Teaching is odd for a few reasons: it requires you, at all moments, to act demonstrating the complete conviction that you know what you are doing and are following a plan. In fields where textbooks are common, this is less of an issue, but for most political theory courses, there are a number of different ways to construct a semester and none is necessarily superior to the others.
The other is that feedback is hard to come by. There are rough indicators, including average attendance and whether people yawn or look at the clock, but these are imperfect signals: a person might skip (or yawn) because your class isn't interesting. They might also be absent because their life has the usual college-student drama. It's possible, as well, to attempt to get feedback during the semester, but there are dynamics that corrupt the quality of that information: the weird interplay for the student between the desire to be honest and the desire to receive a good grade. The problem for the instructor is that there are certain things it takes an entire semester to understand; a middle-of-the-semester evaluation will say less than a retrospective one. (Alan Jacobs somewhere suggests that real evaluations would be given the semester after the course has ended; he's not wrong about that.)*
I've been typing up the written comments I received on my most recent set of evaluations, and it's been a quite lovely experience. I told my students that while end-of-semester evaluations might seem like one undifferentiated thing to them (as it did to me when I was an undergrad), they are composed of two distinct pieces of information that have two different audiences. There's a numerical component which is very important: it's a quick shorthand for hiring committees to know how good of an instructor you are. Everyone who gets an application from me will see those numbers, and they may well play into hiring decisions. The other component is the comments, which, I told my students, very few people will ever bother to look at, and will be read by (quite possibly) no one but me. I instructed them to be brutally honest about things they liked and various shortcomings of the course. I'm still a relatively early in my academic career, and how will I know how I appear to students unless they tell me? So it is surprising and oddly touching to read a number of effusive and lengthy comments: written for an audience of one, long after it will have done the students any good to say nice things to me, and very much appreciated.
*The additional problem is, as Jacobs suggests, that there are some things that don't pay off until much later. The students in my freshmen seminar did not appreciate having to write as much as they did, even though they all managed at least some improvement over the semester. That seems like a fine example of something they might think differently of later on in their college careers.