Measuring Perceived Risk: a Replication and an Application of Equity Theory

Richard H. Evans, Syracuse University
ABSTRACT - The three consumer decision models employed by Peter and Tarpey (1975) were replicated. In addition, an equity theory model was investigated. A comparison of the two sets of data involving the three models revealed similar results in terms of explanation and prediction. The equity theory model, as operationalized, produced results which were similar to the Peter and Tarpey (1975) data.
[ to cite ]:
Richard H. Evans (1982) ,"Measuring Perceived Risk: a Replication and an Application of Equity Theory", in NA - Advances in Consumer Research Volume 09, eds. Andrew Mitchell, Ann Abor, MI : Association for Consumer Research, Pages: 550-555.

Advances in Consumer Research Volume 9, 1982      Pages 550-555

MEASURING PERCEIVED RISK: A REPLICATION AND AN APPLICATION OF EQUITY THEORY

Richard H. Evans, Syracuse University

ABSTRACT -

The three consumer decision models employed by Peter and Tarpey (1975) were replicated. In addition, an equity theory model was investigated. A comparison of the two sets of data involving the three models revealed similar results in terms of explanation and prediction. The equity theory model, as operationalized, produced results which were similar to the Peter and Tarpey (1975) data.

INTRODUCTION

Peter and Tarpey (1975), using six car brands and six risk attributes, examined three decision making models or strategies--minimization of expected negative utility, maximization of expected positive utility and maximization of expected net utility--in terms of model explanation and prediction. Factor analysis was used to test model explanation and multiple regression was employed to measure model prediction. The factor analysis produced two factors which were labelled by Peter and Tarpey (1975): expected performance and psychosocial. [The present study also differs from the Peter and Tarpey (1975) research in that it includes, as suggested by Churchill (1979). reliability information (See Table 1).] The multiple regression calculation indicated that the maximization of expected net utility model explained most of the variance in brand preference (R2) followed next by the negative utility model and the positive utility model, respectively.

The research reported here replicates, in part, the work of Peter and Tarpey (1975) and extends it to include another decision making model or strategy. Consequently, the purpose of the present study was to compare four models or strategies of consumer behavior in terms of their relative explanative and predictive capabilities. Two car brands, Pinto and Malibu, which were used by Peter and Tarpey (1975), were also examined in the present study.

Engel, Blackwell and Kollat (1978) indicate the consumer behavior is motive satisfying behavior. Howard and Sheth (1969) state that 'Motives serve an essential role in explaining overt purchase behavior." Campbell, Dunnette, Lawler and Weick (1970) have classified the various theoretical orientations to motivation theory into two approaches: content and process. The content approach [An example of a content theory would be Maslow's Hierarchy of Needs (Maslow, 1954).] deals with "what it is-within an individual or his environment that energizes and sustains behavior" (Campbell, Dunnette, Lawler and Weick, 1970) while the process approach is ". . . concerned with identifying the major classes of variables and mechanics which underlie how motives operate to determine behavior" (Jacoby, 1976). Three key process theories, according to Campbell, Dunnette, Lawler and Weick (1970) 9 are: stimulus-response theory, expectancy theory and equity theory. Stimulus-response theory has been applied in consumer behavior, for example, by Howard (1968)S Farley and Kuehn (1965) and recently by Srinivasan and Kesavan (1976). The formulations of Rosenberg (1956) and Fishbein (1967) are based on expectancy theory. A great deal of research has appeared in the literature using Rosenberg (1956) or Fishbein (1967) as the foundation (see Wilkie and Pessemier, 1973). Equity theory, however, has been applied in consumer behavior to only a very limited extent. Where equity theory has been applied, theory support has been evident. [Equity theory has been applied by Leventhal, Younts and Lund (1972) dealing with a retail sales relationship, it was mentioned by Scott (1977) relative to the foot-in-the-door technique and it has been applied by Huppertz, Arenson and Evans (1978) in terms of retailing.]

