How we ranked the schools

For a decade, the Chicago Sun-Times has based its exclusive rankings of public schools on average scores on state achievement tests, not on the percent who meet state standards — a measure that’s come under criticism.

Only 2009 reading and math results from the Illinois Standards Achievement Tests and the Prairie State Achievement Exams taken last March and April were analyzed. Results in those subjects can trigger sanctions under the federal No Child Left Behind law.

Elementary school rankings are based on schools that tested at least two grades, third to fifth. Middle school rankings are based on schools that tested two grades, sixth through eighth. High school rankings are based on 11th-grade results. A K-8 school should be ranked among both elementary and middle schools.

The rankings use a statistical method called standardizing to analyze the “scale score” of every reading and math test statewide. The method compares each test score to the state average and creates a school average that’s compared to other schools’ averages. Standardizing levels the playing field in years when one test might be harder to pass than others.

The system allows for more definition among top schools, at which 100 percent of students might pass some tests, because it averages every score, rather than counting only kids who met or exceeded a passing bar.

The rankings include percentiles, reflecting the percentage of Illinois students who scored the same as or worse than the average student at each ranked school.

In the middle of the pack, school averages are much closer together, so differences in rank among middle-scoring schools may reflect very small differences in average scores. In addition, schools in the middle tend to have more ties in rank, so large differences in ranks there may reflect smaller differences in scores than among schools at the bottom or top. Percentiles can help readers detect how much higher or lower their school scored than the ones immediately above and below it.

For each school, only grades in which at least 10 students were tested were analyzed.

Sun-Times database reporter Art Golab performed the analyses.