Musings on misconduct

Abstract

Tools for algorithmically detecting plagiarism have become very popular, but none of these tools offers an effective and reliable way to identify plagiarism within academic software development. As a result, the identification of plagiarism within programming submissions remains an issue of academic judgment. The number of submissions that come in to a large programming class can frustrate the ability to fully investigate each submission for conformance with academic norms of attribution. It is necessary for academics to investigate misconduct, but time and logistical considerations likely make it difficult, if not impossible, to ensure full coverage of all solutions. In such cases, a subset of submissions may be analyzed, and these are often the submissions that have most readily come to mind as containing suspect elements. In this paper, the authors discuss some of the issues with regards to identifying plagiarism within programming modules, and the ethical issues that these raise. The paper concludes with some personal reflections on how best to deal with the complexities so as to ensure fairer treatment for students and fairer coverage of submissions.