Fandom, forgiveness and future support: YouTube apologies as crisis communication

User as audience: theoretical approaches to strategic communication

In the proceedings of an international colloquium during which practitioners and academics debated the impact of social media on corporate communication, social media were hailed as “the new mantra for influence” and recognized as having “a huge impact on corporate reputation” (Kaul and Chaudhri, 2015). However, the recognition that strategic communication is vital to positive relational outcomes between organizations and stakeholders is not new. Grunig's excellence theory of public relations (1992) revealed that good relationships were of value to organizations because they reduced the organization's costs by: lessening the need for litigation, decreasing the likelihood of imposed regulations, reducing negative publicity and providing products and services aligned with stakeholders' needs. Grunig categorized models of organization–stakeholder communication, claiming the two-way symmetrical model was the most ethical and effective. In the two-way symmetrical model, “the public should be just as likely to persuade the organization's management to change attitudes or behavior as the organization is likely to change the publics' attitudes or behavior” (Grunig and Hunt, 1984, p. 23).

Grunig's approach was challenged by researchers who put forth the contingency theory of accommodation, a theory more sensitive to situational variables (Cancel et al., 1997). The theory posits that communication with stakeholders ranges on a continuum from accommodation to advocacy. For example, if a product delivery was delayed because of a natural disaster, such as a tornado, a customer may expect an explanation, but not expect to receive an apology and a discount. However, if the delay was caused because the company failed to maintain adequate stock, the customer may expect an apology and compensation (such as a discount). Deciding the level of accommodation or advocacy is based on a number of variables, such as corporate culture and public expectations (Cancel et al., 1999). More recently, Coombs’ (2011) SCCT focuses on determining the appropriate level of accommodation or advocacy. SCCT, developed from the psychological framework of attribution theory, places audiences' attitudes at the center of crisis response and an integral factor in deciding the level of accommodation or advocacy. Coombs and Holladay (2012) determined social media is an appropriate means to monitor and measure audience response to crisis communication strategies – as it serves as a type of environmental scan.

Choo (2001) categorized types of environmental scanning – ways of seeking and using information about the external environment – as a method of assessing the level of accommodation or advocacy required to meet strategic communication goals. The Internet has become a key environmental scanning tool for strategic communication professionals as technology dictates the need for rapid response, especially in times of misinformation or crises (Coombs and Holladay, 2012; Crawford, 1999; Strauβ and Jonkman, 2017). The focus of media monitoring is increasingly moving online and is an “essential part” of the daily routine of communication professionals (Strauβ and Jonkman, 2017, p. 41). Crawford (1999) encourages communication professionals to “Eavesdrop all you want, learning the attitudes of specific audiences and the issues most important to them” (p. 44).

An important consideration in the social media apology environment is the unknown audience, leading to collapsed context (Boyd, 2010) – not knowing one's audience in order to be socially appropriate and be understood: “without information about audience, it is often difficult to determine how to behave, let alone to make adjustments based on assessing reactions” (p. 50). The audience and its consideration in strategic communication decisions is a key theme in image repair scholarship (Benoit, 1997, 2000, 2014; Burns and Bruner, 2000; Coombs, 2011). Strategic communication decisions, such as when and how to apologize, have consequences on audience perceptions (Benoit, 2014; Bisel and Messersmith, 2012; Coombs, 2007). Much of the research informing these decisions was derived from studies using a mass mediated framework (typically used by communication professionals for image repair strategies) or an interpersonal framework (typically used for relationship repair strategies). However, social media do not fit neatly into either a mass media or interpersonal framework, but are somewhat of a hybrid. Scholars disagree on how to categorize social media and, in particular, how to accommodate the blurring of the boundaries between mass and interpersonal communication advanced by technology (Caplan, 2001; O'Sullivan and Carr, 2017; Procopio and Procopio, 2007; Sheth and Solomon, 2014; Walther et al., 2010a). Because of this duality of function – both mass and personal – communication professionals cannot rely solely on previous research from either framework to guide them in how to apologize via social media for optimal relational outcomes.

