© 2026 Swiftscore, Inc.
An interactive look at which schools outperform what their poverty rate would predict in third-grade reading. Each dot is a school; the schools above the line are doing something worth studying.
CTO of Swiftscore. Writes about what we're learning from districts on the ground.
Simplifies evaluation cycles and supports equitable teacher growth across multiple campuses.
Poverty predicts reading scores. The correlation is one of the most stubborn findings in education data, and it is easy to mistake for destiny. Within every state, though, there are schools serving high-poverty students whose third graders read at rates you would expect from much wealthier ones.
This piece is built to find them. The interactive at the top of the page plots every public school in a chosen state by its student poverty rate and its grade-3 reading proficiency, with a state-specific line of expected performance.
Each dot is a school. The horizontal axis runs left to right, from lower-income to higher-income student bodies. The vertical axis is the share of third graders reading at or above proficiency. The line is the expected proficiency at each income level, fit separately within each state, and it rises with income.
Schools at or below the line are performing about what the data predicts, or less. Schools above the line are the story — same headwinds, better results. The further above, the more brightly the dot is shaded. Hover any dot for the school’s name and how far above expectation it landed.
A raw ranking of proficiency just re-sorts schools by their poverty rate. The wealthiest schools win, every time, and you learn nothing. Fitting an expected line and measuring the gap to it removes the part of the score that poverty explains. What is left is the part a school’s people might be responsible for.
We fit a linear regression of grade-3 reading proficiency on poverty rate within each state, using only schools with at least 30 tested third graders. Each school receives an expected proficiency from that line and a residual — actual minus expected. Reading proficiency comes from the U.S. Department of Education’s EDFacts grade-3 assessments; poverty and enrollment from the NCES Common Core of Data. The approach echoes The 74’s Bright Spots methodology; the dataset is rebuilt from those primary public sources rather than copied from anyone’s processed file.
The questions below are answered in the structured data for this page.