The purpose of the present study was to examine differences in engagement and achievement among 2,695 high students enrolled in (a) a STEM school, (b) a traditional school (no STEM program available), (c) a STEM program within a school, or (d) a traditional program within a school having the STEM program option. The High School Survey of Student Engagement was utilized to measure 3 dimensions of student engagement (cognitive, emotional, and social); grade point average (GPA) and standardized test scores were used to measure academic achievement. Students in the STEM schools and STEM programs academically outperformed students enrolled in the traditional schools or traditional program. Discriminant function analysis yielded 3 statistically significant (p < 0.05) roots. Of these, Root 1 was of a moderate size ([lambda] = 0.808) and worthy of interpretation. All achievement variables contributed appreciably to the root, whereas the contribution of engagement variables was negligible. Canonical correlation analysis was used to illustrate that the relationship between the engagement and achievement measures varied across the 4 instructional settings.