تایید کننده های متعدد و تاییدیه های متعدد :تاثیر تکرار پیام، تجانس منبع و مشارکت بر نگرش نام تجاری (برند)

This paper investigates the effects of source congruence on brand attitudes in two situations: multiple brand endorsements by one celebrity and multiple celebrity endorsers of one brand. Under low involvement conditions, brand attitudes become more negative as a celebrity endorses multiple brands and more favorable with multiple endorsers. In high involvement conditions, strong source congruence overrides the negative effect of multiple brands, and the positive effect of multiple endorsers is found only with strong congruence. We interpret these results as suggestive of a frequency knowledge cue that dominates under low involvement but is non-diagnostic in high involvement scenarios.

One of the most widely recognized models used to describe how consumers respond to persuasive advertising is the ELM (Petty et al., 1983 and Petty and Cacioppo, 1986), which posits two routes through which an advertisement can influence consumers: a high involvement, central route and a lower involvement, peripheral route. Prior research has shown that celebrity advertising can persuade consumers via either route. For example, Kirmani & Shiv (1998) demonstrated that under high involvement conditions the degree of source congruence can be conceptualized as the strength of a persuasive argument and that the effectiveness of celebrity advertising increases directly with source congruence; however, under low involvement conditions, the effectiveness of celebrity advertising is driven by peripheral cues in the ad, such as the attractiveness or likability of the celebrity endorser, rather than the persuasive strength of the arguments contained in the ad ( Kang and Herr, 2006 and Petty et al., 1983). Extending these findings to the focus of the current paper, we conceptualize the contexts of both multiple brand endorsements and multiple celebrity endorsers as different forms of persuasive message repetition, which we define as the repeated use of a celebrity endorser in the context of multiple brand endorsements and the repeated endorsement of a brand in the context of multiple celebrity endorsers. With prior research showing that increased message exposures lead to larger attitude differences between strong and weak message conditions under moderate levels of repetition (Cacioppo & Petty, 1989), we propose that source congruence, repetition, and involvement interact to affect consumer response to celebrity advertising, specifically in regard to brand attitude. First, consider the effect of multiple brand endorsements by the same celebrity on attitudes toward a focal brand. Under high involvement conditions, consumers are known to process source congruence in an advertisement as a persuasive argument (Kirmani & Shiv, 1998). Past research (Petty & Cacioppo, 1984) has also shown that when an attitude object is described by strong versus weak arguments, repetition causes increased differences in attitude. Based upon those findings, we posit that the difference in focal brand attitude between high and low source congruence conditions should increase as the number of exposures to the celebrity endorsing other brands increases (Petty & Cacioppo, 1984), due to the repetition of either a strong or weak argument. This leads to our first hypothesis: H1. Under high involvement, as the number of brands endorsed by a single celebrity increases, the favorability of focal brand attitude following exposure to ads high in source congruence increases relative to the favorability of focal brand attitude following ads low in source congruence. Conversely, under low involvement conditions, peripheral cues—not argument strength—affect brand attitude (Petty & Cacioppo, 1984). Due to the nature of the multiple brand endorsements context, we propose that in addition to source characteristics, consumers may utilize frequency knowledge or “the mere number of positive and negative attributes associated with a brand” (Alba & Marmorstein, 1987, p. 14) as a cue during brand attitude formation. Since consumers act as “cognitive misers” under low involvement conditions (Burnkrant, 1976), they may simply count the number of brands being endorsed by the celebrity and use this frequency knowledge (Alba & Marmorstein, 1987) as a cue to form their brand attitudes. Mowen & Brown (1981) found that consumers had a negative reaction to a celebrity who endorsed multiple brands, which suggests that brand attitudes should become more negative as the number of brand endorsements increases. In a related study, Tripp et al. (1994, Study 2) found that consumers did not spontaneously make higher-order inferences about multiple endorsements, even when they were aware of them. This suggests that a low-involvement cue such as frequency knowledge may be at work: multiple brand endorsements result in an increasing number of negative associations to the brand, even in the absence of specific inferences about the motivations of the brand or endorser. Furthermore, since the degree of source congruence in an ad is likely effortful for consumers to judge, it should not influence brand attitude