• Nenhum resultado encontrado

Consumer innovativeness as a prerequisite of private label success in emerging markets

2. Ad – Brand Discrepancy & the Incongruity Resolution Process

Even though extant research on information incongruity has provided insightful, yet inconclusive findings (Noseworthy, Finlay, & Islam, 2010), there seems to be converging evidence across studies supporting the superiority of moderate incongruity (Maoz & Tybout, 2002). For instance, Pieters and colleagues (2002) found that, compared to typical, relatively unique and unexpected ad executions increase attention to the ad’s verbal and pictorial content, as indicated by the frequency of eye fixations. Likewise, Törn and Dahlén (2008) showed that

consumers devote a longer viewing time to ads that mismatch, rather than match, established beliefs about the advertised brand. In a study that closely followed Mandler’s (1982) paradigm, Meyers-Levy and Tybout (1989) demonstrated that when a new product proposition is moderately discrepant with the product category schema it supposedly corresponds, it is processed to a greater extent and is evaluated more favorably than a congruent or an extremely incongruent product. This inverted-U pattern has also occurred for incongruity between a parent brand and a brand extension, between a product and a celebrity endorsement, or when the music in an ad does not match consumer expectations (e.g., Meyers-Levy, Louie, & Curren, 1994). In the same vein, research in advertising rhetoric has indicated that moderately ambiguous and incongruent rhetorical figures, such as antithesis, metonym, or metaphor, increase the aesthetic value of the ad and create more liking than completely straightforward or extremely figurative ads (McQuarrie & Mick, 2009). In fact, a recent study by Phillips and McQuarrie (2009) has reconciled this proposition with Manlder’s (1982) original postulate. The authors demonstrated that it is not the presence of a figurative element in the ad execution per se that produces the favorable effect, but the degree of its figurativeness, that is, the extent to which the rhetorical element deviates from established norms and expectations. Based on the above discussion, a similar non-monotonic relationship is anticipated in the case of ad – brand incongruity. In more detail, we hypothesize that:

H1: Moderately incongruent ads (a) stimulate more ad processing and (b) lead to more favorable ad attitude, compared to congruent and extremely incongruent ads.

Central to the hypotheses above is the assumption that consumers manage to successfully resolve the incongruity. In other words, consumers need to identify the ad – brand discrepancy and accommodate the incongruent stimulus information within existing brand knowledge in order to make sense of the ad content (Jhang, Grant, & Campbell, 2012). Thus, even though it should become more difficult to makes sense of the ad as incongruity increases, consumers will finally manage to resolve moderate, but not extreme, discrepancies and comprehend the message. Thus, it is suggested that:

H2a: As the degree of ad – brand incongruity increases, consumers find it more difficult to make sense of the ad message.

H2b: Consumers comprehend the content of the congruent and moderately incongruent ads equally, but are less able to comprehend the content of the extremely incongruent ad.

Finally, it is suggested that incongruity enhances the favorability of evaluations through metacognitive experiences of satisfaction generated by successfully carrying out a more intellectually-demanding task (Maoz & Tybout, 2002; Halkias & Kokkinaki, 2011a). Similarly, research in rhetoric argues that there is an intrinsically rewarding value in successfully interpreting an ambiguous or deviant configuration of signs (McQuarrie & Mick, 2009).

Incongruity resolution is necessary for the favorable effect of ad – brand incongruity to occur. That being said, evaluative responses will be mediated by consumers’ ability to accommodate schema incongruity and comprehend the ad message. Overall, it is predicted that:

H3: Ad message comprehension mediates the effect of ad – brand incongruity on attitude toward the ad.

3. Method

The above hypotheses were examined in an experimental study, employing a between- subjects design. In total, 120 undergraduate students participated in the study in exchange for extra course credit. Cell sizes were equally distributed across conditions. Hypotheses were tested through ANOVA and ANCOVA with planned contrasts. The proposed mediation was investigated using Baron and Kenny’s (1986) approach and Sobel’s (1982) test of the indirect effect.

Stimuli and manipulations. Through a series of pretests, we developed three stimulus print ads for a real chocolate brand, each being either congruent, moderately incongruent, or extremely incongruent with the consumers’ schema for the brand. The degree of incongruity was manipulated through the pictorial component of the ad, while all verbal elements and ad format were held constant across conditions (Halkias & Kokkinaki, 2011b). Manipulation effectiveness was checked by averaging participants’ ratings on ad expectancy and ad relevancy (r = .72, p < .001) with regard to their impression of the brand (Heckler & Childers, 1992). Post-hoc comparisons were all significant at the .01 level with F(2, 117) = 26.27, p < .001.

Procedure and measures. The study took place in a computer laboratory. Upon arrival participants were randomly assigned to one of the three experimental conditions, were given a set of instructions, and completed a computer-based questionnaire that contained the all necessary manipulations and dependent measures. Participants’ evaluations for a series of brands, including the target brand, were measured in the beginning of the experiment to control for individual differences in brand attitude prior to the exposure. Measurement of response latency to the experimental stimuli was used to indicate the amount of ad processing (Goodstein, 1993). Ad and brand attitudes were measured on three nine-point items anchored by dislike/like, bad/good, and negative/positive (Cronbach’s α ranged from .89 to .91). Ease of resolution was measured on a nine-point item asking respondents to indicate how many mental would be needed to make sense of the ad concept (few steps/many steps), while their ability to resolve the incongruity was measured on a nine-point item indicating ad message comprehension (makes no sense to me/makes sense to me) adapted from Jhang, Grant, and Campbell (2012).

4. Results

As hypothesized, analysis indicated that ad processing follows an inverted-U pattern across the degrees of incongruity (F(2, 117) = 14.94, p < .001), with participants spending significantly more time processing the moderately incongruent ad (M = 10.25 seconds), compared to the congruent (mean difference = 4.88 seconds, p < .001) and the extremely incongruent ad (mean difference = 3.83 seconds, p < .001). ANCOVA revealed no significant influence of participants’

brand attitude prior to the exposure (F < 1), thus analysis of variance was performed to investigate ad attitudes. Consistent with H1b, a significant main effect was obtained (F(2, 117) = 15.29, p < .001). Planned contrasts showed that moderate ad – brand incongruity leads to more favorable ad attitude (M = 6.82) than congruity (mean difference = 1.35, p < .001) or extreme incongruity (mean difference = 1.81, p < .001). In terms of incongruity resolution, as figure 1 shows, participants found it significantly more difficult to make sense of the ad concept as incongruity increased. However, as H2b predicted, they managed to comprehend the message of the moderately incongruent ad version, whereas they were significantly less able to understand the content of the extremely incongruent ad.

Finally, mediation analysis produced significant results for all four steps of the Baron and Kenny (1986) procedure, indicating complementary mediation (see Zhao, Lynch, & Chen, 2010).

The indirect effect accounted for about 55% of the total effect and a Sobel test showed that it was statistically significant (z = 4.35, p < .001). Full details for the mediation analysis are provided in the figure below.