HD28 .M414 *&& Center for Information Systems Research Massachusetts Institute of Technology Sloan School of Management 77 Massachusetts Avenue Cambridge, Massachusetts, 02139 SOCIAL DETERMINANTS OF INFORMATION SYSTEMS USE: A NETWORK-ANAYLTIC APPROACH David C. Robertson November 1988 CISRWPNo. 185 Sloan WP No. 2099-88 '1988 D. C. Robertson Center for Information Systems Research Sloan School of Management Massachusetts Institute of Technology Social Determinants of Information Systems Use: A Network-Analytic Approach* David C. Robertson Center for Information Systems Research Massachusetts Institute of Technology Sloan School of Management 50 Memorial Drive E53-334 MA Cambridge, 02139 Abstract The social group of which an individual information systems interpreted usage. A and (IS). The social group used. The social field study a part will affect is of group this idea will affect will also how how that individual uses an information system is demands on IS place pressures and was conducted. The concept of structural equivalence, operationalized by blockmodeling, was used to measure the effect of the social group on individual information systems usage. study show that structural equivalence is related to different for different information systems. implications for many important user involvement in the IS areas of IS IS The results of the field usage, but this relationship is These results are believed to have research such as the strategic use of IS, design process, and research on the impacts of IS. * The author would like to thank John Henderson, Andy Boynton, Soonchul Lee, Gordon Walker, John Sviokla, and several anonymous reviewers for their assistance and ideas on this research. I. Introduction The use of an information system usability and usefulness background of the many user, will be determined by many factors, such as social group of which the user is Too often, a part. of these factors have been ignored, and the resulting research suffers. by understanding systems (IS) designing these factors can all we understand how and when are used. The factors affecting IS, applying IS and of the system, the tasks facing the user, the training and the for competitive IS the Only information use are important to consider when advantage, or predicting the impacts of IS. Impacts research especially has suffered from ignoring these factors. The goal of paper this is to explore one specific determinant of an individual's decision to use a system: the social group surrounding the individual. that the social group can exert strong influences on the uses a system. The individual uses an social IS. group way will idiosyncratically affect it is is argued which an individual how and when Thus, similar organizations (and even similar groups same organization) could use the same information system Thus in It in the the very differently. understandable that much research on the impact of information systems on organizations has had mixed and even contradictory results (25, 27). Poor research methods, different operationalizations of the same variables, different levels of analysis, and inadequate theory are variance is also in results (24,36,38). It is argued that the all lack of partly responsible for the convergence in the results due to the fundamental assumptions behind some impacts research: that can cause any impact to occur. with the features of an IS, Rather, affect it is IS believed that the social system, along how and where that IS will be used. This idea is tested using the concept of structural equivalence, and the results of a field study are reported. First, a short review of the IS - 1 impacts literature - is presented. II. A Review of Selected Impacts Research In this section, the research on three different types of quality of working reviewed: life impacts will be IS (QWL), centralization/decentralization, and height of the hierarchy. The purpose of this section is not to survey the research on the impacts of information systems on the individuals and organizations that use them; there have been many good reviews in this representative types of impacts are selected to common in the quality of working on job content and job in the life Some satisfaction. show the mixed increase influence, many research has found that in QWL early research postulated that information systems that IS can enlarge jobs, Some making and studies show would IS on the authority. While increase the ability of to monitor subordinates, thus centralizing authority (31, 42), area of research to be reviewed is (15, 32, 33). research on the impacts of height of the organizational hierarchy. The height of the hierarchy levels of reporting relationships. This area of research and Whisler can cause a on the impacts of subsequent research has found no such simple relationship final IS IS QWL (3, 7). similar pattern of results occurred in research management show (16, 21, 28). relative centralization or decentralization of decision many that are category focuses on the impacts of case studies and generally improve both an increase and decrease The results QWL of clerical jobs by increasing job stress, lowering job content, or deskilling jobs (22, 31, 39, 41). Yet A Rather, three (3, 17, 25). impacts research. in The research decrease area (26), who argued that IS would in this in management. on the numberof 1958 by Leavitt