About this analysis: The Star Tribune used a linear regression analysis to compare each school's proficiency rate to an expected proficiency rate based on the school's percentage of students on free or reduced-price lunch (a common proxy for estimating a school's poverty rate). Only schools that tested at least 25 students, across all grades, were included in the analysis; alternative learning and special education schools were not included. Schools were identified as "falling short" if the actual proficiency rate was 10 percentage points or more lower than the expected rate. Schools were classified as "better than expected" if their rate was 10 percentage points or more greater than the expected rate. All schools between those marks were categorized "as expected."