Political polarization divides the mass public and contributes to a significant amount of social conflict within society. Those that identify with a social group (e.g. a specific political party or social class) tend to prefer policies that safeguard their own group, relative to others. As polarization worsens, it is important to understand how multiple group attachments motivate human behavior. This dissertation examines the consequences of affective polarization in the United States and Latin America. First, although previous literature documents that partisan cues prompt inter-party trust discrimination, it is unclear if these partisan effects are driven by economic divisions. In the first chapter, I disentangle the effects of partisan discrimination from class discrimination, and argues that class identity is tied to partisan identity. The results from an original trust game experiment indicate a large income-based discrimination effect. I also find evidence of cross-partisan envy -- an effect that has not yet been directly tested in the literature.

Second, I examine patterns and consequences of affective polarization in Latin America. While a significant amount of literature documents partisan dislike in the United States and Western Europe, less is known the extent to which partisan animosity exists in a political environment where partisan attachments are traditionally weak. Following recent work on negative and anti-partisanship, I argue that the mass public in Latin-American do hold political identities and that these identities have important political outcomes. Using CSES data, I show that that dislike for mainstream parties increases voter-turnout among nonpartisans.

Third, I use an original conjoint experiment and assess the limits of partisan prejudice. Current literature on partisanship argues that partisans use political identities as short-cuts and heuristics for decision-making. Following the long standing literature on partisanship, I examine the limits of partisan discrimination in a purely apolitical and altruistic domain: charities. My results demonstrate that partisanship not only influences economic and political domains, but that it spillsover into ostensibly apolitical domains such as deciding whether or not to donate to the cancer and research based nonprofits.