaddresses for the mailing and sampling lists used in theCensus Bureau’s data collection programs.

• Supplied information the Census Bureau used in editingreported data.

The principal sources of the administrative recordsused by the Census Bureau were the IRS, SSA, and BLS.Information from these records was compiled by the Cen-sus Bureau under the confidentiality laws governing theCensus Bureau, as well as under the confidentiality regula-tions of the supplying agencies. The statistics collected bythe Census Bureau could not be used for taxation, regula-tion, or investigation. Moreover, only statistical totalscould be released to the public.6

Obtaining the Records

Preparation. Prior to the 1997 census, the CensusBureau reviewed the content of the administrative recordsavailable from other agencies to determine which eco-nomic sectors and firms could be covered by using those

files, rather than by enumeration. This task was compli-cated by the adoption of NAICS, which defined twice asmany major economic sectors as SIC, and which trans-ferred entire industries into new sectors. Nevertheless,administrative records use in the 1997 census paralleledthat of previous economic censuses.

Each of the contributing agencies had confidentialityrequirements of its own governing the use of its records,and special provisions of various kinds were made toenable them to supply the records to the Census Bureau.7

IRS records. In the case of IRS, a clause in the Federalregulations governing that agency (Title 26, United StatesCode) specifically authorized it to provide tax data to theCensus Bureau for the latter’s data collection programs.The Census Bureau, however, made a specific request forthe records that were supplied on computer tape. IRS sup-plied two primary kinds of records—payroll data files (e.g.,total payroll for companies) and business data (i.e., busi-ness receipts and principal type of business activity). Therecords were received weekly, monthly, or annually fromthe following files—

• Annual Business Master File (BMF) containing each legalentity’s EIN, business name, address, and principalindustrial activity code.

• Monthly BMF entity change file supplementing theannual BMF with changes in names, addresses, etc.

• Weekly employer quarterly Federal tax return fileconsisting of all firms that reported Federal payrollwithholding taxes. This file included the EIN, total num-ber of employees as of the March 12 pay period, totalquarterly payroll, taxable tips, and payroll subject to theFederal Insurance Contribution Act.

• Weekly business tax return files, including selectedbusiness-related information, such as company rev-enues, returns and allowances, months in business, andprincipal industrial activity code for all corporations,partnerships, sole proprietorships, and nonprofitorganizations.

Each May, the IRS gave the Census Bureau its currentBMF as a benchmark for the monthly data. The CensusBureau converted the IRS’s principal business activitiescodes into census industry codes.8 Establishments thatcould not be assigned an industrial classification codebased on file information were assigned codes based ontrade name whenever possible, or were assigned to theForm NC-9923, ‘‘Classification Form,’’ mail file for thecensus.

6Title 13, United States Code, prohibits the Census Bureaufrom releasing any information that, directly or indirectly, mightbe used to identify an individual or company. See Appendix A forthe provisions of title 13.

7The Census Bureau reimbursed the cooperating agencies forstaff and other costs associated with the records transfer.

8Time and resource constraints prevented the complete con-version of the IRS Codes into the new NAICS Codes. Therefore,SIC Codes were used in some instances.



SSA records. The records provided by the SSA weredrawn from IRS birth9 classification files: these containedinformation reported to the SSE on EIN applications, suchas SIC classification and type of activity in which the com-pany was engaged. When the Census Bureau made itsannual request to the IRS for administrative records, thelatter agency authorized the SSA to provide the IRS-derived files to the Census Bureau, as well.

BLS records. The Census Bureau and the BLS signed amemorandum of understanding in 1992 regarding the useof the BLS records by the Census Bureau. Under the 1992agreement, the Census Bureau compiled a list of single-establishment EINs for which it needed the BLS industrialclassification codes. The BLS matched the Census Bureau’slist against its own Business Establishment List, andreturned a listing of matched records with the appropriateindustrial classification codes. Unmatched records were‘‘returned’’ to the SSEL without industrial classificationcodes. If no other source for an industrial classificationwas available, unmatched records were coded based ontrade name.

Using Administrative Records in theEconomic Census

The Census Bureau received the administrative recordsfiles on a flow basis and used them first to update theSSEL. The multiestablishment portion of the SSEL wasupdated annually by the Company Organization Survey(COS). During each update, the staff merged and undupli-cated new administrative records, and then matched themto the SSEL. These records were used to update payroll,employment, industrial classification code (SIC or NAICS),and address information.10 Establishments not matched tothe list were added as births.

After converting the activity codes into usable industrycodes, the Bureau used information from the administra-tive records to divide the census mail file into four majorgroups for data collection purposes—

• Multiestablishment firms, ASM sample, and out-of-scopeestablishments.

• Large single-establishment and classification-form casesto be canvassed by mail.

• Small single-establishment firms.

• Nonemployers (i.e. firms with no paid employees during1997).

Additional administrative record information (e.g.,births; SIC Codes; name and address changes; 1997second-, third-, and fourth-quarter payroll; receipts)arrived at Bureau headquarters throughout the first half of1998. For the births, the Census Bureau added the data tothe census mail file and sent the new businesses theappropriate questionnaires. Throughout census process-ing, the Census Bureau also used the administrative datain the SSEL to edit incoming questionnaire data. Adminis-trative records data for nonmail and delinquent cases weresent to the appropriate subject divisions for processing.(Name and address information obtained from administra-tive records between closeout and the end of December1997 was used to mail questionnaires in the third-quarter‘‘birth’’ mailout conducted in April 1998.)

The Census Bureau also extracted selected data fromadministrative records for tabulation in the census datafiles, covered in more detail in Chapter 8, ’’Data Collectionand Processing,‘‘ for more information.

ASSIGNING CENSUS FILE NUMBERS

The Census Bureau assigned a uniqueidentification—the Census File Number (CFN)—to eachestablishment covered by the census. For single-establishment employer companies, the Census Bureauassigned numbers from the administrative records used inpreparing the census mailing list. Each of these recordscontained a nine-digit EIN (assigned by the IRS for tax pur-poses) identifying each company. The Census Bureaumodified this number by inserting a zero as a prefix digitindicating the company was a single-establishment firm.For multiestablishment firms, the Census Bureau used itsown files, adding four digits to the parent company’scensus-assigned six-digit identification number (ID) tocreate a unique identifier for each establishment.

During mail list preparation, the computer added acheck-digit to the ten-digit CFNs to verify accuracy duringprocessing. These CFN/check-digits were printed both innumerics and as a bar code on the address label of eachquestionnaire.

PRINTING AND ASSEMBLING THEMAILOUT PACKAGES

Cover Letters, Instruction Sheets, and OtherMailing Materials

General information. Questionnaires and other packagematerials for multiestablishment companies, as well asthose for single-establishment manufacturing and miningfirms, were printed using four DocuPrint programmablegraphics printers at the NPC. In addition to questionnaires,the census mailing packages included several otheritems—return envelope(s), cover letters, instructionsheets, and, for multiestablishment addressees, inventorylistings (i.e., a complete listing of the location of each

9‘‘Births’’ were businesses that came into being during thespecified reference period. For example, a ‘‘third-quarter birth’’was a business that began during the third quarter of the refer-ence year.

10The Census Bureau staff imputed missing payroll andemployment data by computer for the administrative records por-tion of the data file during the final closeout in late Summer,1998.



establishment in the company). The Census Bureau devel-oped specifications for these materials during 1996 andearly 1997 to ensure that the individual items could beprepared and assembled into the census mailing packageswhen the questionnaires were ready.

Cover letters. The Census Bureau prepared two basiccover letters, one each for single-establishment and multi-establishment companies. The letters informed recipientsthat Federal law required them to complete and return thecensus questionnaires, and that the information would bekept confidential. They also provided the response date.Some special situations required special letters: forexample, mailing packages to manufacturing establish-ments that were part of the 1997 ASM panel included aletter explaining that the survey was being integrated intothe 1997 Economic Census.

Instructions. The Census Bureau tried to include thenecessary instructions on the various questionnaires, buthad to supplement these with separate instruction sheetsin the mailout packages sent to firms with establishmentsclassified in construction; finance, insurance, and realestate; retail trade; services industries; wholesale trade;transportation, communications, and utilities; as well asfor three specific industries within the manufacturing sec-tor (apparel manufacturers, loggers, and publishers). Inaddition, the Census Bureau prepared and distributed aneight-page instruction manual (Form MA-1000-R1) forcompanies on the 1997 ASM panel.

Inventory sheets. Multiestablishment companies’ cen-sus packages included an inventory list of all their estab-lishments, including those not in scope of the census. Thisgave respondents the opportunity to confirm that all thefirm’s establishments were covered by the census, and tocoordinate and control responses by subsidiaries. If one ormore of a company’s establishments were missing fromthe Census Bureau’s inventory, the respondent was askedto call the NPC’s toll-free telephone number to ask foradditional questionnaires.

Printing the questionnaires and other mailing mate-rials. For both the census and the follow-up mailings, pri-vate contractors printed and assembled the bulk of thecensus mailout packages for single-establishment compa-nies that were not in manufacturing or mining. Single-establishment packages for the sectors listed in Table 7-1each consisted of a cover letter, the appropriate question-naire, instruction sheets, and the return envelope. Thecontents of the packages for the census mailout and forthe questionnaire follow-up mailings were identical exceptfor the cover letters.

The Census Bureau issued three primary contractsthrough the Department of Commerce for commercial

printing and assembly of 1997 Economic Census question-naire packages for the retail trades; wholesale trades; ser-vices industries; construction industries; transportationand utilities; and financial, insurance, and real estateindustries.11

The Department of Commerce awarded the major con-tracts for questionnaires for the 50 states in November1996, while those for the Outlying Areas questionnaireswere finalized in July 1997. The contractors began deliver-ing these materials to NPC in March 1997, completingfinal deliveries in December 1997.

Table 7-1.1997 Economic Census Single-Establishment MailoutPackages and Questionnaires Prepared by Contract

Packages ‘‘Loose’’questionnaires

TotalShort questionnaires 4,209,700 32,000Long questionnaires 6,549,100 140,000

Retail tradeShort questionnaires 1,058,300 7,000Long questionnaires 1,668,100 33,000

Service IndustriesShort questionnaires 2,239,600 14,000Long questionnaires 2,621,600 38,000

Wholesale tradesShort questionnaires 556,400 6,000Long questionnaires 1,050,400 42,000

Financial, insurance, and realestateShort questionnaires 355,400 5,000Long questionnaires 767,000 16,000

Construction industries*Long questionnaires 442,000 11,000

*The 1997 Economic Census did not use a short questionnaire forconstruction industries.

Using DocuPrint for the 1997 Economic Census

Background information. Prior to the 1997 EconomicCensus, the Census Bureau relied heavily on private print-ing contractors to augment the Census Bureau’s internalprinting operations. However, for the 1997 census, theCensus Bureau used the DocuPrint 4890 Highlight LaserPrinting System to print questionnaires, other forms, andliterature. This included all material for the multiestablish-ment mailing packages, as well as the single-establishment questionnaires for the manufacturing and

11Two of the primary contractors were predominantly involvedin printing, and used subcontractors to fill and seal the packages.The third contractor was a packaging and mailing company thatsubcontracted the printing of the materials, and then, itselfhandled the assembling and sealing of the packages. In all cases,the printing and package assembly were subjected to CensusBureau quality control procedures to ensure that the materialsmet the Census Bureau’s printing standards and that the packageswere correctly assembled for mailing.



minerals industries. Using the DocuPrint system enabledCensus Bureau staff to monitor the production processmuch more closely.

DocuPrint’s efficiency was especially evident during theassembly of questionnaire packages. NPC staff pro-grammed the four DocuPrint systems at the Jeffersonvillefacility to produce a variety of questionnaire packages andassociated instructional materials. The 1997 EconomicCensus used 474 separate report forms: only inquiriesabout a specific industry, or a closely related group ofindustries, appeared on the questionnaire for that indus-try. The DocuPrint equipment stored all the questionnairesand other forms used in the census mailing packages, as‘‘images’’ that could be accessed at will. The system drewfrom its data base all files needed to customize, print, andassemble the census questionnaires and packages. Thiseliminated the need to manually assemble and inspectpackages prior to mailing.

DocuPrint production. The Administrative andCustomer Services Division’s (ACSD’s) Forms and MailManagement Branch (FMMB) staff electronically designed,edited, and reviewed all census questionnaires from theirworkstations at Census Bureau headquarters. Once initialdrafts had been completed, files containing the specifiedquestionnaires and other materials were transferred to the

DocuPrint data base. Following quality control reviews, theprint files were sent electronically to the four DocuPrint4890 systems at the NPC facility. (ACSD maintained oneDocuPrint 4890 at Census Bureau headquarters for testprinting and for review of FMMB’s materials.) DocuPrintproduction began on September 1, 1997 and continued (ata combined rate, for all four machines, of up to 2 millionpages per day) until mid-December 1997.

Table 7-2 shows the quantities of 1997 questionnairesprinted by the DocuPrint systems in Jeffersonville.

In addition, a total of 73,646 Form NC-9923 ‘‘GeneralSchedules’’ were mailed to single-establishment firms(both those on the initial mail list and to second- andthird-quarter ‘‘births’’).

Address Labels

Description. The adhesive address labels for the censusmailout contained information for check-in and processingof the completed questionnaire. The label was arranged ina maximum of 16 ‘‘lines’’ (including several blank lines)showing identification information, including the estab-lishment’s CFN, the form number of the questionnaire,mailout or follow-up letter number or correspondencecategory code, EIN, industry classification code, stategeographic code, employment size code, county name or

Table 7-2.1997 Economic Questionnaires Printed By DocuPrint

Industry classificationMultiestablish-

mentcompanies

Single-establishmentcompanies*

Single-establishmentcompany Births*

Questionnaire type Long Long Short Long Short

Total 1,411,113 1,764,107 107,751 189,825 6,314

Auxiliaries 50,504 n/a n/a n/a n/a

Construction 11,903 106,438 0 13,096 0

Financial, insurance, and real estate 156,675 141,783 0 12,572 0

Manufactures/ASM 69,431 95,876 47,650 2,257 2,012

Mining 6,330 5,160 2,823 250 157

Retail and accommodations and foodservices 530,905 448,570 0 55,778 0

Service 330,171 633,213 0 70,109 0

Utilities 44,753 64,027 0 8,288 0

Wholesale 131,034 241,677 57,278 21,838 4,145

Guam 292 2024 0 266 0

Northern Mariana Islands 178 781 0 132 0

Puerto Rico 4,959 23,111 0 3,150 0

Virgin Islands of the United States 332 1,447 0 215 0

Totals 1,411,113 1,764,107 107,751 189,825 6,314

*Single-establishment company totals include second- and third-quarter (1997) ‘‘births,’’ that is, newly identified businesses added to the SSEL dur-ing the quarters covered. Single-establishment ‘‘Births’’ are fourth-quarter 1997 births. abbreviation, permanent plant number (PPN), establishment or,for multiestablishments, company name and address, and the like.



Printing. The address labels for single-establishmentcompanies—except manufacturing and mining—firmswere printed on high-speed ink-jet printers at the NPC,using computerized address files transmitted to Jefferson-ville via datalink telephone lines from the Census Bureau’sheadquarters. Approximately 5.4 million address labels,including 1.6 million used in the November classificationmailing, were printed and then affixed to the individualmailing packages using automated labeling equipment.The NPC check-in updated the response list daily, and theCensus Bureau created an address file of nonrespondentsfollowing each response cutoff date. The Jeffersonvillestaff used the updated address files to generate newaddress labels for follow-up mailings to nonrespondents.The NPC used the DocuPrint equipment to produceaddress labels for the classification forms, all multiestab-lishment mailing packages, and for the manufacturing andmining single-establishment packages.

Quality Assurance

Contract printing/package assembly. NPC’s QualityAssurance Section (QAS) monitored questionnaire printingat printing contractors’ sites, and also inspected the con-tractors’ materials received at Jeffersonville. On-site moni-toring required that the QAS staff observe the equipmentoperators’ method of sample selection, check that themachine counters were working properly, and select a ran-dom sample of printed material for inspection from eachmachine. The inspectors checked the contractors’ ownquality assurance plan to ensure proper record keepingand correct sample selection.

Cartons of ‘‘flat forms’’ (i.e., materials not assembledinto packages) were randomly sampled by type andchecked to ensure that the printed materials met specifica-tions. For cartons of assembled mailing packages, the QASstaff matched the information on the label against thepackage specifications. The clerks selected three packagesfrom each carton and verified (1) the color and registrationof the forms and questionnaires, (2) that all inserts werepresent and in the proper sequence, and (3) that all docu-ments in each package matched the envelope. Individualprinted items within each package were checked to verifythat each item was printed to specifications (i.e., color ofshading and ink, quality of image, and dimensions), andthat the envelope met requirements.12 One package fromthe first contractor quality assurance sample for eachpackage type was selected, and inspectors proofread allitems to check for misspelling and incorrect language. The

quality control clerks notified their supervisors of anyerrors detected in the contractor-selected samples, and thesupervisors or lead clerks determined what correctiveaction was required.

At Jeffersonville, the mailing packages from the privateprinting contractors underwent a second quality assurancereview.13 For shipments of one package/form type, twocartons were selected from each pallet for inspection,while inspectors sampled shipments containing more thanone type of package or form by selecting one carton ofeach type of package. The staff checked each cartonselected to verify that it contained the correct type andnumber of packages, and then selected three packagesfrom each carton for content inspection.

Any errors detected were entered on the Quality Assur-ance Control/Inspection Record for the shipment, and thefive preceding and succeeding items in the cartonsampled were inspected. If additional errors were found,the next five preceding and succeeding items wereinspected and any additional errors resulted in 100-percent verification of the contents of the carton. If noerrors were identified, or if the defective items were foundin the original sample but not in the expanded search, thedefective packages were referred for correction and theremainder of the shipment was released.

If additional defective packages or items were identifiedin the expanded review, the shipment was rejected andeither (1) sent back to the contractor for reprinting, (2) setaside for use (of acceptable items) only if supplies of theitem ran low, or (3) used for follow-up correspondenceoperations.

The quality of the contractor-printed and assembledmaterials for the 1997 Economic Census was good. Prob-lems that did occur were corrected by reprinting or byreplacing cartons or individual packages as necessaryfrom the overstock supplied by the contractors.

DocuPrint printing and package assembly. The QASconducted quality assurance inspections of the materialsprinted on the DocuPrint equipment. These included multi-establishment packages, single-establishment question-naires, classification and general questionnaires, consoli-dated forms, follow-up letters, and instructions and flyers.The inspections covered all DocuPrint work units (a workunit was a complete print file— about 1,500 pages ofprinted material). Inspections focused on two major typesof defects—printing and programming errors.14

The DocuPrint quality assurance procedures involvedthree kinds of inspections—

12The questionnaires varied in size from 8 1/2 x 11 inches to10 1/2 x 17 inches and were printed in black ink on 50-lb. offsetwhite book stock, with shading in a variety of ink colors. All thecover letters, instruction sheets, and envelopes were printed inblack ink on white stock.

13Shipments of ‘‘flat forms’’ did not require additional inspec-tions.

14Printing errors (e.g., illegible printing on the label or in theestablishment data, printing not centered on the page [generalmisalignment was a printing problem, misalignment of only partof a printed page indicated a programming error], unreadable bar-code, etc.) were ‘‘noncritical errors’’ that could be corrected byreprinting the file. Programming errors were ‘‘critical’’ errors



• A sample inspection of the first production file printedfor each print group.

• Daily monitoring of the printing operation.

• Inspection of all materials reprinted due to machinereject.15

The QAS chose a ‘‘start with’’ number from a randomnumber table, and every tenth company package oritem thereafter, for inspection until reaching the end ofthe file. The inspectors used laser wands to check eachpackage/item for readable barcodes, then checked forcorrect format and assured that the package containedthe correct questionnaires or other forms. Once the firstproduction file for a file type16 was accepted, daily moni-toring was used for that file type. Inspectors selected onefile from each shift (not one file of each type), and threecompanies’ packages (any company package with machinerejects, plus at least one randomly selected package) fromthe files for verification. If only printing errors weredetected, the file was held and an expanded inspection17

was conducted to determine the extent of the problembefore reprinting. When inspectors identified program-ming errors, the file was rejected and the inspectingstatistician contacted the EPCD staff for instructions totake further action on the file.

The DocuPrint mailing packages were assembled cleri-cally, in the order printed, and underwent clerical verifica-tion. The clerks informed their supervisors of errors andchecked the preceding and succeeding packages. Uponcompletion of this inspection, the clerks checked that thepackages for all companies in the file were present. Fileswith missing packages were held until the requisite mate-rials were added, then were subject to 100-percent verifi-cation. Once the QAS confirmed that all the packages for afile were present, and errors corrected, the file wasreleased.

Procedures for machine-assembled mailing packagesdiffered slightly. For packages prepared using equipment

programmed to assemble the materials by barcode, themachine selected a random sample of each work unit forclerical inspection, while packages assembled using a pro-grammed number of pages of materials for a specific com-pany were inspected on a 100-percent basis. The first 25packages, and 1 in 100 of the remainder in the file, wereinspected for content; the clerk checked that all pro-grammed materials were included, and that the surveycode and sequence number of the packages within the file(i.e., 1-25, then 125, 225, 325, etc.) were correct. Forpackages assembled by page number, clerks checked thatall packages were present, that address informationshowed through the envelope window, and that allsequence numbers were present and correct. Necessarycorrections were made before files were released.

Single-establishment mailout packages printed on theDocuPrint equipment were assembled clerically and sub-ject to 100-percent verification.

Label printing. The quality assurance program for boththe high-speed printer and DocuPrint label printinginvolved (1) sample inspection of the first production fileproduced for each print group, (2) daily monitoring of theprinting operation, and (3) inspection of all materialsreprinted because of printing errors.

For the first production file inspection, staff randomlypicked a ‘‘start with’’ number for each file and checkedthat sheet of labels, then inspected every tenth sheetthereafter. For daily monitoring, staff selected one ‘‘file’’for each printing shift, and inspecting that file (using thesame sample selection method employed for the first pro-duction file inspected). Quality assurance staff checkedeach chosen sheet of labels for printing legibility. Foreach label, the clerks looked for missing alpha or numericcharacters, and for mandatory entries (e.g., barcode,CFN, survey code, form number, mailout/follow-up/correspondence category code, reporting media, EIN,classification code, etc.).

When the review identified printing errors, the labelsheet was placed on hold and an expanded inspection wascarried out, checking succeeding and preceding sheetsuntil the extent of the problem was found. Then, the filewas referred for reprinting. When programming errorswere found (for example, mandatory items missing fromthe labels), the sheet and file were rejected, EPCD staffwas informed, and corrective action was taken.

PRE-CENSUS COMPANY CONTACTS

Information Mailings

In November 1996, the Census Bureau sent an informa-tion mailout to 10,000 of the largest—based on totalemployment and number of establishments reported inthe previous census—SSEL-listed multiestablishment com-panies. These companies represented about half the totaleconomic activity covered by the census. This mailing had

present within the file itself and had to be referred for correctionto the appropriate subject programming staff at Census Bureauheadquarters.

15‘‘Machine rejects’’ were any packages with an unreadable bar-code. These would include not just badly printed codes (e.g.,unclear, partially obscured, offline, etc.,), but such things as ques-tionnaires inserted backwards, or materials in the wrong order sothat the barcode could not be read.

16The file types were as follows: Annual Survey of Manufac-tures single establishments; manufacturing; mineral industries;multiestablishment companies requiring split, electronic, or certi-fied mailing; companies in a foreign location; small (other compa-nies with less than 28 printed sheets in the package); medium(companies with 29 to 130 printed sheets in the package); andlarge (more than 130 printed sheets).

17The ‘‘expanded’’ inspection selected one company packageon each side of the defective package in the work unit andinspected their materials. If further defects were found, theinspectors moved on to the next company on the side(s) of thenewly identified defective package, and continued until no errorswere identified, or until the next/previous selected company wasreached.



three major objectives: (1) to establish initial contact withthe largest companies covered by the economic census,(2) to provide those firms contacted with information onthe forthcoming census and the kinds of information thatwould be requested, and (3) to facilitate response by giv-ing companies a chance to assemble their data. Responseswere used to update the SSEL prior to the census mailout.

Each mailing package consisting of the following:

• A cover letter signed by the Director explaining that the1997 Economic Census would be carried out in about ayear and listing the kinds of questionnaires the recipientcompany would receive in the census.

• Form EC97-PR-1, Preparing for the 1997 Economic Cen-sus, a 24-page booklet presenting basic information onthe plans for the 1997 census (including basic data tobe collected, timing of the enumeration, and samples ofcensus questionnaires), how respondents could use theWorld Wide Web to view census questionnaires andobtain more information on the 1997 Economic Census;telephone and fax numbers, and e-mail addresses forsubject division contacts at the Census Bureau.

• A Contact Exchange Card showing the mailing addressshown in the SSEL for the recipient, and asking respon-dents to (1) make address corrections, (2) provide the

name of the person within the company who would beresponsible for the firm’s response to the census, and(3) provide the name of the supervisor of the company’sresponding official.

The Account Manager Program

Of the 10,000 companies that received the informationmailing, the Census Bureau selected 1,000 for inclusion inits Account Manager Program. The criterion for selectionwas the company’s likelihood of either delaying itsresponse, or not responding at all, to earlier censuses. Toimprove the response of these companies, the CensusBureau assigned census staff to serve as their ’’accountmanagers.’’ Each staffer was responsible for several com-panies, and followed up the information mailout by tele-phone.

Chapter 5, ‘‘Promoting Census and Product Awareness,’’provides details of the Account Manager Program: itshows how the Census Bureau created a system to identifythe correct contact within targeted companies, and to pro-vide account managers with the capacity to reinforce theobjectives of the larger census mailout.



Chapter 8. DATA COLLECTION AND PROCESSING

The Classification Mailout . . . . . . . . . . . . . . . . . . . . . . . . . . . 73Background Information . . . . . . . . . . . . . . . . . . . . . . . . . . . 73The Three Classification Mailouts . . . . . . . . . . . . . . . . . . . . . . 73Processing the Classification Mailout . . . . . . . . . . . . . . . . . . . . 73

The Census Mailout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73Electronic Reporting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

Following Up the Census Mailout. . . . . . . . . . . . . . . . . . . . . . . . 74Background Information . . . . . . . . . . . . . . . . . . . . . . . . . . . 74Mail Follow-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75Single-establishment follow-up . . . . . . . . . . . . . . . . . . . . . . 75Multiestablishment follow-up . . . . . . . . . . . . . . . . . . . . . . . 75

Telephone Follow-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

Processing Data From Administrative Records . . . . . . . . . . . . . . . . . 77

Processing the Questionnaires . . . . . . . . . . . . . . . . . . . . . . . . . 77The Transmittal Tracking System. . . . . . . . . . . . . . . . . . . . . . . 77Receipt and Check-In . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77General information . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Preliminary sorting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Barcode check-in and sorting . . . . . . . . . . . . . . . . . . . . . . . 78Disposition of mail receipts . . . . . . . . . . . . . . . . . . . . . . . . 78

Correspondence Processing . . . . . . . . . . . . . . . . . . . . . . . . . 78Correspondence reading . . . . . . . . . . . . . . . . . . . . . . . . . . 78CFN research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78Correspondence with single-establishment questionnaire processing . . 79Census Bureau-originated correspondence . . . . . . . . . . . . . . . . 79Package assembly and mailout operations . . . . . . . . . . . . . . . . 79

Multiestablishment Completeness and Coverage Processing . . . . . . . . 79Microfilming the Questionnaires . . . . . . . . . . . . . . . . . . . . . . . 80Data Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80Data keying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80Quality assurance for data keying . . . . . . . . . . . . . . . . . . . . . 81Output files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Control File Match . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81Matching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81Control file match referrals. . . . . . . . . . . . . . . . . . . . . . . . . 82

Decentralized Post-Collection Processing . . . . . . . . . . . . . . . . . . . 82General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82Micro Edit Referral Processing . . . . . . . . . . . . . . . . . . . . . . . . 82Background information . . . . . . . . . . . . . . . . . . . . . . . . . . 82Complex editing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82Subject area databases . . . . . . . . . . . . . . . . . . . . . . . . . . . 83Reject and referral processing . . . . . . . . . . . . . . . . . . . . . . . 83

CONTENTS


Decentralized Post-Collection Processing—Con.Macro Analytical Processing . . . . . . . . . . . . . . . . . . . . . . . . . 83Background information . . . . . . . . . . . . . . . . . . . . . . . . . . 83Data tabulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83Subject tabulation files for main publication series . . . . . . . . . . . . 83Corporate repository . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84

Analytical Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85


Data Collection and Processing

THE CLASSIFICATION MAILOUT


The American economy is never static; new businessesappear and old ones disappear or change to adapt to newconditions. The tabulation and publication of statisticaldata requires that data be summarized under specifiedclassifications, so every economic census must identifyand correctly classify new business activities, as well asrevise those not accurately classified in the U.S. CensusBureau’s records. For the 1997 Economic Census, theadoption of the North American Industry ClassificationSystem (NAICS) entailed doubling the number of majortrade areas for classifying economic activity from 10 ‘‘divi-sions’’ under the old Standard Industrial Classification (SIC)Code to 20 ‘‘sectors’’ under NAICS. Since NAICS definedover 300 new industries within the U.S. economy, it wasimperative to correctly classify businesses before attempt-ing to measure their economic activity in the census.

The Census Bureau carried out a special classificationquestionnaire mailing to establishments for which theagency had insufficient information to determine the cor-rect NAICS classification. These generally were small,single-establishment companies that would not havereceived a census long form if their classification informa-tion had been complete.

The Three Classification Mailouts

In early November 1997, the Census Bureau mailedclassification questionnaires to 1,362,914 establishmentsselected from the census mail list. In the general censusmailout, in December 1997, the Census Bureau sent classi-fication report forms to an additional 83,000 single-establishment manufacturing businesses that could not beclassified in detail.

Although the census mail list was compiled prior to theinitial mailout, the Census Bureau continued to compilelistings for new businesses—‘‘births’’—identified from

Internal Revenue Service (IRS) and Social Security Adminis-tration (SSA) records throughout 1997. Thousands of newbusinesses identified in the third and fourth quarters of1997 also needed to be included in the census. In March1998, the Census Bureau carried out a supplementaryclassification mailing, sending classification forms to194,893 businesses added to the census mail file for thethird quarter of 1997. Fourth quarter ‘‘births’’ for 1997were not mailed questionnaires; the Census Bureau usedadministrative records to cover these new businesses.

Processing the Classification Mailout

Responses to the classification mailouts, together withcensus questionnaires, were received and processed at theNational Processing Center (NPC) office in Jeffersonville,IN.1 After check-in, clerks screened the classification formsto determine the industrial classification of the establish-ments and then referred the questionnaires for additionalprocessing. The information on primary economic activityfrom the classification report forms was used to updatethe Census Bureau’s Standard Statistical Establishment List(SSEL). The basic business data (receipts, payroll, employ-ment, and SIC/NAICS classification) from the long-formquestionnaires were keyed and added to the respectiveindustry files.

THE CENSUS MAILOUT

In the second half of December 1997, the NPC mailedout approximately 4.32 million census questionnaires andclassification questionnaires, with response requested byFebruary 12, 1998. In mid-March 1998, NPC mailed anadditional 526,328 questionnaires (and 194,893 classifica-tion forms) to single-establishment companies that hadbegun operations during the second and third quarters of1997. Together, these initial mailings totaled approxi-mately 5.04 million questionnaires. The numbers of ques-tionnaires and classification forms mailed, including initialmailings of single-establishment second- and third-quarterbirths, by type, are shown in Table 8-1.




Table 8-1.Initial Mailout: 1997 Economic Census

Trade or geographic area

Multiestablishment companies Single-establishment companies*

Total

Longques-

tionnaires

Consolidatedques-

tionnaires Total

Classificationques-

tionnaires

Longques-

tionnaires

Shortques-

tionnaires

Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1,411,113 1,400,850 10,263 3,630,777 1,557,807 1,958,704 114,266Auxiliaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50,504 50,504 0 0 0 0 0Construction industries . . . . . . . . . . . . . . . . . . . 11,903 11,903 0 0 0 119,461 0Finance, insurance, and real estate . . . . . . . 156,675 148,474 8,201 255,136 96,395 158,741 0General (forms 9901 & 9923) . . . . . . . . . . . . . 73,646 73,646 0 362,721 362,721 0 0Manufacturing and annual survey ofmanufactures (ASM) . . . . . . . . . . . . . . . . . . . 69,431 69,431 0 230,414 82,575 97,997 49,842

Mineral industries . . . . . . . . . . . . . . . . . . . . . . . 6,330 6,330 0 10,865 2,409 5,437 3,019Retail trade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 530,905 530,905 0 823,931 319,076 504,855 0Service industries . . . . . . . . . . . . . . . . . . . . . . . 330,171 328,109 2,062 1,362,062 662,263 699,799 0Utilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44,753 44,753 0 105,162 32,368 72,794 0Wholesale trade . . . . . . . . . . . . . . . . . . . . . . . . 131,034 131,034 0 327,594 0 266,189 61,405Guam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292 292 0 2,446 0 2,446 0Northern Mariana Islands . . . . . . . . . . . . . . . . 178 178 0 1,010 0 1,010 0Puerto Rico . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4,959 4,959 0 28,182 0 28,182 0U. S. Virgin Islands . . . . . . . . . . . . . . . . . . . . . . 332 322 0 1,793 0 1,793 0

*Single-establishment counts include 721,221 second- and third-quarter 1997 ‘‘births’’, mailed in January and March 1998.

ELECTRONIC REPORTING

The Census Bureau first used electronic reporting tocollect economic census data for the 1987 census. Theagency employed the technique primarily for large retail,food services, and hotel chain enterprises—mailing mag-netic computer tape reels to selected companies andrequesting them to enter the data for their establishmentson the reels and return them to the agency.

For the 1997 Economic Census, the Census Bureauagain offered companies the option of response throughmagnetic tape, but also contracted with a private companyto develop an electronic questionnaire on diskette for usewith personal computers. The latter project produced theWindows-based Computerized Self-Administered Question-naire (CSAQ) for use in the 1997 Economic Census. TheCensus Bureau established stringent requirements for theCSAQ, it had to—

• Be usable by multiestablishment companies engaged ina variety of industrial activities.

• Enable companies to export and manipulate establish-ment identification and address information supplied bythe Census Bureau.

• Allow companies to link to internal spreadsheets anddatabases.

• Include necessary communications software to transmitcensus data.

The Census Bureau selected approximately 600 compa-nies for electronic data collection in the 1997 EconomicCensus—400 were sent the CSAQ diskettes and about 200received computer tape in the December 1997 censusmailing. These companies represented 250,000 separatebusiness establishments (about 3.8 percent of the total

establishments covered by the census). During data collec-tion, the Census Bureau received another 71 inquiries, andsent CSAQ diskettes to these companies. By the time datacollection closed out on September 23, 1998, 671companies—representing 317,593 business establish-ments had responded to the census using the CSAQ ormagnetic computer tape.

FOLLOWING UP THE CENSUS MAILOUT


The 1997 Economic Census employed separatefollow-up plans for single-establishment and multiestab-lishment companies. Both plans involved remailing censusquestionnaires to nonrespondent companies—severaltimes in some cases—as well as telephone follow-up tolarge multiestablishment companies. The NPC conductedthe mail follow-up in four phases, beginning in March1998 and extending through the last week of June 1998.Telephone follow-up began prior to the requested ques-tionnaire response date—February 12, 1998—when Cen-sus Bureau staff called selected large multiestablishmentcompanies to confirm that the census packages had beenreceived. Telephone follow-up continued through June1998.

Response close-out dates were set for each major typeof mailout package—classification, multiestablishment,single-establishment (Annual Survey of Manufacturers[ASM], non-ASM manufacturing and mineral, and othersingle-establishments), and third-quarter births. (See Table8-2 for mailout totals.) After each close-out date, NPC staffused computerized check-in files to generate mailing listsof nonrespondents in each category. The close-out datesfor the four follow-up mailings extended over several daysas follows:



First follow-up February 23-March 6, 19982

Second follow-up March 27-April 4, 1998Third follow-up May 4-May 18, 1998Fourth follow-up June 12-June 26, 1998

Delinquent multiestablishment companies were subjectto mail and telephone follow-up, with telephone follow-uplistings generated during the second, third, and fourthfollow-up closeout. Totally delinquent companies that hadrequested a time extension for response were followed upas well, with one letter mailing and one telephone follow-up.

The lists of nonrespondents produced after each close-out date were used, together with the NPC’s high-speedprinters, to print the address labels for each mail follow-up.

Mail Follow-Up

Single-establishment follow-up. The single-establishment mail follow-up covered third-quarter 1997‘‘births,’’ classification cases, and census and ASM single-establishment mail cases. Each category of respondentswas subject to a specific follow-up procedure. One ‘‘letter’’mailing to ASM single-establishment nonrespondents wascarried out in February 1998, while three questionnairefollow-ups were conducted in March, April, and May. (Afourth follow-up mailing, to single-establishments withestimated annual payrolls of $200,000 or more, was car-ried out in July.) Classification cases were subject to twofollow-up mailings in April and May 1998.

The first classification form close-out date was March30, with the mailing conducted during the following week.Each package included the cover letter form EC-G-L11,requesting response, the classification form and a returnenvelope. This first follow-up mailing sent questionnairesto approximately 1.32 million nonrespondent addresses.The final closeout for the classification cases was May 8,after which the NPC mailed packages containing the formfollow-up letter, another copy of the classification form,and a return envelope to delinquent classificationaddresses.

The follow-up process for nonclassification single-establishment delinquents varied by type of case. Third-quarter births received only one mailing; 115,584 casesnonrespondent as of April 29 were sent packages contain-ing the appropriate questionnaire, a follow-up letter, and areturn envelope. ASM single-establishment delinquentcases were sent a letter in the first week of March as partof the follow-up of all delinquent single-establishmentcases. Thereafter, they were subject to two more mailings,following close-out dates of April 9 and May 18. Each mail-ing used the appropriate report form, together with formletters urging response. All other single-establishment

nonrespondents—except classification cases were—followed up three times after close-out dates of March 4,April 15, and May 22, 1998. Each mailing involved com-plete census packages containing the appropriate reportform, cover letter3, and return envelope.

Close-out dates and package counts for the single-establishment mail follow-up are given in more detail inTable 8-2.

Table 8-2.Single-Establishment Follow-Up Mailings

Packagetype

Endcloseout

datePackages

mailed

First follow-upASM cases . . . . . . . . . . . . . . . . Letter 02/23/98 12,463Non-ASM, excluding classifi-cation and third-quarterbirths. . . . . . . . . . . . . . . . . . . . . Questionnaire 03/04/98 861,848

Classification, excludingthird-quarter births. . . . . . . . . Questionnaire 02/28/98 315,356

Third-quarter births. . . . . . . . . . Questionnaire 04/29/98 5,326

Second follow-upSingle-establishment, exceptclassification andthird-quarter births. . . . . . . . . . Questionnaire 548,121

Classification, excludingthird-quarter births. . . . . . . . . Questionnaire 04/15/98 175,261

Third follow-upSingle establishment . . . . . . . . Questionnaire 05/22/98 415,809

Fourth follow-up*Single establishment . . . . . . . . Questionnaire 07/09/98 109,165

*This follow-up was mailed only to single establishments with esti-mated annual payrolls of $200,000 or more.

Multiestablishment follow-up. Follow-up proceduresfor multiestablishment companies varied (1) by the size ofthe company (the size designations varied by industry ortrade area, based on volume of sales and annual payroll),(2) by whether the firm was totally or only partially delin-quent, and (3) by whether request for a time extension forresponse had been received by the Census Bureau.

The follow-up mailings and listings done were asfollows:

• For large and medium size (‘‘L’’ and ‘‘M’’) companies ineach trade area—

• Totally delinquent firms—those that had made noresponse whatever by the cut off date following eachmailout (see Table 8-3) and with no extensionrequested were scheduled for up to four letterfollow-up mailings. Telephone follow-up listings forthese delinquents were generated following theMarch, May, and June close-out dates.

2Third-quarter ‘‘births’’ were mailed out on March 18, 1998,shortly after the first mail follow-up.

3Cover letters varied by type of establishment; a different ver-sion of each follow-up cover letter was used for manufacturing/mineral industries establishments, and for all other single-establishments (excluding ASM, classification, and third-quarterbirth cases). The content for all letters was similar.



• Partially delinquent companies—i.e., companies fromwhich at least one questionnaire had been receivedand with no time extension requested—were fol-lowed up by letter up to two times and also weresubject to up to two telephone follow-up calls.

• Totally and partially delinquent companies that hadrequested a time extension for response were fol-lowed up as well, with one letter mailing and onetelephone follow-up.

