Sie sind auf Seite 1von 11

Measuring Customer Satisfaction: Fact and Artifact

Robert A. Peterson
University of Texas

William R. Wilson
Vyvx, Inc.

Self-reports of customer satisfaction invariably possess distributions that are negatively skewed and exhibit a positivity bias. Examination of the customer satisfaction literature and empirical investigations reveal that measurements of customer satisfaction exhibit tendencies of confounding and methodological contamination and appear to reflect numerous artifacts. Implications and suggestions for research and practice are discussed.

INTRODUCTION
Without question, customer satisfaction is one of the most widely studied and embraced constructs in marketing. Over the past two decades more than 15,000 academic and trade articles have been published on the topic. As the cornerstone of the marketing concept, customer satisfaction has been embraced by practitioners and academics alike as the highest-order goal of a company. Firms espouse their commitment to satisfying customers in their advertisements, public relations releases, and mission statements. One survey (reported in Chief Executive 1989) found that 90 percent of responding firms had customer satisfaction reflected in their mission statements. From a normative or philosophical perspective, it is not possible to argue against the goal of customer satisfaction. For a business to be successful in the long run, it must satisfy customers, albeit at a profit. Indeed, it can be argued that satisfying customers is the primary obligation of acom-

pany. Hence, customer satisfaction is a defensible and appropriate company objective--the glue that holds various corporate functions together and directs corporate resource allocation. Conceptually, virtually all company activities, programs, and policies should be evaluated in terms of their contribution to satisfying customers. Customer satisfaction appears to most typically be measured through surveys (McNeal and Lamb 1979). Surveys provide formal feedback to a firm and send a positive signal to customers that the firm is interested in them. Their popularity derives from their directness, ease of administration and interpretation, clarity of purpose, and face validity. Although unobtrusive, indirect measures of customer satisfaction (e.g., sales, profits, complaints) are also utilized, they are typically viewed as complementary to direct survey measures. Ratings of customer satisfaction derived from surveys are used to evaluate the performance of company employees (including being incorporated in merit and compensation reviews), enhance sales management and training programs, obtain insights into competitors, and, if favorable, are included in company advertising (Bertrand 1989a, 1989b; Smith 1989).

Purpose
To be able to interpret and effectively utilize customer satisfaction ratings, it is necessary to understand what determines them as well as know what variables and/or factors relate to them. In practice, customer satisfaction measurements are designed to tap the underlying global or "net" satisfaction with a product or service. If this is so, an observed distribution of satisfaction ratings reflects "true" satisfaction. However, to the extent observed satisfaction ratings are determined in part by measurement artifacts or the personal characteristics of the surveyed customers, this suggests caution must be employed when interpreting them or using them in decision making.

Journal of the Academy of Marketing Science ISSN: 0092-0703 Volume 20, Number I, pages 61-71. Copyright 9 1992 by Academy of Marketing Science. All rights of reproduction in any form reserved. This article was accepted by the previous editor.

JAMS

61

WINTER, 1992

MEASURING CUSTOMER SATISFACTION:FACT AND ARTIFACT

PETERSON AND WILSON

Stated somewhat differently, are customers (or, for that matter, "consumers" in general) essentially satisfied with the products and services they purchase and consume, or are measurements of customer satisfaction systematically and artificially influenced by variables or factors other than satisfaction? In other words, to what extent do satisfaction self-reports reflect "true" satisfaction? The goal of this article is to provide at least a partial answer to these questions.

A STRIKING CHARACTERISTIC
Studies of customer satisfaction are perhaps best characterized by their lack of definitional and methodological standardization. Even so, despite differences in the subject matter of the satisfaction measurement (durable, nondurable product; for-profit/not-for-profit service; etc.), nature of the measurement scale (single item, multi-item; direct, indirect), population investigated (student, special group), and research design (laboratory experiment, longitudinal design, etc.), studies of customer satisfaction share a common characteristic: Virtually all self-reports of customer satisfaction possess a distribution in which a majority of the responses indicate that customers are satisfied and the distribution itself is negatively skewed. Moreover, the modal response to a satisfaction question is typically the most positive response allowed. For example, in the context of a meta-analysis of satisfaction with mental health treatment, Lebow (1982) reported the median percentage of "very satisfied" consumers (typically the most positive response) to be 49 percent. Figure 1 schematically illustrates the distribution of customer satisfaction selfreports typically obtained in surveys. Although this characteristic is frequently overlooked, its observation is not new to this article. Hunt (1977), LaBarbara and Mazursky (1983), and Oliver (1981) have all noted it in passing, and Westbrook (1980b) even attempted to develop a scale that would produce a more normal distribution. Nor is the characteristic unique to measurements of customer satisfaction. Self-reported measures of "consumer" satisfaction--whether they are domain-specific measures of job satisfaction, marital satisfaction, satisfaction with health, medical, or educational services, community services, or general measures of life satisfaction

or subjective well-being--are uniformly negatively skewed with the majority of the survey participants reporting satisfaction. For example, following an extensive review of the subjective well-being literature, Diener (1984) concluded that satisfaction ratings were nearly always skewed left, with most responses occurring on the positive end of the scale. Headey and Wearing (1988) arrived at a similar conclusion and stated that "to feel 'above average' is normal" (p. 499). Tanner and Stacy (1985), writing in the context of mental health services, went so far as to term satisfaction ratings "too good to be true" (p. 148). Table 1 presents observed distributions of satisfaction measurements that illustrate the schematic distribution in Figure 1. Table 2 contains example percentages of customers reporting satisfaction with a variety of phenomena. Collectively, the data in Tables 1 and 2 provide empirical evidence for the pervasive skewness and positivity of satisfaction ratings. (Additional support in terms of skewness statistics is contained in Michalos (1986) and Westbrook (1980b).) It is this "striking characteristic" that motivated the present research.

Possible Explanations
Skewness and positivity of self-reported satisfaction ratings are intellectually interesting and pragmatically important issues. Oversimplifying a bit, there are four possible explanations for the observed distributions of customer satisfaction measurements. The distributions could simply be reflecting actual satisfaction with products and services purchased and consumed. It is reasonable to expect that, given the nature of the products and services typically studied, individuals are, for the most part, satisfied with what they purchase and consume. What rational consumer would knowingly purchase a product or service that is not expected to satisfy some need or want? Alternatively, if, as typically hypothesized by researchers such as Anderson (1973), Bearden and Teel (1983), Cadotte, Woodruff and Jenkins (1987), Cardozo (1965), and Oliver (1980, 1981), satisfaction is caused by factors such as expectations and requires considerable cognitive effort, its antecedents may influence the shape and level of the observed distributions. Or, satisfaction may simply possess a distribution that is different from the (normal) distribution of other common psychological constructs (Willerman 1979). Finally, the level and shape of a distribution of satisfaction ratings may be artifacts of the research methodologies employed as well as a function of factors such as the inherent, intrapersonal characteristics of the surveyed customers. Although none of the above explanations can be summarily dismissed, the intent of this manuscript is to explore the possibility that the shape and level of customer satisfaction rating distributions are at least partially due to methodological artifacts and/or individual difference variables. Therefore, in the remainder of this article, possible causes and implications of the "striking characteristic" are examined in detail. Initially, various methodological issues are addressed. This is followed by a discussion of selected variables shown to be related to customer satisfaction. The final section contains a series of implications regarding the inter-

