This investigation of undergraduates' heterogeneous science identity trajectories within a gateway chemistry course identified three latent classes (High and Stable, Moderate and Slightly Increasing, Moderate and Declining) using growth mixture modeling. Underrepresented minorities were more likely to exhibit Moderate-and-Slightly-Increasing science identities versus High-and-Stable patterns. Students with higher perceived competence were more likely classified into the High-and-Stable class compared to the other classes. Students classified into the High-and-Stable class scored significantly higher on the final exam and appeared to be more likely to remain in a STEM major across fall and spring semesters compared to the other two classes. Results suggest that some students' identities shift within a single semester and supporting science perceived competence before college may support students' science identity development.
Keywords: STEM pursuit; expectancy-value; growth mixture modeling; science identity development.