Content area
Full Text
1 Introduction
The 2016 election was the first in which a woman ran for president as the nominee of one of the two major parties while also featuring a Republican nominee who frequently made sexist remarks during his run for office. Trump’s sexism was an especially salient feature of the closing stages of the general election campaign. Trump frequently directed sexist rhetoric at Democratic nominee Hillary Clinton, accusing her of “playing the woman card” and referring to her as a “nasty woman.” A litany of studies conducted since the election find that sexist attitudes were a strong predictor of voting for Trump over Clinton in 2016 (Schaffner, MacWilliams, and Nteta 2018; Setzler and Yanus 2018; Valentino, Wayne, and Oceno 2018; Cassese and Barnes 2019; Stewart, Clarke, and Borges 2019). And political scientists have continued to probe the influence of sexism on American political behavior since the 2016 election (Costa et al. 2020; Schaffner 2020b).
While empirical research in American political behavior has paid substantially more attention to the role of sexist attitudes since the 2016 election, there is significant variation in how these attitudes are measured. In this paper, I evaluate the measurement properties of items comprising the Ambivalent Sexism Inventory (which includes the hostile sexism and benevolent sexism scales) and the modern sexism scale. I evaluate these items on three main criteria—(1) convergent validity, (2) predictive validity, and (3) proximity to politics. I find that a subset of items from the hostile sexism battery have especially desirable qualities with regard to convergent and predictive validity and are also viewed by people as being relatively distant from politics. As such, these items provide a particularly useful way of measuring sexism with statements that are mostly detached from politics, thereby reducing concerns about endogeneity. I conclude by showing that a reduced scale constructed from as few as two of these items maintains similar levels of predictive validity as the full scales, leading to my recommendation that scholars (and the flagship political science surveys) use these items to efficiently and validly measure an individual’s prejudice towards women.
2 Political Science Research on Sexism
The 2016 election clearly influenced the focus of American politics scholarship. This can be seen from a keyword search for the stem “sexis*” among...