How Satisfactory Is Research on Consumer Satisfaction?

ABSTRACT - The past several years have seen remarkable growth in the quantity of published research on consumer satisfaction and dissatisfaction. This paper evaluates the general quality of this new area of consumer research and critiques the three preceding papers. The quality is mixed but appears to be improving.

Citation:

Ralph L. Day (1980) ,"How Satisfactory Is Research on Consumer Satisfaction?", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 593-597.

Advances in Consumer Research Volume 7, 1980     Pages 593-597

HOW SATISFACTORY IS RESEARCH ON CONSUMER SATISFACTION?

Ralph L. Day, Indiana University

ABSTRACT -

The past several years have seen remarkable growth in the quantity of published research on consumer satisfaction and dissatisfaction. This paper evaluates the general quality of this new area of consumer research and critiques the three preceding papers. The quality is mixed but appears to be improving.

INTRODUCTION

This is the fifth consecutive ACR Conference to have at least one session devoted to research on consumer satisfaction, dissatisfaction and/or complaining behavior. Following the first of these sessions in 1975, an international conference (workshop) devoted entirely to consumer satisfaction was organized by Keith Hunt, funded by the National Science Foundation, and conducted by the Marketing Science Institute in Chicago in April, 1976 (Hunt 1977). This was followed by a second conference organized by Ralph Day and conducted at Indiana University in Bloomington in April of 1977 (Day 1977). Subsequent annual conferences took place in Chicago in 1978 (Day and Hunt 1979) and Bloomington in 1979 (Hunt and Day 1980, forthcoming). The next conference, described by Keith Hunt as "a fifth of satisfaction," will take place at Washington University in Saint Louis in October, 1980. The past ACR sessions and the four national "CS/D" conferences have jointly produced approximately 120 published proceedings papers. Although the combination of ACR sessions and CS/D conferences has generated world wide excitement, the paucity of papers published in major refereed journals is startling. Only a handful of research papers on consumer satisfaction have appeared in major American journals during the past five years. The most notable exceptions are Swan and Coombs (1976), Westbrook and Newman (1977), and Oliver (1977).

Because a generally accepted criterion for evaluating the quality of a stream of research is the extent to which it penetrates major research journals, scholars who have made a commitment to research on consumer satisfaction appear to have cause for concern. The following section will address some of the apparent reasons for the small number of papers reporting satisfaction research to be published in journals.

FACTORS INFLUENCING QUALITY

There are a number of possible explanations for the limited success of satisfaction researchers in getting their papers accepted by major research journals. One obvious factor is the newness of the topic. While there were soma earlier papers pointing out the need for satisfaction research and a few experimental studies dealing with aspects of expectation formation and disconfirmation, the present stream of empirical research on satisfaction/dissatisfaction can be traced back no more than six or seven years. The publication of satisfaction research in substantial quantities goes back no more than four or five years. Therefore editors and reviewers are not likely to be familiar with the research area and are not as well qualified to judge it as they would be for more familiar streams of research. Because of the lack of an established research tradition to provide researchers with examples of quality research, one would expect high variability in the quality of research. While it is clear that the newness of the satisfaction research area contributes to the apparent quality problem, there are also several other factors which are not totally a result of newness. These include: the lack of a well developed and widely accepted theoretical framework; the generally pragmatic public policy orientation of much of the research which results in stress on the collection of descriptive data rather then on hypothesis testing; the general lack of adequate funding which has led to many "quick and dirty" studies with small convenience samples; and, a considerable amount of output from researchers whose papers lack rigor in research design and fail to use sophisticated analytical techniques when appropriate. Each of these factors will be discussed below.

