Disentrained Channels

Channel entrainment refers to the synchronization of interactional partners’ interdependent communicative activities (Walther, 1996), so that disentrained channels are media that provide greater temporal resources to interactants, facilitating strategic self-presentation and reducing copresence. Whereas face-to-face and videoconferencing reflect synchronous (and thus more entrained) channels, e-mail and Facebook posts typically reflect asynchronous (and thus more disentrained) channels. These disentrained channels provide interactants greater temporal and cognitive resources to construct self-presentation and interpret others’ messages. Social media disconnect interactants from their messages to varying degrees, providing senders time to encode their own messages while decoding others’. This channel disentrainment fosters opportunities for groups and their members to more carefully construct and consider both their own selves and their interactions, but also presents some challenges to task-based group outcomes.

Disentrained channels have long been noted to affect intergroup communication, particularly through the obfuscation of senders and the strategic presentation of self that may occur. Perhaps the largest manifestation of this effect is the equalization hypothesis, in which the anonymity and obfuscation of status cues given off via disentrained channels can lead to the reduction of status differences and inequality within groups (Kiesler, Siegel, & McGuire, 1984). Several studies have supported the equalization hypothesis (Baltes, Dickson, Sherman, Bauer, & LaGanke, 2002; Bordia, 1997; Scott, 1999). Additional findings indicate that a lack or reduction of identifiability of group members can result in a more equal allocation of contributions among group members stemming from both social (Siegel, Dubrovsky, Kiesler, & McGuire, 1986) and group (Dubrovsky, Kiesler, & Sethna, 1991; Hollingshead, 1996) status differences. Moreover, this equalization phenomenon can reduce symptoms of groupthink-like conformity to group norms or within-group authority (Lee, 2005).

Though channel disentrainment may facilitate some social and participatory gains (Nowak, Watt, & Walther, 2005), disentrainment can also stymie functional and task-related gains of mediated group interactions. Findings of experimental studies using entrained face-to-face and disentrained CMC groups generally reveal that CMC groups report greater social attraction toward group members, satisfaction with intragroup communication, and group cohesion; however, the quality of group’s task-based decisions are typically lower (Walther, 1995; Walther & Bunz, 2005). Similar results have echoed this schism in effects, noting that while disentrained channels may lead to more favorable social attributions of group members, those gains are matched by reductions in the quality of decision outcomes for task-oriented groups (Baltes et al., 2002; Hollingshead, 1996; Straus & McGrath, 1994; Warkentin, Sayeed, & Hightower, 1997). Within the disentrained channels of social media, similar effects may be expected: groups communicating via social media may see positive social outcomes but more laborious decision-making. It is thus important to consider both the perceptual and functional effects that altering the temporal resources available to group members may have on communication.

Value Derived from User-Generated Content

A second component of social media is that individuals using social media do so primarily due to the perceived utility or value they obtain from content within social media generated by other users, rather than by the medium’s superordinate organization or content. This component helps distinguish social media both from static websites (i.e., Web 1.0) and the broader subset of user-generated web tools (i.e., Web 2.0). For example, though users certainly may generate and read comments posted to online news articles, few readers are likely go to the New York Times (Web 1.0) website primarily for those comments. Rather, users go to NYT.com for the content posted by the site’s owner: news articles, editorials, and interest stories posted by the New York Times staff. Contrariwise, just because a site is comprised primarily of user-generated content does not innately define it as a social medium, particularly when users are not able to interact. RottenTomatoes.com, a movie rating service that amalgamates and displays users’ evaluations of movies as a Web 2.0 site, is not driven by user interaction, and thus would not be considered a social medium.

