It is by now well known that political attitudes can be affected by emotions. Most earlier studies have focused on emotions generated by some political event (e.g., terrorism or increased immigration). However, the methods used in previous efforts have made it difficult to untangle the various causal pathways that might link emotions to political beliefs. In contrast, we focus on emotions incidental (i.e., irrelevant) to the decision process, allowing us to cleanly trace and estimate the effect of experimentally induced anxiety on political beliefs. Further, we build upon innovative new work that links physiological reactivity (Oxley et al., 2008a; Hatemi, McDermott, Eaves, Kendler, & Neale, 2013) to attitudes by using skin conductance reactivity as a measure of emotional arousal. We found that anxiety—generated by a video stimulus—significantly affected physiological arousal as measured by tonic skin‐conductance levels, and that higher physiological reactivity predicted more anti‐immigration attitudes. We show that physiological reactivity mediated the relationship between anxiety and political attitudes.