• For smaller (‘‘N’’ and ‘‘S’’)4 multiestablishmentcompanies—

• Totally delinquent ‘‘N’’ and ‘‘S’’ companies with notime extension requested were subject to up to threeletter follow-ups in the first three follow-up mailings.If they were still delinquent after the June 26 close-out date, they were sent a complete questionnairepackage follow-up. Those that had requested timeextensions received one letter follow-up after theMay 4 close-out date.

• Partially delinquent ‘‘N’’ and ‘‘S’’ companies withouttime extensions requested were sent letter follow-upsin May and June, while those that had requested timeextensions for response received only one letterfollow-up if they remained partially delinquent afterthe June close-out date.

The close-out dates and packages mailed in the multies-tablishment company mail follow-ups are shown in Table8-3.

Telephone Follow-Up

The Census Bureau supplemented its mail follow-upwith a telephone follow-up targeting selected delinquentlarge and medium-size multiestablishment companies.Analysts at the Census Bureau’s headquarters referred toworksheets and listings generated at the NPC to conductthe telephone follow-up in three phases, after the April,May, and June 1998 close-out dates for multiestablishmentmail follow-up.

The Suitland headquarters staff contacted the compa-nies referred for telephone follow-up as the workload ofincoming calls permitted. The purposes of the telephonecontacts were to (1) remind companies to file their censusreports, (2) determine when they would be filing, and (3)provide assistance that might be required. Some compa-nies were subject to more than one telephone contactdepending on their filing status.

The first telephone follow-up list—compiled after theApril 3 mail closeout—comprised 2,565 totally delinquentcompanies that had not requested a time extension forresponse. After the May 7 closeout, the NPC produced

worksheets for 2,433 more totally delinquent companies,including 744 companies with unexpired time extensions,plus work sheets and establishment listings for 940 par-tially delinquent multiestablishments (with no time exten-sions requested). The last referral was made followingthe June 11 mail closeout, when 1,687 totally, and 545partially, delinquent firms were referred for telephonefollow-up.

Table 8-3.Multiestablishment Follow-Up Mailings

Packagetype

Endcloseout

datePackages

mailed

First follow-upTotally delinquent,no extension . . . . . . . . . . . . . . Letter 02/2/98 73,333

Second follow-upTotally delinquent,no extension . . . . . . . . . . . . . . Letter 04/01/98 51,669

Third follow-upPartially delinquent,no extension . . . . . . . . . . . . . . Letter Letter 5,326

Totally delinquent,no extension . . . . . . . . . . . . . . Letter Letter 40,772

Totally delinquent, unexpiredextension . . . . . . . . . . . . . . . . . . . Letter Letter 1,709

Fourth follow-upPartially delinquent,no extension . . . . . . . . . . . . . . Letter 06/10/98 3,252

Partially delinquent, unexpiredextension . . . . . . . . . . . . . . . . . Letter 06/10/98 282

Large companies, totallydelinquent, no extension . . . . . . Letter 06/11/98 950Small companies, totallydelinquent, no extension . . . . Questionnaire 06/25/98 31,661

Other, totally delinquent,unexpired extension . . . . . . . . . Letter 06/11/98 846

Census Bureau staff prepared folders for each companyreferred for telephone follow-up. Each folder contained acompany worksheet (and establishment listings for par-tially delinquent companies) and forms for recording tele-phone calls. Nonrespondent companies were called indescending payroll order—i.e., larger companies werecalled first—and the interviewers tried to speak with thecontact person shown on the company worksheet. Theanalysts updated the worksheets with information on eachcontact and noted any actions requested, such as requestsfor time extensions. If no contact was listed, the analystasked for the person responsible for filing governmentreports.5

4It should be mentioned that the designations ‘‘L,’’ ‘‘M,’’ ‘‘N,’’and ‘‘S’’ companies are not statistical data collection, tabulation,and publication categories. These designations are ‘‘in-house’’ways that the Census Bureau used to refer to companies accord-ing to the level of intensity of the follow-up that they will receive.

5The Census Bureau carried out a pre-census mailing to10,000 multiestablishment companies in November 1996 toestablish initial contact for the census. The Census Bureau askedcompanies to give the name of the person who would be respon-sible for census reporting. Chapter 5, ‘‘Promoting Census andProduct Awareness,’’ discusses this subject in detail in the sec-tions on the ‘‘Account Managers’’ and ‘‘Company Contact’’ pro-grams.



PROCESSING DATA FROMADMINISTRATIVE RECORDS

The Census Bureau sampled—at variable rates—eacheconomic sector’s census mail universe of nonemployerand small employer firms to develop estimates of special-ized data (e.g., merchandise-line sales for retail trade;revenue-line sales in transportation; sources of revenuefor finance, insurance, and real estate). The bulk of theinformation on nonemployers and small employers, how-ever, was drawn from the Census Bureau’s SSEL andadministrative records files. The Census Bureau obtainedabout 140 million records on computer tape from the IRSbetween January 1997 and December 1998, and approxi-mately 60,000 SSA birth (new business) records monthlyduring the same period. (Chapter 7, ‘‘Preparations for Cen-sus Mailout,’’ details the Census Bureau’s acquisition ofadministrative records.)

The Census Bureau identified, accumulated, and editedadministrative records of nonemployer firms, andextracted basic data for these operations and for selectedsmall employer companies not included in the mailsample.6 The agency gathered data on sales or receipts,kind of business, and geographic location for nonem-ployer and small employer companies in all the majoreconomic sectors covered in the economic census exceptwholesale trade.7 Records for businesses not meetingminimum criteria for census coverage, as well as forprobable duplicates of employer businesses, wereremoved from the files. Once the nonemployer files foreach trade area had been edited and the informationextracted, the resulting data file was transmitted to theappropriate subject division staff for tabulation andanalysis.

PROCESSING THE QUESTIONNAIRES

The Transmittal Tracking System

The Transmittal Tracking System (TTS) was used by theCensus Bureau to control economic census documentsmoving from one processing unit to another. The paperdocuments—questionnaires and paper correspondencewith attached materials—were transported from unit tounit in wheeled bins. Each processing unit created acomputer-generated transmittal document for each bin of

materials just prior to its referral to the next unit. The TTSretained the computerized record of the progress of eachbin, while the paper copy accompanied the bin as itmoved from unit to unit. The transmittal records anddocuments showed the count of questionnaires, corre-spondence cases, etc., in each bin, and the processingactions completed on them to that point.

Receipt and Check-In

General information. The economic census check-inbegan late in December 1997 at the NPC in Jeffersonville,IN. The check-in system used laser technology to read thebarcodes on the questionnaire address labels and sort thereceipts. After each mail response close-out date, thecheck-in file was used to identify nonrespondent establish-ments and companies for telephone or mail follow-up.

Preliminary sorting. The U.S. Postal Service presortedincoming economic census mail packages using differentmail box numbers for eight specified categories ofreturns.8 The packages were placed in trays by type ofreceipt and then routed to the preliminary sort clerks whochecked to make certain that each tray contained only onetype of receipt. The clerks sorted the receipts contained ineach tray and referred them to NPC processing units asfollows:

• Materials addressed to a specific analyst; forwarded tospecified analyst.

• Multiestablishment undeliverable as addressed (UAA),receipts, single-establishment ASM receipts, single-establishment UAA receipts, census follow-up letter UAAcases, and Census Bureau originated correspondenceUAA cases, were referred to the Directory Analyst Unit.

• Receipts with no visible barcode, and economic censusreceipts with visible barcodes, were referred to RegisterBatch Unit for check-in.

• All other economic census related receipts; referred tothe Open and Sort Unit. (Receipts obviously not censusrelated were referred to the appropriate subject area.)

Receipts referred for check-in were grouped intobatches of up to 10,000 pieces and assigned batch andcheck-in codes that enabled the NPC staff to track eachbatch through processing by using the TTS.

6The definition of ‘‘small’’ employer firms varied slightlyamong the various sectors. ‘‘Small’’ firms were generally single-establishment companies that met specified minimum annualsales (usually $1,000) and maximum payroll. The number ofemployees varied among sectors; in retail trade small-employerfirms had 1 to 3 paid employees; in the sectors derived from theSIC services area the ‘‘small’’ companies had 1 to 4 employees; intransportation and utilities, the limit was fewer than 10 employ-ees; and in manufacturing, fewer than 5 employees.

7Administrative records could not be used to gather economicdata on wholesale trade because the data available for agents andbrokers generally reflected commission receipts rather than grossdollar volume of business conducted.

8These categories were (1) wholesale long and short question-naires; (2) retail, including classification forms; (3) services,including classification forms; (4) construction; utilities; finance,insurance, and real estate; manufacturing and minerals shortquestionnaires; and forms NC-9923, ‘‘General Schedule,’’ andNC-9926, ‘‘Classification Report;’’ (5) manufacturing and mineralslong questionnaires; (6) small and medium multiestablishmentpackages; (7) Puerto Rico, Guam, Virgin Islands of the UnitedStates, and Northern Mariana Islands; and (8) the 1997 VehicleInventory and Use Survey.



Barcode check-in and sorting. Census questionnairesand other materials began their processing with barcodecheck-in. Two 56-pocket laser reader/sorters read the bar-codes, slit open the return envelopes, and sorted the pack-ages. The sorter operator created a ‘‘header’’ record foreach batch by keying the sort number, batch number, andbatch status code from the cover sheet attached to eachbatch. The reader/sorter read the 22-digit barcode, aswell as the census file number (CFN—CFNs are discussedfurther in the ‘‘CFN research’’ section below), questionnaireform number, codes for follow-up letters and correspon-dence, and then sorted the batch by questionnaire type.

Machine rejects—i.e., receipts with unreadablebarcodes—were sorted to ‘‘pocket 1’’ of the machines,removed from the working batch, and rebatched withsimilar receipts. Clerks used hand laser-wand readers toread barcodes. If the barcode could not be read by wand,or if the barcode label was missing, the clerk keyed thebarcode identification number directly to the check-in file.The check-in actions from the laser sorters and the wandand keyboard operators were validated on a data-entrysystem in the NPC processing office, then were transmit-ted to the Economic Statistical Methods and ProgrammingDivision (ESMPD) at the Census Bureau’s headquarters inorder to update the SSEL.

Disposition of mail receipts. Following check-in, mailpackages and individual questionnaires went to theRemove Contents and Sort Unit. Staff opened the enve-lopes and sorted the questionnaires and other materialsinto categories as follows:

• Correspondence sent to the attention of a specificanalyst was referred to that analyst.

• Respondent-originated correspondence was referred tothe Correspondence Suspense or Reading Units forreview.

• Census Bureau-originated correspondence cases werereferred to the subject divisions.

• Classification forms were referred to the Screening Unitfor review and updating of the SSEL.

• Economic-only single-establishment company ‘‘2+’’cases (i.e., responses in which two or more single-establishment questionnaires were returned in a singleenvelope) were held for review by analysts).

• All other single-establishment company cases werebatched into work units and sent for microfilming. Com-pany questionnaires with attached correspondence, andmultiestablishment company 2+ cases were sent to theDirectory Analyst Unit.

• All other multiestablishment company cases weregrouped into manila folders, batched into work units,and routed to the Completeness and Coverage Unit.Single establishment company questionnaires werebatched and referred for microfilming.

Correspondence Processing

Correspondence reading. Clerks checked each piece ofrespondent-originated correspondence for a CFN. Next,they referred to lists of problem descriptions and instruc-tions in order to determine what action should be taken.Correspondence with CFNs were read, categorized, andthen processed as follows:

• Congressional correspondence (i.e., correspondencefrom, copies of correspondence to, or letters with nota-tions to, any Federal government official ranking higherthan the Director of the Census Bureau). These werereferred to supervisors for special handling.

• Freedom of Information Act cases were referred toCensus Bureau headquarters for response.

• Correspondence sent to the attention of a specific per-son, title, office, or department was annotated appropri-ately and referred to the Directory Analyst Unit.

• Correspondence from multiestablishment companyrespondents who reported a change of name or addresswas annotated and referred to the Directory AnalystUnit.

• Correspondence from a multiestablishment companyrespondent who wanted to combine data for all estab-lishments on one questionnaire was annotated andreferred to Central Files after keying.

• Correspondence (1) from single-establishment respon-dents indicating they had more than one establishment,(2) from respondents requesting help completing thequestionnaire, and (3) for other problems that did not fitany of the problem descriptions provided, were anno-tated and referred to the Directory Analyst Unit.

Clerks annotated correspondence with a five-digit codeto indicate actions taken and the unit to which the casewas to be referred. Then, using an interactive routine(described below) for each case, they keyed in such infor-mation as correspondence category code, CFN, timeextension codes, date the respondent’s business wasclosed or sold (if contained in the correspondence), as wellas changes of company ame and/or address.

CFN research. Each correspondence case needed a validCFN; clerks had either to identify a CFN for each case or, ifnone was available, assign one. When a valid CFN wasfound, the processing clerk wrote the CFN and check digiton the correspondence and continued processing—usingthe interactive systems. When a CFN was not located, butthe correspondence included a nine-digit EIN, the clerkadded a zero (‘‘0’’) to the front of the EIN and assigned thisten-digit number as the CFN to the file. When no EIN waspresent, the clerk entered the name and all other identifi-cation information, then annotated and referred the corre-spondence as necessary.



Correspondence with single-establishment ques-tionnaire processing. Correspondence Unit clerks readand categorized correspondence that was attached to asingle-establishment questionnaire and transcribed theCFN from the questionnaire to the letter or other materialsattached. The clerks used problem listings, selected theone that most closely described the subject of the corre-spondence, and carried out the actions indicated. Theyannotated the correspondence with the appropriate actioncodes, circled the CFN and correspondence categorycodes, and transcribed any information given by the corre-spondent to the appropriate item.

Census Bureau-originated correspondence. CensusBureau-originated correspondence was generated fromthree—sources (1) interactive applications run against theSSEL file, (2) records referred by the control file match, and(3) records flagged during complex edit (see below) thatrequired contact with the respondent to resolve a prob-lem. Census Bureau-originated correspondence processingwas handled by the various processing units generatingthe requirements. Interactive computer routines wereused to identify particular problems with individualrecords and the kind of correspondence required to con-tact respondents for corrective or additional information.Clerks keyed the CFN for the specific record, used corre-spondence reference listings to select a correspondencecategory that described the problem, and keyed thespecified category code to the record. The interactive rou-tine indicated whether (1) the record belonged to a single-establishment or a multiestablishment company, (2) if thelatter, whether the master company or a component estab-lishment was involved; and (3) which form letter(s) and/orother materials should be used.

The ESMPD generated correspondence files twice dailyduring questionnaire processing, and the NPC used theDocuPrint equipment to print the materials for the mailoutpackages. Through computerized printing, printed materi-als were sorted into three groups:

• Cases for immediate mailing were routed to the Corre-spondence Mailing Unit for assembly and mailout.

• Packages requiring annotation went to the subject-areaproblem solving units.

• Special cases were referred to the subject analysts forthe industry.

Package assembly and mailout operations. Corre-spondence Mailout Unit clerks used correspondence cat-egory code lists to determine actions needed for eachmailout package, then assembled and addressed the pack-ages to those who had originated the correspondence. Forexample, for a package with a correspondence categorycode of 102 (time extension granted—accountants), theclerk typed the correspondent’s address on the form letterand added the extension date granted to the first line of

the letter. Typically, the correspondence mailing packagescontained one or more of the following items: (1) letter;(2) questionnaire(s) as needed; (3) miscellaneous bro-chures, listings, instructions, flyers, address sheets, orother enclosures ; (4) return envelope; and (5) inventorylist (showing the contents of multiestablishment mailingpackages, if any).

Standard items for correspondence mailout wereprinted using NPC’s DocuPrint equipment. Two computer-generated lists, a Correspondence Reference Listing and aProduction Control Listing, were used for controlling eachrunning of the correspondence printing program. The ref-erence listing identified cases for referral to the appropri-ate problem-solving units for resolution or research. Thecontrol listing was used to manage the work and identifyprinter failures for manual correction.

Multiestablishment Completeness andCoverage Processing

Following check-in, all multiestablishment companyquestionnaires were batched into work units by companyemployment size (based on prior year employment sizefrom administrative records), and underwent complete-ness and coverage keying. Multiestablishment complete-ness and coverage processing verified coverage of allestablishments within a firm, and that additions, mergers,sales of plants or companies and other critical companyand establishment data were reported.

Accessing the SSEL file, which was password-protected,terminal operators took the following actions:

• Correcting company affiliation information, such as—

• Revising address information.

• Adding new or acquired establishments.

• Moving establishment sold to the acquiring company.

• Inserting codes to indicate closed or idle establish-ments.

• Adding new multiestablishment companies and theirestablishments to the file. When a new companyaddress was identified, the clerks updated the SSELand the appropriate questionnaire package wasremailed. If the clerks failed to find a new address,they referred the case to the coverage analysts forresolution.

• Researching undeliverable as addressed (UAA) cases forthe latest address for remailing questionnaires.

• Correcting unusual employment or payroll data relation-ships (as compared to administrative record tax data).

• Correcting establishment reports of the company inwhich the company combined all the data in one report.(This meant splitting the combined data across theestablishment reports based on prior year employmentand payroll relationships.)



Problems identified during keying were typicallyresolved by contacting the company for additional infor-mation, or by using the SSEL interactive routines to pro-cess affiliation changes or to break out establishment datafrom combined reports.

Once processed through the completeness and cover-age unit, the multiestablishment folders were routed tothe Microfilming Unit.

Microfilming the Questionnaires

The Microfilming Unit filmed all economic single- andmultiestablishment questionnaires—except classificationforms—just prior to data-entry keying. The equipmentimprinted an eight-digit serial number on eachquestionnaire—the first four digits identified the microfilmreel and the last four the microfilm frame, for each form.

The questionnaires were held in the Microfilming Unituntil the film had been processed and checked to assure itwas of acceptable quality. Once the film was reviewed andapproved, staff made copies of the reels and providedthem for referral and research purposes, to analysts inJeffersonville and to the subject divisions’ staffs at head-quarters. The Microfile Unit staff then forwarded the paperquestionnaires to be rebatched for data keying.9

Data Entry

Data keying. Data keying involved transcribing datafrom the economic census questionnaires to a machine-readable data file for edit and tabulation. The NPC datakeying staff used a key-to-disk interactive system thatcombined basic clerical review of the questionnaires withthe keying operation. Each work station had a keyboardand video monitor that enabled the keyer to display andedit keyed data, as well as displaying messages alertingthe keyer to unacceptable entries, and corresponding cor-rective procedures. The systems used 19 programs (onefor each of the principal NAICS sectors covered by theeconomic census). The programs assigned ‘‘screens’’ tospecific sections of each type of questionnaire for use inentering data from individual forms. The first severalscreens were used for keying identification information,while the bulk were used for entering data from the ques-tionnaires. The number of screens varied by trade area;for example, finance, insurance, and real estate question-naires data keying required 63, while 29 were needed forkeying data from the manufacturing questionnaires.

Questionnaires, grouped in batches of 35-50 (for longform manufacturing questionnaires), and 100 (for all othertrade areas) were received from the Microfilming Unit.

Classification forms were batched in quantities of 200.Each batch had a control sheet with the batch number(which included a prefix identifying the questionnairetype) and the number of questionnaires in the batch. Thedata entry programs were designed for straightforwarddata entry from the top down for each page and fromfront to back of the questionnaire. Each keyer began key-ing by entering the identification information—i.e., theCFN, form number, and the microfilm reel and framenumber for the questionnaire—and address informationfrom the address label area and leading sections of thequestionnaire. Thereafter, the keyer entered the two-digitsection identifier for each section, followed by the firstthree-digit item keycode10 , and proceeded through thequestionnaire, entering successive codes and dataresponses.

Keyers had instructions for handling common respon-dent errors, such as—

• Illegible entries. Illegible entries were referred to thekeying supervisor, who, if able to interpret theresponse, entered it for the keyer. If an entry could notbe resolved, it was marked ‘‘NIL’’ and the keyer contin-ued keying remaining items on the questionnaire.

• Bracketed entries. Keyers entered the data for thefirst keycode, followed by an ‘‘&,’’ then keyed an ‘‘&’’ foreach additional keycode included in the bracket. Whenan entire item, with or without the total line, was brack-eted, the response was entered on the ‘‘total’’ line forthe item.

• Altered stub description. When respondentschanged, added to, or crossed out a preprinted descrip-tion of a data item, the keyer entered the keycode orline code and the corresponding data followed by a‘‘dash’’ in the first data field containing an entry.

• Range entries. Keyers calculated the midpoint of therange reported and entered that number (rounding upto the next whole number in case of a fractional result).

As the keyers entered the data, the computer per-formed a series of edit checks to ensure that (1) the vari-ous codes keyed were valid, (2) entries were numeric, (3)all mandatory fields contained responses, and (4) the data

9Economic census microfilm typically is retained at the CensusBureau’s headquarters for a period of about 5 years (or until thenext census is completed). Thereafter, each economic subjectdivision’s staff decides either to keep the microfilm in its ownarchive, or send it to the Federal Records Center in Suitland, MD,for permanent storage.

10Keycodes were three-digit numbers preprinted in the datacells of the questionnaires to identify that cell. The data systemedits ensured that no keycode was keyed and accepted withoutcorresponding data. Line codes—three-to-nine digit numbers pre-printed to identify a row of data cells extending across two ormore columns—were keyed, preceding the data for that line. Mostitems requiring multiple responses, such as merchandise andcommodity lines, required this format. The only information fromthe questionnaires that did not usually require keycodes or linecodes was questionnaire identification and coverage information;address changes and new locations; person-to-contact, and tele-phone number; and basic data for sales, receipts, revenues,employment, payroll; and kind-of-business codes.



were within acceptable ranges. The display screens noti-fied keyers of specific problems detected by the computeredit and displayed suggested corrective action. After suc-cessful data entry, the completed batches of question-naires were sent to the Holding Unit, where they wereretained while the data records were processed throughcontrol file matching and computer editing.

Individual questionnaires could be rejected from pro-cessing for various reasons (i.e., invalid form number,invalid kind-of-business code, invalid or duplicate linecodes or keycodes, etc.). Keyers pulled these question-naires from their batches, placing them on top of each,and entered the number of rejects on the batch coversheet.

Quality assurance for data keying. The CensusBureau carried out quality assurance checks of data keyingto ensure that information from questionnaires was accu-rately recorded in the data file. This was done by selectinga systematic random sample of CFNs from each batch ofquestionnaires for verification.

The quality assurance staff rekeyed data from question-naires for selected records from each batch and matchedthe original keyers’ data sets with the verifiers’ sets. Differ-ences were displayed for evaluation. The proceduresdefined errors as either ‘‘keyer’’11 or ‘‘nonkeyer’’ errors.Keyer errors typically involved (1) miskeying that causederrors in the data fields, (2) field or document omission or(3) duplication, unnecessarily keyed fields, etc. Nonkeyererrors usually involved mechanical problems, or supervi-sor or verifier errors. All mistakes identified during theverification process were corrected and reverified beforethe data were released for computer processing.

Output files. Data entry transformed respondent infor-mation from a paper questionnaire into an electronic inputfile. Before these input files could be processed, theyneeded accurate translation into an output file of variablelength records. The input record structure reserved spacefor every possible input field; the output record structurepermitted only fields or repeating groups that containeddata.

Using variable length records required moving the inputdata exactly into sections or fields specified while elimi-nating the filler required to make all records the samelength. A single output record consisted of 11 sections; a‘‘section 1’’ was always present and contained indicatorsthat described the other 10 sections. Sections 2-4 con-tained address change, contact and telephone, and

‘‘splitter’’ information12, while sections 5-11 contained thekeycodes and data for specific responses. The ‘‘repeatinggroup length’’ in each section contained the applicablekeycodes with a specified number of characters for thedata. The repeating group could occur a variable numberof times, depending on the section and the type ofresponse. The section indicators were used to determineeach section’s length within the record, while the sum ofthe sections’ lengths determined the length of the fullrecord.

Once reformatted for output, the data files were readyfor control file matching.

Control File Match

Matching. After data entry, the data records for the cen-sus questionnaires were matched to the SSEL control file.For each company, the control file contained (1) adminis-trative record data, (2) historical census data, and (3) thegeocoded census address reference file containing thestate, county, place, and census tract numerical codes foreach establishment. The Census Bureau carried out thecontrol file match on a flow basis, as the data frombatches of questionnaires were entered into the censusdata file. In this way, the subject division staff could carryout edits as the data for their respective trade areas werekeyed.

The control file match compared the CFNs of freshlykeyed records to the SSEL file, then compared the identifi-cation and address information in each file to the SSEL file,as follows:

• Identified company structure and affiliation referrals forreview.

• Identified duplicate establishment records and potentialnew establishments for review.

• Identified address changes for use in geocoding for dataprocessing and tabulation.

• Compiled data from the questionnaires and the SSELinto a single file for analysis.

After the SSEL update was completed, the records weretransmitted to the subject divisions for complex data edit-ing. Staff also identified problems in addition to thoselisted above—e.g., blank records, records that included

11Each keyer in the Data Entry staff underwent a three-stageprogression from ‘‘training’’ status keyer, to ‘‘prequalification,’’and finally to ‘‘qualified’’ (Q) status. Keyers had to limit their errorrate to 1.5 percent to become qualified and thereafter had tomaintain a ‘‘reject’’ rate—the number of errors that would cause arecord to require rekeying—of less than three per week to remainin qualified status.

12A ‘‘splitter’’ was a firm that formerly operated at a singlelocation, under one employer EIN, but opened operations at morelocations under the same EIN. Splitter information was the detailthe respondent gave the Census Bureau about the new locations,such as name and address, kind-of-business description, receipts,annual payroll, and number of employees. As the splitter was pro-cessed, the firm was redesignated a multiestablishment companywith multiple locations. A single-establishment company did notbecome a splitter if it simply acquired another business operatingunder its own EIN; it did so only if it had activities at more thanone location.



out-of-scope activities— that required correction before arecord could be released for further processing, androuted these records for control file match referral.

Control file match referrals. Data records that failedthe control file match were withdrawn from the data filefor their batch. The computer generated a referral listingthat identified the record and the problem. Clerks handledreferrals for missing data or a blank record either byremailing questionnaires, by telephoning the company, orby inserting data derived from other information on thequestionnaire. Records with other kinds of industry-related problems were referred for review by analysts inthe trade area problem-solving units.

Poorly keyed records, or those with too many correc-tions or changes required, were recycled. The originalmicrofilm reel and frame numbers were taped over andthe questionnaire was sent back, re-microfilmed in aseparate recycle batch, and rebatched for data keying.

Questionnaires that belonged in a different trade areabatch were referred after the ‘‘sending’’ area analystentered a ‘‘bridge’’ code to identify the record. In the‘‘receiving’’ area, the analyst working on the recordassigned a correct new code, and the questionnaire waschecked in again using a laser wand. These wand actionswere processed against the corresponding held records,and the records were released to the appropriate subjectdivision. Cases requiring that the respondent be contactedwere sent to the Correspondence Unit for resolution.

DECENTRALIZED POST-COLLECTION PROCESSING

General Information

Post-collection processing consisted of editing, tabulat-ing, and analyzing of the 1997 Economic Census data.This followed the control file match and encompassed allthe activities required to prepare the data for publication.Unlike collection activities, which were centralized, post-collection activities were decentralized—each subject areawas processed independently.13 Analysts from the subjectdivisions (Manufacturing and Construction [MCD], ServiceSector Statistics Division [SSSD], and Company StatisticsDivision [CSD]), with assistance from analytical andclerical staff at the NPC, carried out the work in twophases—micro edit referral and macro analysis. Micro editreferral involved editing and cleaning data at the estab-lishment level—that is, the micro level through complexediting, as well as reject and referral processing. Macroanalysis took place during review of tabulated data beforepublication.

Micro Edit Referral Processing

Background information. Subject analysts within therespective areas of the Economic Directorate conductedmicro edit referral processing—the major steps of whichare described below and illustrated in Figure 8-1.Before editing began, ‘‘good’’ records from the control filematch were acquired. These were establishment recordsthat had passed computer checks for organizational struc-ture, coverage, and completeness.

Complex editing. As subject areas received records,they ran processes that edited the data. In the complexediting phase of processing, the Census Bureau used aseries of computer programs to subject establishmentrecords to a series of edit modules that had two majorpurposes—assigning valid kind-of-business or industrycodes and evaluating respondent data for consistency andreasonableness.

Assigning a valid kind-of-business or industry classifica-tion code depended on computer evaluation of theresponses to specific items on the census questionnaires.Examples of such items would be merchandise lines soldby a retail establishment, products manufactured by aplant, entries written in by the respondent explaining theestablishment’s activities, and self-designated check-boxclassifications. If all critical information was available, theedit module assigned the correct classification code. If any

13For the 1992 Economic Census, each of the subject areashad developed specifications for classification. There were 12altogether (3 for Manufacturing, 2 for Mineral Industries, 1 forConstruction, 5 for Services, and 1 for the Outlying Areas), eacharea of which had required about 100 pages of specifications. TheGeneralized KB/Industrial Code edit for the 1997 census reducedthe required specialized material to an average of just 3 pagesper area.

Figure 8-1Post-collection Processing:Micro Analytic Processing

Control File MatchDatabase

Get Release Records

Complex Edit

Load Data Bases

Subject AreaDatabases



of this information was missing, the edit module ‘‘flagged’’the record as a ‘‘reject,’’ and referred it for correctionbefore allowing further processing.

For the 1997 Economic Census, the computer alsoassigned a ‘‘bridge’’ code that enabled the Census Bureauto tabulate the record on both the old SIC basis and thenew NAICS.14 After classification codes were assigned, thecomputer edit program verified each record to validate thekind-of-business, industry, geographic, and ZIP Codes.

The complex edit also evaluated response data for rea-sonableness and consistency. Computer routines checkedthat employment data reported on the questionnaireswere consistent with payroll or sales/receipts data,as well as response data from prior censuses. For the1997 census, the Census Bureau introduced the use of‘‘generalized’’ edit modules for these checks. The modulesbalanced items and conducted ratio checks:15

• A balancing module allowed for comparing detail itemsto totals and nested tests of subtotals to broader levels,and for two-dimensional tests to compare additive rowsto corresponding column information.

• The ratio module compared data for individualestablishments—using items from the questionnaires,administrative sources, or historically reportedinformation—to industry parameters derived from datacorrelations in past censuses or through administrativesources. Imputation options were used to estimate non-critical missing data or to replace extreme outlier data.

Subject division staff particularized these complex editmodules to account for industry differences.

Subject area databases. After the establishmentrecords underwent complex editing, computer programsloaded them into the appropriate subject-area micro data-bases, shown in Figure 8-1. Each of these databases con-tained responses from establishment questionnaires. Thedatabases also contained the edited data and flags (seebelow), administrative data from the SSEL, questionnairewrite-ins, and historic 1992 census data (if available) foreach establishment.

Reject and referral processing. The complex editidentified, and flagged for referral, establishment recordsthat had responses or other characteristics outside speci-fied tolerances. The edit also rejected records with (1)

critical data missing, (2) obvious errors in classification, or(3) other significant problems that could not be correctedwithin the parameters of the edit or imputation routines.

Such establishments were handled with either of twoprocedures—post-edit correspondence or manual problemresolution:

• For post-edit correspondence, the Census Bureau mailedtailored requests to respondents to ask for additionalinformation about the data. For example, respondentsreporting broad product lines were mailed lettersrequesting the detailed product lines.

• Cases referred for manual resolution went to subjectanalysts or to clerks. These cases typically involvedrespondent write-in entries for kind-of-business descrip-tions, merchandise lines, or products. Clerks codedwrite-ins to the appropriate bridge code, type ofoperation/construction, or product code.

Both analysts and clerks used video-display terminals toreview the rejects and referrals and followed guidelinesand procedures for evaluating and correcting problems.The interactive editing systems permitted them to correcterrors and reedit the data as required.

Macro Analytical Processing

Background information. The objective of macro ana-lytical processing was both to tabulate data and to reviewthose tabulations to prepare them for data publication.These activities began at the end of the data collectioncycle.

Macro analysis (Figure 8-2.) began with the tabulationof establishment records housed in the subject area data-bases. Micro records were first tabulated as part of themain publication series for the census. Then, they weretabulated a second time for the cross-sector publicationsthat showcase the introduction of NAICS.

Data tabulation. ‘‘Tabable’’ establishment records werethose that passed the complex edits and were not rejectedor flagged for referral. The object of the reject and referralprocessing was to resolve problems so that the establish-ment record would be tabable. It was not unusual, giventhe scope of this activity, for many establishment recordsto remain unfinished at the time macro analysis began.Therefore, at the macro analysis stage, corrections weremade only for the establishment records that were themost problematic. After the collection cycle was com-pleted the data frequently were retabulated to incorporatecorrections to micro records made during the macroreview stage.

Subject tabulation files for main publication series.After tabulation, the totals were loaded into computerfiles, which analysts accessed via macro analytical tools.The Census Bureau tabulated the micro establishmentrecords many different ways—based on data product and

14The Census Bureau assigned two extra digits to the classifi-cation code for each part of the 1987 SIC that had been reclassi-fied to a different NAICS sector or industry. This allowed previouscensus data, collected and tabulated using the SIC, to be com-pared with the 1997 census tabulations using NAICS.

15See Richard S. Sigman, ‘‘Development of a ‘Plain Vanilla’ Sys-tem for Editing Economic Census Data,’’ a working paper pre-sented at the October 1997 United Nations Work Session on Sta-tistical Data Editing.



analytical needs. Each subject area had its own set oftabulation files for its industries. For example, subjecttabulation files included summed data by industry, geo-graphic area, firm size, products produced, materialsused, and merchandise lines sold.

Corporate repository. Since each subject area had itsown tabulation files, the Census Bureau introduced theuse of a corporate repository to allow sharing of summarydata across the areas. Each week, the establishment micro

Publication

Figure 8-2Post-collection Processing:Macro Analytic Processing

Subject AreaDatabases

Tabulation

Microdata fileextraction

Microdata set merge

CorporateRepository

Subject Tabulation Files forMain Tabulation Files

1997 NAICSMicrodata set

1992 NAICSMicrodata set

Tabulation

Tabulation Files forCross-Sector Publication

Macro Analysis Analytical Tools-Subject Table Review Systems-Problem Identification-Multidimensional Reports-Online tools



records were fed from the subject databases into a corpo-rate repository. This allowed analysts in manufacturing,for example, to have access to view data for establish-ments in retail industries.

In feeding records from subject areas to the repository,only a subset of data items were compiled for eachestablishment—e.g., sales/receipts/value of shipments,payroll, employment, operating expenses, and wholesaleinventories. The resulting file was called the ‘‘1997 NAICSmicrodata set.’’ Once this dataset was created, it was com-bined with the 1992 micro dataset that contained alltabable records from the 1992 Economic Census. Theresulting current and historic dataset enabled the CensusBureau to create special cross-sector tabulations andreports that allowed analysts to compare 1997 and 1992data on a NAICS, as well as on the old SIC basis. The cor-porate repository was a SAS dataset.

Analytical Tools

The subject divisions prepared detailed analyticalreview guidelines to instruct analysts on what to look forand on how to use the available tools. Typical reviewpoints covered analyzing large differences; assessingchanges to the tabulation cell composition due to organi-zational changes (establishments bought, sold, or new);nonresponse; and errors in historic 1992 data. Analystsdocumented their findings for tabulated cells confirmed tobe correct. The basic analytical systems used were asfollows—

• Subject table review systems. As with micro editreferral processing, macro analysis had speciallydesigned interactive systems that permitted analysts toview the subject tabulations in table formats. Analystsfor each of the subject matter areas used interactivetable review systems to analyze tabulated cells on aNAICS basis. As analysts corrected micro data throughthe establishment review and correction system, theyused the table review systems to determine the overalleffect of these corrections on the data.

These review systems also were critical to the disclo-sure analysis process. Disclosure analysis preserves theconfidentiality of reported information so that no indi-vidual firm’s operations or identity could be inferredfrom the published data. To identify possible disclo-sures, the Census Bureau established consistent rulesfor application across subject areas. Often when thetabulations resulted in disclosure, it was necessary tosuppress not only the primary (originally) disclosed cell,but additional cells so that the original disclosure couldnot be obtained through subtraction. Suppressing thesesecondary tabulated cells is called complementary dis-closure analysis, and analysts used the table review sys-tems to review and interactively set the complementarysuppression symbols.

• Problem identification tools. To review thousands oftabulated data cells, analysts needed mechanisms tohelp pinpoint specific problem cells. There were twomain methods of identifying cells whose data appearedquestionable—reconciliation and outlier detection. Forreconciliation, analysts compared cell totals to non-census data drawn from the Census Bureau’s currentprograms as well as from other government statisticalagencies. For outlier detection, for the 1997 census,analysts used a new automated graphical data reviewtechnique. A SAS product called ‘‘SAS Insight’’ enabledanalysts to identify suspicious cells by providing visualdisplays showing cells outside the normal distributions.The graphical analysis was used to identify extreme cellchanges for historic (e.g., historic census employmentcompared to current census employment) as well as cur-rent (e.g., current census annual payroll compared tocurrent census employment) ratios. Subject matter areaanalysts also used manual techniques to identify celloutliers. Generally, analysts used predetermined valuesfor their historic and current ratios. For example, sincethe manufacturing sector had prior year comparisonpoints with the Annual Survey of Manufactures, analystssearched ‘‘census to prior year’’ totals exceeding fivepercent for basic statistics items and 10 percent forproducts.

• Online tools. The principal online technique used byanalysts in their work with the tabulated cells and microestablishment records was the capability to search thecomputerized data files. Analysts developed ‘‘cannedsearches’’—i.e., programmed searches that could beexecuted interactively for typical inquiries about thedata. The canned searches selected records based onspecified size codes, but also could be modified forunexpected situations that surfaced during data review.

Specially designed computer interfaces enabled ana-lysts to do special ad hoc tabulations. For example,using the corporate repository, analysts could selectand tabulate establishment records associated with spe-cific companies—even if those establishments crossedsubject areas. The analysts also could use the subjectmicro databases to select and tabulate records by clas-sification code (SIC, NAICS, or bridge), geography, andtype of operation in formats that varied from thosespecified for the publication tables.

• Multidimensional reports. Analysts used ‘‘multidi-mensional reports’’ with the corporate repository toevaluate the new NAICS classification system. Thereports showed SIC industries based on their NAICScomponents and, vice versa, as well as providing ana-lysts with comparative SIC data showing the basic datafrom 1997 to 1992 at the broader levels.



Chapter 9. THE 1997 ECONOMIC CENSUS IN THE OUTLYING AREAS

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

Company Organization Survey . . . . . . . . . . . . . . . . . . . . . . . . . 88

Puerto Rico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88Preparatory Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88Printing questionnaires. . . . . . . . . . . . . . . . . . . . . . . . . . . 88Scope and coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89The mail enumeration . . . . . . . . . . . . . . . . . . . . . . . . . . . 89Mail follow-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90Telephone follow-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

The Virgin Islands of the United States. . . . . . . . . . . . . . . . . . . . . 90General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90Preparatory Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90Printing questionnaires. . . . . . . . . . . . . . . . . . . . . . . . . . . 90Scope and coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90Mail enumeration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90Mail and telephone follow-up . . . . . . . . . . . . . . . . . . . . . . . 91

Guam and the Commonwealth of the Northern Mariana Islands (CMNI) . . . 91General Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Preparatory Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Printing questionnaires. . . . . . . . . . . . . . . . . . . . . . . . . . . 91Scope and coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92Mail enumeration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92Mail and telephone follow-up . . . . . . . . . . . . . . . . . . . . . . . 92

Data Processing for Outlying Areas Questionnaires . . . . . . . . . . . . . . 92Receipt and Check-in . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92Data Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92Editing and Tabulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

CONTENTS


The 1997 Economic Census in the Outlying Areas

INTRODUCTION

The Census Bureau carried out the 1997 Economic Cen-sus in the Outlying Areas under the provisions of Title 13,United States Code, Section 191. This required that thequinquennial economic census cover the Virgin Islands ofthe United States, Guam, the Commonwealth of the North-ern Mariana Islands (CNMI), and the Commonwealth ofPuerto Rico. Although there were similarities in the censusin these areas, there were sufficient differences in theplanning, preparations, and data collection activities foreach area to warrant a separate discussion. The CensusBureau’s Company Statistics Division (CSD)1 was respon-sible for the planning, management, and coordination ofthe census in the Outlying Areas. Data-processing activi-ties took place at the three locations: the Census Bureau’sNational Processing Center (NPC) office in Jeffersonville,IN2, at Census Bureau headquarters in Suitland, MD, and ineach outlying area.

While the 1997 Economic Census for the 50 states wasconducted by applying the new North American IndustryClassification System (NAICS), the governments of PuertoRico, Guam, the Virgin Islands, and the CMNI requestedand received a postponement of the implementation ofthe new system until the 2002 Economic Census. TheCensus Bureau and the respective governments agreed tothis measure to ease the impact on local surveys andnational accounts of the change to the new classificationsystem. Adopting NAICS in the 2002 census would allowtime for the local agencies in the Outlying Areas to con-sider new approaches to their own surveys and to convertthem to NAICS. Therefore, the 1997 Economic Censusquestionnaires for all the Outlying Areas continued to usethe Standard Industrial Classification (SIC) code for datacollection.

