an empirically derived taxonomy of information technology … · 2017. 9. 30. · frames the...

An Empirically Derived Taxonomy of Information Technology Structure and Its Relationshipto Organizational StructureAuthor(s): Kirk Dean Fiedler, Varun Grover and James T. C. TengSource: Journal of Management Information Systems, Vol. 13, No. 1 (Summer, 1996), pp. 9-34Published by: M.E. Sharpe, Inc.Stable URL: http://www.jstor.org/stable/40398201 .

Accessed: 15/08/2013 10:01

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

M.E. Sharpe, Inc. is collaborating with JSTOR to digitize, preserve and extend access to Journal ofManagement Information Systems.

http://www.jstor.org

This content downloaded from 130.127.15.5 on Thu, 15 Aug 2013 10:01:39 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=mes

http://www.jstor.org/stable/40398201?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


An Empirically Derived Taxonomy of Information Technology Structure and Its Relationship to Organizational Structure

KIRK DEAN FIEDLER, VARUN GROVER, AND JAMES T.C. TENG

Kirk Dean Fiedler is an Assistant Professor of MIS at the University of South Carolina, School of Business Administration. He received a B. A. from Wittenberg University and an M.B.A. and an M.S. in information systems and systems science from the University of Louisville, before completing his Ph.D. in MIS at the University of Pittsburgh. His work experience includes several years at Arthur Young & Com- pany, and he has earned CPA certification. Currently, his research interests involve the investigation of technology assimilation and business process redesign. He has published this research in various journals, including MIS Quarterly, IEEE Transac- tions in Engineering Management, California Management Review, Long Range Planning, Journal of Information Technology, Omega, and European Journal of Information Systems. Dr. Fiedler was a finalist in the Decision Sciences Institute's Instructional Innovation Award Competition. He is a member of the Academy of Management, AICPA, DSI, and AIS.

Varun Grover is currently an Associate Professor of MIS in the Management Science Department at the University of South Carolina. He holds a B.Tech, in electrical engineering from the Indian Institute of Technology, an M.B.A. from Southern Illinois University, and a Ph.D. in MIS from the University of Pittsburgh. Dr. Grover has over sixty refereed papers published or forthcoming on organizational impacts of IT, reengineering, strategic information systems, and telecommunications, in journals such as Journal of Management Information Systems, MIS Quarterly, Decision Sciences, IEEE Transactions on Engineering Management, Data Base, Information and Management, Journal of Systems Management, California Manage- ment Review, Communications of the ACM, Long Range Planning, Journal of Infor- mation Systems, and Omega. Dr. Grover is the recipient of an Outstanding Achievement Award from the Decision Sciences Institute. He is also the recipient of the 1992 Alfred G. Smith Award for excellence in teaching. He is currently on the Editorial Boards of four journals and an active referee for twelve more. He recently coedited a book called Business Process Change: Reengineering Concepts, Methods and Technologies. He is a member of DSI, AIS, and TIMS.

James T.C. Teng is Associate Professor at the College of Business Administration, University of South Carolina. He earned his M.S. in mathematics from the University of Illinois at Urbana-Champagne, and his Ph.D. in MIS from the University of

Acknowledgment. This research was supported by the Center of International Business Educa- tion and Research at the University of South Carolina through a grant from the US Department of Education.

Journal of Management Information Systems I Summer 1996, Vol. 13, No. 1, pp. 9-34

Copyright© 1996 M.E. Sharpe, Inc.



10 FIEDLER, GRÒ VER, AND TENG

Minnesota. Dr. Teng's research and consulting interests are in the areas of information management, decision support systems, and management of process and organizational change. He has published over forty articles in journals such as Journal of Management Information Systems, Decision Sciences, Information and Management, California Management Review, INFOR, Data Base, European Journal of Informa- tion Systems, Information Systems Journal, and IEEE Transactions in Engineering Management. In 1 992 he won the Outstanding Achievement Award from the Decision Sciences Institute.

Abstract: This study empirically develops a taxonomy that has implications for matching information technology (IT) and organizational structures. The taxonomy of IT structure is based on the degree of centralization of computer processing, capability to support communications, and the ability to share resources. By using a multistep cluster analysis, both the membership and number of groups are derived from the responses of 313 firms. Four IT structures are identified: centralized (centralized processing, low communication, low sharing), decentralized (decentralized processing, low communication, low sharing), centralized cooperative (centralized processing, high communication, high sharing), and distributed cooperative computing (decentralized processing, high communication, high sharing). Centralized computing is related to functional organizational forms with low integration and centralized decision making. Decentralized computing is related to product organizational forms with decentralized decision making. Centralized cooperative computing is related to functional organizational forms with high integration. Distributed cooperative computing is related to both matrix and product organizational forms with high integration. The ability to identify and understand the implications of IT structure is of critical importance to both academic and management practitioners.

Key words and phrases: information technology structure, organizational decision- making structure, organizational integration, organizational structure, taxonomy.

Thematchingof IT structure,orthe distribution of electronic communication, processing, and storage capabilities, with the needs of the firm is one of the most critical decisions of a corporation [7, 53, 58, 62]. For example, in response to

competition, many corporations are altering their organizational structure by downsiz-

ing their middle management. These leaner structures can put new demands on

managers, which place heightened importance on obtaining an appropriate IT structure to support the flatter management structure. Even the most fundamental business

operations, as they are supported by transaction processing (e.g., sales, payroll), require a suitable IT structure. In spite of the increasing dependence of organizations on IT structure, few classification schemes exist that recognize IT archetypes. A classification scheme would allow an organization to identify its current IT structure and its alternative structural choices. This would be an essential step in the process of matching IT structure with organizational structures [2, 32].

The task of matching IT structure with organizational structures has been complicated in the past decade by the widespread acceptance of new information technologies that have the potential to enable the creation of fundamentally new IT structures. Less expensive and more effective storage and processing technologies have com-



EMPIRICALLY DERIVED TAXONOMY OF IT STRUCTURE 1 1

bined with developments in telecommunications to create new opportunities for

technology deployment. Simultaneous advances in transmission technologies (e.g., data compression, encryption, ISDN) and the increased availability of electronic networks (e.g., Internet, commercial carriers, local area networks) have created

exponential growth in electronically supported communication [54]. Developments in processing technology have created desktop workstations that have analytical capabilities superior to many mainframe computers of the previous decade, while their

purchase and maintenance are a very small fraction of the cost of the latter. Advance- ments in processing have enabled the development of what sometimes appears to be an endless assortment of increasingly user friendly and powerful software applications. New methods for data storage are being combined with communication and

processing capabilities to allow users to seamlessly share data and application programs (e.g., database management systems, image technology) [9, 26, 27].

While computing can still be defined in terms of its functions of communication,

processing, and storage, the capability to provide these resources is expanding so

quickly that many organizations are finding it difficult to capitalize on the new

opportunities [7, 25, 27]. Corporate planners are no longer limited to choices between traditional centralized and decentralized computing configurations. Increased resource sharing and communication can produce decentralized networks that appear to

be a single centralized computer because the location of the resources is transparent to the user [31]. It has even been suggested that that these new technological

capabilities may result in new, more collaborative organizational structures [27]. The

importance of this revolution has not gone unnoticed by senior information systems executives, who have recognized these new opportunities and identified the planning and development of corporate information technology structure (IT architecture) as

the most critical issue of the decade [43]. The research presented in this paper represents an attempt to begin to facilitate the

task of categorizing and matching IT structure to the organization by: (1) identifying the characteristics and number of alternative IT structures and (2) describing and

testing anticipated relationships between IT structure and organizational structure. This study, through the responses of 313 North American senior IS executives,

develops a function-based taxonomy of IT structure and explore its relationship to

organizational structure. The first section reviews classification techniques. Next, the

evolution and use of IT structural typologies are reviewed. This review then forms the

basis for the methodology used to derive the IT structure of the surveyed corporations. A discussion of the anticipated relationship between IT and organizational structure then

frames the presentation of a series of research propositions. This is followed by a discussion

of the research methodology, measurement, and taxonomy validation. The paper con-

cludes with a review of the results and a discussion of the study's implications.

