The myth that there's no evidence that school choice works has more lives than Dracula. Worse, it's often repeated by people who should know better, like the education wonks at Third Way or the ranking Democrat on the U.S. Senate education committee. In a particularly egregious recent example, a professor of educational leadership and the dean of the University of Wisconsin-Madison School of Education wrote an op-ed repeating the "no evidence" canard, among others:
The committee also expands the statewide voucher program. There is no evidence privatization [sic] results in better outcomes for kids. The result will be to pay the tuition for students who currently attend private school and who will continue to attend private school—their tuition will become the taxpayers’ bill rather than a private one. Additionally, the funds for the expansion would siphon an estimated $48 million away from public schools, decreasing the amount of money available for each and every school district in the state.
It is astounding that a professor and a dean at a school of education in Wisconsin would be unfamiliar with the research on the Milwaukee voucher program, never mind the numerous gold standard studies on school choice programs elsewhere. Fortunately, Professor James Shuls of the University of Missouri-St. Louis and Martin Lueken of the Wisconsin Institute for Law & Liberty set the record straight:Read the rest of this post »
There are many good reasons to oppose a federal school voucher program, but a supposed lack of evidence that school choice improves student outcomes isn't one of them. Sadly, Sen. Patty Murray (D-WA), the ranking minority member of the U.S. Senate's education committee, repeated this canard during the debates over a proposed amendment that would have added a federal school voucher program to the No Child Left Behind replacement bill:
What’s more, studies of voucher programs in Milwaukee and the District of Columbia have shown that they do not improve students’ academic achievements, she said. “Study after study has shown that vouchers do not pay off for students or taxpayers,” Murray said.
That's simply not true. According to Dr. Patrick Wolf, coauthor of the only longitudinal study of the effect of Milwaukee's voucher program, “school choice in Milwaukee has had a modest but clearly positive effect on student outcomes.”
First, students participating in the Milwaukee Parental Choice (“voucher”) Program graduated from high school and both enrolled and persisted in four-year colleges at rates that were four to seven percentage points higher than a carefully matched set of students in Milwaukee Public Schools. Using the most conservative 4% voucher advantage from our study, that means that the 801 students in ninth grade in the voucher program in 2006 included 32 extra graduates who wouldn’t have completed high school and gone to college if they had instead been required to attend MPS.
Second, the addition of a high-stakes accountability testing requirement to the voucher program in 2010 resulted in a solid increase in voucher student test scores, leaving the voucher students with significantly higher achievement gains in reading than their matched MPS peers.
In the final year of the study, Milwaukee voucher students in grades 3-9 performed about 15 percent of a standard deviation higher on standardized reading tests, “a modest but meaningful educational difference.” Moreover, the study concluded that Milwaukee district-school students were “performing at somewhat higher levels as a result of competitive pressure from the school voucher program.” And contrary to Sen. Murray's assertion that "vouchers do not pay off for taxpayers," the study found that the voucher program saved the state nearly $52 million in fiscal year 2011 because the vouchers were worth about half of the cost per-pupil at the district schools.
Earlier today, the Colorado Supreme Court ruled that Douglas County’s school voucher program violates the state constitution.
The Douglas County Board of Education unanimously voted to enact the Choice Scholarship Pilot (CSP) Program in 2011, making it the first district‐level school voucher program in the nation. The program granted 500 school vouchers worth up to 75 percent of the district schools’ per‐pupil revenue, which was approximately $6,100 in the last academic year. Students could use the $4,575 vouchers at the private school of their choice and the district retained the remaining 25 percent of the funding ($1,525 per voucher student).
However, the ACLU, Americans United for Separation of Church and State, and several local organizations that wanted to protect district schools from competition filed a legal challenge almost immediately. Although they won an injunction from a trial court, it was later overturned on appeal in 2013. Plaintiffs then appealed to the state supreme court.
In a narrow 4–3 decision*, the Colorado Supreme Court held that the voucher law ran afoul of the state constitution’s historically anti‐Catholic Blaine Amendment, which says:
Neither the general assembly, nor any county, city, town, township, school district or other public corporation, shall ever make any appropriation, or pay from any public fund or moneys whatever, anything in aid of any church or sectarian society, or for any sectarian purpose, or to help support or sustain any school, academy, seminary, college, university or other literary or scientific institution, controlled by any church or sectarian denomination whatsoever…
The court held that “aiding religious schools is exactly what the CSP does.” Even though “CSP does not explicitly funnel money directly religious schools, instead providing financial aid to student,” the court ruled that the Blaine Amendment’s prohibitions “are not limited to direct funding.”
According to the American Federation for Children, Sen. Marco Rubio (R‑FL) and Rep. Todd Rokita (R‑IN) have reintroduced “the Educational Opportunities Act, which would create an individual and corporate tax credit for donations that pay for scholarships for students to attend a private school of their parents’ choice.”
