How Does the Negative Impact of an Athlete’s Reputational Crisis Spill Over to Endorsed and Competing Brands? The Moderating Effects of Consumer Knowledge

Abstract

Despite some of the recent examinations of an athlete’s reputational crisis (ARC), their negative spillover effects on endorsed and competing brands have been overlooked. The purpose of this study is to investigate the relationships between perceived severity, athlete endorser credibility (i.e., incompetence, untrustworthiness), and attitudes towards endorsed and competing brands. To enhance theoretical understanding of the phenomenon, the moderating role of consumer knowledge was also tested. Participants were recruited from Amazon Mechanical Turk (N = 339). A multigroup structural equation model was employed to test the hypothesized model. Results indicated that the severity of an ARC is associated with the perceived incompetence and untrustworthiness of focal athletes. Perceived incompetence is associated with negative evaluation of an endorsed brand. Furthermore, this impact is significantly stronger for consumers with greater knowledge of the athletes than those who are less knowledgeable. Interestingly, competitor brands received negative impact indirectly from the athlete endorsers’ incompetence. This spillover effect is also manifested differently depending on the level of consumer knowledge.

Keywords

negative publicity athlete transgression consumer knowledge memory endorsement

While watching TV shows, reading magazines, and browsing the Internet, people are regularly exposed to numerous media messages that are framed with sport content due to the popularity of sports media (Hutchins & Rowe, 2012; Jackson, 2013; Lewis & Weaver, 2015; Pedersen, Laucella, Kian, & Geurin, 2016). Companies and organizations strive to associate themselves to reputable celebrity athletes to develop favorable relationships with stakeholders in order to enjoy desired outcomes. For example, Nike’s endorsement contract with LeBron James brought a significant economic boost in the basketball shoe line (Badenhausen, 2017). Such endorsement strategies have been common in the sports industry as athlete endorsers can help companies effectively communicate with consumers (Spry, Pappu, & Cornwell, 2011). Previous studies have shown that athlete endorsements can offer numerous benefits including eliciting consumers’ favorable attitude towards endorsed brands (Carlson & Donavan, 2008; Till & Busler, 2000) and stronger purchase intentions (B. Brown, Bennett, & Ballouli, 2016).

However, this marketing communication strategy has been considered a double-edged sword. The media pay keen attention to an athlete’s reputational crisis (ARC) defined as negative incidents that threaten to disrupt an athlete’s reputation (Sato, Ko, Park, & Tao, 2015). It is common in news headlines to hear of an ARC such as Taylor Teagarden’s use of performance enhancing drug (PED), Tiger Woods’s extramarital infidelity, or Manny Pacquiao’s discriminatory remarks. Such negative information can capture consumer attention more than positive or neutral information (Ito, Larsen, Smith, & Cacioppo, 1998). Kroloff (1988) found that negative information receives 4 times more media exposure relative to positive information.

Scholars have provided valuable insights into understanding the characteristics of an ARC (J. S. Lee, Kwak, & Braunstein-Minkove, 2016; Sato et al., 2015), their impact on consumer perceptions toward athletes and their associated endorsed brands (J. S. Lee & Kwak, 2015; Lohneiss & Hill, 2014), and damage minimization and image repairing strategies (K. A. Brown, Anderson, & Dickhaus, 2016; Hambrick, Frederick, & Sanderson, 2015). Although these studies have provided valuable insights, there are research gaps that require further examination. Specifically, no prior research has focused on (1) credibility aspects (i.e., trustworthiness and expertise) influenced by ARCs and (2) the extent to which an ARC spills over to associated brands.

First, the majority of studies have not focused on the impact of an ARC on the specific endorser credibility aspects of expertise and trustworthiness. Expertise is related to athletes’ on-field performance while trustworthiness is associated with social desirability (Ohanian, 1991). One may consider the importance of expertise and trustworthiness differently to differentiate judgments regarding these attributes. For example, Tiger Woods may still receive support due to his expertise even if his trustworthiness might have been tainted by his misbehavior. Others may think that Woods’s misbehavior ruined his trustworthiness as well as characteristics related to his on-field performance (see J. S. Lee, Kwak, & Moore, 2015; however, a fictitious athlete, Ted Franklin, was used in their study). Without investigating the effect of an ARC on each endorser credibility aspect, communication managers will not be able to know whether endorser credibility attributes are tainted.

Second, there is a lack of recognition regarding the spillover effects of an ARC. Spillover effects refer to “the extent to which a message influences beliefs related to attributes that are not contained in the message” (Ahluwalia, Burnkrant, & Unnava, 2000, p. 458). In human memory, various nodes in memory are theorized as being associated and activated simultaneously (Anderson, 1983). Based on this argument, it could be reasoned that Lance Armstrong’s PED scandal, for example, had adverse impacts on Nike as an official endorsed brand but also affected competitor brands of Nike. We argue that competitor brands may also experience adverse impacts due to the product category associations with Nike.

Lastly, related to the spillover effects, people may react differently to athletes’ misbehavior depending on their level of knowledge. Consumer knowledge is defined as the extent to which consumers process experiences and familiarity with a target product (Sujan, 1985). Individuals with organized knowledge about a target object engage in a more elaborated information processing than those with less knowledge (Mitchell & Dacin, 1996; Roy & Cornwell, 2004). Consumers who possess different levels of knowledge about athletes, therefore, may show different reactions when perceiving information about an ARC. Considering the moderating role of consumer knowledge can provide insight on spillover effects.

The purpose of this study is to investigate the relationships between perceived severity, athlete endorser credibility, and attitudes towards endorsed and competing brands. The conceptual model is shown in Figure 1. Specifically, we argue that consumers evaluate credibility of misbehaved endorsers based on the severity of an ARC (first stage of the conceptual model). Regarding the spillover effects (second stage of the conceptual model), drawing on the associative memory network model (Collins & Loftus, 1975), we assess whether the tainted credibility of endorsers can impact directly (i.e., official endorsed brand) and indirectly associated objects (i.e., competing brand). In addition, we argue that the level of knowledge about the target athlete can play a moderating role in the second stage. The outcome of this study sheds light on the specific mechanisms that underlie the spillover effects of an ARC on endorsed and competing brands.

Figure 1.

A proposed model of athlete scandal spillover impact.

Theoretical Background and Hypothesis Development

In this section, we introduce hypotheses based on conceptual and empirical studies regarding an ARC, endorser credibility, and spillover effects. We focus on the source credibility model in the first part and the associative memory network model in the latter section that focuses on spillover effects.

The Severity of an ARC on Credibility of Celebrity Athlete Endorsers

Athletes are utilized as spokespersons for promoting various products and brands. It is reported that the top three professional basketball players in the National Basketball Association (NBA) earned more than $US120 million collectively from their endorsement contracts (Badenhausen, 2017). Companies clearly value the effectiveness of athlete endorsers. While the positive aspects of athlete endorsement have been verified in previous studies (Carlson & Donavan, 2008; Y. Lee & Koo, 2015; Sato, Ko, Kaplanidou, & Connaughton, 2016), scholars have also recognized detrimental impacts when athlete endorsers are involved in an ARC (J. S. Lee et al., 2015; Sato et al., 2015).

