In Search of the Economists' Consumer: the Effects of Product Information, Money, and Prices on Choice Behavior

John J. Wheatley, University of Washington
Richard F. Yalch, University of Washington
John S. Y. Chiu, University of Washington
ABSTRACT - Economists and behavioral researchers take two distinctly different approaches to assessing the effects of price and other factors on consumer choice behavior. The economists' view, which ignores consumers' perceptual biases related to price and product information, was tested by varying the conditions under which consumers made a choice between two similar brands. The results suggest that the parsimonious theory relied upon by most economists may be a misleading oversimplification in some product categories.
[ to cite ]:
John J. Wheatley, Richard F. Yalch, and John S. Y. Chiu (1980) ,"In Search of the Economists' Consumer: the Effects of Product Information, Money, and Prices on Choice Behavior", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 533-537.

Advances in Consumer Research Volume 7, 1980     Pages 533-537

IN SEARCH OF THE ECONOMISTS' CONSUMER: THE EFFECTS OF PRODUCT INFORMATION, MONEY, AND PRICES ON CHOICE BEHAVIOR

John J. Wheatley, University of Washington

Richard F. Yalch, University of Washington

John S. Y. Chiu, University of Washington

ABSTRACT -

Economists and behavioral researchers take two distinctly different approaches to assessing the effects of price and other factors on consumer choice behavior. The economists' view, which ignores consumers' perceptual biases related to price and product information, was tested by varying the conditions under which consumers made a choice between two similar brands. The results suggest that the parsimonious theory relied upon by most economists may be a misleading oversimplification in some product categories.

INTRODUCTION

The Economic Model of Consumer Choice Behavior

Economists, when addressing themselves to the question of consumer choice behavior, frequently take the position that, "on average," individuals are knowledgeable about the products they buy, have well-defined tastes and preferences, and are concerned with getting the most satisfaction possible from the limited amount of money they have to spend. This results in the postulated downward sloping demand curve, or in other words, a decrease in the purchases of any product as its price increases. All other things remaining the same, the argument goes, a price increase is equivalent to a decline in the income of the buyer, i.e., less of a product whose price has increased can be bought with the same amount of money. Other products, which may serve as substitutes for the product whose price has increased, also become less expensive, relatively speaking. Thus, what might be called the monetary constraint effect of higher prices serves as a deterrent to purchase behavior. Any deviation from this hypothesized choice model is generally assumed to be an unlikely situation (Stigler 1947).

The Behavioral View of the Role of Price in Consumer Choice Behavior

Behavioral researchers tend to believe that prices affect consumers in ways other than those included in the economists' model of consumer choice. One way is to serve as an aid to those unable to judge products. Even though higher priced products are not always of higher quality, this can be a useful heuristic (Andrews and Valenzi 1970). For example, Sproles (1977) recently found a positive price-quality relationship existed for 51% of the products he analyzed while only 14% had a negative relationship. Under these circumstances, assuming that a buyer wants, can afford, and does not pay disproportionately for higher quality, always choosing the higher priced of two brands may be more satisfactory than any other decision rule. Thus, higher priced products may be purchased rather than lower priced ones.

Prices are also thought to influence consumers' judgment by affecting their perceptions of product quality. In a large number of studies consumers have been shown to evaluate a higher priced product as being of higher quality than a lower priced one, even though they were actually of equivalent quality. Olson (1977), reviewing 24 of these studies of the price-quality perceptual bias, found them generally supportive of this proposition, but also noted many methodological problems which hindered interpretation of the results. Only two of the 24 studies were reported as involving "brand choice," rather than quality perceptions (Leavitt 1954; Tull, Boring and Gonsier 1964), and closer examination reveals that neither involved monetary transactions. They merely recorded paper and pencil indications of what the subjects would buy on a simulated shopping trip. Thus, given evidence that perceptions are only moderately predictive of attitudes (cf., Mazis, Ahtola and Klippel 1975), and that attitudes are only somewhat related to behavior (Wicker 1969), one must conclude that there is insufficient evidence to reject the economists' consumer choice model which ignores the price-quality perceptual bias.

In light of this observation, a research study was conducted to determine if consumers behave as postulated by the economic model or whether the price-quality perceptual bias is a significant factor in product choice. Three research questions addressing this issue are discussed below.

