Comment on Factors Affecting Evaluations

Richard F. Yalch, University of Washington
[ to cite ]:
Richard F. Yalch (1992) ,"Comment on Factors Affecting Evaluations", in NA - Advances in Consumer Research Volume 19, eds. John F. Sherry, Jr. and Brian Sternthal, Provo, UT : Association for Consumer Research, Pages: 276-278.

Advances in Consumer Research Volume 19, 1992      Pages 276-278

COMMENT ON FACTORS AFFECTING EVALUATIONS

Richard F. Yalch, University of Washington

The conference co-chairs helped my integration task by recognizing a common element, a concern with influences on consumer judgments, in grouping three apparently unrelated studies. Another major element is that the papers assess the influence of new information presented to consumers relative to information consumers previously stored about a particular brand or product. Thus, apparently disparate factors such as context, brand familiarity, self-referencing, and explicit recognition of missing product information similarly affect consumer judgments by altering the importance of message information relative to previously stored information. After reviewing the major findings and some methodological issues, I will present an information processing model elaborating on this perspective.

MAJOR FINDINGS OF PAPERS

The three sets of authors are to be congratulated for identifying important limitations in the literature and designing studies to address overlooked issues. Two papers question whether previously reported levels of association between variables are artificially inflated. Machleit and Sahni (1992) insightfully identify that research relating attitude toward the ad and attitude toward the brand used a context in which the two concepts were measured in close proximity and therefore should tend to appear related. Further, they observed that many studies using familiar brands reported low and insignificant correlations and offered brand familiarity as another important moderator of the attitude toward the ad and attitude toward the brand relationship. Their study verified the moderating roles of context and brand familiarity.

Sirdeshmukh and Unnava (1992) review past research showing that missing information causes consumers to report more negative attitudes toward a product, but express concern that these findings may inflate the importance of missing information because respondents were made aware explicitly that important product information was missing. They hypothesize that without such awareness consumers would be less affected by the missing information. Again, their study validates their suspicions.

Romeo and Debevec (1992) also raise a methodological issue in concluding that past research on the self-reference effect considers self-referencing to be a cognitive process and did not measure affective reactions. Their study includes affective measures and validates that they are affected by self-referencing.

SOME METHODOLOGICAL CONCERNS

Before developing a model synthesizing these results, brief comments are provided on the methods employed by the authors. Although the studies are well-carried out, a few suggestions are offered to aid future research.

Machleit & Sahni (1992) used television ads for two familiar brands (Coke and Michelob) and unfamiliar brand (Mug and Kronenbourg). In reading the descriptions, I noticed that the familiar brands used musical themes whereas the unfamiliar brands used humor. Thus, there is a confound between familiarity and advertising approach. I do not know that the literature speaks clearly why correlations should be higher between attitudes toward humorous ads and brand attitudes than for attitudes toward musical ads and brand attitudes. However, humor may be more affect stimulating and memorable than music, providing a rival explanation for the results.

Sirdeshmukh and Unnava's (1992) asked subjects to evaluate hypothetical product given only the information available. Not surprisingly, they did so. Only in the explicit condition (i.e., when clearly more information could have been provided), and only when provided with less than half the information (four out of ten characteristics), did the subjects evaluate the product significantly less favorably than in the full information conditions. Interestingly, this occurred although subjects perceived that the information was not sufficient in most conditions (four attributes - implicit presentation, four, six and eight attributes in the explicit condition). Thus, rather than conclude that consumers react unfavorably when important product information is missing, one could conclude that they are surprisingly indifferent to missing information.

Some questions arose regarding Romeo and Debevec's (1992) analysis rather than their design. In the design, the level of self-referencing was varied by having subjects read copy with pronouns focusing their attention on themselves, another person or the product. However, in the analysis, the authors used the reported level of self-referencing not the assigned level. This amounts to self-assignment to treatments and may undermine internal validity. For example, it risks confounding self-referencing with other factors such as agreeableness or involvement. That is, highly involved or cooperative individuals might be expected to report high self-referencing, positive attitudes toward the ad and favorable product evaluations.

