Congress has passed a voucher bill that would provide low-income students in Washington, D.C., an opportunity to attend the private school of their choice. The White House has threatened to veto the bill on the grounds that school choice will undermine public education--never mind that both Bill Clinton and Al Gore sent their own children to private schools.

It is not too late for Mr. Clinton to rethink his position on a matter of vital importance to inner-city families. New findings from the Cleveland Scholarship Program, a pilot voucher plan, provide further evidence about the substantial benefits inner-city children in the nation's capital and elsewhere could reap from school vouchers.

Last summer we reported that Cleveland parents enthusiastically supported the program, which in its first year enabled the parents of some 2,000 children to choose a school for them. We also reported significant gains in reading and math scores for 263 students who attended two newly established choice schools, called the Hope schools, which serve 25% of the choice students who hailed from public schools.

Our test-score findings were challenged by teachers unions on two grounds: The gains observed during the 1996-97 school year could be expected to fade away over the summer break; and another evaluation, by Indiana University's School of Education, found no programmatic effects on the test scores of 94 third-grade choice students. Neither of these challenges withstands scrutiny.

To test the "summer-fade" theory, we obtained test-score data from the students returning to the Hope schools in the fall of 1997. Although the test score gains did slip somewhat, students in grades one through three still gained in their first year an estimated 5.1 normal curve equivalent percentile points in math and 3.4 points in reading--gains that were statistically significant.

The Indiana University evaluation now used to cast doubt on the Cleveland program has four major weaknesses:

First, the study analyzed only third-grade test scores; no information is available for students in earlier grades.

Second, to control for student achievement prior to the beginning of the scholarship program, the evaluation used implausible second-grade scores collected by the Cleveland public schools before the choice experiment began. According to these scores, second-grade students from inner-city, low-income, largely one-parent families were performing at nearly the national average (getting a 51.6 percentile score on the vocabulary test, for example). Yet in an independently proctored test administered one year later, the same students scored, on average, at 39.6 points in reading. Clearly, the second-grade test scores used by the Indiana study as a benchmark for assessing the choice program were inflated.

Third, the evaluation excluded Hope school students, despite the availability of comparable test-score results.

Fourth, the Indiana research team used inappropriate statistics that have been shown to underestimate program effects.

With the release of the data from the Indiana University evaluation, it is possible to correct some of its deficiencies by adding the Hope school test results into the analysis, using more appropriate statistical techniques, and reporting results based on analyses that both include and exclude the doubtful second-grade data supplied by the Cleveland public schools.

When these corrections are made and the implausible second-grade test scores are removed from the regression analysis, choice-school effects are positive in all subject domains. Scholarship students score 4.1 percentile points higher in language, 4.7 points higher in science, 2.5 points higher in reading, 2.5 points higher in social studies and 0.56 points higher in math.

If the dubious second-grade scores are included in the analysis, results remain positive in all subjects except math. The school choice advantage in language arts is 2.3 percentile points and 2.7 points in science.

Even the most conservative of these estimates of choice-school effects are comparable to those that we observed in Milwaukee in 1991, the first year of that city's voucher program. The modest first-year gains in Milwaukee were followed by substantially larger effects in years three and four.

One cannot expect to observe more than modest learning gains in the first year of attendance at a choice school. It takes time for children to adjust to a new setting and take advantage of whatever opportunities a school can provide. As Barbara Lewis, an Indianapolis mother whose son attends a choice school, testified before Congress: "I must admit there was a period of transition, culture shock you might call it. He had to get used to discipline and the homework. . . . But Alphonso began to learn about learning, to respect the kids around him and be respected, to learn about citizenship, discipline and doing your lessons. . . . My son has blossomed into an honor roll student."

Mr. Peterson is director of the Program on Education Policy and Governance at Harvard University, where Messrs. Greene and Howell are program associates. Their study will appear in a book forthcoming from the Brookings Institution, "Learning from School Choice."