Extending computer-assisted text analysis techniques to the detection of source code plagiarism and collusion: assisting manual inspection

Traditionally, a distinction has been drawn between software and services to detect text-based plagiarism or collusion, and products to detect such practices in computer programs. The Joint Information Services Committee (JISC) even commissioned two separate studies, the Technical Review of Plagiarism Detection Report (Bull et al, 2001) which focused on text-based assignments, and Source Code Plagiarism in UK HE Computing Schools, Issues, Attitudes and Tools (Culwin et al, 2001), which concentrated on source code plagiarism. The text-based survey evaluated the performance of five products with mixed results, while the source code survey evaluated two products, with different strengths and weaknesses. This separation of concerns is not spelled out explicitly in either report: both seem to be based on an assumption that the practice of plagiarism and collusion differs between textbased documents and examples of computer programs written in a particular programming language.