Where’s the Evidence?

September 17, 2009 • Commentary
This article appeared on National Review (Online) on September 17, 2009.

There is very little empirical support for national education standards.

If you listen to advocates of national education standards — from the Obama administration to the conservative Thomas B. Fordham Institute — you’ll seldom hear anything about “evidence” or “research.” You’ll get plenty of assertions about the craziness of having 50 state standards, and how a modern nation must have one bar for all, but zilch about actual, empirical evidence.

So why are these folks — many of whom regularly decry the absence of “scientifically based” policymaking in almost every other facet of education — hawking national standards with nary a whisper about research showing that this monumental reform will actually work? Because there’s hardly any such research to cite. There are very few extant studies comparing educational outcomes in countries with and without national standards — in other words, studies with “treatment” and “control” groups — and what little research does exist is, at best, ambiguous.

Much of this work has been conducted by one scholar: economist John Bishop. Even Bishop hasn’t focused solely on national standards, but rather on standards coupled with curriculum‐​based external examinations (CBEEs), exams that have real ramifications for students (such as effects on course grades and graduation). German economist Ludger Woessmann has also conducted comparative research on national standards. And that’s about it.

Using international assessments such as the Third International Math and Science Study (TIMSS) and the International Assessment of Educational Progress (IAEP), both Bishop and Woessmann report that CBEEs have significant positive effects. But a handful of studies do not a conclusive body of research make, especially when they suffer from considerable deficiencies.

First, as noted, the studies haven’t looked just at national standards, but at standards linked to exams with direct effects on students. Such high‐​stakes examinations are not contemplated under the leading U.S. national‐​standards effort, the Common Core State Standards Initiative.

Much more problematic is that the number of nations participating in international comparisons has been quite small, and only a tiny number of those have not had CBEEs. That has rendered the results easily skewed by unique, outlying nations. Indeed, in 2004, researchers Hendrik Jürges and Kerstin Schneider found that controlling for outliers in Bishop’s sample eliminated any significant positive impact of CBEEs on test scores.

Finally, there’s a serious possibility that even if better achievement is correlated with centralized standards, another variable may be causing both. It’s possible that specific events or cultural predispositions have driven both centralization and a culture that values academic excellence. Suggestive of the latter, Bishop has found that on the TIMSS mathematics exam, being from an East Asian nation carries a positive effect that is nearly three times more powerful than a CBEE. That’s probably why he concluded that to get a better understanding of the impact of CBEEs, we need “studies which hold national culture constant.”

So now we know why national‐​standards aficionados say so little about the scientific support for their favorite reform: There is little scientific support to discuss. Perhaps, though, some of the comparative research on U.S. states could provide strong evidence.

Alas, dear standardizers, it does not. Yes, what studies there are show “treatment” states generally outperforming “control” states, but tiny sample sizes and outliers again plague the research. Telling is a 2002 work by economists Martin Carnoy and Susanna Loeb, who put together an “accountability index” based on how much pressure states put on students and schools to perform. They found that, generally, the greater the pressure, the better the results. But “generally” is the operative word. Charts showing test‐​score changes by state revealed obvious outliers, and many states with very low accountability scores outpaced many states with high scores. And, of course, there was the problem of controlling for variables like cultural differences among states.

Overall, the only conclusive thing that national and state comparisons reveal is not that centralized standards are a magic elixir, but that nations and states are very different. The Philippines is not France. Connecticut is not California.

The same holds for children, which is why it is truly irrational to think that the key to transforming American education is a single, centralized standard for all 50 states. The key is quite the opposite: letting education work like a free market, enabling unique children to attend schools able to specialize in their needs, and instilling competition and innovation everywhere.

Free‐​market education has strong empirical support. Drawing on more than 25 years of research, the Cato Institute’s Andrew Coulson has found that studies showing a significant advantage for free‐​market education outnumber those favoring government monopolies by a ratio of 17 to 1.

National‐​standards research, frankly, isn’t within light years of that, and neither are the simplistic arguments that national‐​standards supporters trot out to fill the empirical vacuum. But when it comes to national standards, apparently science isn’t supposed to matter.

About the Author