Division of Research Graduate School of Business Administration The University of Michigan September 1982 INSTITUTIONAL DYNAMICS OF ACTION RESEARCH Working Paper No. 315 Noel M. Tichy Steward D. Friedman The University of Michigan FOR DISCUSSION PURPOSES ONLY None of this material is to be quoted or reproduced without the expressed permission of the Division of Research. Paper for: Producing Useful Knowledge for Organizations Conference at the University of Pittsburgh October 28-30, 1982

INSTITUTIONAL DYNAMICS OF ACTION RESEARCH Noel M. Tichy Institute for Social Research and Graduate School of Business Administration University of Michigan and Stewart D. Friedman Institute for Social Research University of Michigan We are interested in keeping action research alive and well. We adhere to the Lewinian notions that the best way to understand a social phenomenon is to try to change it and that the best theories are useful ones. It is through the collaborative efforts of academicians and practitioners in organizations that an advance in scientific knowledge about organizations can be and should be made. But, for maturation of our still young science of organizational behavior to continue, we must take heed of the forces which foster and those which inhibit action research endeavors. In this paper we examine the conditions in academic/research institutions and in work organizations that ought to be recognized in building on the foundation constructed by action researchers since Lewin. Alternative arrangements for action research are considered from the point of view of the host organization and of the academic/research institution in which the researcher is based. Then, we describe one approach to managing the institutional dynamics of action research. Based on a case example, the implications for the design of research institutions, and for particular projects carried out therein, are drawn. THE CASE FOR ACTION RESEARCH Action research involves the creation and refinement of theory based on 1

'1 2 the application of theoretical principles of organizational behavior in an effort to alter a social system. There are two primary reasons why this type of research is a necessity for our field. First, tests of utility of theories about organizational behavior must occur in organizations. Ideas that germinate in contexts non-organizational may theoretically appear useful but may prove to be otherwise. For example, the use of T-groups as a means to develop greater self-awareness in members of organizations failed in part because the cultures of work organizations were not supportive of the changes initiated in T-group experiences. Certainly it is possible for a theory or method to successfully enter organizational realms from elsewhere, but the final judgment of their usefulness must occur inside these realms. Unfortunately, because of the paucity of well-supported and well-done action research, there is not much that is currently useful to work organizations from the field of organizational behavior. The second reason bears on the need for researchers to observe organizational phenomena as non-members in order to create and refine their theories about organizations. Non-member status grants the required objectivity for dispassionate analysis; being in and around organizations stimulates theorizing and refinement of theory. Perhaps the most successful example of this kind of discovery of theory in organizational reality was Trist and Bamforth's (1951) experience in the British coal mines in the late 1940's. Socio-technical systems theory - which has enjoyed successful applications in many spheres - arose from their work in a live organizational setting. Freud's pioneering work in psychoanalysis - a kindred field wherein a theory of human behavior is applied to human problems - evolved continuously because of new insights gained from his clinical practice. Freud was foremost a scientist; he saw patients primarily because it was only by applying his

3 theory that he could derive corrections and reformulations to it. Certainly there are risks inherent in action research. In collaborating with client organizations on change projects the researcher endangers his or her objectivity. The extent to which the researcher "goes native" depends on the pressures from the host organization and from his or her academic/research base to do so. As we discuss below, there are ways to reduce these pressures. The research strategy least likely to lead investigators astray is one that is "balanced", in the sense used by Runkel and McGrath (1972). More than one research method is used including laboratory experiment, sample survey, simulation. observation, and action research. with the choice of method made to exploit the strengths and weaknesses of each (e.g., degree of objectivity attainable by the researcher). An action research effort then, may best contribute knowledge in the context of a larger, programmatic set of research projects. McGrath's (1964) advice on the "logical path for programmatic research" is an idealized course to follow, nevertheless we ought to select research methods and settings, on the basis of their applicability to the stage of theoretical advance at which we find ourselves. Therefore, the action research mode should be an integral part of a balanced research strategy for developing knowledge about organizations. TECHNICAL, POLITICAL, AND CULTURAL FACTORS IN ACTION RESEARCH In the following two sections, we will present the forces for and against action research using a tripartite framework of organizations. For analytical purposes, we trisect organizations according to their technical, political, and cultural subsystems; these are concerned with three core dilemmas, respectively. The first dilemma is the technical design problem. Here, the organization faces a production problem; that is, social and technical

