Patterns in assignment submission times: Procrastination, gender, grades, and grade components

In this study we characterize student procrastination habits and investigate associations between these habits and student performance on graded course components, student beliefs about their own procrastination behavior, and gender.The procrastination habits of calculus-based introductory physics students are measured via the amount of time before the assignment deadline or "completion time" that students submit their work on relatively short (>30 min) weekly online assignments.With the aid of latent profile analysis, we find that one can meaningfully categorize students into 4 completion time classes that clearly distinguish students between their mean completion time, their week-to-week completion time patterns, assignment completion rates, mean course grades, and proportion of women.Consistent with many studies in a variety of contexts, we find that procrastinating students tend to have lower course grades.Closer examination of exam and nonexam grade components reveals that completion time is directly associated with the nonexam component, but its association with the exam component is weaker and completely mediated by the nonexam component grade.This is in contrast to student ACT score, which is strongly associated with exam component but only weakly associated with the nonexam component, and the direct association of ACT with the exam scores is only weakly mediated by nonexam scores.Further, we find that ACT score is at best very weakly correlated with completion time.Taken together this supports the idea that exam and nonexam components are separately predicted by the two somewhat "orthogonal" measures of ACT score and completion time, and we propose that these are measuring so-called cognitive and noncognitive factors, respectively.Regarding gender differences, we found that on average women tended to procrastinate less than men, submitting the assignments on average 8 h earlier than men.Considering previous studies documenting that women tend to score higher than men on nonexam components, we found that completion time completely mediates the gender differences in nonexam components, providing support for the hypothesis that procrastination mediates the gender differences in performance on nonexam components.Finally, we found that the overwhelming majority (90%) of students did not strategically ("actively") intend to delay completion of the assignment, and that students who did indicate actively delaying were 2-3 times more likely to receive a D or E in the course.