Classification Techniques

The technique of classifying related subjects into similar groups for study has been an important aspect of scientific research since Aristotelian applications over



12 FIEDLER, GROVER, AND TENG

2,000 years ago. Groups allow researchers to identify underlying patterns and help them to extrapolate understanding to wider populations. Typologies and taxonomies

represent two fundamentally different approaches to classification. Typologies clas-

sify subjects by forcing deductive assignment into a priori predefined groups, while taxonomies determine membership into a posteriori categories that emerge from

empirical analysis inductively [52, 63]. Even though the typological use of predefined categories has been applied in important research, such as that conducted by Miles and Snow [39] and Porter [50], it has also been subjected to criticism. Because

typological categories are preordained, some subjects may not fit into the available

groups, so the technique may be neither exhaustive nor exclusive. Since typologies are the product of researcher intuition, classification is also exposed to the potential for increased researcher bias or misconception [4, 5, 14, 52].

The groups of a taxonomy are derived (through a multi variate method such as cluster

analysis) from the characteristics of the measured subjects, so the categories are both exhaustive and mutually exclusive. This method is especially useful when one is

examining unexplored phenomena because both the nature and the number of categories can be determined by the population. In spite of the differences between taxonomies and typologies, both methods must be empirically examined to evaluate the

representativeness and generalizability of the classifications to the population they are meant to describe. Unlike the predetermined, idealized categories of a typological methodology that lend themselves to prescriptive hypothesizing, a taxonomy's classifications emerge from analysis so that the characteristics of the grouping can be examined for reasonableness only after the nature of the classification is derived [14, 20].

Developing a taxonomy can be viewed as a multistep process. First, the classification scheme must be developed. The classification criteria should be defined in terms of

clearly demonstrable features based on theory or experience [4, 5, 52]. The inductive nature of taxonomies may begin with experience and result in emergent theory. For

example, the first classification schemes of biology were based on the physical characteristics, or the morphology, of nature. This primarily atheoretical beginning established the basis for the later development of the theory of evolution. The next

step requires the measurement, consistent application, and multivariate analysis of the classification criteria to produce the item groupings. The final step consists of the evaluation of the classification groupings. The groups should be exhaustive, mutually exclusive, and stable. Rich [52] also noted that the classification system should "mirror the real world . . . describe organizational reality in a way that is recognizable to and consistent with the vision of practitioners and researchers alike as a viable reproduction of the diverse world in which we live and study" (p. 777). In the next section, the role of

technological development in the evolution of IT structure is reviewed to form the basis for the identification of appropriate criteria for taxonomy development.

The Evolution of Information Technology Structure

While computing technology has changed considerably since the first vac- uum-tube-based machines, the configuration of the technology has been relatively



EMPIRICALLY DERIVED TAXONOMY OF IT STRUCTURE 13

stable (see Table 1 ). As technology has advanced, the number of possible IT structures has increased gradually. The advances in technology performance and IT structure are

represented in figure 1 [21, 47, 55, 56, 57]. The first computers were isolated

processors that were accessed either directly or through the use of dumb terminals. This centralized computing was the mainstay information technology structure for over thirty years. Most, if not all, of the important sanctioned business operations were carried out using a centralized processor, database, and repository of application programs. However, as information technology became less expensive and more

powerful, end users gained control of their computer applications. Many firms found that there was a processing migration forming isolated islands of decentralized

computing throughout their organizations. Even though there were attempts to connect these islands, these early networks were formed around a central processor that maintained control and was the hallmark of hub-and-spoke computing. Enabled by improvements in the cost and performance of information technology, computer networks were beginning to be developed that would allow direct interaction between the islands without the aid of a central processor, forming a distributed computing environment.

In the 1 990s an increasing number of companies chose to network their computers, creating hub-and-spoke and distributed computing environments that had a greater emphasis on resource sharing [21, 47, 55, 56, 57]. The objective of many of these

network designers was to create an information system that would allow seamless communication and resource sharing throughout the organization. The advancement

in computer-based telecommunication, processing, and data storage, such as ISDN, RISC-based processing, and database management systems, has enabled the creation

of a system oí cooperative computing, or a client-server computing structure, with

shared access to dispersed data and applications [3 1 , 47, 48]. A cooperative computing

system could allow a user to perform a task that would require remotely stored

resources with the same ease as a local operation. In effect, the network of information

technologies performs as if it were a single localized system - the network becomes

the computer. In an environment of widespread networking, it is important to under-

stand the impact of this increased emphasis on resource sharing on IT structures, as

opposed to the traditional use of processing distribution and networking as the sole

determinants of IT structure.

IT Structural Typologies

In the late 1980s it was generally recognized that several types of information technology structures had evolved and were commonly found in organizations. Researchers, such as Ahituv, Neumann, and Zviran [2], and Leifer [32],

proposed typologies for describing these structures. These typologies were based on

their perception of the dominant distinguishing features of information technologies in the late 1 980s. Both classification schemes, as summarized in Table 1 , are defined in terms of the degree of process centralization and networking capabilities. Using these dimensions, Leifer [32] created a typology that divided IT structures into




Table 1 . Information Technology Structure Typologies

Shared

Processing Network data and

Category decentralization connectivity applications Centralized computing Low Low N/A

[2, 10, 21, 29, 31, 32, 38, 47, 65]

Decentralized computing High Low N/A

[2, 21, 29, 32,* 47, 65]

Hub-and-spoke computing [31,* 32] Low High N/A

Distributed computing [2, 10, 21, 29, 32,* 38, 47, 65] High High N/A

Cooperative computing [21, 29, 31,* 47, 65] High High High

* Category is recognized even though the terminology may differ.

Figure 1. Evolution of IT Structure - Computer Cost and Computing Power

centralized systems, designed around an isolated central processor and dumb termi-

nals, stand-alone systems of dispersed isolated computers, decentralized systems of networked peer computers, and hub-and-spoke computing systems, designed around a centralized processor with networked connections. Ahituv, Neumann, and Zviran's

[2] typology divided IT structure into three groups: ( 1 ) centralized systems, which were based on a central isolated processor, (2) distributed systems, which had a number

Computing Cost Computing Power

y/ Cooperative

l^^stributedConputirA

|^^-^í^andSpokeConr|)Uting '

^^^ Decentrafized Computing

1955 Time 1995




Centralized Computing Decentralized Computing Distributed Computing

Figure 2. Information Technology Structures - Traditional Three-Group Typology

of networked processors, and (3) decentralized systems, which had a number of isolated processors. It should be noted that the Leifer [32] and Ahituv, Neumann, and Zviran [2] typologies used the same terms (decentralization and distributed) but defined them differently. Because of the varied acceptance of the terminology, and for the purpose of clarification, this discussion will focus on the more widely accepted Ahituv, Neumann, and Zviran [2] classification nomenclature as noted in figure 2.

Despite the dearth of academic typologies, the critical importance of the subject has been widely recognized by practitioners and educators. An introduction to management information systems would not be complete without discussing IT structure and its relationship to organizational structure [ 1 0, 2 1 , 29, 38, 47, 65]. Almost every author uses the distinction of centralization of computer processing and networking to present the topic of computing structure. Recently, however, several authors, including Ahituv, Neumann, and Riley [ 1 ] and Lee and Leifer [3 1 ], have recognized the ability to share applications and resources as a third dimension of IT structure [2 1 , 29, 47, 65]. The ability to share resources and data has been seen as the hallmark of client-server or cooperative computing [48]. Cooperative computing has been defined by some authors as a special case of distributed computing (decentralized processing and networking) with the added ability to share data and application resources [ 1 , 47]. The prior reviews of typological and historical growth of IT structures are used in the next section to help identify criteria for taxonomy development.

