A previous evaluation of the six-year-old Milwaukee voucher project, which has enabled disadvantaged African-American and Hispanic children to transfer to relatively poorly resourced private schools, suggested that it had produced no identifiable gains. However, the latest study by Jay Green, of the University of Houston, and Paul Peterson and Jiangtao Du, of Harvard University, points out that the earlier evaluation made the mistake of comparing private school students from low-income families with state school pupils from more advantaged backgrounds.
Jay Green and his colleagues believe they have avoided that trap by comparing the maths and reading scores of pupils in three private schools with those of the children who failed to gain places in the schools. They argue that this is a fairer comparison, because the rejected children came from the same neighbourhoods and were not turned down because of their academic or social backgrounds. They were merely unlucky in the lottery that was arranged to allocate places in the three oversubscribed schools that took 80 per cent of the voucher children.
The researchers discovered that after three years in the private schools children were scoring 5 per cent more in maths. By the following year the gap had widened to almost 11 per cent. The reading progress of the two groups was more similar, but after four years the scores of children in private schools were almost 6 per cent ahead of those who had remained in the state schools.
Jay Green and his colleagues point out that if efficiency gains are to be found in private schools such as Milwaukee, which were operating under "considerably less than ideal circumstances", then even greater gains might be obtained in more opportune conditions.
They acknowledge that private schools may make the biggest impact in inner-city areas that have a single school board that does not allow schools to compete.
The team also note that "the benefits of privatisation may be greater for those families who desire an alternative to the public school serving them. Their children may have been particularly at risk in public school."
However, they also argue that the differences in racial attainment levels would largely disappear if African-American and Hispanic children in other parts of the United States could make the type of gains made by the voucher pupils in Milwaukee over the 12 years they attend school.
"A large part of between-group (whites and minorities) reading differences and all of the maths differences could be erased," they predict.
"Similar experiments should be conducted in a variety of contexts, but especially in central (inner) cities where potential efficiency gains seem particularly likely."