COMPANY ORGANIZATION SURVEY

The Company Organization Survey (COS) was con-ducted in the Outlying Areas to update the Standard Statis-tical Establishment List (SSEL). The SSEL served as theprincipal source of business establishment addresses forthe economic census mailing list. The last COS prior to the1997 census was carried out in December 1996.

The Census Bureau mailed form NC-9901, ‘‘Report ofOrganization,’’ questionnaires to multiestablishment com-panies in the Outlying Areas. In Puerto Rico the COSincluded all multiestablishment companies with 50 ormore employees, as well as a sample of those with 50 orfewer employees. The survey covered all multiestablish-ment companies in Guam, the Virgin Islands, and theCommonwealth of the Northern Mariana Islands.

PUERTO RICO

General Information

The Census Bureau and the Puerto Rico PlanningBoard signed the Memorandum of Agreement to conductthe 1997 Economic Census by mail in Puerto Rico onOctober 30, 1997. The agreement stipulated that theCensus Bureau design and translate the questionnairesinto Spanish, plan and conduct the mail and follow-upactivities, process and tabulate the data, and prepare thecensus publications. The Puerto Rico Planning Board wasresponsible for making questionnaire content recommen-dations. In addition, the Board accepted responsibility forpublicizing the census through newspapers, television,and radio, and for designating February 1998 as‘‘Economic Census Month.’’

Preparatory Work

Printing questionnaires. The Forms and Mail Manage-ment Branch of the Census Bureau’s Administrative andCustomer Services Division (ACSD) designed question-naires and letters for the census. The Census Bureauprinted draft versions of the multiestablishment question-naires using the DocuPrint 4890 Highlight Laser PrintingSystem3 at Census Bureau headquarters. Once reviewedand approved for production, drafts of the multiestablish-ment questionnaires were electronically transferred to theDocuPrint database at NPC for production printing. At thispoint, NPC printed the English version of the multiestab-lishment company questionnaire, OA-9820, and theinstruction sheet, OA-9820(I).

For single-establishment companies, the Census Bureaucontracted with a private printer on September 27, 1997.The private printer produced the Spanish and Englishquestionnaires and instruction sheet (OA-9819(I)), cover

1CSD acquired its name in June 1998, see Chapter 2, ‘‘ThePlayers and the Plans’’ for the history of how CSD was formed.

2Until May 1, 1998, the Jeffersonville office was called DataPreparation Division (DPD).

3The Docuprint 4890 system is described in Chapter 7, ‘‘Prepa-rations for Census Mailout.’’



letters (EC-PR-L1 and EC-PRL1(MU4)), follow-up letters forsingle-establishment companies (EC-PR-L11, EC-PR-L21,ECPR-T135, and EC-PR-L31), and envelopes for the Outly-ing Areas.

Scope and coverage. The 1997 Economic Census inPuerto Rico covered construction industries; manufactur-ing; wholesale trade; retail trade; selected serviceindustries; and hotel, motel, and lodging industries. Thequestionnaire for Puerto Rico business establishmentsasked for basic information, as follows:

• Employer identification number corrections.

• Operational status (months of active operation during1997, in operation, temporarily or seasonally inactive,ceased operation, sold or leased to another operator,name and address of other operator).

• Legal form of organization (i.e., individual proprietor-ship, partnership, cooperative association, public orprivate corporation, government, other).

• Dollar volume of business.

• Payroll.

• Principal type of business activity (i.e., wholesale; retail;services; hotels, motels, and other lodging places;manufactures; and construction).

Wholesale establishments were asked to provide thefollowing additional information—

• Type of operation (merchant, sales branch, broker).

• Employment by principal activity (selling, support,administrative, manufacturing).

• Class of customer (builders, exporters, households,etc.).

• Operating expenses.

• Inventories.

• Storage space.

• Commodity sales.

Retail establishments were asked to provide the follow-ing additional information unique to retailing—

• Under-roof floor space.

• Franchise holders.

• Shopping Center.

• Class of customers.

• Merchandise sales.

Service establishments, in turn, also were asked for —

• Merchandise sales.

• Shopping Center.

Hotels, Motels, and Other Lodging Places were asked toprovide—

• Source of receipts (rooms, meals, packaged liquor,merchandise, etc.).

• Number and type of accommodations.

Manufacturing establishments were asked to provide—

• Shipments and other receipts.

• Employment and hours worked by quarter.

• Selected costs (electricity, fuels, materials, contractwork, etc.).

• Inventories.

• Capital expenditures.

• Value of products made and contract work performedby customer location.

Construction establishments were asked to provide—

• Type of construction/activity.

• Dollar value of business (new, maintenance, develop-ment, etc.).

• Employment and hours worked, quarterly.

• Construction work subcontracted out.

• Selected costs (electricity, fuels, communication, repairs,materials, etc.).

• Work for other contractors or builders.

• Ownership of construction projects.

• Assets, capital expenditures, and depreciation.

Data Collection

The mail enumeration. The 1997 Economic Census inPuerto Rico was primarily a mailout operation conductedin concert with the census mailing for the 50 states. TheCensus Bureau did not open an economic census fieldoffice, or train a temporary staff to enumerate and con-duct follow-up visits to nonrespondent establishments.Instead, the NPC and the Census Bureau’s Tucson, AZ,Telephone Operations Center conducted a telephonefollow-up operation of nonrespondents. At the request ofthe Economic Directorate, two establishments that weredeemed essential to the accuracy of the census received afollow-up visit from a staff member assigned to the Cen-sus Bureau’s Census 2000 office in San Juan.

4Multiestablishment companies.5EC-PR-T13 (Title 13, U.S. Code) stated the legal authorization

for the Census Bureau to conduct the economic census in PuertoRico.



Mailout packages consisted of the appropriate question-naire or questionnaires, cover letter, instruction sheet, anda return envelope. These packages were printed andassembled by the aforementioned private contractor, andwere delivered to the NPC in Jeffersonville during thelast 2 weeks of November 1997. The aforementionedDocuPrint system was used for the multiestablishmentcompany mailout.6

The initial mailout consisted of 26,112 single-establishment and 4,959 multiestablishment packages.Mailout to both single- and multiestablishment companieswas completed by the second week of December 1997.Recipients were asked to return their completed question-naires to the NPC by February 12, 1998.

Mail follow-up. Following the initial mailout, nonrespon-dent single-establishment companies were mailedfollow-up letters on a flow basis:

• 19,663 form EC-PR-L11, mailed February 24, 1998 toMarch 6, 1998.

• 16,186 form EC-PR-L21 (with EC-PR-T13 attached),mailed March 31, 1998 to April 9, 1998.

• 12,617 form EC-PR-L31 (with EC-PR-T13 attached),mailed May 7, 1998 to May 18, 1998.

Multiestablishment companies in the 50 states andPuerto Rico were mailed the same follow-up letters. Thesemailouts began in February 1998, and were completed inJune 1998.

Telephone follow-up. The telephone follow-up targetednonrespondent establishments with the intention of com-pleting a census questionnaire by phone. Staffs at the NPCand in Tucson, AZ, called approximately 10,000 delin-quent establishments.

If the information necessary to complete the question-naire by phone was unavailable, the interviewer scheduleda call-back to complete enumerating the establishment. Ifthis was not workable, extensions for response in 2-weekincrements were granted.

If neither follow-up mailings nor phone inquirieselicited a response, the Census Bureau determinedwhether a personal enumeration would be required. Thisdecision was based on the size of the company (payrolland employees) and the effect that the loss of the compa-ny’s data would have on accuracy of the economic census.

THE VIRGIN ISLANDS OF THE UNITED STATES

General Information

After consultation with the Census Bureau, the VirginIslands’ Department of Labor signed a Memorandum ofAgreement to conduct the 1997 Economic Census on

November 21, 1997. This agreement would remain ineffect until the completion of the 1997 Economic Censusor until termination by mutual agreement.

The Census Bureau was responsible for designing andpreparing the census questionnaires and for processing,tabulating, and publishing the data.

Preparatory Work

Printing questionnaires. As with the conduct of thecensus in Puerto Rico, the Census Bureau used the Docu-Print systems at the NPC to print the questionnaires, coverletters, and follow-up letters for multiestablishment com-panies in the Virgin Islands. A private contractor, whosecontract was awarded on September 23, 1997, printedand assembled the questionnaires and other mailout mate-rials for single-establishment companies, and deliveredthe packages to the NPC in November 1997.

Scope and coverage. The 1997 census covered estab-lishments in construction and manufacturing industries,as well as in retail trade, wholesale trade, and selectedservice industries. Addresses for employer businesses inthe Virgin Islands were obtained from the SSEL.

Business establishments located on these islands wereasked to provide—



• Operational status (in operation, temporarily or season-ally inactive).

• Legal form of organization (individual proprietorship,partnership, corporation, government, other).


• 1997 payroll before deductions.


• Kind of business or activity.

• Class of customer.

• Sources of sales or receipts.

• Purchases from other businesses.

• Sources of receipts for hotels, motels, and other lodgingplaces.

• Number and type of accommodations.

• Ownership, control, and locations of operation.

Data Collection

Mail enumeration. As was the case for Puerto Rico, thecensus mailout for the Virgin Islands was carried outsimultaneously with the initial economic census mailingfor the 50 states. During the second half of December

6See Chapter 7, ‘‘Preparations for Census Mailout,’’ for a moredetailed discussion.



1997, 332 multiestablishment company and 1,652 single-establishment company economic census packages weremailed by the NPC. These packages included a cover letter,questionnaire OA-9873, and instruction sheet OA-9853(I).Response was requested by February 12, 1998.

Mail and telephone follow-up. After the initial mailout,nonrespondent single- and multiestablishment companieswere mailed follow-up letters on a flow basis. Table 9.1shows the dates and counts of the follow-up mailing.

Table 9-1.Follow-Up Mailings for the Virgin Islands of theUnited States

Letter Date of mailout Total

EC-VI-L11 . . . . . . . 02/24/98 to 03/06/98 940EC-VI-L21 . . . . . . . 03/31/98 to 04/09/98 866EC-VI-L31 . . . . . . . 05/07/98 to 05/18/98 353

After February 12, 1998, telephone operators in Jeffer-sonville, IN, called approximately 500 delinquent estab-lishments to complete the questionnaires by phone. Per-sonal visits to nonrespondent establishments were notrequired in the Virgin Islands.

GUAM AND THE COMMONWEALTH OF THENORTHERN MARIANA ISLANDS (CMNI)

General Information

The Census Bureau and the Guam Department of Com-merce signed a Memorandum of Agreement for the 1997Economic Census on October 10, 1997. The CensusBureau and the Central Statistics Division, CNMI, signed aMemorandum of Agreement on October 19, 1997.

The Census Bureau provided $16,000 to the govern-ment of Guam and $15,000 to the government of theCNMI to conduct the census in their respective areas.7 Aportion of this funding was earmarked for the trainingand salaries of locally hired individuals who performedtelephone follow-up interviews, and in some instances,personal visits to nonrespondent establishments. If expen-ditures exceeded the amount agreed upon in the Memo-randa of Agreement, the expenditure was renegotiated. Inaddition to funding, the Census Bureau also was respon-sible for (1) designing and preparing the questionnaires,(2) conducting the initial mailout, and (3) follow-up ofnonrespondents, processing, tabulating, and publishingthe data. The governments of Guam and the CNMI wereresponsible for recruiting, hiring, and training staff mem-bers, conducting telephone follow-ups, and publicizingthe census.

Preparatory Work

Printing questionnaires. The Census Bureau used theDocuPrint systems in Jeffersonville, IN, to print the formOA-9863 questionnaires (Guam), the form OA-9883 ques-tionnaires (CNMI), cover letters, instructions (OA-9853(I)),and follow-up letters for multiestablishment businesses.The following questionnaires and accompanying materialsfor single-establishment companies were printed by con-tract printers:

• For Guam, EC-GU-L1, EC-GU-L11, EC-GU-L21, EC-GU-L31.

• For the CNMI, EC-NM-L1, EC-NM-L11, EC-NM-L21, andEC-NM-L31.

The contractor delivered the assembled single-establishment company packages to the NPC at the end ofNovember 1997.

Scope and coverage. The 1997 Economic Census inGuam and the CNMI covered construction and manufactur-ing industries, retail trade, wholesale trade, and serviceindustries. Addresses of employer businesses and indus-tries in Guam and in the CNMI were obtained from theSSEL.

Questionnaires OA-9863 (Guam) and OA-9883 (CNMI)collected information on the following:

• Location of establishment.


• Operational status (in operation, temporarily or season-ally inactive, ceased operation, sold or leased to anotheroperator).

• Legal form of organization (individual proprietorship,partnership, corporation, government).


• Payroll before deductions in 1997.


• Kind of business or activity.

• Class of customer.

• Sources of sales or receipts.

• Ownership, control, and locations of operation.

Businesses in the CNMI were asked the following ques-tions that did not appear on the questionnaires mailed toestablishments in Guam:

• Citizenship status (U.S.-Commonwealth of the NorthernMariana Islands born, U.S. Other, Japan, Philippines,Korea, Other).

• Class of business.

7Prior to the 1997 census, data were collected from establish-ments in Guam and the Northern Mariana Islands by personalinterviews. The 1997 census was primarily conducted by mail.



Businesses in Guam were asked the following questionsthat did not appear on the questionnaire mailed to estab-lishments in the CNMI:

• Purchases from other businesses.

• Sources of receipts for hotels, motels, and accommoda-tions.

Data Collection

Mail enumeration. Census participants received a coverletter, the appropriate questionnaire or questionnaires,instruction sheet, and a postage-paid return envelope.Mailout of the questionnaires to single- and multiestablish-ment companies was completed by the second half ofDecember 1997 (see Table 9-2).

Table 9-2.Initial Mailout in Guam and the Commonwealth of theNorthern Mariana Islands

Geographic areaMulti-

establishmentcompanies

Single-establishment

companies

Guam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292 2,276Commonwealth of the NorthernMariana Islands . . . . . . . . . . . . . . . . . . . 178 908

Mail and telephone follow-up. After the initial mailout,single- and multiestablishment companies in Guam andthe CNMI were mailed follow-up letters on a flow basisbetween February and May 1998 (see Table 9-3).

Table 9-3.Mailout Follow-Up Guam and the Commonwealth ofthe Northern Mariana Islands

Geographic area Letter Date of mailout Total

Guam . . . . . . . . . . . EC-GU-L11 02/24/98 to 03/06/98 1,412EC-GU-L21 03/31/98 to 04/09/98 1,255EC-GU-L31 05/07/98 to 05/18/98 692

Commonwealthof the NorthernMariana Islands . . EC-NM-L11 02/24/98 to 03/06/98 561

EC-NM-L21 03/031/98 to 04/09/98 505EC-NM-L31 05/07/98 to 05/18/98 322

The Census Bureau contracted with the local govern-ments of Guam and the CNMI to conduct telephone follow-ups. Local government employees attempted to completea questionnaire through a phone interview for approxi-mately 500 nonrespondent establishments in Guam and300 nonrespondent establishments in the CNMI. Althoughtelephone follow-ups were successful in completing ques-tionnaires for the majority of nonrespondent establish-ments, a very limited number of personal follow-up visitswere conducted by employees of the respective local gov-ernments.

DATA PROCESSING FOR OUTLYING AREASQUESTIONNAIRES

Receipt and Check-in

Mail receipts from single-establishment companies werechecked in using optical laser scanning equipment thatidentified each case by the bar code on the mailing label.Multiestablishment company mail receipts were openedand the individual questionnaires placed in company fold-ers, then checked in using hand-held laser wand readers.

Questionnaires from the Outlying Areas were checkedin at the NPC. The CNMI questionnaires were separatedout at the NPC and returned to the local government officein Saipan, CNMI for data keying.

Questionnaires completed by phone interviews werechecked in using assigned identification numbers thatwere keyed directly into the database. Clerks then openedall questionnaire packages (except ‘‘Undeliverable asAddressed’’ [UAA] cases) and sorted them into the follow-ing categories:

• Single-establishment company receipts.

• ‘‘2 + packages’’ (envelopes containing two or moresingle-establishment company questionnaires).

• Multiestablishment company receipts.

• Correspondence (unattached and attached).

Two bilingual NPC staff interpreted questionnaires fromPuerto Rico. After all review work was completed, ques-tionnaires were released for batching, microfilming, anddata keying.

Data Entry

Questionnaires returned by respondents from PuertoRico were batched for data entry at the NPC. The CNMIquestionnaires were keyed at the census field office inSaipan, CNMI. The resulting electronic data file and origi-nal questionnaires from CNMI establishments were sent toCensus Bureau headquarters for editing and tabulation.Questionnaires from Guam and the Virgin Islands werekeyed at Census Bureau headquarters.

Editing and Tabulation

Economic census data for Puerto Rico were processedusing a generalized edit system developed for processingdata from stateside establishments within the servicessector. (This is described in Chapter 8, ‘‘Data Collectionand Processing.’’) The system consisted of a series of gen-eralized edit modules that could be used individually or incombination for editing data, and also could be custom-ized for more efficient editing of data for specific subjectareas. Subject division staffs wrote specifications for thecomplex generalized edits and tabulations. The Economic



Statistical Methods and Programming Division (ESMPD) ranproduction activities on their computer platforms, andstaff in the Economic Census Branch of CSD resolved editproblems.

The editing process assigned valid kind-of-businesscodes to the individual data records, and evaluated theresponses contained in the data records for consistencyand reasonableness. The Census Bureau’s staff applied aversion of the generalized kind-of-business/industry codeedit. This edit, customized for the Outlying Areas, checkedand/or corrected the kind-of-business codes in the indi-vidual establishment records. If a relationship was discov-ered to be outside specified limits, the item was ‘‘flagged’’by the computer for adjustment or review by subjectanalysts.

Data from Guam, the Virgin Islands, and the CNMI wereedited and tabulated using the Integrated MicrocomputerProcessing System (IMPS) at Census Bureau headquarters.Although the IMPS tabulated and edited data following thesame principles as those used for Puerto Rico, the pro-cessing system was more flexible, adjusting for the rela-tively small number of Outlying Areas questionnaires.

Once editing was complete, the data were ready fortabulation. After subject specialists reviewed the files toensure respondent confidentiality had been maintained(making it impossible for an establishment to be identifiedby the data), the statistical tabulations were ready forpublication.

PUBLICATIONS

Publications for the 1997 Economic Census in theOutlying Areas contained data on construction, manufac-turing, retail trade, wholesale trade, and service establish-ments. Data for the Outlying Areas were published in bothprinted and electronic formats.

Economic data for Puerto Rico were published in fourreports:

• Geographic Area Statistics (OA97E-1).

• Subject Statistics (OA97E-2).

• Construction (OA97E-3).

• Manufacturing (OA97E-4).

Economic data for the Virgin Islands, Guam, and theCNMI were published in individual printed reports(OA97E-5, OA97E-6, OA97E-7, respectively). The reportspresented data by kind-of-business for construction indus-tries, manufacturing, retail trade, wholesale trade, andservice industries.



Chapter 10. TRANSPORTATION SURVEYS

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

1997 Vehicle Inventory and Use Survey . . . . . . . . . . . . . . . . . . . . 96Precursor to the Vehicle Inventory and Use Survey (VIUS). . . . . . . . . . 96Scope and Content of the 1997 VIUS. . . . . . . . . . . . . . . . . . . . . 96

Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

Uses of the Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97Differences: VIUS vs. Other Economic Data Collection . . . . . . . . . . . 97Definition of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97Sample Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98Mailout and Follow-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

Receipt and check-in . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100Data screening and entry. . . . . . . . . . . . . . . . . . . . . . . . . . 100Data edits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

Data Tabulation and Release . . . . . . . . . . . . . . . . . . . . . . . . . 101

1997 Commodity Flow Survey . . . . . . . . . . . . . . . . . . . . . . . . . 101Precursors to the 1997 Commodity Flow Survey (CFS) . . . . . . . . . . . 101Scope of the 1997 CFS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101Changes Made from Previous Commodity Flow Surveys . . . . . . . . . . 101Uses of the Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102Sample Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102Standard Classification of Transported Goods (SCTG) . . . . . . . . . . . . 102

Mileage calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102Mailout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103Editing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103Imputation of missing data . . . . . . . . . . . . . . . . . . . . . . . . 103Disclosure review. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Publication Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

CONTENTS


Transportation Surveys

INTRODUCTION

This chapter presents information on two surveyprograms conducted during the 1997 Economic Census:(1) the 1997 Vehicle Inventory and Use Survey (VIUS) and(2) the 1997 Commodity Flow Survey (CFS). Title 13,United States Code, Sections 193 and 195, authorizes theSecretary of Commerce to direct the U.S. Census Bureau tocollect economic data such as those gathered by the VIUSand CFS. These surveys produced essential transportationdata that were not available from other sources.

1997 VEHICLE INVENTORY AND USE SURVEY

Precursor to the Vehicle Inventory and Use Survey(VIUS)

Prior to the 1963 Census of Transportation, little wasknown about the Nation’s trucking fleet. To fill this void,the Census Bureau established the Truck Inventory andUse Survey (TIUS) as part of the quinquennial economiccensus. The TIUS collected detailed information from asample of trucks and truck-tractors selected from statemotor vehicle registration records for private and commer-cial vehicles.

Following the 1992 TIUS, the Census Bureau offeredpublic and private data users a chance to recommend newitems and changes to the survey. As a result, the CensusBureau made the following changes to the questionnaire:

• Respondents were asked to report air bags.

• The ‘‘Other’’ category was deleted from the list ofresponses for the vehicle acquisition and disposalitems.

• A supplementary sample of approximately 3,000 truckswas selected for the entire U.S., a sample that consti-tuted approximately 1.5 percent of all registered trucks.

Scope and Content of the 1997 VIUS

Scope. In 1997, the Truck Inventory and Use Survey(TIUS) was renamed the Vehicle Inventory and Use Survey(VIUS) to reflect an expanded survey scope. The 1997 VIUScollected data on the physical and operational characteris-tics of the Nation’s trucking fleet, based on a stratifiedrandom sample of private and commercial trucks regis-tered (or licensed) in each state and in the District ofColumbia in 1997.

While the survey sampled private and commercialtrucks, it excluded vehicles owned by Federal, state, andlocal government agencies, as well as ambulances, buses,

and motor homes. It also excluded some privately or com-mercially owned vehicles that did not have to be licensed(such as farm equipment). Other vehicles were classifiedas out of scope for the VIUS after the Census Bureaureceived the questionnaire responses from the vehicles’owners. Such out-of-scope vehicles included—

• Trucks sold prior to July 1, 1996.

• Farm tractors.

• Unpowered trailer units or converter dollies.

• Trucks reported to have been wrecked prior to the 1997registration year.

Many states allowed pick-ups, minivans, and sport-utility vehicles to be registered as either cars or trucks.For this reason, a search was conducted of the passengercar files: then, pick-ups, small vans, and sport-utilityvehicle records found in these files were moved into theVIUS sample universe.

The annual vehicle registration date varied amongstates. Although a few states used the calendar year forvehicle registration, most staggered the registrationworkload across the 12 months. In addition, most statespermitted ‘‘grace periods’’ to better distribute the annualprocessing of vehicle registrations. Regardless of the indi-vidual states’ respective registration deadlines, vehicleregistration data in the survey were to be reported as ofJuly 1, 1997.

Content. Characteristics of the Nation’s trucking fleetwere collected using two questionnaires—Forms TC-9501and TC-9502. Form TC-9501 was sent to owners of trucksin the pick-up and van strata (about 25 percent of thesample). Form TC-9502 was sent to owners of all othertrucks selected in the sample (about 75 percent). The datarequested on the questionnaires were of two types:

• Physical characteristics—

• Vehicle type.

• Average weight.

• Type of engine.

• Type of braking system.

• Type of equipment.

• Fuel conservation.

• Type and size of body.

• Number of power axles.

• Axle arrangement of trailer units.

• Cab type.



• Operational characteristics—

• Base of operation.

• Number of trucks.

• Truck-tractors and trailers operated from the base ofoperation.

• Area of operation.

• Vehicle miles.

• Weeks operated.

• Miles per gallon.

• Use of vehicle.

• Types of commodities carried, including hazardouscargo.

Uses of the Data

Federal and state transportation agencies used the datafor the following purposes:

• Planning highway cost allocations.

• Planning road improvements.

• Determining truck size and weight regulations.

• Determining energy consumption.

• Determining types of fuels used by commercial andprivate vehicles.

• Improving transportation services for shippers andcarriers.

The private sector also had important uses for the data:

• Planning future vehicle designs and improvements.

• Conducting market studies.

• Assessing the utility and cost of certain types ofequipment.

• Determining fuel demands.

Differences: VIUS vs. Other Economic DataCollection

The VIUS differed from other 1997 Economic Censusdata collection efforts in four major respects. First wasthe reporting unit: in the economic census, the reportingunit was the business establishment. In the VIUS, analo-gous organizational units were the truck owners, eitherindividual or corporate. Coverage was the second majordifference: the VIUS survey cut across industrial classifica-tions. The third major difference was in the use of sam-pling procedures. Most of the other economic census datacollection was conducted on a total enumeration basis,and supplementary data often were obtained by sampling.In contrast, the VIUS exclusively used stratified probabilitysampling. Finally, VIUS data collection did not begin on asingle date. For 1997, the data collection intervalextended from January 1998 to May 1998, as 1997 vehicleregistration files became available from the states.

Definition of Terms

The Census Bureau employed the following terminology:

• Major use: This was based on answers to the question,‘‘Which of the following best describes your business orthe part of your business in which the vehicle wasused?’’ (Question 19, questionnaire TC-9501 and ques-tion 27, questionnaire TC-9502.) Each of the 15 specificmajor-use categories conformed to the generallyaccepted meaning of the term. If possible, the analystsrecoded responses to the ‘‘Other’’ category into one ofthe other 14 categories. The following list containsfrequent responses reported in the ‘‘Other’’ category:

Entry Recoded to (category)

House moving For-hire transportation

Trucks used in conjunc-tion with railroads

For-hire transportation

Armored car services For-hire transportation

Commercial fishing Agricultural activities

Oil field services Mining and quarrying

Specialized activitiescommonly thought to beservice (e.g., plumbing,painting, plastering,carpentry, electricalwork, etc.)

Contractor activities specialtrades

The Census Bureau also recoded the following: U.S.mail service on a for-hire, contract basis, antique trucks(when used as exhibits), and yard tractors. These entrieswere placed in the major-use that was consistent with thecompany’s name and other responses to the survey.

• Hazardous materials. This term identified those truckstransporting hazardous materials in quantities largeenough to require a placard under the Code of FederalRegulations, Title 49, Transportation.

• Products carried. The survey included broad classifica-tions of agricultural, manufacturing, and mineralproducts, as well as special categories of materialscarried by trucks. If possible, the agency recodedresponses to the ‘‘Other’’ category to one of the 29specified categories.

• Range of operation.

• Local (less than 50 miles from the vehicle’s homebase—the farm, terminal, factory, mine, etc.).

• Short range (between 50 and 100 miles).

• Short range-medium (between 100 and 200 miles).

• Long range-medium (between 200 and 500 miles).

• Long range (trips beyond 500 miles from thevehicle’s home base).

• No home base (mostly over-the-road truck tractor orconsumer one-way rental vehicles that did not oper-ate from one, particular home base location).



• Off-road (minimal use of public roads; usually associ-ated with construction and farming activities).

• Truck fleet size. The Census Bureau based the size ofthe truck fleet on the number of trucks and trailersoperated by a truck owner’s entire company. If respon-dents did not provide an answer in the ‘‘Truck FleetSize’’ section, the vehicle was estimated in the ‘‘notreported’’ category.

• Operator classification. This term consisted of thefollowing categories:

• ‘‘Business use.’’ A private owner or a company thattransported its own materials or merchandise.

• ‘‘Personal transportation.’’ A vehicle operated forpleasure driving, travel to work, school, shopping,entertainment, etc.

• ‘‘For hire.’’ Motor carrier (a vehicle operated by acompany whose primary business was to providetransportation services carrying freight belonging toothers) and owner/operator (a vehicle operated by anindependent trucker who drove the vehicle forhimself/herself or on lease to a company).

• ‘‘Daily rental.’’ Vehicles rented or leased under dailyor short-term rental or lease agreements (excludedmotor carriers).

• ‘‘Mixed.’’ A vehicle used for at least two of the aboveoperator classifications (business, personal, for-hire).If the percentage of mileage the vehicle traveled foreach activity was not equal, the response wasrecoded to the classification with the highest percent-age of use.

• Type of carrier. This category was limited to ‘‘for hire’’interstate operators as follows:

• Contract. Offered transportation services to certainshippers under contracts.

• Common. Offered transportation services to thegeneral public over regular and irregular routes.

• Exempt. Transported commodities or provided typesof services that were exempt from Federal regulation,and possibly also operated within exempt commer-cial zones.

• Body type. This term included the type of body thatwas either permanently attached to the truck or mostfrequently used with a truck tractor. Where possible,entries in the ‘‘Other’’ category were recoded to aspecific category. Responses in the ‘‘Other’’ categoryincluded mobile-home carriers and boat transports.

• Minivans. The minivan universe included minivansmanufactured on a truck or passenger car chassis.

• Estimates were requested for the following dataregarding vehicle size:

• Empty weight of the vehicle or vehicle trailercombination.

• Average vehicle weight plus cargo weight of thevehicle or vehicle/trailer combination when carryinga typical payload.

• Gross weight (maximum) at which this vehicle orvehicle/trailer combination operated.

The response to the average weight item enabled theCensus Bureau to classify trucks and truck/trailer combi-nations into the following four vehicle size classes:

• Light 10,000 pounds or less.

• Medium 10,001 to 19,500 pounds.

• Light-heavy 19,501-26,000 pounds.

• Heavy-heavy 26,001 pounds or more.

Sample Design

Using an estimated 75 million motor vehicle registra-tions supplied by the states, the R.L. Polk & Co., under aCensus Bureau contract, selected a stratified probabilitysample1 of 130,500 trucks and truck tractors in accor-dance with the Census Bureau’s sample-size criteria. R.L.Polk & Co. stratified the trucks and truck tractors into thefollowing five body-type categories within each state:

• Pickup (consisting only of pickup trucks).

• Van (consisting of panel trucks, vans, minivans, sport-utility vehicles including jeeps, and station wagons ontruck chassis).

• Single-unit light (consisting of all other single-unittrucks, excluding those in the pickup and van strata,with a gross vehicle weight of 26,000 pounds or less).

• Single-unit heavy (consisting of the remaining single-unit trucks).

• Truck tractor (consisting of only truck tractors).

Upon receipt of the sample from the contractor, theCensus Bureau entered the sample into a master controlfile that was used to mail the two questionnaires and todetermine which vehicles should be included in thefollow-up.

Mailout and Follow-Up

Staff at the National Processing Center (NPC) in Jeffer-sonville, IN,2 mailed enumeration packages, by state, topotential respondents from January through May 1998.

1The term ‘‘probability sample’’ means; (1) there are distinctsamples that can be selected. (2) each sample has a known prob-ability of selection, (3) one of the distinct samples is chosen, and(4) the estimation procedure must lead to a unique estimate ofeach sample.




Table 10-1.Number of Trucks Sampled

Panel state Number of questionnaires

Panel 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . January 16, 1998Alaska . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,336Delaware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,338District of Columbia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1,777Florida . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,706Hawaii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,304Maine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,483Massachusetts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,640New Hampshire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,489New Jersey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,752New York . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,657North Carolina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,731Oregon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,668Utah . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,555Washington . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,625

Panel 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . February 13, 1998Arizona . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,624Arkansas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1,795Colorado . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,587Idaho . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,456Indiana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,618Iowa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,701Louisiana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,563Michigan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,653Minnesota . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,685Mississippi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,459Missouri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,611Montana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,549Nebraska . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,593Nevada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,481North Dakota . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,594Ohio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,628Oklahoma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3,091Pennsylvania . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,622Rhode Island . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,266South Carolina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,572South Dakota . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,539Tennessee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,671Texas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,787Vermont . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,345Virginia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,685West Virginia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,547Wisconsin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,656

Panel 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . April 10, 1998Alabama . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,596California . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,617Connecticut . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,557Georgia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,669Illinois . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,625Kansas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,599Kentucky . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,613Maryland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,588Wyoming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,492

Panel 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . May 27, 1998New Mexico . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2,903



The Census Bureau provided customized mailing andreporting arrangements for owners of large fleets oftrucks selected in the sample.3 Based on the owners’ pref-erence, questionnaires were boxed by ‘‘panel’’—each panelcomposed of truck registrations in selected states—andmailed to the corporate headquarters or other designatedlocation for completion.

Each mailout package consisted of the appropriatequestionnaire, cover letter, and a return envelope. TheCensus Bureau sent questionnaire TC-9501 to owners oftrucks in the pickup and van strata, and questionnaireTC-9502 to owners of all other truck types (see Table10-1). Census Bureau staff identified vehicles on the ques-tionnaire prior to mailing: they included—in the registra-tion information section—the vehicle make, model year,license number, and vehicle identification number. Ownerswere asked to reply only for the identified truck, regard-less of other vehicles they might own or operate. Theywere to return their questionnaires to the NPC within 30days of receipt.

Four days prior to each panel’s mail follow-up, the fileof nonrespondents was sent to the NPC. Then, a secondform TC-9501 or TC-9502 questionnaire was printed, withthe registration information, and mailed with the properfollow-up letter. Recipients of questionnaires TC-9501 orTC-9502 were mailed letter TC-9500-F during the firstfollow-up. During the second follow-up, letter TC-9500-CFwas mailed with the questionnaire.

The Census Bureau used certified mail for the secondfollow-up mailing of the questionnaires. Then, staffconducted a telephone follow-up, by strata, within stateswhose response rate was less than 75 percent. In somecases, the phone call served as a reminder, and resulted ina mail response. If questionnaires were not returned to theNPC, analysts phoned the company’s contact person torequest a response. This continued until the strata wereclosed out for each panel. The data collection process wascompleted in September 1998.

The response rate for certified mail was 38.6 percent.The Census Bureau received responses from 105,054 (orabout 80.4 percent) of companies in the sample. The finalresponse rate for customized reports was 88.7 percent.4

Data Processing

Receipt and check-in. Clerks at the NPC checkedresponses in as the questionnaires arrived. The clerksused laser scanners to read the barcoded Census FileNumbers (CFNs—Chapters 7, ‘‘Preparations for CensusMailout,’’ and 8, ‘‘Data Collection and Processing,’’ providemore detail on CFNs) on each mail address label. This pro-cedure checked each report form into the TransportationAutomated Control System (TRACS), a system thattracked questionnaires through all stages of data process-ing. The NPC received two types of survey-relatedmail—questionnaires and correspondence. Clerks assigneda check-in (source) code for the type of receipt: mail, tele-phone, fax, undeliverable as addressed, received in Wash-ington DC, delinquent, telephone follow-up, congressionalcorrespondence, and other correspondence. After check-in, clerks entered the appropriate status code for eachquestionnaire in the TRACS, and referred the question-naires to the appropriate processing unit. When more thanone questionnaire was received in one package, the clerkskept the questionnaires together through all stages of pro-cessing so that one contact could be made, if required, forall questions.

Data screening and entry. Clerks screened all question-naires for completeness and to ensure that the informa-tion provided by respondents could be keyed. They also(1) looked for ‘‘must items’’ that needed to be completed,(2) converted responses (e.g. pounds to tons), (3)transcribed respondent remarks, (4) clarified illegibleresponses, and (5) rounded figures. If information for anyof the ‘‘must items’’ was missing, the questionnaire wasrouted to the telephone unit, where Census Bureau staffcalled the truck owner to obtain complete information. Allquestionnaire items that could not be completed satisfac-torily through screening or by phone were forwarded to astatistical assistant at the NPC for resolution.

Once the questionnaires were screened, and all prob-lems had been resolved, they were forwarded to themicrofilming unit. Microfilming equipment scanned thebar-code on each address label as each questionnaire wasfilmed, and included it in a file sent to Census Bureauheadquarters with the accompanying reel and frame num-ber. The original microfilm was stored at the NPC and acopy of each reel was sent to headquarters to be used indata editing and data analysis. After microfilming, thequestionnaires were routed for data keying.

The data entry unit keyed (1) corrections made to theregistration information, (2) all data responses providedon the questionnaires, and (3) comments written byrespondents. The NPC transmitted the information to theCensus Bureau’s computer center in Bowie, MD, via dedi-cated telephone lines. Data tapes were retained for pos-sible retransmission until the Census Bureau’s computercenter notified the NPC that the data had been received.

3While the Census Bureau forms typically were mailed to theregistered owner of the vehicle, the Census Bureau contactedowners of large fleets and asked if they would prefer anotherarrangement. Some owners wanted all forms mailed to their cor-porate headquarters to a particular individual who became theCensus Bureau contact for follow-ups. Other owners preferredthat the Census Bureau sort the forms by a given characteristic(such as state of registration or subsidiary) and mail each to a par-ticular address or individual.

4The response rate for customized reports was higher becausethe Census Bureau was dealing at the company level for thecompletion of many forms—in some cases, hundreds. The CensusBureau offered companies options on completing one form to rep-resent others with ‘‘like’’ characteristics.



Data edits. At Census Bureau headquarters, each datarecord underwent a detailed computer edit for inconsis-tencies and reasonableness. During this stage, the CensusBureau sometimes inserted data—based on administrativeinformation supplied by R.L. Polk & Co. or from otherdata-item responses—into the data file to complete blank,or obviously incorrect, responses. The computer flaggeditems that failed edits, and displayed each failed record forreview by an analyst. Once the analyst made corrections,the record was re-edited and displayed again if it stillfailed the edit. In addition to making corrections, analystsalso deleted records—determined by pre-set criteria—tobe out of scope of the survey.

Census Bureau staff developed procedures to imputeannual miles and lifetime miles. Averages were calculatedbased on responses to the five characteristics most closelycorrelated with annual and lifetime miles—model year,vehicle type, acquisition, disposal, and weeks operated.Cubic-inch displacement and number of cylinders werederived from administrative records from each state.Length and average weight (if data were missing) wereimputed for each truck. These imputations were basedupon responses from records with similar characteristics.

Once data records from the 1997 VIUS had been editedand tabulated, the NPC stored the paper questionnairesuntil January 1, 2000, after which they were destroyed.After data tabulation and publication for the 1997 VIUS,the microfilm of the questionnaires was sent to theWashington Federal Records Center in Suitland, MD, forarchiving.

Data Tabulation and Release

For each state, data were tabulated using unweightedand weighted estimates of the number of trucks and theircharacteristics. Headquarters analysts computed corre-sponding sampling variability, checked tabulations forconsistency and logic, and made comparisons between1992 and 1997 data. Prior to the preparation of PortableDocument Format (pdf) files for release on the Internet,certain data were automatically suppressed. Such datawere not published if the estimates did not meet publica-tion standards on either of two bases: the response rate,or the associated or relative standard error.

In addition to the respective printed and Internetreleases of the United States Summary—as well as Internetreleases of state level data—the 1997 VIUS producedmicrodata information for each truck in the sample. Theserecords were modified to avoid identification of individualtrucks or operating establishments, and subsequentlywere made available on a CDROM. The CD-ROM includedstate level and United States’ data tables, tabulating soft-ware, and documentation.

1997 COMMODITY FLOW SURVEY

Precursors to the 1997 Commodity Flow Survey(CFS)

From 1963 through 1977, the Census Bureau producedmeasures on the flows of goods and materials accordingto mode of transportation. These data were gathered in

the Commodity Transportation Survey (CTS). An evaluationof the 1977 CTS showed that major problems were causedby the methodology used to collect the shipment samples,as well as the small sample size used. In 1983, the CensusBureau conducted an abbreviated CTS that required sum-mary data from those surveyed. Because an evaluation ofthe 1983 survey found that the summary data were ofpoor quality, the Census Bureau did not publish 1983 sur-vey results.

In 1990, the Department of Transportation and theCensus Bureau began exploring the possibility of restoringa commodity-based survey. The need for this informationwas recognized in the ‘‘Statement of National Transporta-tion Policy’’ issued by the Secretary of Transportation in1990. This was reinforced by the Intermodal SurfaceTransportation Efficiency Act that Congress enacted in1991. Thus, the Census Bureau designed the CommodityFlow Survey (CFS) to collect and publish statistics that hadpreviously been produced by the CTS. The new surveywas first conducted for 1993; the data were collected bythe Census Bureau with additional funding and supportfrom the Bureau of Transportation Statistics.

Scope of the 1997 CFS

The 1997 CFS collected data on commodities shipped,their value, weight, and mode of transportation, as well asthe destination of shipments. The 1997 survey involved asample of approximately 100,000 domestic manufactur-ing, mining, wholesale, and selected retail establishments.