Perceptions of sincerity and forgiveness influence relational outcomes

There is evidence that an audience's perception of apologies as sincere and their willingness to forgive the transgressor influences relational outcomes. Multiple studies have concluded that apologies that are perceived as sincere are more effective, restore relationships and reputations and can lead to forgiveness (Choi and Chung, 2012; Darby and Schlenker, 1982; Gold and Weiner, 2000; Sandlin and Gracyalny, 2018; Schlenker and Darby, 1981; Schmitt et al., 2004; Tomlinson et al., 2004; McCullough et al., 1997; Waldron and Kelley, 2008). Harrison–Walker's research determined “forgiveness plays a critical role” regarding desirable outcomes and is integral to all models of service recovery regardless of the industry (2019, p. 386). In other words, if audiences detect sincerity in an apology, they are more likely to forgive the transgressor and, once forgiven, the transgressor is more likely to be able to forge positive relationships with the audience. Therefore, an effective apology is key to relational outcomes. However, assessing the effectiveness of social media apologies presents challenges.

Previous studies have approached the effectiveness of social media apologies through content analysis of YouTube comments (Coombs and Holladay, 2012; Sandlin and Gracyalny, 2018; Thelwall, 2017; Thelwall et al., 2012; Walther et al., 2010b). Although there is value in monitoring social media comments (Coombs and Holladay, 2012; Edgerly et al., 2013; Siersdorfer et al., 2010; Thelwall, 2017), it's also important to recognize that the comments reflect the sentiment of the commenting public, and not the majority of the viewers – the noncommenting public. After Cerulo and Ruane (2014) analyzed 183 celebrity apologies using Benoit's (1995) typology of image restoration strategies, they concluded further research was needed, specifically a design that considered “the characteristics of those evaluating the apology” (p. 145) to gauge the impact of evaluators' social profiles, (e.g. similarities between the audience and the offenders and audience affinities for the offender). In a study analyzing the comments posted on public figures' YouTube apologies, the authors cautioned against “assuming that content analysis of comments will give a complete picture of stakeholder perceptions–since commenters' motivations may differ from non-commenting viewers” (Sandlin and Gracyalny, 2018, p. 403).

Attitudes of audiences

Reputation. Previous studies have highlighted the importance of fostering a positive pre-crisis reputation (Aula, 2011; Shim and Yang, 2016; VanSlyke Turk et al., 2012). They've noted that social media can serve to confirm users' pre-existing attitudes when related to online searches, reviews and information seeking (Knobloch-Westerwick et al., 2015; Winter et al., 2016; Yeo et al., 2015; Yin et al., 2016). Earlier studies have indicated that positive attitudes toward the apologizer increase the persuasiveness of a YouTube apology (Manika et al., 2015). Sandlin and Gracyalny (2018) found that positive viewer comments about the public figure's reputation posted adjacent to the YouTube apology video were related to perceptions of apology sincerity, and public figures whose apologies were deemed sincere were more likely to be forgiven by YouTube commenters. However, they noted “a striking lack of evidence that online audience perceptions, as reflected in the comments, were related to the theories of either interpersonal apology or image repair” (p. 9). Instead they considered the possibility of pre-existing attitudes influencing audience perceptions. As an example, they quoted one YouTube commenter who wrote in response to John Mayer's apology:

IT'S SIMPLE [sic]. If you're a true John Mayer fan you'll stick with him no matter what, cause you know that he's got a good heart. … I will always love him no matter what. Whatever stupid thing he might do, he's still a brilliant musician and a wonderful person. Nothing will ever change that fact:) (p. 9).

