Tag: Research

Is School Choice Worth Celebrating? A Look at the Evidence

In honor of School Choice Week, I’ll be answering questions on Facebook tomorrow (4:00pm, Eastern) about the evidence regarding free education markets. When I began studying education policy back in the early 1990s, parent-driven education markets were generally thought of as a new, radical and speculative adventure—uncharted waters where, heaven help us, “thar be monstars.” That was a mistaken view then, and it’s positively absurd now.

As I wrote in Market Education, The Unknown History, the education market of classical Athens, in the 5th century BC, was the first time and place on Earth in which education reached beyond a tiny ruling elite. There was no government participation in education. Teachers competed in the town square to attract paying customers, families called the shots, and the city ended up building a thriving economy and the highest literacy rate in the ancient world. During their heyday, the Athenians invented democracy, most forms of Western literature, and some pretty enduring art and philosophy. Simultaneously, 100 miles away, Sparta established a highly organized system of public boarding schools. It’s legacy? One decent action movie and a name for high school football teams.

Over the next 2,500 years, markets continued to outshine state-run school systems in their ability to serve the needs of families, and they also reduced the social tensions created by state schooling. Near-universal literacy and elementary enrollment among the free population were achieved in the United States by the mid-19th century—before the rise of state school systems—chiefly through private and home schools financed by a combination of parent fees and philanthropy. Even the semi-public “district” schools of the early 19th century charged most parents fees, reserving free and subsidized places for the poor.

Granted, historical evidence is subject to interpretation and charges of selectivity, and so it might not be universally persuasive. But, since 1990, scores of within-country scientific studies have compared education systems ranging from state-run monopolies such as our public schools, to state-funded and regulated private schools, to truly market-like systems in which regulation is minimal and parents choose their schools, as well as paying at least some of the cost directly themselves. I reviewed that body of research a few years ago for the Journal of School Choice and found that it shows private schools tend to outperform state-run schools. More specifically, it shows that the freest and most market-like education systems have the most consistent advantage over state schooling.

There is no credible case against this body of research. I could not find a single study that found a public school system to be more efficient than a market system in terms of student achievement per dollar spent. There weren’t even any insignificant findings for this comparison. Every single study that looked at the efficiency question found statistically significant results favoring education markets over state schooling. It’s rare to see such clear results in the social sciences, but perhaps that’s because there are few areas of life that are still under the thrall of state-run monopolies.

Education markets, when coupled with a mechanism to ensure universal access (such as education tax credits) are a better way to serve our individual needs and to advance our shared ideals. Compulsion and state provision are not only unnecessary, they are counterproductive to our most cherished educational ideals.

Education Policy, The Use of Evidence, and the Fordham Institute

In recent weeks, the Fordham Institute has repeatedly called for government testing and reporting mandates to be imposed on private schools participating in school choice programs (here and here), on the grounds that such “public accountability” improves private school academic outcomes. In defense of this claim, the Fordham Institute cites a study of Milwaukee’s voucher program in which test scores rose following the introduction of such mandates.

Patrick Wolf, director of the research team that conducted the study, has now responded, explaining that his team’s results do not necessarily support Fordham’s claim:

[B]y taking the standardized testing seriously in that final year, the schools simply may have produced a truer measure of student’s actual (better) performance all along, not necessarily a signal that they actually learned a lot more in the one year under the new accountability regime….

What about the encouraging trend that lower-performing schools in the MPCP are being closed down?  [Fordham] mentions that as well and attributes it to the stricter accountability regulations on the program.  That phenomenon of Schumpeterian “creative destruction” pre-dated the accountability changes in the choice program, however, and appears to have been caused mainly by low enrollments in low-performing choice schools, as parents “voted with their feet” against such institutional failure. Sure, the new high-stakes testing and public reporting requirements might accelerate the creative destruction of low-performing choice schools in Milwaukee, but that remains to be seen. [emphasis added]

But there is a deeper problem with the Fordham claim, to which Wolf alludes: a single study, no matter how carefully executed, is not a scientific basis for policy. Because a single study is not science. Science is a process of making and testing falsifiable predictions. It is about patterns of evidence. Bodies of evidence. Fordham offers only a toe.

And Fordham’s preferred policy not only lacks a body of supporting evidence, it is undermined by a large body of evidence. When I reviewed the within-country studies comparing outcomes among different types of school systems worldwide in 2009, I sorted the results into two categories: 1) all studies that compared “public” schools to “private” schools, where those terms were loosely defined; and 2) studies that compared “market” schools to “monopoly” schools. “Market” schools were those paid for at least in part directly by parents and only minimally regulated. “Monopoly” schools were public school systems such as those common in the U.S.

The purpose of these separate categorizations was to see if limited regulation and direct parent funding make a real difference, or if private schools that are paid for entirely by the state and subjected to Fordham’s “public accountability” have the same advantages as their more market-like counterparts.

