The Psychological Situation As a Determinant of Consumer Behavior

Richard J. Lutz, University of California, Los Angeles
Pradeep Kakkar, University of California, Los Angeles
[ to cite ]:
Richard J. Lutz and Pradeep Kakkar (1975) ,"The Psychological Situation As a Determinant of Consumer Behavior", in NA - Advances in Consumer Research Volume 02, eds. Mary Jane Schlinger, Ann Abor, MI : Association for Consumer Research, Pages: 439-454.

Advances in Consumer Research Volume 2, 1975      Pages 439-454


Richard J. Lutz, University of California, Los Angeles

Pradeep Kakkar, University of California, Los Angeles

[This research was supported by an intramural grant from the Academic Senate of the University of California, Los Angeles, and by the Center for Marketing Studies. Graduate School of Management. UCLA.]

[Richard J. Lutz is Assistant Professor of Marketing. and Pradeep Kakkar is a doctoral candidate in Marketing.]

The effects on behavior of the situation in which that behavior occurs have long been recognized, but seldom systematically investigated. The present study reviews some of the recent research on situational effects in marketing and suggests possible extensions. Specifically, a theory developed in environmental psychology by Mehrabian and Russell is applied in a partial replication of Belk's recent research in consumer psychology. This new framework relies on three internal state variables-pleasure, arousal, and dominance--which are conceptualized as mediating the influence of the situation on behavior. Present results indicate that the Mehrabian-Russell framework may be quite useful in the understanding and classification of consumption-related situations.

The effects of "the situation" on behavior have been alluded to frequently in the psychological and marketing literatures. Many articles include situational qualifiers in their Discussion sections, attempting to delineate those situations over which the observed relationships may be generalizable. Yet, as noted by Cottrell (1950), the actual usage of the term "situation" has been ".... quite as great in preserving an illusion of understanding as it is in conveying genuine comprehension" (p. 711). Despite this rather skeptical assessment of the status of situational variables, Cottrell went on to advocate the necessity for a situational perspective in the study of human motivation. Similarly, Frank, Massy and Wind(1972) have recently pointed to the value of "situation specific" variables in market segmentation. Thus, it appears that more systematic investigation of situational variables would make a valuable contribution to both psychology and marketing.

There have been a few scattered-articles on-situational influence in the marketing literature (e.g., Sandell, 1968; Gronhaug, 1972; Hansen, 1972), but only recently has any systematic empirical research been undertaken (Belk, 1974a,b,c). The purpose of the present paper, therefore, is primarily developmental; drawing selectively from recent literature pertaining to situational influence on behavior, recommendations will be made regarding potential directions for situational research in consumer psychology.


A crucial condition which must be met in order to allow progress in situational research is an adequate conceptual definition of the phenomenon. One key distinction has been offered by Belk (1974c):

....situation and environment ....represent distinct sources of influence on consumer behavior and should not be used synonymously. Environment is the broader construct and represents a general milieu of behavior, whereas situation is a more momentary concept (pp. 1-2).

This distinction appears meaningful, useful, and operationalizable and should be maintained in future research. Having made this distinction, Belk (1974b) further defined "situation:"

This study adopts a general view of situation as something outside the basic tendencies and characteristics of the individual, but beyond the characteristics of the stimulus object to be acted uponsituation may then be defined as all those factors particular to a time and place of observation which do not follow from a knowledge of personal (intra-individual) and stimulus (choice alternative) attributes, and which have a demonstrable and systematic effect on current behavior. It should be noted that this definition describes situation in terms of observable aggregate effects rather than in terms of similarities in individual perceptions of situations (pp. 156-7).

In restricting the definition of situation to "observable aggregate effects," Belk differs with most situational theorists. For instance, Cottrell (1950) stated (rather categorically): "We are quite certain that individuals and groups react to their own definitions of situations..." (p. 711), thus suggesting a subjective, rather than objective, definition. Similarly, Rotter (1955) states: "The basic principle for classifying or categorizing a situation is psychological, that is, subjective..." (p. 259), and relates it to Lewin's concept of the "life space." Murray (1952), in his theory of situational press, suggested classification of situations based on the differing effects they exert on the individual. Finally, Hansen (1972) concludes: "Altogether, how the actor perceives the situation is as important as the actual elements found in the physical environment" (p. 47).

