&Quot;Degrees of Freedom&Quot; in Case Research of Behavioral Theories of Group Buying

Elizabeth J. Wilson, Pennsylvania State University
David T. Wilson, Pennsylvania State University
ABSTRACT - Wilson (1986a) has noted a significant lack of systematic progress in building theory about group decision making processes. In the present study, a "degrees of freedom" (Campbell 1975) approach is used to perform a crucial test of four theories of group decision making. Observational case data was used to construct the "degrees of freedom" to examine the goodness of fit of competing and complementary theories. Results of an exploratory study indicate that in the context of a modified rebuy situation, one theoretical model, the bounded rationality model, tends to have more theoretical predictions confirmed by the data than the other three models considered.
[ to cite ]:
Elizabeth J. Wilson and David T. Wilson (1988) ,"&Quot;Degrees of Freedom&Quot; in Case Research of Behavioral Theories of Group Buying", in NA - Advances in Consumer Research Volume 15, eds. Micheal J. Houston, Provo, UT : Association for Consumer Research, Pages: 587-594.

Advances in Consumer Research Volume 15, 1988      Pages 587-594


Elizabeth J. Wilson, Pennsylvania State University

David T. Wilson, Pennsylvania State University

[The authors thank Professor Ahmed Ghoniem for serving as a judge and the Institute for the Study of Business Markets for research support.]


Wilson (1986a) has noted a significant lack of systematic progress in building theory about group decision making processes. In the present study, a "degrees of freedom" (Campbell 1975) approach is used to perform a crucial test of four theories of group decision making. Observational case data was used to construct the "degrees of freedom" to examine the goodness of fit of competing and complementary theories. Results of an exploratory study indicate that in the context of a modified rebuy situation, one theoretical model, the bounded rationality model, tends to have more theoretical predictions confirmed by the data than the other three models considered.


Research on individual choice processes has been plentiful (e.g., Wright 1975; Park and Lutz 1982) and theory development in this area has progressed (Howard and Sheth 1969; Engel, Kollat, and Blackwell 1974; Hansen 1976; Bettman 1979). However, there are no clear theoretical frameworks for consumer group decision making processes. After many studies, the family decision making literature contains equivocal findings about how family groups resolve individual preferences to arrive at a group decision (e.g., Park 1982; Davis, Hoch, and Ragsdale 1986).

The development of behavioral theories of group buying behavior in organizations has fared somewhat better due the work of Robinson, Faris and Wind (1967) and Wind (1966). These researchers proposed the testing of propositions from behavioral theories of the firm (Cyert and March 1963; Cyert, Simon & Trow 1956) using ease research data. The purpose of the present paper is to examine propositions from several organizational theories simultaneously using a"degrees of freedom" approach developed by Campbell (1975).

Examination of competing and/or complementary theories in this manner has been recommended by Platt (1964) with his idea of "strong inference." Sternthal, Tybout and Calder (1987) maintain that relatively more scientific progress can be made by comparing theories rather than simply conducting more empirical studies to increasingly confirm a theory.

Although the case data for this study is from an organizational setting, the methodology can be applied to consumer settings involving family buying decision processes. The data were typed transcripts of personal interviews conducted by the first author. Each buying decision transcript was evaluated by judges. This type of data generation and judging procedure is similar to verbal protocol analysis often used in consumer decision making studies. Therefore, this methodology is applicable to consumer as well as organizational settings.

The product stimulus here, office copiers, is a rebuy decision. Although it may seem that the purchase of a copier is more complex than a rebuy, actual interviews revealed that the decision process tended to be characteristic of a rebuy rather than a new task. Decisions were made relatively quickly (average time of 2 weeks between problem recognition and order placement), the buying center was small with no more than three persons involved at any one point in time. A post-hoc explanation for this is the fact that there are many (5+) brands of desktop copiers available with comparable features and prices.


Since the early work of Wind (1966) and Robinson, Faris and Wind (1967), several researchers have proposed inductive models of organizational buying behavior (Crow, Olshavsky & Summers 1980; Hakansson 1982; Vyas & Woodside 1984). Wilson (1986a, 1986b) notes the need for more theory building in industrial buying and offers insights on theory development from small sample studies.

Heretofore much of the empiricism-to-theory-building approach used in research on organizational buying behavior has been process modelling akin to Bonoma's (1985) drift, design and prediction states of case research. Drift involves learning the concepts and social environment of the phenomenon. Design evolves from drift through the inductive process of the researcher. The researcher offers an inductive set of propositions to explain variances in observations. The prediction or generalization-formation stage builds generalizations for testing, using more cases for different sites.

Now the possibility exists to test the structures (i.e., the inductive set of propositions) of behavioral theories in organizational buying behavior. Such tests calls for examining "thick" (Geertz 1973) case descriptions to learn the degree to which the set of theoretical propositions are supported. A crucial test (Carlsmith, Ellsworth & Aronson 1976) can be made by relative degrees of fit of competing theories, e.g., an economic versus a political theory of decision-making.

Tests on the degrees to which case data fit one or more theories have been described first by Campbell (1966; 1975) as"pattern-matching" and "building degrees of freedom." Dean (1986) has been first to apply the method to test behavioral theories of organizational decision-making.

In a case study done by an alert social scientist who has thorough local acquaintance, the theory he uses to explain the focal difference also generates predictions or expectations on dozens of other aspects of the culture, and he does not retain the theory unless most of these are also confirmed. In some sense, he has tested the theory with degrees of freedom coming from the multiple implications of any one theory. The process is a kind of pattern-matching in which there are many aspects of the pattern demanded by theory that are available for matching with his observations on the local setting (Campbell 1975, pp. 181-182).

Campbell calls for direct tests of the "box scores' of correct and incorrect predictions of different theories applied to case data. "We need a tradition of deliberately fostering an adversary process in which other experts are encouraged to look for other implications of the theory and other facts to contradict or support it" (Campbell 1975, p. 186). Also, Campbell (1975, p. 188) advocates that "one should keep a record of all of the theories considered in the creative puzzle-solving process - [in inductive data-to-theory building]. To represent the degrees of freedom from multiple implications, one should also keep a record of the implications against which each was tested, and the box score of hits and misses."

A Recent Study

To apply Campbell's (1975) recommendations, Dean (1986) developed a prediction matrix to test four behavioral theories of group decision-making. The four theories included the rational model (Allison 1971), the bounded rational model (Cyert & March 1963), the political model (Pettigrew 1973; Pfeffer 1981), and the garbage can model (Feldman and March 1981). For the prediction matrix a specific hypothesis was formulated for each theory across seven facets or phases of decision making (problem definition to final choice). Thus a total of 28 cells are included in Dean's (1986) prediction matrix (4 theories by 7 decision facets).

Dean (1986) developed one to two operational statements for each cell in the 28 cell prediction matrix. Three trained judges used the operational statements to evaluate the box score hits and misses of each theory by comparing the observed decision-making behavior among managers to the operational statements. The managers' decisions were to buy advanced technology, e.g., a computer-assisted design system. Each judge coded each cell as either a C (the theory is confirmed), an N (the theory is not confirmed), or a P (the theory is partially confirmed).

The levels of agreement among the three judges in Dean's study (1986) were perfect (i.e., CCC, PPP, or NNN) or nearly perfect (e.g., CCP, PPC, NNP) for 82% of the 150 cells evaluated (28 cells by 5 buying decision cases), levels substantially above chance expectations. A central finding in Dean's study was that no one theory dominated all or most of the decision facets; the Garbage Can model was most often refuted among the five cases and seven decision facets.