Dimensions of a Taxonomy

Leifer [32] and Ahituv, Neumann, and Zviran's [2] typologies focused on the dimensions of networking and processing centralization, and while these dimensions are still relevant in distinguishing the nature of IT structure, the new emphasis on resource sharing may require special attention. It is critical to understand the relationship between cooperative computing and the networked environments of distributed and hub-and-spoke computing. It is also important to understand the impact of the widespread availability of new technologies (e.g., high-speed processers, LAN, and database management systems) on un-networked centralized and decentralized com-




puting environments [27, 31]. To gain a better understanding of these relationships, an empirically based taxonomic approach can be used to build on the typologies and historical observations that are currently accepted concerning IT structure. A numer-

ically driven, inductive approach would allow the determination of both the nature and the number of types of structures.

A critical first step in empirically deriving a taxonomy is recognizing the salient dimensions of IT structure. The two elementary components are information technol-

ogy function and structure. Information technology has been historically defined in terms of its functions of processing, communication, and storage [7, 62]. Structure has been determined, traditionally, by the degree of centralization and the pervasiveness of networking [2, 32]. The task is then to develop a framework in which to meld these two accepted perspectives, while capturing the resource-sharing dimension of cooperative computing. The recognition of the extent of centralization (structure) of

processing (IT function) would be an example of capturing both IT function and

structure in a single dimension and should be included as an aspect of an IT structural

taxonomy. The structural component of networking or connectivity is related to the functional

aspects of both communication and storage. A dimension that would capture the

degree in which networked (structure) computers could communicate (IT function) with each other would address both the network structure and communication capability of a system. The storage function of computers is the archiving of data and

application program resources. The capability of networked (structure) computers to

share stored data and application programs would be a dimension that would address

the storage function and networking structure that form the basis of cooperative computing.

These attributes can be stated in the following dimensions and will be used to

empirically derive an IT structural taxonomy:

1 . The extent that computer processing is centralized; 2. The degree that computers support communication; 3. The ability of computers to share data and application programs.

The importance of this examination is further enhanced by the availability of new

capabilities in each of these dimensions brought about by recent technological developments. New high-speed, low-cost processors affect the organizational options for computer centralization. Widespread electronic mail, groupware, graphical user

interfaces, and the Internet have a potential impact on computer-supported communication. The development of database management systems, object-oriented databases, and imaging technology could affect the ability of computers to share data and

application programs. These three dimensions are used to classify the IT structures and form the basis for developing a profile of the responding organizations.

The relationship and need to match IT and organizational structures have been a critical concern for both researcher and practitioner for over three decades [ 1 7, 30, 3 1 , 33]. Because the nature of the clusters will be determined by the data, it is not possible to propose hypotheses for the classifications a priori when deriving a taxonomy.




However, it can be anticipated that the emergent groups would be related to organizational structure. Based on prior research and theory, anticipated relationships between the dimensions of the taxonomy and organizational structure can be proposed. The resulting inductive taxonomy can also be contrasted with previous deductive IT structural typologies. To facilitate these comparisons, the organizational characteristics of, degree of organizational integration, formal organizational structure and centralization of decision making are considered. The next section reviews anticipated relationships between the dimensions of the taxonomy and organizational structural characteristics and suggests propositions considering the combined dimensions of the

taxonomy.

IT and the Organization

Centralization of decision making is a fundamental feature of organizational structure. Centralization of major decision making is the extent to which decisions (e.g., capital budgeting, pricing, personnel) are made at the top levels of the

organization. Ahituv, Neumann, and Zviran [2] examined the relationship between

their typology and their categorical measurement of corporate decision making and

concluded that centralization of processing is directly related to centralization of

decision making. Organizations with the most centralized decision-making structure

had a centralized IT structure and organizations with a decentralized IT structure had

the most decentralized decision-making structure. Their typology has only one net-

worked system, the distributed IT structure, which is located in the middle of the two

extremes of decision-making styles. It is possible that IT structures that increase communication and resource sharing

may also affect the structure of the organization's decision making and change the

nature of organizational work. The potentially moderating impact of IT-supported communication on the organizational structure of corporate decision making was

anticipated by Huber [22], who predicted, from prior research, that computer-sup-

ported communication can cause centralized organizations to distribute power and

become less centralized. In decentralized organizations, the same information tech-

nology can provide executives with sufficient information to bring about movement

to centralize organizational decision making. The effect of lower information cost through the availability of computer-supported

communication and data and application sharing may also be anticipated through the

application of agency theory [19]. In an agency model of the firm, principals hire

agents to delegate tasks. The relationship is complicated by information asymmetry, in which the agent may have superior knowledge about the nature of the task and the

degree of its successful completion [6, 64]. As information costs associated with

monitoring compliance of agents are lowered, the ability of centralized management to decentralize decision-making tasks while maintaining control would be increased.

The availability of information concerning the nature of the task would relieve the

requirement of decentralized management to hire the agent. This suggests that only those organizations that have been denied the influences of IT-supported communi-




cation and data and application sharing will maintain organizational structures that are characterized by extreme decentralized or centralized decision making.

A multidimensional taxonomy would suggest that there is an interaction effect between the individual dimensions. The derived IT structure taxonomy could be related to Huber's [22] assertion that increases in computer-assisted communication could cause decentralized organizational structures to become more centralized and centralized organizational structures to become less centralized. Based on Huber's work [22], agency theory, and the findings of Ahituv, Neumann and Zviran [2], IT structural dimensions can be related to decision-making centralization by the follow-

ing proposition:

Proposition 1. Organizations with the most extreme decision-making structures will have IT structures that have (a) reduced capabilities for communication,

application, and data sharing, and (b) corresponding extreme centralized or decentralized computer processing configurations.

Researchers have also suggested that IT has the potential to alter the nature of

organizational work by increasing or decreasing organizational integration [35]. Organizational integration is the degree to which the firm has interdepartmental cooperation. Interdepartmental cooperation would include the lateral sharing of projects, applications, ideas, and information. It is assumed that the sharing of tasks is

accomplished by the sharing of horizontally dispersed computer data and application resources. As computer resource sharing occurs, avenues of interdepartmental cooperation are created that facilitate firm integration. For example, a department's individually maintained computer data and application portfolios would reflect their isolated vision and understanding. As cooperative computing facilitates the sharing of resources, it would be anticipated that a department's isolated perspectives and

computer portfolios would be affected by the needs and understanding of other

departments. The sharing of department computer resources and perspectives would facilitate organizational integration. Other researchers have concentrated on the

positive effect of communication on integration [12, 15, 16, 23, 28, 61]. The positive relationship between IT and organization integration can also be related

to coordination theory. Coordination theory deals with how information, goals, and

operations related to organizational tasks can be shared. In general, increased organizational integration requires increased coordination [34, 35]. Coordination is associated with both the increased efficiency of resource utilization and increased coordination costs. When coordination is costly, organizations minimize integration [36]. However, IT support can lower coordination costs and increase the feasibility of more integrated organizational structures [8]. It may be anticipated that there would be a reinforcing effect between increased opportunities for electronic communication and resource sharing that can be expressed as the following research proposition:

Proposition 2. Organizations with higher levels of interdepartmental integration will have IT structures that have a greater capacity for resource sharing and communication.




