How Online Incivility Affects Consumer Engagement Behavior on Brands’ Social Media

Abstract

Research on consumer engagement in social media is flourishing. However, online incivility is rampant and its effect on consumer engagement is unknown. The current work posits long-term consumer engagement with a brand is decreased when consumer-to-consumer uncivil interactions take place on brands’ social media channels. Using behavioral data from Facebook, the first study documents that a consumer’s incivility to another consumer increases the victim’s engagement in the short term but decreases their engagement over the long term. Further, a brand’s response mitigates these effects. Two follow-up studies using scenario-based experiments provide evidence that consumer injustice perceptions mediate a confrontation coping strategy, while ostracism perceptions mediate an avoidance coping strategy. The experiments also evidence that a brand response mitigates some of the effects of incivility. However, an uncivil interaction from a brand advocate can ostracize a victim despite a brand response. Together, our work furthers consumer engagement and consumer incivility theory while also suggesting that practitioners should manage incivility on brands’ social media pages.

Graphical Abstract

Keywords

consumer incivility consumer engagement brand social media Facebook

Online incivility, or “disrespectful statements for the purpose of attacking,” has become so rampant that it is affecting brands’ social media usage (Kim and Kim 2019, p. 220). For example, reports suggest Starbucks is considering closing its Facebook page due to users’ uncivil comments (Guzman 2021). Research has even begun to evidence that social media fosters outrage and incivility by design (Munn 2020). Despite this prevalence, incivility’s effects on consumers are not well documented. Currently, most brands do not address incivility (Bacile et al. 2018). However, if incivility is harmful, then brands are overlooking the harm. Therefore, the current research explores whether and how consumer-to-consumer (C2C) incivility affects victims’ engagement on social media.

To develop a rationale for why uncivil interactions matter, we draw insight from cognitive theory on stress and coping (Folkman 1984; Lazarus 1999). This theory suggests that incivility is a stressor resulting in coping strategies. We postulate that victims of incivility commonly employ two strategies that affect their engagement: confrontation and avoidance. Whereas confrontation entails immediate engagement to rebuff an uncivil attacker, avoidance happens later as victims ruminate over the incivility when it becomes salient (e.g., when they contemplate responding to a brand’s social media post). Put another way, incivility has dual effects; it increases short-term engagement but decreases long-term engagement.

Study 1 tests the proposed effects using behavioral data from Facebook and Amazon Mechanical Turk (MTurk) workers to identify uncivil interactions. The results indicate that consumers’ uncivil attacks increase a victim’s activity on a brand’s social media channel immediately after the attack, but decrease the victim’s future activity. However, a brand’s response to the victim can nullify these effects. Two follow-up experiments then test consumers’ psychological responses to incivility. More precisely, these studies show that an uncivil attack increases short-term engagement due to perceptions of injustice and the desire to respond defensively, but decreases long-term consumer engagement because of perceptions of ostracism. We find that a brand response can mitigate defensive responses from injustice and decreased engagement from ostracism. However, even when a brand responds, ostracism can persist if the attacker is a brand advocate.

The above theory and results help the current research bridge the consumer engagement, incivility, and coping literature streams, while also offering three contributions. First, the current research reveals how consumer incivility negatively affects customer engagement. Incivility represents a dark side of engagement, an underexplored phenomenon in which consumer empowerment can create negative outcomes. Second, we theorize and evidence the mediating mechanisms from incivility, showing the critical role of injustice and ostracism perceptions. These constructs are new to the engagement literature and potentially explain reactions in other engagement contexts. Finally, our last contribution is the identification of unique harm from a brand advocate’s incivility. This finding is novel and intriguing, especially given that brands promote brand advocates through social media badges, which can serve as a signal of group membership that exacerbates ostracism. Consumer incivility and engagement research has ignored the characteristics of an attacker, which our findings suggest are important determinants of victims’ reactions. In sum, the present research suggests consumer incivility on social media deserves attention from practitioners and academics.

Conceptual Development

Consumer Engagement on Social Media

At its core, consumer engagement represents interactive experiences between consumers and brands (Harmeling et al. 2017; Hollebeek, Srivasta, and Chen 2019). Some research considers engagement as “comprising cognitive, emotional, and behavioral dimensions” and possibly encompassing purchase behavior (Brodie et al. 2013, p. 107; Kumar et al. 2010). Still, other research considers engagement to comprise only consumers’ non-purchase behavior (e.g., word-of-mouth) that contributes to a firm’s marketing function (Harmeling et al. 2017; van Doorn et al. 2010). We adopt this latter approach as our studied phenomenon is behavioral and typically unrelated to purchases.

On a brand’s social media page, consumer engagement consists of consuming, contributing, and creating content, which increases consumers’ brand relationship quality, loyalty, and lifetime value (Hudson et al. 2016; Simon and Tossan 2018). Because consumer engagement is iterative, it can be highly variable and manifest differently in the short- and long term (Brodie et al. 2011). Typically, engagement begets more engagement, but research has documented non-linear effects as consumer engagement can diminish, stop abruptly, or be initiated by firm incentives and promptings (Fehrer et al. 2018; Viswanathan et al. 2017). Thus, examinations of engagement are benefited by examining multiple time periods.

Brands do not control consumer interactions; therefore, consumer engagement can have a dark side when consumers negatively affect each other (Harmeling et al. 2017). Sometimes considered negative engagement (i.e., unfavorable brand-related thoughts, feelings, and behaviors; Do, Rahman, and Robinson 2020), these dark-side actions are often anti-brand. Research on negative engagement has mostly had an anti-brand focus, noting the drivers, outcomes, and moderators of such engagement (Do, Rahman, and Robinson 2020; Hollebeek and Chen 2014; Rahman et al. 2022).

Interestingly, negative engagement is not just brand focused because it can target other consumers (Bowden et al. 2017). Such negative engagement can even be conducted by a brand advocate (Brodie et al. 2013). While the negative engagement by a brand advocate may be a mistake, such as offering misleading brand-related advice, it can also be purposeful, such as attacking other consumers (Hollebeek, Kumar, and Srivastava 2020; Ilhan, Kübler, and Pauwels 2018). Thus, even when incivility comes from a brand advocate, it can be a form of negative consumer engagement to the extent the incivility creates “co-destruction within a service relationship” (Naumann, Bowden, and Gabbott 2020, p. 1471). However, the extent incivility is co-destructive is not fully known. While researchers have explored how online incivility can be harmful in consumer situations, the engagement literature has not incorporated these findings. We now cover research in multiple domains that documents incivility’s effects on consumers.

The Problem of Incivility

As shown in Table 1, incivility research has focused primarily on three contexts: politics, online, and business (i.e., employees and customers), although the experience and outcomes of incivility are remarkably similar across contexts. Research in all contexts conceptualizes incivility as rude and disrespectful behavior. Management research distinguishes incivility from aggression by denoting incivility’s ambiguous intent or low intensity (e.g., Andersson and Pearson 1999). Given our context, we follow online incivility research using “incivility” as an umbrella term encompassing high- and low-intensity behaviors, including harassment, rude critique, and attacking others’ statements (Anderson et al. 2014; Coe, Kenski, and Rains 2014).

Table 1.

Incivility across multiple contexts.

Topic	Exemplar research	Common definition of incivility	Types of incivility studied	Summation of relevant findings
Political incivility	Mutz and Reeves (2005); Brooks and Geer (2007); Sydnor (2019)	“Attacks that go beyond facts and differences, and move instead toward name-calling, contempt, and derision of the opposition” (Brooks and Geer 2007, p. 1)	Interruptions without apology, gratuitous asides, raised voices, eye rolling, inflammatory comments, outrage, and ad hominem attacks	Incivility can increase attitude polarization, political engagement, and amusement, but can also foster negative emotions, decrease political trust, and prompt political disengagement.
Online incivility	Anderson et al. (2014); Coe, Kenski, and Rains (2014); Antoci et al. (2016); Kim and Kim (2019)	“Disrespectful statements for the purpose of attacking” (Kim and Kim 2019, p. 220)	Name-calling, flaming, rude critique, social harassment, hate speech, expletives, vulgarities, aspersions, and pejoratives	Incivility increases anxiety, depression, attitude polarization, and social isolation. Incivility also decreases media consumption on the channel where the incivility occurs.
Business incivility	Andersson and Pearson (1999); Cortina et al. (2001); Lim, Cortina, and Magley (2008)	“Low-intensity deviant behavior with ambiguous intent to harm the target, in violation of workplace norms for mutual respect” (Andersson and Pearson 1999, p. 457)	See below	See below
Employee-employee incivility	Porath, Macinnis, and Folkes (2011); Porath, Macinnis, and Folkes (2011); Hershcovis (2011)	“Insensitive, disrespectful, or rude behavior directed at another person that displays a lack of regard for that person” (Porath, Macinnis, and Folkes 2011, p. 292)	Condescending/derogatory remarks, unreasonable demands, demeaning notes, interruptions, exclusion, berating, and unfair accusations	Employee incivility leads to increased psychological distress, burnout, job withdrawal, and turnover intentions as well as decreased job satisfaction.
Consumer-employee incivility	van Jaarsveld, Walker, and Skarlicki (2010); Wilson and Holmvall (2013); Henkel et al. (2017)	“All forms of rude, disrespectful, condescending, or degrading customer behaviors toward an employee” (Henkel et al., p. 132)	Eye-rolling, sighing, rude/insulting remarks, grumbling, inappropriate gestures, continued complaining, condescension, verbal attacks, and unreasonable demands	Consumer incivility leads to increased emotional labor, exhaustion, and absences as well as decreased job satisfaction, working memory, and sales performance.
Consumer-consumer incivility	Caza & Cortina (2007); Bacile et al. (2018); Okan and Elmadag (2020)	“When rude or offensive comments are made toward an individual” (Bacile et al. 2018, p. 60)	Rude/insulting remarks, interruptions, hostile looks/stares, unfair accusations, flaming, name-calling, challenging others’ statements, and verbal attacks	Consumer incivility causes anger, distress, and withdrawal from brands for victims. Incivility also affects the service climate of brands’ social media page and disrupts service/recovery.

