Editore"s Note
Tilting at Windmills

Email Newsletter icon, E-mail Newsletter icon, Email List icon, E-mail List icon Sign up for Free News & Updates

January 28, 2006
By: Kevin Drum

PUBLIC vs. PRIVATE....Do private schools do a better job of educating our kids than public schools? Lots of people think so. But a new, large-scale statistical analysis of the 2003 NAEP test results suggests that when you control for things like income, race, home environment, and so forth, the performance of private schools actually turns out to be worse or about the same as that of public schools, not better.

The study analyzed only the math portion of the NAEP test, and the results from the 4th grade test are shown below. The red line shows the average public school score, and as you can see from the black bars on the graph, the raw scores for most types of private schools are higher than the public school average. However, much of this difference is due to the fact that private schools attract better kids in the first place, not because the schools themselves are better.

So what would happen if both types of schools had similar student bodies? Those results are shown for private schools in the gray bars in the graph, where test scores are controlled for demographics, and they're considerably lower than the public school average. In other words, if you took two similar kids and sent one to a public school and one to a private school, the kid in the private school would probably do a little worse than his public school twin. (Note that a difference of 10 points is roughly equal to one grade level.)

The 8th grade results are better, with most private schools scoring about the same as public schools. The only exception is the conservative Christian schools, which continue to score considerably lower than public schools although the sample size is small enough that the results aren't conclusive.

I don't imagine that one study will change any minds, but the size and sophistication of this one should at least give us pause. The full report is here. The New York Times summarizes the results here.

Kevin Drum 6:49 PM Permalink | Trackbacks | Comments (118)

Bookmark and Share
 
Comments

Not surprising. But you're right, these data won't change many minds. Particularly those of people who believe private schools are better. Now we know why.

Posted by: Joel on January 28, 2006 at 7:09 PM | PERMALINK

I'm not shocked. No kids, but watched my friends go through the public vs. private. Seemed to be more about the parent's insecurities than the kid's actual needs, or the quality of the school.

We need to get rid of the voodoo surrounding education.

Posted by: NotThatMo on January 28, 2006 at 7:11 PM | PERMALINK

One thing that does get left out of this is that there is also an opportunity effect that won't show up when you control for these variables. My public high school, for example, didn't have AP Physics or Biology, or BC Calculus, which are often standard features for private high schools. The large numbers of affluent students with educated parents at private schools provide educational opportunities which may just not be available at the public school. That said, the study does point out that public schools do a very good job of educating their students and are very much worth supporting.

Posted by: Don Hosek on January 28, 2006 at 7:17 PM | PERMALINK

As a proud public school alumnus now attending a selective private college, let me tell you, it is possible to get an education just as good as at private school in the public school system, if not a better one. I many, many, of my classmates (including the daughter of a cabinet secretary) went to priate schools, and I don't feel like they had any huge advantage over me.

Posted by: Don Zeko on January 28, 2006 at 7:22 PM | PERMALINK

Of course, for many parents there isn't an opportunity to "control for income, race, home environment, and so forth". Your kids go to the public school in their district, or they go to private school. If the public school in your area is full of well-adjusted kids from affluent, upper middle-class families, then probably you won't get any benefit from private schools vs. the available public schools. On the other hand, say you live in a neighborhood where many of the students have less than supportive families and often have severe discipline problems- you get a huge boost by sending your kids to a local parochial school rather than the public school. So in a global sense, private schools may be no better, or sometimes worse. But at the individual level, they can be a huge improvement.

Posted by: MJ Memphis on January 28, 2006 at 7:27 PM | PERMALINK

Try a real comparison. Like our schools vs. Belgium. ( Consult John Sttossel's report on his ABC special for the ugly truth ). It is unbelievable that anyone can really seriously defend American public schools. We are falling behind. The big difference between American schools and Belgium schools?

Belgian schools the money follows the student ( vouchers ) instead of going to the school district.

Belgian schools actually can fire incompetent teachers, instead of like New York City Schools which have to pay the incompetent not to teach because its cheaper than firing them ).

Belgian schools actually have local control of where the funding goes.

In short Belgian schools are run like a capitalistic system when our American public school system is run like a communist system.

The only reason I think perverse statictics are manipulated to make it look like the public school system is doing extremely well is because the main culprit in making American public schools lousy ( the various teacher's unions of America ) is such a strong force in democratic politics.

Posted by: John Hansen on January 28, 2006 at 7:27 PM | PERMALINK

I'd be curious to know how a similar control for demographics would effect intra-country comparisons.

It's a commonly held belief that American students are getting their clocked cleaned by their foreign counterparts.

How much effect would demographics change that picture? That is, if say, Japan, has a much more homogenous population, and poor parents that are less common (just making that up), would controlling for that fact radically alter the comparison of schools between America and Japan?

What about a nation with a broad social welfare infrastructure? Presumably, being poor in that country is easier than being poor in America. Would those kids thus seem to be doing better than American kids because the comparison is different?

If so (and that's a big if), it would seem to indicate that the most pressing issue in America would equality of income, not education, per se.

Curious.

Posted by: teece on January 28, 2006 at 7:31 PM | PERMALINK

Under the W. Post rules, you must delete all comments that say "consult John Stossel." Nothing, but nothing, could be more insulting.

I'd say, depends on the private schools. Top end here in the northeast do provide a better education, if by education you mean ability to perform well on tests.

Posted by: david on January 28, 2006 at 7:34 PM | PERMALINK

Uh, no, John. Comparisons between Belgian schools and American schools are experiments without controls. Until you compare like with like, you are espousing religious, not scientific, beliefs. But that's consistent with your silly comparison between US public schools and communism. Shows you know nothing about communism, probably because you were educated in one of the US private schools.

Posted by: Joel on January 28, 2006 at 7:34 PM | PERMALINK

Kevin, it's never been about the grades or opportunities. Those elements are window dressing.

The real reason, in the end, for most parents, is segregation by choice. (I'm not talking specifically racial segregation - economic and social probably weigh as much or more than racial)


Posted by: merelycurious on January 28, 2006 at 7:36 PM | PERMALINK

Another interesting factor to throw in would be money spent per student.

One of the failures of the American public school systems is what poor performance we get for the dollars spent.

Posted by: John Hansen on January 28, 2006 at 7:38 PM | PERMALINK

The only reason I think perverse statictics are manipulated to make it look like the public school system is doing extremely well is because the main culprit in making American public schools lousy ( the various teacher's unions of America ) is such a strong force in democratic politics.

You need to step away from the bong, or whatever it is that is clouding you thinking, if you truly believe that crap.

A study that shows objective reality is not an effort to "manipulate" something to make it look "extremely well." There is not a shred of rationality in your statement there.

If Belgium is doing much better, lets learn from them. Of course, what if Japan is doing much better, too, and China, and France? Suddenly your conservative mantra of "privatize!" goes out the window, as you can find models not at all like Belgium that are doing good too.

You need to clear up your "thinking."

But really, comparing between countries is very hard. As mentioned above, there is no control. And the changes you want done just happen to be things you magically attribute to Belgium's success, when it could just as easily be something else entirely (like, maybe Belgium has a well-off society, or a huge social welfare system that sets up the necessary situation for kids to learn in [stability and money]. Or Something else altogether).

Posted by: teece on January 28, 2006 at 7:38 PM | PERMALINK

"Another interesting factor to throw in would be money spent per student."

It would be, if that comparison was made between students matched for socioeconomic level, family situation and *disability.* But that would be hard, because private schools aren't forced to take all kids in the district, leaving public schools with the most costly kids.

Just comparing raw numbers is stupid. But that's what you might expect from a private-school-"educated" critic.

Posted by: Joel on January 28, 2006 at 7:44 PM | PERMALINK

Reason tempered by civility is the virtue at the foundation of the best of the modern world. As with any virtue, it requires discipline and respect for what is not yet known and may be never be known by the individual.

Any school can teach this. The best hope for parents should be that their children know more of that virtue than they do.

A civilization that honors accumulation of material goods as opposed to principles makes a mockery of the designation.

Posted by: Michael7843853 G-O in 08! on January 28, 2006 at 7:46 PM | PERMALINK

John Hanson and others:

Public school education, like a vast number of other social concerns, fall into what I call the Jig-Saw Puzzle Paradigm. There are hundreds, if not thousands of variables. Some are obvious, others not. Isolating any single issue may or (usually) may not help with the big picture.

That is why this topic is so frustrating to the "slam bam thankyou mam" mentality of most Americans.

Posted by: Keith G on January 28, 2006 at 7:47 PM | PERMALINK

when you control for things like income, race, home environment, and so forth, the performance of private schools actually turns out to be worse or about the same as that of public schools, not better

Do you think that this means that Lutherans do better when they transfer to public schools, and that blacks do worse when they transfer to private schools? The "control" (actually a regression equation) assigns to the Lutherans the "effect" of the school, so that when the "effect" of the Lutherans is removed, the school is about average.

Consider it this way: a private school can remove repeatedly disruptive children, and when it does so the academic performance of the remainder increases; however, when you "control" for the fact that the class now has studious pupils, it's adjusted performance declines.