Research Questions

1. Will a price-quality perceptual bias affect actual behavior?  Many studies have demonstrated a perceptual bias when consumers are asked to evaluate products differing in prices, but none has sought to determine whether it is sufficient to alter actual product choice. If, in fact, it has little or no influence, then perhaps the economists are correct in concluding that it can safely be ignored. This question is to be answered by determining whether choices between equivalent products occur nonrandomly merely because one brand is identified as having a higher price than the other. Choice of the higher priced brand was encouraged by not requiring the consumers to sacrifice anything in order to select the higher priced brand.

2.  Will the introduction of price as a monetary constraint eliminate the behavioral effect of the price-quality perceptual bias? It can be argued that while the perceptual bias is capable of affecting behavior under conditions in which the subjects do not have to give up anything in order to exercise a preference for a more expensive brand, this preference effect might disappear in a more realistic setting involving sacrifices. This argument recognizes that prices may have two roles in determining the value received for a given expenditure. One is to determine the quantity of goods and services received so one tries to pay less in order to get more products; the other is to suggest something about the product's quality, and thus one is willing to pay more in order to get better quality. This question is to be addressed by introducing money into the choice process and observing whether it results in a significant change in choice behavior. Theoretically, from the economists' view, consumers should select the lowest priced product in order to realize the maximum satisfaction with their choice.

3. Will the combination of the monetary constraint factor and "perfect" information eliminate the behavioral effect of the perceptual bias?   It has been shown that one reason why individuals prefer higher priced products is because the higher price is thought to indicate the product's high quality on some attribute. If consumers have the opportunity to evaluate the products, this effect should be minimized. However, they may be influenced by the higher price into believing that there are hidden qualities not readily apparent on a one-trial test. Thus, the perceptual bias might persist. Economists and many public policy advocates believe if "expert" information is presented to consumers about the true quality of a product then these is no reason for any perceptual bias to continue. This is equivalent to "perfect" information and price's only effect should be to cause consumers to select the lower priced product as a way to maximize their satisfaction. To behave otherwise would violate the economists' model.

This question, which represents the most stringent test of consumer models ignoring the price-quality perceptual bias, is to be answered by comparing the brand choices of consumers in a "perfect" market situation with those predicted by the economists' consumer model. Thus, any choice of the higher priced brand under the condition in which individuals have reasons to know that it is of no higher quality based on expert judgment and their own personal assessment, and when they must sacrifice money in order to acquire the more expensively priced product, would substantiate the importance of psychological factors in consumer choices.

These three research questions were investigated in an experimental study of the effect of price and product information on consumer choices in different "markets." Each question was studied by creating a market situation corresponding to the conditions hypothesized to generate data to test the following null hypotheses:

H1:  In the absence of price as a monetary constraint and "perfect" information, consumers will not evidence a price-quality perceptual bias by selecting the brand labeled with the higher price at a greater than chance level.

H2:  When price serves as a monetary constraint, consumers will not select the lower priced product at a greater than chance level.

H3:  When price serves a monetary constraint and when consumers have "perfect" product information, consumers will not select the lower priced product at a greater than chance level.

METHOD

The research design utilized three treatment groups to test the behavioral effect of the price-quality perceptual bias under increasingly more "perfect" market conditions. Subjects were recruited from a university library cafeteria on the basis of their indicating they were cola users. Cola was selected as the test product because it is likely to be familiar to the subject population, it could be evaluated by the subjects, and pilot tests had established widespread belief in the existence of quality differences among brands. In addition, under chilled conditions (a normal drinking situation), few consumers could detect differences among brands, and thus cola quality seems to be a perception frequently biased by price information. Ultimately, 219 subjects were recruited and randomly assigned to one of three different market conditions.

After their cooperation was solicited, the participants were seated at a table upon which two six-packs of cola were displayed, wrapped to prevent identification. The respondents were asked to fill out a short questionnaire dealing with such matters as frequency and quantity of cola consumed, favorite brand, type of container usually purchased, and prices paid. These responses were later used to determine if the random assignment procedure was successful in getting three equally experienced groups.

Treatments

After responding to the questionnaire, each subject received a brief written message. Those assigned to the "perfect" product information group received information from the Food and Science Department of a prestigious private university, which a pilot test had determined to be an unbiased source of information on food matters. The expert information concerned similarities in the contents of cola drinks and the way in which they are manufactured as well as research evidence from a scientifically conducted experiment with a large sample of cola users which revealed that it was impossible for consumers to perceive brand differences in a blind taste test. Individuals assigned to the other two groups received only filler material.

