Impact of Role Model Gender and Communality on College Women’s Math Performance and Interest in STEM

The current project examined the effect of science, technology, engineering, and math (STEM) role model gender and communality and communal goal endorsement on college women’s math performance and their interest in, perceived belonging in, and perceived communal goal affordance of STEM. The first study was conducted online and did not include math performance, while the second study was conducted in the laboratory and added a math assessment. It was hypothesized that exposure to a female role model would result in greater math performance and interest and perceiving belonging in STEM, exposure to a communal role model would result in increased interest in, perceived belonging in, and perceived communal goal affordance of STEM fields, and gender and communality would interact such that gender would only make a difference for communal role models. In addition, it was hypothesized that the effects of role model communality would be greater for participants with higher communal goal endorsement. The hypotheses were not supported. There was no effect of role model communality, and effects of role model gender were not consistent across the two studies. Contrary to predictions, communal goal endorsement was positively related to interest and perceiving belonging in STEM in both studies. GENDER AND COMMUNALITY OF STEM ROLE MODELS 4 Impact of Role Model Gender and Communality on College Women’s Math Performance and Interest in STEM Although women receive bachelor’s degrees at a higher rate than men, men earn degrees at a greater rate in several areas of science, technology, engineering, and math (STEM) (National Science Foundation [NSF], 2013). Women have high participation in psychology, representing 70 to 79 percent of psychology degrees earned in 2010 (depending on the type of degree), and medium participation rate in biosciences and social sciences, representing between 47 and 56 percent of social science degrees and 52 to 58 percent of bioscience degrees. On the other hand, women have medium to low levels of participation in math and physical sciences and low participation in computer science and engineering, earning 33 to 41 percent of degrees in physical sciences, 30 to 43 percent of degrees in math, 18 to 23 percent of engineering degrees, and 18 to 28 percent of computer science degrees. Furthermore, the percent of bachelor’s and master’s degrees in computer science awarded to women has decreased from 2000 to 2010 from 28 percent to 18 percent of bachelor’s degrees and from 33 percent to 28 percent of master’s degrees. In terms of the makeup of the workforce, men made up 72 percent of the science and engineering workforce in 2010 (NSF, 2013). A recent review examined several potential causes of women’s underrepresentation in STEM fields (Ceci, Williams, & Barnett, 2009). The review indicated that there is little evidence that there are genetic or biological differences in ability, and, indeed, if representation in math-intensive careers was based solely on math ability, women’s representation would double, demonstrating that ability cannot be the primary cause of women’s underrepresentation in STEM. The review concluded that four main factors contribute to women’s underrepresentation in math-intensive careers. First, more men than women score at the highest end on math or quantitative sections of gatekeeper tests such as SAT and GRE. This gender GENDER AND COMMUNALITY OF STEM ROLE MODELS 5 difference on the GRE-Q is particularly problematic because it leads to men being admitted to graduate programs in math-intensive fields at greater rates than women. Second, women with high math competence are disproportionately likely to also have high verbal competence, compared with men, which allows women with high math competence more career options in non-STEM careers. Third, math-proficient women are more likely than math-proficient men to prefer careers in non-math-intensive fields, and for the women who do enter math-intensive fields, they are more likely than men to leave those careers as they advance. Finally, in some math-intensive fields, women are penalized in promotion rates for having children (Ceci et al., 2009). The Ceci et al. (2009) review and another review by Ceci and Williams (2011) both conclude that the main reason for women’s underrepresentation in STEM is women’s preferences for non-STEM careers. Ceci and Williams (2011) suggest that role models may increase girls’ interest in STEM, and another review (Halpern, Aronson, Reimer, Simpkins, Star, & Wentzel, 2007) recommends specifically female role models as a technique for encouraging girls to explore STEM, though it notes that the evidence for this is minimal. The review calls for more research to determine if role models are effective at increasing girls’ interest in STEM and thus women’s representation in those fields, as well as what aspects of role models are key to boosting girls’ and women’s interest in STEM. Research shows that role models must be relevant, and their success must be perceived as attainable in order for them to be motivational (Lockwood & Kunda, 1997). The stereotype inoculation model (Dasgupta, 2011) describes how both expert and peer role models can protect ingroup members’ self-concept from belonging threats in domains in which there are negative stereotypes about the group. Essentially, the presence of ingroup members in an academic or GENDER AND COMMUNALITY OF STEM ROLE MODELS 6 professional context can improve sense of belonging, self-efficacy, attitudes, and identification, and can change the perception of a domain in which the ingroup is negatively stereotyped from that of a threat to that of a challenge. Those changes in turn can increase effort, performance, and active participation, and influence individuals’ career decisions. The stereotype inoculation model also suggests that similarity and identification with role models are important moderators of these effects (Dasgupta, 2011). In addition, role model gender has been found to be important for inspiring females, but not males (Lockwood, 2006). A number of studies have examined the effects of female role models for girls and women in math and other STEM domains (Cheryan, Drury, & Vichayapai, 2013; Cheryan, Siy, Vichayapai, Drury, & Kim, 2011; Marx & Roman, 2002; Stout, Dasgupta, Hunsinger, & McManus, 2011). One study examined the impact of the gender of a math role model on women and men’s math performance and performance state self-esteem, and found that women and men had equivalent math performance and self-esteem with the female role model, but men performed better and had higher self-esteem with a male role model (Marx & Roman, 2002). Two follow-up studies used only female role models, but manipulated perceptions of competence, and tested the effects on math performance, performance state self-esteem, and math self-efficacy. Women performed better with a competent female role model, whereas men performed worse with a competent female role model. One of the studies found that the improvement in women’s performance was due to increased accuracy, while the other found that it was due to both higher accuracy and attempting more problems. Math self-efficacy was higher for female participants in the competent female role model condition, compared to the noncompetent female role model condition. The effect of female role model competency on women’s state self-esteem varied between the two studies, with one study finding that women GENDER AND COMMUNALITY OF STEM ROLE MODELS 7 had lower performance state self-esteem in the competent female role model condition than in the non-competent female role model condition and the other finding the opposite pattern (Marx