Bruce Baker has studied charter enrollments in New Jersey, Pennsylvania, Houston, and New York City.

Matthew DiCarlo observed that the GAO report actually understated the disparity in charter enrollments of students with disabilities, by comparing charters to the nation, instead of to the district where they are located. Urban districts have higher rates of students with disabilities than the national rate.

Bruce Baker notes that some charters inflate their numbers of special education students by taking only those with the mildest disabilities:

A really big issue which I’ve been able to explore only in a few contexts is the breakout of children with disabilities served by charters versus those left behind in public districts. There are cases where it looks like charters are serving comparable total rates of children with disabilities. But, when classification data are available, it almost invariably turns out that the charter schools are serving only (or mostly) those with speech impairment or mild specific learning disabilities. I provide one example here:http://schoolfinance101.wordpress.com/2012/06/05/the-commonwealth-triple-screw-special-education-funding-charter-school-payments-in-pennsylvania/

Where the PA special education funding formula for charters actually encourages taking on low severity special education students, because charters receive the average special education spending rate of the host district for each special education student. In other words, the fiscal incentive in PA is to set up a charter specifically geared toward mild learning disabilities and speech impairment.

I also show this effect in New Jersey here:

http://schoolfinance101.wordpress.com/2012/08/06/effects-of-charter-enrollment-on-newark-district-enrollment/