The Impact of Direction-Of-Comparison on the Formation of Preference

ABSTRACT - Two studies explore the direction-of-comparison effect, the tendency of people to neglect features that are unique to the referent of comparison and to highlight features that are unique to the subject of comparison. The present research finds boundary conditions for these effects. In the first study, we manipulated subjects' processing goal at the time of processing information about two brands of cereal. Further, we studied the direction-of-comparison effect in a stimulus-based judgment context rather than the conventional memory-based judgment context. Results show that the direction-of-comparison effect does not occur under impression formation conditions and is attenuated even under memorization conditions. This result could be due to the stimulus-based judgment context where subjects are free to re-evaluate information at the time of preference formation. In study 2, the subject and the referent of comparison had unique positive as well as negative features. Results reveal that the direction-of-comparison effect does not extend to these situations and that the processing goal does not affect preference. Implications for marketing are discussed.

Citation:

Gita V. Johar and Elizabeth H. Creyer (1993) ,"The Impact of Direction-Of-Comparison on the Formation of Preference", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 284-287.

Advances in Consumer Research Volume 20, 1993      Pages 284-287

THE IMPACT OF DIRECTION-OF-COMPARISON ON THE FORMATION OF PREFERENCE

Gita V. Johar, Columbia University

Elizabeth H. Creyer, New York University

ABSTRACT -

Two studies explore the direction-of-comparison effect, the tendency of people to neglect features that are unique to the referent of comparison and to highlight features that are unique to the subject of comparison. The present research finds boundary conditions for these effects. In the first study, we manipulated subjects' processing goal at the time of processing information about two brands of cereal. Further, we studied the direction-of-comparison effect in a stimulus-based judgment context rather than the conventional memory-based judgment context. Results show that the direction-of-comparison effect does not occur under impression formation conditions and is attenuated even under memorization conditions. This result could be due to the stimulus-based judgment context where subjects are free to re-evaluate information at the time of preference formation. In study 2, the subject and the referent of comparison had unique positive as well as negative features. Results reveal that the direction-of-comparison effect does not extend to these situations and that the processing goal does not affect preference. Implications for marketing are discussed.

Understanding the process of comparison is necessary in order to explain many different types of judgments and choices. We often compare, people, places, products, and a variety of other things with other known items. In fact, as Houston, Sherman, and Baker (1989) note, before we can categorize or evaluate a new item, we must judge its similarity and difference to more familiar items. In particular, comparison processes are an important element of the consumer decision making process. However, prior research in the realm of consumer decision making has tended to neglect the process of comparison between pairs of options.

Contrary to common sense, the process of comparison is not a straightforward matter in which the features of one item are simply compared with the features of the other item (Sanbonmatsu, Kardes, and Gibson, 1991). Tversky (1977) has proposed a feature matching model of the comparison process. According to his model, during the comparison process one object serves as the subject of comparison while the other object serves as the target, or referent, of comparison. Asymmetries in judgments of similarities will occur, depending on the direction-of-comparison. That is, differences in judgments of similarity may result depending on which object serves as the subject of comparison.

For example, consider the following example. You have been very happy with your computer at work. In fact, you would like to by the same computer for your home but the price is very high. During the search process, your office computer will serve as the subject of comparison. The new computer you consider purchasing, called the referent, will then be reviewed and evaluated in reference to your office computer. Consequently, the unique features of your office computer will be highlighted while the unique features of the referent computer are not likely to be noticed during the comparison process.

Hence, the object which serves as the subject of comparison can significantly influence the formation of preference. Numerous studies (e.g., Payne, 1982; Slovic & Lichtenstein, 1983) suggest that consumers rarely have well-articulated preferences. Rather, research suggests that preferences are often not simply recalled from memory but are instead, contingent on the decision context (Payne, 1976, 1982; Fischhoff, Slovic, & Lichtenstein, 1980). Consequently, it is especially important for consumer behavior researchers to understand the circumstances to which the direction-of-comparison effect will generalize. Is the direction-of-comparison effect likely to occur in the aisles of supermarkets? If so, under what processing circumstances is the effect more likely to occur?

