PART 1
• INSERT COURSE OUTLINE
MARKETING CASE REPORT FORMAT
A. Executive Summary: Self-Contained Document, one to two pages
• Statement of Purpose and Issues to be Addressed
• Research Method Used to Address Issues• Salient Findings
(Appears before Table of Contents)
B. Table of Contents• Subject and Page Numbers Including All Exhibit
References
I. Introduction• Background• Purpose and/or Problem Definition• Objectives of Report
II. Methodology• Specific Methodology – Why!!!• Data/Information to be Studied
III. Case Analysis• Application of Specific Methodology to Case• Discussion/Explanation of Analysis• Interpretation of Tables and Charts. (It is not acceptable
to merely refer to Tables, e.g., see Table X)
IV. Findings and/or Conclusions
V. Appendices
VI. Other Requirements• Paragraph and Sub Paragraph headings• Identification of all exhibits which are to be
explained and referenced in text• No Misspellings!!!!• Proper Grammar• Interesting Style• On Time Delivery of Oral and Written Report
Marketing Research and the Four Ps1. Products
New Products Evaluating Packaging and Brand Designs Compassion Studies With Competitor’s Products Consumer Evaluation of Current Products
2. Place (Distribution Channels) Analysis of Different Storage or Transportation Methods Analysis of Alternative Sites Determination Of Inventory Levels Growth Rates of Different Channels
3. Promotion Testing Different Ad. Messages Establishing Sales Territories Selecting Media Evaluating Ad. Effectiveness
4. Pricing
Research on MarketsForecasting DemandProviding Information of General TrendsProviding Information For Segmenting
MarketsDeveloping Customer ProfilesIdentifying New Markets For Existing ProductsIdentifying New Product NeedsForeign Markets
Elements Of The Marketing Mix That Compose A Cohesive Marketing Program
Marketing Manager
ProductFeaturesBrand namePackagingServiceWarranty
PriceList priceDiscountsAllowancesCredit termsPayment period
PromotionAdvertisingPersonal sellingSales promotionPublicity
PlaceOutletsChannelsCoverageTransportationStock level
ProductPromotion
Price Place
PART 2
Introduction to Marketing Research
Dr. Doherty
Tobin College of Business
St. John’s University
Marketing Research
Three Generic ApproachesI. ExploratoryII. DescriptiveIII. Casual/Experimental
Applications: See Tables 1 and 2
Definition:A scientific approach to(a) the collection; (b) analysis; and (c) presentation
of data/information to be used in the management decision making process
The Exploratory Approach
Purpose: Identify Potential Relevant Factors (Don’t try to solve the problem!)
• Develop Hypothesis• Establish priorities for further research• Identify information and data sources• Clarify concepts• Increase analysts’ familiarity with problem(s)• Identify potential causes
The Exploratory Approach
Five Popular Exploratory Approaches:
1) Literature Search
2) Experience Survey
3) Analysis of Selected Cases
4) Focus Groups
5) “Small” Sample/Surveys/Interviews
The Descriptive Approach
Purpose: Test Hypothesis• Analyze Data• Develop Findings/Conclusions
Two Types (Depending on Type of Data)A. Longitudinal (Time Series)
True Panel Omnibus Panel
B. Cross Sectional Field Survey Field Study
True Panel Application
The Brand Switching Matrix or Turnover Table (see your textbook!)
Brand T1 T2A 200 250B 300 270C 350 330D 150 150
Total 1000 1000
Time Period Brand A B C D TotalA 175 25 0 0 200
(.875)B 0 225 50 25 300
(.750)C 0 0 280 70 350
(.800)D 75 20 0 55 150
(.367)Total 250 270 330 150 1000
Time (T1)
Time (T2)
Applications of Turnover Table
Evaluating:a) Price Changes
b) Promotional Campaigns
c) New Packaging
d) New Products
e) Results can be integrated with other databases to determine customer profiles and media habits
Causal/ExperimentalResearch Design
1. Scientific Criteria– Concomitant Variation– Time Sequence– Elimination of Other Causes
2. Controlled Experiment– Reflects 1.– Lab vs. Field– Validation– Two Groups: Experimental and Control
3. Basic Concepts Defined– Experiment : Process– Treatments : Alternatives– Test Units : Entities– Dependent Variables : Measures– Extraneous Variables
• Hold Constant• Randomize Assignment of Treatments• Specific Design• ANCOVA
Types of Evidence That Support a Causal Inference
• Concomitant Variation– evidence of the extent to which X and Y occur together or vary together in the way predicted by the hypothesis
• Time order of occurrence of variables- evidence that shows X occurs before Y
• Elimination of other possible causal factors- evidence that allows the elimination of factors other than X as the cause of Y
X– the presumed causeY– the presumed effect
Types of Experiments
Scientific investigation in which an investigator manipulates and controls one or more independent variables and observes the dependent variable for variation concomitant to the manipulation of the independent variables
Experiment
Laboratory Experiment
Research investigation in which investigator creates a situation with exact conditions so as to control some, and manipulate other, variables.
Field Experiment
Research study in a realistic situation in which one or more independent variables are manipulated by the experimenter under as carefully controlled conditions as the situation will permit.
Types of Extraneous Factors That Can Contaminate Research Results
History—Specific events external to an experiment, but occurring at the same time, which may affect the criterion or response variable.
Maturation—Processes operating within the test units in an experiment as a function of the passage of time per se.
Testing—Contaminating effect in an experiment due to the fact that the process of experimentation itself affected the observed response.
Main testing effect—The impact of a prior observation on a later observation.
Interactive testing effect—The condition when a prior measurement affects the test unit’s response to the experimental variable.
Instrument Variation—Any and all changes in the measuring device used in an experiment that might account for differences in two or more measurements.
Statistical Regression—Tendency of extreme cases of a phenomenon to move toward a more central position during the course of an experiment.
Selection Bias—Contaminating influence in an experiment occurring when there is no way of certifying that groups of test units were equivalent at some prior time.
Experimental Mortality—Experimental condition in which test units are lost during the course of an experiment.
Test Marketing1. Who?2. Objectives
a. Forecasts: Sales, Market Share; CANNALBALISTIC EFFECTS
b. Pretest Market Mixc. Serendipity
3. Key Decisionsa. How Many Cities?
2 To 6 Importance of Regional Differences Degree of Uncertainty
b. Which Cities?Syracuse Leonia Dayton Des Moines
c. Length Of Test? 2 Months to 2 Years Average Repurchase Period Competition Concern First to Market Importance
Test Marketing Cont’dd. What Data?