Fandom. Social media have introduced new ways to gather like-minded audiences and engage fans. An example of the heightened importance of social media is Nielson's efforts to revamp the company's traditional television rating system to incorporate social media activity (Cassella, 2015). Previous studies have indicated that audiences who identify as fans have greater information-seeking intentions related to their fandom and are more engaged in fan-related activities (Groene and Hettinger, 2016; Tsay-Vogel and Sanders, 2017). The convergence of media has created a particularly fruitful space for participatory culture, where fans not only engage but also create materials with wide-reaching consequences. In Spreadable Media, the authors note, “New platforms create social, cultural, economic, legal and political change and opportunities for diversity and democratization” (Jenkins et al., 2018, p. xii). The authors identify the tension between these new participants and their potential to destabilize the corporate communication environment. In addition, Hills argues that “online fandoms cannot merely be viewed as a version or reflection of ‘offline’ fandoms. The mediation of ‘new media’ must be addressed rather than treated as an invisible term within the romanticised ‘new’” (Hills, 2002, p. 135). Fans, whether online or offline, do share a favorable attitude toward the object of their fandom (the person, product or service), and past research has indicated fandom influences behavior.

Attractiveness. Audience perceptions of the attractiveness of the offender is another notable attitude that has been shown to have a variety of implications, especially with respect to social competence and influence (Dion et al., 1972; Eagly et al., 1991). This relationship has been tested regarding offenders' attractiveness and its influence on forgiveness. Phillips and Hranek (2012) found male participants judged the apology as higher in quality when it was offered by the attractive offender, whereas female participants rated the apology as higher quality when it was offered by the less attractive offender.

Therefore, knowing the attitude the audience holds toward the public figure prior to viewing the YouTube apology is important in gaining a deeper understanding of what is more salient to audiences – the apology or the audience's pre-existing attitudes in relation to sincerity and forgiveness. Therefore, we asked the following questions:

Audience demographics and relatedness to apologizer. As far back as the 1890s when American psychologist Williams James wrote, “Neither threats nor pleadings can move a man unless they touch some one of his potential or actual selves” (James, 1890/1981, p. 297), scholars have considered relatedness to the audience as a possible mediator of the message. Social psychological research has generally supported the similarity-attraction theory that posits individuals who are similar to each other are attracted to each other, and this concept goes beyond attitudes and includes demographics (Byrne, 1971; McPherson et al., 2001; Umphress et al., 2007). In studies examining second-hand forgiveness – people who have not been wronged directly but who identify with the victims of a transgression – they found people who identified strongly with the in-group (in this case, the victims) felt “vicariously harmed” and found it difficult to forgive (Brown et al., 2008, p. 1416). Therefore, people who identify demographically with the public figure may be more empathetic toward the public figure and be more willing to forgive. This led us to ask the following question:

Intentions of future support. Future support is key to the health of any organization. There is much evidence that apologies do play a role in post-crises reputation repair, but there are conflicting studies regarding if a repaired reputation necessarily translates to future support and economic health. In Choi and Chung's study (2012), the results indicated that an apology was an effective strategy for repairing the organization's reputation, but it did not increase their purchase intentions. Separate works by Harrison-Walker (2019) and Fombrun (2018) indicated corporate communication strategies – including apologies – could be used to reconcile customers, reduce negative word-of-mouth and positively impact the economic health through renewed stakeholder support. Therefore, we asked:

Covariates: audience's perceived severity of transgression and guilt of apologizer. Finally, the varied conditions under which apologies are made (e.g. guilty vs not guilty, severity of transgression, etc.) make it challenging to measure apology effectiveness. Coombs’ SCCT posits preventable offenses are perceived as more severe than accidents. In addition, perceived severity is intensified if the organization has a history of crises or a poor reputation (Coombs, 2011). Psychological studies have found the severity of the offense impacts forgiveness – the more severe the transgression, the more difficult it is to forgive (Brose et al., 2005; Wade and Worthington, 2003).

Guilt also plays a role in apologies. If the audience perceives the organization as guilty of the transgression, denying it or not responding to the allegations is significantly less effective than the positive response strategy (e.g. the organization says it is making changes); and if the transgressor initially denies guilt, but later is found guilty, anger and reputational damage intensify (Coombs et al., 2016). Since transgression severity and the perceived guilt of the offender have previously been shown to impact response to apologies, we included them as covariates.

