When we build an assessment tool, we can ask what levels the
questions are targeting. It's easier to grade items which just test
memory (bottom level). Far harder to come up with good questions
(and to grade them) involving synthesis and comparison.