Changes Made From Previous Commodity FlowSurveys

Using the results of the 1993 CFS, the Census Bureaudeveloped a more efficient sample design. Whereas thesample size in 1993 was 200,000 establishments, the1997 survey sampled only 100,000 establishments. Inaddition, the sampling period for each questionnaire wasreduced from 2 weeks for the 1993 survey to 1 week for1997.

Other changes for 1997 included—

• Data collection using the four-digit UnitedNations/North American (UN/NA) code—instead of the‘‘yes/no’’ indication used previously—to improve theidentification of hazardous materials being transported.

• Change in the coding system—for major commodityshipped—from the 1993 survey’s Standard Transporta-tion Commodity Classification (STCC)5 system to theStandard Classification of Transported Goods (SCTG)system in 1997 (see discussion below).

5The Standard Transportation Commodity Classification (STCC)is a seven-digit classification developed by railroads and main-tained by the Association of American Railroads.



Uses of the Data

Data from the CFS were used by various Federalagencies for policy and program purposes, as well as byprivate companies for research:

• The U.S. Department of Transportation used these datafor policy development, assessing infrastructure needs,and in evaluating transportation safety issues.

• State Departments of Transportation used the CFSresults in combination with other data sources to evalu-ate existing uses/needs of their transportationinfrastructure.

Sample Design

The 1997 CFS used an approximate 1-in-8 sampledrawn from a universe of 800,000 establishments in themining, manufacturing, wholesale trade, and selectedretail trade industries, including auxiliary establishmentsof multiestablishment companies. The sample wasselected using a stratified three-stage design in which thefirst-stage sampling units were establishments, thesecond-stage sampling units were groups of four 1-weekperiods (reporting weeks) within the survey year, and thethird-stage sampling units were shipments.

Data Collection

Establishments were mailed four questionnaires—onefor each of the reporting weeks. Establishments were toprovide data on approximately 25 of their outbound ship-ments for a 1week period in each of the four calendarquarters of 1997. Establishments with 40 shipments orfewer for the 1-week period were to report all shipments.

Standard Classification of Transported Goods(SCTG).

For the 1997 CFS, commodities shipped were classifiedusing the SCTG. The SCTG is a four-level hierarchical sys-tem, created jointly by agencies of the United States andCanada to produce better measures of transportationactivity from the Commodity Flow Survey. The SCTG isbased on building blocks provided by the HarmonizedCommodity Description and Code System and Canada’sStandard Classification of Goods coding system. (TheHarmonized Commodity Description and Code System is asix-digit coding system used in the United States to valu-ate exports, imports, and customs.)

Mileage calculations. The mileage estimate componentof the 1997 CFS was produced by the Oak Ridge NationalLaboratory’s (ORNL) Center for Transportation in OakRidge, TN. The Center used a series of transportation net-works developed and maintained at the ORNL tocreate estimates of miles traveled by mode(s) of transpor-tation on a shipment by shipment basis.

Mailout. Mailout and data collection were conducted bythe NPC and took place on a flow basis during each of thefour calendar quarters of 1997. Initial mailout began inDecember 1996, and follow-up concluded in February1998. During mailout, establishments were sent a ques-tionnaire, an instruction guide, and a Commodity CodingManual.

The Census Bureau used two versions of thequestionnaire—form CFS-1000 and form CFS-2000. Eachestablishment received form CFS-1000 in each of its first 3reporting weeks. For the fourth reporting week of eachquarter, a subsample (approximately 25,000) receivedform CFS-2000, while the remaining establishmentsreceived form CFS-1000.

The Census Bureau requested that each establishmentreport the following information:

• Physical location (if different from that on the mailinglabel).

• Operational status (i.e., in operation, temporarily or sea-sonally inactive, ceased operation) of the establishmentduring the 1-week sampling record.

• Total number of shipments made during the 1-weekperiod.

The establishments also were requested to provide thefollowing data for each shipment sampled:

• Shipment document number and date of shipment.

• Shipment weight in pounds.

• Value of shipment.

• Commodity description and the five-digit SCTG.

• The UN/NA code for hazardous material shipments.

• All domestic modes of transportation used.

• Whether the shipment was containerized.

• Domestic-destination city, state, and ZIP Code.

• Whether the shipment was an export and, if so, themode of export and the city and country of destination.

In addition to the above items, recipients of the formCFS-2000 were asked to furnish additional data on theavailability and use of transportation facilities and equip-ment as follows:

• Were the following on-site facilities available, and werethey used: rail siding, waterway docks, airport/landingstrips, or pipeline terminals?

• Did you use the following off-site facilities and, if so,what was the distance of the facility from your estab-lishment and mode of transport used to reach it: railsiding, waterway dock, airport/landing strip, andpipeline terminal?



• Did you use rail cars that (1) your company owned orleased, (2) a common carrier owned or leased, or (3)another party owned or leased? Also, what percentageof your total rail shipments were each of these threemethods?

• Did you use trucks (with six or more tires) or trucktractors that your company (1) owned, (2) leased with adriver, or (3) leased without a driver?

• Did you use the following equipment that your com-pany owned or leased: truck trailers, aircraft, barges,and other equipment (specify)?

• Who generally decided on the mode of transportationused for your shipments—your company, receiver of theshipment, or other?

Data Processing

Mailout, check-in, screening, and data capture for the1997 CFS were conducted by the NPC. Additional auto-mated edits were conducted at Census Bureau headquar-ters. Edit problems were resolved at both the NPC andCensus Bureau headquarters. The NPC also was respon-sible for the follow-up, which consisted of a remindermailed to nonrespondents. A telephone follow-up wasconducted, during which the largest establishments wereasked to return their completed questionnaires. Censusstaff did not attempt to collect data by phone because ofthe large amount of data requested by the questionnaires.The overall response rate for the 1997 CFS was approxi-mately 75 percent.

Editing. The Census Bureau headquarters staff created aflow-processing edit program to identify inconsistencies,range errors, and critical omissions. The edit program wasrun on a computer system, following the transmission ofthe data entry files from the NPC to Census Bureau head-quarters. The edit program assigned a series of flags formissing or incorrect data.

Imputation of missing data. Item nonresponseoccurred when a data variable (either value or weight)either was not reported or was determined to be invalidby the edit process. Establishment nonresponse, in turn,took place when an establishment failed to reportshipment data for at least one quarter. Under certainconditions, item nonresponse was handled by imputationprocedures. Establishment nonresponse was handledprimarily by an adjustment of statistical weights.

Disclosure review. To make certain that informationabout individual companies would not be disclosed, theCensus Bureau subjected the data to disclosure analysis.In conformance with the Census Bureau’s confidentialityrules, a computerized disclosure analysis systemidentified—and withheld from publication— data thatmight have revealed the identities of, and informationabout, individual companies. In those instances, the Cen-sus Bureau substituted a reference symbol ‘‘D’’ for the datain the published tabulations.

Publication Program

Data from the 1997 CFS were made available in bothprinted and electronic formats. Tables provided informa-tion on commodities shipped, their value, weight, andmode of transportation, as well as the origin and destina-tion of shipments. The survey also provided selected dataon hazardous material shipments and exports. Commoditydetail for national tables was at the four-digit SCTG level.Data were at the two-, and where possible, three-digitSCTG level for state reports. Data from the 1997 CFS werereleased for the following geographic levels: United States,census regions and divisions, states and selected metro-politan areas and remainder of states.



Chapter 11. ECONOMY-WIDE PROGRAMS

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

1997 Surveys of Minority- and Women-Owned Business Enterprises . . . . . 107Background Information . . . . . . . . . . . . . . . . . . . . . . . . . . . 107Comparability With the 1997 Economic Census . . . . . . . . . . . . . . . 107Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107Use of the Standard Industrial Classification (SIC) Code . . . . . . . . . 107Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107Business unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108Treatment of corporations and employer identificationnumbers (EINs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Explanation of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108Survey Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108Survey of Minority-Owned Business Enterprises (SMOBE) . . . . . . . . . 108Survey of Women-Owned Business Enterprises (SWOBE) . . . . . . . . . 109Sampling frames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109General information . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109Questionnaires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109Mailout and mail follow-up. . . . . . . . . . . . . . . . . . . . . . . . . 109Telephone follow-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

Data Processing and Tabulation . . . . . . . . . . . . . . . . . . . . . . . 110Receipt and check-in . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110Data entry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110Edit and tabulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110Use of administrative records . . . . . . . . . . . . . . . . . . . . . . . 110

1997 Business Expenditure Survey. . . . . . . . . . . . . . . . . . . . . . . 111Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111Comparability With the 1997 Economic Census . . . . . . . . . . . . . . . 111Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111Use of the SIC Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111Reporting units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111Collection methodology . . . . . . . . . . . . . . . . . . . . . . . . . . 111Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

Changes From Predecessor Survey. . . . . . . . . . . . . . . . . . . . . . 111Scope and Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111Explanation of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112Operating expenses . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112Measures of value produced . . . . . . . . . . . . . . . . . . . . . . . . 112

Sample Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113Report forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113Mailout and follow-up . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

CONTENTS


1997 Business Expenditure Survey—Con.Data Processing and Tabulation . . . . . . . . . . . . . . . . . . . . . . . 113Receipt and check-in . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113Data entry. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114Edit and tabulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

Tabulation and Sources of Related Data . . . . . . . . . . . . . . . . . . . 114

Publication Program—Economy-Wide Programs . . . . . . . . . . . . . . . . 114


Economy-Wide Programs

INTRODUCTION

The 1997 Economic Census included an economy-widesurvey—the Survey of Business Owners. The U.S. CensusBureau published results of this survey in several volumes,as the Survey of Minority-Owned Business Enterprises(SMOBE) and the Survey of Women-Owned Business Enter-prises (SWOBE).

The 1997 SMOBE was the seventh in a series initiated in1969; the first was a special project, while the 1977,1982, 1987, 1992 and 1997 surveys were part of theeconomic census. As in the past, this latest effort providedbasic economic data on businesses owned by Blacks,persons of Hispanic origin, Asian and Pacific Islanders,and American Indians and Alaska Natives.1

The 1997 SWOBE was the sixth in a series begun in1972. The first was conducted as a separate program,while the subsequent surveys were part of the economiccensus and provided economic data on businesses ownedby women.

The SMOBE and SWOBE provided data on not only thenumbers of minority- and women-owned firms, but alsoon sales and receipts, annual payroll, and employment.Data aggregates were presented for the United States bytwo-digit Standard Industrial Classification (SIC) code,employment and receipts size, and legal form of organiza-tion; data were provided at the state level by two-digit SICCode.

The 1997 census also included the 1997 BusinessExpenditures Survey (BES), which covered establishmentsin most of the wholesale, retail, and services sectors ofthe economy. The BES, previously called the Assets andExpenditures Survey (AES), was renamed for 1997 whenthe assets and capital investments inquiries were droppedfrom the survey. Data collected in the BES included operat-ing expenses of retailers, merchant wholesalers, and ser-vice firms (as classified under SIC). Statistical results wereissued with operating expenses compiled in other tradesurveys covering communications, transportation, travel,manufacturing, mining, and construction firms.

The SMOBE and SWOBE results were published on a flowbasis in a series of seven publications. The BES resultswere published in a single report, Business Expenses. (Seethe ‘‘Publication Program’’ section below for more details.)

1997 SURVEYS OF MINORITY- AND WOMEN-OWNEDBUSINESS ENTERPRISES


The 1997 SMOBE and SWOBE surveys represented theNation’s principal source of data needed to evaluate theextent and growth of business ownership by minoritiesand women. The data provide a framework for assessingand directing Federal, state, and local government pro-grams designed to promote the activities of disadvan-taged groups.

Persons, private firms, and institutions, in both the pub-lic and private sectors, use these survey data to evaluatebusinesses and markets. They are used by the media fornews stories, as well as by researchers and academia. TheSmall Business Administration and the Minority BusinessDevelopment Agency use the data when allocatingresources for their business assistance programs.

Comparability With the 1997 Economic Census

Introduction. The information collected by the SMOBEand SWOBE was not always comparable to the statisticscollected in the other economic census programs becauseof differences in; (1) the industry classification systemsused, (2) industrial scope and business reporting unitsused, and (3) the treatment of corporations and employeridentification numbers (EINs). These are explained in moredetail below.

Use of the Standard Industrial Classification (SIC)Code. SMOBE and SWOBE continued to use the SIC Codefor industrial classification while the Census Bureau hadadopted the North American Industry Classification Sys-tem (NAICS) for the bulk of the 1997 Economic Census(Chapter 3, ‘‘The Introduction of the North AmericanIndustry Classification System (NAICS),’’ details the historyand differences of the two).

Scope. The 1997 SMOBE and SWOBE covered most SICCode industries. Agriculture production (SIC 01, 02), theU.S. Postal Service (43), Membership Organizations (86),Private households (88), and Public Administration (91-97)were excluded.

Several SIC industries were in scope of the SMOBE andSWOBE, but out of scope of the basic economic census.These included Agricultural Services (07), Forestry (08),

1In addition, the 1997 SMOBE and SWOBE incorporated por-tions of the Enterprise Statistics Program, last conducted duringthe 1992 Economic Census, but dropped from the 1997 EconomicCensus (due to budget restrictions).



and Fishing, Hunting, and Trapping (09). MembershipOrganizations (86) were within the scope of the census,but out of scope of the SMOBE and SWOBE.

Business unit. Most of the economic census was con-ducted on an establishment or physical location basis,individually enumerating and tabulating each establish-ment owned by a firm. In the SMOBE and SWOBE, however,the whole firm was the primary unit for enumeration andU.S. level tabulations.

Treatment of corporations and employer identifica-tion numbers (EINs). For the first time, the 1997 SMOBEand SWOBE included all corporations. Prior to 1997, onlylimited data were available for corporations.

Treatment of EINs also was new for 1997. The CensusBureau had historically defined a business as an entity thatfiled a business income tax return. Businesses that hadreceived one or more EINs may have filed tax returnsunder more than one as well. In past surveys, each EINwas treated as a separate business. For 1997, all opera-tions under the same ownership were defined as one com-pany or business, irrespective of the number of the com-pany’s EINs. This definition eliminated the likelihood ofsurveying the same business owner more than once. Thischange had no effect on the employment and payroll datafor the survey, but slightly reduced the count for the num-ber of businesses and the total receipts.

EXPLANATION OF TERMS

Definitions of the survey terms for SMOBE and SWOBE,which differed somewhat from those used for the 1997Economic Census, were as follows—

• Company. A company was a business consisting of oneor more domestic establishments owned or controlledby the reporting firm at the end of 1997. If a companyowned or controlled other companies, all establish-ments of the subsidiaries were included as part of theowning or controlling company. The five types of enti-ties (legal forms of organization) covered were—

• Subchapter S corporation—Special Internal Rev-enue Service (IRS) designation for a business that waslegally incorporated under state laws, and had 75 orfewer shareholders who, because of tax advantages,elected to be taxed as individual shareholders ratherthan as corporations.

• C corporation—A legally incorporated businessunder state laws. Unlike a subchapter S corporation,a C corporation has no restriction.

• Individual proprietorship—Unincorporated busi-ness owned by one person or by self-employed indi-viduals with no employees. The economic activitymay have been the self-employed activity of a personwho held a full-time job working for someone else.

Moreover, the firm might operate one or more placesof business, such as a chain of restaurants, or haveno fixed business location, such as the firm repre-sented by a self-employed carpenter or salesperson.

• Partnerships—Unincorporated business owned bytwo or more persons.

• Other—Included in this group are cooperatives,estates, receiverships, and businesses classified asunknown legal forms of organization.

• Gross receipts—These included the gross value ofall products sold, services rendered, or other receiptsfrom customers during 1997, less returns and allow-ances. The published reports presented value of salesand services whether or not payment was actuallyreceived during the year and, therefore, the data didnot indicate a cash flow. There were no adjustmentsmade for the costs of doing business.

• Employees. These were individuals who worked eitherfull or part time. Proprietors, partners, or owners whoworked in the business, but were not paid a regular sal-ary subject to withholding for Social Security Adminis-tration (SSA) deductions, were excluded. Also excludedwere family members or others who worked in the busi-ness, but were not subject to withholding taxes.

Survey Methodology

Survey of Minority-Owned Business Enterprises(SMOBE). The SMOBE required determining the minoritystatus of the ownership of each firm in the survey, andthen matching to each firm the data collected in the eco-nomic census or gathered from administrative records.

The Census Bureau classified a firm as minority-ownedbased on the race and ethnicity of the person(s) owningmajority interest. Businesses were identified asminority-owned through a mail canvass to a sample ofcorporations, partnerships, sole proprietorships, and self-employed individuals. Any business was eligible to beselected in the sample, but firms whose owners werepotentially of Hispanic, Asian, Pacific Islander, or NativeAmerican origin had a higher probability of selection.These firms were identified using one of the sources listedbelow—

• Lists of minority-owned businesses published insyndicated magazines, located on the Internet, ordisseminated by trade or special interest groups.

• Previous SMOBE respondents.

• A list of common surnames for these minority groups,based on research using prior SMOBE data.

• The owner(s)’ self-designated race classification on anapplication for a Social Security Number.



• Racial distributions for various ZIP Codes.

• Racial distributions for various state-industry classes(derived from 1992 SMOBE responses).

• Word strings in the company name indicating possibleminority-ownership (derived from 1992 SMOBEresponses).

The sample was stratified by size (using volume ofsales or payroll) and geographically by state. The CensusBureau selected large companies ‘‘with certainty’’—i.e.,they were sure to be selected and represented onlythemselves (i.e. have a selection probability of one and asampling weight of one). The certainty cutoffs varied bysampling stratum, and the Census Bureau sampled eachstratum at varying rates, depending on the number offirms in a particular industry in a particular state. Thesampling rate was lowest in states and industries with thegreatest number of firms. A similar methodology was usedto select a sample from the remaining universe: the pur-pose of this was to estimate the number of firms ownedby persons of minority ancestry when no indication ofminority-ownership was found from any of the sourceslisted above. The mail sample did not include those soleproprietorships that had reported in the 1992 survey andwere still active in 1997; rather, the 1992 responses wereused in lieu of responses to a mail sample.

Survey of Women-Owned Business Enterprises(SWOBE). With the 1992 SWOBE, the Census Bureau dis-continued using administrative record data from the IRSand SSA to identify owners of partnerships and subchapterS corporations. Rather, women-owned corporations andpartnerships were identified from sample data.

For 1997, the Census Bureau used administrative datato determine the gender of the owner(s) of businessesfiling an IRS form 1040C (sole proprietorship or self-employed person—not a joint return). Businesses filinga joint return, where it could not be determined withcertainty whether the husband/wife or both owned thebusiness, were eligible for sampling and were mailed aquestionnaire to determine the owner’s gender.

In past surveys, a firm fell within scope of this programif the sole owner was a woman or if one-half or more ofthe partners or shareholders were women. For 1997, aseparate tabulation was made for those businesses whosestock was equally owned by males and females and thesewere not counted as female-owned. Only businesses inwhich women owned a 51 percent or greater interest werecounted as women-owned firms. Nonprofit and foreign-owned companies, as well as those with publicly heldstock whose ownership is indeterminate were tabulatedseparately.

Sampling frames. There were six sampling frames inthe SMOBE and SWOBE and every case was assigned to oneof the following frames—

• American Indian.

• Asian/Pacific Islander.

• Black.

• Hispanic.

• National (non-Hispanic White male).

• Women.

Data Collection

General information. Phase I of the data collectionbegan in August 1998. The universe included all corpora-tions or partnerships reporting any business activity in1996. A sample of over 1 million businesses from theAmerican Indian, Asian/Pacific Islander, and Hispanicframes were mailed report forms.

Phase II began the following year, with the mailing of anadditional 1.5 million report forms in July 1999 to casesassigned to each of the six sampling frames. This universeincluded corporation or partnership births and soleproprietorships and self-employed persons reporting anybusiness activity in 1997.

Questionnaires. Four questionnaires were used for the1997 SMOBE and SWOBE—the MB-1 and MB-2, in English,and the MB-1(SP) and MB-2(SP), in Spanish. The Spanishlanguage questionnaires were mailed to respondents uponrequest. The Census Bureau sent the MB-1 questionnairesto partnerships and corporations, or to sole proprietor-ships that submitted joint tax returns where, based on theadministrative records, it was probable that the husbandand wife were of different races. The MB-2 was used forsole proprietors and self-employed individuals who were‘‘single filers’’ or who filed joint returns where, based onthe administrative records, there was a low probabilitythat the husband and wife were of different races.

Businesses were asked questions about the gender, raceand ethnicity of the person(s) owning majority interest inthe business, the number of owners, and whether any ofthe businesses’ receipts in 1997 resulted from businessconducted with the Federal government.

Mailout and mail follow-up. The preparations for the1997 SMOBE and SWOBE mailout were similar to those forthe economic census. The Census Bureau engaged a pri-vate contractor to print the 2.6 million MB-1 forms, the 2.9million MB-2 forms, and approximately 5.5 million each ofthe cover letters (which included frequently asked ques-tions about the survey and answers). The contractorassembled the survey packages and delivered them to theNational Processing Center (NPC). The MB-1 packages weredelivered in July, and the MB-2 packages in December,1998. The Spanish versions of each questionnaire,MB-1(SP) and MB-2(SP), were printed using an in-houseDocuPrint high speed color printer. At the NPC, addresseswere imprinted directly on the questionnaires using an inkjet printer.



On August 17, 1998, 1,001,426 MB-1 questionnaireswere mailed to employer corporations and partnershipsactive in 1996 (Phase I). The Census Bureau conducted afinal mailout (Phase II) of 1,529,396 questionnaires on July7, 1999, consisting of 208,019 nonemployer andemployer corporations or partnerships that were active in1997, but had not been active in 1996 (Phase 1), and1,321,377 sole proprietorships and self-employed personsreporting any business activity in 1997. Surnames and SSArace codes were available for limited use in samplingthese firms. Otherwise, the Census Bureau usedmethodology similar to that specified above for soleproprietorships.

The Census Bureau asked companies to return theircompleted questionnaires within 30 days of receipt; non-respondents were included in one or more of the threemail follow-ups. Below are the dates of the follow-ups, andthe types and number of questionnaires mailed in each.

Completed questionnaires were received from 813,870businesses in Phase I, yielding an overall response rate of81.3 percent. The U.S. Postal Service returned 58,935cases in Phase I as ‘‘Undeliverable As Addressed’’ (UAAs)and these cases were assumed to be out of business.

Figure 11-1.Mailouts and Mail Follow-Ups

Mailout Date Questionnaire Number mailed

PHASE I

Initial mailout . . 08/17/98 MB-1 1,001,426

1st follow-up . . 10/05/98 MB-1 373,364

UAA remail . . . 10/05/98 MB-1 11,282

2nd follow-up . . 10/26/98 MB-1 257,188

3rd follow-up . . 11/25/98 MB-1 160,025

PHASE II

Initial mailout . . 07/07/99 MB-1 293,582MB-2 1,235,814

1st follow-up . . 08/12/99 MB-1 143,92308/20/99 MB-2 544,877

UAA remail . . . 09/02/99 MB-1 5,326MB-2 18,095

2nd follow-up . . 09/13/99 MB-1 79,05909/20/99 MB-2 300,916

3rd follow-up . . 10/14/99 MB-1 46,33710/20/99 MB-2 221,287

Telephone follow-up. The Census Bureau conducted atelephone follow-up of 50,639 Phase I nonrespondentSMOBE and SWOBE cases, beginning January 11, 1999, andcontinuing through March 5, 1999. ‘‘Certainty’’ cases com-prised approximately half of the of the cases selectedfrom a pool of large-receipts size companies—the otherhalf was randomly selected from the remaining pool ofnonrespondents. By the closeout of data collection,responses had been obtained by telephone for 26,095 of

these cases; 14,118 firms had refused to respond, or hadnot answered after 4 attempts to contact. The CensusBureau was unable to obtain telephone numbers for theremaining 10,426 cases.

A telephone follow-up to 36,000 Phase II nonrespon-dent SMOBE and SWOBE cases was conducted fromSeptember 10 through September 30, 1999 in conjunctionwith the second follow-up mailing. A total of 26,426 caseswith valid phone numbers were selected with certaintyand 9,574 cases were randomly selected from the pool ofremaining cases for which no phone number existed onthe SSEL. By the closeout, 5,999 late mail returns had beenreceived, leaving a net workload of 30,001. Data wereobtained from 19,144 (63.8 percent) cases, 507 (1.7 per-cent) cases refused to respond, and 10,350 (35 percent)were unreachable (i.e. no phone number, voice mail, noknowledgeable contact, etc.).

Data Processing and Tabulation

Receipt and check-in. The 1997 SMOBE and SWOBEquestionnaires were returned to the NPC office where theywere checked in using automated barcode scanning equip-ment. The check-in file was used to update the surveymail list and to generate address files for mail follow-up ofnonrespondents.

Data entry. Company identification information and datafrom each questionnaire were captured by an electronicimaging system developed by Lockheed Martin for thedecennial census using optical character recognition andoptical mark recognition software. The SMOBE/SWOBE pro-grams were the first production test of the system andhad to work within the time constraints imposed by the2000 Decennial Census. Forms that were returned too latefor processing were keyed by NPC data-entry staff.

Edit and tabulation. The data for the 1997 SMOBE andSWOBE were edited and tabulated on a flow basis asresponses were received. Between February and April2000, staff at Census Bureau headquarters used the com-pany record files and interactive computer systems todevelop the publication tables for each survey. The Com-pany Statistics Division and Administrative and CustomerServices Division staff prepared the accompanying textand worked out the details of the format and design of theprinted reports.

Use of administrative records. Once the minority- andwomen-owned businesses had been identified, the CensusBureau retrieved the following information from its data-base, which contained 1997 Economic Census question-naire responses and information from administrativerecords),2 and matched the information to the properfirm—

2The Census Bureau’s principal sources of noncensus adminis-trative records were the IRS, the SSA, and the Bureau of Labor Sta-tistics. For more information on the kinds of records and dataused, see Chapter 7, ‘‘Preparations for the Census Mailout.’’



• Annual payroll (for a company with paid employees).

• Geographic codes.

• Organizational status (sole proprietorship, partnership,or corporation).

• Number of employees reported during the week includ-ing March 12, 1997.

• Receipts.

• SIC Code.

1997 BUSINESS EXPENDITURE SURVEY

Introduction

The 1997 Business Expenditures Survey (BES) repre-sents the only reliable source of the Nation’s businessexpenses data covering merchant wholesale trade, retailtrade, and service industries. The 1997 BES supplementsbasic economic statistics produced by the 1997 EconomicCensus for the wholesale trade, retail trade, and servicessector industries with estimates of operating expenses.Further, it provides measures of value produced for mer-chant wholesale and retail trades. Essential measurementof the Nation’s economy requires compilation of compre-hensive and reliable data on both economic outputs (e.g.,sales) and inputs (e.g., utility and advertising expenses),and the BES is the sole source of national business inputsdata for covered industries.

The Federal Government is the primary user of the BESdata. In particular, the Bureau of Economic Analysis—theFederal agency that produces gross domestic productestimates and maintains the national economicaccounts—uses information from the survey for develop-ing the national income and product accounts, input-output tables, and economic indexes, and to fill previouslyidentified critical gaps in underlying data in theseaccounts.

Comparability With the 1997 Economic Census

Introduction. The 1997 BES differed from other 1997Economic Census operations in four major respects—(1) use of the SIC Code for industrial classification,(2) reporting units used, (3) data collection methodology,and (4) the data collection schedule.

Use of the SIC Code. As was the case with the SMOBEand SWOBE, the BES continued to use the SIC Code forindustrial classification instead of the NAICS used for thebulk of the 1997 Economic Census.

Reporting units. The economic census collected data ona total enumeration basis, obtaining supplementary databy sampling. In contrast, the BES used stratified probabil-ity sampling exclusively (discussed below under ‘‘SampleDesign’’). While the reporting unit for multiestablishmentfirms in the census was the individual component estab-lishment, the reporting unit for the BES was a combination

of company-within-trade area, subsidiary/division-within-trade area, or Employee Identification Number (EIN)-operations-within-trade area (EINs are described above).This reporting arrangement was designed primarily tominimize respondent burden.

Collection methodology. A third major difference washow the data were collected. Because the BES sample wasshared by the annual business surveys, the initial ques-tionnaire mailings were combined with those surveys’1997 mailings, and shared the respondent cover letter.Mail and telephone follow-ups, however, were conductedindependently.

Schedule. A fourth difference was the data collection andpublication schedule. The response due date for the 1997Economic Census was February 12, 1998, but mid-Marchfor the BES. Reports from most of the 1997 Economic Cen-sus were released on a flow basis from early 1999 throughlate 2000. The BES results were scheduled for release bylate 2000.

Changes From Predecessor Survey

The 1997 BES was the successor to the 1992 Assetsand Expenditures Survey (AES). The new survey question-naire deleted questions on depreciable assets, capitalizedinterest, and capital expenditures. Items were added ask-ing for expenditures for contract labor and computer soft-ware and services. Coverage of selected service industrieswas extended to include nonemployer companies (thosewith no paid employees). The 1992 survey had obtaineddata for nonemployer companies using a business areasample; for the 1997 BES, the Census Bureau used admin-istrative records for basic data for nonemployers.

Scope and Content

This Census Bureau conducted the 1997 BES under theauthority of the 1997 Economic Census, covering businessoperations classified in the 1987 SIC manual. As in prioryears, the industrial scope included merchant wholesaletrade, retail trade, and most service industries. The itemscovered included various types of operating expenses.Data were tabulated and released for the United Statesonly. Content of the BES report forms included inquireson—

• Annual payroll.

• Supplemental labor costs.

• Cost of contract labor.

• Depreciation and amortization charges.

• Taxes and license fees, excluding sales and incometaxes.

• Lease and rental payments.

• Cost of office supplies.



• Cost of materials (wholesale, retail only).

• Cost of contract work (wholesale only).

• Commission expenses (wholesale only).

• Cost of purchased services: communications services;electricity, fuel, and other utilities; cost of computersoftware and other computer related services; repairand maintenance services; accounting and bookkeepingservices; and legal services.

Companies also were asked to include data for auxiliarylocations whose primary functions were to manage,administer, service, or support the activities of the mainestablishments covered.

Explanation of Terms

Operating expenses. A few types of operatingexpenses requested on the BES report form are self-explanatory; others included—

• Annual payroll—All salaries, wages, commissions,bonuses, and allowances for vacation, holiday, and sickleave paid to employees in 1997 before payroll deduc-tions. For corporations, it included amounts paid toofficers and executives; for unincorporated businesses,it excluded payments to or withdrawals from profits byowners or partners. (Annual payroll did not includeemployer contributions to plans required by law andother plans besides vacation, holiday, and sick leave.)

• Employer costs for programs required by law—Allprograms required under Federal and state legislationsuch as FICA, unemployment tax, workers’ compensa-tion, and state disability payments.

• Employer costs for other fringe benefits—Thisincluded programs not specifically required by Federalor state legislation, such as life and health insurancepremiums for employees.

• Cost of contract labor—Cost of labor, as supplied byother firms, under the supervision of the reporting firm.

• Depreciation charges—All amortization and deprecia-tion charges against depreciable assets owned during1997. Amounts representing tax-exempt organizations,if not reported, were estimated even though deprecia-tion accounts may not have been maintained.

• Taxes—All taxes and license fees paid during 1997,excluding income and payroll taxes. These includedtobacco and liquor tax stamps, fines, and propertytaxes, and excluded direct payments to government taxagencies for sales and other taxes collected directlyfrom customers.

• Lease and rental payments—Payments for lease orrental of buildings, office and store space, machineryand equipment. Payments for machinery and equipment

included use of production machinery, office equipment,computer systems, passenger cars, trucks, materialshandling equipment, and all other types of machineryand equipment.

• Cost of purchased communication services—Costof telephone, data transmission, telegraph, telex, tele-type and all other communications services purchasedin 1997.

• Cost of purchased electricity—Some businessesincluded this expense as part of building/office leasepayments.

• Cost of purchased fuels—Cost for fuels used forheating, power, or generation of electricity, also some-times included in lease payments.

• Cost of other utilities—All utilities costs to respon-dents, except purchased fuels and electricity, such aswater treatment, sewage, and refuse removal. Thesecosts were sometimes included in lease payments.

• Cost of materials—Purchases of containers, wrapping,packing and selling supplies used in packaging, pro-cessing, shipping, and selling of goods. Not collectedfor service industries.

• Cost of contract work done for wholesalers byother firms—This included readying merchandise forsale, such as blending, printing, repacking, and otherprocessing.

• Commissions paid to other firms—These were forsales of merchandise made on the account of merchantwholesalers.

• Computer related services—All costs for data pro-cessing and other computer services purchased fromother firms. It included the cost of custom and prepack-aged software.

• Cost of purchased repair and maintenanceservices—The total amount paid for noncapitalizedrepairs and maintenance to buildings, structures,machinery, and equipment, such as motor vehicles andoffice space. It excluded cost of repairs to leased build-ings and equipment covered by regular lease payments.

• All other operating expenses—This covered allinventory storage and shipping costs, insuranceexpense (nonemployee), uninsured casualty losses, andbad debt losses. It excluded interest on loans and salesand excise taxes.

Measures of value produced. These items were com-puted for merchant wholesale and retail trades using datacompiled in the BES and business annual surveys.

• Cost of goods sold was calculated by adding allpurchases of merchandise (net of returns, allowances,and discounts, but including charges from freight, insur-ance, etc.) during the year to the beginning-of-year



inventory, then deducting the end-of-year inventoryfrom the total. Firms were instructed to exclude the costof packaging, processing, shipping, and selling ofgoods from the cost of purchases, but to report theseitems under ‘‘operating expenses.’’ Purchase of taxstamps and payments of excise taxes, often included bytobacco and liquor wholesalers in their purchases, were,for this survey, excluded from cost of goods sold andincluded under ‘‘operating expenses.’’

• Gross margin represented sales less cost of goodssold. Gross margin is equivalent to the cost of all mate-rials (as distinguished from goods to be resold) andservices provided, whether provided by the reportingfirm itself or purchased by it from others. To the extentthat it included cost of contract work done by others onmaterials of merchant wholesale firms, gross marginincluded an element of value added by manufacturing.

• Value added was the gross margin (as defined above)less the cost of supplies, materials, fuel and otherenergy, and the cost of contract work on materials ofmerchant wholesalers.

• Net income produced at market prices was thevalue added less lease and rental payments; cost ofcommunication, advertising, repair, legal, accounting,and computer related services; commissions to otherfirms; and amounts set aside for bad debt losses andother losses not compensated by insurance.

• Net income produced at factor cost was incomeproduced at market prices less depreciation, licensefees, and taxes other than income taxes. It includedpayroll, employer contributions to the Federal InsuranceContributions Act (FICA), and unemployment insurance.

Sample Design

The 1997 BES sample combined probability samplesdrawn for the 1997 Annual Trade Survey, the 1997 AnnualRetail Trade Survey, and the 1997 Service Annual Survey.The sampling frames for the surveys were constructedfrom the Census Bureau’s SSEL as of December 31, 1994.The sampling frame contained two types of samplingunits that represent firms—

• Alpha numbers that represent multiestablishment firmsowning or operating two or more establishments.

• EINs assigned by the IRS for Federal Insurance Contribu-tions Act purposes that represent a cluster of one ormore establishments.

Firms were stratified by kind-of-business and then by ameasure of size based on annual receipts, revenue orsales.

To reduce the estimate variability, sampling units withthe largest measures of size were selected ‘‘with cer-tainty.’’ The Census Bureau determined, within each kind

of business, a substratum boundary (or cutoff) thatdivided the certainty units from the noncertainty units. Allmultiestablishment firms not selected with certainty andall single-establishment firms were subjected to samplingon an EIN basis.

Census Bureau staff analyzed the data from the 1992Economic Census to determine the certainty cutoffs andnoncertainty stratum boundaries, and to allocate thesample among the size strata and kind-of-businessgroups. The sampling rates used to select a randomsample of EINs from each noncertainty stratum variedfrom 1-in-3 to 1-in-1,000.

After the initial sample selection was completed, newlyidentified businesses were represented in the samplethrough a two-phase sample procedure. Each quarter, newEINs on the IRS mailing lists for FICA taxpayers were iden-tified and stratified by kind of business and size (expectedemployment or payroll). A relatively large sample wasdrawn and canvassed for a more reliable measure of sizeand a more detailed kind-of-business code, if needed.Using this more reliable information, the ‘‘births’’ selectedin the first phase were subjected to probability-proportional-to-size sampling with overall probabilitiesequivalent to those used in drawing the initial sample. EINbirths were added to the sample approximately 9 monthsafter beginning operations.

Data Collection

Report forms. Data covering the year 1997 were col-lected during 1998 using report forms that supplementedthe Census Bureau’s Annual Business Surveys. In particu-lar, forms B-151(S), B-151A(S), B-151D(S), B-153(S), andB-153D(S) supplemented the Annual Retail Trade Survey;forms B-450(S) and B-451(S) supplemented the AnnualTrade Survey (of merchant wholesealers); and formsB-500(SA) and B-500(SE) supplemented the Service AnnualSurvey.

Mailout and follow-up. In January 1998, the CensusBureau sent 1997 BES questionnaires to approximately60,000 businesses as a supplement to the Annual RetailTrade Survey, the Annual Trade Survey (of merchant whole-salers), and the Service Annual Survey. The due date forreturning the responses—directly to the NPC—was March12, 1998. On March 24, the Census Bureau mailed the firstof three follow-up letters (with another copy of the ques-tionnaire) to nonrespondents. This was succeeded by asecond follow-up letter and questionnaire on April 28.The third and last follow-up was sent on May 27, 1998.Telephone follow-up began the last week of June andconcluded in early September 1998. At close-out, after allfollow-ups, the total response rate was almost 86 percent.

Data Processing and Tabulation

Receipt and check-in. BES questionnaires received bymail were checked in by the NPC. This was done by read-ing mailing labels using automated barcode scanning



equipment. The resulting files were transferred to the BESmain control file. Questionnaires received by FAX or tele-phone were checked in by clerks with direct access to theBES control file using the Census Bureau’s Automated Con-trol System (ACS) software. In addition to handling reportsreceived by mail, fax, and phone, NPC staff handled UAAcases, as well as telephone and mail inquiries andrequests. Congressional correspondence was referred toheadquarters. Staff entered interactively into the ACS bothcheck-in and associated actions. Problems too difficult orcomplex to resolve by NPC staff were referred to CensusBureau headquarters.

Data entry. After check-in, BES reports were routed toclerks in the screening section of the control unit. NPCclerks screened all questionnaires for completeness and toensure that the information provided by respondentscould be keyed. Questionnaires that were incomplete orinadequate for keying were referred to the on-site analystfor further review and/or for a telephone call to the com-pany. After screening and problem resolution, the clerksbatched the questionnaires into work units and forwardedthem to the microfilming unit. The microfilming equip-ment stamped a reel number and frame number onto eachquestionnaire before filming. The NPC stored the originalmicrofilm and sent a copy of each reel to Census Bureauheadquarters to be used in the data-editing and data-analysis operations. After microfilming, the questionnaireswere sent for data keying.

The data-entry unit keyed all data responses on thequestionnaires, and after verification, transmitted the datafiles electronically to the Census Bureau’s computer centerin Bowie, MD, via dedicated telephone lines. The NPCretained backup copies of the data tapes for possibleretransmission until notified that the data had beenreceived.

Edit and tabulation. As soon as data from the CensusBureau’s 1997 annual surveys of merchant wholesale,retail trade, and service industries were pronounced final,business records for selected data fields from the surveys’databases were copied to the new BES database. Once thiswas completed, records in the BES data entry output fileswere matched and the keyed data formatted into theBES database. The records in the BES database weresubsequently batch edited for balance, consistency, andtolerance. The edit program ‘‘flagged’’ suspicious data forreview by subject analysts using customized, interactive,computer systems.

After editing, BES records underwent imputation formissing data. Computer programs calculated imputationfactors for all combinations of kind-of-business, legal formof organization, and tax status for selected types of orga-nization. Then, imputation rates were computed duringtabulation in order to measure the extent of imputationfor analysis and potential use in the suppression system.

Tabulation of the BES database produced simpleweighted estimates and adjusted estimates. The simpleweighted estimates were obtained by summing theweighted data (reported or imputed), where the weight fora given sampling unit was the inverse of its probability ofselection. Adjusted estimates were derived by benchmark-ing the simple weighted estimates to the SIC-based Busi-ness Annual Survey estimates. For merchant wholesaleand retail trades, these estimates included sales; invento-ries; purchases of merchandise for resale; value added;gross margins; cost of goods sold; net income produced atmarket prices; and net income produced at factor cost.The 1997 BES estimates for retail trade and service indus-tries included coverage for nonemployer firms. Estimatesfor these nonemployers were derived from informationprovided by other Federal agencies.

