Complainers and Noncomplainers Revisited: Another Look At the Data

ABSTRACT - The point of departure for the present paper is previous attempts to profile complaining, dissatisfied consumers. By reviewing the previous literature, three "models" were identified: the "resource", the "learning" and the "personality'' model respectively. An empirical test of the three models reveals only modest differences between complaining and noncomplaining consumers. Marketplace participation was found to be the most meaningful explanatory factor.


Kjell Gronhaug and Gerald Zaltman (1981) ,"Complainers and Noncomplainers Revisited: Another Look At the Data", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 83-87.

Advances in Consumer Research Volume 8, 1981      Pages 83-87


Kjell Gronhaug, Norwegian School of Economics and Business Administration

Gerald Zaltman, University of Pittsburgh


The point of departure for the present paper is previous attempts to profile complaining, dissatisfied consumers. By reviewing the previous literature, three "models" were identified: the "resource", the "learning" and the "personality'' model respectively. An empirical test of the three models reveals only modest differences between complaining and noncomplaining consumers. Marketplace participation was found to be the most meaningful explanatory factor.


Over the past few years, interest in and research on consumer satisfaction and dissatisfaction (CS/D) and complaining behavior have grown rapidly. Symptomatically, a recent review of 73 publications in the field noted that only 12 of them had appeared prior to 1972 (Hunt 1977).

Previous research efforts reveal the underlying, though implicit, assumption that dissatisfied complainers and noncomplainers, respectively, are different. The following quotation clearly indicates this belief: "..., it was hoped that demographic characteristics of such consumers could be found that would distinguish consumers with problems who had not complained from those consumers who had complained" (Kraft 1977, p. 79). Much of the research emphasis has been directed toward profiling the complainers and noncomplainers by stable, sociodemographic characteristics such as age, education, and income. Such characteristics may definitely be useful as resource-related or social indicators (Bauer 1966), but there are definitely reasons why such indicators only exhibit modest descriptive and explanatory power in a consumer behavior context (cf. Sheth 1977).

Previous Research

At the more specific level, the following conclusions may be drawn from previous research:

- Only a fraction of the dissatisfied consumers actually takes overt action and complains. Andreasen and Best (1977) in a telephone survey conducted among 2,400 households found that the consumer (buyer) was dissatisfied in one of five purchases, but in less than 50 percent of these cases did he or she take any action. A recent review of the complaining behavior studies (cf. Robinsen 1978) strongly supports these findings. Although variations in propensity to complain are observed across studies, the overall conclusion to be drawn is that only a fraction of the dissatisfied purchasers are overt complainers.

- As noted above, various sociodemographics have been applied extensively in previous research to profile the complainers and noncomplainers, respectively. The results are, however, in no way conclusive. Liefield, et al. (1975) found that Canadian consumers who had written complaint letters were middle-aged, better educated, earned higher incomes, and more represented by managerial and professional heads of households. Similar findings regarding the use of demanding sources of complaints have been reported by Gr°nhaug (1977), Zaichkowsky and Liefield (1977), Kraft (1977) and Warland et. al. (1975).

With regard to complaining in general (and, in particular, when looking away from complaining through the more demanding sources), the descriptive and predictive power of the various sociodemographics reveal very mixed results. By relating various sociodemographics (age, education, income, place of living, social status, occupation) to the propensity to complain due to a perceived dissatisfaction with a variety of nondurables (milk, meat, coffee, vegetables) and durables (textiles, cars), Gr°nhaug (1977) found the sociodemographics do possess no descriptive or explanatory power. Warland, et al. (1975) found, on the ocher hand, by relating the action due to the "most recent, salient negative experience in the marketplace" (p. 151), that the active complainers in general are younger and above average with respect to social status, income, education, and group membership. Zaltman et al. (1978) found age negatively related to propensity to complain. Granbois et al. (1977) found that sociodemographics overall possess almost no descriptive power. For some products, however, sex was found to possess some descriptive power. Wall, et al. (1977) found in a sample of female buyers of clothing that age was somewhat negatively related to the propensity to complain. Lawther (1978) found that consumers who are less socially integrated were less prone to make overt complaints than were the more integrated consumers,

The apparently inconsistent findings regarding the descriptive and explanatory power of the sociodemographics have led researchers to include other variables in their research. Inclusion of perceptual variables have revealed the almost self-evident finding that the more aware an individual is of defects, the higher the propensity to complain (cf. Kraft 1977, Zaltman et al. 1978). Also, attribution theory in terms of perceived causes to the dissatisfaction has been applied in the context of complaining behavior, revealing very promising and insightful results (Valle and Wallendorf 1977, Krishnan and Valle 1979). Personality characteristics have also been applied in attempt to profile complainers. Wall et al. (1977) found some items related to personality (liking/disliking, ways of perceiving things and self-confidence), to possess some descriptive power among a sample of female buyers related to satisfaction/ dissatisfaction with clothing. In an attempt to describe the personality profiles of consumer complaint letter writers, Zaichkowsky and Liefield (1977) concluded however, "... that consumer complaint letters cannot be distinguished on the basis of personality types" (p. 128). In other words, the attempts to include various personality traits cannot be considered as successful.

The research on consumer satisfaction/dissatisfaction and complaint behavior is still in its infancy. Several objections regarding previous research may be made (cf. +lander 1977, Haines 1979). No attempt will be made to discuss the inherent weaknesses in previous research. However, we would like to address the reader's attention to the lack of explicit arguments why and how the various indicators used should discriminate between complainers and noncomplainers, respectively.

Making the Assumptions Explicit

As noted above, in most of the previous research the assumptions underlying the choice of concepts and models have been more or less implicit. By reinterpreting previous research, the following assumptions (or models) can all be traced.

(1) The resource model: To make overt complaints requires resources. Resources are not equally distributed, and thus the propensity to complain will vary across the members of the society. This type of model (or argument) has determined very much of the consumer and welfare policy. Sociodemographics -- in particular, age, income and education -- have all been treated as individual resource indicators (cf. Atkinson 1975, Bauer 1966, Rein 1976).

(2) The learning model: Learning plays an important role in consumer behavior (cf. Howard 1977). Products and brands are learned; expectations and attitudes are learned; and how to handle purchases and complaints is learned. The basic underlying assumption or model in the CS/D-research is that "knowledge is power", i.e. the experienced, well trained buyer will be better off. Learning in terms of awareness of unfair marketing practices and knowledge about consumer rights has been included in previous research (cf. Kraft 1977, Zaltman et al. 1978, Lawther 1978).

(3) The personality model: Some recent attempts have been made to incorporate personality variables into the research on CS/D (cf. Wall et al. 1977, Kraft 1977). The underlying assumption is that certain personality characteristics, such as high degree of self confidence, is associated with ability to perceive dissatisfaction and handle complaints.

In focusing on complaining behavior, we will -- as in previous research -- consider perceived dissatisfaction as a prerequisite to overt complaints. Based on the previous discussion, variables will be borrowed from the "resource", the "learning", as well as the "personality" model in order to describe and explain complaint behavior.


Source of Data

Data used in this paper are from a large research study which investigated the consumer problems of the elderly. The data were obtained from two structured mail questionnaires which were developed from focus group interviews. The questionnaires were sent to 4,000 consumers between the ages of 25 and 80 who were members of a national panel maintained by Market Facts, Inc. Usable responses covering both waves (spaced by a six-month interval) were 2,849. The response rate to the original 4,000 questionnaires was 71 percent. The response rate for the second wave was 90 percent of the first wave sample.

The Sample

Due to the purpose of the data gathering, the elderly were deliberately oversampled. Compared to the U.S. population, the sample is skewed toward the elderly, which is due to deliberate disproportional sampling due to the focus of the study and the data gathering. Furthermore, the sample shows a slight underrepresentation of high school graduates and a moderate overrepresentation of college graduates. Compared to the U.S. population, the annual income in the sample was somewhat higher. In addition, it should also be noted that compared to the national population, the sample had a lower fraction of individuals whose spouses were no longer living, and fewer males living alone.


Reported below are the measurements used in this study:

Dissatisfaction.  In order to map dissatisfaction, the following measurement was used. The respondents were confronted with a rather ions list of products and services, and the following question was asked: "Please read the products and services listed below and 'X' all of those where you felt bad about your buying experience''. From this operational definition, it follows that no restriction in time is given; in other words, many of the reported "bad buying experiences" may have taken place a long time ago. A summary measure for dissatisfaction across products was worked out by adding up the "bad buying experiences" (range: 0, ..... , 21).

Complaint.  If dissatisfied, the consumer (buyer) may react in several ways. This was mapped in the following way. For the "worst buying experience" a list of 12 different actions, which also included "no action", was given the respondents, and the following question was asked: "Please 'X' below any actions you took because of the problems you had." The various actions listed may definitely put different requirements on the dissatisfied consumer. Actions requiring some sort of overt behavior directed toward the marketer or consumer agency were labeled "activist" complaint strategies. [The "activist" strategy included "complained to the person who sold me the product or service", complained to the company or store", "complained to a consumer agency", "complained to a public agency or my congressman", "complained to a newspaper or magazine", and "consulted or hired a lawyer to protect my interests'.] In this paper the persons applying the "activist" strategy were contrasted with the "passivist", i.e., those who "took no action at all". [Besides the "activist" strategy, strategies related to "word of mouth", "stop buying", and "the passivist" strategy ("no action") were located. By examining the various strategies according to the variables (indicators), the "activist" and "passivist" strategies were really found to represent the extremes.]

Resources. Resources are often thought of in terms of money/economy. However, the individual resources should in no way be restricted to money/economy only. Here a distinction was made between the following types of resources: economy; problem solving capacity; health; time, and social resources. Several indicators were used to map the resource dimensions: economy (income, financial problems); problem solving capacity (education, decision-making skill, planning horizon); time available; health (age, various health problems); social resources (social contacts, perceived problems in getting help rich consumer related problems). [The indicators included the mapping of several "life events" such as the experience of financial crises and health problems.]

Learning .  Learning was measured by mapping buying experience and knowledge, i.e. ability to handle the buying problems.

Personality. Indicators mapping involvement, self-confidence, and perceptions of being treated fairly besides perceived consumer influence were applied in order to map personality.

The measurements applied include self-reports on perceptions and behavior as well as sociodemographics. The dissatisfaction and complaint measures coincide to a substantial degree with measures used in previous research. With regard to the resource measures, the sociodemographics (income, education, and health) have been extensively applied in previous research (cf. Sheth 1977). The "life events" (cf. "financial crisis" and "health problem") have not been applied in previous research, and are included in order to indicate changes. The various indicators related to knowledge, experience, planning horizon, and perceived decision-making skill, have all been widely used in previous consumer research.


Reported below are the major results from this investigation: 30.3 percent of the respondents had not any "bad buying experience". Of those who had at least one such experience, 834 respondents had applied the "activist" strategy, and 432 the "passivist" strategy. These two subsets, the "activists" and the "passivists", will constitute the sample of analysis (n = 1,266).

Table 1 summarizes the results from the bivariate analysis breaking down the various variables by the two strategies applied. The statistical procedure is t-test applied to differences of means; i.e. the differences between the "passivist" and "activist"-scores. The results are surprising in term of lack of associations between complaint strategy and the various indicators. Closer inspection of Table 1 reveals that the active complainers are:

- higher in experience (i.e., buying experience);

- higher in income;

- higher in education; and

- they tend to be younger than do the noncomplainers.

