The Relative Effects of Price, Store Image, and Intrinsic Product Differences on Product Quality Evaluation


George J. Szybillo and Jacob Jacoby (1972) ,"The Relative Effects of Price, Store Image, and Intrinsic Product Differences on Product Quality Evaluation", in SV - Proceedings of the Third Annual Conference of the Association for Consumer Research, eds. M. Venkatesan, Chicago, IL : Association for Consumer Research, Pages: 180-186.

Proceedings of the Third Annual Conference of the Association for Consumer Research, 1972      Pages 180-186


George J. Szybillo, Department of Psychological Sciences, Purdue University

Jacob Jacoby, Department of Psychological Sciences, Purdue University

According to early economic theory, a consumer's choice was based on the assumption that he knew what he was buying. This was probably a valid assumption considering the limited range of consumer goods available at the time. With increasing technology, however, products have become more varied and complex. As a result of these change*, the average consumer today seldom has the time, knowledge, or the means to make a direct determination of product quality (cf. Cox. 1962; Olson. 1972; Shapiro. 1970).

Cox (1962) was one of the first investigators to develop a model of the consumer product evaluation process. He suggested that a product could be viewed as "an array of cues." The consumer's task in evaluating any given product was to use cues from this array for making evaluative judgments about that product. Cues were conceived of as relating to dimensions or attributes of the product, and included such information as price, brand name, store name, friends' and salesmen's opinions, and inter-brand composition differences.

Cox hypothesized that consumers tend to evaluate cues on two dimensions: predictive value and confidence value. Predictive value reflected the probability that the cue was associated with an attribute of the product. Confidence value reflected the certainty that the consumer felt about his ability to interpret ant use that cue. For example, consider an audiophile and an unsophisticated housewife, both of whom are trying to form judgments about the quality of a tape deck. Both might share the belief that internal components represent a high predictive value. The audiophile, however, would probably have a higher degree of confidence than the housewife in judging whether the components were of good quality.

More recently, Olson (1972) has developed an extended conceptualization of the quality perception process. Quality perception is presented as a two-stage process in this model. Briefly, in Olson's Stage One the consumer perceives, distinguishes, and makes judgments of cues which he believes are related to product quality. Since it is difficult for the consumer to carry out objective product quality testing procedures, the average consumer judges product quality on the bases of cues which are "surrogate indicators." These indicators are variables associated with a product which are perceived by that consumer to be related to product quality. In Stage Two, the consumer combines his judgments of the cues into an overall judgment of product quality.

In addition to the two-stage description of the quality perception process, a conceptual framework for quality cues is provided which includes the concepts of: (a) intrinsic vs. extrinsic cues; and (b) cue awareness. Although other investigators (cf. Jacoby, Olson, & Haddock, 1971; Olson & Jacoby, 1972; and Valenzi & Andrews, 1971) have referred to the intrinsic-extrinsic dimension of quality cues, Olson (1972) has presented the most formalized approach to this dimension.

In this system, a cue to product quality is considered in terms of the degree to which it is intrinsically or extrinsically a part of the physical product. A cue is considered intrinsic to the extent that, if changed, a resultant change would take place in the physical product itself. As examples, intrinsic cues to the quality of beer would be flavor and aroma, while extrinsic cues would include price and brand name.

Predictive value, confidence value, and cue awareness are applicable to both intrinsic ant extrinsic cues in this system. In general, cues influence quality perception only if the consumer is aware of them (Olson. 1972).

A number of hypotheses have been derived from this conceptual framework. One of these states that, given both intrinsic and extrinsic cues, the consumer will depend more heavily on intrinsic than extrinsic cues in forming impressions of product quality. More specifically, extrinsic cues will have little effect on judgments of product quality when the predictive value and confidence value of intrinsic cues are high.

Data from two earlier studies tend to support this hypothesis. Valenzi and Andrews (1971) found that actual product composition differences in butter samples accounted for 13% of quality rating variance while price information accounted for only 4%. Jacoby et al. (1971) found that composition differences in beer samples accounted for 35% of the quality rating variance, considerably more than either Price or brand name information.

