Psychology › Social psychology › Social cognition: how we judge people
Stereotype threat
1995· Finding or experiment
Mixed evidenceSome support; effects smaller or less consistent than once thought.
The idea that worry about confirming a negative stereotype about your group can drag down your test performance.
See it on the Psychology map →What it says
In 1995 Claude Steele and Joshua Aronson gave Black and White Stanford students hard questions from the GRE, a graduate-school admissions test. When told it measured intellectual ability, Black students scored lower than White students with similar SAT scores; when told it was a problem-solving exercise, the gap disappeared. Similar results followed for women in math and older adults on memory tests. Hundreds of studies and school programs meant to counter the effect followed.
Why it matters
It offered a hopeful, situational explanation for group score gaps and changed how tests are introduced. It now also shows how a celebrated lab effect can shrink on retesting.
The evidence
Status: mixed evidence. Many lab studies found it, but meta-analyses detect publication bias, and a 2019 preregistered Dutch study of about 2,000 high-school students found no effect on girls' math scores.
Open questions
Why early studies found strong effects and recent large ones do not: publication bias, a changing social climate, or differences in how threat is created.
Connections (2)
▸Stereotype threat was claimed to lower scores on IQ testsMixed evidence
The connection
The 1995 study used hard GRE verbal questions, a cousin of intelligence tests. Media and some textbooks then said stereotype threat explained racial gaps in test scores. But the study had statistically equalized students by prior SAT scores, so a gap existed even without threat. Steele and Aronson themselves did not claim it explained the whole gap.
Who showed it and when
Steele and Aronson at Stanford, 1995; Paul Sackett, Chaitra Hardison and Michael Cullen's critique in American Psychologist, 2004.
How solid is it?
Lab studies support some score drop under threat, but evidence that it affects real, high-stakes tests, where decisions are actually made, is weak.
Open questions
How much of any real-world score gap it explains; a 2004 study of actual AP Calculus exams found asking gender before the test made negligible difference, a conclusion a 2008 reanalysis disputed.
▸Stereotype threat is a proposed effect in Social cognition: how we judge peopleMixed evidence
The connection
Social cognition usually studies how we judge others; stereotype threat turned the lens inward, to how knowing others' stereotypes affects us. The proposed mechanism is worry and self-monitoring that use up working memory. It inspired school interventions, such as brief writing exercises on personal values, which early studies said narrowed achievement gaps.
Who showed it and when
Claude Steele at Stanford, 1990s; Toni Schmader and Michael Johns's 2003 working-memory study; Geoffrey Cohen's 2006 values-writing study in a northeastern US middle school.
How solid is it?
Many lab studies; but bias-corrected meta-analyses and large preregistered trials find smaller, less reliable effects, both for the threat and for the school interventions.
Open questions
Whether anxiety, effort or lowered expectations drive it, and whether successful interventions work through stereotype threat at all.