On the Stability of Brand Images: Identifying the Routes to Information Discounting

ABSTRACT - The study utilizes the discounting paradigm to assess what is the sensitivity of brand images and attitudes to change. It proposes that the instruction to disregard information about a brand may lead to different consequences, depending upon how brand information is stored in memory. Specifically, it is posited that if the brand is unfamiliar, brand information is stored as a bundle of individual beliefs with little or no associations among them and thus, attitudes are mediated by changes in the challenged beliefs. If, on the other hand, the brand is familiar, brand schema is more unitized and the challenged belief is not retrievable individually. In this case discounting is not mediated by a revision in the challenged or related beliefs.



Citation:

David Mazursky (1996) ,"On the Stability of Brand Images: Identifying the Routes to Information Discounting", in AP - Asia Pacific Advances in Consumer Research Volume 2, eds. Russel Belk and Ronald Groves, Provo, UT : Association for Consumer Research, Pages: 19-27.

Asia Pacific Advances in Consumer Research Volume 2, 1996      Pages 19-27

ON THE STABILITY OF BRAND IMAGES: IDENTIFYING THE ROUTES TO INFORMATION DISCOUNTING

David Mazursky, Hebrew University of Jerusalem

ABSTRACT -

The study utilizes the discounting paradigm to assess what is the sensitivity of brand images and attitudes to change. It proposes that the instruction to disregard information about a brand may lead to different consequences, depending upon how brand information is stored in memory. Specifically, it is posited that if the brand is unfamiliar, brand information is stored as a bundle of individual beliefs with little or no associations among them and thus, attitudes are mediated by changes in the challenged beliefs. If, on the other hand, the brand is familiar, brand schema is more unitized and the challenged belief is not retrievable individually. In this case discounting is not mediated by a revision in the challenged or related beliefs.

Further, it is postulated that the impact of a discounting cue is mediated by levels of involvement. If involvement is high and the cue provides a central argument, attitude change is expected. If involvement is low, consumers will tend to focus on peripheral aspectsBnamely they downgrade the importance and reliability of the cue communicator without changing their attitude. Two studies were conducted that generally support these hypotheses.

INTRODUCTION

In some marketing situations, brand beliefs are challenged no by new information about a brand but by evidence discrediting the validity of information previously acquired about that brand. For example, agencies such as the FTC or experts in the relevant domain may advise consumers to disregard (or discount) some of the information provided in a previously advertised commercial because that information has been proven to be false.

Unlike the process of new information acquisition in which brand knowledge is changed with exposure to ads, word-of-mouth testimonials etc., the discounting process is always retrospective; it bears on information that has been previously acquired and processed by consumers. For that process to succeed, the discounting cue has to bring about a revision in the beliefs challenged by that cue and cause a corresponding re-evaluation of the brand so that the modified attitude will reflect the post-discounting beliefs.

The present study suggests that the discounting paradigm can serve as a useful framework for analyzing how brand images are encoded and stored in memory and assessing their sensitivity to a revision or displacement. For example, if brand information is stored in memory as a set of discrete components, a discounting cue may easily activate the challenged belief and cause an attitude change. If, on the other hand, brand information is represented as a utilized knowledge structure, discounting is more likely to be inhibited and not affect brand attitude.

The present study focuses on the process and the mechanism underlying information discounting for varied levels of brand knowledge and familiarity. It also investigates the moderating role of consumer involvement in facilitating (or inhibiting) this process.

Process of Information Discounting

A useful starting point for considering differences in discounting is the comparison between unfamiliar and familiar brands in relation to their representation in memory. Generally, theorists agree that product (or any object) information is stored in memory as a set of components (e.g., price, color) with a network of associations and inferences linking these components (see for example, in Hebb 1949; Hayes-Roth 1977).

The acquisition of new information about an unfamiliar product (or brand) can be characterized as a piecemeal (or discrete) processing (Fiske and Dyer 1985) whereby information is encoded and stored as discrete components with little or no associative links among them (Schul and Burnstein 1985). Product judgments that are based on such representation can best be viewed along the lines of the elementaristic approach (see, for example, Anderson 1981). This approach postulates that objects are represented as a bundle of independent beliefs. Product evaluation (e.g., attitude) emerges as a combination of these beliefs. The elementaristic approach underlies a host of methods that have been proposed in the past to predict product and social judgments (e.g., Fishbein and Ajzen 1975; Dickson and Miniard 1978).

As familiarity with the product increases, the network of associative links becomes more elaborated and integrated (Jacoby et al. 1974; Johnson and Russo 1984). Some components become increasingly more linked and generate information "Chunks". In addition, new beliefs and inferences are formed that reflect the new configuration of the attributes. The integrated cognitive structure and the derived inferential beliefs then form generic knowledge structures that categorize brands and attribute values within a certain product class (Mitchell and Olson 1981). For example, if consumers believe that a computer’s processing speed is high, it is likely that they will infer that its memory capacity is large. In its extreme, when product information becomes coherent and unitized, the entire knowledge structure can be activated in an all-or-none fashion (Fiske and Dyer 1985). This may occur if a brand is familiar and labeled (e.g., by mentioning "IBM PC" in an ad presented to computer users).

