The Great Sorting

Ch. 12

The Great Sorting

In 1974, the General Social Survey asked Americans both how they thought of themselves ideologically (Liberal, Moderate, or Conservative) and which party they identified with. The cross-tabulation produced what would now look like an absurd picture: 25% of Democrats said they were Conservative. 19% of Republicans said they were Liberal. The two parties contained the full ideological spectrum, in roughly overlapping proportions, with only a modest tilt — Democrats leaning Liberal, Republicans leaning Conservative, but with substantial cross-pressured voters in both. Fifty years later, the cross-pressured voter has nearly vanished. The 2024 GSS shows 11% of Democrats calling themselves Conservative and 6% of Republicans calling themselves Liberal. The two parties are now ideologically pure, in the sense that knowing someone's party tells you their ideological self-identification with near-certainty. This chapter is about how the great sorting happened and what it means.

6%
The share of Republicans in the 2024 GSS who identify as Liberal — down from 19% in 1974. The corresponding figure for Democrats identifying as Conservative is 11%, down from 25%. The cross-pressured voter who was a substantial share of both 1970s coalitions is, in 2024, a marginal exception. The two parties have become ideologically sorted to the point where the rare counterexamples — the Liberal Republican, the Conservative Democrat — are now political curiosities rather than typical members of their coalition.

The lead chart: ideological composition of each party

The cleanest single chart for this chapter is the ideological composition of each party over fifty years.

Distribution of Political Party by Ideology

For the US Adult Population

Source: General Social Survey, United States, 1974 - 2024; 196716 Observations
Figure 1. Share of self-identified Democrats by ideological self-identification, GSS 1974–2024. The Liberal share has risen from 36% to 59%. The Conservative share has fallen from 25% to 11%. The Moderate share has fallen from 40% to 30%. Democrats in 2024 are substantially more Liberal-coded than Democrats in 1974, with most of the change concentrated in the post-2000 period.

Distribution of Political Party by Ideology

For the US Adult Population

Source: General Social Survey, United States, 1974 - 2024; 196716 Observations
Figure 2. Share of self-identified Republicans by ideological self-identification, GSS 1974–2024. The Conservative share has risen from 42% to 71%. The Liberal share has fallen from 19% to 6%. The Moderate share has fallen from 39% to 23%. Republicans in 2024 are overwhelmingly Conservative-coded, with the Conservative share now occupying more than 70% of the coalition.

The two charts together tell the chapter's story. In 1974, both parties contained substantial Liberal, Moderate, and Conservative members. The plurality of each party leaned in the expected direction — Democrats more Liberal, Republicans more Conservative — but the modal Democrat was a Moderate (40%) and the modal Republican was a Conservative (42%). The two coalitions overlapped meaningfully.

In 2024, the modal Democrat is a Liberal (59%) and the modal Republican is a Conservative (71%). The overlap has compressed to the small wings of each party. The 1974 Democratic coalition contained 40% Moderates and 25% Conservatives — 65% of the coalition was not Liberal. The 2024 Democratic coalition contains 30% Moderates and 11% Conservatives — 41% is not Liberal. The Republican-side shift is parallel: 1974 Republicans were 58% non-Conservative, 2024 Republicans are 29% non-Conservative.

The mirror view: where does each ideology land?

The complementary view — where do self-identified Liberals and Conservatives go on the partisan question — produces an even sharper picture.

Distribution of Ideology by Political Party

For the US Adult Population

Source: General Social Survey, United States, 1974 - 2024; 196716 Observations
Figure 3. Share of self-identified Liberals identifying as Democrat, by year, GSS 1974–2024. In 1974, 68% of Liberals were Democrats; 18% were Republicans. By 2024, 78% are Democrats; 8% are Republicans. The Liberal-Republican has nearly vanished as a meaningful political category.

