Pattern discovery tools for detecting cheating in student coursework

  • Authors:
  • David J. Hand;Niall M. Adams;Nick A. Heard

  • Affiliations:
  • Department of Mathematics, Imperial College London;Department of Mathematics, Imperial College London;Department of Mathematics, Imperial College London

  • Venue:
  • LPD'04 Proceedings of the 2004 international conference on Local Pattern Detection
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Students sometimes cheat. In particular, they sometimes copy coursework assignments from each other. Such copying is occasionally detected by the markers, since the copied script and the original will be unusually similar. However, one cannot rely on such subjective assessment – perhaps there are many scripts or perhaps the student has sought to disguise the copying by changing words or other aspects of the answers. We describe an attempt to develop a pattern discovery method for detecting cheating, based on measures of the similarities between scripts, where similarity is defined in syntactic rather than semantic terms. This problem differs from many other pattern discovery problems because the peaks will typically be very low: normally only one or two cheating students will copy from any given other student.