In analogous fashion (a close mathematical analogy), when you "control" for the effects of cancer and heart disease, smokers live at least as long as non-smokers.

One process in the school system is that good teachers migrate toward the good schools, and bad teachers migrate toward the bad schools or stay in place; then when you "control" for the quality of the students in the schools, the teachers are all equal. Good students migrate, when they are able, toward good schools; again, when you "control" for the quality of the students, all the schools are about equal (as you reported in your post.)

You could do this the other way around. If you "controlled" for the quality of the schools, all the classes of students would be about equal.

This works for physical attributes as well: when you "control" for the sizes of their parents, all children have about the same height; and when you "control" for the heights of children, all parents are about the same height.

Luckily, correlation and regression do no imply any particular causal relationship.

Posted by: contentious on January 28, 2006 at 7:51 PM | PERMALINK

Keith, it is a difficult challenge, but the study Kevin linked to does a good job of meeting the challenge. They identified the key variables and controlled for them.

"That is why this topic is so frustrating to the "slam bam thankyou mam" mentality of most Americans."

Actually, I think a large part of the problem is innumeracy. Most Americans don't understand numbers, statistics or controls.

Posted by: Joel on January 28, 2006 at 7:52 PM | PERMALINK

It really isn't just prejudice that the study may not change enough minds, I think. It tries, as one might hope, to factor out what students and parents themselves bring to test scores other than talent, to see how talent rather than wealth is best nurtured. But one could argue that this adds a distortion: a private school could provide better education, but naturally those who can afford it benefit disproportionately. So it could argue against the horrible income inequality in America or, perhaps, even for vouchers, were they able really to allow anyone access to any school, rather than only the few for whom a little extra money would make the difference and rather than those already able to afford it but reaping the cash windfall to the wealthy, and were it not then draining educational opportunites for the rest. (Indeed, public schools ARE an effective way of evening the field in this way, just like or better than vouchers, I'd argue.) Moreover, as another writer has noted, private schools can have benefits beyond test scores, in elective courses, AP courses, extracurricular activities, mentors, and other support lines and challenges.

So it's not convincing taking the cream of private schools against urban public schools. It's not an argument for the rich to move their kids to public schools. Rather, it's a key point in disabusing the hopes of people like me who focused on charter schools to resolve problems within cities, with comparable student populations, even if we opposed vouchers. It does a good job of convincing me, against my hopes, that it's not really about creative teaching, something that the market could winnow out, but old fashioned nuts and bolts of dealing with a student population. Money put into higher-paid teachers and smaller class sizes, I fear, despite right-wing protests, still is the way to go.

Posted by: artcrit on January 28, 2006 at 7:53 PM | PERMALINK

Joel

Consider the following analogy please.

EQUAL ACCESS TO FOOD ACT:

1. Everyone will be assigned a local grocery store based on where you live. You are not allowed to shop at any other government store.

2. Employees at the grocery store will have their salaries set by government contract according to years of service independent of how good their service is.

3. There will be a titular manager at each store but he will have no power over the employees because he has no control over salaries.

Anyone would recognize this as a system that is patterned after a communist business model. ( Note: its not completely communist, but is similar in its business model - large bureaucratic control from above, no choice for consumers... ).

Most Americans would reject this as a food distribution system. However substitute EDUCATION for FOOD and school for store and you have a good approximation of the American public school system.

Posted by: John Hansen on January 28, 2006 at 7:55 PM | PERMALINK

"This works for physical attributes as well: when you "control" for the sizes of their parents, all children have about the same height; and when you "control" for the heights of children, all parents are about the same height."

Uh, no, not necessarily. For instance, this is not true of Japanese children over the last 50 years. That's because height is not solely a function of genetics, but is also affected by environment (e.g., nutrition).

Posted by: Joel on January 28, 2006 at 7:55 PM | PERMALINK

"...when you control for things like income, race, home environment, and so forth, ..."

An old statistical trick. When you control for everything else, the thing you want to highlight shows the magic. Kevin has simply shown that teachers are generally well trained.

So, what are we suburban parents running from? Not the teachers. Mostly, we are running from the parents of the other students in urban schools, we parents are running.

So, Kevin, do a study, control everything except parents running from neighborhoods with bad dysfunctional families. You will find that parents which remained intact, and made it to suburban schools with like minded, school involved parents did substantially better than the schools these parents ran from.

Posted by: Matt on January 28, 2006 at 7:59 PM | PERMALINK

John, you are simply ignorant of what communism is. In the US, you have a choice of where to live. Moreover, public schools are not controlled by the federal government. Moreover, your children do not have to go to public schools at all. None of that was true in actual, real "communist" countries. But in your breathless need to use the word "communism" as a substitute for honest thought, you're content with specious analogy. Which is why you just don't get it.

So is it true you were educated in private schools?

Posted by: Joel on January 28, 2006 at 8:00 PM | PERMALINK

Sorry, I look forward to reading your answer later, but gotta take the wife out dancing.

We are both public-school-educated Ph.D.s (hard sciences, basic research).

Bye now.

Posted by: Joel on January 28, 2006 at 8:03 PM | PERMALINK

I have two comments:
I taught in a Houston high school for a couple of years in the 1980s. While there I taught an introduction to computers course in a Magnet School so I had some very bright kids. Most of them were 9th and 10th graders. I also had a 12th grade Norwegian exchange student. He was a very nice and bright young man, but our kids were every bit as capable and knew as much about computers as he did.

My second point is that Sen. Kerry is working on getting support for a filibuster. This site makes no mention of it. Is this education question really more important at this time, than discussing the possible filibuster?

Posted by: Bear Country on January 28, 2006 at 8:04 PM | PERMALINK

If you "controlled" for the quality of the schools, all the classes of students would be about equal.

An old statistical trick. When you control for everything else, the thing you want to highlight shows the magic.

You guys really aren't this dense, are you?

Posted by: teece on January 28, 2006 at 8:06 PM | PERMALINK

Extraordinary claims require extraordinary proof.

There are tons of anecdotal and statistical evidence that private schools are better, so much so that much of the argument against vouchers, etc. has been about the externalities caused by a shift (ie. the negative impact on the poorest and worst students who would remain stuch in public schools) rather than claiming that students who moved to private schools would not benefit.

There are also good logical reasons for expecting students to do better in private schools:

  1. The public system includes some totally dysfunctional schools in which it is hard to believe anyone can learn. The general private system does not - by the time a school gets that bad the parents just remove the kids and send them to public schools.

  2. Private schools cost a lot of money. Many offer scholarships for the poorest families but a parent has to be smart and motivated to jump through the hoops to get them. This should mean that students from very poor families in private schools have extremely motivated parents who are highly involved in their educations. How can this not increase their performance in school?

  3. The simple fact that a parent at almost any income level is willing to pay for his child's education implies a more stable family that cares more about the child's education.

  4. Private schools can remove criminal and disruptive students more easily. This should make it easier to educate the other students.

Given all this it is hard to see how public schools can do better than private schools even if their teaching quality is similar - they are not competing on a level playing field.

This is an extraordinary claim and one single study that now claims that public schools are betteris interesting, but not enough to disprove the previous received wisdom. Instead it needs to be further investigated and subjected to intensive analysis.

I can think of several potential explanations for the anomalous result:

  1. Inclusion of private schools that specialize in troubled students

  2. Inclusion of unacredited private schools that are run by weird cults or other unqualified people, that are not truly aimed at educating students in normal topics, and that would not be eligible for vouchers in any plan being currently discussed.

  3. Selection bias on the part of parents - maybe for many income levels only parents whose children are doing badly in public school are willing to spend the money to move them to private schools.

I think it would be really interesting to see a longitudinal study - go to some poor inner city neighborhood, pick out 2,000 preschoolers, give half of them tuition vouchers, and then follow them for 15 years.

Posted by: Michael Friedman on January 28, 2006 at 8:09 PM | PERMALINK

You need to step away from the bong, or whatever it is that is clouding you thinking, if you truly believe that crap.

You need to clear up your "thinking."

One of the things that is always entertaining about trying to express the conservative side of issues on this board is the extent of the name calling that is endured. Would you rather sit in your bubble chamber and not be challenged.

Please don't assume that someone who thinks differently from you is an idiot.

In reality I think that it is very, very difficult to make adequate controls for variables in social studies. However, people with an agenda will be quick to publish all the results that agree with their agenda. Thus their is sort of a natural selection paradigm about the manipulation of the raw data. The particular manipulation of the data that jibes with their agenda will be quickly published. Manipulation of the data that does not jibe with that agenda will not be considered noteworthy and not be published. So indeed maybe not conciously, but simply in accordance with natural human behavior the data will be manipulated to show a certain result, based on the agenda or the investigators.

It just human nature that causes this, not any deviousness on the part of social scientists.

Note: the same must be said about studies that come from conservative think tanks.

Posted by: John Hansen on January 28, 2006 at 8:15 PM | PERMALINK

PS. For what it's worthy, I grew up in Brooklyn, New York.