Another concern is the construct validity of the affective and cognitive scales for measuring attitude toward the ad. For example, Romeo and Debevec (1992) use exciting and interesting as cognitive evaluations. However, they relate their results to Edell & Burke (1987) who used excited and interested as measures of upbeat feelings. In addition, Romeo and Debevec (1992) claim that relevant, interesting, exciting and entertaining measure the content whereas believable, realistic and informative measure the of the ad. To me, interesting seems more a quality measure and informative more a measure of content. It would be desirable to validate the scale labels.

FIGURE

FACTORS AFFECTING THE INFLUENCE OF EXISTING MEMORY AND MESSAGE INFORMATION ON CONSUMER JUDGMENTS

A MODEL OF THE JUDGMENT PROCESS

Methodological concerns aside, what do these studies have in common? All three support a model of memory that maintains that consumers base their judgments on the information available to them and that the amount used in making judgments is typically much less than their total product knowledge. Available information includes their experiences, exposures to advertising and other marketing stimuli, as well as the information presented to them in the laboratory. As shown in the figure, consumer judgments are based on two sources of information. One is information retrieved from long-term memory and the other is information recently acquired from an advertisement (e.g., Machleit And Sahni 1992; Romeo and Debevec 1992) or product description (e.g., Sirdeshmukh and Unnava 1992). The extent to which one or the other influences judgments depends on its likelihood of retrieval. Retrieval, in turn, depends on factors such as recency (usually the shorter the time between exposure and judgment the more likely information will be used in the judgment task) and elaboration (the more associations to a specific bit of information the more likely it will be used in the judgment task). Note that thoughts can be favorable, unfavorable or neutral and either cognitive or affective (emotional). Let's apply this model to the three studies.

For Machleit And Sahni (1992), it should be obvious that consumer thoughts should be more elaborate for familiar brands like Coke and Michelob than for unfamiliar brands like Mug and Kronenbourg. For familiar brands, consumers can easily retrieve prior thoughts and attitudes and use them when making product judgments. This is especially likely when filler questions are used to reduce the accessibility of attitudes toward the ad (their context effect). However, having consumers respond to attitude toward the advertisement questions immediately before making product evaluations might make these thoughts sufficiently salient that they will influence the product judgments resulting in significant correlations. This was true for Michelob but not for Coke. This might be explained by Coke's being such a familiar brand that prior thoughts are extremely accessible and thus not easily ignored. It also might be attributed to the fact that the particular ad used did not present any new information (emotional or cognitive) about Coke.

For unfamiliar brands, attitudes toward a recently viewed ad will tend to dominate prior thoughts. This results in relatively high correlations between attitude toward the ad and attitudes toward the brand. The dominance of advertising-related thoughts will be maintained even when product judgments are separated from advertising judgments by filler questions. However, one should expect some attenuation as Machleit and Sahni (1992) found (see their Table 1). This raises the important question of how much predictive power attitudes toward the ad will have when measures are greatly separated. For example, do attitudes toward the ad in a copy test predict purchase behavior?

For Romeo And Debevec (1992), self-referencing serves as the device for altering the accessibility and valence of message information. The authors believe that having subjects elaborate on message information by relating it to their personal experiences enhances message learning and makes message information more accessible when product judgments are rendered. In addition, Romeo And Debevec (1992) offer the idea that there may be an emotional component to self-referencing such that self thoughts tend to be positive (we think well of ourselves) and this will further encourage favorable evaluations of the advertised product. The results of their study support these conjectures. However, these results may not always be obtained. For example, research by Brian Sternthal and me (in review) shows that self-referencing has less effect on learning advertising messages than isolated nouns and adjectives. Advertising is connected discourse and hence stimulates a moderate level of elaboration. Further, self-referencing requires that consumers be able to relate the advertised product to themselves. If consumers cannot easily see themselves as users, self-referencing may undermine learning. This occurred in one of Lord's (1980) condition in which individuals found it easier to visualize others using an object than themselves.