Dean's (1986) field research application of Campbell's (1975; 1966) recommendations for theory testing using case studies is an important step in "applied epistemology" (Campbell 1975, p. 191), i.e., the integration of qualitative and quantitative knowing. Note that by building degrees of freedom with operational statements among decision-making facets for several theories and including several judges to evaluate the box scores of hits and misses for each theory, the natural, "inevitable residual ethnocentrism" (Campbell 1975, p. 186) of the social scientist to find data to support "his/her theory," and to ignore contrary evidence is curbed. Thus, residual ethnocentrism is reduced in three ways: (1) applying planned theory-testing thoughts (operational statements) to case data, (2) performing crucial tests of multiple theories, and (3) using several experts, judges, to independently evaluate the associations of the theories to the case data.

The present study is a report on applying the theory-testing case study approach advocated by Campbell (1975; 1966) and demonstrated by Dean (1986) within the context of group buying behavior for a modified rebuy (Robinson, Faris & Wind 1967). In an exploratory study, two operational statements are developed for each of seven decision-making facets. The application of the resulting 14 operational statements to four theories of decision-making provides unique patterns of predictions. The box scores of each theory are estimated for four case studies of group buying behavior by three judges. The results provide normative and theoretical insights into group buying behavior. The next section of the paper is a brief review of some of the concepts and models developed for studying group decision making.


Theory building in organizational decision taking has been advanced primarily by work done in the organizational behavior field (e.g., March and Simon 1958; Cyert and March 1963; Kepner and Tregoe 1965; Pettigrew 1973; Cohen, March, and Olsen 1972). These theories characterize decision activities in terms of particular aspects or facets. Although each theory uses different terms, seven generalized decision facets are recognized (Dean 1986). They are:

1. Problem definition the conceptualization of the decision problem or process by buying center members.

2. Search for solutions-the existence, degree and type of search for alternative solutions to the problem(s).

3. Data collection, analysis, and use-the extensiveness, type and function of attempts to collect and use information.

4. Information exchange the ways in which buying center members share information during the decision process.

5. Individual preference the existence, nature and resistance to change of buying center members' preferences.

6. Evaluation criteria how decision criteria are developed and used.

7. Final choice-how, when, and why choices among alternatives are made.

Four decision process theories are described next in terms of the decision facets above. The discussion is largely adapted from Dean's (1986) work and applied to buying decision processes in a marketing context.

First, the Rational Model of organizational decision making (Kepner and Tregoe 1965; Allison 1971) is grounded in microeconomic theory. Decision outcomes are chosen such that the firm derives maximum benefit (utility). In terms of decision facets, buying groups would be expected to develop comprehensive problem definitions, conduct an exhaustive information search, develop 'a priori' evaluation criteria, and exchange information in an unbiased manner. Individual preferences and final buying center choice should reflect the alternative which offers the maximum benefit to the organization.

Second, the Bounded Rationality Model (March and Simon 1958; Cyen and March 1963) posit that decision makers intend to be rational but are constrained by cognitive limitations, habits, and biases (Dean 1986). Problem definitions are simplified, search is sequential and limited to familiar areas, and information exchange is biased by individual preferences. Preferences originate from either personal or departmental sub-goals of the buying center member. Evaluation of alternatives tends to follow a conJunctive decision rule where criteria are expressed in terms of cutoff levels. Choice depends on which alternative first exceeds the minimum cutoff levels of evaluative criteria.

Third, the Political Model (Pettigrew 1973; Pfeffer 1981) proposes that buying center members compete for the decision outcome to satisfy personal and/or departmental interests. Preferences are based on these interests and formed early. Problem definitions, search, data collection and evaluation criteria are weapons used to tilt the decision outcome in one's favor (Dean 1986). Choice is a function of the relative power of buying center members.

Finally, the Garbage Can Model (Cohen, March, and Olsen 1972) suggests that decisions are analogous to garbage cans into which problems, solutions, choice opportunities, and buying center members are dumped. Problem definitions are variable, changing as new problems or people are attached to choice opportunities. Data is often collected and not used. Preferences are unclear and may have little impact on choice. Evaluation criteria are discovered during and after the process, and choices are mostly made when problems are either not noticed or are attached to other choices.