Lack of Theory

While everyone knows what satisfaction means, it clearly doesn't always mean the same thing to everyone. It has been defined by researchers in a variety of ways including the following: a level of "happiness' resulting from a consumption experience;" a cognitive state resulting from a process of evaluation of performance relative to previously established standards; a subjective evaluation of the various experiences and outcomes associated with acquiring and consuming a product relative to a set of subjectively determined expectations; a two factor process of evaluating a set of "satisfiers" and a set of "dissatisfiers" associated with the product; and, one step in a complex process involving prior attitude toward a product or service, a consumption experience resulting in positive or negative disconfirmation of expectancies, followed by feelings of satisfaction or dissatisfaction which mediate post consumption attitude which subsequently influences future purchase behavior. The above definitions represent a range from the notion of a simplistic "black box" happiness function to a very complex set of concepts which overlap each other and are virtually impossible to operationalize. To say that there is no general agreement among satisfaction researchers on how to define satisfaction would indeed be an understatement.

Given that one has accepted a particular definition of satisfaction, many theoretical and measurement issues remain. At the simplest level one can assume that a consumer buys a product, immediately consumes it, instantaneously evaluates it, and responds with feelings of satisfaction or dissatisfaction. Even in this situation, influences of many kinds can affect the outcome. This basic scenario fits only a small set of product/service situations most of which involve relatively unimportant consumption experiences. For many products, prepurchase activities may be important and may take place over time, consumption of most products takes place over time and under a variety of circumstances; therefore, postpurchase evaluation itself is very complex and takes place at different points in time.

Until definitional problems are mastered and some appropriate theoretical framework for conceptualizing and measuring satisfaction is agreed upon, efforts involving the testing of causal relationships will continue to be difficult to publish except in proceedings volumes of the ACR and the annual CS/D conferences as in the past. While there seems to be some movement toward consensus on appropriate definitions of satisfaction, the rate of convergence is less than rapid.

Pragmatic Orientation

Much of the impetus for the current stream of satisfaction research came from the desire of public policy decision makers with consumer protection responsibilities to obtain better information on consumer dissatisfaction than that provided by volunteered complaint data. It might be more correct to say the impetus came from a small group of consumer behavior "experts" who were able to convince some consumer protection decision makers that volunteered complaint data were totally inadequate for decision making purposes and that survey research data from a representative sample of the population was needed. This led to the design of research which sought to identify those categories of products and services which had relatively high instances of dissatisfaction and/or complaints. The basic idea was to assist consumer protection agencies to allocate their resources to those particular items or categories which generate the highest levels of dissatisfaction.

Surveys for the purpose of assessing satisfaction/dissatisfaction over the full range of consumer goods and services had to deal with categories of similar products and measure generalized feelings of satisfaction with these categories over a specified recall period. As scholars of marketing have long known, most products and services please most consumers. Defects and unsatisfactory performances are the exception rather than the rule even for products with relatively high incidences of dissatisfaction (e.g., new cars, auto repairs, moving and storage). Thus the expectation of good performance for the great majority of consumer goods provides a frame of reference for identifying those products which have defects or are unsatisfactory in some other way. In this context, each consumer provides her own standard for determining her degree of satisfaction or dissatisfaction with each particular product or service.

Recognizing that some consumers will set higher standards than the typical person and others set lower standards, tabulations of self assessed satisfaction ratings over representative samples of consumers provides valuable data for scanning product and service categories for problem areas. Studies in the U.S. (Day and Bodur 1977, Day and Ash 1978, Leigh and Day 1979) and Canada (Ash and Quelch 1980) have demonstrated extremely wide ranges of variation in satisfaction over products and over individuals. Such results provide very interesting and useful information about the performance of manufacturers and marketers. However, these results are basically descriptive and are typically obtained by straightforward research procedures. Therefore, it is to be expected that such descriptive studies might be publishable for their substantive value in conference proceedings and perhaps in managerially oriented journals but not in research journals. A substantial percentage of the CS/D papers which have been published in the past are of interest for their findings rather than for contributions to theory or methodology. It seems likely that papers of a descriptive nature are rapidly becoming less acceptable even for specialized conferences or ACR sessions.

