Ad skepticism: the consequences of disbelief.

Advertising is an industry that involves billions of dollars per year (nearly $250 billion for 2003 [McCann-Erickson 2003]), employing bright minds with access to state-of-the-art theory in human behavior. Despite its sophistication, subtlety, vast resources, and ubiquity, advertising faces

Sorting through ad claims, many of which are difficult to substantiate, requires effort, and consumers are generally not highly motivated to process advertising information (MacInnis, Moorman, and Jaworski 1991). It is, therefore, reasonable that simple disbelief may be one way that consumers cope with the persuasive attempts of advertising. Friestad and Wright (1994) describe how a person's knowledge of persuasion strategies influences his or her responses to persuasion attempts. Both consumers and marketing agents have access to three categories of knowledge: The consumer has knowledge of the topic (or product), knowledge of persuasion (beliefs about how persuasion occurs and what tactics are used or are effective), and knowledge of the agent (advertiser or manufacturer). The marketing agent, in turn, has knowledge of the topic, knowledge of persuasion, and knowledge of the target (the consumer segment). The marketing agent relies on knowledge bases to develop a persuasion attempt. The consumer, relying on a set of knowledge, counters with a persuasion-coping response. Although the relationship between marketer and consumer is not always adversarial, we suspect that the most frequent coping response may be to discount messages recognized as ads.

Empirical evidence suggests that advertising claims (in ads that are processed) are frequently not accepted. Calfee and Ringold (1994) reported the consistent observation in public opinion polls that roughly two-thirds of consumers claim they doubt the truthfulness of ads. Previous research (Obermiller and Spangenberg 1998, 2000) supports the proposition that consumers are socialized to be skeptical toward advertising, and the extent of their skepticism is a determinant of their responses to advertising.

Obermiller and Spangenberg (1998) defined skepticism toward advertising (ad skepticism) as the tendency to disbelieve the informational claims of advertising. They developed and validated a scale to measure the construct (SKEP), and demonstrated empirical support for a hypothesized nomological network. Their work proposed that ad skepticism is a marketplace belief; demonstrated its relation to attitude toward advertising and attitude toward marketing; and identified its consequence to certain personality characteristics, marketplace experiences, consumer socialization, and education (Obermiller and Spangenberg 2000). Ad skepticism was also shown to relate to global measures of evaluative responses to advertising. (1) We extend that work in the present study, focusing on the consequences of ad skepticism on advertising.

AD SKEPTICISM AND RESPONSES TO ADVERTISING

We propose that ad skepticism influences a set of responses to advertising, one of which is to bypass advertising and gather product information from other sources. Other moderated responses include processing of the ads, formation of brand beliefs, attitude toward the ad, responsiveness to certain advertising tactics, and purchase intention. We expect that consumers who are high in ad skepticism would show weaker brand beliefs consistent with advertising claims, more negative attitudes toward ads, different responses to emotional versus informational appeals, and lower intentions to purchase the advertised products. Furthermore, higher skepticism should be related to less attention to advertising and greater attention to nonadvertising sources of information. We justify these hypothesized consequences as follows: Because ad skepticism reflects a tendency not to believe advertising, the disbelief in specific ad claims follows directly. Because virtually all ad claims are positive, less belief in those claims should lead to less positive attitudes and purchase intentions. Because high ad skepticism is reflective of a general distrust in the validity of advertising, consumers with high skepticism should be inclined to ignore ads and seek product information from other sources. Finally, because ad skepticism reflects a disbelief in the information content of ads, consumers with high skepticism should be less responsive to information-based appeals (as compared to consumers with low skepticism).

Our investigation of the influences of ad skepticism is in accord with other research on individual differences as moderators of responses to advertising. Many studies have examined immediate or temporary antecedent variables, some of which may be influenced by the ad itself, such as mood or ad-induced feelings (see Meuhling and McCann 1993). Other studies have also examined effects of more stable consumer dispositions, similar to ad skepticism, including attitude toward the ad/advertiser/advertising (Homer 1990; Miniard, Bhatla, and Rose 1990), beliefs about specific ad elements (Moore, Harris, and Chen 1995), prior brand attitudes (Chattopadhyay and Basu 1990), brand-related beliefs (Mittal 1990), demographics and lifestyles (Green 1999; Grief and Brumbaugh 1999), product involvement (Rose, Miniard, and Bhatla 1990), and various personality traits (e.g., Schoenbachler and Whittler 1996; Zhang 1996). Generally, these studies have found predictable relations: In relative senses, responses to ads are less positive when ad or advertiser attitudes or beliefs are negative, when prior brand attitudes or beliefs are negative, when consumers are older or better educated (possibly a consequence of increased ad skepticism), and when product involvement is lower (though the latter effect has been shown to be moderated by a number of ad factors).

Consistent with these findings, we expect that consumers with relatively higher skepticism toward advertising should exhibit less positive responses to ads.

Skepticism about advertising has both social and individual antecedents. Earlier researchers (among others, Darby and Karni 1973; Ford, Smith, and Swasy 1988; Nelson 1970) developed the notion of information economics in advertising, arguing that consumers tend to value information that is perceived to be useful and valid. Because advertising is associated with selling and tends toward exaggerating, especially as claims are more difficult to substantiate, consumers are socialized to be skeptical. That social antecedent is entirely consistent with our consideration of skepticism as an individual difference. While acknowledging the general, social effects on skepticism, we focus on differences in individuals' willingness to believe advertising claims. Finally, our hypotheses are based largely on the discussion in Obermiller and Spangenberg (1998) of the nomological relation between ad skepticism and these other factors. In the current research, moreover, we examine a greater range of consequences than in previous studies.

