NCGIA National Center for Geographic Information and Analysis Testing Technology Transfer Hypotheses in GIS Environments Using a

Although they are perhaps the most commonly-used and popular research methods, case studies and other qualitative forms of social science research have long been criticized for their lack of generalizability to the larger population and lack of sampling controls. These criticisms may be aptly addressed and solutions constructed by evaluating the rules of scientific method within case research environments. By using more logically consistent, rigorous, and systematic approaches, some of the shortcomings of case study methods may be overcome. This article draws from the management information systems (MIS) and organization behavior (OB) literature to make some suggestions on how to conduct and evaluate GIS case study research. It reviews the requirements of natural science research models, particularly as described by Lee (1989), and provides examples of how the substance of those requirements may be met in the context of GIS case studies. Introduction Case study methodologies have been suggested within the GIS community as appropriate for researching a range of GIS implementation, utilization, and diffusion issues (Zwart 1986, Niemann et. al. 1988, NCGIA 1989, Craig 1989, Azad 1990). These issues include identifying the determinants of adoption outcomes; isolating critical adoption factors and processes for particular classes of users; investigating the stages at which change agent, opinion leader, and champion influences are most critical; assessing use success; determining levels in the organizational structure where GIS products are used and to what extent; identifying the forms of decision making which have utilized GIS; identifying factors and processes leading to rejection of previously embraced GIS; and identifying organizational and societal consequences of GIS (Onsrud and Pinto, 1991). Case studies examine phenomena in their natural settings and typically involve collection of data by several different means from a range of sources. When used as the sole research heuristic device, case studies have been criticized for their limitations in terms of generalizability to the larger population and lack of sampling controls (Piore 1979, Bariff and Ginzberg 1982, Bonoma 1985). It is generally acknowledged in the social science research community that no single research methodology is most appropriate for all research applications (Williams, Rice & Rogers 1988). In addition, it is generally agreed that using multiple forms of research to investigate an issue leads to better and more reliable results than using a single methodology (Yin and Heald 1975, Cook and Campbell 1979, McClintock, Brannon & Maynard-Moody 1979, Kaplan and Duchon 1988). Case studies are often included and occupy a lead position in the suite of methods used by researchers to evaluate intervention strategies within organizations. However, there is a need to clarify the explicit methodological means by which case studies within GIS application environments should be carried out. For purposes of this paper, a case study is an examination of a phenomena in which the primary purpose of the observer has been to carry out research rather than to implement a system or improve an operational environment. That is, since the techniques suggested in this article are intended for individuals testing theories relating to the efficacy of intervention strategies, they will be less useful to practitioners who are implementing systems. Of course, the overall intent of investigating various case study techniques is to aid researchers in building a relevant body of knowledge which eventually will aid practitioners in their system implementation and improvement efforts. Let us assume that the general GIS practitioner community is confronted with an implementation issue in which the most appropriate intervention strategies to use in addressing that issue are difficult to determine. For instance, organizations throughout the general GIS community currently are involved in determining which processes they should follow and which factors they should consider to ensure that their GIS systems are used to their greatest benefit over time. Another example is determining which policies 1 This article is based upon work partially supported by the National Science Foundation under Grant No. SES-88-10917. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. agencies should pursue in regard to public access to their GIS database and which organizational and legal tools should be used in carrying out those policies. When confronted with these and similar problems, the practitioner community engages in trial and error processes in attempting to find out "what works." The brainstorming and testing engaged in by the practitioner community results in a substantial body of valuable knowledge. This knowledge base of successes and dead ends is communicated by various methods throughout and among the community. However, as the experience base grows and system implementations become more diverse and complex, large numbers of conflicting and competing messages may be received on what works or does not. At this point in the diffusion of a technology (i.e. when the appropriate decision routes in addressing an issue are no longer intuitively obvious), social science researchers can play an important role in providing direction to the user community. Consequently, there is a need for social science researchers to construct research questions from the existing experience base and to develop a body of research that builds up falsifiable hypotheses and tests them through rigorous methods following the nature of scientific canons. Utility of Case Study Approaches The traditional phases of accruing knowledge within "learned" settings are often expressed as exploration, hypothesis generation, and hypothesis testing (Glaser & Strauss, 1967). In the exploration phase, researchers descriptively study how organizations have dealt with the constraints imposed upon them. These knowledge capture studies form the basis for developing theories regarding phenomena and for hypothesizing prescriptive strategies. After deriving the theories and generating hypotheses in support of them, the researcher proceeds to the hypothesis testing phase. Conventional thinking in the MIS literature indicates that case study approaches are highly appropriate for the exploration and hypothesis generation phases but are generally ill-suited to the hypothesis testing phase (Roethlisberger 1977, Bonoma 1983, Benbasat 1984). The argument made is that, although disconfirmation of a hypothesis might be shown by a single case, reasonable confirmation of a hypothesis requires analytic deductive testing of a representative and substantial sample (Benbasat, Goldstein & Mead 1987). By this reasoning, data must be gathered in a form suitable for quantitative processing and must be gathered for a significant number of cases to evidence confirmation of a hypothesis reliably (Bariff and Ginzberg 1982, Dickinson, Benbasat & King 1982, Kauber 1986). A recent article by Lee (1989) challenges the conventional wisdom with regard to the use of case studies in the hypothesis testing stage. He argues that the data or results generated from case studies need not be quantitative, statistical, or mathematical in order to be analytically rigorous or "scientific." We believe the application of his approach to the evaluation of GIS implementations could provide a useful and practical means for the GIS community to better isolate those factors and processes which are critical for inclusion in prescriptive implementation and improvement strategies. Croswell (1989) lists numerous obstacles to successful GIS implementation and categorizes them into eleven major groups: apathy/fear of change; funding availability or justification; planning/management support; organization coordination and conflicts; training/understanding of technology-, staffing availability/recruitment; software complexity/maturity of technology; data communication/networking; data structure and source materials; data and software standards/data integration; and miscellaneous. The obstacles listed in each group were identified from a wide range of GIS literature extending from 1985 through 1988. From the diffusion of innovations literature and GIS tracer case studies, Onsrud and Pinto (1991) developed a similar list of factors potentially critical to successful adoption of GIS. Typical questions which arise in regard to such lists include the following: Which of the factors cited may be most critical for a particular class of users or for particular forms of organizational structure? To what extent is it necessary to consider each item? Are there items on the list which might be "perpetuated myth" rather than actual requirements for success? Are there items on the list which may improve the likelihood of initial acquisition but be counterproductive to long term success or vice versa? Appropriate application of case study approaches could aid in finding the answers to these and similar questions. Lee’s Scientific Case Study Methodology "Few propositions in science are directly verifiable as true and none of the important ones are" (Copi 1986 as reported in Lee 1989). Thus, most theories, whether in the natural or social sciences, may be tested only indirectly. The indirect testing of a theory consists of deriving one or more suppositions capable of being tested directly from the theory and then comparing the actual outcomes against those which were predicted. In addition to making this comparison, the scientist must also provide evidence that the outcomes are attributable to the supposition being tested and not to other causes. Lee argues that a single case study meets the requirements of scientific method if it adequately addresses four methodological problems; making controlled observations, making controlled deductions, allowing for replicability, and allowing for generalizability. The discussion which follows draws heavily from Lee (1989) and, with the exception