Action research: a new paradigm for research in production and operations management.
(Lord Chesterfield's Letters to His Son, 4th October 1746)
Introduction - the debate on POM research
It is a sign of the vitality of an academic discipline if it frequently debates its role, its boundaries, and its methods. However, such debates can, on occasion, amount almost to a subject having a "crisis of identity". This appears to have happened each time that production and operations management (POM) has revisited its research agenda and methods. There has been a clear dichotomy, implied by the use of such terms as "traditional" and "non-traditional" methods, or by the opposition of research using "techniques" and research which is "more managerial", or by comparison of what has been researched with what those reviewing the situation feel should have been researched.
This debate on POM research is now very familiar, so we need here only to enumerate four main issues, which are:
(1) POM research has been traditionally based on modelling techniques. Although business and government can point to many useful cost-saving applications of such models, the wider relevance to most operations managers of the published results has been questioned.
(2) The influential developments in practice over the last decade or so (JIT, TQM, benchmarking) have not come from POM academics but from practitioners and consultants.
(3) Prescriptive solutions to well-defined problems have been pursued at the expense of broader contributions to theory. Partly for this reason, POM remains relatively poor in theoretical developments.
(4) There was a need for POM research to address problem "messes", interrelated issues across organizations which were not well structured.
In particular the need to do research which will be of practical value to managers, and which will be integrative rather than focused on a sub-system technique, has raised questions of research methods. One group of authors state: "It is time to expand our limited set of worn-out paradigms and consider new research methods from paradigms used in our sister fields". More recently, the editorial of the inaugural issue of a POM journal declared that "papers do not have to follow one of the traditional research paradigms". Because real world operations often face unstructured problems, which cannot be modelled but must be managed, the aim will usually be to compromise rather than optimize. Thus the traditional techniques of POM research, such as simulation and mathematical modelling, are unlikely to be the main methods for research in those situations. The principal research methods will be - or should be - the various types of empirical research.
The remainder of this article falls into four sections:
(1) Empirical research methods in POM. Surveys and case research are discussed briefly, then action research is defined and distinguished from other kinds of intervention.
(2) Theory building. The potential for theory building using qualitative methods, still relatively little known in POM, is explained using arguments drawn from other disciplines.
(3) Conducting action research. Some practical aspects of action research are described, drawing on the author's experience. The issue of methodological rigour is also addressed here.
Empirical research methods in POM
There are a variety of empirical research methods, and considerable experience in the social sciences of using them. Since operations management, involving as it often does people and groups in organizations, has some of the characteristics of a social science, this experience can be used. Other disciplines have used it, including the fairly new discipline of management information systems (MIS), with which POM also has affinities. But some POM researchers are assumed to be ignorant of these methods. Flynn et al. wrote their useful survey of empirical research methods in POM precisely because "many POM researchers do not have a strong foundation in gathering empirical data". But there is no need for POM to reinvent the methodological wheel; the social scientists and MIS researchers have not only used a range of empirical methods, they have also engaged in considerable debate about their relative merits in different situations. Hence this section will consider the major empirical methods partly by drawing on some MIS and social science literature.
Surveys usually provide a snapshot of practice or attitudes, across a number of respondents at a point in time. Data are collected by postal questionnaire and/or interview and analysed by standard statistical techniques. Subsequent surveys may be used to see how or if the patterns have changed over time. The survey is by far the most used of empirical data collection methods by POM researchers, perhaps because it is efficient in use of research time, or because the scope for hypothesis testing is a link with natural science and thus endorses a particular view of academic integrity. Certainly a major strength of survey research is that, with the appropriate sample size and structure, we can draw conclusions which are generalizable across different firms, industries, or countries. Its weaknesses include the possible bias of the sample of respondents, who, with a questionnaire only survey, will be partly self-selected, and there may be too little opportunity for dialogue to discover what questions should have been asked but were not. Also, for some critics of the method, "little insight is usually obtained regarding the causes or the processes behind the phenomena being studied".
A case study documents or records, in an appropriate degree of detail, the operational activity of a single organization. It has the merit of being integrated, involving all relevant variables, and clearly real world. The obvious defect is the difficulty of making valid generalizations beyond the individual case. Nonetheless major POM single case studies can be influential, especially when they are purposely non-representative, perhaps reporting major innovations in practice. Thus in recent years the studies by Monden of Toyota and Schonberger of Kawasaki have influenced both JIT research and practice.