In the latter case, the cognitive mechanismunderlying brand judgment is likely to be characterized along the lines of the holistic approach (Anderson 1981). According to the holistic perspective, the evaluation of the brand (or product) cannot be derived from an analysis of the individual components. Rather, it depends on the nature of the associations linking the components and the inferences that constitute the unitized schema (Asch and Zukeir 1985).

Several previous investigations explicitly mentioned a familiar brand (e.g., Listerine) in testing the impact of discounting and remedial appeals (e.g., Mazis and Adkinson 1976; Dyer and Kuehl 1978; Sawyer and Semenik 1978). In these studies, the remedial cue represents an attack on a familiar, relatively highly interconnected knowledge structure. According to the holistic perspective, its meaning, as perceived by consumers, may be more easily activated by mentioning the brand name than by providing (or reminding) information about its attributes. In these cases, brand evaluation may be more closely related to the "holistic" meaning associated with the brand name than by the meaning derived as a function of its individual components.

In order to analyze the effectiveness of a discounting cue, it is useful to compare its impact on unfamiliar brands (represented as discrete knowledge structures in memory) with that on labeled and familiar brands (represented as a unitized structure in memory). We discuss these effects in turn.

Consider first the impact of a discounting cue devised to attack a specific attribute of an unfamiliar brand (e.g., "ignore the false claim stating that (unfamiliar) brand X is the least expensive computer within a specific performance category"). Assume, for this purpose, that all the brand information held in the consumer’s memory originates from an ad presented prior to the delivery of the discounting cue. Since brand evaluation, according to the elementaristic approach, is defined as a function of the person’s salient beliefs (Fishbein and Ajzen 1975), differences between pre- and post-discounting evaluations should be mediated by changes in the beliefs about the challenged and related attributes. Indeed, this mechanism postulates that if the challenged belief has been revised following discounting, so will also the overall brand attitude.

Alternatively, consider the impact of an identical discounting cue that if that cue relates to an attribute of a familiar and labeled brand ("IBM PC"). While the cue is relevant to a particular attribute (e.g., price), it may only marginally affect the "holistic" or integrative meaning reflecting the coherent associative structure (e.g., Schul and Burns1985). This may occur because the beliefs and inferences associated with the integrated schema may continue being interpreted in a manner that supports the original (pre-discounting) meaning (Schul and Burnstein 1985). This may occur because the beliefs and inferences associated with the integrated schema may continue being interpreted in a manner that supports the original (pre-discounting) meaning (Schul and Burnstein 1985). Thus, in the face of a highly coherent schema, the impact of a discounting cue is likely to be inhibited. In addition, since attitude is relatively less influenced by the individual beliefs, it is less likely that brand attitude be mediated by changes in the challenged and other beliefs that reflect message arguments.

Evidence from several psychological studies similarly points to asymmetry in discounting success when a cue is applied to alternative representations of cognitive structures: The more integrated the knowledge structure is, the higher the likelihood that discounting be impeded (Ross et al. 1977; Anderson et al. 1981; Anderson 1983; Schul and Burnstein 1985). Along similar lines but using a different discounting approach, the evidence from the marketing literature suggests that in order to successfully discount information about a familiar (and labeled) product, a massive rather than a weak cue is required (Hunt 1973; Dyer and Kuehl 1974).

Disounting Effectiveness: The Moderating Role of Involvement

The discussion provided above suggests that integrated knowledge structures may block the impact of discounting appeals on brand evaluation. In view of the fact, however, that in many realistic contexts discounting appeals are designed to attack familiar and labeled brands (e.g., Mazis and Adkinson 1976; Sawyer and Semenik 1978), it is important to identify the conditions under which this inhibition may, at least to some extent, be attenuated.

Evidence drawn from consumer involvement research seems to suggest that involvement can play an important role in moderating the success of discounting. Particularly, discounting parts of a coherent product schema information may be facilitated if involvement is high and the discounting appeal relates to a salient belief (e.g., price of a personal computer). This proposition follows from the premise that under high involvement, the product is highly relevant with respect to the person’s inherent needs, values and interest (Zaichkowsky 1985). Three complementing mechanisms are posited to account for discounting effectiveness under high involvement conditions. First, on the cognitive level, high involvement with a product has been hypothesized to enhance the perception of attribute difference (Howard and Sheth 1969). To some extent, this mechanism countervails that which inhibits the retrievability of individual’s information components in unitized memory schemata. That is, the challenged attributes under high involvement are more accessible and, thus, subjected to a direct attack by the discounting cue. Accordingly, discounting parts of familiar brand information may succeed when involvement is high. If involvement is low, the mechanism underlying discounting effectiveness is more like that discussed in the earlier section. Namely, if the appeal conflicts with information on a familiar brand and involvement is low, discounting is likely to fail.

Second, on the motivational level, high involvement is associated with increased consumer motivation to devote the cognitive effort required to evaluate the true merits of the product. The higher the involvement, the greater the likelihood that attitude change be accomplished by diligently considering salient product attributes. This process was termed the "central route" to persuasion (Petty and Cacioppo 1983). According to this reasoning, if the discounting cue attacks a salient brand characteristic, then brand evaluation ought to be revised.

On the other hand, a consumer who is not currently considering purchasing the product or brand (i.e., low involvement) may focus on peripheral product aspects such as attractiveness and credibility of message source rather than on central product arguments (i.e., "peripheral route"Petty and Cacioppo 1983; see also Wright 1974). If brand information is represented in memory as a coherent schema, an attempt to discount some of that information may fail; consumers will tend to cling to their old judgments without modifying their evaluation of the brand.

