Аннотация
This study examines how the transition from direct external assessment in Grade 10 to indirect external assessment in Grade 12 is associated with longitudinal patterns of student performance in secondary science education and how internal laboratory-based assessments relate to performance continuity across assessment formats. Longitudinal data were analyzed for Biology ( n = 368), Chemistry ( n = 206), and Physics ( n = 983), including external assessments in Grades 10 and 12 and internal assessments in Grades 11 and 12. Sequential regression models for longitudinal associations, mixed-effects clustering and distributional analyses for performance variation, and robustness and sensitivity tests were conducted, with confidence intervals for all estimates. Resultsshowed consistently greater score dispersion in Grade 12 across all subjects, indicating differences in observed score distributions under the indirect assessment format. Prior external achievement was a significant predictor of Grade 12 outcomes, while internal assessments, particularly in Grade 12, were positively associated with external assessment performance and improved model explanatory power. Mathematics achievement was also positively associated with science outcomes, with the strongest relationships found in Chemistry and Physics. These findings indicate longitudinal associations between prior achievement, internal assessments, and later external outcomes across assessment formats. However, because the Grade 10 and Grade 12 assessments differ in construct emphasis, task structure, and cognitive demand, the results should not be interpreted as evidence of psychometric superiority, construct equivalence, or causal effects. The results highlight the value of combining laboratory-based internal assessment with externally assessed scientific reasoning and interpretation in science education.