Study questions plagiarism-detection software

Computer programs designed to root out possible instances of plagiarism often flag common phrases and could deter students from using valuable sources in papers and essays, according to research conducted by two Texas Tech University professors.

The research focused on two of the most popular plagiarism-detection software programs, Turnitin and SafeAssignment. Susan Lang, one of the plagiarism researchers and an associate professor at Texas Tech, said reports generated by the plagiarism software highlight many phrases that qualify as frequently used jargon, such as "global warming" or "according to a university study."

"If there’s a value in these applications, there’s a value in a very limited context," said Lang, director of the university’s first-year composition program. "I think at one level we suspected that they wouldn’t be as good as potentially advertised. … But we were surprised at just how many things were flagged that would never be labeled plagiarism in context."

The Texas Tech research showed that of 400 student papers flagged for further review, only two were found to be plagiarized.

SafeAssignment and Turnitin compile a database of student research papers, internet news articles, and academic journals. Instructors enter new student papers and compare the work against the programs’ databases. Similarities are highlighted, and instructors are given a report documenting the paper’s originality.

Katie Povesjil, vice president of marketing for Turnitin, said the Texas Tech study contributes to a common misperception about the company’s software.

"It is important to point out that we do not detect plagiarism," Povesjil said. "We find text that matches sources in our databases, and we highlight those matches and identify possible source of the matches. Whether those matches constitute plagiarism must be discerned by human judgment. While our software cannot substitute for human judgment, our software can provide information and context to assist or amplify human judgment."

Turnitin is often misused by instructors, Povesjil said. Some professors institute hard-and-fast policies that give automatic failing grades to any student whose paper has a high "similarity index" when compared with Turnitin’s database. A score of 0 means the student’s work is original. A score of 100 means the entire paper was potentially plagiarized.

"We still recommend that the instructor examine the originality report to understand the context," she said. "Such absolute policies can cause unnecessary drama instead of a teachable moment."

SafeAssignment officials did not return phone calls or eMail messages as of press time.

Lang said Turnitin’s originality report often did not provide enough context for educators. For example, if a Texas Tech professor found a student’s paper was similar to others at Baylor University, the professor would have to contact Baylor to get the similar work and compare the student papers side by side.

"It could conceivably be a waste of time," said Lang, who conducted the research with Kathleen Gillis, director of Texas Tech’s university writing center.

The Texas Tech research team submitted a paper pulled directly from four internet sources and ran it through the plagiarism-detection programs. Neither program identified the web sites that were used to write the paper, Lang said.

The shortcomings of plagiarism-detection software, Lang said, should make campus decision makers pause as operating budgets are hit hard by a slumping economy.

"I don’t know of any institutions … that are sitting there rolling in money, able to write checks for these sorts of things," she said.

Turnitin’s rates vary, but most schools and colleges pay $1 per student, per year, for the service. Turnitin’s software is used by more than 7,000 campuses worldwide, including Texas Tech University, Povesjil said, adding that the university runs 14,000 student papers through the program every year.

"Statistics on [its] usage show a significant reduction in the amount of unoriginal material submitted by Texas Tech students in their papers over time," she said. "So the researchers do not appear to speak for the institution as a whole."

Lee Bobbitt, Texas Tech’s student body president, said plagiarism-detection software could create a distrustful atmosphere in the lecture hall and deter students from using valuable online information sources that could create an alarming originality report when the paper is run through the software. 

"It’s kind of like you expect students to plagiarize … and I know that in a lot of my papers, the sheer number of journals that you read, you’re going to pull quotes out of there and form your ideas around them," said Bobbitt, a senior political science major. "The consequences for plagiarism are huge, and no one wants to get kicked out of school, so that deters students from using some web sites and authors."

Pouring more resources into writing centers and face-to-face programs that show students–especially freshmen–how to avoid plagiarism would be more effective, Bobbitt said.

"I think students would find it more beneficial to be able to develop their skills," she said.




Texas Tech University plagiarism research

Computer programs designed to root out possible instances of plagiarism often flag common phrases and could deter students from using valuable sources in papers and essays, according to research conducted by two Texas Tech University professors.

Read the full story at eCampus News