Why Time Constraints Increase the Gender Gap in Measured Numerical Intelligence in Academically High Achieving Samples
Abstract. Gender differences in the numerical domain vary greatly according to the assessment method used. We suggest that strict time constraints, as employed on most numerical intelligence tests but not on mathematical competency tests, unduly increase the gender gap in measured numerical intelligence if the test focuses reasoning. Two studies were conducted. First, 666 11th and 12th graders were randomly assigned to speeded or nonspeeded versions of verbal, figural, and numerical reasoning tests. Extending the test time reduced gender differences in numerical but not in verbal and figural reasoning. To rule out ceiling effects and to test for potential motivational and emotional effects on test performance, a second sample of 542 students completed both a speeded and a nonspeeded numerical reasoning test as well as several motivational and emotional questionnaires. In the nonspeeded condition, girls increased their performance more than boys. This effect was especially strong for female students with medium and high performances and was largely but not fully explained by emotional and motivational factors. We conclude that girls are prevented from showing their actual potential on speeded numerical reasoning tests.