Notes: The references for citations that only appear in Table 1 are available in the Web Appendix.

Even low-intensity forms of incivility can be highly insidious. In the workplace, rude behavior toward employees, whether from other employees or consumers, prompts conflict-avoidance coping strategies, emotional exhaustion, and workplace burnout (Cortina and Magley 2009; Cortina et al. 2001). Rude social exchanges between students can affect institutional satisfaction (Caza and Cortina 2007). Even witnessing incivility can cause consumers to avoid firms (Porath, MacInnis, and Folkes 2011; Okan and Elmadag 2020). Uncivil C2C interactions online can interrupt service recovery attempts and affect other observers’ evaluations of a brand (Bacile et al. 2018). Further, online C2C incivility affects the perceived service climate of a brand’s social media channel and experiential value (Bacile 2020).

Notably, incivility is not always harmful in consumer contexts. For example, witnessing incivility toward an employee can increase consumers’ feelings of warmth toward the employee (Henkel et al. 2017). In addition, initial research suggests that when brands address incivility during service recovery, it can increase consumers’ justice perceptions (Bacile et al. 2018). The political literature notes that incivility simultaneously increases and decreases political engagement (Sydnor 2019). Thus, multiple bodies of literature document the positive and negative effects of incivility. This literature suggests that incivility may have dual effects on consumer engagement, an idea we further develop below.

How Consumer Incivility Affects Consumer Engagement

To understand the effect of C2C incivility on consumer engagement, we utilize cognitive theory on stress and coping (Lazarus 1999; Folkman 1984). This theory predicts coping reactions to stressors, with the understanding that coping has “two widely recognized functions: regulating stressful emotions and altering the troubled relation causing the distress” (Folkman et al. 1986, p. 993). According to this view, two groups of coping strategies exist: 1) problem-focused, active efforts to eliminate a problem and 2) emotion-focused efforts to reduce the distress caused by a given stressor (Lazarus Richard & Folkman, 1984). These groups represent approach-and-avoidance coping strategies (Carver, Scheier, and Weintraub 1989), which prior has work linked to incivility (Hershcovis et al. 2018).

Two such coping strategies, one from each group, are pertinent to our work: confrontation (i.e., aggressive efforts to change a situation) and avoidance (i.e., behavioral withdrawal from a situation; Folkman et al., 1986; Lazarus Richard & Folkman, 1984). These two strategies result from risks of losing respect, appearing incompetent, or appearing unethical (Folkman et al. 1986), all of which are associated with uncivil C2C exchanges on brands’ social media pages (Bacile et al. 2018). Importantly, Lazarus Richard & Folkman, 1984 posit that people often employ multiple coping strategies, a view we adopt in this study. The first step in coping with a stressor is an appraisal of the situation, including evaluations of the harm in an encounter (i.e., the primary appraisal), source of the harm, and coping potential (i.e., the secondary appraisal, Folkman 1984). We propose that these appraisals take two forms in the current context: consumer injustice and perceived ostracism. We now elaborate on how incivility results in the two coping strategies and why injustice and ostracism mediate these effects.

Confrontation from consumer injustice perceptions

Individuals use confrontation coping when they perceive that a situation can be changed (Lazarus Richard & Folkman, 1984). This strategy entails standing one’s ground, expressing feelings, and getting an attacker to change their mind. Involving a degree of “hostility and risk taking” (Folkman et al. 1986, p. 993), confrontation copes with stress by directly dealing with its cause. For confrontation to be a viable coping strategy, an individual must identify the source of the stress (i.e., the attacker; Folkman et al. 1986). In the current context, a victim can certainly identify an uncivil attacker and may confront the attacker by challenging their statements and rebutting their use of incivility.

The mediating appraisal process of a confrontation strategy should be consumer interpersonal injustice, referred to hereafter simply as “consumer injustice” and defined as a perceived lack of fairness in a consumer’s actions toward another (either consumer or employee; Rupp et al. 2008). Uncivil interactions violate social norms; therefore, incivility should cause recipients to consider the exchange unfair (Hershcovis 2011). Justice research has a rich history in marketing, establishing the central role of organizational justice in understanding responses to employees’ rude behavior (e.g., interactional justice), procedures governing service recovery (e.g., procedural justice), and remedies provided for service failures (e.g., distributive justice; Tax, Brown, and Chandrashekaran 1998). In addition, research on incivility has identified that rudeness fosters perceptions of injustice (Rupp and Spencer 2006), something that has been extended from organizational justice to C2C justice perceptions (Bacile et al. 2018).

Justice perceptions also pertain to confrontation. For example, a lack of justice in organizational contexts predicts employee retaliation (Skarlicki and Folger 1997). Similarly, employees lash out at consumers who treat them unjustly (Skarlicki, van Jaarsveld, and Walker 2008). In the consumer literature, unfairness in consumption situations prompts complaint intentions as a form of confrontation (Gelbrich 2010). In summary, incivility creates injustice perceptions, which foster confrontation coping to manage the harm of incivility. Such confrontation often takes the form of rebuttals, as a victim replies defensively to an attacker, which is a form of consumer engagement (Simon and Tossan 2018).¹ Thus, we propose the following hypotheses:

H1a. Consumer incivility on a brand’s social media page exhibits a positive relationship with victims’ short-term engagement.

H1b. Consumer injustice mediates the relationship between consumer incivility and victims’ increased short-term engagement.

Avoidance from perceived ostracism

Research suggests that confrontation of uncivil behavior often leads to an escalation of reciprocal aggression (Andersson and Pearson 1999; Hershcovis et al. 2018; Hmielowski, Hutchens, and Cicchirillo 2014). Rather than being deterred by conflict, uncivil attackers can enjoy the chance to respond with further incivility (Andersson and Pearson 1999). Thus, confrontation is unlikely to end incivility, which will likely lead a victim to avoidance coping. The goal of “avoidance is to manage the emotions from the stress by distancing oneself from the stress” (Hershcovis et al. 2018, p. 165). Unlike confrontation, avoidance coping arises from unchangeable situations, as represented by a spiraling uncivil exchange (Folkman 1984). Given the inability to curb incivility, individuals commonly use avoidance in response to incivility (Antoci et al. 2016; Cortina and Magley 2009).

Perceptions of ostracism are likely to mediate avoidance coping. Perceived ostracism is defined as “a perception of being ignored, excluded, or rejected by others, which deprives the perceiver of feelings of belongingness” (Caza and Cortina 2007, p. 337). Incivility and ostracism form a subset of mistreatment behaviors, with incivility predicting ostracism (Howard, Cogswell, and Smith 2020). Incivility from peers is particularly ostracizing (Caza and Cortina 2007). More fundamentally, group members selectively apply rudeness toward out-group members as a form of rejection (Cortina et al. 2011).

As a sense of social rejection, ostracism hurts the perceiver’s self-esteem and leads to loneliness and depression (Leary 1990; Lee and Shrum 2012). The rejection of ostracism triggers withdrawal from a group except for the most devoted (Gómez et al. 2011). In the current context, the group is a brand’s consumers and followers on a social media channel who defend the brand by attacking other individuals (Ilhan, Kübler, and Pauwels 2018). A victim is likely to feel ostracized from distress caused by incivility from those perceived as a brand’s customers. Ostracism then fosters avoidance coping through withdrawal from the social media channel. Subsequently, we propose the following:

H2a. Consumer incivility on a brand’s social media page exhibits a negative relationship with victims’ long-term engagement.

H2b. Perceived ostracism mediates the relationship between consumer incivility and victims’ long-term engagement.

Moderators of the Effects of Incivility

There are likely many possible moderators of incivility’s effect, including consumer, brand, relationship, and cultural characteristics. Rather than try to identify all possible moderators, we focus on two that relate directly to the interaction and are managerially relevant: (a) whether a brand replies to the victim of incivility, and (b) whether an uncivil attacker is perceived as a brand advocate or as a troll.

Brand reply

Brands rarely address incivility on their social media pages, although brands often respond to consumers’ compliments, complaints, and questions (Bacile et al. 2018; Baer 2016).² Even when incivility is unaddressed, a brand’s reply to a victim of incivility may mitigate the effects of the incivility for several reasons. First, a brand’s reply can redirect a victim’s attention to the brand’s message. Researchers have posited that a brand’s comment on an instance of incivility negates the disruptive effect of an uncivil comment on service recovery, likely due to the consumer paying less attention to the incivility (Bacile et al. 2020). Hence, the effect of incivility on injustice and ostracism is less impactful if a target pays less attention to it by reading a brand’s comment.

Second, a brand’s response is often part of a service recovery process designed to re-establish fairness (Schaefers and Schamari 2016). Online C2C incivility often occurs when consumers complain because critical comments about a brand provide others the opportunity to defend one brand and ridicule another (Ilhan, Kübler, and Pauwels 2018). Injustice from such encounters becomes part of the overall service recovery process, with a brand’s comment re-establishing some fairness (Bacile et al. 2018). Thus, a brand’s reply may reduce the effect of injustice on a victim’s need for confrontation.

