Gender effects in education revisited: Gender roles, instrumental and expressive traits, and motivational factors for STEM high school course selection

The present study investigates how gender role attitudes and instrumental and expressive traits can predict STEM course choices of male and female students in upper secondary education. Moreover, it is examined whether this relationship is mediated by mathematical self-concept, interest and anxiety. Using two waves of the BIJU dataset (N = 6,507), latent path models showed that gender role attitudes were not associated with the choice of a STEM course. For both genders, mediation effects of expressive and instrumental traits and STEM course choice emerged, mediated primarily by self-concept. Female and male students who described themselves as having more expressive traits exhibited lower self-concept and were less likely to choose a STEM course. When students described themselves with more instrumental traits, they were more likely to choose a STEM course - mediated by mathematical-motivational factors.