Academic Integrity and Remote Testing


The Academic Standards Committee, which is a standing committee of governance at my college, has been charged with updating the language in the student conduct code regarding academic dishonesty.

This is a bigger job than one might expect.

Last week I gave the committee stage time at the collegewide faculty meeting. The meeting became a series of discussions. One set of conversations focused on prevention of cheating in the first place; even there, on a tactical level we had to divide the discussion between exams and similarly timed live assignments on the one hand, and papers or other asynchronous assignments on the other. A third group discussed ways of detecting cheating when it happens. The fourth discussed whether there should be codified “tiers” of penalties based on the severity (and number) of offenses, and, if so, how the tiers should be defined.

That’s a lot, and folks added considerations on top of those to make the discussion even more wide-ranging.

I’m sharing this not to betray any confidences but to outline just how complicated these issues can be.

A few folks took issue with the premise. One argument was that any discussion of academic integrity that doesn’t grapple with grade inflation is missing the big picture. I suspect that’s more true at some elite institutions than at most community colleges, but to the extent that the assumption underlying the argument is that integrity cuts both ways, I have to concede some truth to it. Another suggested that the enforcement metaphor already gives the game away; if we just trusted students, cheating wouldn’t be an issue. I agree at some level that if we have to resort to sanctions, something has already gone awry. But my own reading of human psychology isn’t quite sunny enough to assume that if we got rid of all rules, we’d get rid of all bad behavior. At a basic level, I just don’t buy it.

A more nuanced version of that critique—one with which I’m much more comfortable—holds that there’s a meaningful difference between the “teachable moment” mistake of sloppy citation (or what one professor called “patchwork plagiarism”) and just buying a paper. In the former case, the enforcement frame is unduly harsh; in the latter, it strikes me as necessary.

I suspect most agree that prevention is vastly preferable to punishment. I was heartened to see some of the methods that people use to prevent cheating. The most effective and attractive ways involve building assignments that are both intrinsically engaging and customized. If students are intrinsically motivated, the rest pretty much takes care of itself; at worst, you might get some mistakes of enthusiasm. To the extent that it’s possible to craft assignments that way, of course, I’m all for it. But the idea that you can engage all of the people, all of the time, in every class, just isn’t plausible. There are too many variables. Some safeguards are necessary.

The advent of Zoom courses, with remote exams, has brought new challenges. How do you prevent cheating on a timed test when you aren’t in the room? Here, I was impressed by some of the ideas folks brought forward. We have the Respondus lockdown browser that takes videos of students taking tests and flags moments when students move in ways that the algorithm doesn’t like. Those flags need to be reviewed by human beings—something as innocent as a sneeze could trigger one—but at least they save the time and tedium of sitting through, say, 25 hours of video per section. Someone suggested angling the camera differently, so instead of watching the student’s face (and seeing the background), it would focus on the student’s hands. I wouldn’t have thought of that. Other suggestions included randomizing the order of questions, placing time limits that would defeat efforts to look up answers and avoiding publishers’ test banks, since those are often searchable online.

The question of tiers, though, remains sticky. Tiers of punishment presume that offenses are reported at all; it’s clear that many are not. The tension here is between respecting discretion and assuring equal treatment. I’d hate to see students who did similar things treated differently, especially if those differences fall along racial or other demographic lines. Ideally, reasonably clear criteria in the code could make it more likely that enforcement would be evenhanded. But the details are tricky.

Wise and worldly readers, have you seen a reasonably elegant and effective way to deal with academic integrity issues in classes with remote testing? (I can be reached at deandad (at) gmail (dot) com or on Twitter @deandad.) Thanks!

Source link

Leave a Reply

Your email address will not be published.