It is therefore more likely that under low involvement conditions, consumers will revise their judgments about some peripheral brand aspects following exposure to a discounting appeal so as to maintain cognitive consistency (e.g., Festinger 1957; Goethals and Reckman 1973). Specifically, they may indicate, for example, that the source providing the discounting cue is not reliable. Since consumers under low involvement conditions focus on source credibility and other peripheral attributes, brand attitude following a discounting cue may remain unaffected.

Third, the moderating role of involvement may also be analyzed according to a generalized context effect or "halo" effect (Kaplan 1975). Namely, if a brand is familiar, its representation in memory includes inferences reflecting the attributes and the associations among its attributes (e.g., "IBM probably has some functional advantage over identically compatible brands"). These beliefs and inferences may produce a positive effect reflected in the overall evalution of the brand (Mitchell and Olson 1981; Schul and Mazursky 1990). If a cue intended to discount some of the brand information is subsequently introduced (e.g., "The claim stating that IBM PC is the least expensive brand within that quality category was proved as being false; please ignore that information) it may "contaminate" consumers affective judgments and cause a revision in brand evaluations. Furthermore, this "contamination" may be more pronounced if involvement is high; persuasion via the central route under this condition may induce a stronger impact on all the inferences that are cognitively related to the challenged attribute. Hence, the resulting halo effect will produce a more substantial revision in brand evaluation if involvement is high than when it is low.

In sum, there is mounting research evidence indicating that discounting appeals may be less effective in bringing about attitude change if they conflict with highly coherent memory schemata. According to the hypothesis, however, if consumers are in a highly involved brand purchase (or use) situation, discounting inhibition may be attenuated even if the schemata are well developed.

Further, the underlying mechanism that gives rise to information discounting ascribes different roles to the challenged belief in moderating brand attitude change. If the brand is unfamiliar, brand attitude is mediated by changes in the target belief. On the other hand, if the brand is familiar, attitude change may be mediated by changes in the configuration of message and inferential beliefs associated with the unitized schema rather than by a revision in the challenged belief.

Finally, when involvement is low and the brand is familiar, focus of attention may not involve a direct conflict with the image. Instead, consumers may focus on peripheral cues (or beliefs). In such a case, rather than affecting brand favorableness, the cue may cause, for example, a discounting in (or derogation) the credibility attached to the source providing that cue.

Two studies were conducted to test these hypotheses. The first study was conducted a few weeks before a main frame computer system was replaced by a network of personal computers on a university campus. A relatively enhanced externally valid manipulation of involvement was facilitated on the basis of personal relevance as the subjects were not aware before the study about the installation of the new computers (see the method section below for a more detailed description). The study also manipulated two additional variables: the brand image (familiar vs. unfamiliar brand name) and discounting (discounting vs. message-only condition).

We had two major hypotheses. First, in the low-involvement condition, attitude change was expected to be less pronounced when the brand name was familiar. Discounting inhibition under the familiar brand name condition, however, was expected to decrease the higher the involvement. Second, within the unfamiliar brand condition, the beliefs reflecting message arguments were expected to covary with attitude change to the extent that their insertion as covariates in an analysis of covariance would remove the significance that was obtained when attitude was used as a sole dependent measure. Such covariation was not expected under the familiar brand-name conditions.

The second study focused more extensively on the condition in which discounting was inhibited. It replicated the familiar brandBlow involvement conditions using a different sample (consumers in department stores). It also incorporated measures of peripheral cues such as the credibility of the source providing the discounting cue. It was hypothesized that respondents would not modify their attitude following a discounting cue, but would tend to derogate the credibility of the source providing that cue.

STUDY 1

Method

Subjects

One hundred and sixty-eight students enrolled in an introductory computer course participated in the study. The study was conducted in groups of 10 to 15 subjects in each session. About the same number of females and males participated.

Design

The experiment involved eight between-subjects conditions. Four conditions were created by crossing an involvement manipulation (high vs. low) with the familiarity of brand image (a familiar labeled brand vs. an unfamiliar labeled brand). All the subjects in these groups received a discounting cue. The remaining four groups were assigned to similar treatments except that they were not exposed to the discounting cue, and thus, served as control groups.

Stimulus

Two major advantages motivated the selection of personal computers as the tested product category. First, the information about the university’s plan to install the new personal computers in the near future was known to the researchers but unknown to the subjects (all of whom were required to continue using computers during their studies). This may have represented an advantage in the study’s external validity over alternative simulation methods. Specifically, subjects assigned to the high involvement conditions were asked to express their attitude to a deal that would have implications for their own future studies. Subjects assigned to the low involvement groups, on the other hand, were asked to provide their judgments without knowing the details about the new plan.

With respect to involvement manipulation, the selection of personal computers was deemed appropriate as it held the physical dimension of involvement equal for all groups while varying the personal relevance of the product use decision (Lastovicka and Gardner 1979; Petty and Cacioppo 1983).