In this paper, equity theory [The reason equity theory was investigated was that it was believed that it would produce reasonably high R2 values. The R2 values produced in the. Peter and Tarpey (1975) study were extremely low.] will be compared with the three models advanced by Peter and Tarpey (1975), to assess its effectiveness in terms of explaining and predicting preference. [The relationship between fairness and preference should be positive, i.e., a fair attribute or facet (e.g., price) will be perceived favorably.] The equity formulation adds to or extends the Peter and Tarpey (1975) models in that it involves an element of fairness (Adams, 1963) and, as such, it may be perceived not as a loss minimization, gain maximization or net return model but as a satisficing model (Jacoby, 1976). Further, Jacoby (1976) indicates that ". . . consumer behavior data reflects 'satisficing' . . . rather than maximizing behavior." [One stream of thought regarding decision making is that consumers tend to maximize outcomes. Another is that consumers make "satisficing" decision where objectives are set and the individual attempts to realize or perhaps meet these minimum objectives. Conceptually, decision making on an equity basis is similar to the "satisficing" approach. In other words, an optimum decision in equity theory would be: Oa/Ia = Ob/Ib where Oa pertains to the outcomes of individual a, in equals inputs of individual a, Ob refers to the outcomes of the comparison other b and Ib pertains to the inputs of comparison other b.] It was surmised that the equity model would be similar to the other models in terms of explanation (i.e., it will have two dimensions) but superior to the other models in terms of prediction.

Theoretical Models

Peter and Tarpey (1975) analyzed three consumer decision strategies. The attributes, characteristics or facets used to analyze the decision-making strategies were based on six dimensions of perceived risk. They were: (l) financial risk, (2) performance risk, (3) psychological risk, (4) physical risk, (5) social risk, and (6) time risk. According to Jacoby and Kaplan (1972) each of these attributes are independent and it may be inferred from the Jacoby and Kaplan (1972) paper that the risk facets are essentially inclusive. [Jacoby and Kaplan (1972) analyzed five risk facets. A six risk variable was advanced by Roselius (1971).] Peter and Tarpey (1975) view the consumer to be rational in that ". . . behavior is (a) goal directed, (b) calculated and (c) predicted upon some knowledge of the costs and benefits of alternative choices." Within this framework and considering the above risk facets, Peter and Tarpey (1975) operationalized three models.

The first consumer strategy or model was minimization of expected negativity utility. Algebraically, the model reads:

EQUATION    (1)

where

OPRj = overall perceived risk for brand j

PLij = probability of loss i from the purchase of brand j

ILij - importance of loss i from purchase of brand j

n = loss characteristic

Bauer (1960) and others suggested that consumers employ a strategy that involves the minimization of expected loss or perceived risk. Kogan and Wallach (1964) indicated that perceived risk involved two components ". . . 'a choice' aspect where the emphasis is on probability [of losing] and a 'danger' aspect where the emphasis is on severity of negative consequences." Other researchers (e.g., Cunningham, 1967 and Jacoby and Kaplan, 1972) have suggested that these two components exist in a multiplicative relationship. Consequently, the above negative expected utility model involves both a multiplicative and an additive calculation.

The second model advanced by Peter and Tarpey (1975) was called a perceived return model. The theoretical background for this model is attitude theory which has been widely discussed in the consumer behavior literature. [See Wilkie and Pessemier (1973) for an excellent review article.] This model involves multiple attributes or facets, component multiplication and summation. The perceived return may be formulated as:

EQUATION    (2)

where

OPRej = overall perceived return

PGij = probability of gain i from purchase of brand j

IGij = importance of gain i from purchase of brand

n = return characteristics or facets.

