There are two common reasons for testing. First, the static fashion of determining roughly (and often rather arbitrarily) what students have learned and can repeat from the curriculum. Second, the dynamic fashion of assessing what has been learned so the teacher can adjust to the realities of the students’ experience. A third reason is, of course, to assess students for college admissions or their status as compared to other students, but this standardized mode of testing is so pedagogically problematic and even vulgar that I find it hard to mention it.
The first is by far the most common. It is driven by the need to produce data for administrative supervisors and has minimal value as an aid to student learning.
The second is less common in one sense, but on the other hand, every good teacher is doing it continually. So it is less formally common, but informally, I hope, much more common. It is certainly more necessary.
If teaching involves a relationship, one would hope that the teacher is in a continual interaction with the student so as to adjust her teaching according to teh circumstances and reality of the actual student in her presence instead of a theoretical student who fits somewhere on a bell curve.
I refer the reader to this article at Edweek on how some schools are even formalizing what I shall call “dynamic assessment”: Testing to Teach: Using Assessment to Shape Instruction
Here at CiRCE, we apply dynamic assessment in the way we teach The Lost Tools of Writing.