More than two years ago, philanthropists funded pilot school voucher programs in New York City, Washington, D.C., and the Dayton metropolitan area. The programs, which helped students from low-income families attend private schools, asked the Program on Education Policy and Governance at Harvard to evaluate them.
We agreed to do so on the condition that we be allowed to conduct a randomized field trial, a rigorous type of evaluation historically used to determine whether medical interventions, such as a pill, can be marketed. In a randomized field trial a lottery is used to decide whether or not a patient gets the pill (in this case a school voucher). When you compare those who got the pill (voucher) with those who did not, you are comparing two groups that were similar when the field trial began. Any differences that appear later are due to the pill (voucher).
The results from the first two years of the pilot voucher programs were mixed. We found that the programs had no significant impact on students from ethnic backgrounds other than those who were African American. The results for African Americans, however, were clearly different. The African American students who switched to private schools scored, on average, 3.3 national percentile ranking points higher after one year and 6.6 percentile points higher on combined reading and math tests than their counterparts in public schools. (About three-fourths of the students participating in the evaluation were African American.)
Thus after two years, the voucher intervention erased, on average, about one-third of the differences that exist nationally between the test scores of African American students and those of other students. If the trend observed over the first two years continues in subsequent years, the test gap between African Americans and whites could be eliminated for African American students who use a voucher to switch from public to private school.
We need to do more studies to ascertain why the impact of the voucher program on African American test scores is greater than it is for other students. But a preliminary analysis suggests that the change in the educational environment that accompanies a switch from public to private school is substantially greater for African American students than for students from other ethnic backgrounds. When an African American student moves from a public school to a private school in New York City, he, more than other students, is likely to move to a school that has fewer discipline problems and stronger communication links with parents. Also, teachers demand more homework, the school is smaller, and there are fewer students in the classroom.
These changes occur for all students, but the changes are the largest for African Americans. In short, the chance to go to a private school makes a bigger difference for African American students than for those from other ethnic backgrounds.