Brand managers recognize that an ARC negatively influences the evaluation of endorsers (Chien, Kelly & Weeks, 2016; Sato et al., 2015). The severity of an ARC can magnify the impacts (Hughes & Shank, 2005). Severity can further reinforce the impression of potential harm (Smith, Bolton, Wagner, 1999) which can lead to critical evaluations of wrongdoers (Kahneman, Schkade, & Sunstein, 1998). The severity of an ARC can be determined by performance relatedness (J. S. Lee et al., 2015; Sato et al., 2015). Sato, Ko, Park, and Tao (2015) found that consumers develop unfavorable attitudes toward misbehaved athletes; the negative impact is exceptionally strong when misbehaviors are linked with athletic performance (i.e., use of PEDs).

With regard to the endorser evaluation, scholars have commonly applied the source credibility model to understand the effectiveness of message senders (i.e., endorsers in this study) in persuading message receivers (i.e., consumers in this study) in focal communication topics (Hovland, Janis, & Kelly, 1953; Mudrick, Burton, & Lin, 2017). Source credibility is defined as positive characteristics of message senders that influence the degree to which message receivers accept the message (Hovland et al., 1953).

Trustworthiness and expertise have been considered two important subdimensions of source credibility (Hovland et al., 1953). Empirical studies show that trustworthiness of athlete endorsers positively influences consumers’ purchase intention of endorsed brands (Fink, Parker, Cunningham, & Cuneen, 2012). Expertise of endorsers is also particularly beneficial in persuading consumers when endorsers promote products that are related to their area of expertise (e.g., athlete endorsers promoting sport products; Ohanian, 1991). Through a meta-analysis of endorser effectiveness, Amos, Holmes, and Strutton (2008) found that trustworthiness and expertise are important aspects of endorser credibility that determine the effectiveness of endorsement campaigns. While there is a general consensus that athlete endorsers’ credibility (i.e., trustworthiness and expertise) is an important factor that influences consumers’ attitude toward endorsed brands, little is known about whether ARC severity can influence the credibility of athlete endorsers mutidimensionally. The harmful impacts of misbehaved endorsers depend on the severity of the focal negative event (Kahneman et al., 1998) since unethical behaviors increase the perception of untrustworthiness. Consumers also evaluate athlete endorsers as incompetent when they link the focal misbehavior and incompetence of athletes in their professional domain (e.g., PED issues; J. S. Lee & Kwak, 2015; Sato et al., 2015). Till and Shimp (1998) also yielded consistent findings, demonstrating that negative information of celebrity athletes can contribute to the critical evaluation of trustworthiness and expertise. Hence, the current research first sought to replicate the aforementioned studies:

Hypothesis 1a: Perceived severity is positively associated with perceived untrustworthiness of athlete endorsers.

Hypothesis 1b: Perceived severity is positively associated with perceived incompetence of athlete endorsers.

Spillover Effect

Scholars in the field of marketing communications have actively investigated the spillover effect (Bundy, Pfarrer, Short, & Coombs, 2017; Lei, Dawar, & Lemmink, 2008; Yu, Sengul, & Lester, 2008). The spillover effect refers to “the extent to which a message influences beliefs related to attributes that are not contained in the message” (Ahluwalia et al., 2000, p. 458). The direct spillover from endorsers to endorsed brands can be explained by the spreading activation principle in the associative memory network model (Collins & Loftus, 1975) or by the connectionist model of cognition (Fodor & Pylyshyn, 1988). These accounts similarly posit that human memory consists of neural networks that are built by numerous associated nodes of information. When encountering an information cue, associated nodes of information are automatically activated or spread in memory (Janiszewski & Wyer, 2014; Lei et al., 2008). For example, the attributes of celebrity athletes (e.g., trustworthiness and expertise) and product brands (e.g., design and quality) represent individual nodes of information separately stored in consumers’ memory (Chang & Ko, 2018). The two information nodes, athletes and products, then can be integrated into a single network in memory through endorsements. Consistent and repeated pairing often result in the transferring effects of subjective familiarity (Chang, 2018) and attitudinal (un)favorableness (J. S. Lee et al., 2015) from athletes to products. For example, a 2016 CNN article by Mullen (2016) and Chang (2018) suggested that when a scandal about a celebrity emerges, one of the most common responses by product managers is to cut ties with the transgressed athletes to prevent undesirable spillover. Nike, TAG Heuer, and Porsche suspended endorsement deals with Russian tennis star Maria Sharapova after she admitted failing a drug test (Mullen, 2016). When encountering particular nodes of information relevant to athletes (e.g., Sharapova’s doping scandal and news about her untrustworthiness or incompetence), associated information (e.g., endorsed brands such as TAG Heuer) become not only automatically activated but spread to other relevant nodes of information (e.g., TAG Heuer’s quality and durability). This is suggested by the associative memory model (Collins & Loftus, 1975) and the connectionist model (Fodor & Pylyshyn, 1988). Scholars have frequently identified trustworthiness and competence as important athlete credibility dimensions influencing the evaluation of brands (Fink, Cunningham, & Kensicki, 2004; Spry et al., 2011; Till & Busler, 2000). Hence, we develop the following hypotheses:

Hypothesis 2a: Perceived athlete untrustworthiness is negatively associated with consumers’ attitudes toward endorsed brands.

Hypothesis 2b: Perceived athlete incompetence is negatively associated with consumers’ attitudes toward endorsed brands.

The power of information to spread in memory corresponds to the strength of associations between nodes of information (Anderson, 1983). The strength of association in an endorsement refers to “consumers’ perceived intensity of the connection between celebrity and product brand caused by the shared associative nodes of the two properties in consumers’ memory” (Chang & Ko, 2018, p. 1260). As such, it is likely that the node (i.e., athlete endorser) simultaneously activates the primarily associated nodes (i.e., endorsed brand). Meanwhile, the activated node (i.e., athlete endorser) can also be linked to relatively less directly relevant nodes in memory (e.g., other brands in the same product category) if the strength among nodes is sufficiently strong (Chang & Ko, 2018; Collins & Loftus, 1975). Among the numerous factors influencing the strength of association, Keller (1993) emphasized that the product category can play an important role in establishing strong associations among brands. Roehm and Tybout (2006) also found that the negative impact of brand scandals in the athletic shoe category spilled over to a competitor brand in the same product category. Therefore, we posit that:

Hypothesis 3a: Perceived athlete untrustworthiness is negatively associated with consumers’ attitudes toward competitor brands.

Hypothesis 3b: Perceived athlete incompetence is negatively associated with consumers’ attitudes toward competitor brands.

Researchers argue that endorsed brands can serve a mediating role to bring spillover effects from the misbehaved athletes to competitor brands. This prediction has been made based on the aforementioned memory accounts (Chang & Ko, 2018; Collins & Loftus, 1975; Fodor & Pylyshyn, 1988). Endorsed brands may operate as an associative link to connect athletes involved in an ARC and competitor brands, given that competitor brands may not necessarily be relevant to the athlete. Endorsed brands presumably represent strong associations with both scandalized athletes and competitor brands. Yu, Sengul, and Lester’s (2008) conceptual model with a focus of organizational crisis also support this idea; organizations that are similar to the one who experience a crisis are more likely to receive spillover effects. Since prior studies have not examined spillover effect in the context of athlete endorsement, we, therefore, developed the following hypotheses:

Hypothesis 4a: Perceived untrustworthiness toward endorsed brands spills over to competitor brands via endorsed brands.