The belief that IT and organizational structure are related is an underlying principle behind a great deal of IS research [3, 33, 49]. Both Ahituv, Neumann, and Zviran [2], and Leifer's [32] typologies were originally motivated by the importance of matching IT and organizational structure. However, Ahituv, Neumann, and Zviran [2] had insufficient sample characteristics to evaluate the relationship between their typology and overall organizational form. The value and credibility of a taxonomy of IT structure would be greatly enhanced if it could be used to match organization form and IT structure.

A stream of research over the past twenty-five years has supported the importance of a contingency theory of information processing to match information-handling capabilities with organizational uncertainty [15, 28, 36]. It has been observed by March and Simon [36] and later by Galbraith [15, 16] that organization structures are

related to the task uncertainty of their operations. Tushman and Nadler [6 1 ] noted that

effective organizations can match their information-processing capabilities to fit the

uncertainty associated with their environments. Daft and Lengel [ 1 2] further suggested that organizations needed to consider both the availability and the "richness" of the

information provided by the organizational information system. The current dynamic and competitive business environment combined with the availability of new IT

technologies would suggest that there is an even greater need to match organizational and IT structures to assure the availability of appropriate levels of IT supported information processing and coordination [27, 34, 35]. Traditionally, organizations have been categorized into matrix, product, and functional organizational structures

[15, 16, 21]. In light of coordination theory [34] and a contingency theory of

information-processing and organizational structure [12, 16, 28, 61], it is possible to

anticipate the needed information capabilities for these organizational designs. The

anticipated relationship between organizational forms and IT structure can be ex-

pressed in the following general proposition.

Proposition 3. Organizational structures, as represented by matrix, product, and

functional forms, will be related to different IT structural types.

The functional organizational form is focused on dividing firms into their basic

corporate functions (e.g., accounting, finance, marketing). It is classically character-

ized by hierarchy of authority, unity of command, functional specialization (which would require relatively less additional information), and coordination [13, 15, 16]. Hierarchical structures function best in environments of limited task uncertainty, which is dealt with through rules and control. A successful functional hierarchy would

require comparatively less computer-supported communication and resource sharing. The centralization of control and unity of command would be associated with a central

processing environment [36]. This relationship can be summarized by the following

proposition.

Proposition 3. 1. Organizations with a functional organizational form will have

IT structures that have more centralized processing and a reduced capacity for resource sharing and communication.




The product organizational form is focused on output or the products of the firm. The organization is divided into product lines, which are then organized internally by functions (e.g., Medical Division with Medical Accounting, Medical Finance; Sol- vents Division, Solvents Accounting, etc.) [21]. The segregated focus of the product lines would be associated with dispersed processing facilities. The single product focus would be associated with reduced task uncertainty and need for information

processing [15,16]. For example, solvents accounting would not have to be concerned with evaluating Federal Drug Administration approval and medical accounting might not have to deal with the uncertainty of evaluating liabilities associated with toxic waste. These IT and organizational characteristics can be expressed in the following proposition.

Proposition 3.2. Organizations with a product organizational form will have IT structures that have more distributed processing and a reduced capacity for resource sharing and communication.

A matrix organization design is a combination of product and functional structures. It is characterized by a dual reporting system in which members have to report to both

product and functional leadership. This dual role would require increased information and coordination. The complex horizontal communication channels and resource allocations associated with a matrix organization would be facilitated by increased

communication, resource sharing, and processing distribution [ 1 1 , 1 5, 1 6, 35, 36, 65]. These anticipated relationships can be summarized in the following proposition.

Proposition 3.3. Organizations with a matrix organizational form will have IT structures that have more distributed processing and an increased capacity for resource sharing and communication.

In the next section, the methodology used in IT structural taxonomy development and research proposition examination are discussed.

Research Methodology

A SURVEY INSTRUMENT WAS DEVELOPED TO DERIVE THE DIMENSIONS of the IT

structural taxonomy and to examine the organization structural characteristics that could be associated with IT structure. The first step in the development was centered on grounding the constructs in past research, either by directly adopting validated instruments or by designing items based on developed concepts. This preliminary instrument was validated through a multistage process. Initially, the questionnaire was exten-

sively pretested on MIS academics. The feedback from these sessions was used to create a prototype document that captured the anticipated item wording, ordering, and style of the final instrument. The prototype was administered and followed by in-depth interviews with twelve senior IS administrators in major firms in the United States and Canada. The

multistage process of instrument development resulted in significant restructuring and refinement of the questionnaire and preliminary support for content validity [46].




In this study, the research questions were at the level of the corporation so the constructs are operationalized at the organizational unit of analysis. To measure the dimensions of IS structure, subjects were asked to respond to semantically differentiated seven-point scales. The items identified the degree of centralization of organizational processing, the extent of network communication, and the extent of data and

applications programming sharing. Demographic information was captured using a mixture of open-ended and scaled questions.

Three constructs were used to determine the organizational structure of the subject firms. To measure the formal global organizational structure of the firm, subjects were

asked to choose from a series of descriptive classifications between matrix, product, and functional corporate structures. The second construct addressed the degree of

centralization of organizational decision making. The degree of centralization of

decision making was measured using five items that were semantically differentiated

on seven-point scales. The construct for measuring centralization of organizational decision making was adapted from a validated instrument developed by Miller and

Friesen [40]. Each question was scaled between very decentralized and very central-

ized. The third construct measured the extent of organizational integration using four

items that were anchored between "strongly agree" and "strongly disagree" on a

seven-point scale. The construct measuring organizational integration was adapted from Grover [ 1 8]. The study variables and their measures are summarized in Table 2.

A sample of 900 senior IS executives was selected from a database of 200,000 North

American firms. The sample was limited to for-profit companies with revenues greater than $50 million a year. The initial mailing was followed by a second mailing

approximately a month later to facilitate subject participation. Forty-five addresses

were determined to be invalid, and 313 were received for an effective response rate

of 36.6 percent. A test for nonresponse bias was conducted by comparing the early and late respondents' answers, and no significant difference was detected.

Construct validity of the measurement instrument for organizational integration and

centralization of organizational decision making was evaluated through factor analysis. The constructs loaded as two distinct factors. Organizational decision-making structure had an eigenvalue of 2.65 and an acceptable alpha coefficient of 0.82. The

organizational integration construct had an eigenvalue of 3.2 and an acceptable alpha coefficient of 0.91 [46].

The study was directed at senior IS managers because these subjects would have the

appropriate knowledge to answer questions concerning IT and organizational struc-

ture. To determine if the intended audience responded to the questionnaire, subjects were asked to indicate their job title. The majority of the respondents (86.3 percent) held the title of CIO, VP of information systems, director of information systems,

manager of IS, or supervisor of IS. The remaining 13.7 percent of respondents were

CEOs, project leaders, consultants, managers of systems development, or held some

similar title. The subjects would appear to have the appropriate level of expertise to

supply valid answers to the instrument. The generalizability of a study is determined by the representativeness of the

respondents. The respondents represented a cross-section of industries and sizes. Over




Table 2. Construct Operationalization

Variable Measure

Dimensions of information (1 ) Your organization's computer processing power is primar- systems structure (partially ily (centralized - distributed) adapted from [2, 32]) (2) Individual computers (including PCs) in your organization

are networked and can communicate with (no other computer-all other computers in the organization) (3) Individual computers (including PCs) in your organization can share common data and applications programs through a network with (no other computer- all other computers in the organization)

Formal organizational Please indicate your organizational structure: structure (adapted from (1) Functional (i.e., divided into production, marketing) [15, 16]) (2) Product (i.e., Division by product/service produced)

(3) Matrix (i.e., mixture of above two; subordinates report to multiple managers)

Organizational structure: To what extent are the following decisions centralized at the

degree of centralization of top levels of your organization? (very decentralized - very decision making (adapted centralized) : from [40]) ( 1 ) Capital budgeting

(2) New product/service introduction (3) Entry into major new markets (4) Pricing of major product lines (5) Personnel selection

Organizational structure: Please indicate the extent to which you agree with the follow-

degree of integration ing statements regarding your internal environment: (adapted from [1 8]) (strongly disagree- strongly agree)

(1 ) Our organization encourages exchange of ideas between departments (2) Applications are often shared among departments (3) Information is often shared among departments (4) In our organization, joint development of projects between departments occurs frequently

three-quarters of the companies were associated with manufacturing (24.9 percent), financial/business services (29.1 percent), health care (11.8 percent), or retail-

ing/wholesale (10.5 percent) industries. The remaining participants represented publishing/broadcasting, information service/software, transportation, utility, and construction industries. The size of the responding firms ranged from 21.1 percent small or medium (up to 1 ,000 employees), 56.2 percent large ( 1 ,000-1 0,000 employees) to 1 6.6 percent very large ( 1 0,000 or more employees). The diversity of the sample would strengthen the external validity of the study results.