On the other hand, Barker (1968) and his associates have developed a theory of "behavior settings" which allows specification of situations based upon only physical characteristics and observable behavior patterns, with no consideration of internal factors. The cataloging of an almost infinite array of physical stimuli has been a major drawback to this approach, however, and has been likened to the similarly impossible task of deriving a complete list of human motives. Even among those researchers who have advocated an objective approach to the study of situational influence, a strong emphasis has been placed upon subjective factors as well. Volkart (1951) stated: "The human situation often includes some factors common to both the observer and the actor...[but] also includes some factors that exist only for the actors, i.e., how they perceive the situation, what it means to them, what their 'definition of the situation' is" (p. 2). Similarly, Thomas (1937) posited that: "An adjustive effort of any kind is preceded by a decision to act or not act along a given line, and the decision is itself preceded by a definition of the situation, that is to say, an interpretation, or point of view...." (p. 8), and that to adequately describe a situation, the situation must be studied " it exists in verifiable, objective terms, and as it has seemed to exist in terms of the interested persons" (cited in Volkart, 1951, p. 6).

Thus, there appears to be a definite need to adopt a subjective view of the situation in order to understand its effects on human behavior. As Hansen (1972) succinctly stated:

Even if it were found that situational variations explain all variations in behavior, the question must still be asked how the individual transforms the situational input to behavioral output, which makes it necessary to look at the individual's internal processes (P. 42).

Therefore, the present research adopts the viewpoint that the "situation relevant for the understanding of consumer behavior is the psychological situation, which may be defined as an individual's internal responses to, or interpretations of, all factors particular to a time and place of observation which are not stable intra-individual characteristics or stable environmental characteristics, and which have a demonstrable and systematic effect on the individual's psychological processes and/or his overt behavior.

The above definition is similar in many respects to the one offered by Belk (1974b), but differs in the focus on subjective interpretation of the situation. Additionally, psychological processes preceding behavior are posited as a locus of situational effects, rather than overt behavioral outcomes only. This process orientation has been adopted previously by Wright (1973, 1974) and would seem to be an important area of enquiry, given the current interest in consumer information processing. In essence, the present approach is not intended to be a competitor, but rather a complement, to Belk's earlier work, in which he acknowledged that various components of the situation "....gain meaning and effect only through the perceptions of the individual" (1974c, p. 5). The present definition simply modifies his earlier one to include explicit consideration of perceptual factors and the idea of Process variables.


Most situational theorists have recognized the need for a taxonomy of situations or situational components which would allow generalizable investigations of situational influence. Rotter (1955) suggested that "systematic schemata" be used to predict human behavior from knowledge of the situation and Moos (1973, 1974) and his associates have determined the dimensionality of eight different "environments" in attempting to develop a classification scheme. Other researchers have variously specified time (Cottrell, 1950; Wright, 1973; Belk, 1974c), group opinion (Gorden, 1952), physical surroundings (Barker, 1960; Belk, 1974c, Toffler, 1970), and goal structure (Gronhaug, 1972; Belk, 1974c) as aspects of situations. However, as noted by Volkart (1951), "....situations are multidimensional, and it is difficult to specify all their elements" (p. 19).

Yet it is clear that some sort of taxonomy is necessary in order for situational variables to become meaningful in the explanation of consumer behavior. A recent theory developed in the discipline of environmental psychology may represent the first tentative step in the pursuit of such a taxonomy. Environmental psychology is normally concerned with issues such as the effects of noise, color, temperature and other specific physical stimuli on behavior. Within this tradition, however, Mehrabian and Russell's (1974) proposals relating to the description of environments and the measurement of environmental impact provide a rich framework for the study of the general psychological situation.

The Mehrabian-Russell Theory

The fundamental proposition of Mehrabian and Russell's theory is that the impact of the situation on behavior is mediated by emotional responses, so that any set of conditions initially generates an emotional (affective, connotative, feeling) reaction, which in turn leads to a behavioral response. Further, the universe of all possible emotional responses may be represented by one or a combination of three basic dimensions: pleasure, arousal and dominance. Pleasure as an emotional state is distinguished from "preference, liking, positive reinforcement, or approach-avoidance...since the latter responses are also determined by the arousing quality of a stimulus" (Mehrabian and Russell, 1974, p. 18). It is a composite of feelings such as happiness, contentment, satisfaction, etc. Arousal is an activity orientation and is "a measure of how wide awake the organism is, of how ready it is to act" (Berlyne, 1960, p. 48). Finally, dominance is a reflection of the extent to which the individual feels in control of or overpowered by his environment. The higher the level of dominance perceived in the situation, the more submissive is the state of the individual.