The primary purpose of the present investigation was to provide a specific, a priori test of this hypothesis, using a product possessing different cues than were involved in the earlier studies. While the intrinsic cues employed by Jacoby et al. (1971) and Valenzi and Andrews (1971) were primarily olfactory and gustatory in nature, the intrinsic cues for the product used in this study (ladies nylon hosiery) were primarily tactile.

Extrinsic cues utilized in the present study were price and store image information. Only two studies have investigated the effects of price and store image in combination as cues to product quality (Andrews & Valenzi, 1971; Stafford & Enis, 1969). Both studies were concerned with the specific effects of various levels of the two cues. The present study investigated the relative potency of these cues rather than focusing on the effects of various levels of these cues. Moreover, neither of these studies permitted actual product differences to vary as a third factor in the experimental design. Recent research (Jacoby et al., 1971) suggests that such procedures may over-simplify the quality perception process. Another consideration is that subjects in both these studies were exposed to all experimental conditions in their respective studies. This suggests that confounding demand characteristics may have unknowingly been introduced. In contrast, subjects in the present study were exposed to only a single experimental condition, i.e., there was independence of subjects across cells.

Finally, Shapiro (1970) has recently presented a model of quality perception for products without brand names. The model is based on an analysis of the interrelationships of perceived quality, perceived likelihood of purchase, and perceived worth (i.e., value for the money) for sweaters, chairs, and hosiery. Shapiro found that, for these products, perceived worth had a stronger relationship to perceived likelihood of purchase than did perceived quality. Inasmuch as unbranded product samples were used in the current investigation, the relationships between perceived quality, likelihood of purchase, ant perceived worth were also examined. The specific hypotheses of the study were:

Hypothesis 1. Intrinsic cues (i.e., cues which are directly due to composition differences between brands) would exert a greater effect on quality perception than would extrinsic quality cues (e g , price or store image)

Hypothesis 2. Perceived worth, defined as value for the money, would have a stronger relationship with perceived likelihood of purchase than would perceived quality



The subjects (Ss) were 90 female undergraduates at Purdue University enrolled in Introductory Psychology during the fall 1971 session One restriction was placed on subject participation subjects were required to have lived in the Lafayette area for a minimum of sis months The purpose of this requirement was to insure that the independent variable of store image possessed meaning for the Ss (N B An additional 23 and 14 Ss, respectively, were used to pre-test the instructions and experimental manipulations )


A 2 X 3 X 3 factorial, consisting of two levels of price (i e , present vs absent), three levels of store image (i e , high quality store image, low quality store image, and no information provided as to store image), and three levels of product samples (i.e , three physically different samples of panty hose), was employed Repeated measures were taken across the third factor (Winer, 1971) Fifteen Ss were used in each of the six experimental conditions

Independent Variables

Price (P) information was manipulated by either presenting or withholding the actual prices of the three product samples The prices employed ($0.98, $1.67, and $2.49) were based on the results of one of the pre-tests as well as comparative shopping

Store image (SI) information was manipulated by either presenting or withholding the names of stores purportedly selling the three product samples Two store names were selected on the basis of the pre-test to represent high and low quality store images The difference between the store with the highest mean quality rating (XH = 7.95, on a 9-point scale) and the store with the lowest mean quality rating (XL = 2 56) was highly significant (t = 16 58, df = 22, p < . 001)

Each product sample (PS) of panty hose was identified by either I, J, or R This decision was based on a pre-test which showed no differences in preference between these three letters The letters were randomly assigned to samples for each experimental condition For purposes of a simplified exposition, the letters I, J, and R will be used to designate the low, medium, and high quality samples, respectively, throughout the remainder of this paper

The three samples of panty hose were pre-tested on the attributes of knit, texture, workmanship, cut, ant overall quality Results of the analysis of variance of overall perceived quality indicated a significant main effect for samples (F = 24 97, df = 2, 26, p < 0001), and Newman-Keuls analysis indicated that all differences between pairs of means for the three samples (XI = 34.28, XJ = 54.64, XR = 84.50) were significant beyond the .01 level. These differences were in the anticipated direction.

Dependent Variables

The major dependent variable for this study was overall perceived quality. The minor dependent variables were perceived worth and perceived likelihood of purchase. All were measured using the 100-point scale procedure described in Jacoby et al. (1971).