Lack of Funding

An emphasis on measuring satisfaction in real world settings has characterized the recent stream of satisfaction research. As a result it seems unlikely that very small sample surveys or laboratory experiments using student subjects will be taken seriously even for future CS/D conferences. To have a chance for journal publication, or perhaps any publication at all, future studies should be based either on field experiments or very sophisticated surveys with substantial real world samples. Sources of funding for such studies do not appear to be readily available. Although some of the initial public policy oriented studies were developed for federal agencies, funding for the field work was not forthcoming. For example the 1974 Bloomington and Boulder studies (Day and Landon 1975) used an instrument developed at the FTC but had to be funded locally. The same is true for the 1976 Bloomington study (Day and Bodur 1977). It was designed on a research contract with the Office of Consumer Affairs but had to be funded on a shoestring basis locally. The instrument finally was used nationally in the summer of 1979 with a sample of more than 3,000 but in Canada rather than the United States. Prospects for future funding by the federal government of the United States do not seem promising.

While there has been some funding of satisfaction research by business firms, the practice does not appear to be widespread. Most business firms tend to underestimate the importance of information on consumer satisfaction. Research designed and conducted by consumer research scholars could provide companies with valuable data at low cost while providing scholars with opportunities to make contributions to knowledge. Progress in consumer satisfaction research could be accelerated considerably if business firms, federal agencies or both would fund faculty research projects.

Lack of Analytical Sophistication

There is some indication that an important reason for the failure of CS/D research to penetrate major research journals is the lack of methodological sophistication in papers submitted. In many cases this is simply because the use of sophisticated analysis is not necessary to present the results in an understandable way. This is often the case for public policy oriented studies. However, cases undoubtedly occur when papers would have a better chance for publication if the research design was more elaborate and/or more sophisticated analysis was done. In particular, I am aware that some interesting studies have been turned down by a major journal because of a failure to demonstrate the reliability and validity of the measures being used or to provide more careful operationalizations of concepts.

In some cases, scholars who are interested in satisfaction research may not have been trained in research design or sophisticated analytical techniques of the type that might be expected in papers accepted by the Journal of Marketing or the Journal of Consumer Research. Researchers in this situation should interact with other satisfaction researchers with strong methodological skills or seek to involve colleagues with such skills in joint research. In the early stages of the development of satisfaction research, rough ideas and quick and dirty studies were of interest to others. This will no longer be the case even among the highly sympathetic and tolerant group who have participated in the first four annual CS/D conferences. Increased success in major refereed journals will require considerably higher levels of methodological sophistication than has been typical of CS/D research in the past.

At this point it may seem that a case has been made for an answer of "not at all" to the title question of this paper. To the contrary, the past quality has been quite satisfactory when evaluated with respect to realistic expectations of research in a new field which is fundamentally exploratory in nature. However, it must be recognized that the "honeymoon" period for satisfaction research is over and future research must be evaluated against rising expectations.

COMMENTS ON THE THREE PAPERS

Two of the three papers presented in this session (Westbrook and Cote 1980, LaTour and Peak 1980) are in the mainstream of current satisfaction research and represent efforts to move from purely descriptive studies to tests of causal relationships. These two papers provide examples of the present state of the art in consumer satisfaction research and illustrate some of its problems and opportunities. The third paper (Cosmas, Samli, Meadow 1980) is more like some of the earliest papers on consumer satisfaction of four or five years ago in that it is concerned with conceptualization at a very general level and has no empirical content. The linkage of this paper to the stream of research on consumer satisfaction is not strong. While it is possible to conceptualize quality of life (QOL) as the individual's overall level of satisfaction with "the total human experience" (Day 1978), "life satisfaction" is but one of four approaches to QOL discussed in the paper. Each of the three papers is discussed below, in the order in which they were presented. The comments will focus on the assessment of the contribution of the papers and the urge to summarize has been restrained in the interest of length constraints.

