The Party of Science Switched Sides

Ask Americans of the Nixon era which party trusted scientists, and the survey answer would have surprised today's readers: the Republicans. From 1973 through 1980, 48.7% of Republicans (leaners included) told the General Social Survey they had "a great deal" of confidence in the scientific community, against 43.2% of Democrats — a 5.5-point Republican edge. By 2022–24 that had inverted into a 28.2-point Democratic edge. Republican confidence has roughly halved, to 23.7%, while Democrats hold at 51.9%. No other institution in the GSS's half-century confidence battery has moved that far.

The chart below is the whole story in thirteen panels: every institution the GSS has asked about since 1973, with the weighted share of each party saying "a great deal" of confidence. Most panels show two lines braided together for decades. A few — science, medicine, the executive branch — show the braid coming apart. One, the military, shows a Republican advantage vanishing in a single six-year stretch. And one, the Supreme Court, shows fifty years of near-perfect overlap ending abruptly in 2022.

A slow flip, then a fast one

The science reversal happened in two speeds. The slow part took thirty-five years: the Republican advantage of the 1970s eroded gradually — sociologist Gordon Gauchat documented the conservative decline through 2010 in what is now the canonical paper on the subject — and 2006 was the last year the annual gap leaned Republican, at a trivial 1.3 points. In 2008 the gap turned Democratic, at +7.5 points, and never went back.

The fast part came after 2016. The gap stood at +12.8 points that year. By 2021 it had hit +30.1 points — Democrats at 63.7%, Republicans at 33.6% — and it sat at +28.2 points pooled across 2022–24. The COVID years did to the science gap in five years what the previous flip had needed decades to do: the 2016-to-2021 jump alone exceeds the entire 1973-to-2008 swing.

Two caveats matter here, and neither rescues the old pattern. First, the GSS switched from in-person to mixed-mode interviewing in 2021, so exact post-2020 levels deserve some humility — but the gap is similarly enormous in 2021, 2022, and 2024, across different fielding conditions. Second, the result is not an artifact of counting independent leaners as partisans: restricted to strong and weak identifiers only, the gap runs from −6.5 points in 1973–80 to +27.7 points in 2022–24. The flip is the same either way.

Science is the extreme case, not the exception

Pull back to all thirteen institutions and science looks less like an anomaly and more like the leading edge of a general bifurcation. The mean absolute Democrat–Republican gap across the full battery more than doubled, from 4.8 points in 1973–80 to 10.5 points in 2022–24. Medicine traced science's path in miniature: a 3.6-point Republican edge in the first era became a 14.6-point Democratic edge in the last. The press hit a partisan floor — just 3.0% of Republicans now express a great deal of confidence in it, against 12.8% of Democrats, on an item where the top box is a high bar for everyone.

The mirror image: the military gap vanished

If science is the story of Democrats inheriting an institution, the military is the story of Republicans abandoning one. The Republican advantage on the military was the largest pro-Republican gap in the whole panel for most of this century — it peaked at 27.1 points in 2006 and stood at 21.4 points as recently as 2018, when 74.0% of Republicans and 52.5% of Democrats expressed a great deal of confidence.

By 2024 the gap was gone: 44.2% of Republicans versus 44.5% of Democrats — a difference of 0.3 points, statistically nothing. Republican confidence in the military fell 29.8 points in six years. Some of that drop coincides with the 2021 mode change, so the cleanest like-for-like comparison is within the mixed-mode era: from 62.1% in 2021 to 44.2% in 2024, an 18.0-point slide in three years, while Democratic confidence drifted up. The armed forces went from the most partisanly Republican institution in the survey to one with no partisan tilt at all — not because Democrats warmed to it, but because Republicans cooled, fast.

The last consensus institution

Through all of this, one panel in the grid stayed flat and entangled: the Supreme Court. In 1973–80 the party gap on the Court was 0.0 points — 33.8% of Democrats and 33.8% of Republicans, identical to the decimal. Across every pooled era from the 1970s through 2021, the gap never exceeded 7.0 points in either direction, even as science, the military, and the executive branch polarized around it. The Court was the consensus outlier.

Then it snapped. Democratic confidence fell from 24.9% in 2021 to 8.3% in 2022 — a 16.6-point single-wave collapse, in the first GSS fielded mostly after Dobbs v. Jackson Women's Health Organization. Republicans barely moved over the same interval, from 30.6% to 26.8%. And the drop held: in 2024 Democratic confidence was 10.1%, against 28.7% for Republicans. The pooled 2022–24 gap of −18.5 points is the largest partisan tilt the Court has shown in the survey's history, and it survives the no-leaners robustness cut (−0.9 points in 1973–80, −18.8 now).

How unusual is a 16.6-point one-wave drop? Large, but not unprecedented. Among all 828 consecutive-wave party changes in the panel since 1973, the Democrats' Court collapse ranks ninth; since 2000 it ranks third, behind Republicans' crash in confidence in banks from 2006 to 2008 (−19.5 points) and in major companies from 2000 to 2002 (−19.1 points). The all-time record belongs to Republicans and the executive branch, from 1991 to 1993 — a 30.2-point plunge when the White House changed hands. What sets the Court apart is not the size of the fall but what fell: the one institution that had spent fifty years immune to exactly this kind of partisan repricing. The bifurcation that began with science in the 2000s now has no holdouts among the major institutions of American government.

What this does and doesn't show

These are descriptive trends in a top-box survey item, and three limits are worth stating plainly. The GSS's "great deal / only some / hardly any" scale sets a high bar at the top, which is why the press sits in single digits for everyone; the gaps, not the levels, carry the story. The 2021 redesign moved the survey from in-person to mixed-mode interviewing, so any comparison that crosses 2021 — including the military's slide and part of science's surge — mixes a mode shift with real change; the within-mode movements (the Court's 2021–22 collapse, the military's 2021–24 slide) are the cleanest evidence. And nothing here is causal: the 2022 fielding window brackets Dobbs, it does not isolate it. The GSS is a national sample, so none of this supports state or county claims. What the data do show, robustly across weighting and partisan definitions, is a half-century rearrangement of which party vouches for which institutions — with science the biggest mover, the military the latest defection, and the Supreme Court the last domino.