19
Toward a Better Understanding of the Role of Advertising Message Involvement in Ad Processing Russell N. Laczniak and Darrel D. Muehling Iowa State University ABSTRACT This study investigates the effect of advertising message involvement (AMI) on brand-related beliefs, attitude toward the ad (Aad), and their corresponding certainty dimensions. Support is found for the notion that belief strength is unaffected by involvement level, though brand- related beliefs are held with greater certainty for high- as compared to low-involvement individuals. Aad was also found to be more confi- dently held for high- versus low-AM1 receivers. However, contrary to expectations, AM1 had a direct effect on Aad as well. 0 1993 John Wiley & Sons, Inc. Over the years, numerous models/explanations of the consumer per- suasion process have appeared in the marketing and advertising lit- erature. For example, Greenwald (1968) and Wright (1973) recognized that cognitive responses may have considerable influence on attitude formation and change. Later, other authors (e.g., Lutz, 1985; Mitchell & Olson, 1981; Shimp, 1981) introduced and further developed the no- tion that brand attitudes may be a function not only of attribute-related beliefs, but also of affective reactions to the ad itself, termed attitude toward the ad (Aad). Smith and Swinyard (1982) modeled advertising response sequences and noted that attitudes and beliefs formed as a Psychology & Marketing Vol. lO(4): 301-319 (JulyiAugust 1993) 0 1993 John Wiley & Sons, Inc. CCC 0742-60461931040301-19 301

Toward a better understanding of the role of advertising message involvement in ad processing

Embed Size (px)

Citation preview

Toward a Better Understanding of the Role of Advertising Message Involvement in Ad Processing Russell N. Laczniak and Darrel D. Muehling Iowa State University

ABSTRACT

This study investigates the effect of advertising message involvement (AMI) on brand-related beliefs, attitude toward the ad (Aad), and their corresponding certainty dimensions. Support is found for the notion that belief strength is unaffected by involvement level, though brand- related beliefs are held with greater certainty for high- as compared to low-involvement individuals. Aad was also found to be more confi- dently held for high- versus low-AM1 receivers. However, contrary to expectations, AM1 had a direct effect on Aad as well. 0 1993 John Wiley & Sons, Inc.

Over the years, numerous models/explanations of the consumer per- suasion process have appeared in the marketing and advertising lit- erature. For example, Greenwald (1968) and Wright (1973) recognized that cognitive responses may have considerable influence on attitude formation and change. Later, other authors (e.g., Lutz, 1985; Mitchell & Olson, 1981; Shimp, 1981) introduced and further developed the no- tion that brand attitudes may be a function not only of attribute-related beliefs, but also of affective reactions to the ad itself, termed attitude toward the ad (Aad). Smith and Swinyard (1982) modeled advertising response sequences and noted that attitudes and beliefs formed as a

Psychology & Marketing Vol. lO(4): 301-319 (JulyiAugust 1993) 0 1993 John Wiley & Sons, Inc. CCC 0742-60461931040301-19

301

result of direct trial/experience with the brand may be more strongly held than lower-order ones formed as a result of exposure to advertising alone. Petty and Cacioppo (1986) proposed and empirically tested an elaboration likelihood model that suggests that brand attitudes may be formed via either central (message-based) or peripheral (less cognitive) routes, depending on the consumer’s motivation and ability to attend to and comprehend the ad’s claims.

Most theoretical explanations of the persuasion process have recog- nized that the means by which brand attitudes are formed or changed is likely to be affected/moderated by individuals’ level of involvement. As a result, many of the empirical tests of theoretical relationships have incorporated either a measure or manipulation of involvement. The present study adds to the growing body of involvement research by examining the effects of involvement on two primary influencers of brand attitudes, namely, beliefs and attitude toward the ad.

Interestingly, though past research has examined and supported in- volvement’s moderating effects on the relationship between &, beliefs, and brand attitudes (see, e.g., Gardner, 1985; Muehling & Laczniak, 1988; Park & Young, 1986), little is known concerning the effect in- volvement has on the antecedents of brand attitude. For example, it has been established that Aad is likely to influence brand attitudes in both high- and low-involvement situations, whereas beliefs influence brand attitudes primarily in high-involvement conditions. But research has not specifically examined whether beliefs and Aad, formed as a result of ad exposure, may be more strongly or weakly held as a result of individuals’ involvement with advertised claims. Our research seeks an answer to the question, “Does involvement affect the strength (mean level) of beliefs and Aad-factors known to influence brand attitudes?”

Furthermore, no research to date has examined the possibility that the certainty/confidence with which these beliefs and Aad are held may be influenced by involvement. Though beliefiattitude confidence has been studied in a variety of contexts, for example, ad exposure versus product trial (Smith & Swinyard, 1983, 1988), comparative versus non- comparative advertising (Droge & Darmon, 1987), single versus repeat exposures (Berger & Mitchell, 1989), it is not clear whether higher levels of message involvement will necessarily yield ad attitudes and beliefs held with greater certainty.