Westbrook and Cote

This paper seeks to make a contribution to understanding the processes through which consumers generate feelings of satisfaction or dissatisfaction. The approach is partially based on the work of Richard Oliver (1977, 1979) who has sought to separate the cognitive and affective components of the purchase, consumption, evaluation process. Westbrook and Cote identify and seek to measure three aspects of the process: (a) the cognitive process of comparison of experiences with prior expectations (confirmation/disconfirmation of expectations); (b) relatively stable affective structure; and (c) transient affective states which vary over products, situations, and time. These three aspects are then related to the dependent variable, satisfaction, as a means of evaluating the relative effects of cognitive evaluations, stable affective structures, and transitory affective elements. Data were collected for two product classes (automobiles and footwear) and multiple regression analyses were done for each product class. Respondents were 194 students who completed self-administered questionnaires in class.

Satisfaction was operationalized as a single 0-100% rating scale anchored by "not at all satisfied" and "completely satisfied." Expectancy confirmation was measured on product features (both good points and bad points) identified by the respondent who then rated them on separate five point scales on the extent to which product performance was greater or less than expected. The two scales were summed to form a single measure. The other independent variables were: optimism (eight item summated scale), pessimism (eight item summated scale), life satisfaction (single graphic rating scale), consumer discontent (41 item form), and mood favorability (combined scores on elation-depression and harmony-anger batteries). The first four affective variables represented stable affective states and the final one represented transitory affective states.

While the correlations among the variables were generally in the hypothesized direction, the relationships of the independent variables with the satisfaction measure tended to be quite weak. This was especially true with the variables chosen to reflect affective structure. Inspection of the intercorrelation matrix revealed only two "middle range" correlations and these were related to the cognitive variables confirmation/disconfirmation of expectations. Satisfaction with automobiles was correlated .48 with expectancy confirmation/disconfirmation for automobiles and footwear satisfaction was correlated .46 with expectancy confirmation/disconfirmation for footwear. Given the wide acceptance of the strong relationship between satisfaction and the confirmation/disconfirmation of expectations, even these relationships are quite weak (the r²'s are .23 and .21). Among the 10 intercorrelations of the two satisfaction measures with the five affective variables, only one was significant at the .01 level [automobile satisfaction with life satisfaction (r = .22, r² = .048)] and two at the .05 level [automobile satisfaction with consumer discontent (r = .20, r² = .04) and footwear satisfaction with life satisfaction (r = .16, r² = .026) 3.

As could be expected, multiple regression analyses with the expectancy confirmation/disconfirmation variable included resulted in this variable having the dominant beta weight for both product classes. It was the only variable with a statistically significant beta for footwear. For automobiles, the betas for life satisfaction and consumer discontent were quite small relative to expectancy confirmation/disconfirmation but were significant at the .05 level. With confirmation/disconfirmation in the regression ("complete" model) the adjusted R²'s were .321 for automobiles and .220 for footwear. When only the five affective variables are retained ("reduced" model) the adjusted R²'s drop to .073 and .002. For cars two betas (life satisfaction, consumer discontent) are significant at the .05 level but none are significant for footwear.

The authors say they found "mixed support" for the hypotheses of the study relating to the set of affective variables selected for the study. Given the obviously weak results described above, this conclusion is based more on optimism than evidence. Given the fact that the hypotheses were stated in general terms (applicable to all products and services under all circumstances), one must conclude that the overall results were generally damaging to the author's hypotheses. They chose a sample of two products from thousands of possible products and got very weak support in one case and a complete absence of support in the other. Which of the two product classes is more representative of the total population of products and services? In my opinion, it is footwear since cars are one of a handful of uniquely important products which represent a major financial commitment on the part of the consumer and play a central role in daily activities. On the other hand, footwear is more typical of the great majority of products and services. Even if the results for automobiles in this study had been more favorable, this would not have supported a general conclusion about all products and services.

