In 2012, my graduate student Will Dobbie and I collected unprecedented data from nearly 50 New York City charter schools to see which practices truly boosted student learning. Class size and teacher credentials—political obsessions for decades—mattered little. What mattered most were five concrete, replicable practices:
more instruction time,
high expectations,
frequent teacher feedback,
data-driven instruction and
high-dosage tutoring.
Together, these five tenets explained roughly half the difference between effective and ineffective schools.
Armed with that evidence, we searched for districts willing to test the model—from Haiti to Harlem. Most weren’t interested. But in Houston, superintendent Terry Grier opened the door. Together we applied the Five Tenets in 20 struggling public schools serving nearly 20,000 students. We lengthened the school year by 20%, brought in hundreds of tutors, replaced 95% of principals and half the teachers while retraining the rest, embedded data into instruction, and built a culture of high expectations. It was one of the most ambitious social experiments in American public education.
The results were astonishing. In elementary-school math, students gained the equivalent of four extra months of learning a year—enough to erase the racial achievement gap in less than two years if we implemented these practices in the lowest-performing half of schools nationwide. In secondary schools, where skeptics said reform was impossible, students gained nearly eight additional months of learning in a nine-month school year. These were bigger effects than those produced by the Harlem Children’s Zone. Bigger than Success Academy. Bigger than anything else I’ve seen in my career.
For context, cutting class size yields about three months of extra learning. Teach for America adds two months in math. Head Start delivers about two months in early literacy. The Houston schools doubled those gains—year after year. By the third year, elementary students had accumulated the equivalent of an extra academic year. In middle and high school, it was two. These weren’t “miracle kids” or “superhuman teachers.” The system—not the students—changed.
Math results were jaw-dropping, but reading proved stubborn. Gains were modest in elementary school and nearly zero in secondary school—mirroring what even the best charter schools have found nationwide.