Scale matters

In an altogether excellent piece on medical innovation, Tyler Cowen notes:

The NIH works as well as it does because the money is mostly protected from Congress. It is not a success which can easily be replicated. The more money is at stake, the more Congress wants to influence allocation. We should guard this feature of the system jealously and try to learn from it. If we can.

This is a seriously, seriously underrated factor in public policy analysis, and I include the libertarian variety. The fact that you can do something awesome with $15 million does not mean that you could do something super-awesome with $150 million. It may simply not be possible to broaden what you are doing very much before countervailing forces--such as congressional interference (Exhibit A: the goddamn Acela)--kick in.

Since we've been talking a lot recently about vouchers, education is one area where this is fairly easy to see. You get a pilot program: a curriculum, a teaching method, a high-intensity preschool program (such as the Perry program) for disadvantaged kids. You do a rigorous study of that pilot. It produces terrific results. Naturally, we should roll it out everywhere!

Not so fast. That pilot program has a huge administrative staff whose sole incentive is to ensure that it is meticulously carried out. In the real world, that curriculum will be put into place by an administrator whose priority list is crowded with everything from mollifying the latest lunatic on the school board, to ensuring that she gets out of town for a three day weekend with her new boyfriend who she really thinks may be The One.

That pilot program is staffed with a narrow band of extremely highly qualified teachers, sifted from the best the environment has to offer. In the real world, whoever happens to be standing in front of the classroom come September 5th has to do it, even if they flunked Remedial Math four times and only got this job because the school board needed a body.

That pilot program is rigidly policed for deviations from standard procedure, because deviations will kill the accuracy of the result. In the real world, tranquilizing the kid who just pulled a knife during study hall may take priority.

The pilot program is supported by a crack team that will move heaven and earth to ensure its completion; if funds are tight, they will not sleep until they have procured another grant. In the real world, it's probably less important than redecorating the teacher's lounge.

The pilot program has buy in from all participants; schools, teachers or students who don't like it, don't believe in it, or don't want it anyway, have already naturally dropped out of the sample. They will thus be striving to actually put it into place as closely as possible as described in the prospectus. In the real world 60% of everyone will think this is a moronic idea, and most of the rest will strenuously resent the intrusion on their autonomy.

Result: what worked beautifully in pilot will generally fail miserably in wider execution.

In business, these facts are summed up (over and over) with the dolorous mantra: didn't scale. In the public sector, that realization is still coming very hard.

Comments

A friend summarised your complaint about schooling for me many years ago. "The trouble with educational experiments" he said "is that they all work."

Posted by dearieme | November 14, 2007 12:56 PM

This reminds me of many of the criticisms of Jeffrey Sachs' Millennium Villages.

Posted by Christina | November 14, 2007 1:17 PM

Megan recognizes the general problem with scale, and especially the problem that can arise when the sample being studied is not representative of the population. Tyler Cowen discusses a more narrow dynamic, which I think economists might call “bureaucratic inefficiency.”

An efficient operation is one in which autonomy rights are clearly drawn, and the person who exercises discretion can enforce those rights with little cost. In an inefficient operation, autonomy rights are ambiguous, and the decision-maker must expend much effort and resources to implement a decision.

Employees often have tastes and preferences that run counter to their boss’s. An employee wants to use workplace assets for their own purposes. Maybe he’s lazy, wanting to use his own labor time for recreation. Maybe he wants to use the organization’s budget to provide nicer surroundings for himself. Maybe he wants to promote his religion. And maybe he want to promote what he understands to be the mission of the organization. And he may want to do any of these things regardless of what his boss wants him to do. An employee’s ability to pursue any of these agendas depends upon the amount of bureaucratic friction in the system.

Arguably, national executives and legislatures have an incentive to treat government as their private property, wringing every last bit of personal gain from the system. In return for campaign contributions, politicians sell public assets such as no-bid government contracts, patronage jobs, stays in the Lincoln bedroom, pardons, draft deferments and even office furniture.

Through the theory of the “unitary executive,” George W. Bush has arguable run a very efficient administration. His administration has managed wrest control of more levers away from the bureaucracy than perhaps any other. As a result, there has been ever less room in government for bureaucrats to cultivate pet projects like professionalism and competence, and ever more pressure to operate in a manner consistent with the idea of government being the private property of the executive. As a general or a prosecutor or a budget analyst, you may feel compelled to act according to the dictates of your profession rather than the dictates of your political masters. And when you do, you’re fired. The remaining generals, prosecutors and budget analysts get the picture pretty quickly.

As Robert Reich notes in his book Supercapitalism, this is a dynamic of bureaucracy, not government per se. If your corporate board isn’t very demanding, a CEO can humor his interest in making charitable donations with corporate money. If the board is more demanding – and they will tend to become more demanding over time – those donations dry up. If you’re rich enough, you can honor the restrictive covenant agreement with your neighbors not to sell your house to a Catholic, even if the Catholic is offering the best price. But as financial pressures grow, you may not have the luxury to be so honorable. If you’re not under big pressure to win ballgames at any cost, you’re free to refuse to hire players you don’t like, whether they have a bad attitude or belong to an ethnic group you disfavor. If you have a big need for winning, then these other concerns will have to take a back seat to your need to find the best players possible.

Starting in the 1980s, regulators would compel energy utilities to prepare engineering reports about how to operate more efficiently and promote conservation. Of course, the conclusions arising from these reports would often run counter to the utility’s self-interest. But the engineering culture of competence, intellectual rigor and risk aversion that dominated most utilities was sufficient to cause utilities to provide useful information about themselves. Since then, MBAs have replaced engineers at the helm of most energy utilities, and a business culture of self-interest has largely displaced the culture of engineering competence.

I’m glad to hear that things are working so well at the NIH. In contrast, France is famous for its unbridled bureaucrats, and the results are not always as salutary. Similarly, many people grumble about “unaccountable federal judges” in the US; they are among the few federal employees that feel free to speak against the current administration. Clearly, there is no guarantee that bureaucrats will act with idealism if left to their own devises. But there’s pretty much a guarantee that politicians won’t.

Posted by nobody.really | November 14, 2007 1:51 PM

I don't see the link to Tyler's post in your post. I added it as the URL for this comment. Clicking on my name should take you to his blog post.

Posted by Tim Fowler | November 14, 2007 2:04 PM

The manhattan project and later the Apollo Project are often used as examples of sucessful crash efforts. In both cases the key information was obtained cheaply and only the implementation (and engineering) was put on a crash basis.

The basic physics for the bomb were all done in small labs on small scale. The remaining engineering problems were large but amenable to speedup. Even then muliple teams on parallel paths were used. They also had the advantage of limited political interference. The sites were not selected as pork, but for project reasons.

Apollo had all the basic science done by Issac Newton, and most of the engineering done in the 1950's under Eisenhower. The moon race was then a single high priority development task. In this case we still have the pork related dispersal of NASA facilities.