Procedure

Adult participants in the United States were recruited through Amazon's Mechanical Turk to complete an online survey via Qualtrics. Respondents were required to complete an audio and video equipment check to ensure that they could see and hear YouTube videos before beginning the survey. Participants were randomly assigned to one of two apology conditions; each condition contained the YouTube apology videos of two public figures with varying demographic characteristics. The apologies chosen for the study met the following criteria: (1) made by a highly visible public figure as determined by coverage of the apology on either the top Nielson-rated broadcast news channel (NBC) or cable news channel (FOXNEWS) for that year; (2) made between 2009 and 2014; and (3) posted on YouTube as a video. Because political communication is influenced by partisanship, politicians were excluded from the study. To obtain the public figure apologies, two search engines were used: Google and Waybackmachine.org. First, Google was searched using the terms “celebrity,” “sports,” “corporate,” “entertainment,” “artists,” “company” or “public figure,” followed by the word “apology” and then the year (ex. 2009, 2010, 2011, 2012, 2013 or 2014). From the initial list of available apologies, Waybackmachine.org was used to determine whether the apology appeared on either of the top-rated broadcast news or cable news networks. Of these apologies, only the videos that were posted on YouTube were selected for inclusion. Of the 28 eligible apologies, four were selected based on the public figure's demographic characteristics (i.e. sex, race/ethnicity, age) and length of the apology video (less than 2 min).

Participants were randomly assigned to view two of the four apologies. Participants in condition 1 viewed the YouTube apologies of singer Chris Brown (male/young/non-White) and celebrity chef Paula Deen (female/mature/White). Participants in condition 2 watched the apologies of athlete Serena Williams (female/young/non-White) and Carnival Cruise Lines CEO Gerry Cahill (male/mature/White). Young was defined as under 35 years old; mature was defined as 35 years old and above. Within each condition, the order of apologies was randomized to reduce potential order effects. For each apology, participants first answered questions regarding their pre-existing attitudes (if any) of the public figure and the offense, then viewed the 1–2 min apology video posted to YouTube and finally answered questions about their perceptions of the apology, the public figure's guilt, apology sincerity, their likelihood to forgive and demographic information. All participants were compensated US$0.25 for completing the survey.

Sample

Four hundred and twenty-seven participants completed the survey. Of those, 24 were removed for failing the equipment check, thus the final sample consisted of 403 participants, 211 (52.4 percent) men, 191 (47.4 percent) women and 1 (0.2 percent) other/preferred not to answer. Most participants (n = 276; 73.4 percent) defined themselves as White (non-Hispanic). Participants also identified as African American or Black (n = 35; 8.7 percent), Hispanic/Latinx (n = 23; 5.7 percent), Asian (n = 35; 8.7 percent), Native American (n = 1; 0.2 percent), Native Hawaiian or Pacific Islander (n = 1; 0.2 percent), Two or more races (n = 10; 2.5 percent) and two participants (0.5 percent) preferred not to answer. The majority (n = 182; 45.2 percent) of participants were between 25 and 34 years of age, 75 (18.6 percent) between 18 and 24, 79 (19.6 percent) between 35 and 44, 49 (12.2 percent) between 45 and 54, 14 (3.5 percent) between 55 and 64 and 4 (1.0 percent) participants were between 65 and 74 years of age.

Measures

Participant and public figure demographics. Participants reported their sex, which was coded as 1 for male or 2 for female. They also indicated their race; however, because there were very few non-White participants (e.g. African American, Hispanic, Asian, Native American), race was categorized as 1 for White (73.4 percent) or 2 for non-White (26.6 percent). Age was coded as 1 for participants age 34 and under (63.8 percent) and 2 for participants 35 and older (36.2 percent). Public figure sex, race and age were coded using the same categories.

