You are on page 1of 8

Hypothesis Creation in Organizational Behavior Research'

CRAIG C. LUNDBERG Oregon State University

The field of organizationalbehavior has begun to place undue emphasis on hypothesis testing and thereby underemphasize hypothesis generation. After suggesting several reasons for this, this paper reviews how useful research ideas are discovered and the prerequisites for hypothesis creation, and then catalogues hypothesis generation. Suggested criteria for acting on new research ideas and questions are: extensity, professional encounter, mental experiments, and timing.

In the development of any field or discipline, there comes a period where "respectability" insidiously reigns. During this period, the earlier spokespersons for the field are increasingly supplanted by a cadre of influential younger researchers who are usually sophisticated in methodology a s well a s ambitious. Subtly, choices are made which profoundly affect the field; methodological rigor and precision become prized over phenomenological significance, researching over scholarship, conceptual edifices for scientists only, and reification over elucidation. The field of organizational behavior
Craig C. Lundberg (Ph.D. - Cornell University) is Professor of Behavioral Science and Administration, School of Business, Oregon State University. Received 6/30/75; Accepted 9/9/75; Revised 70/24/75.

appears to be s:;pping into such a period at the present time. 2 Why is this so? The position taken here i s straightforward and, therefore, probably an oversimplification. The contemporary nature of organizational behavior knowledge i s primarily the product of how researchers are trained for their craft. This paper comments primarily on the growing emphasis attributed to hypothesis testing a s opposed to hypothesis creation or generation. It secondarily takes note of a variety of ap1 The author wishes to acknowledgt the challenging critique of John W. Hennessey, Jr. as well as the provocative discussions with Barbara Karmel and Kurt Motamedi in the preparation of this paper.

An eloquent appraisal which concludes that many of these same trends are occurring for industrial and organizational psychology is made by Bass (1).

Hypothesis Creation in Organizational Behavior

proaches for discovering new ideas for research questions and hypotheses, thereby modestly contributing to a portion of the literature that i s at present meager and uneven.

Contemporary Research Thought in OrganizationalBehavior


Contemporary organizational behavior researchers have misplaced their attention in the research process, resulting in a dysfunctional emphasis on hypothesis testing over hypothesis creation. Just a couple of decades ago the field of organizational behavior was a diffuse amalgam of applied sociology, applied psychology and human relations, before coalescing a s organizationa l behavior. Being "well-trained" as a researcher in organizational behavior meant beginning with cases and/or clinical work. A considerable portion of one's early research apprenticeship was spent "in the field," gaining what Henderson (8) has called, "an intimate, habitual, intuitive familiarity with'things" of everyday experiences. After being well grounded in one's chosen phenome-non, one judiciously acquired more systematic and rigorous methodologies (11). Theoretically speaking, these extensive field experiences prompted low and middle-range, conservativeconceptual schemes. To contrast the above with a caricature of contemporary organizational behavior research, the problems selected today a s worthy of investigation most often are derived from review essays and abstract models. Wide-band research approaches are foresaken for a preoccupation with precision in measurement. Basic paradigms are likewise forsaken for methods, and external validity i s widely ignored. An alternative characterization i s the stereotype that has developed among both scientists and practitioners as to what constitutes "serious" organizational behavior research. It goes (somewhat ironically) like this: crucial studies growing out of previous findings are performed with precision and elegance. The results are subjected to the closest scrutiny and analysis, with all alternative inter-

pretations judiciously considered, and accepte_d or rejected in accordance with the canons of scientific rigor. Finally, the now confirmed discovery i s inserted i n a systematized lattice of already available knowledge to complete a forward step toward humanity's mastery of the unknown. The imagery promoted above contrasts in both the style and the strategy of research. On the one hand, research i s seen to be a messy affair, a meandering, blundering, often serendipitous adventure, with the objective of ever more adequate descriptions of reality. On the other hand, research i s seen as a highly formal, preplanned and programmed activity in the quest of certainty. Von Mises (17) provides a label of this in the extreme; "apodietic certainty" is when there i s absolutely certain certainty. This latter image i s seemingly based on the assumption that the more formal and sophisticated the methodological preparation one has for research, the inore productive and creative one will be in performing the research: an assumption which clearly depends on the nature of the preparation. There are two primary reasons why contemporary research preparation for organizational behavior i s inadequate and therefore contributes to an enactment of the stereotype noted above. The first is that the research training offered today lacks effectiveness in several ways. Basic methods tend to be relatively ignored, basics such a s interviewing and observation, and how to cull data from documents. Methods that are taught are probably obsolescent procedures. For example, methodology courses continue to emphasize or maintain a rigorous distinction between dependent and independent variables. Examples with two variable or few variable designs are used, and the assumption of continus i s the ous variables i s constantly reinforced, a setting of equal numbers in equal'intervals, etc. Yet it seems that the complexity of individuals and social systems requiresmot only more use of multivariate time-series designs, but also more attention to parallel processing, bi-directional

