But CPS has selective enrollment public schools that skew the results...

But CPS has selective enrollment schools that skew the results.  What happens when you remove these schools from the analysis?

• The Sun-Times analysis did not include selective enrollment schools.
• I remove them from my analysis when I eliminated all high attainment public and charter schools from the analysis (see note below). In that analysis the public school growth percentile number was 64 while the charter growth percentile was 29.

• The assumption behind the question about selective enrollment schools is not correct, and it's one of the main misinterpretations of NWEA growth data. I hear it every time I talk about MAP data. Selective schools' growth percentiles are calculated against other schools that got the same growth percentile in the first administration of the test.  Put simply, their percentiles are generated by comparing them to other schools across the U.S. with equally high performing students. It answers the question, "How much growth did they get compared to other high performing schools?"  For example one of CPS's selective academic centers has the highest average attainment in Chicago.  However, their growth percentile is only 76, which ranks 143rd in the district.  Another selective school--has the 35th highest math attainment score in the city but is at the 2nd growth percentile, one of the worst in the district, ranking 473rd out of 490.  CPS recently amended its performance policy to protect the mayor's prized selective enrollment schools from the embarrassment their low growth numbers would cause them if they were evaluated the same way neighborhood schools are.

• The MAP "growth" measure is as as close to an apples-to-apples comparison as is available, so there is no need to remove any type of school from the analysis. Still, the Sun-Times removed them anyway and charter schools still feel far short of public neighborhood schools in terms of fostering academic growth in their students.

Note:  I excluded schools that have 50% or more of their students scoring above average in attainment.  That ensured I was left with schools that served a predominantly low performing student body which obviously excluded all selective enrollment schools.  I did that by sorting the spreadsheet by the "% at/above national average" column and including only those schools whose percentage was below 50.  The CPS data specialist then processed those schools into CPS's "School Growth Calculator" to reach our findings.