To exemplify the importance and distinction of this component relative to intergroup communication, consider Facebook without any users. Facebook, the company, can provide the infrastructure and source code to enable Facebook.com, and even generate stories and news summaries; but without users to generate content and (more importantly) interact with other users about that content, Facebook would not be a terribly interesting tool. The utility and value of numerous other social media (e.g., Pinterest, CyWorld) would likewise plummet were no one to provide content and interact with other users about that content. The necessity of interaction in social media is thus a unique property of group communication relative to other channels. In face-to-face interactions, and even in other CMC channels, an individual can be a group member without interaction. An individual wearing the scarf or jersey of a sporting team may identify as a fan of that football team even without interacting with other fans, either of the same team or of their rivals (Kwon, Trail, & James, 2007). Likewise, thousands of individuals may attend a sporting event and vigorously root for the home team, but never talk with other individual fans as in-group members. Even in traditional Web 1.0 and 2.0 tools, users can derive utility and social identity from the website without social interaction with other in-group (or out-group) members. In social media, the absence of communication is not possible, as the content and utility of social media are dependent upon interaction. Thus, even more than in other channels, within social media group communication should be a driving force through which individuals’ social identities are primed and activated, social identification occurs, and in-group norms are established and conformed to. For scholars, this component increases the importance of studying communicative exchanges (rather than isolated messages) in the formation, maintenance, and termination of group processes within social media.

Masspersonal Communication

As the final component of social media, masspersonal communication refers to the convergence of previously bifurcated conceptualizations of mass communication and interpersonal, recognizing opportunities for messages to have properties of both. O’Sullivan and Carr (in press) explicated masspersonal communication as messages wherein:

(a) individuals use conventional mass communication channels for interpersonal communication, (b) individuals use conventional interpersonal communication channels for mass communication, and (c) individuals engage in mass communication and interpersonal communication simultaneously.

(p. 4)

More directly, O’Sullivan and Carr defined masspersonal communication as those message concurrently (i) highly personalized for a specific individual or set of receivers and yet (ii) accessible to a broad and oft-indistinguishable audience. Masspersonal communication is not unique to social media—or even CMC in general—though it has become much more commonplace via social media. A marriage proposal via a Jumbotron is masspersonal, as it is (hopefully) a personalized message accessible to all attendees. In social media, the ability to present a personalized message accessible to a mass audience is even more readily afforded. Twitter reflects a means of readily communicating interpersonally via a mass channel, so one user’s tweet, even if intended for another specific Twitter user, is technically accessible to any other Twitter user to chance upon. Even in a more closed social media environment like Facebook, in which users generally need to mutually acknowledge ties to interact, a status update is visible to the user’s entire network, even when the status is directed toward a specific individual. Thus, individuals are rarely truly interacting interpersonally in social media, but rather do so masspersonally. Several studies have noted that the ways individuals communicate masspersonally differ meaningfully from how individuals communicate either interpersonally or via mass communication.

Masspersonal messages and the media that facilitate them can affect the nature and processes of group communication, most notably by changing the directionality of message transmission and receipt. For example, in a study of online cancer support communities, Love and Donovan (2014) found that young adults commonly used social media features to broadcast support-seeking messages to other in-group community members, but received personalized and private feedback interpersonally from individual members. Similarly, a content analysis of pro-anorexia blogs revealed that bloggers sought and received social support masspersonally. Bloggers used media tools to both (a) interact interpersonally with others through direct messages and (b) communicate via publicly accessible messages that could include other group members with similar support needs in subsequent conversations (Tong, Heinemann-Lafave, Jeon, Kolodziej-Smith, & Warshay, 2013). These findings suggest that group members strategically (though perhaps not consciously) use social media to manipulate and harness the level of identification (i.e., personal or social) with other members. Masspersonal communication within social media can thus be used to either strategically include or exclude others, emphasizing or making salient groups and their processes.

Context Cues and Social Identity

There are several ways the structure, nomenclature, focus, or conventions of a social medium can guide group processes. One of the most fundamental cuing mechanisms is the very nature of an individual social medium. Initially, the very nature of each social medium can cue social identities. The core of a medium may be evidenced in its mission statement, its users’ adoptions and norms, its affordances, or even its terminology. For example, the names of many social media tools explicitly denote—and in so doing activate—various groups and social identities. DeviantArt affords users a means of sharing and critiquing sketches, drawings, painting, and other work among fellow artists. Likewise, CafeMom facilitates interactions among mothers, GovLoop connects and allows discussions among government employees, and HR.com situates itself as a tool for interaction among human resource professionals. In these examples, the naming of the social medium itself makes salient particular groups to guide interactions: One is certainly able to discuss topics unrelated to artwork on DeviantArt, but will quickly find other users more welcoming of users and topics who are also engaged as artists. Likewise, individuals on CafeMom surely identify (both personally and socially) as more than mothers in other contexts, but as individuals seek out and interact with others on the medium, they do so guided by their social categorization as mothers rather than other facets of their identity (e.g., mountain climber, CEO, wife).

