As the parties’ nomination battles were getting under way, many observers of American politics were expecting the 2016 election to be defined by gender. On the Democratic side, it looked to be the first election in US history in which a major political party would nominate a female candidate. On the Republican side, the party was to nominate a hyper-masculine candidate with a penchant for offending women. This particular pairing of candidates thus seemed likely to magnify the tendency of men to vote Republican and women to vote Democratic. As one Republican pollster predicted in January 2016, “We will have a gender gap that looks like the Grand Canyon if we have Trump vs. Hillary.” 1
The gender gap has become a regular part of elections in the US. Its importance arises from the fact that men and women represent large and approximately equal voting blocs. Even small shifts in the vote choices of one sex could easily alter an election outcome. At the same time, the overall gender gap is not as large as some other demographic cleavages. Voting patterns diverge more dramatically along racial and ethnic lines, between the married and unmarried, among those with differing religious attachments, and between urban and rural residents. Nonetheless, conditions in 2016 seemed to facilitate a historic divide along gender lines that might have rivaled these other social cleavages.
The “Grand Canyon” of gender gaps did not materialize in 2016. To explain this non-event, we contend that the gender gap was muted by two factors – one specific to the 2016 context and one more enduring. First, both Clinton and Trump had long histories in the public eye and were familiar figures for many voters. This limited the ability of campaigns or events to reshape public attitudes toward them along gender lines. Second, and more importantly, affective partisanship has become sufficiently important in determining voters’ choices that it leaves little room for demographic or attitudinal divides to be important influences on the vote beyond baseline party loyalties.
To be sure, gender mattered tremendously in the 2016 presidential election, but not in the ways that most people expected. We find that the gender gap was not far out of line with recent presidential elections and that it did not work to Clinton’s advantage. Moreover, based on previous elections, it is not clear that a larger gap would have worked to the Democrats’ advantage. Contrary to the idea that Clinton would win in 2016 by drawing female voters away from Trump, the gender gap was often smallest at times when men were most favorable toward her candidacy, thus boosting her overall standing in the polls. This highlights a second surprising result: it was men rather than women who vacillated most over the course of the campaign. Although both men and women wavered on whether they viewed Clinton as honest, men’s vote intentions were more variable, especially in response to two events: the “Access Hollywood” recording of Trump talking about assaulting women and a letter from FBI Director James Comey about reopening the investigation of Clinton’s email messages.
Expectations Versus Reality of the Gender Gap in 2016
Numerous events during both the nomination campaigns and the general election campaign suggested that a big gender divide would develop in the electorate, in large part because of Donald Trump’s troubling interactions with women. This possibility first became salient in the first Republican debate back in August 2015. In a highly anticipated event that attracted a record 24 million viewers, moderator Megyn Kelly of Fox News addressed a provocative question to Trump. She admitted that while people often love Trump’s lack of a “politician’s filter,” his straight talk “is not without its downsides, in particular, when it comes to women. You’ve called women you don’t like ‘fat,’ ‘pigs,’ ‘dogs,’ ‘slobs,’ and ‘disgusting animals.’” Trump dismissed the criticism as a problem of political correctness and moved on to other topics. Yet this provocative interaction led to an ongoing conflict between Kelly and Trump that lasted for months. Shortly after the event, he stated in a television interview that Kelly was angry and treated him unfairly. As he put it, during the debate, “…you could see there was blood coming out of her eyes, blood coming out of her wherever.” This crude comment, apparently intended to dismiss Kelly’s question as a mere product of her menstrual cycle, set the stage for a campaign chock full of gendered discourse.
Harsh comments about women surfaced regularly as the campaign unfolded over the next year and a half. In reference to fellow Republican contender Carly Fiorina, Trump said in a magazine interview in September, “Look at that face. Would anyone vote for that?...I mean, she’s a woman, and I’m not supposed to say bad things, but really, folks, come on.” 2 In a town hall-style interview before the Wisconsin primary, Trump was asked how the law should treat women who get an abortion when it is not permitted. After some hesitation, he stated that “There has to be some form of punishment” for the woman involved, a comment that was criticized even by fellow anti-abortion Republicans. 3
In a primary night victory speech the following month he suggested that Clinton was only competitive with him because of her sex. As he put it, “Frankly, if Hillary Clinton were a man, I don’t think she’d get five percent of the vote. The only thing she’s got going for her is the woman’s card.” 4 Clinton responded to this jab by issuing a humorous “woman card” in exchange for donations to her campaign. The cards sent to donors stated at the bottom “Congratulations! You’re in the majority.” Rather than avoid gender questions, Clinton hoped to capitalize on her sex and win over female voters who were repelled by Trump.
