What the data shows
Whether charter schools beat the district schools they compete with depends almost entirely on which state you are in. Nationally the two sectors look nearly tied on a poverty-adjusted basis, but that average hides an enormous spread. In New York, charter schools beat their demographics by about 13 percentage points more than the state's district schools do, the widest charter advantage in the country. In Ohio and Pennsylvania, charters trail district schools by 22 and 14 points, dragged down by large and troubled cyber-charter sectors. Across the 25 states with enough scored charters to compare, charters lead district schools in 9 and trail in the rest.
The other consistent pattern is variance. Charter schools are far more likely than district schools to land at both extremes: nationally about 18% of scored charters are top-decile outperformers (versus roughly 10% of district schools), but charters are also over-represented among the bottom-decile underperformers. Charters are a higher-risk bet on average, with a wider gap between the best and the worst.
Why we adjust for poverty
Ranking charters by raw proficiency mostly ranks them by the families that enroll. To make a fair comparison we use BeatsExpectations, a per-state regression of school proficiency on free-and-reduced-lunch share. It predicts the proficiency a school should post given its poverty profile, and the residual, in percentage points, is how far above or below that prediction the school actually lands. The "charter edge" in the table above is the charter sector's average residual minus the district sector's, computed inside each state so the comparison is always like-for-like.
How to read the table
Each row is a state with at least 15 scored charter schools. "Charter avg" and "District avg" are the two sectors' average residuals; "Charter edge" is the difference. Positive means charters beat their demographics by more than district schools do. Click any state to see its individual charter schools ranked from best to worst.
Limits
BeatsExpectations controls for poverty, but not for selection. Charters can use application and lottery processes, can enroll different shares of special-education and English-learner students, and can see different mid-year attrition than the neighborhood schools they are measured against. A positive charter edge is strong evidence that a state's charter sector is adding value, but it is not proof that the same students would have done worse in district schools. States with charter bans or very small charter sectors are absent because the sector average would be too noisy to report.
Methodology
Sources: NCES Common Core of Data 2024-25 (charter flag, FRL share, enrollment) and state-native assessments for the most recent available year. The BeatsExpectations regression is ordinary least squares fit independently per state and recomputed after every data refresh; the code is in the public allk12 repository.
Browse by state
Alabama · Arkansas · California · Colorado · Florida · Georgia · Illinois · Indiana · Louisiana · Maryland · Michigan · Minnesota · Missouri · New Hampshire · New Jersey · New Mexico · Nevada · New York · Ohio · Oklahoma · Oregon · Pennsylvania · Texas · Utah · Wisconsin