Consumer Recognition of Increases in Odd and Even Prices

ABSTRACT - Respondents were shown pictures of products with either odd or even prices. Two days later they saw the same items, but for half of the products, the prices had increased. Respondents were less likely to notice the price increases in the odd priced products than in the even price 4 products. Evidence indicated that this was due both to poorer memory for odd prices and to a bias toward judging that the odd prices were not the ones that had increased.

Citation:

Robert M. Schindler (1984) ,"Consumer Recognition of Increases in Odd and Even Prices", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 459-462.

Advances in Consumer Research Volume 11, 1984      Pages 459-462

CONSUMER RECOGNITION OF INCREASES IN ODD AND EVEN PRICES

Robert M. Schindler, Northwestern University

ABSTRACT -

Respondents were shown pictures of products with either odd or even prices. Two days later they saw the same items, but for half of the products, the prices had increased. Respondents were less likely to notice the price increases in the odd priced products than in the even price 4 products. Evidence indicated that this was due both to poorer memory for odd prices and to a bias toward judging that the odd prices were not the ones that had increased.

INTRODUCTION

Retailers' use of the technique of odd pricing (e.g., pricing an item at $9.99 rather than $10.00) is extremely common (Twedt 1965). However, despite the fact that odd ,rices are such a prominent feature of the retail price environment, the effect of this technique on the consumer has been mostly a matter of speculation.

One common hypothesis is that consumers will tend to pay attention to only the dollar part of a price and drop off the cents part. Thus, if consumers perceive a price such as $4.99 as $4.00, then by pricing a $5.00 item at $4.99, the retailer will lower the perceived price by one dollar at the cost of only one penny. Lambert (1975) tested this hypothesis by asking respondents to give an estimate of the monetary value of a set of products which was either even or odd priced, but found no consistent evidence to support the hypothesis. However, recently, Schindler and Wiman (1983) did find some evidence in favor of this "rounding down" hypothesis. They asked respondents to recall prices the respondents had seen two days earlier and found that respondents were more likely to underestimate odd prices than even prices. Although many respondents recalled a $9.99 price as $9.89 or $9.50 rather than as $9.00, the fact that they were less likely to recall a $10.00 price as "9 dollars and some cents" provides some support for the rounding down hypothesis.

Another common hypothesis is based on the fact that a number of studies, including at least three field studies (Ginzberg 1936 , Dalrymple and Haines 1970; Georgoff 1972), have failed to find any reliable effect of odd pricing on sales. This hypothesis holds that odd pricing actually has little effect on consumer behavior except that it has created familiar "price points." Many retailers would abandon odd pricing were it not that consumers tend to consider these price points as "correct" prices, and will be especially likely to notice, and presumably resist, prices which are raised above these price points (Nwokoye 1975: Whalen 19805.

Gabor and Granger's (1964) finding that consumer demand may peak at odd prices for a product which is normally odd-priced but not for a product which is usually not odd priced provides some support for the price point hypothesis. However, results of the Schindler and Wiman (1983) study raise questions about the price point hypothesis. In addition to finding a greater tendency to underestimate the odd prices, Schindler and Wiman found less accurate recall of odd prices than of even prices. If this accuracy difference is due entirely to a tendency to guess even prices when the actual price is forgotten, then it does not bear on the price point hypothesis. But if this recall accuracy difference is not due entirely to guessing strategies, then it must indicate that consumers are morn likely to forget odd prices than even ones. However, if a consumer has forgotten a price, then he/she cannot notice an increase in that price. Thus, greater forgetting of odd prices should lead to consumers being less likely to notice increases in those odd prices, not more likely to notice them. as suggested by the price point hypothesis.

