ABSTRACT Despite the well-known theoretical and empirical connection between surveillance and individual behavior, research on political participation in the United States rarely incorporates government surveillance into individual-level empirical models. Using two experiments embedded in a nationally representative survey, this study assesses how the degree and target of conflict associated with political activities influences perceptions of online government surveillance. The principal results suggest that large segments of the public believe that the government likely monitors a range of political activities, although not all forms of political behavior are equally vulnerable. Specifically, political activities with links to violence and legitimate activities in opposition to the presidential administration most influence online surveillance perceptions. The results also demonstrate that approval of the president moderates these perceptions. Nonviolent political activities in conflict with the presidential administration increase online surveillance perceptions for disapprovers of the president but not for approvers. More than anything else, our findings demand the incorporation of government surveillance into empirical studies of U.S. political participation.