Hypothesis 4b: Perceived incompetence toward endorsed brands spills over to competitor brands via endorsed brands.

The Moderating Role of Consumer Knowledge

Consumer’s evaluative judgments vary depending on their knowledge level regarding the target object (Sujan, 1985). Hence, consumer behavior scholars often categorize consumers into experts and novices based on their knowledge level regarding the target objects (Roy & Cornwell, 2004). Experts, relative to novices, have highly organized knowledge in specific object domains. Thus, experts and novices are significantly different in terms of decision-making (Mitchell & Dacin, 1996) and information processing (Roy & Cornwell, 2004).

Spillover effects differ based on the level of knowledge and the strength of association among athlete endorsers, endorsed brands, and competitor brands. For example, experts can identify detailed product attribute information (Alba & Hutchinson, 1987), whereas novices tend to elicit overall evaluation based on affect and spontaneous thoughts (Sujan, 1985). Since experts are more likely to incorporate detailed information about misbehaved athletes into judgment, expert consumers can easily identify associations between a focal athlete and endorsed brands. Therefore, spillover effects are hypothesized as less likely, and the magnitude of effects can be weakened, if consumers are more knowledgeable. The last hypothesis was developed as follows:

Hypothesis 5: Consumer knowledge will have a moderating effect. Specifically, spillover effects from (a) untrustworthiness to endorsed brands, (b) incompetence to endorsed brands, (c) untrustworthiness to competitor brands, and (d) incompetence to competitor brands are stronger for novice consumers than expert consumers.

Method

The authors conducted two pretests prior to the main study. The first pretest was conducted to select a target celebrity endorser for the main study. The second pretest (a focus-group interview with undergraduate students and a survey of the target population) was aimed to verify the association between the target endorser and brands in consumers’ memory, which was necessary to test the spillover effects. The details of the pretests are described below.

Pretests and Stimulus Development

The first pretest was conducted to select a celebrity athlete endorser for this study. Ten athlete endorsers who did not have a significant major scandal history at the time of data collection were selected from a list of the top 100 highest paid athlete endorsers of 2016 (as published by opendorse®). Nineteen undergraduate students from a Southeastern University in the United States participated in the pretest. They were asked to rate familiarity (i.e., how familiar are you with the athletes) and likeability (i.e., how much do you like the athletes) of the list of athlete endorsers on a 7-point Likert-type scale. The method for this pretest follows previous studies (S. Y. Lee & Rim, 2017; Martin & Steward, 2001; Sato et al., 2016) and was deemed acceptable as the aim of the pretest was to obtain descriptive results (Doyle, Pentecost, & Funk, 2014). The results indicated that Chris Paul, a basketball player in the NBA, could be selected due to his high familiarity and likeability scores with small standard deviations (SD; M _familiarity = 6.20, SD _familiarity = .51, M _likeability = 5.1, SD _likeability = .87). These ratings (with small SDs) were desirable for selecting the athlete to control for confounding variables such as personal preferences. For example, LeBron James was rated as more familiar to respondents, but the SDs of his likeability score was larger, which may be problematic in terms of confounding effects. Hence Chris Paul was chosen as the target athlete of the study.

The authors conducted a second pretest to assess whether Paul and his endorsed brand (i.e., Nike) are associated in the target population’s memory. An initial focus-group interview with eight undergraduate students was run. Respondents were asked to provide initial thoughts about Paul. Responses were related to general information about him (i.e., basketball player), affiliated team (i.e., Clippers), and his athletic ability (e.g., fast). Importantly, the participants did mention the endorsement contract with Nike. This implied that Paul and Nike are connected in respondents’ memory.

To further validate the association in the sample population, we employed another pretest by using 25 participants recruited from Amazon Mechanical Turk (Mturk). They were asked to choose an endorsed brand of Paul from a list of sports brands (obtained from a Forbes® web page). The majority of participants (n = 20, 80%) chose Nike as Paul’s endorsed brand, followed by Adidas (n = 4, 16%), and Under Armour (n = 1, 4%). These results further demonstrated that Paul and Nike were cognitively associated in memory. Conducting these rigorous pretests was necessary to establish the baseline that focal endorsers and brands are associated in memory, given that spillover effects cannot be tested without verifying these memory associations.

A short news story about Paul was created as if it appeared on ESPN web page. The news story described a fictional PED case. We chose a PED case because such stories often receive keen attention from the press (Bie & Billings, 2015; Hambrick et al., 2015; Osborne, Sherry, & Nicholson, 2016) and are commonly studied as an important issue (Denham, 2004; Liao & Markula, 2016). The basic tenet of the spillover effect is the impact transfers to associated objects even if the information is not presented in publicity (Ahluwalia et al., 2000).

For the purpose of investigating the spillover effect, specific information of focal brands was absent in the publicity. Nike was selected as an endorsed brand because it is Paul’s official endorsed brand. Adidas was selected as a competitor brand based on the previous literature (Pitt, Parent, Berthon, & Steyn, 2010). The study by Laufer & Wang (2017) similarly support the selection of a competitor brand since companies that belong to the same industry tend to be categorized because of similarities in memory. It could be reasoned that consumers can identify Adidas as one of the Nike’s competitor with ease. The stimulus for this experiment is displayed in Figure 2.

Figure 2.

Experimental scenario of the selected endorser’s PED scandal.

Participants and Procedure

Participants for the main study were recruited through Mturk. Respondents received US$0.30 for their participation. A total of 1,173 participants accessed the survey link and first answered a screening test. The screening question asked participants to select athletes whom they knew at the time of study that consisted of four real professional athletes (i.e., Kobe Bryant, Alex Rodriguez, Ichiro Suzuki, and Chris Paul) and a fictitious athlete (i.e., Ted Franklin; adopted from Till and Busler, 2000). Participants who did not know Chris Paul were excluded from the study; and those who selected Ted Franklin were eliminated as we considered them inattentive participants. As a result, 403 participants were included in the main study. They were then asked to respond to the scales of consumer knowledge. Participants then read the short news article about Chris Paul’s PED case and answered questions about perceived severity, untrustworthiness, incompetence, and attitude toward endorsed and competitor brands. Participants were debriefed at the conclusion of the study that the content of the news story was completely fictional and had nothing to do with Chris Paul’s real life.

One hundred and eighty-seven participants were males (54.6%). The majority of participants were young (e.g., 18–24 years old = 23.0%, 25–34 years old = 49.0%) and relatively low income (e.g., less than US$24,999 = 23.9% and US$25,000–US$59,000 = 45.1%). The majority were Caucasian (n = 219, 64.6%) followed by African Americans (n = 49, 14.5%), Asians (n = 41, 12.1%), and Hispanics (n = 20, 5.9%). The sample distribution in this study is deemed consistent with the demographics of Mturk population, but the interpretation of the results should be exercised with caution for any generalizations because Mturk workers are relatively younger and lower income than the general population in the United States (Ipeirotis, 2010). The characteristics of the participants recruited in Mturk may be different than typical NBA spectators, but it deemed appropriate as the context of the study is publicity in general.

