There are several reasons to doubt that Harvard’s massive plagiarism ring is as massive, or as damning, as Harvard officials suggest. It might have not have actually happened in the first place. Here’s what we know so far:
Actual evidence: Very early this morning, The Crimson updated its report to include the circumstances of the take-home, open-book exam that students currently under scrutiny took last spring. The article now indicates that, in the final hours before the exam was due, a single TF issued to a bunch of frustrated students the definition of a term that apparently was never fleshed out, or possibly even discussed, in any of the course’s materials or lectures.
This may explain how many of them employed the “same long, identical strings of words.” That doesn’t make what they did ethical, necessarily, but it weakens the theory that over a hundred students colluded to copy each others’ work.
Precedent: A handful of outlets cited Adam Wheeler’s web of deception as though the array of lies told by a mentally ill man were relevant to over a hundred people using the same language in an exam, simply because Wheeler and they attended the same university. (We’re looking at you, Bloomberg.)
A much better example is an incident that took place in 2010 at the University of Central Florida, where a business professor used statistical analysis to “prove” that 200 students had cheated on an exam. In truth, those students had studied using readily available sample questions, the same questions that the professor explicitly stated he would not use for the exam, but then did. (Which meant that he had presented someone else’s work as his own.)
Believability: While the circumstances at Harvard and UCF are different (at Harvard, a lot of students used the same language; at UCF, a statistically questionable number of students did well on a test), the sheer scale of both invites a comparison, and an investigation, of the testing conditions rather than individual students. You’ll find the same lesson in that Freakonomics chapter about Chicago school-teachers everybody remembers: that symptoms of widespread cheating often uncover the intentions, ill or otherwise, of just a few individuals.
Let’s be clear: Harvard students are capable of cheating. But Harvard is suggesting that over a hundred students (in a 279-student lecture) colluded to share a string of words (not answers), and managed to hide this nefarious cheating, which is totally obvious, from the rest of their classmates and the teaching fellow who was helping many of them hours before they submitted their exams.
This assertion undoes itself. Out of 279 students, how many would cheat fully aware that they’re going to be caught? A few people, maybe. But a hundred?
(Get in touch if you know more. Anonymity guaranteed.)