This study is useful because it states some interesting hypotheses which are worthy of investigation and reports an effort to evaluate them. The results suggest that future research should attempt to more clearly define the concepts of a theory of satisfaction/dissatisfaction, investigate more ways to operationalize these concepts, and utilize more powerful methods for investigating causal relationships. The regression methods used in this study are not well suited to investigating causal linkages. For circumstances where resources or other circumstances limit the researcher to data from surveys, structural modeling (path analysis) or canonical correlation analysis might prove more fruitful. This study has been useful in bringing attention to the product specific elements in the study of satisfaction, suggesting the need for satisfaction studies to be concerned with representative samples of products as well as representative samples of consumers.

LaTour and Peat

Experimental research has played a quite limited role in the recent outpouring of research on consumer satisfaction/dissatisfaction. This paper is a welcome exception; unlike most previous CS/D experimental studies, it was done in the field rather then in a laboratory setting. Field experiments seem more promising for CS/D research since they allow the development of research designs which retain realistic aspects of the consumer's environment while permitting the manipulation of key variables.

In the LaTour and Peat study the products were actual window cleaning products which were adulterated for "poor quality" treatments and were tested by subjects in their own homes. The independent variables were prior experience, manufacturer induced expectations, and experiences of others. Prior experience was simulated in a first experience with the product class in which half of the sample received the product as manufactured and half received adulterated bottles. For the second experience with the product, manufacturer induced expectations were operationalized by written statements indicating either a high quality product or an inexpensive product of low quality. Experience of others was operationalized by average ratings by consumers of four other brands. Dependent variables were: perceived amount of streaking, perceived amount of residual film, satisfaction/dissatisfaction with amount of streaking, satisfaction/dissatisfaction with amount of film, overall satisfaction/dissatisfaction with the product, and likelihood of purchase. In the paper, results for the perceived amount of streaking, perceived amount of film, overall product satisfaction, and intentions to buy are discussed.

The analysis of the results with respect to the perceived streaking and film dependent variables was a 2 (prior experience) x 2 (manufacturer induced expectations) x 3 (experiences of others) multivariate analysis of variance. Significant effects were found for prior experience (whether the first trial involved adulterated or unadulterated cleaner) and for experiences of others (consumers' ratings) but not for manufacturers statements with regard to quality. Those with poor "prior experience'' perceived less streaking than those who received the unadulatered product. Those who received higher "experience of others" ratings reported lower perceptions of streaking and film. These results support the notions that expectations reflect one's own experiences and knowledge of others experiences.

The results with overall satisfaction as the dependent variable were significant only for "prior experience," showing that those who received the adulterated product in the first stage were more satisfied then those who received the unadulterated product. This is somewhat surprising since "experience of others" affected perceptions of product quality which in turn would be expected to influence reported satisfaction. The authors suggest a "cancellation effect" of satisfaction derived from perceiving the same level of quality reported by others. I am less than overwhelmed by this speculation and prefer to assume that one might use different rules in using (combining, weighting) information to make perceptual judgments than when assessing one's own felt satisfaction/dissatisfaction. The results for the "likelihood of purchase" dependent variable paralleled the results with perceptions of product quality, rather than for reported satisfaction. That is, the effects of one's own prior experience and the reported experiences of others are significantly related to the reported likelihood of purchases but the effects of information supplied by the manufacturer are not.

These results are interesting but are not exactly surprising. The experimental design imposed a recent "prior experience" on each subject which is hardly the universal situation in which real world consumers find themselves. Also, the experience is not only recent but is concrete and personal, in contrast with printed information about what manufacturers say or about what unknown others say. The dominance of prior experience appears to be to a considerable extent an artifact of the experimental design.