Third, a brand reply can mitigate the rejection of ostracism. Given that ostracism represents exclusion from a group, a brand is a member of the excluding group whose comment is a form of support to the victim of ostracism. Wittenbaum, Shulman, and Braz (2010) have found that an individual’s sense of belongingness within a group is stronger with a mix of support and non-support, versus only non-support. Put another way, brands can limit ostracism by validating a victim’s comment and suggesting acceptance of the victim’s presence. Given the above reasons, we offer the following moderation hypotheses:

H3a. The positive effect of consumer incivility on victims’ short-term engagement is weaker when there is a brand reply, as compared to when there is no brand reply.

H3b. The positive, mediated effect of consumer incivility on victims’ short-term engagement through perceived injustice is weaker when there is a brand reply as compared to when there is no brand reply.

H4a. The negative effect of consumer incivility on victims’ long-term engagement is weaker when there is a brand reply as compared to when there is no brand reply.

H4b. The negative, mediated effect of consumer incivility on victims’ long-term engagement through perceived ostracism is weaker when there is a brand reply as compared to when there is no brand reply.

Type of attacker

Certain uncivil exchanges arise from badly behaved brand advocates (i.e., “badvocates,” Bacile et al. 2018) who use their brand knowledge to confront complainants rudely and defend their valued brand. More typically, uncivil exchanges result from infrequent visitors (Coe, Kenski, and Rains 2014). Such visitors may be trolling, defined as “malicious online behavior, intended to disrupt interactions, aggravate interactional partners, and lure them into fruitless argumentation” (Coles and West 2016, p. 233). Trolling is considered easy to identify, and victims of online incivility question an attacker’s identity partly to determine whether the attacker is a troll (Coles and West 2016; Suler 2016).

Naturally, a brand advocate represents a group encompassing a brand’s consumers, whereas a troll does not. Group members are the ones who reject others from a group, meaning that incivility from a brand advocate should be particularly ostracizing (Dahl, Niedbala, and Hohman 2018). In accord with this idea, rudeness from group members uniquely increases psychological distress, ostracism, and behavioral withdrawal (Caza and Cortina 2007; Williams 2007). Such effects exist in online environments, wherein ostracism can result from a lack of “likes,” and group membership can be represented by a name and profile picture (Sacco et al. 2014; Schneider et al. 2017). In summary, incivility from brand advocates, as compared to trolls, exacerbates ostracism. As a result, we propose the following:

H5. The negative mediated effect of consumer incivility on victims’ long-term engagement through ostracism is stronger when the incivility is from a brand advocate as compared to a troll.

Study 1

The purpose of Study 1 was to provide a test of H1a, H2a, and H3a—the direct effect hypotheses—using existing behavior data. We collected data from Facebook in 2017 as it was the most popular online social media platform worldwide, with nearly 2 billion users (Fiegerman 2017). Custom written Python code³ communicated with the Facebook Graph Application Programming Interface (API) in 2017 to crawl consumers’ interaction records from the Facebook brand pages of six service companies: McDonald’s, Starbucks, Applebee’s, Macy’s, Target, and Lowes. The first three companies are in the food industry, and the latter three are retailers. These firms’ social media channels met criteria advanced by Kozinets (2002) for investigating online environments: they were relevant segments to our research and they had many messages, message posters, and between-member interactions.

In the following write-up, “comment” refers to an original post by a consumer, “consumer reply” refers to a post by another consumer on the thread of the original comment, and “self-reply” refers to additional posts by the original consumer in the thread of their original comment. Consumers who posted at least one comment on any of the six service brands’ Facebook pages over a 6-month period were the focus of the study. Similar to Dhaoui and Webster’s (2021) methodology, the Python code collected the comments and all direct replies to the comment, including self-replies. For each commenter, we collected 180 days of post behavior (for a dependent variable) and previous behavior (for a control). As such, our entire observation window spans 1 year. In total, we collected 21,552 comments within the focal 6 months but removed 228 non-English and 274 blank comments. Among the remaining comments, 16,293 were from unique users as some users posted multiple comments. We kept only users’ first comments because the dependent variable was future user behavior, and multiple comments from the same user would violate assumptions of independence. Keeping or removing multiple comments did not change the results. An illustration of our data collection and variable construct is available in the Web Appendix (see Figure WA1).

Variable Operationalizations and Measures

Interacting on a brand’s social media page is a basic form of consumer engagement, dictated by consumers’ satisfaction and trust with a brand (van Doorn et al. 2010). We operationalized engagement as comment posts and likes on the brand’s social media page (Simon and Tossan 2018). We operationalized short-term engagement in two ways, examining a commenter’s self-replies and engagement activity for 7 days after their original comment. The 7-day window was necessary because our pilot study evidenced that some exchanges lasted days. Further, previous research supports the time frame (Hansen, Kupfer, and Hennig-Thurau 2018). To assess long-term consumer engagement, we examined a user’s engagement behavior for 180 days (though we did not include the first week after the original comment).

We operationalized incivility as whether a consumer reply was rude (e.g., “You are an idiot!”) or challenged a comment (e.g., “I’ve been there several times and have never seen this before”). A “rude” statement and a “challenging” statement from a stranger were both considered uncivil in the pilot study.⁴ Prior studies have confirmed MTurk workers’ reliability for simple classification tasks (Marge, Banerjee, and Rudnicky 2010). For each consumer reply, we asked two MTurk workers to rate its sentiment in terms of incivility. If two workers could not agree on the rating of a consumer reply, we then asked two additional workers to rate the reply once again (Mturk.com 2017). The first round of rating yielded a disagreement rate of 27.8% on the incivility rating. After one round of re-rating, the disagreement rate dropped to 13.5%, indicating high interrater reliability. When there was a conflict between the two new workers, one researcher manually reviewed the unsettled reply to reconcile the disagreement. In total, 1228 comments had uncivil replies out of 4621 comments that had replies (27%). The incivility variable was marked as a dummy variable, in which a zero meant a comment had no uncivil replies and a one meant a comment had one or more uncivil replies.

We controlled for two confounding influences. First, we collected a user’s previous engagement behavior before their original comment to control for past behavior. Second, we controlled for the sentiment of consumers’ comments because a negative sentiment could represent a complaint and a lower likelihood of future engagement. The TextBlob Python programming module determined the sentiment of the comments using its polarity measure (Loria et al. 2020).⁵ TextBlob is a natural language processing package that has been frequently used to inspect the sentiment of online user-generated content (e.g., Stöckli and Khobzi, 2021). Al-Natour and Turetken (2020) reported that TextBlob is a significant predictor (p < .01) of human judge scores and is more accurate (error rate = 16.27%) than a Lexicon-based tool (e.g., VADER). We used the default sentiment analysis in TextBlob, PatternAnalyzer, which is based on the pattern library for natural language processing (De Smedt and Daelemans 2012).⁶

Study 1 Hypothesis Testing

Aspects of the data influenced the analytical choice for hypothesis testing. First, the data were nested as the comments came from six brands and were grouped within the brands’ posts. Second, the dependent variable, consumer engagement behavior, was a count variable with many zeros and large, positive skew. To account for these features, a generalized linear multilevel model incorporated a negative binomial distribution with Laplace estimation to facilitate convergence (by proc mixed in SAS v. 9.4). Interactions were assessed using multiplicative terms between the presence of an uncivil consumer reply and the brand response variables, which were at Level 1. All continuous variables were grand mean centered.

Table 2 lists three different analyses, one for each dependent variable. In the first two columns (M1 and M2 in Table 2), the presence of one or more uncivil replies increased the number of self-replies (b = 2.849, t = 32.85, p < .001) and the engagement activity in the first 7 days after the comment (b = 0.358, t = 2.65, p < .001), which supports H1a. In the third column (M3), the presence of an uncivil consumer reply decreased engagement behavior after 7 days (b = −0.848, t = −3.06, p < .001), which supports H2a. Lastly, the interaction of a brand’s response and an uncivil consumer reply was at least marginally significant in all three analyses. For the number of self-replies (b = −2.354, t = −13.30, p < .001) and engagement in the first 7 days (b = −0.617, t = −1.73, p = .08), a brand’s response interacted negatively with the presence of an uncivil consumer reply. For engagement behavior after 7 days, a brand’s response exhibited a positive interaction (b = 1.340, t = −2.17, p = .03). The interactions are illustrated in Figure 1, depicting the predicted dependent variable values converted from their log form. Figure 1a and b reveal that an uncivil consumer reply increases the expected rate of a self-reply and short-term engagement, but a brand’s response reduces the expected rate. In contrast, Figure 1c reveals that an uncivil reply reduces the expected rate of long-term engagement unless a brand responds to the consumer. Thus, the data support both H3a and H4a.

Table 2.

Effect of uncivil replies on consumer engagement on a brand’s social media page.

