April 25, 2006

In which our hero ...

... takes on the union drones at EDWise on the subject of class size.

Update: Well, that stayed somewhat civil and the teachers made some good points (of course, they were points I had already conceded). No one had a response for my assertion that the current research failed to show any advantage in decreasing class sizes across the board. That's because there is no valid response. If there were any valid evidence, and the UFT surely has access to classroom data, they would have ran it up the flagpole long ago.

So, I'll stick with my present opinion: first you fix the broken instructional delivery system, then you worry about tweaking the class sizes to make sure the implementation can handle the class size, targeting areas that need reductions the most, like the early elementary years in low-SES schools.


Anonymous said...

The horror here is that you don't live in Irvington.

Anonymous said...

I'm not sure what to think about class size, though my sense is that the research is mixed at best. (Stevenson & Stigler recommend increasing class size, btw.)

What do you make of this study?

Using Mainmonides' Rule to Estimate the Effect of Class Size on Scholastic Achievement

Kilian Betlach said...

Here are some resources your adversaries were apparently unable to supply:






They key of course, is to get the class sizes low. Reducing from 35 to 25 is not enough. The drawback, sadly, is reduced class size is expensive, and causes increased demand for teachers, often at the last moment, which causes ineffectual, uncommited, and/or unqualified teachers into the classroom, in turn lowering acheivement and weakening the effectiveness of the reduction.

I think a better move than across the board reduction is to target it for key demographics -- one of the reasons we need that data disaggregated along such lines. Here I'm speaking primarily of English Language Learners, but also low-income, under-taught kids across the board.

KDeRosa said...


Check out tables 4-6 from The effectiveness of Class Size which shows that the vast majority of these studies do not have statistically signficant results (bad science). In any event, the effect sizes of the good studies was about half the size necessary to be educationally signficant (1/4 standard deviation).

Of course, this doesn't appear to make sense since smaller class sizes would appear to be easier to impart knowledge. This is why I say that the effectiveness of class size reductions is highly dependent on the quality of the instruction. For example, in DI reading classes with low-performers the instructional size needs to be less than 10, preferrably less than 8, before the instruction breaks down.

Kilian Betlach said...


It sounds like the issue is less that class-size reduction ought to be rejected out of hand, and more that the current research is methodologically flawed. By all means, let's fix the latter, because the personal experience of a great many educators (including this one) argues that this is a long overdue idea.

Obviously, teacher quality is a huge factor, but dismissing the reform as a result throws the baby out with the bathwater. I think we are in agreement, however, that large scale reductions are both unnecessary and carry many unforeseen consequences, and any move in this direction ought to begin with low-income (and especially) ELL kids.

KDeRosa said...

TMAO, I'm going to post on this topic later this week. I think we are pretty much in agreement.