Multiple case studies, by looking at several sites, attempt to reach more generalizable conclusions than those provided by a single case. But they inevitably suffer from the number of the variables that change from case to case, and the difficulty of interpretation that this presents. Nonetheless some tabulations can be made and the differences of the situations duly acknowledged. Their real value will be in building theory from the observation of practice. Case research is not always an efficient method; many visits may be needed to develop a comprehensive view of operations and to understand, in all the mass of detail, what is likely to be significant. Also the researcher's personality and experience, as well as his training and intellect, will partly determine his effectiveness in such research situations. As with surveys, the social scientists have produced much practical writing to assist the case researcher (see especially Yin). More recently, McCutcheon and Meredith have advocated the wider use of case studies in POM research and listed representative case research articles from POM journals over the period 1981-1991.
Action research can be seen as a variant of case research, but whereas a case researcher is an independent observer, an action researcher:
...is a participant in the implementation of a system, but simultaneously wants to evaluate a certain intervention technique. ... The action researcher is not an independent observer, but becomes a participant, and the process of change becomes the subject of research.
The intervention technique is adapted as it is used, and the understanding of its scope and limitations develop with each application.
The roots of action research lie in the models of learning developed by the social scientist Kurt Lewin. Lewin was the developer of field theory, which among other things emphasizes the importance of understanding the total situation rather than abstracting a few measurable variables from a situation. Other key ideas of Lewin's are the value of concrete experience in testing abstract concepts, and the role of feedback (a notion he took from electrical engineering):
Lewin and his followers believed that much individual and organizational ineffectiveness could be traced ultimately to a lack of adequate feedback processes.
He saw too that action research derived its effectiveness partly from the immediacy of the feedback that the research intervention receives.
Action research is thought to be most effective for technique development or theory building. Working closely with managers provides a depth of understanding denied to more objective methods. It is thus at the furthest remove from much traditional research, since as with a case study it cannot rigorously relate dependent and independent variables. As it cannot be objective, because the researcher is also part of that which is researched, it is less valuable for hypothesis testing. Nonetheless if we believe that POM is an area in need of a broader conceptual basis, then the theory-building potential of action research should make it a major method in the future.
Action research compared with other types of intervention
The class of POM literature which comes closest to action research is the considerable body of interventions reported by consultants and others in the journals of the American and British Production and Inventory Control Societies (APICS/BPICS) as well as elsewhere. These are usually successful implementations of established techniques, such as the large material requirements planning (MRP) implementation literature. Often the implementation is the point, showing how to adjust a standard approach for a specific setting. This literature normally differs from an action research approach in three ways. First, the goal is not to develop new theory, but to report an application. Thus even when innovative departures are being claimed, there is little theoretical underpinning. Second, they are usually only successes, and the paths and obstacles to success are rarely explored. As with a consultancy assignment the result is what matters, and this may give the impression that the technique is a panacea, and there is little assessment of what the resistance to, or setbacks of, the application might tell us about the technique itself. Third, there is rarely sufficient context or description of the implementation approach to permit the reader to assess generality or make comparisons with other reported situations. The consultants/researchers all work independently of each other and use few common formats or methods. Otherwise such studies would serve as base data for meta-analysis or other secondary research, but they rarely do because researchers cannot have confidence in comparisons based on this literature. Action researchers need to ensure that they distinguish their outputs from this literature and give the research community all that it needs to assess and use their findings.
It is important also to distinguish action research projects from consultancy assignments. A consultant shares a single common goal with the company, the completion of an analysis and/or the implementation of change. The action researcher will have this goal as part of a larger primary goal which the company may not share - the discovery of new knowledge. A consultant will specify the nature of his end result, often in detail, before he begins. He may even specify the means to achieve it. An action researcher needs to keep an open mind on these issues. For the consultant the destination is more important than the journey, the detours, the mode of conveyance. But for the action researcher these may be important and revealing. The consultant's timescale, costs, and attendance days will be specified in a contract. An action researcher needs to be wary of such ties, for one site may prove richer than another and attention may need to be switched elsewhere (to some degree). Any fee may be on a "contribution to research costs" basis. Above all, a consultant is likely to use established techniques, and not seek to develop new ones or to generalize to another setting. The action researcher seeks to develop and test new approaches and ultimately to establish their validity beyond the research setting.