Each of the sis cells was run separately. The subJects were se ted at a long table and separated by table-top partitions to cut off between-subJect communication. A female source introduced the experiment as a psychological study designed to develop reliable instruments for measuring hosiery perceptions.

SubJects were then given the questionnaire booklet, ant the female experimenter demonstrated how to use the rating scales. SubJects were told they would be evaluating three samples of panty hose available in the local area. Hosiery samples were presented ln wrapped gift boxes. Price information ant store image information were given, as appropriate, by labels attached tn the experimental samples and communicated verbally before subJects began evaluating the hose. Samples were checked after each experimental session for damage, replaced if necessary. and new labels assigned.


Manipulation Check

The manipulation check for store image indicated a significant main effect for "quality of store merchandise" (F = 53.02, df = 2, 87, p < .001). A Newman-Keuls analysis indicated that all differences between pairs of means for the three store image conditions (XL = 3.80, XO = 5.73, XH = 7.93) were significant beyond the .01 level. These differences were in the anticipated direction, with the store names selected on the basis of the pre-test to represent low store image ant high store image receiving the lowest and highest mean ratings, respectively.

Model for Data Analysis of Dependent Measure

The model for a repeated measure analysis of variance is appropriate if the following assumptions are satisfied: (a) homogeneity of the variance-covariance matrices, and (b) the symmetry of the pooled variance-covariance matrix (Winer, 1971). For the dependent measure, tests for homogeneity of variance for variation due to subJects within groups ant due to the interaction between samples and subJects within groups were computed. Complete assumptions about the variance-covariance matrices were not tested. Consequently, the analysis used the Greenhouse and Geisser (1959) procedure for all tests involving the samples factor. This procedure is negatively biased in that it uses conservative degrees of freedom, leading to errors in the direction of not rejecting false null hypotheses.

Data Analysis of Major Dependent Variable

The means and standard deviations of overall perceived quality ratings for each experimental cell are presented ln Table 1. Results of the analysis of variance based on these data are summarized in Table 2. Tests for homogeneity of variance yielded insignificant results for variation due to subjects within groups (Fmax.95 = 4.24, df = 6, 14), and due to interaction between samples and subjects within groups (Fmax.95 = 2.95, df = 6, 28).





A strong Product Samples main effect was obtained (F = 394.85, df = 1, 84, p < .0001), ant a Newman-Keuls analysis indicated that all differences between the pairs of means for the three samples (XI = 26.76, XJ = 51.10, XR = 85.15) were significant beyond the .01 level. The analysis also indicated a significant Store Image main effect (F = 5.46, df = 2, 84, p < .01). A Newman-Keuls analysis indicated that only the difference between the means for the low store image (XL = 49.75) ant high store image (XH = 58.72) conditions was significant.

Intrinsic quality cues were anticipated to have a greater effect on quality perception than extrinsic cues. To test this hypothesis, w2 values were computed for the effects of price, store image, ant product samples (Winer, 1971). In terms of these indices, 73% of the total variance was due to the main effect of samples, ant approximately 12 was tue to the main effect of store image. Price information was relatively ineffective in explaining any percentage of the total variance. This finding confirms Hypothesis 1.

The price-present, store-image absent condition of this study was analogous to Shapiro's (1970) experimental setting. The correlations between perceived quality and perceived likelihood of purchase for samples I, J, and R were .07, -.03, and .49, respectively. The correlations between perceived worth and perceived likelihood of purchase for samples I, J, and R were .72, .56, and .64, respectively. Tests were made for differences between dependent correlations (Bruning & Kintz, 1968). Perceived likelihood of purchase was more closely related to perceived worth than perceived quality for samples I and J (tI = 2.68, tJ = 2.60, df - 12, p < .05), but not for sample R (tR = .75). These results partially confirm the second hypothesis.


Consistent with earlier research (Valenzi & Andrews, 1971; Jacoby et al., 1971), intrinsic quality cues were fount to have a greater effect upon quality perception than did extrinsic cues. More specifically, composition differences in hosiery samples had a greater effect (accounting for 73% of the quality rating variance) than tit either price or store image information. In fact, contrasted to previous studies, the variance accounted for by product composition differences in this study is relatively large. This may be an effect of moving from products (i.e., butter and beer) where between-brand differences, although discriminable, have a relatively small range, to a product (i.e., hosiery) where product differences can have a relatively large range.

