The phenomenon of the gender gap — the fact that women as a whole are more supportive of the Democratic Party than men are — masks significant divisions in the American electorate.
Nicholas Winter, a political scientist at the University of Virginia, addresses this issue head on in “Gendered (and racialized) partisan polarization,” a chapter in the 2020 book “Community Wealth Building and the Reconstruction of American Democracy”:
A focus on gender gaps misses important ways that gender structures party competition because the politics of gender does not, generally speaking, divide men from women. Rather, it often engages questions of women’s and men’s roles, rights, and relative power; it pits those defending traditional gender arrangements against those advocating for egalitarian gender arrangements. In other words, the politics of gender divides supporters of gender egalitarianism and feminism — male and female — from gender traditionalists and anti-feminists of both sexes.
Research by Winter and other social scientists demonstrates that competing ideas about the roles of men and women, at home and at work, shape our political life. They do not set men against women as much as produce two opposing coalitions, each made up of both men and women.
It almost goes without saying, but men and women who support traditional gender roles for men and women lean strongly toward the Republican Party; men and women who question traditional gender roles and who are sympathetic to women’s rights lean strongly toward the Democratic Party.
The public reaction to the Supreme Court’s decision overturning Roe v. Wade last month reflects this divide. A June 24-25 CBS/YouGov survey found that women disagreed with the Dobbs decision 67-33 while men disapproved by 51-39. The gap is there, but we need to think about it in different terms.
The emergence of these two coalitions adds to our understanding of the pervasive polarisation between left and right that dominates American politics. And it doesn’t just divide the electorate; it also has a huge impact on public policymaking.
While there are modest gender gaps in partisanship, voting and policy views, Winter wrote, “these pale compared with the differences among men and among women in views on gender roles and feminism. And gender roles and feminism have increasingly structured elite partisan debate.”
In an email expanding on the points he made in his book chapter, Winter wrote that “a voter’s personal masculinity/femininity (and views on same)” interacts with partisanship such that “people (men or women) who support traditional gender roles tend to favour the Republican Party and people who either reject or at least do not valorise traditional gender roles (men and women both) favour the Democratic Party.” The focus “is on the voter’s views about how gender should be organised (i.e., the belief that men should be masculine, act masculine and hold masculine roles; women should be feminine, act feminine, hold feminine roles).”
The gap between the attitudes of Democrats and Republicans toward the women’s rights movement has widened in recent years, Winter notes:
From 1970 through 2016, Democrats rated feminists and the women’s movement higher than did Republicans. This difference was modest — between 5 and 10 degrees — through the 1970s, then increased steadily from 10 degrees in 1980 to almost 20 degrees in the mid-1990s. After closing slightly, partisan polarisation in ratings of feminists reached their most polarised level yet in 2016. That year, Democrats rated feminists at 67 degrees, compared with 43 degrees among Republicans, a difference of about a quarter of the 101-degree rating scale.”
The same pattern Winter describes can be found on a wide range of politically salient issues. The Center for American Women and Politics at Rutgers issued a report, “Gender Gap Public Opinion,” based on poll data from the 2016 and 2020 American National Election Study, the 2018 General Social Survey, the 2020 Cooperative Election Study and the June 2020 AP-NORC Center Poll.
On the question “Should the federal government make it more difficult to buy a gun?” there is a significant 12-point gender gap, with 55.6 per cent of women saying it should be more difficult to buy a gun compared with 43.6 per cent of men. The gap between Democratic and Republican men is, however, more than three times larger, at 51.3 points, while the gap between Democratic and Republican women is larger still, at 54.4 points.
The question “Should federal spending on aid to the poor be increased, decreased, or kept the same?” produced a modest five-point gender gap, with 53.8 per cent of women saying it should be increased compared with 48.8 per cent of men. But the gap between Democratic and Republican men was 52.5 points, and between Democratic and Republican women it was 50.3 points.
The Rutgers centre shows the same configuration for other issues, like the death penalty, affirmative action, transgender rights, military spending and environmental protection.
In a separate email, Winter said that a key question he is now researching is “the degree to which ideas and feelings about gender and gender roles have become increasingly central to our partisan politics.”
Winter pointed to his 2010 article, “Masculine Republicans and Feminine Democrats: Gender and Americans’ Explicit and Implicit Images,” which makes the point that voter conceptions “about the two political parties are mapped onto ideas about the two genders, both in the images citizens consciously hold of the parties and in the implicit connections between these images and their gender concepts.”