4 resources must be arranged so that the organization produces some desired output. Second is the political allocation problem or, in other words, the problem of allocating power and resources. The uses to which the organization will be put, as well as who will reap the benefits from the organization, must be determined. The third is the ideological and cultural mix problem. As social tools, organizations are, in part, held together by normative glue, that is, by the sharing of certain important beliefs by its members. Hence, the organization must determine what values need to be held by what people. These three dilemmas are ongoing problems due to the fact that organizations are always undergoing shifts and change. These three problems are viewed as systems of interrelated sets of components, each organized around a coherent logic. The technical system includes the interrelationship of all those elements required to deal with the production problem. The political system includes all the practices, activities, and elements used to work on the allocation problem. And the cultural system involves the symbols, values, and elements organized to address the dominant ideology problem. In the remainder of this article technical, political, and cultural will refer to factors related to these three systems in both work organizations and academic i nst i tut ions. Technical, political, and cultural forces impinge on the conduct of action research from the point of view of the host organization and of the academic/research institution in which the researcher is based. We examine these forces in turn. The Host Organization's Perspective (see Exhibit 1) Insert Exhibit 1 About Here ___________________________ In the technical sphere, the major force in favor of action research is

5 that data-based change efforts can add value to the bottom line. The introduction of quality circles, for example, may lead to modifications in operating procedure that reduce cost. By examining such projects with action research methodology, the extent of the value can be assessed with some accuracy. On the down side, there are difficulties in the measurement of change, ethical and pragmatic obstacles to using control groups; and a lack of research skills in most organizations. Politically, the information obtained in action research projects can be a very powerful leverage point. In many action research projects there are evaluative conclusions made that may indicate the need for power shifts to occur. The action researcher must remain vigilantly aware of the political implications of his or her efforts for the client, always cognizant of the ethical dilemmas he or she confronts. The political forces operating in the host organization against the conduct of valid research are: (a) the costs are high; (b) the reward system in most organizations is for short-term profits - results from many action research projects are long-term and difficult to quantify; and (c) the action researcher, because he or she may be paid by the client, is in danger of overadvocating his or her position, thus biasing the conclusions drawn from the research. Biased research may be worse than none at all for the client to use in guiding policy decisions. The cultural factors supporting action research in client organizations are: (a) the influence of Japanese management is forcing a reevaluation of our traditionally non-research oriented approach to change. The notion of the "experimenting organization" (Staw, 1977) may be the norm for U.S. managers in the coming years; (b) if action research can serve the interests of society as well as the company in which it is conducted, then the company is seen as making a contribution to the greater good. For many firms, this meets their

6 stated philosophy of being "good citizens". The main factor acting against action research is the typical American corporate image of the "macho man" who takes the ball and runs with it. This attitude is contrary to one in which the measurement and assessment of incremental changes guides future action. The mentality of the "quick fix" pervades corporate culture. Impatience for results - fast solutions - is the American style. The Academic/Research Institution's Perspective (see Exhibit 2) Insert Exhibit 2 About Here The primary technical force that encourages action research is that it provides scientists with opportunities to test theoretical propositions about organizations. Furthermore, experience in field settings is a necessary condition for the generation of ideas about how organizations function. In applied fields such as ours, immersion in the phenomena in "real life" settings is required. However, since field skills are generally not taught in our graduate programs, where instead the emphasis is on purer academic concerns, the cadre of trained action researchers remains relatively small. Another technical problem is that particular methods of research tend to be ingrained in certain schools of thought; action research calls for flexibility in methodology. The political forces encouraging action research are: (a) since resources available for research in the public sector are on the severe decline, scientists must turn to the private sector for support; and (b) once a theory has been subject to field test via action research projects and has proven to be useful, it can be marketed to industry with greater ease than if it were not tried and true. Action research can provide evidence for the utility of organization theory. Political forces against action research are:

