New IBO Study: Charter Students Stick Around

A new report from the NYC Independent Budget Office found that student attrition rates in charter schools are lower across nearly every student subgroup — with the one exception, which contradicts previous research, being calculated from a tiny sample.

The IBO study followed the kindergarten class of 2008-09 at 53 charter schools and 116 district schools in the same neighborhoods, for four years. The report finds that English Language Learners, Students Eligible for Free or Reduced-Price Lunch, Black students, Hispanic students, White students, boys, girls, and students in general, were all more likely to remain at their school four years later if they attended a charter school.

These findings are consistent with a 2012 analysis by WNYC public radio, which found “no exodus from the charters. In fact, the charters had lower attrition rates, on average, than the regular public schools.”

The IBO did find a striking exception to the trend, that only 20 percent of charter school special education students remained for the full four years, compared to 50 percent in the district school comparison group. But something doesn’t add up.

It turns out that the IBO counts only 25 total charter school kindergarteners as “special education students,” and it was this tiny group that produced the 20 percent figure. In another recent study also using NYC Department of Education data, researcher Marcus Winters identified 198 charter school students in special education, in kindergarten, in that same year—more than seven times as many.[1]

That difference may have prevented the IBO from finding lower charter school attrition in every category. When Winters analyzed attrition among the larger group, he found lower rates of attrition among charter school students with special education needs (albeit compared to students in district schools citywide rather than the IBO’s comparison group of district schools located near charter schools).[2]

So what happened? It seems the IBO made a critical decision that restricted its sample of charter school special education students: it looked only at students with a special education “flag” in the data, which for some reason is a small subset of all those listed as having a disability. Only about half of students classified with a learning disability have this flag, for example. The reason is not clear, but it may be related to which students have finalized special education plans or are actively receiving services from the DOE.[3]

As a result, the IBO’s attrition calculations cover only a very small subset of special education students in charter schools – only one in seven students who should have been included.

Yet that doesn’t mean the report is all roses for charter schools. A large majority of those 25 special education students still chose to leave, after all, and it is true that charter schools do not enroll a proportional share of students with the most intensive special education needs. Further research is needed to see how attrition rates may vary by disability type.

At the Charter Center, we’re on the record calling for better data reporting about student mobility patterns.[4] The IBO’s report is misleadingly incomplete on the topic of special education, but still a step in the direction of transparency.


1. Winters, Marcus. Why the Gap? Special Education and New York City Charter Schools. Center on Reinventing Public Education, September 2013. See p. 10, Table 3, Year 1, “Ch” column, lowest two rows. Elsewhere this study mentions the number 178, which excludes students who left NYC over a four-year period.

2. Winters, p. 15: “[T]raditional public school students with IEPs are significantly more likely to leave their kindergarten school than are students with IEPs who attended a charter school in kindergarten.”

3. The IBO also reports that only 18 total charter school students (in 53 schools!) were newly flagged as “special education” during their kindergarten year, but then a dozen times as many (232 students) were flagged during first grade (Table 6). This pattern is utterly implausible in terms of educational practice, but makes sense if the “special education” flag variable depends on some kind of DOE paperwork being finalized.

4. Check out our white paper on Four Simple Ways to Improve New York City School Data.