Warehouse Shipments
Store Audits Consumer Panels Buyer Surveys Trade Attitudes
4. What Action?Trial Rate
Repurchase Rate
High Low
High Go! More Adv.
Low Product Flaw Bust!
BAYESIAN WORK TABLEState of Nature (Sj or Ej)
Prior Prob. P(Sj)
Cond’l Prob. P(Zk/Sj)
Joint Prob. P(ZkSj)
Posterior Prob. P(Sj/Zk)
Z1:
S1 0.6 0.7 0.42 0.858
S2 0.3 0.2 0.06 0.122
S3 0.1 0.1 0.01 0.020
0.049 1.000Z2:
S1 0.6 0.2 0.12 0.364
S2 0.3 0.6 0.18 0.545
S3 0.1 0.3 0.03 0.091
0.33 1.000Z3:
S1 0.6 0.1 0.06 0.333
S2 0.3 0.2 0.06 0.333
S3 0.1 0.6 0.06 0.333
0.18 1.000
Computation of Expected Values From BAYESIAN Work Table
GivenZ1 (Test MKT. Results show Light D)
EV(A1)= 100(.858) + 50 (.122) + -50(.02)= $90.9MEV(A2)= 50 (.858) + 100(.122) + -25(.02)= $54.6MEV(A3)= -50 (.858) + 0 (.122) + 80 (.02)= -$41.3M
Z2 (Test MKT. Results Show Moderate D)
EV(A1)= 100(.364) + 50 (.545) + -50(.091)= $59.1MEV(A2)= 50 (.364) + 100(.545) + -25(.091)= $70.4MEV(A3)= -50 (.364) + 0 (.545) + 80 (.091)= -$10.9M
Z3 (Test MKT. Results Show Heavy D)
EV(A1)= 100(.333) + 50 (.333) + -50(.333)= $33.3MEV(A2)= 50 (.333) + 100(.333) + -25(.333)= $41.6MEV(A3)= -50 (.333) + 0 (.333) + 80 (.333)= $10.0M
Prob. OF Obtaining Each Test MKT. Result
P(Zk) = j=1 k P(Sj) P(Zk/Sj)
P(Z1) = P(S1)P(Z1/S1) + P(S2)P(Z1/S2) +P(S3)P(Z1/S3)
= (.6) (.7) + .3(.2) + .1(.1)
= 0.49
P(Z2) = P(S1)P(Z2/S1) + P(S2)P(Z2/S2) +P(S3)P(Z2/S3)
= .6(.2) + .3(.6) + .1(.3)
= 0.33
P(Z3) = P(S1)P(Z3/S1) + P(S2)P(Z3/S2) +P(S3)P(Z3/S3)
= .6(.1) + .3(.2) + .1(.6)
= 0.18
Prob. OF Obtaining Each Test MKT. Result
FORECASTS Decision Acts Opt. EV Prob.
Z1 A1 90.9 0.49
Z2 A2 70.4 0.33
Z3 A3 41.6 0.18
EV(Research) = 90.9(.49) + 704.(.33) + 41.6(.18)
= $75.26M
EV(U) = ’70.0M
Max Price For Res. = EV(R)-EV(U)= 75.26 – 70.0 = $5.26M
Causal/ExperimentalResearch Design
4. Validation– Internal vs. External
5. Internal– History– Maturation– Mortality– Regression– Instrumentation– Selection Bias– Main Testing Effect– Interactive Testing Effect
6. Four Types of Experimental Research Designs1. Pre Exp (3)2. True Exp (3)3. Quasi (3)4. Advanced Statistical Design (4)
Causal/ExperimentalResearch Design
7. Pre-Exp. Design (3)a. After Only: X O
b. Before After: O X O
c. Static Group Comparisons: X O1 O2
Major Errors: H, SB
Causal/ExperimentalResearch Design
R Exp O1 X O2
R Control O3 O4
X = (O2 – O1) – (O4 – O3)
R Exp X O1
R Control O2
X = O1 – O2
a. Before/After with Randomization (R) and Control (C)
b. After Only with R and C
8.True Experimental Design
R O1 X O2
R O3 O4
R X O5
R O6
c. Solomon 4 Group
EXT = ?ITE = ?X = ?
O1 = 100O2 – 160O3 = 106
O4 = 140O5 = 150O6 = 135
Problem
Causal/ExperimentalResearch Design
9. Quasi Exp (3)A. Single Time Series
O1 O2 O3 X O4 O5 O6
B. Multiple Time Series
O'1 O'2 O'3 X O'4 O'5 O'6
C. Separate Sample Before/After Design:R: O1 X
R: X O2
Main Problem of Quasi Approach: History(Note: 9A is typical of consumer panel investigation data.)
Causal/ExperimentalResearch Design
10. Advanced Statistical Design (4)A. CRD
B. RBD
C. LSD
D. Factorial
Part 2ADecision Making Under
Uncertainty
Criteria for Selecting the Best Option
• MAX/MIN
• MAX/MAX
• MIN/MAX-REGRET
•EXPECTED VALUE
Value of Information• Payoff (Decision) Table
EVENTS (States of Nature)Management
OptionsE1 Ez . . . En
A1 X11 X12 X1n
A2 Xz1 X22 Xzn
An Xn1 Xnz Xnn
Prior Probabilities
(P1) (Pz) (Pn)
AI : Decision ActsEj : Events (or Sj = States of Nature)Eij : Payoff or ConsequencesPj : Prob. Associated with Ej
ILLUSTRATION
E1 E2 E3 E4
A1 80M 40 -10 -50
A2 30 40 30 10
A3 20 30 40 15
A4 5 10 30 20
Regret Table
E1 E2 E3 E4 MAX
A1 0 0 50 70 70
A2 50 0 10 10 50
A3 60 10 0 5 60
A4 75 30 10 0 75
Part 2B Marketing Research Case Study
Bayesian Analysis
Bayesian CaseObjective: Determine Value of Research
Problem
S1 S2 S3
A1 100 50 -50
A2 50 100 -25
A3 -50 0 90
Prior Probs. 0.6 0.3 0.1
P(Sj)
EV(A1)=$70M
EV(A2)=$57.5M
EV(A3)=-$22M
EV(L1)=$70M
EV(C)=$98M
EV(PI)=$28M= EV(C) – EV(U)
EV(C)= .6(100) + .3(100)+ .1(80) = $98M
EV(PI)= EV(C) – EV(U) = $98M - $70M = $28M
Conditional Prob. MatrixActual Results
Test MKT Results
S1 S2 S3
Light D Z1 0.7 0.2 0.1 Should, but not
necessary to, sum to
one.