The result of this breakdown of the literature was stark. Studies looking at truly market-like education systems are twice as consistent in finding a private sector advantage as those looking at “private” schools more broadly construed (and thus including state-funded and regulated private schools).

The pattern of evidence thus seems to contradict Fordham’s belief in the merits of “public accountability” in market education systems. What it favors are policies that promote the rise of minimally regulated education markets in which parents pay at least some of the cost of their own children’s education directly themselves, whenever possible.  That’s just the sort of system likely to arise under education tax credit programs.

School Choice Lowers Crime

New research by Harvard professor David J. Deming studied the crime rates of young adults who participated in a random lottery at the middle or high school level. The lotteries decided whether students were able to attend a school of their choice or whether they were forced to attend their assigned public school. Students who won the lottery committed significantly fewer crimes as young adults than those who lost it. So here is another in the long list of educational outcomes improved by market freedoms and incentives.

Send this to a friend who is still on the fence about the merits of educational freedom.

But Don’t We Really Need Government Research?

It’s a valuable public good, research is, isn’t it? Think of where we’d be without it! I mean, it was government research that came up with the Internet, for heaven sake.

That’s a response to the argument I made last week against government funding of scientific research. Moving away from public funding of scientific research would solve the problem of private companies capturing publication spoils from research that taxpayers funded.

The Defense Advanced Research Projects Agency did indeed come up with and popularize the protocol called TCP/IP, which the Internet uses. (Everyone’s use of the protocol really makes the Internet what it is, of course, but nevermind that.)

To take the Internet as proof that the government is a necessary producer of research and innovation, you have to reject the scientific method. Unfortunately, there are rarely controls in public policy. We can’t find out what would have happened if government policy had taken a different course, so we don’t know anything more about who should fund research from the fact that government-funded research has produced good things in the past.

But what would have happened if U.S. public policy had taken a different course? I’ve thought about the impossible-to-answer question of where we would have been without DARPA and other government influences on telecom. What most people don’t consider, I believe, is the restraining influence the government-granted AT&T monopoly had on telecommunications for most of the 20th century. AT&T developed a “Teletypewriter Exchange” system in 1931, for example, but had no need to develop it, there being little or no competitive pressure to do so. (Its patent on attaching devices to phone wires undoubtedly helped as well, preventing anyone using AT&T’s wires for modem service.)

Had there been competition, I suspect that someone would have come up with the idea of packet-switched networks—that’s what the Internet is—before Leonard Kleinrock did in 1962. Kleinrock was a student at MIT—he wasn’t at DARPA, which didn’t get into packet-switching until about 1966. (Then again, MIT was almost certainly awash in government money—specifically military money—so there you go. Maybe we owe all the good things we’ve got to war, but I doubt it.)

My guess—and it’s only that—is that we would have had the Internet some decades earlier if not for government interventions in telecommunications. We probably would have had multiple, competing “Internets,” actually, adopted more slowly than the Internet we got. (In a chapter of Privacy in America: Interdisciplinary Perspectives, I explored how government has accelerated the development of computing and communications, overpowering society’s capacity to adjust, with negative consequences for privacy.)

Support for government-funded research requires one to elide opportunity costs, the things foregone when one thing is chosen. As I said before, tradeoffs are ineluctable: Money spent on government research takes away from private research, or from other priorities such as reducing debt. In the absence of taxation to support research, the money would go to the public’s priorities as determined directly by the public in manifold spending and investing decision. Taxation and spending on government research is merely the substitution of centralized, political decision-making for a distributed, direct decision-making system. Its supporters are generally going to be beneficiaries of that system—elites, in short.

Even these beneficiaries of the status quo tend to agree that political decisions about funding for scientific research are warped. The solution to that problem, they’ll say, is fixing the political system—that is, creating a political system that is not so political.

Such a breakthrough is as unlikely as the invention of water that is not wet. Perhaps we can put DARPA on both projects.

Open Government Research—-or Maybe Private Ordering

I came across an interesting information policy scuffle yesterday. It’s worth knowing about in general, and I’ll share my liberconoclastic view of things below.

Congressman Darrell Issa (R-CA) has introduced a bill called the Research Works Act. The consensus is that it’s meant to keep government-funded research from being published for free. This would keep the publication of that research going through scholarly and scientific journals, neatly maintaining profits for an industry that society might not need while restricting public access to research the U.S. taxpayer paid for. (I have my doubts that the language of the bill actually successfully does that, but that’s inconsequential.)

Here’s a good opponent-side article on the bill. The Association of American Publishers likes the bill.

On a discussion list, Jonathan Band articulated how the business of government-funded research works. It’s helpful to know if you haven’t focused on this area before:

  1. Federal and state governments, directly or indirectly, pay salaries of researchers.
  2. Federal government awards grants for specific research projects. Average NIH grant is around $500,000.
  3. Researcher performs the research and writes a draft article about it.
  4. Researcher submits the draft article to publisher.
  5. Publisher requires the researcher to transfer the copyright in the draft article (for free) before it will touch the draft.
  6. Publisher emails the draft article to other researchers in the field.
  7. These “peers” review the article for free as part of their contribution to the field. (As noted in step 1, their salaries are paid by government.)
  8. The researcher revises the draft in response to the peers’ comments.
  9. Publisher does copy editing and publishes article. Publishers acknowledge that their costs per article are under $5,000.
  10. Publisher sells subscriptions to research libraries, which ultimately are largely government funded.

