Most Western liberal democracies have a complicated history of immigration, marked by significant policy shifts from laissez-faire approaches to national origin quotas and outright exclusion. The migration literature offers numerous theories to explain these changes, but few consider lawmakers' agency in influencing these outcomes. Instead, scholars often view policymakers as passive adopters of interest group preferences or as reactors to economic and international pressures. My study addresses this gap by examining political elites' own words in legislative debates on the treatment of immigrant groups. I ask: to what extent has prejudice influenced immigration policy over the past century and a half? To answer this, I analyze 124 years of U.S. Congressional debates on immigration, from the 1882 Chinese Exclusion Act to the 2006 Secure Fence Act, using a systematic coding scheme to measure individuals' levels of prejudice and their support for immigration restrictions. Through statistical analysis using a fixed- effects estimation strategy, I find support for the theory, based on insights from cognitive and social psychology, that prejudice significantly influences decision-making about immigrants. Specifically, prejudice leads to votes favoring exclusion. Additionally, I identify and test the causes of these attitudes, finding that elites' prejudice is influenced by the size of the immigrant population in their constituency, which, in turn, affects their voting behavior. These relationships hold despite changes in the normative, socio-political, and demographic landscape over time, though there is some attenuation in more recent debates. This dissertation concludes with a discussion of limitations that might explain the varying consistency of results and offers several suggestions for further research.