In addition, regarding the familiar brand conditions, a large proportion of the population at large was presumed to have at least some prior knowledge and familiarity with several leading brands (e.g., IBM, Apple). Conversely, regarding the unfamiliar brand conditions, an assumption that the presentation of an unfamiliar brand name would not sound unrealistic could be made given the proliferation and availability of many less known compatible PC’s in the current market reality. These facts were assumed to generate a reasonable dichotomy and at the same time minimize artificiality in the familiarity manipulation.

Procedure

The study was conducted four weeks before the semester break (installation of the new computers by the university being planned to take place during the break). The study was presented as a survey inquiring into students’ responses to a certain retailer’s new deal concerning a brand of personal computer. The subjects were handed a booklet which included several pages containing the information necessary for inducing the appropriate manipulation and a questionnaire. The purpose of the study was presented on the front page. Then, a page describing the new installation plan ending with the sentence, "The University authorities are presently considering the deal: Your evaluation of these computers may be helpful in reaching decisions with the students’ best interest in mind" was either (high involvement) or excluded (low involvement). A message describing eight characteristics of the computer in the form of an advertisement (e.g., memory capacity, a statement that a 100MB hard disc is available as part of the deal, price information claiming that it is the least expensive personal computer for this personal category) was subsequently presented. The attributes were selected for the study by comparing the frequency of their appearance in various personal computer ads and brochures. This was done to ensure that these eight attributes indeed represented key decision factors. In addition, there were two versions in which the brand names were altered: In one version, the message was attributed to a familiar brand name and in the otherBto an unfamiliar (in fact, fctitious) brand name.

Then, a distraction task was conducted in which a message describing a new study program was presented. Immediately afterwards, a discounting cue was communicated to the subjects in the "discounting" conditions. The cue was attributed to a computer expert. The information about the price of the computer was the target of the discounting cue. According to this expert, the claim stating that the computer is the least expensive within that category of computers is false. It should therefore be ignored in any decision or judgment that is made about that brand.

Finally, a questionnaire was administered consisting of two parts: the first part pertaining to the distraction task and the second, to the personal computer. The computer part consisted of three sections. The first section inquired about the attitude toward the brand. It included a series of four questions (e.g., very goodBnot good at all, usefulBnot useful at all) rated on seven-point scales. The questions were correlated (Alpha=0.70) and were averaged to form an attitudinal index.

The second section focused on three beliefs bearing on message attributes (memory capacityBdefinitely sufficient to not sufficient at all; ease of operationBvery easy to not easy at all; and the challenged beliefBprice informationBvery inexpensive for that performance category to very expensive for that performance category). The responses in this section were also rated on seven-point scales. Third, recall measures were obtained. Similar to the objectives of Johnson and Russo’s (1984) study, the recall task directed subjects to recall information beyond that which was provided to them in the message. Accordingly, subjects were instructed to recall brand attributes other than those presented in the original message.

TABLE 1

DISCOUNTING OF ATTITUDINAL AND BELIEF JUDGEMENTS - STUDY 1

Results

Recall and Manipulation Checks

To assess the effectiveness of the familiarity manipulation, the recall rate of the non-message attributes was contrasted between the familiar and unfamiliar brand conditions. Since under the discounting conditions, it was suspected that recall measures would be confounded by the discounting manipulation, this comparison was conducted only in the control (no-discounting) conditions. The analysis showed that subjects assigned to the familiar brand conditions recalled more non-message attributes (M=1.52) than those assigned to the unfamiliar brand conditions (M=.97, t(86)=1.97, p<.05).

To test the discounting manipulation, the main effect of the discounting manipulation on the challenged belief was examined. The results show that the discounting cue was indeed significantly effective (F(1,170)=5.72, p<.02). In addition, the main effect of the involvement manipulation was significant (F(1,170)=8.53, p<.01).

Discounting Success

The Moderating Role of Involvement: The analysis first concentrated on the contrasts relevant to the first hypothesis. The hypothesis stated that within the low involvement conditions discounting will be inhibited if the brand is familiar. However, a familiar labeled image, according to the hypothesis, will not always block the impact of a discounting cue; Discounting was expected to succeed if involvement is high.

The means of the relevant contrasts are displayed in columns 1,3 and 4 of Table 1. Testing the hypothesis entailed a comparison among the three conditions first within the no discounting and then within the three discounting condition. Within the no-discounting conditions, it was necessary to ensure that the initial pre-discounting attitudes were similar across the three conditions in order to enable the attribution of attitude change to variation in brand knowledge. Conversely, within the discounting conditions, no differences in attitudes were expected between the high involvement-familiar and the low-involvement-unfamiliar brand conditions (i.e., columns 1 vs 4). But when these two conditions are jointly compared with the low-involvement-familiar brnd condition (i.e., columns 1 and 4 vs 3) a significant difference was expected.

A 2X3 between subject ANOVA in which these three conditions were crossed with the discounting (discounting vs. no-discounting) factor was conducted including the contrasts as outlined above. When the level of discounting was controlled for, no difference within the no-discounting group among the three conditions was obtained (Table 1, first row, F<1). Within the discounting conditions, while no differences were obtained for the first contrast (Table 1Bsecond row, columns 1 vs. 4, F<1), differences have been observed in the second contrast (second row, columns 1 and 4 vs. 3, (F(1,123)=8.78, p<.010), as expected. The interaction effect across the two discounting conditions approached significance (F(2,123)=2.88, p<.06).