	Number of replies a user made to their own comment (i.e., self-replies)	Engagement activity for a 7-day window starting right after an uncivil reply from another consumer	Engagement activity for a 173-day window starting 7 days after an uncivil reply from another consumer
Model	(M1)	(M2)	(M3)
−2 log likelihood	7916.31	10,691.44	6483.39
df	6	6	6
∆ −2 log likelihood from baseline model (∆df)	−1236.87*** (3)	−8.49*** (3)	−10.41*** (3)
Intercept	−3.771 (.903)***	−2.473 (.249)***	−3.571 (.395)***
Uncivil reply _{(0 = no uncivil replies; 1 = one or more uncivil replies)}	2.849 (.087)***	0.358 (.135)***	−0.848 (.277)***
Brand response _{(0 = brand did not respond; 1 = brand responded at least once).}	2.151 (.121)***	−0.011 (.209)	−0.186 (.384)
Brand response $\times$ uncivil reply	−2.354 (.177)***	−0.617 (.358)+	1.340 (.617)*
Post sentiment	−0.379 (.092)***	−0.293 (.099)***	0.582 (.150)***
Previous activity of consumer before uncivil reply	−0.010 (.207)	1.973 (.318)***	0.960 (.089)***

***p < .001; ** p < .01; * p < .05; + p < .1.

Notes: Numbers to the left of parentheses are estimated unstandardized coefficients whereas numbers in parentheses are standard errors. Engagement activity is the number of comments and likes on a brand’s social media page during the specified timeframe. The previous activity variable was 7 days for M1 and M2 and 173 days for M3. Baseline models for −2 log likelihood comparisons include an intercept and controls (i.e., previous activity and post sentiment).

Figure 1.

Study 2: The interaction of the presence of consumers’ uncivil replies and a brand’s response on victims’ self-replies, short-term engagement, and long-term engagement.

Study 1 Discussion

The results provide credence to the research’s focus on C2C incivility’s influence on consumer engagement. Our results suggest an uncivil reply spurs immediate engagement, most likely because a victim is defending themselves. After this flurry of activity, incivility decreases consumers’ engagement on a brand’s social media page. This effect was robust to controlling a consumer’s previous behavior and disposition (based on the sentiment of the consumer’s post). Although Study 1 facilitated observing actual behavior, it did not allow for the testing of the mediating process or for ensuring incivility was causal in the observed effects. For example, an argumentative Facebook user could be the type to foster incivility and confront other users. To overcome these concerns, we subsequently undertook an experimental study.

Study 2

Study 2 Design

Study 2 tested the predicted and observed results in Study 1 in a controlled setting, aiming to assess cognitions in the mediation hypotheses (i.e., H1b–H4b). Study 2 was a 2 × 2 between-subjects experiment with manipulations of a consumer comment (uncivil vs supportive) and brand reply (none versus a brand response). The context of the post was a consumer complaint as research has found that this type of consumer post is likely to elicit uncivil responses (Bacile et al. 2018). Subject recruitment used MTurk, which research has shown to be comparable to data collected from other sources (Casler, Bickel, and Hackett 2013). MTurkers disproportionately use Amazon (Hulland and Miller 2018). Consequently, we chose Amazon for the study context to ensure respondents had experience with the brand and that the engagement measures would be meaningful.

Subjects were exposed to a mockup complaint from the pilot study and asked to imagine the complaint was their own. Subsequently, subjects saw the uncivil reply manipulation and the brand reply manipulation (see Supplemental Appendix for manipulations). We intentionally made the uncivil reply low-intensity to provide a conservative test of incivility and to abide by research that restricts incivility to low-intensity behaviors (e.g., Andersson and Pearson 1999). Manipulations were pre-tested using MTurk participants, who reported on a two-item, 10-point scale (e.g., “How rude is the response?”). The uncivil reply was perceived as more uncivil (M = 7.0, SD = 2.1) than the supportive reply (M = 2.5, SD = 2.5, F = 37.51, p < .001). For the brand response manipulation, we copied the text of Amazon’s reply to a similar complaint posted on their Facebook page, which contained an apology, a request to contact consumer service for further help, and no mention of the uncivil reply (shown in the Supplemental Appendix). This type of reply is a common tactic when brands respond to complaints with C2C incivility (Bacile et al. 2018).

Before the manipulations, subjects answered questions concerning their social media use frequency, social media complaint experience, attitude toward Amazon, and filler questions (e.g., attitude toward Apple). After the comment and brand reply manipulations, subjects responded to scales for the mediators, a measure of shopping enjoyment (as a control), and outcome variables. For short-term engagement, we measured a subject’s intention to respond defensively to the uncivil reply manipulation because defending oneself is constituent of confrontation. For long-term engagement, we measured general consumer engagement intentions. Finally, subjects provided demographic information and read a funneled debriefing.

Subject Recruitment

Following current data quality standards for MTurk, we required subjects to have strong reputations (i.e., greater than 98% approval ratings across 1000 prior assignments) to participate (Peer, Vosgerau, and Acquisti 2014). Compensation in the MTurk market relates positively to data quality (Litman, Robinson, and Rosenzweig 2015). As such, subjects were paid $1.25 for participation, which equated to a higher rate than the minimum wage, given the expected time of the survey (approximately 7 minutes). We assessed inattentiveness by using an instructional manipulation check and examined sample appropriateness by using the questions about social media usage and complaint experience (Cheung et al. 2017).

A total of 325 MTurkers were recruited, with 25 (7.7%) removed due to never using Facebook and 12 (4%) removed due to failing one of two attention checks (an instruction to mark a specific answer or a free text response asking about the scenario). The portion of the sample that passed the attention checks (96%) was higher than that reported in student samples (e.g., 14%–46%; Oppenheimer, Meyvis, and Davidenko 2009). The final sample size was 288.

Study 2 Measures

Existing measures for consumer injustice and perceived ostracism did not fit the study objectives and we did not find any scales to measure short-term engagement through defensive responding. Existing scales were framed either from employees’ standpoint (Cortina and Magley 2009) or in terms of fairness (Bacile et al. 2018) rather than unfairness. As such, we undertook the scale development procedures prescribed by Churchill (1979) to develop measures for consumer injustice, perceived ostracism, and short-term engagement. As part of the procedure, we first examined the literature and culled adaptable items. We also conducted interviews with consumers (students and non-students) who had experience in online complaining to develop more items that could represent the domains of the constructs. We then allowed experts to remove any redundant items or those that did not fit the definitions of the constructs. These procedures produced 10 items for consumer injustice, nine items for perceived ostracism, and five for short-term engagement intentions.

Separate exploratory and confirmatory analyses reduced the item pool to provide psychometrically sound measures. Scale assessment began with a sample of undergraduate marketing students from a large U.S. university. Subjects were to simulate an online complaint, which was then met with an uncivil reply within the survey, followed by the scales of interest. An exploratory factor analysis then removed weak item loadings (< .5) or high cross-loadings (> .33) on an unintended construct (Netemeyer, Bearden, and Sharma 2003). This process resulted in a five-item scale for consumer injustice, a six-item scale for perceived ostracism, and a three-item scale for short-term engagement intentions.

Simon and Tossan’s (2018) eight-item scale measured long-term engagement intentions, comprised of three sub-dimensions (consuming, contributing, and creating). We added a stem that asked subjects about the likelihood of their social media usage in the future (rather than the past). A three-item scale from Dawson, Bloch, and Ridgway (1990) measured shopping enjoyment. Single items measured social media usage and social media complaining experience.

Study 2 Preliminary Analysis

Using Mplus (v. 7.11), we then employed a confirmatory factor analysis (CFA) to ascertain whether our adapted and developed measures exhibited sound psychometric properties. Per conventional CFA procedures, all items loaded on their intended construct and all cross-loadings were constrained to zero (Anderson and Gerbing 1988). Model fit was strong (χ² = 707.789, df = 332, p < .01. χ²/df = 2.1, CFI = .95, TLI = .95, RMSEA = .070, and SRMR = .042). Furthermore, the measures met convergent validity as all loadings were above .7 (range: .78–.99), and the average variance extracted (AVE) for each construct was above .5 (range: .66–.93, Fornell and Larcker 1981). Last, heterotrait-monotrait (HTMT) scores were below a conservative threshold of .75 (range: .02–.45), suggesting discriminant validity (Voorhees et al. 2016). The full results of these analyses and the correlation matrix appear in the Appendix.

Study 2 Hypothesis Testing

For hypothesis testing, we used a custom-coded model in PROCESS to calculate the direct, indirect, and interaction effects (Hayes 2013). The model designated consumer injustice and perceived ostracism as parallel mediators and the brand response as a moderator for only the hypothesized mediated effects (to test H3b and H4b, as illustrated in Figure 2). Separate analyses were used for each of the two outcome variables. Indirect effects were estimated using 5000 bootstrap samples. Dummy variables represented the manipulations, with a one representing an uncivil reply or a brand’s response. Respondent’s pre-existing attitudes toward Amazon, shopping enjoyment, social media complaining experience, social media usage frequency, and demographics were included as covariates for the mediator and outcome variables.

Figure 2.

Study 2: The sequence of effects from an uncivil reply and the moderating effect of brand response on a victim’s engagement intentions. *** p < .001; ** p < .01; * p < .05. Notes: Numbers to the left of parentheses are estimated coefficients whereas numbers in parentheses are standard errors. Coefficients are unstandardized. For simplicity, the illustration does not show effects of control variables (but these effects are shown in the Web Appendix).

Figure 2 illustrates the results of the hypothesis testing (the full results with controls are available in the Web Appendix). The uncivil reply had a positive relationship with consumer injustice (b = 3.83, t = 14.19, p < .001) and perceived ostracism (b = 1.30, t = 4.67, p < .001). Both injustice (b = 0.44, t = 4.17, p < .001) and ostracism (b = 0.19, t = 2.23, p = .03) had a positive relationship with short-term engagement (i.e., defensive response). However, the indirect effect of the uncivil reply on short-term engagement was significant through injustice (a × b = 1.67, 95% CI = 0.83/2.53), but not ostracism (a × b = 0.20, 95% CI = −0.04/0.44). Thus, the data support H1b as injustice uniquely mediated the effect of incivility on short-term engagement. Regarding long-term engagement, neither injustice (b = 0.06, t = 0.72, p = .47) nor ostracism (b = 0.15, t = 1.62, p = .11) had a significant relationship. Instead, long-term engagement was driven solely by a subject’s previous attitude toward Amazon (b = 0.32, t = 4.52, p < .001), shopping enjoyment (b = 0.28, t = 3.57, p < .001), and social media complaint experience (b = 0.80, t = 3.99, p < .001). Thus, the data do not support H2b.