I went to private schools up to 8th grade and then, pretty much at my own insistence, switched over to Stuyvesant, one of NYC's four selective highschools - you had to take a test to get in.

My opinion is that Stuyvesant was a better school than Poly Prep. But I also saw other public highschools at various events and met their students and I think they were mostly worse than Poly.

Posted by: Michael Friedman on January 28, 2006 at 8:16 PM | PERMALINK

Hansen, your analogy is beyond silly, but you know that. Our system was initially designed to accomplish specific goals.

Yes we do need a redesign (example: operating with a pre-industrial calendar); yet when changes are proposed (as in my city), hundreds of constituencies come out of the shadows.

In our case, a research supported year-round curriculum was shot down by a coalition of Little League parents and businesses that depend on surplus teenage labor in the summer. Jig-Saw strikes again.

Notice how in the above case, it was not teacher organizations or the bad old gummit that nixed reform.

Posted by: Keith G on January 28, 2006 at 8:17 PM | PERMALINK

A much larger, more relevant, and, um, braver sample would involve comparing the K-12, publicly-supported systems of different countries. I strongly suspect such a study would demonstrate better results for those systems (such as those in Holland, Japan, Hong Kong, Scandanavia) that employ customer-driven choice and competition, and eschew the American M.O. of assigning each customer a school based on address or race.

Posted by: 99 on January 28, 2006 at 8:20 PM | PERMALINK

You guys really aren't this dense, are you?

teece, you appear not to understand how regression works.

joel, anybody nowadays also includes the year of entry into the study as a covariate; adding that to the regression of child height on parent height would lead to the conclusion that I wrote. Indeed, when year of entry and parent height are both included, nutrition is unrelated.

One of the interesting outcomes of the Milwaukee school district program was that the public school performance improved following the introduction of the voucher program. that's way more important than this regression analysis. Of course, if you include time as a covariate, it "accounts for" the effect of the change.

Posted by: contentious on January 28, 2006 at 8:20 PM | PERMALINK

Joel,

Sorry to dissapoint you, I was publicly educated.

Valedictorian of my public High School, Newport Harbor High School.

Graduated UCLA with a 3.995 summa cum laude.

Ph.d UCLA 1990 in Physics.

Invited paper at the American Physics Society 1991.

Its just convenient for you to try and label someone who disagrees with you as an idiot. Its just a self defense mechanism. Hopefully, you will get over it and be able to have a respectable civil discourse with others.

Posted by: John Hansen on January 28, 2006 at 8:26 PM | PERMALINK

Controlling for things like race, SES, home environment, etc. is standard practice for studies like this. If your goal is to find out how good the school is, you try to control for things that are extraneous to the school itself.

And of course the results are only true in the aggregate. On average, this study is saying that if two similar people went to a private school and a public school, they'd do about equally well -- or maybe a bit better in the public school. It's an indication that on average public schools are not inferior simply by virtue of being public schools.

Of course, it means nothing about any particular school. If you lived in an inner city with only one crappy public school nearby, you might very well do better by attending a nearby private school instead.

Posted by: Kevin Drum on January 28, 2006 at 8:37 PM | PERMALINK

Please don't assume that someone who thinks differently from you is an idiot.

JH, that sentence I quoted of yours is idiotic. You are being judged by your words, not your ideology. You posit a grand conspiracy by teachers unions to pad numbers, thus dismissing this study without even mentioning. It's infantile nonsense you spouted.

As for your point about the authors of this study having an agenda: give me a break, would you? You have the agenda. If there is some fatal statistical flaw in this study, find it. You're not doing that: you're just assuming its wrong because it is not what you want to hear.

contentious: I understand regression just fine. Better than most. You don't have a point, and you know it. You're spouting complete nonsense. If your are interested in the quality of education provided by private vs. public, you must control for other factors that you already know have a profound impact on quality of education.

If you think that's a statistical sleight of hand, you're innumerate. More likely, even if you are numerate, you're just dismissing the study out of hand, because you don't like it. That's not thinking, so please don't try to convince us it is.

Only a fool would argue that the deciding factor in education is public vs. private. Indeed, even the proponents of private school will admit, in a circumspect fashion, that this study is confirms a reality they knew already existed.

Private schools get sent kids that are already inclined to do much better in school. That does not mean that private schools are inherently better at educating kids.

Posted by: teece on January 28, 2006 at 8:39 PM | PERMALINK

Lets consider another problem that clean aggregation of data hides. By aggregating data you always hide something.

The aggregate performance of the American school system is bad compared to other countries. Finding out the reason for that is very difficult.

But worse than that is the problem faced by those with under-performing schools.

Poor people trapped in an underperforming district are by the current system trapped. Releasing the bureaucratic bondage of the system that keeps underperforming schools and teachers in place would certainly help.

I only know anecdotal evidence of vouchers that says it helps to clean up bad schools. Does anyone have anecdotal evidence that vouchers hurt schools?

Posted by: John Hansen on January 28, 2006 at 8:45 PM | PERMALINK

Does anyone have anecdotal evidence that vouchers hurt schools?

Why, yes, I do, JH. Most parents are not willing or able to "choose" their child's school. The school near their house is where they go.

In my hometown, the limited voucher crap we have makes marginal schools worse. Why? Because the minority of parents that can use the voucher system take their kids to the magnate school. The regular school gets left with all the problem kids.

Thus, through no fault of their own (rather by the fault of their parents), the public school screws over those kids that get left in the now-"failing school."

All of the voucher systems I have seen fail to address this point. They (marginally) help parents that were going to already help themselves anyway. The do nothing to address the problems of the poor (generally inner city, generally minority) students who were going to be screwed over by our school system to begin with.

Posted by: teece on January 28, 2006 at 8:54 PM | PERMALINK

teece - One of the things about communication is that some people read into the words something that was put in that the author did not intend.

I apologize for the language of the first post that implied that people were deliberately manipulating the results. I think I explained myself better in the second post. I am not accusing the teachers union of padding the data. Like I said above - natural selection of the results tends for publication of that particular manipulation of the data which favors the agenda. How many times have you ever seen a think tank release a study which contradicted their own views. It just does not happen. Contradictory results are not considered interesting to the average person. I am not accusing social scientists of being devious, just human.

Posted by: John Hansen on January 28, 2006 at 8:57 PM | PERMALINK

http://www.ctredpol.org/vouchers/schoolvouchers.pdf

The above refenced site in scholarly, not anecdotal, and posits that current research is inconclusive. It nonetheless is a quick and interesting read.

Posted by: Keith G on January 28, 2006 at 9:05 PM | PERMALINK

So a public school performs to the same standards as a charter school when they control for a number of variables.

I suppose they neglected to control for the fact that public schools spend more per student than charters.

I suppose I'll actually have to read the report to see whether the study design is well constructed.

Posted by: TangoMan on January 28, 2006 at 9:12 PM | PERMALINK

Give me a break!

Posted by: Stossel on January 28, 2006 at 9:14 PM | PERMALINK

er, American Physical Society, not Physics Society.

Your tell 'em John - We, Belgies, are superior intellects.

Posted by: Hercule Poirot on January 28, 2006 at 9:18 PM | PERMALINK

er, American Physical Society, not Physics Society.

Sorry, for the innaccuracy, I'm curious what your point is.

Posted by: John Hansen on January 28, 2006 at 9:37 PM | PERMALINK

Does anyone have a cross plot of school performance vs. money spent per student? I'll bet there is a pretty good correlation for both private and public schools. If you indexed teacher pay to cost of living you'd probably see an even better correlation. A good portion of this is just demographics -- wealthy white kids with educated parents live in districts with stable tax bases and well funded public schools. But part of it is related to better teachers going to better funded schools.

Menlo-Atherton High is a hell of a lot better school than any private school within a 100 miles of my home.

Posted by: tasmaniandevil on January 28, 2006 at 9:42 PM | PERMALINK

Of course anybody who has seriously and honestly looked into school performance knows that individual/family factors are far more important than school factors.

Trying to statistically control after the fact for demographics is difficult and approximate, though.

I'm not disputing the conclusions of this study. We shouldn't expect large differences from private v. public schools. What is shocking is that we ever use raw test scores to compare schools.

A better way of analyzing school performance is too look at the "value-added" by a school. That is, to compare how many points students improve on tests year after year, rather than their absolute scores.

This mostly controls for demographic factors. It's not a perfect system (e.g. is a difference from 600 to 650 the same as 650 to 700?), but it's vastly better than what is usually done. The most developed system is the one used in Tennessee, where it has been in use for over a decade (and is used for NCLB compliance.)

See, e.g., http://www.shearonforschools.com/TVAAS.html or just google value-added-assessment.

Posted by: Ken Hirsch on January 28, 2006 at 9:44 PM | PERMALINK

Has anyone looked at the original paper, and seen what the "controlling factors" actually are? Those who have commented on it are right that you can make a statistic do pretty much what you want if you "control" for enough things, and make the right assumptions about the controls.

When it comes down to it, parents will make their own choices on schools, based on a hell of a lot more than math grades, and society needs to give them as wide a choice as possible. A highly-centralized solution will not be the best one.