These tests were complemented by an analysis of the simple effects to test the impact of the discounting cue within each of the three conditions. The discounting and no discounting conditions were found to be significantly different under the high involvementBfamiliar and low involvementBunfamiliar brand conditions (F(1,123)=5.13, p<.03 and F(1,123)=6.3, p<.02, respectively) but were not statistically different under the low involvementBfamiliar brand condition (F<1). This analysis indicates that the overall interaction effect can be attributed to differences in discounting effectiveness between the first two versus the latter condition.

TABLE 2

A COMPARISON BETWEEN UNIVARIATE AND ROY-BARGMANN STEPDOWN ANALYSIS FOR FAMILIAR AND UNFAMILIAR BRAND CONDITIONS EFFECT: INVOLVEMENT x DISCOUNTING

Examining the Mechanism Underlying Discounting

According to the hypothesis, patterns of attitude revision within the unfamiliar brand condition are likely to be mediated by the challenged and other beliefs reflecting message arguments. Conversely, under the familiar brand condition, changes in brand attitude may not be mediated by these beliefs.

To test the hypothesis, two analysis of variance designs contrasting discounting with involvement were conducted. The first analysis focused on the unfamiliar and the second on the familiar brand conditions. Two MANOVA designs with Roy-Bargmann step-down tests (Roy and Bargmann 1958) were utilized. This analysis orders the entry of the dependent measures hierarchically so that for the first response variable it is identical to the univariate analysis. The test for the second response variable is identical to the univariate test statistic that would result if the first response variable were treated as a covariate. Similarly, the third response variable is adjusted for the first two variables, etc. A significant test for the kth response variable indicates that it cannot be accounted for by a linear combination of the preceding k-1 response variables.

Accordingly, the three measures of brand belief were entered to the equation along with the attitude measure. The effect of interest was the interaction between involvement and discounting given the prediction made about difference in discounting effectiveness as a function of involvement level. The univariate F tests for the attitude measure indicate that in both analyses the interaction effects were or approached significance (F(1,83=3.5, p<.06, and F(1,82)=5.4, p<.03 for the unfamiliar and familiar brand conditions, respectively).

It is interesting to compare these results to those obtained in the Roy Bargmann test when the measure of individual beliefs are controlled for (see Table 2). Concerning the unfamiliar brand conditions, the interaction effect on the challenged belief was significant (F(1,67)=3.9, p<.05). However, note that when the significance of the interaction on attitude disappeared when attitude was adjusted for by the belief measure. It seems, therefore, that changes in the target and other message derived beliefs covaried with the change in overall brand attitudes.

A different pattern of results was obtained for the familiar-brand conditions when the Roy-Bargmann procedure was utilized. The rate of discounting of the challenged belief information did not reach significance (F<1), while the significance of the interaction effect for the attitude measure remained significant (F1,72)=6.6, p<.02). Under these condtions, it would appear, therefore, that discounting was not mediated by a revision in the target and other message related beliefs.

Discussion

The findings of Study 1 support the hypothesis stating that with low involvement, a coherent memory representation of brand information may block the impact of a discounting cue. The findings also indicate that discounting does not necessarily fail under all circumstances: When involvement is high, the impact of the discounting cue on attitude change may be substantial.

It should be noted that the major dependent variable used for assessing the impact of discounting was attitude. As noted earlier, the discounting process is successful if it brings about a re-evaluation of the brand after the discounting cue is presented. However, a comparison of the impact of the discounting cue on the challenged belief versus that on attitude change also assists us with the interpretation of discounting success under the high involvement condition. Specifically, the mean differences in the challenged belief show no major difference between the low and high involvement conditions (i.e, columns 1 and 3 in Table 1) though significant differences are observed on the attitudinal level. This observation seems to provide some support to the generalized context of "halo" effect as a plausible interpretation of this finding. Specifically, the impact on attitude was partly moderated by beliefs and inferences related to the challenged belief. In the high involvement condition, the cue may have served as a contrasting reference to inferences associated with the target belief and, thus, have a strong impact on attitude change. Conversely, under the low involvement condition, the remaining beliefs and inferences were not activated to the extent that they affected attitudinal judgments.

The findings also shed light on the mechanism underlying discounting when varied levels of brand familiarity are considered. In the case of an unfamiliar brand, attitude change following discounting was mediated by changes in the target belief and other beliefs reflecting message arguments. When the brand was familiar, on the other hand, attitude change was not mediated by message derived beliefs.

An interesting question arises following an assessment of these findings. In the low involvement-familiar brand conditions no significant changes were obtained in the challenged belief or in brand attitude; is it plausible that in spite of the attack on brand information, the discounting cue did not exert any impact on any brand judgments? Put differently, how can a cognitive and affective balance be maintained following the delivery of a discounting cue?

As noted above, since under low involvement the peripheral route is posited as dominant, it may be that the influence of the discounting cue is impeded by derogating the credibility of the source communicating that cue. If this is true, cognitive balance may be maintained even though no brand related judgments had been affected. Study 2 was designed to replicate the low involvement-familiar brand condition. To address the hypothesis, other dependent measures were included inquiring into the importance and credibility of the discounting cue communicator. It was hypothesized that rather than discounting product judgments, consumers will derogate the creditability of the cue communicator and its importance in brand judgments.