Posted by Dave Moelling | November 14, 2007 2:10 PM

Your concerns about scaling, and the incentives for pilot programs to "prove" their effectiveness, are well-placed.

But your assumptions about the public sector aren't serving you well in the rest of this post. Judging by my experience working on NIH grants, the real secret to their success is that they throw boatloads of money down the drain funding ill-conceived and poorly-designed experiments in order to find a few that work. Their procedures for renewing grants are somewhat more rigorous, but there's still a glaring tendency to confuse the quantity of publications for quality.

The research is often so highly specialized that virtually no one, possibly including the primary investigator, is really qualified to judge it on its merits. Congress maintains its distance from the funding decisions because they can't even pretend to have the expertise to judge where the money would best be spent. Secondary education really isn't quite so complicated.

Additionally, every academic scientist has a strong incentive to publish data showing a limited success that requires further funding to pursue. Unsurprisingly, life science journals are packed full of articles that provide a not-very compelling set of data and conclude that follow-up research is necessary.

Attempting any variant of this model in education or public works would be a terrible idea. Allowing insider cliques to disburse funds with no legislative oversight is a recipe for waste and abuse. Congressional oversight, however imperfect, at least shines a light into the dark corners and forces the recipients of public funds to justify their existence in layman's terms. The reason for the NIH to maintain their model is not because this model is appropriate for all public spending, but because 5% of the funded research is so extremely productive and beneficial that we're willing to tolerate all the waste and abuse.

Posted by LaFollette Progressive | November 14, 2007 3:04 PM

You're absolutely right about scale. And it's as powerful an argument as I've ever heard against school vouchers. What you complain about with health care is exactly what you advocate with private school, that any advantages it confers in education can be scaled upwards by a truly staggering degree, to populations enormously different from those taught by private school currently.

The current rejoinder is also the same concerning both education and health care-- that the status quo is so bad, anything is preferable. And that rejoinder is unsatisfying, lazy and irresponsible for both. Neither is a genuine response to the concern, and neither is an intelligent or correct basis for public policy.

Posted by Freddie | November 14, 2007 3:06 PM

Edit: I was primarily responding to Cowen's post, not McArdle's. Should have made that clear.

Posted by LaFollette Progressive | November 14, 2007 3:08 PM

A good example of this principle is early childhood education. A small and lavishly funded program like the Abecedarian project can work, but the benefits of a national program like Head Start have been much more difficult to demonstrate (see http://www.brookings.edu/papers/2003/07childrenfamilies_haskins.aspx and http://www.cato.org/pub_display.php?pub_id=5483 )

And needless to say, we voucher supporters should admit that this applies to vouchers as well. Vouchers may work in small-scale programs in a few urban areas that already have failing public schools and a healthy network of pre-existing private schools, but that doesn't mean that they will necessarily work on a larger scale, or in suburban or rural areas (where the public schools may really be the best thing around), or in areas that don't successfully monitor which schools are getting the voucher money (see http://www.jsonline.com/story/index.aspx?id=333093 ).

Posted by Stuart Buck | November 14, 2007 3:20 PM

but Freddie,
private schools wouldn't be scaled up; there would be multiplicity of differently scaled educational organizations.

Posted by josh | November 14, 2007 4:30 PM

re: NIH as exemplar.

Is a fiction. Walk around the halls for a few days and you'll understand how it doesn't hold a candle to a (private) teaching and research hospital. (For just one of a dozen disasters, look at the state of any NIH labs infotech).

Every agency sits in the same soup of broken congressional incentives (we get what we reward and dont' get what we punish). The system that creates a (failing) intelligence community, DOE labs that can't keep a secret, and high-schools whose best students can't compete with the average Asian in college without a years worth of remediation in math and science has no wildly successful agencies.

A great Principal Investigator has to spend a week every six "feeding the beast" by making their funding agency look good in front of their oversight committee. And this includes NIH. Only private funding agencies come close to matching commercial productivity (Wellcome Trust for one, Gates foundation for another).

Take the NIH budget and allocations and divide it by state. Like the NSF (aka "by the peers for the peers") you'll see to a first order approximation that it tracks congressional representation, with a thumb on the scale for committee memberships and senatorial representation. You'll also find the numbers hard to get.

One nasty side effect of a growing central government in a federal system (v. one that's starved) is our party system insures that a State's congressional team (that fights like cats and dogs in DC) will agree with religious certainty that the state deserves its share of Federal spending. And since rather than intellect and judgment, a representative government is first about compromise leading to consensus in service of a constituency, managed by a process, we have small states struggling to rationally spend "their share" (starting w/ 1/50th). So we get ag programs (to offset, say, Virginia's shipyards), bridges to nowhere, and supercomputer centers in Hawaii when a battleship and its 10,000+ jobs are lost.

NIH is like NASA (and FEMA under Mr. Clinton). They hand out lots of money and have a large PR and congressional affairs department. Could just as well be outsourced to Disney.

Every dollar NIH (and NSF) spends outside of market forces increases the odds I (and others) will die early and more painful deaths than necessary, if rather than making these god-like decisions for other, we let the random-walk of the market decide. Ditto for programs like Social Security. Think of how much more rich the poorest of us would be if we didn't play these games. And it always comes out of the hide of the poorest first (both this country and others). (us) Rich don't care. Shame on the so-called elites that don't have the wit to know better.

Posted by Ari Tai | November 14, 2007 4:36 PM

private schools wouldn't be scaled up; there would be multiplicity of differently scaled educational organizations.

Voucher proponents argue constantly about a vague and amorphous "private school system." We are asked to believe that wide-spread voucher programs would succeed because private school, writ large, is superior to public. If you're going to argue by such general and ill-defined terms, I'm certainly going to confront you on those same terms.

Posted by Freddie | November 14, 2007 4:43 PM

Freddie,

Lets not be pedantic here. Voucher supporters argue that moving some children from larger, bureaucratic, failing public schools into smaller, diverse and autonomous private schools gives the most children as possible a better chance to succeed.

No one is arguing we scale up the local Sisters of the Holy Cross and have them run 10 schools because they run 1 well.

Posted by Skullberg | November 14, 2007 5:10 PM

Yes, have to disagree here: the problem of scale in business is a problem of having one corporation try to dominate a market, not of having a zillion little producers. Indeed, one might argue that public education is the classic "didn't scale"; the biggest school districts are the most dysfunctional.

Posted by Megan McArdle | November 14, 2007 6:53 PM

MM wrote: Yes, have to disagree here: the problem of scale in business is a problem of having one corporation try to dominate a market, not of having a zillion little producers. Indeed, one might argue that public education is the classic "didn't scale"; the biggest school districts are the most dysfunctional.

Which is bolstered by the observation that companies which are large, diverse, and successful often get that way by allowing their entities to function somewhat autonomously, while pooling the big ticket items -- capital, accounting, legal expertise, and intellectual property. Thus an entity like the Dover Corporation can turn over a profit while casting its umbrella over industries as diverse as refuse vehicles, defense contracting, specialty electronics, and health/environment products.