Perhaps no incident was as central as a leaked video of Trump as he was preparing to appear on the television program “Access Hollywood.” The widely-seen recording from 2005 featured Trump bragging about his efforts to have sex with a married woman and his tendency to pursue women physically without their permission. As his suggested to the show’s host, “Just kiss. I don’t even wait. And when you’re a star, they let you do it. You can do anything…Grab ‘em by the pussy.” Trump later dismissed the comments as “locker room banter” and suggested that former president Bill Clinton had acted similarly. Nonetheless, the tape remained a news item for days. It even spurred allegations from a dozen women alleging Trump had sexually harassed or assaulted them over the years. If seemed likely that this repeated sexually offensive behavior would drive even more women away from the Republican candidate and create the Grand Canyon of gender gaps.
The reality of the 2016 election did not match these expectations. Although men and women continued to vote for different parties on average, exit polls showed that 54% of women and 41% of men voted for Clinton, a gender gap of 13 points. As Figure 1 shows, this disparity between the sexes is larger than gaps observed in previous elections, by not by much. It is only three points larger than the gap in 2012 and just two points larger than it was in 2000. 5 Rather than a dramatic break with prior elections resulting from the Clinton-Trump face-off, 2016 represents the continuation of a gender divide that has slowly expanded in recent decades.
Does the Gender Gap Help Democrats Anyway?
Although the gender gap was a bit larger in 2016, it did not appear to help Hillary Clinton. The literal and figurative “gender card” she played did not win her a larger share of female voters or the presidency. She won smaller shares of the popular and electoral votes than fellow Democrat Barack Obama did despite the slightly larger gender gap in 2016. This raises questions about a popular notion that Democrats somehow benefit from a larger gender divide. If this connection is real, it would suggest that a larger gender gap is mostly the result of women moving to the Democratic candidate without a corresponding loss to the Republican candidate among men.
Figure 2 shows the relationship between the gender gap and the Democratic vote in the last 11 presidential elections. In the figure the gender gap is again based on exit polls and is computed as the percentage of women who voted for the Democratic candidate minus the percentage of men who voted for the Democratic candidate. The Democratic vote is measured as a share of the two-party popular vote. As the regression line running through the plot indicates, there is not much connection between the two variables. The correlation is small (r=0.15) and not statistically significant (p=0.66). Removing the 1984 outlier – the only election in the series where the popular vote favored the winner by more than 10 percentage points – makes the relationship even flatter (r=−0.07, p=0.84).
Burden and DeCrescenzo (2016) provide more evidence on this point. Using a longer time series based on American National Election Studies data since 1952 and cross-sectional comparisons across the states using exit polls from 2004 and 2008, they show that the gender gap and the election outcome are not necessarily related to one another. Why might this be the case? The gender gap conveniently summarizes the differences in the vote choices of men and women, but Burden and DeCrescenzo argue that the size of the gender gap does not convey enough information to determine which party gains more votes. Rather, it is crucial to look at the relative turnout rates of men and women in addition to their vote choices to determine whether the Democrats gain more votes from women than the Republicans gain from men.
Voting Intentions of Men and Women Over the Campaign
Although there does not appear to be much of a connection between the magnitude of the gender gap and partisan outcomes across elections, it is possible that the unique circumstances of the Clinton-Trump pairing created different relationships within the 2016 campaign. The 2016 campaign was nothing if not eventful, driven as it was by Trump’s frequent tweets and outlandish statements at campaign rallies plus release of hacked emails released by WikiLeaks. These developments could have affected voters’ preferences, and perhaps differentially for men and women. It often seemed that Clinton surged in the polls just after Trump said something offensive but suffered when explaining her private email server or the hacked emails of other Democratic Party actors and officials.