A simple model of action research
Flynn et al. propose a model of empirical research in operations management (OM) [ILLUSTRATION FOR FIGURE 1 OMITTED]. This model reflects the authors' focus on the survey as the paradigm of empirical research, and they emphasize that "reliability and validity considerations underlie all stages". Action research plays little part in their survey of the available empirical methods, so their model has no obvious scope for relationships with companies, or for iterations within or between stages. By contrast, a model of action research (such as that proposed in [ILLUSTRATION FOR FIGURE 2 OMITTED]) allows for iterations within the implementation phase, since each application may lead to changes in the focus of the investigation, or even to new theory developments.
Theory building - traditional scientific method v. qualitative
research There are clear limits to the use of traditional scientific methods in POM empirical research. An experimental result in physical science is required to be repeatable if it is to be judged valid. It would be difficult to repeat exactly the conditions of a POM investigation, since much would change over time or between organizations. Also the new POM research agenda requires a holistic, integrated approach, to investigate how subsystems fit together. But the scientific approach, following Descarte's second rule, manages complexity by reductionism - breaking the whole into its constituent parts, which can then be understood individually. POM has already followed this reductionist route, which provides insufficient understanding of an operation in all its complexity. The reductionistic belief that the whole is the sum of its "researchable" parts cannot be held simultaneously with a view of operations management as essentially an integrated set of activities taking place in a dynamic social setting. In a situation of conflicting and changing objectives, and of trade-offs which may involve political as well as commercial judgement, the whole cannot be comprehended by looking at each subsystem in turn.
A perspective from organization behaviour
The organization behaviour (OB) research community has also been involved in this same "rigour v. relevance" debate, and for the same reasons. Scientific method did not permit them to capture or interpret the richness of human activity in organizations. Much of their work has been deliberately descriptive, recording what happens in particular contexts rather than testing theories. Yet this descriptive bias has undoubtedly added to knowledge, as in the way observations of managerial work allowed the "planning and control" theory of management to be supplanted. They have divided research methods into "qualitative" and "quantitative":
Qualitative researchers in contrast to their quantitative colleagues claim forcefully to know relatively little about what a given piece of observed behaviour means until they have developed a description of the context in which the behaviour takes place and attempted to see that behaviour from the position of its originator ... such contextual understandings and empathetic objectives are unlikely to be achieved without direct, firsthand ... knowledge of a research setting.
In such settings the scientific notion of rigour will be difficult to maintain, as Miles discovered. Wishing to make objective comparisons across several sites in a research project in the field of education, Miles describes the frustrations of trying to codify diverse situations. He concludes by asking:
What are the possible conceptual and organizational solutions to the steady tension between the unique, contextually specific nature of single sites, and the need to make sense across a number of sites? Must we trade close-up descriptive validity for accurate but "thin" generalization?
In POM we have not yet asked this question, and we need to ask it if we are going to do onsite comparative case or action research.
Still in the organizational behaviour field, Mintzberg takes a more assertive stance in defence of qualitative methods:
Organization theory has, I believe, paid dearly for the obsession with rigour ... Too many of the results have been significant only in the statistical sense ... . What, for example is wrong with samples of one? Why should researchers apologise for them? Should Piaget apologise for studying his own children, a physicist for splitting only one atom?
Mintzberg also emphasizes the role of induction over that of deduction. Deduction draws conclusions from certain premises on the grounds that to deny the conclusion would be to contradict the premises. Induction starts from premises about some observed things of a certain kind and draws conclusions about all the other things of the same kind. Inductive research has two steps: first, detection, the search for a pattern; second, the "creative leap", generalizing beyond the data. Here Mintzberg moves far from traditional scientific method:
There is no one-to-one correspondence between data and theory. The data do not generate the theory - only researchers do that - any more than the theory can be proved true in terms of the data. Our choice, then, is not between true and false theories so much as between more or less useful theories. And usefulness, to repeat, stems from detective work well done, followed by creative leaps in relevant directions.