of the additional references noted, paraphrases extensively from his work. In natural science research models, controlled observations are typically achieved by regulating the test environment through laboratory controls (e.g. control groups along with treatment groups) or statistical controls. Neither of these controls is readily available to the GIS case study researcher. GIS implementation issues by their very nature must be observed in real world settings and therefore setting up or finding a case study control group for comparison purposes is very difficult. Similarly, the study of a single case generates more variables than data points and as such is inappropriate for evaluation through statistical methods. A GIS case study must necessarily identify and take advantage of natural controls. Glaser and Strauss’s (1967) study of "grounded theory" research argues for a comparative model where practicable. Their approach suggests that the case study method should be supplemented by the selection of another, comparable research site. These two case sites should possess similar goals (e.g., expanding the use of GIS in the organization) but have dissimilar characteristics in other aspects. For example, a comparative model to test the effects of top management support for expanded use of GIS would select two similar municipalities: one with strong top management support for GIS and another with little support. It could then be argued that the use of a comparative model has allowed the researcher to set up a form of controlled observation allowing for generalizable theory development. (See Meckstroth (1975) and Lijphart (1975) for more precise definitions and refinements of the comparative model.) Deductive reasoning is typically controlled for in natural science models through mathematical propositions. Since GIS case studies are qualitative in nature, the convenience of mathematical logic and mathematical tools are unavailable to us. However, we are not precluded from using the formal rules of logic from which those mathematical tools derive. "Mathematics is a subset of formal logic, not vice versa." (Lee 1989, p. 40) The expression of logic through verbal propositions is certainly possible and seems a n appropriate and valid alternative for making controlled deductions within GIS use environments. Ragin’s (1987) pioneering work in application of Boolean logic to qualitative comparative method and his subsequent development of a computer-based package for such analysis has proven beyond doubt that the use of logic in case study research is attainable and appropriate. Replicability is controlled for in natural science models by testing an observed phenomena numerous times under the same set of conditions to determine whether the results remain constant. In evaluating GIS implementations the same set of technical, organizational, and social conditions are unlikely ever to be replicated in natural settings. However, it is possible to carry a constant theory from one case study to another. The concept of replicability can be satisfied in GIS case study work by testing the same theory in new environments. Switching environments to test the same theory may involve deriving new predictive outcomes for the testable hypothesis or may even involve completely new testable hypotheses. Regardless, the new observations will either continue to support the theory or falsify it. Thus, GIS researchers may replicate the results of their peers through testing of the same theory even though conditions under which the theory is tested will change; that is, their results in a variety of environments will show that either the theory continues to be supported or is falsified. Allowing for generalizability of research results is little different in a GIS case study environment than in a natural science environment. Generalizability is a quality measure which indicates the extent to which a theory is applicable to a range of settings. In a natural science setting, one test or experiment under a single set of circumstances is not generalizable. Only when a theory has been repeatedly tested across a range of circumstances and not falsified can it be reasonably said that the theory is generalizable through that range of circumstances. Even though a case study may meet the four requirements of scientific methodology, Lee suggests that a case study investigation remains inadequate without inclusion of an evaluation of the extent of its analytic rigor. He suggests that the extent of analytic rigor may be evaluated by addressing the questions which are paraphrased and summarized in Table 1. Rather than discussing the concepts and reasoning behind the questions, we illustrate by example in the next section how GIS implementation theories might be tested using a case study approach. For those desiring a more indepth discussion of underlying concepts, we commend the Lee (1989) article to you. Application of Scientific Methodology to a Hypothetical GIS Case Study Let us presume that XYZ Corporation is a large company generating most of its revenue from the sale of lumber and other forest products from its extensive land holdings. XYZ Corporation acquired GIS technological capabilities several years ago but use of the system has been largely limited to its automated mapping capabilities. Management seldom calls upon the analysis capabilities of the system. Personnel in other divisions of the corporation occasionally request hard copy and digital products from the GIS Division but in general other divisions have been reluctant to use products generated from the GIS or to contribute data to the system. By referring to the "lessons learned" in case histories documented in the GIS literature (e.g. URISA Proceedings) and MIS literature, numerous theories may be postulated in an attempt to explain why a GIS has not been used more extensively within a particular organizational setting. A small sample of the competing theories we might consider in this single case are as follows: 1) Relative Advantage Theory: If the relative advantages of using the GIS over the procedure or system the GIS replaces is small, even though benefits to the overall organization might be great, the GIS’s intended users will resist its use. 2) System Complexity Theory: If the complexity in using a GIS over the procedures or system the GIS replaces is great, even though advantages to individuals, groups, or the organization as a whole may be substantial, the GIS’s intended users will resist its use. 3) Communication Networking Theory: If the degree of interpersonal communication networking in the organization is low (i.e. few informal and formal opportunities to talk with others at all levels and across divisions), resistance to expansions of use or improvements in use of the GIS will be great. 4) Interaction Theory: If the patterns of information flow and control prescribed by using the GIS are incompatible with the information control patterns prescribed by the organizational structure, those individuals and groups losing power in the organization will resist its use. Each of the above theories has been phrased in terms of resistance to GIS use or improvement efforts. The first two theories relate primarily to characteristics of the technology, the third theory relates to a characteristic of the user organization, and the fourth theory focuses on the relationship between the technology and the organization. Numerous additional theories attempting to explain why technologies are being underutilized are suggested in the literature (e.g. Greer 1981, Wellar 1988, Croswell 1989, Obermeyer 1990, Onsrud & Pinto 1991) but those stated here are merely a sample for illustrative purposes. Each of the theories involve phenomena which are very difficult to observe. Nebulous factors upon which the theories are dependent include "... human nature, cognitive styles, personality traits, factors internal to the person, user friendliness, technical deficiency in (the) system, ergonomics, horizontal and vertical dimensions of organizational power, power in terms of symbolism, and ... the distribution of power implied in the design of (the) information system" (Lee 1989, p.37). As a result, none of the above theories can be directly verified as being true. From each of the theories, however, we may predict outcomes for various scenarios which are directly observable and which, if they are observed, will either imply that the theory is true or will evidence the theory as false. (See Table 2) One prediction leading from the "relative advantage theory" is that providing increased advantages to the intended users of the GIS in those divisions which are not using it will reduce resistance to its use. For instance, management has determined that incorporating the thousands of survey measurements made by its surveying division into its GIS database each year would be extremely beneficial to the long term efficacy of the corporate database. However, a case study researcher observes that personnel in the surveying division of the corporation continued to resist all efforts to contribute to or use data from the GIS until an automated survey measurement management system was installed. This system allowed surveyors to carry out their typical daily tasks much more efficiently and provided them with capabilities they never had before, such as the ability to automatically detect and isolate blunders in their work and the ability to efficiently incorporate GPS observations in their adjustment networks. The system also allowed convenient file transfer back and forth with the GIS. The observation that resistance to working with the GIS by the surveyors was greatly decreased by providing means to do their jobs more effectively lends support to the "relative advantage theory." Of course, other theories might just as reasonably explain the overcoming of the resistance. At this point we can say only that the particular theory being evaluated has not yet been falsified. Note that the researcher here took advantage of natural controls. No personnel or organizational changes occurred during this time. This tends to suggest that