Consumer Decision-Making: on the Importance of Price

Citation:

C. S. Craig and James F. Engel (1971) ,"Consumer Decision-Making: on the Importance of Price", in SV - Proceedings of the Second Annual Conference of the Association for Consumer Research, eds. David M. Gardner, College Park, MD : Association for Consumer Research, Pages: 243-255.

Proceedings of the Second Annual Conference of the Association for Consumer Research, 1971     Pages 243-255

CONSUMER DECISION-MAKING: ON THE IMPORTANCE OF PRICE

C. S. Craig, The Ohio State University

James F. Engel, The Ohio State University

W. Wayne Talarzyk, The Ohio State University

[The authors are indebted to Professors Philip Burgess and Lawrence Mayer of the Behavioral Sciences Laboratory for use of their facilities in the execution of this study.]

[C.S. Craig is Assistant Professor of Library Administration and Assistant Director of the Mechanized Information Center. James F. Engel is Professor of Marketing and W. Wayne Talarzyk is Assistant Professor of Marketing.]

An important aspect of consumer decision-making is the influence additional information has on the ultimate purchase decision. The central thrust of this study was to examine the consumer's reaction to price change, i.e., given a stated decision what sort of price change is required to alter that decision. The study focused on: (1) the relationship between a behavioral intention (willingness-to-buy) and the magnitude of the price change (price threshold) required to alter that intention, (2) the critical nature of price awareness as a mediator in brand switching behavior, (3) attitudinal discriminants of price threshold and price awareness, and (4) methodological refinements in price research. As a prelude to the research the conceptual framework which guided the study will now be articulated.

CONCEPTUAL FRAMEWORK

The Three Brand Willingness-to-buy/Price Model (Figure 1) presents the normative constraints of consumer decision-making as examined in the study. In its present state the model is best suited to branded non-durable (food and non-food) consumer products. Items should be of low unit value, have a distinct brand identity, and be frequently purchased.

The model conforms to the experimental methodology in that only one brand price is allowed to vary at a time. Given that constraint, the model begins with the subject's willingness-to-buy rank and evaluates each brand within a product category in decreasing willingness-to-buy order. In some instances the subject always buys the brand she is most willing-to-buy regardless of price conditions. Others may not always buy the same brand. Certain factors may cause this: (1) random vacillation, (2) a high degree of indifference, or (3) a change in some situational variable. A central situational determinant (in the model) is price change. For the subject to act the price change must be attended. Operationally this has been conceived in terms of a binary choice. If price is not attended, then irrespective of price conditions the subject buys brand one (subject to noise and extra-model determinants). Given an attended price change, the subject then determines whether the change is sufficient to cause a situational reordering of the willingness-to-buy hierarchy. Parenthetically, it should be added that a price change in a real world context is neither necessary nor sufficient for a change in buying behavior. However, within the confines of this normative model it is viewed as the primary determinant.

Actual purchase outcomes depend on the magnitude of the price change. Subjects are viewed as possessing price thresholds. Below the threshold subjects will remain brand loyal and at or beyond the threshold switching will occur. Mediators which affect the magnitude of an individual's thresholds are discussed in the following section.

Stages in the Decision Process

For discussion purposes the model has been divided into three stages (indicated by the capital Roman numerals to the left of Figure 1). Stage I addresses habitual versus non-habitual behavior. Although this phenomenon can be attributed to a number of causes, for analytical simplicity it has been relegated to salience of brand in the evaluative criteria. Reasons for brand salience could range from its use as the primary purchase cue to simplify the purchase process to distinct attributes it may possess relative to others.

FIGURE 1

THREE BRAND WILLINGNESS-TO-BUY/PRICE MODEL

Stage II reflects the salience of price in the over-all evaluative criteria. Subjects who do not check price are apt to buy brand one, but the probability associated with that outcome is not as high as the "always buy" in stage I.