Finally, tabulated data were analyzed for reasonable-ness and consistency. Data that did not meet publicationstandards on the basis of this review, sampling error, ornonsampling error, were suppressed in the tables.

Tabulation and Sources of Related Data

The BES tables were combined with previously pub-lished data covering other trade areas and released on theInternet and CD-ROM as the Business Expenses report, partof the 1997 Company Statistics Series. The sources ofthese related data were the Annual Survey of Communica-tion Services, Annual Survey of Transportation and Ware-housing, Arrangement of Passenger Transportation AnnualSurvey, and the 1997 Economic Census coverage of manu-facturing, mineral industries, and construction sectors.

In addition, unpublished data on establishments auxil-iary to manufacturing, mining, and construction werecompiled: these operating expenses included supplemen-tal labor costs, energy costs, taxes and a variety of pur-chased services including advertising, legal and account-ing services, and repairs.

PUBLICATION PROGRAM—ECONOMY-WIDEPROGRAMS

Data compiled from the 1997 industry-wide programsare released as part of the Company Statistics Series. TheCensus Bureau publishes the following seven reports inthe EC97-CS series, in printed volumes—

• EC97CS-1, Company Summary.

• EC97CS-2, Women-Owned Business Enterprises.

• EC97CS-3, Black.

• EC97CS-4, Hispanic.

• EC97CS-5, Asians and Pacific Islanders.

• EC97CS-6, American Indians and Alaska Natives.

• EC97CS-7, Summary.

Each of these reports includes data by industrial classi-fication and/or geographic areas (states, metropolitanareas, counties, and places), size of firm (employment and



receipts), and legal form of organization (individual propri-etorships, partnerships, and corporations). All seven ofthese reports were made available at the Census Bureau’sInternet site ‘‘www.census.gov/csd/mwb,’’ and also werepublished on CD-ROM.

The eighth report in the series—EC97CS-8, BusinessExpenses, was issued on CDROM and was available on theCensus Bureau’s Internet site. This report shows operating

expenses data at the national level for SIC-based merchantwholesale trade, retail trade, service industries, communi-cations industries, transportation and warehousing, andtravel industries. In addition, NAICS-based data are pre-sented for manufactures, mineral industries, and construc-tion industries.



Chapter 12 PUBLICATION PROGRAM

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118Publication Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118The Economic Product Strategy Team . . . . . . . . . . . . . . . . . . . 1181995 Survey of Economic and Agriculture Data Users . . . . . . . . . . 118

Electronic Dissemination . . . . . . . . . . . . . . . . . . . . . . . . . . . 119Background information . . . . . . . . . . . . . . . . . . . . . . . . . . 119CD-ROMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119Internet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120Print-on-demand service . . . . . . . . . . . . . . . . . . . . . . . . . . 120

1997 Economic Census Publications . . . . . . . . . . . . . . . . . . . . . . 120Core Business Statistics Series . . . . . . . . . . . . . . . . . . . . . . . . 120Advance Report. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120Comparative Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . 120Bridge Between NAICS and SIC . . . . . . . . . . . . . . . . . . . . . . . 120Nonemployer Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

Sector Report Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121Industry series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121Geographic area series . . . . . . . . . . . . . . . . . . . . . . . . . . . 121ZIP Code statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121Summary series. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

Company Statistics Series . . . . . . . . . . . . . . . . . . . . . . . . . . 1241997 Economic Census of Outlying Areas . . . . . . . . . . . . . . . . . . 124Puerto Rico—Geographic Area Series (OA97E-1) . . . . . . . . . . . . . 124Puerto Rico—Subject Series (OA97E-2). . . . . . . . . . . . . . . . . . . 125Puerto Rico—Construction (OA97E-3) . . . . . . . . . . . . . . . . . . . 125Puerto Rico—Manufacturing (OA97E-4) . . . . . . . . . . . . . . . . . . 125Virgin Islands of the United States (OA97E-5) . . . . . . . . . . . . . . . 125Guam (OA97E-6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125Northern Mariana Islands (OA97E-7) . . . . . . . . . . . . . . . . . . . . 125

Reference Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125Preview to the 1997 Economic Census . . . . . . . . . . . . . . . . . . 125Guide to the 1997 Economic Census . . . . . . . . . . . . . . . . . . . 1251997 North American Industry Classification System—United States . . 125Accountant’s Guide to the 1997 Economic Census (EC97-PR2) . . . . . . 125Preparing for the 1997 Economic Census, Advance Information YouCan Use (EC97-PR1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126Numerical List of Manufactured and Mineral Products, Reference Series(EC97M31R-NL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126The Economic Census—Two Moments of Truth: 1954 and 1997(EC97X-TMT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

CONTENTS


Publication Program

INTRODUCTION

The U.S. Census Bureau published data for all sectorscovered by the 1997 Economic Census, as well as datafrom the 1997 Commodity Flow Survey (CFS), 1997Vehicle Inventory and Use Survey (VIUS), 1997 Surveys ofWomen- and Minority-Owned Businesses (SWOBE andSMOBE), 1997 Business Expenditures Survey (BES), and the1997 Economic Census of Outlying Areas—Puerto Rico,Guam, the Commonwealth of the Northern MarianaIslands, and the Virgin Islands of the United States. Thesedata were published on the Internet and were made avail-able on CD-ROM. Some, but not all, were published inprinted reports.

Given the focus on electronic products for 1997 censusdata, the standardization of data presentation was critical.This was especially true for data disseminated by Internetor CD-ROM, where there is the potential for instant datacomparison. Thus, sector-by-sector conformity was a keyfeature of the 1997 publication program, in terms of (1)methods of displaying geographic data, (2) units of mea-sure, and (3) allowing the detail of publication to be drivenby the data rather than predetermined. By contrast, certainartifacts of printed reports (such as brackets, underlines,or data-driven footnotes) that previously varied by sectorwere eliminated for 1997. As a result data presentationwas greatly enhanced.

Publication Planning

The Economic Product Strategy Team. To prepare forthe release of 1997 data, the Census Bureau establishedseveral planning teams to study the various aspects of dis-semination and product planning. The Economic ProductsStrategy Team (EPST) was organized to develop plans,strategies, and time frames for publishing the1997 data inelectronic reports.

The EPST recommended that the Census Bureaudevelop a new Core Business Statistics Series and elimi-nate printed reports for all but national summaries.Following reports by the subteams cited below, the EPSTalso recommended that, for the 1997 Economic Census,the Census Bureau—

• Publish 1997 data more quickly than data for 1992.

• Simplify data access, media, and software to appeal to alarger audience.

• Use the Internet as a primary tool for data dissemina-tion.

• Publish CD-ROMs quarterly with new data, as these databecome available.

• Provide a print-on-demand (or fax-on-demand) servicefor those users not prepared to use electronic-only data.

The team also recommended the Census Bureau estab-lish additional working groups to address specific EPSTobjectives. Accordingly, the agency organized two EPSTsubgroups—

• Economic Product Strategy Team-Subgroup 1developed plans for a well coordinated data dissemina-tion system, processing flow, and data content stan-dards for the 1997 Economic Census.

• Economic Product Strategy Team-Subgroup 2developed design concepts for software to access prod-uct data on CD-ROMs and on the Internet. In responseto findings from the user survey discussed below, thesubgroup also developed the concept of ‘‘quick reports’’to easily address the most basic user inquiries aboutgeographic areas or industries, and specified the cre-ation of an Economy-Wide Key Statistics file to makegeography quick reports feasible.

Later, additional teams were organized to develop censusproducts:

• Economic Product Team (EPT) prepared tables, text,Portable Document Format (pdf), and electronic fileguidelines for the 1997 Economic Census report series;finalized the production flowchart; initiated and moni-tored the subject divisions’ specification preparations;supervised the incremental development of metadatafiles; developed a production system; and oversaw theproduction of 1997 data products.

• Economic Products Internet and CD-ROM Teamdetermined the functional requirements for the CD-ROMand Internet access software. It also developed userinterfaces and search/query/output routines forCD-ROM products.

1995 Survey of Economic and Agriculture DataUsers. The EPST recommendations were supported bycustomer feedback obtained from the 1995 Survey ofEconomic and Agriculture Data Users, sent to 5,147 datausers. (Response statistics are shown in Table 12-1.) TheCensus Bureau drew the sample for the survey from itsown mailing list, as well as from sales records of theGovernment Printing Office (GPO). The agency’s mailing



list included subscribers to 1992 Economic CensusCD-ROMs, other users known to Census Bureau special-ists, and various intermediaries including census regionaloffices, State Data Center lead agencies, and CensusInformation Centers.

Table 12-1.1995 Survey of Economic and Agriculture DataUsers Response

Total distribution 5,147Postmaster returns 757Out-of-scope 7Mail response 1,031Response rate 23.5 percent

The National Processing Center (NPC) in Jeffersonville,IN,1 assembled and mailed the survey packages. Eachpackage included a cover letter, questionnaire, andpostage-paid return envelope. A promotional message,‘‘Your Opinion Counts,’’ was overprinted on the outboundenvelope to encourage recipients to respond. Mailout tookplace on August 4, 1995, with a request for response byAugust 15.

There was one follow-up mailing. A private contractorreprinted the survey forms and other materials, assembledthe packages, and mailed them to nonrespondents onOctober 10, 1995. Questionnaires received at the NPC byDecember 1, 1995, were included in the final results.

Written responses supplied by survey participants sug-gested that the majority of data users wanted the CensusBureau to make more data available, in a more timelymanner, regardless of the medium of publication. Respon-dents indicated that, in 1995, they primarily accessed datausing the agency’s printed publications. They alsobelieved that, within 5 years, they expected to access datain an electronic format, and they asserted that electronicdissemination would be most accessible by CD-ROM andthe Internet.

Electronic Dissemination

Background information. The Census Bureau tradition-ally published its major data findings in printed reports.The introduction of computers, and their use by both pub-lic and private data users, led the agency to begin publish-ing data in electronically readable formats. These datawere published first on reels of computer tape, and later,as technology developed, on diskettes and CD-ROMs. TheCensus Bureau also used commercial online vendors, and,finally, the Internet. Although most 1992 Economic Census

reports were published in both printed and electronic for-mats, 1997 data publication was primarily electronic. Elec-tronic media appeal to data users for three basic reasons:

• Electronic dissemination permits easier manipulation ofdata. From 1972 to 1992, while data on tape or CD-ROMmostly duplicated those available in print, users chosecomputerized data for its ability to be manipulated.These users employed tapes and CD-ROMs to rank,compare, analyze, reformat, and extract data. AsCD-ROM software became easier to use, novice usersbegan using these platforms for data analysis.CD-ROM’s provided data for all sectors on a singledevice, whereas a number of separate reports wererequired to disseminate the same data in print. The1997 CD-ROM included search engines to further helpusers find data.

• More data series were available to users in electronicformat. While most data from 1972 to 1992 appeared inboth print and computerized media, the latter wereunique sources for several databases too large to beprinted cost-effectively. These included ZIP-Code Statis-tics and sub-national Merchandise Line Sales. Theincreasing cost of printed media, coupled with increas-ing acceptance of CD-ROMs and the Internet, led to themajority of data being released only in a standardizedelectronic format. In the 1997 Economic Census, morethan 90 percent of the data released were available onlyin electronic format.

• Electronic dissemination allows faster distribution. TheInternet provides immediate access to data as soon asthey are released to the public, thus bypassing delaysassociated with the preparation of printed media, andsurpassing even the speed of CD-ROM delivery.

Economic census products were sold either by the GPOor by the Census Bureau’s Customer Service Center. Themajority of printed products were sold through the GPO.CD-ROMs, reports printed from PDFs, maps, and thoseprinted reports the GPO elected not to sell were soldthrough the Census Bureau.

CD-ROMs. A single CD-ROM holds up to 680 millionbytes of data. Software on each database CD-ROM enabledthe user to select, reformat, merge, and rank numbers,and then export data to a spreadsheet or other applicationfor further manipulation. The CD-ROMs were ‘‘plug andplay’’ and self-documenting. Use of the access software onthe CD-ROM required an IBM compatible,2 Pentium com-puter using a Windows 95, Windows 98, or Windows NToperating system, and having 32 megabytes of RAM, aswell as 10 megabytes of available hard-drive space.

Software programs in 1987 and 1992 were menu-driven, but lacked a graphic user interface. CD-ROMs for1992 included four separate programs for simple data

1Prior to a reorganization effective May 1, 1998, this unit wascalled the Data Preparation Division (DPD). 2Applications for Macintosh platforms were not produced.



viewing, data extraction, printing of multisector profiles,and documentation. By contrast, the 1997 CD-ROM soft-ware integrated all these functions. The addition of asearch function assisted users not yet familiar with thenew North American Industry Classification System(NAICS) structure. Data files on 1997 CD-ROMs were rede-signed to facilitate analysis of the entire economy versusanalysis of a single sector. Standard hierarchic industrycodes were used, rather than internal codes. Units of mea-sure also were standardized across sectors. Publicationcriteria were made more flexible to allow presentationof data for more industries, particularly those most impor-tant in each locality. ‘‘Quick Reports’’ were added tosimplify data retrieval for the novice user.

Data from the 1997 Economic Census were publishedfrom the first quarter of 1999 to the second quarter of2001. The first CD-ROM, issued in September 1999,included only those relatively few industries and statespublished to that date. This product was reissued onceevery quarter, with successive CD-ROMs progressivelymore inclusive. Most CD-ROMS were sold exclusively bysubscriptions. These subscriptions were to includeroughly 8 quarterly issuances of the NAICS-based CD-ROMpackage (6 of which included two discs), and 3 SIC-basedCD-ROMs. ZIP Code data were sold separately on a singleCD-ROM.

Internet. Data users were able to access 1997 EconomicCensus data via the Census Bureau’s Internet website athttp://www.census.gov. Economic Census data were fea-tured at http://www.census.gov/econ97. To access datavia the Internet, data users needed only possess a com-puter, and to subscribe to an Internet service provider.Internet data could be viewed, manipulated, and printedusing browsers on Windows, Macintosh, UNIX, and otheroperating systems.

Data users could access, free of charge, the two dataformats—

• Database. The database access system, known asAmerican Factfinder, had similar functions to thosefound in the CD-ROM software, including a searchsystem and ‘‘quick reports.’’

• Viewable. Viewable formats allowed users to view orprint tables similar to the detailed printed reports thatwere issued after previous economic censuses.

Print-on-demand service. To address the needs ofusers who may not have been prepared to use data inelectronic form, the economic census planners workedwith the Market Services Office (MSO) to develop a print-on-demand service. The MSO’s Customer Services Branchdeveloped the capability to print publications from thePDFs on the Internet, with covers of heavy stock and arudimentary binding. Demand was relatively low becausepublications produced this way were relatively expensive

(roughly 10 cents per page plus $20 for handling and cou-rier delivery), and because most users eventually discov-ered that electronic access met their needs.

1997 ECONOMIC CENSUS PUBLICATIONS

Core Business Statistics Series

Data collected in the 1997 Economic Census were pub-lished in several nationwide, economy-wide reportsbetween 1999 and 20013 (see Tables 12-2 and 12-3).

Advance Report. The Advance Report, issued in March1999, initiated the publication program with a preliminarylook at 1997 census data. Information contained in thisreport provided data at the national level only, by sectorand subsector. This report contained the first data everpublished on a NAICS basis. While the NAICS table waslimited to 1997 data, a separate SIC-based table showedfigures for both 1992 and 1997, thus making possible thecalculation of percent change in sales or employment overthe 5-year period. The NAICS-based data contained withinthe Advance Report were superseded as more detailedgeographic area and industry reports were published. TheSIC-based data were superseded with the publication ofComparative Statistics.

Comparative Statistics. Once basic data were pub-lished on a NAICS basis, national and state totals wererecast according to the 1987 SIC system. The ComparativeStatistics report superseded the preliminary two-digit SICstatistics in the Advance Report. Comparative Statisticscontained two-, three-, and four-digit SIC summaries.4

Bridge Between NAICS and SIC. The Bridge BetweenNAICS and SIC presents 1997 data cross-tabulated by bothold and new classification systems identifying the lowestcommon denominators between the two systems.

Conceptually, the bridge tables reflected the same com-binations shown in the correspondence tables of the 1997NAICS Manual. However, each line was extended to showthe number of establishments, sales or receipts, payroll,and employment associated with a particular NAICS andSIC combination. From the Bridge Between NAICS and SIC,data users were able to calculate percentages for alloca-tion strategies to assemble time series before and after1997. Because the Bridge Between NAICS and SIC provideda key to translation between NAICS and SIC, it is publishedin both printed and electronic formats.

Nonemployer Statistics. Nonemployer statisticssummarized the number of establishments, and sales orreceipts, of companies with no paid employees. These

3At this writing, July 2000, what follows was the CensusBureau’s projected publication schedule.

4No SIC-based data were published for metro areas, counties,or cities, due to the complexity of evaluating the potential for dis-closure of confidential company information stemming from theinterplay of NAICS and SIC classifications.



nonemployers are typically self-employed individuals orpartnerships operating businesses that they have chosennot to incorporate. Although nonemployers account forapproximately 3 percent of the Nation’s business activity,they represent nearly 75 percent of all businesses. As aresult, data from these businesses was crucial for theanalysis of the Nation’s economy.

Payroll tax records are at the heart of the CensusBureau’s system for keeping track of businesses betweencensuses. However, because nonemployers (firms not cov-ered by payroll tax) do not file such tax records, they arenot sent census questionnaires. Therefore, nonemployersare not reflected in other core business statistics ordetailed sector-specific reports.

Sector Report Series

The Census Bureau also published report series for indi-vidual sectors (see Table 12-3). Industry-specific data werepublished in the following reports:

Industry Series. Industry statistics, product statistics,and materials consumed reports were issued for six-digitNAICS industries in mining, construction, and manufactur-ing. While most data were reported in terms of NAICS,these reports show limited data for the SIC Codes, or partsthereof, that corresponded to the subject six-digit NAICS.More complete and definitive data about the relationshipbetween NAICS and SIC Codes were published in theBridge Between NAICS and SIC.

Geographic Area Series. The geographic area serieswere issued for each sector, most containing 52 reports aU.S. summary report and one report for each state and the

District of Columbia. There was no U.S. summary reportissued as part of the geographic area series for mining,construction and manufacturing, and for mining Delawareand the District of Columbia were combined in a singlereport to avoid disclosing information about a small num-ber of companies in the District.

Subject Series. Subject series reports were published foreach sector to add further dimension to the analysis ofindustries. Subject series reports, listed in Table 12-3,included topics such as merchandise line sales, establish-ment and firm size, and miscellaneous subjects.

ZIP Code Statistics. ZIP Code statistics were publishedfor selected sectors. These reports presented counts ofestablishments, by size, by detailed NAICS classifications,for five-digit ZIP Codes. These statistics included the countof establishments in several employment-size categories(fewer than 5 employees, 5 to 9, 10 to 19, 20 to 49, 50 to99, and 100 or more employees). For sectors other thanmanufacturing, establishment counts were also providedby sales- or receipts-size class by NAICS classification, andsummary total sales, employment, and payroll withoutregard to the NAICS. Establishment counts by size classwere not considered to be a disclosure of confidentialinformation, and thus could be shown for areas as smallas ZIP Codes.

Summary Series. After the publication of all industry,area, and subject reports for each sector, a single volumewas published in print that summarized the most widelyused statistics about that sector from all the previouslyissued reports.

Table 12-2.1997 Economic Census Publication Schedule, by Quarter

1999 2000 2001

Quarter of Publication . . . . . . . . . . 1 2 3 4 1 2 3 4 1 2Advance Report . . . . . . . . . . . . . . . . XIndustry Series . . . . . . . . . . . . . . . . . X X XGeographic Area Series . . . . . . . . . X X X X XComparative Statistics . . . . . . . . . . XBridge Between NAICS and SIC . XSubject Reports . . . . . . . . . . . . . . . . X X X XSummary Reports . . . . . . . . . . . . . . XZIP Code Statistics . . . . . . . . . . . . .Nonemployer Statistics . . . . . . . . . . XOutlying Areas . . . . . . . . . . . . . . . . . X X X X XBusiness Expenses . . . . . . . . . . . . . XMinority- and Women-OwnedBusiness Enterprises . . . . . . . . . . X X



Table 12-3Publications by Sector

Sector and Title Geography MediaProjected release date

(as of Mar. 2000)

Economy-WideCore Business Statistics:

Advance Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D,V,P Mar. 1999Comparative Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, U D, V Jun. 2000Bridge Between NAICS and SIC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V, P Jun. 2000Nonemployer Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D 3rd qtr. 2000

Auxiliary Establishments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V Feb. 2000Company Statistics:

Minority- and Women-Owned Business Enterprises . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V, P 3rd - 4th qtr. 2000Business Expenses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000

MiningIndustry Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V Jul. 1999 - Dec. 1999Geographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V Mar. 1999 - Apr. 2000Subject Series: General Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V, P 2nd qtr. 2000Subject Series: Product Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 3rd qtr. 2000Subject Series: Materials and Fuels Consumed . . . . . . . . . . . . . . . . . . . . . . U D, V, P 3rd qtr. 2000Location of Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D 1st qtr. 2000

UtilitiesGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V Jan. 2000Subject Series: Revenue Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4thqtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 1st qtr. 2000Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 2nd qtr. 2000

ConstructionIndustry Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V Jul. 1999 - Jun. 2000Geographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V Mar. 1999 - Apr. 2000Subject Series: Industry Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V, P Jan. 2000Subject Series: Geographic Area Summary . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V, P 2nd qtr. 2000

ManufacturingNumerical List of Mfg/Mineral Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (NA) D, V, P Apr. 1999Industry Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V May 1999 - Jan. 2000Geographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V Apr. 2000Subject Series: General Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V, P 1st qtr. 2001Subject Series: Product Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V, P 1st qtr. 2001Subject Series: Materials Consumed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001Subject Series: Concentration Ratios. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V 1stqtr. 2001Subject Series: Class of Customer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V 1st qtr. 2001Location of Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D 2nd qtr. 2001ZIP Code Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z D 2nd qtr. 2001

Wholesale TradeGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V May 1999 - Jan. 2000Subject Series: Commodity Line Sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V 2nd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 3rd qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V 4th qtr. 2000Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001

Retail TradeGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V May 1999 - Jan. 2000Subject Series: Merchandise Line Sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V 2nd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 3rd qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V 4th qtr. 2000Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001ZIP-Code Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, Z D 2nd qtr. 2001

Transportation and WarehousingGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V Feb. 2000Subject Series: Revenue Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 2nd qtr. 2001Vehicle Inventory and Use Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V Oct. 1998 - Oct. 1999





(as of Mar. 2000)

InformationGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V Oct. 1999Subject Series: Sources of Receipts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001ZIP Code Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, Z D 2nd qtr. 2001

Finance and InsuranceGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V Feb. 2000Subject Series: Revenue Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 2nd qtr. 2001

Real Estate and Rental and LeasingGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V Jun. 1999 - Jan. 2000Subject Series: Sources of Receipts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1stqtr. 2001

Professional, Scientific, and Technical ServicesGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V Jul. 1999 - Dec. 1999Subject Series: Source of Receipts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001Zip Code Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Z D 2nd qtr. 2001

Management of Companies and EnterprisesGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V Feb. 2000Subject Series: Sources of Receipts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001

Administrative and Support, Waste Management and RemediationServicesGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, s, M, C, P D, V Aug. 1999 - Jan. 2000Subject Series: Sources of Receipts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001ZIP Code Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, Z D 2nd qtr. 2001

Educational ServicesGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V Sep. 1999Subject Series: Sources of Receipts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4 qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001ZIP Code Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, Z D 2nd qtr. 2001

Health Care and Social AssistanceGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V Oct. 1999Subject Series: Sources of Receipts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 3rd qtr. 2000

Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001ZIP Code Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, D D 2nd qtr. 2001





(as of Mar. 2000)

Arts, Entertainment, and RecreationGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V Sep. 1999 - Nov. 1999Subject Series: Sources of Receipts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001ZIP Code Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, Z D 2nd qtr. 2001

Accommodation and FoodservicesGeographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V May 1999 - Jan. 2000Subject Series: Merchandise Line Sales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001ZIP Code Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, Z D 2nd qtr. 2001

Other Services (Except Public Administration)Geographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M, C, P D, V Dec. 1999Subject Series: Sources of Receipts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S D, V 3rd qtr. 2000Subject Series: Establishment and Firm Size. . . . . . . . . . . . . . . . . . . . . . . . . U D, V 4th qtr. 2000Subject Series: Miscellaneous Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, M D, V 1st qtr. 2001Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U D, V, P 1st qtr. 2001ZIP Code Statistics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . U, S, Z D 2nd qtr. 2001

Outlying AreasPuerto Rico:

Geographic Area Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PR, C D, V, P 2nd qtr. 2000Subject Series. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PR D, V, P 2nd qtr. 2000Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PR, M, C D, V, P 2nd qtr. 2000Manufactures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PR, M, C D, V, P 2nd qtr. 2000

Virgin Islands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI, C, P D, V, P Nov. 1999Guam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GU, C D, V, P Sep. 1999Northern Mariana Islands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . NM, C D, V, P Mar. 1999

Key to Geography: U = United States, S = States, M = Metropolitan Areas, C = Counties, P = Places Z = ZIP Code area. Key to Media: D =Database, V = Viewable, P = Printed

Company Statistics Series

The Census Bureau publishes eight reports in itsCompany Statistics Series following the 1997 EconomicCensus. Six reports publish data on minority-owned andwomen-owned business enterprises, and the CompanySummary provides corresponding information on allbusinesses:

• EC97CS-1, Company Summary.

• EC97CS-2, Women-Owned Businesses.

• EC97CS-3, Black.

• EC97CS-4, Hispanic.

• EC97CS-5, Asians and Pacific Islanders.

• EC97CS-6, American Indians and Alaska Natives.

• EC97CS-7, Summary.

Each of these reports contain national data by SIC-based industry groups, size of firm (employment andreceipts), and legal form of organization. Totals were also

shown for states, metropolitan areas, counties and places.Because of their broad popularity, these seven reportswere issued in print, as well as in electronic form.

The eighth report in the Company Statistics seriesEC97CS-8, Business Expenses—provided data on operatingexpenses of retailers, merchant wholesalers, and servicefirms (as classified under the SIC) and auxiliary establish-ments of manufacturing, mining, and construction firms(as classified by the NAICS). Operating expenses includesupplemental labor costs, energy costs, taxes and a vari-ety of purchased services including advertising, legal andaccounting services, and repairs. (Auxiliary establishmentsprovide services to other parts of the same company.)

1997 Economic Census of Outlying Areas

Puerto Rico—Geographic Area Series (OA97E-1).This report presented data for the wholesale and retailtrade and service industries. It included kind-of-businessdetail on number of establishments, sales or receipts, pay-roll, employment, proprietors and partners working, andlegal form of organization. Data were presented for Puerto



Rico as a whole, for commercial regions, and for municip-ios. Retail trade data included total and selling floor space,franchise holders, and class of customer for selected kindsof business. Wholesale trade data on inventories, operat-ing expenses, class of customer, and employment byprincipal activity were presented by kind-of-business. Datafor hotels and motels by type of receipts and number ofrooms were published for service industries.

Puerto Rico—Subject Series (OA97E-2). The subjectreport presented commodity and merchandise line salesdata for retail and wholesale trade by kind-of-business forPuerto Rico.

Puerto Rico—Construction (OA97E-3). This report pre-sented summary, industry, and geographic area statistics.The industry chapter presented 1997 data for industrieson the number of construction establishments; value ofconstruction work; employment; proprietors and partnersworking; payroll; hours worked; payments to subcontrac-tors; payments for materials, components, and supplies;payments for power, fuels, and lubricants; payments forrental of machinery, equipment, and structures; valueadded; and capital expenditures during the year. Selecteddata were shown for selected industries by geographiclocation of establishments and by legal form of organiza-tion. Data also were provided for establishments byemployment size.

Puerto Rico— Manufacturing (OA97E-4). This reportincluded data on the number of establishments, employ-ment, proprietors and partners working, payroll, valueof shipments, value added by manufacturers, class ofcustomer, inventories, costs of materials, capital expendi-tures, and country of destination. Statistics were shown byindustry and geographic area.

Virgin Islands of the United States (OA97E-5). Thisreport presented data by kind of business for constructionindustries, manufactures, retail trade, wholesale trade,and service industries. In addition to data for the VirginIslands as a whole, data were presented for St. Thomasand St. John (combined to prevent disclosure problems),St. Croix and the towns of Charlotte Amalie, Chistiansted,and Frederiksted.

Guam (OA97E-6). This report presented data by kind ofbusiness for construction industries, manufacturing, retailtrade, wholesale trade, and service industries. Tables pre-sented data for Guam and its election districts.

Northern Mariana Islands5 (OA97E-7). This reportpresented data by kind of business for construction indus-tries, manufacturing, retail trade, wholesale trade, and ser-vice industries. Tables presented data for the NorthernMariana Islands and the four municipalities.

Reference Series

In addition to the above reports, the Census Bureaupublished a number of reference works to help locate datafor specific industrial, product, or geographic categories,or for additional information about the economic censusprograms.

Preview to the 1997 Economic Census. This bookletwas an overview of the content and dissemination ofproducts containing data collected in the 1997 EconomicCensus. The booklet contained a number of key referencecharts, all organized by NAICS sector: coverage of theeconomic census, data items (e.g., employmentsales/receipts), geographic coverage, and data products.Several special programs (i.e., minority- and women-owned businesses and transportation surveys) were alsodescribed.

Guide to the 1997 Economic Census. This Internet-only reference described the content and products ofthe Economic Census in much greater detail than waspossible in the Preview to the 1997 Economic Census.Its contents were kept up-to-date as new informationbecame available. The Guide to the 1997 EconomicCensus included descriptions of related monthly,quarterly, and annual data. The Guide was accessible atwww.census.gov/epcd/www/guide.html

1997 North American Industry ClassificationSystem—United States. Both printed and CD-ROMversions of the 1997 NAICS Manual were published underthe auspices of the U.S. Office of Management and Budget.In addition, most of the manual’s features were includedat the NAICS Internet site: http://www.census.gov/naics.The manual included—

• Detailed industry definitions.

• Tables showing correspondence between the 1997NAICS and 1987 SIC Codes.

• An alphabetical list of more than 18,000 types of busi-ness activities and their corresponding NAICS code. Asection on ‘‘Why businesses need NAICS’’ was includedin the printed and PDF versions of the manual.

Accountant’s Guide to the 1997 Economic Census(EC97-PR2). This guide was created to assist accountantsin completing economic census forms for their clients, andto stress the importance of responding to the census. Likeother guides, it described products, programs, and uses ofdata. It answered questions frequently asked by accoun-tants, explained what their clients would receive, andprovided sample census questionnaires.

5Commonwealth of the Northern Mariana Islands (CMNI). SeeChapter 9, ‘‘The Economic Census of Outlying Areas.’’



Preparing for the 1997 Economic Census, AdvanceInformation You Can Use (EC97-PR1). This bookletpresented information designed to help businesses pre-pare for the 1997 Economic Census. It contained samplesof census questionnaires; Census Bureau telephone num-bers and e-mail addresses by industry and trade; andinformation on how questionnaires could be obtained andanswered electronically. The booklet also noted data uses,and identified main topics in the census.

Numerical List of Manufactured and Mineral Prod-ucts, Reference Series (EC97M31R-NL). This publica-tion lists the detailed product classes and productspublished for the manufacturing and mining sectors inthe Industry Series reports.

The Economic Census—Two Moments of Truth: 1954and 1997 (EC97X-TMT). This booklet described the roleof the 1953 Watkins Commission Report in reinstating the1954 Economic Census. Used as a promotional item forthe 1997 Economic Census, it drew parallels betweenevents surrounding the 1954 and 1997 censuses.



Chapter 13. COMPLETING THE CENSUS BUREAU’S MISSION THROUGHDATA INTEGRATION

A Review of Our Journey . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

The Two-Fold Path: Collection and Dissemination . . . . . . . . . . . . . . . 129

Data Integration as Completion of the Mission . . . . . . . . . . . . . . . . 129

CONTENTS


Completing the Census Bureau’s MissionThrough Data Integration

A REVIEW OF OUR JOURNEY

The U.S. Census Bureau’s mission statement is quiteclear: ‘‘Be the pre-eminent collector, and provider, oftimely, relevant, and quality data about the people and theeconomy of the United States.’’ The foregoing chapters ofThe History of the 1997 Economic Census have shown howfar the Census Bureau has come in achieving its mission,particularly in collecting and providing economic censusdata.

Chapter 2, ‘‘The Players and The Plans,’’ showed that toplan and carry out the 1997 census with improvements inefficiency over the previous economic censuses there werea series of reorganizations among the divisions under theEconomic Directorate.

To better identify, classify, and measure the businessactivity of our economy, the United States—with the publi-cation of the 1997 Economic Census—was the first ofthree nations to apply the new North American IndustryClassification System (NAICS) (Chapter 3. ‘‘The Introduc-tion of the North American Industry Classification Sys-tem’’). The data quality stemming from this census wasexpected to surpass that of earlier enumerations as aresult of the more comprehensive and refined identifica-tion of economic sectors and sub-sectors.

Always mindful of both data user needs, as well aspotential respondent burden in supplying information onthe census questionnaires, Census Bureau staff wentthrough a lengthy process of determining which questionsto ask for the 1997 census (Chapter 4, ‘‘QuestionnaireDesign and Development’’). This process went to the heartof the agency’s mission to collect ‘‘ relevant and qualitydata.’’

Chapter 5, ‘‘Promoting Census and Product Awareness,’’showed how earlier promotional efforts were builtupon—and combined with use of the Internet— to pro-duce a state-of-the-art promotional campaign. This wasthe first economic census promotional effort that explicitlyconnected census awareness and product promotion.

‘‘Geographic Area Coding,’’ Chapter 6, detailed theexcellent system that has evolved across many censusesfor linking business activity to legally or administrativelyestablished, and statistically defined, geographic entities.

‘‘Preparations for Census Mailout,’’ in Chapter 7, show-cased the Census Bureau’s advances in data collectiontechniques. The new DocuPrint 4890 Highlight Laser Print-ing System was used to print questionnaires, follow-upletters, envelopes, and many other materials essential tocollecting information from millions of businesses. The

system eliminated many of the massive clerical assemblyline activities at the Census Bureau’s National ProcessingCenter and many, though not all, private printing con-tracts were no longer necessary because of the in-housecapacity of the DocuPrint system.

‘‘Data Collection and Processing,’’ Chapter 8, showedthat centralized collection activities were followed bydecentralized post-collection processing. The chapterdiscussed the Census Bureau’s latest techniques for datacheck-in, entry, and control file matching, as well as a hostof post-collection activities that made the data fit for tabu-lation and dissemination. New for the 1997 census wasthe development of a module, aptly named ‘‘Plain Vanilla’’for its generalized edit function, for post-collection dataprocessing.

Chapter 9, ‘‘Collecting the Census in the OutlyingAreas,’’ reflected the advances in census-taking discussedin Chapters 7 and 8. The Outlying Areas included PuertoRico, the Virgin Islands of the United States, Guam, andthe Commonwealth of the Northern Marianas Islands.

Chapter 10, ‘‘Transportation Surveys,’’ covered the 1997Vehicle Inventory and Use Survey and the 1997 Commod-ity Flow Survey, as well as the precursors to those sur-veys. Improvements were made to both surveys basedupon earlier attempts to accurately measure, respectively,the Nation’s trucking fleet, and the flow of goods andmaterials according to mode of transportation.

Chapter 11, ‘‘Economy-Wide Programs,’’ discussed theSurveys of Minority- and Women-Owned Businesses, whichwere economy-wide, as well as the Business ExpendituresSurvey (BES). The History of the 1997 Economic Censushas the distinction of having the most detailed explana-tion of the BES of any census document ever targeted to awide audience. The many associated surveys that leadinto the BES are referenced in this chapter.

Chapter 12, ‘‘Publication Program,’’ described theCensus Bureau’s expanded electronic data disseminationendeavors, with particular emphasize on the Internet andimproved CD-ROM capabilities. The titles of an array ofpre-census promotional and informational items, as wellas a vast offering of products formed the basis of thischapter.

Without question, it can be said that the CensusBureau’s continuing improvements in data collectiontechniques guaranteed that the 1997 Economic Censussurpassed all previous economic censuses in sophistica-tion.



THE TWO-FOLD PATH: COLLECTION ANDDISSEMINATION

The entire purpose of data collection is data dissemina-tion. The Census Bureau’s mission, as mentioned, is two-fold: collection and dissemination. Plans for disseminating1997 data were equally ambitious, as were the new,improved logistics of data collection. The 1997 dissemina-tion strategies were to serve as the precursor to futureCensus Bureau data dissemination endeavors.

The Census Bureau has taken pride in its award-winningwebsite at www.census.gov. PC Magazine sited the U.S.Census Bureau as among the ‘‘Top 100 Web Sites’’ in 1998,and Government Executive Magazine sited the U.S. CensusBureau as one of ‘‘GovExec’s Best Feds on the Web’’ in1999. But while awards for accomplishments capture asingle point in time, a premier data collection and dissemi-nation statistical agency must move with the times inorder to retain its celebrity and prestige.

Data users have been challenging statistical agenciesfor years to make their data easier to obtain, easier tounderstand, and more relevant to real world issues.Survey results show that data use is on the increase, asare needs for using multiple data sets. Moreover, theburgeoning popularity of the Internet has increased bothproficiency and expectations among data users. Amongthese expectations are the need for both tailored informa-tion delivery and dynamic data integration.

DATA INTEGRATION AS COMPLETION OF THEMISSION

It is in the data integration field that the Census Bureauhas the greatest opportunity to add value to the data itcollects, to consolidate its mission of being a pre-eminentdata provider, and to retain its reputation as a highlyrespected statistical agency.

‘‘Data integration’’ simply means the construction of anelectronic architecture that permits data users to accessdata from multiple data sets with minimum ‘‘mouse

clicks.’’ Currently, it is necessary for users first to familiar-ize themselves with how the Federal Government, theCensus Bureau, its programs, or its data files are orga-nized. The user then searches, parses, and compiles infor-mation into digestible formats across various data sets toarrive at an answer to a research question. The technologyexists, however, to free data users from these steps.

At this writing, the Census Bureau is on the threshholdof providing a modernized, customer-driven, cross-program, and cross-agency integrated data access anddissemination system. Although many technical issues willneed to be addressed and resolved before this systembecomes fully operational, the important point, from thestandpoint of the history of the 1997 Economic Census, isthat the Census Bureau has recognized the need for thecreation of such a system. To that end, the Census Bureauwill be incorporating data from the 1997 census, as wellas from Census 2000, into its new ‘‘Integrated StatisticalSolutions’’ system.

This system will provide users with answers to theirquestions—without the user first having to master thestructure of the Census Bureau’s, or any other source’s,data file organization. It should be mentioned that, asdeveloped, the system will incorporate more and moredata sets from within the Census Bureau, as well as datasets from other Federal, state and local sources. Thus,data integration is an important component of the revolu-tion in electronic data dissemination. It will enable datausers to realize the full value of data collected andprovided by the U.S. Census Bureau.

If the economic data the Census Bureau collects every 5years shows one fact, it is this: Success of an industryrests upon keeping up with the times—and with thecompetition. Whoever takes advantage of the technologyof the times and taps into public demands becomes thepre-eminent provider of the service. Entering the newmillennium, the Census Bureau is well-positioned for suchsuccess.



Appendix A.Provisions of Title 13, United States Code,Relating to the 1997 Economic Census

CHAPTER 1. ADMINISTRATIVE

Subchapter I—General Provisions

5. Questionnaires; number, form, and scope ofinquiriesThe Secretary shall prepare questionnaires, and shalldetermine the inquiries, and the number, form, andsubdivisions thereof, for the statistics, surveys, andcensuses provided for in this title.

6. Information from other Federal departmentsand agencies; acquisition of reports fromother governmental and private sources

(a) The Secretary, whenever he considers it advis-able, may call upon any other department,agency, or establishment of the Federal Govern-ment, or of the government of the District ofColumbia, for information pertinent to the workprovided for in this title.

(b) The Secretary may acquire, by purchase or other-wise, from States, counties, cities, or other unitsof government, or their instrumentalities, orfrom private persons and agencies, such copiesof records, reports, and other material as may berequired for the efficient and economical con-duct of the censuses and surveys provided for inthis title.

(c) To the maximum extent possible and consistentwith the kind, timeliness, quality and scope ofthe statistics required, the Secretary shall acquireand use information available from any sourcereferred to in subsection (a) or (b) of this sectioninstead of conducting direct inquiries.

7. Printing; requisitions upon Public Printer;publication of bulletins and reports

The Secretary may make requisitions upon the Pub-lic Printer for miscellaneous printing necessary tocarry out the provisions of this title. He may furtherhave printed by the Public Printer, in such editionsas he deems necessary, preliminary and other cen-sus bulletins, and final reports of the results of theseveral investigations authorized by this title, andmay publish and distribute such bulletins andreports.

