Measuring and Modeling Brand Interest As an Alternative Ad Effect With Familiar Brands

Karen A. Machleit, University of Cincinnati
Thomas J. Madden, University of South Carolina
Chris T. Allen, University of Cincinnati
ABSTRACT - This study develops a measure for a new construct, brand interest, and reports exploratory modeling with the construct in a traditional Aad modeling context. The four-item brand interest scale evidences unidimensionality and discriminant validity. Structural equation modeling suggests that affecting brand interest may be a meaningful advertising objective for familiar brands.
[ to cite ]:
Karen A. Machleit, Thomas J. Madden, and Chris T. Allen (1990) ,"Measuring and Modeling Brand Interest As an Alternative Ad Effect With Familiar Brands", in NA - Advances in Consumer Research Volume 17, eds. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT : Association for Consumer Research, Pages: 223-230.

Advances in Consumer Research Volume 17, 1990      Pages 223-230

MEASURING AND MODELING BRAND INTEREST AS AN ALTERNATIVE AD EFFECT WITH FAMILIAR BRANDS

Karen A. Machleit, University of Cincinnati

Thomas J. Madden, University of South Carolina

Chris T. Allen, University of Cincinnati

ABSTRACT -

This study develops a measure for a new construct, brand interest, and reports exploratory modeling with the construct in a traditional Aad modeling context. The four-item brand interest scale evidences unidimensionality and discriminant validity. Structural equation modeling suggests that affecting brand interest may be a meaningful advertising objective for familiar brands.

INTRODUCTION

Since the early 1980s, many studies have demonstrated an association between attitude toward the ad (Aad) and brand attitude (e.g., Batra and Ray 1986; Edell and Burke 1986, 1989; Gardner 1985; MacKenzie, Lutz and Belch 1986; Mitchell and Olson 1981; Stayman and Aaker 1988). However, a common feature which limits the generalizability of these empirical studies is their focus on unfamiliar or hypothetical brands. On a practical level, while many new brands are continually being introduced, companies depend on mature, familiar brands for a stable base of sales. In addition, these familiar brands often utilize emotional appeals in their ads, and thus may evoke positive feelings or Aad on the part of the viewer.

It is not surprising to find that Aad affects brand attitude when the individual is unfamiliar with the brand; in this instance, the advertisement is the only basis for brand evaluation. Indeed, preliminary empirical evidence indicates that the effect of Aad on brand attitude may be important only for unfamiliar brands. For example, Gresham and Shimp (1985) found significant Aad effects on brand attitude (Ab) for only six of fifteen familiar brands, Edell and Burke (1986) found weaker Aad effects on Ab for familiar brands, and Machleit and Wilson (1988) report non-significant effects of Aad on Ab for two familiar brands.

Given it is expected that Aad will have limited effects on brand attitude for familiar consumers, additional research is needed to address the question of what happens when a consumer is exposed to an ad for a brand with which s/he is already familiar. To expect that exposure to a few ads for a brand will change brand attitude for an experienced consumer seems ambitious. It is known that attitudes of experienced individuals are typically "well-formed" and are resistant to change (cf. Fazio and Zanna 1981), and would be more likely to be affected by additional experience rather than exposure to advertising (Smith and Swinyard 1982, 1988). Thus, to consider brand attitude as the only direct effect or consequence of ad-induced emotion or Aad is limiting. For familiar brands in particular, our understanding of how advertising "works" will need to move beyond the outcome variables traditionally examined in this research stream.

The purpose of this study is to propose and examine an additional "effect" or outcome of advertising exposure that may be affected by Aad. While research has focused on many antecedents of Aad (e.g., MacKenzie and Lutz 1989), the outcome variables examined have been limited to cognitive structure variables such as brand cognitions (MacKenzie, Lutz, and Belch 1986), brand attribute evaluations (Burke and Edell 1989), brand attitude, and purchase intentions; all these variables are likely to be very difficult to change with familiar brands.

