The appropriateness of experiments for studying causal mechanisms is well established. However, the ability of an experiment to isolate the effect of emotion has received less attention, and in this letter we lay out a guide to manipulating and tracing the impact of emotions. Some experimental manipulations are straightforward. Manipulating an emotion like anxiety is less obvious. There is no magic “political anxiety pill” and placebo that can be randomly assigned to participants. While the magic political anxiety pill is still elusive, we advocate using multiple manipulations, extensive pretesting, and mediation models. These approaches have allowed us to situate a discrete emotional experience in a complex political environment.