It is encouraging to see growing support for scholarship tax credit school choice programs, which have been found to simultaneously boost achievement for students who switch to private schools, do the same for students who remain in public schools, and save taxpayers millions of dollars every year–a win‐win‐win scenario. Nevertheless, it is ill advised to pursue such a program (or other school choice programs) at the federal level.
Years ago I summarized those problems when President George W. Bush advocated creating a federal school voucher program. Such programs are not only beyond the mandate accorded to Congress by the Constitution, they bear the risk of suffocating private schools nationwide with a raft of new regulation, defeating their very purpose of increasing the range of educational options available to families with limited means.
In the past few years I have visited Sweden and Chile and studied their federal school chioce programs. Both confirm my earlier worries about national programs. Chile’s entrepreneurial voucher schools grew rapidly at first, but with a recent change of government hostile to the program they have sensed the new climate and stopped expanding.The new government is trying to enact regulations to diminish the scope and freedom of private schooling in Chile.
Meanwhile, something similar is happening in Sweden. Among other things, the government has mandated that all schools hire graduates of government‐certified teacher training programs, despite the well known fact that those programs are currently attracting the lowest‐achieving college students.
National school choice programs have proven to be a prime case of “staff car legislating.” The legislators who enact them are not always the ones in the official staff cars, making the rules. New lawmakers with different preferences ultimately come to power and can wreak havok on a nation’s entire K‑12 education sector.
This problem can be minimized by leaving school choice legislation to the state level, where the Constitution rightfully leaves it. We thus have a “laboratory of federalism”–a variety of different policies across states that make it easier to determine how best to design such programs.
Just over a week ago, Swedes threw out the relatively pro‐market coalition that had goverened the country for the past 8 years, handing power (though not an outright majority) to a new left‐of‐center coalition. Swedish students’ falling scores on international tests were a key cause of public dissatisfaction, and they have been widely blamed on a nationwide voucher‐like school choice program introduced during the early 1990s. But as I point out in an op‐ed in yesterday’s Svenska Dagbladet, the facts simply don’t support that narrative. Here’s the English draft of the op‐ed:
Sweden’s collapsing performance on international tests was clearly a factor in the recent election, and redressing that slide will be a priority for the new government. A good first step in charting the way forward is to understand what has gone wrong and what has gone right in the past. Unfortunately, the most popular narrative about Swedish education trends is badly mistaken.
Many have blamed Sweden’s falling international test scores on the proliferation of free schools, merely because the decline is thought to have followed their large‐scale expansion. This would be a common logical fallacy even if the timing were correct—but it isn’t.
Between 1995 and 2011, Swedish math scores on the Trends in International Mathematics and Science Study (TIMSS) fell by a massive 56 points. But the vast majority of that decline—41 points—had already taken place by 2003. In that year, 96 percent of Swedish students were still enrolled in government schools.
Another international test, the Programme on International Student Assessment (PISA), began in the year 2000 and has the advantage of breaking out the scores for government and private schools. The last PISA test was administered in 2012, by which time government school scores had fallen by 34 points while free school scores had fallen by only 6 points.
Anders Böhlmark and Mikael Lindahl’s long‐term nationwide study helps to explain these trends: increased local competition from free schools actually raises the performance of students in both sectors—on both national and international tests. But, since free schools still enroll a small fraction of students nationwide, the benefits of this competition have yet to be felt in many areas.
Of course, none of this is to suggest that there are no bad private schools. There has never been an education system in history capable of producing only good schools. The best that can be hoped for is that unsuccessful schools close while good schools expand. And that is precisely what has been happening in Sweden.
Much has been made of the failure of JB Education, which attracted too few students to remain financially viable, and was forced to shut down. This was regrettable for everyone directly concerned, in the short run. In the long run, it is better than any realistic alternative. In most countries, including the United States, atrocious government‐run schools are able to continue operating indefinitely because they face no meaningful competition—the poor parents they most often serve simply cannot afford any alternative. These schools are numerous enough that a term has been coined to describe them: “dropout factories.” Swedish families are lucky that they can far more easily escape such schools.
Not only does the Swedish system pressure failing schools to close, it encourages good ones to expand. International English Schools is one of the highest‐performing school networks in the country, even after controlling for the parental level of education and immigrant background of its students. It is also one of the fastest growing, now operating 25 schools serving nearly 18,000 students. IES has plans to continue growing so long as demand for its services remains unmet. But if IES’s emphasis on academics and civil classroom behavior seems too traditional for some families, there are many other options to choose from. Another large and successful network is Kunskapsskolan, which allows students to proceed through the curriculum at their own pace, combining tremendous student autonomy with weekly one‐on‐one meetings with teachers.