8. Authenticated transcripts or copies of certainreturns; other data; restriction on use; disposi-tion of fees received

(a) The Secretary may, upon written request, furnishto any respondent, or to the heir, successor, orauthorized agent of such respondent, authenti-cated transcripts or copies of reports (or por-tions thereof) containing information furnishedby, or on behalf of, such respondent in connec-tion with the surveys and census provided for inthis title, upon payment of the actual or esti-mated cost of searching the records and furnish-ing such transcripts or copies.

(b) Subject to the limitations contained in sections6(c) and 9 of this title, the Secretary may furnishcopies of tabulations and other statistical materi-als which do not disclose the informationreported by, or on behalf of, any particularrespondent, and may make special statisticalcompilations and surveys, for departments,agencies, and establishments of the Federal Gov-ernment, the government of the District ofColumbia, the government of any possession orarea (including political subdivisions thereof)referred to in section 191(a) of this title, State orlocal agencies, or other public and private per-sons and agencies, upon payment of the actualor estimated cost of such work. In the case ofnonprofit agencies or organizations, the Secre-tary may engage in joint statistical projects, thepurpose of which [is] otherwise authorized bylaw, but only if the cost of such projects [is]shared equitably, as determined by the Secre-tary.

(c) In no case shall information furnished under thissection be used to the detriment of any respon-dent or other person to whom such informationrelates, except in the prosecution of alleged vio-lations of this title.

(d) All moneys received in payment for work or ser-vices enumerated under this section shall bedeposited in a separate account which may beused to pay directly the costs of such work orservices, to repay appropriations which initiallybore all or part of such costs, or to refundexcess sums when necessary.

Appendix A A-1History—1997 Economic Census

U.S. Census Bureau

9. Information as confidential; exception(a) Neither the Secretary, nor any other officer or

employee of the Department of Commerce orbureau or agency thereof, may, except as pro-vided in section 8 of this title—(1) use the information furnished under the pro-

visions of this title for any purpose otherthan the statistical purposes for which it issupplied; or

(2) make any publication whereby the data fur-nished by any particular establishment orindividual under this title can be identified;or

(3) permit anyone other than the sworn officersand employees of the Department or bureauor agency thereof to examine the individualreports.

No department, bureau, agency, officer, oremployee of the Government, except the Secre-tary in carrying out the purposes of this title,shall require, for any reason, copies of censusreports which have been retained by any suchestablishment or individual. Copies of censusreports which have been so retained shall beimmune from legal process, and shall not, with-out the consent of the individual or establish-ment concerned, be admitted as evidence orused for any purpose in any action, suit, orother judicial or administrative proceeding.

12. Mechanical and electronic developmentThe Secretary is authorized to have conductedmechanical and electronic development work as hedetermines is needed to further the functions andduties of carrying out the purposes of this title andmay enter into such developmental contracts as hemay determine to be in the best interest of theGovernment.

Subchapter II—Officers and Employees

23. Additional officers and employees

(c) The Secretary may utilize temporary staff,including employees of Federal, State, or localagencies or instrumentalities, and employees ofprivate organizations to assist the Bureau inperforming the work authorized by this title,but only if such temporary staff is sworn toobserve the limitations imposed by section 9 ofthis title.

24. Special employment provisions

(a) The Secretary may utilize the services of non-temporary employees of the Bureau (by assign-ment, promotion, appointment, detail, or other-wise) in temporary positions established forany census, for not to exceed the period during

which appropriations are available for that cen-sus. Whenever the Secretary determines thatthe services of an employee which have beenutilized under this section are no longerrequired in such a temporary position, he may,without regard to the provisions of any otherlaw, return the employee to a continuing posi-tion, with rank and compensation not less thanthat which he held in his last permanent posi-tion in the Bureau: Provided, That no employeeshall, by reason of his service in a temporaryposition under this subsection, lose the protec-tion of any law or regulation with respect to hisseparation, suspension, furlough, or reductionin rank or compensation below the level held inhis last permanent position in the Bureau. Ser-vice by a nontemporary employee in a tempo-rary position under this subsection shall becreditable for step increases (both periodic andlongevity) under title VII of the ClassificationAct of 1949, as amended, as though it were acontinuation of service in his last permanentposition.

(b) As used in this title with respect to appoint-ments or positions, ‘‘temporary’’ shall be con-strued to mean not in excess of one year, ornot in excess of the specific period duringwhich appropriations are available for the con-duct of a particular census, whichever is longer.No employee of the Bureau who holds only atemporary appointment within the meaning ofthis section shall be considered as other thanstrictly temporary for purposes of any otherprovision of law relating to separations, sus-pensions, or reductions in rank or compensa-tion.

(d) The Secretary may fix compensation on a piece-price basis without limitation as to the amountearned per diem, and payments may be madeto enumerators for the use of private automo-biles on official business without regard to sec-tion 4 of the Travel Expense Act of 1949, asamended (5 U.S.C. 837), but at rates not inexcess of the rates provided by that Act.

(e) The Secretary may authorize the expenditure ofnecessary sums for travel expenses of personsselected for appointment for attendance attraining courses held by the Department ofCommerce with respect to any of the work pro-vided for by law.

25. Duties of supervisors, enumerators, andother employees(a) Each supervisor shall perform the duties

imposed upon him by the Secretary in theenforcement of chapter 5 of this title in accor-dance with the Secretary’s orders and instruc-tions.

A-2 Appendix A History—1997 Economic Census

U.S. Census Bureau

(b) Each enumerator or other employee detailed toserve as enumerator shall be charged with thecollection in his subdivision of the facts andstatistics called for on such schedules as theSecretary determines shall be used by him inconnection with any census or survey providedfor by chapter 5 of this title.

CHAPTER 5. CENSUSES

Subchapter I—Manufactures, Mineral Industries,and Other Businesses

131. Collection and publication; five-year periodsThe Secretary shall take, compile, and publishcensuses of manufactures, of mineral industries,and of other businesses, including the distribu-tive trades, service establishments, and transpor-tation (exclusive of means of transportation forwhich statistics are required by law to be filedwith, and are compiled and published by, a desig-nated regulatory body), in the year 1964, then inthe year 1968, and every fifth year thereafter, andeach such census shall relate to the year immedi-ately preceding the taking thereof.

132. Controlling law; effect on other agenciesTo the extent that the provisions of this subchap-ter or subchapter IV of this chapter conflict withany other provision of this title or other law, per-taining to the Secretary or the Department ofCommerce, the provisions of this title shall con-trol; but nothing in this title shall be deemed torevoke or impair the authority of any other Fed-eral agency with respect to the collection orrelease of information.

Subchapter IV—Interim Current Data

182. SurveysThe Secretary may make surveys deemed neces-sary to furnish annual and other interim currentdata on the subjects covered by the censuses pro-vided for in this title.

Subchapter V—Geographic Scope, Preliminary andSupplemental Statistics, and Use of Sampling

191. Geographic scope of censuses

(a) Each of the censuses authorized by this chap-ter shall include each State, the District ofColumbia, the Virgin Islands, Guam, the Com-monwealth of the Northern Mariana Islands,and the Commonwealth of Puerto Rico, and asmay be determined by the Secretary, suchother possessions and areas over which theUnited States exercises jurisdiction, control, orsovereignty. Inclusion of other areas over

which the United States exercises jurisdictionor control shall be subject to the concurrenceof the Secretary of State.

(b) For censuses taken in the Virgin Islands,Guam, the Commonwealth of the NorthernMariana Islands, or any possession or area notspecifically designated in subsection (a) of thissection, the Secretary may use census informa-tion collected by the Governor or highest rank-ing Federal official, if such information wasobtained in accordance with plans prescribedor approved by the Secretary.

(c) If, pursuant to a determination by the Secre-tary under subsection (a) of this section, anycensus is not taken in a possession or areaover which the United States exercises jurisdic-tion, control, or sovereignty, the Secretary mayinclude data obtained from other Federal agen-cies or government sources in the censusreport. Any data obtained from foreign govern-ments shall be obtained through the Secretaryof State.

193. Preliminary and supplemental statistics

In advance of, in conjunction with, or after thetaking of each census provided for by this chap-ter, the Secretary may make surveys and collectsuch preliminary and supplementary statisticsrelated to the main topic of the census as arenecessary to the initiation, taking, or completionthereof.

195. Use of sampling

Except for the determination of population forpurposes of apportionment of Representatives ofCongress among the several States, the Secretaryshall, if he considers it feasible, authorize the useof the statistical method known as ‘‘sampling’’ incarrying out the provisions of this title.

CHAPTER 7. OFFENSES AND PENALTIES

Subchapter I—Officers and Employees

211. Receiving or securing compensation forappointment of employees

Whoever—

(1) receives or secures to himself any fee,reward, or compensation as a considerationfor the appointment of any person as super-visor, enumerator, clerk, or other officer oremployee of the Department of Commerceor bureau or agency thereof, referred to insubchapter II of chapter I of this title; or


U.S. Census Bureau

(2) in any way receives or secures to himselfany part of the compensation paid to anyperson so appointed— shall be fined notmore than $3,000 or imprisoned not morethan five years, or both.

212. Refusal or neglect of employees to performdutiesWhoever, being an employee referred to in sub-chapter II of chapter I of this title, and havingtaken and subscribed the oath of office, neglectsor refuses, without justifiable cause, to performthe duties enjoined on such employee by thistitle, shall be fined not more than $500.

213. False statements, certificates, and informa-tion(a) Whoever, being an officer or employee referred

to in subchapter II of chapter I of this title,willfully and knowingly swears or affirmsfalsely as to the truth of any statementrequired to be made or subscribed by himunder oath by or under authority of this title,shall be guilty of perjury, and shall be finednot more than $2,000 or imprisoned not morethan five years, or both.

(b) Whoever, being an officer or employee referredto in subchapter II of chapter I of this title—(1) willfully and knowingly makes a false cer-tificate or fictitious return; or(2) knowingly or willfully furnishes or causesto be furnished, or, having been such anofficer or employee, knowingly or willfully fur-nished or caused to be furnished, directly orindirectly, to the Secretary or to any otherofficer or employee of the Department ofCommerce or bureau or agency thereof, anyfalse statement or false information with refer-ence to any inquiry for which he was autho-rized and required to collect information pro-vided for in this title—shall be fined not morethan $2,000 or imprisoned not more than fiveyears, or both.

214. Wrongful disclosure of informationWhoever, being or having been an employee orstaff member referred to in subchapter II of chap-ter I of this title, having taken and subscribed theoath of office, or having sworn to observe thelimitations imposed by section 9 of this title, pub-lishes or communicates any information, the dis-closure of which is prohibited under the provi-sions of section 9 of this title, and which comesinto his possession by reason of his beingemployed (or otherwise providing services) underthe provisions of this title, shall be fined not morethan $5,000 or imprisoned not more than 5 years,or both.

Subchapter II—Other Persons

221. Refusal or neglect to answer questions;false answers(a) Whoever, being over eighteen years of age,

refuses or willfully neglects, when requestedby the Secretary, or by any other authorizedofficer or employee of the Department ofCommerce or bureau or agency thereof actingunder the instructions of the Secretary orauthorized officer, to answer, to the best of hisknowledge, any of the questions on anyschedule submitted to him in connection withany census or survey provided for by subchap-ters I, II, IV, and V of chapter 5 of this title,applying to himself or to the family to whichhe belongs or is related, or to the farm orfarms of which he or his family is the occu-pant, shall be fined not more than $100.

(b) Whoever, when answering questions describedin subsection (a) of this section, and under theconditions or circumstances described in suchsubsection, willfully gives any answer that isfalse, shall be fined not more than $500.

223. Refusal, by owners, proprietors, etc., toassist census employeesWhoever, being the owner, proprietor, manager,superintendent, or agent of any hotel, apartmenthouse, boarding or lodging house, tenement, orother building, refuses or willfully neglects, whenrequested by the Secretary or by any other officeror employee of the Department of Commerce orbureau or agency thereof, acting under theinstructions of the Secretary, to furnish the namesof the occupants of such premises, or to give freeingress thereto and egress therefrom to any dulyaccredited representative of such Department orbureau or agency thereof, so as to permit the col-lection of statistics with respect to any censusprovided for in subchapters I and II of chapter 5of this title, or any survey authorized by subchap-ter IV or V of such chapter insofar as such surveyrelates to any of the subjects for which censusesare provided by such subchapters I and II, includ-ing, when relevant to the census or survey beingtaken or made, the proper and correct enumera-tion of all persons having their usual place ofabode in such premises, shall be fined not morethan $500.

224. Failure to answer questions affecting com-panies, businesses, religious bodies, andother organizations; false answersWhoever, being the owner, official, agent, personin charge, or assistant to the person in charge, ofany company, business, institution, establish-ment, religious body, or organization of any

A-4 Appendix A History—1997 Economic Census

U.S. Census Bureau

nature whatsoever, neglects or refuses, whenrequested by the Secretary or other authorizedofficer or employee of the Department of Com-merce or bureau or agency thereof, to answercompletely and correctly to the best of his knowl-edge all questions relating to his company, busi-ness, institution, establishment, religious body, orother organization, or to records or statistics inhis official custody, contained on any census orother schedule or questionnaire prepared andsubmitted to him under the authority of this title,shall be fined not more than $500; and if he will-fully gives a false answer to any such question, heshall be fined not more than $10,000.

225. Applicability of penal provisions in certaincases(a) In connection with any survey conducted by

the Secretary or other authorized officer oremployee of the Department of Commerce orbureau or agency thereof pursuant to subchap-ter IV of chapter 5 of this title, the provisionsof sections 221, 222, 223 and 224 of this titleshall apply—(1) with respect to the answering of questions

and furnishing of information, only tosuch inquiries as are within the scope ofthe schedules and questionnaires and ofthe type and character heretofore used inconnection with the taking of completecensuses under subchapters I and II ofchapter 5 of this title, or in connectionwith any censuses hereafter taken pursu-ant to such subchapters;

(2) only after publication of a determinationwith reasons therefor certified by the Sec-retary, or by some other authorized officer

or employee of the Department of Com-merce or bureau or agency thereof withthe approval of the Secretary, that theinformation called for is needed to aid orpermit the efficient performance of essen-tial governmental functions or services, orhas significant application to the needs ofthe public, business, or industry and is notpublicly available from nongovernmentalor other governmental sources;

(3) in the case of any new survey, only afterpublic notice, given by the Secretary orother authorized officer or employee ofthe Department of Commerce or bureau oragency thereof at least thirty days inadvance of requesting a return, that suchsurvey is under consideration.

(b) The provisions for imprisonment provided bysection 222 of this title shall not apply in con-nection with any survey conducted pursuantto subchapter II of chapter 3 of this title, or tosubchapter IV of chapter 5 of this title.

Subchapter III—Procedure

241. Evidence

When any request for information, made by theSecretary or other authorized officer or employeeof the Department of Commerce or bureau oragency thereof, is made by registered or certifiedmail or telegram, the return receipt therefor orother written receipt thereof shall be prima facieevidence of an official request in any prosecutionunder such section.


U.S. Census Bureau

Appendix B. HISTORICAL DEVELOPMENT OF THE ECONOMIC CENSUSOF THE UNITED STATES

History of the 1997 Economic Census . . . . . . . . . . . . . . . . . . . . . 2

The Economic Censuses in the Nineteenth Century . . . . . . . . . . . . . . 2The Beginnings of the Economic Census. . . . . . . . . . . . . . . . . . . 21810 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21820 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21830 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31840 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

The Expansion of the Censuses . . . . . . . . . . . . . . . . . . . . . . . 3Measuring the Success of the First Censuses . . . . . . . . . . . . . . . 31850 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41860-1879 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41880 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51890 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61900 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

The Twentieth Century . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6The Census Bureau and the Growth of the Economic Census . . . . . . . . 6The Permanent Census Act. . . . . . . . . . . . . . . . . . . . . . . . . 61905 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7The 1909 Economic Census of Puerto Rico . . . . . . . . . . . . . . . . 7The 1910 Census Act and Confidentiality . . . . . . . . . . . . . . . . . 71910 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71915 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Special Censuses in 1917 and 1918 . . . . . . . . . . . . . . . . . . . . 81920-1929 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81930 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91931-1933 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101935 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111937 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111940 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Post-War Economic Censuses. . . . . . . . . . . . . . . . . . . . . . . . . 121947 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121948 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13The Advent of the Electronic Computer in 1951 . . . . . . . . . . . . . 14The Watkins Commission: 1953-1954 . . . . . . . . . . . . . . . . . . . 141954-1955 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141958 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151963 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151967 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161972 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171977 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17The Creation of the Standard Statistical Establishment List (SSEL) . . . . 181982 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181987 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191992 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

CONTENTS

Appendix B B-1History—1997 Economic Census

Appendix B.Historical Development of theEconomic Census of the United States

HISTORY OF THE 1997 ECONOMIC CENSUS

Several obstacle were encountered as the CensusBureau began preparations for the 1997 Economic Census.As a result of budget reductions, the overall scope of the1997 Economic Census was reduced—the principal casu-alty being the Enterprise Statistics Program. At other timesduring the census design and planning, other areas forwhich the Census Bureau had collected data in the pastwere considered for deletion, including the mineral andconstruction industries, the Vehicle Inventory and Use Sur-vey, and the Surveys of Minority- and Women-Owned Busi-ness Enterprises, and the Outlying Areas (i.e., Puerto Rico,Guam, the Virgin Islands of the United States, and theCommonwealth of the Northern Mariana Islands). Fortu-nately, the Congress provided the majority of fundsneeded to continue collecting these vital data.

The increasing obsolescence of the Standard IndustrialClassification (SIC) Codes used by the Federal Govern-ment, and the passage of the North American Free TradeAgreement (NAFTA) produced tremendous pressure for acomplete overhaul of the SIC Code. The decision toreplace the SIC Code was evident by the mid-1990s, whenthe fastest growing industrial sector was ‘‘not elsewhereclassified.’’ This situation resulted in the development ofthe North American Industrial Classification System(NAICS), a cooperative effort between the Census Bureau,the Bureau of Economic Analysis, and the statistical agen-cies of Canada and Mexico (see Chapter 3). The NAICS,which was a single industrial classification system,replaced the three separate national systems used bythe NAFTA signatories. In February 1997, the Office ofManagement and Budget (OMB) directed the CensusBureau to adopt the NAICS for the 1997 Economic Census.

Contact with census data users indicated that publica-tion of census data in less expensive electronic formatswould be more useful than printed-format data and theaccompanying expenses required for their publication.The Census Bureau’s decision to use electronic dissemina-tion as the primary form of data required greater degreeof product standardization than in previous censuses, aswell as user-access software improvements. (Standardiza-tion across product lines was critical for cross referencingpermitted through the manipulation of electronic datafiles.)

The increasing popularity of the Internet suggested thatreports should be made available to data users visiting theCensus Bureau’s Web site. The 1997 Economic Census datawere released on CD-ROM and via the Internet. Reports

from the 1992 census were released in portable documentformat (pdf) and were available either via a CD-ROM or theInternet to ease the transition from printed to electronicreports.

THE ECONOMIC CENSUSES IN THE NINETEENTHCENTURY

The Beginnings of the Economic Censuses

1810. The third census of the United States was the firstthat included inquiries into the economy of the Nation. Inan act passed on May 1, 1810, the Congress directed,‘‘That it shall be the duty of the several marshals, secretar-ies, and their assistants aforesaid, to take, under the direc-tion of the Secretary of the Treasury, and according tosuch instructions as he shall give, an account of theseveral manufacturing establishments and manufactureswithin their several districts, territories, and divisions.’’The act did not outline any specific questions or prescribea schedule, leaving those matters to the Secretary of theTreasury’s discretion. To facilitate the collection of data,the U. S. Department of the Treasury divided manufac-tured products into 25 broad categories, encompassingmore than 220 kinds of goods. As the U.S. marshals andtheir assistants conducted the decennial census, they alsovisited the manufacturing establishments in their assignedareas to obtain economic data generally on the quantityand value of products manufactured. This enumerationwas conducted from August 1810 to July 1811, at a totalcost of $40,000. However, it was not until March 1812that Congress authorized $2,000 for the U. S. Departmentof the Treasury to prepare a statistical report on the kind,quantity, and value of goods manufactured and the num-ber of manufacturing establishments in each state, terri-tory, district, and county. The report, published in May1813, noted that there had been serious undercountingand omissions in the enumeration. Although the censusvalued total manufactures at about $173 million, theactual figure probably exceeded $200 million.

1820. The manufactures census of 1820 was similar tothat of 1810. However, in addition to a question on thelocation of establishments, 14 additional inquiries col-lected information on raw materials employed (kind, quan-tity, and cost), number of employees (men, women, andchildren), machinery (whole quantity, kind of machinery,and quantity of machinery in operation), expenditures(capital, wages, contingent expenses), and production

B-2 Appendix B History—1997 Economic Census


(nature and names of articles manufactured, value,demand, and sales). Federal marshals and their assistantscollected the data and published a digest of the returnsunder the auspices of the U.S. Department of State. Statis-tics appeared for each state, territory, and district, butthere was no attempt to compute U.S. totals because thedata were again incomplete. The deficiencies were attrib-uted to insufficient funds to pay the marshals and the factthat many establishments apparently neglected (orrefused) to provide the required information. Data from1810 and 1820 were not comparable because householdmanufactures (goods produced at home) were counted in1810, but not in 1820.

1830. The 1830 decennial census made no attempt toobtain economic data, partly as a result of incompletenessin the 1810 and 1820 censuses and perhaps because ofthe comparatively slow rate of economic growth in the1820s.

1840. Because of extensive growth in commercial fish-ing, commerce, and mining, and considerable pressure toresume the collection of economic statistics, the 1840census included a census of manufactures, as well asseries of questions about mining and fisheries. Marshalsand their assistants used a single questionnaire, ‘‘Scheduleof Mines, Agriculture, Commerce, Manufactures, Etc.,’’ tocollect data on these subjects. The census divided manu-facturing into 30 categories on the basis of the manufac-tured product (machinery, paper, furniture, etc.), andincluded an ‘‘all other manufactures’’ classification.

Census takers collected statistics on the quantity andvalue of goods produced, amount of capital invested,number of employees, and number of establishments.Mine operators were questioned following the classifica-tion of their mining activities (iron, lead, gold, salt, stone,etc.). These inquiries generally covered quantity and valueof materials produced, employment, and capital investedin mining operations. Commercial fishermen responded toquestions regarding quantity (and sometimes value) offish products, employment, and capital invested.

Commerce and trade activities accelerated during theearly and middle decades of the 19th century. Volume oftrade (total imports plus total exports) increased almosttwofold, from about $126 million in 1821 to approxi-mately $248 million in 1841. A series of questions on thenumber of business enterprises in various categories(commercial houses in foreign trade, commission houses,lumber yards, grocery stores, etc.), capital invested, andnumber of employees were added to the decennial censusin 1840 to measure the extent of commercial activities.

Federal marshals tabulated the returns for manufactur-ing, fishing, commerce, and mining and statistical tableswere published on the ‘‘commerce and industry of thecountry.’’ However, the 1840 census results, even though‘‘corrected’’ in Washington, reflected a considerable under-coverage of the national economy.

The Expansion of the Censuses

Measuring the Success of the First Censuses. Theattempts to collect economic data in the censuses of1810, 1820, and 1840 are considered to be of little valueexcept as indicators of the gross outlines of manufactur-ing development. The inauspicious beginnings of theeconomic census were the result of several interrelatedfactors. In addition to customary handicaps, such asgeographic dispersion, poorly defined boundaries, andinadequate transportation, collection of the economic datawas hindered by the following:

• The Federal marshals who supervised the field opera-tions had other duties and often could not devoteadequate time and attention to the census of economicactivity.

• The marshals’ assistants, the actual enumerators, oftenreceived sketchy instructions or none at all. Althoughthey might not normally have required detailed trainingor instructions to obtain answers to straightforwardpopulation questions (e.g., age and sex of members of ahousehold), they may have had problems elicitinganswers to more complex economic questions (such asthe cost of raw materials consumed in manufacturingduring the year).

• The marshals supervising the enumeration had tocompile and classify the data for their jurisdiction andprepare tables for publication. This decentralizationinevitably introduced some irregularities in the reportedinformation, because the marshals, who were nottrained statisticians, frequently used divergentprocedures in performing these duties.

• Many entrepreneurs, believing the government shouldhave a minimal role in economic affairs, were suspiciousand uncooperative when asked to provide informationabout their business to Federal agents.

To deal with nonresponse, the instructions given themarshals stated that:

• ‘‘...you (the marshals) will perceive the strong neces-sity for acquainting the people in advance with thenature of the inquiries to be made of them, and togive them time for preparation to answer the ques-tions promptly. . . . If (the inquiries) be made knownand generally understood before the enumerationcommences, the answers to the interrogatories willbe prepared in time for the domiciliary visit of theassistant, and the responses promptly made.’’

The instructions went on to say that:

• ‘‘Objections, it has been suggested, may possiblyarise on the part of some persons to give the statisti-cal information required by the act, upon the ground



of disinclination to expose their private affairs. Such,however, is not the intent, nor can be the effect, ofanswering ingenuously the interrogatories. On thestatistical tables no name is inserted -the figuresstand opposite no man’s name; and therefore theobjection cannot apply. It is, moreover, inculcatedupon the assistant that he consider all communica-tions made to him in the performance of his duty,relative to the business of the people, as strictlyconfidential.’’1

In other cases, the census takers found that the respon-dents’ records were inadequate or nonexistent, and theowners or operators simply could not provide more thansketchy estimates.

1850. As a result of the inaccuracies in the previouscensuses and to improve the 1850 census results, Con-gress passed an act on March 3, 1849, establishing theCensus Board. This Board consisted of the Secretary ofState, the Attorney General, and the Postmaster General.The act also provided for a full-time secretary who func-tioned as the director of the census, with responsibility fordesigning and printing the census schedules, and collect-ing and publishing data on manufacturing, mining, fish-ing, and commerce. The Census Board appointed JosephC. G. Kennedy, a statistician and newspaper editor-ownerfrom Pennsylvania, to serve as secretary of the Board in1849. The Secretary of the Interior formally appointedKennedy ‘‘superintendent’’ of the Census Office in 1850.Kennedy was largely responsible for the improved accu-racy of the census results.

Resigning in 1853, Kennedy became U.S. representativeto the International Statistical Congresses in Brussels, in1853 and 1854. The Secretary of the Interior rehiredKennedy in June 1858 to supervise the preparation of thereport on economic statistics. (Upon completing thatassignment, he was appointed superintendent of the 1860Decennial Census.)

The Census Board consulted with prominent statisti-cians in Federal Government and the academic and busi-ness communities to develop six questionnaires,2 one ofwhich (schedule 5) was for the collection of economicdata. This questionnaire was to be completed for each cor-poration, company, or individual having annual productionvalued at $500 or more for the year ending July 1, 1850.

Enumerators received written instructions andexamples of properly completed questionnaires, includingSchedule 5. Schedule 5 asked for the following data:

• ‘‘Name of business, manufacture, or product.’’

• Amount of capital invested in real and personal estate inthe business.

• Quantities, kinds, and values of raw materials used.

• Kind of motive power (water or steam), machinery,structure, or resource used in the manufacturingprocess.

• Average number of male and female workers employed.

• Average monthly cost of male and female labor.

• Quantities, kinds, and values of annual production.

Although Federal marshals still supervised field opera-tions, they no longer compiled and assembled statisticsfor publication. This work was performed by clerks in anoffice in Washington, DC. General census results werepublished in June 1853, and partial data was available formanufactures in September 1854.

Complete economic statistics for the 1850 census werenot released until December 1859. The total value ofmanufactures (including fisheries and the products ofmines), as reflected in the 1850 census, exceeded $1 bil-lion. This represented a 500-percent increase since 1810,compared to a growth in population of only 75 percent.Although there was an undercount, the economic anddecennial censuses were considerably more accurate thanthe previous censuses.

1860-1879. With the exception of some minor modifi-cations in the questionnaires, the censuses of 1860 weresimilar to those of 1850. The Census Office published fourvolumes of census data, including one devoted exclusivelyto economic statistics. A major innovation was Superin-tendent Kennedy’s analysis of the census statistics. Heused the data as a basis for describing ‘‘all the great ele-ments of a nation’s prosperity as they existed in the year1860.’’ In his section on ‘‘products of industry,’’ included inthe Preliminary Report on the Eighth Census, published in1862, Kennedy described the increasing impact of manu-facturing and commerce in the United States, estimatingthat one-third of the entire population in 1860 wassupported, directly or indirectly, by manufacturing. Hecompared various industries, and interpreted the data toshow how and why sections of the economy expanded,stabilized, or decreased, and illustrated interrelationshipsamong the production and socioeconomic factors.

As industrial specialization increased, it became moredifficult to develop general questions applicable to allestablishments. Unique questionnaires, tailored to thecharacteristics of each industry, had to be designed toobtain meaningful statistics. At the same time, theconcentration of economic power brought a growing pub-lic demand for Federal Government regulation of business.

1The 1840 census marked the first evidence of official concernwith the need to regard economic census data as confidential.Subsequent censuses had similar instructions.

2For a time the government used the words ‘‘schedule’’ and‘‘questionnaire’’ interchangeably to refer to the report form com-pleted either by a respondent or a census enumerator. In morerecent years, the Census Bureau has generally used the term‘‘questionnaire.’’



As the Government exercised increasing power overeconomic affairs, it required more and better data uponwhich to base its far-reaching decisions.

During 1869 and 1870, Congress attempted to draftnew census legislation to supersede the 1850 law underwhich the 1850 and 1860 censuses had been taken. Nonew law was approved, however, and the 1870 DecennialCensus was conducted following the 1850 law. Congressagreed to direct that the 1870 schedule for ‘‘products ofindustry’’ be modified to collect additional information onmachinery and machine power sources, cost of labor, andnumber of months the establishment was in operation.The inquiry concerning motive power and machinery wassubdivided so as to show specifically the kind of motivepower, horsepower (if steam or water), and the numberof machines used. The inquiry concerning the averagenumber of hands employed was revised to cover males(older than 16 years), females (older than 15 years), andchildren. Despite these changes, statistics on industry andwealth contained basic tabulations closely resemblingthose of 1850 and 1860. Other innovations for the 1870censuses included the use of tally machines (for the popu-lation schedules) and the introduction of maps and chartsto portray census results in a statistical atlas.

1880. In March 1879, the Congress passed a new censuslaw that implemented major changes in economic datacollection for the censuses of 1880. This law providedfor—

• Tailoring schedules to specialized businesses.

• Expanding the scope of the censuses.

• Using census supervisors and ‘‘experts,’’ appointed bythe President and confirmed by the Senate, to conductthe censuses (instead of Federal marshals, who hadbeen responsible for the enumeration since 1790). Forthe economic census, ‘‘experts’’ and ‘‘special agents’’(college professors, engineers, economic statisticians,and others with training and experience in the appropri-ate subject-matter areas) were authorized to collect,classify, and analyze statistics on manufacturing,mineral industries, or commercial fisheries in 279 largecities and towns.

• Implementing of a schedule of fines and/or imprison-ment for census workers who breached respondentconfidentiality, falsified census records, or participatedin other forms of misconduct.

For the 1880 census, the Census Office expanded thenumber of general economic questions to 29, anddesigned 49 special schedules, containing more than2,000 inquiries, for particular industries in the manufac-turing area. Although there was considerable overlap ofquestions (i.e., the same item appearing on several specialschedules, or on the general schedules and also on one ormore special schedules), more than 700 of the inquirieswere unique to an individual economic sector.

Whereas questions on mining and mineral industrieshad been included on the ‘‘products of industry’’ schedulesin the 1850, 1860, and 1870 censuses, the Census Officeused 42 special schedules to collect data on this subject in1880. These special schedules contained almost 3,000inquiries, including over 1,600 unique items. Those forcopper mines contained more than 120 questions, butthe median number of items on the special schedule forminerals was 53.

There were five special schedules for commercialfisheries. The 87 questions asked for information on theamount and kind of fish taken, type of equipment used,employment, and finances. There was a basic schedule forall commercial fishermen and more detailed schedules fordealers in fresh and salt fish, as well as two schedules forriver fishing activities.

Although some data on telegraph facilities was col-lected in 1850, the 1880 censuses marked the first majoreffort to compile detailed statistics on transportation andcommunications. Congress directed the Superintendent ofthe Census to collect and publish data on steam railroads,steamboat companies, incorporated express companies,and telegraph companies. Several special questionnaireswere used for this purpose. Most of the 375 questionswere designed to compile statistics on the railroads’ finan-cial and physical characteristics.3

The rise of organized labor and trade associationsprompted the Government, for the first and only time inthe economic censuses, to include a series of inquiries onlabor unions and trade societies (number, objectives andmethods, membership, receipts, and expenditures) and onthe number, location, causes, and results of strikes andlockouts. The census used 4 special schedules containinga total of 101 questions to collect this information.

The 1880 census also contained a special survey onwages and prices. Three schedules, with a total of 167questions, collected statistics on wages in the manufactur-ing industries and building trades, and on average retailprices of the ‘‘necessaries of life.’’

A few questions on the insurance industry (primarily onlife insurance) were included in the 1860 census. In 1880,the economic census covered life, fire, and marine insur-ance companies, using 15 special schedules to compile

3Financial characteristics included income, expenses, andanalysis of earnings, while physical characteristics encompassedtrack mileage, amount and kind of rolling stock (locomotive cars),and fuel used to power locomotives, etc. The Census Bureaurequested a detailed report from each railroad on virtually everyfacet of its operations, including timber conservation practicesand data on employees and others killed or injured in accidents.In addition to questions about railroads’ current characteristics,

there was a special inquiry on the history of their construction,from which the Census Office tabulated statistics on mileagebuilt and existent, by groups of states, for individual companies,annually from 1830 to 1880. The special schedules askedfewer and less detailed questions of steamboat companies(21), express companies (30), telegraph companies (24), andtelephone/telegraph companies (16).



statistics on their organization, finances, and insurance inforce. Census staff encountered considerable difficulty incollecting and tabulating the data and, except for a pre-liminary statement in the statistical compendium, noreport was issued on this subject.

The economic statistics compiled in the 1880 censuswere more comprehensive than in any previous census.The use of experts and special agents generally resulted inmore complete and consistent returns, which led to moreaccurate data on the Nation’s economy. The Census Officepublished the compendium summarizing census data inearly 1883, the basic census volumes reporting manufac-turing statistics and transportation-communications datain October 1883, and the mineral industries volume in July1886. Several special monographs (e.g., on water powerused in manufacturing; precious metals; and petroleum,coke, and building stone) appeared in the mid-1880’s. Thestatistical atlas was printed by a commercial publisher,once the Census Office resumed producing atlases follow-ing the 1890 census.

1890. The 1890 Censuses of Manufactures and MineralIndustries followed the basic procedures established forthe 1880 censuses; e.g., specialists and agents visitedestablishments in 1,042 manufacturing and commercialcenters, bulletins announced preliminary census results,and in many instances, the same schedule formats wereused. The 1890 censuses saw the first use of administra-tive records (those kept by governmental or privateorganizations) to compile economic census data. Congressdirected that statistics be collected on the recordedindebtedness (i.e., real estate mortgages) of private corpo-rations and individuals. Special census agents went to realestate recorders’ offices to collect information about mort-gages made during the period 1880 to 1890, whichincluded a description of the property, provisions of themortgage, and the addresses of the mortgagor and mort-gagee. Then, the Census Office mailed schedules (andpostage-paid return envelopes) to the mortgagors (or themortgagees if the mortgagors did not respond), that askedthem to supply additional information about their mort-gages (whether or not they had been fully paid and, if not,how much was owed) and return the completed schedulesto the Census Office.

The census of manufactures covered Alaska for the firsttime in 1890.4 The census considerably expanded thenumber of inquiries on transportation (first asked in

1880), with coverage extended to sailing vessels andrapid-transit facilities in cities (which later included cablerailways, railways operated by animal power, and electricstreet railroads).

1900. The census of 1900,5 which was limited by lawto duplicate that of the 1890 census, included coverage ofmanufactures, mines and quarries, street and electric rail-roads, and, for the first time, central electric light andpower stations. The census compiled manufacturing infor-mation on the number of establishments, capital invested,number of wage earners and total wages paid, cost ofmaterials, and value of products. Most manufacturingestablishments (about 530,000 of the 644,000 totalreturns were from manufacturers) reported information onthe general questionnaire, but there was still some criti-cism from respondents that the 32 special questionnairesrequired too much detailed information.

Manufacturing data in 1,340 cities and towns werecompiled by 1,891 specialists (supervised by 20 ‘‘expertspecial agents’’). To disseminate the results, the CensusOffice prepared 59 unique bulletins on various specialsubjects in manufacturing, including shipbuilding, slaugh-tering and packing, and lumbering. The Census Office alsoreleased bulletins on manufacturing for each of the statesand territories except Alaska and Hawaii.6 Manufacturingdata were published in four volumes and summarized inthe statistical abstract and atlas.

Statistics on mines and quarries, street and electric rail-roads, and central electric light and power stations wereinitially published in bulletins and later in final reports thatwere somewhat less detailed than the bound volumes onmanufactures. Data on mining covered the same generaltopics as before (number of mines and mine operators,wage earners and total wages, cost of supplies, otheroperating expenses, and quantity and value of mineralsextracted), by geographic area and by type of mineral.

For street and electric railroads, Census Office employ-ees assembled statistics on such topics as number of com-panies, length of rail lines, cost of construction and equip-ment, employees, and passengers. Data compiled forcentral electric light and power companies included num-ber of stations, cost of construction and equipment, earn-ings, expenses, and power generated.

THE TWENTIETH CENTURY

The Census Bureau and the Growth ofthe Economic Census

The Permanent Census Act. In March 1902, the Con-gress enacted the Permanent Census Act that established

4In the early 1880’s, Ivan Petroff, a special Census Officeagent, traveled to Alaska (in accordance with instructions fromthe Superintendent of the Census Bureau) to report on the num-ber of inhabitants. As a result of his work, he (1) produced a briefstatistical review of Alaska in geographic divisions, with tables ofpopulation, (2) provided a review of the fur trade, fisheries, min-eral, and agricultural resources, (3) described the geography andtopography of Alaska, (4) furnished a historical sketch of Alaskafrom its discovery to the year 1880, and (5) wrote notes on Alas-kan ethnology. However, this effort did not constitute an eco-nomic census of that territory.

5Statistics for manufacturing were for the year 1900. Data formining, street and electric railroads, and electric light and powerstations were for 1902, but the Census Office released theseresults as part of the 12th decennial census publication series.

6Data for Alaska and Hawaii appeared in the regular CensusBureau reports.



the Census Office as a permanent agency.7 The actincluded a provision directing ‘‘That in the year 1905, andevery 10 years thereafter, there shall be a collection of sta-tistics of manufacturing establishments....’’ This was inaddition to the coverage of manufactures in the decennialcensus, which meant that the new Census Bureau wouldconduct censuses of manufactures every 5 years (quin-quennially). The legislation also mandated several special-ized interdecennial censuses subsequently taken between1902 and 1937 that included street railways and tele-phone and telegraph companies. In 1922, the CensusBureau expanded the ‘‘street railways’’ category to includemotor buses and, in 1932, trolley buses. The ‘‘telegraph’’classification included only land telegraph and ocean cablesystems in 1902, but the Census Bureau enlarged the defi-nition for 1907 to include wireless systems. Censuses ofwater transportation were conducted in 1906, 1916, and1926. Other special enumerations included a census ofexpress businesses in 1907 and a census of commercialfisheries was taken in 1908 and again in 1963 and 1967.8

1905. To conduct the 1905 census, the Census Bureauconstructed a card index containing names and addressesof manufacturing establishments on the basis of 1900census results, city directories, trade publications, stateand local government lists, and similar sources. In October1904, the agency sent preliminary circulars to these estab-lishments, asking them to provide their establishmentsdata (name and address, period of operation, and kind ofbusiness). Based on the results of this precanvass, theCensus Bureau updated the card index and mailed ques-tionnaires to establishments in early December 1904.Beginning in January 1905, 835 canvassers (CensusBureau employees and temporary employees hired for theduration of the operations) visited nonrespondent compa-nies.

The schedules used in the 1905 census were the sameas those used in 1900. The agency redefined the scope ofthe census to cover only manufacturing establishmentsunder the ‘‘factory system,’’ excluding the neighborhoodand household industries and trades9 that had been previ-ously included. To provide comparability between cen-suses, the Census Bureau retabulated the 1900 manufac-tures census results using the new definition. Withneighborhood and household industries and hand tradesexcluded from the 1900 results (for 1899), the number ofmanufacturing establishments shrank from about 509,000

to approximately 205,000, and value added by manufac-ture was reduced from $5.5 billion to $4.6 billion.

