Your Child Left Behind

For years, poor performance by students in America relative to those in other countries has been explained away as a consequence of our nationwide diversity. But what if you looked more closely, breaking down our results by state and searching not for an average, but for excellence?

Imagine for a moment that a rich, innovative company is looking to draft the best and brightest high-school grads from across the globe without regard to geography. Let’s say this company’s recruiter has a round-the-world plane ticket and just a few weeks to scout for talent. Where should he go?

Our hypothetical recruiter knows there’s little sense in judging a nation like the United States by comparing it to, say, Finland. This is a big country, after all, and school quality varies dramatically from state to state. What he really wants to know is, should he visit Finland or Florida? Korea or Connecticut? Uruguay or Utah?

Stanford economist Eric Hanushek and two colleagues recently conducted an experiment to answer just such questions, ranking American states and foreign countries side by side. Like our recruiter, they looked specifically at the best and brightest in each place—the kids most likely to get good jobs in the future—using scores on standardized math tests as a proxy for educational achievement.

We’ve known for some time how this story ends nationwide: only 6 percent of U.S. students perform at the advanced-proficiency level in math, a share that lags behind kids in some 30 other countries, from the United Kingdom to Taiwan. But what happens when we break down the results? Do any individual U.S. states wind up near the top?

Incredibly, no. Even if we treat each state as its own country, not a single one makes it into the top dozen contenders on the list. The best performer is Massachusetts, ringing in at No. 17. Minnesota also makes it into the upper-middle tier, followed by Vermont, New Jersey, and Washington. And down it goes from there, all the way to Mississippi, whose students—by this measure at least—might as well be attending school in Thailand or Serbia.

Hanushek, who grew up outside Cleveland and graduated from the Air Force Academy in 1965, has the gentle voice and manner of Mr. Rogers, but he has spent the past 40 years calmly butchering conventional wisdom on education. In study after study, he has demonstrated that our assumptions about what works are almost always wrong. More money does not tend to lead to better results; smaller class sizes do not tend to improve learning. “Historically,” he says, “reporters call me [when] the editor asks, ‘What is the other side of this story?’”

Over the years, as Hanushek has focused more on international comparisons, he has heard a variety of theories as to why U.S. students underperform so egregiously. When he started, the prevailing excuse was that the testing wasn’t fair. Other countries were testing a more select group of students, while we were testing everyone. That is no longer true: due to better sampling techniques and other countries’ decisions to educate more of their citizens, we’re now generally comparing apples to apples.

These days, the theory Hanushek hears most often is what we might call the diversity excuse. When he runs into his neighbors at Palo Alto coffee shops, they lament the condition of public schools overall, but are quick to exempt the schools their own kids attend. “In the litany of excuses, one explanation is always, ‘We’re a very heterogeneous society—all these immigrants are dragging us down. But our kids are doing fine,’” Hanushek says. This latest study was designed, in part, to test the diversity excuse.

To do this, Hanushek, along with Paul Peterson at Harvard and Ludger Woessmann at the University of Munich, looked at the American kids performing at the top of the charts on an international math test. (Math tests are easier to normalize across countries, regardless of language barriers; and math skills tend to better predict future earnings than other skills taught in high school.) Then, to get state-by-state data, they correlated the results of that international test with the results of the National Assessment of Educational Progress exam, which is given to a much larger sample in the U.S. and can be used to draw statewide conclusions.

The international test Hanushek used for this study—the Programme for International Student Assessment, or PISA—is administered every three years to 15-year-olds in about 60 countries. Some experts love this test; others, like Tom Loveless at the Brookings Institution, criticize it as a poor judge of what schools are teaching. But despite his concerns about PISA, Loveless, who has read an advance version of Hanushek’s study, agrees with its primary conclusion. “The United States does not do a good job of educating kids at the top,” he says. “There’s a long-standing attitude that, ‘Well, smart kids can make it on their own. And after all, they’re doing well. So why worry about them?’”

Of course, the fact that no U.S. state does very well compared with other rich nations does not necessarily disprove the diversity excuse: parents in Palo Alto could reasonably infer that California’s poor ranking (in the bottom third, just above Portugal and below Italy) is a function of the state’s large population of poor and/or immigrant children, and does not reflect their own kids’ relatively well-off circumstances.

So Hanushek and his co-authors sliced the data more thinly still. They couldn’t control for income, since students don’t report their parents’ salaries when they take these tests; but they could use reliable proxies. How would our states do if we looked just at the white kids performing at high levels—kids who are not, generally speaking, subject to language barriers or racial discrimination? Or if we looked just at kids with at least one college-educated parent?

