Claude Steele

Who they are

Steele is credited on this site as the originator of both constructs. The vault credits first authors, and in the 1995 stereotype threat paper he is the first author. The vault records him because a popular book used his stereotype threat construct as settled terminology across several pages and cited him nowhere, and because the research sense of affirmation needs its originator named to be kept apart from the self-help sense.

This page records only what the papers and the published record state. Biography beyond the cited work is not given here, and nothing on it is a judgement of him.

Key papers

The psychology of self-affirmation: sustaining the integrity of the self (1988). A chapter in Advances in Experimental Social Psychology, volume 21, pages 261 to 302. It is a theoretical chapter, not an empirical paper and not a meta-analysis. It proposes that people are motivated to maintain an image of themselves as morally and adaptively adequate, and that this image can be restored by affirming any valued domain, not only the one that was threatened.

Stereotype threat and the intellectual test performance of African Americans (1995), with Joshua Aronson. Published in the Journal of Personality and Social Psychology, 69(5), pages 797 to 811. The first two studies varied whether a difficult verbal test was described as diagnostic of ability. A later study found that making the stereotype salient could impair performance without a diagnostic framing. Per-study sample sizes were not confirmed in the vault's verification pass, so no figure is carried here.

Key ideas

Self-affirmation. The idea is that a threat to the self can be absorbed by reflecting on something else the person values. The reflection need not touch the threatened area and contains no statement about a desired outcome. The research that grew from it has measurable but small effects. A 2015 meta-analysis of 144 experimental tests in health behaviour found small but reliable positive effects. A 2021 meta-analysis of 58 classroom studies found a small benefit for students under identity threat and a negligible one for the rest. A well-powered 2017 replication in schools found none, and its authors concluded that the exercise may be fragile at scale.

Stereotype threat. The proposal is that when people know a negative stereotype about a group they belong to, and are placed in a situation where they might confirm it, their performance can suffer. The theoretical contribution is genuine and durable, and the later state of the literature is not a reason to withdraw credit for the idea.

The record since, stated as the record. In 2004 Paul Sackett, Chaitra Hardison and Michael Cullen showed that the 1995 paper is widely misreported as demonstrating that removing stereotype threat removes the black and white test-score gap. It does not, because the analysis adjusted for prior test performance, so threat adds to an existing gap and does not account for it. That is a finding about how the paper has been read, not a criticism of the paper. Later work puts the pooled effect at a smaller size than the early studies suggested. A registered report across 2,064 Dutch pupils found no overall and no moderated effect, a meta-analysis of 212 adult samples put the effect in realistic testing conditions at 0.14 standard deviations and 0.09 after bias correction, and a meta-analysis of schoolgirls reported 0.22, falling substantially after adjustment. The critics state their own scope limit: they do not question the existence of the phenomenon, only the evidence that it systematically affects cognitive ability tests in high-stakes settings.

---