The propositions of each theory regarding the decision facets are summarized in a prediction matrix in Table 1. Given these predictions, several 'a priori' conjectures can be made regarding the research phenomenon in the context of the present study. First, the buying decision is a rebuy for all buying centers.

Conjecture 1: The models expected to apply in the cases here are the rational model and the bounded rationality model. A "rational" process is expected because of organizational policies regarding buying decisions.

Conjecture 2: The degree of rationality is expected to be relatively low, therefore, the bounded rationality model may hold more often than the rational model. The rationale is that the product stimuli (copiers) are standardized and there is relatively low risk involved in buying compared to adopting new technology (Dean 1986).

Conjecture 3: The political model is not expected to dominate because of the nature of the product. A copier purchase is not likely to induce buying center members to act on the basis of personal or organizational sub-unit goals. The purchase is for standard equipment and has relatively low self-relevance for those other than the primary user.

Conjecture 4: The Garbage Can model may find more support in the context of a rebuy versus a new task decision because buying center members may not feel threatened with the consequences of making a "bad" decision. Thus, they may not act as methodically as they would if the financial and job-related risk was high .


For each of the seven facets of a decision, two operating mechanisms in the form of decision process questions were developed. A 14 x 4 cell prediction matrix was developed from discussion of three judges (authors plus one) regarding the answers to the operating mechanism questions implied by the literature addressing the four theories. Each operating mechanism either confirmed, partially confirmed, or disconfirmed a tenet of the theory. See the prediction matrix in Table 1.


Four buying decisions for office copiers were studied using semi-structured personal interviews with buying center members and archival materials (purchase order requests). Interview transcripts were reviewed by the three judges. The results are profiles of each decision, illustrating the extent to which the facets of each decision were characterized by each of the four theories.

A contact person for each buying center was obtained from the Central Purchasing Office of a large northern university. Buying center members interviewed were secretaries and administrators of four academic departments. Copier purchases were found to be modified rebuys in all cases and consequently had fewer buying center members than new task purchases. For this reason, the two buying center members that were identified as being the most active in the purchase were interviewed.

The personal interviews with buying center members were structured only in that similar questions were asked of each participant. Questions were open ended, and were not designed to operationalize any of the theories reviewed above. As in Dean (1986), this -allowed the structure of the decision process to emerge from participants' own descriptions rather than being imposed by theoretical expectations. See Figure 1 for a list of the interview questions.


Interjudge Reliability

For each buying center case, judges responded either "Y" (theory confirmed), "N" (theory not confirmed), or"P" (theory partially confirmed), to the 14 operating mechanism questions based on the events in the case. With four buying center decisions having 14 operating mechanisms each, a 56 cell matrix resulted for the three judges. The levels of agreement between the judges were categorized similar to Dean (1986) as perfect (YYY, NNN, PPP); near perfect (YYP, YPP, NNP, NPP); some (YYN, YNN); or none (YNP).







The judges were in perfect agreement for 31 cells (55 percent). Agreement was nearly perfect for 13 cells (23 percent). There was some agreement in 11 cells (20 percent) and no agreement in only 1 cell (2 percent). Comparing these levels of agreement to those expected by chance (11, 44, 22, and 22 percent, respectively) a goodness of fit test yields X2= 103.77 (3 d.f., p c .001). This statistic was based on the total observed versus expected responses. An individual analysis for each buying case was not performed due to the sparsity of data in cells. In sum, the distribution of judges responses is substantially different from the distribution of responses expected by chance. There is a high degree of consistency in the responses of the three judges. See Table 2 for details.

