Objective. Taking an institutional approach to the determinants of outcomes for women in science and engineering, we examine the effects on women's percentages among undergraduate majors and among degree recipients of four basic factors: (1) the percentage of faculty who are women in the students' major science/engineering area; (2) the students' disciplines (biology, physical sciences, and engineering); (3) the type of institution in which students are enrolled (“Research I” vs. others); and (4) a time trend (1984–2000).

Method. We use longitudinal, multivariate, and multi-institutional data from the Integrated Postsecondary Data System (IPEDS) and from a mail survey of registrars.

Results. Over the observation period, the women's percentages have risen steadily. The effects of disciplines and departments are stronger than those of institutions. Especially notable is that the percentages of women among undergraduate science/engineering majors and degree recipients are associated with the percentages of women among the faculty in these fields.

Conclusion. The findings contribute empirically to the discussion about the effects of “role models” for the participation and performance of women in science and engineering—and point to the strong effects of departments, compared to institutions, in accounting for degrees awarded to undergraduate women.