To examine this possibility, we collected data from every public survey conducted during the campaign where the pollster reported vote intentions separately for men and women. We limited the data to polls conducted with “likely voters” within the final 200 days of the campaign and where respondents were asked how they planned to vote in the two-way contest between Clinton and Trump. 6
Figure 3 shows this time series. The vertical axis marks the margin by which each poll favors Clinton, that is, the difference between the percent saying they will vote for Clinton minus the percent saying they will vote for Trump. The figure also highlights several important milestones. These include traditional campaign events: the two major party conventions in July and the three presidential debates. But we also note Clinton’s September comment that half of Trump’s supporters belonged in a “basket of deplorables” and her pneumonia health scare the following day in which she was removed from a 9/11 memorial event and subsequently took leave from the campaign trail for several days. The figure also indicates the release of the “Access Hollywood” tape (just before the second debate), and appearance of a letter from FBI Director James Comey to congressional leaders indicating that the agency was reopening its investigation into email messages on Clinton’s private server.
The figure shows that the familiar gender gap held through most of the campaign. Women favored Clinton in almost every poll, often by double-digit margins. By contrast, men favored Trump in most polls, albeit often by smaller margins.
The absence of dynamics is surprising. This might reflect the public’s high level of familiarity with both Clinton and Trump based on their visibility in the public sphere even before their campaigns began. Despite the many events that transpired, women hardly wavered in their support, averaging about 10 points in favor of Clinton from the conventions until election day. The combination of Clinton’s “basket of deplorables” comment and her pneumonia spell in September, the Access Hollywood tape and the accusations that followed, and even the Comey letter all had little impact on women’s vote intentions. In contrast, men moved decidedly toward Trump during Clinton’s difficult week in September. To be fair, Clinton’s standing among men appeared to have taken a hit even before her bad week in September, perhaps in response to emails released by WikiLeaks in August that pointed to special treatment of Clinton Foundation donors when Hillary Clinton served as Secretary of State.
But men gradually moved back toward Clinton as the three presidential debates took place from late September until mid-October. Men actually crossed the threshold to favor Clinton on net in some of the polls immediately after the “Access Hollywood” tape and the second debate. It is not until after the Comey letter about Clinton’s email controversy that men broke back to Trump, apparently by just enough to hand him the election. It is interesting to entertain a counterfactual scenario in which the timing of the Comey letter and the “Access Hollywood” video were reversed. Presumably men would have shifted away from Clinton in the wake of the letter’s release but doubled back to her after the tape emerged. That dynamic among men would have likely resulted in Clinton winning the Electoral College. 7
There is one more surprise in Figure 3. A closer look shows that Trump was not harmed by the gender gap. Clinton actually performed better when the gender gap was smaller, not when it was larger. Because her level of support among women was relatively constant, Clinton’s high and low points in the polls turned primarily on the vote intentions of men. Clinton rode highest in mid-October when men moved in her direction following the tape. Then, after the Comey letter, men’s movement away from Clinton caused her to fall in the polls. As a result, the eventual gender gap observed on Election Day was larger, but her vote share was smaller. In addition, contrary to the popular idea that women are more persuadable and thus determine the size of the gender gap, in 2016 it appeared that men’s vote choices were more volatile and consequential, with a standard deviation of 6.6 compared to 5.7 for women.
