Despite increasing evidence that negative stereotypes adversely affect the career development of women in science, technology, engineering, and mathematics, measuring their effects has been a challenge. This study focuses specifically on the sciences, since some, particularly physics, are notably imbalanced in terms of gender diversity. Our aim was to develop and begin to validate a measure of gender stereotypes in science, referred to as the Stereotype Threat in Science Scale-Gender (STSS-G). Participants consisted of 629 undergraduate women enrolled in biology, chemistry, and physics courses. Results of exploratory and factor mixture analyses yielded support for a 7-item scale consisting of two factors, social identity and identity threat, and five latent classes underlying a heterogeneous population. Results of structural equation modeling indicated that the two latent variables were significantly associated with scores on existing measures of science identity and stereotype threat. The construct, differential, and concurrent validity of the STSS-G are discussed.