under low involvement conditions. Formally stated: H2. Under low involvement, as the number of brands endorsed by a single endorser increases, the favorability of focal brand attitude decreases regardless of the level of source congruence. Next, consider the context of multiple celebrity endorsers, where consumers are being exposed to both brand and message repetition. Similar to the context of multiple brand endorsements, source congruence is more likely to be scrutinized under high (versus low) involvement conditions (Kirmani & Shiv, 1998). Thus, when involvement is high, we propose that the difference in brand attitude between high and low source congruence conditions should increase as the number of celebrity endorsers increases, due to the repetition of either strong or weak arguments (Petty et al., 1983). Formally stated: H3. Under high involvement, as the number of celebrities endorsing the brand increases, the favorability of brand attitude following exposure to ads high in source congruence increases relative to the favorability of brand attitude following ads low in source congruence. Again, similar to the context of multiple brand endorsements, brand attitude should be influenced by peripheral cues under low involvement conditions (Kang and Herr, 2006 and Petty et al., 1983). However, in the context of multiple celebrity endorsers, exposure to celebrity advertising should associate the featured brand to multiple positive attributes (Alba & Marmorstein, 1987), assuming that the source characteristics of each celebrity are positive (Kahle & Homer, 1985). Therefore, if consumers extract the frequency knowledge that exists in the context of multiple celebrity endorsers and use it as a cue during brand attitude formation, then as the number of celebrity endorsers increases, brand attitude should also increase. Finally, with the number of celebrities endorsing a single brand being easier to process and judge than the degree of source congruence contained in each ad, the former rather than the latter should influence brand attitudes under low involvement conditions. Formally stated: H4. Under low involvement, as the number of celebrities endorsing the brand increases, the favorability of brand attitude increases regardless of the degree of source congruence.

The objective of this research was to explore how factors that are well-established in the celebrity advertising literature interact and affect consumer response to celebrity advertising when (a) a single celebrity endorses multiple brands (i.e., multiple brand endorsements) and (b) multiple celebrities endorse a single brand (i.e., multiple celebrity endorsers). In this paper, we have argued that when involvement is high, the effectiveness of a celebrity endorsement depends on the degree of congruence shared between the celebrity and brand (Kirmani and Shiv, 1998 and McCracken, 1989) as well as the degree of message repetition. Across two studies, we showed that increased repetition caused by either multiple brand endorsements (Experiment 1) or multiple celebrity endorsers (Experiment 2) leads to greater divergence in brand attitudes between high and low source congruence endorsements under high (but not low) involvement conditions. Although our manipulation checks showed that participants were aware of source congruence differences in both high and low involvement conditions, we did not find an effect of source congruence on brand attitude under low involvement conditions in either study, which is consistent with our theorizing. Such a finding suggests that source congruence was given less weight during the formation of brand attitude under low (versus high) involvement conditions and further suggests that consumers based their brand attitudes on some other information communicated by the ad(s) under low involvement conditions. While determining the specific underlying process for how brand attitudes are formed under low involvement conditions was not the focus of this paper, we find the pattern of results across the two studies (i.e., a negative effect of the number of brands that a celebrity endorses under low involvement conditions in Experiment 1 and a positive effect of the number of celebrities endorsing a brand under low involvement conditions in Experiment 2) to be an interesting area for future research (see discussion below). Theoretical implications Our results make important contributions to the celebrity endorsement literature. First, we extend Kirmani & Shiv's (1998) ELM-based explanation of celebrity endorser effects and make important connections to earlier research that examined consumer response to multiple brand endorsements by a single celebrity (e.g., Tripp et al., 1994) and multiple celebrity endorsers for a single brand (e.g., Mowen & Brown, 1981). The results of our two studies show that source congruence and message repetition are important determinants of brand attitudes when consumers are highly involved with processing the advertising. The results are consistent with Kirmani & Shiv (1998) in that source congruence is treated as a persuasive argument to be scrutinized under high involvement conditions. Second, our findings demonstrate