FIGURE 1 Conceptual Distribution of Satisfaction Measurements

"High Dissatisfaction"

"High Satisfaction"

JAMS

62

WINTER, 1992

MEASURING CUSTOMER SATISFACTION: FACT AND ARTIFACT

PETERSON AND WILSON

TABLE 1 Example Distributions of Customer Satisfaction Self-Reports


Sample Characteristics

Source
Campbell (1981)

Topic
Marriage

Description
Adults

Size
3,700

Satisfaction Distribution Scale 1 (Completely satisfied) 2 3 4 5 6 7 (Completely dissatisfied) Satisfaction percentage 91-100 81-90 71-80 61-70 51-60 Scale Very satisfied Somewhat satisfied Somewhat dissatisfied Very dissatisfied Uncertain Scale Very satisfied Somewhat satisfied Unsatisfied Very unsatisfied Scale Completely satisfied Very satisfied Fairly satisfied Somewhat dissatisfied Very dissatisfied Scale Very satisfied Quite satisfied Not very satisfied Not at all satisfied Scale Very satisfied Somewhat satisfied Completely unsatisfied Percent 54 27 9 5 2 2 1 Percent 12 38 31 15 4 Percent 76 18 3 2 1 Percent 64 28 6 2 Percent 33 44 18 4 1 Percent 70 20 7 3 Percent 60 36 4

Lebow (1982)

Mental health centers

Users

26 a

Proprietary

Brand X automated payroll system

Small business owners

500+

Proprietary

Brand Y personal computers

Owners

4,500+

Proprietary

Automobiles

Owners

38,000+

SIP Servizio Opinioni (1989)

Cellular telephone service

Users

6,553

Weinstein (1989)

Banks

Adult consumers

1,000+

aLebow performed a meta-analysis of 26 satisfaction studies. The data read: 12 percent of the studies reported a user satisfaction percentage between 91 and 100 percent.

pretation and measurement of the construct of customer satisfaction.

METHODOLOGICAL ISSUES
Intuitively, Figure 1 suggests that the typical shape of a distribution of customer satisfaction measurements is simply due to a ceiling effect--the scales used to measure customer satisfaction do not have a sufficient number of categories to permit survey participants to make fine discriminations, especially at the positive (highest) end. Examination of the data in Table 1 suggests the number of scale categories used and the manner in which they are labelled

may well contribute to both the level and the shape of reported customer satisfaction distributions. Even so, the pervasiveness of the skewness and positivity of customer satisfaction ratings suggests that the phenomenon is more complex than can be accounted for by only a ceiling effect. For example, even when Westbrook (1980b) used, in effect, an 11-point rating scale to measure satisfaction, the resulting distributions were substantially skewed for five of the seven product categories investigated. Moreover, multi-item satisfaction scales exhibit the same skewness and positivity bias that single item scales do, even though their ceilings are technically much higher. To illustrate, Ware and Davies (1983) reported a mean of 13.0 for the Patient Satisfaction Questionnaire (PSQ) General Satis-

JAMS

63

WINTER, 1992

MEASURING CUSTOMER SATISFACTION:FACT AND ARTIFACT

PETERSON AND WILSON

TABLE 2 Illustrative Levels of Expressed Satisfaction Source


Blume (1988) Diblase (1988) Fitzgerald (1990) Floyd (1975)

Topic~Sample

Percentage Satisfied
85 92 84 82 84~ 82h 75-80 70-90 83c

British Airwayscustomers HMO enrollees Sears' customers Children's instructional programs/parents Hall and Doman (1988) Medicalcare metaanalysis Hughes (1977) Clothing and white goods/adults Levois, Nguyen and Mental health service Attkisson (1981) clients Ware, Davies-Averyand Healthcare services metaStewart (1978) analysis Westbrook and Cote (1979) Shoes/students

aMedian value from 68 studies; range = 43-99 /'Average percentage of "completely satisfied" responses from 32 product categories; range = 70-94. cSatisfaction measured on a 0% ("not at all satisfied") to 100% ("completely satisfied") scale.

faction Scale in their study of customers of prepaid health plans. Since this scale consists of four items, each scored 15, the possible range was 4 - 2 0 . Thus the mean is 65 percent of the upper scale bound, suggesting an upward bias in response. Sabourin, Laferriere, Sicuro, Coallier, Cournoyer, and Gendreau (1989) provided even more dramatic evidence. They reported a mean of 28.2 on the Client Satisfaction Questionnaire (CSQ), an 8-item scale whose possible range is 8-32. Hence the mean was 88 percent of the possible score, again supporting a positivity bias and, in this instance, skewness. In brief, then, a ceiling effect explanation, while intuitively appealing, does not appear to be sufficient to account for the skewness and positivity of satisfaction ratings. Although no doubt a contributing factor, it is necessary to look elsewhere when attempting to identify possible determinants of the distribution shape and level. Other methodologically related explanations include response rate bias, data collection mode bias, question form, question context, measurement timing, and response styles. Each of these possible explanations is discussed below.

isfaction percentages with response percentages for 34 studies. The resulting correlation coefficient (r = .05) was not statistically significant. We also conducted a content analysis of the customer satisfaction literature in marketing. From this analysis we identified 15 studies that contained satisfaction percentages and response rates. ~ The correlation between these two variables was .06 (not statistically significant). Although these two analyses possess limitations, including substantial uncontrolled variation (e.g., data from telephone, personal, and mail interviews were combined), the results are consistent: satisfaction percentages are not related to response rate percentages. Thus, while the response rate hypothesis might initially appear viable, on further reflection there is no logical reason why satisfied customers should be more likely than dissatisfied customers to respond to a satisfaction survey. Indeed, the opposite argument could be made. Dissatisfied customers may be more likely to respond, and respond negatively, to a satisfaction survey because dissatisfaction itself is inherently more action-oriented and emotionally intense than satisfaction (e.g., Richins 1983). Alternatively, it could be hypothesized that both highly satisfied and highly dissatisfied customers are more likely to respond to a satisfaction survey than customers who are neither highly satisfied nor dissatisfied. (This would result in a bimodal distribution of satisfaction responses.) These latter two hypotheses, though, run counter to the observed distribution in Figure 1 and the data in Tables 1 and 2.

