Investigating consumers’ motives for consumer brand-cyberbullying on social media

Abstract

In this article we offer the first survey-based study on the motivations that spur consumers to bully others about the brands they support on social media, a phenomenon we term “Consumer Brand-Cyberbullying” (CBC). Analyzing data from 1,203 participants of online brand communities, we find that consumers who seek to be popular and attractive are more likely to engage in CBC, while those who seek to affiliate with close others and help the community are less likely to do so. Consumers who identify with and are loyal to a particular brand are more likely to engage in CBC. Taken together, our study moves us toward a systematic analysis of the relationship between brands and cyberbullying on social media.

Keywords:

• Brand identification
• cyberbullying
• Facebook
• online communities
• online trolling
• social media marketing

Introduction

The following comments appeared on Nike’s official Facebook brand page below a post by the company about a new video commercial¹:

Table

Babs Dowell	NIKE SUPPORTS CHEATING – JUST AS LONG AS IT MAKES THEM MONEY
Andrew Cox	Well, Babs, as long as you don’t make any money at all, be sure to shut your trap, my dear.
Babs Dowell	Andrew Guywithopinion This is a reflection of your ignorance, nothing else. What a nasty piece of work you are. What country are you from? I am sure your country people are not all as ignorant as you
April Low	Andrew Guywithopinion You are extremely rude-not a gentleman that is for sure.
Denitsa Kurt	Well … you did make a bold accusation. Some people respond badly to that. At least be respectful and say the name of the person you are boldly accusing.
Tom Hope	Nike is an industry and reflects capitalism. Now please take your stupid moral issues and go somewhere else where people actually care, nobody asked for your incredibly acute revelation.
Babs Dowell	Slightly more than you-a rude ignorant specimen-male chauvinist pig-point made
April Low	Tom Hope apart from bad English, your puerile comments are a reflection of how stupid you are-look pretty weird too…
Babs Dowell	Tom Hope Oh just found out where you work—will be informing them of your harassment here.

The above exchange illustrates an increasingly frequent phenomenon: individuals offending others by making jokes or rude remarks about them. While terminology and definitions vary (see Slonje, Smith, and Frisén 2013), and disciplines differ in their conceptualization of such behaviors, we follow the literature in Psychology and classify them as “cyberbullying” (Ybarra and Mitchell 2007)—posting of comments, information, and pictures online for others to see with the intent to embarrass or offend.

There are two principal reasons why cyberbullying is a problem for contemporary societies. First, it causes harm. Being the victim of as well as merely a witness to cyberbullying reduces one’s life satisfaction, emotional security, and performance in daily tasks (Rodkin, Espelage, and Hanish 2015). A recent meta-analysis by Kowalski et al. (2014) shows that long-term outcomes for victims may be very severe, including a higher likelihood of depression, anxiety, and drug abuse. Second, in line with the ever widening reach of the Internet and social media, increasing numbers of individuals are exposed to cyberbullying. Recent market research shows, for instance, that 66% of social-media users witness cyberbullying on a regular basis (PEW 2017). Unlike traditional (i.e., offline) bullying, victims of cyberbullying cannot physically remove themselves since mobile devices continually notify and reiterate social-media content (Tanrikulu, Kinay, and Aricak 2015). Moreover, cyberbullying on social media usually lasts longer since materials once posted often are permanently there and the ease of replication makes control of their circulation difficult (Runions and Bak 2015). Finally, cyberbullying tends to be more severe because perpetrators feel less inhibited in computer-mediated communication (Suler 2004). Consequently, a growing number of researchers are investigating why it occurs and what can be done about it (for a review, see Slonje, Smith, and Frisén 2013).

However, while extant work in Psychology, Information Science, and Digital Media Studies concentrates on cyberbullying that occurs in relation to individuals’ race, gender, social norms, political opinions, physical attributes, or personality dispositions (Herring et al. 2002; Moon, Weick, and Uskul 2018; Rodkin, Espelage, and Hanish 2015), researchers in these fields have so far overlooked the possibility that it can also take place in relation to consumer brands. Likewise, the Digital Marketing and Branding literature has largely focused on the positive aspects of online brand communities (e.g., Schau, Muñiz, and Arnould 2009), and theorization on why consumers engage in cyberbullying because of the brands they support has so far been limited to anecdotal suggestions (Breitsohl, Roschk, and Feyertag 2018). In effect, there is a lack of research on cyberbullying in a consumer-brand context in general, and specifically regarding consumers’ underlying motives. Following Breitsohl, Roschk, and Feyertag (2018), we hereafter use the term “Consumer Brand-Cyberbullying” (CBC) to refer to cyberbullying that results from consumers’ bonds with brands.

Two additional factors specific to the consumers’ investment in their preferred brands add extra weight to the importance of studying of CBC. First, to be cyberbullied in relation to one’s brands can be just as damaging to an individual’s well-being as cyberbullying in other identity-related contexts, such as one’s gender or physical attributes. Several studies in Marketing show that consumers use brands as a means to express who they are (Isaksen and Roper 2016; Underwood, Bond, and Baer 2001), often to the extent that a brand’s values and performance define their own values and self-worth (Ferraro, Escalas, and Bettman 2011). Second, CBC appears to be unregulated even though it affects millions of social-media users. Although increasing numbers of consumers join online communities built by and around brands (e.g., Nike’s Facebook brand page), most brands’ corporate owners largely choose not to intervene when aggressive interactions occur (Dineva, Breitsohl, and Garrod 2017). Given that corporate online brand communities such as the one centered on Facebook brand pages have millions of daily visitors, CBC may present a particularly damaging form of cyberbullying compared to that which occurs in smaller, less public online communities.

In this article we draw on multidisciplinary sources to expand the limited knowledge on the relationship between consumers’ bonds with brands and cyberbullying. While extant research does not engage directly with participants of online brand communities and relies on anecdotal and observational suggestions (Breitsohl, Roschk, and Feyertag 2018), we used an online survey with responses from 1,203 Facebook brand-page users, and structural equation modeling (SEM) to test two main theoretical propositions. First, following Life Aspirations Theory (Grouzet et al. 2005), we hypothesize that consumers’ general aspirations in life—those that are rewarding in themselves (intrinsic) and those that require others’ recognition (extrinsic)—capture underlying reasons for how likely they are to bully others about the brands they support. Second, following Consumer Identification Theory (Bhattacharya and Sen 2003), we hypothesize that consumers’ bond with a particular brand—their purchasing loyalty and brand identification—has a bearing on CBC motivation. In other words, we propose that a consumer who cyberbullies others about brands is driven by general, psychological motives as well as by motives related to his/her role as a supporter of a brand.

In what follows, we provide a brief overview of the cyberbullying literature, followed by discussions on the theorization that gave rise to our hypotheses, our methodology, and findings. We conclude with discussions on our contributions to theory and practice, limitations of our study, and suggestions for future research.

Cyberbullying research across disciplines

Cyberbullying has been studied across several research disciplines, including Psychology (Kowalski et al. 2014), Information Studies (Xu, Xu, and Li 2016), Digital Media Studies (Chen, Ho, and Lwin 2017), Computer Science (Rosa et al. 2019), Politics (Bauman 2019) and Sociology (Moloney and Love 2018). Consequently, different research streams have developed in parallel, giving rise to conceptual debates about what constitutes cyberbullying and how it compares to other forms of verbal aggression on social media (Foody, Samara, and Carlbring 2015; Kowalski et al. 2014). For instance, on the difference between cyberbullying and trolling, some researchers suggest they are discrete concepts, as cyberbullying tends to be seen as the more harmful behavior (March and Marrington 2019), whereas trolling may be both lighthearted and serious (Sanfilippo, Fichman, and Yang 2018), if not a positive behavior that can increase online community engagement (Cruz, Seo, and Rex 2018). Other researchers argue that social-media users may not distinguish between cyberbullying and trolling, since the perception of what is harmful and anti-social in online environments is highly subjective and contextual (Chen 2018), given that non-verbal cues here are limited compared to offline interactions (Lapidot-Lefler and Barak 2012).

Similarly, concepts such as online incivility (Ordoñez and Nekmat 2019), flaming (Hutchens, Cicchirillo, and Hmielowski 2015), and online hate (Chau and Xu 2007) tend to overlap conceptually; consensus, across disciplines, on the conceptualization of different forms of anti-social behaviors online is yet to be achieved (Cruz, Seo, and Rex 2018; Peter and Petermann 2018). Therefore, we use cyberbullying as an umbrella term that includes all forms of anti-social behaviors online in this article, and direct readers to recent systematic reviews by Foody, Samara, and Carlbring (2015) and Moor and Anderson (2019) for an overview of different research streams and current debates.

On social-media users’ motives for verbally attacking each other, research has focused predominantly on socio-psychological factors such as personality traits (Moor and Anderson 2019), social and peer norms (Chen, Ho, and Lwin 2017), moral beliefs (Johnen, Jungblut, and Ziegele 2018), empathy (Howard 2019), and negative mood states (Pieschl et al. 2013), to name but a few (for a review, see Chen, Ho, and Lwin 2017; Guo 2016; Kowalski et al. 2014). In the face of missing empirical research on motivations behind brand-related cyberbullying, in a first step, we rely on researched based on two established theories on consumers’ life aspirations and brand identification to develop hypotheses, as will be outlined in the next section.

Consumers’ life aspirations and CBC

According to Kasser and Ryan’s (1993) Life Aspiration Theory, consumers pursue two types of aspirations in life. Intrinsic aspirations refer to the pursuit of intrinsically rewarding need satisfactions, such as having meaningful affiliations, accepting one’s self, and making a valuable contribution to the community. Extrinsic aspirations focus on attaining external rewards or praise and usually include a desire to feel popular, to fit in, and to be attractive to others. In general, intrinsic aspirations tend to trigger positive interpersonal behaviors, whereas extrinsic aspirations are more likely to lead to negative interpersonal behaviors (Kasser and Ryan 1993).

In particular, aspirations for affiliation are likely to negatively influence CBC. According to Cialdini and Goldstein (2004), affiliation relates to family life and good friends, and maintenance of meaningful social relationships (Cialdini and Goldstein 2004). In virtual environments, Claffey and Brady (2017) identified affiliation with other brand followers as the main purpose for engaging in online communities. In relation to cyberbullying, Ang’s (2015) review of existing research indicates that individuals who lack emotional relationships in their social environment and yet long for such bonds are less likely to engage in bullying behaviors, both online and offline. Conversely, in a study on narcissism, Kirkpatrick et al. (2002) found that affiliation (i.e., a desire to feel socially included by others) diminishes aggressive behavioral intentions.

Likewise, we expect self-acceptance to have a negative impact on CBC. Self-acceptance includes aspirations for personal growth, autonomy, and happiness (Kasser and Ryan 1993). Studies show that individuals with high levels of these aspirations are less likely to be perpetrators of cyberbullying (Brewer and Kerslake 2015) and aggression in general (Mofrad and Mehrabi 2015). Further, self-acceptance inhibits negative social comparisons in online communities (Appel, Crusius, and Gerlach 2015), an important contributory factor to feeling rejected, and aggressing peers who appear socially superior (Banerjee and Dittmar 2008).

Similarly, community-oriented aspirations, such as engaging in altruistic activities and contributing to society as a whole, are likely to reduce socially dysfunctional behaviors (Kasser and Ryan 1993). This is significant, given that community aspirations, such as helping others by sharing knowledge and solving problems, is a cornerstone of viable online communities (Schau, Muñiz, and Arnould 2009). A recent meta-analysis by De Wit, Greer, and Jehn (2012) shows that citizenship behavior—the will to go beyond one’s own interest in helping a group or organization—is negatively correlated to individuals’ tendency to start a conflict with someone. Promoting community-driven behaviors also seems to reduce bullying among schoolchildren (Frey et al. 2005). Taken together, we propose the following:

H₁: Consumers’ aspiration for affiliation has a negative effect on CBC.

H₂: Consumers’ aspiration for self-acceptance has a negative effect on CBC.

H₃: Consumers’ aspiration for community engagement has a negative effect on CBC.

In contrast, extrinsic aspirations are likely to increase CBC. The aspiration to be attractive to others, for instance, has already been found to cause greater amounts of aggressive thoughts (Sakellaropoulo and Baldwin 2007), appearance-related cyberbullying (Berne, Frisén, and Kling 2014), and aggressive behavior in general (Bushman and Baumeister 1998). Further, the need to be seen as attractive positively correlates with narcissism (Vazire et al. 2008), a personality trait which is predictive of bullying behavior on Facebook (Craker and March 2016). Therefore, we theorize that consumers who place a high level of importance on their appearance, and use brands to communicate their attractiveness to others in online communities (e.g., Krämer et al. 2017), are likely to bully others who somehow threaten their narcissistic needs (Weiser 2015).

The extrinsic aspiration for conformity fosters aggressive behavior. According to Grouzet et al. (2005), conformity-seeking relates to people’s aspirations to fit in with others and to appear similar to those in their aspirational groups. Studies on bullying behavior show that people aggress others in order to show that they fit in with a particular group and are different from those who are not part of that group (e.g., Shapiro, Baumeister, and Kessler 1991). Further, interviews with schoolchildren and adolescents show that those from low-income backgrounds who cannot afford brands are likely to be victims of bullying as they “do not fit in” (Isaksen and Roper 2016, 652). Breitsohl, Wilcox-Jones, and Harris (2015) also found that conformity-seeking is a frequent phenomenon in online communities as a result of stress and perceived social insulation.

The need to be popular and admired is also likely to produce CBC behavior. Findings from studies on political hate-speech (Sobkowicz and Sobkowicz 2010) and harmful peer interactions (Isaksen and Roper 2012) support the notion that verbal aggression may be used as a social tool to gain admiration from others. Online community users, for instance, are likely to experience jealousy if they feel less popular than others (Appel, Crusius, and Gerlach 2015), which in turn can lead to the derogation of those who are envied (Banerjee and Dittmar 2008). Arguably, consumers in online brand communities may therefore satisfy their need for popularity by directing derogatory comments at other brand users who may have breached a community norm or engaged in brand criticism (see Breitsohl, Roschk, and Feyertag 2018). Taken together, we propose the following:

H₄: Consumers’ aspiration for attractiveness has a positive effect on CBC.

H₅: Consumers’ aspiration for conformity has a positive effect on CBC.

H₆: Consumers’ aspiration for popularity has a positive effect on CBC.

Consumers’ brand bond and CBC

Although studies on the link between cyberbullying and consumers’ bond with brands are scarce, research based on Consumer Identification Theory allows for some tentative propositions. This research suggests that proximity of one’s self-image to a brand’s image manifests in consumers’ identity beliefs as well as by their loyalty behaviors (Algesheimer, Dholakia, and Herrmann 2005; Stokburger-Sauer, Ratneshwar, and Sen 2012). In relation to consumers’ brand identification, Isaksen and Roper (2012) found that individuals who draw on brands to develop and express their identity tend to bully those who do not identify with the same brands. Likewise, Wooten (2006) found that when a consumer feels that someone presents a threat to his/her identity, the consumer tends to respond aggressively. This is likely to particularly occur in online brand communities where members sanction those who do not adhere to or criticize the values of the brand upon which the community is built (Luedicke, Thompson, and Giesler 2010).

Similarly, consumers’ brand loyalty, i.e., the sustained, long-term preference and repurchase of a brand (Chaudhuri and Holbrook 2001), is likely to increase CBC. Previous studies on online communities suggests that the long-term preference of one brand over another can lead consumers to attack those who do not share the same preference (Thompson and Sinha 2008), or brand rivals who openly support competitors (Schau, Muñiz, and Arnould 2009). Further, the longer one has been loyal to a brand, the higher the costs of switching to a competitor since switching would require an acknowledgement of having followed the wrong brand (see Lam et al. 2010). Arguably, high switching costs may therefore lead loyal brand followers to attack anyone whose comments threaten the superiority of their chosen brand. Taken together, we propose the following:

H₇: Consumers’ identification with a brand has a positive effect on CBC.

H₈: Consumers’ loyalty to a brand has a positive effect on CBC.

Figure 1 summarizes our hypothesized relationships.

Figure 1. Conceptual model: Motives for consumer brand-cyberbullying.

Method

To test our hypotheses, we designed an online survey and conducted a two-stage pilot test with Facebook brand-page users, since this is where Breitsohl, Roschk, and Feyertag (2018) identified instances of CBC. In stage 1, we sought qualitative feedback from four marketing academics and 12 postgraduate students, who indicated regular use of Facebook brand pages. In stage 2, we sent the survey to 26 regular Facebook brand-page users without a marketing background, to garner further feedback and eliminate potential terminological misunderstandings. We subsequently posted a link to the final survey in 13 online communities and thereon reposted it for three weeks. More specifically, following Ridings, Gefen, and Arinze (2006), we considered an online community to be suitable for inclusion if it contained a minimum of 10 posts per day by at least 15 different members for each of three consecutive days chosen at random. Due to access restrictions on a number of corporate Facebook brand pages, we chose to include unofficial brand pages hosted both on Facebook and independent forums. A screening question—“Do you follow the official Facebook brand page of the brand which you entered in the previous textbox?”—ensured that all respondents were users of corporate Facebook brand pages independent of where the survey link was posted.

Our survey generated a total of 1,203 completed and utilizable responses. Respondents were predominantly male (71%), aged between 18 and 34 (60%), and reported a monthly income between US$1,651 and US$7,000. A majority of respondents (57%) indicated that they posted on Facebook brand pages at least once per month, and a majority (59%) visited brand pages at least twice per week.

In our survey, we employed established measurement instruments taken from the Branding and Psychology literature—based on five-point Likert scales (1 = “strongly disagree” to 5 = “strongly agree”). We measured consumers’ life aspirations via the refined Aspiration Index (Grouzet et al. 2005), and consumers’ brand bond via constructs from Stokburger-Sauer, Ratneshwar, and Sen’s (2012) study on brand identification and brand loyalty. For CBC, we adopted Parada’s (2000) Peer Relations Instrument and adjusted items to fit the brand-specific context. In doing so, we conducted a pretest which provided 82 marketing students (aged between 18 and 51) with a selection of typical CBC comments. We then gave them a list of terms from Parada’s instrument and asked them to indicate which term or terms best described what they saw in the CBC comments. We took the three terms that participants mentioned most frequently, namely “teasing,” “making rude remarks” and “making jokes.” Table A1 in the Appendix provides the details of all instruments used.

Table A1. Measurement items and psychometric properties.

Construct name	Construct measurement
Affiliation (α=.806; CR=.899; AVE=.691)	People will show affection to me, and I will do likewise to them. I will feel that there are people who really love me. I will express my love for special people. I will have a committed, intimate relationship.
Self-acceptance (α=.795; CR=.898; AVE=.690)	I will choose what I do, instead of being pushed along by life. I will deal effectively with problems in my life. I will feel good about my abilities. I will overcome the challenges that life presents me.
Community (α=.682; CR=.841; AVE=.726)	The things I do will make other people’s lives better. I will help the world become a better place.
Attractiveness (α=.864; CR=.918; AVE=.788)	I will have people comment about how attractive I look. I will keep up with fashions in hair and clothing. I will achieve the “look” I’ve been after.
Conformity (α=.722; CR=.840; AVE=.725)	My desires and tastes will be similar to those of other people. I will “fit in” with others.
Popularity (α=.756; CR=.855; AVE=.748)	I will be admired by many people. My name will be known to many people.
Brand Identification (α=.840; CR=.908; AVE=.767)	I identify strongly with my brand. My brand embodies what I believe in. My brand is like a part of me.
Brand Loyalty (α=.734; CR=.860; AVE=.674)	I stick with my brand because I know it is the best brand for me. I will buy this brand the next time I buy a product. I intend to keep purchasing this brand.
Consumer Brand-Cyberbullying (α=.811; CR=.901; AVE=.752)	I have teased others about their brand. I have made rude remarks to others about their brand. I have made jokes about others regarding their brand.

Note: Before respondents were exposed to the scales for Brand Loyalty and Brand Identification, they were asked the following: “The next statements refer to the online brand community in which you have found the link to our survey. Please enter the name of the brand that you support in this community into the textbox below.” The entered brand name was then automatically carried forward into each scale item, replacing the words “my brand” or “this brand” with an actual brand name that respondents had entered (e.g., Nike).

Results

To analyze the data, we used LISREL 8.7 and structural equation modeling (SEM). SEM is a well-accepted covariance analysis in marketing research for estimating causal models and multivariate data sets (Iacobucci 2009). It enables the researcher to test multiple regressions between all constructs within the same analysis, while using a number of diagnostic tools to account for various measurement errors (Gefen, Straub, and Boudreau 2000; Hair et al. 2017). For these reasons, SEM tends to be the preferred analytical method in survey research and cross-sectional studies (Bagozzi and Yi 2012). We followed the established Anderson and Gerbing’s (1988) two-step approach, consisting of the validation of the measurement model and subsequent analysis of the latent variables via path analysis.

As a first step, we examined the measurement model to test for reliability and validity. An exploratory factor analysis (EFA) confirmed the unidimensionality of analyzed factors, which is important to determine the correlations among the observed variables and related factors in a dataset (Bagozzi and Yi 2012). Subsequently, a confirmatory factor analysis (CFA) showed that all standardized loadings were larger than .7 and significant at p<.001. For a measurement model that adequately fits the data, Baumgartner and Homburg (1996) suggest a number of threshold criteria, stating a required comparative fit index (CFI) above .9, a goodness of fit index (GFI) above .9, and a root mean square error of approximation (RMSEA) below .5. A satisfactory measurement model offers an additional degree of confidence in the relationship between the observed data and the proposed model.

Given that all fit indices of our model met these criteria (χ2 = 1205.17, p = .01, CFI = .97, GFI = .93, and RMSEA = .05), in our estimation, it provided a good fit. SEM studies need to establish several other criteria to validate the adequacy of their measurement constructs, including measures of convergent validity, reliability, and discriminant validity (Bagozzi and Yi 2012). Commonly reported criteria to establish convergent validity are an average variance extracted (AVE) of at least .5, and a composite reliability (CR) of at least .7 (Bagozzi and Yi 2012), and our data exceeded these thresholds. To measure the reliability of our constructs, we computed Cronbach alpha (α) values. For a construct to be reliable, the α needs to be above .7, and this was the case for all our constructs apart from Community. Since the composite reliability for the Community construct was sufficient, and all other coefficients in our study fulfilled the required criteria, we proceeded with the analysis including the Community construct, as suggested by Cronbach and Shavelson (2004). The AVE value for each construct was always greater than the squared correlation estimated between any two factors (see Table 1), suggesting discriminant validity (Fornell and Larcker 1981).

Table 1. Correlation matrix.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
Affiliation (1)	.831	.830	.719	.174	.261	.303	.371	.538	.065
Self-acceptance (2)	.788	.831	.693	.085	.128	.278	.255	.491	.058
Community (3)	.775	.754	.852	.419	.454	.520	.475	.576	.196
Attractiveness (4)	.339	.268	.535	.888	.798	.864	.431	.368	.483
Conformity (5)	.409	.302	.563	.838	.851	.808	.491	.431	.456
Popularity (6)	.442	.422	.616	.830	.760	.865	.450	.393	.494
Brand Identification (7)	.497	.404	.580	.545	.593	.560	.876	.813	.585
Brand Loyalty (8)	.630	.593	.661	.494	.545	.514	.766	.821	.463
CBC (9)	.252	.246	.357	.586	.565	.595	.668	.570	.867

Note: Values below the diagonal show correlations between constructs; values on the diagonal (in bold) show the square root of AVE; and values above the diagonal show correlations between the constructs adjusted for the marker variable. All correlations were significant at .001.

We further tested for common method variance (CMV) to identify unwanted correlations among our focal variables, i.e., correlations that exist because of how we designed our research instrument, rather than for theoretically meaningful reasons. Following Lindell and Whitney (2001), we designated a marker variable—corporate social responsibility (“The company should intervene when there is undesirable behavior in this community”), and used the lowest correlation (r = .20) as a proxy to adjust the correlation matrix. All correlations that were significant before the adjustment remained significant, suggesting that CMV did not affect our findings.

To analyze the hypothesized relationships, we then used SEM based on the maximum likelihood estimator, a commonly used statistical inference framework that has been shown to perform well in SEM (Hair et al. 2017). As can be seen in Table 2, the overall model fit was acceptable. Results generally confirm our hypotheses that consumers’ intrinsic life aspirations have a negative effect on CBC, while extrinsic life aspirations have a positive effect on CBC. Two life aspirations—Self-acceptance and Conformity—did not have a significant effect on CBC. Both factors related to consumers’ identification had a positive effect on CBC. In accordance with past research on life aspirations and consumer behavior (Kasser et al. 2014) we also included several control variables. Both age (ß: .07) and gender (ß: .09) had a significant effect on CBC, while posting frequency and income did not.

Table 2. Motives for Consumer Brand-Cyberbullying (CBC): Structural parameter estimates.

Hypothesized path	ß	t-value	Result
H1: Affiliation → CBC	−.19	−3.56***	Supported
H2: Self-acceptance → CBC	.06	1.19 ^n.s.	Rejected
H3: Community → CBC	−.11	−2.76***	Supported
H4: Attractiveness → CBC	.18	2.60***	Supported
H5: Conformity → CBC	.01	.28 ^n.s.	Rejected
H6: Popularity → CBC	.18	3.38***	Supported
H7: Brand Identification → CBC	.35	8.40***	Supported
H8: Brand Loyalty → CBC	.12	2.85***	Supported
Goodness of fit χ2 = 1119.96 (p=.00); GFI=.94; CFI= .98; RMSEA=.049

Note: ***p < .001; n.s. non-significant.

Conclusions

Theoretical implications

The cyberbullying literature in Psychology as well as related literatures in disciplines such as Sociology and Information Studies have overlooked consumers’ bond with brands as a motivating factor. Our study makes a contribution by locating brand identification as an identity-centric motivating factor for cyberbullying. On a larger canvas, it complements work on political- (Clark 2016; Gil de Zúñiga, Barnidge, and Diehl 2018), race- (Ramasubramanian 2016), religion- (Barzilai-Nahon and Barzilai 2005) and gender- (Gruber 1999; Herring 1999) related issues and identity.

Likewise, the Marketing and Branding literature has largely overlooked cyberbullying in studies related to consumer’s identification with brands. Extant studies are limited to anecdotal suggestions, they note that consumers often attack each other in online brand communities, but do not systematically examining this phenomenon. Breitsohl, Roschk, and Feyertag (2018), examining textual data (i.e., consumer comments) in their study on cyberbullying, are able to provide evidence of its occurrence but not engage the question why CBC takes place. We take this emerging literature a step further with a survey-based study on people’s motivations for engaging in CBC. More specifically, we show that brand loyalty and brand identification can lead to CBC.

Further, our study offers fresh insights into how individuals’ general aspirations in life can lead to brand-related cyberbullying. Our findings indicate that intrinsic and extrinsic life aspirations have diametrically different effects on CBC. We show that consumers who aspire to be popular and attractive to others are more likely to engage in CBC. However, those following life aspirations that privilege their community and affiliation with others are less likely to do so. This generally aligns with Kasser and Ryan’s (1993) early work, which shows that those following intrinsic aspirations are less likely to engage in anti-social behavior than those following extrinsic aspirations (see also Kasser et al. 2014).

Managerial implications

Our study also provides insights for policymakers and marketing practitioners. For the former, we flag the need for them to take notice of cyberbullying in online brand communities. Our findings show that intrinsic life goals reduce the likelihood of CBC—they should therefore be cultivated. Approaching from the other side, educational campaigns that generate an understanding of the negative effects of extrinsic life goals may also help reduce CBC.² For the latter, we propose that management of CBC can be another vehicle for positioning their companies as responsible participants in the marketplace. It could be a corporate contribution toward improving social well-being, as part of their companies’ corporate social responsibility and philanthropic activities.

Limitations and future research

Our study paves the way for several avenues of future research. First, our choice of life aspirations for this study should only be seen as an initial exploratory step. Future research may fruitfully test alternative and complementary constructs for understanding people’s motives for engaging in CBC. For instance, recent work on aggressive dialogues on Facebook indicates that personality traits are another significant predictor of verbal aggression on social media (Bollmer et al. 2003). In particular, the dark triad traits—narcissism, Machiavellianism, psychopathy, and sadism—are worth exploring in this respect (Craker and March 2016). Second, we urge researchers to investigate the role of companies that host online communities. Do consumers expect companies to intervene when CBC occurs in an online brand community? Moreover, what type of intervention is most appreciated by those actively involved in a CBC episode, and those who passively observe it? Dineva, Breitsohl, and Garrod’s (2017) observational study suggests that currently companies follow a strategy of ignoring any form of consumer-to-consumer conflicts. However, future research should employ experimental designs to test different potential corporate interventions, such as censorship (Pfaffenberger 1996), explicit community rules (De Cindio et al. 2003), troll management tactics (Herring et al. 2002) and other forms of recently identified governance strategies (see Feenberg 2019; Helberger, Pierson, and Poell 2018).

Finally, there is scope to expand on two elements of our methodological approach. One, our selection of online communities focused exclusively on Facebook brand pages. Researchers could explore other social-media channels. For instance, it is likely that CBC also occurs on Twitter (see Simunjak and Caliandro 2019), which, unlike Facebook, is more limited in terms of how consumers can communicate, as well as in its interactive features (John and Nissenbaum 2019). There may also be some cross-channel effects, where consumers engage in CBC on several social-media channels and interlink their content (Haythornthwaite 2002). Second, researchers could build on scale items we used to identify CBC. Due to the lack of studies on brand-related cyberbullying, we chose to adopt an existing instrument and use pilot tests to identify the three items that best reflect CBC from a respondent’s perspective—namely “teasing,” “making rude remarks” and “making jokes.” We hope researchers will explore more holistic instruments to capture CBC, perhaps incorporating additional items that reflect new types of cyberbullying, such as threats and constructive criticism (Breitsohl, Roschk, and Feyertag 2018).

Notes

1 Names have been changed for purposes of anonymity; spelling errors have been kept as originally found on the brand page.

2 Studies on the effectiveness of cyberbullying interventions, at present, remain inconclusive (Ang 2015), yet there is some evidence that targeting specific motives of Internet users can render anti-bullying campaigns more successful (Yeager et al. 2015).

Appendix