A Multivariate Analysis of the Perception of Value From Retail Price Advertisements

Albert J. Della Bitta, University of Rhode Island
Kent B. Monroe, Virginia Polytechnic Institute and State University
ABSTRACT - This study attempts to expand the scope of previous research on consumers' reactions to retail comparative price advertising. A large factorial experiment with multiple response measures examined the influence of several variables on consumers' perception of value from retail sale advertisements. The size of sale discount was found to have an interesting effect on consumers' perceptions of value. Also, the method of presenting sale information significantly influenced perceptions.
[ to cite ]:
Albert J. Della Bitta and Kent B. Monroe (1981) ,"A Multivariate Analysis of the Perception of Value From Retail Price Advertisements", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 161-165.

Advances in Consumer Research Volume 8, 1981      Pages 161-165


Albert J. Della Bitta, University of Rhode Island

Kent B. Monroe, Virginia Polytechnic Institute and State University


This study attempts to expand the scope of previous research on consumers' reactions to retail comparative price advertising. A large factorial experiment with multiple response measures examined the influence of several variables on consumers' perception of value from retail sale advertisements. The size of sale discount was found to have an interesting effect on consumers' perceptions of value. Also, the method of presenting sale information significantly influenced perceptions.


The issue of consumers' response to price has stimulated considerable research interest in the past decade (Monroe 1973, 1977; Olson 1977). However, much of this inquiry has been directed at the price/"quality"-perception question. Except for the concept of unit pricing, one area receiving relatively little attention is the issue of presenting price information to consumers. This is unfortunate since numerous alternatives exist for presenting prices, and such information is an integral component of much retail advertising. Consequently, research into the price presentation area could have significant implications for managerial and regulatory decision-making.

Several schemes are useful for conceptualizing the problem. Jacoby and Olson (1977) have viewed price perception from the S-O-R perspective. Here stimulus variables (S) are the actual cues presented to the consumer where organismic variables (O) include those factors internal to the individual influencing the acquisition, processing, and derivation of subjective meaning from objective stimuli. These O-variables could include properties such as price awareness, subjective standards of reference and attitudes. The interaction of acquired stimuli and organismic conditions leads to responses (R) by the individual.

Clearly, presentation of an offered price qualifies as an S-variable which can significantly influence a consumer's perception of the offer. However, it should also be clear that the consumer's experiences and the context of an offering can influence perception of the offered price. Thus, price experiences over tine and the prices of other products may be internalized as an O-variable to influence price perceptions. To illustrate, adaptation-level theory and empirical evidence suggest that the consumer's price experience forms a specific and quantifiable standard of reference for judging subsequent prices (Della Bitta and Monroe 1974; Nwokoye 1975).

In addition to such internalized influences, a variety of factors surrounding the price presentation (store setting, medium used to convey price information, etc.) form a context for judgment. Besides these general contextual variables, a given price presentation may be comprised of a set of more specific contextual cues also capable of influencing the meaning derived from the offered price. Walton and Berkowitz (1980) have classified these cues into comparison and semantic categories. Comparison cues are prices the advertiser references in order to provide a basis for Judging the offered price. This may be a prevailing price in the market, the manufacturer's suggested selling price, or the retailer's regular selling price. Semantic cues involve phrases describing the price ("very low"), price reduction ("significantly reduced") or amount of savings involved ("super savings"). Of course, some semantic phrases are emotionally laden while others are more neutral.

From a managerial and public policy viewpoint, a valid concern is how strongly price presentation formats influence buyers. The managerial concern is to present price information in a favorable manner while hopefully not misleading consumers about the nature of the offering. From a public policy perspective, the concern is over misleading and/or deceptive practices. In fact, in 1958 and 1964 the FTC issued a set of guides for presentation of retail price information.

Relevant Research

Whet is significant to note is the paucity of directly applicable research evidence to guide managers and policy makers. However, the few relevant studies available do provide some perspective. Fry and McDougall (1974) investigated buyers' perceptions of advertised sale and regular prices. Although the study probably suffered from experimental demand effects, the authors concluded that acceptance of advertised regular prices tended to vary inversely with the size of discounts. However, acceptance of the sale price as the "lowest price in town" varied directly with discount magnitude.

