Work in progress - technology in plagiarism detection and management

The detection of intra-cohort plagiarism is often difficult in virtue of the quantity of material that must be compared. This can be limited by imposing constraints on the granularity of comparisons and through heuristic approaches to content comparison. Plagiarism sourced out with a group cohort is more difficult to detect, although Internet-based resources are the principal basis for such comparisons. This paper describes a survey of course work submissions from several computer science classes. The survey purpose was to determine the historical level and extent of plagiarism across electronically submitted assignments. This paper also describes our plans for using automated means of detecting plagiarism against future submissions. An approach to document tagging is described which supports the detection of contamination across documents within a cohort.