Two Unpopular Nominees
It was widely noted that Hillary Clinton and Donald Trump were the two most unpopular major party presidential candidates since at least 1984. According to the Real Clear Politics polling aggregator, Clinton’s final average unfavorable rating was 54.4% compared to Donald Trump’s 58.5%. To put that in perspective, the highest unfavorable rating of any major party nominee recorded by the Washington Post-ABC Poll was 53% for George H.W. Bush in March of 1992. 8
Although Clinton and Trump might have been the least liked nominees in modern history, they were not universally disliked. Their poor ratings were driven mostly by independents and people who identify with the opposing party. By mid-summer 2016, only about one in four respondents had an unfavorable view of both of them. 9
Recent research attributes part of the rise in polarization to an increase in what has been called “affective” or “negative” partisanship (Iyengar, Sood, and Lelkes 2012; Nicholson 2012; Iyengar and Westwood 2015; Abramowitz and Webster 2016). Under this theoretical view, disagreement between Democrats and Republicans stems not just from policy differences, but also from merely identifying with the opposing party. This is part of a decades-long trend in which partisans are no more favorable to their own party but are much more negative in evaluating the other party (Abramowitz and Webster 2016). In the short-term, people do not favor policies more just because they are endorsed by the presidential candidate of their party, but they do disfavor policies more when they are endorsed by the opposing candidate (Nicholson 2012). Distaste for the opposing party also grows with exposure to presidential campaigns (Iyengar, Sood, and Lelkes 2012). This is because negative campaigning facilitates voter confirmation of partisan stereotypes of the opposition.
Why did Republicans who had opposed Trump’s candidacy during the primaries, were offended by his statements during the general election campaign, or found themselves out-of-step with his policies – a group that is disproportionately female – largely vote for him in the end? This apparent inconsistency is less mysterious when the question is turned around. The distaste for the opposing party is at a modern high and partisans are predisposed to hold negative stereotypes of the opposition. The 2016 presidential campaign featured speeches, tweets, ads, and other communications reminding supporters about the negative traits of the opposition – a reinforcement of partisan stereotypes. Trump’s character and rhetoric spawned questions about why Republicans would stick with him, even if they were “party loyalists.” Affective partisanship suggests that party loyalists are not so named simply because of affection for a party’s nominee but also because of a highly negative view of the opposition, namely, Hillary Clinton.
The heightened tendency for voters to dislike candidates of the opposing party coupled with negative views among independents suggested that both campaigns lacked positive support and many voters were instead voting in a negative fashion to elect the “lesser of two evils.” As Trump’s campaign created controversies and invited disavowal from high-profile Republican officials, it was conceivable that Hillary Clinton would be elected president despite her high unfavorability ratings, if only as a negative statement against Trump.
To what degree were voters in 2016 casting affirmative votes for a candidate they actually favored rather than voting defensively against an even less appealing candidate? We explore this using survey questions that asked people whether their votes were primarily for their candidate (“affirmative voting”) or against the opposing candidate (“negative voting”). This mix of motivations bears on our understanding of the 2016 gender gap in two ways. First, just as there were reasons to expect a historic gender gap, there was also reason to believe that an unprecedented share of the electorate would engage in negative voting. And second, the gender gap and negative voting were likely to be substantively linked. An unusually high percentage of women were expected to vote against Trump. It was also possible that men might have voted against Clinton rather than for Trump if they were reluctant about electing a woman as president.
In the final pre-election poll asking this question, the share of supporters expressing an affirmative vote for Clinton in 2016 was 62%. 10 This is on par with prior presidential candidates going back to 1980, with Obama’s two elections being notable exceptions. Trump supporters were not as enthusiastic for their own candidate, with 52% voting affirmatively for him. Unlike Clinton’s relatively average performance on this question, the level of affirmative voting among Trump supporters was close to a record low.
Although the anticipated “Grand Canyon” of a gender gap did not materialize in final vote choices, perhaps an examination of affirmative and negative voting among men and women separately will reveal the gap in a subtler way. Republican women who were repelled by Trump’s candidacy may have still voted for Trump but expressed their support as a “vote against Clinton” at a higher rate than their male co-partisans. At the same time, perhaps Democratic women rallied behind Clinton by affirmatively voting for her whereas Democratic men were more likely to express their support as a “vote against Trump.”
To examine this possible asymmetry, we identified 18 national polls conducted over the course of the 2016 election campaign that posed the affirmative voting question and reported the results by gender. 11 On average, Clinton outperformed Trump with respect to supporters voting affirmatively for their candidate. The only period of time when Clinton and Trump supporters were nearly equal in their affirmative support was after Clinton’s “deplorables” comment and prior to the first debate. Clinton regained the advantage in affirmative support after the first debate.