Hence, the purpose of the research is to provide insight into the generalizability of the direction-of-comparison effect. First, we review the research which has examined this phenomenon. Then, two experiments which extend prior research are described and their results discussed. Finally, the implications of this research are identified and directions for future research are presented.

THEORETICAL BACKGROUND

Interest in the process of comparison, the manner in which the similarities and differences between two objects are identified, has recently increased (Dhar and Simonson 1992; Houston, Sherman, and Baker 1989; Sanbonmatsu, Kardes, and Gibson 1991). An understanding of comparison processes is necessary to fully explain many judgment and choice behaviors. For example, comparison is an essential element of the categorization process. Marketers spend a substantial amount of time and money on product positioning which, at its most basic level, involves a consumer's judgment of a product's category membership (Cohen and Basu 1987). As Cohen and Basu (1987) note, the outcome of the categorization process "is not only a particular identification of a product, but the increased salience of information relevant to that category (and the corresponding suppression of information relevant to other categories) together with the category-based inferences that result (p.455)." Consumers may compare a target brand to an exemplar of the product category during the initial categorization process that precedes more detailed processing of the brand (Fiske and Pavelchak 1986). According to the direction-of-comparison effect, the result of the categorization process could well depend on whether the target brand is the subject or the referent of comparison.

Comparison processes also play a fundamental role in the formation of preference. Tversky (1977) has characterized an object as a set of concrete and abstract features. For example, a consumer's total base of knowledge about a specific brand of peanut butter may include the following features: high quality, creamy, tasty, comes in a jar with a blue label, high calories, and no cholesterol. However, depending on the choice context, only a limited "working" list of features may be recalled. The list of features which are recalled may be significantly influenced by the features of the object of comparison.

Sanbonmatsu, Kardes, and Gibson (1991) note that during the comparison process, the object which serves as the subject of comparison, rather than as the referent of comparison, is critically important because people are attuned to the subjects' features (p. 132). This is because people often neglect features which are unique to referent, instead comparing only features which are present in the subject of comparison. On the other hand, features which are unique to the subject tend to be highlighted by the comparison process (Sanbonmatsu, Kardes, and Gibson 1991). This phenomenon is known as the direction-of-comparison effect. Consequently, preference for a given object is not determined solely by the bundle of attributes that define that object; preference is relative to the object to which it is compared.

Houston, Sherman, and Baker (1989) conducted a series of studies which explored this phenomenon. Respondents were presented with two different objects within a variety of categories. After reading descriptions of the different objects, respondents indicated which item they most preferred. Note that this preference measure was based on recall of the features - the actual features used to described the pair of objects was not available. Houston and his colleagues correctly assumed that the object which was most recently observed, and therefore better remembered, was the subject of comparison. They found that the subject is preferred when it possessed unique positive features. However, when the subject possessed unique negative features, the referent is preferred, even when the referent possessed an equal number of negative features.

Sanbonmatsu and his colleagues (1991) extended this research by examining whether the availability of an overall evaluation of the objects diminished the direction-of-comparison effect. They presented respondents with booklets describing two set of pens and two sets of automobiles. In one set of conditions, the objects shared positive features but had unique negative features. In another set of conditions, the objects shared negative features but had unique positive features, After either memorizing or forming an impression of each of the objects, the booklets were removed. A surprise preference measure was then administered. As expected, the direction-of-comparison effect was diminished when an overall evaluation, or impression, of the objects was formed. The availability of overall evaluations for each pair of objects reduced the need to perform a feature-based comparison and consequently, the unique features of the subject were not highlighted. The results also show that preference for the subject of comparison was significantly greater when it possessed unique positive features than when it possessed unique negative features. Thus, the notion that the features of the subject of comparison, compared to the features of the referent of comparison, has a stronger impact on the formation of preference is supported.