STUDY 2

Method

Subjects

One hundred and forty consumers were handed a booklet while waiting on line before entering a department store. (The study was run in the morning, about an hour to fifteen minutes before the store opened). Allthe three sampled stores were located in a large metropolitan area. A condition for selecting these stores was that they do not carry personal computers or any appliances related to that product. The sample was pre-screened on the basis of personal involvement; consumers indicating that they were actively seeking the purchase of a personal computer (for their own or anybody else’s use) during the study period or had bought one during the past six months were screened out [To ensure that the screening measures selected low involvement consumers, measures of the facets of involvement (Laurent and Kapferer 1985) were assessed. These included perceived importance, probability of mispurchase, pleasure value and sign value. On all these measures, the mean values did not exceed the level of 3 (on seven-point scales where 1 was lowest and 7 - highest).]. According to this criterion, one-hundred and twenty-eight consumers participated. Their age ranged between 18 and 64 and about the same number of males and females participated.

Procedure and Design

Consumers were approached and asked to participate in a survey concerning some of the latest published information about a well known personal computer. They were randomly assigned to one of two discounting conditions by means of the content of a booklet handed to them. Some consumers received a booklet containing a message advocating the purchase of the computer (identical to the ad handed to subjects in the first study), an excerpt from a comparison of various brands published in a well known consumer information magazine (i.e., a discounting cueBsee below) and a questionnaire. The other group of consumers received a similar booklet except that the excerpt from the magazine was omitted. The discounting manipulation was therefore induced by the inclusion (or omission) of that cue.

The excerpt presented the recommendation drawn from a comparison among brands of pc’s. According to the message, there were alternative available computers that use operating systems similar or identical to the focal (familiar) brand x but were less expensive. Consequently, consumers were advised to ignore the ad claim about the computer being the least expensive choice among the computers that were equivalent in performance.

All consumers were then asked to fill out a questionnaire. This included a measure of attitude toward the brand (composed of measures similar to those in Study 1, Alpha=.88), a belief reflecting the challenged argument, and two judgments related to peripheral aspects of the brand. These included the reliability of the discounting cue source (very reliableBnot reliable at all) and the importance consumers attach to that source (very importantBnot important at all) in brand evaluation. As in Study 1, all the questions were rated on seven-point scales.

Results

A MANOVA design with one independent factor (discounting vs. no discounting) and four dependent measures (belief about the challenged argument, attitude, reliability of the source providing the discounting cue and its importance in brand assessment) was conducted. The means and standard deviations of these variables are presented in Table 3.

The univariate analysis of variance revealed different patterns of impact on brand judgments and on source judgments. Particularly, there was no revision in any of the challenged belief (F(1,126)<1) nor in overall brand attitude (F(1,126)<1). On the other hand, subjects exposed to the discounting cue attributed less importance and reliability to the cue communicator than subjects assigned to the message-only conditions

(F(1,126)=9.94, p<.01 in importance judgment and F(1,126)=7.03, p<.01 in the reliability of the discounting cue communicator). Overall, the multivariate tests of significance of the discounting effects were significant (the value obtained according to Pillari’s criterion was .09(F4,123)=3.10, p<.02).

General Discussion

Recent research in psychology and marketing has focused on the processes that underlie the development of social and product schemata. The present study offers the discounting cue paradigm as a useful framework for investigating how strongly integrated and utilized the schemata are and how attitudinal judgments derive from different memory repreentations of product information.

The study compares two contrasting approaches that underlie judgmentsBthe elementaristic and the holistic perspectives. To this end, it focused on the role played by the challenged beliefs in mediating attitude revision following the introduction of a discounting cue. The hypothesis stated that the elementaristic approach is more likely to characterize processing if the schemata are in early stages of development. If the schemata are well developed, on the other hand, then the holistic approach is more likely to underlie decisions and attitudinal judgments. Empirically, the hypothesis was tested by measuring the extent to which removing the effects on message derived beliefs (including the challenged belief) also removed the significance of the effects on attitudinal judgments. This hypothesis was generally confirmed. Specifically, when the effects on beliefs were controlled for, the significance of the effects on attitude was removed under the unfamiliar brand condition but remained in the unfamiliar brand conditions.

TABLE 3

DISCOUNTING OF ATTITUDINAL, BELIEF AND SOURCE JUDGMENTS - STUDY 2 (LOW INVOLVEMENT - FAMILIAR BRAND CONDITION)

Recent research on the structure of schemata has suggested that stages of schemata development can be ordered along a continuum whereby each level represents a meaningful stage in information processing (Fiske and Dyer 1985). The continuum corresponds to acquisition patterns which range from discrete (or piecemeal) encoding and storage to processing in which acquired information activates unitized and integrated schemata in a holistic fashion. In this respect, this continuum serves as a useful framework for categorizing various approaches to judgment and decision processes. On one end of this continuum (representing schemata in early stages of development) we can place the expectancy value models and the traditional information processing perspective. These models generally view the evolution of brand judgments as a function of discrete attributes or components. In contrast, in the well developed schemata or holistic end of this continuum some of the more recent approaches can be placed that include as part of the judgment process, notions previously considered external (and perhaps diminishing in influence) to this process. These consist of product familiarity (e.g., Johnson and Russo 1984; Zajonc and Markus 1982), attitude toward the advertisement (Shimp 1981; Mitchell and Olson 1981; Gardner 1985), vividness (Kisielius and Sternthal 1986), inferential belief formation (Olson 1978) and others. In the listed studies, an attempt is generally made to extend the study of schemata characteristics beyond the focus on discrete brand beliefs. The present study focused on two extreme stages along this continuum and utilized the discounting paradigm to account for differences in the ability to remove the impact of individual beliefs under two alternative schemata structures.

Though two far apart levels along the schemata development continuum were selected, it should be noted that in most marketing situations, new brands typically are not represented as completely discrete components (or alternatively, as completely undeveloped schemata). Learning information about new brands (and even new products) involves the avocation of overlapping schemata (Fiske and Dyer 1985). For example, the ease of learning a computer’s word processing characteristics may depend on familiarity of the consumer with a typewriter and how well developed is her schema of a typewriter. Two consumers who differ in the degree of typewriter familiarity may also develop distinguishable word processor schemata. Thus, while levels of schemata development differed between the unfamiliar and the familiar brand conditions, the attributes learned about the unfamiliar brand were not free of associations to previously learned knowledge about relevant phenomena. Thus, even under the unfamiliar brand conditions, it could be that the presentation of the ad induced, albeit to a lessor degree, the formation of inferential beliefs.

Consumer involvement was identified as a factor that counteracts discounting inhibition when brand schema is unitized. This finding was observed under the high involvement condition. In part, this observation i attributed to the distinct functions of the central and peripheral routes in information processing and that the two routes are moderated by the involvement construct. In this regard, the present study supports, though from a different angle, the existence of such distinct routes. However, the interpretation should be made with caution: The results of Study 1 provide some support for the impact of a generalized halo effect in accounting for the difference between high and low involvement when the brand was familiar.

Although the purpose of the present study was confined to stimuli and discounting effects which were assumed to be prevalent and important in marketing practice, other manipulations and stimuli may be useful to complement the theory relating to the impact of discount cues. For example, both studies focused only on a cue considered to be central and salient in attitude formation. Variation in centrality of the discounting cue as well as variation in the source(s) communicating the cue may be useful to enrich the theoretical framework pertaining to discounting effectiveness. It is also recommended that the findings be generalized by sampling other non-student segments in future research.

Much advertising effort is expended in attempting to influence consumers to form attitudes and images which will be resistant to attacks by competitors and other external influences. Manufacturers, as well as consumer protection agencies may be interested in learning the conditions under which attitudes are resistant to change and those under which brand evaluation is more susceptible to change. From the company’s point of view, such an activity can be beneficial as it may reduce the probability of brand switching (Aaker and Myers 1982) and extend the profitable stage of the product life cycle (Bither et al. 1971). From the point of view of consumer protection agencies, this investigation may assist the relevant authorities with an a priori appraisal of strategies designed to remove deceptive beliefs and attitudes (see a review in Wilkie et al. 1984). In all these cases, the identification of levels of knowledge structure and consumer involvement are critical for assessing the usefulness of alternative marketing strategies.

REFERENCES

Aaker, D.A. and Myers J.G. (1982) Advertising Management, N.J.: Prentice Hall.

Anderson, C.A. (1983) "Abstract and Concrete Data in the Perseverance of Social Theories: When Weak Data Lead to Unshakable Beliefs," Journal of Experimental Social Psychology, 19, 93-108.

Anderson, C.A., Lepper, M.R. and Ross L. (1980) "Perseverance of Social Theories: The Role of Explanation in the Persistence of Discredited Information," The Jornal of Personality and Social Psychology, 39, 1037-1049.

Anderson, N.H. (1981) Foundation of Information Integration Theory, New York: Academic Press.

Asch, S.E. and Zukeir, H. (1983) "Thinking About Persons" Journal of Personality and Social Psychology, 46, 1230-1240.

Bither, S.W., Dolich I.J. and Nell B.E. (1971) "The Application of Attitude-Immunization Techniques in Marketing," Journal of Marketing Research, 8, Feb., 56-61.

Burnstein, E., and Schul, Y. (1983) The Informational Basis of Social Judgments: Memory for Integrated and Non-integrated Trait Descriptions, Journal of Experimental Social Psychology, 19, 49-57.

Dickson, P.R. and P.W. Miniard (1978) "A Further Examination of Two Laboratory Tests of the Extended Fishbein Attitude Model," Journal of Consumer Research, 4, pp. 261-66.

Dyer, R.F. and Kuehl P.G. (1978) "The Corrective Advertising Remedy f the FTC: An Experimental Evaluation," Journal of Marketing, 38, pp. 48-54.

Festinger, L. (1957), A Theory of Cognitive Dissonance. Evanston, IL: Row, Peterson.

Fishbein, M. and Ajzen, I. (1975) Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research Reading, Mass: Addison-Wesley.

Fiske, Susan and Linda Dyer (1985) "Structure and Development of Social Schemata: Evidence from Positive and Negative Transfer Effects," Journal of Personality and Social Psychology, April pp. 839-52.

Gardner, M.P. (1985) "Does Attitude Toward the Ad Affect Brand Attitude Under a Brand Evaluation Set?" Journal of Marketing Research, 22, May, 192-198.

Goethals, G.R. and Rechman, R.F. (1973) "The Perception of Consistency in Attitudes," The Journal of Experimental Social Psychology, 9, 491-501.