But not all good private schools grow. Specifically, non‐profit schools tend not to build large networks, no matter how good they are. As a result, thousands of students who might benefit from their services never get the chance to do so. The only good schools that consistently “scale‐up” in response to rising demand are those operated as for‐profit enterprises. This is not a coincidence. Building a network is both risky and expensive. The profit‐and‐loss system provides both the resources and the incentives that allow and encourage successful enterprises to grow.
Sweden is fortunate to have harnessed that system to spur the growth of its high performing schools. Chile does the same thing, and has become not only the highest‐performing nation in Latin America but also one of the fastest‐improving countries in the entire world on international tests. If Sweden wishes to become a fast‐improving nation educationally, the evidence strongly supports preserving the entrepreneurial freedoms and incentives that promote the growth of successful education networks.
Slate recently published a badly misinformed piece about Sweden's voucher program, which I addressed here. One of the other responses to the Slate piece was written by Swedish economist Tino Sanandaji for NRO. Sanandaji did an excellent job of showing that the voucher program cannot plausibly explain Sweden’s test score decline and usefully explored some of the more likely causes.
Though I agree with much of what Sanandaji wrote, his piece occasionally endorses heavier regulation of the program for reasons that are either not apparent or inconsistent with the evidence. For instance, he rightly observes that the Swedish government requires universities to accept high school grades as a key admissions criterion but does not permit them to take into account differential grading practices across high schools. This, he notes, puts significant external pressure on high schools to inflate grades. But despite acknowledging this, he later refers to “other problems caused by the [voucher/school choice] reform … such as grade inflation,” implying that this “corruption” is “caused by the lack of [state] control.”
And yet the evidence he presents points to the opposite conclusion: that grade inflation is particularly problematic in Sweden because of imprudent government intrusion into university admissions policy. Consider as a contrast the case of the United States, where universities are free to take high schools’ grading practices into account during admissions. We still have differential grade inflation across high schools, but it is less of a concern because universities can adjust for it. As the head of admissions at Brandeis University has observed, “It’s really not that hard [for colleges] to evaluate a school bearing in mind the differences in grading and weighting processes they employ.” In the absence of government meddling, high schools cannot secure admission to good colleges for their students simply by giving them all A’s.
Still more puzzling is Sanandaji’s criticism that “some private schools broke the rules to cherry-pick students.” This is curious because Sanandaji defends free markets on a number of other occasions, and a hallmark of free markets is that they rely on mutually voluntary exchange. So, naturally, schools in a relatively free marketplace want to enroll students they think they can successfully serve, just as families seek schools they believe can successfully serve them.
This does not mean that all private schools in a relatively free market will seek to serve only high-scoring or well-behaved students. In the United States, where the vast majority of private schools are free to admit students based on any criteria they like, many exist specifically to serve difficult-to-educate students that the typical public school is not well-equipped to teach. A study conducted in the mid-1990s found that public school districts were sending hundreds of thousands of students to the private sector for just that reason. Do some other private schools focus on serving high-performing students? Of course. But the largest share seem to place little or no emphasis on students’ prior academic performance, based on survey data from Arizona that I analyzed several years ago.
There's an old legal proverb about how to win a court case: "If the law is on your side, pound the law. If the facts are on your side, pound the facts. If neither is on your side, pound the table." In this factually-challenged attack on school choice, two lawyers at the UNC Center for Civil Rights do a great deal of table pounding.
Despite mountains of evidence to the contrary, the lawyers charge that school choice programs don't work and that they increase racial segregation. For example, they claim:
...in states with [school choice] programs, student achievement at the private schools is no better, and often worse, than in the public schools. In fact, in Milwaukee and Cleveland, whose voucher programs are the country’s longest running, traditional public school students outperform voucher students on available proficiency measures.
Even read in the most charitable light, the lawyers misleadingly compare apples and orangutans. Participants in school choice programs are generally more disadvantaged than the general population, so it is absurd to compare their average performance against the general population, which includes all the students in wealthy "public" school districts (where low-income parents have been arrested for trying to enroll their kids). Government school advocates rightly object when someone compares average private school performance to average government school performance. The private schools outperform government schools on average, but because both parents and the private schools select each other, the comparison breaks down. The same is true here.
A meaningful comparison requires a randomized-controlled trial, which is the gold standard of social science research because the process of randomization allows researchers to compare like against like and to isolate the effect of the "treatment" (in this case, the offer of a school choice scholarship). Fortunately, there have been 12 such studies addressing this very question from highly-respected institutions like Harvard University and the Brookings Institution. Eleven found that school choice programs lead to positive student outcomes, including higher academic performance and higher rates of high school graduation and college matriculation. One study found no statistically significant difference and none found a negative impact.