We propose the construct "brand interest" as an additional outcome variable. Brand interest is defined as the level of interest or intrigue the consumer has in the brand and the level of curiosity s/he has to inquire or learn more about the brand. Unlike attitude, brand interest is not a cognitive evaluation of the brand. Moreover, brand interest may be viewed as a "pre-attitudinal" construct which, when acted upon, may facilitate attitude development or change. For example, a high level of brand interest may lead the consumer to search for more brand-related information or try the brand. For familiar brands, a high level of brand interest may lead the consumer to try the brand again and re-evaluate his/her beliefs; as a result, brand attitude might shift (Smith and Swinyard 1988). Indeed, for uninvolving, familiar brands, generating a level of brand interest through advertising that is high enough to lead to a "re-trial" may be the most effective means of attitude change, and thus a primary objective for the advertiser.

Sustaining and/or reviving interest may be particularly challenging for familiar stimuli (Berlyne 1960; Faison 1977). As Pessemier and Handelsman (1984) conclude, familiarity can lead to variety seeking and varied purchase behavior. When the individual desires the satisfying aspects of variety, alternation among familiar brands within a set is often exhibited as a choice behavior. McAlister and Pessemier (1982) review experiments on switching tendencies and note that among somewhat familiar elements within a favored subset of stimuli, subjects' liking increased with additional exposure to the stimuli, although boredom also increased with additional exposure. By alternating among the familiar stimuli, subjects could balance the liking and the boredom.

In the situation where there are various familiar brands to choose among within a product category, the brands constantly fight the consumer's inherent desire for novelty and varied brand purchases. It is this case in particular where increasing and sustaining brand interest should be critical for the advertiser. Additionally, it is likely that emotionally-oriented ad executions will have an impact on brand interest in this situation; for example, the positive feelings created by an ad could revive the intrigue or interest in the brand.

The question of how ad-produced feelings and Aad affect brand interest and brand attitude for familiar brands is what motivates this study. As a new construct, "brand interest" needs an empirical indicator before it can be integrated into the Aad research stream. Thus, we begin with a scale development exercise for the brand interest construct. We then present some exploratory modeling designed to illustrate the importance of considering brand interest as an alternative outcome of advertisement exposure.

METHODOLOGY

Procedure

Subjects participated in a study which was described as an evaluation of advertising campaigns. A videotape was prepared which included four different ad campaigns of three commercials each. Two campaigns were selected as the test campaigns-Central Trust Bank (a large local bank), and Lowenbrau beer. The Central Trust campaign included humorous portrayals of the problems that one can run into with banking services. The Lowenbrau campaign included three of the emotional, "Here's to good friends" ads. Central Trust Bank and Lowenbrau beer were selected since the subjects should have at least some familiarity with both. In addition, purchasing a brand of beer versus deciding on banking services represented considerable diversity in decision making processes, thereby providing the opportunity to examine the brand interest construct in two very different contexts .

One hundred two undergraduate marketing students were exposed to the videotape during the regular class sessions of two different courses. Questionnaires were distributed after the completion of the 10 minute videotape. Half of the randomly-distributed questionnaires included the measures for Central Trust Bank as the first set of questions, while in the other half the measures for Lowenbrau came first. Thus, all subjects completed the measures for both of the test campaigns.

Measurement

The questionnaire began with the brand interest measure followed by measures of positive and negative affect, and ad evaluation, for one of the ads in the Central Trust/Lowenbrau campaigns. These measures were followed by several "filler" questions (e.g., What is the best advertising campaign you have ever seen?, Why was it your all time favorite?). The filler questions were included to reduce consistency biases on the focal measures. Finally, the subject completed brand attitude and purchase intention scales.

With the exception of the newly developed brand interest scale (discussed below), all measures are consistent with those used in previous research. Positive/negative affect and ad evaluation were measured with the scales provided by Madden, Allen and Twible (1988). The coefficient alpha value for the positive affect measure (good, cheerful, pleased, stimulated, soothed) was .91 for Central Trust and .95 for Lowenbrau. For negative affect (insulted, irritated, repulsed), alpha values were .86 (Central Trust) and .90 (Lowenbrau). Coefficient alpha values for ad evaluation (pleasant/unpleasant, likeable/unlikable, interesting/boring, tasteful/tasteless, artful/artless, good/bad) were .89 for Central Trust and .92 for Lowenbrau.

Brand attitude was measured with three semantic-differential items: good/bad, unfavorable/favorable, poor/excellent. The coefficient alpha values for this scale were .93 for both Central Trust and for Lowenbrau. Purchase intention was also measured with a semantic-differential scale that included the items likely/unlikely, improbable/probable, and possible/impossible with alpha values of .90 and .91 for Central Trust and Lowenbrau, respectively.