For both candidates, women were slightly more affirmative in their vote choice than were men. Of those supporting Clinton, women consistently tracked as voting a bit more affirmatively than their male counterparts, although the margin was not wide. In the final pre-election poll, 64% of women and 60% of men were voting affirmatively for Clinton compared to 34% of women and 37% of men who said they were voting against Trump.
The gender difference among Trump’s supporters were even smaller: 50% of men and 53% of women reported voting for Trump, with 49% of men and 46% of women voting against Clinton. These numbers provide little support for the notion that women voting for Trump would be more likely than men to cast negative votes against Clinton.
The four previous presidential elections showed similarly small gender differences in affirmative support for candidates. Trump’s 52% affirmative vote is lower than every candidate for whom we have data with one exception: John Kerry’s 44% in 2004, a year where Democrats were highly motivated to defeat George W. Bush. A stark difference between these two elections, however, is that Bush’s unpopularity among Democrats was countered by about 75% of Bush’s own supporters expressing affirmative support in 2004, compared to just 62% of Hillary Clinton’s backers doing the same.
What may be more telling is the relatively little movement in “negative voting” expressed by women in the final weeks of the campaign. In contrast, there appeared to be a small upward tick in negative voting among male Trump voters. This dynamic would complement what was suggested by the gender gap data on vote choice in Figure 3, where women remained relatively stable while men displayed greater variance and a late break towards Trump.
A Focus on Clinton’s Honesty
Both campaigns focused their attention on the opposing candidate’s character and fitness to lead. These strategies reflected public dislike of the candidates that was present from the outset of the campaign, but they also probably heightened negative views of the nominees. While the 2016 campaign saw a heavy amount of personal attacks, the campaigns’ messages did not merely offset one another. A coding of each campaign’s television advertisements by Lynn Vavreck found a lot of criticism of their opponents: three-quarters of Clinton ads and half of Trump’s were about the other’s character or traits. 12 Although character attacks are nothing new in presidential campaigns, the prevalence of these types of ads was much higher in 2016 than in the prior presidential campaigns. As the article by Fowler, Ridout, and Franz (2016) in this issue reports, overall TV advertising was not more negative in 2016 than in 2012. But this average disguises differences between the campaigns. Most of Clinton’s ads were focused on personal characteristics, especially Trump’s liabilities, whereas more than 70% of Trump’s ads were focused on policy (an imbalance also apparent in Vavreck’s coding). Clinton’s ads were remarkably more focused on personality than any major party candidate going back to at least 2000.
As an aside, we note that although Trump’s ads were not especially personal in nature, his campaign rhetoric often was. In tweets, at campaign rallies, and in debates he repeatedly derided Clinton’s email practices as criminal and accused her of giving big donors special favors. Trump tagged Clinton with the moniker “crooked Hillary,” and in their final debate called her a “nasty woman.” Although ads are important parts of campaign discourse, Trump’s statements – which were often quite personal and sometimes misogynistic – regularly made news and helped frame the choice between him and Clinton.
Yet survey data suggest that both advertising and rhetorical strategies had modest effects on how male and female voters perceived the candidates and their personal qualities. Over the course of the campaign, several polling organizations asked respondents to weigh in on Clinton’s and Trump’s general favorability, temperament, qualifications to be president, and their honesty and trustworthiness. As Figure 4 shows, despite the barrage of ads and news stories disproportionately focusing attention on personal characteristics and scandals, on these measures both men and women were remarkably consistent in their evaluations of each candidate between the conventions and Election Day.
Unfavorability ratings of Trump were approximately 70% among women and 55% among men. Unfavorability ratings of Clinton were lower and flipped, 60% among men and 50% among women. Although the precise levels varied across the more specific trait questions, this basic ordering among men and women generally replicated regardless of the characteristic being evaluated. The only exception was the question about honesty and trustworthiness. While both candidates were rated by a majority of voters as “not honest” for most of the campaign – with women holding more negative judgments about Trump’s honesty and men holding more negative judgments about Clinton’s honesty – the last few weeks of the campaign saw a sudden shift in voters’ responses on this question. Trump experienced a slight decrease in the percentage of both men and women who labeled him as “not honest” while perceptions of Clinton’s dishonesty increased. By the end of the campaign, women were only four points more likely to apply the term dishonest to Trump than Clinton, but men came to see Clinton as much more dishonest than Trump, by about 12 percentage points.
