[NickGerleman]

I've been a TA for a couple CS classes at Iowa State University. Cheating here at least is pretty rampant. I would guess 1/3 or more of the students cheat in some capacity.

Some of it is really easy to detect too. For a first semester freshman class, I tried using MinHash to cluster student submissions together. Maybe 10% of the assignment have near duplicate pieces of code. More comprehensive methods like MOSS (Measure of Software Similarity) find much more.

It's hard for professors to take time to deal with it. A professor more or less told me that the process to report it to the dean of students would take months of his time, and that he already spent 55 hours a week for his classes. He explained that the time it would take to go after the cheaters would detract from time to make good assignments, labs, and lectures for the students.

It's hard to see a professor for a large class (500 people) having the time to deal with handling cheaters so the problem continues. I'm curious how common this is as a problem at other schools.

[wayn3]

In germany, what you call cheating isnt even considered cheating.

On homework, we openly collaborated and handed in near duplicates without trying to hide it. Some students would just copy the work of others. None of this mattered. Nobody "went after" it. It just gets graded, you get your points and you move on.

Most people accept the fact that grades don't matter, but you need them for your resume. For some reason, universities have moved on from being institutions of knowledge-production to being diploma mills, and grades are their way of succumbing to societies need to make shit measurable. The way we cope with this is to just hand out grades that are utterly meaningless, but keep the system alive.

In the meantime, people carry on with their lives. Some people in university want to learn things, some are just enrolled because its a cheap livestyle choice.

In the end, nobody cares about grades. Whether I actually know my stuff is my problem. Not my professors. No company hires you based on your grades, but a lot of HR personnel will reject you based on them. So you need good grades. But good grades alone wont get you shit. We live in this weird limbo of grades being a necessity but beyond that, dont mean shit. They are the shittiest heuristic known to man.

If grades were a scientific theory, those who came up with them would be ostracized crackpots. Neither meaningful nor in any way falsifiable.

A written exam on quantum field theory can only test so many things. Since its written, you have to ask for things that can be produced in a given timeframe. That leaves you with memoizable calculations. Any professor will tell you that such an exam will tell you nothing about whether the student actually understands the subject matter. A 2 minute discussion of the topic in very high level terms is probably more useful to assess actual aptitude.

Professors KNOW that they cant possibly measure the ability of individuals in classes of more than 15 students. Why would they care about cheating? They are part of a system that doesnt make sense just as much as everyone else.

[raphman]

> In germany, what you call cheating isnt even considered cheating.

Having taught computer science and HCI courses for ten years at German (Bavarian) universities, I can not agree.

Of course, some professors/groups are more interested in finding and addressing plagiarism than others. However, I have only once encountered a professor who didn't really care about a plagiarised thesis.

In general, addressing cheating is much easier in Germany, as professors and teaching assistants usually may (and have to) take appropriate measures themselves without involving deans and examination offices.

Typically, teaching assistants will check submitted assignments both manually and automatically (usually using MOSS or JPlag, sometimes custom scripts). If plagiarism is found, affected students are called in to a meeting with the teaching assistant to present their side. Unless students can present evidence that they are not at fault, they are given a stern talking and a failing grade. If students do not accept the outcome, they can ask for an appointment with the professor (very few do this). If the professor upholds the decision (which they ususally do), students may escalate the case to the dean. This is done very rarely (I know of only one case).

In my experience, staff at German universities definitely cares about countering cheating. We will certainly never catch every cheater.

However, I would even argue that addressing cheating is easier in Germany, as the process is generally more light-weight than at Anglo-American universities. As students face fewer consequences for admitting cheating (just failing the course, not getting expelled), they usually accept these consequences and try not to get caught again.

[wayn3]

I'm not talking about Theses. I'm talking about "übungszettel".

I'm aware of the fact that bavaria takes a bit of a harsher stance on this.

In Heidelberg, nobody gives even the slightest crap about "cheating". But as I said, it isn't considered cheating anyway. Students are expected to collaborate on these things.

[raphman]

> I'm not talking about Theses. I'm talking about "übungszettel".

Sorry for the confusion. While I mentioned theses in passing, my statements were about programming assignments in undergraduate courses where students were expected to turn in individual, original work.

> I'm aware of the fact that bavaria takes a bit of a harsher stance on this.

I'm actually not aware of this fact. In my experience, how cheating is addressed depends much more on the stance of the individual professor than on that of the university or federal state. Extrapolating from your experience at one department within one university to all universities in Germany seems audacious.

[Houshalter]

Would you be alarmed if you found out your doctor had entirely cheated their way through medical school and knew very little medicine?

Great, you figured out that grades aren't a perfect metric. But so what? Grades don't prove a student is great, but they can eliminate a lot of students that are bad. If the bad students just cheat their way through, everyone else down the line suffers for it.

[wayn3]

- if you actually read what i wrote, youd have figured out, on your own, that i didnt say that you should cheat in order to learn less, but to leave grades behind in favor of actually acquiring knowledge.

