Sniffing Out the Cheaters

Portland, Oregon's Jesuit High School (JHS) has a long tradition of being both Catholic and countercultural, says librarian Kathy Fritts. "Teachers tell the kids, 'We know other people cheat, but you don't.' We teach that there's nothing to be gained by cheating. We talk with them about Stephen Ambrose and Doris Kearns Goodwin," two writers in the center of recent plagiarism scandals. "We work it into the whole context of societal dishonesty," Fritts says.

Even so, a few of the college-prep school's 1,000 students have occasionally been caught cheating, especially off the Internet. As a result, Fritts and her colleagues decided during the 2001–2002 school year that there needed to be a mechanism in place to detect copying and pasting on important research papers. So Fritts decided to test the efficacy of two anti-plagiarizing tools. She arranged for a trial subscription to Turnitin.com (www.turnitin.com), and purchased a copy of the Glatt Plagiarism Screening Program (GPSP) from Glatt Plagiarism Services (www.plagiarism.com).

Fritts ran three dozen student papers through Turnitin.com. The service found many matches, but except for one sentence, all of them had been cited correctly. After papers are sent as word-processing documents to Turnitin.com, they are added to the company's ever-growing database, which includes thousands of papers, including those available from "paper mills," such as the Evil House of Cheat. Because Fritts was unable to discover a single plagiarized paper in the 36 she tested, "the vice-principal had to supply us with two papers of proven cheaters, and Turnitin.com nailed them both." However, Turnitin.com, she says, doesn't yet check papers against the major subscription databases. That means that a student paper isn't reliably checked against previously published magazine articles, which may have been lifted from EBSCO, for example. (Turnitin.com is negotiating with the database vendors for access.) But all in all, the service impressed Fritts; she calls it "a potent weapon against dishonesty." That said, the hefty cost of subscribing to the service gave her pause. "We were quoted $750 a year," a rate based on the school's size.

Fritts also considered the Glatt Plagiarism Screening Program, but at first couldn't get the software to run. She eventually decided not to test the program because she wasn't confident that its methodology actually proved whether a paper had been plagiarized. The Glatt software removes every fifth word from a student's paper. Then, eight days later, the student is presented with a truncated version of the paper and asked to supply the missing words from memory. "If the student can't fill in every space with the correct word or something close to it, the student fails the test," says Fritts, and it is assumed that the paper was plagiarized. The paper and test can also be mailed to Glatt, and the service will determine through a proprietary formula whether the paper has been plagiarized. Says Fritts: "It's hard to see the Glatt software as an improvement over making the suspected cheater take a short quiz on his 'original' vocabulary and concepts."

Ultimately, Fritts and the faculty decided that the school should subscribe to Turnitin.com for the 2002–2003 academic year. There were some reservations, however; one English teacher asked a student to deliberately copy a full paragraph from the Net into a paper as a test, and Turnitin.com didn't catch it. But JHS students who are aware of the test seem to have taken it as a warning. "We're getting a lot more questions about how to cite [online] materials [properly]," Fritts says. "The kids are asking me, 'What's the difference between plagiarism and paraphrasing?'"