In case you haven’t heard, the University of Central Florida was recently rocked by a large-scale cheating scandal in a business management course. At one point, over 200 students in the course had turned themselves in to Prof. Richard Quinn or an associate. Prof. Quinn uses (or I should say “used”) tests from a pre-made test bank, and somehow students got hold of the test bank with answer keys prior to the midterm. Every student in the class, guilty or otherwise, was required to retake the midterm, which apparently then showed a normal distribution as opposed to a severely bimodal one on the compromised exam.
UCF puts the videos for Prof. Quinn’s lectures online. Here’s the one where he announces he’s discovered the cheating and describes what’s about to happen. This is 15 minutes long, but you MUST watch it. Seriously. All of it.
Wow. Can I breathe now? Four things:
- That lecture was a masterpiece of restrained forcefulness. You can tell that Prof. Quinn wants to explode all over those people, and yet he doesn’t — and somehow it makes you feel worse than if he had blown up. I’ve been in situations like this before and never come close to keeping my cool the way he did.
- You wonder how much of what he’s saying about “forensics” and the “net tightening” around students is just bluffing, and whether students with the chutzpah to cheat on this scale have the nerve to call the bluff. Can a university IT department really do an NSA-style traffic analysis to determine who cheated?
- I think Prof. Quinn is being unbelievably gracious (you might even say “lenient”) towards the students who ‘fess up to the cheating. Between getting kicked out of school and having to take an ethics course, I think I’d choose the latter any day as long as there’s no penalty. Speaking of which, did anybody consider the poor schmoe who has to teach that ethics course? How crummy of a teaching assignment would it be to teach an ethics course to students who are forced to take it because they got caught cheating? Like teaching a drivers’ ed course to a 200-student class full of known traffic violators.
- Finally, and a little more seriously, I abhor academic dishonesty, and the fault here lies squarely on the students who chose to cheat. However: This should serve as a warning to any professor who chooses to use a publisher’s test bank to give prefabricated tests. Doing so adds so many exploitable seams into your test security that it is practically begging for unscrupulous students to try to find those seams. I know: I’ve written one of those test banks before. The publisher doesn’t keep track of who has a copy; the fact that there are extant copies of the test banks and their keys just floating around out there should be enough to make profs not want to use them. But some still do. The convenience is not worth it.
Casting Out Nines 
Is it really cheating if you use a published test bank to prepare for an exam? It’s certainly cheating if you expect that the exam questions will be selected from the test bank. Couldn’t the students defend themselves by arguing that they did not know that the test bank would contain the exam questions?
I’m also concerned that honest students might have been coerced into false confessions. If I had been one of the students who got an A on that test, I would not be confident that the professor’s “forensics” would exonerate me. I might have confessed to avoid the risk of a false accusation.
How can a professor that cheats himself (by not designing own exam problems) demand students to behave ethically? Had he made his job properly the used “cheating” technique (involution’s point stands) could never have worked. That does not excuse cheating in general, of course.
One of the great difficulties facing the modern academic is to create assessment tasks which can’t be cheated. Clearly using a pre-existing question bank is asking for trouble. I think it’s also a bit unfair that the cheaters get away scot-free – they get a no-questions asked re-sit, and the non-cheaters get their marks thrown out and have to sit the test all over again.
There’s a tradeoff in assessment between time and security. The more cheat-proof an assessment it is, the more resource- (usually time-) intensive it is. This prof is going to the low investment/high cheating probability end of the continuum. At the other end is, I think, oral exams of each student individually. Those are virtually cheat-proof but a major investment in time. And at a large university it’s hard to pull off. Those of us at small colleges have an easier time with more personalized testing.
And somewhere in between is having written exams with “hard” problems (i.e. such that are not multiple choice or pure reproduction), for example proofs, application of algorithms/techniques, programming and so on. You then need only five to ten problems for a two hour exam. Grading it is, of course, more time intense than just counting right crosses.
My exams all looked like that, by the way.
I agree about the bluffing part. However, the test bank was almost certainly distributed in electronic format. ANd for most of those, that format would be email. How many of them were sent or received from university email accounts? That is a pretty straight-forward bit of cyber-forensics that would at least tell you who received the file. Now some who received it wouldn’t have actually used it, but it would be a major piece of evidence against a student (and frankly, an ethical student who received such a file should probably have alerted the professor before the exam was given).
Pingback: Students respond to UCF cheating scandal | Casting Out Nines
Pingback: “The system has failed you” | Casting Out Nines
Pingback: Better testing through “data forensics”? | Casting Out Nines