The Social Science of Online Radicalization


As Charlie Winter noted recently at War on the Rocks, the Islamic State’s robust social media apparatus has been propagating a remarkably effective, multi-faceted communications strategy that incorporates narratives of statehood, military success, and religious legitimacy. The Islamic State’s success in using social media to disseminate its extremist ideas and mobilize tens of thousands of foreign fighters to join the caliphate has raised many questions about the relative efficacy of online radicalization and recruitment. Can social networking sites replace face-to-face communications in fostering the group dynamic that is so important to spurring people to engage in terrorist acts? How do online group dynamics differ from those of face-to-face networks? Does social media accelerate the process of radicalization, so that individuals may be ready to illegally support violent causes more quickly after exposure to extremist ideas than in the past? It is vital that we seek to understand these questions in order to counter the Islamic State’s social media outreach and more effectively respond to other groups that seek to emulate its successes.

Public discussion and debate about these issues has tended to overlook the rich body of research produced by psychologists, sociologists, and communication scholars. Academics have been studying the impact of online or “computer-mediated” communication (CMC) on human behavior since the 1960s, and the literature on the subject, especially three concepts from the field of social psychology — identity demarginalization, group polarization, and the social identity model of de-individuation effects — can do much to inform our exploration of online radicalization.

Identity Demarginalization and Group Polarization: Validating Fringe Identities

Identity demarginalization theory, as articulated by Katelyn McKenna and John Bargh in a 1998 study, explores why some social groups are more drawn to online communication than others. McKenna and Bargh found that individuals with “concealable and culturally devalued identities” were more likely to participate in and value online communities than individuals with mainstream identities. Specifically, their study found that people who posted in online forums dedicated to concealable identities such as homosexuality or drug usage valued the feedback and opinions of other group members more strongly than did members of forums focused on marginalized identities that are easier to perceive, such as obesity and stuttering. “For the first time,” the authors wrote, an individual exploring his or her marginalized identity in an online environment “can reap the benefits of joining a group of similar others: feeling less isolated and different, disclosing a long secret part of oneself, sharing one’s own experiences and learning from those of others, and gaining emotional and motivational support.”

Online communities may be uniquely powerful among groups with concealable identities because the Internet provides more anonymity, reach, and in-group reinforcement than these people are likely to find in mainstream society. Relative anonymity can embolden individuals with concealable marginalized identities to discuss issues that may be taboo in a mainstream social setting; for example, in a study of lesbian, gay, and bisexual (LGB) online groups, one woman noted: “I think it’s much easier to talk about certain things online such as relationships, sexual things and compliments and insults. It’s easier to talk to someone when you don’t have to see the physical reaction and think of a response right away.”

As for geography, in a study examining the website Stormfront and the white nationalist movement, Neil Caren and two colleagues note that the absence of spatial boundaries allows online communities “to draw in otherwise isolated movement participants.” The same phenomenon has been observed in online interactions between salafi jihadists. As analyst and author J.M. Berger remarked in Senate testimony: “It’s different than the 1950s when, if a radical jihadist was in Peoria, he might go his whole life without finding somebody who shares his views. Now it may take 10 minutes.”

Participation in online communities can also lead to a self-reinforcing of what McKenna and Bargh called demarginalization. Their study found that people who actively took part in online discussions not only came to “consider the group identity more important than did those who did not actively participate,” but also intensified marginalized behaviors based on positive reinforcement from other group members. A 2008 study of pro-anorexia online communities found that such forums were “an ideal space for maintaining and validating a pro-anorexic identity.” Pro-anorexia individuals participating in the forums received encouragement and guidance from like-minded members, which reinforced the forum participants’ commitment to their pro-anorexia identities. Joining and receiving positive feedback from a like-minded group online can help group members come to view their concealed identities more positively — sometimes to the point of incorporating them into their public personas.

The Internet’s ability to help individuals reinforce and even publicly embrace once-concealed marginal identities is neither inherently good nor bad: There is obviously an enormous moral gulf between one individual coming to accept the fact that he is gay and a second person coming to positively embrace his identity as a white supremacist. With respect to the Islamic State and other extremist groups, we need to understand the power of online communities to cause individuals to adopt and, crucially, decide to act upon the group’s violent extremist ideology.

