In post-mortems on the 2016 presidential election, one of the leading headlines was that more white women voted for Donald Trump than for Hillary Clinton, which struck many pundits as a puzzle.Footnote ¹ For one, the conventional wisdom is that the gender gap favors Democrats. Even more importantly, this was the first time white women had an opportunity to vote for a descriptive representative for President, and one who made a point of priming gender. She was also running against a candidate with a history of misogyny, and this was on full display during the election, most notably with the release of the Access Hollywood tape. Given the salience of gender in the election, it was therefore a surprise to many that Trump received more votes than Clinton from white women.

We might question though whether this is really much of a puzzle. For one, the empirical reality, as Junn points out (Reference Junn2017), is that white women have been more likely to vote for Republican presidential candidates for decades. Therefore, the gender gap in presidential voting is really driven by greater Democratic support among women of color. Part of the problem then with the narrative of surprise is that it is based on problematic assumptions. For one, it assumes that gender is the only salient identity for female voters. Second, it assumes that all women have the same preferences, weight issues in the same way, and weight the importance of having a descriptive representative in the same way.

Both Junn and Phillips argue that taking an intersectional approach helps us understand some of this “puzzle”. They focus on the intersection of race and gender, and argue that one's position in the racial hierarchy is likely to affect group identity adoption. Given that white women are higher in the racial hierarchy, and often play a role in protecting white men at the apex of that hierarchy (Junn Reference Junn2017), they may perceive “a greater distance between the interests of their group identities than women of color do” (Phillips Reference Phillips2018, this volume). That is, white women may face cross-pressure given that these two identities may not always move in concert, and gender may come second to race in relative weighting. It is therefore not surprising, but expected, that Phillips finds white women weighted descriptive representation less in their voting decisions compared with women of color.

Another important identity that was almost entirely absent in media coverage of the surprise narrative is partisanship, where the Republican Party has an advantage among white women.Footnote ² Using data from the 2016 ANES, if we break down voting patterns among white women even further by partisanship, we see that 86% of white female Democrats supported Clinton while 96% of white female Republicans supported Trump. These voting patterns are therefore a testament to the power of partisanship, and stronger pull of partisan identity among Republicans (Theodoridis Reference Theodoridis2017).

We are at one of the highest points of partisan polarization in U.S. history, and while scholars debate the type and extent of polarization, two aspects seem relevant here. For one, voters are better sorted today, which means that issue preferences, ideology, and party identification are better aligned than they have been historically (Levendusky Reference Levendusky2009), and racial attitudes are an important part of the current partisan divide (Bafumi and Shapiro Reference Bafumi and Shapiro2009; Carmines and Stimson Reference Carmines and Stimson1989). Second, there is an affective, or emotional, element to polarization such that individuals have come to hold more negative views of the other side (e.g., Huddy, Mason and Aaroe Reference Huddy, Mason and Aaroe2015; Iyengar, Sood and Lelkes Reference Iyengar, Sood and Lelkes2002); and, Clinton, in particular, has been loathed by the right for decades.Footnote ³ Against this backdrop, we should therefore not expect white female Republicans to vote for a Democratic female they dislike, who would not act in their policy interest. In short, we should not be surprised that partisanship would trump (pardon the pun) gender in this election cycle, as it has been found to do in other work (e.g., Dolan Reference Dolan2014; Hayes Reference Hayes2011).

In sum, if we factors in these three identities that were highly salient in 2016, it is less of a puzzle as to why Trump won majority support from white women. Given the advantage that Republicans enjoy among white females and the salience of racial threat, only the salience of gender may have counteracted support for Trump (for more on identity cross-pressures, see work by Klar Reference Klar2013). If anything, the fact that fewer white women supported the Republican candidate in this election compared with 2012 is consistent with the complex interplay of identities, and the greater salience of gender this election cycle. In an era of high partisan polarization, it may simply be very unlikely that individuals will cross party lines to vote for a descriptive representative. Rather, we may be more likely to see individuals weighting such concerns in elections where party is removed from the equation, such as primary elections.