The third model researched by Peter and Tarpey (1975) was the net perceived return model. This decision-making strategy was based on earlier research by Lewin (1943) and Bilkey (1953). With this strategy consumers are said to perceive products in terms of ". . . desirable (positive valence) and undesirable features (negative valence)." Peter and Tarpey (1975) operationalized this approach by assuming that ". . . individuals attempt to maximize the 'net valence' which is the arithmetic difference between expected positive and negative utility (i.e., net perceived return)." Symbolically. this strategy may be stated as:

EQUATION    (3)

where

NPRej = net perceived return for brand

PGij = probability of gain i from purchase of brand

IGtj = importance of gain i from purchase of brand n

PLij = probability of loss i from purchase of brand

ILij = importance of loss i from purchase of brand

n = utility facets (characteristics)

The forth model discussed in this paper was based on equity theory. This decision-making strategy was not analyzed by Peter and Tarpey (1975). Equity theory has been discussed in some detail by Jacoby (1976) and reviewed by Pritchard (1969). As Jacoby (1976) indicates "equity theory is essentially a social comparison theory in which an individual evaluates his "inputs into" versus "outputs derived from" a given situation relative to those of another, where this other may be another person, a class of people, an organization, or the individual himself relative to his experiences from an earlier point in time." According to Campbell and Pritchard (1976) equity theory ". . . teals with exchange relationships and the fairness or equity of these exchange relationships." One way to operationalize equity theory is to measure the fairness of each facet or attribute in the decision process. This approach for measuring equity has been used by other researchers (e.g., Huppertz, Arenson and Evans, 1978; Austin and Walster, 1974; Brickman and Bryan, 1975; and Wicker and Bushweiler, 1970). In this study the probability of an attribute occurring will be combined with the perceived fairness of that attribute. The numerical value resulting from the multiplicative and additive equity model will be compared to an overall preference measure (i.e., the dependent variable). Equity theory is relevant to risk measurement in that it adds a new dimension for assessing risk. Peter and Tarpey (1975) examined risk in terms of occurrence probabilities and importance. In this study, equity theory was also employed, and risk was assessed in terms of probability of occurrence and fairness. As an example, high monthly payments (risk attribute) for a car may be perceived to be probable-improbable and fair-unfair. If high monthly payments are deemed probable and unfair, purchase is likely to be negatively affected. The conceptual background for equity theory lies in the work of Thibaut and Kelley (1959) and Homans (1961), however, it was formally developed by Adams (1963, 1965). Algebraically, the equity model may be depicted

EQUATION    (4)

where:

PEj1 = perceived equity for brand j in loss (1) situation

PLij = probability of loss i from the purchase of brand i

Fij = perceived fairness of i from the purchase of brand j

and

EQUATION  (5)

where:

PEjg = perceived equity for brand j in gain (g) situation

PGij = probability of gain i from purchase of brand j

Fij = perceived fairness of i from purchase of brand j

METHODOLOGY

Sample

The sample consisted of approximately 250 juniors and seniors enrolled in a basic marketing course at Syracuse University. A total of 245 usable questionnaires were obtained. Of the 245 questionnaires, 119 were completed on the Pinto and 126 were completed on the Malibu. Peter and Tarpey (1975) also used a sample of students.

Questionnaires

The questionnaires used in this study were identical to those of the original study (i.e., Peter & Tarpey, 1975), with the exception that questions on equity were added. Each questionnaire (Pinto, Malibu) contained a preference (like-dislike) question (dependent variable) and thirty six independent variable questions: (a) six to measure probability of loss [Loss facets refer to high maintenance cost, high monthly payments, poor self-image with car, etc.] from brand purchase; (b) six to measure loss importance from brand purchase; (c) six to measure fairness of loss from brand purchase; (d) six to measure probability of gain [Gain facets pertain to low maintenance cost, improvement in self-image, improvement in social acceptance, etc.] from brand purchase; (e) six to measure gain importance from brand purchase; and (f) six to measure the fairness of gain from brand purchase. The probability questions were measured on a scale with endpoints labelled: probable-improbable and the importance questions were measured on a scale with endpoints labelled: importance-unimportance. To obtain a measure of perceived equity, subjects were asked to rate each at tribute in terms of each situation. Therefore, as there were six attributes and two situations (i.e., 1088, gain), twelve measures of equity were obtained (for each brand). Fairness was operationalized using a 7-point bipolar scale with end points labelled: fair-unfair.