Hayes-Roth, B. (1977) "Evolution of Cognitive Structures and Processes." Psychological Review, 84, pp. 260-78.

Hebb, D.O. (1949) The Organization of Behavior: A Neuropsychological Theory. N.Y.: Wiley.

Howard, J.A. and Sheth J.N. (1969) The Theory of Buyer Behavior, New York: John Wiley.

Hunt, H.K. (1973) "Effects of Corrective Advertising," Journal of Advertising Research, 13, 15-24.

Jacoby, J., Speller, D.E. and Kohn Berning, C.A. (1974) "Brand Choice Behavior as a Function of Information Load: Replication and Extension," Journal of Consumer Research, 13, 15-24.

Johnson, E.J. and Russo, J.E. (1984) "Product Familiarity and Learning New Information," Journal of Consumer Research, 11, June, 542-550.

Kaplan, M.F. (1975) "Evaluative Judgments are Based on Evaluative Information: Evidence Against Meaning Change in Evaluative Context Effects," Memory and Cognition, 3, 375-380.

Kisielius, J. and Sternthal B. (1986) "Examining the Vividness Controversy: An AvailabilityBValence Interpretation," Journal of Consumer Research, 12, March, 418-431.

Lastovicka, J.L. and Gardner, D. (1979), "Components of Involvement," in Attitude Research Plays for High Stakes, eds. John C. Maloney and B. Silverman, Chicago: AMA, 53-73.

Laurent, G. and Kapferer, J. (1985) "Measuring Consumer Involvement Profiles," Journal of Marketing Research, Feb. 22, 41-53.

Mazis, M.B. and J.E. Adkinson (1976) "An Experimental Evaluation of a Proposed Corrective Advertising Remedy," Journal of Marketing Research, May 13, pp.178-83.

Mitchell, A.A. and Olsen, J.C. (1981) "Are Product Attribute Beliefs the Only Mediator of Advertising Effects on Brand Attitude?" Journal of Marketing Research, 18, Aug. 318-332.

Olsen, J.C. (1978) "Inferential Belief Formation in the Cue Utilization Process," in Advances in Consumer Research, 5, H. Keith Hunt, ed. Ann Arbor, Michigan: ACR, 706-713.

Petty, R.E., Cacioppo J.T. and Schumann, D. (1983) "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10, Sep. 135-146.

Ross, L., Lepper, M.R., Strack F. and Steinmetz, J. (1977) "Social Explanation and Social Expectation: Effects of Real and Hypothetical Explanations on Subjective Likelihood, Journal of Personality and Social Psychology, 35, 817-829.

Ross, M. and Michael Conway "Remembering One’s Own Past: The Reconstruction of Personal Histories," in press.

Roy, J. and Bargmann, R.E. (1958) "Tests of Multiple Independence and the Associated Confidence Bounds," Annals of Math. Stat., 29, 491-503.

Sawyer, A.G. and Semenik, R.J. (1978) "Carryover Effects of Corrective Advertising," in Advances in Consumer Research, 5, H. Keith Hunt, ed. 343-351.

Schul, Y. and Burnstein E.(1985) "When Discounting Fails: Conditions Under Which Individuals Use Discredited Information in Making a Judgment," Journal of Personality and Social Pathology.

Schul, Y. and D. Mazursky (1990) "Conditions Facilitating Successful Discounting: Type of Discounting Cue, Message Encoding, and Kind of Judgment," Journal of Consumer Research, 16, March, 442-451.

Shimp, T. (1981) "Attitude Toward the Ad as a Mediator of Consumer Brand Choice," Journal of Advertising, 10 (2), 9-15.

Wilkie, William, Dennis McNeill and Michael Mazis (1984) "Marketing `Scarlet Letter’: The Theory and Practice Corrective Advertising," Journal of Marketing, Spring, No.2, pp. 11-31.

Wright, P.L. (1974) "Analyzing Media Effects on Advertising Responses," Public Opinion Quarterly. 7, Sep. 151-175.

Zaichkowsky, J.L. (1985) "Measuring the Involvement Construct," Journal of Consumer Research, 12, Dec. 341-352.

Zajonc, R.B. and Markus, H. (1982) "Affective and Cognitive Factors in Preferences," Journal of Consumer Research, 9, Sep. 123-131.

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

Authors

David Mazursky, Hebrew University of Jerusalem



Volume

AP - Asia Pacific Advances in Consumer Research Volume 2 | 1996



Share Proceeding

Featured papers

See More

Featured

A Meta-Analysis on the Endowment Effect in Experiments

DANIEL SUN, University of Calgary, Canada

Read More

Featured

The Secrecy Effect: Secret Consumption Polarizes Product Evaluations

Maria A Rodas, University of Minnesota, USA
Deborah Roedder John, University of Minnesota, USA

Read More

Featured

What Converts Webpage Visits into Crowdfunding Contributions: Assessing the Role of Circumstantial Information

Lucia Salmonson Guimarães Barros, Universidade Federal de Sao Paulo
César Zucco Jr, Brazilian School of Public and Business Administration, Brazil
Eduardo B. Andrade, FGV / EBAPE
Marcelo Salhab Brogliato, Brazilian School of Public and Business Administration, Brazil

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.