Abstract
<p>Advanced Placement (AP) enrollment, examination participation, and examination performance are often treated as interchangeable, although they represent different stages of an advanced-coursework pipeline. This study distinguishes realized course-based opportunity, examination reach, conditional attainment, and successful attainment using 1,231 school-year observations from 247 comprehensive New Jersey public high schools, 2020–21 through 2024–25. Models with school and year fixed effects are combined with course-record validation, decomposition, and longitudinal profile analyses. Because students earning qualifying scores are nested within examinees, who are largely drawn from enrolled students, the positive enrollment–attainment association is partly structural. Its magnitude is the informative quantity: a 1-percentage-point within-school increase in reported formal enrollment is associated with a 0.33-percentage-point increase in successful attainment in the baseline model (0.32 with corrected composition controls). Course breadth and enrollment intensity strongly predict formal enrollment and examination reach but are weakly or negatively associated with conditional attainment. Annual-change models similarly show that greater examination reach predicts higher successful attainment after adjustment for enrollment change, whereas broader reach is associated with lower conditional attainment. This pattern is consistent with a compositional mechanism in which marginal examinees are, on average, less prepared than inframarginal examinees; it does not show that expansion reduces instructional quality. Net attainment nevertheless rises because gains in reach outweigh the decline in conditional success. The pooled cross-sectional slope is larger than the within-school estimate, indicating that stable between-school differences account for much of the unadjusted association. The study contributes construct-sensitive measurement and longitudinal evidence on the access–dilution tradeoff in advanced coursework.</p>