automating While some area has found that information systems do tend to flatten the hierarchy (14, 29, 42), other studies actually increased the started the flatten organizations by functions traditionally performed by middle subsequent research was is IS number of levels of found that information systems management (8, -2- 9, 32). This lack of consistent results research is due to many Comparing results. in these three as well as other areas of impacts Poor methods and measures cause much noise factors. different types of information technology also introduces error. Inadequately conceptualized concepts and theory can cause Yet the lack of results is due in in impacts research part to the assumption implicit cause an impact to occur. Kling thing we learn is in due to more than these problems; some impacts in (25), after is results to vary as well. it that computers research: reviewing the literature, states, "the first that computers by themselves 'do' nothing to anybody (pg. 100)." Impacts are mixed because organizational contexts are mixed, and it is the interaction of the organizations and the information system that ultimately determines the impact of that system. Attewell and Rule we suspect that the transformations computing are so multifarious as to effect relations life through encompass the most disparate cause- what appear to be the "same" causes even identical organizations (pg. Thus, mixed impacts can organizational different contexts.... [0]ne might expect quite in different effects to ensue from similar or in (4) state: in a 1 190). sense be viewed as a replicated finding of system can have different impacts in in different organizations. : the same type Individuals in different social groups will have different influences on their usage, and will thus use information systems differently. III. Social If Determinants of Information Systems Usage the usage of an individual using the occurs. how IS is IS, determined then we must There are two ways the social in part by the social define and measure group can have an the individual interprets the system and demands from the group on group surrounding the (2) how effect: (1) by affecting by the social pressures and the individual using the system. -3- this social effect Initially, the social group the manner will affect in individual interprets the system system) with it (6). The group, social which an (i.e. how when IS is he or she decides faced with a and form ideas about which used by affecting the way the tasks it new when and how to technology, will use the experiment should or shouldn't be used to solve. Different people will have different (sometimes radically different) views of the technology and its Over time, these interpretations of the technology capabilities. be tested and refined, and may converge within the will social interpretation process occurs, the tasks performed by the group It is believed that the usage of a new IS within is not a new interact to social finding. The other pressures determine the groups influence the perceptions of the individuals Social psychologists social in have known for years that the if an individual's rivals in tasks, or the firm are using a required new group on the in this IS usage by supervisors. system, then the system may fear, envy, or curiosity. the social group of which the used new from requests to use a particular system arise order to test the proposition that the usage of an social not a determinant of computing usage comes from individual role and demands. These can be used out of is many ways. from others working on interdependent In the in Turkle (37) argued that computers are highly abstract objects that can be interpreted Also, is system can change an individual's perception of unchanging physical objects conclusion either. the nor the opinions of those That computers can be perceived differently by different people (2). as well. the social group. IS in The general idea that social IS group using the system. Rather these two factors proper role of the may change As this (and the ultimate impact of that system) not determined solely by the features of the social group. IS user individual. is a part, we must The concept of paper to dimensionalize the social -4 IS is in first part determined by quantify the influence of structural equivalence will system so that its be influence can be measured Structural equivalence concept that, the be linked to will and IS meaning its be discussed next. Following will usage. Structural Equivalence Structural equivalence patterns If technical (1, 11, 34, 43) is measure of a two people work with the same and administrative help, go others, similarity in interaction to the same others for and report to the same person, then the two are Two structurally equivalent (with respect to those relationships). structurally equivalent individuals can be said to occupy the same or more structural position. Structural equivalence analysis aggregates interaction patterns. is thus a any in B. The of the interaction whom" social structure between any As such, structural equivalence captures than that captured by formal data to structural positions individual in overall distill A and position B A and revealed reflects the informal structure of the organization-- the actual pattern of interactions responsibilities, to The interaction between any two good