STUDY 1: SKEPTICISM AND IMMEDIATE RESPONSES TO ADS

Ad skepticism should influence specific, immediate responses to ads. In particular, we expect more skeptical consumers to like ads less and find their claims less believable, less influential, and less informative. To test these hypotheses, we assessed the ad skepticism of 140 students (65 M.B.A.s, 75 undergraduates), and four weeks later, presented them with a series of 13 print ads. Ad skepticism was measured with the SKEP scale, a nine-item scale that rates advertising on the following dimensions: getting the truth, aim is to inform, is informative, is generally truthful, is a reliable source of information, is truth well told, presents a true picture, provides essential information, and leaves one feeling accurately informed (see Obermiller and Spangenberg 1998 for exact wording). After each ad, the participants rated the ad on several five-point scales: "In general, I like the ad very much," "In general, I find the ad very believable," "I think other people would be very likely to be influenced favorably by this ad," and "I found the ad informative." Participants also indicated their interest in the product category (very interesting to me/ not at all interesting to me) and the relevance of the product category for them (very important to me personally/not at all important to me personally). Finally, participants identified any specific claims they believed were untrue.

The ads were full-color, one-page print ads taken from popular magazines. The products were Whirlpool appliances, gold jewelry, Kenmore dryer, Pony athletic shoes, Amoco Ultimate gasoline, Land O'Lakes butter, Colt automobile, Mont Blanc pen, Murray lawn mower, Mazda Miata automobile, Isuzu Truck service division, Dewar's Scotch whiskey, and Invisible Fencing. The ads were displayed as color slides for about one minute each; participants were instructed to view them carefully and were given one minute to respond to the first three scales (like, believe, and influential). The ads were then presented again while participants completed the remaining scales (informative ad; interesting, important product; untrue claims). Two orders of exposure were used to control for order effects.

Study 1 Results

Average responses to the series of ads were computed. SKEP scores were a sum of the items, corrected for direction. Although the reliability and validity of the scale have been demonstrated elsewhere (Obermiller and Spangenberg 1998, 2000), we assessed and found adequate internal reliability with Cronbach's [alpha] (.825) and confirmatory factor analysis (CFA) (CFI [comparative fit index] = .906, GFI [goodness-of-fit index] = .908, RMSEA [root mean square error of approximation] = .092, RMR [root mean squared residual] = .047, [[chi square].sub.(27)] = 59.00, p < .01). (2)

Correlations between SKEP and the three ad response measures were significant and negative (higher scale values indicated greater skepticism and more positive responses to the ads), as predicted. The pair-wise correlations were SKEP-Like, r = -.23; SKEP-Believable, r = -.37; and SKEP-Influential, r = -.27 (all p's < .01). Participants who were more skeptical toward advertising in general found the 13 ads, on average, less likable, less believable, and less likely to be influential.

The total number of claims identified as untruthful was summed for each participant across the 13 ads as an indication of the extent of perceived falseness of the collection of ads. This total ranged from zero (for 67 participants) to 15. The total number of perceived false claims was correlated with SKEP at r = .50 (p < .01), indicating that more skeptical participants identified more ad claims as false than did less skeptical participants.

Ad skepticism is a general characteristic, and these results support its relation with responses to ads in general (at least, across 13 ads). We should expect, however, that individual respondent or ad variables, such as product involvement and perceived information content, might moderate the responses. Skeptical consumers might be more receptive to messages about products they were interested in and more receptive to ads with less information (since the information is disbelieved). We investigated these hypothesized moderating effects by contrasting products identified as high and low in involvement and information content.

Our measures of "interest in the product" and "perceived personal relevance of the product" were highly correlated (across the 13 ads, average r = .72, p < .01). Based on this close association, the two were combined into a measure we called "product category involvement" (PCI), with an average rating across product categories of 4.03 (on a scale ranging from 2 to 10 points). We identified the ads with the highest and lowest average PCI. The ads for gold jewelry, Isuzu truck service, Dewar's Scotch, and Invisible Fencing were all rated below 3.50. Ads for Pony athletic shoes, Colt automobile, and Mazda Miata automobile all rated above 4.80. The relations between SKEP and ad responses for these two sets of ads are compared in Table 1.

We conducted a similar breakdown of the ads according to the ratings of informativeness (very informative/not at all informative). The 13 ads had an average informativeness rating of 3.24. Three ads had distinctly higher ratings (all above 4.00)--Kenmore dryer, Colt automobile, and Isuzu truck service-and four ads had distinctly lower ratings (all below 2.40)--gold jewelry, Land O'Lakes butter, Mont Blanc pen, and Dewar's Scotch. The relations between SKEP and ad responses were examined for these two sets of ads, with the results also presented in Table 1.

Product category involvement appeared to have had some effect on the relation between ad skepticism and responses to ads. Skepticism had slightly more explanatory power for liking of the ad and perceived influence of the ad when PCI was higher.

On the other hand, perceived informativeness did not appear to moderate the effect of ad skepticism on responses to the ads. It is noteworthy that perceived presence of information in the ads did not result in more positive responses by highly skeptical participants--even for high-information ads, SKEP was negatively correlated with positive responses. To the extent that skepticism measures a tendency to disbelieve the information claims in ads, we might expect that SKEP would be a better predictor of responses to ads for the most informative ads. However, we observed a significant difference only on perceived influence, where participants with high skepticism rated the more informative ads as less influential.