Barnes (1975) varied price and semantic information in a simulated newspaper advertised "sale" for unbranded products. Each presentation was ranked on a priori basis by Barnes in terms of the amount of its information content. The ranking (high to low) was disclosure of: (1) both regular and sale price, (2) sale price and the discount expressed in percentage terms and (3) only sale price. Newspaper and store type were also varied in the experiment. Dependent measures were believability of the advertisement, perceived value for the money and motivation to act. Results indicate a significant information treatment effect, with presentation format (1) above having the greatest influence on the dependent measures. Although this result is consistent with Barnes' a priori ranking of information content, one should note that such a ranking may not be valid in all circumstances. In some situations consumers may find an expression of the magnitude of the price reduction (e.g., percent off) to be more useful than knowing the specific sale price.

Keiser and Krum (1976) compared consumers' reactions to sale advertisements of different forms. In one, only sale price was shown while in a second condition both regular and sale prices were presented with the semantic phrase "one-half price sale." The authors report that the second condition influenced more subjects to perceive a "true price reduction" although the difference was not significant. Perceived truthfulness and willingness to purchase were also not significantly different between the two treatment groups. The possibility of confounding effects in this study should also be mentioned. In the presentation that disclosed both regular and sale prices, the former was set artificially high to examine consumers' reactions to overstated price savings. This element, not in the other treatment condition, may have introduced a bias into the study.

Sewall and Goldstein (1979) queried 114 catalog store shoppers about their understanding of comparison (reference) prices used by catalog showroom retailers. They found that the vast majority clearly understood the meaning of reference price information. Also, 58 percent of the respondents understood or were explicitly skeptical (discounted the credibility) of such reference prices and regarded them as useful in their comparative shopping activities. It should also be noted that subjects can be influenced by such comparative information more than conscious responses to surveys might suggest.

Blair and Landon (1979) examined the effects of reference prices in retail advertisements for a TV and food processor using adult women in a shopping mall setting. One condition contained only the offered price while another presented the offered price, suggested list price and claimed savings. Results indicate that without reference price information subjects generally estimated savings et between 10 to 12 percent. However, advertisements with reference prices (and discounts from 16 to 36 percent) produced significantly larger perceptions of savings. Also the perceived credibility of claimed savings varied inversely with the magnitude of the claim.

Walton and Berkowitz (1980) reported a study of the effect of reference price, presentation format and several other contextual variables on consumers' price perceptions. Three products (aspirin, fan and camera) were tested at two discount levels (20% and 40%) and four information presentation formats: (1) "regular price"/"sale price", (2) "total value"/"sale price", (3) "compare at"/"our price", and (4) "x percent off"/"now only". Several dependent measures were analyzed using multivariate analysis of variance. Two particularly interesting findings are: (1) reference price effects varied across products, and (2) price information format was significant for cameras only. Here the "percent off" format (4) consistently received the lowest mean rating on all dependent measures.

Although informative, the major problem with previous studies is the lack of consistent findings. For example, Barnes' found the regular/sale price format to yield the greatest perception of value for the money. Although Walton and Berkowitz did not find this to be the case, they did observe a significant difference between the "percent off" format and all others. The two studies also show a similar contrast regarding the influence of information treatments on subjects' willingness to purchase. In addition, while Blair and Landon found the regular/sale price format to influence perceived savings, Walton and Berkowitz were unable to distinguish this influence from two other presentations. Discrepancies also appear between the findings of other studies.

Differences in test conditions may explain some of these divergent findings. Studies have varied in terms of the discount levels, information formats and dependent measures used. Further, they have tested different product classes, branding conditions and price levels. Additional research that accounts for some of these contextual influences is warranted.


This study investigated the influence of specific price presentation formats on consumers' perceptions. The focus was on a sale situation that consumers might confront in a retail setting.

Research Method

Prior to the experiment, content analysis of major metropolitan newspapers and catalog advertising revealed the frequency with which various components of sale information were employed. Four components were chosen for investigation based on their frequency of occurrence: regular (non-sale) price, sale price, percentage price reduction, and absolute dollar amount of price reduction. Because previous research suggests that the influence of price presentation format could be effected by the price level and discount size, these factors were also incorporated.

