Is There a Valid Price Quality Relationship?

ABSTRACT - This study attempts to expand the scope of previous research on the influence of price on product evaluations. A 2 x 2 x 3 factorial experiment with multiple response measures examined price as an information cue while controlling for several potential sources of error identified in previous price cue studies. In contrast to many reported studies, no significant price cue effects were found, however, a highly significant ordering effect was identified. In addition, there were significant interactions between place of purchase, information levels and the ordering of price presentation.

Citation:

Richard J. Rexeisen (1982) ,"Is There a Valid Price Quality Relationship?", in NA - Advances in Consumer Research Volume 09, eds. Andrew Mitchell, Ann Abor, MI : Association for Consumer Research, Pages: 190-194.

Advances in Consumer Research Volume 9, 1982      Pages 190-194

IS THERE A VALID PRICE QUALITY RELATIONSHIP?

Richard J. Rexeisen (student), University of Minnesota

ABSTRACT -

This study attempts to expand the scope of previous research on the influence of price on product evaluations. A 2 x 2 x 3 factorial experiment with multiple response measures examined price as an information cue while controlling for several potential sources of error identified in previous price cue studies. In contrast to many reported studies, no significant price cue effects were found, however, a highly significant ordering effect was identified. In addition, there were significant interactions between place of purchase, information levels and the ordering of price presentation.

INTRODUCTION

Price as an information cue has stimulated considerable research interest over the last forty years. Early research (Levitt 1954; Scitovsky 1944; Tull et al 1964) typically examined the effects of price alone on consumer evaluations. More recent studies, however, have tended to consider multiple cues such as store image, multiple products, brand familiarity, and generic product information (Andrews et al 1971; Enis and Stafford 1969; Landon and Shafer 1974: Valenzi and Andrews 19715.

In a summary literature review, Olson (1977) reported that all single cue studies found price to have a significant effect on evaluative judgments or choices. This is not surprising given that price was the only information available to the subjects. Multiple cue studies, however, did not reveal a clear pattern of price cue effects. Olson attributed the inconsistent findings to serious methodological and/or conceptual flaws inherent in many of the price cue experiments.

Olson's (1977) review raises three key issues: l) Is there a valid price quality relationship?, 2) Are there alternative explanations for the reported findings of the price cue literature? and 3) Rave sufficient theoretical constructs been identified to formulate a meaningful conceptual framework for price cue research? The objective of this study is to address and expand upon these concerns.

BACKGROUND LITERATURE

Although single cue studies were highly simplified, they were useful in demonstrating the perceived price quality relationship and as such served as an important heuristic. The typical single cue study manipulated prices in a within subjects design. Subjects, generally students, were asked to make either a brand choice or a quality rating (Levitt 1954; Tull et al 1964; Lambert 1970, 1972; Deering and Jacoby 1972).

Enis and Stafford (1969) and Jacoby et al (1971) were not satisfied with the single-cue effect and hypothesized that cues influencing product evaluations (e.g., place of purchase, brand name, and product familiarity) probably interact with one another. To test this hypothesis, multiple cue studies have generally utilized the factorial research design and have analyzed their results by either a fixed model analysis of variance, or when appropriate, a multivariate analysis of variance. For a summary table of the research dealing with price cue effects on product evaluations refer to Olson (1977).

Olson (1977) recognized, however, that the validity of quantitative relationships between price and perceived quality depends heavily on interval scale properties of the quality measure and the assumption that equal price intervals are perceived as equal by the consumer. A major weakness of many price-cue studies has been that they have not used multiple dependent variables ant/or measures to cross validate their findings. Inadequate measures may partially explain some of the inconsistent findings of the price related literature.

Stafford and Enis (1969) found significant main and interaction effects between product, price and store image. The most interesting result of this study was that the experimental evidence tended to support the hypothesis that the judgment of quality by price can be confounded by nonprice information about the product. Landon and Schafer (1974), and Wheatley and Chiu (1977), however, did not discover significant interactions in their attempt to replicate the Stafford et al (1969) study. Demand characteristics, however, are potentially present in all three studies.

Gardner (1970) reported significant main effects but no interactions between product type or search time factor, and price. In a second study, however, Gardner (1971) failed to find a similar affect involving the price cue. Gardner's second study, however, also examined the influence of brand name on quality ratings. It is possible that brand name, as an information chunk, has a more powerful effect on reported quality measures than does the price cue.