Deriving an IS Structural Taxonomy

Clusthr analysis was used to empirically derive the IT structural taxonomy. Also known as numerical taxonomy, cluster analysis is a multivariate technique for




identifying similar entities. The first step in cluster analysis is to determine the number of clusters, which in this case is the types of IT structures. It might be assumed that the four groups suggested by Leifer [32] were all-inclusive; however, since his

typology was not empirically based and there has been an increased interest in

computer networking and resource sharing, it was appropriate to let the data suggest the number of groups through the use of hierarchical cluster analysis.

The Ward Method of agglomerative hierarchical cluster analysis was used for

empirically determining the number of groups. This method begins by treating each of the over 300 organizational respondents concerning processing centralization, communication, and resource sharing as a cluster. In a series of steps, it then combines the nearest clusters until it has created one single cluster for the entire sample population. The Ward Method determines the distance between two clusters as the sum of the squares between the clusters summed over all variables, which minimizes the total within-group sums of squares. As each cluster is added to another cluster, a coefficient is recorded that consists of the squared Euclidean distance between the two

cases that have been combined. If the coefficient is small, then the two clusters are

fairly similar; if, however, the coefficient is relatively large, then the two added clusters are dissimilar. In the analysis of this study, the coefficient was large for the

additions of the last four clusters, which suggests that there are four types of IT

structures [20, 41, 45]. Because this hierarchical cluster analysis is somewhat subjective, it is important to

validate and examine the stability of the chosen clusters. Initial validation of grouping is carried out by determining that the four clusters are significantly different from each

other using multivariate analysis of variance. In this case, each group's observed F

statistic revealed differences significant at the 0.00 1 level. To gain further confidence in the chosen clusters, nonhierarchical cluster analysis or K-means clustering was used

to determine cluster grouping and characteristics. Nonhierarchical cluster analysis starts with a predefined number of clusters and centers for each of the clusters. The

method then clusters all the cases that fall within a set distance to each of the defined

group centers. In this step, the analysis was directed to group four clusters that were

centered on the coordinates produced during the hierarchical cluster analysis. The

analysis of variance for these groups produced an F statistic that signified differences

at the 0.001 level [20,41,45]. In the final step, another nonhierarchical analysis was carried out with the designa-

tion of four groups, but with starting cluster centers that were independent of the

hierarchical analysis and maximized the distance between groups. Once again the F

statistic from the analysis of variance revealed differences significant at the 0.001

level. The coordinates for the centers of the four groups from each of the three cluster

analyses were then compared with each other to evaluate the stability of the derived

clusters. None of the group centers varied significantly across the cluster analysis or

more than 5 percent. The combination of the validity and stability analysis allows for

increased confidence in the derived IT structural taxonomy. To aid in understanding of the categorization scheme, the total sample was divided into thirds to determine

high, moderate, and low average scores using a cutoff point that was calculated based




Table 3. Cluster Analysis of Information Technology Structures*

Cluster Processing de- Intercomputer Shared data and Number centralization communication applications ofcases

Centralized computing Low Low Low 82 (2.0) (3.7) (2.4)

Decentralized comput- High Low Low 53

ing (5.4) (4.3) (3.3)

Centralized cooperative Low High High 105

computing d -9) (5.7) (5.2)

Distributed cooperative High High High 69

computing (4-9) (6.1) (5.7)

Number in parentheses represents average structural score of subjects assigned to that cluster. * The total sample was divided into high, moderate, and low average scores using a cutoff point based on the normal deviate of a standard normal curve for each of the dimensions.

on the normal deviate of a standard normal curve for each of the dimensions of the

taxonomy.

IT Structural Taxonomy

The cluster analysis produced the four IT structure types shown in Table 3. The first IT structure is characterized by highly centralized processing, low

communication, and low data- and application-sharing capabilities. This structure would seem to be consistent with the characteristics of a centralized computing environment. The second group has dispersed processing with low communication, data- and application-sharing capabilities, which appear to be consistent with a decentralized computing environment.

The third classification has centralized processing, but high capabilities for both communication and data and application sharing. The capacity for data and application sharing suggests that this system is much more than the hub-and-spokc computing structure proposed by Leifer [32]. A structure of centralized processing also indicates that this grouping was not recognized by some of the established typologies. They predicted that the ability to share applications and data would be limited to decentralized processing environments and that cooperative computing would be an extension of distributed computing [1, 47]. Because this grouping appears to be a previously unidentified type of cooperative computing, it is called centralized cooperative computing. The last group is characterized by decentralized processing with high communication and sharing. This classification appears to be an extension of distributed computing, so it is termed distributed cooperative computing.

The widespread availability of new technologies in the 1 990s appears to have altered the traditional distributed and hub-and-spoke computing environments by adding the

ability to cooperate in the sharing of data and application resources, while the




centralized and decentralized computing structures remain relatively stable. The "discovery" of two distinct types of cooperative computing has implications for both researcher and practitioner.

Examination of Propositions

The propositions anticipated that the organizational characteristics of centralization of decision making, degree of organizational integration, and formal

organizational structure would be related to the derived taxonomy of IT structure. The first proposition predicted that organizations with the most extreme decision-making structures will have IT structures that supported the least communication and data and

application sharing, while also having corresponding extreme computer processing configurations. It is anticipated that computer-assisted communication and data shar-

ing would inform leaders in distributed cooperative environments, enabling them to

carry out more centralized management. In centralized cooperative environments, computer-assisted communication would empower lower-level workers, which would direct the organization to become less centralized. This suggests that only those

organizations that did not have IT structures that allowed increased communication and data and application sharing could avoid the impact of technology to move their

organization's decision-making structure toward the more moderate center position. Proposition 1 was examined by carrying out an analysis of variance to determine if

the IT structures were significantly different in terms of the degree of centralization of organizational decision making. The assumption of homogeneity of variance was examined for the measurement of the integration using Levene's Technique and, after a power transformation1 of the data to stabilize variances, was found to be valid. As shown in Table 4, the F value was significant at the 0.001 level for the IT structural

groups. The results shown in Table 5 demonstrate that those IT structural typologies that do not support increased communication have the most extreme organizational decision-making structure, as anticipated in proposition 1. The IT structure with the highest centralization of decision making (mean = 30.76) is the

centralized computing environment. The centralized computing environment has centralized processing but low capacity for communication and resource sharing. The IT structure with the most decentralized organizational decision-making structure (mean = 22.49) is the decentralized computing environment. The decentralized

computing environment has dispersed processing and very limited capabilities for

communication, application, and data sharing. The results in Table 5 suggests that the relationship between organizational deci-

sion-making structures is consistent with proposition 1. While not all of the group organizational structure scores are significantly different, the organizations with IT structures that do not fully support communication and resource sharing have significantly different organizational decision-making structures. Centralized computing environments, which have the lowest capacity for communication and resource

sharing, have a significantly more centralized organizational decision-making structure than do decentralized or distributed cooperative computing environments (0.05




Table 4. Analysis of Variance of Decision Making: A Comparison of Information Technology Structure Groups

Variable df Sum of Mean square F p squares

Centralization of &82 ÕÕÕí decision making

Between groups 3 2,346.65 782.21

Within groups 296 33,959.23 114.73

Assumption of homogeneity of variance was examined using Levene's technique after a power transformation of the data to stabilize variances and found to be valid.

