Academic Ethics
by Jon Shea
May 20, 09:17 AM
0) I have a Control Theory assignment due today. We need to hand in a printout of our source code, and some graphs that show it working. I have a program that can make the right graphs. In fact, my program works for multi-dimensional cases, when we only had to address one-dimensional cases for this assignment. But, my program has a bug, and I know about it. Many of the case it should work for, it doesn’t work for.
I can print out the required graphs, picking situations for which my program works. Handing in these graphs and a print out of my source code, it is very unlikely anyone would every find the bug. My program would probably see perfect. Unethical? Wrong?
1) Often in physics exams you get a question that says “Use
What if you never completely get equation 1 to meet up with equation 2? Some students will get the equations to match up by “accidently” making some algebra errors in convenient places. Unethical?
2) There are a finite number of solved problems in math and physics. Sometimes a problem you are given for homework or take-home exam is a solved example in a different text. If you happen open a solved version of an assigned problem, then what are your responsibilities? Convetional wisdom is that a simple citation will protect you from all fronts.
But, if it is ok to simply copy a solution from a book, why not just photocopy the pages from the book? Surely no one would expect credit for doing that. But why not? If it is important that you understand (whatever that means) the problem, then what is the correct course of action?
3) Since starting graduate school I’ve start explicitly pointing out errors in weaknesses in my homework, and things I think I poorly understand. I do this because I think it is honest, and furthers my education and understanding. It certainly results in worse grades. Should I stop? Ideally, I would bring these issues up with the professor before the homework was due, and address them in my final solutions. Unfortunately, my time endowment is finite, and so the grade/learning optimal solution isn’t always obvious.
May 20, 04:52 PM
The following represents my opinion. As you know, I am not an ethicist.
(0) Handing in the graphs is okay, because they really do come from your code. It’s not fraudulent output. However, since you are aware of a bug, it would also be best if you write a comment to the effect that “This program does not work for certain cases,” and let them see if they can find the bug. If they can’t, and take points off anyway, ask them to justify why. If this seems obnoxious, remember this: The object of “points” is not to reward or punish, but to give you (the student) objective feedback you can use to improve your knowledge and your work. Many teachers forget this basic truth.
(1) Ethically, I think it’s fine to work forward and backward the way you describe; however, I would probably ask my instructor for an opinion. On the other hand, I don’t think there’s anything wrong with working backward until you get something you can reach from the starting point, then use that as a lemma to go forward to the end. Just because you thought backward doesn’t mean your work has to go that way. And, I don’t think this is a problem.
(2) I have often told my students that they are welcome to copy the solution directly from the textbook as long as they cite it. But, as I also warn them, they will receive zero credit for such a solution. In order to receive credit, I demand that my students solve the problem using only what they have internalized, and I try to ask questions accordingly (i.e., questions that can be answered out of their heads).
(3) Decide whether you care more about grades or more about learning. If you care more about grades, stop; you are wasting your time. I’ll save my rant about optimizing grades vs. optimizing learning for another time. However, if you do want to optimize your learning, then always continue. Not only will you help yourself (which is enough to justify the behaviour in itself), but you will also help your teachers (which is a nice benefit, if you ask me). I’ve learned a lot about how to clarify and improve my teaching from listening to my students.
May 21, 09:23 AM
Michael, when I wrote that post it hadn’t even occurred to me that a CS prof would probably comment on them. Thanks. As far as I’m concerned, you’re far more qualified on ethical issues than Randy Cohen (and I read him every week.)
0) The “come from your code” criteria is obviously incomplete. I knew what the correct output was supposed to look like. It is highly probable that I could get away with fake code that simply plotted a graph that looked like the right one. I could probably even obfuscate the deception. Clearly some kind of honest is required.
I guess my main thesis is that grades and learning aren’t parallel objectives, and I sometimes get frustrated by that fact.
As far as weighting grades/learning is concerned, I challenge you to find someone with a lower weighting ratio than myself.
1) I think it is ok to solve these problems however you want, but it does entirely defeat the goal of feeling like you’re following in the footsteps of the greats.
2) I think your system is problematic, because it clearly isn’t enforceable. Unless you’re going to haul people into your office and make them work things out on the blackboard for you (something I don’t think professors do nearly enough of.)
If someone does cite a source (which they should, if they were influenced by it in their answer), then how do you know how much they understand, and how much they copied?
May 22, 04:10 PM
Thank you for the vote of confidence, my friend. I’m not sure it’s really fair to call me a “CS prof,” even though I do teach undergraduate CS courses nowadays. But, be that as it may, I do have some experience with academic honesty issues.
(0) I took the honesty of the submitter as an implicit assumption. Without getting lawyerly about what “faking” really means, I’m sure you’ll agree that inventing data you didn’t really observe and turning in a graph of it would be dishonest, unless of course you labelled it as such (“This graph is what I expected to get, but did not; it was generated with invented data.”)
I may well have a lower grades/learning weight than you do, but I suspect proving it would be like trying to make interpersonal utility comparisons.
(1) Who cares if you’re following in the footsteps of the greats? Unless, of course, the title of the course is, “Learning to Feel Like the Great Mathematicians and Engineers of History.”
(2) Actually, my system works fine, as long as the work product is sufficiently individual. If I were asking for a proof of Lagrange’s Theorem, it’d be easy to look up somebody else’s version and pass it off as your own, with a minimal amount of rewriting. But I teach programming, where it’s really hard to pass off more than a trivial amount of somebody else’s work (whether proof or code) as your own. And that’s easy to circumvent by twisting the questions around. I’m not claiming it works for every subject, but it works for me. :-)
I actually do make people come explain their work to me, when I think they’ve copied it. Except for the fact that it takes so much time and effort, I actually think that would be a better way to obtain grades than giving exams, in almost any case I can think of.
I have lots to say on the subject of grading, but regrettably, this comment is not large enough to contain it all.
May 22, 04:16 PM
Oh my god, Michael, I argue with my profs for Oral Exams every term. I almost pulled it off last term. It is so obviously the best way to do it. Especial for my 4 student classes.
I agree with the rest of your points.