CONCEPTUAL FOUNDATIONS

Summaries of prior involvement research tend to suggest that placing the construct within a particular domain (e.g., ad message involvement, ad execution involvement, product involvement) and noting the number and types of personal connections between individuals and objects (An- drews, Durvasula, & Akhter, 1990; Krugman, 1965; Zaichkowsky,

302 LACZNIAK A N D MUEHLING

1985) seem most appropriate. In this research, our focus is primarily on ad message involvement, defined as a motivational state inducing message processing. This internal state variable has been usefully de- scribed by Andrews et al. (1990) as possessing three major properties: intensity (which refers to the degree of arousal or preparedness of the involved consumer), direction (the target of the involvement intensity level), and persistence (referring to the duration of the involvement intensity). With respect to the directional property, it should be noted that we make a distinction here between ad message involvement (i.e., involvement with the claims presented in the ad) and ad execution involvement (involvement with the contextual, noncontent aspects of the ad) (cf. MacKenzie & Lutz 19891, with the former being the focus of this study. Individuals high in ad message involvement, therefore, are likely to pay particular attention to message claims, to exert more mental effort in examining these claims, and to persist in this mental activity, more so than would individuals lower in ad message involve- ment.

We also recognize that these motivational states are likely to be influenced by relatively enduring personal factors such as individuals’ product class involvement. As such, this study differs from traditional studies which manipulate ad message involvement. Past studies often implicitly assume that innate predispositional factors (such as product involvement and knowledge) can be overridden by experimenter-gen- erated involvement instructions. Yet recent research suggests that such predispositions are central to determining whether an individual may be motivated and/or able to process the message components of an ad- vertisement (MacInnis & Jaworski, 1989). Thus, we chose to place sub- jects into involvement groups on the basis of product class involvement and product knowledge, that is, motivation and ability factors (MacInnis & Jaworski, 1989; MacInnis, Moorman, & Jaworski, 1991; Muehling, Laczniak, & Stoltman, 1991; Zaichkowsky 1986), while also paying particular attention to downstream, consequence factors often associ- ated and expected to be present with high/low levels of ad message involvement. This approach allows for an investigation of advertising processing when individuals are more “naturally” involved with an advertised message, thus reducing some of the intrusiveness and ar- tificiality often associated with studies manipulating involvement states.

The Relationship Between Product Class Involvement and Advertising Message Involvement

Product class involvement is expected to play a large role in influencing individuals’ level of advertising message involvement. High product class involvement is expected to be associated with such things as nar- rower latitudes of acceptance for product attributes, greater knowledge

INVOLVEMENT IN AD PROCESSING 303

of products and product features, more brands associated with the cat- egory memory node, greater ability to elaborate brand advertising claims, and greater confidence in brand evaluations, as well as a number of behavioral characteristics such as greater quantity and frequency of product consumption and heightened readership of and exposure to specialized media vehicles (see Antil, 1984 for a more complete listing of characteristics).

In turn, these product involvement related factors are expected to have a significant influence on ad message involvement, and hence on the message processing occurring at the time of ad exposure. For ex- ample, individuals who are involved with a product class should have a greater propensity to attend to message points presented in an ad because such processing behavior is consistent with their need to be informed. In addition, highly involved individuals may be more likely to attend to an ad’s message (rather than its creative or executional features) for purposes of evaluating the advertised brand because of the greater ego-related significance of subsequent purchases made in the product class (Gardner, Mitchell, & Russo, 1985). Similarly, these in- dividuals (as compared to individuals who are lower in product class involvement) should perceive messages contained in ads for products with which they are highly involved to be more personally relevant to their information needs (Greenwald & Leavitt, 1984). In addition, with enhanced motivation to attend, coupled with greater product knowl- edge, involved individuals should also be more willing and able to elab- orate on and process ad claims deemed personally relevant, due to their better developed product category schema.

Therefore, when an ad features a brand from a product class with which an individual is highly involved, it would be expected to produce heightened levels of ad message involvement, manifesting itself in terms of message attention, brand processing strategies, perceived mes- sage relevance, and cognitive elaboration. In turn, the nature of indi- viduals’ ad message involvement is expected to influence the process by which brand beliefs and ad attitudes are formed and/or changed.

Belief and Attitude Toward the Ad Certainty Consistent with the work of Zanna and Fazio (1982) and Berger and Mitchell (1989), the present research recognizes that ad attitudes may be made up of nonevaluative as well as evaluative components. In an ad processing context, the evaluative dimension corresponds to an in- dividual’s evaluation of the ad associated in memory with the repre- sentation of ads in general. Measures of attitude valence (the degree of favorability or unfavorability) used in advertising/marketing re- search often tap this evaluative dimension.

Confidence, on the other hand, is an example of a nonevaluative dimension (Berger & Mitchell, 1989) and refers to the conviction with

304 LACZNIAK AND MUEHLING

which the attitude or belief is held. As a point of clarification, Warland and Sample (1973) noted that two individuals may both indicate that they agree that taxes are “too high,” but disagree widely on their strength of conviction in judgment. Measures of confidence, therefore, should assess how certain or confident individuals are in their estimates of their beliefs.

Fishbein and Ajzen (1975) made a distinction between beliefs and belief confidence, noting that one may or may not be related to the other: The subjective probability an individual assigns to the likelihood that a brand possesses a certain attribute (belief) is not the same as the confidence with which this probability is assigned. As such, Bennett and Harrell (1975) proposed a multiplicative expectancy value model that included a confidence component, noting that inclusion of this nonevaluative dimension more fully captures subjects’ dispositions.

Smith and Swinyard (1988) have suggested, however, that the re- lationship between belief confidence and belief strength is complex and has not been fully sorted out in past research, noting that a t times measures of these constructs appear to represent the same cognitive dimension, and at other times they appear to be relatively distinct. Although an extensive discussion of this issue is beyond the scope of the present investigation, we believe that further insight may be gained if involvement effects are taken into consideration. Our a priori expec- tation is that involvement should not affect the composition of the con- structs, but may have a moderating effect on their relationship with one another.