Mod. D Z2 0.2 0.6 0.3
Heavy D Z3 0.1 0.2 0.6
Must sum to one 1.00 1.00 1.00
Bayesian Work Table
0.6 0.7 0.42 0.8570.3 0.2 0.06 0.1220.1 0.1 0.01 0.020
0.49 1.000
0.6 0.2 0.12 0.3640.3 0.6 0.18 0.5450.1 0.3 0.03 0.091
0.33 1.000
0.6 0.1 0.06 0.3330.3 0.2 0.06 0.3330.1 0.6 0.06 0.333
0.18 1.000
State of Nature
(Sj or Ej)
Prior Prob. P(Sj)
Cond’l Prob. P(Zk/Sj)
Joint Prob. P(ZkSj)
Posterior Prob. P(Sj/Zk)
Z1: S1
S2
S3
Z2: S1
S2
S3
Z3: S1
S2
S3
Computation of Expected Values from BAYESIAN Work Table
Given:Z1 (Test MKT. Results show Light D)EV(A1) = 100(.858) + 50(.122) + -50(.02) = $90.9MEV(A2) = 50(.858) + 100(.122) + -25(.02) = $54.6MEV(A3) = -50(.858) + 0(.122) + 80(.02) = $-41.3M
Z2 (Test MKT. Results show Moderate D)EV(A1) = 100(.364) + 50(.545) + -50(.091) = $59.1MEV(A2) = 50(.364) + 100(.545) + -25(.091) = $70.4MEV(A3) = -50(.364) + 0(.545) + 80(.091) = $-10.9M
Z3 (Test MKT. Results show Heavy D)EV(A1) = 100(.333) + 50(.333) + -50(.333) = $33.3MEV(A2) = 50(.333) + 100(.333) + -25(.333) = $41.6MEV(A3) = -50(.333) + 0(.333) + 80(.333) = $10.0M
P(Zk) = P(Sj)P(Zk/Sj)
P(Z1) = P(S1)P(Z1/S1) + P(S2)P(Z1/S2) + P(S3)P(Z1/S3)= (.6)(.7) + (.3)(.2) + (.1)(.1)= 0.49
P(Z2) = P(S1)P(Z2/S1) + P(S2)P(Z2/S2) + P(S3)P(Z2/S3)= (.6)(.2) + (.3)(.6) + (.1)(.3)= 0.33
P(Z3) = P(S1)P(Z3/S1) + P(S2)P(Z3/S2) + P(S3)P(Z3/S3)= (.6)(.1) + (.3)(.2) + (.1)(.6)= 0.18
Probability of Obtaining Each Test MKT. Result
k
j 1
Probability of Obtaining Each Test MKT. Result (cont’d)
FORECASTS Decision Acts Opt. Ev Prob.
Z1 A1 90.9 0.49
Z2 A2 70.4 0.33
Z3 A3 41.6 0.18
EV(Research) = 90.0(.40) + 70.4(.33) + 41.6(.18)
= $75.26M
EV(U) = 70.0M
Max Price For Res. = EV(R) – EV(U)
= 75.26 – 70.0
=$5.26M
Case Description
Newco is a manufacturer of natural soft drink beverages. It has recently experienced a decline in market share. To reverse this decline, management is considering a new promotional program that will cost $1 million. Management believes that the program may have three possible effects:
1. Very Favorable: 10% increase in market share; $4 million increase in profits.
2. Favorable: 5% increase in market share; $1 million increase in profits.
3. Unfavorable: (No Effect on Sales) – incremental loss of $1 million, the cost of the program.
Abbey Normal, Director of Marketing Research, estimates the probability of the three events as follows:
S1: Very Favorable Consumer Reaction = 0.30
S2: Favorable Consumer Reaction = 0.40
S3: Unfavorable Consumer Reaction = 0.30
Newco is considering a proposal made by Marketing Testing Experts (MTE), a private consulting firm, to asses the potential effects of the program.
MTE has advised Newco that based on its past experience of assessing promotional programs that the following results on average have been obtained:
MTE proposes a charge of $250,000 for conducting the research.
Customer Reaction
MTE Experience Very Favorable Favorable Unfavorable
Strongly Positive 0.7 0.2 0.0
Moderately Positive
0.3 0.6 0.2
Slightly Positive 0.0 0.2 0.8
Questions:
1. Construct the relevant payoff table.
2. What are the maximin and maximax solutions?
3. What is the solution according to the expected value criterion?
4. What is the value of perfect research information?
5. Should Newco except MTE’s proposal? Why?
6. What price would Newco be willing to pay for the study?
7. What probabilities are critical to the outcome of the study?
8. How could the various probabilities that are needed for such a study be obtained in practice?
Note: There are many computer software packages, that can be run on a PC, mainframe and microcomputer that can be used to solve this problem. See, for example, D.A. Schellinck and R.N. Maddox, Marketing Research: A Computer Assisted Approach, The Dryden Press, 1987.
PART 3
SECONDARY SOURCES OF DATA
FIVEFOLD (5) CLASSIFICATION
1. INTERNAL• P&L• Balance Sheet• Sales Figure• Sales-Call Reports• Invoices• Inventory Records• Prior Research Studies
2. PERIODICALS & BOOKS• Business Periodicals Index (Monthly Publications that
provide a list of business articles appearing in a wide variety of business publications).
• Standard & Poor’s Industry surveys (provides updated statistics and analyses of industries).
• Moody’s Manuals (financial data and names of executives in major corporations).
• Encyclopedia of Associations (provides information on every major trade and professional association in the U.S.
• Marketing Journals• Trade Magazines (Advertising Age, Chain Store Age
progressive Grocer, Sales and MKT. MGT, Stores).• Business Magazines (Fortune, Business Week, Forbes,
Barrons, Harvard Business Review, etc.)