Academy of Management Review - April 7976 relationships and feedback circuits as well a sa much stronger attention to time. '(A similar argument for psychology i s made eloquently by McGuire (121.1 Lastly, some techniques which have led to progress are mistakenly taken a s a kind of master key to inquiry in general; that is, some successful technique i s assumed to be generally fruitful, as Handy and Harwood note, when inquiry is understood as beginning with a "well-formulated" hypothesis and then searching for evidence, or when mathematical transformations are assumed to be the essence of scientific inquiry, or when logical deductions are emphasized (6, p. 3). The second detracting influence i n the training of researchers comes from the literature itself. Reading vast numbers of research reports and studies can be grossly misleading, not because of their content but because of their structure. This applies to both quantitative (15) and qualitative (10) reports. Research reports are, of course, the context of justification and use logic for proof or what Kaplan calls "reconstructed logic" (9). Both the context and the logic which are by custom the proper format for reporting have little to do with the initiation of research projects. In initiating research, one i s interested i n the context of discovery, not justification; and one's intuition (logic-in-use) is more prominent than one's rational logic. Intuition here i s something preconscious and outside the preferred inference structure. Attempting to be overly formal or trying to make phenomena or reality fit preconceptions of order and construction too early (i.e., to resemble research reports) has obvious dysfunctional consequences for discovery. At the extreme this i s termed "ratiocination": the belief that important knowledge can be obtained independently of observation or experience. The above lament does not imply that the training of organizational behavior researchers should regress; rather the plea i s simply that this training reintroduce some of the basic phenomena-grounding activities of earlier days, a s well as promote the acquisition of systematic and rigorous methodologies, and that the latter be more compatible with the phenomena to be understood. In other words, all that i s urged i s that an o l d epistemological cavea't b e recalled-in successful inquiry, neither measurement nor conceptual work exists in isolation from one another nor has primacy over the other.

On idea Generationand Hypothesis Creation


The conventional initiation of research in behavioral science tends to use one or more of the bases succinctly noted by Webb. Curiosity, confirmability, compassion, cost, cupidity, and conformability-or, more sim: C a n I get an answer," ply, "Am I interested," ' "Will it help," "How much will it cost," "What's the payola," "Is everyone else doing it?" (18, p. 223). Webb then goes on to show that:

. . . these bases, used alone or in combination although perhaps correlated positively with a "successful" piece of research, will probably have a zero or even negative correlation with a "valuable" piece of research.
Research depends on ideas, and valuable research comes from ideas for really new questions and hence new hypotheses. Experienced scientists would agree with Taylor's contention that: . ..worthwhile ideas do not come fullblown in all their glorious maturity out of an empty void. The process of getting and developing ideas i s undoubtedly a confused mixture of observation, thinking, asking why, cherishing little unformed notions, etc. (16, p. 172) New research questions seem to result from examining our assumptions and a combination of passive observation, putting questions to nature and active observation. Weick (19) argues that what i s vital for stimulating research i s the transformation of our assumptions i n t o questions. By assumption he means that which we "know" i s there in reality. Thus, for example, instead of studying the relationship of an