7 (a) rewards for academics too tend to be short-term oriented (e.g. promotion decisions must be made within a short time span; "publish or perish" inhibits longitudinal research reporting); and (b) if private industry is paying for the research, then they own the results. As we noted above, this may invalidate conclusions drawn from the research. For the academic institution this is clearly an obstacle to the accumulation of knowledge. The cultural factors supporting action research are: (a) it meets the ethical needs of social scientists to produce knowledge that enhances the life quality of the community-at-large; and (b) some academics have developed the entrepreneurial style necessary to garner support for action research projects. Thus, there is a heritage of action research that can be nurtured within the academic community. On the other hand, there are those who feel that to be associated with applications of science is to be impure. Basic research, unconfounded by the vagaries of "real life" is extolled in many academic circles; selling out is not. The prototypical "tweedy" academe does not mix well with the "pin-striped" corporate executive. TYPES OF INSTITUTIONAL ARRANGEMENTS FOR ACTION RESEARCH For host organizations and for academic/research institutions there is variation on the arrangements made for the conduct of action research. We examine the differences in approach below. Host Organizations Host organizations can assume one of five orientations vis-a-vis the evaluation and monitoring of change. Each one has a set of technical, political, and cultural tradeoffs associated with it. These are summarized in Exhibit 3.

8 Insert Exhibit 3 About Here 1. Informal anecdotal orientation. This approach is probably the most frequent one found in U.S. corporations. There is little technical sophistication required. Information is picked up via members' observations - anecdotes told in very informal meetings is how data is transmitted. There are minimal political risks and the commitment to evaluation is low since nothing formal is going on. Obviously, this orientation is associated with a cultural value that does not place much of a premium on systematic evaluation and monitoring. 2. Managerial audit orientation. The audit approach is one that is fairly common to organizations. Management makes some effort to stand back and assess what is going on; it audits its own activities. An example of such a process might be periodic review meetings where managers are asked to summarize what they are doing in a particular area and then evaluate the plusses and minuses. One company has been using this approach in its quality of work life (QWL) effort. The company has over 500 quality circles, gainsharing experiments, new training and development activities, and the like. Periodic review meetings are held by senior management at which business managers present what they are doing in the QWL area. An attempt is then made to critique what is going on. This differs from the guidance system approach which includes systematic measurement and evaluation. Technically, the audit approach requires fairly low level sophistication. Political risks and commitments are not that critical because the managers themselves control what they bring to the review sessions. They can, and usually do, color their reports to support what the power figures want. The depth of cultural commitment to evaluation is low because there is a great deal of distortion

9 and the information is quite informal under these conditions. 3. Guidance system approach. The guidance system approach differs from the following one in that systematic data is used to guide an overall change strategy, but it does not include experimental variations to learn how different things might work under different conditions. For example, a company decides to undergo a major QWL effort. Instead of experimenting in some of its plants first, it may launch the program in all its plants, starting with a survey and some other data collection techniques, and then using the survey methodology over time to guide the change effort to see what changes have occurred or have not occurred, making adjustments based on the data. But there would not be an effort made to vary the types of interventions, thereby employing a quasi-experimental design. The guidance system approach calls for a moderate degree of technical sophistication; the use of tools like surveys and other systematic data collection techniques is required. Politically, there are a moderate number of risks and commitment of resources needed. By making the measurement of change somewhat systematic it is easier for people to become visible and to be held accountable. Culturally, it must be an organization that values having systematic data put on the table to examine as the change process unfolds. In many ways the guidance system approach has similarities to other control systems found in organizations in the production and the financial area. There is one major cultural difference. Many of there systems are set up to catch mistakes. The guidance system approach is only workable where the predominant reason is not to catch mistakes, and then punish them, but to catch mistakes and problem-solve and learn from them. Catching mistakes and punishing them may be appropriate in some settings where there are routine activities going on. In an organizational change activity it is inevitable

10 that there will be many mistakes; the issue is how to learn from them. Thus, there must be a cultural value that supports non-defensive evaluation. This is like the type of learning and adjustment identified by Pascale and Athos (1981) in Japanese organizations that allows them to make strategic accommodation. For example, Toyota can introduce a car, as it did in its early days in the United States, and fail, as did Datsun and Mazda. But the organization has the capability for learning from these mistakes, not a cultural and political system which punishes and therefore creates defensiveness. Academics have an opportunity in this type setting to conduct theory generating exploratory action research. 4. Experimenting intervention orientation. The experimenting intervention orientation and the following basic R&D approach are the two preferred types of action research. These differ in degree rather than substance. They both require a fairly high technical sophistication, a fairly high political risk and commitment of resource orientation, and a fairly high cultural value commitment to systematic evaluation and monitoring. The major difference is that the basic R&D orientation requires the organization to put aside separate resources and protected environments to experiment and examine phenomena related to organization and management change, whereas the experimenting intervention orientation requires that measurement be done around naturally occurring change in the organization. The history of Volvo's efforts in the quality of work life (QWL) area is an example of a company that had an experimenting intervention orientation to change. In the late 1960's there were a number of experiments done on job rotation, job enrichment, and a limited amount of teamwork. Each of these experiments was extensively measured and there was a commitment to evaluate and assess what had been learned. Furthermore, different interventions were tried in different groups