However, as seen from Table 1, none of the indicators related to health (except age), time, social resources or personality revealed any differences between the complainers and noncomplainers.




Several questions might be raised based on the preceding data presentation, such as: Why are there so few differences between the complainers and noncomplainers, and more basically, what do the results really mean? We will address our attention to the last question.

The indicators for which there are discrepancies are not necessarily unrelated. By inspecting the intercorrelation matrix, the following picture emerges:



Here ordinary product-moment correlations have been applied. In doing so the nominal scaled complainer categories are turned into dummy variables (i.e. "activist" -l, "passivist = 0). Experience is here the summary measure of previous buying experiences. Age and education are in fact ordinal scaled variables. However, only modest differences were detected by applying nonparametric measures of association.

The correlation matrix reveals that "activists' are positively related to experience, education, and income, but negatively related to age, just as demonstrated in Table 1. However, it is also evident that the various descriptive variables are intercorrelated. By calculating the various partial correlations and controlling for the subsequent variables, the following emerges:

ract., exp. (controlled for education, income, age = .1130 (p < .001).

ract., age (controlled for education, income, experience) = .0560 (p = .048).

ract., educ. (controlled for age, incomes experience) = .0166 (p = .558).

ract., inc. (controlled for education, age, experience = .0678 (p = .017).

In other words, the "experience" variable by far possesses the greatest descriptive power. Similar results were found by performing multiple discriminant analysis. However, the predictive power of the derived discriminant function was found to be modest as assessed by comparing the number of cases correctly classified with the by chance result, which also follows from the rather lay correlation coefficients as demonstrated in Table 2, and the lack of association between complainers (or noncomplainers) and most of the predictors as shove in Table 1.

When taking a closer look at the applied "experience" measure, it obviously relates to buying activities. Or, maybe better, the measure represents a rough proxy for marketplace participation. And, according to the present authors, this represents the key to the explanation. First, buying is not the sole activity with which man is preoccupied, nor does this activity have to be the most important one.

In fact, the individual member of the society is confronts with various types of constraints including limited time and span of attention. [Attention as a limited capacity is dealt with in several of the social sciences. For an interesting discussion, cf. Berth and Olsen (1976).] This, in turn, directs our attention to one of the almost forgotten elements in a marketing context, namely, that the consumer role may be one of the many roles played by the individual. He or she will furthermore put more or less emphasis on this role. The relative importance of this role for him or her will be closely related to the time devoted and the activities performed in the context of this role. And this role will be played across all social strata!

Why is this role (marketplace participation) related to, or better, associated with income, education, and age (cf. Table 2)? First, economic resources are a prerequest to participate in the marketplace as a buyer. Second, income and education are positively correlated, and thus by controlling for income, the association between marketplace participation and education drops dramatically. One intriguing explanation may be proposed for the negative relationship between age and complaining, which is that this phenomenon is caused by a cohort [The rationale behind such an effect is that when brought up in a certain time period, the individual is "stigmatized,'' and will thus carry with him or her the impact of such previous learning (stigma).] effect. However, by controlling for marketplace participation, the effect of age on noncomplaining is (as demonstrated above) dramatically reduced.

The marketplace participation argument launched above may also shed some light on previous findings. The results reported by Warland et al. (1975) may partly be explained by variation in marketplace participation. The impact of the sociodemographics (which is in the same direction as the findings presented here) may also to some extent be explained in terms of variations in marketplace participation. Similarly, the impact of sex in findings reported by Wall et al. (1977) and Granbois et al. (1977) may be explained in terms of products primarily dealt with by women, and represent thus an artifact introduced by the products studied (cf. Rosenthal and Rosnow 1969). Symptomatically no differences were detected between sexes when applying our aggregate "marketplace participation" measure. But why are our deflections for the "marketplace participation" measure not more pronounced? The explanation is probably due to the very rough nature of this proxy measure. Refined measurements taking such factors as allocation of time, the variety of activities performed, and importance placed on such activities would probably yield much stronger results.

