We investigate the role of evidence-based information in shaping individuals’preferences for trade policies through a series of survey experiments that contain randomized information treatments. Each treatment provides a concise statement of economics research findings on how openness to trade has affected labor market outcomes or goods prices. Across annual surveys from 2018-2022, each administered to a representative sample of the U.S. general population, we find that information influences trade policy preferences in complex ways. Information highlighting the link between trade and manufacturing job losses significantly raises expressed preferences for more limits on trade. Strikingly, information on the price benefits of trade (or the cost of tariffs) also induces protectionist policy choices, indicating that these preferences do not respond symmetrically to information on the gains versus losses from trade. We find evidence that these expressed preferences are driven in part by how the received information interacts with one's political identity, resulting in prior-biased belief updating, as well as by pre-existingconcerns over the impact on American jobs and over trade with China. Information that solely communicates the benefits of trade is thus unlikely to succeed unless it addresses these prior beliefs and concerns.