Measures

Perceived severity was measured by the perceived negative publicity scale used in previous literature (Lin, Chen, Chiu, & Lee, 2011). An example of perceived severity scales is “I feel the negative news report about Chris Paul is serious.” All items were measured with a 7-point Likert-type scale ranging from 1 = strongly disagree to 7 = strongly agree. In the previous literature, the internal consistency of the scale was adequate (e.g., α = .91; Grégoire & Fisher, 2008).

Perceived untrustworthiness and incompetence were adopted from Ohanian’s study (1991). Perceived untrustworthiness and incompetence each consisted 3 items and were measured with a 7-point Semantic Differential Scale (i.e., perceived untrustworthiness: 1= honest, 7 = dishonest; 1 = sincere, 7 = insincere; and 1 = trustworthy, 7 = untrustworthy; perceived incompetence: expert–not an expert, experienced–inexperienced, and qualified–unqualified). This scale also showed adequate internal consistency in the previous study (α = .79; Lin et al., 2011).

To measure consumers’ attitude toward endorsed and competitor brands, we adapted MacKenzie and Lutz’s (1989) 7-point Bipolar Scale (i.e., bad–good, unfavorable–favorable, and negative–positive). The internal consistency of both source credibility and brand attitude scales has been verified in the numerous previous studies (e.g., Carlson & Donavan, 2008; Fink et al., 2012).

Consumer knowledge was measured by using a 3-item scale adopted from previous literature (Mitchell & Dacin, 1996; Park, Mothersbaugh, & Feick, 1994). An example is “How familiar are you with Chris Paul?” (1 = not familiar at all, 7 = extremely familiar). The scale was later median-split to divide expert and novice groups. The reliability of this scale has also been verified elsewhere (α = .93; Sierra & Hyman, 2011).

At the end of the survey, the authors included a manipulation check item that reads: “What did Chris Paul do?” Participants were asked to choose one from a choice list, which contained (1) off-field violence, (2) the use of PEDs, (3) sexual infidelity, (4) winning record, and (5) philanthropic activities. After eliminating incomplete responses and those who failed to answer the manipulation check question, a total of 339 useful cases were analyzed.

Data Analysis

Before testing the proposed hypotheses, we first assessed the measurement model of the current study to ensure convergent and discriminant validities. Based on the established technique (Bagozzi & Yi, 1988; Fornell & Larcker, 1981), a confirmatory factor analysis, internal consistencies of each factor, and average variance extracted (AVE) were utilized to ensure convergent validity, while discriminant validity was verified by comparing AVE values to the squared correlations among constructs.

We used structural equation modeling (SEM) to test the hypothesized relationships (Hypotheses 1–4), and multigroup SEM was also employed to assess the moderating effect of consumer knowledge (Hypothesis 5). We employed SEM and multigroup SEM techniques because our conceptual model was relatively complex due to the inclusion of many constructs. The techniques can simultaneously examine structural relationships between multiple latent constructs while taking into account measurement error, which scholars cannot achieve by multiple regression (Davis, Douglas, & Silk 1981). Hence, the use of multigroup SEM deemed appropriate especially when examining a hypothesized model including various constructs. Analyses were run by SPSS and AMOS Version 22.0.

Results

Measurement Model

The results of the measurement model indicated an acceptable model fit (χ² = 228.09, df = 120, χ²/df = 1.91, comparative fit index (CFI) = .98, root mean square error of approximation (RMSEA)= .04, standardised root mean square residual (SRMR) = .04, normed fit index (NFI) = .96) . All factor loadings were in acceptable range (.77–.94; Hair, Anderson, Tatham, & Black, 2009); thus, the convergent validity was confirmed.

As noted in Table 1, Cronbach’s αs were ranged from .88 to .95, supporting the internal consistency of the scales (Nunnally, 1978). To test the discriminant validity, each construct’s AVE and its square intercorrelations of any factors in the model were compared (Fornell & Larcker, 1981). As noted in Table 2, each constructs’ AVE values were greater than the square of intercorrelations, providing evidence of discriminant validity (Hair et al., 2009).

Table 1.

Factor Loadings, Construct Reliability, and Average Variance Extracted (AVE).

Constructs	Items	λ	α	AVE
Perceived severity	I feel the negative news report about Chris Paul is serious	.77	.89	.74
	The news report about Chris Paul makes me feel that he has problems	.91
	I feel that the negative news report truly reveals the problems of Chris Paul	.89
Perceived incompetence	Expert–not an expert	.86	.90	.75
	Experienced–inexperienced	.88
	Qualified–unqualified	.85
Perceived untrustworthiness	Honest–dishonest	.89	.91	.78
	Sincere–insincere	.86
	Trustworthy–untrustworthy	.90
Attitude toward Nike (endorsed brand)	Bad–good	.94	.95	.88
	Unfavorable–favorable	.93
	Negative–positive	.94
Attitude toward Adidas (competitor brand)	Bad–good	.94	.95	.86
	Unfavorable–favorable	.92
	Negative–positive	.92
Consumer knowledge	How familiar are you with Chris Paul?	.81	.88	.71
	I know a lot about Chris Paul	.92
	How would you rate your knowledge about Chris Paul relative to the rest of the population?	.80

Note. All factor loadings are significant at the p < .05 level. Measurement model fit: χ² = 228.09, df = 120, χ²/df = 1.91, CFI = .98, RMSEA = .04, SRMR = .04, NFI = .96.

Table 2.

Means, Standard Deviations (SD), Average Variance Extracted (AVE), Correlations, and Squared Correlations Matrix.

Concepts	Mean (SD)	1	2	3	4	5	6
1. Perceived severity	4.75 (1.38)	.74^a	.04	.23	.01	.01	.01
2. Perceived incompetence	2.63 (1.34)	.21	.75^a	.25	.12	.08	.08
3. Perceived untrustworthiness	4.35 (1.39)	.48	.40	.78^a	.01	.01	.04
4. Attitude toward endorsed brand	5.47 (1.24)	.03	−.29	−.10	.88^a	.32	.02
5. Attitude toward competitor brand	5.22 (1.13)	−.02	−.27	.01	.56	.86^a	.01
6. Consumer knowledge	4.44 (1.24)	.01	−.25	−.20	.14	.09	.71^a

Note. The figures below the AVE line represent correlations between the constructs. The figures above the AVE line represent squared correlations between the constructs.

^a AVE.

Structural Equation Model Test

Structural equation modeling was employed to test Hypotheses 1–4. The hypothesized model demonstrated a good fit with the data (χ² = 208.25, df = 83, χ²/df = 2.51, CFI = .97, RMSEA = .07, SRMR = .08, NFI = .95). The path coefficients from perceived severity to untrustworthiness (Hypothesis 1a; β = .53, p < .01) as well as to incompetence (Hypothesis 1b; β = .28, p < .01) were significant. The causal relationship between untrustworthiness and endorsed brands (Hypothesis 2a; β = .03, p = .55) did not meet significance, whereas the direct effects of incompetence on endorsed brands were significant (Hypothesis 2b; β = −.33, p < .01). The direct effect of untrustworthiness on competitor brands was not significant (Hypothesis 3a; β = .11, p = .09) while that of incompetence on competitor brands emerged significant (Hypothesis 3b; β = −.17, p < .01). A summary of the structural equation model test is presented in Table 3.