Many parents around my area, with a large Hispanic population, do not restrict themselves to the "closest" school if given a choice by the district. My kids didn't go to the closest one, although it needed some research to pick out, and some paperwork for permission.

Posted by: tbrosz on January 28, 2006 at 9:55 PM | PERMALINK

How about someone do a long term study on how different types of schools affect someone's potential to go to college when their parents didn't, get a better job than their parents, get married to spouse more beautiful and intelligent than their parents, and live longer than demographic considerations would predict. Who cares about test scores?

Posted by: tasmaniandevil on January 28, 2006 at 10:01 PM | PERMALINK

I live in an affluent community in southern california where the public schools are better than the available private schools. I moved my kid from private school to public school because it was a better school. It's a small school district that has it's own 501-c-3 organization that does nothing but raise money for the schools year round. Want to know the difference between good schools and bad schools? Money.

Posted by: nina on January 28, 2006 at 10:03 PM | PERMALINK

Gee Kevin, Don't you think a smart guy like yourself would spend a few minutes to look at the NAEP data -- and try to figure out why this "study" didn't compare the ACTUAL scores (Public vs Private vs Charter) for each of the subgroups they used to "weight down" the private school scores.

The question is: how well did each demographic fare in their respective schools -- not how to weigh down the overall score of a school by counting the noses of the subgroup.

That is to say, how well did students in each subgroup (black, white, green, etc) do in Public vs. Private schools.

The NAEP has the data. Why didn't this study use it? D'OH!

OBTW, a far more meaningful analyis could be made if we could see the year-to-year improvement of each subject. That would speak more to the quality of the school than of the student.

And you wonder why I think you're a moron?

Posted by: Norman on January 28, 2006 at 10:12 PM | PERMALINK

Alas, the attack on public education is not new. It is new that the radical PR machine has been able to put the idea that private education is superior to public education into the minds of working Americans. One hundred years ago public education, free to all children, was a triumph of American democracy. We can see that the turn, as with so much in the Republican party, began in the 1930s. The changing attitude of William Randolph Hearst (the Murdoch of his day) is particularly telling. Hearsts mother, like many American aristocrats, had been a great supporter of public education and had, with her fortune, built the University of California. She supported scholarship, free kindergartens and the Parent-Teacher Association, as well as, many public libraries.

William Randolph Hearst had other ideas. His reaction to the New Deal and everything it stood for was a sharp departure from the anit-wealthy populism of his youth. He ran a Dont Soak the Rich campaign in reaction to the progressive income tax (now after a long struggle rich men like Hearst are well on their way to having it repealed). If you read Hearst's papers the enemy was everywhere, including the schools. He hired agents to intimidate professors and sent spies into classrooms (much like David Horowitzs operation), he railed against academics in his papers and suggested deportation, blacklisting and loyalty oaths. Unlike today these tactics did not have many sympathizers even among conservatives who tended to condemn him. So harsh was Hearsts Red Baiting that many thought he had fascist sympathies. Indeed he ran syndicated columns by Hermann Goering and Alfred Rosenburg in his newspapers.

Much of the Republican agenda has been scripted by think tanks and provocateurs funded by a handful of super-wealthy families with attitudes similar to Hearsts. Because it is a reactionary ideology it is a kind of inverse communism with all the ideological zeal and urgency of Stalin. It substitutes radical socialism that holds society is everything with a radical individualism where, not surprisingly, there is no society.

Posted by: bellumregio on January 28, 2006 at 10:54 PM | PERMALINK

A few points--

Controlling for demographics and location is not the same thing as controlling for good students (otherwise it would be tautological). They used math scores because math achievement is less dependent on home life than other subjects.

There are several countries that do very well on the international comparisons that don't have much or any school choice or voucher systems--Finland and Singapore come to mind. Both are well-regarded and seem to regard their teachers as valuable professionals, rather than greedy "special interests." Who would be a teacher in this country, when you look forward to constantly being accused of being shiftless and greedy.

Also, when I hear about how inefficient public education is, and then I think about the $20,000/yr that my boss is spending on his child's private education, I just have to laugh. There is *no* secular private school in my area that costs as little per year as the public districts spend per student. The parochial schools can get the price down in the ballpark of public schools, but they often have the advantage of inexpensive labor. (Of course I'm being a little unfair here--public schools often get their capital costs paid off the books via bond issues).

Posted by: me2i81 on January 28, 2006 at 11:22 PM | PERMALINK

Isn't it self-evident that one of the benefits of putting a kid into private school is to improve the demographics of their peer group?

Posted by: TWAndrews on January 28, 2006 at 11:22 PM | PERMALINK

Want to know the difference between good schools and bad schools? Money.

If that were the case New York City schools would be producing the best educational results on the planet.

In fact, it is certainly the case that the socio-economic status of the parent(s) (in particular their level of educational attainment) is the most important factor in determing how well kids do in school (not how much money the schools themselves spend). Schools filled with the progeny of the well-educated tend to produce good results. If you forced the kids of Scarsdale and Newark to switch facilities, the former would still be Harvard-bound, and the latter would mostly not be.

Posted by: 99 on January 28, 2006 at 11:35 PM | PERMALINK

The only exception is the conservative Christian schools, which continue to score considerably lower than public schools

this amuses me.

Posted by: Nads on January 28, 2006 at 11:51 PM | PERMALINK

Kevin Drum: Controlling for things like race, SES, home environment, etc. is standard practice for studies like this. If your goal is to find out how good the school is, you try to control for things that are extraneous to the school itself.

The problem is that things like the characteristics of the pupils are not extraneous to the schools themselves. In other words, without random assignment of kids to schools, the "measured effects" of schools and the "measured effects" of the demographic characteristics of the kids in the schools are not orthogonal.

The observed effects of the demographics and the schools are assigned to the demographics and subtracted from the total, leaving the school effects underestimated,

But think of the questions I asked. Do you really think that if some black kids transferred from public to private schools their learning would decrease in line with the reduced quality of the schools. How about if some Lutherans transferred from Lutheran to public schools -- would their learning improve?

Posted by: contentious on January 29, 2006 at 12:02 AM | PERMALINK

teece: You don't have a point, and you know it. You're spouting complete nonsense.

read my comment to Kevin about confounding. the authors are overcorrecting for the effects of talented kids in the good schools.

Since it is known that good schools produce better students, you could correct the demographiic effects for the good schools. In doing so, you would find that, corrected for quality of schools, Chinese actually underperform Hispanics.

Overcorrecting for quality of schools reverses or eliminates the effect of ethnicity; overcorrecting for the quality of the ethnic groups reverses or eliminates the effect of schools.

As long as better students migrate to better schools the relative importance of schools and demographics can not be disentangled (or unconfounded) by regression analysis.

What you need to look at are the consequences of specific changes in specific places. Kansas City, MO dramatically improved the quality of its school system with no effect on measured learning. Milwaukee instituted a voucher system and got enhanced learning in all schools. A study of charter schools that Kevin posted last year showed that charter schools did as well as the competing public schools but at lower cost.

Posted by: contentious on January 29, 2006 at 12:15 AM | PERMALINK

Sample sizes used for this study:

. . . . . . . .#Students . . . . #Schools
-Public - - - 182,328 . . . 6,797
-Catholic - - - -2,285 . . . 216
-Lutheran - - - - -555 . . . 88
-Christian - - - - -651 . . . 78
-Other - - - - - -1,227 . . . 157
-Charter - - - - 3,101 . . . 149

Posted by: TangoMan on January 29, 2006 at 12:15 AM | PERMALINK

Do you really think that if some black kids transferred from public to private schools their learning would decrease in line with the reduced quality of the schools.

The National Bureau of Economic Research reports:

Families originally living in public housing were assigned housing vouchers by lottery, encouraging moves to neighborhoods with lower poverty rates. Although we had hypothesized that reading and math test scores would be higher among children in families offered vouchers (with larger effects among younger children), the results show no significant effects on test scores for any age group among over 5000 children ages 6 to 20 in 2002 who were assessed four to seven years after randomization. Program impacts on school environments were considerably smaller than impacts on neighborhoods, suggesting that achievement-related benefits from improved neighborhood environments are alone small.
Posted by: TangoMan on January 29, 2006 at 12:19 AM | PERMALINK

It is difficult to draw any meaningful conclusions from a single study, but it in no way speaks to the cultural differences between public and private schools.

Except for the question of social diversity (access to different types of people, by which I don't necessarily mean people of different ethnic and racial backgrounds; the private school I attended was far more diverse in that sense than any public schools I had attended), it seems difficult to argue that private schools are not a more decent, tolerant environment where creativity and critical thinking are encouraged.

I don't remember if Kevin linked to it or not, but there was a recent study suggesting an astonishingly high percentage of gifted kids drop out of public schools. I suspect when you factor in the percentage of gifted kids being left behind and underperforming in public schools, that percentage rises much higher.

