Comparison of Plagiarism Detection Performance between some Commercial and Free Software

Abstract

Introduction: The act of plagiarism is represented by using someone else’s information or research without the author’s consent and/or without the author’s full acknowledgement. Detection of plagiarism can be easily made using computer software that identifies fragments of texts as not original. Aim: This study aimed to highlight and compare the performance in detecting a specific type of plagiarism (copy-paste) in different types of documents between free and commercial software. Material and Method: A document of 808 words was created using eight fragments of texts from eight different sources. Two other versions of the document were then created: one with approximately 43% of the text similar and another with the entire text paraphrased. Seven software programs were used for the similarity analysis of each of the three texts (4 commercial and 3 free). Results: The original document showed differences in detection performances (Turnitin – 97% observed similarity, and Plagiarism detector – 93% observed similarity, had the highest performance). In the document with 43% similarity, the performance was affected across all programs, but Plagiarism detector had the best performance (43% observed similarity). None of the evaluated software could detect the original sources in the entire paraphrased document. Conclusion: Among the tested software, Turnitin proved to be the best commercial software and Plagiarism detector the best free software for testing academic documents similarity, differences between them being minimal.