Jacoby et al (1971) tentatively hypothesized that a product's physical characteristics may enter into a complex interaction with information cues (such as price) and reported quality measures. Valenzi and Andrews' (1971) study found significant main and two-way interaction effects between physical composition cues and price cues. The Valenzi experiment appears to add credibility to Jacoby's hypothesis.

Other studies, however, have failed to identify any price-cue relationship. Szybillo and Jacoby (1972) reported no significant main or interaction effects on quality evaluations and brand name and store image information. This study represents one of the few published experiments where the findings are non-significant with regard to price-cue effects. The results of this study, however, were also not conclusive because of potential demand characteristics in the price present conditions.

While most multiple cue studies find significant price cue effects, some researchers have found that price remains the dominant cue (Andrews et al 1971) and others have found that it declines in importance (Jacoby et al 1971; Rao 1971). To date, this conflict has been difficult to reconcile. Gardner (1970) has suggested that price effects on quality perceptions are product specific which may account for some of the confusion in the literature.

Olson (1977) identified four major problems with the previous price cue research: 1) confoundings were common (unintentional manipulation of two or more variables within a single independent variable), 2) demand characteristics were potentially present in many of the price cue studies, 3) design factors say have resulted in different price effects (price as a between- or within-subjects factor), and 4) many studies did not use multiple dependent variables or multiple measures of the dependent variables (a measure of convergent validity). Olson suggested that these methods logical weaknesses have confounded the interpretability of the price-cue literature.

In summary, there is convincing evidence that in the absence of other cues, price acts as a significant information cue. In contrast, multiple cue studies have not demonstrated consistent price cue effects. Olson (1977) suggested that some of the inconsistencies may be attributable to serious methodological flaws or measurement problems. This experiment is designed to control for the sources of error identified by Olson and explore possible alternative explanations for the price cue effect.

CONCEPTUAL FRAMEWORK

Berkowitz and Walton (1980) have noted that the encoding process is especially important to an individual's price perception. Fair price theory (Kamen and Toman 1970) and adaptation level theory (Monroe 1977) suggest that the consumer evaluates subsequent price cues after comparison with an adaptation level price. Monroe (1977) has further observed that an individuals standard for judging stimuli is actually like a sliding scale that is influenced by previous stimulus values.

While these theories are informative and are of considerable value in single cue studies, they do not adequately explain the complex interactions reported in multiple cue studies. Further, the inconsistent findings reported in the price literature have seemingly confounded the identification of reliable theoretical constructs. Therefore, rather than propose a conceptual framework based on inadequate or unreliable constructs, the purpose of this study will be to test, rather than explain, the price-quality relationship.

The focus of this study will be on the remaining two issues raised by Olson (1977): 1) Is there a valid price quality relationship? and 2) Are there alternative explanations for the price cue effect? The specific research hypotheses are:

H1: Quality ratings will be differentially affected by price.

H2: Product information will moderate the price cue effect

H3: Place of purchase will differentially affect quality ratings.

H4: Order of price presentation will not differentially affect quality ratings.

Hypotheses 1, 2, and 3 were extracted from the pricing literature and pertain to the validity of the price-quality relationship. Hypothesis 4, however, was based on the observation that pricing studies have typically not reported any randomization of price presentation formats (i.e., low vs. high price being shown first). This suggests that the randomization of the price format is either so common-sensical as to exempt its mention, or implicitly that price presentation will not differentially affect quality ratings. Monroe (1977), however, provides some evidence that ordering effects to occur. The purpose of H4 is to test the ordering effect within the contest of the price quality relationship.

METHODOLOGY

Subjects were one hundred and thirteen undergraduates enrolled in introductory marketing at the University of Minnesota during the Winter quarter of 1981. Subjects were drawn on a convenience basis and were given extra credit for their participation.

A 2 x 2 x 3 factorial experiment design was utilized. The respective factors consisted of two information levels (with or without product information provided), two stores (product was identified as either coming from a department store or a carpet specialty store). Price levels were treated as a within subjects factor, while information and place of purchase were considered as a between subjects factor. Figure 1 illustrates the treatment conditions.

FIGURE 1

Carpeting was selected as the experimental product because of its use in other price related experiments (Enis and Stafford 1969; Landon and Schaffer 1974; Shapiro 1974; Wheatley and Chiu 1977) and because college students were presumed to have little experience making quality assessments of carpeting. Ali carpet samples were identical but controls, as later discussed, were utilized to minimize student's guessing that the samples were in fact identical.

