Wednesday, March 19, 2008

Fake graduation rates and other predictable outcomes of no child left behind

Bush's educational program was said to be based on good outcomes in Texas. Of course it wasn't so. Turns out schools in Texas were cooking the books to get better numbers. Strong incentives, like "improve or die" reliably produce this kind of result.

The easiest way to cook the books for a particular school is to get the low performing students to move elsewhere. Then to make the overall district look better, don't invest in tracking where they "move" to. They did that in Texas too.

I'm picking on Texas, but the same thing will happen everywhere that these kinds of incentives are applied. It works for physicians too. If you pay us less for patients who don't keep their blood sugars tuned up, you'll find that those patients will "leave". There must be fifty ways to help a patient leave ...

Today the NYT tells the story for Mississippi, but I'm betting Minnesota and Vermont are playing the same game, albeit with more subtlety ...
States’ Data Obscure How Few Finish High School - New York Times
March 20, 2008l

JACKSON, Miss. — When it comes to high school graduation rates, Mississippi keeps two sets of books.

One team of statisticians working at the state education headquarters here recently calculated the official graduation rate at a respectable 87 percent, which Mississippi reported to Washington. But in another office piled with computer printouts, a second team of number crunchers came up with a different rate: a more sobering 63 percent...

...“We were losing about 13,000 dropouts a year, but publishing reports that said we had graduation rate percentages in the mid-80s,” Mr. Bounds said. “Mathematically, that just doesn’t work out.”

... federal figures obscure a dropout epidemic so severe that only about 70 percent of the one million American students who start ninth grade each year graduate four years later.

California, for example, sends to Washington an official graduation rate of 83 percent but reports an estimated 67 percent on a state Web site.

... New Mexico defined its rate as the percentage of enrolled 12th graders who received a diploma. That method grossly undercounts dropouts by ignoring all students who leave before the 12th grade.

The law also allowed states to establish their own goals for improving graduation rates. Many set them low. Nevada, for instance, pledged to get just 50 percent of its students to graduate on time. And since the law required no annual measures of progress, California proposed that even a one-tenth of 1 percent annual improvement in its graduation rate should suffice.

.. Most troublesome to some experts was the way the No Child law’s mandate to bring students to proficiency on tests, coupled with its lack of a requirement that they graduate, created a perverse incentive to push students to drop out. If low-achieving students leave school early, a school’s performance can rise...

... In Mississippi, the official formula put the graduation rate for the state’s largest district, Jackson Public Schools, at 81 percent. Mr. Bounds, the state schools superintendent, said the true rate was 56 percent.

At Murrah High School, one of eight here, the official graduation rate is 99 percent, even though yearbooks show that half of Murrah’s freshmen disappear before becoming seniors...
The obvious story here is that you get what you pay for. There will always be a way to game the system though, which is why you can't replace professional culture with incentives, just as you can't create a civil society through police action. Obviously both incentives and policing can be pretty important, but they can't replace professional pride and culture or a basic culture of civil behavior.

The less obvious story is that about 30% of Americans don't complete High School.

So I'd like to know why so many don't finish High School, but I'd first like to know what the "optimal" graduation rate should be. That's the question that leads to the most interesting and important discussions.

No comments: