On Satisfaction Responses and the Effects of Product Labels

Robert A. Westbrook, University of Arizona
ABSTRACT - A study of consumer satisfaction with philanthropic contributions to health service organizations and another of consumer response to cigarette warning labels are reviewed. Several opportunities for cross-pollination are identified, chiefly pertaining to directions for future development of consumer satisfaction theory and measurement.
[ to cite ]:
Robert A. Westbrook (1984) ,"On Satisfaction Responses and the Effects of Product Labels", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 311-313.

Advances in Consumer Research Volume 11, 1984      Pages 311-313


Robert A. Westbrook, University of Arizona


A study of consumer satisfaction with philanthropic contributions to health service organizations and another of consumer response to cigarette warning labels are reviewed. Several opportunities for cross-pollination are identified, chiefly pertaining to directions for future development of consumer satisfaction theory and measurement.


A number of interesting insights and directions for future research on consumer satisfaction are suggested by the Moore and Shuptrine (1984) and Bhalla and Lastovicka (1984) papers. While the exploration of such issues is always relevant in any line of scientific inquiry, it is especially appropriate in satisfaction research where needs include richer conceptualizations, improved measurements, and alternative theoretical paradigms. Certainly these reeds are rot unique to studies of consumer satisfaction indeed, they are descriptive of consumer research more broadly as well. Consumer satisfaction, however, occupies a central role in the raison d'etre of consumption. Accordingly, the observations contained in this paper will focus primarily on knowledge development in the consumer satisfaction arena. Brief reviews are presented to place into proper perspective both the Moore and Shuptrine and Bhalla and Lastovicka papers, followed by a discussion of some of the larger issues facing consumer satisfaction research that are suggested by the papers. Fuller assessments of the present state of knowledge and the agenda for future research may be found in reviews by Day (1983) and Swan (1983).


In this paper, Moore and Shuptrine (1984) offer an empirical test of a model of consumer satisfaction and its antecedents and consequences. The model was originally proposed by Oliver (1980) and reflects the dominant conceptual approach to the explanation of consumer satisfaction among contemporary researchers. This approach, it is worth noting, views sentiments of consumer satisfaction in the context of the broader process beginning before the decision to make a particular purchase and continuing into and throughout the period following the purchase, during which consumption occurs and any redress-seeking or complaining takes place. Specifically, the model proposes that consumer satisfaction judgements are determined directly by two separate and identifiable cognitive elements; (1) the pre-choice expectancies, or beliefs about the probable outcomes of the anticipated purchase, and (2) the post-choice disconfirmation beliefs about to the extent to which actual outcomes correspond to prior expectations. To date, the model has received empirical support in a variety of contexts, and Moore and Shuptrine thus provide additional empirical validation for the model's prepositions, a useful contribution to establishing the generality of the basic model. Moreover, they have demonstrated the robustness of the model under alternative measurement methods for the satisfaction construct and the disconfirmation construct. This, too, comprises a useful addition to current knowledge, inasmuch as there has been little agreement to date on measurement methods in satisfaction research.

Despite these contributions, Moore and Shuptrine's study leaves a number of crucial questions unanswered. Perhaps most significant is the issue of whether it is meaningful to conceptualize satisfaction in relation to consumer philanthropic activity such as contributing to the American Cancer Society and American Heart Association. Is satisfaction a natural outcome of the act of contributing, at least in the same sense as its occurrence after the consumption of economic goods and services? The study to provide a convincing answer as to whether indeed the pressured sentiments of satisfaction actually took place, and were not simply the product of the researcher's instruments. The issue of whether satisfaction always occurs in regard to produce/service purchases and consumption has been discussed by Day (1977) and Swan and Trawick (1979). Clearly, the issue is deserving of greater conceptual and methodological attention then the authors have devoted to it.