Table 5. A Multiple Comparison Test of Centralization of Decision Making

Distributed Decentralize cooperative

Mean1 computing computing Num.

Decentralized computing 22.49 51

Distributed cooperative 25.46 66 computing

Centralized cooperative 27.25 * 102

computing

Centralized computing 30.76 * * 81

* 0.05 significance level determined using Tukey's HSD. Power transformation of data to stabilize variances.

level of significance using Tukey's honest significant difference (HSD) analysis). Decentralized computing environments, which support low levels of computer communication and resource sharing, have a significantly more decentralized organizational structure than do centralized computing environments or centralized cooperative computing (0.05 level of significance using Tukey's HSD analysis).

Proposition 1 is supported by the analysis. As predicted, those organizations that were not affected by the moderating influences of IT-supported communication and data and application sharing had the most extreme decision-making focus and were significantly different.

The second proposition predicted that those organizations that had the highest capacity for resource sharing and communication would have the most integrated organizational structures. Proposition 2 was examined by carrying out an analysis of variance to determine if the IT structures were significantly different in terms of organizational integration. The assumption of homogeneity of variance was examined for the measurement of the integration using Levene's Technique and, after a power transformation of the data to stabilize variances, was found to be valid [42, 44]. As




Table 6. Analysis of Variance of Organizational Integration: A Comparison of Information Technology Structure Groups

Sum of Variable df squares Mean square F p

Integration 4.3875 ÕÕÕ5

Between 3 1,829.40 609.80 groups

Within groups 298 41417.88 138.99

Assumption of homogeneity of variance was examined using Levene's Technique after a power transformation of the data to stabilize variances and found to be valid.

shown in Table 6, in the comparison of integration, the F value was significant at the

0.005 level for the IT structural groups. The analysis summarized in Table 7 shows that, as IT structures support increased

resource sharing, organizations have a more integrative organizational structure. This

suggests that the relationship between organizational integration and the IT structural

typology is consistent with proposition 2. The centralized IT structure, which has the

lowest average capacity for resource sharing and communication, is found in the

organizations with the least organizational integration (integration = 22.44). Distrib-

uted cooperative IT structures have the highest capacity for resource sharing and

communication and the most integrative organizational structure (integration = 28. 1 8). While not all the integration scores are significantly different, centralized computing environments have significantly less integration than do the centralized cooperative or distributed cooperative IT structures (0.05 level of significance using Tukey 's HSD

analysis). Decentralized computing is not significantly different from any of the IT

structural types. However, as predicted, it is less integrated than the cooperative structures and more integrated than centralized computing structures. Even though decentralized computing has low capabilities for communication and data and application sharing, it may be high enough above centralized computing to be associated

with increased, but not significant, levels of organizational integration. The analysis lends partial support to proposition 2.

Proposition 3 predicted that matrix, product, and functional organizational structures

will be related to different IT structural types. To evaluate this proposition, the IT

structural grouping was related to the organization's formal organizational structure

using chi square and phi statistics. As shown in Table 8, the analysis produced a

resulting chi square of 1 5.82, with 6 degrees of freedom, which was significant at the

p < 0.0 1 5 level. In addition, the phi statistics = 0.23, which when taken with the other

statistics indicates that there is a moderate and significant relationship between IT

structure and formal organizational structure.

Even though chi-square analysis does not evaluate individual cell significance, further insight into the relationship between IT and formal organizational structure

can be gained by comparing the study cell frequency with that expected by random




Table 7. A Multiple Comparison Test of Organizational Integration

Centralized Meana computing Count

Centralized computing 22.44 80

Decentralized computing 25.42 50

Centralized cooperative 28.17 * 104

computing

Distributed cooperative 28.18 * 68

computing

* 0.05 significance level determined using Tukey's HSD. a Power transformation of data to stabilize variances.

Table 8. IT Structure and Formal Organizational Structure

Centralized Distributed Centralized Decentralized cooperative cooperative computing computing computing computing

Functional 55 24 71 33

Product 10 14 12 12

Matrix 17 10 20 23

% = 1 5.82; df = 6; p < 0.01 5; (j> = 0.23.

selection [51]. In Table 8, those cells that have a subject frequency higher than

expected by chance are in bold type. Functional organizations had 10 percent more centralized IT structures and 13 percent more centralized cooperative IT structures than expected. These findings support the prediction of proposition 3. 1 that functional

organizational forms will have IT structures that have centralized processing, but do not support the contention that the IT structures would have a minimal capacity for resource sharing and communication. The increased number of functional organizations that have centralized cooperative computing structures may suggest that IT is supplying increased information and coordination capabilities to traditional organizational structures or even enabling the development of new organizational structures [33].

Product-oriented firms had 45 percent more decentralized and 1 1 percent more distributed cooperative IT structures than anticipated by chance. The larger proportion of decentralized IT structures supports proposition 3.2, which predicted that product organizational forms would have IT structures that have decentralized processing and minimal resource- and communication-sharing capabilities. However, the increased

percentage of distributed cooperative IT structures, which have decentralized processing and increased resource-sharing and communication capabilities, suggests that some product-oriented firms may require more information and coordination than

anticipated by proposition 3.2. Matrix organizations had 46 percent more distributed




cooperative IT structures than would be expected. This finding supports proposition 3.3 's contention that matrix organizations would tend to have IT structures that have decentralization of processing and an increased capacity for resource sharing and communication.

Discussion

A STREAM OF BOTH EMPIRICAL AND THEORETICAL PAPERS over the past thirty-five years has suggested that there is a relationship between IT and organizational structure

[30, 33]. During this period, however, IT capabilities have drastically expanded while

only a few IT structural typologies have been described by researchers [2, 32]. This

study empirically developed a taxonomy for IT structure based on the degree of centralization of computer processing, capability to support communications, and the

ability to share resources. By using a multistep cluster analysis, we were able to derive both the membership and number of groups from the responses of the 3 1 3 firms. Four

unique and stable groups were identified; centralized (centralized processing, low com-

munication, low sharing), decentralized (decentralized processing, low communication, low sharing), centralized cooperative (centralized processing, high communication, high sharing), and distributed cooperative computing (decentralized processing, high commu-

nication, high sharing). The derived taxonomy appears to be exhaustive, mutually exclu-

sive, stable, and consistent. The results also suggest that the widely accepted three-group

typology (centralized, decentralized, and distributed cooperative) fails to recognize the

emergence of centralized cooperative IT structures as an option for system architecture.

As summarized in Table 9, three research propositions were proposed. The first

proposition, which was supported, predicted that IT-supported communication and

resource-sharing capabilities would be associated with less extreme centralized or

decentralized organizational decision-making structures. When IT structures sup-

ported communication and resource sharing (centralized cooperative and distributed

cooperative structures), organizational decision making was less extreme, but still directly related to processing decentralization. Only when IT structures were not supportive of

communication and resource sharing (centralized and decentralized structures) did organizations have highly centralized or decentralized decision-making structures.

