To stay, switch, or leave: A four-year longitudinal study of the situated and stable social influences on women’s STEM major choices

•Four years of data show “if,” “when,” and to “where” women in STEM changed majors.•Odds of changing to non– or other STEM major were highest through the second year.•Within-person fluctuations in stereotype threat increased the odds to change.•Within-person/between-person science identity lowered the odds of non-STEM change.•Stereotype threat weakened the science identity—non-STEM change relationship. Research on situational antecedents for women’s persistence is critical to advancing gender equity in science, technology, engineering, and mathematics (STEM) fields. To disentangle the influences of stable and situated aspects of motivational antecedents, we used survival analysis to predict if, when, and to where undergraduate women change majors (i.e., staying in, switching across, or out of STEM) from between-person average and within-person fluctuations in Tripartite Integration Model of Social Influence (TIMSI) motivational constructs (science self-efficacy, identity, and community values) and stereotype threat. Women (N = 413) STEM majors in their first or second year of college were recruited from nine U.S. universities and followed over four years. Women were most likely to leave STEM in the first year of college and were most likely to change STEM majors within the first two years. Major change was predicted by (a) between-person average and within-person fluctuations in science identity, (b) within-person fluctuations in stereotype threat, and (c) an interaction between average stereotype threat and fluctuations in science identity. These findings emphasize the importance of distinguishing between-person and within-person aspects of motivational antecedents of STEM choices and developing tailored motivational interventions for short- and longer-term periods.