Stereotype threat as validity threat: The anxiety–sex–threat interaction

Stereotype threat has been invoked to explain underperformance on a variety of groups for whom the stereotypes allege inferior cognitive abilities. In math testing, stereotype threat has been used to explain sex differences in test performance. This paper describes an experimental study on a large sample (n = 313), in which the role of anxiety and mental rotation on math sex-related differences under stereotype threat has been analyzed in the tradition of Aptitude–Treatment–Interaction (ATI) designs. Results show that the manipulation effects are evident only in interaction with anxiety, which suggests that the construct validity of stereotype threat is itself seriously threatened. The analogy with the validity threat known as reactivity to the experimental situation is elaborated, and a more promising approach (the study of threat-related attentional bias in anxious and non-anxious individuals) suggested.