School boards, by contrast, do not have strong competition to keep their methods efficient, and instead tend to develop bureaucratic warts as the district goes larger. In time, nobody can get anything done because the command structure becomes increasingly obtuse and byzantine.

Posted by anony-mouse | November 14, 2007 7:30 PM

Jesus. I would really like to live life without a self-critical apparatus like many of you do. It would be so much easier.

Listen-- I am not being pedantic. I have asked again and again and again in this space what specific features of private education are going to enable it to educate an impoverished child in a single parent home living in an inner city in the shadow of crime and drugs. I have never gotten an answer. At all. I'm told over and over again that private school vouchers are the answer, but I am denied a) concrete evidence that they improve anything and b) an even basic description of the mechanism through which they would improve that education. No one has anything approaching a coherent answer. What I and others are told is "I don't know how it works, it just works"; and "anything is better than the status quo." Using precisely Megan and others' thinking, it is perfectly consistent for me to ask whether this inexplicable and yet supposedly tangible benefit will, in fact, scale upwardly by a staggeringly large amount. Since the proposed benefits and the way they are extracted is so vague, amorphous and ill-defined, I see no reason why the question of scale isn't appropriate.

And, in fact, it is uniquely appropriate because the selection bias is so enormous in private and public education. You have nothing approaching consistent populations between public and private school, which makes questions of scale even more important. As I have pointed out ad infinitum, the children who suffer from disadvantages such as poverty, special needs and behavioral problems (who don't go to private school, if you really need telling) are not going to shed those problems when they walk through the private school door. Why do you think private schools are going to be successful in educating them? The answer: you don't. You have, in fact, no data whatsoever to suggest that they are, and more damningly, you can't come up with a cogent description of what advantage they would have in doing so. And yet I am expected to endorse those articles of faith when discussing an issue that concerns massive public expenditure and the future of our children.

And, finally, precisely the same justification you offer for vouchers can be used with the NIH model. Of course we don't mean to scale it completely, but to use its model in many smaller scale systems spread broadly.

But then, I suppose any argument is only valid to the degree to which it adheres perfectly to your ideological bias.

Posted by Freddie | November 14, 2007 7:46 PM

I have asked again and again and again in this space what specific features of private education are going to enable it to educate an impoverished child in a single parent home living in an inner city in the shadow of crime and drugs.

Short answer: The ability to expel troublemakers. That is, the ability to leave some children behind.

The principal mechanism through which education will be improved is the pushing of that selection bias you mention into the lower economic levels: pull kids with responsible parents away from those who are likely to drag them down.

On top of that, there may be some gains in flexiblity; if someone knows a good way to reach special needs kids on the price of a voucher, for instance, parents of those kids may flock to schools using that method, while a school board may tie the whole thing up in process and budget fights for years. This might or might not actually happen.

PS: the future of "our" children? Probably not the future of the children of anyone here. Not mine, certainly.

Posted by Rob Lyman | November 14, 2007 7:56 PM

Short answer: The ability to expel troublemakers. That is, the ability to leave some children behind.

Plus, the ability to enroll your kids in schools with strict rules and discipline that the public school system cannot or will not (and arguably SHOULD not) offer.

Posted by Dan | November 14, 2007 8:25 PM

Not so fast. That pilot program has a huge administrative staff whose sole incentive is to ensure that it is meticulously carried out. In the real world, that curriculum will be put into place by an administrator whose priority list is crowded with everything from mollifying the latest lunatic on the school board, to ensuring that she gets out of town for a three day weekend with her new boyfriend who she really thinks may be The One.

That pilot program is staffed with a narrow band of extremely highly qualified teachers, sifted from the best the environment has to offer. In the real world, whoever happens to be standing in front of the classroom come September 5th has to do it, even if they flunked Remedial Math four times and only got this job because the school board needed a body.

But but but, I thought private schools were supposed to be better because they are private. Is this finally an admission that this is really not the case?

I cheerfully note for the nth time btw, that note study has ever been shown that private schools are better because they are private, or indeed that they are better at all. (All of those clueless people from the last go-round who want to parade their statistical ignorance and claim that 'statistically significant' means 'significant' . . . by all means. I'll feel free to sneer and deride again.)

On top of that, there may be some gains in flexiblity; if someone knows a good way to reach special needs kids on the price of a voucher, for instance, parents of those kids may flock to schools using that method, while a school board may tie the whole thing up in process and budget fights for years. This might or might not actually happen.

PS: the future of "our" children? Probably not the future of the children of anyone here. Not mine, certainly.

Posted by Rob Lyman | November 14, 2007 7:56 PM

Which, of course, we already knew. But I'm glad that some of the pro-voucher people are at least finally admitting it has nothing to do with better teaching and everything to do with selection bias. Oddly enough, this is something the public schools have done in the past, and something I'm 100% for right now. Get rid of the trouble makers. They don't want to be there, and we don't need them.

Finally, I note that the results of last Tuesday show that the American people still don't like vouchers, and in fact the only way the advocates get them is if they are shoved down our throats in a very undemocratic way. Don't voucher advocates taking any lessons away from this fact? Other of course, than the 'fact' that the almighty Teacher's Union has brainwashed the American public, that is.

Posted by ScentOfViolets | November 14, 2007 8:27 PM

To put a kinder face on Rob's comment: vouchers would open up (some form of) private education to children whose parents cannot afford it on their own but nonetheless exhibit the values necessary for success in school.

There definitely will be gains in flexibility. Whether those will actually accomplish anything or not is up for grabs.

Before you worry about those "left behind", consider that they are astonishingly unlikely to go anywhere in life no matter what education is offered - as you say, Freddie, it's not as though private education will be likely to assist them. Why not let the last few who cannot afford otherwise escape it? Why let them be dragged down with the sinking ship?

Posted by Devilbunny | November 14, 2007 8:38 PM

Freddie, the voucher argument still isn't a scaling issue. No matter how passionately you want to make semantic tie between 'scaling' an bureaucracy and 'scaling' the number of private schools, they simply are apples to oranges. I know you deperately want to make this another one of your anti-voucher threadjacks, so.... show me any study that proves we'll improve the DC school system with any method. Until you do, I'm willing to accept ANY change that results in better accountability, parental control and direct feedback loops: since those things are universally shown to help.

SoV, since you apparently are fine not educating every child, and explicitly denying education to "the troublemakers", I don't see how you can make any ethical or moral arguments about vouchers. I am honestly appalled at your lack of compassion or responsibility.