HYPOTHESES

Models of persuasion, such as the elaboration likelihood model (ELM) of Petty and Cacioppo (1986) often emphasize that the processes by which attitudes are formed, not attitudes per se, are likely to be influ- enced by involvement. In fact, the ELM is not couched in terms of involvement’s direct effects on persuasion, but rather, views attitude formation or change to be a result of the influence of central and/or peripheral cues whose importance is increased or decreased as a result of involvement levels. The MacKenzie and Spreng (1992) finding that motivation (involvement) influences relationships between central and peripheral cues, rather than their mean levels, supports this notion. Similarly, MacInnis et al. (1991) recently noted that the favorability1 unfavorability of brand attitudes is not likely to be directly influenced by individuals’ level of processing (involvement), but rather, depends upon individuals’ reactions to various factors such as the nature of the arguments made in the ad, and past experiences and predispositions.

From this, we assume that involvement’s effects on brand beliefs and Aad may occur in a similar manner. Heightened involvement with an

INVOLVEMENT IN AD PROCESSING 305

ad’s message is not expected to increase individuals’ subjective proba- bility that a brand possesses a certain attribute (i.e., beliefs), nor is i t expected to affect their evaluations of an ad (Aad). Chattopadhyay and Nedungadi (19901, for example, observed no relationship between in- volvement and attitude toward the ad.

Our reasoning for this expectation acknowledges that highly involved individuals are more likely to cognitively elaborate on ad message points than will their low involvement counterparts. In other words, the production of self-generated thoughts is expected to be increased when individuals are highly involved. But the valence of these thoughts (i.e., whether individuals support or counterargue points made in the ad) should not be a direct result of involvement level. Although latitudes of acceptance for product attributes may be narrower for people deemed high in involvement, individuals’ reactions to the nature of the ad, and more specifically, the specific message contained in the ad (e.g., the attributes discussed, the method of presentation, etc.), will be instru- mental in shaping subsequent brand beliefs. In other words, brand- related judgments (beliefs) are more likely to be influenced by factors such as the strength of message arguments, the credibility of the source, and/or preexisting beliefs about the brand, rather than by the level of attention directed toward the ad.

A similar finding is expected for attitude toward the ad: Involvement may increase the attention drawn to various aspects of the ad, but whether individuals’ ad attitudes are positive or negative is dependent upon their idiosyncratic reactions to various ad elements. Such a notion is consistent with expectations derived from cognitive response theory (Brock & Shav- itt, 1983; Wright 1980). As a result, it is hypothesized that:

H1: Brand-related beliefs will not be significantly different for high advertising message involvement individuals than for low adver- tising message involvement individuals.

H2: Attitude toward the ad will not be significantly different for high advertising message involvement individuals than for low adver- tising message involvement individuals.

However, ad message involvement is expected to have an influence on the certainty with which beliefs and Aad are held. When individuals are in a high-involvement state, they are expected to be more connected to (Krugman, 1965) and elaborate on message aspects of the ad (Greenwald & Leavitt, 1984). Typically, high-involvement processes are also associated with deeper levels of processing (Leigh & Menon, 1987) and with the creation of self-generated thoughts (Greenwald & Leavitt, 1984). There- fore, it seems plausible to expect that, after ad exposure, individuals who have made judgments about the advertised brand (in terms of brand- related beliefs), or about the ad in general (Aad), are likely to have more mental reference points (memory traces) in high-involvement situations

306 LACZNIAK AND MUEHLING

(as a result of this elaborative processing activity), than will individuals who are in a lower advertising message involvement state.

Studies by Greenwald (1968) and Slamecka and Graf (1978) support the notion of ownness bias; individuals view their own thoughts as more unique, original, creditable, and of greater value than the thoughts of others (e.g., advertisers’ messages). Accordingly, if an object has high goal relevance and has been thought about by the individual, responses to questions concerning the object should be made with greater certainty (Warland & Sample, 1973). As such, confidence or certainty ratings are expected to be based on the relative centrality of the object to the indi- vidual, with high-involvement situations yielding judgments more central in nature.

Confidence in judgments about the brand or ad is expected to be en- hanced during high-involvement processing for several reasons. First, deeper processing (which occurs in high-involvement situations) makes more information about the object available to the individual (Fazio & Zanna, 1981). Thus, individuals may feel more confident in their judg- ments because they have accessed more information, thereby reducing the chance of making an incorrect judgment. Smith and Swinyard (1988) have suggested that quantity of information is an important antecedent of belief confidence.

In addition, self-generated thoughts (which are expected to be more prevalent in high-involvement situations) are more likely to be salient because they contain bits of information the individual has independently elected to access. Judgments based on such information may be held with greater certainty in an attempt to maintain a satisfactory level of cognitive consistency (cf. Abelson, Aronson, McGuire, Newcomb, Rosenberg, & Tan- nenbaum 1968). To do otherwise may be counterproductive for the indi- vidual.

Therefore, it is hypothesized:

H3: Belief confidence will be significantly different for high advertising message involvement individuals than for low advertising message involvement individuals, with high involvement individuals being more confident.

H 4 Attitude-toward-the-ad confidence will be significantly different for high advertising message involvement individuals than for low advertising message involvement individuals, with high involve- ment individuals being more confident.

METHOD

Subjects Students enrolled in an introductory marketing course at a major mid- western university were employed as subjects in the study (n = 110).

INVOLVEMENT IN AD PROCESSING 307

This sample was chosen largely because of convenience factors. How- ever, students also represent a potential target market for stereo speak- ers and it was expected that they would exhibit reasonably high levels of variability of involvement and knowledge with this product.