RESULTS

The Brand Interest Measure

Eleven Likert-type scale items were generated for measuring the brand interest construct (see Table 1). The items were designed to capture the curiosity that individuals have about the brand, and their interest in learning more about the brand. In addition, several of the items assessed interest in the brand that may have been influenced by specific ad exposure.

Exploratory factor analysis. Principal components analysis for both Central Trust and Lowenbrau each resulted in a two-factor solution with all scale items loading on one factor except item number 1. A principal axis factor analysis with varimax rotation produced a slightly different result. For Central Trust, two factors emerged; the second factor included scale items number 3, 6, and 11 in addition to number 1. In the same analysis for Lowenbrau, a communality of a variable exceeded 1.0, thus terminating the factor extraction. Note that for Central Trust, the three additional items loading on the second factor (see items 3, 6, and 11 in Table 1) are the ones that reflect the ability of specific ads to create some interest in the brand, rather than the general level of interest the individual has in the brand. It appears as though two "forms" of brand interest may exist: one which is the level of interest that the familiar consumer has at any point in time, and another which is "ad-induced," and represents the impact of specific ads in altering brand interest. Depending on the research context, the more specific "ad-induced" brand interest factor might be relevant to the researcher. However, here we examine further the more general brand interest factor.

Confirmatory factor analysis. Given the above exploratory analysis, seven of the eleven items were retained for confirmatory testing (items 1, 3, 6, and 11 were eliminated). The seven items were evaluated with standard LISREL fit criteria as well as the criteria for discriminant validity (Dillon and Goldstein 1984; Fornell and Larcker 1981). Two additional scale items (4 and 7) were removed from the scale to improve the fit of the factor model. At this point, fit indices of the five remaining scale items were excellent. However, in tests of discriminant validity between brand interest and brand attitude, item number 2 of the brand interest scale had high normalized residuals with the brand attitude scale items. Indeed, by examining the wording of item 2, it appears that the respondent was being asked to make more of an evaluative judgment about the brand (consistent with the brand attitude measure) rather than making an assessment of his/her level of brand interest. Thus, item number 2 was also eliminated and the final brand interest scale was refined to include the four items shown in Table 2.

TABLE 1

ORIGINAL BRAND INTEREST SCALE ITEMS

TABLE 2

MAXIMUM LIKELIHOOD FACTOR LOADINGS

The fit values from the confirmatory analysis appear in Table 3. All of the values for Central Trust indicate an excellent fit of the four item scale to the data; the chi-square value is nonsignificant and the goodness of fit indices are very high. Additionally, the average variance extracted value is greater than the .50 rule of thumb (Dillon and Goldstein 1984), and the coefficient alpha value for Central Trust is well within an acceptable range.

For Lowenbrau, all fit indices are within acceptable ranges with the exception of the adjusted goodness of fit index (AGFI). However, the high average variance extracted values, the high factor loadings (Table 2), and the high coefficient alpha values (Table 3), all indicate that this four item scale is internally consistent and reliable for both Central Trust and Lowenbrau.

Discriminant validity. To be construct valid (Peter 1981), the brand interest measure must show discriminant validity from other more traditional measures. Here we examine measures of brand attitude, purchase intentions, positive affect, negative affect, and ad evaluation for evidence of discrimination from brand interest.

Table 4 reports the discriminant validity checks. Consistent with the rules of thumb provided by Dillon and Goldstein (1984), the average variance extracted values are all greater than .50 and, most importantly, are all greater than the squared structural link between the pairs of constructs. This analysis provides strong support for the discriminant validity of the brand interest measure.

TABLE 3

CONFIMATORY FACTOR ANALYSIS--FOUR ITEM BRAND INTEREST MEASURE

TABLE 4

DISCRIMINANT VALIDITY

Exploratory Modeling

Structural equation modeling was used to explore the relationships among affective responses to the ad (positive affect, negative affect, and ad evaluation), and brand interest, brand attitude, and purchase intentions. The model which was tested is shown in Figure 1. Note that this model includes the traditional links between positive affect, negative affect, ad evaluation and brand attitude, in addition to the brand attitude/purchase intention link. Brand interest is modeled as a potential mediator of ad exposure with effects on brand attitude and purchase intention.