We further investigated the moderating effects of PCI and perceived informativeness by analyzing responses for individual ads. For each ad, we examined the following equation:

[AR.sub.i] = [alpha]([PCI.sub.i]) + [beta](SKEP) + [delta]([PCI.sub.i] x SKEP),

where [AR.sub.i] was a measure of ad response for ad i. [PCI.sub.i] was the measure of the participant's product category involvement, and SKEP was the participant's ad skepticism score. A statistically significant contribution of the interaction term ([delta]) would indicate a moderating effect of PCI on the influence of ad skepticism.

Across the 13 ads, the liking, believable, and influential measures were highly intercorrelated; the 39 bivariate r's were all between .50 and .79. For efficiency, we created a single measure of ad response (sum of like, believable, and influential items; analyses of separate measures yielded essentially the same results). Regression analysis of the equation resulted in statistically significant models for 12 of the 13 ads. The interaction term was statistically significant for two ads, as illustrated in Figure 1, which shows ad response as a function of ad skepticism and product category involvement, split at their mean values. In both cases, PCI had little effect for participants with low skepticism. For participants with high skepticism, however, ad response was relatively negative if PCI was low, but relatively positive when PCI was high. Thus, for some products, participants with high skepticism had as positive responses to ads as participants with low skepticism, provided the products were important or interesting to them.

[FIGURE 1 OMITTED]

We repeated the regression analysis, using perceived information content in the ad as the moderating variable, testing for an interaction between SKEP and perceived information. The interaction term was statistically significant for three of the ads at p < .05, and another at p < .06; these interactions were not consistent, however. For two of the products, participants with low skepticism responded better to high information, and for the other two, participants with high skepticism responded better to high information.

Study 1 Conclusions

Skepticism toward advertising related as expected to important responses to advertising--attitude toward the ad, believability of ad claims, perceived influence of the ads, and perceived untruths in the ads. Evidence of a moderating effect of product category involvement or perceived information level was mixed. The influence of skepticism on ad responses was slightly stronger for more involving products. For some of those ads, involvement with the product tended to mitigate against the negative effects of higher skepticism. The effect of perceived information content was less clear: Although the moderating effect of skepticism on the perceived influence of ads, as a group, was stronger when more information was perceived, perceived informativeness had inconsistent effects on individual ads.

The moderating effect of PCI and perceived informativeness were investigated with two types of analysis. Both methods had shortcomings. One was to contrast the strengths of the correlation of SKEP with response measures for the ads that, over the sample, were regarded as high and low in PCI and informativeness. This method averaged responses over groups of three or four ads. Not all participants agreed on the involvement or informativeness of all ads, however. The second method eliminated the potential problems associated with averaging by focusing on individual ads, but at the cost of statistical power and the combination of dependent measures into a composite. In general, analysis of the relations was limited by the quasi-experimental design; product involvement and perceived informativeness were measured, not manipulated.

STUDY 2: SKEPTICISM AND BEHAVIORAL RESPONSES TO ADVERTISING

The first study indicated that ad skepticism influenced thoughts and feelings about ads and advertised products. Study 2 extended the investigation to more distant and behavioral responses. To clarify the ambiguous evidence from Study 1 regarding the role of skepticism in response to perceived information content of ads, we included direct measures of responses to ad appeals based on information and entertainment. We expected that because participants with high skepticism expect less information value from advertising, they would pay less attention to ads, rely less on ads in decision making, make more efforts to confirm ad claims when they did process them, and buy fewer products in response to ads. Study 2 tested the relation between SKEP and these consequences of advertising.

Participants were 104 adult professionals aged 26 to 53 who responded to an initial questionnaire containing the SKEP scale and demographic information. Three to six weeks later, the same participants were contacted with a second questionnaire containing the items shown in Table 2. We hypothesized that high skepticism toward advertising would be related to avoiding advertising (more "zipping and zapping" [items 1 and 2] and less attention [item 6], less reliance on ads in decision making [item 3], more reliance on Consumer Reports and specialty magazines [items 4 and 5], more reliance on friends for confirmation of ad claims [item 7], and less purchasing prompted by advertising [items 11 and 12]). Despite Study 1 results, which showed no systematic effects of perceived information content of ads, we expected participants with high skepticism to prefer more informative and more straightforward ad executions (items 8-10). Given our findings above, this expectation was more tentative than the others. If highly skeptical consumers are disposed toward processing the information content of ads, we should expect that even though they may not believe the claims of ads, they prefer ads with information appeals (perhaps they feel they are best equipped to deal with this type of ad). Such a preference was suggested by Kanter and Wortzel (1985), who proposed that "cynical" consumers prefer straightforward informative appeals.

Study 2 Results

The reliability of the SKEP scale was again assessed by Cronbach's [alpha] (.923) and CFA (CFI = .946, GFI = .887, RMSEA = .108, RMR = .094, [[chi square].sub.(27)] = 59.80, p < .01). (3)

Correlations between SKEP and the 12 dependent measures are shown in Table 2. The results support the hypothesis that ad skepticism is associated with avoiding advertising; higher skepticism was related to zipping, zapping, and ignoring ads (items 1-3). The results also supported the hypothesis that skeptics were less likely to make purchases solely on the basis of advertising information (items 4, 8, and 9). On the other hand, the results were mixed regarding skeptics' uses of nonadvertising information. No significant associations were found for use of Consumer Reports or other specialty magazines (items 5 and 6), but skeptics did report they were more likely to check out the truth of ad claims by asking friends or trying the product (item 7). Finally, there were no significant associations between SKEP and trust in or preference for celebrity spokespeople, infomercial structure, or noninformation appeals (items 10-12).

