ReducingStereotypeThreat.org

Stricker & Bejar, 2004

In this study, the difficulty of items constituting a standardized test were varied to assess the impact on stereotype threat. Given that research suggests that stereotype threat is more pronounced on difficult than easy tests (Spencer, Steele, & Quinn, 1999), these researchers hypothesized that reducing the difficulty of items on the GRE General Test might reduce or eliminate stereotype threat effects. The sample included White and Blacks students of both genders who were college seniors or first-year graduate students and who agreed to take a computerized GRE General test and to answer some questions about the test. Students completed either the standard test (control) or a test that was comprised of items one standard deviation lower in terms of difficulty (experimental). Analysis of test performance showed that Whites performed better than Blacks and men performed better than women, and this effect was not qualified by test difficulty. Despite the lack of performance effects, White students who took the easier test did report less state anxiety, less test anxiety, higher state self-esteem, and higher effort than did those students who took the standard test. Black students taking the two tests did not differ on these measures. Women who took the easier test reported less test anxiety than those who took the standard test, but men who took the different tests did not differ on this measure. Men who took the easier test reported expending more effort than those men who took the standard test. These results failed to replicate the findings regarding test difficulty and stereotype threat obtained in lab settings.