Information about carpeting was developed from a pretest of 38 college students in a different section of introductory marketing during the same period. The pretest subjects were asked what types of information they would like to have prior to making a carpet purchase and what they thought the range of prices would be for a square yard of carpeting. Anticipated price ranges were used rather than actual price ranges to more accurately reflect the student's perceptions of carpet price ranges. Product information was developed (i.e. information on tensity, thickness, composition) based on reported preferences for the information. Student; were also asked if they knew much about carpeting and the consensus was that they did not.

Subjects were randomly assigned to treatment conditions (Dept. store with no information, Dept. store with information, Carpet store with no information, or Carpet store with information). Neither the "with or without" information groups were given price information. Subjects provided with information were given three minutes to read the information and given the opportunity for more time if they requested it. No subject requested additional time to read the information.

Within each treatment condition subjects were presented first with a carpet sample that had no price on it. Carpet samples were designated only as A, B, or C. Subjects were told that the brand names had been moved to add realism. The no price treatment had been recommended by Olson (1977) and provided a base line for subsequent analYsis.

The second carpet sample was either high or low priced ($7 or $29 per square yard) based on a random assignment within each treatment condition. Subjects were then asked to make various evaluations. Order of presentation was noted for each subject for use subsequently as a blocking factor. The third carpet sample was high priced if the former was low and vice-a-versa. Subjects could look only at one sample at a time and were instructed not to change any answers or look back at a previous response. Carpet samples were turned upside down to make sure only one carpet sample could be looked at to avoid direct comparisons. Questions about each sample were on separate pages of the questionnaire to minimize multiple response contamination.

Multiple dependent variables were utilized and multiple measures were taken of each variable. Dependent variables included: 1) a quality rating, 2) value for the money, and 3) perceived worth. Semantic scales, ratio and Likert type questions were utilized as a basis of testing convergent validity for each respective dependent variable. Subjects were questioned after the experiment to ascertain whether demand characteristics were present.

ANALYSIS AND RESULTS

Preliminary analysis revealed that the three dependent variables did not exhibit substantial intercorrelations (.08<r<.41). This suggested that each variate should be analyzed in its own right rather than employing a multivariate analysis of variance (MANOVA). Multiple measures of each variate, as noted above, were significantly correlated with r's generally above .80. Therefore, it was concluded that each response variate exhibited the desired convergent validity and merited a comprehensive analysis.

A univariate analysis of variance (ANOVA) was computed using the regression approach, blocking on order of presentation. Table 1 summarizes the results for each ANOVA run.

TABLE 1

SUMMARY ANALYSIS OF VARIANCE DATA (F SCORES)

This first point of interest in Table 1 is that neither place nor information levels resulted in significant main effects for any of the three dependent variables. However, highly significant ordering effects were observed for both the perceived quality and value for the money variate. Perceived worth was in the same direction but the ordering effect was only marginally significant.

Table 1 also reveals that there were no interaction effects between priceS place or information levels for the quality or perceived worth variates. This allows direct assessment of the main treatment effects as opposed to examining the influence of treatments as individual values of the other factors. Since information levels and place resulted in significant interactions for the response variate value for the money, differences were examined further. Mean response patterns for place and information are shown in Figure 2.

FIGURE 2

MEAN RESPONSES TO INFORMATION TREATMENTS

Figure 2 suggests that Department stores benefit (receive better reported value for the money ratings) from having information made available to the customers. In contrast, carpet sales and service only stores received lower ratings on value for the money in the information present condition. It is interesting to note that value for the money exhibited this pattern but not quality ratings for place of purchase. Therefore, hypothesis 3 can not be accepted.

Figures 3 and 4 illustrate the interactions between information and order effects which were highly significant (p=.005) for the quality dimension and marginally significant (p-.079) for value for the money. What is especially interesting to note is the direction of the mean responses to the ordering effect. Low price first subjects rated overall quality higher than did high price first subjects. The differences were highly significant as indicated by the main effect for order. Figure 3 suggests no ordering effects for the without information treatment group while a significant change for the with information treatment group.

Adaptation level (AL) theory, Helson (1964) and others, would have predicted that in the absence of price information, all subjects should have rated quality an average "4". This was exhibited by the "without information" group. What should be of interest to researchers, however, is that the "with information" group did not respond as predicted by AL theory, nor as would have been expected by the price quality literature. Recall that the "with information" group was not provided price information.

FIGURE 3

MEAN RESPONSES TO ORDER TREATMENTS

FIGURE 4

MEAN RESPONSES TO ORDER TREATMENTS

The differences between responses to price levels and order treatment was significant for the perceived quality measure As shown in Figure 5 quality ratings remained unchanged the deteriorated for the "high price first" group. In contrast the "low price first" group improved their ratings when the price was raised.

