Online detection of source-code plagiarism in undergraduate programming courses

Plagiarism in programming courses has increased in recent years. Therefore, courses need to be adapted in concept and organization. A relentless pursuit of plagiarizing or the ineffective attempt to completely suppress all communication in the classroom are certainly as inadequate approaches as the installation of an offline plagiarism detection tool without a proper feedback mechanism for the students. Ideally, teamwork is allowed, but each student submits her/ his own solution. A “real-time” plagiarism detection tool gives immediate feedback on a submitted solution during class time. Based on typical indicators, which course instructors use to manually detect plagiarism, a new software system is developed. This system allows submissions of programs, which are automatically tested for proper functionality and checked against other submissions. An evaluation of the system with data sets from an undergraduate C-programming course reveals a high rate of plagiarisms while false detection (coincident similarity) rarely occurred.