Paul E. Peterson and William G. Howell, Harvard University
June 12, 2003
It's round three of the ongoing Harvard-Princeton debate over the effects of a privately funded New York City voucher program. In the first round, Harvard researchers Paul Peterson, William Howell, and others conducted an original study on the effects of the voucher program on low-income children who changed schools with its assistance, compared with a control group of similar youngsters who did not change schools. Based on their analysis of the data, Peterson et al. concluded that participation in the voucher program had positive effects on the test scores of African Americans but no effects, positive or negative, on the scores of Hispanic and white students. These initial conclusions offered tentative support to the idea that vouchers might provide poor students, particularly African Americans, with a better alterative to public schools. Princeton researchers Alan B. Krueger and Pei Zhu reanalyzed the data and concluded that the initial study was methodologically flawed because of the way it determined whether a student was black and because the researchers failed to include students for whom there was no baseline data. In their reanalysis, Krueger and Zhu found that that "the safest conclusion is probably that the provision of vouchers did not lower the scores of African American students." After the New York Times made a big deal of this, Peterson and Howell undertook their own reanalysis of the data. This time, they employed both their initial methodology and Krueger and Zhu's methodology. They (again) found that "all the standard ways of estimating voucher effects showed significantly positive effects" for African American students and (again) that "no effects, positive or negative, were observed for Hispanic students or for students who were members of other ethnic groups." Besides confirming Peterson and Howell's initial findings, this report explains the shortcomings in Krueger and Zhu's methodology and conclusions. The study is being released today at the National Press Club, and will be available online at http://www.ksg.harvard.edu/pepg/.