It would also appear that voters’ opinions of each candidate’s personal qualities were also remarkably stable over time. This is not surprising given how familiar the public was with both candidates before the campaign season began. Of the four questions we examined, the notable exception was again the “honesty and trustworthiness” item, which showed more volatility in voters’ responses than questions about other traits.
We previously noted that men wavered more than women in their intended vote choices, as indicated by their higher standard deviation. As Tables 1 and 2 show, standard deviations between men and women did not differ significantly for either candidate on attitudes about favorability, temperament, or qualifications. When looking at the percentage of respondents who identified Clinton as honest, however, women were more volatile, with a standard deviation of 5.70 compared to 3.60 for men. It is important to note that on this question, respondents also have the option to choose “don’t know” or “unsure.” It is therefore possible that among voters being moved over the course of campaign on the question of honesty, they likely oscillated between affirming a candidate as honest and having no opinion or between not honest and no opinion, rather than moving back and forth between rating a candidate as honest and not honest. While the variance around calling Clinton “honest” was greater for women than men, the opposite was true when the dependent variable was the percentage labeling Clinton “not honest.” In this case, the standard deviation for men was 5.24 compared to just 3.16 for women.
Standard Deviations of Candidate Evaluation: Clinton.
|Temperament – Yes||0.96||1.07||19|
|Temperament – No||3.40||2.83|
Bolded rows indicate differences between men and women that are statistically significant at p<0.01 (variance-ratio test, two-tailed).
Standard Deviations of Candidate Evaluation: Trump.
|Temperament – Yes||3.75||3.57||19|
|Temperament – No||3.57||3.92|
Following the movements observed in the campaign time series in Figure 3, we speculate that ongoing questions about Clinton’s health, her personal finances and those of the Clinton Foundation, and her email server – all punctuated by campaign events and Trump’s “crooked Hillary” framing – caused both male and female voters to revisit their views about her honesty. But the effects differed by gender in that men were trying to decide whether they viewed her as not honest or were merely unsure, whereas women were trying to decide whether they believed she was honest or were unsure. Despite the parade of scandalous revelations about Trump’s finances, legal battles surrounding Trump University and other business endeavors, and past and present treatment of women that Clinton highlighted, neither men nor women seemed to reconsider their views of his honesty as the campaign progressed.
The Gender Gap in the Era of Partisan Polarization
As we note above, although there was a widespread expectation of a “Grand Canyon”-sized gender gap in 2016 among journalists, pundits, and even political scientists, the gap between the expectations and reality for the gender gap may have been larger. While the gender gap in 2016 was 13 percentage points (according to exit polls), that is not far from the gender gaps observed in past elections. The 2016 gap might have technically set a record, but it was not exceptional compared to other recent elections, nor was it acutely activated by the events of the campaign that commenters believed would bring gender to the fore. Why, in a campaign with such atypical gender politics, was the gender gap in voting and candidate evaluations so unremarkable?
The scholarly literature on voting and gender in US elections offers important but underappreciated insights that help us bridge the gap between expectations and reality regarding the 2016 election. We highlight two major points here. First, because the gender gap in voting primarily reflects a gender gap in partisanship, and because the level of partisan voting is at unprecedented highs in recent years, there is little wiggle room for gender to affect the voting preferences of men and women after accounting for their partisan attachments. This point is demonstrated in Dolan’s (2014) extensive treatment of the issue. Second, although journalists and commenters believed that men and (especially) women voters would cast their votes in response to the explicit role of the politics of gender in 2016, past research on gender and public opinion finds that gender differences in issue attitudes fall more strongly along social welfare issues, while gender differences on so-called “women’s issues” and other matters of gender in politics are more muted.
