A talented, innovative workforce in science, technology, engineering, and mathematics (STEM) is a critical component of sustained economic growth and global competitiveness. The development of this workforce is a primary concern among policymakers, industry leaders, and academics. Although many students express an interest in STEM in secondary school, many of them eventually choose not to pursue a degree or career in a STEM field. This trend has been linked to inadequate achievement, but also to lack of confidence, inconsistent interest, and shifting motivation. It is important that we understand the development of precollege socialcognitive factors affecting persistence to help identify whether some trajectories might have more desirable outcomes than others, and points at which intervention efforts might best be targeted. Growth mixture modeling was used in the current study to uncover unobserved developmental subgroups of students’ attitudes toward science and positive core self-concept through their middle and high school years. Three distinct subgroups of change patterns were found for each of mastery motivation, attitudes toward science utility, and science self-concept. Science Self-Concept subgroups demonstrated significant and reasonably distinct associations with relevant science achievement, postsecondary, and career outcomes, where the results for Mastery Motivation and Science Utility subgroups were mixed. Science Utility and Science Self-Concept subgroups of developmental trajectories both exhibited plausible and appropriate associations with parent and demographic factors as well as initial student, parent, and teacher expectations about college and career.