Friday, April 23, 2010

Pundits on kids

E.D. Hearse takes on the most widely read morons in America:
She also challenges the claim by economists, and taken up by influential writers such as Malcolm Gladwell and Nicholas Kristof, that if only students were exposed to excellent teachers for three to five years in a row, the achievement gap between advantaged and disadvantaged students would be overcome. It is true that some math teachers have been able to make up as much as 33 percent of the lag for some students in a single year—whence the economists’ extrapolation to the three-to-five-year figure.
Not, of course, that one can assume either Gladwell or Kristof knows what an extrapolation is. I generally recommend the article -- a review of Diane Ravitch's [a.k.a. Dan Ravage] new book on how school choice has failed -- as it is sporadically sensible and has a lot of interesting factoids. E.g. here's Richard Hofstadter on child nature (on which see also Grobstein):

To believe that Dewey’s synthesis was successful required a certain credulity about the pre-established harmony between child nature and democratic culture which not everyone could share. It seemed…that one would have to give up either the emphasis on child nature or the emphasis on educating for democracy.

I would however part ways with both Hearse and Ravage when they rhapsodize about the "neighborhood school" as a pillar of the community. (I also find Hearse's whining about international league tables a little tiresome. Bloody awful, isn't it, that other countries educate their kids?) Most of the real problems with the school system have to do with schools for the poor; these tend to be in areas where there isn't a meaningful sense of community -- whose residents are in any case largely transient -- and the approach of the successful, no-excuses sort of inner-city schools has been to give kids an alternative social anchor from their "communities." Beyond that I don't think K-12 education can be sensibly addressed in isolation from the Big Sort -- it's a good thing to get middle-class kids into the public school system as their parents function as watchdogs, but it's not sufficient because it doesn't deal with the districts where most parents are too illiterate and/or busy to find out what their kids are learning at school.

Another bit of information that I think is crucial to this entire picture is this study (I'm quoting from an email I wrote a while ago):
[...] there's Richard Elmore's paper on high-performing schools. What Elmore found was that high-performing inner-city schools were different from both suburban schools and crap inner-city schools, which were rather like each other. In suburban schools, the instructors assumed (correctly) that parents would play the primary role in the education of students, and were hands-off; in crap inner-city schools, the instructors (incorrectly) made the same assumption, and let the kids smoke crack in class; in good inner-city schools, instructors realized that they had to make up for the lack of useful parenting, and put a huge amount of work into making sure the students were kept engaged.


andy said...

I have been hearing lately, much to my dismay, that the empirics of school choice aren't holding up so well. I am not totally ready give up on the concept yet though, mainly because the "illiterate/busy" parents are those least likely to be able to work the political mechanisms necessary to keep local schools in check, even if they cared and weren't busy. I also never thought it would be that difficult for parents to determine how good a school is- just look at where the graduates end up- but perhaps I am misoverestimating the difficulty of such a thing. I also think that a system in which schools are more independent would be one in which schools would be more likely to adjust their structures and curricula to the local students, as opposed to the current where the same type of school has been put everywhere regardless of efficacy.

The nefarious role of the teacher's unions cannot be overlooked either, and it is worst where the parents don't have the wherewithal to keep the teachers in check. Perhaps a favorable change would be to give principals more discretion over management of teachers and the structure of the teaching, and then have the school board hold them accountable for the overall performance of the school. Such a system I think would get the incentives right, and it could be made to work for schools in any district. You can't overlook the fact that our public schools tend to suck for the middle classes as well, but our universities and other mechanisms, like private tutors, help to fill in the gaps. Thus the situation for the middle isn't as desperate as for the poor, but it is still tremendously inefficient.

Zed said...

Yes education is messy. For a long time I thought the solution was to force school districts to be socioeconomically diverse (or to merge them into megadistricts that were so) so that _some_ parents in any district would give a fuck. This is however a bad idea if poor kids have different needs than well-off kids; the well-off parents would only make things serve their kids' needs. As this leaves us with no obvious solutions I'm in favor of experimentation but not sure how it should be managed -- school choice per se is probably insufficient as I don't think the relevant parents are tuned-in enough to be a market mechanism -- but I'm not optimistic about central planning either. (OTOH the govt. does run things like the NSF, the NIH, etc. that do a good job of funding research without being central planners in the slightest.) Maybe the right place to start is with changing and retraining the teachers, as Hearse suggests, and as you say the unions will rabidly oppose this.

One thing I do approve of is centralized floor-like curricular targets for schools to achieve, while leaving the pedagogical details up to the schools; to this extent I'm pro-NCLB; where I think NCLB goes awry is its hamfisted way of dealing with bad schools. The notion that there's something evil about "teaching to the test" seems incoherent to me; the point is to structure tests sensibly enough that the best strategy for doing well on them involves _some_ degree of learning the material.

The situation with suburban schools is very different; the outcomes are not terrible and the basic problem is that these schools soak up money that could go elsewhere without really hurting outcomes. Of course this is politically impossible.