Among the extensive evidence offered by Mehrabian and Russell to substantiate their fundamental proposition is the body of literature relating to the semantic differential technique. There is a close parallel between the dimensions discussed by Mehrabian and Russell and by Osgood, Suci and Tannenbaum (1957): pleasure corresponds roughly to evaluation, arousal corresponds to activity, and dominance relates (inversely) to potency. The crucial element common to both frameworks is that the dimensions are not specific to modalities; i.e., the response dimensions are not anchored to any particular channel of experience, thereby allowing comparison of responses to varied types of stimuli.

To develop measures for the three mediating emotional variables, Mehrabian and Russell presented a series-of several hundred hypothetical situations to subJects in three separate studies and obtained responses on several semantic differential scales designed to measure the three basic emotions described earlier. By factor analyzing the data, it was found that three factors corresponding to pleasure, arousal and dominance could be identified. The six items loading highest on each dimension were used to construct a simple 18-item instrument (shown in Appendix A) for self-report measurement of emotional response to any conceivable situation. For instance, a boring situation should elicit responses low in pleasure, arousal and dominance; a hostile situation, on the other hand, may generate a high score on arousal, a low score on pleasure, and a low score on dominance.

According to Mehrabian and Russell, not only can emotional response to any situation be represented by these three factors, but these same factors can also be used to categorize the situations which generate those emotional states. In other words, stimuli can be described according to the responses they generate. Thus an individual shopping hurriedly for a party to which he has invited some important guests might be described as being in a high arousal situation, in contrast to a virtually endless description of the observable situation. All physical and psychological stimuli are seen as resulting in some combination of the three basic emotions, so that the consumer's ultimate overt response is, at least in part, determined by these intermediate responses. Defining situational stimuli in terms of responses is totally consistent with the definition of the psychological situation presented earlier in this paper.

Not only does the Mehrabian-Russell approach allow the description of individual situations, it- also facilitates the vital task of comparing different situations. Since any psychological situation can be measured along these three dimensions, situations can be compared through an analysis of the corresponding coefficients (or scores). This can be done at the individual level or for aggregates of individuals, so that conceivably one could obtain reactions of various market segments to consumption-related situations and thereby draw implications for marketing strategy.

Application of the Mehrabian-Russell theory to the study of the psychological situation in consumer psychology, then, offers potential solutions for two of the pressing problems of situational research. On the one hand, the Mehrabian-Russell framework can be viewed as at least a partial explanation of the influence of situations on behavior, whereas most past research has utilized a "black-box" approach, focusing only on inputs and outputs, with no consideration of mediating psychological variables. Secondly, the Mehrabian-Russell framework may satisfy the need for a taxonomical scheme for situations. Thus, the remainder of this paper will be devoted to an empirical application of this framework in an attempt to assess its potential usefulness in the explanation and classification of situational influence.


Partial Replication

In order to facilitate the evaluation of the Mehrabian-Russell framework, the present research was designed as a partial replication of Belk's (1974b) study of snack situations. The intent was to show how application of this new theory could lead to richer insights into the phenomenon of situational influence.

Belk (1974b) presented 100 subjects with ten different snack consumption situations, asking them to respond to a 10-item behavioral differential inventory in each situation. To determine the effects of situations, products, and individual differences, he then conducted a 100 (persons) x 10 (products) x 10 (situations) mixed effects ANOVA. The present research utilized the same situation descriptions and the same behavioral differential, but did not present every situation to each subject, thus ruling out the comparison of individual differences.

Situation Descriptions

Both Belk (1974b) and Mehrabian and Russell (1974) used short paragraphs to describe situations to subjects. For the purposes of comparison, ten situations from Mehrabian and Russell were included in this research, in addition to Belk's ten snack situations. Paragraphs defining the twenty situations are presented in Tables 3 and 4.


Following each situation description, subjects responded to the 18-item PAD instrument. For the ten snack situations, the behavioral differential instrument described by Belk was also included. For the ten "general" situations taken from Mehrabian and Russell, no behavioral differential was included, since the items would have made no sense in those situations (Table 2 shows the ten behavioral differential items.).