Hypothesis Creation in Organizational Behavior

individual's utility function to another construct, one would ask: i s it the case that individuals have utility functions? If utilities don't exist, what, if anything, i s guiding preference ordering and decisioning? Suppose that what occurs i s that goals get progressively clarified in decisioning. If anything like this i s going on in people, researchers are missing it; and the reason they are missing it i s because they assume that individuals have utilities. Observation i s neither a mysterious nor an erudite process. Basically, it boils down to keeping one's eyes and ears open and making a cecord. Passive observation occurs by chance or spontaneously. Active observation i s induced by some sort of preconceived idea, however roughhewn i t may be.(& p. 7). Questions put to nature then stimulate active observation. One must poke into all sorts of silly things and risk looking foolish, even stupid, at times. Active observation often comes down to a simple playfulness, having fun with ideas, theories and oneself. Weick, for example, prescribes the following ways to discover new questions of organizations: fondle building blocks, burn caribou bones, count statues, communicate non-verbally, and construct flimsy objects (19). Playfulness i s necessary to counteract the fact that we are all brought up to think we know our own minds and what they are up to at any given moment (7, p. 77). While we obviously cannot list the questions put to nature, we can speak to their form. There s follows: seem to be six basic question forms a What i s an X? X i s asserted. I s it so? Where does X occur; what i s the distribution of X? What are the similarities and differences be, tween X 1 and X 2 . . . X ? What i s associated with X? What causes XI or what does X cause? In these questions, X i s a descriptor of some phenomenon. Actually, these questions indicate two quite different kinds of research work. The first four define descriptive efforts (the identifi-

cation and classification of various elements). The last two define theoretical work (discovering the relationships among elements). Before discussing hypothesis creation, there are four prerequisites for suchactivity: acquiring a "knowledge of acquaintance" of the phenomena, really knowing the subject, possessing an ingrained paradigm, and the ability to "galumph." Acquiring and reacquiring a knowledge of acquaintance (14, p. 7), that i s a firsthand familiarity, of one's focal phenomenon offers a grasp of it, useful for countervailing more abstract and analytical knowledge. Possessing a thorough knowledge of the subject may appear obvious, although it i s often side-stepped by the more eager researcher. There i s an old saying that a discovery i s an accident finding a prepared mind. Clearly creative insight occurs more frequently with a thorough knowledge of one's subject area than a s a bolt from the blue. The third prerequisite, possessing an ingrained paradigm, is less well appreciated. What is suggested here is the ability to think unconsciously in accord with a fundamental model. For example, many sociologists are structural functionists, whether they are aware of it or not. Perhaps in organization behavior the paradigm is that of a cybernetic open system. "Galumphing" (131, the last prerequisite, i s the psychological process of voluntarily placing obstacles in one's own path, where such deliberate complication of a process becomes interesting for itself and not under the dominant control of goals. Galumphing as a prerequisite to hypothesis creation implies a strict avoidance of task-oriented efficiency. When a question put to nature i s formalized by specification of constructs or concepts and their supposed relationship, this conjecture i s called a hypothesis, and hypothesis formulation or generation can come about in many ways. In the listing which follows, the emphasis will be on approaches other than theory testing. The approaches will be discussed in three sets: first, those that are most clearly exploratory efforts; then those approaches that exhibit an intentional search; and lastly, those efforts that extend

Academy of Management Review - April 1976 antecedent research, that is, by the coupling of one project to another. mind. A deliberate cessation of effort to solve a problem often leads unexpectedly to a new perception. More pertinent to the discussion, it sometimes leads to the discovery of new problems or questions that are likewise surprising.

Exploratory Approaches The first approach i s probably improperly termed exploratory, for i t concerns happenstances or accidents. Accidents are not just the unexpected, but include those events resulting from a failure or a blunder in a process which provides an opportunity for seeing some phenomenon in another way; Accidents have been dignified in science by the term "serendipity." The paradoxical incident offers another opportunity for arriving at an interesting hypothesis. Something i s noted that does not make sense according to our general understanding of social events; that is, i t appears paradoxical and efforts to account for it can lead to new discoveries. For example, the problem of performance decline was unsatisfactorily "explained" by motivation models and this led to the recognition of activation/arousal. The intensive case study i s an approach long acknowledged as a fruitful source of hypotheses. The case need not be an unusual or exceptional one, for it seems that almost any case studied intensively might serve. The early history of organizational behavior, i.e. its "human relations" phase, held numerous examples. One wonders what the impact on the field would be if another period of such case studies occurred. The analyzing of a practitioner's or craftsman's rule of thumb i s another source of hypotheses. The procedural rule of thumb i s assumed to work, and the research task i s to think of theoretical implications for its effectiveness. Of course, they may be as suggestive by their failures a s by their successes. Many prominent ideas in the field, e.g. "satisficing", came about in this way. The last truly exploratory approach can be labeled "thinking around". This refers to the activity of letting go of one's focus on whatever problem or specific phenomenon one i s working on and permitting oneself to speculate and fantasize about any related matters that come to