Posted by Skullberg | November 14, 2007 9:18 PM

Freddie:
Some of the questions you claim have not been answered have in fact been answered. For instance, you and many others keep saying that special-needs kids will be left out in the cold under a voucher system because private schools will skim off the smart, well-behaved students and stick the public schools with all the difficult cases. I have already pointed out twice that private schools are already taking many of the most difficult cases. Here's some of what I wrote 12 days ago in this comment (follow the link to see the rest):

"As of 2005, more than 88,000 disabled students nationwide were educated in private settings at taxpayer expense, an increase of 34 percent over a decade, according to the National School Boards Association.
"Often school districts acknowledge that they cannot provide an adequate education, and willingly pay for private tuition."
(Often, but not always: sometimes the parents have to sue.)
In other words, severely learning-disabled students are already eligible for vouchers in most states, because the private schools can and do handle cases the public schools can't or won't.

[end quoted section]

In other words, vouchers are already considered good enough for special-needs kids with problems so severe that the public schools can't handle them. If private schools can handle those kids, a fortiori they can handle the ones with lesser problems or none at all.

Posted by Dr. Weevil | November 14, 2007 9:30 PM

Freddie says I have asked again and again and again in this space what specific features of private education are going to enable it to educate an impoverished child in a single parent home living in an inner city in the shadow of crime and drugs

The profit motive.

I cheerfully note for the nth time btw, that note study has ever been shown that private schools are better because they are private, or indeed that they are better at all.

I cheerfully note for the nth time that you are looking at the attempted measurement of a partial equilibrium effect at best, and hence not addressing the voucher versus public administration debate in any significant manner.

All of those clueless people from the last go-round who want to parade their statistical ignorance and claim that 'statistically significant' means 'significant' . . .

So then are you admitting then that the studies show statistically significant differences in some measures?

Posted by Bill Abbott | November 14, 2007 9:39 PM

Dwight -- I don't recall that you ever made a distinction between "statistical significance" and "significance" before. But maybe you did. I suppose what you're trying to say is that the numerous studies showing that Catholic schools do a better job of educating urban minorities are merely statistically significant, but the percentage difference is so small that it's not really significant in real-world terms. Is that it? If so, can you demonstrate that this is the case?

Freddie -- there are plenty of people who have offered quite a few different theories as to why private schools could (not necessarily would in all cases, but could) have an advantage. You may not agree with the theories, but it's odd to suggest that nothing has even been mentioned before.* Here are several possibilities:

1. Potential for stricter discipline.

2. Potential to hire better teachers and/or not to get stuck with bad teachers who have tenure thanks to a union contract. (See, e.g., http://www.tntp.org/files/TNTPPressRelease.pdf ).

3. Potential to be more responsive to parents (look back at the Washington Post article that inspired Megan to start writing about vouchers, i.e., at how much trouble that guy had in getting anyone in the public school system to even talk to him).

4. Because of 3, potential to inspire more parental involvement, both on an individual and community level. This one factor that Coleman and Hoffer identified in their famous book on why urban Catholic schools were superior, at least as of the 1980s, along several dimensions.

5. Potential to be able to use a more effective and rigorous curriculum (e.g., Singapore math) without being quashed by state bureaucrats who haven't approved the purchase of that curriculum, or by the many interest groups that get involved in the curriculum selection process.

6. Potential to impose increased academic demands more generally. In looking at the famous HSB data, Coleman and Hoffer (pp. 44-45) found that the "most striking difference between public and private school curricula is the much greater likelihood of academic program placement in the private schools." (Nearly 50% of Catholic students were in specialized "academic" programs, while only 3.3% of public school students were in such programs; most public school students were in "comprehensive" or more general programs.)

7. Potential to drastically reduce dropout rates. See, e.g., Sander 2001, p. 23. Similarly, Coleman and Hoffer found that the black dropout rate in public high schools was 17.2%, while the black dropout rate in Catholic high schools was a mere 4.6%. (p. 127). They also point out that this is not what you would expect, given that Catholic high schools also had substantially higher achievement gains for black students. Normally, you would think that a more academically demanding high school could easily have a higher dropout rate, not a rate that is nearly 4 times lower. (Coleman and Hoffer have a very lengthy passage in which they try to test for selection effects here.)

* * *

Incidentally, where are y'all getting the impression that private schools are superior only because of selection effects? In my experience, there are often some perverse selection effects going on with religious private schools. I've known troublemakers and delinquents who went to a private Christian school; apparently their parents thought that if they signed their kid up for a religious school, that might be the one thing that could straighten him out. And sometimes the religious schools think of accepting a bad kid as a way of reaching a lost soul. Less anecdotally, Coleman and Hoffer found that "students who transfer from public sector elementary schools to the private sector, particularly to the Catholic sector, contain a high number who were doing poorly, scholastically or behaviorally, in public elementary school." (p. 112).

Also, Coleman and Hoffer found that "it is not true that the private sector schools get rid of their failures by expelling them." (p. 85). Their reasoning was that if private schools make kids meet a higher test score threshold to keep from getting expelled, then kids expelled from private schools should by definition have higher average test scores than kids expelled from public schools. But the exact opposite was the case.

Of course, there are surely a lot of kids are motivated to seek academic success and who end up in a private school because of it. But I don't know that there's any evidence about how many of those kids there are, compared to the number of families who try out private education precisely because their kid is flunking out of the public school or possibly falling into drugs or delinquent behavior. None of the voucher opponents have even tried to provide any such evidence.

Posted by Stuart Buck | November 14, 2007 10:36 PM

Note, by the way, that I routinely use the word "potential" in that list. That's because none of the factors that I list are necessary and universal truths. They could be true at a given place and time, but not true in another place and time (depending on a wide array of circumstances). The point is not to demonstrate conclusively that all private schools are in fact superior in those ways. The point is to answer Freddie, who seems to think that no one has ever even offered an idea as to what private schools might conceivably be able to do differently.

Posted by Stuart Buck | November 14, 2007 10:47 PM

'It doesn't scale' is as obvious in the public sector as it is in the private. But there is a strong incentive to those pulling the levers of public power to listen to the private sales person and their lobbyist and ignore the public employee. Campaign contributions and the presumption that the public sector employee is incompetent/deluded are two of many contributing factors. The scathing comments at MR regarding Cyntia Kenyon are manifestations of the latter.

And while we are talking about scale, the truly collossal public sector trainwrecks are executed by the public sector while some elected official under the influence looks the other way. Those billion dollar boondoggles, be they software systems in California or security systems in Iraq, are private sector fiascoes that we pay for. Those of us who have walked both side of the fence have a healthy appreciation of the competency and honesty of most public sector employees. But these days fewer and fewer public sector employees are responsible for much beyond routine paperwork.

NIH, our school systems, and our military are exceptions. If NIH were turned into a few contract managers funneling 28bn to contractors I think the results would be disastrous. I fear the same with schools. And the Blackwaterization of our military is truly scary. A competent public sector is both possible and necessary for a democracy. A government that is nothing more than a contract manager is a very poor protector of the public trust.