Data Collection Mode Bias


Data collection mode bias refers to the influence that the mode of data collection has on satisfaction measurements. Thus, different results might obtain as a function of whether data are collected by personal, telephone, or mail interviews. In one sense mode bias reflects the intrusiveness of the interviewing process, per se, with personal interviews being the most intrusive and mail interviews the least. Obviously it is not possible to investigate mode bias in an absolute sense. However, there are indications that higher levels of satisfaction are obtained when personal or telephone interviews are used than when mail or selfadministered interviews are used. Although he did not directly study satisfaction, Sudman (1967, p. 57) found that the percentage of survey participants stating they were happy in a personal interview (36 percent) was higher than in a self-administered questionnaire (23 percent). LeVois, Nguyen, and Attkisson (1981) observed a difference of ten percent in satisfaction ratings between a questionnaire administered orally versus being self-administered. Based on a repeated-measures design and a large, nationally representative sample in excess of 5,000 new car buyers, we have found that telephone interviews consistently result in satisfaction percentages for automobiles approximately 12 percent higher than those obtained through mail interviews. Collectively, the literature and our results suggest mode bias needs to be considered when interpreting customer satisfaction ratings. Satisfaction data collected using different modes are not comparable; on average, oral administration of satisfaction questions (personal or telephone interviews) appears to increase satisfaction ratings by approximately

Response Rate Bias


It has been hypothesized (e.g., Lebow 1982) that satisfaction ratings are inflated due to response rate biases. According to this hypothesis, individuals who are more satisfied are more likely to respond to a satisfaction survey than are individuals who are less satisfied. If this hypothesis is correct, satisfaction measurements would tend to be biased upward to reflect the fact that dissatisfied customers are not as likely as satisfied customers to respond to a survey. To test this hypothesis, we correlated the percentages of survey participants reporting they were satisfied with the response rate percentages for studies in two different disciplines. First, utilizing data obtained from Lebow (1983) on satisfaction with mental health treatment, we correlated sat-

JAMS

64

WINTER, 1992

MEASURING CUSTOMERSATISFACTION:FACTAND ARTIFACT

PETERSON AND WILSON

10-12 percent relative to self-administration. Even so, while the satisfaction level differs significantly across data collection modes, negatively skewed distributions of satisfaction ratings are consistently observed. Hence, data collection mode appears to influence the level of reported satisfaction but not the distribution shape of satisfaction ratings.
Question Form

The manner in which questions are asked about satisfaction may influence both the response level and distribution. Most researchers measure satisfaction using a positively presented scale (e.g., "How satisfied are you with ~_?") anchored with "very satisfied"/"completely satisfied" and "very dissatisfied"/"completely dissatisfied." There is a literature, primarily attributable to Tversky and Kahneman (1981; Kahneman and Tversky 1979, 1984), which suggests that when a decision is framed in positive terms, it will be perceived more positively than when the same decision is framed in negative terms. Tversky and Kahneman have repeatedly demonstrated that alternative decision frames produce systematic changes in responses and choice and even reversals in judgments, despite the fact that the frames are equivalent. There is additional literature on information processing (e.g., Levin 1987; Levin, Johnson, and Davis 1987) suggesting that the label given to a stimulus often elicits associations through an activation process. These associations in turn affect the resulting judgment. Consequently, framing the satisfaction question in positive terms is likely to lead to more favorable associations than framing it in negative terms. These associations, in turn, increase the likelihood that survey participants will answer in a positive manner, thereby indicating more sarisfaction than if the question were framed negatively. To investigate this notion, we conducted two studies relating to the manner in which a satisfaction question is posed. In the first study, we asked a closed-end question about satisfaction with a household's primary vehicle in a telephone interview. An identical question was asked in each of two versions of the questionnaire. The only difference was that in one version participants were asked how satisfied they were, whereas in the other they were asked how dissatisfied they were; response categories were identical in the two question versions. The sample consisted of randomly selected middle-class automobile owners in two large metropolitan areas. Demographically the two groups were equivalent. The results of asking each version of the question are presented below.

As the results show, the level of satisfaction elicited from asking the question in a positive manner was significantly greater than when the question was asked in a negative manner ( = 9.27, p < .05). Ninety-one percent of the surveyed individuals reported they were either very or somewhat satisfied with their vehicle when asked the "satisfaction" question. The corresponding percentage for those asked the "dissatisfied" question was 82 percent. Moreover, although both distributions were negatively skewed, consistent with Figure 1, the response distribution for the "dissatisfied question" was a bit less skewed than that for the "satisfied question." A second study was conducted by means of a mail survey of a national consumer panel. Again a split ballot approach was used and, as before, the two groups investigated were equivalent demographically. Although the context of the study differed from most customer satisfaction studies ("satisfaction" in this study is probably best interpreted as "expected satisfaction"), the results still bear on the issue at hand. In particular, survey participants were requested to Imagine that you are trying to decide on whether to go to a particular restaurant. Suppose, however, that there is a 50 percent chance that you will be [satisfied/dissatisfied] with the restaurant. After reading the statement, survey participants were queried as to whether they would pay $30 to eat at the restaurant. From a rational choice perspective, the two satisfaction conditions are identical. A 50 percent chance of being satisfied is the same as a 50 percent chance of being dissatisfied. Hence, the response distributions should be identical for the two groups. However, based on the previously cited research of Tversky and Kahnemann and Levin, it was predicted that relatively more survey participants would decide to eat at the restaurant in the satisfied condition than in the dissatisfied condition. The results support the prediction. Twenty-six percent of the individuals asked the satisfied question (N = 240) stated they would eat at the restaurant, whereas t 1 percent of those asked the dissatisfied question (N = 215) said they would eat at the restaurant. This difference is statistically significant (t = 4 . 1 , p < .01). Taken together, the two studies reiterate the need for caution when interpreting answers to satisfaction (dissatisfaction) questions. Posing a satisfaction question in a positive form appears to lead to greater reported satisfaction than posing it in a negative form and may influence the shape of the underlying response distribution.

Response Category
Very satisfied Somewhat satisfied Somewhat dissatisfied Very dissatisfied N

Percentage Response Question Form "Satisfied" "Dissatisfied"


57.4 33.6 5.0 4.0 243 53.4 28.7 8.5 9.4 240

.Question Context

There is a considerable literature on question-context effects, whether questions asked (and answered) earlier in a questionnaire influence the answers to subsequent questions (e.g., Schuman and Presser 1981, chapter 2). However, apart from McClendon and O'Brien (1988), who found a question-order effect in the context of neighborhood satisfaction, Smith (1979), who found question-context effects

JAMS

65

WINTER, 1992

MEASURINGCUSTOMERSATISFACTION:FACTAND ARTIFACT

PETERSON AND WILSON

in the realm of subjective well-being, and Strack, Martin, and Schwarz (1988), who found that the question-order effect was mediated by conversational context in a study of life satisfaction, little research has been conducted on question-context effects with regard to the measurement of satisfaction. We undertook an exploratory investigation of the question-context effect by means of a telephone survey. As before, data were collected from random (demographically equivalent) samples of middle-class adults by means of a split-ballot design in two metropolitan areas. In one version of the questionnaire, survey participants were queried about their satisfaction with their household's primary vehicle. In the other version, survey participants were first asked about their satisfaction with all the products and items in their household, then asked the vehicle satisfaction question. Results from the study are given below. They indicate that question order had a significant effect on responses (X2 = 9.04, p < .05). Although the results are not conclusive, they suggest that asking a general satisfaction question prior to a specific vehicle satisfaction question slightly increases the tendency for a "very satisfied" response to the vehicle question. The response percentage for the "very satisfied" category was marginally greater in the two-question condition than in the single-question condition. Even so, as before, both response distributions were negatively skewed.