An additional way groups and group processes may be activated within social media is through the emergence of groups or communities tied to certain features, affordances, or norms often either established a priori or that emerge naturally through group behaviors to form a community. Eckert and McConnell-Ginet (1992) defined a community of practice as “An aggregate of people who come together around mutual engagement in an endeavor. Ways of doing things, ways of talking, beliefs, values, power relations—in short, practices—emerge in the course of this mutual endeavor” (p. 464). Exploring the emergence of a community of practice within the photo-sharing social medium Flickr, Smock (2012) noted that site affordances (e.g., sharing of camera equipment or settings like aperture, exposure time, focal length used to take the displayed photo) enabled advice-seeking, feedback provision, and general interaction among fellow shutterbugs within the broader Flickr service. In such cases, individuals can form or join communities within a specific service due to the group’s focus, as the focus may signal a general shared interest or common group identity. For example, frequent contributors to Star Wars–themed Wookiepedia¹ may interact via the wiki’s talk pages, wherein they do so knowing contributors are predisposed and communicating in a context in which their Star Wars fandom is primed and guiding interactions with other Star Wars fans.

Finally, groups may be formed and clearly denoted within broader social media contexts in ways that explicitly articulate the nature and relevance of group affiliation. For example, the “star wars and star trek lovers” group on Facebook² exists as a self-selected subset of Facebook users who have opted in to the forum to discuss their pro-Star fandom. The naming of the group immediately defines and bounds group membership to a subset of science-fiction fans, establishing fans of the Star Wars and Star Trek franchises as in-group members, while establishing others (including fans of other sci-fi franchises like Dr. Who and Babylon 5) as out-group members. Even before joining the group or interacting with other group members, individuals are made aware of the nature of the group and relevant intragroup and intergroup traits germane to the group’s membership. In this way, the nomenclature or context of the social medium itself may activate intergroup communicative processes over alternative forms of communication.

“Physical” Cues

Prominent early CMC theory suggested the visual anonymity facilitated by mediated interactions would drive intergroup interactions, as individuals would be deindividuated without a corporeal representation and thus guided by social identities (Reicher, Spears, & Postmes, 1995). Some scholars (e.g., Spears & Postmes, 2015) remain adamant that visual anonymity is necessary and sufficient to engage intergroup—rather than interpersonal—processes. Yet social media often make use of multiple visual elements, particularly for self-presentation, introducing phrases like “profile photo” and “selfie” into the mainstream vernacular. Moreover, research has increasingly demonstrated that the presence of these visual cues—both individuating and personalizing—within social media channels from Instagram to LinkedIn does not innately negate the opportunity for perceptions, and communication, guided by interpersonal processes (Carr, Vitak, & McLaughlin, 2013). Rather, a more tempered and nuanced approach is merited, as these “physical” cues about an individual—ranging from depictions of clothing and other artifacts to the actual representation of a user’s self—can activate and influence group interactions similar to offline cues.