Statistical Analysis

The data was first analyzed using Cronbach's Alpha (Cronbach, 1951) to measure the reliability of each model component (PL, IL, PG, IG, F (loss) and F (gain) in terms of each brand. To assess if a significant difference existed between a loss situation and a gain situation in terms of fairness the appropriate data was next analyzed in terms of a t-test. Next the data was analyzed in terms of factor analysis to assess the explanatory capability of each model. Using the varimax rotation option, loadings were obtained on each of the six brand characteristics relative to each brand. Originally, it was assumed that all factors would be delineated with an eigenvalue less than 1. However, this was found not to be possible and eigenvalues close to the initial criterion of 1 have been included. Eigenvalues less than 1 are reported in the text of the paper. Further, the data was analyzed using multiple regression analysis. The factor scores served as input in the multiple regression analysis to eliminate multicollinearity among the independent variables. The resulting R2's which were obtained from the multiple regression analysis were compared on a model by model basis to determine which one best represented the variation in brand preference.

RESULTS

Coefficient alpha values (Cronbach, 1951) are shown in Table 1. Nunnally (1967) suggests that reliabilities be tween .50 and .60 are acceptable in early stages of basic research. Therefore, except for Malibu PL, the reliability measures for each model component are within or above an acceptable standard.

Table 2 shows the mean fairness scores (Fij) over the two situations (i.e., loss and gain). Table 2 indicates that the subjects perceived from each car brand over the six risk facets a loss situation to be unfair and a gain situation to be fair. A bias in terms of what is fair or unfair has been reported by others, e.g., Messick and Sentis, 1979. Ideally, a loss and gain situation should be perceived as unfair or inequitable. In a loss situation the customer losses and in a gain situation the seller losses. In a fair (equitable) situation neither lose.

In this study the subjects perceived a fair situation to be when the seller was losing (gain situation).

TABLE 1

CRONBACH'S ALPHA: MODEL COMPONENTS BY BRAND

TABLE 2

MEAN FAIRNESS SCORES (Fij) OVER TWO SITUATIONS: LOSS AND GAIN

Table 3 shows the matrices of the varimax rotated factor loadings for the loss characteristics for the two automobile brands - Pinto and Malibu. Factor 1 tends to load heavily on expected performance characteristics (i.e., financial, performance, physical and time) whereas factor 2 has high loadings on the psychosocial dimension (i.e., social and psychological variables). A partial exception to the above paradigm is the Pinto brand which also has a high loading for the physical risk characteristic on factor 2. These findings relative to perceived loss are similar to those of Peter and Tarpey (1975).

Table 4 depicts the factor loading matrix for the perceived gain characteristics. Two factors emerged from the analysis (the second factor has an eigenvalue of .7855 and .8808 for the Pinto brand and the Malibu brand, respectively). In keeping with the approach used by Peter and Tarpey (1975), these eigenvalues were considered close enough to the cut-off point of 1 to be included in the analysis. Factor 1 involves the expected performance decision (i.e., financial performance, physical and time) and factor 2 includes the psychosocial dimensions (i.e., social and psychological). The results shown in Table 4 are similar to those determined by Peter and Tarpey (1975).

