The Influence of Adaptation Levels on Subjective Price Perceptions

Albert J. Della Bitta, University of Rhode Island [Assistant Professor of Marketing.]
Kent B. Monroe, University of Massachusetts [Associate Professor of Marketing.]
[ to cite ]:
Albert J. Della Bitta and Kent B. Monroe (1974) ,"The Influence of Adaptation Levels on Subjective Price Perceptions", in NA - Advances in Consumer Research Volume 01, eds. Scott Ward and Peter Wright, Ann Abor, MI : Association for Consumer Research, Pages: 359-369.

Advances in Consumer Research Volume 1, 1974    Pages 359-369

THE INFLUENCE OF ADAPTATION LEVELS ON SUBJECTIVE PRICE PERCEPTIONS

Albert J. Della Bitta, University of Rhode Island [Assistant Professor of Marketing.]

Kent B. Monroe, University of Massachusetts [Associate Professor of Marketing.]

[The assistance of M. Venkatesan in the design and running of the experiments and Roy Ageloff in the programming for data analysis is gratefully acknowledged. Financial assistance was provided by the Research Council, University of Massachusetts, and the Consumer Research Institute.]

As previous research on consumer information processing has shown, the price cues are important, but the relative importance varies depending on the purchase context [Alexis, 1968; Bettman, 1970]. But how this information is processed by the buyer, or under what conditions the price cue is determinant of the purchase decision is not known. More important is the implied assumption of most of the germane research that the price cue is uni-dimensional. That is, the only price information processed by the buyer is the dollar quotation. However, recent research reveals that this assumption is basically naive, in that buyers' subjective perceptions of price may depend on, among other cues, the last price paid, the range of prices for similar alternatives, the lowest alternative price, the highest alternative price, conscious concern or awareness of prices, and the frame of reference for evaluating the alternative offers.

ADAPTATION-LEVEL THEORY

A psychological concept that appears to be fruitful for organizing research into the question of how buyers process shopping information is adaptation-level. According to adaptation-level theory [Helson, 1964] an individual's behavioral response to stimuli represents modes of adaptation to environmental and organismic forces. These forces are not random, and the pooled effect of three classes of cues, focal, contextual, and organic, determines the adjustment or adaptation level (AL) underlying behavior. Focal cues are the stimuli the individual is directly responding to, i.e., the immediate focus of attention. And contextual or background cues are all other stimuli in the behavioral situation providing the context within which the focal cues are operative. This adaptation process results in behavioral responses that are commonly expressed along a continuum ranging from rejection to acceptance with a neutral zone or point of indifference in the transitional region(s) between rejection and acceptance.

Perceptual judgment of a stimulus depends on the relationship between the physical value of that stimulus and the physical value of the current AL. It has been found that when stimuli outside the normal range are introduced AL shifts in the direction of the extreme values [Parducci, 1954]. In addition, the original stimuli tend to be rejudged as smaller, lighter, or shorter than they had been before the extension of the stimulus range by larger, heavier, or longer stimuli [Parducci, 1956]. And, an extension of the stimulus range produces a greater shift in judgment than does a restriction of the range [Parducci, 1954]. Each stimulus, whether singled out for judgment or merely presented as background pulls the AL toward its own value [Engel ar.d Parducci, 1961]. Finally, judgments are made as though the judged stimulus was being compared with the whole series of stimuli [Parducci and Hohle, 1957].

In a pricing context, adaptation-level theory suggests that price perception depends on the actual price and the individual's reference price or AL. Although the hypothesis of price serving as an adaptation level for price judgments has not been previously tested, evidence reviewed in Monroe [1973] supports the plausibility of this hypothesis. Essentially, previous price research has documented the existence of reference prices that affect price judgments, but these findings have been ancillary to the objectives of the reported research.

