A New York State Supreme Court justice ruled this week that the state must allow the inclusion of a ballot question regarding class size reduction on the upcoming November ballot. The city had argued that although the question had gathered 115,000 signatures necessary for inclusion on the ballot, the city had the right to "bump" the question off in favor of one written by the mayor.

While we have no opinion on the fine points of New York State constitutional law, we are interested in the movement to reduce class sizes. The New York City law would limit class sizes to 18 at the earliest grades, 22 in the middle grades, and 25 in high school. A similar provision passed by the voters of Florida could cost the state as much as $27 billion over 8 years.

The problem with class size reduction isn't just that it's expensive, it's that there is scant evidence that it has the impact that advocates seek, especially in the middle and upper grades. Most of the push for the reduction in class sizes is based on the results of one study, the Tennessee STAR analysis, which show small but statistically significant gains in teacher quality for children who were poor and in lower grades. But, as researchers such as Caroline Hoxby and Jay Greene note, the entire country has been the testing ground for a huge experiment in class size reduction, as class sizes have dropped significantly in the last 30 years. Nevertheless, test scores and graduation rates for our 17-year-olds have remained flat.