As discussed above, there is strong empirical evidence that the preference formation process can be influenced by determining which object is the subject of comparison and which object is the referent of comparison. However, two key questions for consumer researchers remain unanswered. The first question is C will the direction-of-comparison effect still occur when the judgment is stimulus-based, rather than memory-based? Many choice decisions occur in a context in which information about the objects is available. For example, consumers can compare the ingredients and other nutritional information of two competing brands by simply reading the package labels. While this scenario has been recently examined by Dhar and Simonson (1992), these researchers did not explicitly compare the condition where information was present at the time of making judgments and the condition where the information was not present.

A related issue is whether differences in processing goal (i.e., impression formation vs. memorization; see Sanbonmatsu et al 1991) will matter when the information used to describe the objects of comparison is available at the time of choice. Prior research suggests that consumers under high levels of involvement have an evaluation goal in place and hence process information in greater detail (Mitchell, 1983). This processing goal results in the formation of an overall evaluation based on a consideration of all product attributes, which may then be available at the time of choice (Carlston, 1980; Kardes, 1986; Lingle & Ostrom, 1979). In other words, if a consumer's goal is to form an impression of the object, then an evaluation is made at the time of processing. When these consumers later choose a brand, this overall evaluation influences the choice rather than the individual brand attributes. However, will this evaluation be used when product features are readily available?

Past research has focused on situations where (a) only unique features of both the subject and the referent of comparison are presented (cf. Dhar and Simonson 1992); or (b)the subject and the referent share features of the same valence and have unique features that are differently valenced (cf. Sanbonmatsu et al 1991). Hence, another issue that needs to be addressed is whether the features of the subject of comparison have a greater impact on preference formation when both objects have both unique positive and negative features. That is, for many product categories, different brands have different strengths and weaknesses. For example, one brand of computer may come equipped with an internal modem while lack a high speed processor while another brand of computer may have a high resolution monitor but lack a math co-processor.

In the following two studies, we examine the boundaries of the direction-of-comparison effect. In the first experiment, we determine whether processing goal, that is whether the subjects memorize the product attributes or form an overall impression of the objects, will results in differences in the strength of the direction-of-comparison effect when the features used to define the objects are available at the time of choice. In the second experiment we explore the comparison process when the objects of comparison have both unique positive and unique negative features.

EXPERIMENT 1

Method

Subjects. Forty eight students enrolled in the introductory marketing course at private, northeastern university, whose participation partially fulfilled a course requirement, served as subjects. Subjects were run in two groups of approximately equal size.

Stimuli. Descriptions of two equivalent breakfast cereals, each described by eight features were developed. Four negative features and four positive features described each brand. The cereals shared the four negative features; the four positive features were unique to each brand. These descriptions are presented below.

Brand Fir:

75 milligrams of sodium (low)

0 protein (low)

160 calories (high)

1 gram of fat (low)

8 grams of complex carbohydrates (low)

5000 I.U. Vitamin A (high)

18 milligrams of iron (high)

1 gram of dietary fiber (low)

 

Brand Sec:

0 milligrams of cholesterol (low)

0 protein (low)

160 calories (high)

3 grams of sugar (low)

8 grams of complex carbohydrates (low)

60 milligrams Vitamin C (high)

400 I.U. magnesium (high)

1 gram of dietary fiber (low)

Pretesting established that the four unique positive features of one brand were equal in desirability to the four unique features of the other brand.

Procedure. The experiment was described as a problem to learn more about consumer behavior. Booklets containing the instructions and stimuli were distributed to the subjects. In the impression formation condition, subjects were asked to form an overall evaluation of each of the brands. In the memorization condition, subjects were simply asked to memorize the brand's features. Subjects were given 1 minute to either form an impression or memorize the features of the first brand. After 1 minute, subjects were told to turn the page and repeat this task for the second brand. Then after another minute, a surprise preference measure was administered.

