Data Mining Temporal Work Patterns of Programming Student Populations

This paper reports the second stage of a study of the correlations between the temporal work patterns of computer programming students and their success or failure as measured by programming project assignment grades and related metrics. The first stage confirmed the importance for most students of getting an early start on a programming project, and it also uncovered the fact that some student groups perform well with late starts, suggesting the likelihood that they engage in the productive practice of active procrastination. The second most important factor for success is the average length of assignment work sessions. Session lengths from 60 to 120 minutes appear to be optimal for most students. Other contributing factors include total time spent on a project and working more day than night sessions. This second stage more than doubles the amount of data collected and analyzed. It finds that procrastination and session length remain prominent, while secondary factors become slightly less prominent. Its primary contribution is the analysis of within-student patterns for students who perform significantly better on some assignments than others, finding that for these students, starting early and maintaining appropriate work session lengths and times of day correlate with better performance