This bold doctrine, which echoes Emerson's dictum that "the value of a principle is the number of things it will explain" contrasts greatly with the current position in POM research methodology. Either because of the reward system in academic institutions, or because of the relative narrowness of traditional POM research, we still cling to what we see as the safe haven of epistemological rectitude. Even Flynn et al.'s call for empirical research is principally devoted to the conduct of surveys, and appropriate scales to use and tests to apply. But surveys will not get us into companies, at least not with the frequency that will enable us to detect Mintzberg's "patterns" or produce insights into the complex world of the operations manager.
Creative theory building in POM
The body of theoretical knowledge in POM is still very small. We need to emulate the emphasis by the social scientists/MIS researchers on creativity in theory building in order to develop our discipline. Swamidass suggests how POM theories can be constructed and classified using empirical data and inductive reasoning At present though we have a negative attitude to theory; our "body of knowledge is essentially theoretically oriented" and "relatively little of such an approach is used in practice". But we must not, therefore, abandon theory, only the notion of theory as immutably true and scientifically provable. The economic order quantity (EOQ) body of theory was "true", in the sense of logical and internally consistent, and to some extent was applied in practice. But Japanese practice, dramatically reducing set-up time, undermined one premise of EOQ theory, namely that there was significant cost in placing a new production order on a system. What was a parameter - lengthy set-up times - became a problem to be solved by engineers, and the ground of the theory shifted. This changing of parameters into problems (characteristic of the Far Eastern business approach), of theory as something which is useful for a time but will be overtaken, seems antithetical to scientific thought. If POM academics are to have an impact on practice, they must embrace this creative tension between theory and practice. In particular, they must build theory from observation of, and sometimes involvement in, the practice of POM practitioners. New theories, to influence practice, must be developed out of practice.
Creativity in theory building may come in POM the way it has in other disciplines we have looked at, by skilled observation and constructive intervention. There are risks in this approach, as there are in what the social scientists call "grounded theory" or "being open to what the site has to tell us". This means evolving the framework rather than imposing one at the outset. There will be a conflict between the apparent freedom of such an unstructured approach and the need for clarity and focus. Too much focus too soon may deny the researcher a truly comprehensive understanding. But having no assumptions or working hypotheses at all at the outset would lead to confusion, not least in explaining the research aims to industrial collaborators. And, as Pasteur said, "In the field of observation, chance favours only the prepared minds". The experience of empirical researchers suggests that whatever the framework at the outset, it will - and should - evolve as data are gathered and ideas are developed. As POM is so concerned with commercial enterprises, we should accept that there is little profit without risk. And as more POM researchers gain experience in these "riskier" research methods, a methodological debate will develop such as has developed fruitfully in other disciplines.
Conducting action research
The views given in the preceding section come from the author's experience of conducting an extended piece of action research. In this section some practical aspects of the action research process followed in this project are described. This will permit POM researchers unfamiliar with the process, whether they are neutral or sceptical, to form their own opinion of its value.
The action research project into orderbook modelling
This research was in the fields of manufacturing planning and control (MPC), in small and medium-sized batch manufacture. The project was to design and install computer-based information systems for the management of operational priorities at the order progressing level. In batch manufacture the progress of customer orders from raw materials, allocation through manufacturing departments and inspection stages to despatch involves the co-ordination of a number of different functions. To facilitate this co-ordination we developed the "orderbook model", a single integrated data file containing all the key elements for identifying, grouping and progressing all the current orders of a company. In appearance and use this was rather like a spreadsheet, and in MIS terms was a type of relational database. The research aim was to see if there was a limited number of appropriate designs for such databases, and if so, what the key dimensions of the designs were. The project involved developing designs, via action research, in 17 companies. The taxonomy of orderbook models thus developed has been described in detail elsewhere.
This project and its conduct was influenced by many of the issues which the POM research debate outlined earlier had emphasized. It gave a central role to a management information system, and thus bridged POM issues and MIS/DSS concerns. Its goal was integration, and we took a broad "systems view", investigating procedures, paperwork and the roles and responsibilities of different line managers. The companies were very much faced with "problem messes" in the sense of having to find the best compromise in a changing situation rather than being able to structure problems in a way that permitted optimization. It is not necessary for the purposes of this article for the reader to know more about the details of this project. Suffice to say that the following section on conducting action research derives from the experience of the project, and all the specific examples given of the relations between company and researcher are drawn from it also.