Intentional Search Approaches


Prior research efforts often provide the opportunity for hypothesis creation from their byproducts. Anyone who searches his or her own research or the research of others will find unexpected behaviors or patterns. How often research reports contain phrases which begin "In addition . . .'I or "It was also noted that . . ." These incidental observations or those recorded by others can often be mined for new questions or hypotheses. Similarly, there are methodological by-products, such as the idea of using a technique in a different context or with a different problem, or new ways of assessing or controlling variables in the often hard and ingenious work of research. The intentional use of analogy i s another approach. Here one simply takes the properties, patterning or functioning of some familiar subject and asks whether some other topic exhibits any similarity. Well-known examples include the computer as an analogy to cognitive processes, the biological analogies to organizational growth and development, and the economic transaction as an analogy to social exchange. Less in evidence in organizational behavior i s the approach called hypothetico-deductive. This approach involves putting together two or more common sense principles or empirical findings and deriving from their conjunction some predictions of interest. This hypothesis generating procedure has become increasingly popular and possible with the advent of computer simulation. The contextual twist i s one approach that seems underappreciated in organizational behavior. Thinking about replicating a study in a very different context can often point to new interactions or condition general findings. Moving research into the laboratory from the field or

Hypothesis Creation in OrganizationalBehavior vice versa, as well a s replicating a project in a different. business function, industry, institution or culture, can be provocative of new ideas. Seldom do research findings lend themselves to single explanations. Thus, another source of hypotheses i s by thinking through the possible additional interpretations of any set of data. A closely related technique for provoking new hypotheses i s to try to account for conflicting results. Another creative method is to attempt to account for the exceptions to general findings. Still another approach related to these efforts of developing new hypotheses by examination of prior work is to attempt to reduce observed complex relationships to simpler component relationships. Finally, in the cataloguing of intentional search approaches to hypothesis creation, the manipulation of scientific statements must be mentioned-specifically, the two forms of questions listed earlier, defined as theoretical, and their more complex kin. Scientists tend to put their knowledge in the form of statements, and any untested statements are hypotheses. The manipulation of statements of association simply involves substituting some new factor or variable for one already established (including the specification of new conditioning variables). For "causative" statements, manipulation could consist of reversing the statement (e.g., satisfaction and productivity), substituting variables, the postulation of intervening variables, etc. This approach, called the manipulation of statements, i s close to the next set of approaches to be discussed, the explicit coupling of research to research. the findings, observations or capabilities of others (18). Actually these approaches can be rather quickly identified. Independent variables can be substituted, added to, controlled better or differently, or reduced in number.Similarly, dependent variables may be replaced, added to or otherwise modified (i.e. altering the level of measurement or refining the selected behavior). There are also the obvious surplus or additional relationship variables, the so-called intervening variables. The point here i s that these changes in statements deliberately reflect other research and therefore "couple" prior research to that being initiated.

An Amendment on Implementing New Research Ideas


Assuming the discovery of a new and interesting idea for either a question to put to nature or for a hypothesis: then what? While experienced researchers will exhibit stylistic uniqueness, the following advice seems to have wide acceptance. Any new idea must be examined in light of a critical requirement-for while any new idea may produce a doable or even a successful research project, to increase the probability that it will result in a valuable project, one must ask of it, will the results be generalizable? That is, in how many and what kind of specific circumstances will the relationship expected to be confirmed in the research be likely to hold? This i s the requirement of "extensity." If the idea produces findings applicable to a vast heterogeneity of events in time and space, the findings are likely to have a great value. If a question or hypothesis seems to possess extensity, two activities are urged: professional encounter and the mental experiment. The professional encounter takes the two forms of talking with colleagues and searching the relevant literature. Such talk can generate leads to other investigators working on the same or similar problems, pertinent parts of the literature, alternative explanations, etc.

Extending-Coupling Approaches
Research often evolves from research. The two prior sections have emphasized the more obviously creative approaches. This section notes approaches to hypothesis generation which most explicitly build on previous research, forming a well-defined- line of investigation (5). In general, this means extending research by making a finding (or technique) interactive with