Experimental Ads

A professionally developed full-page ad for a fictitious brand of stereo speaker was used as the experimental stimulus. A neutrally valanced (Rossiter & Percy, 1980) fictitious brand (Maxium) was featured in the ad to control for effects of prior brand familiarity (Muehling & Laczniak, 1988).

The experimental ad, primarily factual in nature, contained supe- riority claims for the speaker brand and a half-page photograph of the product. Claims focused on five attributes (sound quality, attractive cabinet style, small size, wide range of sound, and low price) determined to be salient in a free-elicitation pretest (cf. Edell & Staelin, 1983). The ad was placed near the center of a 12-page booklet containing 8 pages of articles from a student-oriented magazine and three other full-page ads (for an automobile, pair of running shoes, and bicycle helmet).

Procedure

In an auditorium setting, subjects were presented experimental book- lets and were instructed to look at the booklets as they would normally read magazines. Based on pretests, a five-minute time period was al- lotted for all subjects to view the booklets. Upon completion of the viewing, booklets were collected and a postexposure questionnaire was administered. Numerous consumer response measures were gathered with the questionnaire. Those relevant to the present investigation included a cognitive-response elicitation exercise, several self-report ad message involvement measures, and measures of brand beliefs, belief confidence, attitude toward the ad, attitude-toward-the-ad confidence, as well as measures of product class knowledge and product involve- ment.

Cognitive responses were independently classified by three judges (who were unaware of the study’s purpose). A majority decision rule was used to classify thoughts into one of five categories: message re- lated, product related, brand related, ad related, or other (see Table 1 for descriptions). At least two judges agreed on a total of 97% of the classifications. The few remaining responses were categorized by the authors.

Study Measures Advertising Message Involvement (AMI) As opposed to artificially manipulating AMI, subjects were placed into “high” or “low” groups on

308 LACZNIAK AND MUEHLING

Table 1. Cognitive Response Categories Category Description

Message related Explicit references to specific attributes that were specified in either the verbal or visual content of the ad

General affective comments about the brand featured in the ad

Comments about the product class in general, not about the brand or any of its relevant attributes

Comments about the style, theme, execution, or format of the ad, including its creative aspects

All other thoughts, such as those relating to the task, those unlikely to have been generated during exposure but subsequently generated, and those unrelated to the message, brand, product, or ad

Brand related

Product related

Ad related

Other

the basis of their (more enduring) preexisting levels of product involve- ment and product knowledge- factors often linked with the motivation and ability antecedents of AM1 (cf. Andrews, 1988; Batra & Ray, 1986; Celsi & Olson, 1988; Muehling et al., 1991). Product class involvement was assessed using Zaichkowsky’s (1985) 20-item, 7-point bipolar ad- jective scale (Cronbach alpha = 0.97). Product knowledge was measured with a 15-item multiple-choice stereo knowledge test, derived from the work of Jacoby and Hoyer (1981). The knowledge score was com- puted by summing the number of correct responses to test questions (KR-20 = 0.68). Because the product involvement and knowledge mea- sures were correlated ( r = 0.43; p < 0.011, a more comprehensive in- volvement measure was derived by standardizing and summing the scores of the two measures. Subjects were placed into high- and low- involvement groups on the basis of a median split. Although there are certain limitations associated with allowing subjects to self-select into comparison groups, this approach has been used in other studies in- vestigating personal variables that are difficult to experimentally ma- nipulate (e.g., product class involvement and knowledge) (Snyder, 1974; Snyder & DeBono, 1985).

Belief Strength Belief strength was measured on 7-point scales an- chored by “extremely unlikely” (1) and “extremely likely” (7) (similar to the method advocated by Fishbein & Ajzen, 1975). For each of the five ad claims, subjects were asked to determine the likelihood that the advertised brand possessed the particular feature (e.g., sound quality, wide range of sound, etc.).

Belief Confidence Belief confidence was measured by asking subjects how “certain/confident” they were of their beliefs concerning each brand attribute. Seven-point bipolar adjective scales, with endpoints: Not at

INVOLVEMENT IN AD PROCESSING 309

all certain- Very certain, Not a t all confident-Very confident, were used. An index combining the two items was computed and used as the belief confidence measure for each attribute (Berger & Mitchell, 1989). Correlations for the belief confidence items corresponding to sound qual- ity, cabinet styling, size, sound range, and price were 0.89, 0.95, 0.96, 0.96, and 0.97, respectively.

Attitude Toward the Ad Mean responses to five sets of 7-point bipolar adjective scales (Not attractive-Attractive, Unappealing-Appealing, Dull-Dynamic, Depressing-Refreshing, Not enjoyable-Enjoyable) were used to assess subjects’ attitude toward the ad. Cronbach’s alpha for the summative index was 0.92.

Attitude Toward the Ad Confidence Attitude toward the ad confidence was measured by asking subjects to indicate “the degree of certainty1 confidence in their attitudes toward the ad,” with 7-point bipolar ad- jective scales (Not at all certain-Very certain, Not at all confident- Very confident). The correlation between these items was 0.93.