Experimental evidence of the applicability of adaptation-level theory to price perception was generated by Doob, et al. [1969]. They performed five field experiments where a new brand was introduced at a "low introductory" price in one set of stores, while in a matched set of stores the brand was introduced at the normal selling price. After a short period of time the low introductory price was raised to the normal selling price and sales were monitored in both sets of stores during the entire experimental period. The hypothesis derived from cognitive dissonance theory was that initially the low introductory price would produce more sales than the control condition, but that after the low price has been raised to the normal selling price sales would become higher for the control condition.

The results of these experiments are summarized in Table 1. In all experiments the direction of the hypothesis was substantiated, and in three of the product tests the hypothesis was significantly substantiated. In a combined test of the hypothesis over all products the results were significant in favor of the hypothesis (p.< .0002) .

Although one might hastily conclude that the sales results in the experimental conditions simply reflect a downward-sloping demand curve, the time series pattern of sales in the experiment belies such a simple explanation. While the control sales curves exhibit a generally monotonic growth in sales, the experimental sales curves drop and remain below the control sales curves after the price is changed. In fact, for the products and prices studied, aggregate sales volume was greater for the high-price condition.

The behavioral phenomena in the Doob study can be explained by dissonance as well as by adaptation-level theory: "when mouthwash is put on sale at $.25 customers ... may tend to think of the product in terms of $.25.... When, in subsequent weeks, the price increases to $.39, these customers will tend to see it as overpriced, and are not inclined to buy it at this much higher price." [Doob, 1969, p. 350].

TABLE 1

"EFFECT OF INITIAL SELLING PRICE ON SUBSEQUENT SALES"

METHOD

The data were generated as a part of the psychophysical experiments for determining price thresholds previously detailed in Monroe and Venkatesan [1969]. Briefly, subjects rated prices for eight products on a seven-point, absolute judgment rating scale. Each subject rated 14 prices for each product and judged relative expensiveness by indicating the degree the price was perceived to be a high or low price. Each product was judged by four groups of subjects separated into two sets of prices: high or low prices. Within each price set, one group of subjects was serially presented with prices ranging from the lowest price for that set to the highest price; the second group of subjects was serially presented with the prices ordered from the highest to the lowest price. Thus, for each product there were four price treatments: ascending low, descending low, ascending high, descending high. Details of the experiment are summarized in Table 2.

TABLE 2

ADAPTATION LEVEL EXPERIMENTAL DESIGN

HYPOTHESIS

According to adaptation-level theory, the perceptual judgment of any stimulus depends on the ratio of the physical value of that stimulus and the physical value of the subject's current adaptation level (AL). To the degree that the judged stimulus is greater than the current AL the stimulus is judged "high" (or higher than the standard to which it is being compared). The theory defines AL as the physical value of the stimulus which would be judged neutral (or equivalent to the standard). AL theory suggests that stimuli are ordered (judged) as members of a set of stimuli, and these judgments reflect the adaptation made by an individual to the set of stimuli. That is, the pooled effects of present and past stimulation establish a subjective standard with respect to which comparative judgments are made. Thus, all judgments are relative to the level of adaptation that has been established under a given set of conditions.

Adaptation-level theory yields the hypothesis that if a series of stimuli are presented for judgment in order of increasing magnitude, the stimuli in the series will tend to produce higher categories of judgment than if the series were presented in order of decreasing magnitude. This prediction results from the fact that for any stimulus value in the ascending series the mean of all the preceding stimuli is lower than the mean of the stimuli which would have preceded it if the series had been presented in descending order. Essentially, the subject compares each stimulus with the other stimuli he has seen. Because each successive stimulus is larger than the others he has seen when the order is ascending, he judges it with a "higher" category than he would use if it were smaller than the others (i.e., in a descending series).

