Examining the Online Posting Behaviors and Trajectories of Incel Forum Members

Involuntary celibates, or incels, have been of heightened interest to scholars and practitioners due to their ongoing engagement in misogynistic and violent discourse. Yet the incel subculture is complex, requiring targeted strategies to develop effective interventions. Incels are not a homogenous group and vary in their engagement with the views expressed within incel spaces online. It is imperative that researchers contextualize differences among participants in online incel communities to better identify the behaviors and risk factors that may shape their risk of offline violence. Such knowledge can inform the production of tailored intervention strategies to affect individuals with varying levels of commitment and motives for participation in this movement.

Research on incels has increased but there is underdeveloped knowledge of how incels’ acceptance or expression of belief may be tied to their posting activities within online incel community spaces. The present study investigates patterns in incels’ online posting behaviors and whether acceptance of subcultural beliefs is reflected in variations of user engagement and posting behaviors over time. A sample of over 500 unique users was drawn from a well-known incel-moderated forum and analyzed using group-based trajectory modeling (GBTM) and multinomial logistic regression. User posts were analyzed over time to identify heterogeneity among users and determine whether there are emergent patterns that characterize distinct groups within the broader forum population based on their level of engagement, such as posting frequency and username. Several conclusions can be drawn from this study.

First, the findings demonstrate that there is some variation in the posting habits of forum participants, with a small number of posters accounting for a high percentage of posts over the lifespan of this sample (see Table 1 and Figure 1). This finding supports that of prior research on the general nature of online community engagement, and broader patterns of offending behaviors as a whole.

Second, the usernames associated with posters had a significant tie to their frequency of posting. Individuals with usernames reflecting the larger incel subculture were at a significantly greater risk of assignment to the high engagement trajectory group. This may be a reflection of greater acceptance of incel beliefs or a commitment to an incel identity, as noted in other studies of usernames and posting behaviors among deviant or extremist communities online.

Third, while insignificant, users with subculturally relevant usernames (i.e. usernames reflecting sentiment aligning with incel grievances and ideology) showed a greater risk of assignment to the low engagement group than the moderate. Users who appeared to be more invested in incel subculture were more often engaged in high and chronic posting engagement or desisting in their posting behaviors almost immediately after joining the forum (see Table 2).

This dichotomy may support the hypotheses of other extremist and online community scholars, in that users who were deeply engaged in extremist subcultures may enact low or nearly nonexistent posting behaviors as a form of self-preservation. Alternatively, users who are more deeply engaged in incel subculture may be drawn to less populated or more extreme third party sites.

The findings of this analysis are exploratory but provide direction for future research and presents policy implications for counter-radicalization schemes and tactics to disrupt extremist incel community engagement. First, it is important to note that many members of the community appear to disengage after a relatively brief period of time. It is unknown if they cease participation due to a lack of interest, an inability to connect to the community or other factors. Thus, attempts to de-platform these communities, such as blacklisting the sites from web hosting services, may have a limited impact on the community as a whole. Instead, there may be greater value in developing strategies to counter the acceptance of violent narratives among the small, persistent posters within forums and other online spaces. Given the emphasis on blackpilled and misogynistic ideologies among subcultural participants, there may be value in developing targeted ads and content that attempt to counter those messages and redirect users to alternative online resources. Scholars have similarly suggested that counter messaging aimed at gendered narratives may help to counter disinformation and the acceptance of conspiracy theories. Such P/CVE campaigns are still being explored and could be inherently useful to reduce engagement with incel communities. Second, this analysis did not examine the content of posts or changes in users’ expression of views over time. Future research is needed to explore the degree to which incel participants’ language changes in virulence relative to the onset, persistence, and exit from posting over time. Online communities and their userbase change frequently, thus the results of this analysis may not be applicable to all incel communities online or remain valid over time. Additionally, the patterns noted here may not be observed in other online platforms, like social media sites which have different tolerances and policies targeting hate or extremist rhetoric.

Lastly, results from GBTM and multinomial regression suggest that subcultural engagement impacts users’ risk of assignment to trajectory groups reflecting three distinct latent patterns of posting behaviors over time. These results also underscore the importance of tailored interventions, as opposed to widespread efforts to counter radicalization and extremism online, such as de-platforming. Rather, these efforts may have the opposite effect for those who pose less risk, driving them deeper into the subculture as they continue seeking out comradery and support resources. As a result, widespread efforts to disrupt or de-platform incel communities online may not be as effective as targeted interventions and may run the risk of strengthening incel identification of outsiders as adversaries. Future research should consider factors that influence posting behaviors within online extremist forums while working to determine methods for disrupting engagement or de-escalating trajectories for targeted subgroups within the forum. The present study found that there are distinct behavioral subgroups within this particular forum, but it is unlikely that these behavioral patterns are replicated across other online incel spaces and future research is necessary to map the ways in which posting behaviors and affiliated risk factors vary based on the online platform.