As it turned out, even these relatively privileged students do not compete favorably with average students in other well-off countries. On a percentage basis, New York state has fewer high performers among white kids than Poland has among kids overall. In Illinois, the percentage of kids with a college-educated parent who are highly skilled at math is lower than the percentage of such kids among all students in Iceland, France, Estonia, and Sweden.

Parents in Palo Alto will always insist that their kids are the exception, of course. And researchers cannot compare small cities and towns around the globe—not yet, anyway. But Hanushek thinks the study significantly undercuts the diversity excuse. “People will find it quite shocking,” he says, “that even our most-advantaged students are not all that competitive.”

Reading the list, one cannot help but thank God for Massachusetts, which offers the United States some shred of national dignity—a result echoed in other international tests. “If all American fourth- and eighth-grade kids did as well in math and science as they do in Massachusetts,” writes the veteran education author Karin Chenoweth in her 2009 book, How It’s Being Done, “we still wouldn’t be in Singapore’s league but we’d be giving Japan and Chinese Taipei a run for their money.”

Is it because Massachusetts is so white? Or so immigrant-free? Or so rich? Not quite. Massachusetts is indeed slightly whiter and slightly better-off than the U.S. average. But in the late 1990s, it nonetheless lagged behind similar states—such as Connecticut and Maine—in nationwide tests of fourth- and eighth-graders. It was only after a decade of educational reforms that Massachusetts began to rank first in the nation.

What did Massachusetts do? Well, nothing that many countries (and industries) didn’t do a long time ago. For example, Massachusetts made it harder to become a teacher, requiring newcomers to pass a basic literacy test before entering the classroom. (In the first year, more than a third of the new teachers failed the test.) The state also required students to pass a test before graduating from high school—a notion so heretical that it led to protests in which students burned state superintendent David Driscoll in effigy. To help tutor the kids who failed, the state moved money around to the places where it was needed most. “We had a system of standards and held people to it—adults and students,” Driscoll says.

Massachusetts, in other words, began demanding meaningful outcomes from everyone in the school building. Obvious though it may seem, it’s an idea that remains sacrilegious in many U.S. schools, despite the clumsy advances of No Child Left Behind. Instead, we still fixate on inputs—such as how much money we are pouring into the system or how small our class sizes are—and wind up with little to show for it. Since the early 1970s, we’ve doubled the amount of money we spend per pupil nationwide, but our high-schoolers’ reading and math scores have barely budged.

Per student, we now spend more than all but three other countries—Luxembourg, Switzerland, and Norway—on elementary and secondary education. And the list of countries that spend the most, notably, has little in common with the outcomes that Hanushek and his colleagues put into rank order. (The same holds true on the state level, where New York, one of the highest-spending states—it topped the list at $17,000 per pupil in 2008—still comes in behind 15 other states and 30 countries on Hanushek’s list.)

However haltingly, more states are finally beginning to follow the lead of Massachusetts. At least 35 states and the District of Columbia agreed this year to adopt common standards for what kids should know in math and language arts—standards informed in part by what kids in top-performing countries are learning. Still, all of the states, Massachusetts included, have a long way to go. Last year, a study comparing standardized math tests given to third-graders in Massachusetts and Hong Kong found embarrassing disparities. Even at that early age, kids in Hong Kong were being asked more-demanding questions that required more-complex responses.

Meanwhile, a 2010 study of teacher-prep programs in 16 countries found a striking correlation between how well students did on international exams and how their future teachers performed on a math test. In the U.S., researchers tested nearly 3,300 teachers-to-be in 39 states. The results? Our future middle-school math teachers knew about as much math as their peers in Thailand and Oman—and nowhere near what future teachers in Taiwan and Singapore knew. Moreover, the results showed dramatic variation depending on the teacher-training program. Perhaps this should not be surprising: teachers cannot teach what they do not know, and to date, most have not been required to know very much math.

Early last year, President Obama reminded Congress, “The countries that out-teach us today will out-compete us tomorrow.” This September, Ontario Premier Dalton McGuinty, visiting a local school on the first day of classes, mentioned Obama’s warning and smugly took note of the scoreboard: “Well,” he said, “we are out-teaching them today.”

Arne Duncan, Obama’s education secretary, responded to the premier’s trash-talking a few days later. “When I played professional basketball in Australia, that’s the type of quote the coach would post on the bulletin board in the locker room,” he declared during a speech in Toronto. And then his rejoinder came to a crashing halt. “In all seriousness,” Duncan confessed, “Premier McGuinty spoke the truth.”