Some practical aspects of action research in POM
Since there is no standard handbook of how to do POM action research, we began our project with some ideas gleaned from the MIS and OB literature, some experience in case research, and developed some practical rules of thumb as we gained experience. This brief section explains a few of these "rules", not in a prescriptive spirit, but simply to add to the sum of "traveller's tales" other researchers can select from:
* Agree the problem areas with the collaborating companies, but do not prespecify the solution. Managers will have (different) views on this, and the inductive/grounded theory approach requires you to keep an open mind, especially early on. "In the process of most preconceived research - such as that for testing hypotheses - data, which cannot be either forced or selectively picked, is discarded rather than used to correct the category". You cannot afford to discard the data that do not fit, since your goal is to generate theory from such data.
* Seek multiple viewpoints. This is a good stratagem in case research, and just as valuable in action research. Since you need a comprehensive understanding of the situation, and are likely to be concerned with unstructured messy problems, no one individual will see the whole picture. In particular, take a wide view of "the operation", which includes non-production areas such as design, estimating, purchasing, and so on.
* Record in simple semi-standard formats which permit you to compare different sites, but which do not prevent you from pursuing lines of investigation you did not foresee. (When the "miscellaneous" category has become the most used you know it is time to reshape the format.) Figure 3 shows the chart we used for recording what each department did at each stage of the life of an order. We would put a verbal description in each box as appropriate, and number the boxes to show sequence. Such simple devices give freedom but within a consistent framework.
* Have the company check your write-ups. This is not simply to ensure you have understood what you have heard, but to give managers a chance to enhance their descriptions. In the UK at least managers do not always have time to write much to researchers, so further interviews or phone conversations are sometimes necessary.
* Prefer data to opinion. We were sometimes told, by different managers in the same company, that they had "too much inventory", "ran a very tight inventory control policy", and "risked running out of materials". Also check these claims against detailed data as well as the aggregate versions managers use. Checking sample inventory control sheets, routing descriptions, statistical process control (SPC) charts, etc., and discussing them with those who use them reveals important practices hidden in more aggregate reports.
* Remember that opinions are also data. In the previous example, each of the three viewpoints was "correct", given the role of the manager concerned. As with case research, one task is to record different views of a situation before making the intervention which distinguishes action research.
* For many of the above reasons, it is difficult to use an assistant to do much writing-up or coding. As the grounded theorists have also found, only the principal researcher(s) will usually have sufficient grasp of the whole situation.
* Give some thought to the right frequency of site visits. For our particular project, about two half-day visits per month per site seemed to provide the right momentum. You will not be forgotten, you will not be a nuisance, and the interval gives time for the company and yourself to do the necessary development work in between. Then each visit can see some progress. Also each visit can feasibly end - as it should - by agreeing the tasks to be achieved on both sides by next time.
Doubtless these "rules" would need to be varied for different projects, but the message behind such a list will be the same: in action research, the research process itself needs to be proactively managed. Indeed, the quality of the results may depend as much on research project management as on research design or results analysis.
Methodological rigour in the conduct of action research in POM
If POM is to develop new theory by researching broader issues using "nontraditional" methods such as action research, the question of research validity must be raised, as it has been in MIS and social science. Traditional POM research has the merit of perceived rigour, because it is based on methods of investigation into natural science. Under certain assumptions and conditions, aspects of physical systems (inventory or scheduling, usually) can be investigated "in the laboratory" (or computer suite). Hypotheses can be tested and data subjected to statistical analysis. But as we noted earlier, some have felt that this rigour has been achieved at the expense of relevance. Can relevance be gained without the sacrifice of rigour?
As we have seen, action research is a fairly close relative of case research, and raises similar questions of methodology. Persuasive discussions of the validity and reliability issues in case research can be found in Yin and McCutcheon and Meredith. Here we need only note that these authors and the earlier ones they cite have by now established case research as a truly scientific research method if properly designed and conducted. Our concern though is with action research, which because it involves the intervention of the researcher, raises an additional methodological problem beyond those of case research.
