Former Bush White House adviser (and NCLB drafter) Sandy Kress turned in a very compelling New York Daily News op-ed on Monday arguing that President Obama has gone “wobbly” on education accountability.

In the piece, Kress presented impressive NAEP data illustrating the big gains that minority and special needs students have made since the late 1990s.

What has caused these and other similar gains? Most researchers say the biggest factor was that in the late 1990s, states began to implement policies holding schools accountable for improving education for children. Further, in 2001, the Congress extended those policies to schools in all states through the bipartisan No Child Left Behind Act.

Today, if schools shortchange students, especially subgroups of disadvantaged students, improvement in the operation of the school is required. Student problems can no longer be swept under the rug. Because of “consequential accountability,” business as usual is no longer acceptable….

Now, here’s the second big secret: For all of its promise to bring about education reform early in the term, the Obama administration wants to turn back the clock on accountability…

Under the framework being proposed for the reform of the law, the administration would require that, unless a school is among the very worst in the nation, it would no longer be required to improve even if it continues to fail its black, Hispanic and other disadvantaged kids. Further, in the case of schools that do not improve, special tutoring and public school choice would no longer be required.

This is tough stuff. Kress is implying that if we back away from NCLB’s federally-mandated “consequential accountability” (and “subgroup accountability”) for the vast majority of American schools (which is indeed what the Administration’s blueprint proposes–and where Fordham comes down, too), the progress our neediest kids are making will stop or slide backwards.

But is he right? Are our kids on a rapid upward trajectory because of No Child Left Behind’s unique approach? Let’s take a closer look at his claims:

Did you know that in 2008, black 9-year-olds were reading two grade levels ahead of where they were in 1999? According to the National Assessment of Educational Progress, this is true.

Did you know that in 2008, Hispanic 9-year-olds were handling math problems two grade levels ahead of where they were in 1999? This is also true.

Did you know that in 2009, fourth-grade students with disabilities* were reading almost two grade levels above where they were in 2000? Again, this is true….

These claims are true as far as they go. For instance, according to the NAEP, the average reading score for Black 9-year olds rose from 186 in 1999 to 204 in 2008–an increase of 18 points. (At 10 points per grade level that comes close enough to the “two grade levels” of progress Kress claims.) Hispanic 9-year olds increased their average reading scores from 213 in 1999 to 234 in 2008–an increase of 21 points. Fourth-grade students with disabilities increased their reading scores from 167 in 2000 to 189 in 2009.

But here’s the kicker: almost all of these gains had occurred by 2004. For Black 9-year-olds, 78 percent of the improvement took place in the five years between 1999 and 2004, compared to 22 percent in the four years between 2004 and 2008. For Hispanic students, 81 percent of the gains occurred between 1999 and 2004, compared to 19 percent between 2004 and 2008. For fourth-grade students with disabilities, 91 percent of the gains occurred in just two years: between 2000 and 2002.

In fact, this tremendous improvement in the late 1990s and early 2000s is one of the great mysteries of education policy. Nobody knows for sure why it happened. As Kress indicates, Eric Hanushek and others have found plausible evidence to credit accountability-based reforms. But it’s impossible to know whether it was “accountability” in general–or NCLB in specific–that drove the scores upward. No Child Left Behind was enacted in January 2002 and started to be implemented that fall; students had at most a year and a half of the NCLB “treatment” before sitting for the 2004 NAEP. So it’s hard to argue that NCLB gets a lot of credit for these improvements.

This is particularly true when looking at the scores for students with disabilities. No Child Left Behind couldn’t have been a factor in their incredible progress from 2000-2002. And in fact, students with disabilities have made almost no gains in reading since 2002–even though NCLB focuses specifically on boosting the achievement of this subgroup of students.

What this says to me is that accountability in general resulted in a significant boost in student achievement in the late 1990s/early 2000s. And it’s likely that most of those gains occurred before NCLB was implemented. So it’s hard to argue, as Kress does, that it was NCLB’s specific approach to accountability that gets the credit for the progress–or is essential to making more progress today. There’s an urban myth that “going back to the days before NCLB” would necessarily be a bad thing. But the days right before NCLB were actually very good!

Furthermore, there is evidence (in our stagnating student achievement results) that the effects from accountability are fading.

So if Kress and others want to use these data to say: “Don’t mess with NCLB’s accountability framework because it’s working”–well, to me, the evidence shows that NCLB and test-based accountability had their day in the sun, and made a big difference, but now it’s time to try something else if we want to see progress continue.

* Kress’s Daily News op-ed inadvertently had this as eighth graders. He is working to fix that online.

-Mike Petrilli