Countering Gender Stereotypes to Increase Adolescent Girls’ Interest in Enrolling in Computer Science Courses

• In Event: Tracing the Source of Gender Disparities in STEM: Children’s Macro- and Micro-level Sociocultural Contexts

Abstract

Introduction
Motivating adolescent girls to enroll in computer science courses is critically important. Girls represent 29% of Advanced Placement Computer Science test-takers and earn only 19% of computer science bachelor’s degrees (NSF, 2019; Code.org et al., 2020). The presence of negative stereotypes about girls contributes to gender disparities in STEM (Spencer et al., 2016). Gender-interest stereotypes reduce girls’ interest in computer science (Master et al., 2021). Gender-interest stereotypes involve beliefs that boys are more interested than girls in computer science. To neutralize these stereotypes, we need to understand how to counter cues that communicate stereotypes. One cue may be gender representation within courses. If girls see that only 20% of students in a course are girls, they may assume girls are less interested than boys, and that they themselves will not be interested. Explicitly counteracting gender-interest stereotypes could reassure girls that they may enjoy these courses despite their underrepresentation. Middle school students are choosing career paths, making this an ideal stage to influence interest in computer science before choosing high school electives (Maltese & Tai, 2010). This study examined whether providing information that countered gender-interest stereotypes would increase girls’ interest in enrolling in a computer science course.
Method
Participants were 530 8th grade students (63% White, 16.8% Asian, 8.1% Black, 5.5% Multiracial, 0.2% Native American, 6.6% no response; 4.9% were Hispanic; 47% girls, 53% boys). Participants learned about two computer science courses in a within-subjects design. For both courses, participants learned girls were underrepresented (i.e., five girls and 20 boys). For the course that countered stereotypes, participants were told, “girls and boys were equally interested at the end of the course.” For the other course, stereotypes were not countered, so participants were informed that students were not asked about their interest. Participants reported their interest and belonging (3-items for each scale), with alphas > .92 for each course. Participants chose which course they would prefer to take. Students were randomly presented with the courses via an online survey. Hypotheses and analysis plan were preregistered on the Open Science Framework.
Results
For interest, there was a significant interaction between condition and gender, F(1,507)=4.09, p=.044. The simple effect of condition for girls was significant, p=.011, dz =.16; girls were more interested in the course with stereotypes countered. For belonging, there was a marginally significant interaction between condition and gender, F(1,501)=3.11, p=.078. The simple effect of condition for girls was significant, p<.001, dz =.25; girls felt greater belonging in the course with stereotypes countered. In terms of choice, a chi-square test between gender and condition was not significant, X2(1,N=510)=0.52, p=.47, but girls were above chance in choosing the countered course (63% of girls), p<.001. Belonging mediated effects on girls’ interest, indirect effect=0.07, 95%CI [0.03,0.13].
Discussion
We found that countering stereotypes (by providing positive information about girls’ interest when they are underrepresented) boosted girls’ belonging and interest. However, this effect was small. The discussion will center around other ways to increase the impact of countering stereotypes to boost belonging and interest in real-world settings.

Authors

• Jennifer Thompson, University of Houston
  Presenting Author

• Taylor Mikhail Alexander, University of Houston

• Weihua Fan

• Allison Master, University of Houston