Figures 2 and 3 present only the statistically significant model parameters for Central Trust and Lowenbrau. The models for Central Trust (chi-square=3.87, d.f.=3, p=.28; GFI=.99; AGFI=.91; RMSR=.03) and Lowenbrau (chi-square=.52, d.f.=3, p=.92; GFI=.99; AGFI=.99; RMSR=.01) fit the data extremely well. Note in Figures 2 and 3 that the affective responses to the ad have significant effects on brand interest and not on brand attitude. Here, brand interest appears to be mediating the effect of the Aad variables on brand attitude.

To compare the model in Figure 1 to a more traditional Aad network, an alternative model was also tested. This alternative was identical to that in Figure 1 except the brand interest construct and all corresponding links were excluded from the analysis. As a result, positive affect, negative affect, and ad evaluation all had direct links to brand attitude which, in turn, had a direct link to purchase intention. This alternate model also fit the data well for both Central Trust (chi-square=4.0, d.f.=3, p=.262; GFI=.98; AGF =.92; RMSR=.03) and Lowenbrau (chi-square .32, d.f.=3, p=.96; GFI=.99; AGFI=.99; RMSR=.005 ). For Central Trust, both positive affect and ad evaluation had significant effects on brand attitude (coefficients of .20 and .28, respectively). For Lowenbrau, only positive affect had a significant link to brand attitude (.36).

FIGURE 1

STRUCTURAL MODEL

FIGURE 2

STRUCTURAL COEFFICIENTS -- CENTRAL TRUST BANK

FIGURE 3

STRUCTURAL COEFFICIENTS -- LOWENBRAU BEER

DISCUSSION

The findings of nonsignificant direct Aad effects on brand attitude (as illustrated in Figures 2 and 3) implicate the mediating potential of the brand interest construct; however, an important caution must be noted in the interpretation of this modeling. Brand interest is modeled as directly affecting brand attitude, and acting as a mediator between Aad and Ab. However, the way in which brand interest and brand attitude are related is really an open issue. For example, a heightened level of brand interest may lead to a "re-trial" of the brand, followed by a shift in attitude. In such a case, brand interest should not be modeled as having a direct effect on brand attitude. Obviously, empirical assessment of such a process would require longitudinal data collection to disentangle the specific manner in which brand interest may ultimately impact brand attitude.

Whether Aad directly influences brand attitude for well-known brands remains an open question. The significant, direct effects of the Aad variables on brand attitudes in the alternative models is not consistent with previous studies which have shown little relationship between Aad and brand attitude for familiar brands. However, after checking familiarity levels of our subjects, we found that many were not very familiar with Central Trust and/or Lowenbrau. Familiarity was measured on a 1-7 semantic-differential scale (1=unfamiliar): mean familiarity ratings were 4.4 for Central Trust and 5.2 for Lowenbrau. Furthermore, a sizeable number of subjects were unfamiliar with the brands: 39% of the subjects had a familiarity rating of 3 or below for Central Trust, and 28% of the subjects rated Lowenbrau at a level of 3 or below. While this range of familiarity levels was desirable for developing the brand interest measure, it does not allow us to examine brand interest's role as a mediator for uniformly well-known brands, where one would expect the direct Aad/Ab relationship to be nonsignificant. Thus, a key research question remains: will brand interest serve as a mediator of advertising effects for highly-familiar brands? Additional research is needed to resolve this important issue.

Examining various consequences of changes in brand interest can also provide interesting research opportunities. It might be expected that heightening a consumer's brand interest for a familiar brand would lead to "re-trial" of that brand; thus, one would predict that brand interest might have some direct influence on purchase intentions. However, as shown by Figures 2 and 3, the model coefficients for a brand interest effect on purchase intentions are nonsignificant. These nonsignificant parameters may have resulted from the way in which the purchase intention construct was operationalized. For Central Trust, the respondent was asked the likelihood that s/he would open a new savings account with the bank. For Lowenbrau, the respondent was asked how likely s/he would be to purchase Lowenbrau to serve at a party. Since these intention measures represent a stronger purchase commitment than simple "re-trial" of a brand, it is not surprising that the brand interest/purchase intention links are nonsignificant. Research is thus needed to evaluate brand interest effects on specific measures of "trial intention." Indeed, examining the consequences of brand interest on constructs other than brand attitude may provide a more realistic picture of the effects of advertising exposure. For example, a temporary shift in brand interest due to ad exposure may also result in changes in variables such as variety seeking tendencies, and evoked set sizes, that are less "enduring" than attitude. The consequences examined in the Aad research stream should be expanded beyond brand attitude, especially when one is attempting to understand how advertising works for mature, well-known brands.