Stage III of the model examines price thresholds in terms of their derived existence, not in terms of moderators that affect threshold magnitude. Salience of price is the aggregate descriptor for the individual components. Components would include: (1) attitude toward price and price related topics, (2) socio-economic characteristics, (3) behavior, and (4) perceived differences between alternatives. The relationship is shown in Figure 2.

FIGURE 2

MODERATORS WHICH AFFECT SALIENCE OF PRICE

The model (Figure 1) attempts to relate the variables under investigation. It purposely does not include such variables as coupons, specials, or in-store display since they are not manipulated. Alternatives such as postponed purchase or purchase at an alternative store were not modeled since these choices were not allowed in the experimental setting.

Communication of Price Information

In viewing the communication of price information the current research has adapted a model formulated by McGuire. [For a discussion of the communication process, see William J. McGuire, "Personality and Attitude Change: An Information-Processing Theory," in Psychological Foundations of Attitude, ed. by A. Greenwald, T. Brock, and T. Ostrom (New York: Academic Press, 1968), Pp. 171-96.] Figure 3 (see following page) depicts the four steps in the communication process examined in the study. A fifth, retention, is postulated by McGuire but is not considered here. In a market context consumers are exposed (exposure) to many price changes on any given shopping trip. A consumer may or may not receive (reception) the specific price change. The probability of reception could be increased through attention-getting devices such as package alteration and point-of-purchase display. Reception is also influenced by acceptance (i.e., attitude toward price and price related matters). The acceptance of the stimulus (price change) determines in part what action will be taken by the shopper. Thus, Figure 3 may be viewed stochastically with each step being necessary but not sufficient for the next.

FIGURE 3

STEPS IN THE COMMUNICATION OF PRICE INFORMATION

PRICE RESEARCH: BEHAVIORAL DIMENSIONS

Behavioral research conducted on price appears to fall into four basic paradigms: (1) price-quantity research (Applebaum and Spears, 1950; Hawkins, 1957; Greig, et al., 1958; Stout, 1969), (2) price-quality (Leavitt, 1954; Tull, et al., 1965; Olander, 1967; McConnell, 1968; Stafford and Enis, 1969, (3) price-consciousness research (Wells and LoSciuto, 1966; Gabor and Granger, 1961, 1964), and (4) price-sensitivity research (Pessemier, 1960, 1963; Pessemier et al., 1968; Abrams, 1964; Gabor, Granger and Sowter, 1970). The limitations of the existing literature can best be seen by considering it in light of the communication model. The typical price-quantity paradigm views consumers as faceless units of demand, since only the relationship between exposure and action is considered. Prices are changed and records kept of subsequent sales volume. Exposure and action take place in a natural environment. However, little or nothing is known about the organism making the purchase (reception and acceptance are ignored).

Studies following the price-quality paradigm are plagued with demand characteristics. [For a discussion of demand characteristics, see E. Aronson and J.M. Carlsmith, "Experimentation in Social Psychology," in The Handbook of Social Psychology, ed. by Gardner Lindzey and E. Aronson, II (2nd ed.; Reading, Mass.: Addison-Wesley, 1968), pp. 61-66.] Although acceptance is implicitly considered, they force reception and then look at action. Apart from the contrived communication the results are vitiated most by demand characteristics, i.e., the subjects tend to do what they perceive the experimenter wants and the situation dictates them to do.

Price-consciousness research has identified the critical variable of price awareness through observational studies and recall techniques. The major limitation has been that these studies have viewed the phenomenon in isolation. Price-sensitivity research, like price-quality research, views the relationship between forced reception and action. Subjects are informed of price changes and then asked their brand choices. The demand characteristics (i.e., subjects perceive they should switch brand choice and do so even when they would remain brand loyal in the field) and the forced reception confound the findings.

Given the problems associated with prior price research the current study attempts a resolution. The first part of this study in essence follows the price-sensitivity paradigm. With those results established, a new approach is tried to serve as a basis for comparison.