estimate position "who between organizational members. a social A formal title. network title characteristic different imperfectly defines the knowledge, and behaviors of the individuals holding that title. A structural position (with respect to task interactions) will separate those individuals who have similar interaction patterns, characteristics. For example, if persons Consultant," while person C has the database issues go to persons B which may better and title C, A and reflect actual individual B both have the "Manager," but then B and C will all HI: first "Database requests for help on be structurally equivalent. Thus, structural positions will be expected to be different from the title titles. This leads to hypothesis: Structural equivalence will capture an effect different from that captured by title. Structural equivalence has been linked to many important concepts two of which literature (19, 30), will be reviewed here. Following that, a third Walker possible interpretation of structural equivalence will be discussed. examined the network of interactions across the in (40) five types of task relations in an organization and found that structurally equivalent individuals perceived product goals in similar ways. Structural position was linked to cognition and was found to be "a stronger and more stable predictor of differences in cognition than either the type of function an individual had orthe type of product worked on (pg. 103)." A second concept Boorman and White social role. pressures that structural equivalence can be linked to (11) argued that structural equivalence and demands placed on them. from the Coleman doctor's time to adoption was a valid measure of interaction with those in Burt (13) supported this idea gammanym study. argued that, while a doctor other structural positions using a his structural position start using his Burt found that a may have frequent new drug, it is only when the drug that emotions such as fear, envy, and worries about loss of status begin to have an effect. Thus structurally equivalent individuals behave similarly because the social influences and on them are in best predicted by the time to adoption of other structurally equivalent doctors. Burt in that of role. Individuals with similar interaction patterns will have similar social reanalysis of the data others is is demands similar. Structural equivalence (with respect to task interactions) could also be of as reflecting individual competence. individuals receive It thought can be argued that structurally equivalent and send the same task-related interactions because they have the same competencies: they perform the same types of tasks or possess the same types of task-related knowledge. -6 and Structural Equivalence Use IS Structural equivalence (with respect to task relationships) can be expected to be associated with the usage of stated above. First, just similarly, structurally similarly. new information systems for each of the three reasons as structurally equivalent people perceive product goals equivalent people would be expected to interpret an IS Structurally equivalent people (with respect to task relationships) receive the same types of task-related information. Since the application of information systems to tasks (and the definition of tasks in terms of information systems) can vary around an organization, different structural positions could receive different interpretations of H2: new information systems. Thus: Individuals In different structural positions will interpret systems in different pressures and is ways. predict usage of a new that structurally equivalent people would receive the same A second reason why information system structural equivalence demands on individual begin to use a system out of fear of equivalent people and use information behavior. new If would people structurally equivalent to an system, then that individual loss of status, curiosity, may work with or envy. or report to the may begin Furthermore, structurally same people, thus increasing the likelihood thatthe individuals are requested or required to use a given H3: to use the IS. Individuals in different structural positions will have different social pressures and demands, and thus It is will use information systems in different ways. important to distinguish between hypotheses H2 and H3, both conceptually and empirically. Conceptually, the difference internally or externally motivated. is related to whether usage is Hypothesis H2 argues that structurally equivalent individuals internalize similar interpretations and thus use information systems similarly. Hypothesis H3 argues that usage results from factors external to -7- the user: requests or requirements for use, or use triggered by fear of loss of status, curiosity, or envy. Empirically, hypotheses If structural equivalence demands and related to the amount primary in IS way size of that relationship. usage through If, its relationship to social amount of training received however, structural equivalence relationship to the interpretation of the its in that interpretations are communicated. There formal training will individual. Training will thus explain usage explained by training more strongly A the Training is little related to social third reason why much of the variance removed, the demands and may have in IS usage When the residual variance