11 so that they were done as quasi-experiments, with explicit attention given to having some control groups. By the early 1970's the QWL effort at Volvo had been expanded. There was even a new assembly plant designed from the ground floor up at Kalmer, where workers, union, management, and staff people participated in an innovative design of the production system. Throughout this effort there was an experimenting mode of work. Measurements were taken on both the people-side and the productivity-side as the plant was opened and the innovations were implemented. Some things were tried one way, and then another way with systematic measurement to help determine what were the best ways of proceeding. Ultimately, this line of activity, including the transformation of plants like the Skovda Engine plant, moved Volvo toward the model of autonomous work groups and the socio-technical system as a way of life. The evolution of this approach, however, took a ten-year period in which the basic orientation of Volvo could be described as an experimenting intervention orientation. 5. Basic research and development orientation. This approach is one in which the organization is committed to approaching organizational change in the same way one approaches good research and development (R&D) on a product; that is, there is systematic testing and evaluation of the change as it unfolds. This includes doing experiments to determine different ways of accomplishing change goals. It could involve setting up special laboratorylike conditions to test out new concepts and ideas. An example of this might be the General Foods' Topeka plant which was set up as a prototype or model. The Topeka plant ended up being a very successful experiment in the short run, showing how innovative, socio-technical design of new plants could lead to high productivity and high satisfaction among the employees. It succeeded technically but failed due to political and cultural dynamics. As an R&D

12 activity, it did not contribute to strategic change at General Foods. In looking at this failure we can better understand the conditions needed to create a basic R&D orientation in a company. From a traditional point of view, the General Foods' Topeka example did work. They had the necessary technical expertise to treat the plant as an experiment from which they could learn and derive principles for dissemination to other parts of the organization. Where it ran into trouble and where it was counter to the General Foods organization was in the cultural and political area. Politically, problems were created due to the fact that other plants and other plant managers became jealous. A disproportionate amount of attention and resources went to this plant which made other plants feel competitive. There was pressure to either do things better than Topeka or to isolate it and make it into an aberration. The latter course was taken, thus politically preventing disruptions in other plants. The others did not have to copy the change. Furthermore, there was not the political support at high levels of management for treating this as an experiment. Therefore, the plant became isolated and seen as a non-traditional and alien-type appendage. The political problems were clearly linked to cultural ones. The culture did not value treating organizational change as an R&D activity, rather you either made a policy to do something one way or you did not. The result was that those who supported the Topeka plant had to overadvocate it as a panacea, thus threatening the more traditional plants. There was not a culture which enabled them to problem-solve and sort out what worked and what did not, and how to apply lessons to other parts of the company. As a result of these political and cultural forces, they lost the opportunity to conduct meaningful R&D on approaches to managing plants that are in the General Foods network. The ideal basic R&D orientation will occur in an organization where there

13 is the technical sophistication represented in the General Foods example. And, where there is a deep cultural commitment to conducting R&D on organizational innovations and changes; and where there is an alignment of the political structures which allows people to take risks, fail, learn from mistakes, and ultimately develop new orientations. Such political and cultural orientations are very difficult to create, and to our knowledge there is no company that we would currently categorize as having a basic R&D orientation to organization and management. Even though the set of organizations having or willing to have an action research orientation toward evaluating and monitoring organizational change activities is small, we predict that as we move through the turbulent 1980's where increasingly organizations face major transformations, the set will enlarge. The enlargement of the set can be enhanced by the role academics take in encouraging more collaborative action research. Academic/Research Institutions Academic/research institutions concerned with organizational behavior can be considered on two continua: first, from "basic" to "applied" in terms of their relative emphasis on research and the advance of science. Some institutions may provide analyses of case studies that are useful to business firms, for example, but of little interest to the scholar who is developing organization theory. On the other hand, the "ivory tower" may be a good place to spin yarns of theoretical import that may end up having little to do with the stresses and strains of everyday organizational life. Somewhere in between we find the academic/research institute that is simultaneously in close touch with organizational reality and in the arena of theory deve opment. The second dimension refers to the extent to which the action researcher