The above outlined theory provides the testable hypothesis: extreme prices presented first in a series of price stimuli anchor price perceptions thereby pulling t AL toward the extreme end and resulting in depressed or accentuated price judgments. This hypothesis suggests that if subjects are initially exposed to relatively high or low prices for a given product, their frame of reference, or AL, is pulled toward the high or low price stimuli. This AL then serves to anchor subsequent judgments for other prices. Once the high AL has been established, then prices lower than the original anchoring prices will actually be perceived as "cheaper" or "less expensive" than in a low anchor context. On the other hand, once a low AT has been established, then prices higher than the original anchoring prices will be perceived as "more expensive than a high anchor context.

RESULTS

Determining the Stimulus-Response Function

Since the adaptation level only fixes a point or region on the stimulus continuum, prediction of all responses must be made using a stimulus-response function covering the whole continuum. The shape of this stimulus-response function depends on the stimuli being judged, the experimental task, the psychophysical method, the method of analyzing the data, as well as the adaptation level.

To determine the psychophysical function for the data generated a function derived by Helson [1964] was used. The function assumes that Xi is a stimulus value from a continuum of a measurable attribute, price, and that A is the value of adaptation level under the given conditions of stimulation. It is also assumed that a finite number of stimuli elicits judgments that are representative of the continuum [Helson, 1964, p. 191]. In addition, the following specific assumptions provided the foundation for deriving the function [Helson, 1964, p. 192]: (1) Judgment of a stimulus depends on its distance (Xi-A) from the AL, (2) Psychological distance of the Xi from AL is a function of the number of JND's in the interval (Xi-A). (JND refers to the psychophysical concept of just noticeable difference.) (3) The JND is a function of the judged stimulus and AL, i.e., AA = f(Xi,A). From these assumptions the desired psychophysical function reflected in the judgment of the stimulus as derived by Helson [1964]is:

(1)   J = K(Xi-A)/(Xi+A)

where J is the judgment on a linear numerical scale, K is a scale factor if (1) the judgmental categories are transformed into a linear numerical scale for computation purposes, and (2) K is the topmost value of the numerical scale. These conditions can be satisfied by transforming equation (1) into linear form [Helson, 1964, p. 196]:

(2)   (0-5K+J)/(1.5K-J) = Xi/A'+b

where A = A'-bA'.

Mean product-price profiles were computed for each treatment condition and were fitted by a least squares regression of the form

(3)   Yi = b + (1/A')Xi

where Yi = (0.5K+mij)/(1.5K-mij), and mij is the sample mean judgment rating for the ith price stimulus and jth product. Typical curves derived by this method are shown in Figures 1 and 2. As the figures indicate, the curves are negatively accelerated and not symmetric.

FIGURE 1

JUDGMENTS OF AFTER SHAVE PRICES SHOWING ANCHOR EFFECT OF LOW PRICES

FIGURE 2

JUDGMENTS OF AFTER SHAVE PRICES SHOWING ANCHOR EFFECT OF HIGH PRICES

Determining the AdaPtation Levels

Adaptation level prices were obtained for each product and each price treatment condition by using A = A'-bA'. The obtained adaptation levels are shown in Table 3. From the stated hypothesis, it is predicted that (1) the descending AL price will be higher than the ascending AL price for comparable product-price treatments, and (2) the high price ascending or descending AL will be higher than the low price ascending or descending AL, respectively. Looking at Table 3, there are 16 directional possibilities for each of these predictions. Comparing the ascending AL prices with the descending AL prices, 12 of the 16 comparisons satisfy this prediction, two are equal, and in two cases, hair spray and dress shoes (each in low-price treatments), the results are contrary to the prediction. Comparing low ascending or descending AL prices to high ascending or descending prices reveals 13 of the 16 comparisons conforming with the directional prediction.

TABLE 3

COMPUTED ADAPTATION LEVEL PRICES

TESTING THE HYPOTHESIS

Even though the product-price profiles obtained from the regression analysis indicate that the descending series Judgments are lower than the ascending series judgments for comparable product-price treatments, are these differences significant? Thus, to determine the significance of the findings, the data were analyzed using Morrison's [1967] "Profile Analysis for Two Independent Groups". Three tests were performed: (1) Are the population mean profiles similar, i.e., are the line segments of the ascending and descending series parallel? (2) Are the population profiles at the same level, i.e., are the differences between the line segments significant? (3) Are the population means of the ascending and descending series different, i.e., are the different mean responses in a series due to the different stimuli values? This is the equal treatment effects test for repeated measurements.

