A preprint posted to arXiv (arXiv:2603.26896v1, not yet peer-reviewed) offers one of the more granular looks at how Americans misperceive racial diversity. Using a purpose-built national survey that asked respondents to estimate the share of people of color (PoC) in four nested geographies—neighborhood, city, state, and nation—the authors document that overestimation is both scale-dependent and group-dependent. The probability of overestimating rises steadily from local to national levels. People of color are more likely than white respondents to overestimate their own group at both neighborhood and national scales. Sample size and full demographic breakdowns are not detailed in the abstract, a limitation that restricts assessment of representativeness and statistical power. The study is cross-sectional, so causality between exposure and perception cannot be firmly established.

For white respondents, the drivers split by scale: direct interethnic social contact is the strongest correlate of overestimation locally, while perceived frequency of PoC coverage in news dominates at the national level. Across all groups, frequent traditional news consumption correlates with reduced overestimation, whereas frequent social media use correlates with higher overestimation. These patterns suggest an 'illusion of diversity' that may erode support for equity-promoting policies by creating the false impression that representation targets have already been met.

This preprint goes further than most prior work by mapping the same individuals across multiple geographic scales and separating social versus media mechanisms. Mainstream reporting on demographic misperceptions has typically quoted only national-level overestimates—often citing figures suggesting Americans believe the country is roughly 40% Black—yet rarely examines how the bias grows with distance from lived experience or differs by the respondent's own race.

Synthesizing with related sources illuminates broader patterns. A 2018 peer-reviewed study in Social Psychological and Personality Science by Craig and Richeson documented that white Americans who overestimate minority population sizes show significantly lower support for affirmative action and immigration. Similarly, a 2021 Pew Research Center analysis of demographic perceptions found that those exposed to rapid local change often exaggerated national diversity, but stopped short of distinguishing social-contact from media effects. The current arXiv work connects these dots: the availability heuristic, amplified by vivid social encounters or algorithmic social-media feeds, systematically distorts judgment. Heavy social-media users encounter curated, high-visibility minority success stories and protest imagery, inflating perceived prevalence.

The societal implications are rarely quantified in coverage of polarization. When citizens—especially white respondents—believe their neighborhoods and nation are already highly diverse, the perceived need for structural remedies diminishes. This cognitive bias feeds directly into recent backlash against DEI programs and race-conscious policies, helping explain why support for such measures can decline even as objective diversity increases. The study also hints at an underappreciated feedback loop: overestimation may reduce empathy for continued disparities precisely because the 'diversity goal' appears achieved.

Limitations are notable. Self-reported media exposure and social-contact measures can suffer from recall and social-desirability bias. The preprint does not explore longitudinal change or experimental interventions that could test whether correcting perceptions shifts policy attitudes. Still, by distinguishing local social contact from national media effects, it reveals mechanisms that most journalism on racial attitudes has overlooked. The result is a clearer picture of how everyday experience and information diets sustain the perceptual gaps that polarize American society.