
I suspect each has its purpose. A 5 question test can be a useful memory retention exercise with recently recently presented content by helping students committ the right information to memory vs content that is not as important. A 20 question test is a better assessment of broader sets of content. But even a 20 question test can be manipulated to be learning tool. By careful attention to a progressive question sequence, a teacher can build a kind of dvidiing line between what a student knows vs what they don't know. Subsequent tests would then be used to see if the student has moved that line.