John Homer Promotion Portfolio

This paper reports on a quantitative evaluation of data collected from student programming assignments in an introductory programming course, using 44,548 submissions by 192 students over six semesters. Student programs are gathered and automatically scored for correctness; students are permitted to repeatedly submit possible solutions until a specific deadline, and each submission is immediately scored and feedback given about correctness or any errors. We report on the process followed by students (total time spent per problem, number of submissions, rate of completion, etc.) as well as the results (program length, complexity, state space). Our study reveals some common characteristics in “typical” CS1 student solutions and provides some reflection on the assignment framework.