Closely related to the question of whether "philanthropic satisfaction" actually occurs is the matter of its prior determinants and consequences on subsequent behavior. This question, in essence, is whether the model employed by the researchers is indeed reasonably isomorphic to the actual underlying cognitive/affective processes. To assert that it is reflective of actual precesses is equivalent to arguing that most consumers contribute because of expectations they hold as to benefits or outcomes they stand to realize from the donation or act of giving. Yet much of this type of charitable behavior would appear to occur because of pressure from peers and/or employers, or simply out of compliance with socially desirable behavior norms. If giving is determined by these forces, an attempt to explain satisfaction as a function of selected cognitive elements such as expectation, disconfirmation, and attitude, will more than likely simply turn up cognitive consistency response biases by survey respondents. Thus, the observed correlational evidence for the postulated model may not actually support its validity in this unusual context, but rather reflect spurious relationships owing to social desirability and consistency factors. To overcome these difficulties, the authors night usefully provide evidence from depth interviews as to the evaluations and judgements preceding and following the act of giving, showing that the proposed model does in fact appear to operate, at least in a qualitative sense.

A final criticism of the study concerns the propriety of its conclusions on several substantive and methodological issues. The authors claim their results suggest revisions to the basic expectancy disconfirmation model in that (a) expectations and disconfirmation may not be independent, as proposed by Oliver (1980), and (b) that disconfirmation may directly influence post-choice attitude, a linkage not postulated in the basic Model. Such empirically-biased revisions in the model, which is itself theoretically grounded, should be approached extremely cautiously when multicollinearity appears a problem in the empirical measurements. The authors recognize the latter difficulties, but unfortunately a correlation matrix is not presented, so we are unable to judge how serious the problem actually is. Moreover, the authors have operationalized the concept of expectations somewhat differently than in Oliver's (1980) study, thus reducing the comparability of results for purposes of proposing revised theoretical linkages.

Methodologically, the authors argue for the superiority of their 3-item Likert-type satisfaction measure in preference to the Delighted-Terrible scale (Westbrook 1980), as well as their "Difference" measure of disconfirmation. Examination of the path coefficient reported in the paper reveals that the authors' satisfaction measure may have less discriminant validity with respect to the concept of attitude than the D-T scale. Moreover, the latter has received impressive validation in a wide variety of product and non-product-related satisfaction domains. Problems with before-after "difference" measures of disconfirmation are well known to researchers employing panel designs, and may suffer from a "regression to the mean" effect. Oliver (1977) has already cautioned against the use of these types of measures in satisfaction research. The authors' own findings also indicate serious multicollinearity problems with the measure.


Bhalla and Lastovicka (1984) in this paper seek to evaluate some of the cognitive effects of changes in the content and format of the federally-required health warning message appearing in cigarette print advertising. There can be little question of the significance of the bread problem area investigated, and the authors' efforts to design impact-evaluation research for public policy purposes are likewise noteworthy. To provide a conceptual framework for examining the cognitive effects of the various informational changes contemplated, Bhalla and Lastovicka invoke script theory, which despite its promises has yet to receive commensurate application in consumer research.

While the controlled laboratory environment of the study adds to the internal validity of conclusions based or the experimental manipulations, the usefulness of the findings is greatly impaired by the forced, single exposure of subjects to only modestly revised warning messages. Much to the authors credit, these limitations are acknowledged. However, the limitations so jeopardize the external validity of the research results that little may be concluded as to the ultimate efficacy of content and or format message modifications for practical purposes. While Bhalla and Lastovicka's results suggest that format change might be a requirement for obtaining any consumer learning of a new message, it is unclear whether this stipulation would also occur under conditions of multiple exposure. If Krugman's (1972) "three exposure" hypothesis has any merit, then a number of repetitions of the new message in any format might be required before significant note is taken by consumers. Although the realism of the laboratory experimentation in the study is commendable, its replication in a field setting should appear essential prior to any policy decision.

The research has potential value as a test of selected script-theoretic notions which seem well suited to the "advertising readership" problem at hand. However, a most important unanswered question in this regard is some evidence as to whether consumers do typically engage in scripted behavior when reading print advertising in general or cigarette advertising in particular. Recent studies (e.g. Bozinoff and Roth 1983) suggest that it may be possible to empirically elicit scripts in various consumer behavior contexts. Had such findings been presented in Bhalla and Lastovicka's study, its contribution to theory development would be appreciably greater. Falsifying hypotheses based on a theory only loosely linked to the behavior in question is at best a limited contribution.