In sum, the brand interest construct may foster research on advertising's effect for familiar brands. Research challenges await in discovering the antecedents and consequences of shifts in brand interest. By developing a scale and illustrating the relevance of the construct, this study provides a foundation for such a research program.

REFERENCES

Batra, Rajeev and Michael L. Ray (1986), "Affective Responses Mediating Acceptance of Advertising," Journal of Consumer Research, 13 (September), 234-249.

Berlyne, D.E. (1960), Conflict, Arousal, and Curiosity, New York: McGraw-Hill.

Burke, Marian Chapman and Julie A. Edell (1989), 'The Impact of Feelings of Ad-Based Affect and Cognition," Journal of Marketing Research, 26 (February), 69-83.

Dillon, William R. and Matthew Goldstein (1984), Multivariate Analysis: Methods and Applications, New York: John Wiley & Sons, Inc.

Edell, Julie A. and Marian C. Burke (1986), "The Relative Impact of Prior Brand Attitude and Attitude Toward the Ad on Brand Attitude After Ad Exposure," in Advertising and Consumer Psychology, 3, Jerry Olson and Keith Sentis, eds,, New York: Praeger Publishers, 93-107.

Faison, Edmund W.J. (1977), "The Neglected Variety Drive: A Useful Concept for Consumer Behavior," Journal of Consumer Research, 4 (December), 172-175.

Fazio, Russell H. and Mark P. Zanna (1981), "Direct Experience and Attitude-Behavior Consistency," in Advances in Experimental Social Psychology, 14, Leonard Berkowitz, ed., New York: Academic Press, 161-202.

Fornell, Claes and David F. Larcker (1981), "Evaluating Structural Equation Models with Unobservable Variables and Measurement Error," Journal of Marketing Research, 18 (February), 39-50.

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

Gresham, Larry G. and Terence A. Shimp (1985), "Attitude Toward-Advertisement and Brand Attitudes: A Classical Conditioning Perspective," Journal of Advertising, 14 (1), 1017, 49.

Machleit, Karen A. and R. Dale Wilson (1988), "Emotional Feelings and Attitude Toward the Advertisement: The Roles of Brand Familiarity and Repetition," Journal of Advertising, 17 (3), 27-35.

MacKenzie, Scott B., Richard J. Lutz and George E. Belch (1986), 'The Role of Attitude Toward the Ad as a Mediator of Advertising Effectiveness: A Test of Competing Explanations," Journal of Marketing Research, 23 (May), 130-143.

MacKenzie, Scott B. and Richard J. Lutz (1989), "An Empirical Examination of the Structural Antecedents of Attitude Toward the Ad in an Advertising Pretesting Context," Journal of Marketing, 53 (April), 48-65.

Madden, Thomas J., Chris T. Allen, and Jacquelyn L. Twible (1988), "Attitude Toward the Ad: An Assessment of Diverse Measurement Indices Under Different Processing 'Sets'," Journal of Marketing Research, 25 (August), 242-252.

McAlister, Leigh and Edgar Pessemier (1982), "Variety Seeking Behavior: An Interdisciplinary Review," Journal of Consumer Research, 9 (December), 311 -322.

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

Pessemier, Edgar and Moshe Handelsman (1984), 'Temporal Variety in Consumer Behavior," Journal of Marketing Research, 21 (November), 435-444.

Peter, J. Paul (1981), "Construct Validity: A Review of Basic Issues and Marketing Practices," Journal of Marketing Research, 18 (May), 133-145.

Smith, Robert E. and William R. Swinyard (1982). "Information Response Models: An Integrated Approach," Journal of Marketing, 46 (Winter), 81 -93 .

Smith, Robert E. and William R. Swinyard (1988), "Cognitive Response to Advertising and Trial: Belief Strength, Belief Confidence and Product Curiosity," Journal of Advertising, 17 (3), 3-14.

Stayman, Douglas M. and David A. Aaker (1988), "Are All the Effects of Ad-Induced Feelings Mediated by Aad?," Journal of Consumer Research, 15 (December), 368-373.

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