Theory "Scorecard" Results

For the sake of brevity, detailed results of one buying decision process are presented next in terms of the total number of "hits" assigned by the three judges for each of the four theories (Table 3). Profiles for the remaining buying center cases are detailed in Tables 4-6 and can be similarly interpreted. A summary of the results is provided in Table 7.

A Detailed Look at Buying Decision #1: Results of the comparative theory test for buying decision #1 are in Table 3 and are interpreted as follows. For Judge A, 8 judgments out of a possible 14 judgments confirmed the predictions of the rational model of decision making. In other words, for the rational model there were 8 hits and 6 misses (8114 = .57) among the 14 operating mechanisms. Raw scores for the rows may not sum to the same total because multiple hits were possible given overlapping aspects of the theories (see the Prediction Matrix in Table 1).

For the three judges combined, 17 judgments out of 42 (3 x 14) confirmed the predictions of the rational model; that is, the rational model had 17 hits and 25 misses, or a 40 percent hit rate based on the raw (absolute) scores. The bounded rationality model hit in 62 percent of the judgments for this case while the political and the garbage can models both had 31 percent hit rates. The difference in the absolute proportion of hits for the rational model and bounded rationality model is statistically significant (z = 2.01, p < .05). It follows that the proportion of hits for the bounded rationality model is also substantially different from the political and garbage can models. See Table 3.



The total number of hits across all models (n = 69) is obtained by adding the column totals of hits for the four theories. By chance, one would expect the theories to have equal expected hit rates (25 percent) or 17.25 hits per theory (.25 x 69) in this case. Comparing the total observed distribution of hits to the expected distribution yields = 6.53 (3 d.f., p < .10). Thus, the distribution of hits for the four theories is significantly different from that expected by chance.

Further, the bounded rationality model has the highest relative hit rate (38%); this is computed as a normalized score-the total hits for a model over the total hits for the case (26169 = .38). These findings support Dean's (1986) contention that while no particular theory may dominantly account for a decision process, one theory may appear to have a better fit than others. See Table 3.

Data for the remaining three buying decisions are presented in Tables 4-6. In Buying Decision #2, the bounded rationality model again has a significantly higher hit rate versus the other models. A test of the difference in proportions between the two models with highest absolute hit rates (bounded rationality and garbage can) yields z = 1.80, p < .10. The observed distribution of hits is significantly different from that expected by chance (X2= 8.90, 3 d.f., p < .05). The bounded rationality model has the highest relative hit rate (38 percent) in Decision 2. See Table 4.



In Buying Decision #3, the rational model has the highest absolute and relative hit rate with the garbage can model having the second most hits. The difference in the proportion of hits between these two models is not statistically significant (z = 1.48). The distribution of observed hits is substantially different from that expected by chance (X2= 14.24, 3 d.f., p < .001). See Table 5.

Finally, in Buying Decision #4, the bounded rationality model again has the highest absolute and relative hit rate. Compared to the two second best models, the political and garbage can models (40 percent hits each), the bounded rationality model had an absolute hit rate of 66 percent. The difference in these/ proportions is substantial (z . 2.4, p < 05). The difference between the observed hit rate and that expected by chance is also significant (X != 6.0, 3 d.f., p < .05). See Table 6.





The findings for the four cases are summarized in Table 7. The bounded rationality model had the highest absolute hits (99) with the garbage can model in second place with 72 hits. The percentage values represent the column totals for each model. For example, across all cases and judges, each model has 168 judgments (14 operating mechanisms x 3 judges x 4 cases). For the rational model, 71 hits out of 168 is 42 percent.

The difference in the proportion of absolute hits for the bounded rationality model (.59) versus the garbage can model (.43) is substantial (z = 2.93, p < .05). It follows that the bounded rationality model had a substantially higher proportion of hits versus the rational and political models as well.

The total number of hits for across all judges and models is 293 from the column totals. By chance, each theory should have 73.25 hits if the probability of hits among theories is assumed equal. he distribution of observed hits is substantially different from that expected by chance (X2= 15.90, 3 d.f., p < .005). Finally, the bounded rationality model has the highest relative frequency of hits versus the other models.




