More About Oklahoma and Pre-k Evidence
Cato's Adam Schaeffer, responding to a post I wrote two week ago, has more--lots more--to say about pre-k effectiveness (or, from his point of view, the lack thereof). Before we start talking about the evidence on pre-k more generally, though—which is the real bone of contention here, right?—let’s close out the debate that started this: Do trends in Oklahoma’s NAEP scores since the early 1990s indicate that the state’s pre-k program is ineffective? My answer is still “no.”
It is true that, in the early 1990s, Oklahoma 4th graders outperformed the national average on the NAEP in reading and math. But today they’re performing below the national average. Does that mean pre-k does work? No: Between 1998 and 2002, Oklahoma’s 4th grade reading NAEP scores plummeted. That’s also when the national average NAEP score in 4th grade math passed Oklahoma’s.
Now, I would like to know what the heck happened in Oklahoma between 1998 and 2002 (unfortunately, this was before annual NAEP testing, so we have no data points between 1998 and 2002)—but it sure as heck didn’t have anything to do with universal pre-k. Why? Because Oklahoma didn’t start universal pre-k until 1998, so the first kids to benefit from pre-k weren’t even in 4th grade yet when Oklahoma was falling behind.
(Remember: Children enrolled in pre-k in one year won't be in fourth grade until 5 years later, so any score impacts from enrollment increases would have a 5 year time lag.)
Since 2003, Oklahoma 4th graders’ achievement in reading and math has been on the upswing, and it appears to be closing the gap with national averages in math. One could just as easily use this data to argue that universal pre-k arrested Oklahoma’s late-1990s decline in 4th grade reading achievement. For reasons discussed, at length, by both me and Adam, trying to use NAEP data to make an argument like that would be silly—which shows why Schaeffer’s attempt to use Oklahoma’s NAEP data to show pre-k isn’t effective is equally silly.*
What we do know is that Oklahoma’s pre-k program is effective in improving the math and literacy skills with which participants enter kindergarten—and that it’s doing so to a greater extent than Head Start does. Given how difficult it is for any educational intervention to produce evidence of student learning gains, that's a pretty impressive finding. The far more important question, however—and this is one area where Schaeffer and I do agree—is whether those abundantly documented short-term learning gains translate into long-term benefits. And that’s a topic for another post, to come shortly.
*There's plenty of irony in Schaeffer's efforts to make any argument at all based on NAEP data. Cato is on record opposing national standards and government testing requirements in general—if Schaeffer and his colleagues had their way, the data that sparked this debate wouldn’t even exist to begin with!