Table 3.

The Results of Structural Equation Model.

Concepts	β	Critical Ratio	Hypothesis
1. PS → untrustworthiness	53** (.06)	9.49	Hypothesis 1a supported
2. PS → incompetence	0.28** (.06)	4.69	Hypothesis 1b supported
3. Untrustworthiness → Nike	0.03 (.05)	0.60	Hypothesis 2a rejected
4. Incompetence → Nike	−0.33** (.06)	−5.59	Hypothesis 2b supported
5. Untrustworthiness → Adidas	0.11 (.04)	2.31	Hypothesis 3a rejected
6. Incompetence → Adidas	−0.17** (.05)	−3.22	Hypothesis 3b supported
7. Nike → Adidas	0.54** (.08)	10.44
Indirect Effects	β	Z
Untrustworthiness → Nike → Adidas	0.02 (.03)	0.60	Hypothesis 4a rejected
Incompetence → Nike → Adidas	−0.18** (.03)	−4.26	Hypothesis 4b supported

Note. Standardized coefficients and standard errors appear in parentheses. Model fit: χ² = 208.25, df = 83, χ²/df = 2.51, CFI = .97, RMSEA = .07, SRMR = .08, NFI = .95. PS = perceived severity.

*p < .05. **p < .01.

The Sobel mediation test was conducted by assessing indirect effects of endorser credibility to attitude toward Adidas through Nike. The result indicated that indirect spillover effect from untrustworthiness was not supported (Hypothesis 4a; z = 0.60, β = .02, p = .55). However, the indirect effect from incompetence to attitude toward Adidas through Nike was significant (Hypothesis 4b; z = −4.26, β = −.18, p < .01). The results of all sample SEM are shown in Table 3.

Multigroup Structural Equation Model Based on Consumer Knowledge

To execute the multigroup SEM for testing Hypothesis 5, we assessed whether the measurement model is statistically invariant between expert and novice groups based on the established procedures (Chen, 2007; Netemeyer, Bearden, & Sharma, 2003). When the configural invariance model and factor loading constrained model were compared, we concluded that constraining factor loadings do not reduce model fit. There was no substantial difference in CFI (.003) and RMSEA (.001) based on the existing guidelines (i.e., changes of <.005 in CFI and <.010 in RMSEA indicate invariant factor loading; Chen, 2007) which ensured the factorial invariance.

We then tested the differences of structural coefficients between expert and novice groups. The difference in the χ² statistic was significant, χ²Δ(Δdf = 20) = 95.65, p < .01, indicating that there is at lease a significant difference between the two groups. Inconsistent with our prediction, the results did not show significant group differences on the paths from untrustworthiness to attitude toward the endorsed brands (Hypothesis 5a; β_expert = .02, p = .02; β_novice = .05, p = .55) as well as to competing brands (Hypothesis 5c; β_expert = .10, p = .16; β_novice = .12, p = .06). The path from perceived incompetence to attitude toward the endorsed brand was marginally significant between experts and novices (Hypothesis 5b; β_expert = −.21, p < .05; β_novice = −.39, p < .01; χ² group difference = 3.07, p = .08). Although the impact from perceived incompetence to the competing brand was significant (Hypothesis 5d; β_expert = −.16, p < .05; β_novice = −.16, p < .01), the group difference in path coefficients for expert and novice consumers was not significantly different, χ²Δ(Δdf = 1) = 0.45, p = .50.

To further examine the spillover effects, Sobel tests were performed for both novices and experts respectively to examine the indirect effects of endorser credibility on attitudes toward a competitor brand (Adidas) through an endorsed brand (Nike). The result indicated that indirect spillover effects from untrustworthiness on Adidas through Nike were not supported for experts (z = 0.33, β = .01, p = .74) and novices (z = 0.50, β = .03, p = .62). Nevertheless, the indirect effect from incompetence to attitude toward Adidas through Nike was significant for both experts (z = −2.35, β = −.09, p < .01) and novices (z = −4.01, β = −.25, p < .01). Furthermore, the strength of the indirect effect deemed much stronger for novices. The results of multigroup SEM are summarized in Table 4 and Figure 3.

Table 4.

Results of Multigroup Structural Equation Modeling.

	Expert		Novice
Concepts	β	Critical Ratio	β	Critical Ratio	Hypothesis
1. RS → PS	−.21* (.10)	−2.54	.04 (.08)	0.42
2. PS → untrustworthiness	.56** (.07)	7.43	.53** (.08)	6.22
3. PS → incompetence	.32** (.07)	4.00	.25** (.09)	2.86
4. Untrustworthiness → Nike	.02 (.06)	0.21	.05 (.10)	0.60	Hypothesis 5a
5. Incompetence → Nike	−.21* (.08)	−2.51	−.39** (.09)	−4.83	Hypothesis 5b
6. Untrustworthiness → Adidas	.10 (.06)	1.39	.12 (.07)	1.86	Hypothesis 5c
7. Incompetence → Adidas	−.16* (.07)	−2.06	−.16* (.06)	−2.35	Hypothesis 5d
8. Nike → Adidas	42** (.08)	5.45	.63** (.06)	9.14
Indirect Effects	β	Sobel Test Value	β	Sobel Test Value
Untrustworthiness → Nike → Adidas	01 (.03)	0.33	.03 (.06)	0.53
Incompetence → Nike → Adidas	−.09** (.04)	−2.35	−.25** (.06)	−4.01

Note. Standardized coefficients and standard errors appear in parentheses. Group significant differences between expert and novice are in bolded font. RS = relationship satisfaction; PS = perceived severity.

*p < .05. **p < .01

Figure 3.

The results of multigroup structural equation modeling.

Discussion

The current study provides a more systematic understanding as to how consumer evaluation of misbehaved athletes spills over to endorse and competing brands when the athletes are involved in an ARC. The findings extend the existing athlete endorsement and crisis management literature in several ways. First, consumers’ perceptions of untrustworthiness and incompetence increase with the level of perceived severity, supporting Hypothesis 1a and b. These findings are in accordance with the prior study suggesting that individuals evaluate perpetrators harshly based on the level of negativity regarding focal incidents (Kahneman et al., 1998). Empirical studies also provide evidence that perceived severity is strongly associated with evaluations of wrongdoers (Coombs & Holladay, 2002).

It is also worth noting that perceived severity is deemed more strongly associated with endorsers’ untrustworthiness rather than incompetence. Previous studies have shown that severity is magnified based on the types of transgressions, levels of impact on society, the number of perpetrators, and illegality (Chien et al., 2016; Hughes & Shank, 2005). Although the current study did not aim to identify factors that increase perceived severity, the findings indicated that severity is more strongly associated with trustworthiness, rather than expertise, of athletes who are involved in an ARC.