A 2 x 5x 8 factorial experimental design was used. The respective factors were comprised of two price levels, five discount levels (10% to 50% in 10 percentage point increments) and eight presentations of regular price (RP), sale price (SP), percent off (PO) and amount off (AD) information. The specific treatment levels are (1) SP; (2) RP, SP; (3) RP, AO; (4) RP, PO; (5) RP, SP, AO; (6) RP, SP, PO; (7) RP, AO, PO; (8) RP, SP, AO, PO.

A calculator was used as the experimental product because of its familiarity and potential interest to college students who were employed as subjects. Two specific models were used et regular suggested price levels prevailing in the market at the time ($120 and $50). The 400 participating subjects were randomly assigned (five per cell) to the 80 treatment conditions. All subjects were instructed to assume that they sew the provided advertisement which described the product and exposed then to the price presentation format. They were then asked to read the advertisement and respond to a series of dependent measures and personal background/debriefing questions.

The majority of price investigations have only examined s single dimension of buyers' response to price. However, since price presentations have potential for affecting a variety of buyer reactions, single variable studies forego opportunities to explore the potential richness of buyers' multidimensional responses to price presentations (Engel, Blackwell and Kollat 1978; Jacoby 1978; Wind and Denny 1974). This study focused on three response variates: perceived savings, perceived value for money at the sales price and perceived acceptability of the offer. These variables were selected because of a desire to concentrate attention on the perceived worth construct identified in other work (Szybillo and Jacoby 1974; Walton and Berkowitz 1980) and the attention given to the three variables in over nine previous investigations. Each of the three dependent variables was measured on a seven-point equal-interval scale anchored by descriptive phrases appropriate to the variable involved.

Analysis and Results

Preliminary analysis revealed that the three dependent variables exhibited substantial intercorrelations (r > .58). The three variables were also evaluated for internal consistency using Cronbach's (1951) alpha criterion. The resulting alpha value of 0.81 is quite large for this type of basic research (Nunnally 1967), suggesting high consistency among the response variates.

In cases where multiple response measures are correlated, separate univariate analysis of variance (ANOVA) runs on each dependent variate can lead to incomplete and/or inappropriate conclusions (Green 1978; Tatsouka 1971; Wind & Denny 1974). For this reason a multivariate analysis of variance (MANOVA) procedure as summarized in Table 1 was selected. Wilks' Lambda is an appropriate statistic for examining differences among group centroids when two or more dependant measures serve as the joint set of evaluation criteria, The distribution of the F-statistic is proportional to the distribution of Lambda and is employed as the test statistic.

A first point of interest in Table 1 is that all interaction terms are nonsignificant. This allows direct assessment of main treatment effects as opposed to examining the influence of treatments at individual values of the other factors.



As shown in Table l, price level treatments did not produce a significant difference in subjects' perceptions. However, examination of the group centroids suggests some tendency for subjects to perceive leas value for the money in the more expensive model of the test product.

Table 1 also reveals that discount size and information treatments generated significant differences in subjects' perceptions. For these situations it is appropriate to engage in further analysis to isolate the nature of the differences. While a variety of techniques can assist in this exploration (Green 1978; Hair et. al. 1979; Tatsuoka 1971) these data were further analyzed using a simultaneous confidence interval approach to multiple comparison rests (Mortises 1976). This approach to follow-up tasting for a MANOVA is highly conservative since it holds a maximum alpha level for all possible comparisons of a given type. Therefore, the actual alpha level on a given test can be quite small, requiring a considerably large value for the computed test statistic. For this reason, and the exploratory nature of the research, differences that approached significance at the five percent probability level are discussed along with those that actually achieved significance.



As can be seen in Figure 1, the general pattern of the three dependence variables is to increase as the magnitude of the discount increases. Significantly higher savings were perceived at the 50, 40 and 30 percent discount levels than at the 10 percent level. The 50 percent level also produced significantly higher perceptions of savings than the 20 percent treatment condition.

Larger discounts also increased the general level of perceived offer accept, ability and value for money at the sales price. As Figure 1 shows, the one exception is at the 20 percent discount level where, relative to the 10 percent level, perceived offer acceptability and value for the money dropped slightly. In fact, the only group differences that approached or achieved significance on these measures were the difference in responses to the 20 percent and 50 percent-off treatments.

