Welcome to the Profit of Education website. Continuing the conversation begun in the book Profit of Education, we discuss the latest economic evidence on education reform.

Class size redux

Voters and politicians love class size reduction. This love affair persists in the face all evidence that class size reduction is not a sensible reform. It’s not hard to ballpark the costs of smaller classes. You need more teachers, more classrooms, etc. Roughly, cutting class size 20 percent raises costs 20 percent. Will you get better results from smaller classes? Probably, but only a very little bit better. Ain’t worth it.

Matt Chingos has taken a look at the Florida statewide class size reduction put in place last decade. Florida’s schools did improve. But then Florida made lots of other changes too. So what Chingos did was compare outcomes in schools that were required to reduce class size to outcomes in schools that already had smaller classes. His summary:

…mandated class size reduction in Florida had little, if any, effect on student achievement in math and reading. The district-level analysis, which focuses on grades six through eight, yields no statistically significant effects but does not have sufficient statistical power to rule out small positive effects. I am able to examine test scores in grades three through eight using the school-level analysis, and find no evidence of positive effects and some evidence of negative effects. In general, the standard errors are small enough that even modest positive effects can be ruled out.

Let me add one more note, based on adding some of my own calculations to Chingos research. The long-run cost of the mandated class size reduction was about $1,600 per pupil. That’s close to 15 percent of total school spending.

Our kids could use more science and less make-ourselves-feel-good in school reform.

This entry was posted in Uncategorized and tagged , . Bookmark the permalink.

One Response to Class size redux

  1. Jesse Rothstein says:

    I must say, I find this study pretty unconvincing as an evaluation of whether class size reduction is “worth it” (which, for what it is worth, isn’t at all the way the author describes it). He’s pretty explicit that he’s comparing schools/districts that were given additional resources and forced to reduce class size with others that _were given the same extra resources_ but were already compliant with the class size mandate so were free to use the resources however they want. So if the former group didn’t improve relative to the latter group, that just means that districts given freedom to use resources as they see fit find uses that are as productive as is class size reduction — it doesn’t at all mean that class size reduction doesn’t work.

    Moreover, the results are pretty imprecise. Even with a bit of cherry picking, in his district-level analysis (which I find the most credible) he is only just able to reject the magnitude of effects that were estimated from STAR, and that only in reading and not in math. And of course STAR compared class size reduction to a control condition _without_ extra resources, so one would expect the effects to be the same only if districts given unrestricted funds totally fail to use them productively.

    Finally, I have my doubts about the control groups. Roughly, his “treatment” group is South Florida, and his “control” group is northern Florida and the panhandle. These are not exactly comparable places. It looks as if in the pre-policy period the treatment group was paying much higher salaries and making up for it with larger classes. I’m not sure I’d expect either that the trends should have been the same in the two groups or that the marginal productivity of extra revenue is the same in the two places.

Leave a Reply

Your email address will not be published. Required fields are marked *