The Great, Or Actually Not-So-Great, Graduation Rate Debate

Now I know how all those suckers who pay good money to watch a heavyweight bout that ends in the first round feel…or perhaps more likely how you’d feel if you’d just watched George Foreman fight Little Orphan Annie.

For the past few months the head of the Economic Policy Institute, Lawrence Mishel, has been touting a graduation rate doomsday weapon that would discredit the existing consensus about grad rates and make everyone reconsider the issue. Regular readers will recall that last we heard from Monsieur Mishel he was hawking an EPI produced booklet purporting to objectively examine the whole AFT-charter school episode without disclosing that the AFT pays part of his salary. Teachers College Press got a little egg on their face for publishing it without disclosing the funding issue.

A week or two back Mishel finally took on graduation rate experts Jay Greene, Chris Swanson, Gary Orfield*, etc…in Ed Week. But the debate is embarrassingly one-sided, decide for yourself: Here’s Mishel’s salvo and here is a response from Greene, his colleague Greg Winters, and Chris Swanson, formerly of the Urban Institute now at Ed Week. Worth fully reading both. Perhaps the actual Mishel paper, which Greene, Winters, and Swanson say they could not get their hands on, will be more impressive and we’ll revisit accordingly if that’s the case (and for the record, I would love it if the grad rate numbers were better than we think).

Right now, the bottom line is that to buy what Mishel is selling this time you have to believe that all these folks, not just favorite ed establishment villain Greene (a) clumsily overlooked this “gold standard” of data that Mishel found (b) didn’t understand it and/or (c) are all out to unfairly castigate the public schools.

The debate is more about data sets than methods: Mishel uses the Current Population Survey and National Education Longitudinal Study data while Greene, Swanson, etc…use Common Core of Data (CCD) from the Department of Education. A second is a subsidiary methods issue, Greene etc…make calculations based on the CCD using their own methodology (there are some differences between Greene’s method and Swanson’s method) while Mishel uses the reported results in these other datasets with some adjustments for GEDs. And, that’s a third issue with this debate more generally: There is definition of terms confusion not only in terms of different ways of looking at the dropout issue but also about on-time graduation versus completion. The rates of high school completion one gets depend on whether GEDs are included, in other words students who don’t finish in four years but do subsequently earn a degree. Again, read both Ed Week pieces, too much to excerpt here.

It’s worth noting is that even though the Greene method, Swanson method (pdf), and John Robert Warren’s method (pdf) vary, they all land roughly around the same numbers providing some convergent validity in terms making public policy (though there is a debate about the differences among them and other methods). Put another way, despite Mishel’s apparent efforts to say there is less of a disaster going on than it might appear, there is a very serious dropout problem. As states develop better data systems to track students this methods debate will become less important though new issues will invariably arise. But for now these methods are the best at hand. That’s one reason why the National Governor’s Association adopted a compact (pdf) incorporating them. In other words, barring some new evidence at the 11th hour, this debate is largely over.

*Mishel (and the AFTies who predictably jumped all over this) conveniently forget to note that Harvard Civil Rights Project co-director Gary Orfield is right there with Greene, Swanson, et al…kinda makes you think this might be a little political…Orfield complicates the villainous conservatives out to ruin the public schools storyline too much. Funny side note: In their excitement the AFTies even confused Paul Peterson and Jay Greene (all those bad men look alike to us!).

Leave a Reply

Your email address will not be published.