RESULTS

Justification of Grouping Procedure

To justify the procedure placing individuals into high- and low-involve- ment groups, several ad response measures were employed. These mea- sures included a 5-item index of self-reported message attention (e.g., “How much attention did you pay to the written message in the ad?” 1 = None, 7 = Very Much; Cronbach alpha = 0.921, a 5-item index of perceived use of a brand evaluation strategy (e.g., “I paid attention to what was stated in the stereo speaker ad so that I could evaluate the advertised brand:” 1 = Not at all, 7 = Very much; Cronbach alpha = 0.971, a 5-item index of self-reported ad message involvement (e.g., “When I saw the ad for stereo speakers, I concentrated on its contents:” 1 = Strongly disagree, 7 = Strongly agree; Cronbach alpha = 0.951, and a 10-item index of perceived message relevance (e.g., “When I saw the ad for the stereo speakers, I felt the information in it might be important to me:” 1 = Strongly disagree, 7 = Strongly agree; Cronbach alpha = 0.98). In addition, a count of the number of message-related cognitive responses was used to assess message involvement levels (Batra & Ray, 1983). MANOVA results were significant (Wilks’s lambda = 0.78; F5,99 = 5.47, p < 0.001), indicating the appropriateness of in- vestigating involvement differences on each manipulation-check vari- able.

Results of ANOVAs indicated that subjects classified as high involve- ment ( n = 50) generally reported having paid more attention to the

310 LACZNIAK AND MUEHLING

experimental ad (2.84 vs 1.80; F1,103 = 16.93, p < 0.001), were more likely to have reported using a brand evaluation strategy (3.07 vs 1.92; F1,103 = 13.24, p < 0.001), and found the experimental ad's message to be more involving (3.47 vs 2.07; F1,103 = 24.02, p < 0.001) and relevant (3.40 vs 2.27, F1,103 = 12.63, p < 0.01) than those classified as being in the low-involvement group ( n = 55) . In addition, subjects classified as high involvement generated more message-related cognitive responses than did those classified as low involvement (1.00 vs 0.49; F1,103 = 4.43, p < 0.05) subjects.

Relationship Between Product and Ad Message Involvement

As an additional means of demonstrating the relationship between prod- uct class involvement and ad message involvement, a correlation anal- ysis was conducted. The individual measures of product involvement and knowledge, as well as the combined index created from these items, were correlated with the ad response (advertising message involvement) measures used to justify the involvement grouping procedure. Table 2 represents the results of this analysis.

As can be seen from the table, in all but one case, the measures of product involvement and knowledge were significantly correlated with items designed to tap ad message involvement levels. Significant cor- relations ranged from 0.23 to 0.45.

Study Results Beliefs andBelief Confidence Table 3 provides the results of our anal- yses examining the effects of involvement on beliefs (Hypothesis 1) and belief confidence (Hypothesis 31, respectively. With regard to involve- ment's effect on beliefs, high- and low-involvement subjects' brand at- tribute related beliefs were not significantly different from one another

Table 2. Correlation Between Product Involvement and Ad Message Involvement Measures (n = 105)

AM11 AM12 AM13 AM14 AM15

PI1 0.38" 0.34 0.39" 0.40" 0.15 Know 0.23' 0.31" 0.37a 0.308 0.26' Combined 0.36" O.3ga 0.45" 0.41a 0.24b

"p < 0.01. AM11 = self-reported message attention AM12 = brand evaluation strategy AM13 = self-reported message involvement AM14 = perceived message relevance AM15 = total message-related cognitive responses PI1 = product class involvement measure Know = product class knowledge measure Combined = combined product involvement index (PI1 and Know)

hp < 0.05.

INVOLVEMENT IN AD PROCESSING 311

Table 3. Belief Strength and Belief Confidence Means and ANOVA Results

Involvement Level

Low High Attribute ( n = 55) ( n = 50) F

Belief strength Sound

Attractive quality 4.36 4.40 ns

cabinet 4.49 4.82 ns Small size 3.73 4.54 6.69* Sound range 4.42 4.32 ns Low price 4.04 4.52 ns

Sound quality 3.35 4.28 7.04a

Attractive cabinet 3.45 4.63 11.10"

Small size 3.58 4.52 6.65" Sound range 3.05 3.82 4.80b

Belief confidence

Low price 3.24 4.27 7.67*

for four of the five advertised attributes (the only exception being the belief corresponding to the size of the Maxium brand speaker). In other words, and in support of Hypothesis 1, belief strength was largely un- affected by advertising message involvement level.

Consistent with our a priori expectations, cognitive elaboration of message points (as measured by the total number of message-related cognitive responses generated) appeared to have very little influence on individuals' brand beliefs. Only one belief (regarding speaker size) out of the five measured was significantly correlated with the cognitive elaboration measure ( r = 0.22, p < 0.05).

However, involvement did have a significant impact on the confi- dence with which these beliefs were held. Each of the brand attribute related beliefs held by subjects in the high advertising message in- volvement group was held with greater confidence than the correspond- ing beliefs held by subjects in the low-involvement group. These find- ings provide empirical support for Hypotheses 3.

Furthermore, message-related cognitive elaboration did appear to have an influence on the certainty/confidence with which belief judg- ments were held. All but one of the belief confidence items (that for sound quality) were significantly correlated with the measure of message-related cognitive response.

Combined, the results suggest that, although the probabilities in- dividuals assign to the likelihood that a brand possesses certain attri- butes are largely unaffected by involvement levels, the confidence with which these brand/attribute linkages are assigned and held is signifi-

312 LACZNIAK AND MUEHLING

cantly influenced by involvement, with high-involvement processors being more certain of their beliefs. Consistent with expectations, in- creased cognitive elaboration of message points appear to have con- tributed to belief certainty while having little influence on belief strength.

Attitude Toward the Ad and Attitude Toward-the-Ad Confidence With regard to involvement’s effects on Aad and Aad confidence, mixed support for the hypotheses was obtained. Contrary to expectations, involvement did have an influence on subjects’ evaluations of the ad (4.13 vs 3.17; 8’1,103 = 15.76, p < 0.01). Subjects in the high-involvement group eval- uated the test ad more positively than did subjects in the low-involve- ment group, not supporting Hypotheses 2. As will be discussed later, this finding may be a reflection of the strong claims used in the exper- imental ad.

