Ch. 11
The Long View — Five Decades of Change
This book has presented twenty-four chapters of GSS evidence. Across roughly fifty institutional-confidence and attitudinal items, across the major demographic axes the GSS tracks, a fairly small number of patterns recur. This chapter is the long-view summary of those patterns. It names them, it identifies the historical periods when each was at its sharpest, and it argues that they fit together into something tighter than "American politics has become more polarized" — though not as tight as "everything has moved together." Each pattern lives in its own chapter; this chapter is the cross-chapter index.
Pattern 1: the institutional-confidence collapse
The dominant story of Part II is that institutional confidence has fallen — across nearly every institution the GSS tracks, in nearly every wave since the mid-1970s. The drops range in size from the catastrophic (Congress fell from 24% to 5.6% over fifty years, a 78% relative collapse) to the moderate (the military, religion, banks, business, the press, the medical profession, the scientific community have each lost 20–35 points on their high-confidence readings).
The collapse is not partisan in the standard sense — it shows up across both Democratic and Republican respondents on most items. It is, however, asymmetric across institutions. The institutions that lost the most confidence were not the institutions of greatest contemporary political controversy. They were Congress, the press, and the medical profession. The institutions of greatest contemporary political controversy — the Supreme Court, the scientific community — actually held their confidence relatively well through 2018, then collapsed in specific recent events (Dobbs for the Court, COVID-19 for science).
Most-cited chapters: §2.2 Government, §2.4 Media, §2.3 Knowledge, §2.5 Economic, §2.6 Social/Cultural.
Pattern 2: the partisan flip on institutional confidence
The partisan structure of each institution's confidence reading has also shifted, and in many cases the direction of the partisan order has flipped. The executive branch is the canonical case: Republicans in the 1970s were the higher-confidence party (with Nixon in office); Democrats are now the higher-confidence party (with Biden in office, by a margin of 16 points). The pattern is fifty years deep and operates on each presidential transition.
But the partisan flip on the executive is the least interesting partisan flip in the data, because it is a feature of the political system rather than a feature of the institution. The more interesting flips are the institutions whose partisan signature has reversed in ways that are not directly explained by which party holds office:
- The scientific community — Republicans were more confident in 1973 (R+5); Democrats are now more confident by 26 points.
- Medicine — Republicans were more confident in 1973 (R+4); Democrats are now more confident by 14 points.
- The press — Modest partisan gaps until the early 2000s; now Democrats hold a 20+ point lead.
- Major companies — Republicans were the higher-confidence party for fifty years; the parties converged within 2 points by 2024.
The pattern is: institutions whose authority comes from expertise have flipped partisan-coding from Republican-favorable to Democratic-favorable. Institutions whose authority comes from belief (religion, the military) have lost confidence on the Republican side without gaining it on the Democratic side. The contemporary partisan map of American institutional trust does not look like the 1973 map. It is roughly the inverse of it.
Pattern 3: cultural liberalization with widening partisan gaps
The most striking pattern in Parts II and IV is that the country has, on virtually every cultural question the GSS tracks, become more liberal — and yet the partisan gap on those same questions has widened.
- Homosexuality always wrong: 78% (1991 peak) → 33% (2024). Partisan gap: 1973 = 3 points, 2024 = 33 points.
- Discrimination causes Black/White outcome differences: country leaned anti-discrimination 1977 to 2010; now divided. Partisan gap: 1977 = 7 points, 2024 = 48 points.
- Lack of motivation explains outcome differences: 64% (1977) → 31% (2024). Partisan gap: 1977 = 4 points, 2024 = 23 points.
- Confidence in the scientific community: country fell from 41% to 36%. Partisan gap: 1973 = −5, 2024 = +26.
- Premarital sex always wrong: 37% (1972) → 17% (2024). Partisan gap stable around 10–20 points.
The mechanism behind the simultaneous liberalization and gap-widening is asymmetric movement. On every cultural item where both coalitions liberalized, Democrats liberalized faster — typically 1.5 to 2× the Republican rate. The partisan gap widens not because Republicans dug in while Democrats moved, but because both coalitions moved at different speeds. The 33-point gap on homosex is the sum of a 25-point Republican liberalization plus a 54-point Democratic liberalization. The 48-point gap on discrimination is the sum of a 14-point Republican move plus a 27-point Democratic move.
Most-cited chapters: §4.1 Abortion, §4.2 Civil Liberties, §4.3 Sexual Morality, §4.4 Racial Attitudes, §2.3 Knowledge Institutions.
Pattern 4: the diploma divide on information-and-expertise questions
The college-vs-non-college gap is the dominant axis on questions involving information, expertise, and the legitimacy of expert institutions. It is large, persistent across the entire fifty-year GSS series, and shows up on:
- Confidence in the scientific community (15–25 point college lead, throughout).
- Confidence in major companies and banks (5–10 point college lead).
- Generalized social trust (18–26 points college lead within each income tier).
- Tolerance for unpopular speech (atheist, homosexual, communist speakers) (10–20 points college lead, narrowing in recent years).