The finding that so few subjects were able to correctly recall the revised message content under any format is not actually a surprising finding. First, the phenomenon of cognitive intrusion errors in scripted behavior has been well documented (Whitney and John 1983). Here subjects may simply have "intruded" their own scripts for reading or viewing cigarette ads, which would of course contain the present warning message. Second, the readership of cigarette ads suggests very low levels of attention to textual material--according to Starch studies, typically fewer than 10% of an ad's readers report reading some of the textual materials in the advertisement, even in pictorial executions.


As ar area of inquiry, consumer satisfaction has attracted increasing attention, and Bunt's (1483) bibliography includes over 400 citations on the subject. In this light, the two papers reviewed above raise a number of important and timely issues about the practice, scope and future directions of consumer satisfaction research. Though the intent of the authors was perhaps more modest, their empirical investigations provide an impetus to consider these issues in relation to the developmental status of the study of consumer satisfaction research more broadly.

First and foremost is the question of the scope of satisfaction, i.e. when and with respect to what consumption phenomenon it occurs. To date, satisfaction has been explored in regard to products and services of many kinds, retail outlets, shopping activities, and the larger marketplace and market system. Typically, satisfaction is assured to follow consumption or performance of a behavioral act as a matter of course. It may not, especially if the consumer has for some reason not yet formed an evaluation of his/her experiences. Such evaluations may be triggered by the presence of relatively high or low levels of affect during consumption. They may also require some amount of perceived volitional control over the behavior whose consequences are being evaluated. These speculations are deserving of further study by satisfaction researchers, but an answer to the questions of when satisfaction arises and when it does not, is essential to the development of theory. Research is also needed to develop methods of measuring satisfaction which do not by their own application create the phenomenon we are interested in studying.

A second issue concerns the need to develop additional theoretical accounts of how satisfaction is determined. Extant models, such as the expectancy disconfirmation model (Oliver 1980), the equity model (Fuppertz 1978), the comparison-level model (LaTour and Peat 1980), and the value-percept model (Westbrook and Reilly 1983), all assume a highly cognitive evaluation process leading to satisfaction responses. Perhaps the antecedents of satisfaction are more affective in nature, such as the assessment of relative preponderance of positive versus negative emotional states during prior consumption activities. If much of consumer behavior is determined by low levels in cognitive activity, as many researchers have suggested, then the deliberate, evaluative processes posed by our satisfaction models may be inappropriate. Indeed, as typified by Moore and Shuptrine's results, the postulated model succeeds in explaining only a fraction of the variance in observed satisfaction responses, even with reasonably reliable measures of constructs. Here is where the Bhalla and Lastovicka paper offers some cross-pollination potential. Are many consumer satisfaction responses themselves actually "scripted behavior," reflecting simply the absence of any significant purchase-related problems? Viewing the dependent variable in this way may lead to further insight into its very makeup, antecedents, and consequences.

A third issue facing satisfaction research is the standardization of construct measurement. As the Moore and Shuptrine paper reflects, progress is being made in the increasing commonality of measures across studies and investigated. However, further advances in this area are essential if empirical results are to be meaningful compared and theoretical notions properly evaluated and refined.

The fourth and final issue of interest suggested by joint consideration of the papers concerns the identification of sources of consumer satisfaction/dissatisfaction. That is, from what aspects of the consumption experience do these sentiments arise? In particular, what are the respective roles played by (a) the physical product or tangible aspects of the service purchased, (b) advertising positioning and imagery, (c) retail sales personnel and store environment characteristics, (d) retail pricing, (e) package labelling, consisting of usage instructions, warnings, etc. The latter is of particular interest in the context of this review. For some products, the labelling may be instrumental in assuring proper usage of the product which in turn is required for achieving high level of satisfaction (e.g., food products, appliances). For others, package labelling may create or reinforce product imagery which constitutes a substantial portion of the satisfaction obtained from consumption experiences (e.g., cosmetics, perfumes). From the labelling perspective, perhaps the low levels of attention to the cigarette warning message noted by Bhalla and Lastovicka reflects not only scripted behavior, but also selective avoidance by smokers owing to the warning message's potential to reduce the satisfactions afforded by smoking. The implication of such a view is that disconfirmation perceptions may not be solely causally-prior to satisfaction judgements, but perhaps partly a product of them as well.