As for the moderating effect of a brand’s response, the presence of a brand response reduced the relationship between injustice and short-term engagement (b = −0.27, t = −2.77, p < .01), but not incivility with injustice (b = −0.15, t = −0.45, p = .66). Importantly, the index of moderated mediation did not contain zero (index = −1.04, 95% CI = −1.80/−0.27). These results support H3b. Furthermore, the indirect effect of incivility on short-term engagement through injustice was smaller when there was a brand response (a × b = 0.74, 95% CI = 0.18/1.32), as compared to when there was none (a × b = 1.65, 95% CI = 0.91/2.41). This effect is illustrated in Figure 3, where Figure 3a shows that a brand response reduced the indirect effect of an uncivil consumer reply on short-term engagement intentions.⁷

Figure 3.

Study 2: The indirect effect of a consumer’s uncivil reply on a victim’s short-term and long-term engagement intentions, through injustice and ostracism, as moderated by a brand’s reply.

Regarding the final moderating hypothesis, a brand’s response decreased the relationship between incivility and ostracism (b = −0.85, t = −2.17, p = .03), but not the link between ostracism and long-term engagement (b = 0.05, t = 0.32, p = .75). Given ostracism’s non-significant relationship with long-term engagement, the index of moderated mediation contained zero for the indirect effect of an uncivil reply on long-term engagement (index = −0.10, 95% CI: −0.34/−0.15). Thus, the data do not support H4b. Figure 3b reveals that incivility did not indirectly decrease long-term engagement intentions, regardless of a brand’s response.

Study 2 Discussion

Study 2’s experimental results provide some support for Study 1’s behavioral results Table 3. An uncivil interaction (versus a civil one) produced greater perceptions of consumer injustice and ostracism. Subsequently, a defensive response was more likely to spur immediate engagement. One reason for the non-significant relationship between ostracism and long-term engagement may be that the uncivil commenter did not adequately reflect a group member. Although we theorized that a brand advocate would enhance the effect of an instance of ostracism, it is possible that ostracism occurs only from brand advocates. Therefore, a follow-up experiment investigated this possibility by directly testing H5.

Table 3.

Hypotheses and their results across studies.

		Is the hypothesis supported?
		Study 1	Study 2	Study 3
H1a	Consumer incivility on a brand’s social media page exhibits a positive relationship with victims’ short-term engagement.	Yes
H2a	Consumer incivility on a brand’s social media page exhibits a negative relationship with victims’ long-term engagement.	Yes
H3a	The positive effect of consumer incivility on victims’ short-term engagement is weaker when there is a brand reply, as compared to when there is no brand reply.	Yes
H4a	The negative effect of consumer incivility on victims’ long-term engagement is weaker when there is a brand reply as compared to when there is no brand reply.	Yes
H1b	Consumer injustice mediates the relationship between consumer incivility and victims’ increased short-term engagement.		Yes
H2b	Perceived ostracism mediates the relationship between consumer incivility and victims’ long-term engagement.		No
H3b	The positive, mediated effect of consumer incivility on victims’ short-term engagement through perceived injustice is weaker when there is a brand reply as compared to when there is no brand reply.		Yes
H4b	The negative, mediated effect of consumer incivility on victims’ long-term engagement through perceived ostracism is weaker when there is a brand reply as compared to when there is no brand reply.		No
H5	The negative mediated effect of consumer incivility on victims’ long-term engagement through ostracism is stronger when the incivility is from a brand advocate as compared to a troll.			Yes

Study 3

Study 3’s design was similar to that of Study 2 in terms of context and the uncivil reply, but it featured three key differences. First, all subjects saw the firm response to see whether an ostracism effect (H2) was possible for a brand advocate commenter despite a brand response. Second, the commenter manipulation was introduced (troll vs advocate). Third, we included three marker items from Steenkamp and Maydeu-Olivares (2021) with the mediator scales (i.e., injustice and ostracism) to reduce common method variance in the mediated relationships.

To develop the troll and brand advocate manipulations, we undertook a qualitative pilot study (explained in the Web Appendix). Interviewees in the study identified attackers as troll-like if they exhibited 1) an ambiguous profile, 2) little brand expertise, 3) challenging or combative motives, and 4) schadenfreude (i.e., pleasure in insulting victims). As such, the troll condition used a default, anonymous profile picture and a fabricated profile name (“I have NoName”). The brand advocate condition used the same name and profile image from Study 2 but included a “Top Fan” badge. Facebook implemented the Top Fan badge in 2019 to recognize those users who are most active on a brand’s page. Using a method similar to that of Schaefers and Schamari (2016), we included a brief note to subjects to explain that a Top Fan designation meant a user had contributed often on a brand’s page and had knowledge about a brand.

A pre-test showed that participants thought both conditions represented customers of the brand (M _{brand advocate} = 3.8 M _troll = 3.6, F(1,36) = 0.32, p = .58), measured on a 5-point scale, “To what extent do you think the commenter was a customer of Amazon?”). However, participants perceived the brand advocate condition as a brand advocate (M = 3.3) on a single-item, 5-point scale (“To what extent do you think the commenter was a loyal customer protecting Amazon?”) but not the troll condition (M = 2.0, F(1,36) = 10.09, p < .01).⁸ All measures used in Study 3 were the same as those used in Study 2.

Study 3 Subject Recruitment and Measures

The recruitment procedures of Study 3 used those of Study 2 to enlist 325 MTurkers, but 22 subjects (6.8%) had not used Facebook, and 29 more (9.6%) failed an attention check. The final sample was 274. Sufficient measurement properties were achieved for model fit (χ² = 729.002, df = 332, p < .01. χ²/df = 2.2, CFI = .96, TLI = .95, RMSEA = .066, SRMR = .060), loadings (.80–.98), AVE (.64–.93), and HTMT scores (.05–.51). The results of these analyses and a correlation matrix are displayed in the Appendix.

Study 3 Results

For hypothesis testing, we used a custom-coded PROCESS model to calculate the necessary direct, indirect, and interaction effects (Hayes 2013). The model was the same as that of Study 2, except the w matrix was coded to test H5a and H5b simultaneously. Figure 4 shows the results (see the Web Appendix for full results with controls). As expected, incivility exhibited a positive relationship with injustice (b = 3.46 t = 22.02, p < .001) but not ostracism (b = 0.38, t = 1.47, p = 0.14). Here, incivility’s relationship with ostracism was smaller than in Study 2, likely because a brand’s response moderated the relationship in Study 2, and all subjects saw a brand response in Study 3. Similarly, consumer injustice had no relationship with a defensive response (b = 0.02, t = 0.24, p = .81), meaning the uncivil reply’s indirect effect on a defensive response through injustice was non-significant (a × b = 0.07, 95% CI: −0.48/0.66). These results accord with those of Study 2, given the presence of a brand’s response.

Figure 4.

Study 3: The sequence of effects from an uncivil reply and the moderating effect of type of attacker on a victim’s engagement intentions. *** p < .001; ** p < .01; * p < .05; + p < .1. Notes: Numbers to the left of parentheses are estimated coefficients whereas numbers in parentheses are standard errors. Coefficients are unstandardized. For simplicity, the illustration does not show effects of control variables (but these effects are shown in the Web Appendix).

Importantly, the type of commenter moderated the relationship between incivility and ostracism (b = 0.68, t = 1.86, p = .06) and between ostracism and long-term engagement (b = −0.29, t = −2.62, p < .01). Simple slopes analyses revealed that when a commenter was a brand advocate, ostracism exhibited a significant negative relationship with engagement (b = −0.21, t = −2.50, p = .01), and incivility had a strong, positive relationship with ostracism (b = 1.06, t = 4.02, p < .001). The indirect effect was negative, and the confidence interval did not contain zero when an uncivil commenter was an advocate (a × b = −0.22, 95% CI = −0.52/−0.02), in contrast to when a commenter was troll-like (a × b = 0.03, 95% CI = −0.04/0.11). Further, the index of moderated mediation was negative (index = −.25), and the confidence interval did not contain zero (95% CI = −0.54/−0.04). The indirect effect is illustrated in Figure 5, which shows that incivility from a brand advocate decreased long-term engagement more than incivility from a troll.⁹ Thus, the data support H5. We now holistically discuss the results of all the studies.

Figure 5.

Study 3: The indirect effect of a consumer’s uncivil reply on a victim’s long-term engagement intentions, through perceived ostracism, as moderated by attacker type (troll versus brand advocate). Notes: Because the uncivil reply had a significant direct effect on long-term engagement intentions and a significant indirect effect through consumer injustice, the graphed values were calculated using all direct and indirect effects.

General Discussion

Our research, supported by qualitative data, Facebook data, and experiments, suggests that C2C incivility is not benign. Victims of C2C incivility exhibit confrontation and avoidance coping responses, increasing short-term engagement and decreasing long-term engagement. Two perceptions, consumer injustice and ostracism, mediate these coping responses. A brand response lessens the coping responses, but an uncivil reply from a brand advocate fosters ostracism even when a brand responds. The following are implications derived from our proposed theory.

