The Effects of Ad Affect on Advertising Effectiveness

Danny L. Moore, University of Florida
J. Wesley Hutchinson, University of Florida
ABSTRACT - The nature of the relationship between affective reactions to advertising and advertising effects is examined. Immediately following exposure to print ads subjects indicated that they would be more likely to consider products associated with ads eliciting positive affective reactions than negative reactions. After seven days delay products associated with ads eliciting both positive and negative reactions generated greater change in brand consideration than neutral ads. The role of memory in determining the immediate and delayed effects of ad affect is discussed.
[ to cite ]:
Danny L. Moore and J. Wesley Hutchinson (1983) ,"The Effects of Ad Affect on Advertising Effectiveness", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 526-531.

Advances in Consumer Research Volume 10, 1983      Pages 526-531

THE EFFECTS OF AD AFFECT ON ADVERTISING EFFECTIVENESS

Danny L. Moore, University of Florida

J. Wesley Hutchinson, University of Florida

ABSTRACT -

The nature of the relationship between affective reactions to advertising and advertising effects is examined. Immediately following exposure to print ads subjects indicated that they would be more likely to consider products associated with ads eliciting positive affective reactions than negative reactions. After seven days delay products associated with ads eliciting both positive and negative reactions generated greater change in brand consideration than neutral ads. The role of memory in determining the immediate and delayed effects of ad affect is discussed.

INTRODUCTION

Theories of advertising effects based on multiattribute models assume that advertising influences brand attitudes by modifying consumers' beliefs about product attributes. While there is substantial support for the hypothesis that product beliefs affect brand attitudes (Holbrook, 1978; Lutz, 1975; Mitchell & Olson, 1981; Wilkie & Pessemier, 1973), there is growing evidence that product beliefs are not the sole mediator of advertising effects on brand attitudes. Current evidence suggests that affective reactions to the ad account for a significant amount of brand attitude variation beyond that accounted for by product-related cognitions (MacKenzie & Lutz, 1982; Mitchell & Olson, 1981). In short, there is emerging support for a model of advertising that includes both product beliefs and advertising reactions as independent mediators of brand attitude.

A diagram of two alternative paths through which advertising may affect brand attitudes is given in Figure 1. The first path in Figure 1 corresponds to the multiattribute model and postulates that beliefs about product attributes mediate the effect of advertising on brand attitudes (AB). The second path in Figure 1 represents the independent contribution of affective reactions to the ad (AAd). The primary purpose of the present paper is to examine this second path. In particular, we are interested in how affective reactions to advertising (AAd) influence attitudes toward the brand (AB), and the nature of the relationship between AAd and AB.

FIGURE 1

TWO MEDIATORS OF ADVERTISING EFFECTS ON ATTITUDE TOWARD THE BRAND

The Influence of AAd on AB

Prior speculations concerning the functional relationship between AAd and AB suggest either a linear or curvilinear relation. Shimp (1981) proposes that AAd is linearly related to AB such that the more positive consumers' reactions are to the ad, the more positive their reactions to the brand. The tacit assumption is that reactions to the ad generalize to the brand--perhaps through some conditioning process (Gorn, 1982). This hypothesis is not particularly novel in the advertising literature and is frequently referred-to as the "Superiority-of-the-Pleasant" hypothesis (Silk & Vavra, 1974).

The hypothesis that AAd is curvilinearly related to AB follows from the notion that extreme affective reactions can either impair or facilitate memory and attitude change. Such reactions may impair advertising effects by distracting consumers from processing brand-related information. If so, then one would expect an inverted-U relationship between AAd and AB, such that extreme affective reactions, regardless of valence, should suppress memory for the brand and inhibit change in AB.

