Stereotype Threat in School Assessment: What the Evidence Says and What Schools Can Do

In 1995, psychologists Claude Steele and Joshua Aronson at Stanford University published a paper in the Journal of Personality and Social Psychology titled "Stereotype threat and the intellectual test performance of African Americans." The study documented a striking phenomenon: students who were reminded of a negative stereotype about their group before taking a test performed significantly worse than students from the same group who were not given that reminder.

The effect was not caused by ability. The students in both conditions had the same academic background and prior test scores. The performance difference was caused by the psychological burden of being aware that their performance might confirm a negative stereotype, a burden that consumed cognitive resources and disrupted performance on the test itself.

Subsequent research expanded the finding significantly. Stereotype threat has been documented across a wide range of groups and contexts: women in mathematics, students from lower socioeconomic backgrounds in academic settings, older adults on memory tasks. The phenomenon is robust, well-replicated, and has direct implications for any context in which assessment is high-stakes and group membership is salient.

What this means for school admissions

School admissions assessment is exactly the kind of high-stakes context where stereotype threat operates. An applicant who is aware that students from their background are not typically successful at this type of school, or who has received messages about their academic ability that prime a sense of inadequacy, may perform below their actual capability in the assessment environment.

This is particularly relevant for international school admissions, where the applicant pool is highly diverse in terms of cultural and educational background. An assessment that is designed for one cultural and linguistic context, administered in ways that signal its alignment with that context, and framed in competitive terms may inadvertently suppress the performance of students whose backgrounds are less aligned with the dominant culture of the school.

What assessment design can do

Research on stereotype threat has also documented interventions that reduce its effects. Several have direct applicability to assessment design in school admissions.

One of the most consistently effective interventions is reframing the assessment purpose. When test-takers are told that the assessment is a research tool being developed and refined, rather than a definitive measure of their ability, performance differences associated with stereotype threat are substantially reduced. For school admissions, this suggests that framing the assessment as one piece of evidence in a holistic review, rather than as the decisive test, may improve fairness.

A second intervention is normalising the experience of difficulty. Steele's research found that when students were told in advance that the test was challenging and that feeling challenged was a normal part of the experience, stereotype threat effects were reduced. In the Evalent assessment, each section opens with a short video that introduces the task and normalises the effort involved, this is one function that the video guides serve.

A third intervention is reducing the salience of group identity at the point of assessment. Assessment contexts that do not ask students to identify their ethnicity, nationality, or first language immediately before beginning the test avoid priming the group identities that activate stereotype threat. Demographic questions, where needed, are better placed after assessment completion than before it.

The limits of assessment design

Assessment design can reduce but not eliminate the effects of stereotype threat. The broader culture of a school, whether students from diverse backgrounds see themselves represented in the faculty, the curriculum, and the school's public identity, has a larger long-term effect on whether students from underrepresented groups feel that they belong and perform accordingly.

For admissions teams, the implication is that assessment results should be interpreted with an awareness of the contexts that shaped them. A student whose performance may have been affected by assessment anxiety, language background, or the unfamiliarity of the testing format deserves to have those factors considered alongside the raw scores.

This is an argument for assessment that generates contextual signals, not just scores, and for assessors who are equipped to interpret those signals thoughtfully.

See how Evalent approaches evaluation fairness.