This isn't really political -- I mean, I found the link on DailyKos, and I'm sure the conclusions can be used to further certain political aims, but the paper itself is pure statistical analysis. But as a long-time proponent (and product) of the American public school system, I found this very interesting. Basically, a couple of professors at the University of Illinois analyzed data from the 2003 4th and 8th grade math tests by the type of school (public or private, and the various types of private schools), but adjusted the data to reflect the socio-economic makeup of the students at each school. In other words, they took into account that public schools often have to deal with more disadvantaged (and more severely disadvantaged) students than those at private schools. I'm about the furthest thing from a mathematician, so the computing method was hard for me to analyze, but it seems relatively sound.

Some of the conclusions:
• Public schools significantly out-scored Catholic schools (by over 7 points in 4th grade, and almost 4 points in 8th grade).
• Of all private school types studied, Lutheran schools performed the best. 4th grade scores in Lutheran schools were roughly 4 points lower than in comparable public schools, but were (a statistically insignificant) 1 point higher at the 8th grade.
• The fastest growing segment of the private school sector, conservative Christian schools, were also the lowest performing, trailing public schools by more than 10 points at grades 4 and 8.
• Charter schools scored a significant 4.4 points lower than non-charter public schools in 4th grade, but scored (a statistically insignificant) 2.4 points higher in 8th grade.

The authors are quick to point out faults in their own data (like a lack of complete information for conservative Christian schools, leading to a relatively small sample set compared to the other school types), and admit just one study isn't enough to prove their point. But among the interesting possibilities it raises is that charter schools and school vouchers may be far less effective solutions to our educational problems than they are usually made out to be. Read the full paper (or at least the abstract at the beginning), at http://www.ncspe.org/publications_files/OP111.pdf

This dovetails nicely with a scary story on PBS's NewsHour late last week about a school in the Bronx that is stuck on the "underperforming" list because of the way the No Child Left Behind Act test score criteria is calculated. The principal was almost in tears as he pointed out that, had their scores started far lower than they were when they first went on the list three years ago, the scores that would have resulted in their removal would actually be lower than their test scores currently were. This is a school that has been incredibly successful in a very poor neighborhood with many ESL students (part of the problem, since ESL students are expected to meet the same criteria as non ESL students in 4th grade), and they are basically being punished for the makeup of their student body. Any study that considers socio-economic factors when measuring educational progress can only help them.