Abstract
Extant research in endorsements mainly examines the impacts of scandals on consumers' attitudes toward tarnished endorsers and the endorsed brands (e.g., Bartz, Molchanov, and Stork 2013; Um 2013). Less is known about the impacts of scandal characteristics on endorsers and brands. Therefore, this study investigates the influence of scandal characteristics, specifically in severity and relevance, on consumers' attitudes toward brands and athlete celebrity endorsers in question. With an experimental study of two hundred participants and several pretests, we demonstrate that the impacts of scandals on endorsers and brands are a function of transgression relevance and severity. Severe scandals relevant to the professional expertise of athlete celebrity endorser are most detrimental. Based on the cue-diagnosticity theory (Skowronski and Carlston 1987, 1989), diagnostic cues are unexpected and extreme and, thus, receive more weight, and generate more impact, on the impression formation of a subject. Specifically, a positive diagnostic cue is more extreme to, and generates stronger positive impacts on, the negative category, yielding extremity and positivity biases. That is, a positive diagnostic cue is more influential, and activates more positivity biases, on the member of the negative category. The subject originally in the negative category is, thus, re-categorized as a member of the positive category because of the positive diagnostic cue. In contrast, a negative diagnostic cue is more extreme to, and generates more negative impacts on, the positive category, yielding extremity and negativity biases. That is, a negative diagnostic cue is more influential, and facilitates more negativity biases, on the member of the positive category. The subject originally in the positive category is, thus, re-categorized as a member of the negative category because of the negative diagnostic cue. In general, an effective athlete celebrity endorser is expected to have a very positive image. Thus, transgression information is unexpected for an athlete celebrity endorser. Specifically, severe transgression information is unusual for an endorser and a more extreme cue indicating of a worse athlete celebrity endorser. Based on the cue-diagnosticity theory, high- (vs. low-) severity transgression information is a more diagnostic cue exerting more negative impact on expertise and endorser evaluations directly and brand evaluations indirectly. Thus, high-severity transgression information instigates more negative impacts on the perceived expertise of athlete celebrity, which in turns weakens the evaluations of athlete celebrity endorsers and then the endorsed brands (hypothesis 1). Moreover, extant research documents that information relevant to judgments is more influential (Ahluwalia, Unnava, and Burnkrant 2001). Thus, information highly (vs. lowly) relevant to endorser expertise is a diagnostic cue instigating more influence on expertise and endorser evaluations directly and brand evaluations. Thus, expertise-relevant transgressions instigate more negative impacts on the perceived expertise of athlete celebrity, which in turns weakens the evaluations of athlete celebrity endorsers and then the endorsed brands (hypothesis 2). As a result, severe and relevant transgression information is the most diagnostic cue for, and is most detrimental to, endorser and brand evaluation. Thus, transgressions, which are high-severity high-relevance, are most detrimental to athlete celebrity endorsers and the endorsed brands, followed by high-severity low-relevance and low-severity high-relevance transgressions, and then low-severity low-relevance ones (hypothesis 3). Based on a pretest, doping and adultery were selected to represent the relevant and irrelevant transgressions, respectively. A fictitious NBA basketball celebrity, J. R., was portrayed with a statement mimicked the profile of the NBA celebrity of Kobe Bryant. Moreover, a fictitious brand of J-R signature shoes and a premier athletic shoes maker, Nibok, were portrayed with a statement mimicked Nike's LeBron 11 signature shoes endorsed by LeBron James. Four statements were developed and pretested to represent the four experimental conditions. One hundred and eighty-one U.S. residents participated in the 2 (severity: high vs. low) x 2 (relevance: high vs. low) between-subject experimental design. The participants were informed that the purpose of study was to investigate consumers' opinions about athlete celebrities. The participants started by rating the statements about J. R. athlete celebrity and Nibok's J-R signature shoes followed by rating the severity and relevance of a transgression statement and, then, re-rating J. R. and the J-R signature shoes. Waves of two-way ANOVA and simple effects tests on expertise, endorser, and endorsed brand evaluations, respectively, were performed to verify the hypotheses. The analysis revealed that transgression severity weakened perceived expertise differently only when the transgression was high-relevance, which suggested that transgression relevance is more dominating than transgression severity on weakening perceived expertise (Hla and H2a supported). Further, the weakening effects of transgressions on endorser attitudes were exclusively determined by transgression severity. Transgression relevance was not influential on endorser evaluations (H1b supported; H2b not supported). Moreover, for high-relevance transgressions, the high-severity doping weakened more brand attitudes than the low-severity doping and, for the low-relevance transgressions, the high-severity adultery weakened more brand attitudes than the low-severity adultery. Moreover, for the high-severity transgressions, the high-severity doping weakened more brand attitudes than the high-severity adultery, whereas, for low-severity transgressions, the low-severity doping weakened brand attitudes identically to low-severity adultery (Hlc and Н2с supported). In comparison, the high-severity high-relevance transgression (i.e., high-severity doping) weakened brand attitudes the most, followed by high-severity low-relevance transgression (i.e., high-severity adultery), and then the low-severity high- and low-relevance transgressions (i.e., low-severity doping and adultery). As a result, the third hypothesis was supported. Furthermore, path analyses examining the casual relationship among transgression, endorser, and brand evaluations revealed that these three factors were all highly correlated. However, the correlation between transgressions and brand attitudes was decreased when the mediator of endorser attitudes was included, while the tion between and the terminal variable of endorsed brand evaluations remained significant. The results indicated that the casual and effect relationship between transgressions and brand attitudes was partially mediated by endorser attitudes. In other words, the decrease in brand attitudes was directly caused by both transgression relevance and severity and the declining of endorser attitudes. In conclusion, endorsed brand evaluations are moderated by transgression severity and relevance, whereas endorser evaluations are exclusively affected by transgression severity. In other word, transgression severity affects endorser and endorsed brand evaluations, whereas transgression relevance moderates endorsed brand evaluations only. Specifically, severe and relevant transgressions are most detrimental to endorsed brand evaluations.