One of the interesting things about working on both K-12 and higher education policy is observing the differences between them. In K-12, there's a long-running and very active research and policy conversation about class size. Books have been written, studies conducted, debates organized, policies implemented, all focused on the relationship between student / teacher ratios and learning. People argue about whether there's a specific threshold class size associated with learning gains (15:1? 18:1?), whether there are differential effects for different student groups, whether marginal dollars are better spent hiring more teachers or better teachers.
All of this occurs despite the fact the class sizes in K-12 don't actually vary all that much. The vast majority are probably somewhere between 15 and 35 students. In higher education, by contrast, class sizes in similar courses can range from less than 10 to 500 or more. In theory, that should permit for even more robust inquiry into the impact of college class size on teaching and learning, particularly since student / faculty ratios are a commonly-accepted and often-used measure of institutional quality.
Yet virtually no such research exists. And this is true for lots of other basic elements of the higher education teaching enterprise. This is the subject of my new column at InsiderHigherEd. Read it here.
4 comments:
Kevin,
You're making a straw-man argument to knock down here, implying that there is no research on undergraduate achievement. Maybe there's little on class sizes in undergraduate classes, but a quick Google Scholar search on "college mathematics journal" and "meta-analysis" implies that there is serious research going on.
Since many higher-ed management issues (often including class sizes) are beyond the level of a department's faculty, the items of interest to those faculty (the people who would be putting in the time on the research) are pragmatic: how do you teach the fundamental theorem of calculus? What demonstrations are going to make sense and help students learn? The fact that the published research doesn't coincide with a particular interest of yours doesn't mean that people are entirely uninterested but that there's a gap. That's life in the research world. I keep a list in my head of topics on which I'd love to see great research. That doesn't mean that the relevant fields are full of incompetent or uncurious people, and that's pretty much what you said in the IHE piece. I think it was largely unfair.
Just left a comment on IHE and then saw this in my reader. As I mentioned over on IHE, there has been quite a bit of work on class size in the economics education literature, so I'm skeptical of the statement that 'virtually no research exists'. Of course, the research that does exist is no more conclusive than in the K-12 world - depending on who you ask, STAR was conclusive or it wasn't. But that actually feeds into my other point in my IHE comment which is that more data may not persuade anyone, at least not in a way that matters. As Sherman points out, decisions about class size are rarely made by those teaching the classes, and those who do make those decisions simply may not be convinced by any data you throw at them.
Sherman,
I think you're the one indulging in straw man argumentation here. I didn't "imply" that "there is no research on undergraduate achievement." That would be absurd. I just said there's not nearly enough such research, and offered examples to support that contention. I don't have any particular interest in pulling punches, here or elsewhere, so if I had thought words like "incompetent" or "incurious" were appropriate I would have used them. Instead, I think this is mostly about incentives and existing organizational arrangements -- a point you pretty much make for me when you note that the people in a position to study class size aren't in a position to influence it, and thus don't study it. Right -- and that's a big problem.
Plus all the classes in such a study would not only have to teach similar things but also all give the same test (assuming test scores are the outcome variable)
Post a Comment