METHODOLOGY

The research was conducted using Columbus area housewives in a laboratory setting. Subjects were divided into groups (A and B). Data collected on group A subjects included: shopping behavior information, demographics, attitudinal information (regarding shopping and price) willingness-to-buy data, and data on experimentally induced brand switching in reaction to price changes. [See Exhibit A for a list of predictor and criterion variables.] The same data were collected on group B subjects except for brand switching behavior. Instead, subjects in group B were sent on a simulated shopping trip.

Data on willingness-to-buy (behavioral intention) were collected on the V-scale. [See Exhibit B (Rating Instructions) for willingness-to-buy induction. See Exhibit C for description of V-scale.] There were three product categories (catsup, cake mix and tooth paste) and three "national" brands in each. After all products (brands within product category) had been rated, subjects were again presented with the brands. This time the prices of the brands were systematically manipulated until a subject (group A only) altered brand choice.

The magnitude of the price change required to induce brand switching (price threshold) was recorded for all three conditions: (1) increase in the price of the brand the subject was most willing-to-buy, (2) decrease in the price of the brand the subject was second most willing-to-buy, and (3) decrease in the price of the brand the subject was third most willing-to-buy.

Group B was sent on a simulated shopping trip. Extreme care was taken not to sensitize this group of subjects to price. The experimental procedure followed for group A was altered. Data were collected for willingness-to-buy alone (no price threshold data were collected). A suitable guise was adopted to insure normal shopping behavior. Subjects were told that the experimenter was interested in the time it took them to shop for certain products. They were instructed to shop as they normally do, taking as little or as much time as usual. On the simulated shopping trip subjects were given real money and a shopping list. They were allowed to keep their purchases and the change .

RESULTS AND DISCUSSION

Price Changes and Behavioral Intentions

From the data in Part A, a relationship was found to exist between willingness-to-buy (a behavioral intention) collected on the V-scale and price threshold (magnitude of price change required to induce switching). Data on willingness-to-buy were found to be positively correlated with price threshold (see Table 1).

As the distance between brands on the V-scale increased so did the magnitude of the price change required to induce brand switching. Thus, the more a subject preferred (was willing-to-buy) a brand relative to the other brands, the greater the price change necessary to alter that preference.

TABLE 1

CANONICAL CORRELATIONS BETWEEN WILLINGNESS-TO-BUY AND PRICE THRESHOLD FOR ALL PRODUCT CATEGORIES

Attitude and Price Sensitivity

Knowledge of differential brand switching behavior in reaction to price changes is of little value to the marketer unless consumers possessing these characteristics can be identified. With that problem in mind the analysis focused on a means to identify these consumers.

Subjects were divided into two managerially relevant groups (high and low price sensitive or those who switched brand preference when a one cent difference existed and those who waited until the difference was greater than one cent). Fifteen attitudinal variables were used as predictors of group membership in a multiple discriminant analysis. These variables served as effective discriminators, correctly classifying (range across product categories was 78 to 82 percent correct) the subjects into their respective groups (see Table 2).

TABLE 2

CLASSIFICATION MATRICES SUMMARIZING RESULTS OF PRICE SENSITIVITY DISCRIMINANT ANALYSIS USING FIFTEEN ATTITUDINAL VARIABLES AS PREDICTORS FOR ALL PRODUCTS

Demographics and Price Sensitivity

Multiple discriminant analysis was also used with non-attitudinal variables. The same criterion variable was used with demographic and shopping behavior variables as predictors. Subjects were correctly classified (range 80 to 88 percent correct) into their respective groups (see Table 3).

Simulated Shopping Trip

The problems associated with the part A methodology (specifically, forced reception and demand characteristics) prompted part B of the study. Implicit in part A of the study (and prior price studies) was the forced reception of price change information. However, on a shopping trip the consumer may not be aware of price differentials and behave as if there were none. The crux of the matter then becomes, if a price has been changed to the threshold, will the consumer attend it? Part B of the study attempted to resolve this.