14 pursues his or her work alone or with others as part of an interdependent team. In some settings it may not be possible for one to work without others whereas in other places the very thought of seeking assistance or collaboration is taboo. Exhibit 4 illustrates four idealized types of arrangements in which academic or research interests are pursued. The types are described in terms of their relative emphasis on action vs. research and individual vs. group efforts. Each type bears its particular weakness and its strengths. We present them here to aid in our description of an institutional arrangement that may offer some guidance for encouraging more action research in the organization behavior field. Insert Exhibit 4 About Here A MODEL AND SOME GUIDELINES FOR AN INSTITUTIONAL BASE FOR ACTION RESEARCH We have already identified the technical, political, and cultural forces which make it difficult to conduct action research in academic settings. In spite of these forces, there are some examples of research institutes which have been able to sustain long term commitments to action research. In England, the Tavistock Institute has more than 30 years experience as a center for action research resulting in many contributions to organization theory. The Tavistock Institute is not part of a university system and thus is more akin to what has occurred in the U.S. in other social science fields, namely, a Brookings Institute or a Center for Policy Research. These provide one type of model, but one that we feel does not have as much generalizability as a university based one. One university based model which may have greater generalizability in U.S. settings is that of the Institute for Social Research (ISR) at the

15 University of Michigan. It too has over 30 years of experience in action research and through the research works of such ISR staff as Rensis Likert, Robert Kahn, Stanley Seashore, Arnold Tannenbaum, Basil Georgopolis, Edward Lawler, David Bowers, and the scores of Ph.D. students trained, has been a major force in shaping the organization field. Before continuing with ISR as a model, some background is required. We are not proposing that ISR as a total entity is replicable nor is it necessary to contemplate such an undertaking to accomplish action research aspirations. ISR is a large social and behavioral science research center within the University of Michigan. It has an annual budget of over $15 million dollars and a senior research staff of between 80 and 100 with an additional support staff and research assistant staff of around 200. There are multiple centers within the Institute focusing on different problems and disciplines including behavioral economics, political behavior, and organization behavior. ISR generates its own research funding without the aid of University subsidy. This provides a great deal of autonomy but at the cost of giving up security. The part of ISR we will focus on is the organization behavior group made up of between 8 and 12 senior researchers. The number varies from year to year depending on projects and commitments of faculty to their various academic departments. Most ISR staff hold joint appointments in academic departments ranging from 20 to 80% of their time. Because ISR is a separate entity at the University of Michigan its members wear two hats. The chairs or deans of academic departments are parallel to the Director of ISR. As we shall point out, this is an important element in why action research is able to thrive at ISR. The organization behavior group at ISR is able to support research in all

16 four of the modalities presented in Exhibit 4. The technical, political, and cultural conditions which exist at ISR and the potential lessons for other academic settings desiring to foster this type of research arrangement is presented below. Technical Factors: Because ISR's mission, strategy and its structure is all focused on research, it has an infrastructure in place with state of the art capabilities in field interviewing, sampling, data managers, accountants who can handle complex research funding arrangements, publication facilities, as well as a cadre of graduate research assistants from the business school, organizational psychology, social psychology, and sociology. The free standing nature of ISR within the University of Michigan eliminates conflicting goals for the institute in terms of teaching versus research. Another significant technical factor is physical space. ISR has its own six story building which houses all of its staff. It provides a self contained unit which fosters interaction, integration, and a sense of permanence. This sense of physical and technical permanence provides an important momentum to research. Technical implications for other academic institutions desiring to foster action research include the importance of (a) creating a self-contained unit, with a clear research mission and non-conflicted goals; (b) permanent research support staff committed to the mission of the unit; (c) physical space to house the activities of the unit. Political Factors: The ISR based action researcher must acquire funding by way of grants and contracts with external agencies, public and/or private. His or her financial security is affected not by the decisions of department chairpersons as it is by the researchers ability to acquire research grants. As a result, the researcher is buffered from many internal academic pressures