The bounded rationality model had a significantly greater proportion of confirmed predictions (hits) in three of the four buying cases. In Buying Decision #3, the rational model had the greatest absolute hits but that proportion was not statistically different from the garbage can model in second place. The garbage can and rational model of decision making were almost tied for second place in the combined analysis (see Table 7). The garbage can model had a slightly greater absolute proportion of confirmed predictions versus the rational model (.43 versus .42, respectively) but both were significantly less than the absolute proportion of hits for the bounded rationality model (.59). The political model ranged from being tied for second place in Decision #4 to last place in Decision #3.

In conclusion, several observations can be made about the conjectures stated earlier. Regarding Conjecture 1, the bounded rationality model does seem to have the "best fit" in the context of a rebuy decision to acquire a copier. Tenets of the bounded rationality model were consistently supported by the judges' responses to the operating mechanisms. Dean's (1986) finding that tenets of one theory may dominate results is supported by the findings of the crucial test performed in the present study. Tenets of the other theories were certainly supported, to a degree, by the observational data but the bounded rationality model had the greatest number of absolute and relative hits in the overall box score (see Table 7).

In Conjecture 2, the rational model was proposed to be more applicable than the political or garbage can models. This conjecture was not supported by the judges' responses. The rational model had the greatest hits in Decision #3 but no consistent support for the remaining decision cases.

Support for aspects of the political model was more evident than proposed in Conjecture 3. Thus, there was some degree of political maneuvering by buying center members even in a relatively risk-free, mundane decision. The conclusion from the present results and from Dean's (1986) study is that the political model may have relevance in both rebuy and new task decisions. These processes should be considered for inclusion in future models of group decision making.

Finally, Conjecture 4 is supported by the results of the present study regarding the garbage can model. Aspects of the garbage can model found more support in the cases here than in Dean's (1986) study. Buying center activities were characterized by some aspects of the garbage can model such as uncertain preferences, no 'a priori' development of evaluative criteria, and uncertain problem definition.

Future Research and Limitations

Although the current study is conducted in an organizational buying context, the methodology may be useful in building and testing theory in consumer group choice situations. A number of case studies of family or consumer (small) group choices could be collected and the underlying theories of choice tested using the "degrees of freedom" approach illustrated in this paper. The output would be a measure of the appropriateness of competing and/or complementary theories. Such research would serve as input to further experimental or survey studies of consumer group decision making.

A limitation to the study is the use of the authors as judges. More work in this area is planned by the authors and independent judges will be employed at several stages of the research process. For example, one set of judges will be asked to design the prediction matrix according to their interpretation of the theories compared. A different set of judges will be enlisted to go through the interview transcripts to record the "box scores." A second limitation is the fact that only one product was studied. Different products (of varying task complexity) bought by organizational and families should be examined.

Future studies of comparative theory testing are recommended for all areas of group decision making. Adversary replication (Campbell 1975) is strongly encouraged. Copies of each buying decision case and judges responses are available from ole authors for examination by others that would care to conduct an adversary replication. Such work would offer objective insights for future development of models and theories of group decision making.


Allison, G. (1971), Essence of Decision: Explaining the Cuban Missile Crisis, Boston: Little, Brown, and Company.

Bettman, James (1979), An Information Processing Theory of Consumer Choice, Reading, MA: Addison-Wesley.

Bonoma, Thomas (1985), "Case Research in Marketing: Opportunities, Problems, and a Process," Journal of Marketing Research, 22 (May), 199-208.

Campbell, Donald (1966), "Pattern Matching as an Essential in Distal Knowing," in K. Hammond, ed., The Psychology of Egon Brunswik New York: Holt, Rinehart, and Winston, 81-106.