The highly significant interaction between price levels and order effect was in the expected direction. Figure 6 illustrates that respondents rated carpeting a better value for the money when it was shown with a lower price for either treatment group. The reported interaction for this dimension is an artifact of the experimental design; price level for B and C is a function of order format.

FIGURE 5

MEAN RESPONSES TO PRICE LEVELS

FIGURE 6

MEAN RESPONSES TO PRICE LEVELS

Sex differences were analyzed using individual t-tests for each dependent measure and no significant differences were found except that women rated style information more highly than men (p-.006). T-tests were computed rather than using sex as a blocking factor in the ANOVA run because of cell size limitations. A separate univariate analysis was computed for sex and type of information requested but no significant differences (p-.125) were found.

Price. aside from the ordering effect, did not result in any significant changes in quality ratings. Therefore, hypothesis one can not be accepted and hypothesis four is rejected because of significant ordering effects. Hypothesis two is no longer meaningful because price did not have an effect in either the information present or absent conditions, except as noted for the ordering effect; Finally, subjects were questioned about the experiment at its conclusion and there did not appear to be significant demand characteristics. The significant ordering effect would, in fact, argue against the presence of demand characteristics.

DISCUSSION

Results from this experiment have significant implications for both researchers and practitioners. To a large extent the potential sources of error identified in earlier studies have been controlled or accounted for. This is especially true for confounding effects, demand characteristics and dependent variable validity considerations. The design factor consideration was only partially addressed by this study as this experiment was not a fully crossed, repeated measures design.

Implications for Researchers

The findings of this study do not support the traditional notion of the price quality relationship. First, this study joins a handful of studies (Gardner 1971; Jacoby et al. 1971; Roa 1971; Szybillo and Jacoby 1972) which did not find significant main and/or interaction effects for price-quality. Second, this study suggests that ordering effects, if uncontrolled, may confound the interpretation of main effects. The resulting error could be attributed to method variance.

The significance of the main ordering effect is magnified by its multiple interactions with place of purchase and information levels. This finding would strongly suggest that price levels must be randomized for the within subjects design or significant confounding effects could be present. A review of the literature does not indicate that previous within subjects research, which constitutes a majority of price studies, randomized price levels. In addition, significant order interactions could also confound interpretations of between subjects designs.

Aside from the issue of method variance, the ordering effect has interesting research implications. When information was present, high price first resulted in lower overall quality ratings. This affect is in the opposite direction than would have been predicted by the classical price-quality literature. However, the theory of assimilation and contrast does suggest that when a price falls outside a subjects latitude of acceptance it will be discounted further.

A review of the mean responses to price levels also reveals some interesting ordering effects. In the no price treatment, all subjects rated quality of the carpeting similar. When the high price was shown first to subjects, no change was recorded in their quality rating. However, there was significant lowering of the quality rating when the low price carpeting was shown next. In contrast, subjects significantly raised their quality ratings when the low price was shown first and maintained these ratings when the price was raised. Perhaps the affect reflects an inflation mentality.

Considerations for Practitioners

Retailers should also recognize the significance of these findings. First, retailers should recognize the long term effects that one pricing decision will have on subsequent decisions. This experiment would suggest that quality perceptions are adversely affected by lowering the price. This study does not, however, address the influence reduced quality perceptions have on the willingness of a consumer to buy.

Second, pricing strategies based on the traditional price cue research should be cautiously employed. If a high price strategy is selected with the objective of raising quality perceptions, the firm may be committed to that strategy if quality ratings are to be maintained.

SUMMARY

This experiment was designed to address three questions: 1) Is there a valid price quality relationship?, 2) Are there alternative explanations for the reported price cue effect? and 3) Have sufficient theoretical constructs been identified to formulate a meaningful conceptual framework for price cue research? The treatment conditions were place of purchase, generic product information, and price levels. Price presentation format was analyzed as a subsequent blocking factor. Response measures were perceived quality, value for the money and perceived worth.

The author regrets that a conceptual framework could not be developed for this study. Clearly, theoretical constructs need to be identified before experimental results can be meaningfully generalized. Nevertheless, this study did demonstrate the importance of considering the effect of one pricing decision on subsequent decisions.

There does appear to be a valid price quality relationship. However, in contrast with previous studies the relationship identified in this study could be explained by a significant ordering effect.