The 1909 Economic Census of Puerto Rico. TheCensus Bureau conducted the first economic census ofPuerto Rico (for manufactures) for the year 1909. With theexception of 1929, a census was conducted in Puerto Ricoat 10-year intervals through 1949. Censuses of manufac-tures then were taken concurrently with censuses of retailand wholesale trade and selected service industries, forthe years 1954 and 1958. . In 1952 and 1956, the Com-monwealth Government of Puerto Rico conducted cen-suses of manufactures with more limited coverage thanthe stateside census. Since 1963, the census in PuertoRico has been part of the stateside economic censusprogram

The 1910 Census Act and Confidentiality. The 1910census act strengthened confidentiality protections, par-ticularly as they related to economic data. As a matter ofadministrative policy, responses on schedules for earliereconomic censuses had been considered confidential, butthe 1910 law specified that information furnished by busi-ness, manufacturing, and mining establishments:

• ‘‘...shall be used only for the statistical purposes forwhich it is supplied. No publication shall be made bythe Census Office whereby the data furnished by anyparticular establishment can be identified, nor shall theDirector of the Census permit anyone other than thesworn employees of the Census Office to examine theindividual reports.’’

The Census Bureau acknowledged the importance ofconfidentiality by noting in the volume containing 1910manufactures census results that:

• ‘‘It is essential to the success of the manufacturescensus that every concern should be assured explicitlyby law that its business will not be disclosed to com-petitors, to the general public, to state and local offi-cials, or even to officials of the Federal Government out-side of the Census Bureau. Only with such pledge ofconfidential treatment can the Census Bureau expectmanufacturers to furnish data promptly and accurately.’’

1910. For the 1910 census of manufacturing establish-ments, the Census Bureau used 1 general and 60 specialquestionnaires. The general schedule was materially sim-pler than that employed for the censuses of 1900 and1905, and the same was even more true of most of thespecial questionnaires. Establishments completed andreturned their questionnaires to the Census Bureau. Spe-cial agents conducted a nonresponse follow-up. For minesand quarries, one schedule collected data for use by theCensus Bureau and the U.S. Geological Survey. (This wasdone to decrease respondent burden, since both agenciescollected similar data). The Census Bureau assigned 1,227

7The name changed to the ‘‘Bureau of the Census’’ in 1903,when it became part of the new Department of Commerce andLabor. Legislation continued to refer to ‘‘the Census Office,’’ how-ever.

8The census of commercial fisheries was discontinued after1967, due to a lack of interest by data users.

9Examples of these industries include custom tailoring, car-pentering, taxidermy, and, in gerenal, contract work done forindividual customers. The census also excluded establishmentsthat produced products with an annual value of less than $500.



special agents and 76 Census Bureau employees to enu-merate factories, mines, and quarries. In sparsely settledareas, decennial census enumerators visited the manufac-turing and mining establishments.

For the first time, the 1910 Economic Census collecteddata from custom sawmills and gristmills and steam laun-dries. Under the definitions used for the time of the twoprevious censuses of manufactures, mills that did notproduce for sale, but only sawed lumber or ground grainfor toll, did not fall under the factory system. The censuscovered steam laundries because they had become animportant industry. As a result of consolidating somecategories, the number of separate industry tabulationspublished was reduced.

An amendment to the 1902 census act, passed Febru-ary 25, 1910, required the Census Bureau to enumeratethe number of animals slaughtered for food purposes andthe number of hides produced during the year. This neces-sitated a canvass of all butchering establishments, manyof which would not have been included under the generalrules defining the factory system.

1915. The 1915 Census of Manufactures sought theassistance of prominent manufacturers and of representa-tive commercial and trade groups to conduct the census.Additionally, the Census Bureau enlisted the Congress, theDepartment of Agriculture, the Bureau of Corporations,and state statistical organizations. The director of theCensus Bureau and the Chief Statistician for Manufacturesvisited cities (Philadelphia, New York, Boston, St. Louis,etc.) to obtain suggestions on the form and content of theschedules from economic census participants. CensusBureau employees tabulated 1914 data for 344 industriesand 271 industry subgroups. Plans called for the promptpublication of bulletins containing preliminary results, butthe United States’ entrance into World War I in April 1917,delayed the preparation of the printed statistics until late1918.

Special Censuses in 1917 and 1918. Because of theurgent need for industrial data during World War I, therewere a number of special economic censuses somethingnot done previously. For example, the agency compiledstatistics of New York’s daily landings of fresh catchesfrom U.S. fishing boats and receipts by rail and steamerduring the last 4 months of 1917. In late 1917 and early1918, the Census Bureau took a number of mail censusesfor the War Trade, War Industries, Shipping, FederalReserve, and Commercial Economy Boards, the FoodAdministration, and the Council of National Defense.These emergency censuses covered such commodities asiron and steel; wool machinery and woolen manufactures;kapok fiber, jute and silk; leather stocks and manufacturedleather goods; antimony and graphite crucibles; commer-cial greenhouses; materials used in the manufacture ofexplosives; and dental gold production.

1920-1929. The growing need for detailed economicstatistics prompted the Congress to direct the CensusBureau (in the act providing for the decennial census of1920) to conduct censuses of manufactures on a biennialbasis, to collect and publish economic statistics for theyears 1921, 1923, 1925, 1927 and for every tenth yearthereafter. Data were collected that would be of economicand sociological importance, such as (1) the size of estab-lishments and hours of labor, (2) the absolute and relativemagnitude of the various branches of industry and theirgrowth and decline, and (3) the industrial importance(with increase or decrease) of individual states and largecities.

In planning the 1921 census of manufactures, CensusBureau officials consulted the National Association ofManufacturers, the Census Bureau Advisory Committee(which included representatives of the American EconomicAssociation), civic associations (such as chambers of com-merce), trade associations, and government and private-sector statisticians. The recommendations made by theseorganizations and individuals helped the Census Bureaudesign questionnaires, publicize the census, and generallyencouraged industry’s acceptance of the project.

To reduce the expense of the census and expedite theprocessing and publication of results, the Census Bureauomitted certain items that had been included in previouscensuses of manufactures (capital invested, age and sexdistribution of employees, rent and taxes, primary horse-power used, and kind and quantity of fuel used in manu-facturing). Additionally, only limited statistics were col-lected on number of wage earners and value of productionfrom manufacturing establishments reporting productsvalued at less than $5,000 for 1921.10 In 1921, about 22percent of manufacturers had products valued at less than$5,000; however, these plants accounted for less than 1percent of all wage earners and total production.

The Census Bureau conducted a mailout/mailback cen-sus in 1921 to collect economic data on 358 separateindustries (98 of which were subdivided to show greaterdetail) on such things as the number of proprietors or firmmembers, number of salaried employees, number of wageearners, amounts paid in salaries and wages, amount paidfor contract work, and cost of materials. To tabulate thedata from manufacturers, mines, and quarries, the CensusBureau, for the first time in an economic census, used apunchcard tabulating system similar to the one introducedduring the 1890 decennial census for tabulating popula-tion data.

The Census Bureau issued preliminary summary reportsof census results as press releases, each relating to a par-ticular industry or group of industries. A summary bulletinpresented statistics for the United States by industries,and for all industries combined by state. Mining data were

10Previously, data had been obtained from establishments withannual production valued at $500 or more.



published to reflect geographic distribution of operations,land controlled by mining operators, characteristics oforganization, scale of operations, and amount and kind ofpower used. The general report of economic data includedanalytical tables and tabulations, selected by mineralindustry and state, so as to facilitate comparisons with themining statistics published annually by the U.S. GeologicalSurvey.

The procedures and coverage of the 1923 census werevirtually the same as those of 1921, but the mail opera-tion was more successful than in 1921, due in part to theexpanded cooperation of chambers of commerce. In manylarge cities, chamber representatives were sworn in ascensus agents and supervised the census in their area.Approximately 65 percent of the returns were received bymail and, by June 1924, almost 95 percent of establish-ments had been canvassed by mail or by personal enu-meration. The Census Bureau collected and tabulated datafor 333 industries, of which 87 were subdivided to pro-vide greater detail. Preliminary statistics first appeared inpress releases and final data were published in industrybulletins. A one-volume compendium report was pub-lished in January 1926.

The 1925 census covered 324 industries. The CensusBureau again sought the cooperation of local chambers ofcommerce and professional groups to encourage responseto the census. Respondents returned about 75 percent ofthe schedules by mail. Census staff prepared pressreleases of preliminary results, plus the usual industrybulletins, and released the one-volume final report inDecember 1927.

For the 1927 census, manufacturers of confectioneries,ice cream, and sheet iron completed questionnaires only iftheir annual production was valued at $20,000 or more.(This was in recognition of the fact that many of thesmaller firms were primarily engaged in retail trade, notmanufacturing.) Over all, the census encompassed 335industries, and approximately 65 percent of the respon-dents returned their schedules by mail. The previouspress-release, industry-bulletin, and final-volume publica-tion sequence was followed, with the one-volume finalreport published in April 1930.

1930. The economic component of the 1930 DecennialCensus was broader in scope than any previous census,encompassing censuses of manufactures and mineralindustries, construction industries, distribution (whichincluded retail and wholesale trades and special topics),and hotels. These censuses covered activities for the year1929.

The construction census was a response to the post-World War I boom in this sector of the economy. In 1920,less than 850,000 workers were employed in contractconstruction, and total private construction for that yearwas valued at $5.4 billion. By 1928 (the last full yearbefore the start of the Depression), more than 1.6 million

workers were employed in this field, and total private con-struction was valued at $9.2 billion.

When planning the first census of construction indus-tries, the Census Bureau consulted an advisory committeecomposed of representatives of national contractors’ asso-ciations and individual construction companies. This coop-erative effort produced one basic questionnaire designedto collect information on the following:

• Organization of the establishment.

• Number of salaried employees and total salaries paid.

• Number of skilled and unskilled workmen employed.

• Total annual wages.

• Length of working day and week.

• Expense for equipment, operation, and overhead.

• Total value of materials and building equipmentinstalled.

The construction census was a mailout/mailback opera-tion. The Census Bureau prepared a directory of construc-tion establishments from lists of names and addressesprovided by contractors’ associations, private statisticalagencies, builders’ exchanges, chambers of commerce,city officials and postmasters (who were asked to submitthe names and addresses of known construction establish-ments in their jurisdictions). Census Bureau clerksobtained additional names and addresses from telephonedirectories, city directories, and other sources. Ultimately,the census list included about 144,000 names andaddresses. The questionnaire was to be completed byestablishments engaged in construction business of anykind (except industrial concerns, public utilities, munici-palities, or common carriers that maintained constructioncrews to repair or maintain their own property). The Cen-sus Bureau followed up nonrespondents with two mailingsof reminder letters. In cities with populations of at least100,000, enumerators employed by the decennial censuscontacted nonrespondent establishments.

Detailed statistics were compiled only for establish-ments that reported gross receipts of at least $25,000 dur-ing the 1929 calendar year. A punchcard system mechani-cally tabulated construction census results (and, in fact,virtually all 1930 census data), but a series of clericalcross-checks and reviews by experienced statisticiansensured maximum accuracy and consistency. The resultswere published in a single bound volume in December1932.

The 1930 census of distribution consisted only of afield canvass. The retail trade census covered operationsof establishments involved in some manner of retail trade,including restaurants and some service businesses (suchas garages that sold merchandise in addition to their ser-vices). The agency classified retail establishments on thebasis of the following types of operation:



• Single-store establishments.

• Two- and three-store independents.

• Local branch systems.

• Local, sectional, and national chains.

• Miscellaneous types of operations.

Enumerators personally visited each of approximately1.5 million stores included in the census to complete oneof the six questionnaires designed for this operation. Incities with populations of at least 10,000, special enu-merators took the census of distribution. In smaller citiesand rural areas, the decennial census enumerators visitedretail outlets. The census of retail trade compiled data onthe number of stores, personnel, payroll, stocks, sales,operating expenses, seasonal employment characteristics,credit business, receipts from sales of meals and automo-tive services, value of returned goods and allowances, andtype of organization. The Census Bureau published pre-liminary results as press releases. A one-volume finalreport that included a summary of data by states, coun-ties, and incorporated places, and separate tables for eachstate was released in February 1933.

The census of wholesale establishments used the samefield enumeration procedures to complete one of fourquestionnaires for each of approximately 168,000 estab-lishments. The wholesale trade classification included allestablishments engaged in the purchase, sale, or distribu-tion of goods on a conventional wholesale basis, plusother special categories, such as cash-and-carry wholesal-ers, drop shippers (middlemen who secured orders frombuyers and had merchandise shipped directly from themanufacturer to the buyer), manufacturers’ salesbranches, and cooperative marketing associations. Virtu-ally all merchandising establishments not in the retailgroup were covered by the census of wholesale trade.

The census of wholesale trade compiled data on thefollowing:

• Number of establishments.


• Salaries and wages.

• Stocks.

• Net sales, credit sales, sales to ultimate consumers, andsales to industrial consumers.

Clerks tabulated the census data statistics by kind ofbusiness (chemical products wholesaler, drug wholesaler,etc.), by geographic area (division and state), and by typeof organization (proprietorship, partnership, etc.). Specialtabulations yielded wholesale statistics for cities withpopulations of at least 100,000 (where almost half of theestablishments were located). The Census Bureau pub-lished preliminary results as press bulletins and releasedthe final bound volume in December 1933.

The census of hotels, which included only establish-ments with at least 25 guestrooms, was originally plannedas a mailout/mailback operation. The Census Bureau com-piled a directory of names and addresses of about 27,000hotels, and mailed questionnaires in February 1930.Because of changes in ownership, duplications, and classi-fication problems, a field canvass was necessary, usingdecennial census enumerators. Ultimately, clerks compileddata for approximately 15,500 hotels (70 percent of whichreturned questionnaires by mail, while enumerators con-tacted the remainder). The agency published tabulationson the number of hotels, number of rooms, seating capac-ity of dining rooms, receipts, employment, salaries andwages, and number of proprietors and firm members.These data were classified by plan of operation (American,European, and mixed), type of occupancy (transient, per-manent, or mixed), and geographic division and state.

The procedures, coverage, and publication program forthe 1930 manufactures census closely resembled those ofthe biennial censuses for 1921 through 1927. There were165 questionnaires used to collect data for 238 industrycategories. One major difference in coverage was thatwhereas the first four biennial censuses of manufacturesincluded only firms reporting annual production valued ofat least $5,000, the Census Bureau returned this cutoff to$500 for the 1930 census.

The scope of the 1930 census of mines and quarries,differed considerably from the 1920 census of mineralindustries. The 1930 census excluded the petroleum andnatural gas industries, and did not collect data for capital,land holdings, rents, royalties, taxes, or detailed break-downs by kind of employees. The census covered thesand and gravel, glass-sand, and molding-sand industries,and the quarrying of limestone carried on in connectionwith the manufacture of lime and cement. The 1930 ques-tionnaires included new inquiries as to distribution ofsales, equipment purchased, and mobile power equip-ment; and consolidated and reclassified a number ofindustries. The data-collection methods and the publica-tion program remained unchanged from previous cen-suses.

The 1930 census publication program included severalspecial reports on economic topics, including distributionof sales of manufacturing plants; products of manufactur-ing industries; materials used in manufacturing; and loca-tion of industrial plants. With the addition of the censusesof distribution, construction industries, and hotels, thecontinuation of the manufactures and mineral industriescensuses, and the coverage of special topics, the 1930census was broader in scope than any of the previous eco-nomic censuses.

1931-1933. The Congress reduced the Census Bureau’sfunding and personnel authorizations for the 1931 and1933 censuses of manufactures in an effort to reduceFederal expenditures and balance the budget. The 1931



census covered 310 industries, but discontinued someinquiries (e.g., on salaried employees, power equipment,and coal consumption). The Census Bureau mailed ques-tionnaires in January 1932, and approximately 60 percentwere returned by mail. Officials of local chambers of com-merce and, in large industrial centers, census employees,conducted a telephone follow-up of nonrespondents.Because of decreased funding there were few field follow-ups. Budget restrictions also delayed the publication ofcensus results. The individual industry bulletins werereleased over a 2-year period (1933-1934), while the finalone-volume report was published in 1935.

The agency reduced the number of special question-naires for 1933, making greater use of the general ques-tionnaire and of a short-form questionnaire for smallerestablishments. The census used the mailout/mailbackenumeration, with telephone follow-ups. About 75 percentof the questionnaires were returned by mail. The pub-lished volume containing final census results warned datausers that fewer personal follow-ups had resulted inincomplete coverage in some areas.

President Franklin D. Roosevelt’s ‘‘New Deal’’ policies,and their economic data requirements, triggered severalspecial economic census projects, such as the 1933 busi-ness census. The census included retail distribution,wholesale distribution, and a new category, ‘‘Services,Amusements, and Hotels.’’ The services classification con-sisted of personal services (including barber shops andbeauty parlors), mechanical repair services (such as radioshops), and miscellaneous services (such as parking lots).The agency collected data exclusively in a field enumera-tion funded by the Civil Works Administration so as to pro-vide temporary employment for 4 million people duringthe winter of 1933-1934.

1935. The 1935 Census of Manufactures provided aboutthe same level of detail as in the 1930 census. Althoughthe Census Bureau developed a mailing list and mailedquestionnaires in January 1936, companies in all cities andtowns were instructed to hold their questionnaires untilenumerators visited their establishments. Enumeratorscollected data from every establishment, including thosethat had not received questionnaires by mail.

The 1935 Census of Business covered retail trade,wholesale trade, the construction industries, and serviceestablishments (personal, business, repair, custom, andmiscellaneous services) in an extensive field operationfunded by the Works Progress Administration. Addition-ally, the census included a miscellaneous business cat-egory encompassing many types of establishments notcanvassed in any previous census. This miscellaneous cat-egory included—

• Advertising agencies (mailed questionnaires andreceived a follow-up visit).

• Radio broadcasting stations (mailed questionnaires, onefollow-up letter, and received a follow-up visit).

• Banks (data collected by mail by the Federal ReserveBoard, Comptroller of the Currency, and Federal DepositInsurance Corporation. Follow-up visits were supervisedby the Census Bureau).

• Insurance companies (mailed questionnaires andreceived a follow-up visit).

Data from real estate agencies, financial institutionsother than banks (e.g., stock brokerages and finance com-panies), hotels, places of amusement, transportation (e.g.,buses, trucks for hire, and warehousing), nonprofit organi-zations, property management firms, and miscellaneousbusiness were collected in a field enumeration.

The goal of the 1935 census was to canvass every rec-ognizable place of business and provide an appraisal ofthe effects of the economic depression. Some of the datawere tabulated at a special Census Bureau branch that wasopened to provide work in Philadelphia, where the unem-ployment rate was higher than in Washington, DC.

The Census Bureau published the final business censusresults in 14 volumes (3 each for retail trade, wholesaletrade, selected service industries, and construction; 1 fortransportation and warehousing; and 1 for the miscella-neous topics), plus a series of special reports. The agencyalso published a one-volume census of manufacturesreport. The sheer scope and complexity of the operation,the limited time available for planning, and the difficultyin supervising the large contingent of field workers andclerks resulted in misclassifications, undercounting, tabu-lation difficulties, and other problems.

1937. In 1937, The Census Bureau conducted only thecensus of manufactures. The Congress funded the opera-tion through regular census appropriations rather than asan emergency public works project. Although the numberof special questionnaires was reduced, the amount ofdetail concerning products was greater than in any previ-ous census. For example, there were questions added con-cerning finished-product and work-in-progress inventories.Enumerators canvassed 351 industries, using 1 general, 1administrative, and 143 special questionnaires.

The Census Bureau derived a mailing list from 1935census files, trade directories, license lists, and othersources. It mailed questionnaires in January 1938, anddispatched two follow-up letters to nonrespondents inFebruary and March. Enumerators visited establishmentsbetween April and June 1938, after which the agencymade a final attempt to obtain outstanding questionnairesby mail.

The results of the 1937 Census of Manufacturesappeared first as press releases with preliminary data,then as pamphlets with final results, and finally, as a vol-ume, which was released in December 1939.

1940. The 1940 Decennial Census included the cen-suses of business (retail and wholesale trades; selectedservice establishments, places of amusement, hotels, and



tourist camps; and construction), manufactures, and min-eral industries and collected data for the year 1939. Thebusiness census was exclusively a field enumeration. In1940, the Census Bureau did not canvass the majority ofthe establishments classified in the 1935 miscellaneousbusiness category (e.g., banks, advertising agencies, andradio stations). The agency shifted coverage of places ofamusement and hotels to the selected service industrygroup. The 1935 retail classifications were modified for1940, and a special table on ‘‘reconciliation of classifica-tions’’ was published to facilitate comparisons of 1935and 1940 tabulations. The Census Bureau changed thescope of the census of selected service establishments byadding the miscellaneous business classifications (asnoted above) and by internal adjustments.11 For the con-struction industries, the Census Bureau eliminated the sizecutoff for tabulating detailed statistics used in previouscensuses (annual business of at least $25,000), and thescope of the wholesale trade category remained essen-tially the same as in previous censuses. The agency alsoextended the censuses of retail and wholesale trades andselected service industries to Puerto Rico in 1940 (and forthe subsequent economic census years, except for 1948).

Coverage of the census of manufactures was similar tothat of 1937, but the enumeration shifted from amailout/mailback operation with follow-up visits to a fieldenumeration. There was a new inquiry on capital expendi-tures for plants and equipment, and the question on per-sonnel requested data on the gender and detailed break-downs for various categories of nonmanufacturingemployees.

The census of mineral industries used a mail enumera-tion. The Census Bureau enumerated the bituminous coalindustry by mail with the close cooperation of the Bitumi-nous Coal Division of the Bureau of Mines, U.S. Depart-ment of the Interior. The Bureau of Mines field officesdistributed and collected the questionnaires from coalindustry establishments.

In addition to the usual press releases and pamphlets,Census Bureau employees compiled data in 10 volumes (5for the census of business, 3 for the census of manufac-tures, and 2 for the census of mineral industries). Thecountry’s entry into World War II interrupted the final tabu-lation and preparation of planned special reports. Some ofthese (e.g., subject reports in the retail area) were aban-doned. The Census Bureau did not publish the last volumeuntil June 1943.

During World War II, the Government discontinued theperiodic economic censuses in favor of war-related sur-veys that provided statistics for the Office of Price Admin-istration, the War Manpower Commission, the Office ofDefense Transportation, and other agencies in charge of

defense efforts. Title 14 of the Second War Powers Act,passed March 27, 1942, allowed the Secretary of Com-merce to dispense with or curtail any regular census of theU.S. Department of Commerce in order to undertake moreurgent statistical work deemed vital to the war effort.Executive Order 9152 of April 29, 1942, specifically can-celed the 1941 Census of Manufactures. The 1943 Censusof Manufactures was similarly dispensed with. The censusin 1945 was suspended because the Congress did not toappropriate the necessary funds.

Post-War Economic Censuses

1947. The first economic census taken after World War IIwas the 1947 Census of Manufactures. This was amailout/mailback operation12 taken in accordance withthe prewar law authorizing biennial economic censuses.Title 14 of the Second War Powers Act (but not the entireact) expired on March 31, 1947, so the 1947 Census ofManufactures was carried out in 1948, and censuses in1950 covered 1949.

There were a number of new features in the 1947Census of Manufactures:

• The agency based the classification of industries on the1945 Standard Industrial Classification (SIC) Manual.(Prior to World War II, the Census Bureau had developedits own classifications).

• The Census Bureau used the Old Age and SurvivorsInsurance (OASI) records of the Social Security Adminis-tration (SSA) to ensure more complete coverage in thecensus. Staff reconciled the OASI establishment andindustry classifications and those of the census. Thisimproved coverage was an important initial step in fur-ther integrating Census Bureau and other Federal agen-cies’ statistics.

• The Census Bureau sent precanvass cards to about525,000 possible manufacturing firms before mailingthe questionnaires. The address listing operation con-tained inquiries on company name and address, corpo-ration affiliation, manufacturing process used, types ofproducts, and number of employees. The response tothese questions enabled the Census Bureau to eliminateaddresses that should not have been listed as manufac-turing establishments, and to determine in advancewhich of the 212 types of industry questionnairesshould be sent to each establishment. The Post Officeassisted with the Census Bureau’s address listing opera-tion by verifying the presence or absence of an estab-lishment at specified addresses.

• A simplified questionnaire for small establishments wasused to reduce reporting burden and processing cost.(Simplified questionnaires were previously used in1933).

11The agency recommended that data users should not com-pare the 1935 and 1940 aggregates, because of the numerousadditions and deletions.

12Field enumerators canvassed approximately 15,000 sawmillsthat had highly mobile and sporadic operations.



• The Census Bureau collected statistics on work hoursfrom all industries, and gave more prominence to thetotal number of employees rather than to productionworkers. Size distributions in presenting data were interms of the total number of employees rather than thenumber of production workers.

• The census collected questionnaires from firms manu-facturing products valued at $5,000 or more during thecensus year. This provided coverage more comparableto other Federal programs.

• In addition to statistics for individual products ordinarilypublished in the census of manufactures, the CensusBureau grouped value figures into approximately 1,000product classes. These classes were used when thenumber of reporting establishments was too small topermit showing data for individual products.

• The Census Bureau discontinued the publication of dataon cost of materials and value of products for theUnited States, major industry groups, and all geographicareas, because of the unknown amount of duplicationcontained in these data. The Census Bureau did publishthese data for most individual industries, and releaseddata on value added by manufacture and number ofemployees for industry groups and individual indus-tries.

• The agency carried out a field canvass of selected areasbelow the state level to establish the completeness ofthe census, the nature of problems involved in cover-age, and types of establishments missed or misclassi-fied.

• The agency tabulated and published data for 147 stan-dard metropolitan areas.13

From January to March 1948, the Census Bureau usedabout 325,000 census questionnaires to collect data from435 manufacturing industry categories. Census Bureauemployees telephoned nonrespondents and companieswho returned incomplete questionnaires. After the staffremoved duplicate questionnaires and out-of-scope firmsfrom the file, approximately 141,000 manufacturingestablishments were tabulated.

As in the past, the 1947 census results appeared firstas preliminary reports, then as final reports, and finallywere assembled in three volumes including a general sum-mary, statistics by industry, and statistics by state.

1948. On June 19, 1948, Congress passed Public Law80-671 (incorporated into Title 13, United States Code,when the Census Bureau’s censuses and surveys were

codified in 1954). This law authorized the Census Bureauto conduct economic censuses for 1948 and every fifthyear thereafter.

The first application of the new law came with the 1948business census, which included retail trade, wholesaletrade, and selected service industries.14 Significantchanges in the scope of the 1948 business censusincluded—

• The exclusion of the contract construction industry; thelimitation of the census to those areas where coveragewas mandatory under the new law (the 48 States, theDistrict of Columbia, Alaska, and Hawaii, but not PuertoRico).

• A reduced emphasis on employment and payrollinformation.

• The compilation of data on the number of trucksoperated by business establishments.

The census collected only a limited amount of informa-tion from most small single-establishment firms, butobtained more detailed data from a sample consisting ofevery 10th small independent retailer, large retailers(those independent retailers with 1948 sales volume inexcess of $100,000), and multiestablishments.

The Census Bureau established 308 temporary fieldoffices for the 1948 census and carried out an addresslisting operation (which excluded medical, dental, and lawoffices; government offices; and farms). Between May andNovember 1949, enumerators left 1 of the 12 availablequestionnaires (5 for selected service establishments, 5for wholesalers, and 2 for retail stores) at most single-establishment firms with instructions that it be completedand returned to the Census Bureau by mail. For multies-tablishments, enumerators completed the questionnairesfor all establishments by interview, unless the establish-ments’ management specifically requested other arrange-ments.

The second phase of the 1948 business census con-sisted of data collection, coding, and editing. Clerksrecorded the receipt of questionnaires from each estab-lishment in listing books for each enumeration district.Nonrespondents received a reminder card and then afollow-up visit. If questionnaires contained incomplete orincorrect data, enumerators visited the establishments tocomplete or correct the questionnaire. For the first time,preliminary editing of questionnaires took place in theCensus Bureau’s field offices and enumerators were paidon a per diem rather than a piece-rate basis.

After the field canvass, the Census Bureau reenumer-ated 2,500 representative small (substate) areas. This sur-vey revealed an undercoverage of 8.2 percent of service

13These were the forerunners of the present Office of Manage-ment and Budget-designated statistical areas. These are one ormore counties around a central city or urbanized area with50,000 or more inhabitants. Contiguous counties were included,if they had close social and economic ties with the area’s popula-tion nucleus.

14The law authorized censuses of mineral industries and trans-portation in 1949, but congress did not appropriate funds forthem and they were not taken until 1954 (minerals) and 1963(transportation; see discussion under 1963.)



establishments and 3.6 percent of retail establishments. Ingeneral, the undercoverage reflected a failure to list andcanvass small businesses, particularly those with noemployees.

Before the temporary offices closed, there were severalcoverage checks, including matching operations withSocial Security Administration (SSA) lists and comparisonsof county totals with those of neighboring counties orcounties with similar economic characteristics.

As usual, the Census Bureau published preliminaryresults and final reports, the latter in bound volumes(three for retail trade, two for wholesale trade, and two forservice industries). In addition to the usual tabulations(receipts, sales, payroll, number of establishments, etc.),special tabulations were published on retail store sales bymerchandise line and wholesale sales by commodity line.The publications presented statistics for 147 standardmetropolitan areas (MSAs), in addition to states, counties,and cities.

The Advent of the Electronic Computer in 1951. In1951, the Census Bureau acquired the first large nonmili-tary computer, UNIVAC I, in time to tabulate part of the1950 Decennial Census. Data from punchcards wereloaded onto magnetic tape. Once the data had been digi-tally stored, UNIVAC I was used extensively for calculating,editing, imputation, and for operating high-speed printersthat prepared offset copy.

The availability of computers and the centralization ofcensus operations for multiestablishments at CensusBureau headquarters enabled the agency to launch theenterprise statistics program. This program grouped datafrom establishments under common ownership or controlinto tabulations showing various economic characteristicsbased on the classification of the owning or controllingcompany.

The Watkins Commission: 1953-1954. Although theCongress appropriated funds for planning and preparatoryoperations in fiscal years 1952 and 1953, it disallowed thefiscal year 1954 budget request for the actual census-taking. Money was allocated only for special surveys ofmanufactures and business. As a result, the CensusBureau terminated work on the economic censuses in May1953. This action provoked considerable alarm in manyGovernment departments and agencies, as well as in thebusiness and academic communities. In response to com-plaints about the cancellation of the census, the Secretaryof Commerce appointed a number of professors, businessexecutives, economists, and other specialists not affiliatedwith the Census Bureau to a committee known as the Wat-kins Commission in October 1953. This Committee,chaired by Dr. Ralph J. Watkins, Director of Research forDun and Bradstreet, Inc., was charged with conducting anintensive review of the economic census. In March 1954,the Watkins Commission recommended that the economiccensuses be resumed. Congress enacted Public Law

83-467 in June 1954, providing for censuses of manufac-turing, mineral industries, and other business (includingthe distributive trades and service establishments) in theyear 1955 relating to the year 1954 instead of a censusesin 1954 relating to the year 1953. The economic censusesbecame an integrated economic statistical program inwhich data for retail trade, wholesale trade, manufacturingestablishments, and construction, mineral, and serviceindustries were collected for the same benchmark years.

1954-1955. The 1954 Census of Business (retail andwholesale trades and selected service industries) markedthe Census Bureau’s first attempt since 1890 to compilecensus statistics from administrative records. Whileemployers were enumerated via the mailout/mailback pro-cedure, the Census Bureau used 1954 income tax recordssuppled by the U.S. Internal Revenue Service (IRS) toderive selected data items (such as employment, payroll,and sales) for retail nonemployers with 1954 sales of aleast $2,500 and for service nonemployers.

Since the income tax records did not contain the kind-of-business classifications necessary to determine whichquestionnaire should be mailed to a particular establish-ment, the Census Bureau had to match the IRS list withother lists containing kind-of-business classifications (e.g.,the SSA’s employer master file). The names on these listscould usually be linked through the use of the employeridentification number (EIN) assigned to each case by theSSA, and matched mechanically on punchcard collatingequipment.

For mining companies, the Census Bureau supple-mented the mailing list with information provided by suchagencies as the Bureau of Mines and the Federal PowerCommission. Approximately 280,000 manufacturing, 1.8million business (retail and wholesale trade and selectedservices), and 32,000 mining establishments completedquestionnaires. The Census Bureau derived data for morethan 1 million retail and selected service companies fromtax returns, saving and estimated savings of $3 million indata-collection costs.

To secure accurate and up-to-date information for multi-establishment companies and to correct and consolidatethe IRS list, the Census Bureau conducted amailout/mailback precanvass in the fall of 1954. It askedmultiestablishment companies for data about names andaddresses, employment, types of activities, etc.

The Census Bureau excluded wholesale firms withoutemployees and retail and service nonemployers with salesand receipts below the cutoff points from the census. Non-employers, although large in number, accounted for only asmall percentage of retail sales volume and servicereceipts for 1954. Past experience demonstrated thatthese small firms were most likely to be missed in an enu-meration, since many were operated from the owner’shome and/or on an intermittent basis.

Prior to the mailout of the economic census question-naires, the Census Bureau prepared a mailing list using



administrative records and, for large companies, the mail-ing list compiled for the 1953 Annual Survey of Manufac-tures (ASM).15 Mailout of the economic census questionsoccurred in early 1955. The mailout and follow-up opera-tions involving large companies (generally those with atleast six employees) were supervised by Census Bureauheadquarters’ staff. Field offices controlled mailout andfollow-up operations for smaller companies. Respondentsreturned the questionnaires by mail, and nonrespondentsreceived follow-up letters. Census Bureau employees tele-phoned or conducted follow-up visits to those establish-ments that had not responded to the follow-up letters.Coverage-control procedures ensured that all establish-ments on the mailing list were accounted for.

As in past censuses, the Census Bureau published 1954census results in preliminary bulletins, final reports, andbound volumes. In addition to the usual summary, indus-try, subject, and area reports, there were special tabula-tions for central business districts.16 Using the 1954 defi-nition of a central business district, the agency alsopublished retroactive retail trade and selected service datafor 1948.

1958. The scope, coverage, questionnaires, procedures,and tabulations for the 1958 censuses closely resembledthose of 1954. The Census Bureau mailed questionnairesfrom January to May 1959. Innovations for the 1958 cen-suses included—

• The use of more sophisticated computers. Processingoperations performed by computer expanded to includemailing list controls, more extensive editing, certainkinds of coding, and other functions that had been pre-viously performed manually or with punchcard equip-ment.

• Establishment of a census operations office in Jefferson-ville, IN, responsible for preparatory operations, prepa-ration of mailing packages, mailout, receipt, check-in,clerical editing and coding, and card punching.

• A new statistical geographic area for the business cen-sus was introduced—the ‘‘major retail center’’ (MRC)—which was an outlying business area, such as a shop-ping center. The Census Bureau published data for theMRCs in a series of 97 reports, each covering a standardmetropolitan statistical area (SMSA) and including infor-mation for central cities and their central business dis-tricts.

• The economic censuses included Guam and the VirginIslands of the United States for the first time.

1963. Because of inadequate transportation data, PublicLaw 80-671 authorized a census of transportation in1949. However, because the Congress did not appropriatefunds for the census and plans for a similar census in1953 were canceled, the 1963 Economic Census was thefirst to feature a census of transportation.17 FollowingCongressional hearings in July 1961, funds were appropri-ated to collect detailed transportation data on a nation-wide basis.

The Census Bureau’s primary objective in conductingthe census of transportation was to collect data on theNation’s transportation without duplicating data alreadyavailable from other government or private sources. Toachieve this objective, four surveys, each aimed at a spe-cific gap in knowledge, were designed—

• The Passenger Transportation Survey. This surveyproduced statistics showing national and regional pas-senger patterns for 1963 and their relationship to socio-economic and geographic factors. The Census Bureaucollected data in four quarterly personal interviews froma probability sample of about 6,000 households.

• The Truck Inventory and Use Survey (TIUS). Thissurvey collected data on the Nation’s truck resources,such as the number of trucks classified by physical char-acteristics, occupational use, intensity of vehicle utiliza-tion, and geographic distribution of vehicles. The Cen-sus Bureau mailed questionnaires to a sample of about100,000 truck and truck-tractor owners selected fromstate motor vehicle registration records.

• The Commodity Transportation Survey (CTS). Thissurvey collected data concerning the geographic distri-bution of tons and ton-miles of commodities, by type,shipped by the manufacturing sector. The surveyobtained basic information from a probability sample ofbills of lading or other shipping records maintained incompany files.

• The Motor Carrier Survey. This survey produced sta-tistics concerning ‘‘for-hire’’ carriers not subject to eco-nomic regulation by the Interstate Commerce Commis-sion. The Census Bureau conducted a completemailout/mailback enumeration of bus companies andpublic warehouses that also provided trucking services,plus a mailout/mailback enumeration of a probabilitysample of trucking firms.

15The ASM was first conducted in 1949 to provide data forintercensal years. The survey covered all large establishments(those having 100 employees or more from 1954 through 1967and, from 1972, those with 250 employees or more) and a repre-sentative sample of smaller establishments. Large establishmentsremained in the ASM; however, the Census Bureau replacedsmaller establishments every 5 years. During census years, theASM questionnaire became the first two pages of the question-naire used for the economic census.

16Central business districts were newly defined geographicentities encompassing the downtown business areas of largecities.

17Earlier economic censuses had collected data on transporta-tion, but these had covered only specific areas of transportation,such as water transportation (periodic censuses between 1880and 1926) and street and electric railways, together with affiliatedmotorbus lines (included in the censuses of electrical industriestaken at 5-year intervals between 1902 and 1926).



At the request of the U. S. Department of the Interior’sBureau of Commercial Fisheries, the 1963 Economic Cen-sus resumed the collection of data from commercialfisheries following a 55-year hiatus. The census included amailout/mailback enumeration of commercial fisheriesthat had supplied employment data to the Social SecurityAdministration (SSA). The questionnaires contained inquir-ies on employment, payroll, receipts, characteristics ofvessels, and landed catch. The Census Bureau used asample of administrative records to compile limited statis-tics for part- and fulltime fishermen, without paid employ-ees.18

The 1963 censuses of retail and wholesale trades,selected service industries, manufactures, and mineralindustries closely resembled those conducted in 1958.The Census Bureau used newer computer systems andexpanded the use of electronic equipment to—

• Computerize geographic coding of establishmentaddresses.

• Substitute magnetic tape for punchcards in developingthe census mailing lists from IRS lists.

• Automate procedures to control mailout and follow-upoperations.

• Expedite mass transfer of data from Jeffersonville, IN, tothe computer facility at Census Bureau headquartersusing a high-speed telephone-transmission system.

• Computerize a system for work and progress reportingof census operations.

• Produce computer-programmed news stories for releas-ing census results and to automatically mail these tonews media outlets.

The Census Bureau obtained selected information, suchas value of receipts and sales, from nonemployers’ taxreturns. The agency also used administrative records toassemble mailing lists and obtain preliminary industryclassifications for employer firms to be included in themailout/mailback enumeration. In total, the Census Bureauasked about 3 million establishments to complete ques-tionnaires (which were mailed between November 1963and March 1964). The agency derived statistics forapproximately 1.5 million nonemployer establishmentsselected from administrative records. (For a detailedaccount of the 1963 Economic Censuses, see 1963 Eco-nomic Censuses: Procedural History.)

1967. In 1967, Congress modified Title 13 of the U.S.Code Census, changing the reference years for the eco-nomic censuses to those ending in ‘‘2’’ and ‘‘7.’’ This wasdone to distribute more evenly the staff and computerworkload generated by the economic and demographiccensuses.

The 1967 Economic Censuses thus included censusesof retail and wholesale trades, selected services, construc-tion and mineral industries, manufactures, commercialfisheries, transportation, and the enterprise statisticsprogram.

The Census Bureau expanded the scope of the 1967Economic Census, to include—

• A census of construction industries (the first since1939), which, for the first time, included Puerto Rico.

• Adding architects and engineers, law firms, andarrangement of passenger transportation (travel agentsand tour operators) to the census of selected serviceindustries.

• Adding the National Travel Survey (NTS),19 TIUS, and theCTS to the census of transportation.

• Doubling the CTS sample size in the major populationcenters to improve the quality of the point-of-origin topoint-of-destination commodity flow data, and extend-ing the survey’s scope to include printing and publish-ing establishments (except newspaper and periodicalpublishers).

Data for all nonemployers (establishments without paidemployees) in retail trade and selected service and con-struction industries were obtained from IRS income taxrecords. In addition, the Census Bureau expanded the useof these records to compile statistics for selected single-establishment small employers (firms with few employeesduring 1967), as follows:

• For retail trade and selected service industries, differentpayroll cutoffs for various kinds of businesses deter-mined which firms would be excused from completingquestionnaires.

• The number-of-employees equivalent to the payroll cut-off was 10 employees for manufacturing firms and 5 formineral industries establishments.