Demographic similarity/difference. To create the measure of sex similarity/difference, participant–public figure dyads were coded as 1 if the participant and public figure were of the same sex and 2 if they were of the opposite sex. Race similarity/difference was categorized as 1 if the participant and public figure were either White or non-White and 2 if one of the members was White and the other non-White. Finally, age similarity/difference was coded as 1 if both members were aged 34 or under or both were aged 35 and above; it was coded as 2 if one member was 34 or under and the other was aged 35 or above.

Attitudes toward the public figure and the offense. At the beginning of the survey, participants were asked a series of Likert-type questions to assess their attitudes toward the public figure and the offense before viewing each YouTube apology video. Only participants who were familiar with the public figure/organization completed these measures. Fandom was assessed using a five-point Likert-type scale (“Are you a fan of (public figure)?”; 1 = definitely not; 5 = definitely yes); Reputation was measured by asking participants “Do you think he/she is a good person?” (1 = definitely not; 5 = definitely yes). Participants were also asked if they had heard of the offense (briefly described in the questionnaire) and to assess the Severity of the event (1 = not at all severe; 5 = extremely severe).

Perceptions of apology sincerity, forgiveness and future support. After watching each YouTube apology video, participants were asked another series of Likert-type questions to measure their perceptions of the public figure, the apology, their likelihood to forgive him/her and their likelihood to purchase/support their work or organization in the future. The public figure's physical Attractiveness was measured by asking participants to rate him/her on a five-point scale (1 = not at all attractive, 5 = very attractive). Guilt was assessed by asking participants whether they believed the figure was guilty of the offense (1 = definitely not; 5 = definitely yes). Sincerity was measured by asking participants, “After watching the apology video, how sorry do you think he/she is? (1 = not at all sorry; 5 = extremely sorry). Forgiveness was assessed by asking, “How likely would you be to forgive him/her?” (1 = not at all likely; 5 = extremely likely). Finally, participants indicated their level of Future Support by rating their likelihood to watch/listen to/buy the products of the public figure or their organization in the future (1 = not at all likely; 5 = extremely likely).

Audience attitudes and perceptions of sincerity

A hierarchical regression analysis was conducted to examine the impact of Reputation, Fandom and Attractiveness on perceptions of sincerity with the measures of offense severity and perceived guilt entered in the first step as covariates. An examination of the correlations is shown in Table I. The regression analysis showed that the variables entered in Step 1 were significantly associated with audience perceptions of sincerity, F (2,479) = 14.44, p < 0.001, R = 0.24, adj. R² = 0.06. The model significantly improved when audience attitudes toward the public figure were added, F (5,476) = 15.89, p < 0.001, R = 0.38, adj. R² = 0.15, F_change = 15.96, p < 0.001. Thus, Fandom, Reputation and Attractiveness emerged as significant positive predictors of perceived apology sincerity (see Table II).

Audience attitudes and willingness to forgiveness

To examine the impact of audience attitudes on forgiveness, a second hierarchical regression analysis was conducted. Again, perceived guilt and offense severity were entered in the first step as covariates, with Fandom, Reputation and Attractiveness entered in Step 2. The regression analysis showed that the covariates entered in Step 1 significantly associated with forgiveness, F (2,478) = 68.69, p < 0.001, R = 0.47, adj. R² = 0.22. The model improved once the attitudes toward to the public figure were added, F (5,475) = 52.71, p < 0.001, R = 0.60, adj. R²= 0.35, F_change = 32.95, p < 0.001, indicating that 35 percent of the variance in forgiveness was predicted by the public figure's perceived guilt, severity of the offense and the audience attitudes toward the public figure (see Table II).

Demographic “sameness”

The third research question asked how audience demographics, including demographic similarity to the public figure, affected perceptions of apology sincerity and forgiveness. Results showed that there were no significant differences between men and women in overall perceptions of sincerity (t (400) = −0.40, p = 0.69) or forgiveness (t (399) = −0.56, p = 0.58). There were also no significant differences between White and non-White participants in perceived apology sincerity (t (400) = 1.06, p = 0.29) or forgiveness (t (399) = 1.52, p = 0.13). Finally, there were no significant differences between participants under the age of 35 or those 35 years old and above in perceptions of sincerity (t (400) = −0.40, p = 0.69) or forgiveness (t (399) = −0.56, p = 0.58).