will be pressures. structural position could predict competence and task assignment. task relationships) is is be informally conducted by those associated with structural equivalence's relationship to interpretation. variance then IS, of training received should affect this relationship. the organization, thus the training received around the also considering training. related to usage only through pressures, then the inclusion of the would not change the is is H2 and H3 can be separated by IS usage relates to Structurally equivalent people (with respect to similar competencies. People with similar competencies would tend to perform similar tasks and would thus use information technology tasks, but is in similar good ways. A given information system can be used to solve at solving relatively fewer. For example, while a spreadsheet can be used as a word processor, few would decide that fashion. Thus two different individuals, many when it should be used in that faced with two similar tasks and given the same limited set of information systems with which to solve those tasks, will tend to choose the same technology. depend on the task at hand- available, others relationship to some may have few competence tasks The range of choices available will may have many good information systems or none. Thus if structural equivalence predicts task assignment, then 8 it through its could be expected that it will predict some types of usage as well This leads to the final two hypotheses to be tested: H4: Different structurally equivalent groups will have different task assignments. H5: People with similar task assignments will use IS in similar ways. Methodology IV. The hypotheses were tested people were surveyed in in one office of a management consulting firm. 93 70 usable replies were received for a January of 1986. response rate of 75.3% Measuring Information Systems Measuring the usage of the available information systems was done by identifying (through interviews with five members of the organization positions in the organization) the important systems. the organizations were all The IS in first various most widely used by personal computer-based software programs. The five programs, followed by an acronym, the uses of each, and the date the organization started using each program are presented in Table Table 1. 1 The Information systems Technology System Designer's Acronym 5DW Uses Identified Computer-Assisted Software Engineering (CASE) tool PMW Forecasting project-related resource Workbench Project Manager's Workbench Initiation Date Summer 1985 Summer 1984 requirements and expenses; project tracking Mac Paint MAC Presentation graphics; data flow Early 1984 diagrams Lotus 123 123 Project management (same as PMW); Early 1983 project-related financial analysis Word Processing WP Word processing (no graphics) -9 Before 1981 For each type of technology, the respondents were asked to rate their frequency of usage on a six corresponding to no usage or experience and point scale with 5 corresponding to a great deal (more than 16 hours per week). Training was rated from all (no training) to 4 (one to two weeks of training). The correlation matrix for the usage variables is shown in Table 2. Table 2 SDW Use SDW SDW MAC MAC Tng. Use Tng. Use PMW PMW Tng. Use 123 123 WP Tng. Use Tng. and the until performed on either or both of the two groups. analysis the size of groups all is seven members or The second algorithm, CALCOPT, moving members from one group fit can be obtained. refines the CONCOR to another until developed by Boorman, and enhanced by Walker fit The interactions used in structured interviews interactions which amount in were applied partition by successively no improvement in this is in blockmodel study was originally The function used to (see [40]) of the blockmodel partition described in (10). the blockmodel were obtained through a series of which respondents were asked to name the types of completing their jobs. crucial in of communication across final questionnaire, is less. The version of CALCOPT used evaluate the goodness of CONCOR all In types of interactions addition, the total was measured. In the the most important interactions were measured by two questions (one each concerning the sending and receiving of these types of communication). The major task-related interactions appeared on the questionnaire) (in the order in which they are: • Receive feedback on job performance • Receive help on problems related to a specific technology (i.e. database systems, telecommunications, etc.) • Receive help on problems related to a specific industry (for example, problems that are unique to the oil or the banking industry). • Receive information aboutthe coordination of projects. • Trade information information about the how projects are progressing (with respect to deadlines, specifications or budget). • Trade information about the • Give feedback on job performance. • Provide help on problems related to a specific technology. • Provide help on problems related to a specific industry. political situation at 11 the client site. In • Give information about the coordination of projects. • Total communication. who addition, the respondents were asked to identify those individuals controlled resources critical the individuals All in to the success of their project(s). the office were listed in Holland