The experiment reported here was designed to test whether consumers' tendency to notice a price increase in odd prices differed from their tendency to notice a price increase in even prices. Respondents were shown pictures of products which were given either odd or even prices. Two days later, the respondents were shown the same pictures with half of the even prices increased and the other half unchanged, and with half of the odd prices increased and the other half unchanged. The respondents were asked to indicate for each product whether or not the price had changed from -he price they had seen two days earlier. The price point hypothesis predicts greater awareness of price increases among odd prices than among even prices. On the other hand, if the lower recall accuracy of odd prices is due to a greater tendency to forget odd prices, then the prediction would be lower awareness of price increases among odd prices than among even prices.

METHOD

Sample

One hundred and twenty-seven undergraduate business students at Northeastern University served as respondents in this study. The experiment was conducted in six Introduction to Marketing classes during two class sessions, the second coming two days after the first. During the second session, data were collected from only those students who were present for the first session. This resulted in a sample size of 92 for the second session.

Design

The materials used in this study were slightly modified from those used by Schindler and Wiman (1983). Pictures of twenty products judged to be relevant to students were selected from a mail-order catalog and mounted on 7 l/2" by 5 l/2" cards. Using a table of random numbers, a standard order of presentation of the twenty cards was established and they were bound together in this order with looseleaf rings. The price of each product was lettered on a rectangular white sticker and affixed to the lower center portion of each card.

Three card sets were prepared, which were identical except for the prices affixed. If a product were priced at, say, $25.00 in one cart set, then that product would be priced at $24.99 in the second card set, and $24.98 in the third set. Odd prices with $.98 endings were included along with odd prices with $.99 endings in order to be able to demonstrate that any differences which were found between even prices and prices with $.99 endings were due to the "oddness" of the $.99 ending rather than to the one cent difference between the even and $.99 price.

The price ending ($.00, $.99, or $.98) assigned to a product in a particular card set was determined randomly for each of the twenty products, thus yielding a mixture of the three pricing conditions in each card set. Each set was given a cover card with the letter A, B, or C prominently displayed, and each respondent received only one of the three sets. Thus, each respondent was exposed to all three pricing conditions, but the particular products which exemplified each condition depended on the card set the respondent received.

The card sets used during the second session were identical to those used during the first session except that in each set, the p-ices were increased in ten of the twenty products. The increases ranged from 15% to 50: and were designed to (l) maintain the "roundness" of the even prices, and (2) cause the same number of digits to change in the odd and even priced forms. For example, a price of $400 would be increased to $500. Both are round numbers, and the increase involves a change of only one digit in both the even pricing conditions ($400.00 - $500.00) and the two odd pricing conditions ($399.99 - $499.99 and $399.98 $499.98). Note that the price increase never altered the type of price ending that a product had in a particular card set, and that each product had the same size price increase when it appeared with the even ending as when it appeared with the $.99 and $.98 endings. The result of this is that, over the three card sets, the size of the average price increase for each of the three price ending conditions was equal.

Also, the twenty products were randomly split into two groups of ten. For each card set, half of the respondents received it during the second session in a version where prices were increased in the first group of products. The other half of the respondents received it in a version where prices were increased in the second group of products. Table l illustrates, for one card set, the product prices and their increases for both versions of that card set used during the second session.

TABLE 1

THE PRODUCTS AND THEIR PRICES IN CARD SET A

Procedure

At the beginning of the first session, respondents were told they were participating in a study of "shopping behavior," and were each given a response sheet. The response sheet for the first session included a list of the twenty products in the order that they appeared in the card set.

To the right of each product name was a series of six purchase probabilities, beginning with 0% and proceeding at intervals of 20% up to 100%. Respondents were instructed to examine the product pictured on each card and to note the price indicated on the sticker. They were then told to assume that they were in the market for such a product and to indicate their willingness to buy the pictured product at the price displayed on the card by circling the percentage which indicated their purchase probability. Two of the six marketing classes run were given card set A, two were given card set B, and two were given card set C. Thus, each card set was seen by approximately the same number of respondents.