The study also examined how demographic similarity affected audience perceptions of sincerity and forgiveness. For similarity in sex, paired t-tests found no significant differences in participants' perceptions of sincerity in apologies made by same-sex (M = 3.02, SD = 1.26) or opposite-sex (M = 2.94, SD = 1.20) public figures, t (402) = 1.04, p = 0.30, r = 0.17. There were also no significant differences in participants' reported forgiveness of same-sex (M = 3.22, SD = 1.24) or opposite-sex (M = 3.21, SD = 1.25) public figures, t (402) = 0.16, p = 0.875, r = 0.19. There was, however, a significant interaction between participant sex and sex composition for forgiveness, F (4, 396) = 4.97, p = 0.001, such that male participants were more likely to forgive female figures, (i.e. opposite-sex dyads, M = 3.41, SD = 1.19) than male figures (same-sex dyads, M = 3.08, SD = 1.12). Female participants were also more likely to forgive female figures (i.e. same-sex dyads, M = 3.39, SD = 1.25) than male figures (opposite-sex dyads, M = 2.99, SD = 1.25).

Regarding race similarity, paired t-tests found no significant differences in participants' perceptions of sincerity for apologies made by same-race (M = 3.01, SD = 1.21) or different-race (M = 2.95, SD = 1.23) public figures, t (402) = 0.75, p = 0.45, r = 0.12. Likewise, there were no significant differences in forgiveness of same-race (M = 3.29, SD = 1.18) or different-race (M = 3.14, SD = 1.28) public figures, t (402) = 1.87, p = 0.06, r = 0.19.

With respect to age similarity, the results of paired t-tests showed no significant differences in participants' perceptions of sincerity for apologies made by similar-age public figures (M = 2.99, SD = 1.22) or those made by a different-age demographic (M = 2.98, SD = 1.22), t (402) = 0.13, p = 0.90, r = 0.15. There were also no significant differences in forgiveness of similarly aged public figures (M = 3.22, SD = 1.26) and public figures of a different age (M = 3.22, SD = 1.22), t (402) = 0.03, p = 0.98, r = 0.18.

Future support

The final research question (RQ4) asked about the relationship between audience forgiveness and intentions of future support of the public figure's work or organization. Results indicate a significant positive relationship between participants' willingness to forgive public figures and their likelihood to support them in the future, r = 0.64, p < 0.001. In addition, because prior support (i.e. Fandom) may strongly influence the likelihood of future support, a hierarchical regression analysis was conducted to determine whether Forgiveness contributes to the prediction of Future Support over and above that which can be accounted for by Fandom. Thus, Fandom was entered as a covariate in Step 1, and Forgiveness was entered in Step 2. The regression analysis showed that Fandom does indeed positively predict Future Support, F (1,365) = 249.54, p < 0.001, R = 0.64, adj. R² = 0.40, accounting for 40 percent of the variation in Future Support. Adding Forgiveness to the model explained an additional 17 percent of the variance in Future Support, above and beyond that which was accounted for by Fandom, F (2,364) = 244.78, p < 0.001, R = 0.76, adj. R² = 0.57, F_change = 142.96, p < 0.001. Partial regression coefficients are displayed in Table III. The results suggest that forgiveness, independently of fandom, also predicts the likelihood of future support.

Sincerity

One aspect of an effective apology is sincerity (Gold and Weiner, 2000; McCullough et al., 1997; Waldron and Kelley, 2008), and this study contributed new insights by demonstrating that audience attitudes are related to perceptions of sincerity of YouTube apology videos of public figures. In the study, audiences who perceived the public figure as having a good reputation, considered themselves a fan of the public figure or considered the public figure attractive were more likely to perceive the public figure's apology as sincere. Also, the more severe the offense and perceived guilt of the offender, the less likely audiences were willing to forgive. Both males and females were more likely to forgive females. The two audience characteristics with the strongest relationships to perceptions of sincerity were reputation and fandom.

