The Great Sorting

Ch. 9

Gender & Political Identity

In 1972, when the GSS first asked Americans about their party identification, women and men were essentially identical. 56% of women called themselves Democrats; 56% of men did. Republican identification was 28% among women, 29% among men. There was no gender gap to speak of. Eighteen years later, in 1990, women were 7 points more Democratic than men. The contemporary 2024 gap is 14 points — the largest in the GSS's history. The "gender gap" of American political journalism is a real phenomenon, observable in the country's longest-running political-attitude survey, and traceable to a specific period: the late 1970s and early 1980s. What is not true is that the gap has steadily widened across every decade since. It opened during the Reagan era, then stabilized at about 10–15 points, and has held in that band for thirty years.

14pts
The 2024 gender gap on Democratic margin (D−R): women lean Democratic by 7 points (D=40%, R=33%); men lean Republican by 7 points (D=32%, R=39%). The 1972 reading was a 0-point gap. The 1990 reading was 12 points. The contemporary gap is in the same range it has occupied for thirty years — not a recent widening but a persistent feature of the post-Reagan political landscape.

The gender gap, opening: 1972 to 1990

The cleanest chart for this chapter is the trajectory of Democratic identification within each gender group across fifty years.

Distribution of Political Party by Gender

For the US Adult Population

Source: General Social Survey, United States, 1972 - 2024; 225195 Observations
Figure 1. Share of US adults identifying as Democrat, by gender, GSS 1972–2024. In 1972 the two lines were within 0.3 points of each other (F=56%, M=56%). The gap opened in the 1980s as men drifted Republican faster than women did. By 1990 women were 7 points more Democratic than men; by 2010 the gap was 7 points; by 2024 it has widened slightly to 8 points (F=40%, M=32%).

The Democratic-line chart shows the gap opening between 1972 and 1990 — a roughly 18-year process — and then holding roughly stable through the contemporary period. The "Reagan-era gender gap" is the more accurate framing than "the gender gap that has steadily widened over fifty years." Whatever caused the 1980s divergence — Reagan-era social policy, the salience of abortion politics post-Roe, the cultural reorientation of the parties — finished its work by about 1990 and has not produced large additional movement.

The Republican-line chart shows the mirror image.

Distribution of Political Party by Gender

For the US Adult Population

Source: General Social Survey, United States, 1972 - 2024; 225195 Observations
Figure 2. Share of US adults identifying as Republican, by gender, GSS 1972–2024. The two lines were within 1 point in 1972 (F=28%, M=29%). The gap opened in the 1990s and has held roughly stable since. The 2024 reading shows men at 39%, women at 33% — a 6-point male-Republican advantage.

Combined, Figures 1 and 2 produce the contemporary "Democratic margin" gap: women are 7 points more Democratic than Republican; men are 7 points more Republican than Democratic. The 14-point difference between those two margins is the contemporary "gender gap" most journalism refers to.

What drove the gender gap

The cleanest single-finding interpretation of the gender-gap data: the gap is driven by men, not by women. The Democratic line on women has fallen from 56% in 1972 to 40% in 2024 — a 16-point decline. The Democratic line on men has fallen from 56% to 32% — a 24-point decline. The gap opened because men moved away from the Democratic Party faster than women did.

Table 1. Democratic-margin (D − R) by gender across fifty years. Both groups have become less Democratic, but men more so. The gap opened in the 1980s and has held roughly stable since 1990.
YearWomen D − RMen D − RGender gap
1972+28+27+1
1980+21+22−1
1990+7−5+12
2000+16−2+18
2010+19+6+13
2018+12−3+15
2024+7−7+14

The shape of the gap is what most accounts of the gender realignment understate. The conventional framing is "women have moved Democratic over the past fifty years." The data shows the opposite — women have become less Democratic, just more slowly than men have. The contemporary Democratic majority among women is one-quarter the size of the 1972 Democratic majority among women.

The ideology gap is smaller than the partisan gap

The ideology breakdown by gender produces a smaller and less directional version of the same pattern.

Distribution of Ideology by Gender

For the US Adult Population

Source: General Social Survey, United States, 1974 - 2024; 197319 Observations
Figure 3. Share of US adults identifying as Liberal, by gender, GSS 1974–2024. The female Liberal line has held remarkably stable around 28–32% across fifty years. The male Liberal line has also held stable around 26–31%. The contemporary gender gap on Liberal identification is small — about 2 points more Liberal among women than men.

