[hardy263]

In my university programming class, we submitted assignments using turnitin, the same software you saw in the screenshot. We were allowed to see how much of our programs matched other people's programs, and where it was copied from. Scores of up to 70% were not uncommon because the whole class had the same assignment and example code to work from. The only problem is if your score was near 99%, which would probably raise some eyebrows.

But since we could see what parts were considered to be plagiarized, we could easily change our program and resubmit to avoid the filter. It would only deter the students who were too lazy to change their variable names to something else.

[Jach]

Did anyone submit a pre-processed version of the code? (Assuming your assignments were in C.) If only one person did it they'd probably have high-uniqueness, but if >1 did it they'd all have high-sameness.

[hardy263]

No, it was done in C#, and even if it was in C, it's not worth the trouble to preprocess the code anyway. The assignment answers were usually simple but short. Something like sorting a linked list using insertion sort. If you have no idea how linked lists work, it's likely you would copy it. But if you understood linked lists, and already knew insertion sort, you could probably write it on the spot within 20 minutes.

[Jach]

Wow, so you never had to write complete programs for assignments? That "write a sort function" question seems more suitable to what I'd expect on an exam. You find out who knows/doesn't know a lot faster than a homework problem where you may be able to program it at the time but don't understand in general only to fail on the next exam, or maybe you think you know it already so you just copy from somewhere and change variables, but don't really know it, and so you also fail at test time.

Bleh. The whole point of catching plagiarizers has been to determine whether work was done in some expected way, not to make sure learning happened or anything useful came from the work.