The "Law of Extremes" hypothesis, in contrast to the distraction formulation, predicts a J-shaped relationship between AAd and advertising effectiveness measures (Silk & Vavra, 1974). Both liked and disliked ads should generate greater brand awareness and more favorable brand attitudes than neutral ads, with liked ads maintaining a slight edge over disliked ads. There at least two processes that could produce the predicted J-shaped relationship. First, ads that elicit extreme affective reactions may attract attention. If attention is directed to processing persuasive advertising copy, and if reactions to the ad do not transfer directly to the brand, then both liked and disliked ads should be more effective then neutral ads. Note that this line of reasoning assumes that ads eliciting strong affective reactions are more involving. A second process that could result in a J-shaped relation between AAd and advertising effectiveness assigns a critical role to brand familiarity--brand familiarity is used to refer to the salience of a brand in memory relative to other brands in the product category. The underlying rationale is that extreme affective reactions facilitate memory for the advertised brand. This increased familiarity, in turn, may increase liking for the brand (Zajonc, 1980); especially if reactions to the ad are not salient at the time of brand attitude measurement. Thus, affective reactions to the ad, regardless of valence, may facilitate the development of favorable attitudes toward the advertised brand if such reactions a) improve memory for the brand or facilitate processing of persuasive advertising copy and b) are dissociated with AR.

Silk and Vavra (1974) note that AAd cannot generalize to AB if the Law of Extremes is to hold. Otherwise, negatively valenced ads would produce a less robust advertising effect than neutral ads. Thus, the crucial questions are when and why should AAd become dissociated with a brand? Silk and Vavra propose that the association between a brand and advertising is forgotten over time and that this dissociation could lead to a "sleeper" Effect. Immediately following exposure to an ad, AAd is assumed to be closely associated with the brand, and therefore, partially determines AB. After some delay AAd is no longer spontaneously associated with the brand, but increased brand familiarity or processing of brand information generated by ads on the extremes of the AAd continuum produce more favorable attitudes toward the brand. In short, AAd may produce a direct effect on AB through a simple associative process or an indirect effect by facilitating memory for the brand or by facilitating processing of persuasive advertising copy.

Hypotheses

The brief discussion above suggest five basic hypotheses concerning the effect of AAd on AB. Each of these hypotheses is discussed below. The first three hypotheses discussed assume that the immediate and delayed effects of AAd do not differ. A distinction between immediate and delayed effects is made in two of the five hypotheses.

Generalization Hypothesis. According to the generalization hypothesis, affective reactions to the ad are associated directly with the brand through a conditioning process. If so, then AB should increase linearly with AAd.

Distraction Hypothesis. If distraction is the mediator of AAd effects, then advertisements eliciting strong affective reactions, regardless of valence, should impair brand memory and attitude change. This hypothesis suggests that AB is an inverted-U function of AAd.

Distinctiveness Hypothesis. If strong affective reactions to advertising increase memory for the advertised product, then brand attitudes may be more favorable for brands associated with ads eliciting intense affective reactions relative to ads eliciting little or no affective reaction. The assumption is that brands associated with positive or negative ads are distinctive in memory, and consequently, increased familiarity with the brand will lead to more favorable attitudes toward the brand. An implicit assumption of the distinctiveness hypothesis is that reactions to the ad and reactions to the brand are separated in memory. The distinctiveness hypothesis suggests a U- or J-shaped relation between AB and AAd.

Familiarity-based "Sleeper" Hypothesis. If AAd generalizes to AB, and if the influence of AAd on AB is initially strong, but weakens over time, then it is possible that increased brand familiarity resulting from ads that elicit strong affective reactions will eventually lead to increased favorability toward the advertised brand. Essentially, this hypothesis predicts that immediately after exposure to an ad the generalization hypothesis will hold, and that over time a J-shaped relationship will appear between AB and AAd. In addition, the familiarity-based sleeper hypothesis predicts that brands associated with either negative or positive ad affect will be remembered better than brands associated with neutral ads. However, overtime the direct influence of ad affect on AB should decay and the impact of brand familiarity should increase.

Affect-based "Sleeper" Hypothesis. The rationale underlying this hypothesis is that reactions to the ad and brand evaluations can initially be separated in memory. Moreover, it is assumed that AAd has little influence on AB immediately following exposure. Instead, initial brand evaluations following ad exposure are based upon beliefs about brand attributes and brand familiarity. Thus, if ads eliciting strong positive or negative reactions are attended to more than neutral ads, then one might expect a J-shaped relation between AB and AAd immediately following ad exposure. After some delay period, however, reactions to the ad and reactions to the brand may become confused in memory, and consequently, AB will be a positive, linear function of AAd. The affect-based sleeper hypothesis, in contrast to the familiarity-based hypothesis, assumes that the influence of product attributes and brand familiarity on AB is initially more important in determining AB than ad affect.

METHOD

Overview and Stimuli

Our chief concern was to determine whether reliable delayed effects of ad exposure would appear when naturally occurring stimuli were used. Silk and Vavra (1974) review studies dealing with the effect of affective reactions to advertising on effectiveness measures and find that most, if not all, of these studies assess ad effects immediately following exposure. Therefore, the effects of ad affect on memory and brand attitudes was measured at three time periods in the present study: l) Immediately following exposure, 2) two days after exposure, and 3) seven days after exposure.

Print ads were used as stimuli to simplify procedures and provide a preliminary test of the hypotheses before proceeding to a more detailed test that would include the medium as a factor. A total of 80 print ads were used in the study. These ads were chosen such that four ads from each of 20 product categories were represented in the stimulus sample. The ads were acquired from a wide variety of magazines and represented a diverse set of products (e.g., guitars, clothing, motorcycle helmets, rugs, bathroom fixtures, etc.).

Procedure

Forty-six students enrolled in a Marketing Research class at the University of Florida participated in the experiment-.- Participation was required as a class exercise.

The experiment was conducted in two stages. In the first stage all participants completed an initial brand questionnaire (Brand Test l). The questionnaire contained two questions about each of the 80 brands. The first question was a multiple choice test about knowledge of the brand's product category. The second question assessed brand attitudes indirectly by asking respondents to indicate the likelihood that they would consider the brand if they were in the market for a product in the category associated with the brand name. Immediately following Brand Test 1 participants were divided into two groups and each group was shown one-half of the 80 ads. Each advertisement was presented on a slide for approximately 10 seconds. After exposure to an ad participants completed two questions about the ad. First, they indicated whether or not they remembered seeing the ad previously, and secondly, whether they had a positive or negative emotional reaction to the ad. Responses to both questions were obtained on five-point scales ranging from definitely seen before to definitely never seen before for the first question, and from positive to negative for the second question. Immediately following this first Ad Test, respondents were given a second brand test about the full set of 80 produces. The questions on Brand Test 9 were identical to those on Brand Test 1. However, the order of brand presentation was randomized across brand tests. In summary, the first stage of the experiment involved three parts: 1) Brand Test 1, 2) Ad Exposure and Ad Test 1, and 3) Brand Test 2.

The second stage of the experiment represented the delay manipulation. Participants were asked to report back to the experimental rooms either two days following the ad exposure or seven days later. During this second session a third brand test was administered followed by a second ad exposure and ad test. Again, the questions on Brand Test 3 and Ad Test 2 were identical to those on previous tests but, the order of brand presentation was randomized for each test.

Dependent Measures

The major dependent variables were change in brand knowledge and brand consideration from baseline, i.e., Brand Test l. Ad recognition scores were also analyzed but, since this measure is lot central to the hypotheses, it will not be discussed further. Affective reactions to the ad collected on Ad Test l served as a classification factor in the analyses.

RESULTS

Analyses of the dependent measures were conducted in an analysis of variance where Delay Group (two-day delay or seven-day delay), Test (Brand Test 2 vs. Brand Test 3), and Ad Affect (positive, somewhat positive, neutral, somewhat negative, and negative) served as factors. The two primary dependent measures, i.e., brand consideration And brand knowledge, were computed by subtracting responses on Brand Test 1 from Brand Test 2 and 3.

Brand Consideration

The analysis of brand consideration change scores revealed a main effect for Ad Affect, F(4,154)=17.96, p<.0001. This effect reflected a linear trend in brand consideration change scores such that the more positive reactions were to the ad, the more likely the brand would be considered if subjects were in the market for the product. However, the effect of Ad Affect was modified by a significant Test x Ad Affect interaction, F(4, 154) = 3.85, p <.006, and a significant Delay Group x Text x Ad Affect interact: on, F(4, 154) = 2.56, p <.05. The significant Test x Ad Affect interaction indicated that change in brand consideration for levels of Ad Affect was not stable across Brand Tests. Inspection of the mean brand consideration change scores revealed that immediately following exposure to the ads brand consideration and ad affect were linearly related; the more positive reactions to the ad were, the more likely the brand would be considered. However, after a delay period, the relationship between ad affect and brand consideration was curvilinear. Brands associated with the negative and positive ads showed greater brand consideration change scores than brands associated with the neutral ads. In both cases the change scores were positive indicating that positive and negative ads increased the extent to which subjects would consider the brand for purchase. Thus, a J-shaped relationship between Ad Affect and change in brand consideration developed after a delay between exposure and assessment of brand consideration.

The three way interaction between Delay Group, Test, and Ad Affect indicated that the J-shaped relationship between Ad Affect and brand consideration was not constant across delay groups. Figures 2a and 2b illustrate the nature of this interaction. Examination of Figure 2a reveals that Ad Affect was linearly related to brand consideration change scores on both Brand Test 2 and Brand Test 3 for the two-day delay group. This suggests that with moderate delays ad affect generalizes directly to brand affect. Figure 2b shows that after seven days delay the relationship was no longer linear. Subjects in the seven-day delay group showed greater positive change scores in brand consideration as Ad Affect increased from negative to positive on Brand Test 2. On Brand Test 3 a U-shaped relationship between Ad Affect and change in brand consideration developed. Seven days after exposure negative and positive ads produced greater positive changes in brand consideration then neutral ads. Presumably, this could occur because the advertising copy was processed more thoroughly or because intense reactions to advertising make brand names more memorable. While we cannot separate these two explanations in the present paper, the brand knowledge data allowed us to address the plausibility or a familiarity-based sleeper hypothesis.

FIGURE 2

CHANGE IN BRAND CONSIDERATION FOR BRAND TEST 2 (NO DELAY) AND BRAND TEST 3 (DELAYED MEASUREMENT) BY DELAY GROUP

Brand Knowledge

Change scores for brand knowledge were constructed by subtracting the proportion of correct brand-product associations on Brand Test 1, i.e., before exposure to the ads, from the proportions obtained on Brand Tests 2 and 3. If ads generating positive or negative reactions are more effective than neutral ads, then the results for change in brand knowledge should reveal greater increments in brand knowledge for the positive and negative ads compared to neutral ads.

Analyses of change in brand knowledge indicated a U-shaped relation between Ad Affect and brand knowledge. However, a significant Delay Group x Text x Ad Affect interaction, F(4, 154) = 9.57, p <.05, reflected the fact that only the seven-day delay groups showed the expected U-shaped relation. Figures 3a and 3b graphically display the three-way interaction. As can be seen in Figure 3a, the change in brand knowledge was linear with respect to Ad Affect in the two-day delay group for both a Brand Test 2 and Brand Test 3. For the seven-day group, a U-shaped relation between Ad Affect and brand knowledge was observed at Brand Tests 2 and 3 (see Figure 3b). Thus, the results for brand knowledge are not entirely consistent with the brand consideration data. it appears that subjects remembered more about brands associated with positive or negative ads in the seven-day delay group. Whether or not this increased knowledge affected brand consideration judgments remains unclear. Nevertheless, the data do suggest that the "sleeper" effect for brand consideration may be due, in part, to the greater knowledge about brands associated with positive and negative ads.

FIGURE 3

CHANGE IN PROPORTION OF CORRECT BRAND-PRODUCT ASSOCIATIONS FOR BRAND TEST 2 (NO DELAY) AND BRAND TEST 3 (DELAYED MEASUREMENT) BY DELAY GROUP

