Abstract:
Women are less likely than men to choose mathematically intensive STEM careers (Eurostat, 2017; National Science Foundation, 2017). Because there are usually high status career options in these fields, such gender differences are critical for gender equity in rights and opportunities (European Commission, 2013; Noonan, 2017). Eccles et al.’s (1983) expectancy-value theory provides a powerful framework for investigating the question of why fewer women than men enter these careers. According to Eccles et al. (1983, 2009), gender differences in mathematically intensive STEM careers can be linked to early emerging gender differences in expectancy and value beliefs in mathematics, which are themselves rooted in gendered socialization processes. There is ample empirical support for these assumptions, and a variety of individual and environmental factors have been found to be associated with women’s lower participation in mathematically intensive STEM fields compared to men (see Wang & Degol, 2013; Wigfield et al., 2015). Nevertheless, important questions remain concerning the influence of critical individual and environmental factors. The present dissertation sought to address some of these questions: Using Eccles et al.’s (1983) expectancy-value theory as a guiding framework, the effects of two environmental factors that have the potential to reduce women’s participation in mathematically intensive STEM careers were investigated, namely gender-math stereotypes in the media and high school coursework requirements. Furthermore, expectancy and value beliefs, the most proximal constructs associated with STEM careers in Eccles et al.’s (1983) expectancy-value theory, were compared to prominent predictors in the area of work and vocational psychology (i.e.,
vocational interests) in terms of their relative predictive power on gendered STEM careers.
To this end, three empirical studies were conducted as part of the present dissertation: Study 1 investigated the influence of gender stereotypes embedded in a children’s television program on girls’ and boys’ stereotype endorsement, math motivation, and math performance. For this purpose, a randomized study was conducted with a total of 335 fifth graders. Watching a short clip of gender-math stereotypes embedded in a television program increased girls’ and boys’ stereotype endorsement. However, there was no effect on girls’ math motivation, and only a small effect on boys’ motivation. Furthermore, no effects on math performance were found for either girls or boys. Thus, there was only partial support for short-term effects of gender stereotypes embedded in a television program on girls’ and boys’ math motivation and performance.
Study 2 investigated whether encouraging young women to take advanced math courses in high school might bring more women into STEM careers. Young women are less likely than men to choose advanced math courses in high school, which has been linked to gender differences in STEM careers (Ceci, Ginther, Kahn, & Williams, 2014; Watt & Eccles, 2008). Therefore, the effects of a statewide educational reform in Germany that required all students to take advanced math courses were investigated. To this end, data from 4,730 students who participated in high school courses before the reform were compared with data from 4,715 students who participated in high school courses after the reform. The reform was associated with different effects for young women and men: Gender differences favoring young men were smaller in math achievement, but larger in math self-concept and STEMrelated
vocational interests (i.e., realistic, investigative) in the cohort after the reform. Gender differences favoring young men in the choice of STEM majors at university did not differ between the two cohorts. Thus, it seems that encouraging young women to take advanced math courses in high school does not automatically increase gender equality in mathematically intensive STEM fields.
Study 3 examined the relative predictive validity of expectancy-value constructs and vocational interests for gender differences in mathematically intensive STEM careers. Both sets of constructs are highly predictive of gender differences in mathematically intensive STEM careers (Schoon & Eccles, 2014; Su, Rounds, & Armstrong, 2009), but their relative predictive validity is unclear so far. To address this question, longitudinal data from 4,984 students in Germany at the end of high school and two years later were reanalyzed. Both expectancy-value constructs and vocational interests predicted math achievement and the choice of STEM majors, but there were important differences in their predictive validity: Whereas expectancy-value constructs were more predictive of math achievement, vocational interests were more predictive of the choice of different STEM majors. Gender differences at
the mean levels of expectancy-value constructs and vocational interests partly explained gender differences in math achievement and the choice of a STEM major. Furthermore, whereas the predictive power of expectancy-value constructs for math achievement and the choice of different STEM majors was invariant over gender, the predictive power of vocational interests differed slightly between young women and men. Thus, expectancy-value constructs and vocational interests seem to contribute differently to young men’s and women’s STEM career pathways.
The findings of the three studies are summarized and discussed within the broader research context. All three studies contribute to a better understanding of gender differences in mathematically intensive STEM careers, but also highlight the complexity of factors
involved.