It is not unreasonable to think that many smart kids are happier in private schools where they are far less likely to harassed, bullied, and treated as outcasts. The culture of public schools is highly illiberal, elevating asshole jocks and dumb, mean girls to the top of social pile.

Giving the smartest kids (of all races and classes) access to places where they feel most comfortable may be the best reason to support vouchers.

Posted by: The Blue Nomad on January 29, 2006 at 12:19 AM | PERMALINK

Average public school expenditures per student can be found here.

In constant dollars, we are spending twice the amount of money per student than we were in 1970.

This little chart shows what teachers have been paid. In the same time period, since 1970, in constant dollars, their salaries have inched up from $42,489 up to $45,822. You can figure out the percentage increase if you want. In some intervening years it even dipped.

I have put two kids through public school, and have yet to run across a crappy teacher. Most put their own money into classroom supplies and things that the school can't or won't provide.

You figure out where the hell all our education money is going, and you'll figure out what's wrong with the system.

This chart has the average tuition paid by private school students. I don't know how that's changed over time, but it would be good to find out.

Only the non-sectarian private schools exceed the public school per-student costs.

Posted by: tbrosz on January 29, 2006 at 12:46 AM | PERMALINK

Two other points:

1) The study doesn't look at differences in high school performance.

2) It is easy to quantify income differences but this study attempted to look at "home life" factors as well. I am immediately and overwhelmingly dubious of the validity of the latter. One suspects that's where the massaging was.

Posted by: The Blue Nomad on January 29, 2006 at 12:49 AM | PERMALINK

This is the single silliest post I have ever read here. To state that, sure, private schools score better, but that's only because they have better students --- I mean, its so astonishingly ridiculous I'm almost speechless.

There may well be some people in the United States who believe that public and private schools start with the same statistical student body, and thus, any differences in end performance is a function of the quality of the school, but come on! Those people are, ahh, I can't even continue.

The policy decision is not how to allocate the better students among the schools, its how to assist the kids who are no lucky enough to be born to rich, educated parents who focus on education!

Posted by: hank on January 29, 2006 at 12:52 AM | PERMALINK

It's never a bad idea at times like this to find out what kind of work the authors have done in the past, and perhaps learn a bit about whether or not they have an ideological ax to grind.

It sems Mr. Lubeinski wasn't so fond of charter schools even before this study. From a 2003 publication by "The American Association of School Adminstrators."

"Contributor Christopher Lubienski clearly documents concerns about putative charter school innovations."

It seems he also wasn't so fond of voucher programs before undertaking this study.

"There is mixed evidence on academic gains for children in these programs," he said. "But there is some concern that increased segregation results, leading to overall declines in social and academic indicators."

And homeschooling is mind you only for wacko right-wing Christians.

Posted by: The Blue Nomad on January 29, 2006 at 1:08 AM | PERMALINK

I like this one even better.

Homeschooling is a social threat to public education, Chris Lubienski, who teaches at Iowa State University's college of education, told Time magazine."

Posted by: The Blue Nomad on January 29, 2006 at 1:23 AM | PERMALINK

Any minute now, someone is going to tell us that since average private school tuition is so much less than the average public school expenditure per pupil, this proves that money is not part of the equation.

Please don't assume that someone who thinks differently from you is an idiot.

Anyone, and I mean, anyone, who directs actual thinking human beings to "consult" a John Stossel "report" for any purpose short of ridicule is, by definition, an idiot. Whether the afflicted party is only temporarily stupefied by blind ideological presumption and dogmatic certitude is, to be charitable, yet undetermined.

Posted by: R.Porrofatto on January 29, 2006 at 1:50 AM | PERMALINK

Everyone seems to think that "private school" means the elite neighborhood school that they remember the children of ambitious parents having attended. In fact, there are thousands of private schools across the country, many of the mediocre and some of them quite poor. The statistics in this study expose the existence of these poor private schools, which normally fall "under the radar," since for most of us, those aren't the private schools that we attended, our friends attended, or want our children to attend-- but they are out there.

The elite, top-notch private schools that we know about and think about when we think of "private schools" are very, very expensive (twice as much as some of the highest-estimates of per-pupil expditures in public school systems) and not particularly common.

Providing vouchers in inner cities may have certain social and demographic benefits, however.

Posted by: Constantine on January 29, 2006 at 2:43 AM | PERMALINK

"Better kids"?!?!

Posted by: Nancy Irving on January 29, 2006 at 4:40 AM | PERMALINK

I wish I had had this information when the vice principal at my daughter's public school asked, when I complained, if I had considered parochial schools.

I have no doubt that it's true, having sent my kids to good, bad, and in-between public schools in three different states.

Posted by: KathyF on January 29, 2006 at 6:15 AM | PERMALINK

I teach at the public high school where John Stossel got his education. I'm beginning to think that maybe we are inferior.

Here's something to think about: parents choose private schools based on 'school climate' and the availability of specific programs to fit their kids' needs much more often than test scores or the overall effectiveness of the academic curriculum.

Here's another thing to think about: Even after you control for various demographic factors, you haven't really controlled for the fact that there are some differences between who chooses a public school and who chooses a private school.

Posted by: reino on January 29, 2006 at 7:26 AM | PERMALINK

Tbrosz -- If you go to the same source you are quoting -- a very good source thank you for providing it a long time ago -- you find that we are spending the same % of gdp on k-12 education now that we were spending 30 to 50 years ago.

This actually takes the discussion off on a different tangent. There is essentially no difference between public and private schools teaching methodology. No one has improved on the age old approach of one-on-one teaching with a maximum of some 20 to 30 teachers per student.
Both private and public schools are they same in this reguard. But that means there is no productivity in education that we can measure.
In theory, you could create some weighted result by student achievement but it would have all the problems discussed above in comparing schools.

So we are left with the problem that if you are going to pay teachers competitive salaries the real expenditures on education are going to have to rise over time because of a lack of productivity growth. Moreover, I see no reason why this would not be true for both public and private schools.

This study is similiar to many others in that it finds no significant difference between public and private schools once you adjust for the background of the students. But that is our problem. The public education system does a fine job of educating children from average and above average homes. The problem is we are now trying to educate kides from poor home environments. But no one has demonstrated that this can be done without spending much more on these kids. So the consequence is that the voucher advocates are promising a free lunch, and that is the real problem with their argument.

Yes, vouchers could help an occassional kid, but that does not really help solve the problem. As was pointed out above, removing the exceptional kid from a poor school just makes the school worse for those left. This is just another example of how the voucher argument is just a free lunch argument.

Posted by: spencer on January 29, 2006 at 8:41 AM | PERMALINK

In other news, if you take the ham out of a ham and egg sandwich, it won't be a ham and egg sandwich.

Posted by: John Isbell on January 29, 2006 at 9:32 AM | PERMALINK

"You figure out where the hell all our education money is going, and you'll figure out what's wrong with the system."

Well, around here we've got a) layers and layers of useless administrators and b) schools being built for way over what they should cost on contracts that weren't put out for multiple bids.

Posted by: MJ Memphis on January 29, 2006 at 9:39 AM | PERMALINK

Private schools pay teachers much less than public schools. Hence the quality of the teachers is probably lower than that in public schools. Therefore, the education in private schools is probably less stellar.

Posted by: Carol on January 29, 2006 at 9:53 AM | PERMALINK

when you control for things like income, race, home environment, and so forth

All the people who think that vouchers will give them more choices are going to find out that the "free market" is going to lock them out of school choice.

Clients of private schools are paying for exclusivity.

Posted by: Stephen on January 29, 2006 at 10:47 AM | PERMALINK

tango man, thanks for the links.

It is good to know that the vouchers did not result in lower test scores. I thought that they wouldn't.

Posted by: contentious on January 29, 2006 at 11:07 AM | PERMALINK

Clowns.

4th grade is not the problem; in fact in 4th grade, American education is as successful for the most part as the rest of the industrialized world. No link, but it is statistically true. The problem really is the 12+ adolescent public education where newly hormone driven students are either being supervised in a caring and nurturing environment, or , statistically speaking, going to public schools dominated by union members who cannot be fired.

But, how about the courage of your convictions; if there really is no difference, vouchers will have no effect, right. Surrreee.

Anti-voucher advocacy is the equivalent of child-abuse and since it is child-abuse in exchange for money and perquisites, it's as non-sexually close to pimping children as it comes.

Stop the madness; put children first, i.e., before union advocacy. Be "pro-choice" for children. Be "liberals" for children's liberty to go to a school of his or her choice. Be "progressive" in incentivizing public school bureaucrats and teachers to make progress and to cause education generally to make progress. Be "socialists" by putting quality of education and society ahead of making sure people who make on average $50,000+ for 180 work days year cannot be fired no matter how burned out or otherwise unmotivated and uneffective they are.

Stop being the 21st century George Wallace at the school house door; where he stopped African-Americans from going to the better "white" schools, you are preventing poor children from going to the better private schools. You are as wrong as George Wallace was and like your fellow Democrat, George Wallace, history will remember your perfidy in regard to our precious children.