TABLE 3

MATRICES OF ROTATED FACTOR LOADINGS FOR SIX PERCEIVED LOSS CHARACTERISTICS FOR TWO AUTOMOBILE BRANDS

TABLE 4

MATRICES OF ROTATED FACTOR LOADINGS FOR SIX PERCEIVED GAIN CHARACTERISTICS FOR TWO AUTOMOBILE BRANDS

Table 5 depicts the matrices of the rotated (varimax) factor loadings for the six net return characteristics. As with the loss model and the gain model, a two factor solution was requested. The eigenvalue for second factor was .9524 for the Pinto brand and .9367 for the Malibu brand. The factor loading matrices in Table 5 did not show the same two dimensional pattern that was observed in Table 3 and Table 4. The Pinto brand in Table 5 has heavy loadings on the social, psychological and physical characteristics in factor two. On the other hand, the Malibu brand, in terms of factor two, only has a heavy loading on the psychological variable. When considering the perceived net return model, the two dimensional solution involving the expected performance and psychosocial paradigm did not appear to be evident. In other words, relative to the perceived net return model, the results of the present study did not equal those of Peter and Tarpey (1975) in terms of attribute dimensionality. [Of the six brands studied by Peter and Tarpey (1975), three did not factor out to produce a clear dichotomy in terms of the two dimensions--performance and psychosocial.]

TABLE 5

MATRICES OF ROTATED FACTOR LOADINGS FOR SIX PERCEIVED NET RETURN (GAIN-LOSS) CHARACTERISTICS FOR TWO AUTOMOBILE BRANDS

Table 6 shows the matrices of the rotated (varimax) factor loadings for the equity model. The eigenvalues for the second factor in the gain situation were below the cutoff point of 1. [The eigenvalue for the second factor with Pinto (gain) and Malibu (gain) was .7631 and .8098, respectively.] The perceived equity model did not follow the attribute dimensionality paradigm that was observed by Peter and Tarpey (1975).12

Table 7 shows the results of the multiple regression calculation using factor scores as input. The perceived net return model had the highest mean R2, followed by the perceived loss model, the perceived equity (gain) model, the perceived gain model, and the perceived equity (loss) model, respectiveLy. In the Peter and Tarpey (1975) study the mean R 's were ranked from highest to lowest as follows: net perceived return, perceived loss and perceived gain, respectively. The rank order of the R2's is identical in both studies in terms of the net return, loss and gain models.

In both studies the model that appeared to have the poorest predictive capability was perceived gain.

TABLE 6

MATRICES OF ROTATED FACTOR LOADINGS FOR SIX PERCEIVED EQUITY CHARACTERISTICS FOR TWO AUTOMOBILE BRANDS

TABLE 7

BRAND PREFERENCE AS A FUNCTION OF PERCEIVED LOSS, PERCEIVED GAIN, PERCEIVED NET RETURN AND PERCEIVED EQUITY: A COMPARISON OF R2'S OBTAINED FROM THE REGRESSION OF ORTHOGONAL FACTOR SCORES

CONCLUSIONS

The basic purpose of this study was to investigate four alternative decision making models in terms of their relative ability to explain and predict brand preference. In part, the study by Peter and Tarpey (1975), was replicated. Models are usually compared in the literature on the basis of predictive criteria. However, as Wilkie, McCann and Reibstein (1974) indicate, interest in various models is also based on diagnostic or explanatory performance.

Peter and Tarpey (1975) examined three models -- perceived loss, perceived gain and perceived net return. Essentially, with each model they found that attribute dimensionality could be explained in terms of two dimensions: expected performance and psychosocial. Peter and Tarpey (1975) found the perceived net return model to be the best predictor of brand preference.

In the present study a forth model called perceived equity was added to the three researched by Peter and Tarpey (1975). In the present study, the perceived gain model and the perceived loss (Malibu brand) model exhibited the two dimensional attribute pattern found by Peter and Tarpey (1975), the perceived loss (Pinto brand) model, the perceived net return (Pinto brand) model and the perceived equity (Pinto brand) model in the gain situation only partially yielded a two dimensional pattern and the perceived net return (Malibu brand) model and the three equity models--loss (Pinto and Malibu brands) and gain (Malibu brands)--did not produce the expected performance and psychosocial paradigm. In summary, the present study did not produce a clear cut picture of risk characteristic dimensionality as was the case in the Peter and Tarpey (1975) research. In the present study it was found that the perceived net return model was the best predictor of brand preference--the same as Peter and Tarpey (1975).

Several factors may account for the differences in results. First, although both studies utilized students as subjects, the students may have been different between the two universities involved. Second, the time lapse between the two studies was approximately five years. The results of the two studies could have also been affected by the fact that non-ratio scale measurement was involved in each study (see Schmidt and Wilson, 1975).

The two studies produced very low R2 values. A possible reason for this finding is that the facets (attributes) were not determinant (see Alpert, 1971).

Generally speaking, the two studies produced similar results.in terms of predicting brand preference but different results in terms of explaining brand preference. The equity model produced less than expected results in terms of both explanation and prediction. This may have been due to the method of operationalization used in the model in this study. 13

REFERENCES

Adams, J. S. (1963), "Toward an Understanding of Inequity," Journal of Abnormal and Social Psychology, 67, 422-436.

Adams, J. S. (1965), "Inequity in Social Exchange," in Advances in Experimental Social Psychology, ed. L. Berkowitz, New York: Academic Press, 267-299.

Alpert, M. I. (1971), "Identification of Determinant Attributes: A Comparison of Methods," Journal of Marketing Research, 8, 184-191.

Austin, W. and Walster, E. (1974), "Reactions to Confirmation and Disconfirmation of Expectancies of Equity and Inequity," Journal of Personality and Social Psychology, 30, 208-216.

Bauer, R. A. (1967), "Consumer Behavior as Risk Taking," in Risk Taking and Information Handling in Consumer Behavior, ed. Donald F. Cox, Boston: Harvard University, 389-398.

Bilkey, W. J. (1953), A Psychological Approach to Consumer Behavior Analysis," Journal of Marketing, 18-25.

Birchman, P. and Bryan, J. H. (1975), "Evaluation of Theft, Charity, and Disinterest Transfers that Increase or Decrease Equality," Journal of Personality and Social Psychology, 31, 156-161.

Campbell, J. P. and Pritchard, R. D. (1976), "Motivation Theory in Industrial and Organizational Psychology," in Handbook of Industrial and Organizational Psychology, ed. M. D. Dunnette, Chicago: Rand McNally College Publishing Co.

Campbell, J. P., Dunnette, M. D., Lawler, E. E., III, and Weich, B. E. Jr. (1970), Managerial Behavior and Human Performance, New York: McGraw-Hill.

Churchill, G. A. Jr. (1979), "A Paradigm for Developing Better Measures of Marketing Constructs," Journal of Marketing Research, 16. 64-73.

Cronbach, L. J. (1951), "Coefficient Alpha and the Intern l Structure of Tests," Psychometrika, 16, 297-334.

Cunningham, S. M. (1967), "The Major Dimensions of Perceived Risk," in Risk Taking and Information Handling in Consumer Behavior, ed., Donald F. Cox, Boston: Harvard University, 82-108.

Engel, J. F., Blackwell, R. D. and Kollat, D. T. (1978), Consumer Behavior 3rd Ed., new York: Holt, Rinehart and Winston, Inc.

Farley, J. V. and Kuehn, A. A. (1965), "Stochastic Models of Brand Switching," in Science in Marketing. ed. G. Schwartz, New York: John Wiley and Sons, Inc.

Fishbein, M. A. (1967), "Attitude and the Prediction of Behavior," in Readings in Attitude Theory and Measurement, ed. M. A. Fishbein, new York: John Wiley and Sons.

Homans, G. C. (1961), Social Behavior: Its Elementary Forms, new York: Harcourt, Brace and World.

Homans, G. C., Unpublished research conducted by the author using a different equity formula has produced R2's that are in the .5 to .7 range.

Howard, J. A. (1968), "Learning and Consumer Behavior," in Perspective in Consumer Behavior, eds. H. H. Kassarjian and T. S. Robertson, Glenvire, Ill.: Scot. Foresman.

Howard, J. A. and Sheth, J. N. (1969), The Theory of Buyer Behavior, New York: John Wiley and Sons, Inc.