Even if it enhances message learning, self-referencing might not produce favorable affective reactions. In Yalch and Sternthal (in review), when individuals had access to personal experiences indicating that a product was not of high quality, individuals using high self-referencing were less favorable toward the products then individuals using only a moderate amount of self-referencing.

In Sirdeshmukh and Unnava (1992), the issue is whether consumers resort to inferences when a message does not provide enough information to make a valid judgment. The researchers' messages consisted of product ratings on a series of characteristics. Subjects received information on four, six, eight or ten characteristics and were either made aware that the product could be evaluated on ten characteristics (explicit condition) or not made aware of this (implicit condition). In the implicit condition, attitudes did not vary with the amount of information provided, whereas attitudes became more favorable when more information was provided in the explicit condition. However, the major difference between implicit and explicit conditions occurred only when four characteristics were described. In this case, subjects provided substantially more negative product evaluations. It is not clear why this occurred except that subjects may have become suspicious why less than half the possible information was available. Their negative evaluations may have been based less on inferences drawn from the information provided and more on attributions regarding the manufacturer's integrity.

Note, that once evaluative information for six characteristics were provided or when subjects were not informed that important information was withheld, their product evaluations tended to be similar. This occurred although subjects acknowledged that the information was not sufficient (e.g., sufficiency rating was less than three on a seven-point scale when information was provided for six attributes in the explicit condition - see Table in Sirdeshmukh and Unnava 1992).

Why are evaluations largely unaffected by missing information? An explanation is that judgments tend to be based only on a limited amount of information even when much information is provided. That is, when making judgments, individuals rarely access and process all pertinent information as they might in a recall task. Instead, they bring a limited amount into short-term memory (four to five bits) and they reach a judgment. It is possible that Sirdeshmukh and Unnava (1992) are observing the effect of consumer suspicions about why the additional information is not provided. This might include concerns about the quality of the testing agency or the manufacturer's reluctance to cooperate with the testing agency. It would be an interesting follow-up to redo the study and offer different explanations for the missing information. Attitudes should vary depending on whether it was preventable and who is the cause.

In conclusion, the three studies provide confirmation for an integrative view of memory and judgments. They show that the relative effect of product information acquired externally versus information retrieved from long term memory is predictable by the relative accessibility of each. Having individuals rehearse message information by asking them to evaluate the message immediately before making product judgments, by instructing them to self-reference it, or by listing it on the same sheet of paper as the product evaluations enhances the influence of the message information. On the other hand, including a time delay between advertising evaluations and product evaluations, using familiar products, not having individuals self-reference the material, or providing more or less information than needed will minimize the influence of message information.

REFERENCES

Edell, Julie A. and Marian Chapman Burke (1987), "The Power of Feelings in Understanding Advertising Effects," Journal of Consumer Research, 14 (December), 421-433.

Lord, Charles G. (1980), "Schemas and Images as Memory Aids: Two Modes of Processing Social Information," Journal of Personality and Social Psychology, 35 (February), 63-78.

Machleit, Karen A. and Arti Sahni (1992), "The Impact of Measurement Context on the Relationship between Attitude Toward the Ad and Brand Attitude for Familiar Brands," in Advances in Consumer Research, Vol. 19, John Sherry and Brian Sternthal, eds., Provo, UT: Association for Consumer Research, in press.

Romeo, Jean B. and Kathleen Debevec (1992), "An Investigation of Self-Referencing's Influence on Affective Evaluations," in Advances in Consumer Research, Vol. 19, John Sherry and Brian Sternthal, eds., Provo, UT: Association for Consumer Research, in press.

Sirdeshmukh, Deepak and H. Rao Unnava (1992), "The Effects of Missing Information on Consumer Product Evaluations," in Advances in Consumer Research, Vol. 19, John Sherry and Brian Sternthal, eds., Provo, UT: Association for Consumer Research, in press.

Yalch, Richard F. and Brian Sternthal (in review), "The Effect of Self-Reference on Message Learning and Judgments." University of Washington.

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