Posted by Nat | November 14, 2007 11:22 PM

This post and argument would be more complete if they also mentioned the converse: there are many programs which cannot be scaled down. There is no point building a subway system with only one line. There is no point launching a program to eradicate gangs at just one school. There is no point launching a television ad campaign if you only have enough money to run the ad once. (Unless you can run it during the Superbowl.) There is no point building a better cell phone if you can't support it with a nationwide distribution network. There is no point invading Iraq if you aren't going to devote enough resources to policing and rebuilding it properly. And so on.

Many of our arguments here have to do with whether programs fit into one kind of problem, or the other.

Posted by brooksfoe | November 15, 2007 3:30 AM

Dr Weevil wrote: In other words, vouchers are already considered good enough for special-needs kids with problems so severe that the public schools can't handle them. If private schools can handle those kids, a fortiori they can handle the ones with lesser problems or none at all.

True, but my impression is that the payments to private schools for these special needs children are often on the order of $35,000-45,000 per child per year. A fortiori goes away in the real world unless you are talking comparable numbers

Bill Abbot
"the profit motive" - proving Freddie's point, you don't specify a mechanism by which private school swill be superior, you rely on magical thinking
"partial equilibrium" - what? how is a staightforward comparison of public vs. private school outcomes partial equilibrium? And since when does all research require a general equilibrium model?
"statistical significance" - well, if you make a sample size large enough, you can almost always get a significant F-statistic for even tiny differences between two populations. So while the result may be "statistically significant," in real terms of actual impact it is of quite minimal signifcance.

Posted by Gene | November 15, 2007 9:01 AM

Freddie's questions and the lack of answers for mechanism makes me think of South Park:

1. Implement School Vouchers
2. ????
3. Kids are Educated!

Posted by Disgusted Beyond Belief | November 15, 2007 10:40 AM

Here is a funny and related story. PETA letter to Homeland Security:

November 14, 2007

The Honorable Michael Chertoff
Secretary of Homeland Security
U.S. Department of Homeland Security
Washington, DC 20528

Dear Secretary Chertoff,

At a time when Americans fear for their safety, when they turn trusting eyes to you in the sincere belief that you will do all that is in your power to serve and protect our country and its people, why has Homeland Security chosen to betray that trust? Are you aware that the National Institute of Allergy and Infectious Diseases (NIAID) is using chimpanzees—our closest animal relatives—to develop antibodies to anthrax and other dangerous pathogens even though non-animal antibodies have already been approved by the Food and Drug Administration or are in late-stage clinical trials? This appears to be a stunning case of the left hand of government not knowing what the right hand is doing. And it means that tax dollars that should be used to provide real protection for Americans are being squandered.

PETA and its more than 1.8 million members and supporters certainly back efforts to develop treatments for the 15 pathogens targeted by NIAID, a number of which have been identified as potential bioterrorism threats under Project Bioshield. But torturing chimpanzees to accomplish this, when antibody treatments for most of the pathogens have already been developed, is wasteful, cruel, and astonishingly foolish. So far, antibody treatments for anthrax, smallpox, hepatitis A and B, rabies, botulism, West Nile, and lupus have already been developed without the use of chimpanzees. Two companies have been contracted to provide their anthrax and smallpox drugs to the Strategic National Stockpile, and at least three other companies have antibody-based anthrax drugs undergoing clinical trials.

Although PETA doesn't receive a dime from the government to protect our country from misuse of tax dollars and stupidity, we uncovered these redundant experiments through public records. I must assume that you didn't know, or you would have acted to stop it; now that you do know, I urge you to put an immediate end to this completely unnecessary abuse of chimpanzees.

If you won't act to protect chimpanzees, who may be caged for decades and endure repeated and excruciating bone marrow extractions, then please act to protect our precious resources from the vile greediness of experimenters who apparently have no idea or do not care that they're doing yesterday's science.

I look forward to hearing from you without delay.

Very truly yours,

Ingrid E. Newkirk
President and Founder

Posted by Hugo Pottisch | November 15, 2007 11:09 AM

I wouldn't say that selection bias is the only reason private schools might help kids under a voucher program, but it certainly is the one that scales up the most easily.

SOV, I wouldn't say it has "nothing" to do with better teaching--incentive pay and the threat of firing if your classes don't learn would probably be good motivators for some of the tenured deadwood--but again, addressing the scaling issue, it isn't obvious (to me) that we can manufacture huge numbers of teachers as good as those at the best private schools. So to me better teaching would be a smaller effect that you would expect from pure extrapolation, but I wouldn't place it at zero.

And as long as Hugo is barreling off on a tangent, I'd like to point out that "renewable energy" scales very poorly. A grid driven mostly by wind and solar is not a stable grid.

Posted by Rob Lyman | November 15, 2007 12:22 PM

"the profit motive" - proving Freddie's point, you don't specify a mechanism by which private school swill be superior, you rely on magical thinking

I will admit to being a bit flippant here. This demand for a "specific mechanism" is a common and absurd expectation that demonstrates the asker's lack of understanding about the nature of institutional analysis in specific, and markets in general.

It is similar to asking someone 40 years ago who might be defending a private market in computer technology to predict specific future technological innovations.

Though in other threads I have addressed hypothetical superior innovations that a market system is likely to produce, as others have done here, the real issue is the comparison between market mechanisms and top-down bureaucratic mechanisms at generating and implementing innovation. If you know one institutional structure is superior in this regard, the ability to predict specific innovations is beside the point.

"partial equilibrium" - what? how is a staightforward comparison of public vs. private school outcomes partial equilibrium?

There is no consideration of the gains to educational performance which widespread voucherization would produce through better markets for input factors (teachers, curricula, etc...), regulation (private, of course), evaluation; nor of the potential increase in returns to education.

And since when does all research require a general equilibrium model?

It doesn't, nor did I say it did. It is merely certain claims which do.

"statistical significance" - well, if you make a sample size large enough, you can almost always get a significant F-statistic for even tiny differences between two populations. So while the result may be "statistically significant," in real terms of actual impact it is of quite minimal signifcance.

Yes, I understand what statistical significance means. I was asking for an admission that statistically significant results were obtained.
But to what you mention, it is quite possible that even small differences can compound over time to give large differences.

Posted by Bill Abbott | November 15, 2007 12:58 PM

If NIH were turned into a few contract managers funneling 28bn to contractors I think the results would be disastrous. I fear the same with schools. And the Blackwaterization of our military is truly scary. A competent public sector is both possible and necessary for a democracy. A government that is nothing more than a contract manager is a very poor protector of the public trust.

I certainly agree that the practice of 'privatization' that we're seeing is largely just a way for politicians to funnel money to their top contributors/cronies.

BUT, vouchers are inherently different because we are not talking about multi-billion dollar contracts to for-profit or non-profit schools to take on whole populations of children. We're talking about entrusting parents to use voucher money to pick schools for themselves. So, instead of having one monolithic system replaced with another, you are instituting a wide-open market for education.