Theoretical Implications

Our work offers three contributions to theory, with the first being the explanation and evidence of how incivility affects consumer engagement. Harmeling et al. (2017) noted that engagement may be problematic because brands give control to consumers, whose actions may be unwanted by the brand and engagement participants. Qualitative research has even documented how positive engagement from consumers can impair the engagement of others (Bowden et al. 2017; Hollebeek and Chen 2014). In accord with this view, the current research reveals that even when incivility is not anti-brand, it can diminish victims’ long-term engagement. One consumers’ benefit-inducing engagement can be another consumer’s engagement roadblock. Social media naturally fosters incivility and as the aforementioned Starbucks case suggests, research is sorely needed to address this phenomenon (Guzman 2021; Munn 2020).

The second contribution is the mechanisms that explain incivility’s dual effects on engagement. The present research documents how online incivility provokes coping responses in consumers which results in perceptions of injustice and ostracism. Researchers have applied coping theory to other consumer contexts, such as customers’ reactions to service failure (Strizhakova, Tsarenko, and Ruth 2012). However, the limited research on consumer incivility draws primarily on justice theory and equates the outcomes of incivility victims and witnesses as first and third parties evaluate justice (e.g., Bacile et al. 2018; Okan and Elmadag 2020). Coping theory suggests that victims of incivility are in unique roles as they cope with direct threats to their self-esteem and competence. Future research can likely apply more of the coping framework by examining how other coping strategies (e.g., social support) affect outcomes of interest (e.g., negative word-of-mouth). Further, we expect injustice and ostracism to be useful in evaluating other incivility interventions and consumer engagement outcomes.

Our third contribution is theorizing the unique role of brand advocates. Research on consumer incivility has not examined many differences across incivility perpetrators and research on engagement is just now conceptualizing how brand advocates can create negative engagement (Hollebeek, Kumar, and Srivastava 2020). The current study suggests the nature of the attacker is critical to understanding the long-term effects of incivility. We did not study consumer incivility toward employees, but our theory suggests brand advocates are pertinent to that domain. In fact, our theory suggests any indicators that position an uncivil attacker as representative of a group will magnify the effect of their incivility. Our pilot study identified four characteristics people used to determine whether someone was a troll or a brand advocate: profile information, brand knowledge, combative comments, and schadenfreude. These characteristics may be unique to our online context, suggesting that future research may usefully aim to identify those characteristics in other contexts that determine the insider status of uncivil perpetrators.

Managerial Implications

This work offers an opportunity for academic research to lead industry practices. Uncivil consumer comments are overlooked by managers, possibly due to new media marketing frameworks not recognizing all the harmful outcomes (e.g., Hennig-Thurau et al. 2010). The prevailing belief—both in practice and theory—is that more consumer engagement on social media is positive for consumers and brands. Our results refute that view, showing how an increase in short-term engagement may result from consumers defending themselves against other consumers acting uncivilly. The results subsequently demonstrate how long-term engagement diminishes for victims of incivility. Common web metrics such as the number of shares, followers, or page visits overlook this phenomenon.

Managers now have options if they are aware of this perspective. Based on our results, we now know victims use a coping strategy to reduce future engagement after the incivility leads to ostracism. As a first step in avoiding adverse outcomes, managers (or their technology) can flag uncivil threads and the self-replies of a victim. A brand can then communicate with a victim who may be defensively replying to their attacker. At the very least, our research shows that a brand response that ignores incivility can still mitigate some of the outcomes from the incivility. However, the persistence of ostracism from a brand advocate suggests that companies may need to further develop competencies to handle consumer interactions on social media.

For service managers, our work provides a pivotal link between C2C interactions in traditional and digital service environments. Practitioners know the importance of managing customer interactions in traditional service environments (e.g., restaurants). Our findings imply that such strategies of service management are relevant to digital service environments. For example, Pranter and Martin (1991, p. 46) suggest that service workers takes on “legislator” (rule creator), “teacher” (rule communicator), or “police officer” (rule enforcer) roles to manage C2C interactions. A case in point of a legislator role is Ford’s “Participation Guidelines” on their Facebook page, which includes how it will address posts with “anything obscene, vulgar, sexually explicit, illegal, harmful, threatening, abusive, harassing, humiliating, defamatory, libelous, invasive of someone else’s privacy, or objectionable.” Additional management roles such as communicating and enforcing stated policies involving C2C incivility may be beneficial.

Limitations

Study design choices carry limitations that affect the research. One limitation is the focus on Facebook, the social media market leader. Facebook’s users tend to be older than users on other social platforms (e.g., Pinterest and Snapchat) and may react differently to incivility than younger populations. Relatedly, rudeness is subjective, and certain people and cultures are more acclimated to uncivil interactions than are others. As such, individual and cultural characteristics may affect reactions to incivility. Another limitation is our variable operationalizations. For example, our focus on engagement as non-purchase behavior ignores other possible aspects of engagement, such as absorption and attention. In addition, we did not link our outcomes to purchase behavior; therefore, the effect of incivility on the totality of engagement is unknown. Similarly, we did not examine whether incivility affects the consumer-brand relationship beyond engagement, such as through brand relationship quality. Although regulatory engagement theory (Higgins and Scholer 2009) suggests that such a situation is likely, this possibility deserves further examination. Finally, incivility can have positive consequences in employee-to-consumer contexts (Ward and Dahl 2014). Thus, incivility may have benefits in C2C contexts.

Supplemental Material

Supplemental Material - How Online Incivility Affects Consumer Engagement Behavior on Brands’ Social Media

Supplemental Material for How Online Incivility Affects Consumer Engagement Behavior on Brands’ Social Media by Jeremy S. Wolter, Todd J. Bacile, and Pei Xu in Journal of Service Research

Supplemental Material

Supplemental Material - How Online Incivility Affects Consumer Engagement Behavior on Brands’ Social Media

Supplemental Material for How Online Incivility Affects Consumer Engagement Behavior on Brands’ Social Media by Jeremy S. Wolter, Todd J. Bacile, and Pei Xu in Journal of Service Research

Footnotes

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 iDs

Jeremy S. Wolter

Todd J. Bacile

Supplemental Material

Supplemental material for this article is available online.

Notes

Author Biographies

Jeremy S. Wolter (PhD, 2013, Florida State University) is an Associate Professor of Marketing at the Harbert College of Business, Auburn University. His research focuses on consumer-based marketing strategy through service strategy, customer loyalty, and customer identity.

Todd J. Bacile (PhD, 2013, Florida State University) is an Associate Professor of Marketing in the College of Business at Loyola University New Orleans. His research interests include digital marketing and services marketing concepts. He also serves as an expert witness in legal cases within the domain of digital marketing.

Pei Xu (PhD, 2014, University of Kentucky) is an Associate professor of Business Analytics in the Harbert College of Business at Auburn University. Her research focus is on understanding the economic and social impact of emerging information technology with a variety of quantitative methods.

References

Al-Natour

Sameh

Turetken

Ozgut

(2020), “A Comparative Assessment of Sentiment Analysis and Star Ratings for Consumer Reviews,” International Journal of Information Management, 54 (May), 102132.

Anderson

Ashley A.

Brossard

Dominique

Scheufele

Dietram A.

Xenos

Michael A.

Peter

Ladwig

(2014), “The “Nasty Effect:” Online Incivility and Risk Perceptions of Emerging Technologies*,” Journal of Computer-Mediated Communication, 19 (3), 373-387.

Anderson

James C.

Gerbing

David W.

(1988), “Structural Equation Modeling in Practice: A Review and Recommended Two-Step Approach,” Psychological Bulletin, 103 (3), 411-423.

Andersson

Lynne M.

Pearson

Christine M.

(1999), “Tit for Tat? The Spiraling Effect of Incivility in the Workplace,” Academy of Management Review, 24 (3), 452-471.

Antoci

Angelo

Delfino

Alexia

Paglieri

Fabio

Panebianco

Fabrizio

Sabatini

Fabio

(2016), “Civility Vs. Incivility in Online Social Interactions: An Evolutionary Approach,” PLoS ONE, 11 (11), e0164286.

Bacile

Todd J

(2020), “Digital Customer Service and Customer-to-Customer Interactions: Investigating the Effect of Online Incivility on Customer Perceived Service Climate,” Journal of Service Management, 31 (3), 441-464.

Bacile

Todd J.

Krallman

Alexandra

Wolter

Jeremy S.

Beachum

Nicole Dilg

(2020), “The Value Disruption of Uncivil Other-Customers During Online Service Recovery,” Journal of Services Marketing, 34 (4), 483-498.

Bacile

Todd J.

Wolter

Jeremy S.

Allen

Alexis M.

Pei

(2018), “The Effects of Online Incivility and Consumer-to-Consumer Interactional Justice on Complainants, Observers, and Service Providers During Social Media Service Recovery,” Journal of Interactive Marketing, 44, 60-81.

Baer

Jay

(2016), Hug your Haters: How to Embrace Complaints and Keep your Customers. New York: Portfolio/Penguin.

10.

Bowden

Jana Lay-Hwa

Conduit

Jodie

Hollebeek

Linda D.

Luoma-aho

Vilma

Solem

Birgit Apenes

(2017), "Engagement Valence Duality and Spillover Effects in Online Brand Communities," Journal of Service Theory and Practice, 27 (4), 877-897.

11.

Carver

Charles S.

Scheier

Michael F.

Weintraub

Jagdish K.

(1989), “Assessing Coping Strategies: A Theoretically Based Approach,” Journal of Personality and Social Psychology, 56 (2), 267-283.