Since the information format treatment produced significant differences these responses were also explored through construction of simultaneous confidence intervals. Mean response patterns for the three dependent measures are presented in Figure 2. Information treatments have been arranged on the X-axis so that responses are generally increasing with movement along the axis. Also, since information treatments differed in non-quantitative aspects, their presentation on the X-axis represents categories of a factor rather than values of a continuous variable.



The difference between responses to information treatment one ("sales price" only) and three ("regular price and amount off") was accepted as significant for the perception of savings measure. As shown in Figure 2, treatment condition three accounted for a higher level of perceived savings. It also produced a nearly significantly greater perception of offer acceptability than did treatment one, and it accounted for the second largest mean difference on the perceived value for money measure. Further, the difference between mean responses to treatment four ("regular price and percent off") and three ("regular price and amount off") approached significance for perceived value for money with treatment three producing higher responses.

It is also useful to examine the pattern of responses as the character of information changes while the number of inputs in the advertisement remains constant. Figures 3 and 4 portray mean responses to two and three informational inputs respectively. As Figure 3 shows, the pattern of responses to two inputs is quite consistent across the three dependent measures. The highest responses are under conditions of "regular price and amount off" information. This was followed by responses to the "regular price and sales price" treatments while the "regular price and percent-off" treatment generated the lowest level of perceptions.





Figure 4, where responses to three informational inputs are portrayed, again shows a quite consistent pattern across the dependent measures. However, this pattern is not consistent with responses shown in Figure 3 where provision of "amount off" information as the second input led to the highest level of mean responses. That is, for three informational inputs the provisions of "amount off" information in conjunction with "regular price" and "sales price" tended to depress responses below those obtained when "percent off" information was presented with "regular price" and "sales price" information.


This experiment was designed to investigate how individuals respond to comparative price advertisements under various conditions. The conditions involved different original price levels, discount magnitudes and formats for presentation of price information. Response measures were perceived savings, value for the money at the sale price and offer acceptability.

Price level ($120 vs. $50) did not have a statistically significant effect on subjects' perceptions. However, there was a tendency to perceive more value for the money in the less expensive unit. This may suggest that the lower priced-lined unit more clearly matched the perceived needs of subjects than the higher priced model which contained more sophisticated features.

The magnitude of price discounts did produce significant differences in perceptions. Although all three dependent measures showed a very similar pattern of responses, the perceived savings variate accounted for most of the significant difference. This result closely resembles, and further confirms, the findings of Walton and Berkowitz (1980) regarding the effects of price discounts.

It is interesting to examine the character of savings perceptions across levels of price discounts. Significant differences occurred between the 10% and 30-50% levels and between the 20% level and 50% level. However, response differences between the 30%, 40% and 50% levels were not significant. This result might be attributable to subjects being suspicious of the larger discount offers. However, a separate question asked subjects to indicate the degree to which they suspected the truthfulness of the advertised prices. No significant differences in the perceived credibility of offers was detected across discount levels. Therefore, the argument that subjects' suspicions accounted for different responses to discount levels is not strongly supported. Conversely, it is interesting to observe that the threshold of significant differences in perceived savings occurs in the neighborhood of the 15% price reduction that many retailers believe must be achieved to attract consumers to a sale.

The method of presenting sale price information produced significant differences in subjects' perceptions. However, the nature of these differences appears to be complex, Presentation of only sales prices produced significantly lower perceptions of savings than did the presentation of regular price and dollar amount-off information. This result is consistent with Blair and Landon's (1979) finding that presenting only sale price led to significantly lower perceived savings than presenting reference price information with the sales price. It is also interesting to note that the average perceived savings response to all conditions employing reference price information (4.56) exceeded the average response to presenting sales price only (3.94). This shows further consistency with Blair and Landon's findings. Additionally, although the difference was not significant, format six that presented regular price, sale price and relative price reduction information led to greater perceptions of savings than when only sales price was presented. This mirrored Keiser and Krum's (1976) findings for these information presentation conditions.

Although price information formats did not significantly influence perceptions of offer acceptability and perceived value, the pattern of these responses closely reflected the perceived savings responses just described. Again, results are consistent with Barnes' (1975) findings where presentations using regular and sales prices produced significantly higher perceptions of value than expressions involving only sales price (with semantic phrase) or sales price with relative price reduction information.

