Back in 2017, the families of children in some of California’s worst-performing public schools sued the state for failing to teach low-income black and Hispanic children how to read. This led to a legal settlement in which the state’s 75 worst-performing elementary schools agreed to invest in evidence-based reading instruction—that is, in training teachers to use techniques, such as phonics, for which there is strong evidence that they work.

According to a new working paper from two Stanford researchers, the extra training helped. Students’ reading scores improved when compared to students from other poorly performing schools. The score increases were roughly as valuable as an additional 25 percent of a school year.

Considerable research shows that children learn to read best by using phonics—essentially, by “sounding out” words they don’t know. Unfortunately, many schools over the past several decades have eschewed phonics in favor of methods like “three-cueing,” which encourages children to guess words they don’t know using context clues, not the actual letters in a word.

Even with these gains, most of the students in these schools are still struggling. At the end of the study, two-thirds of them still failed to meet state reading standards.

“I wouldn’t call the results super large,” Princeton sociologist Jennifer Jennings told The Herishinger Report, an education news site. “I would call them cost-effective”

Small though these results are, they are much larger than those prompted by most other interventions. The paper notes that the change had a “larger effect size than almost 90 percent of educational interventions serving more than 2,000 students.”

This change was funded by the settlement—the training costs came to about $1,000 per student—but it’s not hard to imagine how similar results could be met without a big funding increase. School funding tends to have a very weak relationship with school quality. This isn’t because money doesn’t matter; it’s that many schools waste funds on things like additional non-instructional staff and steep teacher benefits. (In 36 states, schools with higher poverty rates already receive more funding per pupil than schools with higher-income student bodies.)

Retraining teachers is likely well within the budgets of many poorly performing schools—as long as administrators choose to devote their resources to a proven method.