All experimental subjects were next offered two glasses of cola marked Xt and Wb, two letter labels predetermined to be neutral in terms of their effect on product choices. The Xt brand was identified as having a retail price of $1.59 per six 12-ounce cans while Wb retailed for 994> per six pack. These prices were selected because, at the time, they represented the high and low prices being charged by local retailers. The two samples were served in a random fashion from bottles that were covered with opaque plastic to prevent identification of the actual brand used or the shape of the container. In actuality, the two brands were the same cola.

Since choice behavior corresponding to the three conditions could not easily be observed in the market place, a technique suggested by McConnell (1968) was used to simulate it. Rather than have subjects pay for the cola they selected, which might have resulted in a substantial loss of participants from the study, it was decided to manipulate the monetary constraint factor, i.e., the sacrifice one makes when one purchases a higher priced product, by taping 104 to each can of the less expensive brand to represent the actual price difference per can.

The subjects in the two groups selecting under a monetary constraint were then asked to choose which of the following combinations of Xt and Wb they would like to receive for their participation:

0 cans of Xt and 6 cans of Wb plus 604

1 can of Xt and 5 cans of Wb plus 504

2 cans of Xt and 4 cans of Wb plus 404

3 cans of Xt and 3 cans of Wb plus 304

4 cans of Xt and 2 cans of Wb plus 204

5 cans of Xt and 1 can of Wb plus 104

6 cans of Xt and 0 cams of Wb

Thus, two groups were able to trade off the selection of the higher priced brand with the option of receiving more money along with the lower priced brand. The subjects in the first group (no monetary constraint) did not have the option of receiving money with any of the alternatives presented to them. They selected only products with price information displayed on them, a procedure followed in most behavioral science studies of price-quality relationships (Olson 1977).

Finally, all participants were asked to indicate in their own words how they decided to make the choice they had just made, followed by three questions concerning respondent demographics and one asking what the subjects personally thought the purpose of the study was. The following summarizes the different experiences of the subjects based on their group assignment:

TABLE

It should be noted that price as a cue to product quality was present in all three groups, they all tasted the two products, and except for the manipulation of two factors, price as a monetary constraint and "perfect information," all three groups were handled in exactly the same fashion. A fourth group to complete the full factorial design, "perfect" information with no monetary constraint, was not included in the study because it did not represent a research question of any theoretical interest or practical significance since it is not typically encountered in the market place.

RESULTS

Analysis of the premanipulation questions revealed that there were no significant differences between the three groups in terms of experience with the product. The demographic characteristics of the three groups were also similar so the randomization procedure appeared to be successful.

Table 1 presents a summary of the behavioral choice results. If the subjects' selections were random, i.e., not influenced by the prices or experimental manipulations, the expected number of cans of both Xt and Wb would be equal or three each. The standard deviation of a uniform distribution over zero to six is (Hodges and Lehmann 1964):

EQUATION

This will be used to determine if the choices by any of the groups varied from that expected by chance. The behavioral choice data of the subjects are next used to answer the three research questions posed and this is followed by an analysis of the subjects' open-ended explanations for their behavior.

TABLE 1

NUMBER OF CANS SELECTED BY MARKET CONDITION

Analysis of Behavioral Data

1. Will a price-quality perceptual bias affect actual choice behavior?   In their pretaste questionnaire responses, most subjects indicated that they believed that there were perceivable quality differences in cola drinks (mean = 4.5 on 1 to 7 point scale with 7 labeled strongly agree). The choice results for Group I demonstrate that the reported belief in quality differences did not lead to more selections of the higher priced brand, Xt, than one would expect by chance despite the absence of a monetary sacrifice. The choice of 3.3 cans of the higher priced brand was only slightly greater than the 3.0 cans expected by chance, and this difference is not statistically significant using the t-test. The t-test, it should be noted, is appropriate here because while the individual choices are discrete the mean number of cans selected is a continuous variable (Snedecor and Cochran 1971, p. 95). Thus, the first null hypothesis cannot be rejected.

The results for the first group support the belief that a price-quality perceptual bias may not necessarily affect choice behavior. However, in this study, participants had the opportunity to compare directly the two products and this may have overcome most of the perceptual bias. This suggests that to a large extent, price-quality perceptual biases may be reduced by allowing consumers the opportunity to evaluate products before making their choices. Unfortunately, this is rarely done in retail stores for many of the packaged goods assumed to be associated with perceptual biases and few consumers appear to conduct this type of test on their own. This result also suggests that, in this study, any demand artifact, i.e., a desire to behave in accord with the experimenter's hypotheses, was minimal.