3. COMMERCIAL DATA• A.C. Nielsen Co.
1) Retail Index Service (data on products and brands sold through retail outlets)
2) Scan track (Supermarket scanner data)Electronic Test MKTa. Scanner Cards for Panel Membersb. Demographicsc. TV Viewers Habit of Panel Members
3) Media Research Services (Television Audience)4) Neodata Service Inc. (Magazine Circ.)5) Home Services – National Purchase Diary Panel
• MRCA – National Purchase Diary Panel National Menu Census (data on home food consumption)
COMMERCIAL DATA (CONTINUED)• Claritas – buying habits of 250,000 U.S. neighborhoods• Information Resources Inc. – provide supermarket scanner data
1. (InfoScan); also2. Promotio Scan – IMPACT of supermarket promotions
• SAMI/BURKEProvides reports on warehouse withdrawals to food stores in selected market areas (SAMI reports) and supermarket scanner data (SAMSCAN)
• SIMMONS Market Research Bureau (MRB Group)Provides annual reports covering television market, sporting goods, proprietary drugs.
Giving demographic data by sex; income; age and brand preference (selective market and media reaching them)
• OtherAudit Bureau of Circulation ArbitronAudit and SurveysDunn and BradstreetNational Family OpinionStandard Rate and Data ServiceStard
4. GOVERNMENT PUBLICATION• Statistical Abstract of MKT Sources (updated annually)
Provides summary data on: demographic, economy, social and other aspects of the U.S. economy and society.
• County and City Data Book (updated every three years)-Presented statistical information for counties, cities and other
geographical units regarding:
- population, education, employment
- aggr. And med. Income – housing
- bank deposit, retail sales, etc.
• U.S. Industrial Outlook-Projections of industrial activity by industry and includes data on:
production
sales
shipment
employment
• Marketing Information GuideProvides a monthly annotated bibliography of marketing information.
• Other- Annual Survey of Manufacturers- Business Statistics- Census of Manufacturers- Census of Retail Trade, Wholesale Trade and Selected Service Industries- Census of Transportation- Federal Reserve Bulleting- Monthly Labor Review- Survey of Current Business- Vital Statistics Report
5. COMPUTERIZED DATA BASEDefinition: A collection of numeric data and/or textual information that is available on computer readable form.
e.g.: BibliographicABI/INFORMPredicastNumeric
1. 1990 Census DataDonnelly MKTDRI
2. Nielsen Retail Product MovementSAMI
3. SPI (Strategic Planning Institute) -250 CompaniesPIMS
Work Index: Sponsored by Cornell University’s School of Industrial Labor
Relations and Human Resource Executive magazine, this site provides links to resources on labor relations, benefits, training, technology, staffing, recruiting, leadership, legal issues and related topics.
MarketingAdvertising World Links to resources in selected areas of marketing
and advertising.American Association of Advertising Agencies Provides membership
information, recent bulletins, and links to related resources.American Marketing Association Provides information on
membership, publications, and conferences.Guerrilla Marketing Online Provides access to recent articles in
marketing and links to relevant sites.
Marketing Cont’dInstitute for the Study of Business Markets (ISBM) Features
current information about seminars and research projects. Includes marketing links.
John W. Hartman Center for Sales, Advertising & Marketing History (Duke University Libraries) Center promotes study of sales, marketing, and advertising history. Features “Ad*Access,” an image database of over 7,000 advertisements printed in U.S. and Canadian newspapers between 1911 and 1955. Database allows keyword searching.
Project 2000 Home Page Provides access to working papers, course syllabi, and related links.
Yahoo – Business and Economy: Marketing Provides links to marketing web sites
Marketing Information: A Bibliography
Statistical Sources
Business Resources on the Web: Economic Statistics, Government Statistics, and Business Law Maintained by Boise State University’s Albertsons Library, contains extensive links to statistics sources for the economy, population, international trade, statistics by state, etc. Primarily dedicated to statistics sources, but also contains a business law component
Fisher College of Business Financial Data Finder Links to financial and economic data on the web and elsewhere.
PART 3A
Profiling Customers
Dr. Doherty
Tobin College of Business
St. John’s University
Industrial
• Dun’s Market Identifiers (DMI)– D&B’s market information service. A record of
over 7 million establishments updated monthly
• Enhanced DMI extends 4 digit S/C codes to 6 and 8 digits to allow clients to target specific customer groups
Consumer• Geodemographers
– R.L. PoleProduct for Retailers: Vehicle Origin Survey
Samples cars parked in retailer parking lots and identifies (from the Vehicle Registration Database) their home location. Can also match location with Census data and via their TIGER files provide a demographic profile of customers
– ClaritasUses 500+ demographic variables in its Prigm (Potential
Ratings for Zip markets) database to classify 250,000 neighborhoods
40 types based on consumer behavior and lifestyle
(shotguns, pickups, patios and pools, etc.)
Consumer
• Diary Panels– NPD (13,000 HHs)
30 Product Categories– 29 Miniature Panels– Quota Sampling– Applications
» Brand Shares» Brand Switching Behavior» Frequency of Purchase and Amounts» Evaluation of Price and Promotions» Changes in Channels and Distribution» Size of Market
Consumer• Store Audits
– Nielsen Retail Index(Drug stores, Mass media indexes and liquor stores)
– Now Use ScannersBeginning Inventory and Net purchase (from wholesalers and
manufactures) – Ending Inventory = Sales• Audit Includes
– Sales– Purchases by retailers– Inventories– Number of Days of Supplies– Out-of-stock stores– Prices (retail and wholesale)– Special factory packs– Promotions and Advertising
Consumer
– Disaggregate data by• Competitors• Geographic area• Store type
– Nielsen’s Scantrack supplements its Retail index (since 1970’s)
• 11 digit WPC code• Evaluates
– Promotions– Price changes– Channel trends– Product trends
• 40,000 HHs using scanner wands
Consumer
• Behavior Scan (provided by Information Resources)– 3,000 HHs provided scanner cards– Supermarkets and Drugstores provided with
scanner– With coorperation from Cable TV Companies It
links view habits with purchase (Black Boxes)– Distinguishes Users from nonusers of products
WRT …/promotions
Consumer• Television
– Nielsen TV Index• Audimeters attached to TV sets and tied into a central computer. Replaced
by People Meters in 1988.• Aggregate ratings by 10 socioeconomic groups and demographic
characteristics, including territory, ed. Of head of H.H., age of woman in house, etc.