DISCUSSION

The results of the present study indicate that affective reactions to advertising may have different immediate and delayed effects. Immediately following exposure to print ads subjects showed greater change in brand consideration as affective reactions became more positive. This same effect was observed two days after exposure to the ads. Thus, for short delays, change in brand consideration was a positive, linear function of ad affect. After seven days delay the relationship between change in brand consideration and ad affect was J-shaped, indicating that following a sufficient delay period the initial adverse effects of ads eliciting negative reactions may disappear and result in a greater probability of brand consideration;- These results would appear to support a familiarity-based "sleeper" hypothesis. However, the relationship between change in brand knowledge and ad affect observed in our study does not provide a strong confirmation of the familiarity hypothesis. If, as the familiarity hypothesis suggests, sleeper effects are caused by increased salience of a brand in memory and forgetting of the initial ad reaction, then the results for brand knowledge should have exhibited a UT-shaped relation with ad affect in all conditions. That is, ads eliciting a positive or negative reaction should have increased subjects knowledge about the brand more than neutral ads. Under this assumption, the fact that only subjects in the seven-day delay conditions showed a U-shaped relationship between ad affect measures and brand knowledge provides only marginal support for the familiarity-based sleeper hypothesis. It is possible, however, that the brand familiarity involved in sleeper effects is more closely related to brand name recognition than to brand knowledge. That is, familiarity may be a very basic psychological factor that is somewhat independent of whatever specific informational associations with the brand exist in memory. In this case, since brand knowledge is not a particularly good measure of this type of familiarity the obtained results are less problematic.

Several critical questions remain to be addressed in our work on the delayed effects of ad affect on brand attitudes. The most important of these questions concerns why and when "sleeper" effects occur in advertising. Research on the sleeper effect in the persuasion literature has been guided by the dissociative cue hypothesis (Cook, Gruder, Hennigan, & Flay, 1979). According to this hypothesis, non-message cues (e.g., message source) can initially facilitate or suppress attitude change in response to a highly persuasive message. Over time, however, the link between non-message cues and the attitudinal object presumably weakens while the persuasive message is retained. Consequently, forgetting non-message cues that initially suppress attitude change eventually leads to more favorable attitudes toward the message conclusion but, forgetting non-message cues that initially facilitate attitude change eventually leads to less favorable attitudes toward the message conclusion.

While the sleeper effect remains a source of controversy in the persuasion literature (Gillig & Greenwald, 1974), we feel that a somewhat modified version of the dissociative cue hypothesis may suggest some interesting delayed effects of advertising. Specifically, we have equated ad affect with non-message cues in an advertising setting and proposed that initial negative reactions may suppress favorable reactions to the brand, while initial positive reactions to the ad may facilitate favorable reactions to the brand. An important distinction between the sleeper hypothesis we advocate and the dissociative cue hypothesis lies in the assumed process underlying delayed attitude change effects in advertising. If ad affect draws attention to the product, then increased familiarity with the product may eventually lead to more favorable attitudes toward the product, regardless of the valence of the affective reaction to the ad. Unlike the dissociative cue hypothesis, we do not assume that the message, i.e., the advertising copy, is retained or even processed. The critical assumption in our work is that ad affect is forgotten more readily than brand familiarity. In short, immediately following exposure to an ad brand attitudes may be directly linked to ad affect but, after some delay brand attitudes may be more a function of brand familiarity than the initial affective reaction to the ad.

There are several important conditions under which the delayed and immediate effects of AAd and AB may not differ. Silk and Vavra (1974) suggest that repetition will strengthen the associative link between ad affect and the advertised product. Hence, repeating ads that generate negative reactions may increase brand familiarity but, any incremental effects of brand familiarity on brand attitudes will be overshadowed by the association between ad affect and the brand. The source of ad affect may be another important limiting condition of the sleeper effect in advertising. If negative reactions to the ad derive from the advertised product itself or from a poorly constructed advertising message, then it is unlikely that affective reactions will be dissociated with the brand. Therefore, research concerning the effects of repetition and the source of affective reactions to the ad is necessary before any firm conditions about the delayed effects of ad affect can be drawn.

CONCLUSIONS AND LIMITATIONS

Despite the significant differences between the immediate and delayed effects of ad affect on brand consideration, the results of the present study should be interpreted cautiously. Perhaps the most important limitation of the present study is that ad affect was measured rather than manipulated. Consequently, the basic findings are correlational and subject to caveats typically associated with correlational designs. In particular, it is impossible to determine whether affective reactions to the ad were the primary cause of the observed changes in brand consideration and brand knowledge. Additional analyses revealed that there was a significant correlation between ad affect and brand consideration ratings collected on Brand Test 1, i.e., before exposure to the ad; ads associated with brands subjects would consider elicited more positive reactions than ads associated with brands that were not rated highly on the consideration scale. Since this relation was observed for brand consideration measures obtained prior to ad exposure, it is reasonable to assume that reactions to the brand may have, in part, determined reactions to the ad. Clearly, research that separates the independent contributions of ad affect and brand reactions is needed.

