Since around 2000, the National Assessment of Educational Progress has included charter schools in its testing program. Thus it is possible to single out charter schools and compare them with traditional public schools. An important limitation of such comparisons is that it is not possible to check whether the sample of students in the 150 charter schools tested were ahead or behind the traditional school sample in previous years.

To do this comparison, first go to the NAEP Data Explorer. Then click on the button for Main NDE. In Step 1, select reading or math for the subject and 4th or 8th for the grade. Under Jurisdiction, select National Public and the year(s) you are interested in.

Go to Step 2 (select variables). Under Student Factors, select Natl School Lunch Prog eligibility and Race/ethnicity. Under School Factors, select School identified as charter. Then go to Step 3 and edit the reports if you wish.

Finally go to Step 4 (Build Report). The example described here will generate four reports: three comparing values for each of the three variables (by ethnicity, by free lunch, and charters to non-charters) and a cross tabulation of all three.

The advantage of the cross tabulation is that dividing the students into similar ethnic and income subcategories makes it more likely that the students in charter schools are similar to those in the traditional public school. Differences may remain, however.

For example, using 2009 data for mathematics, charter students have an average score of 231 compared to 239 for traditional public schools, a difference well beyond the margin of error. Yet when the students are compared by ethnic and income subgroups, the scores are statistically the same. Thus it appears that charters serve a disproportionate number of students from demographic groups that tend to score lower on the tests.