• Radio– Arbitron
• Panel of HHs are randomly selected who have agreed to complete diaries. Radio marketing are rate 1-4 times age during the “Sweeps” period (April/May). Focus on age, sex, and individual (USHH) behavior
• Print Media– Starch Readership Service– Evals. 50,000 ads in 1000 print media (mag., bus. Publications,
newspapers); u=75,000 person interview– Recognition method: 3 degrees
1. Noted. Remembers any part of ad2. Associated (1) plus recalls brand or advertise3. Read Most recalls 50% or more of the written material
Multimedia Services• Simmons Media/Mkt Service
– Prob. Sample of 19,000+– Cross references product usage and media exposure– 4 different interviews with each respondent
• Magazine, TV, Newspaper, Radio– Results disaggregated by sex– Self –administered questions covering 500 product categories– TV view behavior gathered by means of a personal diary; Radio via both
personal and telephone interviews– Demographics collected– Application Segmentation and targeting by firms
• Mediamark– Similar service, problem sample of 20,000– Tends to establish audience rate 10% higher than Simmons (see p 252)
• Mail Panels– NFO Research
• Quota Sample of 400,000 HHs• Rebuilt every two years• Self-adm q
– Market Facts, Inc,• Quota Sample of 275,000• Cross Tabulation of Aug. Criterion Variable (Adv. Sales, etc) with anyone or number of
demographic variables (Age, sex, automobile,…, pets ordered, etc)
PART 3B
Determining Market Potential
Dr. Doherty
Tobin College of Business
St. John’s University
Determining Market Potential
• Multiple-Factor Index Method(“Annual Survey of Buying Power” published by
Sales and Marketing Management )
Purpose: Measure the relative consumer buying power in different region, state, and metropolitan areas.
Determining Market PotentialBi = 0.5yi + 0.3ri + 0.2pi
where
Bi : % of total national buying power found in area iyi: % of national DI in area iri: % of nat’l retail sales in area ipi: % of nat’l population in area i
Example 1: drug sales
Suppose N.Y. State has: yi = 5.0%, ri = 10.0%, pi = 8.0%Bi = 0.5(5.0) + 0.3 (10.0) + 0.2(8.0) = 7.1Thus, 7.1% of the nation’s drug sales would be expected to occur in NY.
If the total drug sales are $50 Billion, sales in the NY market should be
$50B x .071 = $3.55B
Determining Market PotentialBi = 0.5yi + 0.3ri + 0.2pi
where
Bi : % of total national buying power found in area iyi: % of national DI in area iri: % of nat’l retail sales in area ipi: % of nat’l population in area i
Example 2: Actual 1992 Values for NY
yi = 8.0%, ri = 6.7%, pi = 7.2%Bi = 0.5(8.0) + 0.3 (6.7) + 0.2(7.2) = 7.45Thus, 7.45% of the nation’s drug sales would be expected to occur in
NY. If the total drug sales are $50 Billion, sales in the NY market should be
$50B x .0745 = $3.725B
U.S. Population, effective buying income, and retail sails for selected states, 1991
1991 Total Population (thousands)
Percentage of U.S.
1991 Total EBI ($000)
Percentage of U.S.
1991 Total Retail Sales
Percentage of U.S.
Middle Atlantic 37,947.9 14.9621 632,218,683 16.9542 266,597,624 14.6370New Jersey 7,813.5 3.0807 155,172,906 4.1613 63,209,987 3.4704
New York 18,166.3 7.1626 298,926,889 8.0163 122,445,952 6.7227Pennsylvania 11,968.1 4.7188 178,118,888 4.7766 80,941,685 4.4439
Region State
1991 Regional State Summaries of …
Population Effective Buying Income Retail Sales
Source: Adapted from “1992 Survey of Buying Power,” Part I. Sales and Marketing Management (August 24, 1992), pp. B-2, B-3, B-4.
PART 4
Measuring Attitude: Five Approaches
Dr. Doherty
Measuring Attitude: Five Approaches
1. Self Reports– Most Common Procedure
2. Observation of Behavior3. Indirect Techniques
– Word Association– Sentence Completion– Storytelling– Graphics Interpretation
4. Performance of Objective Tasks5. Physiological Reactions
– Galvanic Skin Response Technique– Pupilometer
Qualitative Research Techniques 1. Focus Group
Skilled moderator leads a small group (6-12) of participants in an unstructured discussion of a particular topic.
A. Advantages1) Flexibility
2) Controllable
3) Group Interaction
4) Openness (encourages participants to be honest and direct)
5) Opportunity for quick execution
Qualitative Research Techniques1. Focus Group
Skilled moderator leads a small group (6-12) of participants in an unstructured discussion of a particular topic.
B. Disadvantages1) Lack of scientific validity
2) Prone to bias (moderator)
3) Offers false sense of security (Results should be considered inconclusive)
4) Measurement difficulties
5) Subject to “Squeaky Wheel Syndrome”
Qualitative Research Techniques2. Depth Interviews
Structured or Unstructured, one-on-one interview.
A. Advantages1) Offers greater comfortability for sensitive
topics
2) More detailed and revealing
3) Easier to schedule
4) Can handle more complex topics (e.g. Interviewing financial experts)
Qualitative Research Techniques2. Depth Interviews
Structured or Unstructured, one-on-one interview.
B. Disadvantages1) No interaction effects
2) Expensive
3) Inconsistency among interviewers and levels of energy (Diminishing Returns)
4) Interpretational errors produce inconsistency and unreliability
5) Lack statistical validity
Qualitative Research Techniques3. Projective Techniques
Based on the theory that people may not be aware of their innermost attitudes and/or may not wish to express certain attitudes.
Qualitative Research Techniques3. Projective Techniques
A. Techniques1) Word Association Ex. Detergents
Respondents Stimulus Words A B Washday Everyday Ironing Fresh and sweet Clean Pure Air Soiled Scrub Don’t Clean Filth This neighborhood Dirt Bubbles Bath Soap and Water Family Squabbles Children Towels Dirty Wash
Qualitative Research Techniques3. Projective Techniques
A. Techniques2) Picture Interpretation
• Thematic Apperception Test (TAT)
Respondent is shown abstract visual stimuli and describes what is going on in the pictures and what will happen
Qualitative Research Techniques3. Projective Techniques
A. Techniques3) Sentence Completion
Ex. Toothpaste• I brush my teeth because _________.• I use my brand of toothpaste because