Study 2 Conclusions

Study 2 related SKEP scores to self-report measures of responses to advertising. The results generally conformed to the hypothesis that skepticism leads to less attention to and reliance on advertising. Evidence of an association between skepticism and preference for other sources of information, however, was mixed, which may reflect different levels of measurement specificity. The two items that showed nonsignificant correlations (items 5 and 6) referred to specific sources, relative to the more general measure in item 7--checking out the truth through friends or trial. Since previous evidence has shown no relation between trust in advertising and trust in Consumer Reports (Obermiller and Spangenberg 2000), it is unlikely that consumers with high ad skepticism are automatically distrustful of objective, nonadvertising information sources. Perhaps reliance on magazine reviews is determined less by skepticism and more by product interest and involvement. On the other hand, across interest levels, asking a friend or trying the product are trusted alternatives to seeking information from an ad.

The results did not support the hypothesis that skeptics have a preference for or against any type of ad appeal (items 10-12). These results support the notion that SKEP is not a measure of consumer cynicism. It has been suggested that cynical consumers may be especially vigilant against overtly manipulative attempts (Campbell 1995; Kanter and Wortzel 1985). Participants with high skepticism showed no disinclination toward celebrity spokespeople, infomercial structure, or "soft-sell" ads (emphasis on pictures, music, or humor). (4)

STUDY 3: SKEPTICISM AS A MODERATOR OF AD APPEAL

Because Studies 1 and 2 provided only limited insight into the relations between ad skepticism and ad appeals, Study 3 was designed as a test of the hypothesized relation between ad skepticism and informational versus emotional appeal. We characterized these two types of appeals according to Puto and Wells (1984). Informational ads predominantly provide relevant brand data, clearly and logically. Emotional (Puto and Wells use the term "transformational") ads predominantly provide an emotional experience that is relevant to the use of the brand. According to Puto and Wells, successful ads must be effective on either the emotional or informational dimensions (possibly both).

We hypothesized that ad skepticism would more strongly affect responses to informational (relative to emotional) appeals. Since ad skepticism is defined in terms of believability and several of SKEP's items refer specifically to the information content of ads, we expected more skeptical consumers to respond more negatively to and be less persuaded by informational ad appeals. Our expectation is consistent with Friestad and Wright (1994), who proposed that emotional appeals are developed by marketers specifically to circumvent consumers' skeptical resistance to informational arguments. Thus, we hypothesized that ad skepticism would moderate the effectiveness of ad appeals, such that higher skepticism would be associated with less positive responses to informational ads (relative to emotional ones).

The stimuli were eight television commercials selected to represent informational and emotional appeals. The advertised products were Allstate life insurance, Ernest and Julio Gallo wine, Goodyear tires, Hallmark cards, Joy dishwashing liquid, Puffs tissues, Sudafed cold medicine, and Vaseline petroleum jelly. The first four products were selected as emotional appeals (death of father leads to loss of home, "family" atmosphere at the winery, brothers reunited after many years, and big brother returns home for Christmas, respectively). The latter four represented informational appeals (how Joy removes baked-on foods, most people prefer Puffs for colds, Sudafed works faster, and Vaseline has many uses). As a check on the face validity of the appeals, a pretest group of 27 participants viewed the ads and rated them on the Puto and Wells Information and Transformation Scale. The results supported the categorization. All eight ads rated higher on the expected dimension. Furthermore, all four of the "emotional" ads rated lower in average information than all four "informational" ads. (For the four "emotional" ads, the mean of the eight information dimension items was 2.75 versus 3.13 [on seven-point scales] for the "informational" ads [p < .05].) Three of the four "emotional" ads rated higher on the transformation dimension than all four informational ads. (For the four "emotional" ads, the mean of the 15 transformation dimension items was 2.97 versus 2.83 [on seven-point scales] for the "informational" ads [p < .05].) The Allstate ad, as the anomaly, was rated near the midpoint of the range of transformation dimension scores of the four informational ads. Thus, pretest results supported face validity of our categorization--four predominantly informational and four predominantly emotional ads.

Participants were 54 M.B.A. students who responded to the SKEP scale during class time. Four weeks later, the participants were shown the eight ads. Two presentation orders were used, each of which mixed the informational ads with the emotional ads. After each ad, participants responded to two, seven-point brand attitude items (very bad/very good and very undesirable/very desirable) and to a thoughts-listing question. (The number and valence of thoughts were coded by two research assistants who were blind to the hypotheses, with disagreements resolved by one of the authors.) Valence for each thought was coded as negative, neutral, or positive.

Study 3 Results

After finding no statistically significant differences, the data were pooled across presentation orders. The SKEP scale was again assessed with Cronbach's [alpha] and CFA. In this study, the scale demonstrated relatively lower internal reliability ([alpha] = .808; CFI = .771, GFI = .821, RMSEA = .148, RMR = .079, [[chi square].sub.(27)] = 58.36, p < .01). (5) The low internal reliability results in Study 3 were mitigated by evidence of good reliability in the other studies and other published reports, as well as the subsequent structural equations analysis results. To test the hypothesis that skepticism would moderate ad responses, we examined the structural model illustrated in Figure 2. The X variables are the nine SKEP items. The set of _G items are the "good-bad" dependent measures, and the _D items are the "desirable-undesirable" dependent measures.

[FIGURE 2 OMITTED]

Study 3 allowed us to assess the effect of ad skepticism on evaluation of ads with either informational or emotional appeal. Ad evaluations were measured with the two attitude measures (good-bad and desirable-undesirable). The structural equation model used to examine the relations is shown in Figure 2. The model permitted correlation between the two evaluation measures for given products. The results of the analysis, using AMOS 4.0 (Arbuckle and Wothke 1999), are presented in Table 3.

