Medical researchers go to a lot of trouble to test a new drug. They record exactly what they're administering, how often, and in what quantity. They solicit volunteers and randomly give the drug to some but not others. Thanks to decades of these randomized experiments, "House," "Doc Martin," and even your local GP have at least a clue as to what works and what doesn't.
As I've just argued elsewhere, most education policy advocacy is quackery by comparison. Analysts routinely claim to evaluate one policy by looking at evidence from another. When they do present relevant evidence, it is often inaccurate, incomplete, and misleading. Many education policy analysts either do not understand or do not care what constitutes meaningful evidence.
If they worked in the field of medicine, you would not let these people within fifty yards of your children, but they've been shaping the way children are taught for over a century. Education has suffered as a result. Despite a near tripling in the inflation-adjusted per pupil cost of a K-through-12 education, graduation rates are lower today than they were two generations ago, and students seem no better prepared academically.
So what's the alternative? Can you reject the quackery and demand the same quality of research from the education policy community that you do from the medical community? The answer, to a surprising degree, is yes... but there are some interesting complications.
The greatest challenge is that there is so little variation in education policy within the United States that our ability to evaluate alternatives is constrained. There are now charter school, voucher, and education tax credit programs in numerous states, but these programs are quite small. Charter schools are the largest, but even they enroll less than five percent of students. To draw firm conclusions we need to see a variety of policies operating on a larger scale.
An obvious solution is to look at the experiences of other nations, but this poses a challenge of its own: how do we know if the outcomes we observe can be attributed to a nation's policies rather than to economic, cultural, or demographic factors?
In principle, we could control for these other factors by mimicking medical experiments, randomly imposing a policy on one set of countries (the "treatment" group), while leaving a second group of countries as-is (the "control"). Not really feasible. Fortunately, medical researchers ran into this difficulty long ago—and found a way around it. Doctors can't impose restricted diets and increased exercise on entire national populations in order to measure the health effects, but they realized that when such changes occur naturallythey can still study the results. These are called "natural experiments" and they exist all over the world and throughout history, not just in medicine but in education.
For instance, many countries have two or three different types of school systems operating side-by-side. By studying the effects of these within-country variations for a large number of nations, and over a vast swath of history, we can isolate the impact of the policies themselves.
Because this approach draws on very large bodies of evidence, the source citations alone for a study of this kind would be many times longer than the present commentary. But while the evidence itself is hard to compress into this space, the findings are not. When we review natural experiments in education policy fromthe 5th century BC to the present, and in dozens of countries in the modern world, clear patterns emerge. It turns out that education is generally most effective, efficient, harmonious, and responsive to families when educators are freed from government regulations, families choose from among a variety of schools, schools vie with one another to attract and serve children, and parents pick up at least some of the cost directly themselves—in essence, a free education marketplace.
But the historical and international evidence also indicates that government funding of private schools tends to bring with it a pall of regulation that grows over time; and schools hamstrung with this red tape underperform those that give educators and families more freedom. Though the regulatory burden is usually heaviest in older and larger programs, it can be seen even in small modern U.S. voucher programs.
The upshot of all this is that vouchers are likely to smother and homogenize the private sector in the long term, causing it to resemble the bureaucratized state-run school system that voucher advocates so ardently wish to reform. Catch 22.
But the news is not all bad. State-level education tax credit programs are another way of broadening access to the kind of education marketplace supported by the historical and international evidence. The early research suggests that they do indeed raise achievement, and improve efficiency as that evidence leads us to expect. But, unlike vouchers, they do not appear to hobble educators with red tape. That does not mean it would be wise to enact education tax credits at the federal level, but it is a path that nine states have already begun to follow and the results so far are promising.
These, at any rate, are conclusions I have drawn in systematically studying scores of school systems from classical Athens and Sparta to modern Chile and America. Much more such work can and should be done. And it might be, if Americans demand the same level of seriousness from the education policy community that they do from the medical community.