Students Evaluating Teachers Doesn’t Just Hurt Teachers, It Hurts Students

The vast majority of healtcare problems are dependent on using the right known technique (medication, surgery, etc.) and of course the problem is finding that, so diagnosis. Handing out morphine is not a solution in the genral case.

Long-term health is sadly largely independent of acute healthcare providers' work. It depends on genetics, environmental factors, behavior (exercise, diet, lifestyle) and a stochastic factor of acute problem solving (as the body breaks down more and more depends on doctors, and the fast and correct identification and application of the best healthcare technique).

So healthcare has two categories, the first is the common flu, I have some gastro issues, refer to specialist, manage chronic illnesses (diabetes, pain, low/high blood pressure) and the other is usually the problem for specialists, when they have no or wrong idea what's going on and thus an acute problem becomes a year(s) long trail of tears.

Now, that said, I'm not advocating for ignoring the edge cases and just use this one simple magic formula to decide who can play doctor. But we are still stuck in the dark, because we focus on the complexity, the long term, the human component, the expertise, yet fail to do the most basic thing to get at least somewhat reliable data.

And I'm aware of the problem of fishing for signals in data, but without it we can't even start the trial and error process, we can't differentiate between the two cases, and on average let's say we end up doing nothing.

Furthermore, we know that the obvious solution is to refactor the system. But that rarely presents a real possibility for education and human long term health. (You know, the small stuff :)) It'd be easy to ban smoking and alcohol and give less damaging stuff for people instead and spend more to treat addicts, and introduce negative income tax as a form of basic income to help poverty to reduce stress and help mental health to reduce number of smokers. Similarly it'd be easy to reform schools, and do all of the above to help parents to have more time for their kids so they get a better nurturing youth overall, which makes them ambitious and motivated to learn, study, explore and understand. And it'd be easy to do all this, but that's not likely. Hence the proposal for simple but maybe vastly effective changes.

Believe me, I'll be first in line to shame their academic standards. However...

> In any case, consider that this system is in fact a solution to the proposed problem.

It might. I think that it creates other sinister problems in the process. You have to ask whether you think that the purpose of going to school is to better yourself versus competing against your neighbors. If a person does poorly in class because of someone else's performance, that's pretty fucked up. Likewise, if someone does well in class because of someone else's performance, that's also fucked up. The system you appear to be describing perversely encourages sabotage and cheating, because learning is secondary to "winning", because you're fucked by other people succeeding. My intuition is that systems that treat grades like a competition produce a combination of accidental failures and assholes who treat other people poorly.

(In a sense, I guess the whole thing is optional - after all, it is anonymous, and we don't check that every student has filled the whole thing out. It also switched from in-class to an online form during the time I was teaching, and the response rate went down quite a bit as a result, but this was after the situation I mentioned above.)

Even without this historical evidence, however, it does seem unfair to grade this way for exactly the reasons in BugsJustFindMe's response. It might make sense in larger first-year classes, but in the more advanced smaller classes (say, 30 students) it's likely that you're already down to a set of students who try really hard and potentially all deserve an A. There's a lot of correlation between (perceived) difficulty of the course material and the students who take the class, so there's no reason to expect that you have the same distribution of students from those beginning first-year courses in your "Advanced Stochastic Processes" course.