As expected and in support of Hypotheses 4, the ad attitudes of in- dividuals in the high ad message involvement group were held with greater certaintykonfidence than the ad attitudes of low-involvement individuals (5.45 vs 4.13; F1,103 = 4.80, p < 0.05). The confidence with which attitude toward the ad judgments were made, however, was not apparently influenced by cognitive elaboration of message points, in that the correlation between total message-related cognitive responses and the Aad confidence index was not significant ( r = 0.08). Therefore, although the results generally support the notion that involvement influences Aad confidence, the source of this influence cannot be directly attributed (in this study) to differences in cognitive activity directed toward the ad’s message.

DISCUSSION

The purpose of this study was to determine the effect of advertising message involvement on belief strength, attitude toward the ad, belief confidence, and attitude toward the ad confidence. As implied in pre- vious studies (e.g., MacInnis et al., 19911, our results provide empirical support for the notion that belief strength should not differ between high- and low-involvement ad receivers. This finding suggests that high levels of ad message involvement do not necessarily lead receivers to believe that a brand is more likely to possess a certain level of a par- ticular attribute.

However, our findings tend to suggest that beliefs formed by high- involvement receivers are held with greater certainty than those of low- involvement receivers. As such, they may resemble the higher-order beliefs (Smith & Swinyard, 1983, 1988) formed by persons who have engaged in product trial (i.e., relative to beliefs formed by less-involved individuals). Such a notion is reasonable in that higher levels of AM1

INVOLVEMENT IN AD PROCESSING 313

typically lead ad receivers to focus and elaborate on an ad‘s message for the purpose of evaluating the advertised brand (Gardner et al., 1985; Laczniak, Muehling, & Grossbart, 1989). Such focusing and elaboration results in deeper processing of advertised message points (Leigh & Menon, 19871, which places highly involved individuals in a better position to rely on their own reactions to an ad message when formu- lating beliefs (because elaboration is typically derived from the asso- ciation of incoming information with past product experiences stored in memory). As a result, it is not entirely surprising that the certainty with which these receivers formulate their beliefs is demonstrably stronger than those of less involved receivers.

Contrary to expectations, the results of this study also suggest that AM1 may have a direct effect on individuals’ Aad. In the present study, such a finding may be due, in part, to the fact that the ad contained relatively strong message claims (i.e., superiority claims dealing with the salient attributes of an ideal brand). Therefore, under high AMI, closer investigation of the ad may have led subjects to respond more positively, because the ad provided them with enough detail to make a more informed evaluation of the brand. This would be consistent with Lutz’s (1985) notion that ad evaluations may be formed via central, as well as peripheral, routes.

However, results did confirm expectations that higher involvement leads ad receivers to hold their ad attitudes with greater certainty. This may be the case since more involved receivers’ ad evaluations are likely to be based on more criteria and a more meaningful analysis than those of less involved persons.

Collectively, our findings have additional implications for research- ers and managers alike. As beliefs and Aad are commonly depicted as antecedents of brand attitudes (Muehling & Laczniak, 1988; Shimp, 1981), our results may provide an explanation for previous findings that deeper levels of processing produce more enduring, stable, and accessible brand attitudes (MacInnis, et al., 1991). Beliefs and Aad, held with more certainty, may result in more strongly held brand attitudes. This may be the case because high levels of AM1 likely allow receivers to more clearly define their evaluations of an object (Fazio & Zanna, 1981), as compared to low-involvement receivers whose lower-order be- liefs are likely the result of a cursory analysis of ad content. Thus, advertisers may need to pay special attention to the information needs of highly involved receivers. Although such receivers are not likely to formulate stronger beliefs toward the brand, they are likely to hold their postexposure beliefs with greater confidence. Because such beliefs are held with greater certainty, they may be resistant to change. As a result, marketers need to ensure that claims directed a t highly involved receivers will be sufficiently persuasive in order to facilitate the for- mation of strongly held beliefs.

314 LACZNIAK AND MUEHLING

In addition to past research findings that suggest that AM1 influences the relationships of brand attitudes with its antecedents (i.e., Aad and beliefs), the present study suggests that AM1 may influence the cer- tainty with which these responses are held. This finding may provide an explanation for why beliefs influence brand attitudes in high-AM1 conditions; beliefs formed by these subjects are held with greater cer- tainty. In addition, given the AM1 grouping procedure employed in the study, results also provide implicit support for the idea that AM1 may be a consequence of ad receivers’ motivation and ability to process ad information (Batra & Ray, 1986; MacInnis & Jaworski, 1989; MacInnis, et al., 1991).

However, although our operationalization of AM1 derives from pre- dispositional factors, it is not without limitations. Studies striving for more naturally occurring involvement levels often do so at the expense of less experimental control. As a result, cause-and-effect statements regarding key constructs may be challenged. Thus, this study may be best classified as falling within an investigational paradigm as opposed to an experimental one (Miller, 1971); alternative explanations may not be ruled out. Yet the study did follow the lead of previous research that allowed subjects to self-select into experimental groups on the basis of personal factors (e.g., Snyder, 1975), and did so with a relatively complex involvement measure that yielded empirically justified group- ings. This approach provides future researchers with an alternative means of operationalizing AMI. Those interested in testing the mod- erating effects of this construct could do so without imposing manipu- lations that rely on artificial means to override the natural predispo- sitions of individuals to become more (or less) involved with an ad’s message. At a minimum, this study empirically links two important domains of involvement.