On moral and lived-experience questions, the diploma divide is not the dominant axis. Religion (church attendance) dominates education on sexual morality. Race dominates education on police-force scenarios and on racial-attribution questions. Age dominates education on questions about cohort-replacement-driven attitude change.
The book's overall reading: the diploma divide is real, large, and durable on information questions, but it is not a master variable for political polarization. Where the partisan gap is small, the diploma divide is often the larger axis; where the partisan gap is large, the diploma divide is usually secondary.
Most-cited chapters: §6 Trust, §8 Educational Realignment, §2.3 Knowledge Institutions, §3.5 Science Space Aid.
Pattern 5: the demographic-coalition realignment
The contemporary American partisan coalitions are demographically very different from the 1972 coalitions. The largest single movement: the 35-point swing in White-voter Democratic margin (§10). White voters were Democratic-leaning by 22 points in 1972 and are Republican-leaning by 13 points in 2024. The Republican Party of 2024 is built on a coalition that gained White voters but did not gain Black or other-group voters.
The other major demographic movement is the 24-point college-vs-non-college Democratic-margin gap (§8). The Republican Party of 2024 is built on a coalition that lost college voters and gained non-college voters, with most of the movement coming from non-college voters becoming more Republican rather than from college voters flipping to Democrats.
The third demographic axis — the gender gap (§9) — opened in the 1980s and has been stable in a 10–15 point band since 1990. It is real but is not the contemporary widening that the post-2010 diploma-divide and post-2010 White realignment are.
The fourth axis — marital status (§9) — is a newer phenomenon. The 21-point gap between married and never-married voters is now larger than the gender gap, and it did not exist in 1972.
Most-cited chapters: §8 Educational Realignment, §9 Gender, §10 White Voter.
Pattern 6: partisan-motivated perception
The most-discussed finding for understanding how polarization works on the ground is the partisan-motivated-perception effect: people report different perceptions of their own lives depending on which party holds office. The book's cleanest example is financial satisfaction (§7.1): Republican-coalition financial satisfaction swung 11 points across the 2018 (Trump) → 2024 (Biden) transition, with no plausible underlying economic mechanism to explain the swing.
The pattern recurs on other items. Confidence in the executive branch is the cleanest case in §2.2 — the partisan flip on that variable is essentially mechanical. The presidential-control effect on overall well-being and life-satisfaction reporting is documented in the political-science literature and consistent with the contemporary GSS data, though the GSS does not directly measure presidential-attribution.
This pattern is what makes interpreting some of the other patterns in this book hard. If respondents' reports of their own happiness and financial situation move with the political environment, what is the survey actually measuring on those items? The cleanest answer, on this book's data, is "a mixture of underlying experience and political mood, with the political-mood component large enough to matter."
Most-cited chapters: §7.1 Life Satisfaction, §7.2 Anxious Liberal, §2.2 Government.
The periodization
Read across all six patterns, the contemporary political-attitude landscape was substantially produced in three distinct historical periods:
| Period | Name (informal) | Major patterns produced |
|---|---|---|
| 1972–1992 | Reagan-era realignment | Cultural liberalization begins (gay-rights, women's-rights movements); gender gap opens; Southern realignment of White voters begins; initial bipartisan consensus on Congress and federal institutions begins to fracture |
| 1992–2008 | Polarization era | Diploma divide opens; partisan media ecosystem fragments; Bush-era institutional disputes; ACA-era polarization of health-policy attitudes; early signs of post-2010 partisan widening on cultural items |
| 2008–2024 | Coalition acceleration | Educational realignment completes; non-college White-voter movement to Republican Party; Trump-era partisan widening on race, science, environment, immigration; Dobbs-era collapse of Democratic confidence in the Court; COVID-era collapse of Republican confidence in science and medicine |
What the data does not show
Three findings the contemporary discourse asserts but the GSS does not clearly support.
The gender gap is not "steadily widening." It opened in the 1980s and has been stable in a 10–15 point band since 1990. The 2024 reading is 14 points; the 2000 reading was 18 points; the 1990 reading was 12. The "men have left the Democratic Party" story is real but is not new in 2024.
The college-educated have not "shifted Liberal." The college Liberal share has fallen modestly (42% → 37%) over fifty years. The college Democrat share has fallen too (50% → 48%). The diploma divide opened because non-college voters became more Conservative and more Republican, not because college voters became more Liberal and more Democratic.
The country has not "lost trust in institutions" uniformly. Confidence collapsed catastrophically on Congress, the press, and medicine. Confidence is roughly stable on the military and on organized labor. Confidence on the scientific community held until 2018, then dropped on one coalition while staying stable on the other. The aggregate "institutions are losing confidence" claim is true on average and very misleading at the institution level.
What this chapter sets up
The next chapter, §12, makes the central interpretive argument the book has been building toward: that the contemporary partisan coalitions are not merely more polarized than the 1972 coalitions but are better sorted — the correlation between party identification and ideological self-identification has risen from about 0.4 in 1972 to about 0.75 in 2024. The cross-pressured voter (conservative Democrat, liberal Republican) of 1972 has nearly vanished. The next chapter is about how that sorting happened and what it means.