Distribution of Ideology by Gender

For the US Adult Population

Source: General Social Survey, United States, 1974 - 2024; 197319 Observations
Figure 4. Share of US adults identifying as Conservative, by gender, GSS 1974–2024. The male Conservative line has risen modestly (31% → 38%) over the half-century. The female Conservative line has risen more modestly (28% → 33%). The contemporary gender gap on Conservative identification is 6 points — men more Conservative than women — and has held roughly stable since the 1990s.

Read together, Figures 3 and 4 show that the contemporary ideological gender gap (Liberal − Conservative differential) is about 8 points — women's Liberal-Conservative margin is −4 in 2024, men's is −12. The 8-point ideological gender gap is smaller than the 14-point partisan gender gap. This is a consistent finding across this book's chapters: people sort onto parties more sharply than they sort onto liberal-vs-conservative self-identification. The partisan label is doing more lifting than the ideology label.

The marriage gap: larger than the gender gap

The proposal observed that the marriage gap — the divide between married and never-married voters — has grown alongside the gender gap. The contemporary data show this is the chapter's larger story.

Distribution of Political Party by Marital Status

For the US Adult Population

Source: General Social Survey, United States, 1972 - 2024; 225351 Observations
Figure 5. Share of US adults identifying as Democrat, by marital status, GSS 1972–2024. The four marital-status groups have followed different trajectories. Never-married voters have stayed roughly Democratic-leaning (40% Democrat in 2024). Married voters have drifted Republican (now D=34%, R=42%). Widowed voters were strongly Democratic in 1990 (D+28) and are roughly even now. Divorced/separated voters are roughly even. The 2024 spread between Married (D−R=−8) and Never-Married (D−R=+13) is 21 points — larger than the gender gap (14 points).

The marriage-gap finding is striking on three dimensions.

First, the 2024 Married-vs-Never-Married Democratic-margin gap (21 points) is larger than the 2024 gender gap (14 points). Marriage is, by this measure, a bigger axis of contemporary political polarization than gender is.

Second, the marriage gap is not fifty years old in the way the gender gap is. The 1972 reading showed never-married voters slightly less Democratic than married voters; the contemporary gap is the opposite direction at much larger magnitudes. The "marriage drifts Republican" / "single drifts Democratic" sorting is a post-1990s phenomenon.

Third, the proposal flagged that the interaction — married men vs. single women — produces the largest political divide visible in survey data. The GSS data shown here aggregates by gender and by marital status separately rather than crossing them, but the directional logic is consistent: the most Democratic-leaning cell would be the never-married women cell, and the most Republican-leaning cell would be the married-men cell. The 21-point marital gap plus the 14-point gender gap would together produce a married-men-vs-single-women spread approaching 35 points — among the largest demographic divides anywhere in the GSS.

What this chapter changes

Three observations.

First, the gender gap is a Reagan-era artifact, not a contemporary widening. The 1972 reading shows no gender gap. The 1990 reading shows a 12-point gap. The 2024 reading shows a 14-point gap. The chapter is not "the gender gap is widening"; the chapter is "the gender gap opened during the 1980s and has held in a 10–15 point band since." The contemporary "men are drifting Republican" story is not new in 2024 — it has been true since the late 1980s.

Second, the gap is driven by men, not by women. Both groups have become less Democratic since 1972. Men have moved faster. The contemporary Democratic majority among women is one-quarter the size of the 1972 majority. The framing "women have moved Democratic" is not what the GSS shows. The framing "men have moved Republican faster than women have" is.

Third, marriage now divides the country more than gender does. The contemporary married-vs-never-married Democratic-margin gap (21 points) exceeds the gender gap (14 points). The marriage divide is a contemporary phenomenon — not visible at all in 1972 — that has grown alongside other demographic sortings of the American electorate, and is now larger than the most-cited gender finding. Any political-coalition analysis that does not include marital status alongside gender is missing the larger of the two related sortings.

The next chapter turns from gender to the demographic axis that is conventionally treated as the largest: the racial composition of the two parties, and what has happened to the White voter — historically a Democratic-majority population — over the same fifty years documented in this chapter.