12.

Casler

Krista

Bickel

Lydia

Hackett

Elizabeth

(2013), “Separate but Equal? A Comparison of Participants and Data Gathered Via Amazon’s Mturk, Social Media, and Face-to-Face Behavioral Testing,” Computers in Human Behavior, 29 (6), 2156-2160.

13.

Caza

Brianna Barker

Cortina

Lilia M

(2007), “From Insult to Injury: Explaining the Impact of Incivility,” Basic and Applied Social Psychology, 29 (4), 335-350.

14.

Cheung

Janelle H.

Burns

Deanna K.

Sinclair

Robert R.

Sliter

Michael

(2017), “Amazon Mechanical Turk in Organizational Psychology: An Evaluation and Practical Recommendations,” Journal of Business and Psychology, 32 (4), 347-361.

15.

Churchill

Gilbert A.

Jr. (1979), “A Paradigm for Developing Better Measures of Marketing Constructs,” Journal of Marketing Research, 16 (1), 64-73.

16.

Coe

Kevin

Kenski

Kate

Rains

Stephen A.

(2014), “Online and Uncivil? Patterns and Determinants of Incivility in Newspaper Website Comments,” Journal of Communication, 64 (4), 658-679.

17.

Coles

Bryn Alexander

West

Melanie

(2016), “Trolling the Trolls: Online Forum Users Constructions of the Nature and Properties of Trolling,” Computers in Human Behavior, 60, 233-244.

18.

Cortina

Lilia M.

Kabat-Farr

Dana

Leskinen

Emily A.

Huerta

Marisela

Magley

Vicki J.

(2011), “Selective Incivility as Modern Discrimination in Organizations: Evidence and Impact,” Journal of Management, 39 (6), 1579-1605.

19.

Cortina

Lilia M.

Magley

Vicki J.

(2009), “Patterns and Profiles of Response to Incivility in the Workplace,” Journal of Occupational Health Psychology, 14 (3), 272-288.

20.

Cortina

Lilia M.

Magley

Vicki J.

Williams

Jill Hunter

Langhout

Regina Day

(2001), “Incivility in the Workplace: Incidence and Impact,” Journal of Occupational Health Psychology, 6 (1), 64-80.

21.

Dahl

Ethan

Niedbala

Elizabeth M.

Hohman

Zachary P.

(2018), “Loving the Group That Denies You First: Social Identity Effects of Ostracism before Inclusion,” Personality and Social Psychology Bulletin, 45 (2), 284-299.

22.

Dawson

Scott

Bloch

Peter H.

Ridgway

Nancy M.

(1990), “Shopping Motives, Emotional States, and Retail Outcomes,” Journal of Retailing, 66 (4), 408-427.

23.

De Smedt

Tom

Daelemans

Walter

(2012), “Pattern for Python,” Journal of Machine Learning Research, 13, 2063–2067.

24.

Dhaoui

Chedia

Webster

Cynthia M

. (2021), “Brand and Consumer Engagement Behaviors on Facebook Brand Pages: Let's Have a (Positive) Conversation,” International Journal of Research in Marketing, 38 (1), 155-175.

25.

Diem Khac Xuan

Rahman

Kaleel

Robinson

Linda J

(2020), "Determinants of Negative Customer Engagement Behaviours," Journal of Services Marketing, 34 (2), 117-135.

26.

Fehrer

Julia A.

Woratschek

Herbert

Germelmann

Claas Christian

Brodie

Roderick J.

(2018), "Dynamics and Drivers of Customer Engagement: Within the Dyad and Beyond," Journal of Service Management, 29 (3), 443-467.

27.

Fiegerman

Seth

(2017), “Facebook Is Closing in on 2 Billion Users,” (accessed November, 2018), [available at https://money.cnn.com/2017/02/01/technology/facebook-earnings/index.html].

28.

Folkman

Susan

(1984), “Personal Control and Stress and Coping Processes: A Theoretical Analysis,” Journal of Personality and Social Psychology, 46 (4), 839-852.

29.

Folkman

Susan

Lazarus

Richard S.

Dunkel-Schetter

Christine

Anita

DeLongis

Gruen

Rand J.

(1986), “Dynamics of a Stressful Encounter: Cognitive Appraisal, Coping, and Encounter Outcomes,” Journal of Personality and Social Psychology, 50 (5), 992-1003.

30.

Fornell

Claes

Larcker

David F.

(1981), “Evaluating Structural Equation Models with Unobservable Variables and Measurement Error,” Journal of Marketing Research, 18 (1), 39-50.

31.

Gelbrich

Katja

(2010), “Anger, Frustration, and Helplessness after Service Failure: Coping Strategies and Effective Informational Support,” Journal of the Academy of Marketing Science, 38 (5), 567-585.

32.

Gómez

Ángel

Morales

J. Francisco

Hart

Sonia

Vázquez

Alexandra

Swann

William B.

(2011), “Rejected and Excluded Forevermore, but Even More Devoted: Irrevocable Ostracism Intensifies Loyalty to the Group among Identity-Fused Persons,” Personality and Social Psychology Bulletin, 37 (12), 1574-1586.

33.

Guzman

Joseph

(2021), “Starbucks May Quit Facebook Because of Hateful Comments from Users, Report Says,” (accessed May, 2021), [available at https://thehill.com/changing-america/respect/552659-starbucks-may-quit-facebook-because-of-hateful-comments-from-users].

34.

Hansen

Nele

Kupfer

Ann-Kristin

Hennig-Thurau

Thorsten

(2018), "Brand Crises in the Digital Age: The Short- and Long-Term Effects of Social Media Firestorms on Consumers and Brands," International Journal of Research in Marketing, 35 (4), 557-574.

35.

Harmeling

Colleen M.

Moffett

Jordan W.

Arnold

Mark J.

Carlson

Brad D.

(2017), “Toward a Theory Of Customer Engagement Marketing,” Journal of the Academy of Marketing Science, 45 (4), 312-335.

36.

Hayes

Andrew F.

(2013), Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach. New York, NY: Guiford Press.

37.

Henkel

Alexander P.

Boegershausen

Johannes

Rafaeli

Anat

Lemmink

Jos

(2017), “The Social Dimension of Service Interactions: Observer Reactions to Customer Incivility,” Journal of Service Research, 20 (2), 120-134.

38.

Hennig-Thurau

Thorsten

Malthouse

Edward C.

Friege

Christian

Gensler

Sonja

Lobschat

Lara

Rangaswamy

Arvind

Skiera

Bernd

(2010), “The Impact of New Media on Customer Relationships,” Journal of Service Research, 13 (3), 311-330.

39.

Hershcovis

M. Sandy

(2011), ““Incivility, Social Undermining, Bullying…Oh My!”: A Call to Reconcile Constructs within Workplace Aggression Research,” Journal of Organizational Behavior, 32 (3), 499-519.

40.

Hershcovis

M. Sandy

Ann-Frances

Cameron

Loie

Gervais

Jennifer

Bozeman

(2018), “The Effects of Confrontation and Avoidance Coping in Response to Workplace Incivility,” Journal of Occupational Health Psychology, 23 (2), 163-174.

41.

Higgins

E. Tory

Scholer

Abigail A.

(2009), "Engaging the Consumer: The Science and Art of the Value Creation Process," Journal of Consumer Psychology, 19, 100-114.

42.

Hmielowski

Jay D.

Hutchens

Myiah J.

Cicchirillo

Vincent J.

(2014), “Living in an Age of Online Incivility: Examining the Conditional Indirect Effects of Online Discussion on Political Flaming,” Information, Communication & Society, 17 (10), 1196-1211.

43.

Hollebeek

Linda D.

Chen

Tom

(2014), "Exploring Positively- Versus Negatively-Valenced Brand Engagement: A Conceptual Model," Journal of Product & Brand Management, 23 (1), 62-74.

44.

Hollebeek

Linda D.

Srivastava

Rajendra K.

Chen

Tom

(2019), "S-D Logic–Informed Customer Engagement: Integrative Framework, Revised Fundamental Propositions, and Application to CRM," Journal of the Academy of Marketing Science, 47 (1), 161-185.

45.

Hollebeek

Linda D.

Kumar

Srivastava

Rajendra K.

(2020), "From Customer-, to Actor-, to Stakeholder Engagement: Taking Stock, Conceptualization, and Future Directions," Journal of Service Research.

46.

Howard

Matt C.

Cogswell

Joshua E.

Smith

Mickey B.

(2020), “The Antecedents and Outcomes of Workplace Ostracism: A Meta-Analysis,” Journal of Applied Psychology, 105 (6), 577-596.

47.

Hudson

Simon

Huang

Roth

Martin S.

Madden

Thomas J.

(2016), “The Influence of Social Media Interactions on Consumer–Brand Relationships: A Three-Country Study of Brand Perceptions and Marketing Behaviors,” International Journal of Research in Marketing, 33 (1), 27-41.

48.

Hulland

John

Miller

Jeff

(2018), “Keep on Turkin’?,” Journal of the Academy of Marketing Science, 46 (5), 789-794.

49.

Ilhan

Behice Ece

Kübler

Raoul V.

Pauwels

Koen H.

(2018), “Battle of the Brand Fans: Impact of Brand Attack and Defense on Social Media,” Journal of Interactive Marketing, 43, 33-51.

50.

Kim

Yonghwan

Kim

Youngju

(2019), “Incivility on Facebook and Political Polarization: The Mediating Role of Seeking Further Comments and Negative Emotion,” Computers in Human Behavior, 99, 219-227.

51.

Kozinets

Robert V.