The findings reported and explanations provided do also have consumer policy implications. First, the elderly, the poor and the individual low in education do not necessarily react more passively to perceived dissatisfaction -- as demonstrated above -- than do the individuals higher in resources, which clearly seems to contradict public policy practice. In fact, for groceries the low income consumers were found in several studies to be higher in price knowledge than were the high income consumers (cf. Goldman 1977, Gabor and Granger 1961), which is in concordance with the fact that the lower income consumers necessarily have to be more conscious due to less economic resources. This "apparent" lack of the "expected" negative relationship between being active in complaining and low in income and education and high in age may also be due to the following explanation. The more involved in the marketplace, the higher the probability of being exposed to buying problems (r = .224; p < .001). In addition, the following argument related to social class and consumption style seems to be relevant: Income and education are both well-known indicators of social class. Furthermore, basic values and consumption patterns vary across the various social strata. Expressive values and thus emphasis on expressive consumption is more frequent among the lower classes as compared to the predominance of instrumental values and behavior in the upper social classes (cf. Segal and Felson 1972). Intuitively, expressive consumption is less demanding, the consequences may be easier to foresee, and the purchases are thus less likely to produce dissatisfaction. Thus the "expected" higher fraction of dissatisfied, passive consumers is not showing up in the research. On the other hand, if the consumption aspiration level of the lower social segments is raised, this may lead to higher degree of perceived dissatisfaction.


Andreasen, A.R. and Best, A. (1977), "Consumers Complain -- Does Business Respond?" Harvard Business Review, 55, No. 4, 98-101.

Atkinson, A.B. (1975), The Economics of Inequality, Oxford: Clarendon Press.

Bauer, R.A. (1966), Social Indicators, Cambridge, MA: N.I.T. Press.

Betaken, J.R. (1973), "Perceived Risk and Its Components: A Nodal and Empirical Test", Journal of Marketing Research, 10 (May), 184-190.

Gabor, A. and Granger, C.W.J. (1961), "On the Price Consciousness of Consumers", Applied Statistics, 10, No. 1 (March), 170-178.

Goldman, A., "Consumer Knowledge of Food Prices as an Indicator of Shopping Behavior", Journal of Marketing, 41, (October), 65-75.

Granbois, D., Summers, J. and Frazier, G.L. (1977), "Correlates of Consumer Expectations and Complaining Behavior", in Consumer Satisfaction, Dissatisfaction, and Complaining Behavior, R.L. Day, ed., Bloomington, IN: Indiana University, 18-25.

Gr°nhaug, K. (1977), "Exploring Consumer Complaining Behavior: A Model and Some Empirical Results", in Advances in Consumer Research, Vol. IV, W.D. Perreault, ed., 159-165.

Haines, G. (1979), "Three Papers on Consumer Satisfaction/ Dissatisfaction: A Comment", in Advances in Consumer Research, Vol. VI, W.L. Wilkes, ed., Ann Arbor, MI: Association for Consumer Research, 450-452.

Howard, J.A. (1977), Consumer Behavior: Application of Theory, New York: McGraw-Hill.

Hunt, H.K. (1977), "Consumer Satisfaction and Dissatisfaction: A Developing Methodology", in The New Role of Marketing Professionals, P.J. LaPlace, ed., Chicago: American Marketing Association, 241-244.

Kraft, F.B. (1977), "Characteristics of Consumers as Complainers and Complaint and Repatronage Behavior", in Consumer Satisfaction, Dissatisfaction, and Complaining Behavior, R.L. Day, ed., Bloomington, IN: Indiana University, 79-84.

Krishnan, S. and Valle, V.A. (1979), "Dissatisfaction Attributions and Consumer Complaint Behavior", in Advances in Consumer Research, Vol. VI, W.L. Wilkie, ed., Ann Arbor, MI: Association for Consumer Research, 445-449.