and the questionnaire. Leinhardt (23) recommend this to reduce the measurement error inherent questionnaire. The responses were coded as binary data, with a presence of a The tie. total result of the 1 " the indicating the communication question was coded so that indicated that the individuals communicated once per The " in a "1" week or more. blockmodel analysis of the survey data was a blockmodel with fourteen structural positions. These positions were used as the categories for the ANOVA tests of Hypotheses H2 and H3. Measuring Task Task assignment was collected using measures defined and used by the consulting firm. The firm categorizes the projects it undertakes so that individuals with different competencies can be assigned to the projects that require their Thus the different project types types are listed in Appendix A. will reflect different types of tasks. Individuals were asked skills. The project to rate the percentage of time they spent on each project type. V. Results The and Discussion results of the Hypothesis tests are H1 was tested using results shown in Table shown in Tables 3 through 6. Hypothesis a crosstabulation of titles versus structural positions. 3 show that respondent titles have little The relation to structural positions. The ways tests of in which hypotheses H2 and H3 are shown structural position could in Table be related to 12 4. IS There are two possible usage; either through Table 3 Table 4 Anova Tests of Hypotheses H2 and H3 (Categories are Structural Positions) Information reduced by including training significantly system, WP, the receive similar IS demands and usage, through its become statistically pressures to use these three systems way that structural equivalence could be related to relationship to task assignment, structural positions as categories (Table The for the oldest the same structural position in showed variables included 5). support for the tests, with the Hypothesis H5 was tested using stepwise the final model are in little ANOVA Hypothesis H4 was tested using multiple hypotheses. In fact, inclusion of training causes the relationship to test of the other possible regression. the analysis. Thus we can conclude that individuals significant. The in shown in Table 6. Of the usage measured, only two were predicted by task assignment. five types of The System Designer's Workbench (SDW) was associated with assignment to Data Modeling (DM) projects. This application of this relatively tools. The usage cf the is new VI. documents system. MAC was Information Systems Planning "boilerplate" as expected, as those projects Thus were the only formal this project has few choices for IS associated with assignment to Strategic projects. (SISP) for this project It were stored is possible that most of the in this format. Conclusion The hypothesis that the structural position of an individual is related to how or she uses information systems received limited support from the data. relationship is believed to occur in two ways: attention to the social influences and indirectly, through the directly, demands social system's influence he This through the individual's of those around him or her, and on how the individual interprets the information systems available. The organization studied is small and highly connected, and the individuals within are very experienced with information technology. The results of a similar study in a different firm would be expected -15 to be different. If the firm is larger or Table 5 ANOVA Tests of Hypothesis H4 (Categories are Structural Positions) Sum Type of Task SISP Main Effects: Residual: CSDI Main Effects: Residual: MRDP Main Effects: Residual: DM Main Effects: Residual: DSS Main Effects: Residual: HRM Main Effects: Residual: FA Main Effects: Residual: AUDIT* IE Main Effects: Residual: MFG Main Effects: Residual: GOVT Main Effects: Residual: BANK Main Effects: Residual: MOA Main Effects: Residual: AOS Main Effects: Residual: SYSSEL Main Effects: Residual: of Table 6 Stepwise Regression Test of Hypothesis H5 (Variables Dependent R-Squared F-value DM 0.152 12 17** SISP 0.071 632* Vanable(s) SOW MAC PMW <None> <None> <None> 123 WP * Sigmficantatthep<0. 05 ** Significant at the p<0 level. 01 level. connected then the variation groups could be higher. would model) Significant Variable less in final in usage between different structurally equivalent Firms with less technologically experienced also be expected to have higher variation in local usage patterns. High experience with information systems would lead an individual to be on those around him or her The findings of for interpretation of a this study employees less dependent new system. have implications for many of the debates in MIS. For example, the competitive advantage caused by a strategic information system will be determined in part by the unique product features of the system implication of this study quite different the system. between is that the perception different users, (SIS) One (5). and use of these features may be and between the users and the providers of Thus different users may perceive the competitive advantage of an information system quite differently. The methods A new SIS in this may study create may also be used to improve our understanding of new patterns of interaction within and between organizations. Blockmodeling (and other network analytic methods) can help meaning from the complexity understanding of how SIS are of the interaction data, and can thus help distill in changing the relationships between organizations. 