Perhaps the most notably manifest “physical” cues in many social media are pictorial self-representations. Numerous social network sites invite users to post pictures of themselves to use as individuating cues to that specific user—an invitation users typically accept, frequently posting photos as profile images (Lampe, Ellison, & Steinfield, 2007). Yet even these highly individualizing cues can activate group processes, giving off social identity cues. For example, users may express a group identity through dress or other clothing attire, either intentionally or unintentionally indicating their social category to others. An individual who selects a profile picture displaying wearing a Boston Red Sox hat and jersey gives off cues to her or his sports fandom (specifically baseball) and team identification which may activate group processes, should either sports or baseball teams be salient (Carr, Varney, et al., 2016; Earnheardt, Haridakis, & Hugenberg, 2011). Likewise, an individual may give off cues to her Islamic religious affiliation by posting a profile photo of her in a hijab (Kavakci & Kraeplin, in press), activating intergroup processes should religion be a salient social cue. Thus, how individuals pictorially depict their physical selves in social media can activate group processes, just as in face-to-face processes activated by physical—and sometimes arbitrary—cues such as clothing (see Bigler, Brown, & Markell, 2001; Goldman, Giles, & Hogg, 2014).

Less explicit and identifying than photos, the avatars used to represent individuals in social media can activate group processes. For example, Lee (2004) placed individuals into three-member groups in an avatar-mediated environment for a decision-making task. Groups were experimentally manipulated so that members were represented by either identical (i.e., three anthropomorphic bears) or different avatars (i.e., a koala, bear, and parrot), and asked to engage in high- and low-risk advice-giving tasks with two virtual confederates. Lee’s findings revealed greater conformity to group norms and confederates’ decision positions when group identities were primed and group members were represented by similar avatars/agents. Moreover, participants reported feeling innately more depersonalized (i.e., part of the social group) merely by utilizing identical virtual physical representations, which further led to conformity to the group.

Interestingly, even though the avatars used in social media need have no ties or warrant to the user’s offline self, online avatars can still guide group categorizations and effects. For example, an avatar’s gender can activate gendered stereotypes, guiding interactions based on individuals’ online presentations rather than offline selves. In a foundational study of this effect, Lee (2007) arbitrarily assigned participants to either male or female avatars and asked them to engage in decision-making tasks about typically masculine or feminine topics. Findings evidenced that the visual cue significantly affected sex inferences about the interaction partner, so that participants assumed female avatars were used by women and male avatars by men. More central to group effects, individuals were more willing to accept partners’ recommendations when the partner’s avatar’s gender was consistent with the masculinity/femininity of the topic, suggesting that even gendered avatars may activate normative assumptions and behaviors commensurate with gendered social identities, even when users are aware avatar assignment may be random. More recently, these effects were supported in more dynamic, graphical-interface environments of massive multiplayer online games (MMOs). Exploring player behavior in the popular MMO World of Warcraft, Yee and colleagues (2011) found that players represented by female avatars were more likely to engage in feminine play types (e.g., playing healing roles, more collaborative engagement) than those represented by male avatars, regardless of players’ offline gender. Consistent with additional results (e.g., Lim & Reeves, 2009; Martey & Consalvo, 2011; Martey, Stromer-Galley, Banks, Wu, & Consalvo, 2014), it seems virtual cues to identities may at times be a more substantive cue of social identity than offline characteristics, subsequently driving social identification and group effects online.

Linguistic Cues

Language has long been used as means of signifying and activating group membership and processes. At a foundational level, the construction of in-group and out-group boundaries and relationships can be conducted as simply as through the use of “us-them” diametric (Duszak, 2002a). The categorization of social groups can lead to intergroup and intragroup interaction through denoting distinct social identities (Brewer, 1999). However, language can be used to likewise connote distinct in-group and out-group boundaries, even when such groups may not clearly exist. As Duszak (2002b) notes, language can be used to “other” (p. 2) individuals perceived to be different from ourselves and the salient social group to which those individuals are compared. The explicit use of us/them verbiage can activate intergroup processes and interactions, even in dyadic exchanges. Simply framing an interaction as “us versus them/you” can infer and appeal to greater social groups to which the two interactants may respectively belong and the similarities or differences among them. In turn, the interactants then may likely do so as group members, rather than as individuals, due to the linguistic frame guiding their interaction.