Group polarization theory expands on some of the themes highlighted in identity demarginalization. Group polarization refers to the propensity for groups to become more extreme in their outlook through mutual reinforcement. As with identity demarginalization, this process is neither inherently good nor bad: It can produce both virtuous and vicious cycles, but tends to have greater effects on more cohesive groups either way. Numerous studies have concluded that groups who interact online experience a greater degree of group polarization than groups that interact face-to-face. One key reason may be the general absence of visual and verbal cues in online communications. As the LGB forum participant quoted earlier in this article observed, individuals interacting online can speak freely without having to worry about the facial or physical expressions of their peers. This reduces inhibitions, and participants become willing to “contribute more novel arguments and engage in more one-upmanship behavior,” which may drive group polarization.

Social Identity Model of Deindividuation Effects: The Group Identity Takes Over

The social identity model of deindividuation effects (SIDE) provides a framework for understanding processes like group polarization and demarginalization in an online setting. SIDE is a revision of classic deindividuation theory, which suggests that group immersion and anonymity within a group result in a loss of self-awareness and an increase in anti-normative behavior. SIDE, in contrast, concludes that in an online context, anonymity and group immersion do not foster anti-normative behavior. Rather, they cause participants to ignore differences between in-group members, and to more closely embrace a group identity. As Tom Postmes and his colleagues explained in Communication Research (1998), the SIDE model found that individuals who adopt a group identity are receptive to group cues, and are thus more susceptible to adopting the behavior of that group, regardless of whether such behavior is normative or anti-normative in society as a whole.

The SIDE model has significant implications for online communications. Several studies have suggested that the Internet reduces the importance of personal characteristics and interpersonal differences, and increases the salience of group identity and group norms. Anonymity in interactions may also accentuate distinctions between members of the group and non-members, and intensify intergroup hostility.

Other factors besides anonymity may also facilitate the shift from individual to group identity in online groups. In a 2011 study, Haines and colleagues modified the SIDE model, concluding that group identity becomes more salient in online interactions when a group identifier (for example, avatars distinguishing in- and out-group members) is visible. In the study, Haines found that individuals with some (even limited) awareness of the opinions of other group members were more likely to conform to group norms than were individuals who had no awareness of other group members’ opinions. Thus, Haines concluded that group influence decreases in completely anonymous online situations — for example, where “no labels are attached to comments” — due to lack of awareness of others’ opinions. But group influence increases when common group identifiers exist. This finding is particularly relevant for platforms like Twitter and Facebook, where group identifiers can be reflected in avatars and other symbols attached to a user’s profile.

Toward a Richer Understanding of Online Radicalization

Identity demarginalization, group polarization and SIDE all have considerable explanatory power in informing the discussion about online radicalization. All three theories demonstrate that certain characteristics of online communications, including reduced social cues and anonymity, often strengthen group influence at the expense of individual identity. Group polarization and identity demarginalization explain the social mechanisms by which political extremism and other fringe identities can become validated and more pronounced in online communications.

With online communications and social media platforms constantly evolving, there remain significant knowledge gaps concerning their influence on human behavior. More research needs to be undertaken about the applicability of CMC-related theories to Web 2.0 platforms like Twitter and Facebook. Additionally, more research is needed on how and when disengagement from online communities occurs (which can in turn provide clues about mechanisms for disengagement from support for violent extremism). Social science research on online and human behavior is a vital resource for scholars and practitioners interested in understanding and countering violent extremists’ online radicalization and recruitment activities, and should figure more prominently in public discussion.


Daveed Gartenstein-Ross is a senior fellow at the Foundation for Defense of Democracies and the chief executive officer of Valens Global, a consulting firm that focuses on violent non-state actors. An adjunct assistant professor in Georgetown University’s security studies program, he holds a Ph.D. in world politics from the Catholic University of America. Nathaniel Barr is an analyst at Valens Global.


Photo credit: James Rhodes