"From the action researcher's perspective, the challenge is to define and meet standards of appropriate rigour without sacrificing relevance". The fact that action research is non-traditional in POM does not excuse it from the methodological requirements of other approaches. Action researchers must take pains to ensure, as far as is possible given the central role of intervention, that their research method is rigorous and their results general. Appropriate rigour in our project meant developing a consistent approach, employing tools which enabled the commonalities and differences of our various sites to be accommodated and acknowledged as our examples accumulated. This approach was documented fully enough to be used by other researchers, making the intervention technique itself transferable. Subjectivity, the main methodological weakness of action research, was further mitigated by having the team of three researchers use different pairings, to reduce personal bias in onsite work. Also managers themselves were co-designers of the information systems and those managers were of course different at each company. Put simply, subjectivity is minimized by having more subjects.
The action research project aimed at developing a general method for designing priority management systems, and to establish a limited set of system designs. A degree of generality was achieved by having several applications in different industries. Although the situations varied, the approach was consistent. Eventually, enough examples were created to suggest that there was not an infinite number of designs, but a limited set of types, a taxonomy. As with case research, the more situations that are studied, the better the prospect for generalizable results. Also frequent access to different viewpoints in the same company led to the development of theory about priority management. As Argyris et al. suggest, managers in these situations can articulate a rationale for their actions which, repeated or varied in different settings, can be very stimulating to a researcher interested in developing theory.
In the action research paradigm the quality of the learning is bound up with the effectiveness of the intervention and the role of company management in collaborative developments:
Intervention is the action science analogue of experimentation. When clients involve themselves in change experiments, they engage in non-trivial learning, and they think and reflect seriously on what they are doing.
This source of learning is not available to other methods, including case studies.
This attempt to administer a treatment can be a powerful device, especially if it occurs across different settings. Apart from ensuring relevance to practice, there will be other gains from the involvement of the company itself and the variety of modes of communication between researcher and the collaborating organization. Interviews, checking transcripts, posting data disks, meeting to examine progress will all play a part. This helps give validity; any misunderstandings or wrong assumptions have multiple opportunities to get exposed and corrected. The involvement of the company also gives momentum; having to supply something management needs (the main contrast with a case study) ensures that the contacts are suitably frequent. But above all, the method provides insight, working with different levels in the organization, over a period of many months, but being a non-threatening outsider, can mean becoming the confidant of all parties. For example, we increasingly felt the pressure for a new information system to become both a sword and a shield: a sword for senior management, who wished to have data to expose any lack of responsiveness among production departments; and a shield for the managers of those departments to provide evidence to defend their decisions. This type of insight would not be available to the case researcher, since it comes from acting as a change agent rather than as recorder of the existing situation.
The debate on POM research has claimed that POM academics have failed to influence practice largely because traditional research topics and methods have not produced results which are:
* broadly relevant to practitioners;
* applicable to unstructured or integrative issues;
* contributions to POM theory.
This article has put forward the claims of action research to help overcome these three deficiencies. While most empirical research (and much traditional research) should be able to provide managerially relevant outputs, the relevance of action research is usually guaranteed by working with management on an issue the enterprise itself wants to address. Only site-based methods (case research and action research) will give much insight into unstructured "messy" problems, and action research is seen (at least outside POM) as especially valuable for theory building in complex situations. The grounded, iterative, interventionist nature of action research ensures closeness to the full range of variables in settings where those variables may not all emerge at once. The methodological problems raised by a qualitative approach in such settings have been tackled in MIS and OB, and POM action researchers can fruitfully draw upon findings from those disciplines. Last, action research requires us to be creative, because it is usually conducted to develop a new approach or solution to a situation for which there is no existing prescription. For that creativity to lead to new theory, however, an holistic and inductive approach will have to replace the reductionist and deductive approach of traditional scientific method.
This article would not perhaps have appeared in an international POM journal a few years ago because of the focus on particular methods and topics, and the dichotomy (even hostility) between traditional and non-traditional research. But although the article has concentrated on action research as one way forward, any dichotomy with traditional research is false. There are surely roles for all paradigms and what our field really needs is genuine pluralism. Different research methods need not be mutually exclusive - multimethod research designs will still be especially valuable, the various paradigms compensating for each others' weaknesses. In the short term, however, there will need to be a shift to more empirical work, which will have implications for the training of POM researchers.
But the focus on applied work and action research projects that we need now may just be part of a cycle. Perhaps as production systems, through CIM and associated developments, become more controlled there will be an enlarged role for traditional analysis of well structured problems, while empirical studies move into less defined areas. As the British writer Cyril Connolly states:
Truth is a river that is always splitting up into arms that reunite. Islanded between the arms, the inhabitants argue for a lifetime as to which is the main river.