Academy of Management Review April 7976

11

A n extensive literature search, an arduous and often dull task which i s often done inadequately, i s more than collecting citable references. It is also a place to test questions and hypotheses without going into the field or the laboratory. How does this occur? One way is to go to the classical pieces of research in one's field, which may or may not have any apparent relationship to one's focus, and then read those research reports carefully once again, trying to see whether the phenomenon that one i s interested in was in fact in that previous research. It just didn't happen to be of direct interest. Second, if one takes key terms and asks what the synonyms are and then goes to the literature and looks up those synonyms, one may find all kinds of things hidden away that would not have been otherwise noticed, especially what Blumer calls additional "sensitizing concepts" (3, p. 5). The second activity urged i s called the mental experiment. It consists of putting one's feet up and staring at the ceiling and imagining the research project in some detail-envisioning in a detailed and descriptive way what i s involved and what one finds. The phrase "mental experiment" i s appropriate because the project i s best fantasized a s an experiment: if I did or didn't do something, what phenomena would appear? This kind of mental work lets one get a better hold, not only on the phenomena but also on the processes and the implicit model. The last suggestion relates to time. There i s a great pressure to rush into print these days, particularly among younger researchers. This can be dangerous. A caricature of an older style of

scholarship makes the point. Here a researcher talked to colleagues, carefully and thoroughly looked at the literature, thought through one after another mental experiment and then, and only then, wrote down a set of notes, roughly defining the research, probably slighting methodology and the finer details, concentrating on the phenomena and the questions. The key activity at this point i s to put those notes away for a while. Go on and do something else; then after the passage of time, pull out the notes and read them with somewhat fresher eyes a s well a s pass them around to colleagues. Then, if it still seems reasonable and makes sense, formalize the research proposal.

Concluding Comments
Organizational behavior research today i s fast becoming complex, costly and overly formalized. Some redirections in research training and a renewed emphasis on hypothesis creation are called for. While it i s likely that we cannot all construct basic paradigms or initiate crucial experiments, we can at least try to be less trivial. When i t comes to doing research that i s enduring and critical, we can, paraphrasing Webb, s we can, believe in new ways, learn as much a more effectively discover new ideas, seek a s great extensity in our variables a s we can, and never lose sight of the phenomenon (18). Doing research in organizational behavior i s just like doing science in general; and that, according to Bridgman, i s simply "doing one's damnedest with one's mind" (4, p. 460).

REFERENCES
B. M. "The Substance and the Shadow," American Psychologist, Vol. 29 (1974), 870-1386. 2. Bernard, C. An Introduction ro the Study of Experimental Medicine (New York: Dover Publications, 1957). 3. Blumer, H. "What is Wrong with Social Theory," American SociologicalReview, Vol. 19 (1954),3-10. 4. Bridgman, P. W. Yale Review (19541,444-461. 5. Cartwright, Dorwin. "Determinants of Scientific Progress: The Case of Research on the Risky Shift," American Psychologist,Vol. 28 (79731,222-231.
1. Bass,

6. Handv, Rollo, and E. C. Harwood. Useful Procedures of Inquiry (Great Barrington, Mass.: Behavioral Council, 1973). 7. Hebb, D. 0. "What is Psychology All About," American Psychologist,Vol. 29 (1974),71-79.
8. Henderson, L. I. Introductory lectures. Unpublished lectures in "Concrete Sociology", 1938.

9. Kaplan, A. The Conduct of lnquiry (San Francisco: Chandler Publishers, 1964).

Hypothesis Creation in Organizational Behavior


10. Lofland, John. "Styles of Reporting Qualitative Field Research," The American Sociologist, Vol. 9 (1974), 101-111. 11. Lundberg, C. C. "New Directions for Personnel Research," Personneljournal, Vol. 41 (19621,497-504. 12. McCuire, W. J. "The Yin and Yang of Process in Social Psychology: Seven Koan," Journal of Personality and Social Psychology, Vol. 26 (19731,446-456. 13. Miller, 5 . "Ends, Means, and Galumphing: Some Leitmotifs of Play," American Anthropologist, 1973,87-98. 14. Roethlisberger, F. I., et al. Training for Human Relations (Division of Research, Graduate School of Business Administration, Harvard University, 1954). 15. Signorelli, Anthony. "Statistics: Tool or Master of the Psychologist?" American Psychologist, Vol. 29 (19741, 774-777. 16. Taylor, D. W., et al. "Education for Research in Psychology," American Psychologist, Vol. 4 (1959),167-179. 17. von Mises, Ludwig. Human Action: A Treatise on Economics (New Haven: Yale University Press, 1949). 18. Webb, W. 0. "The Choice of the Problem," American Psychologist, Vol. 16 (19611,223-227. 19. Weick, K. "Methodology and Systems Theory." Unpublished paper, 1974.

You might also like