17 having a great deal of autonomy and hence power. The link to the academy is not in terms of funding, it is in terms of collegial interaction. This reward system encourages entrepreneurship, yet it comes with a political cost, namely security. ISR tenure is based on the condition that ISR has funds in its overhead to help carry an individual. One can survive for a time without "covering" oneself, but not for too long. On the other hand, once a research scientist has acquired funding at ISR, there is greater time to do research as it does not conflict with teaching which is not done at ISR. With this as the political backdrop, lets focus on a few of the specific political factors which make ISR function effectively. First, there are nonconflicting goals such as teaching versus research found in most departments. This conflict is especially relevant to business schools which derive their revenues from large numbers of MBAs requiring a great deal of faculty attention. Second, resource allocation is unconflicted. In teaching academic departments there is often a crunch around secretarial, duplicating and support services with teaching taking priority. At both the Columbia Business School and the Michigan Business School, secretarial resources were given first priority to teaching needs, research was second place. Not only is the support staff not expert in research activities, they are given conflicting priorities adding to the hassle level of researchers. Third, ISR is selfcontained politically. It has its own funding base, and buys out the time of the faculty from academic departments or, as is the case for the majority of the ISR senior staff, the academic departments buy time from the staff whose primary home is ISR. The obvious other side of this coin is security. All ISR support and faculty research time is carried on soft money. The researcher must be sufficiently entrepreneurial to operate in this envi ronment.

18 Political implications for other settings include creating a unit which (a) has research as its primary goal allowing the allocation of scarce resources to unambiguously reflect this priority; (b) politically selfcontained so that researchers' time can be protected to do research. The explicit accounting of researchers' time necessary to operate such a unit also puts pressure on the individual researcher to be productive. Teaching demands are not a legitimate excuse for not delivering on one's research commitments as the research unit does not have teaching goals. One can visibly and explicitly succeed or fail in two separate arenas, research and teaching. The price to be paid for such a system is loss of security on the part of faculty. This is tempered some by appointments in which the faculty's home base is in an academic department where the individual may have tenure even though 60 to 70% of their annual time is allocated to ISR work. Cultural Factors: The culture of ISR reflects its historically bound tradition of producing useful knowledge. The founding fathers of ISR were of one mind in acknowledging the ethical imperative incumbent on social scientists to work for the betterment of humankind. Since action research is often a collaborative venture, the norms that encourage joint endeavors at ISR fit with the needs of the organization behavior field. Rensis Likert was well known to foster "supportive relationships" among colleagues. This attitude prevails at ISR today and is deeply imbedded into the culture. To conduct research of the highest quality, to be able to work cooperatively or along, and to serve society, these are the values which dominate the ISR culture. They are fully congruent with the values underlying action research in its various modalities. Cultural implications for other settings include having: (a) value position regarding the role of social science research in contributing to the

19 solution of real world problems; (b) set of norms supportive of collegial cooperation and collaboration. This culture needs to be enacted via the leadership of several key individuals who can act as role models. In sum, the technical, political, and cultural factors at play at ISR mutually reinforce each other to provide an environment which fosters action research. The implications for other settings are not to try and replicate ISR but to set up structures that create the technical, political, and cultural conditions suited to their own setting. One example of such an endeavor is the Center for Organizational Effectiveness at the University of Southern California which former ISR member, Edward Lawler has established. It is a good example of a small scale ISR which meets most of the technical, political, and cultural conditions identified above. Such units provide the environment for action research. An example of a typical action research project is presented below. It is the type of project easily carried out at ISR but would be extremely difficult to contemplate in a traditional academic department. A Case Example The following is a description of an action research project that, from the client organization's perspective, can be considered as an instance of the experimenting intervention orientation. From the academic/research institution's viewpoint, it is a collaborative effort that is confronting both theoretical and "real life" organizational problems. In a study currently underway at ISR, three senior research scientists, along with four graduate students, are studying a quality of work life (QWL) type program. In three plants of a 37 plant paper products corporation, a survey of attitudes (a version of the Michigan Organizational Assessment, MOA, created at ISR) was administered in Fall, 1981 for two reasons: (1) to use as

