Table 1.
Exam percent scores (% mean ± standard error, and range) for the entire sample and for students grouped by their z-scores on the exams.
Fig 1.
Exam 1 score strongly and significantly predicts subsequent exam scores (defined as the average percent score of Exams 2, 3, and 4) and explains approximately 74% of the variance (R2 = 0.737).
The 95% confidence interval for the regression line is also shown.
Table 2.
The three-factor structure from exploratory factor analysis on the SRL1 survey responses (n = 340).
Fig 2.
Students with higher exam 1 scores reported higher factor scores for use of course materials and metacognitive strategies, but not for housekeeping strategies, compared to students with lower exam 1 scores (n = 345).
Median scores for (a) housekeeping strategies, (b) use of course materials, and (c) metacognitive strategies by exam 1 z-score group. An asterisk in parentheses indicates a statistically significant difference (Kruskal-Wallis H test); lettered superscripts show which groups significantly differ from one another (post-hoc Mann-Whitney U tests; see S3 Table).
Fig 3.
Sankey diagram representing exam performance trajectories across the semester, for students in Groups 1, 2, or 3 on exam 1 with complete data across the SRL surveys (in aggregate, n = 258).
Trajectories are shown with exam 1, exam 2, and the average of exams 3/4 as points of interest. From exam 1 to exam 2, improvement is shown in green (moving up at least one z-score group), maintaining similar performance in blue (remaining in same z-score group), and decreasing performance in orange (moving down at least one z-score group). From exam 2 to the average of exams 3/4, students who either maintained their improvement or improved on exams 3/4 are shown in green, those who maintained similar performance or returned to their baseline performance by exams 3/4 are shown in blue, and those who maintained lower performance or decreased on exams 3/4 are shown in orange. The thickness of the lines is proportional to the number of students on each trajectory.
Fig 4.
Students who improved from exam 1 to exam 2 (n = 67, panel “a”) significantly increased their use of housekeeping strategies and course materials from the SRL1 to SRL2 survey; students who did not improve (n = 191, panel “b”) significantly increased their use of course materials only. Median scores for self-reported use of housekeeping strategies, course materials, and metacognitive strategies on surveys SRL1 and SRL2, for the subsets of students from Groups 1, 2, and 3 who (a) improved from exam 1 to exam 2 (n = 67) versus (b) did not improve from exam 1 to exam 2 (n = 191). Housekeeping strategies and use of course materials each contain three strategies (maximum possible score of 15) whereas metacognitive strategies contain four (maximum possible score of 20). An asterisk in parentheses indicates a significant difference in strategy use for a given subset of students (Wilcoxon signed-ranked tests).