[barefeg]

1. PhDs don’t “do exams”. You can’t cheat your way to publishing original research (exceptions of course) and have external expert researchers review it and accepted on a reputable journal. Given that that’s a requisite for obtaining your degree, there’s no point on a final exam. Most defenses nowadays are partly ceremonial (exceptions of course). Both the material and your ability to do research has been checked months before by your advisors and graduation committee.

2. Master’s degree courses follow all kinds of schemas for examination. From oral exams, to in person no extra material allowed, to open book, and take home. Being on both sides (taking the exams and creating them) I can say that it really doesn’t matter if you allow people to take the exam home and collaborate among each other. These types of exams are designed to really test a deep understanding and ability of the material. There have been exam questions where an entire class of >20 students are not able to solve it. People that are really good are able to have a shot at it and maybe make some progress, and that differentiates the good from the exceptional.

3. Nowadays undergraduate degrees are a commodity so it feels they need to make sure only the good students get one. But in the end it doesn’t really matter, since most employers (exceptions of course) will want to see how much value the candidate add, which does not correlate with having a degree or good grades

[lutorm]

You can’t cheat your way to publishing original research

Of course you can, making up results is a time-honored tradition!

It does take a certain amount of skill to do it believably, though.... ;-)

[snotrockets]

I felt as if exams in my masters (in Mathematics) were different: they mostly check you understand the definitions, maybe remember some basic tactics for approaching problems. You can't really check the students can solve "real" problems, as those are too complex for the limited scope of an exam.