The second proposition, which was also partially supported, predicted that companies with IT structures that supported resource sharing and communication would have

more integrated organizational structures. The third proposition predicted that matrix,

product, and functional organizational forms would be related to different IT structural

types. The proposed relationship was statistically significant. Three specific predic- tions were partially supported: (1) matrix organizations would be associated with

distributed cooperative computing structures, (2) product organizations would be

associated with decentralized computing structures, and (3) functional organizations would be associated with centralized computing structures.

Implications

It is important to recognize that the ultimate purpose of the IT structure is

to support the firm, and this can best be achieved if the capabilities and characteristics




Table 9. Summary of Propositions

1 . Organizations with the most extreme decision-making Supported structures will have IT structures that have a reduced capacity for communication and resource sharing, and corresponding centralized or decentralized computer processing.

2. Organizations with higher levels of interdepartmental inte- Partially supported gration will have IT structures that have a greater capacity for resource sharing and communication.

3. Organizational structures as represented by matrix, prod- Partially supported uct, and functional forms will be related to different IT structural types.

of the IT structure match the requirements and the nature of the organization. The

study's findings of stability of clusters, predictive validity (i.e., anticipated relation-

ships to organizational decision making, integration, and forms), and interpretability with prior typologies add credence to the robustness of the derived taxonomy. The

derived IT structural taxonomy could be useful in describing and facilitating the

matching of IT and organizational structures, which would have implications for both

practitioners and academics. As part of the discussion of these implications, it is

worthwhile to review a summary of the relationships between the taxonomy classifi-

cations and organizational characteristics, which are found in Table 10.

The emergence of two types of IT structures that have high capabilities for commu-

nication and resource sharing may have special significance. These structures represent the current push to use modern IT for client-server computing systems, which

have been recognized as an enabler for organizational change processes, such as

business process redesign [48]. The commonly accepted nature of client-server

computing would be captured in the distributed cooperative computing structure. Not

surprisingly, organizations with this IT structure are integrated and have a decentralized decision-making focus. These organizations have primarily either a matrix

structure or are product-oriented firms. However, the emergence of a centralized

cooperative computing structure could have a variety of implications, especially since

it has not been previously identified in prior typologies and it is the most prevalent IT

structure (n = 1 08) in the study. For example, the centralized cooperative systems may be able to avoid some of the difficulties that traditional distributed client-server

systems have with data security, backup, and system maintenance [48]. Centralized cooperative computing is the most common IT structure of functional

organizations, even though these organizations may not be as "glamorous" as those with more cutting-edge structures. Because IT structure may be altered more readily than organizational structure, these structures could be the precursors of organization change [32, 48]. There may be some evidence to suggest that those functional firms that have centralized cooperative systems and those product firms with decentralized

cooperative systems are modifying their organizational structures toward more information-intensive structures, as might be suggested by coordination theory, agency




Table 10. Summary of Relationships between IT Structural Taxonomy and

Organizational Characteristics

Centralization of Organizational Organizational decision making integration structure

Centralized 30.76 22.44 Functional computing (11%)*

Decentralized 22.49 25.42 Product computing (45%)*

Centralized 27.25 28.17 Functional cooperative computing (1 3%)*

Distributed 25.46 28.18 Matrix (46%)* cooperative computing Product (1 1 %)*

* Percentage of cases higher than expected.

theory, and an information-processing view of the firm. As predicted by a contingency theory of information processing and organizational structure [12, 28, 61], when organizations are faced with increased uncertainty, their traditional mechanisms for dealing with

uncertainty begin to break down. If the organization cannot reduce the uncertainty, ways of providing more information must be developed by "investment in vertical information

systems," or employing joint tasking through the "creation of lateral relationships" [11, 12, 15, 16]. The development of these mechanisms would be facilitated by the matching of cooperative IT structures that provided richer information through increased capabilities for electronic communication and resource sharing [ 1 1 , 1 2, 34, 35].

Ultimately, the development of these resources may play a role in the transformation

of classical organizational forms into more coordination-intensive organizational structures (e.g., adhocracy, networked, and virtual organizations) [27, 33, 60]. The

fluid decision-making structure of an adhocracy is one alternative to a traditional

centralized or decentralized decision-making structure. In theory, an adhocracy would

allow both workers to be empowered and senior management to make operational decisions as the organizational structure responds to its dynamic environment [35]. As internal barriers for communication vanish, formal organizational structures may be replaced as a means for task completion by a web of interrelationships that represent networked firms [60]. When external organizational barriers collapse, new cooperative relationships between firms may create the widespread existence of virtual

organizations [33]. If cooperative computing systems are lowering coordination costs,

facilitating the development of more "collaborative" organizations, then these product and functional organizations may be on the verge of organizational evolution [27].

Directions for Future Research and Study Limitations

The ability to match IT and organizational structures is important. How-

ever, before definitive prescriptions can be made, further research is needed into the




nature of the causal relationship between IT and organizational structure. The rela-

tionship between IT and organizational structure has been a critical concern for over three decades [17, 30, 37, 59]. While this study focused on the "matching" of IT and

organizational structures without addressing the causality issue, the question as to whether IT affects organizational structure (technological imperative), organizational structure affects IT (organizational imperative), or whether there is a complex interactive effect among management, IT, and organizational and environmental factors, is an

important issue. In the future, researchers may wish to examine the differences between those functional and product organizational forms with traditional centralized and decentralized computing structures and those with cooperative structures. There may be an opportunity to gain understanding of the causal interactions between IT structure and the organization to determine if those organizations with cooperative IT structures follow an evolutionary path toward more coordination-intensive structures. It is also important to determine whether that transition is planned or a

consequence of the environment. Further study is needed, through longitudinal or

experimental research, to determine the direction and extent of the causal relationship between IT and organizational structure.

Further research is needed into the ultimate benefits of matching IT and organizational structure. This study was limited by the implicit assumption that the current

matching of structures is desirable. Studies are needed to measure the effectiveness of

different combinations of IT and organizational structures to determine prescriptions for optimizing organizational performance. Advances in information technology will also create the need for refinement and expansion of the classification scheme. It may be helpful to use a more robust measure of organizational types to determine how IT structures are related to new organizational types (e.g., virtual corporations, negotiated firms, etc.). As management is increasingly pressured to adjust organizational structures through downsizing, business process redesign, or developing new relationships with employees, customers, suppliers, and outsourcers, it is increasingly important to determine the role IT structure may have in enabling the successful fulfillment of

organizational goals.

Note 1 . Power transformation is a technique used to stabilize variances. A power transformation

raises each of the data values to a specific power. In this case, each of the values was squared [42, 44].

REFERENCES

1 . Ahituv, N.; Neumann, S.; and Riley, H. Principles of Information Systems for Manage- ment, 4th ed. Dubuque: Brown Communications, 1 994.

2. Ahituv, N.; Neumann, S.; and Zviran, M. Factors affecting the policy of distributing computing resources. MIS Quarterly, / J, 4 ( 1 989), 389-40 1 .

3. Allen, B., and Boynton A. Information architecture: in search of efficient flexibility. MIS Quarterly, 75,4(1991), 435-445.

4. Bailey, K. Monothetic and polythetic typologies and their relation to conceptualization, measurement and scaling. American Sociological Review, 38, 1 (1973), 18-33.

5. Bailey, K. Methods of Social Research, 2d ed. New York: Free Press, 1982.




6. Baiman, S. Agency research in managerial accounting: a survey. Journal of Accounting Literature, 2 , 1 ( 1 982), 1 54-2 1 0.

7. Bakopoulos, J. Toward a more precise concept of information technology. Proceedings of the Sixth International Conference on Information Systems ( 1 985), pp. 1 7-24.

8. Brynjolfsson, E. Information assets, technology, and organization. Management Science, 40,12(1994), 1645-1662.

9. Brynjolfsson, E.; Malone, T.; Gurbaxani, V.; and Kambil, A. Does information technology lead to smaller firms? Management Science, 40, 12(1 994), 1 628-1 644.