In this current era of partisan polarization, voters’ partisan attachments play a central role in the foundation of the gender gap in voting. Kaufmann and Petrocik (1999) argue that the emergence of the gender gap in voting is primarily the result of gender differences in party identification. They find that while a majority of women have consistently identified as Democrats since the 1950s, the gender gap in voting tracks the gradual conversion of men from being mostly Democratic to near parity between Republican and Democratic identifiers from the 1960s to the 1980s. Further, a number of authors of various perspectives note the growing consistency between partisanship, individual issue beliefs, and liberal-conservative ideology in recent decades (Carsey and Layman 2006; Abramowitz and Saunders 2008; Bafumi and Shapiro 2009; Levendusky 2009). Between the growing influence of party identification and its increasing relationship to issues and ideology, as well as partisanship’s influence on candidate evaluations (Hayes 2005), there are few sources of leverage for parties and campaigns to exploit in encouraging voters to cast swing votes.
Partisanship aside, popular commentary on the gender gap in 2016 also overlooked what political scientists know about the issue areas where men and women disagree. Although politicos and journalists pay a great deal of attention to the politics of gender, gender in the electorate itself does not form as strong of a partisan cleavage as do social characteristics such as race, ethnicity, and religiosity. Despite the prominence of “women’s issues” such as abortion rights and birth control access, public opinion scholars have long observed men and women to have fairly similar attitudes on these issues on average. Instead, the “issue content” of the gender gap – where it exists at all – falls more strongly along social welfare policy, economics, and foreign policy, as well as the relative salience given by men and women to these issue areas (Chaney, Alvarez, and Nagler 1998; Manza and Brooks 1998; Kaufmann and Petrocik 1999; Box-Steffensmeier, De Boef, and Lin 2004; Kaufmann 2006).
What does this mean for 2016? If Hillary Clinton wanted to widen the gender gap by boosting her vote share among women, focusing on Trump’s character and treatment of women may have been a relatively ineffective messaging strategy. The politics of gender were made unavoidable by Trump’s comments about and to women, the scandalous “Access Hollywood” recording, and the accusations of Trumps’ prior sexual assaults that followed. These events appeared to have affected the course of the campaign trajectory, although not in a simplistic fashion. The developments did not affect all voters equally and the order in which they took place was crucial to the outcome on election day. Men were more influenced than women by these campaign revelations, but any advantage Clinton got as a result of them was wiped out by men’s return to Trump’s side in the final days of the campaign in an apparent response to the Comey letter.
Perhaps Clinton could have made more of Trump’s trouble with women, but the literature suggests that women might have responded more favorably to economic policy messages than to overt feminist appeals. Further, while some evidence shows that women desire less militaristic approaches to foreign policy than men do, Clinton’s foreign policy messages were unlikely to activate this difference of opinion. Although Clinton enjoyed an obvious advantage over Trump in foreign policy due to her experience as Secretary of State, her go-to example of her foreign policy judgment was her role in advocating for the mission that killed Osama bin Laden. As a result, persuadable voters (particularly women) may have perceived Clinton’s foreign policy perspective as more hawkish than dovish, undermining her natural advantage among women who prefer diplomacy before displays of military strength.
Gender was a central factor in the 2016 campaign. How could a contest between a hyper-masculine misogynist and the first female major party candidate not be? Yet expectations for a “Grand Canyon” sized gender gap were not met in large part because they were out of step with previous scholarship on the gender gap and scholarship on partisan voting behavior more broadly. The gender gap in vote choice from past elections already reflected gender differences in party preferences and in policy views, broadly speaking. And due to the recent growth in the constraint between partisanship and discrete issue preferences, neither partisanship nor voters’ issue preferences are manipulable in the way that journalistic conjectures often assume.
Abramowitz, Alan I., and Kyle L. Saunders. 2008. “Ideological Realignment in the U.S. Electorate.” Journal of Politics 60: 634–652.
Abramowitz, Alan I., and Steven Webster. 2016. “The Rise of Negative Partisanship and the Nationalization of U.S. Elections in the 21st Century.” Electoral Studies 41: 12–22.
Bafumi, Joseph, and Robert Y. Shapiro. 2009. “A New Partisan Voter.” Journal of Politics 71: 1–24.
Box-Steffensmeier, Janet M., Suzanna De Boef, and Tse-Min Lin. 2004. “The Dynamics of the Partisan Gender Gap.” American Political Science Review 98: 515–528.