that multiple brand endorsements processed under high involvement conditions can lead to the polarization of attitudes toward the brand itself, a finding that is consistent with prior attitude research (e.g., Petty & Cacioppo, 1984) but had not been shown in a celebrity endorsement context. From a theoretical standpoint, this finding illustrates that consumers consider the endorsement portfolios of both the celebrity and the brand when responding to celebrity advertising. Thus, it is important to consider more than just a single celebrity paired with a single brand when predicting likely consumer response to celebrity advertising. However, this raises questions about how consumer knowledge about either of these portfolios may affect consumer response, a point that we will discuss further in the future research section. Conversely, under low involvement conditions, the findings show that an increasing number of brands endorsed by a single celebrity has a detrimental effect (Experiment 1) while an increasing number of celebrity endorsers for a single brand has a beneficial effect (Experiment 2) on consumer response to celebrity advertising. Such results suggest that, in addition to source characteristics, consumers may utilize frequency knowledge (Alba & Marmorstein, 1987) as a peripheral cue during brand attitude formation under low involvement conditions, which is consistent with the ELM and Chaiken's Heuristic-Systematic Model (HSM; Chaiken et al., 1989 and Eagly and Chaiken, 1993). Previously, frequency knowledge has been identified as a “very effective peripheral route to persuasion,” (Alba & Marmorstein, 1987, p.24), but a frequency knowledge explanation has not been studied in a celebrity endorsement context. Managerial implications This research provides empirical support for the managerial practices of paying more for an exclusive endorser and trying to find an endorser who fits well with the brand's positioning. However, to the extent that brand managers can determine the likely level of audience involvement, they might be able to economize somewhat. Our results indicate that under high involvement conditions, a celebrity who endorses multiple brands hurts a brand only if the celebrity fits poorly with the other brands in his or her portfolio. If managers are confident that ads will be viewed under high involvement, then the fit of other brands endorsed is more important than whether other brands are endorsed at all. Conversely, if ads are generally viewed under low involvement conditions, managers might be best served by hiring as many exclusive endorsers as possible with no particular regard to their congruence with the brand. Since a celebrity endorsing multiple brands can lead to more negative brand attitudes, finding exclusive endorsers is important. Also, using many endorsers apparently helps to showcase the brand in a good light, and repetition of the brand appears to drive brand attitude under low involvement conditions through the increased use of frequency knowledge cues. Limitations Several limitations exist in the current research. First, our research utilized fictitious brands, and thus how the established effects translate to a real brand is unclear. Presumably, the established findings would be similar to effects found with new brands. However, the implications for established brands are less clear. While it is possible that a well-known brand may find little benefit or detriment from celebrity endorsers, recent research (Dimofte & Yalch, 2011) has demonstrated that automatic meaning transfer (through mere association effects) can be enhanced with greater familiarity. Thus, the effects of multiple endorsers and multiple endorsements may actually be greater for familiar brands under certain conditions. Second, the studies were conducted in a laboratory environment with undergraduate student participants. While it could be argued that many students might not find the chosen categories (expensive watch, SUV, briefcase) particularly relevant to them, random assignment should eliminate any systematic differences across conditions that could cause our results. However, we acknowledge that replicating the findings with different brands and categories might be worthwhile for future research. It seems possible that if a brand or category is particularly relevant, consumer expertise may moderate the effects, minimizing the impact of any endorsement for consumer with high expertise and amplifying it for consumers with low expertise. Third, we chose to create conditions with one versus three ads, which creates a possible confound in our manipulation. Our intent was to examine the effects of multiple endorsements by a single celebrity and multiple endorsers for a single brand relative to a context where there is only one celebrity endorsing one brand. We operationalized increased exposures as more brands endorsed and as more brand endorsers. It would be interesting to investigate whether a brand shown with equal exposures would fare better with multiple endorsers or a single endorser. Finally, our research was intended to investigate the influence of a moderate amount of repetitions of the brand message. Our