_________.• My toothpaste tastes like _________.• When I brush my teeth, I _________.
Qualitative Research Techniques3. Projective Techniques
A. Techniques4) Third-person technique and role playing
5) Cartoons• Blank bubbles appear above the cartoon
characters• Ex. New car models
Qualitative Research Techniques3. Projective Techniques
B. Disadvantages of Projective Techniques1) Subjectivity of scoring procedures low
reliability
2) Low validity
3) Absence of substantial evidence of “Basic Assumption,” namely, that respondents project their true feelings on ambiguous stimuli
4) Small samples and unstructured formats limit generalization
Basic Measurement/Scale Concepts
Measure:Assignment of numbers to characteristics of objects
Object:A material or physical configuration. Can be seen and/or touched
Characteristics:Qualities associated with objects that give such objects identifying
traits
Measurement Scale:A plan that is used to assign numbers to characteristics of objects
Construct:The “something” that is being measured
Scales of Measurement
Scale Basic Comparisons
Typical Examples Measures of Average
Male-femaleUser-nonuserOccupationsUniform numbers
Preference for brandsSocial classHardness of mineralsGraded quality of lumber
Temperature scaleGrade point averageAttitude toward brandsAwareness of advertising
Units sold Geometric meanNumber of purchasers Harmonic meanProbability of purchaseWeight
Nominal Identity Mode
Ordinal Order Median
Interval Comparison of intervals
Mean
Ratio Comparison of absolute magnitudes
Equal-Appearing Interval Sort of the Statement into Categories
A B C D E F G H I J K Scale Q1 2 3 4 5 6 7 8 9 10 11 Value Value
f 0 8 10 30 60 60 14 12 60 0 01 p 0.00 0.04 0.05 0.15 0.30 0.30 0.07 0.06 0.03 0.00 0.00 5.4 1.7
cp 0.00 0.04 0.09 0.24 0.54 0.84 0.91 0.97 1.00 1.00 1.00f 0 0 0 0 0 6 16 28 44 66 4
2 p 0.00 0.00 0.00 0.00 0.00 0.03 0.08 0.14 0.22 0.33 0.20 9.6 1.8cp 0.00 0.00 0.00 0.00 0.00 0.03 0.11 0.25 0.47 0.80 1.00f 0 0 0 0 10 10 14 32 84 34 16
3 p 0.00 0.00 0.00 0.00 0.05 0.05 0.07 0.16 0.42 0.17 0.08 8.9 1.5cp 0.00 0.00 0.00 0.00 0.05 0.10 0.17 0.33 0.75 0.92 1.00f 0 0 8 16 36 58 48 24 10 0 0
4 p 0.00 0.00 0.04 0.08 0.18 0.29 0.24 0.12 0.05 0.00 0.00 6.2 2.0cp 0.00 0.00 0.04 0.12 0.30 0.59 0.83 0.95 1.00 1.00 1.00
Statement
Sorting Categories
Centile Formula
i
P
pcLV
w
bc
Semantic Differential Scale
1. Origin: Research designed to investigate the underlying structure of words used to describe objects, events, processes, attitude, etc.
2. Rational: Three independent (orthogonal) dimensions can be used to describe an object using a bipolar adjective scale.
Semantic Differential Scale
3. Three Uncorrelated Dimensions1) Potency
Strong - Weak Shallow - Deep Powerful - Powerless
2) Evaluation Good – Bad Sour – Sweet Informative – Uninformative Helpful – Unhelpful Useless – Useful
3) Activity: Dynamic – Static Orderly – Chaotic Aggressive – Non aggressive Dead – Alive Slow - Fast
Semantic Differential Scale
4. Marketing Application– Develop profiles for products, firms, markets
or whatever is being measured– Studies often use adjective that are not
anonyms or single words and use phrases to anchor scales
– 7-Point Scale is common
Semantic Differential Scale
5. Marketing Application– Purification Stage (often times skipped)– Item Analysis. Product Moment Formula is
used to compare score of each item with total score. Or,
– T-test of significance between mean scores of “low” and “high” total scores groups on an item-by-item basis.
Example of Semantic Differential Scale
Extremely
Somewhat
QuiteNeith
er
Somewhat
QuiteExtr
emely
Not Trustworthy ____:____:____:____:____:____:____ Trustworthy
Attractive ____:____:____:____:____:____:____ Unattractive
Not Expert ____:____:____:____:____:____:____ Expert
Knowledgeable ____:____:____:____:____:____:____Not Knowledgeable
Likert Scale
• Allows an expression of intensity of feeling
• Purification Stage (same as SD scale)– Representative Sample of Target Population
• Final Selection of Questions– Same as SD Scale
• Generally a 5-Point Scale
• Mixes Statements as to Positive or Negative Expression
Example of Likert Scale
Strongly
Disagree
AgreeStro
ngly Agree
Neither A
gree nor Disa
gree
Disagree
1.The celebrity endorser is trustworthy.
____ ____ ____ ____ ____
2.The celebrity endorser is attractive.
____ ____ ____ ____ ____
3.The celebrity endorser is an expert on the product.
____ ____ ____ ____ ____
4.The celebrity endorser is knowledgeable about the product.
____ ____ ____ ____ ____
Stapel Scale
• Adjectives or descriptive phrases are tested rather than bipolar adjective pairs.
• Generally, a 10 point scale is used. Points, on scale are identified by number.
• Results my differ according to the manner in which statement is phrased.
Example of the Stapel Scale
-5 -4 -3 -2 -1 +1 +2 +3 +4 +5
1.The celebrity endorser is trustworthy.
2.The celebrity endorser is attractive.
3.The celebrity endorser is an expert on the product.
4.The celebrity endorser is knowledgeable about the product.
Basic Rating Scales (3)
1. Itemized Rating Scale:Most commonly used. Attitudes are measured by the
choice of positions on a continuum.
2. Graphics Rating Scale:Attitudes are expressed along a line or graphic
continuum running from one extreme to the next.
3. Comparative Rating Scale:Uses an explicit reference point for comparison.
• Rank order• Pairwise comparison
Examples of the Rating Scales:Itemized Rating Scale
Please evaluate each of the following attributes of compact disc players according to how important the attribute is to you personally by placing an “X” in the appropriate box.
Not ImportantSomewhat Important
Fairly Important
Exremely Important
1.Quality of sound reproduction
2.Physical size of CD Unit
3. Brand name 4. Durability of unit
Examples of the Rating Scales:Graphic Rating Scale
AttributeNot Important
Extremely Important
1. Quality of sound reproduction
2. Physical size of CD Unit
3. Brand name
4. Durability of unit
Please evaluate each of the following attributes of compact disc players according to how important the attribute is to you personally by placing an “X” at the position on the horizontal line that most accurately reflects your feelings.
Examples of the Rating Scales:Comparative Rating Scale
Please divide 100 points between the following attributes of compact disc players according to the relative importance of each attribute to you.
