Search

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. (Contains 1 figure and 4 tables.)

Sex-related differential studies on mathematical abilities have hardly taken into account the mediator role of the verbal factor, which contrasts with the interest shown in the mediator role of visuospatial aptitude. We predicted that if sex-related differences were found, mental rotation would mediate mathematical abilities typically favoring males (geometry and verbal problems) and lexical access would mediate the one favoring females (arithmetic). Data from 455 participants with a median age of 13 were analyzed, showing that sex-related effect sizes in mathematical criteria were small, ranging from d = -0.16 to 0.18 (corrected by attenuation), as expected for unselected samples. Lexical access scores were consistent predictors for every mathematical subdomain, and mental rotation added to the prediction only for geometry and word problems.

Two psychometric tests measuring mental rotation (MR) and visualization (VZ) were administered to 209 male and 390 female high school seniors. Results clearly indicate that MR is a plausible mediator variable for sex differences in VZ when such differences do exist. Theoretical, methodological, and practical consequences of these results are discussed. (SLD)