Response Category
Very satisfied Somewhat satisfied Somewhat dissatisfied Very dissatisfied N

Percentage Response for Vehicle Satisfaction Question Vehicle satisfaction Two question question only format
56 34 5 5 140 71 20 7 2 102

Measurement Timing
Although the evidence is somewhat equivocal, there does appear to be a relationship between the level of customer satisfaction obtained in a study and the timing of the measurement. In general, customer satisfaction appears to be highest immediately subsequent to purchase but to decrease somewhat over time. McMahon and Forehand (1983), in their study of satisfaction with mental health treatment, found that even though satisfaction ratings remained quite high on an "absolute" level, there was an erosion in satisfaction from immediate post-treatment evaluations to later follow-ups. Fisk et al. (1990) were more specific. In their study of medical patient satisfaction, they reported an eight percent degradation in satisfaction ratings obtained several months after a hospital stay relative to those obtained at the conclusion of the hospital stay. Corroborating, albeit indirect, evidence is provided by Okun, Olding, and Cohn

(1990). They conducted a meta-analysis of the impact of interventions on global (life) satisfaction of the elderly and concluded that while personal interactions have an immediate effect on satisfaction, the effect essentially dissipates within one month. We have investigated automobile satisfaction measurements taken at two points in time, 30 days subsequent to the purchase and 90 days subsequent to the purchase. Based on several thousand observations obtained from a nationally representative sample of new car buyers and a repeatedmeasures methodology, we found there was about a 20 percent decline in satisfaction ratings over the 60-day time period. Even so, skewness and positivity were evident as usual in both sets of ratings. From the limited information available, it appears that satisfaction measurements are influenced by timing. Indeed, Engledow (1977, p. 91) goes so far as to claim that "satisfaction is a time dependent variable." It is not clear whether initial satisfaction ratings are inherently "inflated" or subsequent ratings "deflated." The apparent deterioration in satisfaction ratings over time may simply be a type of regression-toward-the-mean effect or adaptation operating since initial reported satisfaction is relatively high. Or, it may be that different aspects of satisfaction are being measured at different points in time. Satisfaction measured immediately subsequent to purchase may be largely dependent on cognition or introspection, or even reflect decision regret or an attempt to reduce cognitive dissonance (e.g., Hunt 1977), whereas satisfaction measured further from the purchase may be more affect-based. Oliver (1981), for example, concluded that satisfaction decays into attitude over time, implying that different variables are being measured at different points in time. Fisk et al. (1990) speculated that satisfaction ratings decay over time as new information about a product or service is obtained and consumers reflect back on their experiences. Hall and Dornan (1988) hypothesized that the longer the temporal (measurement) framework, the lower the satisfaction ratings will be due to the manner in which information is processed. They posited that negative experiences are more memorable (more available in memory) and consequently have, over a period of time, more weight in satisfaction evaluations. Since extending the time between purchase and measurement allows for the occurrence of more events and experiences (both positive and negative), doing so may increase the probability of lower satisfaction ratings. More research is obviously required on the influence of measurement timing. Such research would, of necessity, be longitudinal in nature and provide insights into the most appropriate times to measure satisfaction. Oliver (1981) believed that "better" measures of satisfaction are obtained immediately after purchase whereas Diamantopoulos, Tyrian, and Hughes (1988) stated that satisfaction should not be measured until customers have an adequate period of time to experience the purchase. Wilton and Nicosia (1986) concluded that satisfaction measurements taken immediately after a purchase were "more likely to overestimate the significance of satisfaction in the consumption process, although such measures are probably also likely to yield the highest construct validity" (pp. 9-10).

JAMS

66

WINTER, 1992

MEASURINGCUSTOMERSATISFACTION:FACTAND ARTIFACT

PETERSON AND WILSON

Response Styles
Several researchers (e.g., E1-Guebaly et al. 1983; Levois, Nguyen, and Attkisson 1981) have posited that customer satisfaction measurements are influenced by such personal characteristics (response styles) as social desirability and acquiescence. In particular, Sabourin and her colleagues (Sabourin, Bourgeois, Gendreau, and Morval 1989; Sabourin, Laferriere, Sicuro, Coallier, Cournoyer, and Gendreau 1989) have argued that consumer satisfaction ratings of mental health treatment are confounded by social desirability responding (as measured by self-deception and impression management scales). Similarly, Ware (1978) found acquiescence to account for a quarter of the variance in satisfaction with medical care. Although these findings are intriguing, widespread support for them is lacking, principally because they have been derived from atypical samples and contexts. When more general samples are considered, especially with regard to life satisfaction, social desirability responding does not appear to be a significant influence. Larsen, Diener, and Emmons (1985) found only minimal relationships between the Marlowe-Crowne social desirability scale and 11 measures of subjective well-being. And, following an extensive literature review, Diener (1984) concluded that social desirability did not significantly influence measures of subjective well-being. No general evidence is available regarding acquiescence. Simultaneously, though, Tanner and Stacy (1985, p. 148) argued for the existence of an elevator variable, a "tendency to suspend critical judgment and answer in an overwhelmingly positive or favorable direction" when making satisfaction evaluations. They characterized this variable as a type of response style that is logically distinct from social desirability and acquiescence. Analogous to the influences of social desirability and acquiescence found in other contexts, those found in satisfaction measurement require further research and scrutiny. Such research must be carefully designed and executed to avoid the types of contamination frequently observed in this genre of research.

tomer satisfaction, we conducted a telephone survey of 186 randomly selected adults in a major southern city. Survey participants were asked a series of questions of which several related to subjective well-being, satisfaction with a vehicle, and mood. Subjective well-being was measured by means of Andrews and Withey's (1976) 7-point delightedterrible life satisfaction scale ("How do you feel about your life as a whole?"). Vehicle satisfaction was measured using a 4-point scale anchored by "very satisfied"--"very dissatisfied." Mood was measured using the Peterson and Sauber (1983) mood scale. This scale has been frequently used in studies of mood and/or affect in consumer behavior research (e.g., Hill and Ward 1989). Correlations between the three variables are shown below.
Correlation Coefficient Variable VS LS

Vehicle satisfaction (VS) Life satisfaction (LS) Mood (M)

.27 .19

.33

Mood
Finally, there may be a "lurking" variable--mood--that influences satisfaction ratings. Schwarz and Strack (as cited in McClendon and O'Brien (1988)) hypothesized that survey participants often have not thought about a study topic until they are asked a specific question. If this is true, the answers to several types of questions may well be related to the survey participant's mood at the time the question is asked and answered. In particular, Lloyd, Cate, and Henton (1984) considered satisfaction to be a momentary assessment of feelings about a particular phenomenon. If so, then mood may influence measurements of satisfaction. This suggestion is supported by Diener's (1984) review of the subjective well-being literature and empirical studies conducted by Schwarz and his colleagues (Schwarz and Clore 1983; Strack, Schwarz, and Gschneidinger 1985) and Isen and her colleagues (Isen, Clark, Shalker, and Karp 1978). To investigate the relationship between mood and cus:

