[asoneth]

With respect to cost, I'm not sure any of these points refute the hypothesis that charters cost less per pupil primarily because they are able to offload the most expensive students. And if that is true then increasing the number of charter schools will simply increase the concentration of the most expensive students into fewer schools that have less funding to deal with them. At that point we can either acknowledge that expensive students require an order of magnitude more money or we let them pass through to the welfare and justice systems. In either case it doesn't seem like we're saving society money by switching to charters, we're just shifting it to other parts of the budget.

With respect to educational opportunities that seems more compelling. But as Sowell alludes, I suspect you could get a similar effect by allowing public schools to enforce stronger discipline and make it easier to suspend or expel difficult students. On that note I'm surprised to hear that teachers unions are opposed to that now. I've certainly heard that unions are an obstacle to getting rid of poor-performing teachers, but I rarely hear of them being an obstacle to dealing with difficult students. At least a couple decades ago the AFT was actively pushing to make it easier to expel students for things like drugs or weapons in school.

[sreque]

So I think an underlying assumption you have that I want to scrutinize is that misbehaving students are costlier from a monetary perspective, or that they require more money spent to handle them, or that spending more money will result in better outcomes for these students. This assumption may or may not be true, but I think it is valuable to question it and expect some evidence to support it. For instance, it is quite possible that stricter and more rigorous discipline in schools could benefit misbehaving students far more than any spending increase in the school system. Likewise, there may be certain cultural norms or expectations in charter schools that have far more impact on misbehaving students than spending increases.

I think the far more important point, however is that some charter schools are working phenomenally well for the underprivileged and doing so with less funding, and that there are valuable lessons to be learned from that fact, which lessons are at risk of being ignored or lost. However, instead of either trying to learn from these charter schools or allow more of them to be created, teacher's unions and the government officials they support via campaign funds are openly hostile to them, as the book details. These adults are clearly acting in their best interests, not in the interests of the children they are claiming to serve.

With regard to government policy in general, there seems to be a complete disincentive to analyze policy in retrospect honestly, determine successes and failures, and learn from the past in order to influence future decisions. With government policy, intent often matters more than results, as intent earns votes. This can easily lead to perverse incentives.

As an example, it's easy to claim good intent when proposing to spend more money on schools. But, if we want to actually help children out, results matter more than intent, and it seems very clear that, past a certain point, pouring more money into the public school system has little to no impact on educational outcomes. Instead, we should both be looking at other factors, and we should be allowing for more competition so that we can iterate on more ideas more rapidly. The relative monoculture of the public school system, combined with perverse incentives among both the government and teacher's unions, seem unhealthy for society and for our children.

[asoneth]

To be clear, I am not arguing in favor of teachers unions or short-sighted politicians. I am not arguing that our education dollars are being well-spent since it is clear that we are spending more to get worse results. I am not arguing that increased funding has a bigger impact than improving discipline.

But it does seem logical to me that given a set of students, some of them will end up costing more than others. To give a trivial example, a student who repeats a grade costs more to educate than a student who skips a grade, just by virtue of spending an extra two years in the system. Similarly, it seems logical to me that students who can't afford lunch, need additional ESL or special-needs teachers, are physically handicapped, are violent, etc will cost more than the average student. (And sure, kicking that violent kid out lowers your costs, but it simply shifts that cost to the juvenile detention system.) Assuming the above is true, any school that is able to attract slightly more low-cost students or discourage slightly more high-cost students (however they are able to do so) should enjoy substantial cost savings.

> some charter schools are working phenomenally well for the underprivileged and doing so with less funding

I do not dispute that there are positive examples. From the statistical analysis I can find the overall effects seem mixed: in some states charter-school students do better (on average) than public school students, in some states worse, and in some states like Ohio there's simply more variance between charter school students. And since I do not know is what this effect is due to I'm not sure to what degree this approach scales. For example, do charter schools work equally well in a region where there are no more public schools left for them to dump difficult students onto?

One way to test this would be to randomly assign a population of students into a public school system and a charter school system, give them the same funding, and require that each system accommodate even the most difficult students who have special needs or disciplinary issues. (Not necessarily in the same classes, just that they are responsible for their education.) If charter schools have better educational outcomes in such an experiment then I would find that truly compelling.

I think it's absolutely useful that states are taking such different approaches to charter schools though, and I think eventually that will give us the data we're looking for.