A separate group of subjects was sent on a simulated shopping trip. Without the subjects' knowledge the prices of specific brands were set at their thresholds. [Thresholds were determined from part A data. Price changes were set equal to the V-scale distances. Thus, if Brand A> Brand B and AB=3, it was assumed that an increase of three cents in the price of A would result in B>A. Part A data suggest that this would hold in at least 50 percent of the cases (range 50 to 75 percent).] If awareness were not a factor then these subjects could be expected to switch brand choice at least 50 percent of the time. There are, of course, variables other than awareness that could account for the differences. However, given the microcasm constructed awareness appears to be the most tenable.

On the simulated shopping trip subjects purchased the predicted brand (the brand they should have purchased after price manipulations were made) less than 25 percent of the time. The relationship is shown in Table 4.

In part A exposure and reception were both 100 percent. With forced reception at least 50 percent of the subjects altered brand choice when the price change was equal to their V-scale distance. When reception of price change was not assured less than 25 percent of the subjects altered brand choice when the price change was equal to the V-scale distance.

TABLE 3

CLASSIFICATION MATRICES SUMMARIZING THE RESULTS OF PRICE SENSITIVITY DISCRIMINANT ANALYSIS USING NINE DEMOGRAPHIC AND SHOPPING BEHAVIOR VARIABLES AS PREDICTORS FOR ALL PRODUCTS

Attitudes and Price Awareness

As with price sensitivity, it comforts the marketer little to know differential price awareness exists unless those consumers can be identified. To address this problem subjects were divided into two groups, price aware and non-price aware (check price--did not check price on the simulated shopping trip). Attitudinal variables served as discriminators between the two types of observed behavior in a multiple discriminant analysis. As shown in Table 5, subjects were correctly classified (range across product categories was 83 to 93 percent correct) into their respective groups.

TABLE 4

DETERIORATION IN COMMUNICATION OF PRICE INFORMATION FOR ALL PRODUCT CATEGORIES

IMPLICATIONS

Perhaps the positive relationship between willingness-to-buy and price threshold is not too surprising. However, the marketer embarking on a price promotion must be cognizant of differential price thresholds and their magnitude. If his brand is desired appreciably less than the alternatives then the magnitude of the price change required to induce switching may adversely affect profitability even if unit sales increase. By gathering willingness-to-buy data on his brand and competitive brands the marketer gains some indication of the probable effect of a price change. In instances where price promotion appears the least desirable strategy, dollars are better spent increasing the desirability of the brand before attempting price cuts. Thus, willingness-to-buy data can serve as a proxy variable for price threshold.

Price awareness appears to have a moderating effect on price sensitivity. Given a price change of a specific magnitude a certain percentage will alter their purchase accordingly, but if and only if they attend the price. The marketer attempting to increase sales through a price cut must, in addition to reducing prices, foster price awareness in his target market.

The results of the study suggest the utility of using the laboratory to take behavioral measures under controlled conditions and then using those as criterion variables. Any number of non-behavioral measures may then be taken and used as predictors in subsequent analyses. However, the microcasm constructed must be sufficiently realistic so that the behavior taking place is similar (hopefully identical) to real world action. Using this approach the marketer can eliminate much noise from the system and begin to understand the consumer.

The V-scale presents other methodological implications. It is a direct and concrete method for collecting data. Any confusion regarding the brands is avoided and the subject is able to visualize the relationship. Products are presented as they are normally encountered while shopping, thus adding a degree of realism. Closely related ta the V-scale is the stated dimension, willingness-to-buy. Willingness-to-buy seemed more consistent with brand switching behavior than the preference measured used during the initial phases of the pilot study. However, the reader is cautioned that this was not tested directly; thus, only anecdotal support is offered.

Finally, the analysis suggests the utility of using attitude measures as a means of developing customer profiles. A discriminant function would be developed and attitudinal variables used to determine the probable number of consumers who would switch brand choice when prices were changed slightly. Additionally, non-attitudinal variables (demographic and shopping behavior) could serve as discriminators (in a separately developed discriminant function) between high and low price sensitive consumers.

