Racial Attitudes & Explanations for Inequality

The discrimination question: how the gap got there

The lead chart in this chapter is the partisan view of racdif1_rec, the discrimination question. To keep the framing consistent with the GSS catalog, the values plotted are the share of respondents saying "no, the differences are not mainly due to discrimination" — the rejection of the structural explanation. Lower numbers mean more respondents accept the structural explanation; higher numbers mean more reject it.

The 1977 starting point is the part of Figure 1 that surprises most readers. The standard contemporary framing of this question — Democrats believe in structural explanations of racial inequality; Republicans don't — describes the 2024 chart accurately but is anachronistic when applied to 1977. In the late 1970s, both parties' base voters mostly rejected the discrimination explanation. The 7-point partisan gap was small, real, and in the same direction it would later widen, but the position both parties occupied in 1977 was the position the Republican Party now occupies more or less alone.

The trajectories that produced the contemporary gap are asymmetric.

The topline movement is modest (5 points more accepting of the structural explanation over fifty years) and barely registers on the chart. The partisan story is what carries the finding: Democrats migrated from a position close to Republicans toward a position roughly opposite to it, while Republicans intensified their original position.

The race breakdown adds a critical layer.

The race-of-respondent gap is one of the most stable findings in the GSS's racial-attitudes battery. Across every wave, in every decade, Black Americans have been substantially more likely to attribute racial outcome differences to discrimination than White Americans have. The 25-point 2024 race gap (Black 32% reject vs. White 57% reject) holds whether we look at Democrats or Republicans separately — Black Republicans are more likely than White Republicans to accept the structural explanation, even though they remain less likely than Democrats of either race.

The education and age breakdowns sharpen what the partisan picture shows. By 2024, college-educated respondents are 9 points more accepting of the structural explanation than non-college respondents; the youngest cohort (18–34) is 14 points more accepting than the 35–49 cohort and is the only age band where a majority accepts the explanation.

The motivation question: the Republican counter-explanation

The racdif4_rec question is the inverse pair of racdif1. "Are the differences mainly due to most Black people just not having the motivation or will power to pull themselves up out of poverty?" This is the canonical individual-attribution explanation for racial outcome differences. It has moved sharply, in the opposite direction from racdif1, over the same fifty years.

The country has substantially moved away from the motivation explanation. The 1977 reading of 64% endorsing it is striking in retrospect: a majority of Americans, across both parties, agreed that the differences in Black-White outcomes were mainly explained by individual motivation. By 2024 only 31% of Americans agree.

Underneath the topline, the partisan structure tells the same story as racdif1 in mirror image.

What is striking about Figures 1 and 4 read together is that both coalitions have moved on both questions. Republicans in 2024 reject the discrimination explanation more strongly than they did in 1977, but they also endorse the motivation explanation less strongly. They have not simply consolidated their 1977 position. They have moved toward the rejection-of-structural-explanation pole on one axis and away from the endorsement-of-individual-explanation pole on the other. The "Republican coalition holds the individual-explanation framework" reading is correct for 2024 but conceals that the actual contemporary Republican position is more skeptical of both framings than the 1977 Republican position was.

The Democratic shift is sharper. On both questions, Democrats moved 27 points and 43 points respectively — among the largest single-coalition movements on any attitude item in this book.

The race breakdown on racdif4 is consistent with the race breakdown on racdif1.

The biological-racism question: the rejected explanation

The third item in the racdif battery — racdif2_rec, "due to less inborn ability" — asks whether respondents endorse a biological explanation for racial outcome differences. This is the explanation that has most clearly lost support over the half-century, and the one with the smallest partisan gap.

The contemporary bipartisan rejection of biological-racism framings — 8% nationally endorse, with Democrats and Republicans within 3 points of each other — is the single piece of good news in this chapter. The country, across both coalitions, has moved away from the explicitly biological framing of racial inequality. Even allowing for social-desirability bias (respondents may be unwilling to endorse a stigmatized framing they privately believe), the trajectory from 26% in 1977 to 8% in 2024 is too large to be explained by reporting effects alone.

The work-ethic stereotype: a measure of explicit prejudice

The workblks_rec item is in a different family from the racdif battery. Instead of asking about explanations for inequality, it asks whether the respondent endorses an explicit stereotype: "On a scale from 1 to 7 where 1 means hardworking and 7 means lazy, where would you rate African-Americans?" The catalog measures the share endorsing the lazy end of the scale.

The work-ethic stereotype reading has, like the inborn-ability reading, fallen substantially over the period the GSS observes it. Whether this represents real attitude change or shifting social-desirability norms about what respondents are willing to admit to a survey is an open question. The most likely answer is "both" — the explicit-stereotype norms have shifted enough that the 2024 readings are best read as an upper bound on the share of the population that still privately endorses these views, not as a clean measurement.

The 2024 partisan structure is what one would expect by now: Democrats lower (14%), Republicans higher (27%), gap of 13 points. The 1990 gap was small (3 points); the contemporary gap is the result of the same asymmetric-movement pattern documented for racdif1 and racdif4. Democrats moved 28 points; Republicans moved 19 points.

The four-item table

Read together, the four items in the GSS's racial-attribution battery produce a consistent partisan picture. The table below summarizes the 2024 readings and the 1977 (or earliest-available) baseline for each.

What this chapter changes

Three findings to carry into the rest of the book.

First, the contemporary partisan structure on race is recent. The 1977 racdif1 and racdif4 readings were within single digits across parties. The 1990 workblks reading showed a 3-point partisan gap. The gaps that exist now — 48 points on discrimination, 23 points on motivation, 13 points on the work-ethic stereotype — are the product of fifty years of asymmetric coalition movement, not of one coalition holding a longstanding position and the other arriving late. The Republican coalition of 1977 sat where the contemporary moderate-Democratic position approximately sits.

Second, the explanatory-framework story is real, but it is more "Democrats migrated" than "Republicans entrenched." On every racial-attribution item the GSS tracks, the Democratic line moved by roughly 1.5–3 times what the Republican line moved. The widening partisan gap is mostly a story of one coalition reorganizing its framework while the other intensified its existing framework somewhat. This is the asymmetric-liberalization pattern that recurs across the cultural questions in §4.3 (sexual morality) and §3.5 (foreign aid), now in a different domain.

Third, the race of the respondent matters at least as much as the party of the respondent. On racdif1 the Black-White gap is 25 points, on the same order as the Democrat-Republican gap when broken out within the same race. The lived-experience effect is large, real, and durable across every wave of the question. Any reading of partisan polarization on racial-attitude questions that does not also account for the racial composition of the parties' coalitions — Democrats more diverse, Republicans overwhelmingly White — is missing one of the two main axes that structure the data.

The next chapter turns from racial-attribution questions to a closely related domain: police, authority, and the criminal-justice system. The same partisan-flip dynamic the proposal observed on crime spending (§3.3) reappears in a more recent and sharper form on the policing items the GSS tracks.