It seems like the perfect parable for advocates of school vouchers: A wealthy benefactor offers to pay private school tuition for any pupil wishing to flee "the worst public school" in Albany, N.Y. In response to the exodus of one in five of its students, the school hires an energetic new principal and adopts a curriculum reputed to help low-income students.

School voucher advocates have been quick to point to Giffen Elementary School as evidence that their market-oriented approach will unleash competitive forces that stimulate public schools to improve. According to the theory, failing public schools will lose students, and with them tax revenues, prompting them to become better and more efficient in order to survive. Giffen represents "the first time we've seen wholesale changes in response" to vouchers, said Brian Backstrom, who administers the Albany program.

But Giffen, it turns out, makes a poor case study in educational competition. Because of a New York state law placing a floor on state aid for each district, Giffen didn't lose any money as a result of the voucher program. In addition, Giffen, which has long had high student turnover, already has gained more students through families moving into the neighborhood than it lost through the voucher program. Officials at the school say its efforts to improve aren't a result of competitive pressure but a response to the embarrassment of being named the city's worst school.

The school vouchers movement gained momentum earlier this week when Florida Gov. Jeb Bush (R) signed into law the first statewide program aimed at improving failing schools. And yesterday the Ohio legislature voted to restore a Cleveland program that had been struck down on a technicality by the state supreme court last month.

But a decade into school voucher experiments in cities across the country, there is still little evidence to support the underlying theory--first articulated by Milton Friedman in the 1950s--that economic competition will force public schools to get better. At the same time, there's also little evidence for voucher foes' prediction that the programs could destroy public school systems.

To date, experiments with public vouchers in Milwaukee and Cleveland and with privately funded programs in 40 other cities have been too small to cause the effects predicted by those on either side of the voucher debate. In addition, in cities such as Albany and Milwaukee, state laws or local politicians have prevented public schools from losing money to voucher programs, insulating the schools from the consequences--good or bad--of competition.

"There's not been enough of a mass movement out of a public [school] system to create change," said Fritz Steiger, president of CEO America, an Arkansas-based clearinghouse of private voucher programs.

"There's talk of big, significant change," said Frederick M. Hess, a University of Virginia professor who has studied voucher programs. "There's zero evidence of that in Milwaukee, Cleveland and Edgewood," an impoverished district in San Antonio with a private voucher program.

Private vouchers, generally funded by business leaders and usually called "scholarships," pay most of the tuition for low-income students in such cities as Albany, San Antonio and the District to attend any private school that will admit them. But so far private benefactors have donated enough money to lure only a tiny percentage of any district's students to private schools.

Sponsors of private vouchers recognize limits on the contributions available to support their programs and say their goal is to create models that could build a case for larger systems of government-financed vouchers. Virginia Gilder, for example, the New York benefactor who put up several hundred thousand dollars for the Albany scholarships, conceded that generosity like her own could never match government as a potential source to pay private school tuitions.

"You cannot have private philanthropy compete with that," Gilder said. "We can set an example."

Government voucher programs are more controversial because of legal disputes over the constitutionality of public funds going to religious schools and because they potentially could take more money away from public schools. Private vouchers reduce funding indirectly by reducing enrollment, which is one of the factors that determines state assistance; with government vouchers, the full state aid for each individual student follows him or her to the private school. (Public schools are also beginning to face economic competition from charter schools, which are public but are managed outside of local bureaucracies.)

But the government-funded voucher programs in Milwaukee and Cleveland are still relatively small in scope. Moreover, state officials have made efforts to prevent these programs from draining funds from public schools, resulting in a lack of evidence on the possible impact of educational competition. (Older government voucher programs in Maine and Vermont don't exert competitive pressure on public schools, because their purpose is to pay private school tuition for students in rural areas where there is no public school.)

The oldest and largest public voucher program with a market-oriented bent is in Milwaukee, where the number of students attending private schools at public expense jumped to 6,000 from 1,500 after the Supreme Court last year declined to review the constitutionality of the state allowing vouchers to be used for tuition at religious schools.

Even with the quadrupling of voucher students, Milwaukee's public school system lost fewer than 6 percent of its 106,000 students. The enrollment decline did cost the district $26 million, the amount of state aid delivered to private schools for tuition. But the district was let off the hook: The school board compensated for the financial loss by raising property taxes, not by cutting the budgets of schools that lost pupils.

And although Milwaukee doesn't keep track of which schools the voucher students have abandoned, it appears that the program has drawn relatively small numbers of students from many of Milwaukee's 157 schools, diffusing the potential competitive impact on any one school.

"I don't think there's a predominant public school where they're coming from," said Sister Virgine, principal of Urban Day School, a popular destination for voucher students. "They come from all over the city. It's one here, one there."

In Cleveland, about 3,700 students used public vouchers to attend private schools in the last academic year, a decrease in public school enrollment that cost the district nearly $9 million in state funds. But a spokeswoman for the Cleveland schools indicated that the district absorbed the loss in its $600 million budget without major difficulties and noted the offsetting effect of overall district enrollment increasing from 74,000 to 77,000 since the voucher program began three years ago.

The Edgewood district in San Antonio does grudgingly acknowledge some competitive impact from its private voucher program, possibly because the anticipated loss of about $4 million in state aid this fall represents a larger proportion of its $86 million budget. To offset the loss, the district has delayed school roof repairs and begun to reduce staff by attrition.

Edna Perez-Vega, spokeswoman for the Edgewood schools, said the district has increased efforts to communicate with parents about academic offerings and accelerated implementation of a remedial program for underachieving ninth-graders.

"In some ways, we have became better marketers this year," Perez-Vega said. "It has encouraged us to pursue more aggressively our secondary reform program we had been working on."

But such steps have not impressed Steiger, who described the Edgewood school board as unresponsive and resistant to educational reform. "I think it's too early to see substantial changes, but I think you're going to see some as more students leave," he said. "Competitive forces are going to kick in."

Hess of the University of Virginia said that scale will ultimately make the difference in districts where a voucher program operates. "It has to be big enough to be actually threatening," Hess said. "You actually have to threaten people with extinction."

School Choice

The highlighted states contain private voucher programs. Wisconsin and Ohio already have publicly funded voucher programs, and one is scheduled to begin in Florida.

Key events

1971

The Supreme Court rules that public funding is permissible to private schools if it has a secular purpose, neither promotes nor restricts religion and avoids creating "excessive government entanglement with religion."

1990

Wisconsin legislators approve a school voucher program in Milwaukee that provides funding for students' tuition at nonreligious private schools. In 1995, the program is expanded to include parochial schools.

1991

An Indiana businessman offers to pay half the tuition for as many as 500 children so they can attend the school of their parents' choice. Other private voucher programs soon follow.

1995

The Ohio legislature passes a voucher program for the city of Cleveland that pays for private and parochial school tuition.

1999

In April, the Florida legislature approves the first statewide voucher program, which will pay for students in failing public schools to attend other public, private or parochial schools.

SOURCE: CEO America, Facts on File

CAPTION: Arianna Muhammad, 4, looks to classmates for guidance during prayer on Aug. 31, the first day of classes at a Milwaukee religious school affected by voucher law.