How Much is Enough? Formative Assessment Dynamics: Quantifying the Relationship Between Intermediate Quiz Performance and Final Exam Scores

Abstract

While the educational value of formative assessment is widely acknowledged, the precise amount needed to effectively predict student performance on summative assessments remains unclear. This study investigates the relationship between intermediate formative assessment performance and final exam scores, addressing the critical question of how much assessment is needed for accurate prediction. Using a large dataset encompassing over 20,000 student enrollments across 127 course runs of 15 online biomedical sciences courses, we examined the correlation between intermediate assessment scores and final exam performance. Our results show that after completing about 40% of the formative assessments in a course, student scores demonstrate a strong correlation (Pearson r > 0.7) with their final exam scores. The correlation after taking additional formative assessments reaches a maximum of approximately 0.75. This finding was consistent across different course types and lengths, suggesting that the relative amount of assessment taken, rather than the absolute number, is key. Surprisingly, we found that random sampling of assessments was even more predictive than chronological sampling, suggesting that the proportion of questions used, relative to the total number of assessment questions, is more important than their specific sequence. These findings contribute to a deeper understanding of the predictive capabilities of formative assessment, and enable educators to identify at-risk students earlier, optimize assessment design, and develop more efficient and targeted interventions.