• For construction industries with employees, the CensusBureau used administrative records only to select themail sample. Data for more than 1 million small-employer establishments, canvassed by mail in previouscensuses, were compiled from tax returns.

In total, the Census Bureau used administrative recordsfor 2.9 million establishments and asked 1.9 million estab-lishments to complete questionnaires.

The census of commercial fisheries was expanded tocollect statistics on the number of vessel operators, catch,fishing gear, and various vessel characteristics. The Cen-sus Bureau obtained some basic information (grossreceipts and industry classification) from the IRS and SSA

18Because the mailing list used to collect data from commer-cial fisheries was later found to be incomplete, a supplementalvessel survey was conducted in 1964. 19Called the Passenger Transportation Survey, in 1963.



for part- and full-time fishermen without paid employees.The Census Bureau did not publish the 1967 data fromadministrative records because of classification problems.

New techniques, computers, and auxiliary electronicequipment were used to handle data from the 1967 Eco-nomic Census. Other significant changes in processingmethods included—

• Expansion of the geographic coding file to facilitatecomputer coding of establishments located in smallcities.

• Development of specifications and computer programsto perform complementary disclosure analysis (toensure that data for individual establishments remainedconfidential).

• Increased use of computer editing.

(For a detailed account of the 1967 Economic Censuses,see 1967 Economic Censuses: Procedural History.)

1972. For the 1972 Economic Census, the CensusBureau mailed 2.9 million questionnaires between Decem-ber 1972, and February 1973. Additional information wascollected from administrative records for approximately2.6 million establishments. The 1972 Economic Censusincorporated several changes. These included—

• Collecting construction industries data in Guam and theVirgin Islands of the United States.

• Precanvassing all out-of-scope activities of companiesthat reported on an establishment basis in previouseconomic censuses.

• Increasing the sample size for the National Transporta-tion Survey (from 18,000 in 1967 to 24,000 householdsin 1972).

• Classifying the 1972 data according to the new 1972Standard Industrial Classification (SIC) Manual. (Forselected data items, bridge tables showed the 1972data classified under both the 1967 and the 1972 SICsystems).20

• Inaugurating a survey of minority-owned businessenterprises (SMOBE). The SMOBE presented tabulationsby major SIC industry or industry group by race (Black,Asian American, American Indian, and ‘‘Other’’) andHispanic origin for MSAs. After 1972, these data werecollected along with data for counties and places with

specified numbers of minority-owned firms. The pub-lished data from the SMOBE were usually limited to legalform of organization and receipt- and employment-sizeof firm.

The first reports in key publication series from each ofthe censuses appeared within 1 year of the period coveredby the 1972 censuses (6 months earlier compared to therelease of the 1967 census data). Publication of the majorseries was completed 9 to 12 months earlier than for1967. Recognizing the increasing need for digital data, theCensus Bureau introduced a program to issue public-usecomputer tape files of economic census statistics.

(For a further description, see 1972 EconomicCensuses: Procedural History.)

1977. The 1977 Economic Censuses covered retail trade;wholesale trade; service, construction, and mineral indus-tries; manufactures; and transportation; the SMOBE; theenterprise statistics program; and a survey of women-owned business enterprises (SWOBE). The Census Bureaumailed more than 3 million questionnaires betweenDecember 1977 and April 1978, and obtained data forabout 4 million small firms from administrative records.The SWOBE collected more specific information about thedemographic and economic characteristics of women-owned businesses and their owners.

For the 1977 censuses, the 1976 RecordkeepingPractices Survey obtained information on the ability ofrespondents to provide data not previously requested inthe Census Bureau’s economic censuses and surveys. As aresult, the Census Bureau decided to—

• Collect additional data needed to improve the grossnational product accounts.

• Expand the coverage of the service industries census.

• Develop a common set of commodity lines for all typesof operations in the wholesale trade census.

• Develop a better geographic coding system to improvethe classification of data by areas.

• Improve the quality of the national travel survey.

• Carry out evaluation programs.

The Census Bureau gave priority to expanding productline detail for the 1977 censuses of manufactures andused mineral industries to meet the growing need formore data to compare domestic output to imports andexports. Section 608, ‘‘Uniform Statistical Data on Imports,Exports, and Production,’’ of the 1974 Trade Act mandatedthat a comparability study be made of the commodityclassification systems used for imports, exports, anddomestic production. This 1975-1976 study was a jointeffort of the Census Bureau and the International TradeCommission, and resulted in proposed modifications thatwould make data collected for individual products or

20The revisions to the SIC system added several industries,eliminated statistically insignificant industries, and containedmodified definitions of other industries within the scope of theeconomic censuses by shifting products and services from oneindustry to another within each SIC division.



groups of products more compatible with each other.Some of the product-line classifications for the 1977 cen-suses were developed or revised as a result of the study.The censuses incorporated new product-line classifica-tions, identified during the trade act review, for whichquestionnaires had not already cleared the Office of Man-agement and Budget (OMB). Those changes identified afterclearance were not used for the 1977 Economic Census,but were included in the 1982 censuses.

Unlike the 1967 and 1972 Censuses of Transportation,the 1977 coverage included nonregulated motor carriersand public warehousing, as well as the NTS, TIUS, andCTS. While the Census Bureau based the last three surveyson probability samples, data for nonregulated motor carri-ers and public warehousing firms with employees wereobtained in an enumeration of establishments in the samemanner as the other economic censuses. The CensusBureau made a number of changes and improvements inthe probability surveys (described in Chapter 10 of theHistory of the 1977 Economic Censuses).

The Creation of the Standard Statistical Establish-ment List (SSEL). In 1977, changes were made in thepractice (begun in 1954) of having a precanvass of allknown multiestablishment companies. This precanvasswas done once every 5 years, so the company and estab-lishment address records soon became obsolete. Follow-ing the 1972 Economic Census, the Census Bureau createdthe Standard Statistical Establishment List (SSEL), andbegan an annual Company Organization Survey (COS)designed to update the address files in the SSEL. The SSELis a central, multipurpose computerized name and addressfile of all known single- and multiestablishment employers(and nonemployer agricultural establishments) in theNation. Through the combined use of the SSEL file numberand the employer identification number, the CensusBureau could link and identify the affiliation of parentcompanies, subsidiary firms, and their establishmentsthroughout all phases of economic activity.

1982. The 1982 program consisted of censuses of retailtrade; wholesale trade; service, construction, and mineralindustries; manufactures; and transportation; the surveyof minority-owned businesses; the survey of women-owned businesses; the survey of characteristics of busi-ness owners (an expanded version of the 1977 special sur-vey of women-owned businesses); and the enterprisestatistics program.

Prior to the census mailout, the Census Bureau carriedout several inventory test surveys for use in revising theinventory question in the 1982 censuses of wholesaletrade, construction and mineral industries, manufactures,and the 1982 Enterprise Statistics Program. The surveys’objectives were to measure response rates and accuracyfor several plausible and controlled data-collection meth-odologies. Based on the response evaluations, the agency

adopted a revised inventory inquiry for the 1982 Eco-nomic Censuses (and its annual surveys). This meant thatonly the term ‘‘last in/first out reserve’’ was used and thatit was related to a standard definition. The question alsoasked for the value of inventories not subject to lastin/first out costing.

Between June 1 and July 31, 1981, the Census Bureautested proposed revisions to the 1982 TIUS. The test dem-onstrated that although the use of two different question-naires increased the Census Bureau’s processing burden, itprovided the respondent with a questionnaire moreclosely related to the characteristics of the sampledvehicle. Other results led to the revision of the sequenceand wording of many of the questions in the 1982 surveyquestionnaire to improve clarity and reduce reportingerrors.

The 1981 Recordkeeping Practices Survey sought todetermine if establishments maintained alternative docu-ment systems that were better suited to gathercommodity-flow data requested by the CTS. The results ofthe survey prompted the following changes in the CTS:

• Instructions to respondents emphasized using the sales-invoice document system versus the traditional bills oflading. (Bills of lading would be accepted if respondentsdecided not to use sales invoices.)

• Respondents received stronger worded instructions forusing the serial number sequence.

• The Census Bureau planned to make special provisionsfor establishments that wanted to use automated data-processing systems in responding; however, this werenot implemented.

Other changes to the 1982 Economic Census included—

• The 1982 census did not include a NTS or collect dataon nonregulated motor carriers as was the case in the1977 census. Data on public warehousing and arrange-ment of passenger transportation appeared in the cen-sus of service industries reports.

• The 1982 census excluded some of the service indus-tries that had been covered for the first time in the1977 census (e.g., hospitals; elementary and secondaryschools; colleges, universities, and professional schools;junior colleges and technical institutes; labor unions;and political organizations) because the information forthese service industries was available from othersources, such as the U. S. Department of Education andthe Bureau of Labor Statistics. The need to reduce cen-sus cost also was a factor.

• For the first time, the Census Bureau obtained dataon retail trade, wholesale trade, manufacturing, andconstruction and service industries for the Common-wealth of the Northern Mariana Islands.



The Census Bureau mailed more than 3 million eco-nomic census questionnaires and classification formsbetween mid-December 1982 and March 1983. Theagency began the mailout of questionnaires for the 1982TIUS in October 1982 and completed the operation in June1983. The CTS was conducted in 1984, which collected1983 data. In addition, data for about 3 million small firms(which were not sent questionnaires) were derived fromadministrative records.

The 1982 CTS did not take place in 1983 (for 1982),but in 1984. This allowed the Census Bureau to test andconsider alternative data-collection methods. The agencydivided establishments into three groups to test one ofthree techniques: (1) The systematic method, (2) the ‘‘first-15’’ method, and (3) the summary method. (For a descrip-tion of these three methods, see Chapter 2 of the 1982history.) The 1983 test did not identify a workable meth-odology for shipment sampling. As a result, the CensusBureau decided on a less detailed survey for 1983; how-ever, because the quality of the resulting information wasunacceptable, the agency did not publish any statistics.

Despite known inadequacies that would prevent com-plete publication processing, the Census Bureau adoptedan experimental computer program Table Image ProcessorSystem (TIPS) for the 1977 Economic Censuses publicationprogram. The Census Bureau designed this system tophoto-compose (in conjunction with its own computer-output-to-microfilm device and the Government PrintingOffice’s video computer-output-to-microfilm system) largevolumes of tables for the census reports. Prior to the 1982censuses, the agency examined the experimental TIPS andidentified 34 improvements. Three of the most importantrefinements were: The ability to (1) place any kind of tableanywhere on a page, (2) produce multiple tables on apage, and (3) produce multibanked tables (those with theboxhead repeated horizontally or the stub repeated verti-cally within the same table and on the same page). Sinceonly about one-third of the required refinements could berealized by modifying the TIPS, the system was notcapable of meeting the 1982 Economic Censuses’ publica-tion requirements. The Census Bureau undertook a majorredesign of the photo-composition system in 1981, devel-oping the Table Image Processor Systems II (TIPS II), whichremedied most of the original system’s deficiencies.

(For a detailed account of the 1982 censuses, seeHistory of the 1982 Economic Censuses.)

1987. The general plan for the 1987 Economic Censusprogram was identical to that of 1982. The censusesencompassed approximately 12.4 million establishments,and collected data from 3.7 million establishments usingmail questionnaires and approximately 8.7 million smallfirms using administrative records from Government agen-cies.

The 1987 Economic Census featured several changes todata collections, including—

• The addition of several record-keeping inquiries toquestionnaires for industries where partial fabricationwas common, e.g., the apparel; motor vehicle; foot-wear; electronics (computers and semiconductors);motor and generator; toy; and sporting goods indus-tries.

• Special inquiries on metalworking operations from thecensus questionnaires were dropped.

• The Census Bureau published concentration data at thefour-digit industry level but not at the five-digit productclass level because of budgetary constraints.

There were a number of significant changes in specificcensuses. For retail trade, there was a 20-percent increasein the number of merchandise-line questions. This pro-vided data users with more detailed information on thekinds of merchandise sold by different kinds of retailestablishments. The 1982 major retail center (MRC) seriesof reports marked the sixth and last in a series that, since1958, had traced the movement of retail businesses fromthe central business districts (CBD’s) to outlying shoppingcenters. Since the migration of retail business to the sub-urbs appeared to be complete, it was decided that CBDdata would no longer be published separately. Conversely,the substantial increase in the number of suburban shop-ping centers changed both the nature and the cost of theprogram. Costs rose significantly because of the increasein the number of centers and because the boundaries ofeach center had to be delineated on site, with each storelisted by name and type of business. Each store had to beclerically matched to census listings for the geographicarea in which the center was located. Even the changesmade for the 1977 censuses (which revised the size crite-ria and delegated the delineation and enumeration processto local Census Statistical Areas Committees or other localorganizations) failed to keep enumeration costs at a man-ageable level. Consequently, the agency decided to dis-continue the MRC program after the 1982 censuses.

There were several changes and additions made tothe 1987 census of retail trade’s publication program, asfollows:

• Retail trade data for employer establishments was pro-vided for each five-digit ZIP Code. These data were clas-sified by employment-size groups and sales-size ranges,and published in ZIP-Code Statistics Series reports viaelectronic media only (i.e., CD-ROM, diskettes, and com-puter tape).

• Nonemployer establishment data were published in anew Nonemployer Statistics Series rather than in theGeographic Area Series reports. The Nonemployer Sta-tistics Series included a separate report for each of fourgeographic regions on the number of establishments



with and without payroll and sales for each type ofretail classification, for each state, MSAs, counties, andplaces with 2,500 or more inhabitants.

• A new Special Reports Series presented the Selected Sta-tistics report. This new series contained selected aggre-gate data by kind of business, including ranks andratios, not provided in the other final reports. Data werereported for the United States, states, and MSAs.

The census of service industries collected data onhealth services, as well as educational services; social ser-vices; museums, art galleries, and botanical and zoologi-cal gardens; membership organizations (except religiousorganizations); rooming and boarding houses; etc. TheCensus Bureau obtained data on privately owned andoperated hospitals and,21 for the first time, ongovernment-owned and -operated hospitals.

The census collected additional information on interna-tional service transactions. In 1982, there were tabulationsof exported service receipts for architects, engineers, andsurveying services; management, consulting, and publicrelations services; equipment rental and leasing services;and computer and data processing services. Four addi-tional industries reported data on exported services adver-tising agencies; accounting, auditing, and bookkeepingservices; research, development and testing services; andlegal services.

The service industries data were published for indi-vidual ZIP Codes (for employer establishments only) andincluded data for nonemployer businesses.

The census of transportation was similar to that con-ducted in 1982 except that the CTS was not conducted.The CTS was canceled in January 1988 as a result of thedeficit-reduction guidelines mandated by the 1986Gramm-Rudman-Hollings Deficit Reduction Act.

Several changes were made to the census of construc-tion industries for 1987. These included—

• Collecting data on the value of construction workdone.22 Prior to 1987, the Census Bureau collected dataon receipts as the primary measure of constructionactivity.

• Increasing the number of construction questions from 8to 24. Questions were tailored to particular industries toreduce respondent burden and collect more detaileddata, especially for the type of construction work per-formed.

• Expanding efforts to capture construction activities ofestablishments not classified as construction establish-ments by the SIC system.

• Adding force-account construction23 questions to thecensus of manufactures questionnaires for the chemical,petroleum, and steel industries.

• Adding questions to the census of manufactures ques-tionnaires covering steel industries and installation ofmetal and wood buildings, and to the census of retailtrade questionnaires covering hardware and buildingsupply stores, to obtain information on secondary con-struction activity.

The census of manufactures collected aggregate dataon the foreign content of domestically-produced productsfor the first time. Industry and government agencies con-sulted about the manufacturing questionnaires asked thatthe Census Bureau measure the cost of foreign-madematerials consumed by domestic manufacturing plants.Many data users asked for information on the foreign con-tent of each material input consumed in an establishment,but the agency found that this information was not easilyreportable and, in many cases, not available. The 1987ASM asked approximately 55,000 establishments for infor-mation on the aggregate amount of materials, parts, andsupplies purchased from foreign countries.

The Census Bureau also published more comprehensivedata in 1987 on exporting manufacturers and improvedproduction statistics to be more comparable with foreigntrade statistics. This was done not only because of therevision of the SIC system for 1987, but also because inJanuary 1989 the United States adopted the HarmonizedCoding system as the official classification for import andexport statistics.

The 1987 censuses witnessed the following significantdata-processing changes—

• Use of minicomputers gave analysts interactive accessto micro-records and summary data, reduced theamount of paper listings, and improved data quality.

• The Census Bureau upgraded its automated photo-composition system the TIPS II. Laser printers in each ofthe subject-matter divisions (Business, Construction,and Industry) significantly reduced the time required toproduce photo-composed proof copies of data tableswhich contributed to the release of the final publica-tions several months earlier they had been in 1982.

• The Economic Programming Division was established inJanuary 1987 to better utilize computer programmingpersonnel. The staff and programming activities for the

21These data were collected for 1997, but not for 1982.22Receipts could be different from the value of work done,

since work can occur in one year and receipts in the prior or suc-ceeding year. In practice, receipts for most contracytors approxi-mated ‘‘value.’’ For key industries, such as operative builders anddevelopers, receipts, and work done might be different. In addi-tion, receipts did not include work a contractor performed for itsown account and use, which could be substantial.

23Force-account construction is construction work performedby an establishment primarily engaged in some business otherthan construction, by its own employees for its own account anduse.



census or census-related surveys were consolidated intoone division to allow computer staff to move amongprojects as required.

Effective June 1983, the OMB established a new set ofMSAs, consolidated metropolitan statistical areas (CMSAs),and primary metropolitan statistical areas (PMSAs). Thesereplaced the two former types of metropolitan areas (stan-dard metropolitan statistical areas and standard consoli-dated statistical areas). The Census Bureau used the2-level system to report data for 21 metropolitan areas in1987.

For the census of manufactures, the Census Bureaudiscontinued the preliminary geographic series. It consoli-dated the 443 industry preliminary reports (published in1982) into 83 bulletins covering related manufacturingindustries. The preliminary U.S. Summary Report includeddata at the four-digit SIC level, while the state reports con-tained information only at the two-digit SIC level. The finalindustry series was also available in 83 bulletins, followedby the final geographic area and subject series.

The census of construction industries continued to pub-lished preliminary industry, final industry, and final geo-graphic area reports.

No preliminary industry reports were published for thecensuses of retail trade, wholesale trade, and serviceindustries; instead, two-page press releases appeared foreach state. These press releases highlighted data pub-lished in the 584 reports 125 preliminary reports and 459final publications. The press releases were publishedapproximately 6 weeks before the final state reports werereleased.

(For a detailed account, see History of the 1987Economic Censuses.)

1992. The 1992 Economic Census covered retail trade;wholesale trade; service industries; transportation, com-munications, and utilities; finance, insurance, and realestate; construction industries, manufactures, and mineralindustries; and the TIUS.24 The program also included the1992 Survey of Minority-Owned Business Enterprises, the1992 Survey of Women-Owned Businesses, the 1992 Sur-vey of Characteristics of Business Owners, and the 1992Enterprise Statistics Program. The census and surveyswere conducted in the 50 states and the District ofColumbia. The 1992 Economic Census of Outlying Areascollected data on retail and wholesale trades, serviceindustries, manufactures, and construction industries inPuerto Rico, the Virgin Islands of the United States, Guam,and the Commonwealth of the Northern Mariana Islands.

The 1992 census marked the most significant expan-sion of the census in half a century, including coverage offinancial, insurance, real estate industries, communica-tions, and utilities. In addition, the following transporta-tion industry groups were added—

• Railroad Transportation. Data for railroads collected byother agencies were included in the Census Bureau’spublications, but excluded from the 1992 EconomicCensus data collection operation.

• Local and Suburban Transit and Interurban HighwayPassenger.

• Transportation.

• Transportation by Air.

• Pipelines, Except Natural Gas.

Altogether, the 1992 Economic Census covered approxi-mately 95 new industries, expanding coverage to approxi-mately 98 percent of the Nation’s economic activity.25

The Census Bureau introduced various industry-specificoutput measures for some of the newly covered indus-tries. For most industries, the basic output measures var-ied from sales (for retail establishments) to operatingreceipts (for taxable service establishments), revenue (fortax-exempt establishments), value of shipments (formanufacturers), or value of construction (for constructionindustries). Several of the newly covered industries had,as operating revenue, items that were nonoperating formost industries, such as rents, interest, investmentincome, grants, contributions, and gifts. This requiredadjustment in the data to be collected. For the finance,insurance, and real estate industries, the Census Bureauasked for total revenue, and used a supplemental questionon the major sources of revenue to separate the compo-nents to allow users to tailor output measures for theirown use. The item on ‘‘sources of revenue’’ also providedimportant information that could be used by the agency toidentify secondary activity. Similarly, the 1992 EconomicCensus questionnaire for transportation, communications,and utilities asked for operating revenues basis data. Thequestionnaire for broadcasting and cable industriesincluded a supplementary question on nonoperating rev-enue.

In the past, the IRS supplied the Census Bureau withgross receipts or sales (less returns and allowances). In1992, the Census Bureau also asked the IRS to supplyadministrative gross rents and interest income for partner-ships and corporations from its administrative records.These data did not comprise total revenue (e.g., dividendand royalty income and net capital gains were notincluded), but did include additional major sources of rev-enue for the newly covered industries.

24The Census Bureau used the term ‘‘economic censuses’’ tocover all of its economic data-collection activities between 1954and 1987. For 1992, on the advice of the Advertising Council, theagency used the term ‘‘economic census’’ primarily to assist itspromotional activities and because data users frequently did notunderstand what ‘‘economic censuses’’ meant.

25Coverage of the 1987 Economic Census was aproximately75 percent of the Nation’s economic activity.



In general, census data were published for the Nation,states, metropolitan areas, counties, and places. Someretail trade and service industry data were available onlyat the ZIP-Code level. For the newly covered industries in1992, data were made available for the Nation and forstates. Selected metropolitan area data also were providedfor some new industries in the transportation, communi-cations, and utilities and financial, insurance, and realestate industries group.

Because there had been a lack of data on the contribu-tion of services to the volume of exports in past censuses,the Census Bureau asked respondents to report informa-tion on exported services for additional service industriesas well as for selected financial services, including deposi-tory banking, business credit and other nondepositoryinstitutions, securities, brokers and dealers, and electricutilities.

In the census of manufactures, the number of pur-chased services was expanded from three to eight, andthe Census Bureau collected additional information on anauxiliary questionnaire. The receipts inquiry for the censusof construction industries was completely restructured toimprove data quality and response. Inquiries on manufac-turing output and materials consumed were streamlinedto reduce reporting burden and improve response rates.As a result of the 1988 Women-Owned Business Act(P.L.100-533), the Census Bureau collected information onwomen-owned corporations for the first time for 1992.

In an effort to improve and accelerate response, theCensus Bureau—

• Used more effective direct-mail techniques.

• Standardization of the size of most economic question-naires. The majority were 8 x 14 inches; however, a fewquestionnaires remained at 8 x 11 inches.

• Improved instructions and streamlined transmittalletters.

• Increased the emphasis on mandatory reporting in1992. Outgoing envelopes were overprinted with a mes-sage indicating that a Census Bureau questionnaire wasenclosed and that response was required by law.

• Implemented more effective follow-up strategies, e.g.,using questionnaire follow-ups for single-establishmentcompanies and multiestablishment companies.

• A toll-free telephone number was added to all question-naire mailings should respondents have questions.

For the 1992 Economic Census, a proactive companycontact program was implemented to encourage theNation’s largest establishments to respond. In December1991, the agency mailed an information booklet to the10,000 largest establishments, informing them of the1992 Economic Census. ‘‘Special handling’’ was offered tothe 1,000 largest establishments, which included accessto a single Census Bureau employee to whom any ques-tions could be directed. Companies requesting a contactperson were asked to reciprocate by identifying a contactwithin their respective organizations.

In November 1992 (about 1 month before census mail-out), analysts called each company contact person at the1,000 largest companies to alert them that the question-naires would soon be mailed. A second call was placed inJanuary 1993, to ensure that the firms had received thecensus packages. Firms that did not respond by April1993 were contacted again by telephone. Census Bureauanalysts encouraged reporting and/or collected data bytelephone from those companies that had not reported bylate June 1993.

Finally, the Census Bureau developed and used for thefirst time an electronic data interchange standard for useonly in the census of retail trade. The agency met withsome of the largest retailers that had extensive electronicdata interchange experience to encourage them to partici-pate.



Appendix C.Principal Advisory Committees onthe 1997 Economic Census

PUBLIC ADVISORY COMMITTEES

A public advisory committee is any group of persons−not composed wholly of officers or employees of the Fed-eral Government−organized under Federal Governmentalauthority for the purpose of obtaining advice, recommen-dations, or other types of assistance. From April 1, 1994to March 31, 2001 (the period in which the majority of the1997 Economic Census operations took place), the U.S.Census Bureau had five public advisory committees thatwere responsible to the Director, the membership of whichwas made up entirely of non-Census Bureau persons. Oneof these committees, the Advisory Committee of Profes-sional Associations, was concerned with one or moreaspects of the economic census. The Committees’ mem-bers were professional experts drawn from the businessand academic associations that had earlier participated infour separate advisory committees. They made recommen-dations concerning the scope, content, and methodology,and proposed tabulations and publications for the cen-suses. They also made other recommendations to help theagency ensure that the censuses would provide accurate,meaningful data. Representation on these bodies changedduring the census period.1

The Advisory Committee of Professional Associationsoperated under provisions of the Federal Advisory Com-mittee Act, which went into effect on January 5, 1973.This act stipulated that the Director of the Office of Man-agement and Budget (OMB) would prescribe administrativeguidelines and management controls. In addition the Com-mittees were governed by Executive Order 11769(February 21, 1974) and by OMB Circular A-63, revised(March 27, 1974). All these public advisory committeesproposals or formal recommendations were reported tothe Secretary of Commerce, together with appropriateresponses by the Census Bureau indicating what actions, ifany, would be taken as a result of these proposals.

The advisory committees were established on the pre-sumption that they could make significant contributions tothe Census Bureau’s program and objectives. Only the Sec-retary of Commerce could establish a public advisory com-mittee within the Department, and no such body could be

formed or used unless the Secretary formally determinedthat the establishment of such a committee was in thepublic interest and was connected with the performanceof the Department’s duties specified by law. Except whereotherwise fixed by law, a public advisory committee termi-nated 2 years after its formation unless the Department ofCommerce determined, in writing, not more than 60 daysprior to the termination date that its continued existencewas in the public interest. Until 1972, the Secretary wasempowered to renew committee members’ appointmentsas often as the Census Bureau might request; after thattime, members of Census Bureau committees were gener-ally limited to a maximum of 6 years of continuous ser-vice. None of the committee members received compensa-tion, other than per diem and travel expenses, forattending meetings.

The Committee, composed of the three subgroupsdescribed below usually met every 6 months.2 Their mem-bership is shown for the period April 1, 1995 to March 31,2001.

THE AMERICAN STATISTICAL ASSOCIATION

In November 1918, Secretary of Commerce William C.Redfield invited the presidents of the American StatisticalAssociation (ASA) and the American Economic Association(AEA) to appoint a joint public advisory committee toassist the Census Bureau in organizing and taking the1920 Decennial Census. The joint committee met for thefirst time in February 1919. This so-called General Advi-sory Committee functioned until 1937 on a permanentbasis, with its members drawn from the two associations.At that time, it was reconstituted, with all its membersappointed by the ASA for 3-year terms.

This oldest standing advisory committee of the CensusBureau was designed to be representative of all statisticalfields in which the Census Bureau operated. The commit-tee’s functions were defined in 1964 as ‘‘(a) criticallyobserving and appraising the Census Bureau program as awhole, and the progress of its various segments, (b) con-sidering priority issues that arise between and withinfields, (c) examining formulations of guiding principles, (d)advising on questions of policy and procedure, especiallyalternative sources of data, and (e) responding to the Cen-sus Bureau’s requests for opinions and judgements in thewhole area of its operations.’’

1Presidential Executive Order 12838 (signed in February 1993)was implemented in 1994 transforming the American StatisticalAssociation, American Economic Association, American MarketingAssociation, and Population Statistics Advisory Committees intoone group called the Census Advisory Committee of ProfessionalAssociations. This single Advisory Committee is composed of thememberships from these professional associations and was orga-nized into four individual subgroups.

2A fourth subgroup of the committee, the Population StatisticsAssociation, was not as directly concerned with the economiccensus as the other three subgroups.

Appendix C C-1History−1997 Economic Census


Members who served during the planning and conductof the 1997 Economic Census were as follows:

• Robert Bell, to 2000 (Chairperson 1997)

• David Binder, to 2000

• G. David Faulkenberry, to 1997

• Joseph Garrett, 1996 - 2000

• Malay Ghosh, 1996 - 2000

• F. Thomas Juster, 1997 - 2000

• Richard Kulka, 1990 - 1995

• Nancy A. Mathiowetz, 2000 -

• William O’Hare, 1996 -

• Judith Rowe, to 1995

• Elizabeth Stasny, 1994 - 2000

• Lynne Stokes, to 2000

• Roger E. Tourangeau, to 1999

• Joseph Waksberg, 1991 - 1996 (Chairperson, 1995)

THE AMERICAN ECONOMIC ASSOCIATION

Various members of the American Economic Association(AEA), together with representatives of the ASA, served ona single Census Bureau advisory committee from 1919 to1937. In 1960, when the Census Bureau was entering newareas of interest to economists and it appeared desirableto have a direct channel of contact with the principal pro-fessional organizations representing economists, a sepa-rate committee of AEA members was established. (From1937 to 1960, the AEA was not represented by any par-ticular committee).

The president of the AEA appointed all members whoserved for 3-year (4-year prior to 1972) terms. While mostof the committee’s attention was given to the economiccensuses, it also reviewed and commented on plans forother censuses, evaluation studies, and program planningwith the larger framework of the Census Bureau’s func-tions.

Committee members who served during the planningand conduct of the 1997 Economic Census were as fol-lows:

• Ernst Berndt, 1996 - 2001

• Roger R. Betancourt, 1997 -

• Lynn E. Browne, 1998 - 2000

• David Card, to 1996 (Chairperson, 1995)

• Ellen R. Dulberger, 1992 - 1997 (Co-chairperson 1997)

• William C. Dunkelberg, 1992 - 1997 (Co-chairperson1997)

• Gail Fosler, 1990 - 1995

• Michael Gort, to 2000

• Bronwyn Hall, 1990 - 1995

• Lee Lillard, to 2000

• Rebecca A. Maynard, 1998 - 2000

• Walter Oi, to 1996 (Chairperson 1993)

• Ariel Pakes, 1996 - 2001

• Frederic M. Scherer, 1997 - 1999

• Daniel T. Slesnick, 2000 -

• Philip L. Swann, 2000 -

• Robert Willis, to 2000

THE AMERICAN MARKETING ASSOCIATION

In 1945, the American Marketing Association (AMA)passed a resolution emphasizing the need for marketinginformation to promote the Nation’s economy and outlin-ing a suggested statistical program for the Census Bureau.The agency felt that a regular advisory committee pro-vided the most efficient way to learn the needs for market-ing statistics and to determine whether these needs couldbe met. Accordingly, in 1946, the Census Bureaurequested that a committee be established for this pur-pose, and the president of the AMA appointed its mem-bers for 3-year terms.

Committee members who served during the planningand conduct of the 1997 Economic Census were as fol-lows:

• Tony Adams, 1996 - (Chairperson 1997)

• Laurie Ashcraft, 1992-1998

• Kenneth Bernhardt, 1992-1997

• Dorothy Clark 1989-1995

• Keith Cox 1991-1996

• Michael Etzel, 1997 - 2000

• Beth Fischer 1993-1998

• Catherine Bock-Jocz 1995 - 2001

• James Myers 1993-1995

• Connie Pechmann, 1998 - 2001

• Robert Peterson, 1998 - 2000

• Arthur Redmond, 1997 - 2000

• Mary Lou Roberts 1992-1997

• Debra Semans 1993 - 1999

• Carol Shea, 1998 - 2000

• Rosann Spiro 2000 -

• Sybil Stershic, 1999 -

• David Stewart 1993 - 1998 (Chairperson, 1996)

• Doris Walsh 1995 - 1997

C-2 Appendix C History−1997 Economic Census


Appendix D.Consultation on the Census Questionnaires

More than a thousand agencies, trade associations, trade journals, and other organizations, including companies report-ing in the censuses, were requested to recommend questionnaire content for the 1997 Economic Census. The list belowshows by major North American Industry Classification System (NAICS) economic sector those organizations thatresponded. To protect confidentiality, companies in the censuses are not listed.

UTILITIES (SECTOR 22)

Other Organizations

American Public Power AssociationAmerican Water Works AssociationAssociated Gas DistributorsAssociation of Oil PipelinesEdison Electric InstituteNational Association of Water Companies

WHOLESALE TRADE (SECTOR 42)

Federal Agencies

U. S. Department of CommerceBureau of Economic Analysis

United States Department of LaborBureau of Labor Statistics

Other Organizations

American Home Sewing and Craft AssociationAmerican Petroleum InstituteAmerican Supply and Machinery Manufacturers’Association

Automotive Service Industry AssociationBeauty and Barber Supply InstituteBusiness Technology AssociationCleaning Equipment Trade AssociationDental Dealers of AmericaFresh Produce CouncilInternational Association of Plastic DistributorsInternational Sanitary Supply AssociationMaterial Handling Equipment Dealers AssociationNational Association of Electrical Distributors, Inc.National Beer Wholesalers AssociationNational Decorating AssociationNational Electronic Distributors AssociationNational Fisheries InstituteNational Food Distributors AssociationNational Marine Manufacturers AssociationNational Paint and Coatings AssociationNational Wholesale Druggists’ AssociationNorth American Equipment Dealers AssociationNorth American Wholesale Lumber AssociationPetroleum Marketers Association of America

WHOLESALE TRADE (Sector 42)−Con.

Other Organizations−Con.

Quality Bakers of America CooperativeSteel Manufacturers AssociationSteel Service Center InstituteTobacco Merchants Association of the United StatesUnited Fresh Fruit and Vegetable Association

RETAIL TRADE (SECTOR 44-45)

Federal Agencies


U. S. Department of LaborBureau of Labor Statistics

U. S. Department of AgricultureEconomic Research Service

Other Organizations

American Booksellers AssociationAmerican Furniture Manufacturers AssociationAmerican Home Sewing and Craft AssociationAmerican Petroleum InstituteAmerican Pharmaceutical AssociationAmerican Professional Needlework RetailersAmerican Professional Pet DistributorsAmerican Specialty Toy Retail AssociationArt Business NewsAutomotive Parts and Accessories AssocationAutomotive Service Industry AssociationBMT PublicationsBook Industry Study GroupCahners EconomicsDairy Management IncorporatedDirect Selling AssociationElectronic Industries AssociationFlower NewsFootwear Industries of AmericaGift Association of AmericaGreater Blouse, Skirt, and Undergarment Association,IncorporatedHardware Age

Appendix D D-1History−1997 Economic Census


RETAIL TRADE (Sector 44-45)−Con.


Hearth Products Association Hobby IndustryAssociation of AmericaHome Improvement Research InstituteInternational Sewing Machine AssociationInternational Sleep Products AssociationJewelers CircularJewelers of AmericaManufactured Housing InstituteMarket Statistics, IncorporatedMotor Age MagazineNational Association of Music MerchantsNational Association of Resale and Thrift ShopsNational Automatic Merchandising AssociationNational Automotive Dealers AssociationNational Coal AssociationNational Decorating Products AssociationNational Glass AssociationNational Home Center NewsNational Home Furnishings AssociationNational Marine Manufacturers AssociationNational Retail FederationNational Retail Hardware AssociationNational Sporting Goods AssociationNonPrescription Drug Manufacturers AssociationOpticians Association of AmericaPainting and Decorating Contractors of AmericaPhoto Marketing Association InternationalProgressive GrocerRecreational Vehicle Dealers AssociationSociety of Independent Gasoline MarketersSporting Goods Manufacturers AssociationThe American Institute of Food DistributorsUnited Dairy Industry AssociationUnited Fresh Fruit and Vegetable AssociationWorld Floor Covering Association

TRANSPORTATION AND WAREHOUSING(SECTOR 48-49)

Other Organizations

American Bus AssociationAmerican Public Transit AssociationAmerican Trucking AssociationAmerican Warehouse AssociationHousehold Goods Forwarders AssociationInternational Association of Refrigerated WarehousesInternational Marina InstituteNational Customs Brokers and Forwarders Association ofAmericaNational Marine Manufacturers AssociationNational School Transportation AssociationNational Waterways Conference, Inc.Regional Airline Association

TRANSPORTATION AND WAREHOUSING(Sector 48-49)−Con.


Transportation Brokers Conference of AmericaUnited Bus Owners of AmericaUnited States Tour Operators Association

INFORMATION (SECTOR 51)

Federal Agencies



Other Organizations

Association of Telemessaging Services International, Inc.Interactive Multimedia AssociationNational Association of BroadcastersNational Telephone Cooperative Association

FINANCE AND INSURANCE (SECTOR 52)

Federal Agencies

Securities and Exchange Commission

Other Organizations

Alliance of American InsurersAmerica’s Community BankersAmerican Bankers AssociationAmerican Council of Life InsuranceCommercial Finance AssociationFinancial Institutions of Marketing AssociationsHealth Insurance Association of AmericaIndependent Bankers Association of AmericaInvestment Company InstituteLife Insurance Marketing and Research AssociationMortgage Bankers Association of AmericaNational Association of Credit Union Service OrganizationsNational Association of Health UnderwritersNational Association of Insurance BrokersNational Association of Life UnderwritersNational Association of Securities Dealers, Inc.National Association of State Credit Union SupervisorsNational Council of Real Estate Investment FiduciaryNational Credit Union AdministrationNational Futures AssociationNational Society for Real Estate and FinanceNew York Stock ExchangeRisk and Insurance Management SocietySecurities Industry AssociationSecurities Transfer AssociationSecurity Traders AssociationSociety of Chartered Property and Casualty Underwriters

D-2 Appendix D History−1997 Economic Census


FINANCE AND INSURANCE (Sector 52)−Con.

Other Organizations—Con.

Society of Professional Benefit AdministratorsSurety Association of AmericaSheshunoff Information Services

REAL ESTATE AND RENTAL AND LEASING(SECTOR 53)

Other Organizations

Building Owners and Managers Association InternationalManufactured Housing InstituteNational Council of Real Estate Investment FiduciaryNational Society for Real Estate and FinanceNational Truck Leasing SystemTruck Renting and Leasing Association

PROFESSIONAL, SCIENTIFIC ANDTECHNICAL SERVICES (SECTOR 54)

Federal Agencies



Other Organizations

American Congress on Surveying and MappingAmerican Consulting Engineers CouncilAmerican Institute of Certified Public AccountantsAmerican Society of Interior DesignersIndustrial Design Services AssociationInternational Interior Design AssociationMarketing Research AssociationNational Technical Services AssociationPhoto Marketing Association InternationalProfessional Photographers of America, Inc.

ADMINISTRATIVE AND SUPPORT, WASTEMANAGEMENT AND REMEDIATION SERVICES(SECTOR 56)

Federal Agencies



Other Organizations

American Society of Travel AgentsAmerican Telemarketing Association, Inc.Associated Credit BureausTravel Industry Association of America

HEALTH CARE AND SOCIAL ASSISTANCE(SECTOR 62)

Federal Agencies



Other Organizations

American Academy of PediatricsAmerican Association of Homes and Services for the AgingAmerican Association of Tissue BanksAmerican Chiropractic AssociationThe American College of Obstetricians and GynecologistsAmerican Dental AssociationInternational Society for Clinical Laboratory TechnologyNational Association of Meal Programs

ARTS, ENTERTAINMENT AND RECREATION(SECTOR 71)

Federal Agencies



Other Organizations

American Association of Botanical Gardens and ArboretaAmerican Association of MuseumsAmerican Symphony Orchestra LeagueInternational Association of Amusement Parks andAttractionsNational Association of Theatre OwnersNational Bowling CouncilNational Sporting Goods Association

ACCOMMODATION AND FOODSERVICES(SECTOR 72)

Federal Agencies



Other Organizations

American Hotel and Motel AssociationInternational Foodservice Manufacturers AssociationNational Association of RV Parks and CampgroundsNational Restaurant AssociationProfessional Association of Innkeepers, International

Appendix D D-3History−1997 Economic Census


OTHER SERVICES (EXCEPT PUBLICADMINISTRATION) (SECTOR 81)

Federal Agencies



Other Organizations

Automotive Oil Change AssociationNational Funeral Directors AssociationTextile Care Allied Trades Association

D-4 Appendix D History−1997 Economic Census


PO

L/00

-HEC

Histo

ry o

f the 1

99

7 E

co

no

mic

Cen

sus

history of the 1997 economic census · history of the 1997 economic census issued july 2000...

Documents