For the second session, each group of respondents received the same card set they had seen during the first session, except that the prices for half of the products had been increased. The respondents were told that "some of the prices" had increased and were asked to indicate whether each price was the "same" or "higher" by circling one of those two words next to the product name on the second session response sheet. Then they were asked to rate, on a 5-point scale ( l = "not sure" and 5 = "very sure") how sure they were of their judgments of whether each price had or had not increased. Finally, the respondents were asked once again to indicate their willingness to buy each of the pictured products under the same assumption and using the same scale as in the first session.

RESULTS

Each respondent's first session mean willingness to buy was computed for each of the three price ending conditions. An analysis of variance of these means indicated there were no statistically significant differences between these three pricing conditions (F = 0.50, p > .60). A similar analysis of variance on the second session willingness-to-buy measures also indicated that there were no statistically significant differences between the three pricing conditions (F = 1. 35, p > .20).

Table 2 shows how often respondents indicated that a price increased when it in fact did increase. When prices had odd endings, respondents were less likely to notice the price increase than when the prices had even endings (x.2 = 18.25, p < .001).

TABLE 2

PERCENT OF "INCREASE" AND "NO INCREASE" RESPONSES FOR PRICES WHICH DID INCREASE

However, as can be seen in Table 3, respondents were also less likely to indicate an increase occurred in an odd price than in an even price even for those odd and even prices which remained unchanged between the first and second sessions (X2 = 5.88, p = .05). Thus, there was evidence of a general bias toward indicating that an odd price had not increased.

TABLE 3

PERCENT OF "INCREASE" AND "NO INCREASE" RESPONSES FOR PRICES WHICH REMAINED UNCHANGED

To determine whether this response bias could be entirely responsible for the respondent's lower likelihood of recognizing an increase in an odd price, or whether poorer memory for the first session prices of odd priced items was also involved, a value of d' was calculated for the respondent's increase/no-increase judgments for each of the three pricing conditions. Developed in the context of signal detection theory (Coombs, Dawes, and Tversky 1970; McNicol 1972), d' is a measure of the sensitivity of the observer to a signal (in this case, the "signal" is the presence of a price increase) which is independent of the size of any response bias present. The calculations indicated that the d' for the prices with even endings was 1.108. For the prices with $.99 endings, d' =0.917, and for the prices with $.98 endings, d' = 0.915. A one-way analysis of variance on the d' scores for each respondent indicated that this effect of price ending on d' was statistically significant (F = 13.65, p < .005). Thus, this signal detection analysis indicates that the respondent's lower likelihood of recognizing an increase in an odd price than one in an even price is at least partially due to a greater tendency to forget the first session prices when they had odd endings.

The respondents' ratings of how sure they were of their increase/no-increase judgments also implicates the role of memory factors in the lower probability of recognizing an increase in an odd price. If the respondents were as likely to forget first session prices with even endings as those with odd endings but simply guessed differently for odd and even prices, then there is no obvious reason why they should differ in their confidence in their increase/ no-increase judgments for odd and even prices. However, if memory for first session odd prices was poorer than that for first session even prices, then respondents would have to more often take guesses as to whether odd prices increased or not, and thus would be expected to have less confidence in their increase/no-increase judgements for odd prices. The results support the latter alternative. The mean sureness rating for increase/no-increase Judgments of even prices was 3.79, the mean sureness for prices with 9.99 endings was 3.75, and the mean sureness for the prices with $.98 endings was 3.58. A one-way analysis of variance indicated that the effect of price ending on sureness was statistically significant (F = 5.97, p < .005).

DISCUSSION

The results of this study indicate that, contrary to the prediction of the price point hypothesis, consumers are less likely to notice an increase in an odd price than in an even price. Further, these results provide evidence that consumers are more likely to forget an odd price than the corresponding even price over a two-day period, and that this poorer memory for odd prices plays a role in the lower tendency to recognize that they may have increased. However, since the respondents in this study were dealing with pictures of products and not making actual purchases, these results must be regarded as tentative and subject to confirmation in a more realistic setting.

