Why do men and women tend to perform differently on analytical portions of standardized tests? Psycho/social research often speculates that women's performance “might be more affected by such variables as role expectations or unjustified fears of incompetence” (Basinger, 1997, p. 2; see also Sternberg & Williams, 1997). This “unjustified fear” is similar to what Steele and Aronson (1995) call “stereotype threat” found among African American test takers. With a small number of subjects, and in laboratory conditions, Steele and Aronson found significant differences in test scores when they made only small changes in the directions for taking the test and in the explanations given to their subjects. Their research showed that, when African American college-level students were asked to take a test that had no direct consequence for them, their performance was equal to or better than that of majority test takers in the same group. However, when similar groups were told the outcomes of the same tests would affect them academically, performance levels among African American test takers dropped dramatically. According to Steele and Aronson, the perceived stereotypes associated with testing and other lab or classroom performances of women and minorities created this effect. Their findings suggest that hidden variables in the testing environment may have long-term effects on women and minority test takers.

Steele and Aronson's work clearly points to the impact of hidden variables such as these on test scores.