Bhalla, Gaurav and John L. Lastovicka (1¦84), "The Impact of Changing Cigarette Warning Message Content and Format," in Thomas C. Kinnear, ed., Advances in Consumer Research, 11 Ann Arbor: Association for Consumer Research.

Bozinoff, Lorne and Roth (1983), "Recognition Memory for Script Activities: An Energy Conservation Application," in Andrew Mitchell, ed., Advances in Consumer Research, 9, Ann Arbor: Association for Consumer Research, 655-660.

Day, Ralph L. (1477), "Extending the Concept of Consumer Satisfaction," in William D. Perreault, ed., Advances in Consumer Research, 4, Ann Arbor: Association for Consumer Research. 149-54.

Day, Ralph L. (1983), "The Next Step: Commonly Accepted Constructs for Satisfaction Research," in Ralph L. Day and P. Keith Hunt, eds., International Fare in Consumer Satisfaction and Complaining Behavior, Bloomington: Indiana University School of Business.

Hunt, H. Keith (1983), "A CS/D & CB Bibliography 1982," in Ralph. L. Day and H. Keith Hunt, eds ., International Fare in Consumer Satisfaction and Complaining Behavior, Bloomington: Indiana University School of Business.

Huppertz, John W. (1979), "Measuring Components of Equity in the Marketplace: Perceptions of Inputs and Outcomes by Satisfied and Dissatisfied Consumers," in R.I. Day and H.K. Hunt, eds., New Dimensions of Consumer Satisfaction and Complaining Behavior, Bloomington: Indiana University School of Business, 140-143.

Krugman, Herbert E. (1972), "Why Three Exposures May be Enough," Journal of Advertising Research, 12, 11-14.

LaTour, Steven F. and Nancy Peat (1980), "The Role of Situationally-Produced Expectations, Others Experiences, and Prior Experiences in Determining Consumer Satisfaction," in Jerry C. Olson, ed., Advances in Consumer Research, 7, Ann Arbor: Association for Consumer Research, 588-92.

Moore, Ellen and F. Kelly Shuptrine (1984), "Disconfirmation Effects or Consumer Satisfaction and Decision Making Processes," in Thomas C. Kinnear, ed., Advances in Consumer Research, Ann Arbor: Association for Consumer Research.

Oliver, Richard L. (1977), "A Theoretical Reinterpretation of Expectation and Disconfirmation Effects on Post-Exposure Product Evaluations: Experience in the Field," in R.L. Day, ed., Consumer Satisfaction. Dissatisfaction and Complaining Behavior, Bloomington: Indiana University School of Business, 2-9.

Oliver, Richard L. (1980), "A Cognitive Model of the Antecedents and Consequences of Satisfaction Decisions," Journal of Marketing Research, 17(November), 460-469.

Swan, John E. (1983), "Consumer Satisfaction Research and Theory: Current Status and Future Directions," in Ralph L. Day and H. Keith Hunt, eds., International Fare in Consumer Satisfaction and Complaining Behavior, Bloomington: Indiana University School of Business.

Westbrook, Robert A. (1980), "A Rating Scale for Measuring Product/Service Satisfaction," Journal of Marketing, 44, 68-72.

Westbrook, Robert A. and Michael D. Reilly (1983), "Value Percept Disparity: An Alternative of the Disconfirmation of Expectations Theory of Consumer Satisfaction," in R.P. Bagozzi and A.M. Tybout, eds., Advances in Consumer Research, Ann Arbor: Association for Consumer Research.

Whitney, John C. and George John (1983), "An Experimental Investigation of Intrusion Errors in Memory for Script Narratives," in Andrew Mitchell, ed., Advances in Consumer Research, 9, Ann Arbor: Association for Consumer Research, 661-66.