"Dividing Points" among cheating students

Question

My AP Computer Science teacher has somewhat of a creative way to deal with students submitting apparently identical code. Instead of handing out lots of zeros, he divides the points by the number of cheaters, phrasing his system as "if you have the same answer, you deserve to split the credit." For example, a 5-person cheating ring on a 10-point project would yield 2 points per person (it rounds down to the best precision on odd splits).

Do you think this is ethical? Is this a wise way to deal with cheating?

On one hand, it is a lot softer than most policies. On the other, it is somewhat logical, and seems maybe fairer than the standard penalty system.

Answer 1

To add to the existing arguments:

The punishment for cheating should always be higher than the direct advantage gained from it. For example, if somebody cheats at one task in an exam, the penalty should not only be failing that task but at least the whole exam, if not much higher. The reason for this is simply that you cannot possibly detect all attempts at cheating and thus the punishment should aim at being sufficciently high to make the average outcome of cheating negative if the detection rate is taken into account.

Of course, it is quite difficult to even estimate the detection rates for cheating and thus to adjust the punishment accordingly, but the punishment you are describing is certainly too low.

This aspect is not only important for discouraging cheating, but also from an ethical point of view, as cheating should not yield an advantage over the honest student.

Answer 2

I don't think it is fair. Say you work very hard on your program, to get a 9. While you go to the toilet, I steal your code and submit it. I get a 4 for doing nothing, you get a 4 after a lot of successful work. If this happens the day before the deadline and I had nothing, I would have handed in nothing, and thus get a 0. So there is a clear benefit in cheating: I get more marks than even if I pull an all nighter and put together a crappy code that barely holds together.

I guess the reason to punish everybody equally is multiple:

• Discourage from people lending code to classmates.
• It is much easier than to try to figure out what happened and who is the original author.

But, the way I see it, there are a few objections

• There are still situations where a cheater may benefit from this scheme (half good is better than nothing); a cheater is being awarded points for doing exactly nothing.
• The original author may be innocent. The code could have been stolen. It is unfair to blindly punish without, at least, inquiring in the matter.
• Serial cheaters may get away with it and still get enough points to pass.

Edit:

What if the homework is given voluntarely? Consider the case where there are weekly assignments, with a high work load. Part of the point is for students to work and get efficient in problem solving. But, if instead of solving all of them, we share the work, change it here and there, perhaps they will not notice, I still work half of it, and maybe I get full recognition; maybe only half.

Now, say a given week you only had time to do half of them. Gambling is beneficial: if you go the ethical way, you get half the marks. If you cheat, you can either get half the marks, or get away with it if you are skilled or the grader is low on coffee.

Answer 3

I'll only give a legal point of view on this issue for French higher education.

In French universities, even if the kind of rule you describe is enforced by some teachers, it is in theory forbidden to alter notation for cheating: one should only grade according to content, without any penalty. If cheating is suspected by the teacher, she should fill a procès-verbal form describing what she suspects happened, and giving all relevant info. This is then passed to a commission that will interview her and the suspected students, and will decide what happens next. This commission has a scale of possible action, going from nothing to five years' suspension from all French higher education institutions.

This process is seldom used for projects and evaluations among small groups (contrôle continu). However it should be strictly enforced for terminal exams, as otherwise student can easily go to court.

Answer 4

It's actually a fairly nice solution to a more relevant problem: the right pedagogy for different kinds of students.

The formula is a good, but should be stated thusly: "Students are allowed to collaborate on homework, but your scores will be divided by the number of collaborators, so pick your team well." This is in tune with the environment that students will likely encounter after their degree, but neither enforces any policy nor penalizes any others.

Then the teacher should grade on the bell curve. This gives a passing score to those who worked together, and a great score for those who were truly above average. In some way this is superior to students who try to get an A by a hyper-competitive attitude towards the rest of the class.

During the test, they will still have to show that their personal level of mastery, without the help of their peers.

Answer 5

The goal of a policy on cheating isn't being "fair", it's to discourage cheating.

In academia we don't want to have academic papers who bend the truth a little bit, but we want to have academics who tell the whole truth and for whom cheating is not an option. As a result you don't want to do anything that encourages cheating.

Answer 6

This seems like a bad policy. With this type of policy it leaves the teacher having to figure out who copied from whom and how many people were in each "group". Usually if Alice's work is copied by Bob, and Bob's "work" is copied by Carol, then Bob and Carol are penalised the same. In your policy, you would have to figure out how much to penalise Bob and Carol. It seems like it would be better to allow group work with the stipulation that the points get divided in a manner either specified by the students or the teacher.