TOH

Posted by: The Objective Historian on January 29, 2006 at 11:43 AM | PERMALINK

tango man,

unfortunately, that's just the abstract. I would doubt the abstract published by "Social Science Research News" because of the enormous Democratic bias among social science researchers, and my familiarity with the inadequacies of the data analyses of most social science researchers. that said, I do still thank you for the link. I intend to buy the article.

Posted by: contentious on January 29, 2006 at 12:04 PM | PERMALINK

Lots of thoughts on this study....

1) There are lots of different types of private schools, so lumping them all together is a bit specious. Independent private schools undoubtedly do a lot better job than religiously affiliated schools, both for funding reasons and because the goal of the latter is more oriented towards transmission of values rather than providing a top-notch education.

2) Motivated students are going to do well regardless of which type of school they go to.

3) At the elementary level, private schools and public schools are pretty similar. Where the difference really lies is at the high school level, which this study apparently did not investigate. As some else already pointed out, private high schools offer far more electives and other types of learning opportunities (school year abroad, field trips, etc.) than public schools.

4)Likewise, private high school faculty are generally much more knowledgable about their subject areas than public school faculty. At the private high school I went to, every single teacher had a Master's in their subject area, and more than a few had doctorates.

5)High school is also where you start to see a greater divergence in class sizes. At the private high school I went to, I never had a class with more than 18 students in it. At the public high school I taught at ten years later (one of the best public high schools in CA), I never taught a class with fewer than 26 students, and about half my classes had more than 30 students.

6)The advantages which private schools offer are not the types of advantages that are necessarily measurable on standardized tests like the NAEP. Often they are things like better facilities (especially for science classes), the chance to undertake more complex learning activities, better writing programs, social networking, etc.

All anecdotal evidence, to be sure, but food for thought nevertheless....

Posted by: MattW on January 29, 2006 at 12:27 PM | PERMALINK

"The Objective Historian" begins his post with "Clowns" and proceeds, in what follows, to demonstrate his firsthand knowledge of that subject.

MattW writes: "The advantages which private schools offer are not the types of advantages that are necessarily measurable on standardized tests like the NAEP."

That's as may be, but if you cannot measure it, we are left with anecdotes and opinions. The study Kevin posted is a good attempt to make quantitative comparisons. These cannot be wished away by subjective impressions, or by tedious invective of the sort spewed by TOH, above. As a scientist, I'm convinced by data.

Posted by: Joel on January 29, 2006 at 12:58 PM | PERMALINK

Let's face it -- a lot of this is based on our traditional Fear of Negroes (which in recent years has been expanded to Fear of Non-Asian Minorities), back to the old silly miscegenation bugaboo. This is especially prevalent in years after elementary school, which corresponds with the onset of puberty. (How many affluent white parents do you know who will send their kids to a public elementary school to make themselves look liberal and "progressive," then search like hell for a private school once he or she begins the journey out of childhood?)

If I were president, I would demand that all my executive branch employees send their kids to public schools -- and not one of those super-rich schools like Walt Whitman in Bethesda where a disproportionate number of grads go to elite private colleges. For that matter, I would insist that as a condition of employment, their children attend public universities (and yes, that includes the likes of Berkeley and Michigan -- believe it or not, there are good public colleges out there). I don't want my workers cocooned in an Ivy ivory tower from the conditions faced by American society; live like the rest of us.

Posted by: Vincent on January 29, 2006 at 1:05 PM | PERMALINK

MattW wrote this: 2) Motivated students are going to do well regardless of which type of school they go to.

That's only true partially.

Motivated students will still do better in schools that have higher standards and fewer classroom disruptions. Even a motivated student can't learn calculus in a school that doesn't offer it, and might not think to study the Kumon math curriculum.

joel, surely as a scientist you have learned to be sceptical of summaries of analyses proffered by social scientists at state universities working off federal grants. It's one of the few groups reliably far more biased than the Bush sdministration in the evaluation of complex problems and complex data. Not as bad as Intelligent Designers, but you really ought to be sceptical.

Posted by: contentious on January 29, 2006 at 1:08 PM | PERMALINK

What I've learned as a scientist, contentious, is not to indulge in ad hominem attacks on research studies. It does not trouble me, the way it does you, that the work was done at a state university or that it was federally funded. Indeed, much of the research in my field (biochemistry of gene expression) has been done at state universities and the vast majority (at both public and private institutions) was federally funded. I've learned to be skeptical of people, like you, who can't be bothered to think critically about the substance of a report, but instead carp about the authors' employer or funding.

Posted by: Joel on January 29, 2006 at 1:29 PM | PERMALINK

Yesterday, we had the good fortune to receive several comments from a former valedictorian at the tony Newport Harbor High School in Newport Beach, CA. Very wealthy district - Not to paint with a broad brush, but it reminded me of comments from another valedictorian of NHHS - Wrote a column complaining about having to compete for enrollment into the UC system with say, a valedictorian with a 4.0 average from Compton, CA High School who received the 4.0 only because he or she had not knifed the teachers.

Just love comments from "valedicks" of the Toniest public schools. We need more San Marino, Rancho Santa Fe, La Jolla, types discussing the Comptons and East LAs of the world. They have just so much insight into the problems of the inner city schools.

Posted by: thethirdPaul on January 29, 2006 at 1:52 PM | PERMALINK

Speaking as a non-white child of the private schools I have to say this is an interesting but not really very important point.

As has been said earlier on this blog, since you can't let the below-average students fall through the cracks in a public school, you can't give the smarter students the education appropriate to THEIR level, harming them. Hence if you are smart and wealthy enough, you go to a private school where they are generally better equipped to do that.

Of course to get the most out of that, you'd take all the smart kids and put them together and put all the stupid kids together regardless of income. But that has problems of its own.

Posted by: MNPundit on January 29, 2006 at 1:57 PM | PERMALINK

". . . since you can't let the below-average students fall through the cracks in a public school, you can't give the smarter students the education appropriate to THEIR level, harming them."

You can't? Why not?

My 17 yo daughter attends a public school. The student body is ca. 85% black. Most of the kids come from single-parent families.

My daughter is taking 5th year latin (college level) and calculus (college level), as well as honors lit and history courses. I haven't been able to detect the harm you so confidently describe. Could it be that you're, uh, generalizing here a little bit?

I don't believe the trade-off between helping below-average students and fostering excellence is an obligatory one.

Posted by: Joel on January 29, 2006 at 2:07 PM | PERMALINK

What I've learned as a scientist, contentious, is not to indulge in ad hominem attacks on research studies. It does not trouble me, the way it does you, that the work was done at a state university or that it was federally funded.

I didn't say it troubled me. I said that such summaries are always biased toward Democratic party objectives (or words to that effect.) This is quite well documented, and anybody not sceptical is naive.

My comment about the technical inadequacy of the analysis, the confounding of school effects with demographic effects, is something you should consider. I have taught this to University graduate students in the behavioral sciences.

Posted by: contentious on January 29, 2006 at 2:24 PM | PERMALINK

This is a very amusing dicsussion.

Vouchers are a way to take monies from public schools and give them to privileged families that are already attending public schools. All the folderol about normalizing for income and family environment is bloviating.

There might be an occasional inner city student, who through his own drive isn able to somhow secure a seat in one of the elite private schools and be successful bu7t it won't happen programatically because the private schools will absorb the public funds and continue to charge the families of the priveleged students the same amount they currently charge. Why?

Because what the parents are paying for isn't quality of education, it is precisely what they are currently delivering... education and social contacts with the social and economic elite which brings their children all sorts of benefits... lifetime relationships with the offspring of movers and shakers... an association with folks who take their education seriously enough to pay a premium for it... and of course, none of the riff-raff populating the public schools.

The only way to deliver this is to place a monetary barrier and enforce that barrier from breaching from undeisrables by maintaining rigid entry criteria.

This is all well and good, but the minute a nickel gets diverted from the public system to enrich the priveleged, we destroy any semblence of a meritocracy in favor of public financing of whe weaklthy a priveleged.

Posted by: CK Dexter Haven on January 29, 2006 at 2:30 PM | PERMALINK

" . . . such summaries are always biased toward Democratic party objectives . . . This is quite well documented . . . ."

Cite, please.

"My comment about the technical inadequacy of the analysis, the confounding of school effects with demographic effects, is something you should consider. I have taught this to University graduate students in the behavioral sciences."

Since I've missed those classes, perhaps you'd care to summarize the substance of your argument, together with some citations to the literature of the field?

Posted by: Joel on January 29, 2006 at 2:30 PM | PERMALINK

What's really interesting about this data is that
it proves private schools educate children for 1/3 the cost of public schools and produce the same educational results.

But you're right, these data won't change many minds. Particularly those of people who believe public schools are better. Now we know why.

Posted by: m on January 29, 2006 at 2:58 PM | PERMALINK

"What's really interesting about this data is that
it proves private schools educate children for 1/3 the cost of public schools and produce the same educational results."

What's really interesting about this comment is that it proves the writer didn't read past the first post.

But I was right, the data won't change many minds. And it can't have any effect at all on the mindless.