Comparing the relative strength of price and store image, the results indicate that while the effect of price was not significant, awareness of store image did influence quality perception. Apparently, knowing the samples came from a store which carries high quality merchandise increases quality ratings relative to knowing that the samples came from a store which carries low quality merchandise. Although the store image effect was statistically significant, its practical importance is mitigated by the amount of variance it explained. Once again, the price main effect was insignificant when price was embedded in a multi-cue framework.

Shapiro's (1970) hypothesis that perceived worth would be more strongly related to perceived likelihood of purchase than would perceived quality received partial support. Given continued support, several substantial implications for practitioners would ensue.

Further research on the quality perception process is needed. Various writers have posited hypotheses regarding consumer characteristics and their relationship to quality perception (e.g., Jacoby et al., 1971; Olson, 1972; Shapiro, 1970). Few, if any, of these hypotheses have been empirically examined. The conclusions of the present study and prior studies may require modification as more information is obtained regarding how such variables as frequency of u age, frequency of purchase, general self-confidence, perceived quality differences between brands, and importance of the product category affect the quality perception process.


Andrews, I. R., & Valenzi, E. R. Combining Price, Brand Name, and Store Cues So Form an Impression of Product Quality. 79th Annual APA Proceedings, 1971, 6, 649-650.

Bruning, J. L., & Kintz, B. L. Computational Handbook of Statistics. Glenview, Illinois: Scott, Foresman, and Company, 1968.

Cox, D. F. The Measurement of Information Value: A Study in Consumer Decision Making. In W. S. Decker (Ed.), Emerging Concepts in Marketing. Chicago: American Marketing Association, 1962, 413-421.

Greenhouse, S. W., & Geisser, S. On Methods in the Analysis of Profile Data. Psychometrika, 1959, 24, 95-112.

Jacoby, J., Olson, J. C., & Haddock, R. A. Price, Brand Name, and Product Composition Characteristics as Determinants of Perceived Quality. Journal of Applied Psychology, 1971, 55, 570-579.

Olson, J. C. Product Quality Perception: A Model of Quality Cue Utilization and an Empirical Test. Unpublished doctoral dissertation, Purdue University, 1972.

Olson, J. C., & Jacoby, J. Cue Utilization in the Quality Perception Process. Purdue Papers in Consumer Psychology, Paper No. 122, 1972.

Shapiro, B. P. The Effect of Price on Purchase Behavior. In D. L. Sparks (Ed.), AMA Fall Education Conference. Chicago: American Marketing Association, 1970. P. 42.

Stafford, J. E., & Enis, B. M. The Price-Quality Relationship: An Extension. Journal of Marketing Research, 1969, 6, 456-458.

Valenzi, E. R., & Andrews, I. R. Effects of Price Information on Product Quality Ratings. Journal of Applied Psychology, 1971, 37, 106-191.

Winer, B. J. Statistical Principles in Experimental Design. New York: McGraw-Hill Book Company, 1971.



George J. Szybillo, Department of Psychological Sciences, Purdue University
Jacob Jacoby, Department of Psychological Sciences, Purdue University


SV - Proceedings of the Third Annual Conference of the Association for Consumer Research | 1972

Share Proceeding

Featured papers

See More


How Categories Transform Markets through Non-Collective, Non-Strategic Collaboration

Pierre-Yann Dolbec, Concordia University, Canada
Shanze Khan, Concordia University, Canada

Read More


When Novices have more Influence than Experts: Empirical Evidence from Online Peer Reviews

Peter Nguyen, Ivey Business School
Xin (Shane) Wang, Western University, Canada
Xi Li, City University of Hong Kong
June Cotte, Ivey Business School

Read More


Ecce Machina Humana: Examining Competence and Warmth in Consumer Robots The two fundamental social judgment dimensions-competence and warmth-are as relevant for judging consumer robots as for humans. We find that competence has an increasing positive eff

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.