For example, the nature of disclosures may subtly indicate group formation and norms. In a study of students’ online discussions in a chat forum, Deitz-Uhler, Bishop-Clark, and Howard (2005) identified normative trends in self-disclosures among group members (e.g., “I am manic-depressive.”). Interestingly, these personal self-disclosures did not increase linearly over time, as may be predicted by interpersonal processes of relational formation (Knapp, 1978). Instead, self-disclosures ebbed and flowed so that the proportion of an individual’s self-disclosing statements reflected others’ frequency of self-disclosure statements throughout the eight class sessions. Thus, the norms of both on- and off-topic selection and disclosure appear to reflect manifestations of group identification. Individuals demonstrated linguistic convergence in the masspersonal discussion forum, observing and converging the nature of their disclosures to reflect that of other in-group members.

Beyond self-disclosure and topic norms, minor and often imperceptible linguistic cues can be viewed as a means of identifying with a group through linguistic convergence. In a study of computer-mediated groups, Lea and Spears (1992) experimentally examined the use and effects of paralinguistic cues in task-group discussions. Between-condition results revealed participants in high group salience, deindividuated conditions reported positive relationships between paralinguistic cues and personal attributions of fellow group members (i.e., likability, uninhibitedness, and dominance). Particularly as compared to other conditions in which use of paralinguistics negatively predicted attributions (see also Carr & Stefaniak, 2012), Lea and Spears (1992) interpreted these findings as indicating that individuals in mediated groups ultimately develop and adhere to even conversational norms, additionally attributing more positive attributions to individuals conforming to those norms, at least over time. In other words, individuals perceiving themselves to be in a group conform to the group’s linguistic norms—even at the level of paralinguistic cues such as spelling, ellipses, and paralanguage (e.g., “um” and “errr”)—and evaluate more positively other group members who likewise linguistically conform. More recently, McEwan (2016) analyzed several of the most popular threads on Reddit for the language used by contributors among the various discussions. Groups of Redditors were more cohesive, stable, social, and interactive, and their language styles and word choice converged. Within social media, this linguistic convergence and cuing of group norms and memberships should be expected, particularly within the many social media dominated by text-based interactions. Individuals can activate and signal their group membership and social identities by adhering to linguistic and conversational norms, adopting the in-group’s dialogic style.

Social identities can also be created or denoted in social media via specific linguistic cues or markers combined with technical or system features, which can then serve as additional within-group and between-group boundaries. One of the most notable and currently visible manifestations of these specific linguistic cues is the hashtag—a keyword or phrase preceded by the “#” symbol—first popularized in Twitter but since diffused to other social media platforms. Though individuals can use hashtags as searchable terms, their adoption or appropriation by specific groups—either intentionally or ad hoc—can be used as a means of signifying group identification and engaging in larger intragroup interactions (Bruns & Burgess, 2011). For example, the hashtag #BlackLivesMatter emerged in July 2013. Following the police shooting of Michael Brown in Ferguson, MO in August, 2014, the hashtag rose to prominence as individuals used Twitter to discuss police violence, particularly toward black individuals. Circumventing traditional media channels and narratives, individuals took to Twitter and, using the #BlackLivesMatter (or #BLM) hashtag began to coordinate and organize national activities among geographically separated group members (Freelon, McIlwain, & Clark, 2016). In the case of #BlackLivesMatter, use of the hashtag was a galvanizing cue to explicitly signal users’ social identities and group affiliation.

Hashtags have likewise linguistically signaled group membership and boundaries, but without the deliberateness of #BlackLivesMatter. Exploring discussions of coffee on Twitter, Zappavigna (2014) explored the ad hoc clustering of Twitter users’ communication structures based on their shared use of the hashtag #coffee. Network analysis identified unintentional groupings of individuals on Twitter discussing the caffeinated beverage, but additionally delineated subgroups based on specific focus or interest in coffee. For example, within the larger #coffee group were embedded network clusters of individuals who (a) focused on coffee as a morning ritual; (b) were coffee gourmands, swapping precise roasting and brewing techniques; and (c) just drank coffee as a caffeinated kick-start throughout their day. These clusters developed around their own shared hashtags and communication structures, and in so doing evidence the natural boundaries of various subgroups within the broader #coffee group on Twitter. Ultimately, Zappavigna concluded that these subordinate clusters of coffee aficionados as well as the superordinate coffee-lover category represented ad hoc groups or communities online. Individuals do not necessarily go onto Twitter self-identifying as members of a coffee group or coffee-related subgroup. Instead, coffee groups naturally emerged through discourse and common language and signals such as hashtags. Consequently, groups may be defined—and their members signal their social identities and subsequent group affiliations—through shared use of common hashtags (or other sociotechnical linguistic features) and interaction with those hashtags. These processes can transcend linguistic cues, occurring in social media whose initial content is dominantly visual (e.g., Instagram, DeviantArt) as groups emerge naturally via common images or pictorial cues. In these ways, language may not only be used to maintain groups, but to bound and identify groups themselves.

