How do representations of violent conflict differ across social media platforms? We constructed and analyzed comparable datasets of public messages and images from elite- and meso-level Syrian actors posting on three popular social media platforms. Our findings show that complementary if divergent discussions of violence remain central even amid a period of relative de-escalation. Narratives on Twitter reaching an international audience contextualize violence within the conflict’s master cleavages, while on Telegram, they address a more local audience and emphasize the violence’s day-to-day impacts. A site with stricter surveillance, Facebook features more loyalist narratives. Paired with a sample of users’ responses to an open-ended questionnaire, our results show that users across platforms diversify their presentation of violence to reach domestic and international audiences and to accommodate technical affordances, with consequences for how both audiences and researchers understand ongoing conflict.

How does work for the state shape local protest against authoritarian regimes? Autocrats can manipulate how both repressive and conventional state workers are selected, compensated and surveilled in ways that can strengthen their hold on power. But because autocrats prioritize control of their repressive staff, we should expect employment in the repressive apparatus to more directly undermine protest occurrence. To examine this argument, I pair interviews with state personnel and new data on geolocated security installations with existing data on public employment and Syria’s 2011 protest movement. I show that high levels of public employment were associated with limited protest only in communities near security installations, where employees were more likely to be repressive staff especially unlikely to engage in or support local collective disobedience. Further from installations, high levels of public employment reflected civilian employment and were not associated with limited protest. The results point to a novel channel through which control of the repressive apparatus influences autocratic resilience to mass unrest and motivate further attention to when civilian employment generates compliance.

Why are some groups of individuals able to develop and sustain autonomous organizations in non-democratic settings? I point to an under-theorized resource: the networks individuals develop working within the autocratic state. By fostering high levels of localized participation and bottom-up collective action, clustered, dense networks can frustrate incumbent attempts to undermine mobilization and organization via top-down manipulation. I field and analyze a networks elicitation survey to formalize how workplace networks differ across otherwise similar state institutions in Jordan. I then draw on qualitative evidence to trace the foundational influence of workplace networks in allowing workers in one sector to sustain organizational autonomy while rendering their peers' organizations susceptible to co-optation and capture. This research underscores how even highly visible networks can sustain mobilization and organization while pointing to the constraints incumbents face when manipulating the state to entrench their power.