Reputation management is often the purview of communication professionals, and there is much research that speaks to the benefits of fostering a positive reputation (Aula, 2011; Coombs, 2011; Shim and Yang, 2016; Šontaitė-Petkevičienė, 2014; VanSlyke Turk et al., 2012). This study adds to that research by demonstrating that reputation is also related to audience perceptions of apology sincerity. The findings affirm a positive reputation and fandom are valuable in times of crises. If the audience viewing the apology video already considers the public figure as having a good reputation, they are more likely to perceive the apology as sincere. If a client's reputation is less than stellar, their apology may be deemed less sincere and of lower value in terms of reputation repair. This is an important consideration as communication professionals consider image repair strategies for their clients. The relatedness of reputation to the perception of sincerity affirms the importance of actively pursuing positive pre-crisis reputation management strategies.

Fandom, future support and forgiveness

In addition, this study demonstrates that a positive reputation could serve as a cyclical multiplier – a positive reputation fosters perceptions of sincerity and sincerity fosters forgiveness (Sandlin and Gracyalny, 2018) and helps restore reputations (Choi and Chung, 2012). The study also indicated that forgiveness relates to intentions of support (i.e. buying the product again, refraining from spreading negative word of mouth, etc.). This is important because even participants who did not identify as fans were more likely to have intentions of future support if they forgave the transgressor. Therefore, this study demonstrated that sincere apologies, with their potential to enhance forgiveness, can foster audiences' intentions of future support and, according to previous studies, can result in a positive impact on the reputation and economic health of an organization (Fombrun and Low, 2011; Harrison-Walker, 2019; Tsarenko and Tojib, 2012). Therefore, communication strategists must recognize and convey the significant impact of an apology opportunity.

Our study also demonstrates a strong relationship between fandom and perceptions of sincerity and forgiveness. Therefore, communication professionals may want to prioritize engaging fans if a situation arises where an apology is necessary and/or an apology is posted on social media.

Attractiveness and audience attitudes matter

The finding that perceived attractiveness is related to the perception of sincerity is consistent with much earlier findings in psychological research indicating that attractiveness has many positive social implications (Dion et al., 1972; Nisbett and Wilson, 1977; Thorndike, 1920). This “halo effect” (Thorndike, 1920) surrounding attractiveness stereotypes also held true in our study. The perceived attractiveness of the apologizer, as rated by the audience, was related to perceptions of sincerity. Again, this is important because a sincere apology is more likely to be forgiven (Gold and Weiner, 2000; Darby and Schlenker, 1982; Schlenker and Darby, 1981; Schmitt et al., 2004) – even in mediated communication channels such as YouTube (Sandlin and Gracyalny, 2018). The audience's perception of the severity of the offense and its relationship in reducing the likelihood of forgiveness is also consistent with previous psychological research (Fincham et al., 2005). Together, these findings demonstrate that audience’s characteristics and attitudes, especially with regard to the apologizer, impact their perceptions of YouTube apologies.

Our findings, highlighting the role of audience attitudes related to perceived sincerity and willingness to forgive, is consistent with Coombs (2011) SCCT. In SCCT, audiences' attitudes are central to crisis response, and reputation is an intensifying factor (positive or negative), as is the nature of the offense. The study also affirms the idea that how audiences view the crisis is important. In our study audiences considered the severity of the offense and guilt of offender; in Coombs' work, they considered who was responsible. In both studies, the perceptions of the audience were strongly related to their response to the apology. Our research extends Coombs' approach by indicating that is it important not only to understand audiences' attributions post-crisis but also to gauge attitudes (reputation, fandom and severity of transgression) as an ongoing practice. This study emphasizes the need for apology opportunities to be carefully considered in relationship to how they will be perceived by online audiences because apologies have the potential to impact relational outcomes and, in doing so, an organization's future.