Why do some protest movements escalate from peaceful demonstrations to violent confrontation while others never change tactics? I develop a dynamic model in which an organizer chooses between a turnout-sensitive tactic (peaceful mass mobilization) and a tactic with higher baseline power but less turnout scalability (violent or confrontational action). Turnout is endogenous: citizens decide whether to participate based on public beliefs about regime strength, and failed protests shift those beliefs pessimistically. The organizer balances immediate effectiveness against the informational consequences of failure. Three equilibrium regimes emerge. When violence lacks meaningful baseline capacity, peace is optimal throughout. At intermediate values, peace is used first but movements escalate after weak, uninformative failures. When violence is both powerful at baseline and highly scalable, the organizer leads with it, and the continuation is path-dependent. I map these regimes to features of the political and physical environment, generating predictions about tactical persistence and escalation and a structural account of the observed correlation between nonviolent tactics and campaign success. On the technical side, the model also offers a resolution to equilibrium multiplicity in dynamic global games by having agents receive private cost rather than information signals.

Negotiating factional conflict is crucial to successful policymaking: coalition governments, political parties, and authoritarian elites must all overcome internal disagreements in order to move forward. Actors in such conflicts sometimes employ hardball tactics to strategically rule out outcomes they dislike. Using a dynamic bargaining model, I explore how the threat and usage of these tactics impact coordination between actors with conflicting interests. In the model, two players who prefer different reforms must jointly agree on only one in order to overturn a mutually unfavorable status quo. Neither knows for certain whether the opponent prefers the status quo over their less-preferred outcome. Players willing to compromise on their opponent's preference rationally delay agreement, balancing the incentive to preempt the opponent against the benefit of waiting to gather better information. Delay is prolonged when actors cannot easily glean one another's willingness to compromise. One such factor is the frequency with which private willingness to compromise is publicly revealed. Thus, higher-leak environments are beneficial to welfare, as the additional delay incentivized by leaks deters mistakes of preemption.

An influential body of scholarship argues that authoritarian regimes design "hard" propaganda that is intentionally heavy-handed in order to signal regime power. In this study, by contrast, we link the power of propaganda to the emotional power of “soft” propaganda such as television dramas and viral social media content. We conduct a series of experiments in which we expose over 6,800 respondents in China to real propaganda videos drawn from television dramas, state-backed social media accounts, and state-run newscasts, each containing nationalist messages favored by the Chinese Communist Party. In contrast to theories that propaganda is unpersuasive, we show that propaganda effectively manipulates anger as well as anti-foreign sentiment and behavior, with heightened anti-foreign attitudes persisting up to a week. However, we also find that nationalist propaganda has no effect on perceptions of Chinese government performance or on self-reported willingness to protest against the state.

How do citizens react to authoritarian responsiveness? To investigate this question, we study how Chinese citizens reacted to a novel government initiative which enabled social media users to publicly post requests for COVID-related medical assistance. To understand the effect of this initiative on public perceptions of government effectiveness, we employ a two-part empirical strategy. First, we conduct a survey experiment in which we directly expose subjects to real help-seeking posts, in which we find that viewing posts did not improve subjects' ratings of government effectiveness, and in some cases worsened them. Second, we analyze over 10,000 real-world Weibo posts to understand the political orientation of the discourse around help-seekers. We find that negative and politically critical posts far outweighed positive and laudatory posts, complementing our survey experiment results. To contextualize our results, we develop a theoretic framework to understand the effects of different types of responsiveness on citizens' political attitudes. We suggest that citizens' negative reactions in this case were primarily influenced by public demands for help, which illuminated existing problems and failures of governance.