Some charter schools are enormously effective. Can specific practices from successful such charter schools be transplanted into regular public schools, bringing their success along with them? A real field experiment run by Roland Fryer suggests that enormous gains can be made in this way.
Fryer engineered an experiment where major reforms were made in a small number of very badly performing Houston schools. The schools were each paired with similarly performing schools where reforms were not implemented. Bottom line? The treated schools had incredible gains relative to the untreated schools. The gains were on the order of a large fraction of black-white test score gaps.
1. Extended learning time.
- Elementary school students were “strongly encouraged” to attend Saturday classes “tailored to each student’s needs” (35 days a year). Instructional time during the week was extended by reducing break time between periods.
- Secondary schools added ten days to the school year as well as an additional hour Monday through Thursday. Secondary school students were also encouraged to attend Saturday classes.
2. Human capital changes
- Almost all the principals were replaced with new, handpicked, experienced leaders.
- Between a third and a half of teachers were replaced. Here too, the experimenters had discretion to decide who would stay and who would go.
- Under the reform, teachers were given ten times more evaluations and feedback than in non-treated schools.
- Teachers received specific summer training as well as Saturday classes throughout the year to help increase rigorous instruction, as well as on classroom management techniques.
3. High-dosage tutoring
- Fourth, six, and ninth graders received intense math tutoring. The fourth grade tutoring was three-on-one. The other two grades were treated two-on-one. The improvement in math in these grades were much greater than in the other grades, where the experimenters did not have enough money to implement high-dosage tutoring.
4. Data-driven instruction
- Teachers received very regular data on the performance of individual students. In addition, benchmark tests were taken in December, February, and March after which teachers met one-on-one with students to set individual performance goals.
5. Culture of high expectations
- Schools had a clear set of goals. Schools and parents signed contracts indicating agreement with high expectation policies. Schools spoke the language of “no excuses.”
The report of the experiment does include one troublesome element. Fryer is quite clear that while there were large gains in math, there weren’t in reading/language arts. Fryer writes, “injecting best practices … has little effect on reading achievement.” One reason may be that the intensive tutoring intervention was math-only. But Fryer cites other studies which show similar failures of interventions for reading. It may be that language formation needs intervention at an earlier age.
Good teachers, good leaders, high expectations. And, importantly, a great deal of attention to individual student needs. In many ways, these are exactly the characteristics that educators have known for a long time work quite well. Next time, more on cost-effectiveness and scalability.