The correlations reveal statistically significant relationships (p < .01) respectively between life satisfaction and mood, life satisfaction and vehicle satisfaction, and vehicle satisfaction and mood in the predicted directions. Three aspects of the results are worth noting. First, given the nature of data collection, the results are most likely subject to common methods variance since all three variables were measured in the same questionnaire. Second, the correlation between mood and vehicle satisfaction is slightly less than that between mood and life satisfaction, as would be expected from the literature. Third, although these findings are tentative, they are intuitively logical and suggest the need for further research on the role of mood as a nuisance variable in satisfaction measurement. Interestingly enough, independent studies have obtained similar mood correlations to those reported here. Westbrook (1980a), using a student sample, obtained an identical correlation between mood and life satisfaction (.33), but did not obtain a significant relationship between either mood and automobile satisfaction or mood and footwear satisfaction. Larsen, Diener, and Emmons (1985), also using a student sample, obtained a correlation of - . 3 3 between the reactivity subscale of Underwood and Froming's (1980) Mood Survey and Diener et al's (1985) Satisfaction with Life Scale.

CORRELATES OF SATISFACTION
The previous section reveals that measurements of customer satisfaction are influenced by a variety of methodological factors. This section presents several correlates of satisfaction ratings observed in a wide variety of situations. The purpose in doing so is not to present a comprehensive treatment of variables relating to customer satisfaction ratings, but simply to complement the previous

JAMS

67

WINTER, 1992

MEASURINGCUSTOMERSATISFACTION:FACTAND ARTIFACT

PETERSON AND WILSON

discussion of methodological considerations and better place the construct of customer satisfaction in perspective. Although not subject to extensive study, there are indications that several variables influence customer satisfaction ratings in addition to the stimulus object (product or service) and commonly studied antecedents such as expectations. For example, several researchers have reported significant relationships between general levels of (global, life) satisfaction and customer satisfaction. Andreasen (1984) obtained a significant correlation between stress caused by negative changes in life status and customer satisfaction. LeVois, Nguyen, and Attkisson (1981) calculated a correlation of .45 between a measure of life satisfaction and a measure of satisfaction with mental health treatment. Westbrook (1980a) found a significant relationship between overall life satisfaction and satisfaction with an automobile (but not between life satisfaction and footwear satisfaction). Westbrook and Newman (1978) reported a significant relationship between satisfaction and perceived personal competence, and Fomell and Robinson (1983) successfully related organizational variables to customer satisfaction ratings. Similarly, Anderson, Engledow, and Becker (1979) reported a relationship between attitude toward business and customer satisfaction, and Prakash and Munson (1985) found a significant relationship between personal values and satisfaction (mediated by two levels of expectations). Furthermore, preliminary research suggests that there may be a significant relationship between the number of shopping or choice alternatives (brands, stores, etc.) available to consumers and customer satisfaction, although the relationship may be very complex. Czepiel, Rosenberg, and Akerele (1974) reported that satisfaction is probably a function of competitive set, with the number of choices and satisfaction being inversely related. However, on the basis of a laboratory experiment, Reibstein, Youngblood, and Fromkin (1975) concluded that perceived decision freedom increased as a function of the subjects' choice set size, but that satisfaction was unrelated to the size of the choice set. Our research on automobiles revealed that consumers who shop several dealerships report they are less satisfied after purchasing a vehicle than are consumers shopping only a single dealer. (This finding--which has implications for brand loyalty research--is corroborated by Swan (1969) in a different context.) Finally, Jacoby and his colleagues (Jacoby, Speller, and Kohn 1974; Jacoby, Speller, and Berning 1974) found that the more (brand) information consumers had when making a decision, the more satisfied they said they were with the decision. Conclusions regarding the relationship between demographic characteristics and customer satisfaction have been inconsistent (see Francken and van Raaij (1985) for a brief review). This is, in part, due to the wide variety of products and services studied as well as the actual lack of relationships between demographic characteristics and customer satisfaction (Westbrook and Newman 1978). Intuitively (and theoretically), there is little basis to argue that relationships should exist for most products and services. Even so, one relationship merits further examination. With regard to automobiles, there appears to be a consistent relationship between age and customer satisfaction. Our research shows

that, regardless of nameplate, reported satisfaction of automobile owners over the age of 55 is 32 percent higher than that of owners under the age of 35. (Why this is so remains to be explained.) While the findings in the section require corroboration (and resurrect the state-trait debate), collectively they illustrate the diverse nature of influences impacting customer satisfaction and question its ubiquitousness. Operationally, customer satisfaction is a complex and elusive phenomenon.

DISCUSSION
Despite the plethora of variables, perspectives, and methodologies investigated, it is not possible to succinctly identify specific causes of the striking characteristic of customer satisfaction ratings that motivated the present research. The phenomenon is too complex for a simple answer; it is probably due to all of the factors set forth at the beginning of this article. Even so, the primary conclusion of this article is that measurements of customer satisfaction are very contextdependent. To a large extent the level (especially) and shape (partially) of customer satisfaction rating distributions are functions of the research methodology employed, apart from the product or service being studied. As a consequence, the measurement of customer satisfaction is less than precise, possibly confounded with other variables, subject to considerable methodological contamination, and likely reflective of the influence of numerous factors other than those explicitly incorporated into measuring devices. Without reiterating the specific findings reported above, it would appear that satisfaction ratings observed to date likely represent more than satisfaction with the product or service being studied. While this statement is also probably true for other psychological variables, the pervasiveness and consistency of the findings relating to customer satisfaction reported here suggest that satisfaction is unique among individual difference constructs. Simply increasing the number of scale categories or refining the category labels used to measure satisfaction is probably not sufficient to overcome the deficiencies of current scales. Andrews and Withey (1976) tried to develop a scale to compensate for the observed skewness and ceiling effects obtained when measuring subjective well-being. Their scale, the delighted-terrible (DT) scale, was modified and applied to measuring customer satisfaction by Westbrook (1980b) and has been used in studies of hospital patient satisfaction by Swan et al. (1985). Even though Westbrook reported that his DT scale appeared to perform better than a percentage-based scale, it still produced skewed distributions with very high satisfaction reports. Although increasing the number of categories may increase scale reliability and sensitivity, doing so may not improve either validity or utility. Parducci (1965) and WedeI1 and Parducci (1988) have demonstrated that changing the number of scale categories changes the context of the scaling task and sometimes produces unexpected consequences, whereas Schwarz et al. (1985) have shown the general influence of category labels on scale responses.