The only assumptions underlying these tests are that the variables have a multivariate normal distribution Eth a non-singular variance-covariance matrix ,. Through the use of Hotelling's T statistic the usual requirement of symmetry of z for repeated measures experimental designs is not necessary [Winer, 1962, p. 632]. The first test is a no groups-by-tests interaction hypothesis and must result in accepting the hypothesis of parallelism of the line segments before the remaining tests can be made. If the parallelism hypothesis is rejected then interactions are present and it no longer becomes meaningful to talk about the main effects of price. Essentially, the hypothesis is

(4)   H0: CM" = CMD

where M is a 1 X (p-1) vector of successive population mean values for a treatment series; p is the number of stimuli values in a treatment series; A is ascending and D is descending; and C is a (p-1) X p transformation matrix of the form

(5)   EQUATION

The test statistic is

(6)   EQUATION

where N refers to the number of subjects in the particular series and

(7)   T2 = [N"ND/N"+ND)][X"-XD]'C'(CSC')-1C[X"-XD]

where S is the sample variance-covariance matrix, the unbiased estimate of E, and Xi are (1 X p) sample mean profile vectors as described above.

If the parallelism hypothesis is accepted, then the second test investigating the hypothesis of equal series levels may be used. For the purposes of this research, the objective is to show significant differences in price judgments; therefore, the mean price judgments for an ascending price series should be greater than the mean judgments for the comparable descending price series:

(8)   J'M" > J'MD

where J' = [1,...,1] is the p-component vector with unity in each position.

The test statistic is:

(9)   t = J'[M"-MD]/[J'SJ(1/N" + 1/ND)]1/2

with NA+ND-2 degrees of freedom.

The third test investigates whether higher prices are judged relatively higher than lower prices, e.g., whether high prices are perceived as high. Although this hypothesis has little substantive relevance in this particular research, the tests were conducted and, as expected, showed highly significant different response means for the different price stimuli. Table 8 summarizes the results of these three tests.

TABLE 4

PROFILE ANALYSIS

Of the 16 parallel tests, 3 tests rejected the parallelism requirement, and 3 tests were not made because singular matrices resulted. The equal series levels and equal treatment effects tests were conducted for the remaining 10 comparisons. The mean judgments of the ascending price series were significantly higher than the mean Judgments of the descending price series in 6 of the 10 tests, and directionally supportive in 3 other tests. And as indicated above, the 10 equal treatment effects all showed significant differences.

When singular matrices result one strategy is to investigate the data, and by eliminating obvious means, totals, and differences produce nonsingular matrices. Because of the anchoring effect of the initial prices presented to the subjects, the only data suitable for elimination were the judgments for the first prices presented to the subjects in a price treatment series. Thus, new profile analyses were run on the six treatments not surviving the original parallelism test. The dress shoes low price treatment remained singular, and the other treatments passed the parallelism requirement. Table 5 summarizes the results of these additional tests.

TABLE 5

PROFILE ANALYSIS - SECOND TRY

Original analysis of the data confirmed the prediction from adaptationlevel theory: if a series of stimuli are presented for judgment in order of increasing magnitude, the stimuli in the series will tend to produce higher categories of judgment than if the series were presented in order of decreasing magnitude. The product-price mean profiles obtained from the regression analysis resulted in higher judgments for the ascending price series than the comparative descending price series. Computation of the adaptation-level prices also confirmed the research hypothesis in that 27 of the 32 comparisons were in the hypothesized direction.