(2002), “The Field Behind the Screen: Using Netnography for Marketing Research in Online Communities,” Journal of Marketing Research, 39 (1), 61-72.

52.

Kumar

Aksoy

Lerzan

Bas

Donkers

Venkatesan

Rajkumar

Wiesel

Thorsten

Tillmans

Sebastian

(2010), "Undervalued or Overvalued Customers: Capturing Total Customer Engagement Value," Journal of Service Research, 13 (3), 297-310.

53.

Lazarus

Richard S

(1999), Stress and Emotion: A New Synthesis. New York: Springer.

54.

Lazarus Richard

Folkman

Susan

(1984). Stress, Appraisal, and Coping. New York, NY: Springer Publishing Company, Inc.

55.

Leary

Mark R

(1990), “Responses to Social Exclusion: Social Anxiety, Jealousy, Loneliness, Depression, and Low Self-Esteem,” Journal of Social and Clinical Psychology, 9 (2), 221-229.

56.

Lee

Jaehoon

Shrum

L. J

. (2012), “Conspicuous Consumption Versus Charitable Behavior in Response to Social Exclusion: A Differential Needs Explanation,” Journal of Consumer Research, 39 (3), 530-544.

57.

Litman

Leib

Robinson

Jonathan

Rosenzweig

Cheskie

(2015), “The Relationship between Motivation, Monetary Compensation, and Data Quality among Us- and India-Based Workers on Mechanical Turk,” Behavior Research Methods, 47 (2), 519-528.

58.

Loria

Steven

Pete

Keen

Matthew

Honnibal

Roman

Yankovsky

David

Karesh

Evan

Dempsey

Wesley

Childs

Jeff

Schnurr

Adel

Qalieh

Lage

Ragnarsson

Jonathan

Coe

Calvo

Adrián López

Nitish

Kulshrestha

Jhon

Eslava

James

Harden Tyler

Kolb

Jeff

Ong

Daniel

Moschella

Jamie

Korolev

Roman

Ram

Rachum

(2020), “Textblob,” (accessed September, 2020), [available at https://textblob.readthedocs.io/en/dev/authors.html#].

59.

Marge

Matthew

Banerjee

Satanjeev

Rudnicky

Alexander

(2010), “Using the Amazon Mechanical Turk to Transcribe and Annotate Meeting Speech for Extractive Summarization,” In Proceedings of the NAACL HLT 2010 workshop on creating speech and language data with Amazon’s Mechanical Turk, 99-107.

60.

Mturk.com (2017), “Tutorial: Reconciling Worker Responses,” (accessed May, 2021), [available at https://blog.mturk.com/tutorial-reconciling-worker-responses-280c96f1a696].

61.

Munn

Luke

(2020), "Angry by Design: Toxic Communication and Technical Architectures," Humanities and Social Sciences Communications, 7 (1), 53.

62.

Naumann

Kay

Jana

Bowden

Gabbott

Mark

(2020), "Expanding Customer Engagement: The Role of Negative Engagement, Dual Valences and Contexts," European Journal of Marketing, 54 (7), 1469-1499.

63.

Netemeyer

Richard G.

Bearden

William O.

Sharma

Subhash

(2003), Scaling Procedures. Thousand Oaks: Sage Publications.

64.

Okan

Mehmet

Elmadag

Ayse Banu

(2020), “Witnessing Verbal Aggression: Role of Customers’ Self-Conscious Emotions,” Journal of Services Marketing, 34 (2), 253-268.

65.

Oppenheimer

Daniel M.

Meyvis

Tom

Davidenko

Nicolas

(2009), “Instructional Manipulation Checks: Detecting Satisficing to Increase Statistical Power,” Journal of Experimental Social Psychology, 45 (4), 867-872.

66.

Peer

Eyal

Vosgerau

Joachim

Acquisti

Alessandro

(2014), “Reputation as a Sufficient Condition for Data Quality on Amazon Mechanical Turk,” Behavior Research Methods, 46 (4), 1023-1031.

67.

Porath

Christine

MacInnis

Deborah

Folkes

Valerie S.

(2011), “It's Unfair: Why Customers Who Merely Observe an Uncivil Employee Abandon the Company,” Journal of Service Research, 14 (1), 302-317.

68.

Pranter

Charles A.

Martin

Charles L.

(1991), “Compatibility Management: Roles in Service Performers,” Journal of Services Marketing, 5 (2), 43-53.

69.

Rahman

Mahabubur

Faroque

Anisur R.

Sakka

Georgia

Ahmed

Zafar U.

(2022), "The Impact of Negative Customer Engagement on Market-Based Assets and Financial Performance," Journal of Business Research, 138, 422-435.

70.

Rupp

Deborah E.

Silke

McCance A.

Spencer

Sharmin

Sonntag

Karlheinz

(2008), “Customer (in)Justice and Emotional Labor: The Role of Perspective Taking, Anger, and Emotional Regulation,” Journal of Management, 34 (5), 903-924.

71.

Rupp

Deborah E.

Spencer

Sharmin

(2006), “When Customers Lash Out: The Effects of Customer Interactional Injustice on Emotional Labor and the Mediating Role of Discrete Emotions,” Journal of Applied Psychology, 91 (4), 971-978.

72.

Sacco

Donald F.

Bernstein

Michael J.

Young

Steven G.

Hugenberg

Kurt

(2014), “Reactions to Social Inclusion and Ostracism as a Function of Perceived In-Group Similarity,” Group Dynamics: Theory, Research, and Practice, 18 (2), 129-137.

73.

Schaefers

Tobias

Schamari

Julia

(2016), “Service Recovery Via Social Media: The Social Influence Effects of Virtual Presence,” Journal of Service Research, 19 (2), 192-208.

74.

Schneider

Frank M.

Zwillich

Britta

Bindl

Melanie J.

Hopp

Frederic R.

Reich

Sabine

Vorderer

Peter

(2017), “Social Media Ostracism: The Effects of Being Excluded Online,” Computers in Human Behavior, 73, 385-393.

75.

Simon

Françoise

Tossan

Vesselina

(2018), “Does Brand-Consumer Social Sharing Matter? A Relational Framework of Consumer Engagement to Brand-Hosted Social Media,” Journal of Business Research, 85, 175-184.

76.

Skarlicki

Daniel P.

Folger

Robert

(1997), “Retaliation in the Workplace: The Roles of Distributive, Procedural, and Interactional Justice,” Journal of Applied Psychology, 82 (3), 434-443.

77.

Skarlicki

Daniel P.

van Jaarsveld

Danielle D.

Walker

David D.

(2008), “Getting Even for Customer Mistreatment: The Role of Moral Identity in the Relationship between Customer Interpersonal Injustice and Employee Sabotage,” Journal of Applied Psychology, 93 (6), 1335-1347.

78.

Steenkamp

Jan-Benedict E. M.

Maydeu-Olivares

Alberto

(2021), “An Updated Paradigm for Evaluating Measurement Invariance Incorporating Common Method Variance and Its Assessment,” Journal of the Academy of Marketing Science, 49 (1), 5-29.

79.

Stöckli

David R.

Khobzi

Hamid

(2021), “Recommendation Systems and Convergence of Online Reviews: The Type of Product Network Matters!,” Decision Support Systems, 142, 113475.

80.

Strizhakova

Yuliya

Tsarenko

Yelena

Ruth

Julie A.

(2012), ““I’m Mad and I Can’t Get That Service Failure Off My Mind”: Coping and Rumination as Mediators of Anger Effects on Consumer Intentions,” Journal of Service Research, 15 (4), 414-429.

81.

Suler

John R.

(2016), Psychology of the Digital Age: Humans Become Electric. Cambridge University Press, New York.

82.

Sydnor

Emily

(2019), Disrespectful Democracy: The Psychology of Political Incivility. New York: Columbia University Press.

83.

Tax

Stephen S.

Brown

Stephen W.

Chandrashekaran

Murali

(1998), “Customer Evaluations of Service Complaint Experiences: Implications for Relationship Marketing,” Journal of Marketing, 62 (2), 60-76.

84.

van Doorn

Jenny

Lemon

Katherine N.

Mittal

Vikas

Nass

Stephan

Pick

Doreén

Pirner

Peter

Verhoef

Peter C.

(2010), “Customer Engagement Behavior: Theoretical Foundations and Research Directions,” Journal of Service Research, 13 (3), 253-266.

85.

Viswanathan

Vijay

Hollebeek

Linda D.

Malthouse

Edward C.

Maslowska

Ewa

Kim

Su Jung

Xie

Wei

(2017), "The Dynamics of Consumer Engagement with Mobile Technologies," Service Science, 9 (1), 36-49.

86.

Voorhees

Clay M.

Brady

Michael K.

Calantone

Roger

Ramirez

Edward

(2016), “Discriminant Validity Testing in Marketing: An Analysis, Causes for Concern, and Proposed Remedies,” Journal of the Academy of Marketing Science, 44 (1), 119-134.

87.

Ward

Morgan K.

Dahl

Darren W.

(2014), “Should the Devil Sell Prada? Retail Rejection Increases Aspiring Consumers' Desire for the Brand,” Journal of Consumer Research, 41 (3), 590-609.

88.

Williams

Kipling D.

(2007), “Ostracism,” Annual Review of Psychology , 58 (1), 425-452.

89.

Wittenbaum

Gwen M.

Shulman

Hillary C.

Braz

Mary E.

(2010), “Social Ostracism in Task Groups: The Effects of Group Composition,” Small Group Research, 41 (3), 330-353.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.43 MB

0.06 MB