1 0. Cash, J.; McFarlan, F.; McKenney, J.; and Applegate, L. Corporate Information Systems Management, 3d ed. Homewood, IL: Irwin, 1992.

11. Daft, R. Organization Theory and Design. St. Paul, MN: West Publishing, 1983. 12. Daft, R., and Lengel, R. Organizational information requirements, media richness and

structural design. Management Science, 32, 5 (1986), 554-571. 13. Davis, G., and Olson, M. Management Information Systems: Conceptual Foundations,

Structure and Development, 2d ed. New York: McGraw-Hill, 1985. 1 4. Doty, H., and W, Glick, W. Typologies as a unique form of theory building: toward improved

understanding and modeling. Academy of Management Review, 19,2 (1 994), 230-25 1 . 15. Galbraith, J. Designing complex organizations. Reading, MA: Addison-wesley, 1973. 16. Galbraith, J. Organizational Design: An Information Processing View. Reading, MA:

Addison- Wesley, 1973. 17. George, J., and King J. Examining the computing and centralization debate. Communi-

cation of the ACM, 34, 1 ( 1 99 1 ), 62-72. 18. Grover, V. An empirically derived model of the adoption of customer-based inter-

organization systems. Decision Sciences, 24,3 ( 1 993), 603-640. 19. Gurbaxani, V., and Whang, A. The impact of information systems on the organizations

and markets. Communication of the ACM, 34. 1 (1991), 59-73. 20. Hair, J.; Anderson, R.; Tatham, R.; and Black, W. Multivariate Data Analysis, 3d ed.

New York: Macmillan, 1992. 21 . Hicks, J. Management Information Systems, 3d ed. New York: West Publishing, 1993. 22. Huber, G. A theory ot the ettects ot advanced înîormation tecnnoiogies on organizational

design, intelligence and decision making. Academy of Management Review, 15, 1 ( 1 990), 47-7 1 . 23. Huber, G., and Glick, W. Organizational Change and Redesign. New York: Oxford

University Press, 1993. 24. Johnson, D. Organizational Communication Structure. Norwood, MA: Apex, 1993. 25. Kaestle, P. A new rationale for organizational structure. Planning Review, Iti A('WV),lv-il. 26. Keen, P. Competing in lime: Using telecommunications for competitive Aavantage.

New York: Ballinger, 1988. 27. Keen, P. Shaping the Future, Business Design through Information Technology. Boston:

Harvard Business School Press, 1 99 1 . 28. Keller, R. Technology- information processing fit and the performance ot R&D project

groups. Academy of Management Journal, 37, 1 (1994), 167-179. 29. Laudon, K., and Laudon, J. Management information òystems, 3d ed. New Y one:

Macmillan, 1994. 30. Leavitt, H., and Whisler, T. Management in the 1 980 s. Harvard Business Review, 36, 6

(1988), 41-58. 3 1 . Lee, S., and Leifer, R. A framework of linking the structure of information systems and

organizational requirements for information sharing. Journal of Management Systems, 8, 4

(1992), 27-44. 32. Leifer, R. Matching computer-based information systems with organizational structures.

MIS Quarterly, I2,'(' 988), 63-73. 33. Lucas, H., and Baroudi, J. The role of information technology m the organization design.

Journal o/Management Information Systems, 10, 4 (1994), 9-23. 34. Malone, T. What is coordination theory? Working Paper no. 82, Center for Information

Systems Research, MIT, 1988. 35. Malone, T., and Rockart, J. Computers, networks and the corporation. Scientific Ameri-

can, 261, 3 (1991), 128-136. 36. March, C, and Simon, H. Organizations. New York: John Wiley, 19!>8.




37. Markus, L, and Robey, D. Information technology and organizational change: concep- tions of causality in theory and research. Management Science, 34, 3 (1988), 583-598.

38. Martin, E.; DeHayes, D.; Hoffer, J.; and Perkins W. Managing Information Technology. New York: Macmillan, 1991.

39. Miles, R., and Snow C. Organizational Strategy, Structure and Process. New York: McGraw-Hill, 1978.

40. Miller, D., and Friesen P. Innovation in conservative and entrepreneurial firms: two models of strategic momentum. Strategic Management Journal, 3, 1 (1982), 1-25.

41 . Milligan, G. A review of Monte Carlo tests of cluster analysis. Multivanate behavioral Research, 16, 3 (1981), 379-407.

42. Neter, J.; Wasserman, W.; and Kutner, M. Applied Linear Statistical Models, 3d ed. Boston: Irwin, 1990.

43. Niederman, F.; Brancheau, J.; and Wetherbe, J. Information systems management issues of the 1 990s. MIS Quarterly, 75,4(1991), 475-495.

44. Norusis, M. SPSS Base System user's Guide. Chicago: SPSS Inc., 1993. 45. Norusis, M. SPSS Professional Statistics 6.1. Chicago: SPSS Inc., 1994. 46. Nunnally, J. Psychometric Theory. New York: McGraw-Hill, 1967. 47. O'Brien, J. Management Information Systems: A Managerial End User Perspective, 2d

ed. New York: Irwin, 1993. 48. O'Hara, M., and Watson R. Automation, business process reengineering and client server

technology: a three stage model of organizational change. In V. Grover and W. Kettinger (eds.), Business Process Change: Reengineering Concepts, Methods and Technologies. Harrisburg, PA: Idea Group, 1995.

49. Orlikowski, W., and Robey, D. Information technology and the structuring of organizations. Information Systems Research, 2, 2 ( 1 99 1 ), 1 43-1 69.

50. Porter, M. Competitive Strategy. New York: Free Press, 1980. 5 1 . Raho, L.; Belohav, J.; and Fiedler, K. Assimilating new technology into the organization:

an assessment of McFarlan and McKenney's model. MIS Quarterly, 11, 1 (1987), 46-57. 52. Rich, P. The organizational taxonomy: definition and design. Academy of Management

Review, / 7, 4 ( 1 992), 758-78 1 . 53. Segars, A., and Grover, V. Communications architecture: towards a more robust under-

standing of information flows and emergent patterns of communication in organizations. European Journal of Information Systems, 3, 2 ( 1 994), 87-1 00.

54. Sproull, L., and Kiesler, S. Computers, networks and work. Scientific American, 261, 3 (1991),116-123.

55. Statistical Abstracts of the United States. Washington, DC: U.S. Department of Com- merce, Bureau of Commerce, 1986.



58. Sträub, D., and Wetherbe J. Information technologies for the 1990's: an organizational impact perspective. Communication of the ACM, 11,32 ( 1 989), 1 328-1 339.

59. Tavakolian, H. Linking the information technology structure with organizational competitive strategy: a survey. MIS Quarterly, 13,3,(' 989), 309-3 1 7.

60. Teng, J.; Grover, V.; and Fiedler, K. Business process reengineering: charting a strategic path for the information age. California Management Review, 36, 3 ( 1 994), 9-3 1 .

61. Tushman, M., and Nadler, D. Information processing as an integrating concept in organizational design. Academy of Management Review, 3, 3 (1978), 613-624.

62. Wardle, D. The evolution of information systems architecture. Proceedings of the Fifth International Conference on Information Systems, 1 984, pp. 205-2 1 6.

63. Warriner, C. Organizations and Their Environments: Essays in the Sociology of Orga- nizations. Greenwich, CT: JAI Press, 1984.

64. Williamson, O. The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. New York: Free Press, 1985.

65. Zwass, V. Management Information Systems. Dubuque: Wm. C. Brown, 1992.



an empirically derived taxonomy of information technology … · 2017. 9. 30. · frames the...

Documents