The study results also reinforce the notion that strength and confi- dence (of both brand beliefs and Aad) may represent distinct constructs. Future research dealing with ad processing issues would benefit by considering both constructs. For example, many ad processing models (such as the dual mediation hypothesis of Lutz, 1985, and the cognitive responsekognitive structure model of Olson, Toy, & Dover, 1982), deal only with belief (and Aad) strength measures. However, our results tend to suggest that such models could be usefully augmented by including confidence measures as well (perhaps by using strength x confidence indices as was suggested by Bennett & Harrell, 1975 and Berger & Mitchell, 1989). In addition, the moderating effect of belief certainty on the relationship between brand cognitions and brand attitudes is worthy of further investigation, given that the linkage may be strength- ened (weakened) when belief confidence is high (low). At a minimum, confidence measures may allow researchers to improve the predictive validity of the belief and Aad concepts (Antil, 1983).

~

INVOLVEMENT IN AD PROCESSING 315

REFERENCES

Abelson, R. P., Aronson, E., McGuire, W. J . , Newcomb, T. M., Rosenberg, M. J., & Tannenbaum, P. H. (1968). Theories of cognitive consistency: A sourcebook. Chicago: Rand McNally.

Andrews, J. C. (1988). Motivation, ability and opportunity to process infor- mation: Conceptual and experimental manipulation issues. In M. J . Houston (Ed.), Advances in consumer research (Vol. 15, pp. 219-225). Provo, UT: Association for Consumer Research.

Andrews, J. C., Durvasula, S., & Akhter, S. H. (1990). A framework for con- ceptualizing and measuring the involvement construct in advertising re- search. Journal of Advertising, 19(4), 27-40.

Antil, J. H. (1983). Use of response certainty in attitude measurement. In R. P. Bagozzi & A. M. Tybout (Eds.), Advances i n consumer research (vol. 10, pp. 409-4151, Ann Arbor, MI: Association for Consumer Research.

Antil, J. H. (1984). Conceptualization and operationalization of involvement. In T. C. Kinnear (Ed.), Advances in consumer research (Vol. 11, pp. 203- 209). Provo, UT: Association for Consumer Research.

Batra, R., & Ray, M. (1983). Operationalizing involvement as depth and quality of cognitive response. In R. P. Bagozzi & A. M. Tybout (Eds.). Advances in consumer research (Vol. 10, pp. 309-313). Ann Arbor, MI: Association for Consumer Research.

Batra, R., & Ray, M. (1986). Situational effects of advertising repetition: The moderating influence of motivation, ability and opportunity to respond. Journal of Consumer Research, 12(March), 432-445.

Bennett, P., & Harrell, G. D. (1975). The role of confidence in understanding and predicting buyers’ attitudes and purchase intentions. Journal of Con- sumer Research, 2(September), 110-117.

Berger, I. E., & Mitchell, A. A. (1989). The effect of advertising on attitude accessibility, attitude confidence, and the attitude-behavior relationship. Journal of Consumer Research, 1 G(December1, 269-279.

Brock, T. C., & Shavitt, S. (1983). Cognitive-response analysis in advertising. In L. Percy & A. G. Woodside (Eds.), Advertising and Consumer Psychology (pp. 91-1 16). Lexington, MA: Lexington Books.

Celsi, R. L., & Olson, J. C. (1988). The role of involvement in attention and comprehension processes. Journal of Consumer Research, 15(September), 2 10-224.

Chattopadhyay, A., & Nedungadi, P. (1990). Ad affect, brand attitude, and choice: The moderating roles of delay and involvement. In E. Goldberg, G. Gorn, & R. W. Pollay (Eds.), Advances i n Consumer Research (Vol. 17, pp. 619-620). Provo, UT: Association for Consumer Research.

Droge, C., & Darmon, R. Y. (1987). Associative positioning strategies through comparative advertising: Attribute versus overall similarity approaches. Journal of Marketing Research, 24(November), 377-388.

Edell, J., & Staelin, R. (1983). The information processing of pictures in print advertisements. Journal of Consumer Research, 10(June), 45-61.

Fazio, R. H., & Zanna, M. P. (1981). Direct experience and attitude-behavior consistency. In L. Berkowitz (Ed.), Advances in Experimental Psychology (Vol. 14, pp. 161-202). New York: Academic Press.

316 LACZNIAK AND MUEHLING

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: A n introduction to theory and research. Reading, MA: Addison-Wesley, 1975.

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.

Gardner, M. P., Mitchell, A. A., & RUSSO, J . E. (1985). Low involvement strat- egies for processing advertisements. Journal of Advertising, 14(2), 4-13.

Greenwald, A. G. (1968). Cognitive learning, cognitive response to persuasion, and attitude change. In A. G. Greenwald, T. C. Brock, & T. M. Ostrom (Eds.), Psychological foundations of attitudes (pp. 147-170). New York: Academic Press.

Greenwald, A. G., & Leavitt, C. (1984). Audience involvement in advertising: Four levels. Journal of Consumer Research, 581-592.

Houston, M. J., & Rothschild, M. L. (1978). Conceptual and methodological perspectives on involvement. In S. C. Jain (Ed.), Research frontiers in mar- keting: Dialogues and directions (pp. 184-187). Chicago: American Market- ing Association.

Jacoby, J., & Hoyer, W. D. (1981). What if opinion leaders didn’t know more? A question of nomological validity. In K. B. Monroe (Ed.), Advances in con- sumer research (Vol. 8, pp. 299-303). Ann Arbor, MI: Association for Con- sumer Research.