Distribution of Ideology by Political Party

For the US Adult Population

Source: General Social Survey, United States, 1974 - 2024; 196716 Observations
Figure 4. Share of self-identified Conservatives identifying as Democrat or Republican, by year, GSS 1974–2024. In 1974, 48% of Conservatives identified as Democrat and 41% as Republican — Conservatives were actually slightly more Democratic than Republican (a relic of the Solid South). By 2024, 12% of Conservatives are Democrats and 72% are Republicans. The Conservative-Democrat coalition that dominated mid-century Southern politics is now a marginal share of the contemporary electorate.

The Conservative-Democrat in 1974 was not a marginal political type. They were 48% of self-identified Conservatives — most Conservatives in 1974 were Democrats. The Southern Democratic Party of the post-Civil-Rights era was substantially built on Conservative-Democrat voters. The dramatic compression of that share to 12% by 2024 is the deepest single demographic movement this book documents.

The sorting in one table

The cross-tab can be presented compactly across decades:

Table 1. Ideological composition of each party at four points across the GSS series. The diagonal (Liberal-Dem, Conservative-Rep) has thickened substantially; the off-diagonal (Conservative-Dem, Liberal-Rep) has thinned. The Moderate share has shrunk on both sides, with the displaced voters mostly migrating to the matching-direction extreme rather than crossing parties.
YearGroupLiberalModerateConservative
1974Democrats36%40%25%
1974Republicans19%39%42%
2010Democrats47%37%16%
2010Republicans10%27%63%
2024Democrats59%30%11%
2024Republicans6%23%71%

The change is asymmetric on direction but symmetric on shape. Both parties moved away from a balanced mix toward an ideologically dominant single category. The mechanism on both sides is the same: Moderates have shrunk as a share of each coalition, and the displaced Moderates went to the matching-ideology pole rather than across the aisle. The 1974 Democratic Party was 40% Moderate; the 2024 Democratic Party is 30% Moderate. The 10-point Moderate compression went Left on the Democratic side (Liberal share rose from 36% to 59%). The Republican Moderate share fell from 39% to 23%, a 16-point compression, and went Right (Conservative share rose from 42% to 71%).

What the sorting changes

Three things follow from the sorting documented in this chapter.

First, the contemporary correlation between party and ideology is very high. The book's earlier chapters have documented the correlation rising from about 0.4 in the 1970s to about 0.75 in the 2020s. The chapter's data is consistent with that: knowing someone's party identification in 2024 lets you predict their ideological self-identification with high confidence. The two variables have fused in a way they had not in 1974.

Second, the cross-pressured voter has nearly vanished. The 1974 GSS showed 25% of Democrats as Conservatives and 19% of Republicans as Liberals — a substantial cross-pressured population in both coalitions. The 2024 GSS shows 11% and 6% — a small marginal population. The "moderate Republican from the Northeast" or the "conservative Democrat from the South" is now a political type whose voters are scarce enough that they no longer constitute a meaningful intra-party constituency.

Third, polarization is now structurally reinforcing. When the two parties contained ideological diversity, partisan disagreement could be moderated by within-coalition cross-pressuring. A 1974 Republican could find Conservative Democrats to negotiate with on cultural issues; a 1974 Democrat could find Liberal Republicans to negotiate with on civil-rights issues. The 2024 coalitions do not contain those mediating voters in meaningful numbers. The compromise positions of mid-century American politics — Northeast Republican governors signing civil-rights legislation; Southern Democrats negotiating Great Society programs — required a within-coalition spectrum that no longer exists.

What this chapter sets up

The remaining chapters in Part V take three different cuts at the same finding. §13 turns from ideological sorting to issue-level sorting (how the contemporary partisan coalitions agree internally on specific issues) and to geographic patterns. §14 asks whether the left-right axis the GSS uses to measure ideology is sufficient — whether the contemporary American political space might actually need multiple dimensions to describe. §15 closes the book with the implications of the chapter-12 sorting argument for the next decade of American politics.