By the standard of measures of fit, this recursive model does not fit the data well, which merely suggests that the model is underspecified. It is clear that more than ad appeal and ad skepticism are needed to provide complete predictions of responses to ads. The crucial finding for our purpose was the strong relations between SKEP and the attitude measures for the two types of ad appeal. In response to informational appeals, the higher the participant's ad skepticism, the more negative the attitude toward the advertised products (standardized regression weight = -.556). On the other hand, in response to emotional appeals, the higher the participant's ad skepticism, the more positive the attitude toward the advertised products (standardized regression weight = .615). Simple correlational analysis of the valence of thoughts listed showed a similar effect: r = -.52 for the average thought evaluation of information appeal ads, and r = +.42 for the emotional appeal ads (both the latter p's < .01). (The number of thoughts listed was unrelated to ad skepticism, either across all ads or categorized by type of appeal.)

The nature of this interaction can be illustrated by consideration of the mean attitude scores for the four conditions, splitting at the SKEP median: For emotional ads, the high and low skeptic groups had average attitudes, respectively, of 5.1 and 4.5. For informational ads, on the other hand, those respective attitudes were 4.1 and 4.8. Contrast tests indicated that informational ads evoked different responses from high and low skeptics (4.1 versus 4.8; [t.sub.(52)] = 3.41, p < .01), and high skeptics differed in response to informational versus emotional ads (4.1 versus 5.1; [t.sub.(52)] = 3.68,p < .01). Other contrasts were not statistically significant. Thus, those participants who had lower ad skepticism did not differ significantly across ad appeals. Highly skeptical participants, on the other hand, were more positive in response to emotional appeals as compared with informational appeals, and were more negative in response to informational ads compared with less skeptical participants. These results are consistent with the hypothesis: Informational appeals were ineffective with participants who were highly skeptical toward ads. This ineffectiveness was apparent both as a contrast between informational and emotional appeals for highly skeptical participants and as a contrast between the responses of high and low skeptics to informational appeals.

Study 3 Conclusions

Results supported the hypothesis that ad skepticism would moderate responses to ad appeals, such that participants with high skepticism would be less responsive to informational appeals. In general, participants with low skepticism did not differ across ad appeals. Furthermore, participants with high skepticism did not differ from participants with low skepticism in response to emotional appeals. The significant interaction was driven by the poor response of participants with high skepticism to informational appeals. Our hypothesis was based on the expectation that consumers with high skepticism would reject the informational claims, leaving the informational appeals with little persuasive power. The results also suggest that high skeptics are not only less persuaded by informational ads than are low skeptics; they are also less persuaded by informational ads than they are by emotional ads.

Study 3 provides a test of the nomological validity of the ad skepticism network of Obermiller and Spangenberg (1998). The experimental manipulation of ad appeal provides a stronger test of ad appeal than the correlational analyses of self-report data in the previous studies. The results also support the central premise of Friestad and Wright's (1994) persuasion knowledge model, which is that advertisers and consumers are engaged in a continuous interaction. Highly skeptical consumers have likely become skeptical over time, in response to numerous interactions in the marketplace that have led them to disbelieve ad claims. Advertisers, in turn, have developed strategies for approaching these skeptical consumers, including emotional appeals, whose success does not require acceptance of informational claims.

GENERAL DISCUSSION AND CONCLUSIONS

The hypotheses we tested were motivated in large part by the nomological network proposed by Obermiller and Spangenberg (1998) and their call for further investigation of the relation between ad skepticism and advertising variables. Our results are generally consistent with that work, providing further evidence of the construct validity of the SKEP scale and extending the legitimacy of their nomological framework. Skeptics were shown to be generally less positive in response to advertising; they like it less, believe it less, and believe it is less influential. They avoid advertising when they can and report less connection between advertising and purchasing. Although skeptics indicated more reliance on friends and trial, they showed no greater use of specific objective sources of product information.

The major contribution of the present study is the observed failure of informational appeals with highly skeptical participants. The advertising skeptic regards advertising as not credible and, therefore, not worth processing. The skeptic's perspective differs from the common sense of a consumer cynic. As more commonly conceptualized, the cynical consumer is critical of advertising because of its manipulative intent and indirect appeals. Such consumers may prefer simple, direct, informative advertising; skeptics, on the other hand, do not. Although it may seem counterintuitive, advertisers are not apt to "win over" skeptics by presenting them with simple informational appeals. Study 1 looked at the moderating effect of skepticism for ads that were perceived to differ in informational content and showed no evidence that participants favored ads that were perceived to contain more information. In Study 2, skeptics reported no preference for or against noninformational appeals. Study 3 manipulated informative versus emotional appeals and found that informative appeals were less persuasive with skeptics. Thus, advertisers may be advised to avoid direct informational approaches with skeptics in favor of emotional appeals, which were shown to work equally well for high and low skeptics, and no worse than informational appeals for low skeptics.

Skeptics are not immune to advertising, however. In addition to an openness to emotional appeals (Study 3), Study 1 provided limited evidence that highly skeptical consumers respond well to ads for products for which they are highly involved. Future research with improved measures or manipulations of product involvement may investigate this question further. The implications for advertisers, however, may be limited, suggesting merely that marketers have a special need to find ways to show how their products are important and relevant to skeptical consumers. Furthermore, to the extent that skeptics rely on friends for information, advertising may have indirect effects by influencing their friends. In general, our results reinforce the need for advertisers to employ a variety of approaches to succeed with the broad spectrum of consumer responses.