Preference was assessed on a 12 point scale anchored by "I strongly prefer (referent brand)" (1), and "I strongly prefer (subject brand)". The higher the value, the greater the preference for the most recently presented brand which was assumed to be the subject of the comparison (i.e., the focal option). Designating the second brand as the subject of comparison follows the procedure used by both Agostinelli, Sherman, Fazio, and Hearst (1986) and Houston, et al. (1989). Since, subjects are unaware of the purpose of the study, they are likely to be focused on the most recent item of the pair during the judgment process. The descriptions of both brands of cereals were provided subjects below this measure. The order of presentation of the brands was counterbalanced across conditions.

Results

As hypothesized, results reveal that the direction of comparison effect is attenuated when subjects have an impression formation goal compared to a memorization goal, at the time of processing information about the brand (F(1,46) = 3.12, p < .10. However, the means reveal that under both conditions, the subject of the comparison (i.e., the second brand presented), is not preferred much greater than the scale midpoint (6.5) that would be suggested by indifference between the two brands (memorization mean preference = 6.89; impression formation preference = 5.60).

These findings imply that the direction of comparison effect holds only when the judgment is memory based and that the effect does not carry over to stimulus based judgments. Thus, the advantage of the subject of the comparison diminishes when both brands can be re-evaluated at the time of forming preferences. The second experiment was designed to test the boundaries of the direction-of-comparison effect when the subject and the referent of comparison have unique positive as well as negative features, in memory-based judgment situations.

EXPERIMENT 2

Method

Subjects. Forty six students enrolled in the introductory marketing course at private, northeastern university, whose participation partially fulfilled a course requirement, served as subjects. Subjects were run in two groups approximately the same size.

Stimuli. Descriptions of two breakfast cereals, each described by eight features were developed. Four negative features and four positive features described each brand. However, unlike study 1, the cereals shared two positive and two negative features. Of the four features unique to each brand, two were positive and two were negative. Thus, unlike Sanbonmatsu et al (1991), the two brands shared both positive and negative features and had unique features that were positive and negative. Pretesting established that the unique features of one brand were equal in desirability to the unique features of the other brand.

Procedure. The procedure used in experiment 1 was repeated in experiment 2. Booklets containing the instructions and stimuli were distributed to the subjects. In the impression formation condition, subjects were asked to form an overall evaluation of each of the brands whereas in the memorization condition, subjects were simply asked to memorize the brand's features. After either forming impression of the brands or memorizing their features, subjects responded to a surprise preference measure. Unlike experiment 1, the descriptions of the objects were not available at the time of judgment. The order of presentation of the brands was counterbalanced across subjects.

Results

Results show that the direction-of-comparison effect does not extend to situations where the unique features of the subject brand are both positive and negative. In both the impression formation and the memorization conditions the mean preference was close to the scale midpoint (impression formation mean = 6.94; memorization mean = 6.10). Further, the processing goal did not significantly affect preference (F(1,44) = 0.91, p N.S.).

DISCUSSION

The findings of the two experiments reported here extend past research on the direction-of-comparison effect. Specifically, the results of Study 1 suggest that this effect does not extend to situations where the brand information is present at the time of judgment, regardless of the processing goal at the time of information processing. An impression formation goal further attenuates the effect. However, it should be noted that we have no way of knowing for sure whether subjects actually used the information which was available at the time of judgment. Perhaps, judgments were based on initial impressions of the brands. Study 2 suggests that this effect will occur mainly when the unique information from the subject of comparison is sufficiently diagnostic. When the subject of comparison has both positive as well as negative unique features, consumers will tend to turn to the unique features of the referent brand for resolution. Hence it appears that the direction-of-comparison effect is not robust to stimulus-based judgments or to situations where the subject of comparison has both positive and negative unique features.

