Testing of support tools to detect plagiarism in academic Japanese texts

Abstract

Plagiarism has been among the top forms of academic misconduct. Detective, reactive and proactive measures are taken to mitigate plagiarism in scholarly works. Text-matching tools play a significant role in the detection of plagiarism. Many studies have tested the performance of text-matching tools in detecting plagiarism from various perspectives. However, no study addressed the performance of such tools in ideographic languages, particularly Japanese. Considering the sharp increase in the number of academic Japanese text and plagiarism incidents in the Japanese context, it is essential to explore to what extent text-matching tools catch similarities in Japanese texts and respond to the needs of Japanese users. Within this scope, this study set out to explore the coverage and usability performance of text-matching tools in the Japanese language. We tested the coverage performance of 10 text-matching tools with five types of intentionally plagiarized documents. Also, we tested the usability performance via a feature checklist. The testing results suggested that the tools generally give a relatively higher performance on the usability side rather than the coverage aspect. Most tools have minimal coverage performance in the Japanese language. In the end, we provided takeaways for vendors, policymakers and educators.