Given “the masculine associations of leadership, we might expect that on balance the masculinisation of the Republican Party and feminisation of the Democratic Party may have conferred advantages to the Republicans, at least at the presidential level,” Winter wrote. But, he continued:
Cultural ideas about masculinity and femininity and about their connections with politics are complex enough that Democratic candidates may have more latitude than simply to try to outman the Republican Party. For example, while observers have commented on Barack Obama’s relatively feminine appearance and approach, he does not seem to have suffered from this image as much as his recent Democratic predecessors.
Using data from the 2020 Cooperative Election Study, Winter examined “the impact of beliefs about gender roles” and in an email noted that
views on gender roles and/or measures of sexism matter for abortion and have a decent impact on a range of abortion policies; and on the prevalence and importance of sexual harassment; and on LGBTQ+ issues, including same-sex adoption, trans adoption, trans military service, requirements for coverage of gender-confirming medical care, and (especially strongly) on banning gender identity discrimination.
Winter contends:
Beliefs about gender roles and sexism moderately impact support for government child care spending and parental leave, and strongly impact views on gender pay equity laws. On environment — gender traditionalists were a lot more supportive of withdrawing from Paris Climate and TPP agreements, and somewhat more opposed to a range of environmental policies including raising fuel efficiency; and they were opposed to stronger E.P.A. regulation of water, air and carbon emissions.
Winter reports that views on gender roles had a notable impact on opinions about Donald Trump, on how people voted over support for impeachment, and that such views have a substantial impact on immigration, on a path to legal status for undocumented immigrants and on support for Dreamers — but not much impact on things like border patrol spending and reducing legal immigration.
Winter made a point of noting that “all of those effects are estimated with controls for partisanship, racism and other important predispositions, and are therefore over and above the impacts of those other, correlated beliefs.”
In a variation on the same theme, Monika McDermott, a political scientist at Fordham, argues in her 2016 book, “Masculinity, Femininity and American Political Behavior,” that what she calls “gendered personalities” play a key role in shaping “party preference, vote choice, and ideology” and — importantly — “all occur independently of biological sex.”
In summary, she wrote, “The more ‘masculine’ an individual is, the more likely he or she is to affiliate with the G.O.P. and vote for Republican Party candidates.” Regardless of the sex of the individual, McDermott argued, “the more ‘feminine’ traits a person possesses, the more likely that person is to affiliate with, and vote for, the typically ‘feminine’” Democratic Party, adding that “‘femininity’ and ‘masculinity’ operate the same way in men and women, defying traditional research.”
In fact, McDermott observed that “once gendered personalities are accounted for, the longstanding sex gap in partisan preferences disappears.” She described what she calls a “femininity dimension” based on measures of such characteristics “as understanding, sympathetic, warm, loves children, compassionate, gentle, eager to soothe hurt feelings, affectionate, sensitive to needs of others and tender.” And she described a “masculinity dimension” based on measures of such traits “as willing to take risks, forceful, strong personality, assertive, independent, leadership ability, aggressive, dominant, willing to take a stand and defends own beliefs.”
Using these measures, McDermott notes that men, as expected, are higher in the “masculinity dimension” and that women are higher in the “femininity dimension.” More significantly, McDermott reports, there are large numbers of men and women who do not fit the stereotype: 41 per cent of men score above the median on the femininity dimension, and 35 per cent of women score above the median on the masculinity dimension.
While “the sex difference in Democratic leanings is nine points,” McDermott observes that there are
much larger gaps in relative party identification among different gendered personalities. On the femininity dimension, those with scores above the median level of femininity identify with Democrats more than Republicans by 23 points, while those below the median on this scale are one point more Republican than Democratic — an overall 24-point gap in relative party identification between low femininity versus high femininity. On the masculinity dimension, those with a greater number of masculine traits show only a three-point advantage in Democratic affiliation, compared to less masculine individuals who display a 21-point Democratic advantage — an 18-point gap by masculinity.
Over the past 50 years, McDermott notes, there has been a radical shift in public attitudes on the role of women. She cites changing results on a poll question that
asks whether women should have an equal role, or whether their place is in the home, on a one- to seven-point scale. Higher values indicate a more traditional role attitude. As the data show, traditional role attitudes — at least toward women — have dropped substantially, moving from a populace almost evenly divided in opinions in 1972 (3.5 point average) to one clearly in favour of an equal role for women (1.8 point average) in 2008.
At the same time, McDermott writes, “men’s possession of both masculine and feminine personality traits has increased. Both sexes are now, on average, more likely than in the past to possess the personality traits of the sex role opposite the one once dictated to them by tradition.”
