Monitoring Novice Programmer Affect and Behaviors to Identify Learning Bottlenecks

We analyze student affect data in order to locate bottlenecks in an introductory programming course. By tracking students’ affective states and behaviors over five laboratory sessions distributed over nine weeks, we find that students exhibit a significantly greater amount of confusion when expected to implement object-oriented constructs such as constructors and object interaction. When asked to undertake an exercise similar in scope with a previous exercise, students expend less time and effort on the exercise. They exhibit less flow and frustration, and spend more time on off-task behaviors.