Subjects were 315 undergraduate students at UCLA and California State University at Northridge. Each subject responded to one "general" situation and one "snack" situation. Treatment booklets were constructed so that the order and frequency of presentation of all twenty situations were balanced. The booklets were distributed, completed, and collected during a regular class session.


Comparison with Belk's Findings

In order to provide at least a partial comparison of the current results with Belk's (1974b) earlier findings, a 10x10 (Situations x Products) fixed effects ANOVA was conducted on subjects' behavioral differential ratings. [More appropriate data analytic strategies might view the Products dimension as a repeated measures factor, or as a set of dependent variables in a MANOVA design. The decision to treat it as a fixed factor was based on the desire for comparability of the present findings with Belk's earlier results.] Each S's responses were in the form of stated intentions to purchase each- of 10 snack products in one of the 10 purchase situations. Thus, a given subject's responses corresponded to one level of the Situations factor and all ten levels of the Products factor. Results of the ANOVA are displayed in Table 1, together with the corresponding results from Belk (1974b).



As shown in the Table, results for the two main effects were quite similar in magnitude in the two studies. Situations explained very little variance, while products explained somewhat more across situations. While Belk does not present cell mean values, the means shown in Tables 2 and 3 for Behavioral Intention suggest that only two or three treatment levels on each factor are accounting for most of the two main effects.

The interaction of Situations and Products accounted for considerably less variance than reported by Belk. This may in part be due to the differences in methods employed in the two studies, as subjects did not respond to each of the ten situations in the present research. Other research reported by Belk (1974c) has tended to show a rather large Situation x Product interaction for beverages, meats and leisure activities; therefore, the present results may be reflecting some method variance. On the surface, there is no apparent reason why the present method should necessarily lead to a suppression of the interaction effect. In fact, the most plausible interpretation of the discrepancy in results may be that demand characteristics were operating in Belk's (1974b) experiment, causing subjects to exaggerate supposed shifts ill consumption behavior across situations. [In one sense, Belk's (1974b) procedure can be viewed as a series of experimental treatments (i.e., situation descriptions) followed by measures of behavioral intentions. Using Campbell and Stanley's (1963) notation, this would be-diagrammed as X1O1X202X303...X10O10. This type of design is especially susceptible to test-treatment interaction, which may partially explain the Product x Situation interaction effect.] Careful attention should be focused on this point in future development of methods for studying situational influence.

Use of the PAD Instrument

In order to assess the stability of the PAD instrument, scores for the ten "general" situations in the present study were compared with scores for the same ten situations as reported by Mehrabian and Russell (1974). Table 4 shows the ten situation descriptions, together with group mean scores on Pleasure, Arousal and Dominance. Numbers in parentheses are the corresponding values reported in Mehrabian and Russell. Correlation coefficients computed between the present scores and the earlier ones revealed that the instrument is fairly stable, particularly for the Arousal measure (r=.99). Pleasure (r=.86) was somewhat less stable, with Dominance (r=.82) being the least stable of the three dimensions.



Intercorrelations among the three dimensions computed across the entire sample (n=315) showed that, while all three correlations were statistically significant, the strength of association was quite low (rpD=.16; rAD=.20; rAD=-.13). Thus it was concluded that the PAD instrument was performing in accordance with expectations, and further data analysis was undertaken.

Figure 1 displays the positions of the twenty situations investigated on the Pleasure, Arousal and Dominance dimensions, in the form of data cubes. Figure 1A shows the stability of scores for the "general" situations, while Figure 1B offers a comparison of the ten general situations with the ten "snack" situations. More polarization of response is evidenced on all three dimensions for the general situations, as compared with the snack situations. While this finding can hardly be regarded as evidence for a general phenomenon, it does serve to illustrate the possibility that purchase situations may be low involvement situations to the extent that they do not generate strong emotional reactions. This would support other research which has shown products to be relatively unimportant as compared with issues (Hupfer and Gardner, 1971). Nevertheless, the current-situations were not randomly sampled, and as conclusion can be drawn regarding the issue of involvement, or more accurately, response polarity.