The evidence that skeptics were more persuaded by emotional than informational ad appeals may appear inconsistent with the observation that skeptics do not prefer emotional appeals. On the one hand, this may simply mean that asking people what they want is not a good way to predict what will be effective. On the other hand, an additional possibility is that ad skepticism, for the highly skeptical, is a salient aspect of the consumer self. When people view themselves as "skeptical," perhaps they feel the need to respond in a manner consistent with that personality characteristic, and correspondingly manage external impressions (Tedeschi, Schlenker, and Bonoma 1971) or prefer to affirm a position that makes them feel best about themselves (Steele 1988). For skeptics, that may mean that a general "No" to advertising is a common response to any prompt, even though some ad appeals and ads for some products may, in fact, be effective.

We should acknowledge an important limitation of this research, which is the time horizon. We examined reports or behaviors immediately subsequent to the ad. Consumer response to advertising may take place well after exposure to the ad. Of particular relevance on this point is some evidence that, over time, people may come to believe what they initially regarded as unbelievable (Gilbert, Krull, and Malone 1990). Moreover, Maloney (2000) has reported that disbelief, per se, may represent a truly negative response to advertising, but "curious disbelief" may not. To the extent that consumers do not believe ad claims but remain curious about them, they may be positively inclined toward further investigation or trial. The relation between ad skepticism and curious disbelief should be investigated further.

One final comment pertains to the social implication of advertising skepticism. It has widely been presumed that some level of consumer skepticism is "healthy" for the marketplace because it encourages honesty from advertisers. The studies reported here made no attempt to establish norms for skepticism, so we cannot draw conclusions about whether the levels of skepticism we observed were "healthy." We suggest, however, that skepticism is both a cause that encourages advertisers to be honest and an effect of consumer experience of dishonesty. In the latter case, too much skepticism is an impediment to an efficient market. Useful information cannot succeed in the marketplace if consumers do not believe it and act upon it. Future research might investigate the public policy implications of skepticism by investigating its relation to negative market behaviors, such as shopping inefficiency (Are skeptics worse shoppers?); complaining (Do skeptics complain more, with less justification?); late adoption (Are skeptics late to switch to improved products?); ignoring credible claims (Do skeptics reject all informational claims, or just puffery?); and rejection of public service information (Do skeptics ignore product warnings, recall notices, etc.?).

REFERENCES

Arbuckle, James L., and Werner Wothke (1999), Amos 4.0 User's Guide, Chicago: SmallWaters Corporation.

Bagozzi, Richard P., and Youjae Yi (1988), "On the Evaluation of Structural Equations Models," Journal of the Academy of Marketing Science, 16 (1), 74-94.

Calfee, John E., and Debra J. Ringold (1994), "The Seventy Percent Majority: Enduring Consumer Beliefs About Advertising," Journal of Public Policy and Marketing, 13 (Fall), 228-238.

Campbell, Margaret (1995), "When Attention-Getting Advertising Tactics Elicit Consumer Inferences of Manipulative Intent: The Importance of Balancing Benefits and Investments," Journal of Consumer Psychology, 4 (3), 225-254.

Carmines, Edward G., and John P. McIver (1981), "Analyzing Models with Unobserved Variables," in Social Measurement: Current Issues, G. W. Bohrnstedt and E. F. Borgatta, eds., Beverly Hills: Sage.

Chattopadhyay, Amitava, and Kunal Basu (1990), "Humor in Advertising: The Moderating Role of Prior Brand Evaluation," Journal of Marketing Research, 27 (November), 466-476.

Darby, Michael, and Edi Karni (1973), "Free Competition and the Optimal Amount of Fraud," Journal of Law and Economics, 16 (1), 67-88.

Ford, Gary T., Darlene B. Smith, and John L. Swasy (1988), "An Empirical Test of the Search, Experience, and Credence Attributes Framework," in Advances in Consumer Research, vol. 15, M. Houston, ed., Provo, UT: Association for Consumer Research, 239-243.

Friestad, Marian, and Peter Wright (1994), "The Persuasion Knowledge Model: How People Cope with Persuasion Attempts," Journal of Consumer Research, 21 (June), 1-31.

Gilbert, Daniel T., Douglas S. Krull, and Patrick S. Malone (1990), "Unbelieving the Unbelievable: Some Problems in the Rejection of False Information," Journal of Personality and Social Psychology, 59 (4), 601-613.

Green, Corliss (1999), "Ethnic Evaluations of Advertising: Interaction Effects of Strength of Ethnic Identification, Media Placement, and Degree of Racial Composition," Journal of Advertising, 28 (Spring), 49-64.

Grier, Sonya, and Anne Brumbaugh (1999), "Noticing Cultural Differences: Ad Meanings Created by Target and Non-Target Markets," Journal of Advertising, 28 (Spring), 79-93.

Homer, Pamela, and Lynne Kahle (1990), "Source Expertise, Time of Source Identification, and Involvement in Persuasion: An Elaborative Processing Perspective," Journal of Advertising, 19 (1), 30-39.

Kanter, Daniel, and Lawrence Wortzel (1985), "Cynicism and Alienation as Marketing Considerations: Some New Ways to Approach the Female Consumer," Journal of Consumer Marketing, 2 (Winter), 5-15.

MacInnis, Debra, Christine Moorman, and Bernard Jaworski (1991), "Enhancing and Measuring Consumers' Motivation, Opportunity, and Ability to Process Brand Information from Ads," Journal of Marketing, 55 (October), 32-53.