17- SIS. the Another implication of the work is concerned with the design process and the importance of user involvement. Many have argued that involvement process of the future users of the system of the benefits of involvement target user group. system is If, may be however, is critical even if the design One to the success of that system. a fuller interpretation of the system by the social forces play a large role in used, then the usage of the system by different, in some other when and how user group may be a quite the second group faces similar tasks. Thus the benefits of user involvement may be much larger for the user group involved than they are for future users. in the book A vivid example of this phenomenon Industrial Future research Democracy at Sea in this in a different setting in described (35)*. area could focus on the features of associated with similar impacts is different organizations, IS that are likely to be and features whose impacts could vary. Future research should also expand and further test the taskrelated hypotheses. It assignments would result H5). could be that a different conceptualization of task in positive results for the task-related hypotheses (H4 Finally, future research could try to organizations cause high or low variations and understand what characteristics of in interpretation between different groups. Trice variable and Treacy (36) note that usage of an information system through which information systems impacts occur. usage of information systems is If is it the intervening is true that the determined by the interplay of the features of the information system and the characteristics of the organization, then the focus of impacts research must shift to include both in any analysis. In fact, what must be book focuses on a merchant marine ship which had been redesigned to encourage greater democracy and better quality of work life. The results of the redesign were found to be very successful for the crew and officers who participated in the design process. The book chronicles the behavior of the second group to use the ship. This second group in effect redefined the features of the ship to implement a more traditional social structure. Few of the gains of the previous group were evident. * This 18 , considered two (20). not either the is The unit of IS or the organization, but the relationship between the analysis should no longer be the information system nor the organization, but the information system with respect to a particular organization. No longer would an context X." IS be described as "easy to use," but rather "easy to use The determinants of the decision to use an information system in will include the features of the system, the characteristics of the individual, and the social context in which the individual isembedded. The purpose of this research organizations. have lead to a is to challenge some assumptions on the role of IS in The technologically deterministic assumptions of impacts research great deal of mixed and contradictory and attempted to show the deterministic position. It is validity of hoped that a results. This study has adopted an alternative view, an organizationally debate on this issue will result in a clearer understanding of the role of information systems in organizations. understanding can add much to the debates on many other 19 issues in MIS. A clearer Appendix A Task Assignments Strategic information systems planning (SISP). Computer systems design and implementation (CSDI). Managerial review of data processing (MRDP). Decision support system design and implementation (DSS). Information systems selection (SYSSEL). Advanced office systems design (AOS). Manufacturing (MFG). Data modeling (DM). Human resource planning (HRM). Industrial engineering (IE). Managerial and organizational analyses (MOA). Finance and accounting (FA). Organizational audits (OA). Banking (BANK). Other. -20- 1 Arable, P., Bibliography Boorman, S., and Levitt, P Constructing Blockmodels: and Why. Journal of Mathematical Psychology, 17, 1 How (February 1978), 21-63 Asch, S.E. Studies of Independence and Conformity A Minority of 2 Against an Unanimous Majority,. Psychological Monographs, 70,9 (1956). One 3. Attewell, P. Microelectronics and Employment. Paper presented at the Conference on Microelectronics in Transition, University of California, Santa Cruz, 1983 (as reported in Attewell and Rule, 1984). 4. P. and Rule, J. Computing and Organizations: What We Know Don't Know. Communications of the ACM, 27, 12 (December 1984), Attewell, and What We 1184-1192. 5. Strategy: Bakos, J.Y. and Treacy, M.E. Information Technology and Corporate A Research Perspective. MIS Quarterly 10,2 (June 1986), 107-1 19. 6. Barley, S. The Professional, The Semi-Professional, and the Machines: The Social Ramifications of Computer Based Imaging in Radiology. PhD Dissertation, Management, Massachusetts Sloan School of Institute of Technology, 1984. Bjorn-Andersen, N., Eason, K., and Robey, D. Managing Computer Impact: International Study of Management and Organization. Norwood, NJ: Ablex, 7. An 1986. 8. Blau, P.M. Basic Books, 1971. 