This study represents a step in the right direction. However, the authors and others should resist the urge to generalize its findings until extensions and replications have been done. The process of choice, consumption, and evaluation of products is extremely complex when viewed within a single product class and varies widely over products and services, individuals, and circumstances. Simplistic theories derived from artificial circumstances can be dysfunctional to long term theory construction. Experimentation is a powerful research tool but experimental studies must be designed and interpreted with caution. Hopefully, this paper will stimulate further field experiments among CS/D researchers.

Cosmas, Samli and Meadow

This paper is concerned with a different body of literature than the two satisfaction/dissatisfaction papers presented in this section. The CS/D literature is devoted primarily to the measurement of satisfaction with specific products and services and to a lesser degree with the measurement of satisfaction with marketing institutions such as retail stores and satisfaction with marketing practices such as advertising. The literature on the quality of life (QOL) is concerned with the evaluation of all aspects of human life of which the purchase, consumption, and evaluation of products and services is but a small part.

Although consumer satisfaction is a highly limited concept in comparison to quality of life, there are some parallels. In fact, an alternative term to QOL is life satisfaction. If one seeks to measure overall satisfaction with life as a single concept, then this process can be viewed as similar to measuring satisfaction for a product. The difference is that the product becomes "the total life experience" at some point in time and this is an incredibly complex concept to deal with.

This paper is frustrating to read because the authors say in the opening line of their abstract that quality of life is "an important concept" but are never willing to deal with it as a single concept. Instead of telling the reader what concept they are trying to define, the authors refer to several lines of research and/or streams of literature with varying degrees of relevance to a unified concept of QOL. Not only do they avoid defining a single concept but they avoid the question of whom the concept will be defined for. The individual? Some group of people? The residents of a particular area or political unit? The paper succeeds in avoiding the three questions it should be trying to answer: (1) What does "quality of life" mean?; (2) For whom should the concept be defined?; (3) How should it be measured?

Rather than address any of these questions, the authors organize the paper around the following five streams of literature:

1.  Social indicators

2.  Economic measures

3.  Life satisfaction

4.  Life styles

5.  Measures of equal opportunity to consume

These streams of literature are discussed individually, as if there were no overlap. No effort is made to integrate the streams of literature or synthesize them. The first three streams of literature overlap to a considerable extent. In fact it is clear that life satisfaction encompasses the subject matter of social indicators and economic measures. As a matter of fact, this is graphic-ly summarized in Figure 3. The literature of life style analysis is not clearly or closely related to QOL. Rather, it deals with the identification of unique patterns of life rather than with the influence of the chosen life style on QOL. The final item on the above list relates to a very small body of literature concerned with the protection of consumer rights, especially their rights to information. While the granting or blocking of rights can affect an individual's QOL, this is as true for a multitude of other human rights as for consumer rights and it is not clear that consumer rights have special treatment (or that they are not already covered in the first three streams of literature).

In summary, I feel that this paper did not fulfill its opening promises of a synthesis of the literature and the presentation of a useful conceptualization of the quality of life concept. Moreover, the reference list fails to include one of the most useful and best known books in this area (Andrews and Withey 1976).

CONCLUSIONS

The papers in this session provided a reasonably representative sample of the nature and quality of research currently being done on consumer satisfaction/dissatisfaction. The Westbrook and Cote paper provided an example of survey research; the LaTour and Peat paper reported a field experiment; and the Cosmas, Samli, and Meadow paper was a literature review and conceptual paper on life satisfaction. While each paper can best be regarded as a pilot study or first effort, the results were interesting and potentially useful. Research on CS/D appears to be gradually shifting away from descriptive studies of levels of satisfaction toward the development and testing of theoretical structures. Hopefully, CS/D research will improve in quality over the next few years and increase its penetration of major research journals.

REFERENCES

Andrews, Frank M. and Withey, Stephen B. (1976), Social Indicators of Well Being, New York: Plenum Press.