This study also found that tainted expertise can have a detrimental impact on the endorsed brand. This direct spillover effect has been well-documented in previous literature (Fink et al., 2012; J. S. Lee & Kwak, 2015). Although incompetence was associated with endorsed brands, the impact of untrustworthiness was not found to be significant. Hence, Hypothesis 2a was rejected whereas Hypothesis 2b was supported.

Similarly, regarding the relationships between credibility and competitor brands, the direct spillover effect on the competitor brand was only supported for incompetence. Therefore, Hypothesis 3a was rejected while Hypothesis 3b was supported. The results demonstrate that the associations between the endorser’s expertise and Nike (Hypothesis 2b) as well as Adidas (Hypothesis 3b) were sufficiently strong as these associations could be enhanced due to the congruency between the endorser’s athletic attribute and the product (Keller, 1993).

It is also worth noting that Adidas was more favorably evaluated when they perceive Chris Paul as an untrustworthy figure (Hypothesis 3b). Because the activation of directly and indirectly associated nodes can occur almost simultaneously (Lei et al., 2008), we speculate that consumers’ immediate judgement of the tainted endorsed brand (i.e., Nike) might have made the innocent competitor brand (i.e., Adidas) look better. However, this interpretation is speculative and the detailed psychological process regarding the result requires further research.

Concerning the unsupported negative spillover from untrustworthiness to Nike (Hypothesis 2a) as well as Adidas (Hypothesis 3a), we speculate that the fan effect introduced in the associative memory network model (Anderson, 1983) could provide an alternative explanation. It describes that the more associations each node has, the more challenging it is to activate one specific association. Nike and Chris Paul as nodes are connected in consumers’ mind through repeated pairing, but each node may be connected with many different nodes. Nike and Adidas might not have been salient enough in memory when untrustworthiness was activated. This idea can also be supported by the spreading activation perspective (Chang & Ko, 2018; Janiszewski & Wyer, 2014). Consumers’ judgments of untrustworthiness might not serve a relevant information cue for sport-oriented brands (i.e., Nike and Adidas) to be activated. If that is the case, more trust and reputation centered brands (e.g., insurance companies) may be influenced by untrustworthiness of the athlete endorsers.

Regarding our Hypotheses 4a and b, the results of Sobel tests provide evidence that only perceived incompetence indirectly influences competitor brands via endorsed brands. Therefore, Hypothesis 4a was rejected while Hypothesis 4b was upheld. These findings indicate that the memory associations between the incompetence of an athlete and their endorsed brands can spread to competitor brands even if they are not present in publicity. Although perceived untrustworthiness increases with the severity of an ARC, it did not show that untrustworthiness spilled over to competitor brands through endorsed brands. This might be because the current study only focused on sports equipment brands and product categories can serve as an important information cue for human judgment (Keller, 1993). Athletes’ untrustworthiness might have spilled over to brands that belong to a product category requiring highly trustworthy characteristics like car insurance and medical devices. When sports equipment brands are main focus, perceived incompetence is the catalyst that brings negative impact to competitor brands.

With regard to the differences between expert and novice consumers, the findings indicate that only one path coefficient from incompetence to Nike was significantly higher for novices than experts, supporting Hypothesis 5b. This result is in line with consumer knowledge literature. Experts who have highly organized knowledge about Chris Paul would be able to analyze detailed information about the target (Roy & Cornwell, 2004). Hence, we argue that cognitive associations among endorsers, endorsed brands, and competitor brands can be accurately organized in experts’ minds; and this decreases the degree of negative spillover effects. On the other hand, novices have little information about the athlete to retrieve from memory. They may simply carry the impact of the most readily available information (i.e., Paul’s incompetence) to make heuristic-based judgments of associated objects.

Finally, the authors ran separate Sobel tests for experts and novices to further understand indirect spillover effects from credibility to Adidas. The results indicate that the negative impact on Nike spills over to Adidas when consumers perceive Paul as incompetent, both for experts and novices. One may assume that experts’ judgments should not influence their evaluation of Adidas due to the organized knowledge structure (Alba & Hutchinson, 1987). However, the results indicate that the spillover effect can manifest indirectly if the cognitive association between brands in the same product category is exceptionally strong.

Brand managers must minimize the scandal damage and block the detrimental impact spills over to stakeholders. A common approach that competitor brands can execute to avoid the influence of spillover effect is the denial response strategy (Yu et al., 2008). The results suggest that the associations among the endorser, Nike, and Adidas could be different between novices and experts. Novices do not necessarily have organized knowledge structure about the endorser. The associations among the endorser, Nike, and Adidas might have been fuzzy in memory. In such a situation, individuals tend to engage in categorization (Laufer & Wang, 2017). Adidas as a competitor brand should not only deny but also emphasize nonsport attributes to differentiate themselves from the other two (i.e., endorser and Nike), especially when novices are the main concern. On the other hand, experts have an organized knowledge about the endorser, the stronger associations among the endorser, Nike, and Adidas should exist in memory. Since experts may not need to rely on a sport-oriented categorization, Adidas should execute the denial strategy and emphasize more specific attributes of Adidas that can differentiate from Nike. Finally, speaking of the origin of spillover effect, our data suggest that it is important to pay particular attention to perceived incompetence of athletes who are involved in an ARC. Perceived incompetence is a strong catalyst of the spillover effect and a key factor that can spread negative impact to stakeholders in the same product category, at least in a case of PED cases.

Limitation and Future Research Directions

There are several limitations to the study. First, the methods utilized in this study limit the applicability of findings. The authors focused only on a performance related ARC in an individual athlete setting. The generalizability of the findings is constrained since the results could be different if consumers obtain different ARC information (e.g., sexual infidelity, tax evasion, and assault; Fink et al., 2012; Sanderson, 2010) in an individual sport setting such as golf or tennis. Furthermore, although the authors conducted a pretest to assess the association between the target athlete and his endorsed brand, this test could also have been included in the main study to further support the memory association. Doing so would have made the findings of this study more generalizable.

In line with the generalizability concern, the findings about the spillover effects in this study were due exclusively to the sport industry category (i.e., Nike and Adidas are both in the sport industry). It is indeed possible that spillover effects can manifest because of other factors (e.g., industry type, simplicity of the industry, and impact receiver characteristics; Laufer et al., 2017; Yu et al., 2008). For example, if the focus of this study was nonsport product brands like automobile or insurance companies, it might have shown different patterns of results.

Researchers should examine the range of the ARC damage in future studies. The authors assume that the ARC impact may potentially spill over to different product categories. To further understand the range of spillover effects, future research should be conducted to explore what kinds of information cues may contribute to categorizing stakeholders. For example, phonetic similarity can play a role to categorize stakeholders (Lei et al., 2008). If an athlete endorser of Honda commits an ARC, would there be a negative impact spillover to a phonetically similar brand like Hyundai? These questions await future studies.

Conclusion

Considering the gaps in the endorsement literature, the current study makes both theoretical and practical contributions. Understanding how consumers react when an ARC occurs is an important step for minimizing damage and repairing the image of misbehaved athletes and their associated brands (K. A. Brown, Murphy, & Maxwell, 2017; Coombs & Holladay, 2002; Sanderson, 2008). The current study sheds light on the psychological mechanisms underlying consumer perceptions of ARC severity, endorser credibility, and associated brands. The severity of an ARC has differential impacts on trustworthiness and expertise of athletes.