Frankly, I don't understand what objection there is to such an idea, given that the post-secondary system in this country runs along the same lines. Of course we don't demand that the government distribute vouchers to pay tuition to colleges, but we also haven't made college compulsory either.

Posted by Christina | November 15, 2007 1:46 PM

Nat wrote: And the Blackwaterization of our military is truly scary.

Why? "Blackwater" may make for a convenient media curse word whenever a report is filed on an incident that went other than according to plan, but for this to be "scary", one needs to make a credible argument that the use of US military elistees for these positions and US-owned military equipment would have made a substantial difference in the outcome.

Posted by anony-mouse | November 15, 2007 1:48 PM

Everything always comes back to vouchers, heh.

Why would vouchers be effective?
a) because faced with the threat of the parents moving their children to other schools, the administration will have significant incentive to focus on educating the children, instead of other stuff
b) Because troublemakers can be left behind. Not underperformers - troublemakers who reduce the educational effectiveness of every student in the class.
c) Because the effort required to move the child into a new school is evidence that the parent is more committed, and will be more involved in the child's education than at a public school that doesn't care about the kids.
d) Because a multitude of private schools will use a multitude of teaching methods, and they will share what works and what doesn't work, and change (or go out of business) (see a). I don't have to know what those changes are to know that this will happen. Right now, public schools have no meaningful incentive to adapt to new teaching methods, because there is no meaningful consequence to failure.

So in summary:
1) Consequence of failure
2) Removal of destructive children
3) Enable parental involvement
4) Experimentation & adaptive teaching methods

Yes, I know the destructive children are still children, but they shouldn't have the ability to ruin the education of the other kids in their classes, and right now, they do. They need to be separated out, and handled as special cases.

Posted by jb | November 15, 2007 2:21 PM

Simply awful. The normally reliable MM stumbled badly in this posting. Vouchers are not about a large unproven public works project.

It's about giving parents access to the resources the public would otherwise spend directly, so they can participate in a private marketplace for educational service. A marketplace now woefully underserved by the private sector, owing to the government's huge tax on consumers (no rebate if you send your child to private school), distortion of labor markets for teaching services (including labor laws through which private teachers might organize), and failure to subsidize wealth constrained, lower income consumers to purchase educational services.

Unshackle the marketplace and get out of the way. Public education makes as much sense as public power and public food. Both equally "necessities", but which no rational observer would put in the hands of collectivized production.

Posted by apetra | November 15, 2007 2:32 PM

Why? "Blackwater" may make for a convenient media curse word whenever a report is filed on an incident that went other than according to plan, but for this to be "scary", one needs to make a credible argument that the use of US military elistees for these positions and US-owned military equipment would have made a substantial difference in the outcome.

Because they're mercenaries. Do you recall what Machiavelli said about mercenaries? Do you trust people who serve you only as much as you can pay them to perform vital military functions?

Yeah, I know I'm arguing from authority, so don't bother pointing it out. I just think Machiavelli happens to be pretty good authority on this matter. We're still assigning "The Prince" in Poli Sci 101.

Posted by Mercenary | November 15, 2007 4:04 PM

Blackwater employees are not mercenaries. Historically, mercenaries are soldiers who will gladly work for either side if the price is right. Do you think even a single one of the Americans working for Blackwater would work for the other side for any amount of money? I don't, and I would fear for your safety if you ever told a Blackwater employee that to his face.

Posted by Dr. Weevil | November 15, 2007 4:30 PM

Mercenary: First off, you didn't make an argument, much less a credible one, that there would have been a difference in any of the controversial encounters if US enlistees had been there.

Second, historically, mercenaries (and for that matter ordinary conscripts) fought for booty, liquor, and women, not for salaries; the willingness to fight for "either side" came from the desire to share in the spoils of victory. There are no spoils of that sort available in Iraq, regardless of which side you're on.

Posted by Rob Lyman | November 15, 2007 4:37 PM

Yeah, I know I'm arguing from authority, so don't bother pointing it out. I just think Machiavelli happens to be pretty good authority on this matter.

I happen to consider myself an authority on why Machiavelli is a crank. No, I don't need evidence, just my insouciant certainty that my opinion is Right And Proper without the burden of making proof. In fact, I don't even need to hint as to why my claims are relevant to the context of the modern theater. They just are.

Hey, you're right. This IS fun.

Posted by anony-mouse | November 15, 2007 4:45 PM

Historically, mercenaries are soldiers who will gladly work for either side if the price is right. Do you think even a single one of the Americans working for Blackwater would work for the other side for any amount of money?

Yes.

I don't, and I would fear for your safety if you ever told a Blackwater employee that to his face.

I already have. In fact, I make it a point to do so whenever I run into a private military contractor. I say, "Hey, mercenary, how's the contract-killing business going, bitch?"

And, somehow, some way, I'm still alive.

Rob Lyman:

First off, you didn't make an argument, much less a credible one, that there would have been a difference in any of the controversial encounters if US enlistees had been there.

I didn't make that argument because it wasn't my point. I wasn't saying that U.S. conscripts wouldn't have fired on those 17 people, I was saying that mercenaries are inherently untrustworthy because they kill people for money.

Second, historically, mercenaries (and for that matter ordinary conscripts) fought for booty, liquor, and women, not for salaries

A distinction without a difference, since: (1) "Booty" not being available to most modern professions save undersea salvager, "salary" substitutes very effectively; (2) "Salary" purchases women; and (3) also liquor.

There are no spoils of that sort available in Iraq, regardless of which side you're on.

The multi-million dollar contracts awarded to PMCs (private military companies) are the spoils. Only in the historical sense is the definition of "spoils" limited to to spanish doubloons, jugs of whiskey, comely lasses and the like.

Posted by Mercenary | November 15, 2007 4:56 PM

I happen to consider myself an authority on why Machiavelli is a crank.

Tell you what, buddy. If your treatise on why Machiavelli is a crank is considered a classic and read worldwide hundreds of years after your demise, then maybe I'll waste my time defending one of the geniuses of political thought against a scrub like you.

But that's not likely, and here's why:

(1) You haven't written any treatise;

(2) You never will; and

(3) I'll already be dead by that point, and won't care.

Posted by Mercenary | November 15, 2007 5:00 PM

I didn't make that argument because it wasn't my point.

Then perhaps next time, to minimize confusion, you will avoid quoting someone calling for Argument A, beginning your reply with "Because," and then denying that you were making Argument A.

mercenaries are inherently untrustworthy because they kill people for money.

So do US enlistees and urban police officers, who merely receive a smaller salary.

Further, salary is not a substitute for booty. Booty is available only to winners, salary is available even when you lose and retreat ignominously. This is the root of the willingness of the mercenary to fight for either side: he only gets paid meaningfully if he wins, so switching sides when the battle isn't going well is attractive. This creates a feedback loop much like a bank run, which any prudent general should fear.