JAMS

68

WINTER, 1992

MEASURING CUSTOMERSATISFACTION:FACT AND ARTIFACT

PETERSON AND WILSON

Implications and Recommendations


The most straightforward implication that arises as a consequence of the typical distribution of satisfaction ratings is technical in nature and relates to both the analysis and the interpretation of customer satisfaction reports. Given a skewed distribution (of any variable), the arithmetic mean is no longer an appropriate measure of central tendency since it excludes considerable information about satisfaction. Indeed, in this instance, "average satisfaction" is a meaningless concept. Further, because of the skewness (and attendant range restrictions), correlations between customer satisfaction measurements and other variables are likely to be attenuated. Succinctly stated, the more skewed the distribution, the more tenuous is any correlation between customer satisfaction ratings and other variables, and hence the more likely is any "true" relationship to be underestimated. Skewness also causes the outcomes of statistical tests, such as t-tests of means, to be of questionable validity. Tests involving measures of central tendency either need to be nonparametric (distribution-free) or, if parametric, preceded by appropriate data transformations. The most informative statistical tests may be comparisons of distributions rather than of central tendencies. The shape and level of a typical distribution of customer satisfaction ratings have two other, somewhat more subtle, implications. First, across-the-board attempts to improve customer satisfaction subsequent to a purchase are not likely to produce measurable increases in customer satisfaction ratings because of their initially high level. Therefore, because of the erosion phenomenon, managerial emphasis should be on maintaining the initial level of measured satisfaction. Alternatively, it would appear to be more fruitful to examine the tail of a satisfaction-rating distribution--those customers indicating that they are dissatisfied--than to focus on satisfied customers. By limiting the role of a global measure of customer satisfaction to identifying dissatisfied customers, and then acting appropriately, greater benefits are likely to accrue to a firm than simply constructing a "report card" of customer satisfaction. Second, evaluative conclusions such as "customer satisfaction is high [emphasis added] since 65 percent of all customers report being satisfied" (Weinstein 1989) or that a "vast majority [emphasis added] of workers (89 percent) are satisfied" (Kohut and DeStefano 1989) need to be reconsidered in light of the present research. Given the data in Tables 1 and 2, which can be thought of as baseline satisfaction data, such statements are potentially misleading. The data imply, in a sense, that self-reports of customer satisfaction can only be interpreted and evaluated relative to a standard or norm. In particular, customer satisfaction ratings should not be viewed as absolute measurements. For example, customer satisfaction with a particular brand at a specific point in time should be evaluated relative to satisfaction with the brand at a prior point in time or relative to satisfaction with other company brands as well as with competing brands and perhaps even with products or services that are not competing directly (via a process of benchmarking). Ultimately, actuarial-type tables of satisfaction ratings (i.e., norms) should be constructed for a variety of prod-

ucts, services, segments, and research methodologies. Such tables would allow the meaningful calibration of satisfaction measurements, facilitate cross-product and crosssegment comparisons (cf. Maddox 1982), and enhance their utility in decision making. In addition to pointing out the effects of certain methodological factors on customer satisfaction measurements, this research reported statistically significant relationships between satisfaction ratings and a host of other variables. These variables have multiple implications for the interpretation and application of customer satisfaction measurements. Consider mood. Mood is obviously a nuisance variable to the extent it confounds, contaminates, or decreases the reliability of customer satisfaction measurements. Minimally, attempts should be made to explicitly control for variables like mood, either through the research designs employed or post-hoc statistical analyses (e.g., analysis of covariance). Until this is done, customer satisfaction ratings will be of limited practical utility.

CONCLUSIONS
Satisfying customers is a fundamentally sound principle. Unfortunately, measurements of customer satisfaction are not especially informative or diagnostic, principally because of their striking distributional characteristic. Indeed, examination of the satisfaction literature inevitably culminates in a pervasive yet inescapable conclusion--it is not clear what customer satisfaction ratings are measuring. Customer satisfaction ratings may well reflect the Hawthorne effect: Attempts to measure customer satisfaction will, in and of themselves, serendipitously increase satisfaction, regardless of the product or service being investigated. This article demonstrated that customer satisfaction ratings are influenced by a variety of methodological factors. Paradoxically, self-reports of customer satisfaction appear to be simultaneously inflated yet constrained by methodological artifacts. Unless viable unobtrusive measuring devices become available, it is probably not possible to measure "true satisfaction." "True satisfaction" is probably so intertwined with both intrapersonal characteristics and methodological considerations that it may never be possible to disentangle them. Observed satisfaction ratings may always overstate satisfaction because "individuals may need to distort upwardly their evaluations of outcomes that reflect on the wisdom of their own behavior or judgment" (Hall and Dornan 1988, p. 643) or reflect attempts to minimize decision regret. This, though, does not mean attempts to directly measure customer satisfaction should be discontinued. On the contrary, it means that more effort is required to improve the measurement of customer satisfaction.

ACKNOWLEDGMENT
This research was supported in part by the IC 2 Institute, The University of Texas at Austin. The opinions expressed in this article are those of the authors and do not necessarily reflect the views of the Institute.

JAMS

69

WINTER, 1992

MEASURING CUSTOMER SATISFACTION: FACT AND ARTIFACT

PETERSON AND WILSON

NOTES
The particular studies included in the analysis are available from the authors. The number of studies is relatively small because few reports of satisfaction studies contain data on both variables and many marketing studies report mean satisfaction levels and/or do not report response rates, or are based on experiments.

REFERENCES
Anderson, Rolph E. 1973. "Consumer Dissatisfaction: The Effect of Disconfirmed Expectancy on Perceived Product Performance." Journal of Marketing Research 10 (February): 38-44. Anderson, Ronald D., Jack L. Engledow, and Helmut Becker. 1979. "Evaluating the Relationships Among Attitude Toward Business, Product Satisfaction, Experience, and Search Effort." Journal of Marketing Research 16 (August): 394-400. Andreasen, Alan R. 1984. "Life Status Changes and Changes in Consumer Preferences and Satisfaction." Journal of Consumer Research 11 (December): 784-794. Andrews, Frank M. and Stephen B. Withey. 1976. Social Indicators of Well-Being. New York: Plenum Press. Bearden, William O. and Jesse E. Teel, 1983. "Selected Determinants of Consumer Satisfaction and Complaint Reports." Journal of Marketing Research 20 (February): 21-28. Bertrand, Kate. 1989a. "With Customers, the Closer the Better." Business Marketing 74 (July): 68-69. . 1989b. "Putting Customers First." Business Marketing 74 (December): 30-36. Cadotte, Ernest R., Robert B. Woodruff, and Roger L. Jenkins. 1987. "Expectations and Norms in Models of Consumer Satisfaction." Journal of Marketing Research 24 (August): 305-314. Campbell, Angus. 1981. The Sense of Well-Being in America. New York: McGraw-Hill. Cardozo, Richard N. 1965. "'An Experimental Study of Consumer Effort, Expectation, and Satisfaction." Journal of Marketing Research 2 {August): 244-249. Czepiel, John A., Larry J. Rosenberg, and Adebayo Akerele. 1974. "Perspectives on Consumer Satisfaction." In AMA Combined Proceedings, pp. 119-123. Ed. Ronald Curham. Chicago: American Marketing Association. Diamaotopoulos, A., A. C. Tynan and R. Hughes. 1988. "Determinants of Customer Satisfaction Among Luxury Car Buyers: An Empirical Analysis." Irish Marketing Review 3 (March): 149-162. Diblase, Donna. 1988. "Most HMO Enrollees Satisfied: Survey." Business Insurance 22 (November 7): 57. Diener, Ed. 1984. "Subjective Well-Being." Psychological Bulletin 95 (May): 542-585. Diener, Ed, Robert A. Emmons, Randy J. Larsen, and Sharon Griffin. 1985. "The Satisfaction with Life Scale." Journal of Personality Assessment 49 (February): 71-75. E1-Guebaly, N., J. Toews, A. Leckie and D. Harper. 1983. "On Evaluating Patient Satisfaction: Methodological Issues/' Canadian Journal of Psychiatry 28 (February): 24-29. Engledow, Jack L. 1977. "Was Consumer Satisfaction a Pig in a Poke?" Business Horizons 20 (April): 87-94. Fisk, Trevor A., Carmhiel J. Brown, Kathleen Cannizzaro, and Barbara Naftal. 1990. "Creating Patient Satisfaction and Loyalty." Journal of Health Care Marketing 10 (June): 5-15. Fitzgerald, Kate. 1990. "Sears' Plan on the Ropes." Advertising Age 61 (January 8): 1, 42. Floyd, Marilyn J. 1975. "Are Parents Satisfied with their Children's Education?" Florida Journal of Educational Research 17: 16-22. Fornell, Claes and William T. Robinson. 1983. "Industrial Organization and Consumer Satisfaction/Dissatisfaction." Journal of Consumer Research 9 (March): 403-412. Francken, Dick A. and W. Fred van Raaij. 1985. "Socio-Economic and Demographic Determinants of Consumer Problem Perception.'" Journal of Consumer Policy 8 (September): 303-314. Hall, Judith H. and Michael C. Dornan. 1988. "Meta-Analysis of Satisfaction with Medical Care: Description of Research Domain and Analysis