20 a baseline indicator against which to measure change, and (2) to inform the efforts of the consultants (non-ISR staff) who were to guide the action. Following the survey administration and feedback, which was conducted by a pair of ISR team members at each site (one senior person, one graduate student research assistant), the change agents entered and began their work. There were three change agents: one worked one site, the other two collaborated on the other two sites. The survey data guided the change effort, but to varying degrees at the different locations. Throughout the past year ISR has been monitoring events at each site by way of weekly telephone interviews with key people in the plants. We have also been receiving archival data (e.g. presentation handouts, attendance records) from the sites. In Fall, 1982 a second survey, as well as personal interviews and observations, will be carried out by ISR to: (1) assess the changes and attitudes resulting from the program, and (2) to guide its future development. These data will be part of the information base which a corporate steering committee will draw on to decide whether or not the pilot program at the three locations ought to be expanded to include other plants. In addition to the surveys, the corporate group will have the reports of the ISR evaluation and monitoring team, documentation provided by the plants and by the change agents, as well as their own impressions formed during site visits and through informal contacts with key players. Here then is a classic case of a data-based change program. It is an instance of experimenting intervention. Results of the pilot study in the three plants will play a large part in determining how best to employ QWL techniques in the other 34 plants of the company. Yet the data from the ISR surveys and interviews will serve not only the interests of the client, they are the stuff from which scientific contributions to organization theory are

21 currently being generated. At the time of the first survey, the ISR team developed a set of approximately 20 research questions for use as a framework for coding and analyzing the data. For each question or topic (e.g. role of the plant manager, effects of external forces, and outcomes of the program) we consider its technical, political, and cultural aspects separately. The change effort will be evaluated in terms of these three sets of factors. It is hoped that our analysis will provide insight into the dynamics of organizational change. The above case is offered as an illustration of how the ISR structure supports action research endeavors. The first contact with the client was initiated by the client. A representative of the organization, aware of ISR's work in the QWL area, expressed an interest in pursuing such a program. This occurred in 1978. An ISR research scientist, after a brief, initial diagnosis, told the client that the organization was not yet ready to engage in a QWL activity, not until it had dealt with significant problems in its pay system. The client agreed and requested a re-evaluation of the organization's readiness for a project. Negotiations with senior ISR staff took place, and a contract was agreed on. For ISR, the responsibilities were to suggest potential change agents, who would be cleared by corporate; to collect and feed back survey data at T, T2, and T3, to monitor progress at each site; to meet regularly with the corporate steering committee as advisors to the project; and to report findings and to make recommendations regarding the dissemination of QWL activities throughout the corporation. To accomplish these tasks, the ISR research scientists formed a team that consisted of two faculty from the Graduate School of Business Administration (one of whom has a joint appointment at ISR and the Graduate School of Business Administration) and one

22 research scientist at ISR, three in total. In addition, three students from the doctoral program in organizational psychology were hired on as assistants. Administrative support staff, including clericals, accountants, and data managers, were signed on to the project as well. By contracting with change agents external to ISR, the research effort took a step closer to the ideally objective stance required for valid assessment. The studies in the QWL series at ISR have followed this model: research team, change agent, and client remain autonomous though interdependent entities. The research instruments (i.e., T1 and T2 surveys) were designed with the participation of the client organization, which also signed off on the research topics that were decided on after T1. Many of the survey items (drawn from the MOA) have been used in similar projects conducted by ISR. The data from this project thus builds on an ever-expanding dataset that has become a most valuable resource for researchers in our field. Numerous and varied theories of organizational life can be, and have been, brought under empirical scrutiny with these data. Some of the technical, political, and cultural dynamics in the client organization and in the academic/research institution in this case are reviewed below. The project will affect the client organization and its technical system by the introduction of a program that may well lead to productivity gains and improved worker satisfaction. If the opposite result obtains, the organization will have learned that such a program may not be a wise investment for the future. In either case, learning will occur. Politically, the project may become a platform on which certain key players can argue for more control in directing the corporations's future. On the contrary, some may lose power if the project they backed is deemed a failure. Culturally,