Profile analysis of the original data resulted in five highly significant differences between ascending and descending price series, one mildly significant difference, three directionally supportive but insignificant differences, of the 10 comparisons surviving the parallelism condition of the analysis. For the six comparisons not meeting the parallelism criterion, examination of the data indicated some of the price stimuli were well outside the subJects' acceptable price ranges, and therefore, produced similar judgments in either ascending or descending price series. These similar judgments produced profiles that were convergent at one end of the price stimuli. Thus, for these situations the profile analysis was conducted on the judgments for the middle 7-10 prices. Of these six comparisons. four significantly confirmed the research hypotheses.

The results of the analyses are in close agreement to the known anchoring effects of physical stimuli that have been confirmed in numerous psychophysical studies. Although other pricing studies have observed effects due to subjects' frames of references, this study is the first to directly examine these contextual effects. Not only are the findings consistent with similar psychophysical studies, but they are also consistent with the Doob [1969] study examining the effect of introductory selling price on subsequent sales.

Confirming that exposure to price stimuli can affect subsequent judgments of relative expensiveness has several implications. One of the immediate implications concerns the finding that exposure initially to relatively high prices diminishes the degree that subsequently lower prices will be judged expensive. Such a finding is supportive of relatively high introductory prices for new products, and questions the policy of introducing a new product with a low introductory price or even give-away product samples. Adaptation-level theory suggests that raising price from a lower introductory price may lead to perceptions that the product is "more expensive" than it actually is. This perceptual phenomenon is especially significant when it is realized the same "regular price" will be Judged "less expensive" than it actually is if the introductory price is higher than the regular price. Although it is tempting to infer actual demand effects due to these perceptual phenomena, there is no knowledge as yet on the causal relation between perception of expensiveness and actual purchase behavior. Nevertheless, if a buyer judges a product as relatively more expensive than the reference price, such a Judgment might be a purchase deterrent.

Further, the contrasting effect produced by the AL phenomenon could influence the result of a planned sale. The sale price of a product if priced substantially below the regular price can lead to buyer disbelief and demand reduction. Such a price reduction may produce a contrast effect resulting in the buyer perceiving the "sale" product as different, and possibly inferior, to the regular product, even though the product has not changed.

From a research point-of-view, establishing that price judgments are influenced by the sequence that the stimuli are reacted to suggests the price researcher must exercise caution when designing a price study. Even if a random presentation of stimuli is used, the particular sequence followed may introduce some reference point bias.

REFERENCES

Alexis, M., Haines, G.H. & Simon, L. Consumer information processing: the case of women's clothing. Proceedings, Fall Conference, American Marketing Association, 1968, 197-205.

Bettman, J.F. Information processing models of consumer behavior. Journal of Marketing Research, August 1970, 7, 370-6.

Doob, A., Carlsmith, J.M., Freedman, J.L., Landauer, T.K., & Tom, S. Effect of initial selling price on subsequent sales. Journal of Personality and Social Psychology, 1969, 11, 345-50.

Engel, G., & Parducci, A. Value of background in the specification of the stimulus for judgment. American Journal of Psychology, December 1961, 74, 569-75.

Helson, H. Adaptation-level theory. New York: Harper & Row, 1964.

Monroe, K.B. Buyers' subJective perceptions of price. Journal of Marketing Research, February 1973, 10, 70-80.

Monroe, K.B., & Venkatesan, M. The concept of price limits and psychophysical measurement: a laboratory experiment. Proceedings, Fall Conference, American Marketing Association, 1969, 345-51.

Morrison, D.F. Multivariate statistical methods. New York: McGraw-Hill Book Co., 1967.

Parducci, A. Direction of shift in the judgment of single stimuli. Journal of Experimental Psychology, March 1956, 51, 169-78.

Parducci, A. Learning variables in the judgment of single stimuli. Journal of Experimental Psychology, July 1954, 48, 24-31.

Parducci, A. & Hohle, R. Restriction of range in the judgment of single stimuli. American Journal of Psychology, June 1957, 70, 272-5.

Winer, B.J. Statistical principles in experimental design. New York: McGraw-Hill, 1962.

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