Siloing Groups

One effect social media may have is siloing groups, as social media provide groups fora in which to engage in intragroup communication within their specific in-groups, limiting the potential for interaction with disparate other groups. Brundidge’s (2010) inadvertency thesis posits that the weakened boundaries among social structures and groups facilitated by social network sites and social media lead to accidental exposure to disparate groups online, specifically those political in nature. Offline, groups tend to cluster homogeneously, so that members of Elks lodges, bowling leagues, and card clubs often reflect similar political ideology, and thus members are not exposed to the perspectives of other groups. Yet Brundidge supposed that social media may encourage exposure to and interaction across group boundaries, as social networks often cut across social boundaries and groupings (Hogan, 2010). Early findings supported Brundidge’s inadvertency hypothesis, noting that Facebook use predicted exposure to diverse political viewpoints (e.g., Kahne, Middaugh, Lee, & Feezell, 2012; Kim, 2011). However, more recent work has generally refuted the inadvertency hypothesis. A growing body of literature suggests a siloing effect, wherein discussions of civics frequently occur in limited and politically homophilous groups (Kushin & Kitchener, 2009; Miller, Bobkowski, Maliniak, & Rapoport, 2015; Valenzuela, Kim, & de Zúñiga, 2012). This siloing potentially occurs as users employ system features to limit exposure to statements of those not sharing their political views or terminate online ties entirely. For example, Hayes and colleagues (2015) found that people may use Facebook’s system features to reduce—or block entirely—the amount of others’ content to which they are exposed when those others’ political views are dissimilar from their own and posted frequently. For more severe differences in political group affiliation, some individuals may even terminate social media connections entirely (i.e., unfriend; Yang, Barnidge, & Rojas, 2017).

One consequence of this siloing of groups within social media may be media switching, wherein a group isolated in a social medium uses it to coalesce group identity and norms, until a sufficient point wherein the group can spill into other channels, including offline interaction, to further manifest their group identity. One example of this is the political mobilization of young voters, who have historically been a disenfranchised and politically inactive group (Gibson, 2009). Particularly during the early integration of social media into elections in 2008, both within the United States (Vitak et al., 2011; Xenos, Vromen, & Loader, 2014) and abroad (Larsson & Moe, 2012; Moeller, de Vreese, Esser, & Kunz, 2014), social media were identified as tools to engage and mobilize young voters, ultimately transcending social media to increase the offline political engagement of this historically disenfranchised group (Carr, Hayes, Smock, & Zube, 2016).

Similar effects have been noted, particularly for special interest groups, whereby social media can serve as echo chambers (Colleoni, Rozza, & Arvidsson, 2014) in which intragroup communication isolates and intensifies group attitudes and norms. After forming online, this group can then spill into offline contexts, whereupon the social identity and values intensified online lead to strong intergroup perceptions and conflict offline. Importantly, this echo chamber intensification of polarized group beliefs transcending social media and emerging into offline action is not inherently antisocial or culturally null. Certainly, incidents of cyberbullying (Hinduja & Patchin, 2013) and increased affiliation with urban gangs (Patton, Eschmann, & Butler, 2013) anecdotally exemplify that social media echo chambers can allow groups to isolate online and ultimately result in harm, both online in situ and offline as polarized group identities spill out into offline abuse and violence. However, the Guatemalan justice movement (Harlow, 2012) and the Arab Spring (Eltantawy & Wiest, 2011) likewise evidence the potential for prosocial, civically engaged, and culturally impactful effects of social media. Social media can provide a safe place for geographically disparate individuals to come together to first identify and delineate a common group identity. Once a common identity is formed, the social medium may then be used to facilitate coordinating activities for the enactment of offline actions of the group.