The artificial nature of the task may also have been responsible for the failure to find any effect of price ending on a willingness-to-buy measure. Other recent laboratory tests of the effects of pricing variables on a willingness-to-buy measure have also found either no effects (e.g. Schindler and Wiman 1983) or only limited effects (e.g. Della Bitta, Monroe, and McGinnis 1981). This underlines the necessity of expanding this laboratory research into situations where the researcher can be more confident of observing realistic levels of consumer concern about price information.

Future research might also be devoted to the question of what it is about odd prices which could account for the findings that they are more difficult to remember. One possibility is that odd prices may be encoded by the consumer's perceptual system to form an internal representation consisting of two or more chunks (Miller 1956), while even prices may be encoded into an internal form consisting of only one chunk (Jacoby and Olson 1977). For example, $19.99 may be encoded into the chunk "19" and the chunk "99", while $20.00 may be encoded into the single chunk "20". If both $19.99 and $20.00 received the same amount of cognitive resources (e.g. rehearsal time, space in memory, etc.), then the probability of at least one of the two chunks of the representation of $19.99 being unretrievable would be greater than the probability of losing the one chunk of the representation of $20.00. If future research supports this possibility, it would indicate that consumers would be less likely to recognize increases in any non-round number, not only ones which are a few pennies below a round number .

An additional area for future research involves the finding of the respondents' bias toward judging that an odd price is one that has not increased. It may be the case that consumers associate odd prices with discounts and low prices, and such associations may prove to be another psychological mechanism of an effect of odd pricing on consumer behavior.

Thus, the present results add to Schindler and Wiman's (1983) evidence for the rounding down of odd prices evidence that odd pricing provides a second benefit to the seller: lower consumer awareness of a price increase. However, this effect would probably also work in the other direction and make the consumer less likely to notice a decrease" an odd price. For the retailers this suggests (i) that odd pricing may be especially useful in advance of expected price increases, and (2) that inclusion of a reference price is especially important when communicating to the consumer that an odd price has been lowered.

REFERENCES

Coombs, Clyde H., Dawes, Robyn M., and Tversky, Amos (1970), Mathematical Psychology: An Elementary Introduction, Englewood Cliffs, New Jersey: Prentice-Hall.

Dalrymple, Douglas J. and Haines, George H., Jr. (1970), ".; Study of the Predictive Ability of Market Period Demand-Supply Relations for a Firm Selling Fashion Products," Applied Economics, 1(4), 277-285.

Della Bitta, Albert J., Monroe, Kent, B., and McGinnis, John M. (1981), "Consumer Perceptions of Comparative Price Advertisements," Journal of Marketing Research, 18, 416-427.

Gabor, Andre and Granger, C.W.J. (1964), "Price Sensitivity of the Consumer," Journal of Advertising Research, 4, 40-44.

Georgoff, David M. (1972), Odd-Even Price Endings, East Lansing: Michigan State University.

Ginzberg, Eli (1936), "Customary Prices," American Economic Review, 26, 296.

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

Lambert, Zarrel V. (1975), "Perceived Prices as Related to Odd and Even Price Endings," Journal of Retailing, 51,13- 78.

McNicol, D. (1972), A Primer of Signal Detection Theory, London: Allen and Unwin Ltd.

Miller, George A. (1956), "The Magic Number Seven, Plus or Minus Two," Psychological Review, 63, 81-97.

Nwokoye, Nonyelu G. (1975), "An Experimental Study of the Relationship Between Responses to Price Changes and the Price Level for Shoes," in Advances in Consumer Research, Vol. II, M.J. Schlinger, (ed.), Ann Arbor: Association for Consumer Research. 693-703.

Schindler, Robert M. and Wiman, Alan R. (1983), "Consumer Recall of Odd and Even Prices," Working Paper 83-10, College of Business Administration, Northeastern University, Boston, Massachusetts.

Twedt, Dik W. (1965), "Does the '9 Fixation' in Retail Pricing Really Promote Sales?" Journal of Marketing, 29, 54-55.

Whalen, Bernard F. (1980), "Strategic Mix of Odd, Even Prices Can Lead to Increased Retail Profits," Marketing News, (March 7), 24.

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