When sports equipment brands (i.e., Nike and Adidas) are the main concern, perceived incompetence significantly hurts evaluations of endorsed brands. Importantly, even if competitor brands are not directly associated with endorsers, they can receive a detrimental impact indirectly transferred from the endorsed brands. This indirect spillover from perceived incompetence to competitor brands is particularly strong for consumers who do not possess organized or “expert” knowledge about athletes. It implies that marketing communication managers need to implement crisis response strategies based on consumers’ knowledge level.

Communication managers from competitor brands should focus on how to dissociate their brands from tainted brands. Crisis response strategies should be developed separately for experts and novices. Furthermore, to develop well-designed communication plans, managers should emphasize unique brand attributes that are distinct from the brands endorsed by athletes. This is a challenging but important task.

Footnotes

Acknowledgments

The authors appreciate anonymous reviewers and editors for valuable comments on this article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Shintaro Sato

References

Ahluwalia

Burnkrant

R. E.

Unnava

H. R.

(2000). Consumer response to negative publicity: The moderating role of commitment. Journal of Marketing Research, 37, 203–214.

Alba

Hutchinson

(1987). Dimensions of consumer expertise. Journal of Consumer research, 13, 411–454.

Amos

Holmes

Strutton

(2008). Exploring the relationship between celebrity endorser effects and advertising effectiveness: A quantitative synthesis of effect size. International Journal of Advertising, 27, 209–234.

Anderson

J. R.

(1983). A spreading activation theory of memory. Journal of Verbal Learning and Verbal Behavior, 22, 261–295.

Badenhausen

(2017) The 2017 NBA All-Stars: Players who earn the most money from endorsements. Forbes. Retrieved from https://www.forbes.com/sites/kurtbadenhausen/2017/02/19/the-nba-all-stars-2017-players-that-earn-the-most-money-from-endorsements/#3b1040f78f66

Bagozzi

(1989). On the use of structural equation models in experimental designs. Journal of Marketing Research, 26, 271–284.

Bie

Billings

A. C.

(2015). Too good to be true? US and Chinese media coverage of Chinese swimmer Ye Shiwen in the 2012 Olympic games. International Review for the Sociology of Sport, 50, 785–803.

Brown

K. A.

Anderson

M. L.

Dickhaus

(2016). The impact of the image repair process on athlete-endorsement effectiveness. Journal of Sports Media, 11, 25–48.

Brown

Bennett

Ballouli

(2016). Examining the effects of advertisement setting and actor race on African Americans’ intentions to consume baseball. Sport Marketing Quarterly, 25, 139–151.

10.

Brown

K. A.

Murphy

Maxwell

L. C

. (2017). Tried in the court of public opinion: Effects of involvement in criminal transgressions on athlete image. Communication & Sport, 6, 283–307. doi:10.1177/2167479517697426

11.

Bundy

Pfarrer

M. D.

Short

C. E.

Coombs

W. T.

(2017). Crises and crisis management: Integration, interpretation, and research development. Journal of Management, 43, 1661–1692.

12.

Carlson

B. D.

Donavan

D. T.

(2008), Concerning the effect of athlete endorsements on brand and team-related intentions. Sport Marketing Quarterly, 17, 154–162.

13.

Chang

(2018). When infamy becomes fame: The positive side of negative athlete publicity. Journal of Sport Management. doi:10.1123/jsm.2017-0082

14.

Chang

Y. J.

(2018). The effects of association strength on attention and product evaluation: Reconsidering endorsement effectiveness. European Journal of Marketing, 52, 1257–1279.

15.

Chen

F. F.

(2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling, 14, 464–504.

16.

Chien

P. M.

Kelly

S. J.

Weeks

C. S.

(2016). Sport scandal and sponsorship decisions: Team identification matters. Journal of Sport Management, 30, 490–505.

17.

Collins

A. M.

Loftus

E. F.

(1975). A spreading-activation theory of semantic processing. Psychological Review, 82, 407–428.

18.

Coombs

W. T.

Holladay

S. J.

(2002). Helping crisis managers protect reputational assets: Initial tests of the situational crisis communication theory. Management Communication Quarterly, 16, 165–186.

19.

Davis

H. L.

Douglas

S. P.

Silk

A. J.

(1981). Measure unreliability: A hidden threat to cross-national marketing research? The Journal of Marketing, 45, 98–109.

20.

Denham

B. E.

(2004). Sports Illustrated, the mainstream press and the enactment of drug policy in Major League Baseball: A study in agenda-building theory. Journalism, 5, 51–68.

21.

Doyle

J. P.

Pentecost

R. D.

Funk

D. C.

(2014). The effect of familiarity on associated sponsor and event brand attitudes following negative celebrity endorser publicity. Sport Management Review, 17, 310–323.

22.

Fink

J. S.

Cunningham

G. B.

Kensicki

L. J.

(2004). Using athletes as endorsers to sell women’s sport: Attractiveness vs. expertise. Journal of Sport Management, 18, 350–367.

23.

Fink

J. S.

Parker

H. M.

Cunningham

G. B.

Cuneen

(2012). Female athlete endorsers: Determinants of effectiveness. Sport Management Review, 15, 13–22.

24.

Fodor

J. A.

Pylyshyn

Z. W.

(1988). Connectionism and cognitive architecture: A critical analysis. Cognition, 28, 3–71.

25.

Fornell

Larcker

D. F.

(1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50.

26.

Grégoire

Fisher

R. J.

(2008). Customer betrayal and retaliation: When your best customers become your worst enemies. Journal of the Academy of Marketing Science, 36, 247–261.

27.

Hair

J. F.

Anderson

R. F.

Tatham

R. L.

Black

W. C.

(2009). Multivariate data analysis (6th ed.). Upper Saddle River, NJ: Prentice Hall.

28.

Hambrick

M. E.

Frederick

E. L.

Sanderson

(2015). From yellow to blue: Exploring Lance Armstrong’s image repair strategies across traditional and social media. Communication & Sport, 3, 196–218.

29.

Hovland

C. I.

Janis

I. L.

Kelley

H. H.

(1953). Communication and persuasion. New Haven, CT: Yale University Press.

30.

Hughes

Shank

(2005). Defining scandal in sports: Media and corporate sponsor perspectives. Sport Marketing Quarterly, 14, 207–214.

31.

Hutchins

Rowe

(2012). Sport beyond television: The Internet, digital media and the rise of networked media sport. London, England: Routledge.

32.

Ipeirotis

P. G.

(2010). Demographics of Mechanical Turk (NYU Working Paper, NoCEDER-10-01) . Retrieved from, SSRN, http://ssrn.com/abstract=1585030

33.

Ito

T. A.

Larsen

J. T.

Smith

N. K.

Cacioppo

J. T.

(1998). Negative information weighs more heavily on the brain: The negativity bias in evaluative categorizations. Journal of Personality and Social Psychology, 75, 887–900.

34.