Relating Situational and Behavioral Responses

One of the primary advantages of the Mehrabian-Russell framework is that it allows quantification of situational descriptions along the three emotional response dimensions, thus facilitating the use of powerful multivariate statistical procedures in assessing situational influence. That is, rather than treating the ten snack situations in a nominal fashion (as in the ANOVA procedure), interval scale values can be attached to certain situational dimensions (i.e., pleasure, arousal, and dominance).

To measure the degree to which the three proposed mediating variables (P, A, and D) were related to stated behavioral intentions, a canonical correlation was computed between the three emotional mediators and the 10 BI measures, across all snack situations (n=315). One significant canonical root emerged (chi square = 50.29, d.f. = 30, p <.025), showing a correlation of .33 between the criterion and predictor vectors. This relationship, while not particularly strong, is more substantial than under the ANOVA model where P,A and D mediators were not considered. In that case, the main effect for Situations explained only about one percent of the variance in BI; the canonical model indicates that subjects' reactions to situations can explain somewhat more of the variance. This lends further support to the notion of utilizing the psychological situation in examining situational influence.





Despite the apparent increase in explanatory power resulting from the use of the pleasure, arousal and dominance variables, it is evident that the situation, in and of itself, is not a powerful Predictor of consumer behavior. This is to be expected. Considerable research has shown that personality variables consistently explain no more than 5-10% of the variance in consumer behavior (KassarJian, 1971). The variance explained by situational variables most likely will be of the same order of magnitude. However, the interaction of situational, personality, and other classes of variables should lead to increased comprehension of consumption behavior.


Earlier it had been stated that the two primary advantages of the Mehrabian-Russell framework were its potential for providing a partial explanation of situational influence and its taxonomical power. The latter claim received some support in the present research, which indicated that, across situations, subjects' responses demonstrated reliable differences along the three mediating dimensions of pleasure, arousal, and dominance (as evidenced in Figure 1A). If a large number of purchase situations were to be considered, the P,A, and D scores could be used in a clustering algorithm to allow classification of similar situations. This was not undertaken in the present case, however, due to the relatively small number of purchase situations being investigated.

The explanatory capabilities of the Mehrabian-Russell theory could not be fully investigated under the present research design. The three proposed mediating, explanatory dimensions can be regarded as generic dimensions of situations, which should explain generic situational influence. A set of ten snack product buying intentions does not meet this criterion. In future research, it would appear that decision processes and/or cognitive structure would be more appropriate dependent variables upon which to focus. Wright (1974) has recently called for the description of generic types of "decision environments" to aid in the prediction of the structure of consumer choice processes. Combination of generic situational dimensions with generic dependent variables representing choice processes will be a valuable area of research.





The specific dimensions of pleasure, arousal and dominance are also interesting because of their relationship to other trends in consumer research. For instance, the pleasure dimension may be quite useful in explaining context effects in attitude change situations (Razran, 1940; Janis, Kaye and Kirschner, 1965) or the effects of humor in advertising.

In Hansen's (1972) excellent treatment of the consumer choice process, he proposes that the discrepancy between "optimal" and "actual" levels of arousal is the primary determinant of the structure of the choice process. Further, arousal properties of the environment are conceived as the major impetus in the initiation of choice processes. Use of the Mehrabian-Russell framework to measure arousal may allow operationalization and testing of Hansen's theory, which offers the richest process orientation of the many "comprehensive" models of consumer behavior.

Finally, the dominance dimension of the Mehrabian-Russell theory closely parallels the work on "information overload" (Jacoby, et- al., 1974) and "overchoice" in the marketplace (Settle and Golden, 1974). In each of these situations, the consumer is supposedly intimidated by some aspect of his environment, causing him to engage in "dysfunctional" behavior. Similarly, research on internal/ external locus of control would seem to benefit from consideration of the dominance dimension.

At a managerial level, situational research would appear to hold great potential for the construction of point-of-purchase displays. By pretesting to determine consumer's reactions along the pleasure, arousal and dominance dimensions, the optimum display for attracting customers could be identified.

Finally, the value of multiple approaches to the study of situational influence cannot be overstated. The Mehrabian-Russell framework may be useful for deriving generalizable statements at a generic level, but other approaches may Me better suited to other objectives. Perhaps retail management could benefit from the application of Barker's (1968) "behavior settings" analysis. Similarly, Insel and Moos'(1974) approach to identifying specific situational dimensions may be quite useful in distinguishing among the various types of situations proposed by Hansen (1972). Regardless of the method employed, however, situational research must maintain an interactive focus by examining the relationship between situational and other variables. Otherwise the "situation" will quickly become one more addition to the already lengthy list of isolated constructs which are supported to. but do not. explain consumer behavior.


