Mancini doth protest a bit too much however, since the concern with the KIPP chain of replicant schools has always been more about attrition, than skimming. It's always been difficult for charter schools to aggressively skim because most are required to admit students through a lottery. Some do it anyway, by claiming that they don't have the resources to teach English language learners or kids with disabilities or behavioral problems, a claim that neighborhood public schools aren't allowed to make.
Miron is actually much kinder and gentler and even somewhat apologetic towards KIPP and the Mathematica researchers than I am. He could have pointed out for example, that KIPP only allowed the researchers to look at 22 carefully selected hand-picked schools after which they concluded that KIPP didn't skim their students. They claim that data wasn't available on the others. Why not I wonder? Mabe Gary will take some of these questions up at a later date.
According to Edweek,
"...the researchers attempted to get data for all 35 KIPP schools that were opened by 2005, but data from school districts and states were only available for 22 of those schools."But skimming at the admissions end was never really as much an issue as was student attrition, with low-scoring kids being pushed out and not replaced by other low-scoring kids.
The KIPP study did find that the charter school chain did and does have much smaller populations of ELL and special ed students which make it difficult to compare KIPP with the public schools that are receives in large numbers of these students.