Hostile Sexism and the 2016 Presidential Election

Abstract

We use Google Trends data over the 2004-2015 period to identify hostile sexism and examine its effect on support for Hillary Clinton in the 2016 general election. An area’s sexist search volume is a significant negative predictor of Clinton’s two-party vote share. Although we find no evidence that hostile sexism was more prevalent among conservative and less educated whites prior to the election, we do find evidence that it had a larger impact in areas with this demographic. We argue that demographic groups targeted by Trump may have been more receptive to his rhetoric. Our main contribution to the literature is showing that sexism had a politically consequential effect. We calculate that sexism cost Clinton 2.6 percentage points of her two-party vote share. In state-level simulations that are made possible by our use of Google Trends data, we show that Clinton would have won an additional 190 electoral college votes if every state had the same level of sexism as the least sexist state.

Keywords: sexism, elections, Google Trends

JEL Classification: D72, J71