2. Will the introduction of price as a monetary constraint be sufficient to eliminate the behavioral effect of the price-quality perceptual bias?  Since the first group gave only slight evidence of being influenced by the price information on the packages, one would expect that most subjects would select the lower priced brand once they were able to receive money for doing this. However, this did not occur. The subjects in the second group selected the lower priced brand only slightly more frequently than those in the first group (3.3 versus 2.7) and this was not statistically significant (t = 1.5, df = 145, p > .05, one-tailed test) or significantly lower than that expected by chance. Thus, the second null hypothesis cannot be rejected and one must conclude that the imposition of price as a monetary constraint was not sufficient to cause subjects to select the lower priced brand.

3. Will the combination of the monetary constraint factor and "perfect" information eliminate the behavioral effect of the perceptual bias?   It was expected that when subjects could realize an actual money savings by selecting the lower priced brand and when they knew that there were no meaningful quality differences between the two brands, the overwhelming choice would be the lower priced brand. The results offer only partial support to this hypothesis. The 3.9 cans of the lower priced brand selected by the subjects in the third group was significantly lower than the number selected by the subjects in the first group, who lacked information and could not realize a real savings (t = 3.0, df = 145, p < .004, one-tailed test); it was also significantly lower than expected by chance (t = 3.1, df = 71, p < .004, one-tailed test). Further, an impressive 50% of the subjects in this group selected six cans of Wb and received the full 604 bonus. However, it remains to be explained why the other 50% did not do likewise.

It may be observed that the mean difference between the first and third groups, which is attributable to both the monetary constraint and product information, amounted to 1.2 cans while the difference between the first and second groups, which is attributable to the monetary constraint alone was .6 or half of the effect when the two were combined. In other words, assuming there was no significant interaction, the information effect on choice behavior was equal to the monetary constraint effect of price.

Analysis of the Open-Ended Responses

After the subjects made their choices, they were asked to explain in their own words how they decided to make the choice that they had just made. Analyses of these open-ended responses revealed that taste was the most frequently mentioned reason for the selection made in all three groups.

Over 70% of the subjects in Group I, where price did not serve as a monetary constraint, gave taste as their basis for choosing while not a single respondent mentioned price. On the other hand, while slightly over 50% of the subjects in Group III which was price constrained mentioned taste, almost 30% said that they were influenced by price considerations.

Apparently responding to the "expert" information that they had been furnished, a larger proportion of the subjects in Group III also said that they could not tell any difference between the two samples than was the case for Groups I and Ii which, of course, had not received this information.

A proportion of the subjects became so absorbed in the taste testing that their choice with respect to the number of cans of Xt and Wb was apparently influenced by a desire to continue with the process on their own. That is, between 20 and 25% of the subjects in each treatment group chose three cans of Xt and three cans of Wb, and a substantial number of them indicated that they did so because they wanted to test the differences between the two samples by themselves. This was empirically determined to have no effect on the statistical significance of the results.

Finally, all of the subjects were asked what they personally thought the purpose of the research study was. None of the answers indicated any awareness of the research hypotheses under investigation.

LIMITATIONS

It must be remembered that this was a laboratory experiment rather than a field experiment. This means that the results reported here should be viewed cautiously. For example, all of the consumers involved in this experiment were placed in a forced choice situation involving only one product. It seems very likely that in a real world setting, characteristics of buyers such as their income level would cause at least some persons not to choose to buy their favorite brand of cola at all if its price were to be increased.

As Shapiro (1975) has pointed out, persons with high levels of education, such as college students, have a tendency to rely less on price as a cue to product quality than those with lower educational achievement. This might explain the small behavioral effect of the price-quality perceptual bias in the first two groups. Care should be taken in extrapolating the results of experiments involving mostly student subjects, although in this case they do constitute a reasonably important segment of the market for colas.

Only two variables, price and product information, were utilized to influence choice behavior. While both the prices used and the information provided were realistic, most consumers are capable of utilizing additional information in making either judgments about products or purchase decisions. In the field, the significance of these factors might also be diminished or enhanced because of a tendency on the part of consumers to perceive the elements of the buying situation in a selective manner, disregarding or emphasizing other potentially useful bits of information.