“In other words,” Band concludes, “the public invests $500,000 in the creation of the article, and the publisher invests under $5,000. Yet, the publisher recoups all the profits from the sale of the article. Profit margins for STM publishers exceed 40%.”

I’m inclined to share these concerns. It appears to be a classic example of regulatory controls—in this case, on information—creating supra-normal rents for a particular business sector.

My conclusion is a little different, though. You see, to me, what Band describes is a situation where researchers—who nobody is paying their own money to hire—are doing research that nobody is paying their own money to produce, which results in journal articles that nobody is paying their own money to read. Privatized profit from government-funded research is as anathema to me as the next open government advocate, but I would solve the problem by letting private ordering decide where research dollars go.

Is this a retrograde argument against research? Who could possibly be against research? Publicly funded research is like nutritious vegetables for a healthy modern society!

Well, I’m against researchers, research, and research results that nobody pays their own money for because it’s demanded by political actors responding to political cues. I would rather have research dollars meted out through private ordering, because then research dollars would go to where they’re most likely to produce the scientific and intellectual gains society actually wants.

Tradeoffs are ineluctable: Money spent on government research takes away from private research, or from other priorities such as reducing debt, or reducing taxes so I can spend my money on things like donating to charity or to the impoverished individual of my choice.

Does Scholar Self-Interest Corrupt Policy Research?

The New York Times recently ran a story portraying the Gates Foundation as the puppeteer of American education policy, bribing or bullying scholars and politicians into dancing as it desires. Rick Hess, of the American Enterprise Institute, feels that the story misrepresented his position on the potentially corrupting influence of foundations, making it sound as though he were referring to the Gates Foundation in particular when in fact he was referring to the impact of foundations generally.

Hess told the Times, among other things, that

As researchers, we have a reasonable self-preservation instinct. There can be an exquisite carefulness about how we’re going to say anything that could reflect badly on a foundation. We’re all implicated.

Next Monday, the Cato Institute will publish a study titled: “The Other Lottery: Are Philanthropists Backing the Best Charter Schools?” In it, I empirically answer the titular question by comparing the academic performance of California’s charter school networks to the level of grant funding they have received from donors over the past decade. The results tell us how much we should rely on the pairing of philanthropy and charter schools to identify and replicate the best educational models. Considerable care went into the data collection and regression model. As for the description of the findings, it’s as simple and precise as I could make it. I doubt it will be hailed as exquisite.

Science: ‘All Kids Different’

It didn’t get a lot of attention, but in last week’s State of the Union address President Obama celebrated the spread of national curriculum standards that’s been fueled largely by the federal Race to the Top. Of course, he didn’t actually call them “national standards” because no one is supposed to think that these are de facto federal standards that states have been bribed into adopting. The point, though, was clear to those in the know:

Race to the Top is the most meaningful reform of our public schools in a generation. For less than one percent of what we spend on education each year, it has led over 40 states to raise their standards for teaching and learning. These standards were developed, not by Washington, but by Republican and Democratic governors throughout the country.

Despite the celebration of national standards by both the President and lots of other supporters, there is essentially zero evidence that such standards will produce better educational outcomes.  Much of that has to do with the reality of democratically controlled, government education: Those who would be held accountable for getting kids to high standards have the most clout in education politics, and they naturally fight tough standards. It also has a lot to do with human reality: All kids are different. It’s an inescapable observation for anyone who has ever encountered more than one child, but the national-standards crowd prefers to ignore it.

Maybe science will help them see the light. According to the BBC, new research comparing identical and fraternal twins reveals that genetics – something that exists before standards and schooling – has a lot to do with how much and how quickly someone learns:

The researchers examined the test results of 12-year-old twins - identical and fraternal - in English, maths and science.

They found the identical twins, who share their genetic make-up, did more similarly in the tests than the fraternal twins, who share half their genetic make-up.

The report said: “The results were striking, indicating that even when previous achievement and a child’s general cognitive ability are both removed, the residual achievement measure is still significantly influenced by genetic factors.”

In light of this confirmation of the obvious, isn’t it clear that a single timeline for what all children should know and when they should know it makes little sense? And doesn’t it point to the best system being one that gives kids individualized attention?

Of course it does, but that would require “experts” of all stripes to stop trying to impose their solutions on all children. It would also, ultimately, necessitate a system in which parents would choose what’s best for their children, and educators would specialize in all sorts of different curricula, delivery mechanisms, and teaching techniques.  

Unfortunately, few in the education policy world are willing to adopt that utterly logical – but power relinquishing – solution.

Pages