Barker, R.G. Ecological psychology: Concepts and methods for studying the environment of human behavior. Palo Alto: Stanford University Press, 1968.

Belk, R.W. Application and analysis of the behavioral-differential inventory for assessing situational effects in buyer behavior. In S. Ward and P. Wright (Eds.), Advances in consumer research. Volume I. Urbana, Ill.: Association for Consumer Research, 1974(a), 370-80.

Belk, R.W. An exploratory assessment of situational effects in buyer behavior J Journal of Marketing Research, 1974(b), 11, 156-163.

Belk, R.W. Situational influence in consumer behavior. Unpublished manuscript, Temple University, 1974(c).

Berlyne, D.E. Conflict, arousal and curiosity. New York: McGraw-Hill, 1960.

Campbell, D.T. and J.C. Stanley. Experimental and quasi-experimental designs for research. In N.L. Gage (Ed.), Handbook of research on teaching. Chicaro: Rand McNally. 1963.

Cottrell, L.S., Jr. Some neglected problems in social psychology. American Sociological Review, 1950, 15, 705-12.

Frank, R.E., Massy, W.F., and Wind, Y. Market segmentation. Englewood Cliffs, New Jersey: Prentice-Hall, 1972.

Gronhaug, K. Buying situation and buyers information behavior. European Marketing Research Review, 1972, 7, 33-48.

Hansen, F. Consumer choice behavior: A cognitive theory. New York: The Free Press, 1972.

Hays, W.L. Statistics for psychologists. Chicago: Holt, Rinehart and Winston, 1964.

Insel, P.M., and R.H. Moos. Psychological environments: Expanding the scope of human ecology. American Psychologist, 1974, 29, 179-88.

Jacoby, J., Speller, D.E. and Kohn, C.A. Brand choice behavior as a function of information load. Journal of Marketing Research, 1974, 11, 63-9.

Janis, I.L., Kaye, D. and Kirschner P. Facilitating effects of "eating-while-reading" on responsiveness to persuasive communications. Journal of Personality and Social Psychology, 1965, 1, 181-6.

Kassarjian, H.H. Personality and consumer behavior: A review. Journal of Marketing Research, 1971, 8, 409-419.

Mehrabian, A., and Russell, J.A. An approach to environmental psychology. Cambridge: t.I.T. Press. 1974.

Moos, R.H. Conceptualizations of human environments. American Psychologist, 1973, 28, 652-65.

Murray, H. Toward a classification of interaction. In T. Parsons and E.A. Shils (Eds.), Toward a general theory of action. Cambridge: Harvard University Press, 1952.

Osgood, ,.E., Suci, G.J., and Tannenbaum, P.H. The measurement of meaning. Urbana: University of Illinois Press, 1957.

Razran, G.H.S. Conditioned response changes in rating and appraising sociopolitical slogans. Psychological Bulletin, 1940, 37, 481.

Rotter, J.B. The role of the psychological situation in determining the direction of human behavior. In M.R. Jones (Ed.), Nebraska symposium on motivation. Lincoln: University of Nebraska Press, 1955, 245-69.

Sandell, R.G. Effects of attitudinal and situational factors on reported choice behavior. Journal of Marketing Research, 1968, 5, 405-8.

Settle, R.B., and L.L. Golden. Consumer perceptions: Overchoice in the marketplace. In S. Ward and P. Wright (Eds.), Advances in consumer research. Volume I. Urbana, Ill.: Association for Consumer Research, 1974, 29-37.

Thomas, W.I. Primitive behavior. New York: McGraw-Hill, 1937.

Toffler, A. Future shock. New York: Random House, 1970.

Volkart, E.H. Introduction: Social behavior and the defined situation. In E.H. Volkart (Ed.), Social behavior and personality: Contributions of W.I. Thomas to theory and social research. New York: Social Science Research Council, 1951. 1-32.

Wright, P.L. The harassed decision maker: Time pressure, distraction, and the use of evidence. Unpublished working paper, University of Illinois, 1973.

Wright, P. The use of phased, noncompensatory strategies in decisions between multiattribute products. Unpublished working paper, Stanford University, 1974.