How do individuals’ policy preferences emerge and change? We synthesize political economy and behavioral approaches to produce a framework that explains how people change their policy preferences when there is uncertainty about the distributive effects of public policies. Our theory proposes a process where people learn from direct experience about how they are affected by policy issues, which leads them to update their preferred response. Climate change serves as a case to test the theory. We use an econometric model of global warming to derive what individuals’ policy preferences over action to reduce climate damages might be if they were fully informed about global warming’s effects and acting according to self-interest. Then, we leverage geospatial data on climate disturbances to capture experiential shocks. Separate analyses using subregional and panel survey data find that climate shocks cause individuals to become more supportive of action to address climate change in line with how they might be materially affected by global warming. Personal experience that leads to learning about who wins and loses from public policies helps to explain when changing beliefs cause shifts in policy attitudes.