TABLE 5

CLASSIFICATION MATRICES SUMMARIZING RESULTS OF PRICE AWARENESS DISCRIMINANT ANALYSIS USING FIFTEEN ATTITUDINAL VARIABLES AS PREDICTORS FOR ALL PRODUCTS

CONCLUSIONS

Although much more research must be conducted on the influence additional information has on the ultimate purchase decision, certain conclusions relative to the consumer and price changes can be drawn at this time.

1. Willingness-to-buy (a behavioral intention) and price threshold are positively related. The more a subject is willing-to-buy a brand, the greater the magnitude of the price change required to change that decision.

2. Differential price awareness exists (across subjects and product categories) and can have a detrimental effect on price promotion.

3. Attitudinal and non-attitudinal variables served as good discriminators between groups of subjects who were high price sensitive and low price sensitive.

4. Attitudinal variables served as good discriminators between groups of subjects who were price aware and groups of subjects who were non-price aware.

5. Any study of marketing phenomenon ought to consider all stages in the communication process, particularly if an omitted stage will alter the ultimate action.

EXHIBIT A

LIST OF VARIABLES

EXHIBIT B

RATING INSTRUCTIONS

EXHIBIT C

DESCRIPTION OF THE V-SCALE

EXHIBIT D

ATTITUDE QUESTIONS

REFERENCES

Abrams, Jack. "A New Method for Testing Pricing Decisions," Journal of Marketing,-XXVIII (July, 1964), 6-9.

Applebaum, William, and Spears, Richard. "Controlled Experimentation in Marketing Research," Journal of Marketing, XIV (January, 1950), 505-170

Gabor, A., and Granger, C.W.J. "On the Price Consciousness of Consumers," Applied Statistics, X (November, 1961), 170-88.

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

Gabor, A., Granger, C.W.J., and Sowter, A.P. "Real and Hypothetical Shop Situation in Market Research," Journal of Marketing Research, VII (August, 1970), 355-59.

Grieg, W.S., Strand, F.O., and Larzelere, H.E. "Relative Retail Sales and Elasticity of Demand for Dehydrated Mashed Potato Products," Agricultural Economics Department Cooperative Extension Service, Michigan State University, East Lansing (July, 1958)

Hawkins, E.R. "Methods of Estimating Demand," Journal of Marketing, XXI (April, 1957), 428-38.

"How Much Do Customers Know About Prices?" Progressive Grocer, February, 1964, pp. 104-106.

Leavitt, H.J. "A Note on Some Experimental Findings About the Meaning of Price," Journal of Business, XXVII (July, 1954), 205-10.

McConnell, J. Douglas. "An Experimental Examination of the Price-Quality Relationship," Journal of Business, XLI (October, 1968), 439-44.

Olander, Folke. "The Influence of Price on the Consumer's Evaluation of Products and Purchases," New Developments in Pricing Strategy, University of Bradford Management Centre (January, 1967)

Pessemier, Edgar A. "An Experimental Method for Estimating Demand," Journal of Business, XXXIII (October, 1960), 373-83.

Pessemier, Edgar A., and Teach, Richard D. "Simulation, Scaling and Prediction Brand Purchasing Behavior," Proceedings of the American Marketing Association, Chicago, August, 1968, pp. 206-212.

Pessemier, Edgar A., Bruger, P., Teach, R., and Tigert, Douglas. "Using Laboratory Brand Preference Scales to Predict Consumer Brand Purchases," Lafayette, Indiana, 1968. (Mimeographed)

Stafford, James E., and Enis, Ben M. "The Price-Quality Relationship: An Extension," Journal of Marketing Research, VI (November, 1969) 5 456-58.

Stout, Roy G. "Developing Data to Estimate Price-quantity Relationships," Journal of Marketing, XXXII (April, 1969), 34-36.

Tull, D.S., Boring, R.A., and Gonsior, M.H. "A Note on the Relationship of Price and Imputed Quality," Journal of Business, XXVII (April, 1964), 186-91.

Wells, William D., and LoSciuto, Leonard A. "Direct Observation of Purchasing Behavior," Journal of Marketing Research, III (August, 1966), 226-33.

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