The teachers I've known don't really care about measuring knowledge. They're looking for a reasonable way to motivate engagement with the class, that's not too disruptive of the overall flow of the course. One professor told me, "A student who has made an effort to work through the homework problems a couple times should be able to easily get a B on the exam.
Testing also acknowledges that you're competing for your students' attention, and if you give no assessments, your students will rationally focus all of their effort on the courses that do. Preparation for the test becomes a reasonable measure, not of your knowledge, but of how much effort you need to apply to a course. Since students have been taking exams for years, each student knows how to calibrate their own level of effort.
As a student, after some trial and error, I developed a pretty good routine for getting A's in the two kinds of classes I was taking: Those that were dominated by solving problems and proofs, such as math and physics, and those that were based primarily on written assignments, such as art history.
I can tell you that almost every teacher who is not burnt out, does care about how we measure knowledge, mainly because they have to. The big difficulty is that there are two roles for teachers, on one hand you are a mentor and supposed to impart knowledge onto your students (the teaching part). On the other hand you are a gatekeeper, you are supposed to check that the thresholds for some qualification are met. Now if we had an ideal way to measure knowledge those two roles would not really be in conflict with ech other, but because we don't teachers have the difficult job of trying to teach a subject and at the same time find a good way to see if the students actually learnt what the were supposed to. All that with a limited amount of time that is available.
What do they do with the information? The threshold in most courses is to be able to pass the next course. The students who won't do that, tend to drop out, or switch to an easier major, of which there are many.
Teachers do tend to change their content and methods if a large number of students are failing exams, but I think it's based more on a hunch, than on hoping that test scores will yield analytical quality data. This is the sense I get from talking to a lot of teachers. My only teaching experience was one semester at a big ten university, a long time ago.
The teachers I've known don't really care about measuring knowledge. They're looking for a reasonable way to motivate engagement with the class, that's not too disruptive of the overall flow of the course. One professor told me, "A student who has made an effort to work through the homework problems a couple times should be able to easily get a B on the exam.
Testing also acknowledges that you're competing for your students' attention, and if you give no assessments, your students will rationally focus all of their effort on the courses that do. Preparation for the test becomes a reasonable measure, not of your knowledge, but of how much effort you need to apply to a course. Since students have been taking exams for years, each student knows how to calibrate their own level of effort.
As a student, after some trial and error, I developed a pretty good routine for getting A's in the two kinds of classes I was taking: Those that were dominated by solving problems and proofs, such as math and physics, and those that were based primarily on written assignments, such as art history.