competing theories involving organizational factors or people factors were not the cause of decrease in resistance by this group of individuals. In GIS case study observations, there is a need to formally analyze the causes of changed conditions as well as analyze continued resistance despite changed conditions. A "system complexity theory" might also be evaluated through the facts presented. The automated measurement management system now being used by the surveying division in conjunction with providing data to the GIS is far easier to use in several ways than the previous software and procedures. One prediction from the system complexity theory is that improving the user friendliness of the surveying measurement interface with the GIS would decrease resistance to use of the GIS. Thus the system complexity theory is also supported by the facts presented. Both the relative advantage theory and the system complexity theory appear to be at least as explanatory as the other. The observed scenario fails to distinguish which of the theories, if either, is most critical in explaining the change in resistance. However, both theories have passed attempts made at their falsification and knowledge has been gained in the context of a formal conceptual framework. Let us assume that the case study researcher observes also that personnel within the corporation’s pulpwood division are often requested to provide yield estimates for various land parcels. For a typical request, personnel can provide within an hour’s notice an accurate estimate of the pulpwood yield for a typical parcel through use of conventional maps, airphotos, and a hand held calculator. Estimates with approximately the same accuracy and requiring approximately the same lead time can be made using far more complex procedures by one of the division’s employees trained to use the GIS. The system complexity theory predicts that employees within the pulpwood division will resist use of the GIS. This prediction is compared with that which is actually observed. In this instance, the GIS is almost always chosen to make such estimates and the prediction is falsified. This result supports the contention that the underlying theory upon which the prediction is based can not be generalized to the circumstance observed. Regardless of whether the prediction holds true or false under the facts stated, the researcher has eliminated many other potential explanations for the results by searching out an instance where the results achieved by using the new procedures were functionally equivalent to those using the old procedures. That is, the results cannot be explained by theories dependent on changes in personnel, alterations in management structure, or changes in communication patterns because these all remained constant in the setting observed. By using this "natural control" situation, the researcher provides evidence, although not conclusive, that the observed phenomena is due to the proposition being tested and not attributable to other factors. Through the methodology outlined, the case study researcher explores one theory after another by generating one or several predictions for each theory and then actively searching for natural control situations to test those predictions within the organization being studied. The study of a single organization might provide results similar to those summarized in Table 2. The table shows that in this particular case study four theories were evaluated. In testing the third theory, a situation was observed in which the disease control manager’s resistance to cooperating with the new innovation could not be explained by the communication networking theory. The observer was unable to isolate control conditions to explain the resistance of this particular manager or of the personnel within his division such that predictions relative to the other three theories could be tested. However, if we consider the four theories to be competing with each other in explaining the cause of resistance to greater use of the GIS throughout the organization, the overall results of the case study suggest that evaluation of the relative advantage theory or the interaction theory may be more fruitful in developing a successful intervention strategy for lessening resistance by the disease control manager and his personnel than providing yet more formal and informal opportunities for communications by that division with others in the organization. Analyzing the Rigor of the Case Study The extent to which the case study described satisfies the four requirements of scientific methodology may be analyzed by evaluating the degree of confirmation of each theory addressed. Confirming a theory by successfully testing it against several of its predictors is more rigorous than testing it against a single predictor. Thus, for this single case, the "interaction theory" is more rigorously supported than any of the others because the observer was able to identify five control conditions in which predictions could be tested and all five observations supported the theory. Although the "relative advantage theory" was also supported by the case study, it was tested against only one of its predictors. The case study supplied at least some evidence of falsification of the other two theories but likewise the limited number of predictions against which those theories were tested makes that evidence less rigorous than it might be. The degree of logical consistency of a theory is also improved by increasing the number of predictors for the theory. Logical consistency requires that a theory must be tested against at least two independent predictors so that their outcomes may be compared with each other for consistency. Thus, in our example, minimal rigor in logical consistency in testing the "relative advantage theory" and the "communication networking theory" has not yet been supplied. That rigor may be supplied by testing the same theories in other case study settings. Minimal requirements for testing the “system complexity theory" have been supplied but the results are inconsistent as shown. Therefore, that theory fails the logical consistency test. However, the generalizability of the conclusion that the theory has been falsified is highly questionable until the same theory has been tested repeatedly in other case study settings. The "interaction theory" is the most rigorously supported in its logical consistency because it was tested against the largest number of predictors and the five outcomes were entirely consistent with each other. Contending that one theory is more predictive of outcome than competing theories may be more rigorously supported by increasing the number of theories against which the theory of interest is tested. In the illustrated case study a total of four theories were tested. A few other rival theories were considered but left untested because no control situations could be found within the organizational setting with which to test those theories. Other theories were unconsidered because they were not hypothesized prior to the study or were left untested because of limited resources for the study. The requirement that a theory must survive attempts at its falsification may be more rigorously supported by increasing the number of predictions through which the theory may be proven wrong. Thus, increasing the number of predictions for a theory also increases its degree of falsifiability. Another way of viewing the scientific rigor of a case study is to evaluate the degrees of freedom encompassed by the study (Campbell, 1975). In statistics, the number of independent random variables that constitute a statistic indicates the degrees of freedom. The larger the degrees of freedom, the more statistically reliable the results. If we make an analogy with the case study approach outlined, the degrees of freedom in testing a theory can be evaluated in terms of the number of predictions, the number of organizational settings, and the number of rival theories against which the theory of interest has been tested. High numbers in each of these three categories suggest more reliable confirmation of a theory than lower numbers. Table 3 shows the degrees of freedom for each of the four theories evaluated in our single hypothetical study. Although it is inappropriate to add the numeric values shown in the table across columns, the degrees of freedom approach may be a convenient and appropriate means of compiling the results of numerous independent case studies and a convenient means of gauging the reliability of behavioral and institutional theories against competing theories. As shown in Table 3, our hypothetical example has focused on a single organizational setting. Reliance on a sole data source confounds the evaluation or compensation for observation effects and biases by the investigator. This threat to internal validity may be substantially reduced by evaluating a cluster of cases. McClintock et. al. (1979) describe the "case cluster" concept. Since the cluster approach highly supports and complements Lee's approach, case study designs should comport with its underlying precepts if atte mpts will be made at extracting generalizable principles from the resulting case data. It is important to note in Table 3 that the observer did not "invent" five predictions for the interaction theory and only one or two for the others. The observer thoroughly searched for "control situations" within the organization and was able to find only the number indicated for each theory. To increase the number of predictions for each theory, the researcher would need to observe additional settings since presumably all control situations in the organization investigated already have been exposed. One issue relative to generalizability not addressed by the research procedure outlined is how a researcher might be able to identify the bounds within which a theory applies. As stated previously, a theory cannot be considered generalizable through a range of circumstances unless it has been tested across that range and not been falsified. This, however, requires gathering consistent background information for all cases which would be used later in comparative analyses (e.g. multi-function vs. single function systems, low budget vs. large budget organizations, engineering vs. planning applications, transactional processing vs. project focused systems, etc.). The case study methodology outlined above does not address this issue because it is the intent of this research process to test the validity of generally applicable theories rather than those which are constrained to narrow circumstances. In addition, we contend that although case study research can provide hints and evidence of the limits of applicability of a theory, large scale surveys and multivariate statistical analysis are far more suited for accomplishing comparative analyses than are case studies. Surveys make possible the testing of much larger samples. Quantifying Benefits and Assessing Impacts In attempting to maximize the use of GIS within an organization, it might be argued that it is necessary to measure or otherwise gauge the extent of use in the organization in order to determine when the maximum has been reached. Although quantifying performance, use, and the benefits of use are important research areas (e.g. Goodchild & Rizzo 1987, Raymond 1987, Brown & Friedli 1988, Dickinson & Calkins 1988, Rivard & Huff 1988, Igbaria & Nachman 1990), they are not the subjects of the research suggested by this article. The set of problems addressed by this article and their relation to quantification issues may be illustrated by example. Let us presume an organization has already acquired GIS operational capabilities. The executive for that organization is interested in maximizing use of the GIS by having the five divisions within the organization all contribute data to the GIS and share the data and products generated from it. The executive knows from experience and observation of successful GIS operations (or, alternatively, from a quantified analysis of projected operations) that if these five groups begin to cooperate and work with each other, substantial and numerous benefits will occur. She is little concerned with explicitly identifying or quantifying those benefits, whether tangible or intangible, because she is already convinced benefits will accrue (or, alternatively, time and motion studies or some other quantitative approach has already projected positive results). The executive’s primary interest is in getting the five divisions of the organization to cooperate with each other and support the common GIS. In accomplishing this goal she needs to know which organizational and behavioral theories are relevant to her situation, which intervention strategies are likely to work, and on which strategies and theories should she focus her efforts (e.g. In order to improve and expand uses of the GIS in her organization should she currently concentrate on relative advantage theory issues, system complexity theory issues, some other set of issues, or what combination of issues?) Thus, the research suggested by this article focuses on understanding the applicability of competing theories of individual and organizational behavior in GIS environments. Although it is a related research need, quantifying the benefits and impacts of GIS should not be construed as a goal of the research methods set forth. In testing competing theories by the processes outlined in this article, the observer looks for changes in behavior or looks for continued behavior despite attempts at intervention. Because the observer is not attempting to "measure" anything, the very substantial problem of finding a common unit for measuring benefits or other effects is not a major issue in the theory testing procedures outlined. Conclusions Numerous maxims have been suggested regarding what allows institutions to acquire, implement, and successfully employ GIS. The problem for researchers and practitioners is that many of these maxims either contradict each other or are so situation-specific as to be nongeneralizable to a larger population. As a result they may be of little use in providing practical guidance for gaining a greater understanding of the issues involved in successful GIS implementation and use. In the GIS community to date, we have seen very few theory-focused empirical studies on issues of acquisition and use of GIS. Rather, the literature is more typically dominated by ad hoc application descriptions from past successes and failures. In other words, many of the studies reported in the literature could be more properly labeled "case histories" rather than case studies. The distinction between the two approaches is profound. Case histories are reported in a post hoc manner, after the effects of the GIS implementation have made themselves felt. The result is often an unconscious bias (due to the effect of time or outcome) interjected into the researcher’s report. A scientific approach to case studies, however, offers tremendous challenges as well as the potential for impressive results. Scientific theory must be falsifiable, logically consistent, be at least as explanatory as any competing theory, and the theory, while falsifiable, must survive attempts made at its falsification. The criteria for scientific application of case study analysis are rigorous. Adhering to the criteria can, however, result in better-grounded theory as well as greater confidence on the part of other researchers and practitioners that the results obtained are supportable and can be used as the basis for additional research and theory-building. Multivariate statistical analyses are likely to remain a predominant heuristic in confirming theories because they make use of larger data bases, employ more rigorous controls, and lessen the likelihood of reporting significant findings where none actually exist (Kraemer and Thiemann 1987, Lipsey 1990). However, case studies can make an important contribution to the research process in a number of ways. First, case studies are an extremely useful tool for early theory development. They offer the researcher the opportunity to explore phenomena of interest in their natural setting, subject to the wide variety of organizational and behavioral factors that can have an impact on desired outcomes. As a second benefit, case studies can provide a "richness" to the research process that is impossible to find in most quantitative methodologies. Survey research is generally conducted in an artificial environment, often through mailings of questionnaires to a pre-selected sample. The benefit of case studies is that they allow researchers to gain insight into the organizational environment within which the study is conducted. Further, they offer the advantage of providing additional qualitative information to complement the statistical analysis. Finally, as case study research methodologies gain in sophistication and popularity in the GIS community, the compilation of such studies in a consistent fashion will have the effect of confirming or falsifying theories over time. In other words, as we build up a data base of case studies using a consistent research approach, meta analysis will allow researchers to develop additional theory based on the contributions of each individual case example. To conclude, we would offer an invitation to other academics and researchers within the GIS community to test the methodological framework set forth in this paper in one or more active GIS use settings. Until we are able to develop a unified method for case analysis, many of the studies reported will remain of minimal use to the field due to the highly idiosyncratic nature of the methodologies employed. It is our own goal to begin using the methodological framework set forth in this article in evaluating a series of GIS use settings during the second half of this year. By compiling the case study work of a variety of researchers in a consistent manner, we believe a lasting contribution can be made to the profession. Acknowledgements Many of the foundation ideas contained in this article are drawn from Lee, 1989. References Argyris, C., 1980, Inner Contradictions of Rigorous Research, N.Y.: Academy Press. Azad, B., 1990, Implementation of Geographic Information Systems, PhD Proposal, Department of Urban Studies and Planning, MIT, forthcoming. Bariff, M.L. and M.J.Ginzberg, 1982, MIS and the Behavioral Sciences: Research Patterns and Prescriptions. Database, vol 14 no 1, pp. 19-26. Benbasat, 1., 1984, "An Analysis of Research Methodologies" in The Information Systems Research Challenge, F. Warren McFarlen (ed.), Boston: Harvard Business School Press, pp. 47-85. Benbasat, I., D.K. Goldstein, and M. Mead, 1987, The case research strategy in studies of information systems. MIS Quarterly, September, pp. 369-386. Bonoma, T.V., 1983, A Case Study in Case Research: Marketing Implementation, Working Paper 9-585-142, Harvard University Graduate School of Business Administration, Boston Bonoma, T.V., 1985, Case research in marketing: opportunities, problems, and a process. Journal of Marketing Research, vol. XXII, pp. 199-208. Brown, P. and Friedli, D., 1988, Assessing Organizational Preparedness for a Comprehensive Distributed LIS/GIS. Proceedings of the Urban and Regional Information Systems Association, Los Angeles (Washington D.C.: Urban & Regional Information Systems Association) vol. 3, pp. 22-36. Campbell, D.T., 1975, "Degrees of Freedom" and the Case Study. Comparative Political Studies, vol. 8, no. 2, pp. 178-193. Cook, T.D. and Campbell, D.T., 1979, Quasi-Experimentation: Design and Analysis Issues for Field Settings, Boston: Houghton Mifflin. Copi, 1., 1986, Introduction to Logic. New York: Macmillan Craig W.J., 1989, URISA’s research agenda and the NCGIA. Journal of the Urban and Regional Information Systems Association, vol. 1, pp.7-16. Croswell, P.L., 1989, Facing reality in GIS implementation: lessons learned and obstacles to be overcome. Proceedings of the Urban and Regional Information Systems Association, Boston,MA (Washington D.C.: Urban & Regional Information Systems Association) vol. 4, pp. 15-35. (Revised and reprinted in Journal of the Urban and Regional Information Systems Association, 1991, vol. 3, no. 1, pp. 43-56.) Dickinson G.W., 1. Benbasat and W.R. King, 1982, The MIS Area: Problems, Challenges, and Opportunities. Database, vol 14, no 1, pp. 7-12. Dickinson, H.J. and H. Calkins, 1988, The economic evaluation of implementing a GIS. International Journal of Geographic Information Systems, vol. 4, no. 4, pp. 307-327. Glaser, B.G. and A.L. Strauss, 1967, The Discovery of Grounded Theory: Strategies for Qualitative Research. New York: Aldine Goodchild, M.F. and B.R. Rizzo, Performance evaluation and work-load estimation for geographic information systems. International Journal of Geographic Information Systems, vol. 1, no. 1, pp. 67-76. Greer, A.L., 198 1, Medical technology: assessment, adoption, and utilization. Journal of Medical Systems, vol. 5, pp. 129-145. Igbaria, M. and S.A. Nachman, 1990, Correlates of user satisfaction with end user computing: an exploratory study. Information and Management, vol. 19, pp. 7382. Ives, B., M.H. Olson and J.J. Baroudi, 1983, The measurement of user information satisfaction. Communications of the ACM, vol. 26, pp. 785-793. Kaplan, B. and D. Duchon, 1988, Combining qualitative and quantitative methods in information systems research: a case study. MIS Quarterly, December, pp. 57 1 -586. Kauber, P., 1986, What’s Wrong With a Science of MIS? Proceedings of the 1986 Decision Science Institute, Honolulu, HA, pp. 572-574. Kraemer, H.C. and S. Thiemann, 1987, How many Subjects? Statistical Power Analysis in Research. Newbury Park, CA: Sage Publications. Lee, A.S., 1989, A scientific methodology for MIS case studies. MIS Quarterly, March, pp. 33-50. Leonard-Barton, D. and E.M. Rogers, 1981, Horizontal diffusion of innovations: an alternative paradigm to the classical diffusion model. Sloan School of Management, MIT. Lijphart, A., 1975, The ComparableCases Strategy in Comparative Research. Comparative Political Studies, vol 8, no. 2 (July), pp. 158-177. Lipsey, M.W., 1990, Design Sensitivity: Statistical Power for Experimental Research. Newbury Park, CA: Sage Publications. Mayo, J.S., 1985, The evolution of information technologies, in Information Technologies and Social Transformation. Washington D.C.: National Academy of Engineering, National Academy Press. McCardle, K.F., 1985, Information acquisition and the adoption of new technology. Management Science, vol 3 1, pp. 1372-1389. McClintock, C.C., D. Brannon, and S. Maynard-Moody, 1979, Applying the Logic of Sample Surveys to Qualitative Case Studies: The Case Cluster Method reprinted in J. Van Mannen (ed.), 1983, Qualitative Methodology. Newbury Park, CA: Sage Publications, pp. 149-178. Meckstroth, T.W., 1975, "Most Different Systems" and "Most Similar Systems": A Study in the Logic of Comparative Inquiry. Comparative Political Studies, vol 8, no. 2 (July), pp. 132-157. Niemann, B.J., S.J. Ventura, D.D. Moyer, J. Licht, W.A. Chatterton, and H.M. Jacobs, 1988, Research needs: the interaction of land and geographic information system technology and society. Proceedings of the 26th Annual Conference of the Urban and Regional Information Systems Association, Los Angeles (Washington D.C.: Urban & Regional Information Systems Association) vol 1, pp. 1-16. National Center for Geographic Information and Analysis (NCGIA), 1989, The Research Plan of the NCGIA, International Journal of Geographic Information Systems, vol 3, no 2, pp. 117-136 Obermeyer, N.J., 1990, Bureaucratic Factors in the Adoption of GIS by Public Organizations, Computer Environment and Urban Systems, vol. 14, no. 3. Onsrud, H.J. and J.K. Pinto, 1991, Diffusion of Geographic Information Innovations, International Journal of Geographic Information Systems, forthcoming. Ragin, C.C., 1987, The Comparative Method: Moving Beyond Qualitative and Quantitative Strategies. Berkeley: University of California Press. Raymond, L., 1987, Validating and applying user satisfaction as a measure of MIS success in small organizations. Information and Management, vol. 12, pp. 173179. Rivard, S. and S.L. Huff, 1988, Factors of success for end-user computing. Communications of the ACM, vol. 31, pp. 552-561. Roethlisberger, F.J., 1977, The Elusive Phenomena, Harvard Business School, Division of Research, Boston Strauss, A., 1987, Qualitative Analysis for Social Scientists. N.Y.: Cambridge University Press. Wellar, B., 1988, Institutional maxims and conditions for needs-sensitive information systems and services in local governments. Proceedings of the 26th Annual Conference of the Urban and Regional Information Systems Association, Los Angeles, California (Washington D.C.: Urban & Regional Information Systems Association) vol. 4, pp. 371-378. Williams, F., R.E. Rice, and E.M. Rogers, 1988, Research Methods and the New Media (New York: The Free Press). Yin, R.K. and K.A. Heald, 1975, Using the Case Survey Method to Analyze Policy Studies, Administrative Science Quarterly, vol 20, pp. 371-38 1. Zaltman, G. and R. Duncan, 1977, Strategies for Planned Change (New York: John Wiley & Sons). Zwart, P., 1986, User requirements in land information system design some research issues. Surveying and Mapping, vol. 46, pp. 123-130. Chapter Three Call to Participate in Testing Thirty Technology Transfer Theories: A GIS Case Study Research Project Theories to be Tested The following pages contain a sampling of technology transfer lessons, hypotheses, and theories derived from the GIS, MIS, and organization theory literature. You are invited to test these hypotheses within operational GIS settings. Through your case study, attempt to isolate reasonable control situations which tend to evidence support or falsification of any or all of the following theories. No attempt has been made to arrange the theories in any order. The proposed case studies to be carried out under this project involve observations of ongoing GIS operations. Therefore, we have tried to incorporate theories in the following list which may be more relevant to improving and expanding the uses of GIS in an organization than to initially acquiring and implementing a GIS. Some of the hypotheses in the list of thirty address broad theories while others test rather narrow concepts. Many of the theories overlap and conflict with each other and some might be considered subsets of others. Many of the hypotheses reflect "lessons" or "rules of thumb" expressed in the literature and do not cover concepts of sufficient breadth to be labeled as "theories" in a formal sense. For instance, the models developed by McCardle (1985) or Mayo (1985) to explain an organization’s decision to adopt an innovation are more appropriately classified as "theories" than most of the statements on the list which follows (see previous chapter for references). Regardless, the term "theory" is used for all items in our listing for convenience. Through the process of testing the full range of these and similar hypothesized statements, the ultimate goal is to derive broad-based theories which are supportable and useful. Each hypothesized theory on the list has been worded in terms of "resistance" to GIS use. We chose to do this because in practice, after a GIS has been initially implemented in an organization, there is typically a need to identify and address the additional impediments which are keeping the GIS from reaching its maximum use throughout the organization. A consistent focus on resistance for every hypothesis should also ease later comparative analyses. However, most of the hypotheses listed also may be reworded readily in a positive sense. For instance, although "lack of strong executive and management support is a likely impediment to expanded use of the GIS in an organization," a reciprocally-related hypothesis would state that "strong management support acts as an incentive to the intended users of the GIS to overcome interpersonal and organizational impediments to enable their expanded use of the GIS." In selecting items for inclusion in the following list, we also tried to keep a focus on potential technological, organizational, and personal impediments without attempting to speculate on the most appropriate means for overcoming those impediments. Strategies for overcoming verified impediments (e.g. altering the technology, organization structures, management arrangements, implementation processes, and personal characteristics) are thus only peripherally considered in this list. Theory 1 Relative Advantage: If the relative advantages of using the GIS over the procedure or system the GIS replaces is small, even though benefits to the overall organization might be great, the GIS’s intended users will resist its use Theory 2 System Complexity: If the complexity in using a GIS over the procedures or system the GIS replaces is great, even though advantages to individuals, groups, or the organization as a whole may be substantial, the GIS’s intended users will resist its use. Theory 3 Opportunity to Experiment: If prior to making substantial commitments of their time and resources, opportunities to try out and experiment with the GIS technology in their organizational setting and to view operational applications of the technology in similar organizational settings are low, the GIS’s intended users will resist use or expanded use of the GIS. Theory 4 Internal Communication Networking: If the degree of interpersonal communication networking in the organization is low (i.e. few informal and formal opportunities to talk with others at all levels and across divisions), resistance to expansions of use or improvements in use of the GIS will be great. Theory 5 Standard Operating Procedure Compatibility: If the standard operating procedures of an organization are incompatible with the transactional processing prescribed by the GIS technology, those whose procedures are altered or disrupted will resist use or expanded use of the GIS. Theory 6 Extent of Change in Standard Operating Procedures: If the degree of change in standard operating procedures necessary to accommodate the GIS is great, resistance by those whose procedures are altered or disrupted by use of the GIS will be great. Theory 7 Management Framework Rigidity: If the ability of an organization to restructure its chain-of-command or alter its mid-level management structure is low, resistance to expansions of use or improvements in use of the GIS will be great. Theory 8 Internal/External Cooperation: If an organization or division within an organization does not operate internally in a way that facilitates cooperation, the GIS’s intended users within the organization or division will resist cooperating with those external to the organization or division in networking or using a common GIS. Theory 9 Personal Differences: If the degree of segregation of professional disciplines and educational backgrounds among the divisions of an organization is great, resistance to expansions of use or improvements in use of the GIS will be great. Theory 10 Executive Support: If strong executive or management support is lacking or management fails to treat and protect its GIS assets (e.g. personnel, database, hardware, software, & GIS-derived products) as a strategic organizational resource, resistance to the continued or expanded use of the GIS will grow over time. Theory 11 Manager/Technocrat Team: If an organization does not have in its management structure at least one key manager who is a GIS advocate and at least one GIS administrator who is highly competent technically, resistance to the continued or expanded use of the GIS will grow over time. Theory 12 Personal Attitudes: If an organization does not have creative and positive individuals leading and moving forward the use of GIS in several of the organizations critical divisions, resistance to the continued or expanded use of the GIS will grow over time. Theory 13 Interaction: If the patterns of information flow and control prescribed by using the GIS are incompatible with the information control patterns prescribed by the organizational structure, those individuals and groups losing power in the organization will resist its use. Theory 14 Personal Advantage: If the personal benefits to individuals resulting from use of the GIS in an organization are low, even though benefits to the overall organization might be great, the affected individuals will resist use or expanded use of the GIS. Theory 15 Personal Risks: If the perceived or real negative ramifications to individuals who fail in their attempts to use a GIS are not far outweighed by the rewards possible by risking and achieving success, the intended users will resist use or expanded use of the GIS. Theory 16 Technological Opportunity: If the technological sophistication of the intended GIS users in the organization is low but through expansion of GIS use to their area of responsibility they are given the opportunity to learn and advance themselves, resistance by such individuals to using the GIS will be low. Theory 17 Technological Threat: If the technological sophistication of those who are unlikely to use the GIS is low, these individuals will resist use or expanded use of the GIS by the organization. Theory 18 Education: If GIS technologies are unable to be appropriately and effectively applied by mid-level managers and technical staff due to their inadequate familiarity with the technical details of the technology (i.e. database, software, and hardware concepts), resistance to the use or expanded use of the GIS will be great. Theory 19 Memory of Past Failures: If past failures in incorporating technological innovations in the organization have been recent, severe or extensive, resistance to the use or expanded use of the GIS will be great. Theory 20 Early Involvement: If key decision makers, line managers, technical staff and intended users in the organization are not consulted and allowed to contribute early on in the planning of a GIS acquisition, implementation or expansion, they will resist its use or expanded use. Theory 21 Unrealistic Expectations: If key decision makers, line managers, technical staff and intended users in the organization have an unrealistic view of the resources and time required to maintain or expand use of the GIS, resistance to the continued or expanded use of the GIS will grow over time. Theory 22 Consensus Building: If the strategy for implementation or expansion of use of the GIS is by executive directive rather than by team building and consensus, the intended users will resist its use. Theory 23 Inertia: If the use or expanded use of GIS is optional for intended users or the GIS is implemented other than through a project mandate, the intended users will resist its use. Theory 24 Committee Management: If the strategy for implementation or expansion of use of the GIS is through "committee decision making" rather than by executive directive or program mandate, committee members will resist change and resist use or expansion of use of the GIS. Theory 25 Fair & Rational Consultant: If recommendations regarding technical, operational, and management structure changes necessary to accommodate the GIS technology are made by individuals internal to the organization rather than by or with the involvement of "independent, knowledgeable, and unbiased" consultants, resistance to suggested changes will be great. Theory 26 External Communication Networking: If the degree of interpersonal communication networking by external GIS consultants and/or GIS developers with the organization’s intended users and key decision makers is low, resistance to expansions of use or improvements in use of the GIS will be great. Theory 27 Goal-Oriented Plan: If active expansion of GIS use does not follow a well-structured and logical plan with clearly focused end goals, the intended users of the GIS will resist its use. Theory 28 Project Progress Reporting: If the degree of reporting on progress and provision for feedback is low during active expansion of GIS use, those intended users not adequately involved will resist its use. Theory 29 Extensive Time Lapse: If the primary advantages in using a GIS will not become evident until a time which is distant, successful and visible applications must be accomplished in the interim or managers and the intended users of the GIS will resist its use. Theory 30 Inadequate Time Allowance: If the pace of expansion of the GIS is such that the time allowed for thoughtfulfeedback is inadequate for the intended users and for those in the organization otherwise affected by the conversion,resistance to the use or expanded use of the GIS will be great. ReferencesThe primary references used in developing the above list are the same as those contained in the reference list of the attachedpaper entitled "Case Study Research Methods for Geographic Information Systems." The concepts reflected in the list can typically bereferenced to numerous sources and therefore references were not provided on an item by item basis. Chapter FourAn Investigation of Technology TransferTheories at the XYZ Paper Company Steven Frankstevef@mecan l.maine.eduDepartment of Surveying Engineering /-NCGIAUniversity of Maine AbstractCase study methods have been suggested as a tool for testing GIS technology transfer theories indirectly. The goal of theresearch reported in this paper was to identify and analyze control situations which would allow testing of pre-specified technologytransfer hypotheses within a GIS organizational environment. The organization chosen for study was the XYZ Paper Company(fictitious name for an actual company), a New England forest products concern which has used GIS technology since 1980. Theorganizational and policy settings of the company were examined through interviews with company employees, documents providedby the company, and by personal observation. The results of hypotheses testing are provided.Case study methods have been suggested as a tool for testing GIS technology transfer theories indirectly. The goal of theresearch reported in this paper was to identify and analyze control situations which would allow testing of pre-specified technologytransfer hypotheses within a GIS organizational environment. The organization chosen for study was the XYZ Paper Company(fictitious name for an actual company), a New England forest products concern which has used GIS technology since 1980. Theorganizational and policy settings of the company were examined through interviews with company employees, documents providedby the company, and by personal observation. The results of hypotheses testing are provided. IntroductionCase study research methods are often used to study phenomena in their natural settings. Onsrud, et al. (1992) have suggestedusing case studies to investigate technology transfer theories involving geographic information systems. They recommend using casestudy methodology proposed by Lee whereby theories may be indirectly tested: "The indirect testing of a theory consists of derivingone or more suppositions capable of being tested directly from the theory and then comparing the actual outcomes against those whichwere predicted (Onsrud, et al. 4)." The integrity of such comparisons is determined by the ability of the researcher to isolate otherpossible causes for the observed behavior through the identification and observation of control situations within the ongoingoperational GIS setting. Repeated success of a theory across a range of circumstances leads to the conclusion that said theory can begeneralized over that range of circumstances (Onsrud, et al. 6). BackgroundThe XYZ Paper Company is a large paper and wood products company located in New England. Approximately threemillion acres of forest lands in New England and eastern Canada are managed by the XYZ Paper Company and an affiliated woodproducts company using Intergraph GIS software on a VAX/VMS computer. The GIS Section is located within the Lands Division ofthe XYZ Paper Company, one of three major divisions within the company. XYZ personnel estimate that less than one percent of thecompany’s annual budget is used for system maintenance and upgrade. The Lands Division at the XYZ Paper Company is charged with managing company forest lands holdings for the purposes ofsupplying wood fiber to its mills. After benchmark studies involving three different vendors, the Division acquired Intergraphhardware and software in 1980 for use as an automated mapping system. The system was upgraded in 1985 when, after more vendorcomparisons, the present system was installed. The system contains approximately 10 million acres of base mapping compiledprimarily from USGS quad sheets at 1:24,000 scale. Three million of these acres -lands actually managed by the XYZ PaperCompany -are mapped into the system in greater detail, primarily keeping track of forest cover types and growth information. Thesystem is interfaced through two workstations. The automated mapping capacity of the system has been used to supply map products throughout the company sinceimplementation of the system. The analysis capabilities of the GIS are little used except in the Lands Division. Among the tasksperformed by the GIS Section are: prepare annual maintenance reports for tax purposes; prepare current inventory reports for tax purposes; provide support for silviculture programs (i.e., planting, nurturing, and harvesting of trees) by maintaining classified forestcoverage records and data; assist in long term analysis and planning for wood fiber supplies; assist with various land management problems;assist with special projects as needed; archive historical spatial records. The 3 million acres managed by the XYZ Company contain additional detail which were originally compiled and digitizedfrom existing paper-based company maps. Forest cover data is updated by combining ground "cruise" information (gathered bycompany foresters) with digitized and scanned 1:63,360 aerial color infrared photography taken on an annual basis. Classified forestcoverages obtained from aerial photography are supplied by private contractors and verified in-house before being entered into thesystem. Additional spatial information is collected from ground surveys performed by private surveying firms. The system is updatedas needed or as new information becomes available. The only current access to the system is directly through the GIS Section. Some routine analyses are handled as part of theeveryday workload. Those wishing nonroutine analyses consult with GIS Section personnel about the feasibility of their requests.Users seem to be generally satisfied and minimally restricted by the current system access priorities. However, some users havecomplained that certain analyses cannot be performed due to scale and data structure restrictions within the system. Companymanagement and GIS Section personnel feel that it would not be cost effective to collect the necessary data or to restructure thedatabase to satisfy these complaints at this time. Plans are under way to expand the system to desktop personal computers located infield offices, which already have personal computers in place and are using them for some GIS-related record keeping. These personalcomputers will lack the full analytical abilities of the VAX/Intergraph located in the GIS Section office (these limits are technologicaland cost related). GIS Section personnel and company management believe this expansion will relieve the GIS Section of manylow-level automated mapping requests. The organization of the XYZ Paper Company has undergone several structural changes since the system was installed in1980. High-level corporate maneuvering has tied the company to another local paper and wood fiber company, the ABC Company.Because the ABC Company does not have a GIS, its forest holdings coverage has been added to the XYZ Paper Company GIS. TheABC Company is currently acquiring its own GIS. When the ABC Company GIS is installed, the XYZ Company will stopmaintaining this data. The current economic recession and the need to become "lean and mean" have also affected the Company, reducingmanpower within the GIS Section by half. The GIS Section appears to have weathered these changes without major damage.However, current system expansion has slowed almost to a standstill during these cutbacks. Study MethodSeven employees of the XYZ Paper Company were interviewed for this study. Interviews were conducted at the XYZ PaperCompany offices during normal working hours. The interviewing process consisted of a series of open-ended questions regarding eachinterviewee’s experiences and opinions about the system. Observations were made at several workplaces and the accessibility of eachworkplace to the GIS section was observed. The organizational structure of the company was reviewed by studying charts provided byXYZ Paper Company and organization policies related to the GIS operations were reviewed through documents provided by thecompany and by interviewing personnel.Of the persons interviewed, only two had direct access to the system GIS workstations. Another was responsible for datacollection and quality control, but normally lacked access to the system. Three persons regularly request or are provided withanalytical services from the system. The seventh person interviewed was a high level manager in the Lands Division. All had collegedegrees related to the forestry industry and had worked for the company for periods ranging from ten to thirty one years. Theory Testing ObservationsThirty technology transfer theories were hypothesized based on case histories drawn from the GIS literature and varioustechnology transfer studies reported in the GIS literature. This process was accomplished by GIS Case Study Research coordinator Dr.Harlan Onsrud. The thirty hypotheses were studied prior to the interviewing and observation processes. Upon familiarization with thetheories, the goal was to search for unique control situations which might allow testing of one or several of the prespecifiedhypotheses. Within the organizational setting studied, I was able to isolate control situations for only four of the hypotheses. However,for some of the hypotheses several control situations were identified. Situations were observed which allowed the following subset oftheories to be tested: Theory 3 Opportunity to Experiment: If prior to making substantial commitments of their time and resources, opportunitiesto try out and experiment with the GIS technology in their organizational setting and to view operational applications of thetechnology in similar organizational settings are low, the GIS’s intended users will resist expansion. Theory 4 Internal Communication Networking: If the degree of interpersonal communication networking within theorganization is low (i.e., few formal and informal opportunities to talk with others at all levels and across divisions),resistance to expansions of use or improvements in use of GIS will be great.Theory 10 Executive Support: If strong executive or management support is lacking or if management fails to treat orprotect its GIS assets (e.g. personnel, database, hardware, software, and GIS-derived products) as a strategic organizationalresource, resistance to continued or expanded use of the GIS will grow over time. Theory 12 Personal Attitudes: If an organization does not have creative and positivethinking individuals leading andmoving forward the use of GIS in several of the organization’s critical divisions, resistance to the continued or expanded useof the GIS will grow over time. Opportunity to ExperimentSituation 1: The XYZ Paper Company examined two major GIS systems in 1979. Benchmark tests were conducted tocompare a Comarc system against an Intergraph system. Since they had the ability to experiment with promising systemswithin their own operational setting, the "opportunity to experiment" hypothesis suggests that there would be less resistanceby users than if the system had been forced on the users by the management without the ability to experiment. Afterbenchmark tests the intended users supported the purchase of one of the systems and the XYZ Paper Company did in factpurchase an Intergraph system in 1979. The result tends to evidence support of the theory. Situation 2: The XYZ Paper Company again examined two major GIS systems in 1985 with the thought of upgrading itscurrent Intergraph system. Benchmark tests were conducted again to compare Intergraph against Arc/Info. The "opportunityto experiment" leads to the prediction that the users would not resist an upgrade to their system. In 1985, the company indeedpurchased new Intergraph hardware and software for their GIS. The result again tends to evidence support for the theory. Situation 3: Around 1987, personnel in the Taxes Section of Lands Division visited a wood products company in California.It was observed that the California company was using its GIS to prepare tax documents and records for local governmentagency requirements with great success. The XYZ Paper Company then experimented with preparing tax documents andrecords on its system. The "opportunity to experiment" again leads to the prediction that the system users would support theexpansion of GIS operations to include tax document and record preparation. In 1987 or 1988, thisexpansion did occur at the request of users. The result once again tends to evidence support for the theory. Situation 4: As a result of high-level corporate maneuvering, the XYZ Paper Company became affiliated with another localpaper and wood processing company. The Lands Department within the affiliated company lacked a GIS. The LandsDepartment of the affiliated company observed the GIS of the XYZ Paper Company and had a major project performed onthe XYZ Paper Company’s GIS. The predicted outcome would be the lower user resistance to acquiring a GIS by theaffiliated company. The affiliated company is in the process of purchasing a GIS. Meanwhile, the land holdings of theaffiliated company are now included in the GIS of the XYZ Paper Company. The result once again evidences support for thetheory. Rival theories: In situations I and 2 a strong GIS champion was credited by many in the department as being instrumental toacquisition and upgrading of the system. This person left the XYZ Paper Company in 1986, shortly after the GIS was upgraded, butbefore situations 3 and 4 arose. This provides evidence that the opportunity to experiment may have been a factor in each of thedescribed situations regardless of the presence of a strong champion. Executive support for the GIS Section appeared to be at the samelevel throughout the situations described, although the number of personnel in the GIS Section was much higher in situations I and 2which occurred before situations 3 and 4. Internal communication networking remained relatively constant throughout the time perioddescribed. Internal Communication NetworkingSituation 5: Personnel within the Lands Division of the XYZ Paper Company responsible for particular duties within thecompany (i.e., management of pest infestation, road construction, taxes) or with particular tracts owned by the company areencouraged to deal directly with personnel within the GIS Section to determine possible methods of analyzing problems using thesystem. Theory application would predict that because of this ready access, resistance to use or expanded use of the GIS within the Lands Division is likely to be smaller than if these communication channels did not exist. While the foresters lack direct access to theGIS, they seem to have little hesitancy in consulting with the GIS Section about new problems that arise. No evidence was gathered todetermine if foresters with better opportunities for communication made more frequent use of the GIS. However, this behavior and theoverall increased GIS use by foresters tends to evidence support for the theory. Situation 6: Personnel within other Divisions of the XYZ Paper Company, such as the Recreation Division and PulpwoodProcessing Division, have almost no low-level communication with the Lands Division. The predicted outcome of such behaviorwould be that greater resistance to the expansion of GIS use would occur within these divisions. Although these divisions do receiveGIS products, there has been no change of the basic nature of those products since they were first introduced in 1980. Although nodirect evidence from those divisions was gathered, there appears to be very little enthusiasm for expanding GIS use in the divisions.The result again tends to evidence support for the theory. Rival theories: There are many opportunities for personnel within the Lands Division to experiment with system applications,leading us to believe that internal communication networking was not the only critical factor in the above described situations.However, lack of low level communication with other divisions of the XYZ Paper Company coupled with the lack of expansion of theGIS to those divisions does gives this theory much credence. The personnel within the Lands Division have remained relatively thesame over the years of system operation, as has the level of executive support. Thus theories which might be affected by changes topersonnel or management do not seem to be relevant to the outcome of situations 5 and 6. Executive SupportSituation 7: The XYZ Paper Company GIS expanded at a rapid rate from 1980 to 1988, requiring a system upgrade in 1985.Corporate alliances and economic conditions have forced many changes in the system since 1988. New management policy seems tofocus on company survival rather than on company growth. While still highly valued within the Lands Division, the GIS Section isseen by higher level management as a part of the organization which must operate under the same limitations and budget constraintsas the rest of the company. GIS Section personnel were reduced from six to three persons. A former policy of cycling GIS personnelthrough field offices was halted. The "executive support" theory argues that since executive support has decreased, resistance to theexpanded use or growth of the GIS would increase. However, the system has expanded within the Lands Division and has gown toinclude coverage of an affiliated paper company’s forest holdings in 1989. Plans are under way to expand the system to personalcomputers located in field offices within the Lands Division. The evidence seems not to support this theory. Rival theories: Opportunities to experiment, internal communications networking, and personal attitudes are difficult to relateto this theory. It would appear that each of these theories can be valid without strong executive or management support beingevidenced. Personal AttitudesSituation 8: The person in charge of special woodlands monitoring and coordinating general forest activities, who holds aMaster’s Degree in Forestry, is held in high regard by her coworkers. While she professed to some disillusionment andfrustration with the system’s technological and accuracy limits, she appears to use the GIS more than any other individualoutside of the GIS Section. (It must be noted that this person has an office in the same building as the GIS Section, unlikeseveral other users of the system.) Most of those with whom she works are located in field offices. "Personal attitude" theoryapplication would indicate that resistance to GIS use would decrease among those with whom she worked. And, in fact, thereis continued use and expansion of use of the GIS among the foresters with whom she coordinates activities. This outcometends to evidence support for the theory. Situation 9: The person in charge of database management in the GIS Section is also highly regarded by his coworkers. Thisperson also holds a Master’s Degree in Forestry. He is not only responsible for the conceptual evolution of the system (whiledelegating many of the technical aspects to others in the Section), but he is also responsible for developing or coordinatingnew applications for the GIS. Coworkers feel that they can consult with him on any problem that has the potential to besolved with help from the GIS. Recent personnel cutbacks in the GIS Section have caused this person to spend less time withnew developments and assist more with the day-today maintenance of the GIS. Theory application would again indicatelower user resistance to the continued or expanded use of the GIS. GIS use has continued and expanded within the LandsDivision of the XYZ Paper Company, although new applications appear to be occurring at a much slower rate in recent times.Again, the outcome appears to evidence support for the theory. Rival theories: Again, the opportunity for others to experiment cannot be ruled out as a critical factor in GIS expansion. Theinternal communication networking at the XYZ Paper Company must also still be considered as a reason for the pattern of expansion in the company. Executive support has seemed to be constant over the period these two people have been with the company and doesnot seem to affect the outcome of our predictions for this theory. ConclusionsThe objectives of this research were twofold. The first objective was to use the methods set forth by Lee (in Onsrud, et al.1992) and expanded by Onsrud, et al. (1992) in a GIS setting to test the utility of the case study methods proposed. The experiencegathered from this single case suggests that the indirect case study research process can be a useful and valid method for testingtheories. However, there were tendencies by both the observed and the observer to bias responses and/or predictions. Continuedexperience in indirect case study methods should allow the observer to recognize possible biases and improve his or her researchtechniques. By extension, I believe that the validity of the results generated by this method are substantially dependent on theresearcher’s ability to make thoughtful observations. This skill may be gained only through experience. Experience may be gained by:1) observation and analysis; 2) personal and peer review of observation and analysis techniques; and 3) repetition of observation andanalysis with new cases. This iterative process should allow the observer to hone the skills necessary to produce credible and usefulresults. A thorough technical understanding of GIS is also highly useful to enable the observer to distinguish behavioral andorganizational issues from technical issues and to observe their interplay. The second objective of this study was to gather actual evidence which would help to support or falsify technology transfertheories which might be generalizable and useful across a range of GIS organizational settings. Control situations were isolated in theorganizational setting of a private paper company which allowed the testing of four technology transfer theories. Evidence supportingor falsifying the theories was gathered using indirect case study research methods. The results are summarized in table 1. A measureof the degrees of freedom of the situations observed are tabulated in table 2. As stated above, the primary limitations to the validity of the results appears to be the experience of the observer, not the casestudy procedures. There appear to be very few people in the GIS community with the requisite theory-focused case experience to testmethods set forth by Onsrud, et al. (1992) against other methods. Experience and the results of many case studies are required to buildthe evidence base that will provide meaningful results over time. Although this single case study makes only a very small incrementalcontribution to actually testing theories, building scholastic-based case study expertise within the GIS community would be valuableand should be stressed. AcknowledgmentsI would like to thank all those at the XYZ Paper Company who agreed to be interviewed and observed for this study. Inparticular, I would like to thank the head of the GIS Section whose enthusiasm and cooperation made this study possible. This study represents work related to research being carried out at the National Center for Geographic Information andAnalysis and was partially supported by a grant from the National Science Foundation (Grant No. SES-88-10917). The research wasconducted under the auspices of the GIS Case Study Research Project coordinated by Dr. Harlan Onsrud of the University of Maine. ReferencesOnsrud, Harlan J., Jeffrey K. Pinto, and Bijan Azad (1992) Case study research methods for geographic information systems. URISAJournal, 4(l): 32-44. Chapter FiveHuman Factor Determinants of GIS Implementation Success:Perception, Experience, Attitude, and CommunicationBehaviorCase Study of Four Governmental Agencies inCumberland County, NC