CONCLUSIONS

The results of this study suggest that the economists' virtually total emphasis on the monetary constraint effect of price on purchase behavior is possibly excessive. The almost equally strong concern of consumer researchers with the effect of price and other variables upon perceptions exclusively seems inappropriate, too.

Product information turned out to have a significant effect on behavior. Price information, serving as a quality cue exclusively, had the effect of influencing behavior slightly in one direction, but when price served also as a monetary constraint behavior was reversed. Specifically, when price served only as an information cue as it did in Group I, choice behavior tended to be in the direction of selecting the higher priced alternative. On the other hand, when price served both as an information cue and a monetary constraint, as it did in Group II, choice behavior was influenced in the direction of selecting the lower priced alternative. The latter behavior was further strengthened by the presence of product information (Group III) but never reached the magnitude expected by the economic model.

Probably the most important findings were that the magnitude of the behavioral effect caused by product information was as large as the monetary effect of price, and that even in a "perfect" market individuals did not completely avoid selecting the higher priced brand. Since price is one of only two factors that economists usually take into account when explaining consumer choice, this raises important questions about the usefulness of conventional economic price theory as a universal guide to price setting by sellers. Economists, of course, acknowledge that choice or purchase behavior depends on many factors but these "other factors" tend to be viewed by them as unimportant enough so that they can be ignored. The evidence from this experiment suggests not that economic price theory is illogical or invalid but that it may be a dangerous oversimplification. In some circumstances, the "other factors" seem to matter a great deal.

It would seem appropriate to suggest that, in view of these results, attempts should be made to replicate the findings reported here with different products, price manipulations and other non-monetary factors that might be expected to affect price and income constrained consumer choice behavior. Extant research findings on factors that appear to influence consumer perceptions of products such as brand, retail outlet, and the physical characteristics of products, provide interesting hypotheses about the elements of what economists include under the rubric of "tastes" that could also affect such behavior. Notwithstanding the general impression that factors such as brand and store image do matter in a realistic setting, there is, surprisingly, virtually no confirming empirical evidence in the consumer behavior, marketing, or economics literature to support this viewpoint. As a consequence, very little is known of the magnitude of the effects that such factors may have on consumer choice in these circumstances.

On a more optimistic note, appropriate experimental designs may ultimately succeed in establishing the nature of a linkage between perception and constrained choice behavior. Rao (1972) has already suggested that consumers simply try to maximize perceived quality per unit price paid.

REFERENCES

Andrews, I. and Valenzi, E. (1970), "The Relationship between Price and Blind-Rated Quality for Margarines and Butters," Journal of Marketing Research, 7, 393-395.

Hodges, J. and Lehmann, E. (1964), Basic Concepts of Probability and Statistics, San Francisco: Holden Day.

Leavitt, Harold (1954), "A Note on Some Empirical Findings about the Meaning of Price," Journal of Business, 27, 205-210.

Mazis, Michael, Ahtola, Olli, and Klippel, R. Eugene (1975), "A Comparison of Four Multi-Attribute Models in the Prediction of Consumer Attitudes," Journal of Consumer Research, 2, 38-52.

McConnell, J. Douglas (1968), "The Price-Quality Relationships in an Experimental Setting," Journal of Marketing Research, 5, 300-304.

Olson, Jerry (1977), "Price as an Information Cue: Effects on Product Evaluations," in Consumer and Industrial Buying Behavior, eds., Arch Woodside, Jagdish Sheth and Peter Bennett, New York: Elsevier North-Holland.

Rao, Vithala (1972), "A Model for Brand Choice Under Price Quality Hypothesis," American Marketing Association Proceedings, No. 34, 366-371.

Shapiro, Benson (1973), "Price Reliance: Existence and Sources," Journal of Marketing Research, 10, 286-294.

Snedecor, George and Cochran, William (1971), Statistical Methods, 6th ed., Ames: The Iowa State University Press, 95.

Sproles, George (1977), "New Evidence on Price and Product Quality," Journal of Consumer Affairs, 11, 63-77.

Stigler, George (1947), "Notes on the History of the Giffen Paradox," Journal of Political Economy, 55, 152-156.

Tull, D., Boring, R. and Gonsier, M. (1964), "A Note on the Relationship of Price and Imputed Quality," Journal of Business, 37, 186-191.

Wicker, Allan (1969), "Attitudes Versus Actions: The Relationship of Overt and Behavioral Responses to Attitude Objects," Journal of Social Issues, 15, 41-78.

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