Alfred Thompson posted today about cheating on CS class projects. It was in response to Garth Flint's post on finding interesting projects which in turn referenced earlier posts by Alfred and me.

Garth laments that it's hard to find projects that are both interesting and meaty but where solutions can't easily be searched for online. Alfred notes that cheating will happen and that it's an ethics issue. This is why I try to create a culture of sharing and acknowledging credit (that is, citing sources) but I'm not naive enough to believe there isn't any cheating in my classes.

I don't go nuts looking out for cheating but there are a number of things I do to try to help keep the students honest.

1. Class time. More often than not students have some time in class to work on any given project. This lets me get a handle on what they actually know since I can interact with the student or team and observe them interacting with each other. Even if a particular project is all done out of class, over the course of the semester you can get a pretty good idea as to what a student will produce on their own.
2. Project quizzes. For a big project, I'll sometimes give a quiz on the day it's due. The quiz will be tied to the project - describe a data structure or algorithm you used. If the student wrote the project it's free points, if not they'll have problems. Of course, this only requires that a student understands their project not that they actually wrote it but at least that's something.
3. Git logs. I get my students on GitHub early and that lets me look at the commit logs for any given project. This usually can't be used to find cheating but it lets me see what was committed when.
4. Be reasonable in terms of workload and expectations. This is a really important one. Back at Stuy, it was made clear to me many times that many students feel forced to cheat due to teachers expecting far too much from them. If you're take 6 classes a day and each gives you two hours of homework somethings got to give. I always try to give a fair amount of work and only if I think it's necessary and makes sense and as long a a student is up front about any issues that arise I'll try to work with them on deadlines and requirements.

I could use a code plagiarism tester like Moss but I usually don't bother.

I also try to keep things in perspective. Usually things come clean in the end. Most classes have multiple measures - tests, homework, projects, classwork and more. Even if a student copies a project and gets away with it, it's just one measure and it likely won't move the needle all that much. True, a kid who's doing well might feel the need to cheat on one assignment (see item 4 above) but that's not a hill I'm willing to die on.

Even if a kid in a 10th grade intro CS class cheats their way through they'll unlikely take any more CS as they'll be woefully unprepared. In that case, the one higher grade they got on their transcript due to cheating on CS projects probably won't make much of an overall difference and it certainly won't hurt the kids who take all your school's CS courses and apply to college to be CS majors.

None of this is to say that cheating doesn't happen and I do believe we should keep an eye out and address it if and when it rares its ugly head. I'm just saying keep things in perspective.

When I pick projects I don't pay much attention to if they can be found online and I'm not the tightest screw when it comes to enforcing the rules in my classes. I know that's led to my being taken advantage of over the years and yes, that does bother me but I'd rather put my time and energy into the kids trying to do it right than those cutting the corners.