Quality of sound reproduction
Physical size of CD Unit
Brand name
Durability of unit
______
______
______
______
100%
Q-Sort Technique
• Similar to Thurstone approach. Respondents place questions into different piles to form a known probability distribution, e.g., normal or log normal
• Subjects reflect their attitude toward an object
• Focus is on individuals and not the object(s)• Used for cluster and segmentation
applications
Consumer Decision Making ModelsAttribute Analysis of Valence and Salience Properties
Product AttributesComputer Memory, Software, Price, etc.Hotel ?Mouthwash ?Lipstick ?
BrandMemory Capacity
Graphics Capacity
Software Diversity Price
A 10 8 6 4B 8 9 8 3C 6 8 10 5D 4 3 7 8
1. Product Examples
2. Illustration: PC
3. Decision ModelsA. Ideal Brand Model
B. Constrained Brand Model
C. Conjunctive ModelMinimum attribute levels screen out competition brands to yield reduced set. Ex. PC brands equals or exceeds (7,6,7,2)
N
iijkikjk PWA
1
N
iikijkikjk CPWD
1
Constrained Brand ModelEx.: (6,10,10,5)
3.8 |5-7|.1 |10-7|.2 |10-3|.3 |6-4|.4D(d)
0.6 |5-5|.1 |10-10|.2 |10-8|.3 |6-6|.4D(c)
1.7 |5-3|.1 |10-8|.2 |10-9|.3 |6-8|.4D(b)
3.1 |5-4|.1 |10-6|.2 |10-8|.3 |6-10|.4D(a)
PART 5
Questionnaire: Anatomy
Dr. Doherty
Tobin College of Business
St. John’s University
Questionnaire: Anatomy
Definition: A formalized schedule (document) that is designed to achieve three purposes:
1. Obtain Relevant Information;
2. Direct the Questioning Process; and
3. Set the format for recording and evaluating data.
Eight Step Process
Step 1: Define Marketing Problem1) Write a paragraph
2) List data to be collected
3) Anticipate use of data
4) State objectives
5) Develop a Plan of Analysis
6) Client “Sign Off”
Eight Step Process
Step 2: Interviewing Process1) Personal
– Structured vs. Unstructured– Interviewer Administered vs. Self Administered
2) Telephone
3) Mail
4) Internet
Eight Step Process
Step 3: Evaluate Question Content
Four Rules:1) Will the Respondent understand the
question?
2) Will the Respondent have the information?
3) Will the Respondent provide information?
4) Will the Analyst understand the Respondent’s response?
Eight Step Process
Step 4: Q/A Format1) Open Ended
a. Free Responseb. Probingc. Projective (e.g. association, construction, sentence
completion)
2) Close Endeda. Dichotomousb. Multichotomousc. Scalesd. Rankinge. Check List
Eight Step Process
Step 5: Determine Wording of Question
Three Rules:1) Unambiguous
2) Simple and Familiar Words
3) Specific Words or Options
Ex.) Why did you fly to Chicago on U.S. Airlines?
Eight Step Process
Step 6: Sequence of Questions1) Screening (if necessary)
2) Gain Confidence and Interest
3) Groups Like Topics Together
4) Funneling
5) Demographics at End
6) Thank You!
Eight Step Process
Step 7: Physical Characteristics of Questionnaire (especially by mail)
Step 8: Pretest - Revise - Formalize - Finalize1) Personal
2) Planned Method of Administration
Guidelines for Question Wording
• Use simple words and questions
• Avoid ambiguous words and questions
• Avoid leading questions
• Avoid implicit alternatives
• Avoid implicit assumptions
• Avoid generalizations and estimates
• Avoid double-barreled questions
Communication MethodsF
orm
: •Standardized questions•Standardized responsese.g. fixed alternative questions
•Non standardized questions•Nonstandardized responses. e.g. depth interviews
•Simple Administration•Simple Analysis•Suitable for facts or clear-cut opinions due to forced alternatives
•Flexible•Difficult interpretation•Interviewer influenced•Better for exploratory research
•Standardized questions•Standardized responses
•Standardized stimuli•Non standard responsese.g. projective techniques
Simple administration•Simple analysis•Difficult interpretation•Least used method
•Difficult analysis•Subjective interpretation•Suited to exploratory research
UN
DIS
GU
ISE
DC
hara
cter
istic
s:F
orm
:C
hara
cter
istic
s:
DIS
GU
ISE
D
STRUCTURED UNSTRUCTURED
Comparison of mail, telephone, and personal interview surveys
BASIS OF COMPARISON MAIL SURVEYS TELEPHONE
SURVEYS
PERSONAL INTERVIEW SURVEYS
Cost per completed survey
Usually the least expensive, assuming adequate return rate
Moderately expensive, assuming reasonable completion rate
Most expensive because of interviewer’s time and travel expenses
Ability to probe and ask complex questions
Little, since self-administered format must be short and simple
Some, since interviewer can probe and elaborate on questions
Much, since interviewers can show visuals, probe, establish rapport
Opportunity for interviewer to bias results
None, since form is completed without interviewer
Some, because of voice inflection of interviewer
Significant, because of voice and facial expressions of interviewer
Anonymity given respondent
Complete, since no signature is needed
Some, because of telephone contact
Little, because of face-to-face contact
Comparison of Three Communications Media on Ten Factors
FACTOR MAIL PERSONAL TELEPHONEBias freedom (from interviewer) 1 3 2Control over collection 3 2 1Depth of questioning 3 1 2Economy 2 3 1Follow-up ability 3 2 1Hard-to-recall data obtainable 1 2 3Rapport with respondent 3 1 2Sampling completeness 3 1 2Speed of obtaining reponses 3 2 1Versatility to use variety of methods
2 1 3
MEDIUM
© 1987 by Prentice-Hall, Inc.A division of Simon & SchusterEnglewood Cliffs, NJ 07632
PART 6
STATISTICAL ANALYSIS
Sampling PlansNon Probability Probability
•Convenience•Judgment•Snowball•Quota
•Simple Random Sampling•Systematic Random Sampling•Stratified Random Sampling
–Proportionate–Disproportionate
•Cluster Random Sampling–One Stage–Two Stage
•Area–One Stage–Two Stage
From A, B, and C
Rewriting (1)
Also, from (1)
Solving for Sample Size
x
xZ
x
xt
ˆ
xZx xZx ̂
n
EZ xE
2
22
E
Zn
or
or
where Examples:
Let =100, Z=2, and E=10
n=(22 x 1002) 102 = 400
Let =100, Z=2, and E=5
n=(22 x 1002) 52 = 1600
(1)
(2)
(3)
Major Principles
A) )(xE
B)n
x
C) CLT
Determinants of Sample Size (3)
• Variance of Population
• Error Allowance
• Probability of Realizing Error Allowance
22
E
Zn
From A, B, and C: Binomial
Similar to (3), for Binomial
2
2 )1(
E
PPZn
A) )(PE
B)N
PPP
)1(
note: )1( PPx 22 )1( PPx
xP
n
PPZP
)1(
(4)
(5)
Example:
Let P =0.2, Z=2, and E=0.02
160002.0
)8.0(2.022
2
n
Suppose that P =0.3 from (5)
0229.3.