Burden, Barry C., and Michael G. DeCrescenzo. 2016. “Mobilization, Persuasion, and the Partisan Fallout of the Gender Gap in U.S. Voting.” Paper presented at Toronto Political Behavior Workshop.
Carsey, Thomas M., and Geoffrey C. Layman. 2006. “Changing Sides or Changing Minds? Party Identification and Policy Preferences in the American Electorate.” American Journal of Political Science 50: 464–477.
Chaney, Carole Kennedy, R. Michael Alvarez, and Jonathan Nagler. 1998. “Explaining the Gender Gap in U.S. Presidential Elections, 1980–1992.” Political Research Quarterly 51: 311–339.
Dolan, Kathleen A. 2014. When Does Gender Matter? Women Candidates and Gender Stereotypes in American Elections. New York, NY: Oxford University Press.
Fowler, Erika Franklin, Travis N. Ridout, and Michael M. Franz. 2016. “Political Advertising in 2016: The Presidential Election as Outlier?” The Forum 14: 445–469.
Hayes, Danny. 2005. “Candidate Qualities through a Partisan Lens: A Theory of Trait Ownership.” American Journal of Political Science 49: 908–923.
Iyengar, Shanto, Gaurav Sood, and Yphtach Lelkes. 2012. “Affect, Not Ideology: A Social Identity Perspective on Polarization.” Public Opinion Quarterly 76: 405–431.
Iyengar, Shanto, and Sean J. Westwood. 2015. “Fear and Loathing across Party Lines: New Evidence on Group Polarization.” American Journal of Political Science 59: 690–707.
Kaufmann, Karen M. 2006. “The Gender Gap.” PS: Political Science & Politics 39: 447–453.
Kaufmann, Karen M., and John R. Petrocik. 1999. “The Changing Politics of American Men: Understanding the Sources of the Gender Gap.” American Journal of Political Science 43: 864–887.
Levendusky, Matthew. 2009. The Partisan Sort: How Liberals Became Democrats and Conservatives Became Republicans. Chicago, IL: University of Chicago Press.
Manza, Jeff, and Clem Brooks. 1998. “The Gender Gap in US Presidential Elections: When? Why? Implications?” American Journal of Sociology 103: 1235–1266.
Nicholson, Stephen P. 2012. “Polarizing Cues.” American Journal of Political Science 56: 52–66.
Ronald Brownstein, “Why the ‘Trump Gap’ Is Growing,” The Atlantic, January 6, 2016.
Paul Solotaroff, “Trump Seriously: On the Trail with the GOP’s Tough Guy,” Rolling Stone, September 9, 2015.
Matt Flegenheimer and Maggie Haberman, “Donald Trump, Abortion Foe, Eyes ‘Punishment’ for Women, Then Recants,” The New York Times, March 30, 2016. (He later recanted the statement while confirming his general pro-life position).
Jose A. DelReal and Anne Gearan, “Trump: If Clinton ‘Were a Man, I Don’t Think She’d Get 5 Percent of the Vote’,” Washington Post, April 27, 2016.
These comparisons ignore sampling error that might render some of the differences insignificant.
Similar patterns are apparent if we expand the data to include surveys of registered voters and polls that ask about the three-way contest (including Libertarian Party nominee Gary Johnson) or the four-way contest (including Johnson and Green Party nominee Jill Stein).
Clinton would have won the Electoral College by picking up 79,000 more votes in three states – Michigan, Pennsylvania, and Wisconsin – that traditionally vote Democratic.
Emily Guskin, “Donald Trump is the Most Unpopular Presidential Candidate Since the Former Head of the Ku Klux Klan,” Washington Post, March 21, 2016.
Frank Newport and Andrew Dugan, “One in Four Americans Dislike Both Presidential Candidates,” Gallup blog, July 12, 2016.
The Economist/YouGov Poll. November 4–7, 2016.
These include 12 Economist/YouGov polls and three each from CNN/ORC and Marist.
Lynn Vavreck, “Why This Election Was Not About the Issues,” The New York Times, November 23, 2016.