In earlier work, we found that women who reported experiencing the "discussion of underrepresentation" in their high school physics classes were more likely to report a career interest in the physical sciences in college. To explore this effect in more detail, we conducted two double-blind, random-assignment experiments on students enrolled in introductory, algebra-based college physics. In the first, students were randomly assigned to read one of two short essays (one focused on the underrepresentation of women in physics, the other reporting on AMO physics research led by a woman) followed by a uniform set of reflection questions. In the second, another cohort of students was primed on their beliefs about gender differences in physics before undergoing the same treatment (one of two essays and reflection questions). In this paper, we compare the impacts on students' general science identity, physics identity, and performance gains on selected FMCE problems.