Issues Raised by Use of Turnitin Plagiarism Detection Software

Turnitin Plagiarism Detection Software Problems Last week, I prepared an opinion on theft detection software with some other members of the GVSU writing department. I and other teachers are worried that teachers from other disciplines may not know about the theft test service. Below is the full text of the statement posted on our campus.

To combat the increased use of the Internet in the classroom, educators have proposed a tool to detect plagiarism called Turnitin. According to an interview with NPR Cody Turner (2014) on the controversy behind anti-theft software, professors agreed that software like Turnitin should be used to introduce plagiarism for university students. Most of the errors found in this software are reference errors that can easily be fixed. Some students did not understand the estimate correctly and submitted the assignment without confirming that it was not correctly estimated. The biggest problem of anti-plagiar software is that it can not distinguish between coincidence theft and deliberate theft. Most professors believe that their students never intentionally copy articles, but this software does not have the ability to distinguish between two articles (Turner 2014).

Even using Web-based detection does not mean that theft is now only a law enforcement agency or a technical problem. While the theft detection system such as Turnitin's originality check compares student papers with publishers accessible to large student papers database and systems, the teacher analyzes the match identified in the report and steals it It is judged whether or not it corresponds to. The teacher also has to decide how to discuss the problem with the students.

Many high school students and university students are familiar with services such as Turnitin, which is a common tool used by teachers to analyze students' petty writings. Turnitin does not precisely clarify how to detect theft, but the study showed how ML can be used to develop a theft detector. Historically, the theft detection of regular text (essay, book, etc.) relies on having a large database of references to compare with the student's text; however, ML is not in the database It helps to detect source theft. Or older information sources that are not digitized. For example, two researchers used ML and predicted that the source code is stolen with an accuracy rate of 87%. They examined the different style elements that each programmer might have, such as the average number of lines of code, the number of indents per line, and the frequency of code comments.