Maloney, John C. (2000), "Curiosity Versus Disbelief in Advertising," Journal of Advertising Research, 40 (6), 7-13.

McCann-Erickson (2003), Bob Coen's Insider Report, available at www.kelseygroup.com/econtext archives/econtext_040128.htm.

Miniard, Paul, Sunil Bhatla, and Randall Rose (1990), "On the Formation and Relationship of Ad and Brand Attitudes: An Experimental and Causal Analysis," Journal of Marketing Research, 27 (August), 290-303.

Mittal, Banwari (1990), "The Relative Roles of Brand Beliefs and Attitude Toward the Ad as Mediators of Brand Attitude: A Second Look," Journal of Marketing Research, 27 (May), 209-219.

Moore, David, William Harris, and Hong Chen (1995), "Affect Intensity: An Individual Difference Response to Advertising Appeals," Journal of Consumer Research, 22 (September), 154-164.

Muehling, Darrel, and Michelle McCann (1993), "Attitude Toward the Ad: A Review," Journal of Current Issues and Research in Advertising, 35 (Fall), 25-58.

Nelson, Paul (1970), "Information and Consumer Behavior," Journal of Political Economy, 78 (2), 45-57.

Obermiller, Carl, and Eric R. Spangenberg (1998), "Development of a Scale to Measure Consumer Skepticism Toward Advertising," Journal of Consumer Psychology, 7 (2), 159-186.

--, and--(2000), "On the Origin and Distinctiveness of Skepticism Toward Advertising," Marketing Letters, 11 (4), 311-322.

--, and--(2002), "Skepticism Toward Advertising as a Moderator of Responses to Advertising Source and Appeal Effects," working paper, Seattle University, Albers School of Business.

Puto, Christopher, and William Wells (1984), "Informational and Transformational Advertising: The Differential Effects of Time," in Advances in Consumer Research, vol. 11, Thomas Kinnear, ed., Provo, UT: Association for Consumer Research, 572-576.

Schoenbachler, Denise, and Tommy Whittler (1996), "Adolescent Processing of Social and Physical Threat Communications," Journal of Advertising, 25 (4), 37-54.

Steele, Claude M. (1988), "The Psychology of Self-Affirmation: Sustaining the Integrity of the Self," in Advances in Experimental Social Psychology, L. Berkowitz, ed., New York: Academic Press, 261-302.

Tedeschi, James T., Barry R. Schlenker, and Thomas V. Bonoma (1971), "Cognitive Dissonance: Private Ratiocination or Public Spectacle?" American Psychologist, 26 (7), 685-695.

Zhang, Yong (1996), "Responses to Humorous Advertising: The Moderating Effect of Need for Cognition," Journal of Advertising, 25 (Spring), 15-32.

NOTES

(1.) An anonymous reviewer raised the question of the relation of "need for cognition" (NFC) to ad skepticism. Measures of causes of ad skepticism were not included in the present study, but were investigated in earlier work (Obermiller and Spangenberg 1998). In that earlier work, we also expected that NFC should correlate with ad skepticism. The evidence has been decidedly mixed, however. In several unpublished studies, we found only a weak positive correlation between need for cognition and ad skepticism, with considerable variance. After some thought, we have concluded that need for cognition is a more basic psychological trait that reflects a need for mental stimulation. Ad skepticism may be influenced by basic psychological dispositions, but it is also influenced by consumer socialization; thus, consumers may become skeptical without regard for need for cognition. As we found, highly skeptical consumers are apt to reject advertising. On the other hand, high need for cognition is likely to lead to processing of advertising, particularly in situations where an ad is the only available stimulus.

(2.) A note on the reliability analyses for all three studies: Typically, for scale reliability, Cronbach's [alpha] should exceed .80. The benchmarks recommended for comparative fit index (CFI) and goodness-of-fit index (GFI) are values in excess of .90 (Bagozzi and Yi 1988). Root mean square error of approximation (RMSEA) indices were mediocre to poor for all three studies. Standardized root mean square residual (RMR) values were above the benchmark of .05, except in Study 1. Finally, all [chi square] values were higher than appropriate to indicate good fit. Because this statistic is sensitive to sample size, some analysts recommend the ratio of [chi square] to degrees of freedom. By this measure, there is adequate fit; the ratio is below the benchmark of 3.0 (Carmines and McIver 1981, p. 80).

(3.) See note 2 above.

(4.) In an unpublished study (Obermiller and Spangenberg 2002) that also investigated the question of skepticism and ad appeal, we tested source effects (expert, attractive, similar) and argument type (case or base information) for moderating effects of skepticism. As expected, participants with high skepticism were generally less positive toward all sources and both argument types, but we found no moderating effects of skepticism on either of those variables.

(5.) See note 2 above.

Carl Obermiller (Ph.D., Ohio State University) is a professor of marketing and Marketing Department chair at the Albers School of Business, Seattle University.

Eric Spangenberg (Ph.D., University of Washington) is dean and Maughmer Chair at the College of Business, Washington State University.

Douglas L. MacLachlan (Ph.D., University of California, Berkeley) is a professor in the Department of Marketing and International Business, University of Washington.

TABLE 1
Relationship Between SKEP and Advertising Responses for Low and
High Product Category Involvement (PCI) and Ad Informativeness

                                          Low ad        High ad
Correlations        Low PCI    High PCI   information   information

SKEP-Likable        -.09 (a)   -.18 (a)   -.11          -.14
SKEP-Believable     -.27       -.29       -.26          -.29
SKEP-Influential    -.17 (a)   -.26 (a)   -.14 (b)      -.30  (a)
SKEP-False claims    .47        .42        .44           .44

(a) Z = 2.00, p < .05.