data cannot address what the boundary conditions would be. For example, we acknowledge that even under high involvement and high source congruence conditions, an extreme number of endorsements by the same celebrity (e.g., twenty) could lead to wear-out effects that might not differ significantly from low source congruence conditions. Conversely, wear-out effects could be caused by the sheer numbers of celebrity endorsers, though our opening Rolex example (over 20 endorsers) suggests that tolerance for new endorsers might be relatively high. Ultimately, these are questions for further research to address. Future research We believe that the current research sparks several areas for future research to pursue. First, brand attitude is just one dependent measure that might be examined in a context involving multiple brand endorsements or multiple celebrity endorsers; brand managers are also concerned about the degree of memorability, consideration, and purchase intention stemming from exposure to celebrity advertising. While one may predict that the pattern of consumer response to celebrity advertising will be equivalent across various measures, the contexts of multiple brand endorsements and multiple celebrity endorsers create the potential for brands and celebrities, respectively, in an endorsement portfolio to interact with one another and thus to impact the effectiveness of celebrity advertising at point of purchase. Therefore, future research might consider investigating whether an endorsement portfolio differentially affects brand attitude and other key measures. Future research could also investigate the extent to which the aspects of celebrity endorsement investigated in this paper (i.e., number of brands, involvement and congruence) influence the links between brand meaning, self-identity and goals suggested by Kirmani (2009). Second, the ELM theoretical perspective suggests that consumers likely utilized a peripheral cue to form their brand attitudes under low involvement conditions in our studies. However since we statistically controlled for all source characteristic effects in our analyses, yet still found a significant effect, another type of peripheral cue (i.e., frequency knowledge) may have been operating. We believe that the contexts of multiple brand endorsements and multiple celebrity endorsers are unique relative to the context of a single celebrity endorsing a single brand and thus may provide consumers with a new peripheral cue, such as the number of brands endorsed/celebrity endorsers, to utilize during brand attitude formation under low involvement conditions. To explore this notion, we conducted a small post hoc study (n = 67)1 and found evidence to suggest that consumers perceive multiple brand endorsements (multiple celebrity endorsers) as unfavorable (favorable) for the brand, which is consistent with our theorizing and with the results found under low involvement conditions in our experiments. However, it is unclear whether the process through which this frequency knowledge flows is consciously attended or if it is one of the many unexplored nonconscious consumer processes (Chartrand & Fitzsimons, 2011) that influence consumer behavior. Therefore, future research should examine whether the contexts of multiple brand endorsements and multiple celebrity endorsers provide consumers with new peripheral cues, like frequency knowledge, to deploy during brand attitude formation under low involvement conditions and the extent to (and conditions under) which these new cues are conscious or nonconscious (Chartrand & Fitzsimons, 2011). For example, future research could examine whether elaboration is a prerequisite to celebrity meaning transfer to brands in light of recent research that suggests the plausibility of unconscious meaning transfer (Galli & Gorn, 2011). While this research has addressed the topic of multiple brand endorsements and multiple celebrity endorsers, other areas within the broader realm of celebrity advertising remain largely unexplored. In today's society, actors and actresses are increasingly trying to undertake varied roles, and it is rare to find one with as well-defined a set of celebrity cultural meanings as James Garner, the foil for McCracken's (1989) analysis. Additionally, multiple sources of possible contamination of an image have recently come into existence (e.g., Facebook, Twitter, tabloid magazines, tabloid TV shows, etc.), all of which make it more challenging to manage a celebrity's cultural meaning. Does this result in confusion and fragmentation of cultural meaning in the mind of the consumer? Finally, perhaps the transfer of cultural meanings between brands and celebrities may be bidirectional. The most powerful brands in the world are spending billions of dollars to create well-defined brand images. Is it possible for celebrities to “game” the system to define or redefine themselves? For example, would a young actor like Leonardo DiCaprio be able to bolster his image of being a patriotic American by endorsing Chevy trucks instead of Ralph Lauren's clothing line? Only further research into this interesting area can answer this with any degree of certainty.