A second critical limitation of the present study concerns the measures used. Throughout this paper we have assumed that brand consideration reflects attitude toward the brand and that brand knowledge reflects brand familiarity. It might be argued that brand consideration is coo distantly related to real purchase behavior. This is a classic problem with measures of brand attitude and we hope to replicate the effect with more proximal measures such as purchase likelihoods or, preferably, actual sales response in controlled field studies. As was discussed above, brand knowledge may not be a particularly good measure of brand familiarity. A priori associating the brand name with the appropriate product class seemed to be a minimal type of knowledge that would be closely related to brand familiarity. It could be that the type of familiarity that is important in the sleeper effect is more related to the initial memorial response to the brand name, and not to knowledge about the brand, per se.

Quite apart from the limitations of the study, the results suggest that the immediate and delayed effects of ad affect may differ. If this basic result can be replicated, it will hold a number of interesting implications about the basis for advertising effects. Specifically, reactions to advertising may influence brand attitudes directly or indirectly. A direct effect of AAd on AB may occur when ad affect is conditioned to the brand name. We would expect that the direct influence of ad affect will increase with repetition. Ad affect may also influence attitude toward the brand indirectly by increasing brand familiarity. If increased familiarity leads to more liking for the brand, then any variable that facilitates retention of the brand name should produce more favorable attitudes toward the brand. While there is support for the conjecture that familiar brands are held in higher regard then less familiar brands (Axelrod, 1968), research on brand familiarity effects is needed. Finally, the results of the present study underscore the role of memory in determining AB. In particular, two critical determinants of AB appear to Se the strength of associations between the brand and affectively valenced stimuli and the salience of the brand in memory. Research concerning the relative forgetting curves for various types of information that influence AB would shed light on potential differences between the immediate and delayed effects of ad affect.

REFERENCES

Axelrod, John N. (1968), "Attitude Measures That Predict Purchase," Journal of Advertising Research, 8, 3-17.

Cook, Thomas D., Gruder, Charles L., Hennigan, Karen M., and Flay, Brian R., (1979), "History of the Sleeper Effect: Some Logical Pitfalls in Accepting the Null Hypothesis," Psychological Bulletin, 86, 667-619.

Gillig, Paulette M. and Greenwald, Anthony G. (1974), "Is It Time to Lay the Sleeper Effect to Rest?," Journal of Personality and Social Psychology, 29, 132-139.

Gorn, Gerald J. (1982), "The Effects of Music in Advertising on Choice Behavior: A Classical Conditioning Approach," Journal of Marketing, 46, 94-101.

Holbrook, Morris B. (1978), "Beyond Attitude Structure: Toward the Informational Determinants of Attitude," Journal of .Marketing Research, 15, 545-556.

Lutz, Richard J. (1975), "Changing Brand Attitudes Through Modification of Cognitive Structure," Journal of Consumer Research, 4, 49-59.

MacKenzie, Scott B. and Lutz, Richard J. (1982), "Monitoring Advertising Effectiveness: A Structural Equation Analysis of the Mediating Role of Attitude toward the Ad," Working Paper Series, Center for Marketing Studies, University of California, Los Angeles, Paper No. 117.

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

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

Silk, Alvin J. and Vavra, Terry G. (1974), "The Influence of Advertising's Affective Qualities on Consumer Response," in Buyer/Consumer Information Processing, G. David Hughes and Michael L. Ray, eds. Chapel Hill, N.C.: University of North Carolina Press. 157-186.

Wilkie, William L. and Pessemier, Edgar A. (1973), "Issues in Marketing's Use of Multiple-attribute Attitude Models," Journal of Marketing Research, 10, 428-441.

Zajonc, Robert B. (1980), "Feeling and Thinking: Preferences Need No Inferences," American Psychologist, 35. 151-175.

NOTE

The research reported in this paper was supported by a summer Research Grant to the first author from the Center for Econometric Decision Sciences and the College of Business Administration at the University of Florida.

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