Although racial identity is usually assumed to be unchanging, recent research shows otherwise. The role of politics in racial identification change has received little attention. Using panel data with waves around the last two presidential elections, this paper reveals survey evidence of race change and its strong relationship with vote switching patterns. Across several models and robust to various controls, switching from a non-Republican vote in 2012 to a 2016 Republican vote (i.e., non-Romney to Trump) significantly predicts nonwhite to white race change. Among nonwhites who did not vote Republican in 2012, switching to a Republican vote in 2016 increases the probability of adopting a white racial identity from a 0.03 baseline to 0.48 (1490% increase). The systematic relationship is not due to measurement error, does not appear for the 2008-2012 election period, and makes theoretical sense in light of 2016 campaign rhetoric and trends in political-social identity alignment.
The primary data we use comes from the Democracy Fund's Voter Study Group (VSG)
longitudinal dataset. The survey firm YouGov conducted this online panel containing survey interviews of the same 8,000 Americans in December 2011, November 2012, and December 2016, weighted to general population benchmarks. The VSG is the only known publicly-available panel dataset with (1) waves entirely before and after the 2016 election and (2) contemporaneous measures of vote and race around the last two elections, and thus provides the data best suited to testing the hypotheses. The independent variable, vote switching, draws on post-election measures of vote choice in 2012 and 2016. The dependent variable is change in racial identity from 2011 to 2016.
Race change in the direction of adopting a white identity in 2016 occurs largely among people who identified as "Other," "Mixed," "Hispanic," or "Native American" in 2011, while Blacks and Asians hardly alter their identities. This pattern of adopting a white identity is consistent with findings from racial fluidity literature.
Changes in racial identity and vote choice are correlated. Of the nonwhite voters who
switched from a non-Republican vote in 2012 to a Republican vote in 2016, 9.84% changed their reported race to white. Among nonwhite voters who did not switch their vote, only 4.23% changed their reported race to white. Changes in vote from non-Republican to Republican and changes in race from nonwhite to white are related beyond chance (χ2=6.37, df=1, p = 0.01). However, changes in vote from non-Democrat to Democrat and changes in race from white to nonwhite are not related (χ2=.46, df=1, p = 0.50).
race-vote-change.pdf
drive.google.com