> Good on SL for setting expectations but I also want to comment a bit. If research is all about being comfortable with uncertainty, we have an adverse selection problem. I went through undergrad with this “comfort with uncertainty, just try things out” mentality and it hurt me.
> Specifically, it hurt my grades, which in turn has hurt my applications to grad school etc. I’ve had to do work post-undergrad to make up for it and demonstrate I am actually a good student. The mindset I had to take on is one where I seek perfect certainty that I know the answer to every problem on every problem set in full so I can score close to 100% to lock in an A. It’s to the point where I was “emotionally drained” from trying to be perfect or near perfect on everything.
> I am sure other students who make it to grad school have a similar mindset, and consequently a similar feeling of emotional fatigue from anxiety. I don’t have a better solution since the grading system is supposed to assess understanding of the material which would be important for grad school. But, it’s an adverse selection problem.
> I agree. The problem is that research requires both technical skill and ability to deal with uncertainty. And one of these is easily measured (grades), while the other is, if anything, mildly decreasing in grades.
So the problem doesn't seem to be the students but use of grading as a system of measuring the progress of students who are on a research-heavy track of study.
Grade inflation exacerbates this problem. The straight A student of today might have been the straight B student of 40 years ago. Meanwhile, the student of 40 years ago that was exceptional in some subjects but mediocre in others can today only get an A for exceptional work, but may get a B for what used to be C level work. Truly exceptional work does not pay in grades anymore, only reliability.
Now, reliability is a very important skill, but it should not be the only thing measured .
That's a good point; I nearly doubled the next highest score on the midterm in a class of mine. I was not even remotely the only student in the class to get an A.
>So the problem doesn't seem to be the students but use of grading as a system of measuring the progress of students who are on a research-heavy track of study
It can be a tricky problem. In more academic fields, I think there's a tendency to not weight grades very heavily in graduate applications, instead focusing on undergraduate research, recommendations, connections to research in your department (obviously, a research group directly wanting a student, and having funding for one, is all but a guarantee of admission), and maybe courses taken. Unlike for undergraduate applications, you'll often have few enough applications after filtering out obviously-unsuitable applicants that you can have discussions about specific individuals, and there will often be enough information about them as individuals that you can make decisions based on that, rather than grades.
But at the same time, at least in the physics programs I was involved in, I got the sense that out of concern for grades potentially affecting futures or discouraging comfort with uncertainty and exploration, once classes were at a point where everyone was probably going to be going to graduate school, many ended up being de facto pass-fail: for the most part, everyone who showed a good understanding of the material would be given an A, and everyone who didn't would get a tap on the shoulder at some point and a suggestion that the professor would do whatever was necessary to let them drop (usually far past the ordinary drop deadline). I also recall that, for example, Kip Thorne simply refused to teach classes that were not pass-fail.