These boundary conditions of the direction-of-comparison effect have implications for marketing. These findings suggest that manipulating the subject of comparison through advertising would work only when consumers do not have access to information about competing brands at the time of making the decision. Hence, the direction-of-comparison effect can be most fruitfully applied in the case of experience goods and credence goods where one-on-one comparison of features is not possible at the time of purchase. However, in the case of search goods, manipulating the focal option may not work to its advantage. For example, consumers may compare the nutritional value of two types of cereal at the time of purchase, regardless of a brand (e.g., Kellogs) making itself the focal option through point-of-purchase material or by comparative advertising against another brand (e.g., Total).

Second, although brands advertise only their unique positive features, most brands in today's market do not have the luxury of having unique features that are only positive. Hence the types of stimuli used in Study 2 add ecological validity to this area of research. It appears that the direction-of-comparison effect can be used by practitioners only when the unique features of the subject brand are positive and hence diagnostic. The findings in this research are limited in that the failure to support the direction-of-comparison effect (that is failure to reject the null hypothesis), does not imply that it does not exist. Future research is required in order to validate these findings.

REFERENCES

Agostinelli, G. Sherman, S.J., Fazio, R.H. and E.S. Hearst (1986), "Detecting and Identifying Change: Additions versus deletions," Journal of Experimental Psychology: Human Perception and Performance, 12, 445-454.

Carlston, D. E. (1980), "The Recall and Use of Observed Behavioral Episodes and Inferred Traits in Social Inference Processes," Journal of Experimental Social Psychology, 16, 779-804.

Cohen, J.B. and K. Basu (1987), "Alternative Models of Categorization: Toward a Contingent Processing Framework," Journal of Consumer Research, 13, 455-472.

Dhar, R. and I. Simonson (1992), "The Effect of the Direction of Alternative Comparison on Consumer Preferences," Journal of Marketing Research, in press.

Fischoff, B., P. Slovic, and S. Lichtenstein (1980), "Knowing What You Want: Measuring Labile Values," in Cognitive Processes in Choice and Human Behavior, ed. T. Wallstein, Hillsdale, NJ: Erlbaum.

Fiske, S. and M.A. Pavelchak (1986), "Category-based versus Piecemeal-based Affective Responses: Developments in Schema-triggered Affect," in The Handbook of Motivation and Cognition: Foundations of Social Behavior, eds. R.M. Sorrentino and E.T. Higgins, New York: Guilford.

Houston, D.A., S. Sherman, and S.M. Baker (1989), "The Influence of Unique Features and Direction of Comparison on Preferences," Journal of Experimental Social Psychology, 25, 121-141.

Kardes, F.R. (1986), "Effects of Initial Product Judgments on Subsequent Memory-Based Judgments," Journal of Consumer Research, 13, 1-11.

Lingle, J. H. and T.M. Ostrom (1979), "Retrieval Selectivity in Memory-based Impression Judgments," Journal of Personality and Social Psychology, 36, 180-194.

Mitchell, A.A. (1983), "Cognitive Processes Initiated by Exposure to Advertising," in Information Processing Research in Advertising, ed. R. Harris, Hillsdale, NJ: Erlbaum.

Payne, J.W. (1976), "Task Complexity and Contingent Processing in Decision Making: An Information Search and Protocol Abalysis," Organizational Behavior and Human Performance, 16, 366-387.

Payne, J.W. (1982), "Contingent Decision Behavior," Psychological Bulletin, 92, 382-402.

Sanbonmatsu, D.M., F.R. Kardes, and B.D. Gibson (1991), "The Role of Attribute Knowledge and Overall Evaluations in Comparative Judgment," Organizational Behavior and Human Decision Processes, 48, 131-146.

Slovic, P. and S. Lichtenstein (1983), "Preference Reversals: A Broader Perspective," American Economic Review, 73, 596-605.

Tversky, A. (1977), "Features of Similarity," Psychological Review, 84, 327-352.

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