Jackson

(2013). Reflections on communication and sport: On advertising and promotional culture. Communication & Sport, 1, 100–112.

35.

Janiszewski

Wyer

R. S.

(2014). Content and process priming: A review. Journal of Consumer Psychology, 24, 96–118.

36.

Kahneman

Schkade

Sunstein

(1998). Shared outrage and erratic awards: The psychology of punitive damages. Journal of Risk and Uncertainty, 16, 49–86.

37.

Keller

(1993). Conceptualizing, measuring, and managing customer-based brand equity. The Journal of Marketing, 57, 1–22.

38.

Kroloff

(1988). At home and abroad: Weighing in. Public Relations Journal, 44, 8–10.

39.

Laufer

Wang

(2017). Guilty by association: The risk of crisis contagion. Business Horizons, 61, 173–179.

40.

Lee

J. S.

Kwak

D. H.

(2015). Consumers’ responses to public figures’ transgression: Moral reasoning strategies and implications for endorsed brands. Journal of Business Ethics, 137, 101–113.

41.

Lee

J. S.

Kwak

D. H.

Braunstein-Minkove

J. R.

(2016). Coping with athlete endorsers’ immoral behavior: Roles of athlete identification and moral emotions on moral reasoning strategies. Journal of Sport Management, 30, 176–191.

42.

Lee

J. S.

Kwak

D. H.

Moore

(2015). Athletes’ transgressions and sponsor evaluations: A focus on consumers’ moral reasoning strategies. Journal of Sport Management, 29, 672–687.

43.

Lee

S. Y.

Rim

(2017). Company–nonprofit partnerships, negative spillover, and response strategies. International Journal of Strategic Communication, 11, 194–208.

44.

Lee

Koo

(2015). Athlete endorsement, attitudes, and purchase intention: The interaction effect between athlete endorser-product congruence and endorser credibility. Journal of Sport Management, 29, 523–538.

45.

Lei

Dawar

Lemmink

(2008). Negative spillover in brand portfolios: Exploring the antecedents of asymmetric effects. Journal of Marketing, 72, 111–123.

46.

Lewis

Weaver

A. J.

(2015). More than a game: Sports media framing effects on attitudes, intentions, and enjoyment. Communication & Sport, 3, 219–242.

47.

Liao

Markula

(2016). The only thing I am guilty of is taking too many jump shots: A Deleuzian media analysis of Diana Taurasi’s drug charge in 2010. Sociology of Sport Journal, 33, 169–179.

48.

Lin

C. P.

Chen

S. C.

Chiu

C. K.

Lee

W. Y.

(2011). Understanding purchase intention during product-harm crises: Moderating effects of perceived corporate ability and corporate social responsibility. Journal of Business Ethics, 102, 455–471.

49.

Lohneiss

Hill

(2014). The impact of processing athlete transgressions on brand image and purchase intent. European Sport Management Quarterly, 14, 171–193.

50.

MacKenzie

S. B.

Lutz

R. J.

(1989). An empirical examination of the structural antecedents of attitude toward the ad in an advertising pretesting context. The Journal of Marketing, 53, 48–65.

51.

Martin

I. M.

Stewart

D. W.

(2001). The differential impact of goal congruency on attitudes, intentions, and the transfer of brand equity. Journal of Marketing Research, 38, 471–484.

52.

Mitchell

Dacin

P. A.

(1996). The assessment of alternative measures of consumer expertise. Journal of Consumer Research, 23, 219–239.

53.

Mudrick

Burton

Lin

C. A.

(2017). Pervasively offside: An examination of sexism, stereotypes, and sportscaster credibility. Communication & Sport, 5, 669–688.

54.

Mullen

(2016). Maria Sharapova loses endorsement deals after failing drug test. CNN Money. Retrieved from http://money.cnn.com/2016/03/08/news/nike-maria-sharapova-nike-suspension-doping/index.html

55.

Netemeyer

R. G.

Bearden

W. O.

Sharma

(2003). Scaling procedures: Issues and applications. Thousand Oaks, CA: Sage.

56.

Nunnally

J. C.

(1978). Psychometric theory. New York, NY: McGraw-Hill.

57.

Ohanian

(1991). The impact of celebrity spokespersons’ perceived image on consumers’ intention to purchase. Journal of Advertising Research, 31, 46–54.

58.

Osborne

Sherry

Nicholson

(2016). Celebrity, scandal and the male athlete: A sport media analysis. European Sport Management Quarterly, 16, 255–273.

59.

Park

Mothersbaugh

Feick

(1994). Consumer knowledge assessment. Journal of Consumer Research, 21, 71–82.

60.

Pedersen

P. M.

Laucella

Kian

Geurin

(2016). Strategic sport communication (2nd ed.). Champaign, IL: Human Kinetics.

61.

Pitt

Parent

Berthon

Steyn

P. G.

(2010). Event sponsorship and ambush marketing: Lessons from the Beijing Olympics. Business Horizons, 53, 281–290.

62.

Roehm

Tybout

(2006). When will a brand scandal spill over, and how should competitors respond? Journal of Marketing Research, 43, 366–373.

63.

Roy

D. P.

Cornwell

T. B.

(2004). The effects of consumer knowledge on responses to event sponsorships. Psychology & Marketing, 21, 185–207.

64.

Sanderson

(2010). Framing Tiger’s troubles: Comparing traditional and social media. International Journal of Sport Communication, 3, 438–453.

65.

Sanderson

(2008). “How do you prove a negative?” Roger Clemens’s image-repair strategies in response to the Mitchell Report. International Journal of Sport Communication, 1, 246–262.

66.

Sato

Y. J.

Kaplanidou

Connaughton

D. P

. (2016). Consumers’ comparative evaluative judgment of athlete endorsers. Journal of Sport Management, 30, 553–565.

67.

Sato

Y. J.

Park

Tao

(2015). Athlete reputational crisis and consumer evaluation. European Sport Management Quarterly, 15, 434–453.

68.

Sierra

J. J.

Hyman

M. R.

(2011). Outlet mall shoppers’ intentions to purchase apparel: A dual-process perspective. Journal of Retailing and Consumer Services, 18, 341–347.

69.

Smith

A. K.

Bolton

R. N.

Wagner

(1999). A model of customer satisfaction with service encounters involving failure and recovery. Journal of Marketing Research, 36, 356–372.

70.

Spry

Pappu

Cornwell

B. T.

(2011). Celebrity endorsement, brand credibility and brand equity. European Journal of Marketing, 45, 882–909.

71.

Sujan

(1985). Consumer knowledge: Effects on evaluation strategies mediating consumer judgments. Journal of Consumer Research, 12, 31–46.

72.

Till

Busler

(2000). The match-up hypothesis: Physical attractiveness, expertise, and the role of fit on brand attitude, purchase intent and brand beliefs. Journal of Advertising, 29, 1–13.

73.

Till

B. D.

Shimp

T. A.

(1998). Endorsers in advertising: The case of negative celebrity information. Journal of Advertising, 27, 67–82.

74.

Sengul

Lester

R. H.

(2008). Misery loves company: The spread of negative impacts resulting from an organizational crisis. Academy of Management Review, 33, 452–472.