When examining the effect of various price presentation formats while controlling for the amount of informational inputs, no significant differences occurred, although a rather inconsistent trend of responses was observed. For two informational inputs, presenting regular price and price reduction information in absolute terms yielded the highest response levels across all three variates, followed by presentation of regular and sales prices. Presenting regular prices and the discount expressed in percentage terms yielded the lowest response levels. However, for presentations involving three inputs the regular price, sale price, per-cent-off format generally yielded the highest response levels, while the regular price, percent-off, amount-off format was only marginally higher than the regular price, sale price, amount-off format. No clear response pattern emerges from this analysis especially when comparisons between two and three informational inputs are attempted. Therefore, results suggest that while a variety of presentations involving more information than just current sales price may favorably influence consumers' perceptions of savings, offer acceptability and value for money, the relative effectiveness of these inputs is presently unclear. Further research is needed across additional subject pools, product categories and price presentation formats.


Barnes, James G. (1975), "Factors Influencing Consumer Reaction to Retail Newspaper 'Sale' Advertising," in Edward Mazze (ed.) 1975 Combined Proceedings, Chicago: American Marketing Association, 471-477.

Blair, Edward A. and Landon, E. Laird, Jr. (1979), "The Effects of Reference Prices in Retail Advertisements," Working Paper, University of Houston.

Cronbach, L. J. (1951), "Coefficient Alpha and the Internal Structure of Tests," Psychometrika, 6, 297-334.

Engel, James, Blackwell, Roger D. and Kollat, David T. (1978), Consumer Behavior, Hinsdale, Illinois: The Dryden Press, 579.

Della Bitta, Albert J. and Monroe, Kent B. (1974), "The Influence of Adaptation Levels on Subjective Price Perceptions,'' in Scott Ward and Peter Wright (eds.) Advances in Consumer Research Vol. 1, Urbana, Illinois: Association for Consumer Research, 359-369.

Fry, Joseph N. and McDougall, Gordon H. (1974), "Consumer Appraisal of Retail Price Advertising," Journal of Marketing, 38, 64-74.

Green, Paul E. (1978), Analyzing Multivariate Data, Hinsdale, Illinois: The Dryden Press.

Hair, Joseph F., Jr., et. al. (1979), Multivariate Data Analysis, Tulsa, Oklahoma: Petroleum Publishing Co.

Jacoby, Jacob. (1978), "Consumer Research: A State of the Art Review," Journal of Marketing, 42, 87-96.

Jacoby, Jacob and Olson, Jerry C. (1977), "Consumer Response to Price: An Attitudinal, Information Processing Perspective," in Yoram Wind and Marshall Greenberg (eds.) Mewing Ahead with Attitude Research, Chicago: American Marketing Association, 73-86.

Keiser, Stephen K. and Krum, James R. (1976), "Consumer Perceptions of Retail Advertising with Overstated Price Savings," Journal of Retailing, 52, 27-36.

Monroe, Kent B. (1973), "Buyer's Subjective Perceptions of Price," Journal of Marketing Research, 10, 70-80.

Monroe, Kent B. (1977), "Objective and Subjective Contextual Influences on Price Perception," in Arch G. Woodside, et al. (eds.) Consumer and Industrial Buyer Behavior, New York: North-Holland, 287-296.

Morrison, Donald F. (1976), Multivariate Statistical Methods 2nd edition, New York: McGraw-Hill Book Co.

Nunnally, J. (1967), Psychometric Methods, New York McGraw-Hill Book Co., 226.

Nwokoye, Nonyelu G. (1975), "Subjective Judgments of Price: The Effects of Price Parameters on Adaptation Levels," in Edward Mazze (ed.), 1975 Combined Proceedings, Chicago: American Marketing Association, 545-548.

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

Sewall, Murphy and Goldstein, Michael H. (1979), "The Comparative Price Advertising Controversy: Consumer Perceptions of Catalog Showroom Reference Prices," Journal of Marketing, 43, 85-92.

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

Tatsuoka, Maurice M. (1971), Multivariate Analysis, New York: John Wiley & Sons.

Walton, John R. and Berkowitz, Eric N. (1980), "Contextual Influences on Consumer Price Responses: An Experimental Analysis," Journal of Marketing Research, forthcoming.

Wind, Yoram and Denny, Joseph. (1974), "Multivariate Analysis of Variance in Research on the Effectiveness of TV Commercials." Journal of Marketing Research, 11, 136-142.