23 the very initiation of a QWL project indicates at least a sense that the company's values may need examination if not alteration. The implied move towards a more participative style of management is encouraged by the project's methods and aims. From ISR's viewpoint, the opportunity to examine in depth a change effort at three separate sites, with two distinct change agent approaches, using refined research tools comparable to other studies, is quite a boon. In addition to enriching the databank, this project provides a training experience for students with some background in the area. Politically, the ability to draw support from the private sector is crucial in these hard economic times. And as the theory and practice of organizational behavior matures, the likelihood of attracting further support increases: proven theories and their methods are more attractive to clients than new, untested ideas about organizational change. Culturally, the project reinforces the norm of doing service to society. The effort to foster more participative management styles fits with ISR's historical concern for improving the effectiveness of and the quality of life in organizations. IMPLICATIONS AND SUMMARY In this paper we have stated a case for the importance of action research vis-a-vis the growth of our field. The technical, political, and cultural dynamics of host organizations and academic/research institutions were considered as they affect action research endeavors. Alternative arrangements for action research in these two kinds of entities were reviewed. ISR's structure was presented as one model for action researchers. Finally, an example of a recent action research project was described to illustrate the utility of the ISR model and to point to the technical, political, and cultural gains that may accrue to host organizations and to academic/research

24 institutions engaging in action research. The challenge before the organizational behavior field is to remain vital in a climate of constricting research resources while at the same time there is great potential demand for our services in helping organizations cope with change. As scientists we need to be able to generate knowledge, as participants in society we need to contribute; action research is a bridge which can provide the integration.

4 25 References McGrath, J. E. Toward a "theory of method" for research on organizations. In W. W. Cooper, H. L. Leavitt, and M. W. Shelley (Eds.), New Perspectives in Organizational Research. New York: Wiley, 1964, 533-557. Pascale, R. T. and Athos, A. G. The Art of Japanese Management: Applications for American Executives. New York: Simon and Schuster, 1981. Runkel, P. J. and McGrath, J. E. Planning a study: Settings and strategies for research. In P. J. Runkel and J. E. McGrath (Eds.), Research on Human Behavior. New York: Holt, Rinehart, and Winston, 1972. Staw, B. M. The experimenting organization: Problems and prospects. In B. M. Staw, Psychological Foundations of Organizational Behavior. Santa Monica, Calif.: Goodyear, 1977, 466-486. Trist, E. L. and Bamforth, K. W. Some social and psychological consequences of the long-wall method of coal-getting. Human Relations, 1951, 4, 3-38.

Exhibit 1 Forces For and Against Action Research in the Host Organization FOR AGAINST TECHN I CAL POLITICAL CULTURAL - Data-based change efforts can add - Control groups difficult to employ. value to the bottomline. - Measurement difficulties. - Development needed in staff skills. - Information is power; can be used Costs are high. by coalitions to advance positions. - Rewards are for profit and for shortterm. - Action researcher loses objectivity, conclusions not valid; overadvocacy - Japanese influence towards greater - Macho image of home run hitter. experimentation. - Quick fix mentality - Contribution to society; good citizenship.

Exhibit 2 Forces For and Against Action Research in the Academic/Research Institution FOR AGAI NST TECHNICAL POLITICAL CULTURAL - Opportunities to test hypotheses; - Field skills not taught in graduate to generate theory. schools - Applied fields require immersion in - Methods of research ingrained. "real life" settings - Limited resources from public - Rewards are for short-term. sector; need to go to private - Industry control of results. industry for support. - Tested theories can sell. - Norm of making contribution to - Applied work is'"se'lling out". society. - Academes and corporate executives - Heritage of action research have different styles. growing. _ | _ w as | | vow~~~~. -

Exhibit 3 Generic Orientations to Evaluating and Monitoring Organizational Change Type of Orientation 11,, Basic Research and Development Experimental Intervention Orientation Guidance System Orientation Managerial Audit Orientation Informal Anecdotal Orientation Cultural Depth of Commitment to Systematic Evaluation HIGH LOW _,, TRADE-OFFS Political Risk and Commitment of Resources HIGH LOW,.... -. Technical Sophistication and Capability HIGH LOW,,, W I. - ----

~' Exhibit 4 Behavior in Four Generic Types of Institutional Setting for Action Researchers BASIC APPLI ED INDIVIDUALISTIC COLLABORATI VE Research scientist Action worker developing theory in engaged in change ivory towers. efforts. Team of research Team of action scientists dealing workers tackling with theoretical organizational issues. problems.