Wednesday, June 4, 2008

Teachers and Student Race

I've been MIA for a few days while I busily jumped through hoops. I just read through all the blogging I missed, and here's the exchange I found most interesting:

Eduwonkette pointed out that a wide body of research has found that teachers are less likely to stay in schools with a large population of racial minority students.

Mike Petrilli responded by asking if that meant that teachers are racist and points out that KIPP schools and some others have had success attracting teachers to teach in schools with overwhelmingly racial minority student bodies.

I have two things to add:

1. No, it doesn't mean teachers are racist. Recent work has found that teachers tend to end up in schools where they feel more comfortable -- where students and teachers are from the same geographic area, social class, race, etc. as themselves. That makes sense. People like be surrounded by familiar things.

2. Pointing out that KIPP attracts teachers to teach racial minorities is a bad example for two reasons. First: I don't have statistics handy, but I don't think KIPP's teacher retention rates are too stellar. KIPP and some other high-flying charter schools rely on young idealistic teachers (including many in TFA) who are willing to devote their lives to the school for a few years before they move on to something else. Secondly, even if KIPP retained the talented teachers it recruits, that wouldn't be proof that there are enough people willing to teach in these schools across the entire country. I'll agree with his point that the working conditions probably matter more than the racial make-up of the student body, but that doesn't mean that we shouldn't consider the racial make-up of a school.


Roger Sweeny said...

Last Sunday, Dennis Fermoyle made the following criticism:

Jay Greene's blog has a piece telling us that students involved in Milwaukee's voucher program graduate at a higher rate than those not in the program. Well, as Gomer Pyle once said, "Surprise, surprise!" [snip]

I don't think too many people would argue with the proposition that children of parents who care about education will generally do better than those who don't, and that includes graduating at higher rates. It also seems obvious that parents who go to the trouble of enrolling their kids in a voucher program probably do so because they care about their kids education. If we put all those kids together in a private school setting where kids who don't try and don't behave can be kicked out, how could they possibly do worse?

I fear you have made the same mistake when you say, "Eduwonkette pointed out that a wide body of research has found that teachers are less likely to stay in schools with a large population of racial minority students."

But the Hanushek, Kain, and Rivkin paper that eduwonkette cites to very specifically says that race may not be what's causing the loss:

But, we might speculate that these [racial] estimates at least partially proxy for more general working conditions (even though our analysis does not permit disentangling the various potential aspects of working conditions). For example, if schools with high minority concentrations have more disciplinary problems, rigid bureaucracies, poor leadership, high student turnover, and general safety concerns, improvement in such directions may reduce teacher turnover.

On average, "schools with a large population of racial minority students" are worse places to teach. It's hardly a surprise that teachers are more likely to leave.

Corey Bunje Bower said...

While only one paper was cited, a number of others have found similar things. I can definitely see the argument that other things matter more than the racial make-up of the student body in terms of teacher retention, but teachers also self-sort by class and race when selecting schools in which to teach.

Roger Sweeny said...

While only one paper was cited, a number of others have found similar things.

And I suspect they all have similar limitations. I doubt that there is a data set which would allow you to control for all the things other than race that cause teachers to leave.

The data, limited as it is, tells us that there is a correlation between large numbers of minorities and a lack of retention of white teachers. It simply does not tell us why.

If you compared residents' IQ with their zip code, you might well find that the "smartest" zip code is 02138. Does that mean that there is something about 02138 that makes people smarter, that if you want to have smart kids, you should move to 02138?

From just that data, you might conclude so. But once you find out that 02138 includes all the Harvard undergraduate dorms and the residences of many of its graduate students, a very different picture emerges. Living in 02138 doesn't make you smarter. Smart people preferentially move to 02138.

Without better data, the "black students cause white teachers to leave" hypothesis isn't much better supported than the hypothesis that you can make yourself smarter by moving to 02138.

Corey Bunje Bower said...

Yes, but white teachers are less likely to teach in schools with large populations in the first place.