Keywords: immigration, public opinion, multi-level model, Latin America

How do community characteristics condition the influence of immigration flows on public opinion? I argue that Social Identity Theory can explain how communities might react differently to immigration flows, as those with closer perceived ties to the immigrants—operationalized as migrant-sending status and racial composition—should view immigration more positively. I test my argument using Colombia in response to the 2016 Venezuelan immigration crisis. Colombia is a traditionally migrant-sending nation in the Global South that 1) exhibits substantial variation in the racial composition and migrant-sending status of its municipalities and 2) currently hosts nearly 3 million Venezuelan immigrants. After aggregating a wide range of opinion data (i.e., the Gallup World Poll, LAPOP AmericasBarometer, Latinobarometro, and World Values Survey), I employ Multilevel Regression and Poststratification models to estimate Colombian immigration attitudes across time and geographies. I then assess my hypotheses--that municipalities perceiving Venezuelan immigrants as more similar to themselves will hold more inclusive immigration attitudes--using fixed effect models. I find that perceptions of similarity improve subnational immigration attitudes, as communities 1) sending more migrants during Colombia’s internal conflict and 2) with similar racial compositions to Venezuelan immigrant flows are more welcoming. By assessing how perceptions of similarity shape immigration attitudes, this paper contributes to literature on intergroup relations and immigration attitudes, suggesting that efforts to foster the recognition of shared identities between host communities and immigrant groups may reduce xenophobia.