Yes, it raises understanding. But takes longer.
The average student understands a bit more a bit better, the bottom 25% understands a lot a bit better, but the top 25% understands less just as well as before.
It's one way of closing the "achievement gap": You pull up the bottom 25%, but at the same time the top 25% stagnates a bit.
American education has a problem: It can't figure out what its purpose is. Is it so equalize opportunity? Produce the best trained graduates possible? Ensure that the majority are fairly well trained, or that everybody who starts graduates minimally trained? Each goal has a downside: If the goal is equal opportunity, then the goal isn't education but social engineering and the content and methods reflect that. If you want the best trained grads possible, then you want weeding out. If you want the majority well-trained, then you'll trim the top and bottom. If you want everybody trained to at least the same level, you're likely to get few trained above the minimum because that's not a priority (and it's damned hard to make sure everybody learns, given differences in cognitive functioning, background, study skills, etc.)
The one good thing in this kind of approach that benefits nearly all students is breaking up the lecture--you lose a bit of continuity, but you let the brain recoup a bit to allow greater focus. That can be taken to extremes, and usually just avoiding 80 minute class periods and giving people a break half-way through a 50-minute lecture class is sufficient.
In any event, if you have your tests properly aligned with what's to be learned and the questions are normed and validated then raising understanding will result in higher test scores. (It's just that you can get those several ways, each of which reflects a different set of goals.)
= new reply since forum marked as read