Association for Public Policy Analysis & Management

The underrepresentation of women in STEM fields is one of the most pressing problems in higher education. The disparity exists around the world and contributes substantially to gender earnings inequality because STEM jobs are typically high paying. Of particular concern is that gender bias, discrimination and fear of competition contribute to fewer women taking courses and graduating in STEM fields. These negative influences may be heightened when female students interact frequently with male students in male-dominated STEM fields. Female college students in STEM report high levels of gender bias from male peers, and STEM-related gender bias from classmates is negatively related to STEM motivation and career aspirations. Female students also often report leaving STEM because of a negative climate characterized by intense competition, lack of support and discouraging peers, and female students are negatively affected by competitive environments in STEM classes whereas male students are not affected.

To investigate these concerns, we conducted a large-scale experiment of interactions between female and male students in an essential, gateway STEM course and one in which students work naturally as partners – first-year Chemistry laboratory classes. To our knowledge, it is the first experimental study to directly test whether female students are negatively affected in course performance and STEM continuation when interacting one-on-one with male peers in the classroom.

We selected general chemistry labs to study gender interactions for four primary reasons. First, students are graded independently on tests and assignments, but work closely together in pairs in the labs for the entire term instead of chance and infrequent interactions that occur in large lecture halls. Second, the Introduction to Chemistry sequence which includes the labs is the gateway requirement to the Sciences and a broader set of STEM majors. Third, chemistry labs develop a broad set of skills including laboratory technique, mathematical modeling and statistics. Finally, female students are underrepresented in Chemistry, which is similar to most other STEM fields. Overall, Chemistry labs provide an important setting to test whether female students experience negative interaction effects from male students early in their in their STEM college experience.

For the experiment, students in every lab section associated with the first-year sequence in Chemistry at a large public research university were randomly assigned a partner. The study involves 3,902 students over four academic years (total N=5,537). The experiment specifically tests whether the academic outcomes of female college students are negatively affected by being paired with male students, and whether those interactions depend on the student’s ability, partner’s ability, and whether they have a female graduate student teaching assistant. We also determined downstream effects of various pairings by examining whether pairing with male students inhibits continuation to more advanced chemistry classes or ultimate selection of a STEM major. Randomly assigning students to lab partnerships allows us to avoid the common estimation bias resulting from self-selection (i.e. students might choose to work with their friends, other students like themselves, or students who can help them the most) and several additional statistical limitations.