CURRENT ISSUE :: DECEMBER 2002 :: EDUCATION

Copy Cop

'Accidental Entrepreneur'
Helps Schools
Tackle Plagiarism

By BRAD REAGAN
Special to The Wall Street Journal

John Barrie is a tattletale—and proud of it.   

In the mid-’90s, as a graduate student in biophysics, Mr. Barrie took an interest in the growth of online “paper mills”—Web sites that allow lazy students to download term papers for a small fee. He decided to take technology from his field of research—analyzing brain waves—and apply it to catching cheaters. Four years ago he turned the operation into a for-profit company, iParadigms, of Oakland, Calif.

Now, the 34-year-old’s company is the digital plagiarist’s worst nightmare. His company and its Web site, Turnitin.com, help teachers at hundreds of U.S. high schools and universities, as well as Oxford and Cambridge universities in Britain, monitor some seven million students.

“They should call me the accidental entrepreneur,” Mr. Barrie says. “One day I woke up and, according to some people, I was the country’s No. 1 plagiarism cop.”

We talked with Mr. Barrie about his growing business and why he thinks it could change the ethics of generations of students.

WSJ: How did you get the idea for this technology?  

MR. BARRIE: While a grad student back in 1994 [at the University of California at Berkeley], I thought it would be cool to let students read and peer-review other students’ papers over the Internet. What I stumbled on with that experiment was that there was a significant level of cheating with those students. And I stumbled onto it because I had a steady stream of students coming into my office telling me [through hearsay] that their fellow students were cheating—taking papers off the Internet or using old papers.

I used some of the brain-wave technology I was working on to check for plagiarism. I was analyzing huge databases of brain waves to look for irregularities. In the case of plagiarism, I’m looking for regularities in text—overlapping and matching texts.

WSJ: How does it work?

MR. BARRIE: Once an institution licenses our service, faculty members create Hotmail-like accounts on our service. They then tell students to submit digital versions of their papers to the service. At the point our computer receives the student paper, it creates a digital fingerprint of that paper and we compare that fingerprint to three databases.

The No. 1 database [contains a copy of every Web page on] the Internet that we update daily to the tune of 30 million pages per day [including Internet-based term papers]; it’s very similar to the principle behind any search engine.

No. 2 is a database of books and journals, including all of the classics—“Moby-Dick,” Nietzsche’s “Beyond Good and Evil,” etc. And database No. 3 is a database of every paper we’ve received from every single one of our clients from Day 1.

Once the computer compares the paper against those three databases, the professor can see an exact copy of that paper with [selected passages] underlined and color-coded. That underlined text was found in one of [the database] sources verbatim. We don’t make the judgment [of whether something has been plagiarized.] You make the call. It’s going to be a cold day before any computer makes the call of whether something has been plagiarized.

WSJ: Is there a cat-and-mouse game between your company and the sites that sell unoriginal works?

MR. BARRIE: No, I would say that 85% of the cases of plagiarism that we see are straight copies from the Internet—a student uses the Internet like a 1.5 billion-page cut-and-paste-able encyclopedia. I would say another 13% come from other student term papers. Maybe half a percent comes from these cheat sites. I believe that the media have significantly overestimated the value that free, Internet-based term-paper mills hold for students. I also believe that we will crush the remaining traditional term-paper mills [which charge for papers] like a pestilent bug.

For-profit term-paper mills cannot stay in business without selling the same paper to multiple students. Since we compare every paper we receive to every other term paper that has [been] submitted to us, the student who buys a term paper may escape detection by us only until we receive a second copy of that paper. …