Who is Like a Scientist? A Self-Prototype Matching Approach to Women's Underrepresentation in STEM Fields

Abstract

Why are women approaching parity in some science, technology, engineering, and math (STEM) majors but still lagging behind in others? I construct a multidimensional scaling map to understand which clusters of majors appear most unattractive to female students, and found that fields seen as scientific rather than cultural and as related to artificial kinds rather than nature were seen as least interesting by women. I then investigated the stereotypes which make those majors seem unappealing. I approach this from a self-prototype matching context, meaning that I predicted that participants would be most interested in majors whose stereotypes matched their self-image. A multi-level model was employed to examine the within-subjects patterns of response to various majors given stereotypes about those majors. Participants did in fact prefer majors whose stereotypes were similar to their self-stereotypes, and this tendency mediated gender differences in interest. In particular, stereotypes about agency, communion, and technological or "geeky" recreational interests were strong mediators. I also examined the consistency across subjects of these effects. Some stereotypes (such as agency) were high-consensus, meaning that subjects consistently took them into account, expressing more interest in majors that were similar to them in those stereotypes and less in majors which were not. Others (such as sociability) were low-consensus, meaning that some subjects took them into account very strongly and others very weakly. Taken together, these results suggest that researchers studying stereotypes should expect the effects of those stereotypes to vary between subjects, and to depend particularly on those subjects' self-perception. Researchers studying the low-consensus stereotypes here should be particularly aware of this possibility, but all should take into account that the same intervention may have opposite effects for participants with opposite self-ratings. This dissertation advocates for interventions which emphasize the diversity of STEM fields and suggest that more than one prototypical class of student can succeed in them.