Posted by: Joel on January 29, 2006 at 3:00 PM | PERMALINK

Since I've missed those classes, perhaps you'd care to summarize the substance of your argument, together with some citations to the literature of the field?

The substance of my argument was already summarized in my comments on confounding.

R. R. Hocking, Methods and Applications of Linear Models, NY:Wiley, 1996.

It's a start.

there's also Cohen and Cohen, Applied Linear regression in the social sciences.

Posted by: contentious on January 29, 2006 at 3:22 PM | PERMALINK

Good job. And as for your other comment:

" . . . such summaries are always biased toward Democratic party objectives . . . This is quite well documented . . . ."

I trust it is as well documented as your assertions regarding confounding. I look forward to reading the citations you provide to support this assertion.

Posted by: Joel on January 29, 2006 at 3:32 PM | PERMALINK

Its funny that only on an anonymous message board can a dollop of reality rise above the general din of anecdotal nonsense.

Despite my despair last night that Kevin had sloppily asserted that this study has some argumentative value concerning the "quality" of public scholls vs. private schools, this study is quite helpful in pointing out the actual problem.

There are two problems with our educational system, not one. The first problem is the overall education rate of the entire student population. Frankly, although this is discussed endlessly, it is not a "problem" in the sense that the overall education rate is only an average, clearly, some students sets are doing quite well.

This study proves that the second problem is really the problem, basically that problem being "is it the right decision, from a policy perspective, to allow ones education to be mostly driven by the luck of who your parents happen to be?"

To say that "well, the private schools only out-perform the public schools because of the students, not because of the actual schools" may be illuminating to some, but certainly not to anyone who has actually sent their child to a private school.

Its hard to believe that a study was needed to show this, what's next, a study to show that toast always hits the floor butter side down?

But, taking the study at its absolute face value, which is that the real difference is the students, not the schools, raises the obvious, next question -- what are we, as a society, going to do about our worst students? The best students, and their families, are clearly already looking out for themselves.

Posted by: hank on January 29, 2006 at 3:38 PM | PERMALINK

contentious,

that said, I do still thank you for the link. I intend to buy the article.

Here's the full paper.

Posted by: TangoMan on January 29, 2006 at 3:58 PM | PERMALINK

I think the question raised by Hank about what are we going to do about our poorest students is the interesting one.

We can stay with the status quo in which inner city schools with poor management and worse teachers are allowed to shuffle the poor through the system, or we can take aim at the problem.

I am a conservative white man living comfortably in West LA. But I would not mind a special tax to pay teachers extra money to take on the challenges of teaching in the inner city public school. These teachers would be recruited for their ability to not only teach, but to relate to kids from broken homes. Their salaries would be substantially above those who get to teach in rich neighborhoods.

Sounds like a great plan. Sorry, you could never get it past the teachers union. The real obstacle to ministering to the poor is not the desire to help. Its those who will not face the problem with real solutions. Someone who can turn around lives of young people from broken homes deserves to be paid out of the union scale. The union would not allow it though.

Posted by: John Hansen on January 29, 2006 at 4:29 PM | PERMALINK

Alright, this is one study, so let's not get excited. We need to see the weight of multiple studies to arrive at any conclusions. That said, it's probably true that private schools don't do much better than public schools when you take into account demographics. However, this discussion also overlooks the fact that public schools are typically much more expensive than private schools. Per-capita student costs at public schools is roughly 11,000 bucks. Compare that to per capita costs at private schools: easily half that.

This is from Herbert Walberg, a professor of education at the University of Illinois:

Paul Peterson and I investigated Catholic and public schools in three New York City boroughs: Brooklyn, Manhattan, and the Bronx. To make costs comparable, we subtracted the costs of government-funded special programs from each public schools expenditures, including compensatory programs for children in poverty, bilingual education for children with limited proficiency in English, and special programs for various categories of special needs such as learning disabilities and mental retardation. Costs of transportation and food services were also subtracted from public school outlays. We deducted the public school costs of the central office and the community school boards that oversee and regulate public schools. With these adjustments, Catholic schools costs per student were 46.8 percent that of public schools. Even so, Catholic school achievement in reading and mathematics exceeded achievement in public schools among students in high, middle, and low ranges of poverty. Most striking, however, was that the adverse poverty effect was substantially diminished in Catholic schools.
So, let me get this straight: private schools, at leat the one's studies here, cost roughly half as much as public schools and produce roughly the same results. Sounds like a bargin to me.

Posted by: Sam on January 29, 2006 at 4:39 PM | PERMALINK

Tango Man, thanks again.


For better mastery of the statistical details, I recommend 3 books by S. R. Searle, beginning with Linear Models, 1971. All are published by Wiley

Posted by: contentious on January 29, 2006 at 4:42 PM | PERMALINK

joel, what do you want documented? That university faculty are mostly Democrats and that the social science faculty are especially so? Or that Democrats are as liable to bias as Republicans?

If you are sceptical of summary reports from the Cato Institute, as you ought to be, then you ought to be sceptical of summary reports from university social science faculty.

Posted by: contentious on January 29, 2006 at 4:51 PM | PERMALINK

In my district, the teachers union is hardly the all-powerful-greedy-obstructionist organization that Hansen and the WSJ, for example, seem to think is to blame for all the problems in public education. In fact, they aren't very effective at all, since the working conditions for teachers around here are getting worse all the time. The problem I've seen is not an inablility to fire really bad teachers but an inability to keep the really good ones.

In my experience as a parent, the two really terrible teachers my kid had were both at private schools. Of the two exceptionally good ones, one was at a private school, but since then has moved to a public school for the higher salary. The other was at a public school and has since left the teaching profession.

Posted by: motherbear on January 29, 2006 at 5:07 PM | PERMALINK

The way they got those results was to "adjust" for for things like income, race, home environment

If you are free to make up any "adjustments" you want to data, you can make the statistics say anything you want.

http://donsingleton.blogspot.com/2006/01/public-school-students-score-well-in.html

Posted by: Don Singleton on January 29, 2006 at 5:17 PM | PERMALINK

Tango Man, I don't want to trash the whole study, which is pretty good. But consider this:

An offer of the experimental treatment led to more school district
changes and to attending schools that ranked somewhat higher on state exams, that had lower
percentages of students who were free lunch eligible (i.e., from families with incomes below 130
percent of the poverty level), and that had fewer minority students. The average change in
school performance on state exams was fairly modest, with the schools of the experimental
group ranking on average at the 19th percentile compared to the 15th percentile for the control
group.

The actual change in neighborhoods of those who did change was slight: from a low-ranked public shcool to a public school not quite ranked as low. The effect on school achievement was correspondingly small but in the anticipated direction. this is approximately as strong an effect as the heart-attack prevention effect of aspirin, but is being tested in a much smaller sample. Only if you already have a strong bias against the idea that the quality of schools matters can you take this as not showing the effect of the school.

Of course, it needs to be replicated many times in order for any conclusions to be drawn. Then, again following the aspirin analogy, a meta-analysis might permit a strong conclusion.

Posted by: contentious on January 29, 2006 at 5:33 PM | PERMALINK

"As has been said earlier on this blog, since you can't let the below-average students fall through the cracks in a public school, you can't give the smarter students the education appropriate to THEIR level, harming them. Hence if you are smart and wealthy enough, you go to a private school where they are generally better equipped to do that."

Bingo.

You said much more concisely what I was attempting to say above.

I would add too that contrary to the widespread belief among some that public schools are in some sense the glue that holds the social fabric together public schools in this country (like those in other pluralistic countries) do no such thing. In homogenous countries like Finland and Japan, even the elites send their kids to public schools; that doesn't happen in America, Britain, Canada, the Netherlands, etc. As such, they have always been bottom-line institutions here, designed less to nurture creativity and critical thinking (including if not especially among the best and brightest) but to train good workers and consumers. Public schooling is by design an illiberal thing in America, even if liberals today are its primary defenders.

Posted by: The Blue Nomad on January 29, 2006 at 5:37 PM | PERMALINK

The way they got those results was to "adjust" for for things like income, race, home environment

If you are free to make up any "adjustments" you want to data, you can make the statistics say anything you want.

http://donsingleton.blogspot.com/2006/01/public-school-students-score-well-in.html

Posted by: Don Singleton on January 29, 2006 at 5:38 PM | PERMALINK

Sure, especially "home environment." That's classic mushy social science.

Posted by: The Blue Nomad on January 29, 2006 at 6:03 PM | PERMALINK

contentious,

I thought you might find the study interesting in light of your comment:

. . kids transferred from public to private schools their learning would decrease in line with the reduced quality of the schools.

This study was just one of many which investigate the influence of various environmental factors on student performance and they all point to minor effects. The aggregate effects of a slew of environmental factors don't amount to a lot of influence thus lending support to the position that it is the students themselves that are most responsible for their success. This study moved students and their families to a higher SES environment, and thus the schools were a little better, the teachers were better paid, the fellow students had less noxious attitudes, etc and yet this environmental change didn't produce remarkable results in bringing the transplanted students up to the standards of the other students in that environment.