(b) Z = 2.68, p < .01.

TABLE 2
Measures' of Behavioral and Distant Responses
to Advertising and Simple Correlations' with SKEP

                              Correlation
Item                           with SKEP    Interpretation

1 . When I watch television      r=.61      Higher skepticism related
that has been recorded on                   to more zipping through
VCR, if I am able, I fast-                  taped commercials.
forward through the
commercials.

2. When I watch television       r=.48      Higher skepticism related
and I have a remote                         to more zapping commercials
control, I switch channels                  with a remote.
to avoid the commercials.

3. When I read a                r=-.47      Higher skepticism related
magazine, I look at                         to lower tendency to note
the advertisements as                       ads in magazines.
well as the stories.

4. When I buy a new product     r=-.48      Higher skepticism related
or any product for the                      to less reliance on ads
first time, I rely on                       for information.
advertising as a source
of information.

5. When I buy a new product       r=0       No relation between
or any product for the                      skepticism and reliance
first time, I rely on                       on Consumer Reports for
Consumer Reports as a                       information.
source of information.

6. When I buy a new product      r=.16      No relation between
or any product for the                      skepticism and reliance on
first time, I rely on                       specialty magazines for
reviews from specialty                      information.
magazines (for example,
PC Magazine, Runners'
World, Northwest Gardener)
as a source of information.

7. If a claim is made in an      r=.40      Higher skepticism related
advertisement for a product                 to greater tendency to
that I am considering                       check the truth of ad
buying, I check out the                     claims.
truth of the claim either
by asking friends who have
used the product or by
trying out the product
myself.

8. After seeing a               r=-.58      Higher skepticism related
commercial for a new                        to lower tendency to new
product, I have buy                         product after prompting
immediately decided                         from advertising.
to buy the product.

9. List below as many           r=-.18      Weak relation between
products as you can             (p<.10)     skepticism and recall of
remember buying for                         fewer purchases based on
which the only source                       advertising.
of information was
advertising.

10. I tend to distrust           r=.01      No relation between
ads that use movie stars,                   skepticism and distrust
famous models, or sports                    of celebrity spokespeople.
celebrities as
spokespeople.

11. I tend to distrust ads       r=.09      No relation between
that look like stories (in                  skepticism and distrust
print) or programs (on                      of infomercial ad appeals.
television or radio)
instead of obviously
looking like ads.

12. I prefer ads that have      r=-.09      No relation between
clear, simple information                   skepticism and preference
to those that try to                        for clear, simple
entertain me with mostly                    informative ads.
pictures, music, or humor.

(a) Item 9 was coded as the number of products listed;
all other items were Likert-type seven-point scales.

(b) Correlations with absolute values above r = .28
were significant at p < .01.

TABLE 3
SKEP Influence on Response to Informational or Emotional Ads

                                                         Standardized
                         Regression  Standard  Critical   regression
                           weight     error     ratio       weight

Structural model
SKEP [right arrow] EMOT     .750       .278     2.693        .615
SKEP [right arrow] INFO    -.716       .267    -2.680       -.556

Measurement models
SKEP [right arrow] X1      1.000                             .675
SKEP [right arrow] X2      1.329       .323     4.121        .652
SKEP [right arrow] X3       .992       .241     4.114        .651
SKEP [right arrow] X4       .781       .212     3.677        .574
SKEP [right arrow] X5      1.042       .229     4.547        .734
SKEP [right arrow] X6       .808       .217     3.722        .581
SKEP [right arrow] X7       .580       .170     3.419        .529
SKEP [right arrow] X8       .285       .181     1.570        .235
SKEP [right arrow] X9       .675       .214     3.156        .486
EMOT [right arrow] GD      1.000
EMOT [right arrow] HD       .959       .347     2.768        .594
EMOT [right arrow] AD       .461       .319     1.442        .250
EMOT [right arrow] ED       .429       .362     1.187        .197
EMOT [right arrow] GG      1.176       .242     4.865        .734
EMOT [right arrow] HG       .868       .315     2.754        .589
EMOT [right arrow] AG       .783       .299     2.618        .530
EMOT [right arrow] EG       .377       .339     1.113        .184
INFO [right arrow] VD      1.000                             .553
INFO [right arrow] SD      1.427       .403     3.537        .698
INFO [right arrow] PD       .950       .303     3.132        .562
INFO [right arrow] JD      1.160       .355     3.269        .603
INFO [right arrow] VG      1.067       .170     6.270        .629
INFO [right arrow] SG      1.548       .393     3.933        .839
INFO [right arrow] PG      1.093       .299     3.654        .712
INFO [right arrow] JG      1.255       .338     3.718        .739

Notes: CFI (comparative fit index) = .838; GFI
(goodness-of-fit index) = .662; RMSEA (root mean
square error of approximation) = .097; RMR (root
mean square residual) = .186; [chi square] (df =
265) = 398.291; [chi square]/df = 1.50; p value =
<.001.

Reliability of the SKEP Scale

                           Study 1     Study 2   Study 3

Cronbach's [alpha]           .8250       .9226     .8080
Sample size              140          105        54
CFI                          .906        .946      .771
GFI                          .908        .887      .821
RMSEA                        .092        .108      .148
RMR                          .047        .094      .079
[chi square] (df = 27)     59.004      59.798    58.357
[chi square]/df             2.18        2.21      2.16
p value                     <.001       <.001     <.001