Treffer: Automatic Plagiarism Detection with PAIRwise 2.0

As part of the research carried out at the University of California, Santa Barbara's Center for Information Technology and Society (CITS), the Paper Authentication and Integrity Research (PAIR) project was launched. We began by investigating how one recent technology affected student learning outcomes. One aspect of this research was to study the effect plagiarism detection had on student attitudes and behaviors in regards to plagiarism detection. As a means of executing this research, we developed PAIRwise, a plagiarism detection system for instructors, students, and researchers. In this paper we describe its design and explore its capabilities with regards to efficiency and effectiveness. Our findings show that PAIRwise can handle large document collections efficiently and find instances of verbatim plagiarism as well as similar passages with modifications to verb tense. Researchers have also approached us with proposals to use PAIRwise for purposes other than plagiarism detection. The original system, PAIRwise 1.0, however, was slow, hard to maintain and not easily used by researchers from disciplines outside of computer science. Therefore, this release of PAIRwise, version 2.0, has many capabilities and applications beyond plagiarism detection. (Contains 5 figures and 1 note.)

As Provided