Huppertz, J. W., Arenson, S. J. and Evans, R. H. (1978), "An Application of Equity Theory to Buyer-Seller Exchange Situations," Journal of Marketing Research, 250-259.

Jacoby, J. and Kaplan, L. B. (1972), "The Components of Perceived Risk," in Proceedings 3rd Annual Conference Association for Consumer Research, ed. M. Venkatesan, Association for Consumer Research.

Jacoby, J. (1976), "Consumer and Industrial Psychology: Prospects for Theory Corroboration and Mutual Contribution,: in Handbook of Industrial and Organizational Psychology, ed. M. D. Dunnette, Chicago: Rand McNally College Publishing Co.

Kogan, N. and Wallacy, M. S. (1964), Risk Taking: A Study in Cognition and Personality, New York: Holt, Rinehart, and Winston.

Leventhal, G. S., Younts, C. M., and Lund, A. R. (1972), "Tolerance for Inequity in Buyer-Seller Relationships," Journal of Applied Social Psychology, 2, 308-318.

Lewin, R. (1943), "Forces Behind Foot Habits and Methods of Change," National Research Council Bulletin. Quoted in J. P. Peter and L. X. Tarpey, Sr., (1975), "A Comparative Analysis of Three Consumer Decision Strategies," Journal of Consumer Research, 2, 29-37.

Maslow, A. H. (1954), Motivation and Personality, new York: Harper And Row.

Messick, D. M., and Sentis, R. P. (1979), "Fairness and Preference," Journal of Experimental Social Psychology, 15, 418-434.

Nunnally, J. C. (1967), Psychometric Theory, new York: McGraw-Hill Book Company.

Peter, J. P., and Tarpey, L. Z. Sr. (1975), "A Comparative Analysis of Three Consumer Decision Strategies." Journal of Consumer Research, 2, 29-37.

Peter, J. P. and Ryan, M. J. (1976), "An Investigation of Perceived Risk at the Brand Level," Journal of Marketing Research, 8, 184-188.

Pritchard, R. D. (1969), "Equity Theory: Review and Critique," Organizational Behavior and Human Performance, 4, 176-211.

Roselius, T. (1974), "Consumer Rankings of Risk Reduction Methods," Journal of Marketing, 35, 56-61.

Rosenberg, M. J. (1956), "Cognitive Structure and Attitudinal Affect," Journal of Abnormal and Social Psychology, 53, 367-372.

Schmidt, F. L., and Wilson, T. C. (1975), "Expectancy Value Models of Attitude Measurement: A Measurement Problem," Journal of Marketing Research, 12, 366-368.

Scott, C. A. (1977), "Modifying Socially-Conscious Behavior: The Foot-In-The-Door Technique," Journal of Consumer Research, 4, 156-164.

Srinivasan, V., and Kesavan, R. (1976), "An Alternative Interpretation of the Linear Learning Model of Brand Choice," Journal of Consumer Research, 3, 76-83.

Thibaut, J. W., and Keller, H. H. (1959), The Social Psychology of Groups, new York: Wiley.

Walster, E., Berscheid, E., and Walster, G. W. (1976), "Nev Directions in Equity Research" in Equity Theory: Toward A General Theory of Social Interaction, ed. L. Berkowitz and E. Walster, New York: Academic Press.

Wicker, A. W., and Bushweiler, G. (1970), "Perceived Fairness and Pleasantness of Social Exchange Situations: Two Factorial Studies of Inequity," Journal of Personality and Social Psychology, 5, 63-75.

Wilkie, W. L., and Pessemier, E. A. (1973), "Issues in Marketing's Use of Multi-Attribute Attitude Models," Journal of Marketing Research, 10, 428-441.

Wilkie, W. L., McCann, J. M., and Reibstein, D. J. (1974), "Halo Effects in Brand Belief Measurement: Implications for Attitude Model Development," in Advances in Consumer Research, ed. S. Ward and P. Wright, Urbana, Ill.: Association for Consumer Research, 280-290.

----------------------------------------