)0115(.23.1600
)7.0(3.023.
From A, B, and C: Binomial
Similar to (3), for Binomial
2
2 )1(
E
PPZn
A) )(PE
B)N
PPP
)1(
note: )1( PPx 22 )1( PPx
xP
n
PPZP
)1(
(4)
(5)
Example:
Let P =0.2, Z=2, and E=0.02
160002.0
)8.0(2.022
2
n
Suppose that Pfound =0.3 from (5)
0229.3.
)0115(.23.1600
)7.0(3.023.
Stratified Sampling1) Proportionate
22
2
iiWE
Zn
Where:
Allocation:
Note:
N
NW i
i
nN
Nn i
i
22
222
11
2
2
... kk
N
N
N
N
N
N
E
Zn
n
W iix
2
Stratified Sampling2) Disproportionate
22
2
iiWE
Zn
Allocation:
Note:
nN
Nn k
iii
iii
1
2
22
11
2
2
...
k
k
N
N
N
N
N
N
E
Zn
n
W iix
2
Stratified Sampling IllustrationN=1250E=8.0090% Confidence Level: Z=1.64
Industry
750 20 0.60 15000 0.50500/ 1250
30 0.4015,000/ 30,000
0.50
iN i iW iiN ii
ii
N
N
1N
2N
1) Proportionate
)360240(042025.
)900(4.)400(6.64
6896.2
)30(4.)20(6.8
)64.1( 222
2
n
10)25(4.
15)25(6.
215.25
2
1
n
n
n
Stratified Sampling IllustrationN=1250E=8.0090% Confidence Level: Z=1.64
Industry
750 20 0.60 15000 0.50500/ 1250
30 0.4015,000/ 30,000
0.50
iN i iW iiN ii
ii
N
N
1N
2N
2) Disproportionate
)576(042025.
121264
6896.2
)30(4.)20(6.8
)64.1(
2
2
2
2
n
121224
12)24()30(500)20(750
)20(750
24
12
1
nnn
n
n
PART 7
STATISTICAL DISTRIBUTIONS
Sales Performance of REPS under Three Different Sales Training Programs
80
758580
375425400Total
818984
718270
738881
687679
829086
IIIIII
x
x
SUMMARY (Anova: Single Factor)Groups Count Sum Average VarianceColumn 1 5 400 80 38.5Column 2 5 425 85 35Column 3 5 375 75 38.5
ANOVASource of Variation SS df MS F P-value F crit
Between Groups 250 2 125 3.348214 0.0699094 3.88529Within Groups 448 12 37.3333
Total 698 14
3.348 3.885
Accept H0 Reject H0
348214.33333.37
125
)315/(448
)13/(250
)/(
)1/(
KNSSE
KSSBF
698
)8081(...)8079()8086(
)(
222
2
N
ij xx
Step I: SST
Step II: SSB
250
])8075()8085()8080[(5
)(
222
1
2
K
jjj xxn
448154140154
154)7581()7571()7573()7568()7582(
140)8589()8582()8588()8576()8590(
154)8084()8070()8081()8079()8086(
)(
22222
22222
22222
2
N
jij xx
Step III: SSE
Note: SST = SSB + SSE
698 = 250 + 448
Step IV: Fcalc. Value
Accept H0. No significant difference among samples at 5% level
88.3 of Value Table
35.33.37
125
124482
2501
05122
j .i F
N-KSSEK-SSB
Chi-Square1. Definition
2. Applications
ler
ji ij
ijij
E
EO,
1,1
22 )(
rrkr2r1
22k2221
11k1211
P...PP
P...PP
P...PP
:H 0
rr
22
11
E0
E0
E0
:H 0
A. Contingency Table (r by le) B. Goodness of Fit Test
Chi-Square3. IllustrationProblem: Children's Commercials:Does the level of Understanding (Levels I, II, and III) vary
with a child's age (5-7 vs 8-10 vs. 11-12)
33k3231
2232221
1131211
PPP
PPP
PPP
:H 0
Sample Test: Level of AGEUnderstanding 5-7 8-10 11-12 Total
I 55 37 15 107II 35 50 60 145III 10 13 25 48
TOTAL 100 100 100 300
Chi-Square4. Solution
160300/48PPP
483300/145PPP
357300/107PPP
3k3231
232221
131211
:0H
Level of AGEUnderstanding 5-7 8-10 11-12 Total
I 55 37 15 107(35.7) (35.7) (35.7)
II 35 50 60 145(48.3) (48.3) (48.3)
III 10 13 25 48(16) (16) (16)
TOTAL 100 100 100 300
1,000
Chi-Square
4. Solution (continued)
488.9
9.3616
)1625(...
7.35
)7.3537(
7.35
)7.3555(
24,05.
2
222
calc
022 reject , Since Hcriticalcalc
Dependent Samples: t-Test
nd
d
ˆ
H0: Consumers are Indifferent Between Alternatives, that is, 0D
d
Ddt
Test Statistic:
n
dd
1
)(ˆ
2
n
ddd
where:
n = number of sample (retail outlets)
Store TUMS I TUMS II d
1 130 111 19 1692 82 76 6 03 64 58 6 04 111 103 8 45 50 48 2 166 56 61 -5 121
TOTAL 493 457 36 310
2)( dd
Dependent Samples: t-Test
Illustration:
Dependent Samples: t-TestIllustration (continued)
66
36
n
dd
87.75
310
1
)(ˆ
2
n
ddd
21.345.2
87.7ˆ
nd
d
0:0 DH
015.205.: 5.0,1 nt
87.121.3
6ˆ
d
Ddt
0Htt criticalcalc reject cannot , Since
END