Posted by: TangoMan on January 29, 2006 at 6:19 PM | PERMALINK

AND THE REASON NOT TO TRY VOUCHERS?

I've read a lot of dire predictions but they all amount to fear and desperation for this simple reason; it's worth a try and any rational person should realize that vouchers might provide the solution through both competition and choice. If the dire predictions here mentioned manifest, we can end the program. But the only reason not to try that I've read is . . . it might succeed. This is where so-called "socialists" show their true colors; it appears the "just society" business was a lot of talk - this union-fetishism is a political smash and grab job and an amateur one at that. I wouldn't care but your not "grabbing" jewels or watches; you're grabbing our children's future. Shame.

TOH

Posted by: The Objective Historian on January 29, 2006 at 6:43 PM | PERMALINK

I quite agree with poster #3 that this reveals a pleasantly high bottom for public schools, not a pleasantly high top.

Posted by: Michael Andersen on January 29, 2006 at 7:04 PM | PERMALINK

Tango Man, the line that you quoted from me was supposed to be parody of the paraphrase of the study conclusion that Kevin Drum wrote, where he concluded that, controlled for other factors, public schools might be better than private schools.


. kids transferred from public to private schools their learning would decrease in line with the reduced quality of the schools.

Obviously, that isn't what anybody expects.

yet this environmental change didn't produce remarkable results in bringing the transplanted students up to the standards of the other students in that environment. As I wrote, a slight change produced a slight improvement. Next on the reading list: Cohen, Statistical Power Analysis for the Behavioral Sciences. As an aside, why is an average improvement from the 15th to the 19th percentile not remarkable? Nobody moved to actual "good" schools.

An analogy: apply a tiny amount of fertilizer on a plot and you'll obtain a tiny enhancement of growth. It's hardly incompatible with the idea that appropriate fertilizer use would have produced "remarkable" results. It's a study with low statistical power.

On another topic, what did you think of my claim that the better teachers migrate to the better schools? Every year there are openings in the better schools, and teachers from all over apply to them; the schools choose the better applicants in the pool. Everybody knows what the best schools are and who the best teachers are; they just pretend not to. Very poor teachers are not fired; they are redistributed away from the schools where the parents complain the most to the schools where the parents complain the least (or least effectively.)

Posted by: contentious on January 29, 2006 at 7:05 PM | PERMALINK

Nobody moved to actual "good" schools.

I think that many people would label Princeton High School, Princeton, NJ, to be a pretty good school with pretty good teachers, in a pretty good neighborhood not plagued with gangs, rampant violence and drugs, yet despite these positive environmental factors:

Last month, the school was cited for the second year in a row, this time because 37 percent of black students failed to meet standards in English, and 55 percent of blacks and 40 percent of Hispanics failed in math.

[ . . . ]

The high school sends 94 percent of its graduates to four-year colleges and offers 29 different Advanced Placement courses. Over all, 98 percent of Princeton High School students exceed the math and English standards required by No Child Left Behind.

[ . . . ]

"If the gap can't be narrowed in Princeton," she said in an interview in her office last week, "then where can it be narrowed? There can't be a question here of resources, or of community support, or of quality of staff. So if we can't impact the students who are not born into privilege, then where can it happen?"

On another topic, what did you think of my claim that the better teachers migrate to the better schools?

I think that there's merit to the claim but we need to recognize that this mobility isn't perfectly elastic in that there are often barriers to mobility. I haven't seen any studies which purport to show that teachers have a hugely significant effect on student achievement. Sure, they can do a lot in creating an atmosphere conducive to learning but they are not the limiting factor. Putting experienced teachers in underperforming classrooms isn't going to be the miracle that many people hope it will be. Putting effective disciplinarians in charge of those classrooms will likely yield better results. A teacher who can garner respect and thus create a supportive environment for learning will, I submit, have far more impact than a teacher who is innovative in technique but a pushover in classroom management. This very dynamic was dramtically captured by a Lehrer Newshour report last month in which a hotshot prinicpal from the suburbs was parachuted into a troubled school and within 6 weeks abandoned all of the progressive ideas that had served him so well in the past. He was mugged by reality. If you're curious about that story I'll try to dig up the PBS link.

Posted by: TangoMan on January 29, 2006 at 7:37 PM | PERMALINK

I was referring to the study. The NYTimes article is a different experiment. I wouldn't claim that the improved schools can narrow the gap completely, only that they can have an improving effect. that was what was shown in the pdf that you posted.

I think we're done here. We'll probably meet again on another thread.

Posted by: contentious on January 29, 2006 at 10:10 PM | PERMALINK

Contentious posts:

"joel, what do you want documented? That university faculty are mostly Democrats and that the social science faculty are especially so? Or that Democrats are as liable to bias as Republicans?"


Neither one. I'm sorry this is so hard for you. You wrote "" . . . such summaries are always biased toward Democratic party objectives . . . This is quite well documented . . . ." If it is, as you asserted, quite well documented that such summaries are always biased toward Democratic party objectives, then lets see some of this alleged documentation. Not documentation of political affiliations of faculty. Not political bias of Democrats or Republicans. Just answer the question I asked, without trying to change the subject.

"If you are sceptical of summary reports from the Cato Institute, as you ought to be, then you ought to be sceptical of summary reports from university social science faculty."

This is just stupid. The Cato Institute is a libertarian think tank. If you believe that University faculty are to the Democratic party what Cato Institute fellows are to the libertarian party, you've been drinking too much of the Koolaid for me to help you now.

Posted by: Joel on January 29, 2006 at 10:45 PM | PERMALINK

Skeptical. I apologize, I just couldn't take it any more. It's "skeptical" with a 'k'.

Posted by: spellcheck on January 29, 2006 at 11:30 PM | PERMALINK

So, let me get this straight: private schools, at leat the one's studies here, cost roughly half as much as public schools and produce roughly the same results. Sounds like a bargin to me.

Public schools have to take everybody and private schools can pick and choose.

Private schools are cheaper because they don't have to deal with the problem students (learning disabilities, discipline, language problems, etc).

Posted by: Stephen on January 30, 2006 at 9:20 AM | PERMALINK

The way they got those results was to "adjust" for for things like income, race, home environment

No, its to control for them. That's a fairly rigorously defined statistical procedure which is the only way to get useful information about the effect of a particular difference seperate from other known sources of variation. Anything that doesn't control for already established sources of variation -- which, unfortunately, is quite a bit of the "statistics" that get reported in newspapers -- is pretty much entirely worthless.

Posted by: cmdicely on January 30, 2006 at 10:00 AM | PERMALINK

Stephen,

Private schools are cheaper because they don't have to deal with the problem students (learning disabilities, discipline, language problems, etc).

If you had read the whole excerpt, you would see that the study I quoted controlled for the costs of these problems, at least to some extent. The private schools studied still had less than half the per pupil costs than public schools in those areas.

It may be true that private schools can be more selective than public schools, but I fail to see why this could fully explain how private schools are so much more productive than public ones. Caroline Hoxby, a professor of economics at Harvard, attempted to calculate the productivity of private and public schools in areas that have tried school vouchers. She found that private schools were nearly 250 percent more productive than public schools even when controlling for student backgrounds.

Posted by: Sam on January 30, 2006 at 10:12 AM | PERMALINK

I thought private schools usually attract problem students.

Posted by: aaron on January 30, 2006 at 10:49 AM | PERMALINK

I think your comment about the two students (one in public, one in private) is flawed. The largest advantage of a private school is the student body (generally more driven to do well in school). The real point of this study as I see it isn't that the real difference between public and private schools isn't the teachers or administration, but the students. Surrounding students with other students who care about school tends to help everyone.

That at least was my experience with a private high school.

Posted by: ibsailn on January 30, 2006 at 11:49 AM | PERMALINK

The quality of private schools vary. Our experience with Catholic school showed that the socio-economic composition of the parish (whose members have first dibs of enrollment) matters just as much as the socio-economic composition of public school boundries.

In our case, the school was better than public school in some ways, but due to the limitation of the student body was not as strong as the best public schools in the same geographic area, two of which were pretty much a match academically of any private school.

The greatest single factor in student success in any school is what kind of a family life he or she has. Bright kids from crappy families (i.e. no support for academic achievement or serious family problems) will flounder in almost all schools. A school, no matter how well funded, can only do so much for its students. And schools with lots of "problem kids" (read: families with problems) will perform the poorly as a whole.

Posted by: Jeff II on January 30, 2006 at 1:20 PM | PERMALINK

This doesn't surprise me at all. Public schools have their performance measured by standardised tests such as NEAP, so they teach to them. Private schools have their performance measured by how many students they get into top Universities, so they teach to that. Of course a private school education doesn't raise your NEAP score - it isn't supposed to.

I have blogged about this based on my own experience of an expensive private education.

Posted by: Jonathan on January 30, 2006 at 6:00 PM | PERMALINK




 

 

Read Jonathan Rowe remembrance and articles
Email Newsletter icon, E-mail Newsletter icon, Email List icon, E-mail List icon Sign up for Free News & Updates

Advertise in WM



buy from Amazon and
support the Monthly