of Overall Satisfaction Levels." Social Science and Medicine 27 (6): 637-644. Headey, Bruce and Alex Wearing. 1988. "'The Sense of Relative Superiority--Central to Well-Being." Social Indicators Research 20 (October): 497-516. Hill, Ronald Paul and James C. Ward. 1989. "'Mood Manipulation in Marketing Research: An Examination of Potential Confounding Effects." Journal of Marketing Research 26 (February): 97-I04. Hughes, Donald A. 1977. "An Investigation of the Relation of Selected Factors to Consumer Satisfaction." In Conceptualization and Measurement of Consumer Satisfaction and Dissatisfaction, pp. 300-332. Ed. Keith Hunt. Cambridge, MA: Marketing Science Institute. Hunt, H. Keith. 1977. "CS/D--Overview and Future Research Directions." In Conceptualization and Measurement of Consumer Satisfaction and Dissatisfaction, pp. 455-489. Ed. Keith Hunt. Cambridge, MA: Marketing Science Institute, 455-489. "Improving Customer Satisfaction: What Works? What Doesn't?" 1989. Chief Executive 12 (September/October): 90-92. Isen, Alice M., Margaret Clark, Thomas E. Shalker, and Lynn Karp. 1978. "Affect, Accessibility of Material in Memory, and Behavior." Journal of Personali~ and Social Psychology 36 (January): 1-12. Jacoby, Jacob, Donald E. Speller, and Carol Kohn Berning. 1974. "Brand Choice Behavior as a Function of Information Load: Replication and Extension." Journal of Consumer Research 1 (June): 33-42. Jacoby, Jacob, Donald E. Speller, and Carol A. Kohn. 1974. "'Brand Choice Behavior as a Function of Information Load." Journal of Marketing Research 11 (February): 63-69. Justice, Blair and George McBee. 1978. "% Client Satisfaction Survey as One Element in Evaluation.'" CommuniO, Mental Health Journal 14 (Fall): 248-252. Kahneman, Daniel and Amos Tversky. 1979. "Prospect Theory: An Analysis of Decision Under Risk." Econometrica 47 (March): 263-291. Kahneman, Daniel and Amos Tversky. 1984. "Choices, Values, and Frames." American Psychologist 39 (April): 341-350. Kohut, Andrew and Linda DeStefano. 1989. "Modem Employees Expect More from their Careers; Job Dissatisfaction Particularly High Among the Young." The Gallup Report 288 (September): 22-28. LaBarbara, Priscilla A. and David Mazursky. 1983. "A Longitudinal Assessment of Consumer Satisfaction/Dissatisfaction: The Dynamic Aspect of the Cognitive Process." Journal of Marketing Research 20 (November): 393-404. Larsen, Randy J., Ed Diener, and Robert A. Emmons. 1985. "An Evaluation of Subjective Well-Being Measures." Social Indicators Research 17 (July): 1-17. Lebow, Jay L. 1982. "Consumer Satisfaction with Mental Health Treatment." Psychological Bulletin 91 (March): 244-259. - - . 1983. "Research Assessing Consumer Satisfaction with Mental Health Treatment: A Review of the Findings." Evaluation and Program Planning 6 (3,4): 211-236. Levin, Irwin P. 1987. "Associative Effects of Information Framing." Bulletin of the Psychonomic Society 25 (March): 85-86. Levin, Irwin P., Richard D. Johnson, and Maria Davis. 1987. "How Information Frame Influences Risky Decisions: Between-Subjects and Within-Subjects Comparisons." Journal of Economic Psychology 8 (March): 53-54. LeVois, Maurice, Tuan D. Nguyen, and Clifford Attkisson. 1981. "Artifact in Client Satisfaction Assessment: Experience in Community Mental Health Settings." Evaluation andProgram Planning 4 (April): 139-I50. Lloyd, Sally A., Rodney M. Cate, and June M. Henton. 1984. "Predicting Premarital Relationship Stability: A Methodological Refinement." Journal of Marriage and the Family 46 (February): 71-76. Maddox, R. Nell. 1982. "The Structure of Consumers' Satisfaction: CrossProduct Comparisons." Journal of the Academy of Marketing Science 10 (Winter): 37-53. McClendon, McKee J. and David J. O'Brien. 1988. "Question-Order Effects on the Determinants of Subjective We/l-Being." Public Opinion Quarterly 52 (Fall): 351-364. McMahon, Robert J. and Rex L. Forehand. 1983. "'Consumer Satisfaction in Behavioral Treatments of Children: Types, Issues, and Recommendations." Behavior Therapy 14 (March): 209-225. McNeal, James U. and Charles W. Lamb. 1979. "Consumer Satisfaction as a Measure of Marketing Effectiveness." Akron Business and Economic Review 10 (Summer): 41-45.