Krugman, H. (1965). The impact of television advertising: Learning without involvement. Public Opinion Quarterly, 29(Fall), 349-356.

Laczniak, R. N., Muehling, D. D., & Grossbart, S. (1989). Manipulating mes- sage involvement in advertising research. Journal of Advertising, 18(2), 28- 38.

Leigh, J. H., & Menon, A. (1987). Audience involvement effects on the infor- mation processing of umbrella print advertisement. Journal of Advertising,

Lutz, R. J. (1985). Affective and cognitive antecedents of attitude toward the ad: A conceptual framework. In L. F. Alwitt & A. A. Mitchell (Eds.), Psy- chological processes and advertising effects: Theory, research, and applica- tions (pp. 45-63). Hillsdale, NJ: Lawrence Earlbaum Associates.

MacInnis, D. J., & Jaworski, B. J . (1989). Information processing from adver- tisements: Toward an integrative framework. Journal of Marketing, 53(0ctober), 1-23.

MacInnis, D. J., Moorman, C., & Jaworski, B. J. (1991). Enhancing and mea- suring consumers’ motivation, opportunity, and ability to process brand in- formation from ads. Journal of Marketing, 55(0ctober), 32-53.

MacKenzie, S., & Lutz, R. J . (1989). An empirical examination of the structural antecedents of attitude toward the ad in an advertising pretesting context. Journal of Marketing, 53(2), 48-65.

MacKenzie, S. B., & Spreng, R. A. (1992). How does motivation moderate the impact of central and peripheral processing on brand attitudes and inten- tions? Journal of Consumer Research, 18(4), 519-529.

Miller, G. R. (1971). Research setting: Laboratory studies. In P. Emmert & W. D. Brooks (Eds.), Methods of research in communication (pp. 77-104). New York: Houghton Mifflin.

16(3), 3-12.

INVOLVEMENT IN AD PROCESSING 317

Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitude? Journal of Marketing Research, 18(August), 318-332.

Muehling, D. D., & Laczniak, R. N. (1988). Advertising’s immediate and de- layed influence on brand attitudes: Considerations across message-involve- ment levels. Journal of advertising, 17(4), 23-43.

Muehling, D. D., Laczniak, R. N., & Stoltman, J. (1991). The moderating effects of ad message involvement: A reassessment. Journal of Aduertising, 20(2),

Olson, J. C., Toy, D. R., & Dover, R. A. (1982). Do cognitive responses mediate the effects of advertising content on cognitive structure? Journal of Con- sumer Research, 9(3), 245-262.

Park, C. W., & Young, S. M. (1986). Consumer response to television com- mercials: The impact of involvement and background music on brand atti- tude formation. Journal of Marketing Research, 23( 1),11-24.

Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes to attitude change. New York: SpringerIVerlag.

Rossiter, J. R., & Percy, L. (1980). Attitude change through visual imagery in advertising. Journal of Advertising, 9(2), 10-16.

Shimp, T. A. (1981). Attitude toward the ad as a mediator of consumer brand choice. Journal of Advertising, 10(2), 9-15 + .

Slamecka, N. J., & Graf, P. (1978). The generation effect: Delineation of a phenomenon. Journal of Experimental Psychology: Human Learning and Memory, 4(6), 592-604.

Smith, R. E., & Swinyard, W. R. (1982). Information response models: An integrated approach. Journal of Marketing, 46(Winter), 81-93.

Smith, R. E., & Swinyard, W. R. (1983). Attitude-behavior consistency: The impact of product trial versus advertising. Journal of Marketing Research,

Smith, R. E., & Swinyard, W. R. (1988). Cognitive response to advertising and trial: Belief strength, belief confidence and product curiosity. Journal of Advertising, 17(3), 3-14.

Snyder, M. (1974). Self-monitoring of expressive behavior. Journal of Person- ality and Social Psychology, 30(4), 526-537.

Snyder, M., & DeBono, K. G. (1985). Appeals to image and claims about quality: Understanding the psychology of advertising. Journal of Personality and Social Psychology, 49(3), 586-597.

Warland, R. H., & Sample, J. (1973). Response certainty as a moderator vari- able in attitude measurement. Rural Sociology, 38(2), 174-186.

Wright, P. (1973). The cognitive processes mediating acceptance of advertising. Journal of Marketing Research, 38(2), 192-205.

Wright, P. (1980). Message-evoked thoughts: Persuasion research using thought verbalizations. Journal of Consumer Research, 7(September), 151- 175.

Zaichkowsky, J. L. (1985). Measuring the involvement construct. Journal o f Consumer Research, 12(3), 341-352.

Zaichkowsky, J. L. (1986). Conceptualizing involvement. Journal of Advertis- ing, 15(2), 4-14.

29-38.

20(August), 257-267.

~ ~~ ~~

318 LACZNIAK AND MUEHLING

Zanna, M. P., & Fazio, R. H. (1982). The attitude behavior relation: Moving toward a third generation of research. In M. P. Zanna et al. (Eds.), Consistency i n social behavior: The Ontario symposium (Vol. 2, pp. 283-301). Hillsdale, NJ: Lawrence Earlbaum Associates.

This research was partially funded by a summer research grant provided to the first author from the College of Business, Iowa State University. Russell N. Laczniak is Associate Professor of Marketing, College of Business, Iowa State University, Ames, IA 50011. Darrel D. Muehling is Associate Professor of Marketing, College of Business and Economics, Washington State Univer- sity, Pullman, WA 99164-4730.

INVOLVEMENT IN AD PROCESSING 319