The usage of students as experimental subjects may be viewed as a weakness in the study, however, their usage is consistent with traditional price research. A second criticism of the study may be that it was product specific. The issue of product generalizability is indeed important and should be addressed further. The primary reason the issue was not considered in this study was because multiple products potentially introduce confounding effects which this experiment was not designed to control.

Finally, future price research would benefit from considering the ordering effect identified in this study. Research into product generalizability, sample frame appropriateness, and conceptual framework development are all areas which should prove highly productive.

REFERENCES

Andrews, E. R., and Valenzi, E. R. (1971), "Combining Price, Brand and Store Cues to Form an Impression of Product Quality," Proceedings, 79th Annual Convention, American Psychological Association, p. 649.

Berkowitz, Eric N. and Walton, John R (1980), Contextual Influences on Consumer Price Response: An Experimental Analysis," Journal of Marketing Research, (August). pp. 340-358.

Deering, Barbara J., and Jacoby, Jacob (1972), "Price Intervals and Individual Price Limits as Determinants of Product Evaluation and Selection," Proceedings, Third Annual Convention of the Association for Consumer Research, M. Venkatesan, (ed.), pp. 145-66.

Enis, Ben M., and Stafford, James E. (1969), "The Price-quality Relationship: An Extension," Journal of Marketing Research, 6, (November), pp. 456-58.

Gardner, David M. (1970), "An Experimental Investigation on the Price Quality Relationship," Journal of Retailing, 46, (Fall), pp. 39-40.

Gardner, David M. (1971), "Is There a Generalized Price/ Quality Relationship," Journal of Marketing Research, 8 (May), pp. 241-243.

Gardner, David M. (1971), "Is the Price/quality Relationship Important," Faculty Working Papers, College of Commerce and Business Administration, University of Illinois, Paper No. 186.

Helson, Harry, (1964), Adaptation-Level Theory, Harper Row, NY.

Kamen, J. N., and Toman, R. J. (1970), "Psychophysics of Price," Journal of Marketing Research, 7 (February), pp. 27-35.

Landon, L., and Schaefer, K. S. (1974), Risk Style and the Price-Quality Relationship, Paper presented at the Southwestern Marketing Association Meeting, Dallas, TX: March 1974.

Levitt, Harold J. (1954), "A Note on Some Experimental Findings About the Meaning of Price," Journal of Business, (July), pp. 205-210.

McConnel, J. Douglas (1968), "An Experimental Study of the Development of Brand Loyalty," Journal of Marketing Research, 5, (February), pp. 13-19.

Monroe, Rent B. (1977), "Objective and Subjective Contextual Influences on Price Perception," Consumer and Industrial Buying Behavior, (eds.), Arch G. Woodside, Jagdish N. Sheth, and Peter D. Bennett, New York: North-Holland, pp. 287-296.

Olson, Jerry (1977), "Price as an Information Cue: Effects on Product Evaluations," (eds.) Arch G. Woodside et al., in Consumer and Industrial BuYer Behavior, New York: North-Holland, pp. 267-286.

Rao, Vitala R. (1971), "Salience of Price in the Perception of Product Quality: A Multidimensional Measurement Approach," AMA 1971, Spring and Fall Conference, p. 576.

Rao, Vitala R. (1972), "Marginal Salience of Price in Brand Evaluations," (ed.), M. Venkatesan, in Proceedings of the Association for Consumer Research, pp. 125-144.

Scitovsky, Tibor (1944-5), "Some Consequences of the Habit of Judging Quality by Price," The Review of Economic Studies, 12, pp. 100-105.

Shapiro, Benson B. (1973), "Price Reliance: Existence and Sources," Journal of Marketing Research, 10 (August), pp. 287-291.

Szybillo, George J., and Jacoby, Jacob (1974), "Intrinsic and Extrinsic Cues as Determinants of Perceived Quality," Journal of Applied Psychology, 59, pp. 74-78.

Tull, D. S., Boring, R. A., and Gonsior, M. R. (1964), "A Note on the Relationship of Price and Imputed Quality," Journal of Business, 38. (April). PO. 186-191.

Valenzi, E., and Eldridge, L. (1973), "Effect of Price Information, Composition Differences, and Rating States on Product Quality Ratings," in Proceedings of the 81st Annual Conference of the Psychological Associations pp. 829-830.

Wheatley, John J., and Chiu, John S. (1977), "The Effects of Price, Store Image, and Product and Respondent Characteristics on Perceptions of Quality," Journal of Marketing Research, 14 (May), pp. 181-186.

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