Along parallel lines, Roland Levant, a professor of psychology at the University of Akron, noted in reply to my emailed query that there is a “relationship between men’s and women’s traditional male ideology and their endorsement of conservative political ideology.”
Traditional male ideology, Levant and five colleagues wrote in their 2021 paper, “The Politics of Men’s and Women’s Traditional Masculinity Ideology in the United States,” is “particularly evident through endorsement of hegemonic norms that condemn any thoughts, feelings, and behaviours for men that could be considered feminine.” Perhaps most significantly, Levant and his co-authors found that there was no gender gap between the political views of those men and women who favour traditional masculine ideology: “Women who believed that men should generally avoid traditionally feminine behaviours (e.g., be tough, stoic, dominant and hypersexual) were equally likely to endorse conservative political ideology as men who endorsed the same beliefs.”
There are two other challenges to the traditional view of the gender gap. First, the 2020 paper “The Changing Politics of American Men, Updated,” by Karen Kaufmann and John R. Petrocik, political scientists at U.C.L.A. and the University of Missouri.
“The gender gap,” Kaufmann and Petrocik write:
continues to reflect male-only changes in party identification. Public officials, political operatives, and the media misstate, sometimes specifically but often only by implication, the nature of the gap. Commentary highlights the lower level of support among women for the Republican Party in both the vote and party identification. Their support for the G.O.P. is considerably lower than it is among men, as this paper reaffirms. However, the gap emerged and has continued to grow because of changed attitudes and behaviour among men.
The very different trends in partisanship among men and women are clear in the data Kaufmann and Petrocik assembled. From 1948 to 2020, Democratic support among women remained relatively constant, ranging between 51 per cent and 59 per cent, according to surveys conducted by American National Election Studies, while Republican identification remained virtually unchanged. Among men, in contrast, Democratic identification fell from roughly 60 per cent in 1948 to 42 per cent in 2020, while Republican identification rose from 25 per cent to 44 per cent. In an email, Petrocik described some of the key points:
The marriage gap far exceeds the gender gap. That is, unmarried women are among the most Democratically inclined group in the society. The marriage gap (among women) approaches the racial gap.
In 2016, Petrocik wrote, TargetPoint Consulting, a Republican firm, found:
Republicans are losing with women by about 10 points, and winning with men by 1 — giving us a gender gap of 11 points. On the other hand, Republicans are winning by 7 points with married people and losing by 23 with single people — adding up to a much larger 30-point marriage gap.
In addition, Petrocik noted,
there is also a large economic sector gap among employed women. Women with professions such as teachers, social workers, and other “people-oriented” occupations were strongly tilted toward the Democrats. Women in business and similar economic activity looked much more like men.
The second challenge to traditional thinking about the gender gap is presented in “The Gender Gap Is a Race Gap,” a 2019 paper by Jane Junn, a political scientist at U.S.C., and Natalie Masuoka, a political scientist at U.C.L.A.
“Scholarship on women voters in the United States,” Junn and Masuoka write,
has focused on the gender gap, showing that, since the 1980s, women are more likely to vote for Democratic Party candidates than men. The persistence of the gender gap has nurtured the conclusion that women are Democrats. This article presents evidence upending that conventional wisdom.
Although the 2016 election clearly demonstrated stronger support among women for Hillary Clinton than for Donald Trump, Junn and Masuoka point out that “exit poll and survey data revealed that, among white women, an estimated 52 per cent voted for Trump compared to 43 per cent who supported Clinton.”
Network exit polls in 2020 revealed even stronger support for Trump among white women, 55 to 44.
Junn and Masuoka write, “While white women support Republicans to a lesser degree than white men, they still vote majority Republican and have done so consistently in presidential elections for more than a half-century.” This pattern, they continue, “has been apparent in national election data for decades, but has been obscured by the way analysts have approached voting behaviour.”
The partisan race gap — the difference between white and nonwhite voters — according to Junn and Masuoka, is nearly 10 times the size of the partisan gender gap (as measured between white men and white women).
All of this will continue to mould and confound party allegiance, but for now let’s give Winter the last word, drawing from his earlier book, “Dangerous Frames: How Ideas About Race and Gender Shape Public Opinion”:
Race and gender are extremely well suited for shaping social and political perception and evaluation, and that very subtle language can trigger powerful effects. In a sense, then, we are subject to the power of our own mental categories and to the power of communication to evoke those categories, and avoiding those effects is extraordinarily difficult, if not impossible.
This article originally appeared in The New York Times.
Thomas B. Edsall has been a contributor to the Times Opinion section since 2011. His column on strategic and demographic trends in American politics appears every Wednesday. He previously covered politics for The Washington Post.
