One hundred eleven college students participated in this two-session experiment designed to explore the origins of anti-nuclear war activism. In Session 1 participants completed questionnaires assessing a number of background characteristics and attitudes, the most important of which for this study's purpose was a measure of perceived political efficacy in the nuclear realm. One week later both the salience of the nuclear weapons issue and participants' sense of personalized risk were experimentally manipulated in a 2 × 2 × 2 (salience × risk × measured efficacy) design. Dependent measures were a behavioral intentions questionnaire and an actual opportunity to sign a petition which participants did not recognize as part of the experiment. High salience significantly increased both anti-nuclear war behavioral intentions and actual behavior as predicted. Individuals who felt highly efficacious were also significantly more likely to take action than others as predicted. A similar relation between efficacy and behavioral intentions was not found, although a salience by efficacy interaction was. Personalized risk influenced only behavioral intentions. Finally, the relative impact of these three variables and of other background and attitudinal variables measured in the first session was explored using regression techniques.