Ash, Stephen B. and Quelch, John A. (1980), "Consumer Satisfaction, Dissatisfaction and Complaining Behavior: A Comprehensive Study of Rentals, Public Transportation, and Utilities," forthcoming in H. Keith Hunt and Ralph L. Day, editors, Proceedings, Fourth Annual Conference on Consumer Satisfaction, Dissatisfaction, and Complaining Behavior, Bloomington: Division of Research, Indiana University.

Cosmas, Stephen C., Samli, A. Caskun and Meadow, H. Lee (1980), "Toward an Interdisciplinary Framework for Examining Quality of Life," Jerry C. Olsen, editor, Advances in Consumer Research, Vol. 7, Association for Consumer Research.

Day, Ralph L. (1977), editor, Consumer Satisfaction, Dissatisfaction and Complaining Behavior, Division of Research, Indiana University.

Day, Ralph L. (1978), "Beyond Social Indicators: Quality of Life at the Individual Level," in Fred D. Reynolds and Hiram C. Barksdale, editors, Marketing and the Quality of Life, Chicago: American Marketing Association.

Day, Ralph L. and Ash, Stephen B. (1979), "Consumer Response to Dissatisfaction with Durable Products," in William Wilkie, editor, Advances in Consumer Research, Vol. 6, Association for Consumer Research, 438-40.

Day, Ralph L. and Bodur, Muzaffer (1978), "Consumer Response to Dissatisfaction with Services and Intangibles,'' in H. Keith Hunt, editor, Advances in Consumer Research, Vol. 5, Association for Consumer Research, 263-72.

Day, Ralph L. and Hunt, H. Keith (1979), editors, New Dimensions of Consumer Satisfaction and Complaining Behavior, Bloomington: Division of Research, Indiana University.

Day, Ralph L. and Landon, Jr., E. Laird (1976), "Collecting Consumer Complaint Data by Survey Research," in Beverlee E. Anderson, editor, Advances in Consumer Research, Vol. 3, Association for Consumer Research, 263-68.

Hunt, H. Keith (1977), editor, Conceptualization and Measurement of Consumer Satisfaction and Dissatisfaction, Boston: Marketing Science Institute.

LaTour, Stephen A. and Peat, Nancy C. (1980), "The Role of Situationally Induced Expectations, Other's Experiences, and Prior Experiences in Determining Consumer Satisfaction," Jerry C. Olsen, editor, Advances in Consumer Research, Vol. 7, Association for Consumer Research.

Leigh, Thomas W., and Day, Ralph L. (1979), "Satisfaction, Dissatisfaction, and Complaint Behavior with Nondurable Products," in Ralph L. Day and H. Keith Hunt, editors, New Dimensions of Consumer Satisfaction and Complaining Behavior, Bloomington: Division of Research, Indiana University, 170-183.

Oliver, Richard L. (1979), "Product Satisfaction as a Function of Prior Expectation and Subsequent Disconfirmation: New Evidence," in Ralph L. Day and H. Keith Hunt, editors, New Dimensions of Consumer Satisfaction and Complaining Behavior, Bloomington: Division of Research, Indiana University, 66-71.

Oliver, Richard L. (1977), "Effect of Expectation and Disconfirmation on Postexposure Product Evaluations: An Alternative Interpretation," Journal of Applied Psychology, 62, 480-86.

Swan, John E. and Coombs, Linda J. (1976), "Product Performance and Consumer Satisfaction: A New Concept," Journal of Marketing, 40, 25-33.

Westbrook, Robert A. and Cote, Jr., Joseph A. (1980), "An Exploratory Study of Affective and Attitudinal Influences on Consumer Satisfaction," Jerry C. Olsen, editor, Advances in Consumer Research, Vol. 7, Association for Consumer Research.

Westbrook, Robert A. and Newman, Joseph C. (1978), "An Analysis of Shopper Dissatisfaction for Major Household Appliances," Journal of Marketing Research, 15, 456-66.

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