9. and Schoenherr, R. The Structure of Organizations. New York: CM., McKinley, W., and Tracy, P.K. Technology and Manufacturing. Administrative Science Quarterly, 21, 1 (March Blau, P.M., Falber, Organization in 1976), 20-40. 10. Boorman, S.A. and Levitt, P.R. Blockmodelling Complex Statutes: Mapping Techniques Based on Combinational Optimization for Analyzing Economic Legislation and its Stress Point Over Time. Economics Letters, 13, 1 (1983), 1-9. 11. Boorman, S. and White, H., "Social Structure from Multiple Networks Role Strurtures/'Amer/can./ourna/of'Soao/ogy, 81,6 (May 1976), 1384-1446. II. 12. Breiger, R.L., Boorman, S.A., and Arable, P. An Algorithm for Clustering Relational Data, With Applications To Social Network Analysis and Comparison with Multidimensional Scaling. Journal of Mathematical Psychology, 12, 3 (August 1975), 328-383. 13. Burt, Ronald. Structural Equivalence. 1335. Business 14. 1984, 118-142. Week. Cohesion versus Social Contagion and Innovation: American Journal of Sociology, 92, 6 (May 1987), 1287- Office Automation Restructures Business -21 - October 8, Computerization as a Predominate Technology: Its influence Academy of Management Journal. Carter, N. 15. on the Structure of Newspaper Organizations. 27, 2 (June 1984), 247-270. Crawford, A. B. Jr. Corporate Electronic Mail - A Communication-Intensive Application of Information Technology. MIS Quarterly 6,3 (September 1982), 1-13. 16. K. and Malone, T.W. Information Technology and Work MIT Center for Information Organization. Systems Research Working Paper No. 165, Crowston, 17. December, 1987. In Case Glenn, E.N. and Feldberg, R.L Proletarianizing Clerical Work. Monthly Review, 1979. on the Labor Process, A. Zimbalist(ed.). NewYork: Studies 18. DiMaggio, 19. Paul. A Structural Analysis of Organizational Fields: in Organizational Behavior, Vol. 8 (1986), 335-370. Blockmodel Approach. Research Downs, G.W. and Mohr, LB. Conceptual Issues in the Study of Innovation. 20. Administrative Science Quarterly, 21,4 (December 1976), 700-714. 21. Effects Foster, L W. and Flynn, D. M. Management Information Technology: Its on Organizational Form and Function. MIS Quarterly 8, 4 (December 1984), 229-235. Gregory, J. and Nussbaum, K. Race Against Time: Automation in the 22. Office, Technology, and People, 1,2-3 (September 1982), 197-236. Office. 23. Holland, Measurement W. and Leinhardt, P. Error in Sociometry. The Structural Implications of S. Journal of Mathematical Sociology, 3, 1 (July 1973), 385-411. Keen, P. G. W. MIS Research: Reference Disciplines and a Cumulative Tradition. Proceedings of the First International Conference on Information 24. Systems. E. R. McLean (ed.). December 8-10, 1980, 174-186. Kling, R. Social Analyses of Computing: Theoretical Perspectives in 25. Empirical Research. Computing Surveys, 12, 1 (March 1980), 61-110. 26. Leavitt, H. Business Review, 36, and Whisler, T. L Management 6 (November-December 1958), 41-48. J. in the 1980's. Recent Harvard Markus, M.L. and Robey, D.F. Information Technology and Organizational 27. Change: Conceptions of Causality in Theory and Research. Management Science, 34, 5 (May 1988), 583-598. Matteis, P. The New Back Office Focuses on Customer Service. 28. Business Review, 57, 2 (March/April, 1979), 146-159. Mumford, E. and Banks, O. 29. Routledgeand Kegan Paul, 1967. 30. Nelson, Reed Organizational Structure: Elliott. The Computer and the The Use of Blockmodeling A Methodological (1986). 75-85. -22- Proposal. Clerk. Harvard London: in the Study of Organization Studies, 7, 1 Noble, D Social Change m Machine Design: The Case of Automatically 31. Controlled Machine Tools, and a Challenge for Labor. Politics and Society, 8, 3/4 (1978), 313-347 Pfeffer, J. and Leblebici, H. Information Technology 32. Structure. Pacific Sociological Review, 20 (1977), 241-261 and Organizational . Robey, D F Computer Information Systems and Organization 33. Communications of the ACM, 24, 1 (October 1 981 ), 679-687. 34. Equivalence: Sailer, L.D. Structural and Application. Social Networks, Schrank, 35. Press, 1983. Robert (ed.). 1, 1 Meaning and Definition, Structure. Computation (August 1978), 73-90. Industrial Democracy at Sea. Cambridge, MA: The MIT 36. Trice, AW. and Treacy, ME. Research, (forthcoming 37. Turkle, S. in Database) Computer Utilization as a Dependent Variable in MIS . as Rorschach. Society, 17, 2 (January/February 1980), 15-24. Turner, J. A. Improving the Quality of Information Systems Research. 38. Proceedings of the First International Conference on Information Systems, E. R. McLean (ed), December 8-10, 1980, 91-97. Turner, J. A. Computer Mediated Work: 39. The Interplay between Technology and Structured Jobs. Communications of the ACM, 24, 12 (December 1984), 1210-1217. 40. Walker, G. Network Position and Cognition in a Computer Software Firm. 1 (March 1985), 103-130. Administrative Science Quarterly, 30, 41. Wallace, M. and Kalleberg, A. Industrial Transformation and the Decline of Craft: The Decomposition of Skill in the Printing Industry, 1931-1978. American Sociological Review, 47, 3 (June 1982), 307-324. 42. Whisler, T.L. The Impact of Computers on Organizations. Praeger, 1970. 43. Networks New York: White, H., Boorman, S. and Breiger, R. Social Structure from Multiple Blockmodels of Roles and Positions. American Journal of Sociology, 81, I. 4 (January 1976), 730-780. -23 Date Due i.ii. : MIT 3 LIBRARIES DUPl TOflO I 0DSb7272 ~>