JAMS

70

WINTER, 1992

MEASURING CUSTOMER SATISFACTION: FACT AND ARTIFACT

PETERSON AND WILSON

Michalos, Alex C. 1986. "An Application of Multiple Discrepancies Theory (MDT) to Seniors." Social Indicators Research 18 (November): 349373. Okun, Morris A., Robert W. Olding, and Catherine M. G. Cohn. 1990. "A Meta-Analysis of Subjective Well-Being Interventions Among Elders.'" Psychological Bulletin 108 (September): 257-266. Oliver, Richard L. 1980. "A Cognitive Model of the Antecedents and Consequences of Satisfaction Decisions." Journal of Marketing Re. search 17 (November): 460-469. ~ . 1981. "Measurement and Evaluation of Satisfaction Process in Retail Settings." Journal of Retailing 57 (Fall): 25-48. Parducci, Alien. 1965. "Category Judgment: A Range-Frequency Model." Psychological Review 72 (November): 407-418. Peterson, Robert A. and Matthew Sauber. 1983. "A Mood Scale for Survey Research." In AMA Educators' Proceedings, pp. 409-414. Ed. Patrick B. Murphy, et al. Chicago: American Marketing Association. Prakash, Ved and J. Michael Munson. 1985. "Values, Expectations from the Marketing System and Product Expectations." Psychology and Marketing 2 (Winter): 279-296. Reibstein, David J., Stuart A. Youngblood, and Howard L. Fromkin. 1975. "Number of Choices and Perceived Decision Freedom as a Determinant of Satisfaction and Consumer Behavior." Journal of Applied Psychology 60 (August): 434-437. Ricerea SuUa Qualita' Percepita Dogli Utenti Del Servizio Autotelefono. 1989. SIP Servizion Opinioni. Richins, Marsha L. 1983. "Negative Word-of-Mouth by Dissatisfied Consumers: A Pilot Study." Journal of Marketing 47 (Winter): 68-78. Sabourin, Stephane, Lise Bourgeois, Pierre Gendreau, and Monique Morval. 1989. "Self Deception, Impression Management, and Consumer Satisfaction with Mental Health Treatment." Psychological Assessment: A Journal of Consulting and Clinical Psychology 1 (June): 126-129. Sabourin, Stephane, Noalla Laferriere, Francesca Sicuro, Jean-Claude Coallier, Louis-Georges Coumoyer, and Pierre Gendreau. 1989. "Social Desirability, Psychological Distress, and Consumer Satisfaction with Mental Health Treatment." Journal of Counseling Psychology 36 (July): 352-356. Schuman, Howard and Stanley Presser. 1981. Questions and Answers in Attitude Surveys. New York: Academic Press. Schwarz, Norbert and Gerald L. Clore. 1983. "Mood, Misattribution, and Judgments of Well-Being: Informative and Directive Functions of Affectire States." Journal of Personality and Social Psychology 45 (September): 513-523. Schwarz, Norbert, Hans J. Hippler, Brigitte Deutsch, and Fritz Strack. 1985. "Response Scales: Effects of Category Range on Reported Behavior and Comparative Judgments." Public Opinion Quarterly 49 (Fall): 388-395. Smith, Ivan C. 1989. "Auditing Customer Satisfaction." ManagementReview 78 (May): 52-54. Smith, Tom W. 1979. "Happiness: Time Trends, Seasonal Variations, Intersurvey Differences, and Other Mysteries." Social Psychology Quarterly 42 (March): 18-30. Strack, Fritz, Leonard L. Martin, and Norbert Schwarz. 1988. "Priming and Communication: Social Determinants of Information Use in Judgments of Life Satisfaction." European Journal of Social Psychology 18 (October-November): 429-442. Strack, Fritz, Norbert Schwarz, and Elisabeth Gschneidinger. 1985. "Happiness and Reminiscing: The Role of Time Perspective, Affect, and Mode of Thinking." Journal of Personality and Social Psychology 49 (December): 1460-1469. Sudman, Seymour. 1967. Reducing the Cost of Surveys. Chicago: Aldine Publishing Company.

Swan, John E. 1969. "Experimental Analysis of Predecision Information Seeking." Journal of Marketing Research 6 (May): 192-197. Swan, John E., Jesse C. Sawyer, Joseph G. Van Matre, and Gall W. McGee. 1985. "Deepening the Understanding of Hospital Patient Satisfaction: Fulfillment and Equity Effects." Journal of Health Care Marketing 5 (Summer): 7-18. Tanner, Barry A. and Webb Stacy, Jr. 1985. "A Validity Scale for the Sharp Consumer Satisfaction Scales." Evaluation and Program Planning 8 (2): 147-153. Tversky, Amos and Daniel Kahneman. 1981. "The Framing of Decisions and the Psychology of Choice." Science 211 (January): 453-458. Underwood, Bill and William J. Froming. 1980. "The Mood Survey: A Personality Measure of Happy and Sad Moods." Journal of Personality Assessment 44 (Augus0: 404-414. Ware, John E. 1978. "Effects of Acquiescent Response Set on Patient Satisfaction Ratings." Medical Care 16 (April): 327-336. Ware, John E. and Allyson R. Davies. 1983. "Behavioral Consequences of Consumer Dissatisfaction with Medical Care." Evaluation and Program Planning 6 (3,4): 291-297. Ware, John E., Allyson R. Davies-Avery, and Anita L. Stewart. 1978. "The Measurement and Meaning of Patient Satisfaction." Health and Medical Care Services Review 1 (January/February): 1-15. Wedell, Douglas H. and Allen Parducci. 1988. "The Category Effect in Social Judgment: Experimental Ratings of Happiness." Journal of Personality and Social Psychology 55 (September): 341-356. Weinstein, Michal. 1989. "Consumers Still Like Service, But Their Enthusiasm Erodes." American Banker Consumer Survey 6: 18, 20. Westhrook, Robert A. 1980a. "Intrapersonal Affective Influences on Consumer Satisfaction with Products." Journal of Consumer Research 7 (June): 49-54. 9 1980b. "A Rating Scale for Measuring Product/Service Satisfaction." Journal of Marketing 44 (Fall): 68-72. Westbrook, Robert A. and Joseph A. Cote. 1979. "An Exploratory Study of Non-product-Related Influences upon Consumer Satisfaction." In Advances in Consumer Research, pp. 577-581. Ed. Jerry C. Glson. Arm Arbor, MI: Association for Consumer Research. Westbrook, Robert A. and Joseph W. Newman. 1978. "An Analysis of Shopper Dissatisfaction for Major Household Appliances." Journal of Marketing Research 15 (August): 456-466. Willerman, Lee. 1979. The Psychology of Individual and Group Differences. San Francisco: W. H. Freeman and Co. Wilton, Peter C. and Franco M. Nicosia. 1986. "Emerging Paradigms for the Study of Consumer Satisfaction." European Research 14 (1): 4-11.

ABOUT THE AUTHORS


Robert A. Peterson holds the John T. Stuart III Centennial Chair in Business Administration and the Charles Hurwitz Fellowship, both at the University of Texas at Austin. His Ph.D. is from the University of Minnesota. William R. Wilson is Executive Vice President of Vyvx, Inc., a subsidiary of the Williams Brothers Companies. He was previously on the faculties of Rice University and the University of Michigan. His Ph.D. is in psychology from the University of Michigan.

JAMS

71

WINTER, 1992

Das könnte Ihnen auch gefallen