Monday, August 28, 2006

Charter School Nonsense

Good to see the WSJ editorial page quickly rebutting the nonsense that appeared on the NYT's editorial page yesterday.  The WSJ correctly points out that the study is flawed in many ways and that "The only way to measure school effectiveness is through longitudinal studies that compare the same students over time."  I then expected the editorial to say something like this (from my earlier emails):
Bryan Hassel of Public Impact has looked at 44 such studies that include both snapshot and longitudinal perspectives. Of the 18 snapshot studies, 12 found that charters did worse than non-charters, and 6 found that charters did the same or better. Of the 26 change studies, 16 found that charters did better than non-charters, 6 found roughly comparable results between charters and non-charters, and 4 found that charters did worse. Of the 16 positive change studies, the most common theme was that charter students start out behind and catch up and surpass non-charter students over time.
Unfortunately, however, the WSJ instead makes the same mistake as the study, using snapshot data to defend charter schools:

Studies in California, New York, Massachusetts, Florida and elsewhere have repeatedly shown charter school students outperforming their counterparts in traditional public schools -- sometimes dramatically. In Michigan in 2004, 46% of black eighth-graders in charter schools passed the state math assessment test, compared with just 21% of black eighth-graders statewide.

Writing in the New York Post last week, Peter Murphy of the New York Charter Schools Association said a state report issued this summer found that "in 2005, a majority of charter schools had a higher percentage of students passing the state exams in English language arts and mathematics at the elementary and middle school levels" than did their respective school districts.

This is a minor quibble, however.  Most importantly, its conclusion is exactly right:
All charters aren't successful, but the bad ones tend to close in due course, which is a good thing and more than can be said for failing traditional public schools. As for the rest, they are providing a fast-growing option for underprivileged children. This irks unions, school boards and others with a vested interest in a public school monopoly that's failing to educate millions of kids.

Charter School Nonsense
WSJ editorial, August 28, 2006; Page A12

The Bush administration's desire to enlighten parents and taxpayers about alternatives to failing public schools is admirable. But it'll have to do better than the misleading report issued last week by the federal Department of Education, which purports to show that charter schools trail traditional public schools in student achievement.

The study used 2003 test-score data from the National Assessment for Education Progress (NAEP) to determine that fourth-graders attending regular public schools scored 4.2 points higher in reading, and 4.7 points higher in math, than their counterparts in charter schools, which are also public schools but operate independently and without many of the union and bureaucratic rules. The study claims to have taken family income into consideration, and that's important because a large number of charter schools cater to mostly low-income students.

But the use of federal-lunch program participation as a poverty indicator is problematic and likely skewed the results. According to the Center for Education Reform, a school choice advocacy group, as many as 1-in-4 charters nationwide would qualify for free- and reduced-lunch programs but don't take advantage of them. Cost and bureaucratic red tape are two primary reasons.

For example, to participate in a federal lunch program, a school must hire certified food service workers. Many charters operating on tight budgets choose to use parents or volunteers in the cafeteria and steer their limited resources into the classroom. By relying on a flawed proxy for poverty, the government's methodology penalizes this sort of efficiency.

 Subscribe in a reader