Blackwater and other PMCs have many problems, but you haven't managed to identify one of them yet.

Posted by Rob Lyman | November 15, 2007 5:09 PM

Oh, and there are no US conscripts.

Posted by Rob Lyman | November 15, 2007 5:11 PM

So do US enlistees and urban police officers, who merely receive a smaller salary.

No. Police officers are paid to serve the public trust and uphold the law. Enlistees are paid to serve and protect the nation. They are paid to perform duties, not to kill people. Killing is incidental.

PMCs are also paid to perform duties, not necessarily to kill people, but the performance of these duties is not based upon loyalty to anything beyond the PMCs' own desire for profit. Hence, they are mercenaries.

This is the root of the willingness of the mercenary to fight for either side: he only gets paid meaningfully if he wins, so switching sides when the battle isn't going well is attractive.

The only change in the incentives for mercenaries of yore and the PMCs of today is from "Who gets paid" to "Who gets paid more." The wish of mercenaries is to be paid, and if both sides are paying, the wish then becomes to be paid the maximum amount possible. Market economics, no?

You're right about the U.S. not having conscripts, though. My bad.

Posted by Mercenary | November 15, 2007 5:23 PM

OK, we're making progress here...

I doubt your account of the loyalties of Blackwater contractors; in principle, soldiers could sell out their comrades to terrorists, and cops could be on the take. I know of no evidence that Blackwater employees are any more likely to do this than those who wear government-issued uniforms, and one might think their high salaries would tend to be tougher for terrorists to beat.

Furthermore, if you want to bang on about Machiavelli, you need to place his remarks in the context in which he was writing. In his day, mercenaries, and indeed putatively "loyal" domestic soldiers, got a miserly wage, lousy rations, and infections that killed more people than enemy steel. The principle reason to be a mercenary was for your share of the towns that you razed. This is different from the "mercenaries" of today, making your historical references of about as relevant as Alexander Hamilton's pre-FDIC thoughts on proper banking reserves.

Posted by Rob Lyman | November 15, 2007 5:45 PM

Merecenary wrote: I already have. In fact, I make it a point to do so whenever I run into a private military contractor. I say, "Hey, mercenary, how's the contract-killing business going, bitch?" And, somehow, some way, I'm still alive.

A quick statistitcal breakdown suggests that the reason you are still alive is because 60% of them simply don't care what you think, 15% are plainly offended but also more honorable than you, 10% think you're drunk or high at the time, and the other 15% have reduced you to a 5:53 mile. Eventually, however, those shoes will wear out, and then what will you do?

Posted by anony-mouse | November 15, 2007 7:47 PM

Yes, I understand what statistical significance means. I was asking for an admission that statistically significant results were obtained. But to what you mention, it is quite possible that even small differences can compound over time to give large differences.

I don't see that you have any knowledge of what 'statistical significance' means. In fact, you don't seem to have any idea of what a p-value is, or the power of a test, etc.

But you can prove me wrong; simply discuss what it means to be 'statistically significant' at p=0.05, but not at p=0.049. For bonus points, you can explain why the value in this instance was chosen to be 0.05, and why a parameter can be significant even though it is not statistically significant 0.05, and why a parameter would not be significant even though it is significant at the p=0.025 level. You being a scientist and all.

You could also explain why you never responded to this:

I disagree with your view that what you see as the status quo beliefs in the social and political arenas deserves a privileged position regarding burden of proof, as you suggest is normal in the natural sciences.

I really have no idea where you're coming up with this. If it's from something I said, please quote it.

or this:

Obviously we are talking past each other. You seem more concerned with personal persuasion, whereas I was more focused on the pursuit of "truth" or better understanding.

I've asked you this several times now - where did you get this from what I wrote? Specifically. I think I've quite explicitly stated that these proof standards evolved as the best way to discover the truth.

If you can't find those quotes that back up your assertions, I quite understand. And the burden of proof is still on you ;-)

Posted by ScentOfViolets | November 15, 2007 10:24 PM

I don't see that you have any knowledge of what 'statistical significance' means. In fact, you don't seem to have any idea of what a p-value is, or the power of a test, etc.

That is no doubt because I made no attempt to display such knowledge. I don't see how asking someone for an admission that they accept certain results as statistically significant requires that I demonstrate my knowledge.

You could also explain why you never responded to this:

Because I'm fairly certain that my point has been made to those able to understand it.

Posted by Bill | November 16, 2007 2:37 AM

I really don't believe you, Bill. In either case. Let's turn it around - let's say you really did know basic statistics, that I really did say something like you said I did. Does anyone doubt that Bill would waste any time posting voluminous rebuttals? Crickets.

No, you've just been rather silly in defense of an ideology. Again.

And no, for the record Bill, if you make an assertion, the burden of proof is on you to prove it, not on anyone to disprove it.

I might add yet again, that the recent crash-and-burn of vouchers in Utah (In Utah!) is ample evidence that libertarians simply aren't being persuasive enough. And the general public likes these tactics (prove me wrong) no more than I.

So why don't libertarians get another set of arguments, or respond to their critics in ways that the voters will appreciate?

Posted by ScentOfViolets | November 16, 2007 8:58 AM

Dwight -- this seems to be entirely a propos of nothing. Why are you dragging in an irrelevant old debate about the statistical significance of something or other?

Posted by Stuart Buck | November 16, 2007 9:20 AM

Sigh.

All of those clueless people from the last go-round who want to parade their statistical ignorance and claim that 'statistically significant' means 'significant' . . .

So then are you admitting then that the studies show statistically significant differences in some measures?

Posted by Bill Abbott | November 14, 2007 9:39 PM

Check the date. And no, the difference between 'significant' and 'statistically significant' is ...significant. As you well know, and as you will never admit that you argued otherwise, many posts to the contrary.

Posted by ScentOfViolets | November 16, 2007 10:08 AM

Dwight -- I'm talking about the last several posts between you and Bill, where you seem to be trying to rehash some obscure and meaningless point about something Bill said elsewhere (not clear where) about something (undefined). Do you have a point to make that's relevant to the discussion here?

Posted by Stuart Buck | November 16, 2007 11:04 AM

When someone uses the term historically, in any context, it's usually a dead giveaway to the fact they're blowing smoke.

The term mercenary doesn't necessarily demand a soldier-of-fortune be willing to fight for either side; instead it means the person's primary motivation is financial. In fact, it wasn't uncommon for officers in the Royal Navy and Army, several centuries back, to sell their services to foreign countries. For example, Robert Cochrane (the real-life inspiration for Patrick O'Brian's tales) had a career as a mercenary naval admiral for Chile, Brazil, Peru and Greece.

Essentially, the canard a mercenary has to fight for the side paying the most isn't based on reality. The reality is a mercenary can and pick and choose his wars and abandon them when they wish.

Posted by Jadegold | November 16, 2007 12:28 PM

Stuart, do you even pay attention to what you post? You just asked about 'statistical significance'. It's right there, it won't go away, and saying you didn't ask, that you asked about something else instead makes you look . . . wierd.

The other point, Stuart, since you're too dim to grasp the implications, is that if I ask questions, legitimate questions, and am ignored, well, I don't feel exactly obligated to answer questions coming from the other direction.

Get it now?

And yes, that exchange is recent as well.

Any other points you need explained?

Posted by ScentOfViolets | November 16, 2007 12:58 PM

Well, I did ask this question of you, Dwight, but you have never tried to answer it at all:

I suppose what you're trying to say is that the numerous studies showing that Catholic schools do a better job of educating urban minorities are merely statistically significant, but the percentage difference is so small that it's not really significant in real-world terms. Is that it? If so, can you demonstrate that this is the case?

What you need to do: Spell out an actual argument (with citations) about what you think is lacking in the significance (statistical or otherwise) of the many studies on vouchers and/or Catholic schools.

Posted by Stuart Buck | November 16, 2007 1:11 PM

Stuart, I've pointed out that the studies you've posted don't say what you think they do, at least the ones that I've looked at. Your responses have been, let us say, less than adequate, in fact, ignorant, disrespectful, and insulting.

So why should I bother to look at anything else you have? Why should I put myself through any sort of grief? Particularly since apparently you don't know basic experimental methocology, at least on the statistical side; anything I say in that department you literally could not comprehend. At least, that is what I've observed so far.

Now, you want to make nice, apologize, admit you were way off base, that my points were valid, that those studies I looked at were in fact not supportive of your arguments, I might reconsider. But otherwise, frankly, no thanks. It's just not worth it.

Posted by ScentOfViolets | November 16, 2007 3:46 PM

Dwight -- so you refuse to discuss any actual studies. Hmm. Well, your point that statistical significance is different from real-world significance is certainly true enough, although since you refuse to apply that elementary concept to any actual evidence, it really doesn't get you anywhere. And I can't blame you for declining to address educational studies after what happened when you tried to dismantle the Cleveland voucher study (it must be rather embarrassing for a math instructor not to understand that a percentage change in raw scores is quite different from a change in percentile rank).

Posted by Stuart Buck | November 16, 2007 4:26 PM

Sigh. Get help Stuart. And get some education. Look at it this way: if you're really interested in learning what these types of reports say, and what the statistical methodologies are behind them, you're going to have to learn it sooner or later. Why not just take a course? It's not hard stuff, honestly, and you'll have a better knowledge base to argue from.

And, segueing seemlessly to the related point, I will note again that it is people like Stuart who were probably responsible for the defeat of the voucher referendum in Utah. Y'know, when I say that the arguments are unpersuasive, and that the manners and attitudes of those advancing the arguments are dissuasive, y'all might actually consider the notion that I'm not being a statist tool and that I am actually saying that the arguments are unpersuasive and the people advancing them need to work on their demeanors. Nothing crafty or underhanded or sneaky or manipulative about that observation at all. You shouldn't need the results of yet another failed attempt to smack you in the face.

Posted by ScentOfViolets | November 16, 2007 6:25 PM

Freddie - Re: "I have asked again and again and again in this space what specific features of private education are going to enable it to educate an impoverished child in a single parent home living in an inner city in the shadow of crime and drugs."

The specific feature that produces better results over time is competition. The child, and his or her parent, given the choice between different schools can pick a better one.

Maybe neither some specific "impoverished child in a single parent home..." nor his specific single parent care enough to actively select for better schools, but if he and his parent do not than other parents still will. Unfortunately forcing poor schools to improve (or weeding them out and replacing them with better ones) takes time, so if this specific student (and the parent) doesn't work hard for a better education he might not get one, but at least future students in his position will as competitive pressures improve the whole system.

RE: "Using precisely Megan and others' thinking, it is perfectly consistent for me to ask whether this inexplicable and yet supposedly tangible benefit will, in fact, scale upwardly by a staggeringly large amount."

No specific school, or organization, or education scheme has to scale upwards much at all. Instead you can have competition between schools, schemes, and organizations.

Posted by Tim Fowler | November 16, 2007 6:28 PM

That is not a specific feature, and I suspect you know it. That is the mechanism by which specific features come into being.

Your reply is as nonsensical as if I'd asked what makes an iPhone a better phone and you replied 'competition'.

Sorry, but if your goal is to sell me an iPhone, well, you just lost a sale.

Posted by ScentOfViolets | November 16, 2007 6:53 PM

Tim Fowler:
I think it's important to note that competition in private schools is not like competition in the 100yd dash, with everyone trying to do exactly the same thing in the same way with slightly different results. Private schools already succeed by providing variety, and they would provide even more variety if they were better financed. In the public sector there are constant arguments over things like uniforms vs. dress codes vs. anything goes, phonics vs. whole word, traditional math vs. trendy math, athletics vs. academics, candy machines & soda machines vs. healthy lunches, and so on. In the private sector, parents get to pick which of these they care about, along with other choices not generally available in the public sector: coed vs. single-sex, heavily religious vs. lightly religious vs. nonreligious, strict discipline vs. get-in-touch-with-your-feelings chumminess. Some kids respond very well to uniforms and rules and sink-or-swim academic competition, some respond well to other things. In the private sector, parents can try to match the child to the school. It would be an invasion of privacy, but I could give numerous specific examples of why kids I have personally taught changed from public to private, from private to public (#1 reason: didn't want to work so hard), and from one private school to another and sometimes back again. Some families send their kids to different schools because the kids are so different.

As with the analogy I gave a few weeks ago (to deafening silence from the anti-voucher side), it's like eating in the college cafeteria vs. eating off-campus. We know for a fact that off-campus meals are generally better than college cafeteria meals because students who have already paid wholesale for meals in the cafeteria will (whenever they can afford to) pay again to eat off campus. Anyone who argued that cafeteria meals were 'just as good' would be an idiot. Similarly, the fact that so many pay so much for private schools proves that they are in fact better than public schools in the eyes of the parents.

Posted by Dr. Weevil | November 16, 2007 7:51 PM

Just to clarify two points:

1. The end of my 1st paragraph is ambiguous. I meant cases where parents send e.g. their brainy unathletic child to one school, and their athletic unacademic child to a different school.

2. At the end of the 2nd paragraph, I should have written "in the eyes of the parents, who aren't perfect but are better able to judge what their children need than anyone else".

Posted by Dr. Weevil | November 16, 2007 8:14 PM

Dwight, I don't mind your desperate lashing out; it's just a sign that you are indeed embarrassed for having confused percentile ranks with raw percentages. (What you said about the Cleveland voucher study is exactly like thinking that if a 6-foot-tall man is at the 50th percentile, a 7-foot-tall man can't be in the 99th percentile, because he's only 16.7% taller.)

Posted by Stuart Buck | November 16, 2007 9:52 PM

Do you have the faintest idea that you are making my point for me? The same one I made h