Notably, this increased polarization of groups via social media can be used strategically by taking advantage of affordances and features within social media to activate social identities, and in so doing strategically motivate groups based on their social identities. One such strategic activation of intergroup differences to motivate action and behaviors was exemplified by Smith, Hitt, Park, Walther, Liang, and Hsieh (2016), who sought to use social media to increase organ donation registration. Smith and colleagues developed a social media–based campaign targeted at members of two universities whose affiliations were highly salient and oppositional. The campaign used social media advertisements to frame donor registration as a competition between the two universities (and thus social groups). This intergroup frame led to significant increases in donor registration over more personalized adverts. In this applied example, social media were used to strategically activate social identities and motivate behavior by identifying conformity to in-group behavior (e.g., “Other Wolverines like you are registering”), while concurrently opposing out-group behaviors (e.g., “Don’t let the Buckeyes win”). Ultimately, though social media may close off groups and reduce the opportunity for inadvertent communication, this isolation is not inherently either negative or positive. Rather, this siloing of groups and reduction of intergroup communication is merely one effect that can occur in social media, to various ends.

Spanning Groups

An alternate effect social media may have is spanning groups, as they provide a means by which groups and their members may seek out and communicate beyond their specific in-group, increasing the potential for interaction with disparate others. One mechanism that has been identified as a means by which intergroup communication may be facilitated or encouraged in social media is Allport’s (1979) contact hypothesis. According to Allport, the differences an individual perceives between a particular in-group and out-group can be lessened by positive interpersonal contact with a member of that out-group. By first interacting with the individual interpersonally, positive attributions about the individual are subsequently transferred onto other out-group members. Amichai-Hamburger and McKenna (2006) initially posited that CMC represents an excellent channel to enact the contact hypothesis, as mediated channels reduce the anxiety often accompanying face-to-face intergroup interactions. Subsequently, Walther and colleagues (2015) applied and tested this contact hypothesis to ameliorate intergroup differences and prejudice within an Israeli teacher-education course whose students represented three disparate social groups: religious Jews, secular Jews, and Muslims. Findings of the pretest-posttest study were consistent with the contact hypothesis, revealing students participating in initial sessions of the course via a course-management system (a limited-access social medium) reported lower levels of prejudice and greater change from pre-enrollment levels following online interactions guided by interpersonal—rather than intergroup—identification. Social media can thus serve as effective means of practically applying the contact hypothesis to help lessen intergroup differences by facilitating communication that is initially interpersonal, but that may evolve into intergroup communication as group prejudices are ameliorated.

An additional mechanism by which social media may help span intergroup boundaries and thus foster communication among otherwise distant social groups is through their frequent collapse of social boundaries and contexts. Unlike offline interactions, in which one is usually engaged in a singular social context (e.g., work, home, the bar) at any given time to guide interactions and govern social behaviors, social media often allow individuals to span typically isolated contexts. On Facebook and Twitter—where one is connected with family members, clergy, coworkers, and friends alike—individuals can interact mass personally, bridging the typically separate social groups to which they belong. Initial discussions of network formation posited that forbidden triads (e.g., if person A is close friends with person B and person C, it is unlikely persons B and C are unacquainted) should be rare (see Burt, 2005). However, as populations migrate and individuals compartmentalize facets of their social identities, such forbidden triads may be common. For example, your boss is not likely to interact with your grandmother, even though you know both well. Yet social media have helped close some of these forbidden triads, both through communicative and social mechanisms (e.g., being exposed to a masspersonal message directed to a mutual friend) and system features (e.g., Facebook and Twitter suggesting network ties) (Golder & Yardi, 2010). Consequently, both social and system structures within social media may facilitate and encourage intergroup communication even more than offline or older CMC channels.