Why do some dissident groups employ violent tactics, while others prefer nonviolence to pursue similar goals, and why do they switch tactics after embracing one? To address these questions, I analyze a formal model of tactical choice by the leadership of a dissident group and endogenous mobilization of citizens. I show that violent and nonviolent tactics are chosen in very different circumstances. Nonviolence is relatively more effective and is more likely when citizens have strong incentives to mobilize due to high grievances but expec! t to face low repression. Violence is relatively more effective when citizens face weak incentives to mobilize, either due to low grievances or high repression. Therefore, the conditions under which nonviolent tactics are preferred are also the conditions under which both tactics are more effective. Further, groups can change focus from nonviolence to violence in the face of increased repression, and counterintuitively, they can also shift from violence to nonviolence as grievances are exacerbated.