Lawther, K. (1978), "Social Integration of the Elderly Consumer: Unfairness, Complaint Actions, and Information Usage", in Proceedings of the 1978 Educators' Conference, Chicago: American Marketing Association, 341-345.

Liefield, J.P., Edgecomb, F.H.C. and Wolfe, L. (1975), "Demographic Characteristics of Canadian Consumer Complainers", Journal of Consumer Affairs, 9, 73-80.

March, J.G. and Olsen, J.P., eds. (1976), Ambiguity and Choice in Organizations, Bergen' Scandinavian University Books.

+lander, F. (1977), "Consumer Satisfaction -- A Sceptic's View," in Conceptualization and Measurement of Consumer Satisfaction and Dissatisfaction, H.K. Hunt, ed., Cambridge, MA: Marketing Science Institute, 455-488.

Rein, M. (1976), Social Science and Public Policy, Baltimore: Penguin Books, Inc.

Robinson, L.W. (1978), "Consumer Complaint Behavior: A Review with Implications for Further Research", paper presented at the Third Annual Conference on Consumer Satisfaction/Dissatisfaction and Complaining Behavior.

Rosenthal, R. and Rosnow, R.L. (1969), Artifact in Behavioral Research, New York: Academic Press.

Segal, D.R. and Felson M. (1972), "Social Stratification and Family Economic Behavior", in Family Economic Behavior, L.B. Sheldon, ed., Philadelphia: Lippincott Company.

Sheth, J.N. (1977), "Demographics in Consumer Behavior", Journal of Business Research, 5 (March), 129-138.

Valle, V.A. and Wallendorf, M. (1977), "Consumer Attributions and the Cause of Their Product Satisfaction and Dissatisfaction", in Consumer Satisfaction, Dissatisfaction and Complaining Behavior, R.L. Day, ed., Bloomington, IN: Indiana University, 26-30.

Wall, M., Dickey, L.E. and Talarzyk, W.W. (1977), "Predicting and Profiling Consumer Satisfaction and Propensity to Complain", in Consumer Satisfaction, Dissatisfaction and Complaining Behavior, R.L. Day, ed., Bloomington, IN: Indiana University, 91-101.

Warland, R.H., Herman, H.O. and Willits, J. (1975), Dissatisfied Consumers: Who Gets Upset and Who Takes Action", Journal of Consumer Affairs, 9, 148-163.

Zaichkowsky, J. and Liefield, J; (1977), "Personality Profiles of Consumer Complaint Letter Writers", in Consumer Satisfaction Dissatisfaction and Complaining Behavior, R.L. Day, ed., Bloomington, IN: Indiana University, 124-129.

Zaltman, G., Srivastava, R.K. and Deshpande, R. (1978), "Perceptions of Unfair Marketing Practices: Consumerism Implications", in Advances in Consumer Research, H.K. Hunt, ed., Ann Arbor, 141: Association for Consumer Research, 247-253.



Kjell Gronhaug, Norwegian School of Economics and Business Administration
Gerald Zaltman, University of Pittsburgh


NA - Advances in Consumer Research Volume 08 | 1981

Share Proceeding

Featured papers

See More


Assuming Ordinality: Best-to-Worst Inferences in Vertical Lists

Mathew S. Isaac, Seattle University
SHAILENDRA PRATAP JAIN, University of Washington, USA

Read More


F11. Anti-Consumption for Sustainability: The Environmental Impact of Anti-Consumption Lifestyles, Environmentally Concerned Individuals and Ethical Consumers

Laurie Touchette, HEC Montreal, Canada
Marcelo Vinhal Nepomuceno, HEC Montreal, Canada

Read More


A10. Opting Opt-in or Out? Effects of Defaults on Perceived Control and Valuation of Personal Data

Iris van Ooijen, University of Twente

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.