Campbell, Donald (1975), " 'Degrees of Freedom' and the Case Study," Comparative Political Studies, 8 (July), 178-193.

Carlsmith, J., P. Ellsworth, and E. Aronson (1976), Methods of Research in Social Psychology, Reading, MA: Addison Wesley.

Cohen, M., J. March, and J. Olsen (1972), "A Garbage Can Model of Organizational Choice," Administrative Science Quarterly, 17, 1-24.

Crow, Lowell, Richard Olshavsky, and John Summers (1980), "Industrial Buyer's Choice Strategies: A Protocol Analysis," Journal of Marketing Research, 17 (February), 34-44.

Cyert, R. and J. March (1963), A Behavioral Theory of the Firm, Englewood Cliffs, NJ: Prentice-Hall.

Cyert, R., Simon, H. and D. Trow (1956), "Observations of-a Business Decision," Journal of Business, 29 (October), 237-238.

Davis, Harry, Stephen Hoch, and Easton Ragsdale (1986), "An Anchoring and Adjustment Model of Spousal Predictions," Journal of Consumer Research, 13 (June), 25-37.

Dean, James (1986), "Decision Processes in the Adoption of Advanced Technology," unpublished manuscript, The Pennsylvania State University.

Feldman, M. and J. March (1981), "Information in Organizations as Signal and Symbol," Administrative Science Quarterly, 26, 171-186.

Engel J., D. Kollat, and R. Blackwell (1974), Consumer Behavior, New York: Holt, Rinehart and Winston.

Geertz, Clifford (1973), The Interpretation of Cultures, - New York: Basic Books.

Hakannson, H. (ed.) (1982), International Marketing and Purchasing of Industrial Goods, New York: Wiley.

Hansen, Flemming (1976), "Psychological Theories of Consumer Choice," Journal of Consumer Research, 3 (December), 117-142.

Howard, John and Jagdish Sheth (1969), The Theory of Buyer Behavior, New York: Wiley.

Kepner, C. and Tregoe, B. (1965), The Rational Manager, New York: McGraw-Hill.

March, J. and H. Simon (1958), Organizations, New York: Wiley.

Park, C. Whan (1982), "Joint Decisions in Home Purchasing: A Muddling-Through Process," Journal of Consumer Research, 9 (September), 151-162.

Park, C. Whan and Richard Lutz (1982), "Decision Plans and Consumer Choice Dynamics," Journal of Marketing Research, 19 (February), 108-115.

Pettigrew, Andrew (1973), The Politics of Organizational Decision Making, London: Tavistock.

Pfeffer, Jeffrey (1981), Power in Organizations, Boston: Pittman Publishing.

Platt, John (1964), "Strong Inference," Science, 46 (October), 347-353.

Robinson, P., C. Farris, and Y. Wind (1967), Industrial Buying and Creative Marketing, Boston, MA: Allyn and Bacon.

Sternthal, Brian, Alice Tybout, and Bobby Calder (1987), "Confirmatory Versus Comparative Approaches to Judging Theory Tests," Journal of Consumer Research, 14 (June), 114-125.

Vyas, Niran and Arch G. Woodside (1984), "AD Inductive Model of Industrial Supplier Choice Processes," Journal of Marketing, 48 (Winter), 30-45.

Wilson, David T. (1986a), "Qualitative Approaches to Organizational Buying Behavior Theory Development," unpublished manuscript, The Pennsylvania State University.

Wilson, David T. (1986b), "Merging Adoption Process and Organizational Buying Models," forthcoming in Advances in Consumer Research, M. Wallendorf and P. Anderson, eds. Association for Consumer Research.

Wind, Yoram (1966), "Industrial Buying Behavior Source Loyalty in the purchase of Industrial Components," unpublished Ph.D. dissertation, Stanford University.

Wright, Peter (1975), "Consumer Choice Strategies: Simplifying Vs. Optimizing," Journal of Marketing Research, 12 (February), 60-67.