- before a doctor is even allowed to practice medicine, he has to go through 5 years of education followed by 3 years of apprenticeship, followed by 5 years of specialization. by that time, hes logged thousands of hours of supervised doctoring and he only "graduates" from that phase if all his superiors write stellar reference letters from him.

i dont really care if hes cheated on his first year anatomy exam. not particularly.

beyond that, doctors are held to a higher standard than other people for a reason. medical mistakes are usually final. but grades are a very shitty way of figuring out whether someone is compassionate, ethical and honest enough to be good doctors.

- modern society is engineered for efficiency. i consider it any persons right to make the best of it, for themselves. whats best for an individual is very often not the best for society as a whole. if you follow the script of modern society, you end up miserable. im not a fan of that. ideally, nobody would cheat those stupid exams, but i wont hold it against anyone. ideally, i dont want 10 million indians immigrating to every first world country, either, but i wont hold it against any one individual who tries.

(i dont have a problem with indians, its just that there are a lot of them, and it feels kind of overwhelming at times)

[ThomaszKrueger]

Parent of a surgeon here. My son's medical school did away with grades altogether - it's pass / fail. They still have to pass the board exams every year.

[kurtisc]

Is it cheating for a professional software developer to use Stack Overflow? If they copy and paste, perhaps, but not if they use it to get a better understanding of something. It's very difficult to grade understanding.

Doctors go through a lot to graduate. Universities are welcome to put all students through the same level of scrutiny, working closely with a professional who can decide if they are unfit for the course.

Aside from thesis defences, they usually choose not to, because contact hours are prohibitively more expensive than a room full of exams and a few TAs.

[Houshalter]

It's not any more cheating to look at stack overflow, then it is for a doctor to look up more information in a medical reference book. But neither should rely on those resources to do their job for them, with no understanding of the underlying domain. A developer should easily be able to write FizzBuzz without using Stack Overflow.

[tscherno]

https://en.wikipedia.org/wiki/Gert_Postel

[eigengrau]

> In germany, what you call cheating isnt even considered cheating.

I can also attest that this isn’t universally the case. Our faculty’s introductory programming classes explicitly forbade this, and we went so far to even run automated anti-plagiarism heuristics over the solutions handed in.

I’m not sure how strongly I feel about the policy (or whether it’s still current). There definitely were tendencies in such classes that students who already came in with prior programming experience to be eager to »help out« more inexperienced students by sending them their solutions in verbatim, which led to other students copying at least certain program constructs without grokking them and without having exercised more general problem solving skills associated with the task.

[giancarlostoro]

In my school one of our professors forces students to do the tests all on paper. We also have a testing center where it's harder to cheat at. There's always people in there watching you, and you likely will be seated somewhere random and nowhere near anyone studying in the same field as you.

I have to say the professor who forces testing on paper really gets the better students in my opinion because they're less likely to cheat (if they even try). They have more reason to try harder to pass.

We also have on our classrooms specialized remote desktop software to see students screens, I've seen someone in one of my classrooms (was not a CS / Programming course, but it was in the same lab) get caught that way, trying to cheat. It was sad, I never saw that kid again after that.

[aurenen]

I used to be a tutor for my university's CS department. Cheating increased as our influx of CS students increased. Lots of kids who didn't actually enjoy CS, but had heard it was a lucrative career. It was incredibly frustrating to deal with them in class, and when they asked for "help" in the walk-in labs.

The biggest problem was the department heads unwilling to take on the issue. Most who get caught cheating just fail, but aren't reported for it, and the one lecturer who did take it seriously never got anywhere because when he would report it, nothing ever happened. The more students in CS, the more funding the school gave the department...

[eric_b]

Somewhat off topic, but I doubt very much there are any professors at state schools specializing in research that spend anywhere near 55 hours a week on their classes. Maybe he meant per month.

[nyc640]

Because introductory CS enrollment has ballooned so much recently, many professors for large introductory CS classes nowadays are full-time lecturers that do not conduct any research. Many full-time research faculty only teach smaller, upper-level classes.

In my experience, this works out pretty well since many of the full-time lecturers do dedicate a lot of their time to teaching (55 hours is totally believable depending on teaching load) and do a really good job of it.

[NickGerleman]

I can confirm that he does. He is a full time lecturer for 3 or 4 classes a semester and produces a staggering amount of material. He is careful to never repeat assignments and each one requires detailing the project, creating automated tests, creating all sorts of frameworks for them(eg, import this jar and pass the instance of your game logic to this black box and you'll have an interactive Swing application to play the game). He plans out his lectures thoroughly and has 2-4 hours of them a day. This doesn't even cover office hours or meetings.

I never really realized how much work a good lecturer does until I was a TA under them. I have so much more respect for people teaching introductory classes now.

[lostboys67]

So make a very public example of one student :-)