Perhaps the POM profession has become like Connolly's islanders, and we should accept that all the tributaries of POM research ultimately contribute to the flow of increased knowledge. But some channels silt up, while others broaden and carry more traffic. If the traffic we wish to carry is to include practitioners, then the stream of empirical research, and of action research in particular, must now be recognised as the main river.
1. Meredith, J.R., Raturi, A., Amoako-Gyampah, K. and Kaplan, B., "Alternative research paradigms in operations", Journal of Operations Management, Vol. 8 No. 4, October 1989, pp. 297-320.
2. Singhal, K., "Introduction: shaping the future of manufacturing and service operations", Production and Operations Management, Vol. 1 No. 1, Winter 1992, pp. 1-4.
3. Flynn, B.B., Sakakibara, S., Schroeder, R.G., Bates, K.A. and Flynn, E.J., "Empirical research methods in operations management", Journal of Operations Management, Vol. 9 No. 2, 1980, pp. 250-84.
4. Galliers, R.D., "In search of a paradigm for information systems research", in Mumford, E. et al. (Eds), Research Methods in Information Systems, Elsevier Science Publishers, Amsterdam, 1985, pp. 281-97.
5. Monden, Y., Toyota Production System - A Practical Approach to Production Management, Industrial Engineering and Management Press, Norcross, GA, 1983.
6. Schonberger, RJ.,Japanese Manufacturing Techniques: Nine Hidden Lessons in Simplicity, The Free Press, New York, NY, 1982.
7. Yin, R.K., Case Study Research: Design and Methods, rev. ed., Sage Publications, London, 1989.
8. McCutcheon, D.M. and Meredith, J.R., "Conducting case study research in operations management", Journal of operations Management, Vol. 11 No. 3, 1993, pp. 239-56.
9. Benbasat, I., Goldstein, D.K. and Mead, M., "The case research strategy in studies of information systems", MIS Quarterly, September 1987, pp. 369-86.
10. Kolb, D., Experiential Learning: Experience as the Source of Learning and Development, Prentice-Hall, Englewood Cliffs, NJ, 1984.
11. Wood-Harper, T., "Research methods in information systems: using action research", in Mumford, E. et al. (Eds), Research Methods in Information Systems, Elsevier Science Publishers, Amsterdam, 1985, pp. 169-91.
12. Mintzberg, H., "An emerging strategy of 'direct' research", Administrative Science Quarterly, Vol. 24, December 1979, pp. 582-89.
13. Van Maanen, J., "Reclaiming qualitative methods for organizational research: a preface", Administrative Science Quarterly, Vol. 24, December 1979.
14. Miles, M.B., "Qualitative data as an attractive nuisance: the problem of analysis", Administrative Science Quarterly, Vol. 24, December 1979, pp. 590-600.
15. Swamidass, P.M., "Empirical science: new frontier in operations management research", Academy of Management Review, 1991, pp. 793-814.
16. Wild, R., "Decision making in operations management", Management Decision, Vol. 21 No. 1, 1983.
17. Glaser, B. and Strauss, A., The Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine, Chicago, IL, 1967.
18. Westbrook, R.K., "Orderbook models for priority management: a taxonomy of data structures", Journal of Operations Management, Vol. 11 No. 2, June 1993, pp. 123-42.
19. Glaser, B., Theoretical Sensitivity: Advances in the Methodology of Grounded Theory, Sociology Press, Mill Valley, CA, 1978.
20. Argyris, C. and Schon, D.A., "Participatory action research and action science compared a commentary", in Whyte, W.F. (Ed.), Participatory Action Research, Sage Publications, London, 1991, pp. 85-96.
21. Westbrook, R.K., "Priority management: new theory for operations management", International Journal of Operations & Production Management, Vol. 14 No. 6, 1994.
22. Argyris, C., Putnam, R. and Smith, D.M., Action Science, Jossey-Bass, Oxford, 1985.
|Printer friendly Cite/link Email Feedback|
|Publication:||International Journal of Operations & Production Management|
|Date:||Dec 1, 1995|
|Previous Article:||Developing effective strategies to prioritize set-up reduction in a multi-machine production system: a throughput approach.|
|Next Article:||The role of management accounting in the development of a manufacturing strategy.|