When the federal Education Department recently reported that children in private schools generally did no better than comparable students at public schools on national tests of math and reading, the findings were embraced by teachers' unions and liberals, and dismissed by supporters of school voucher programs.

But for many educators and policy makers, the findings raised a haunting question: What if the impediments to learning run so deep that they cannot be addressed by any particular kind of school or any set of in-school reforms? What if schools are not the answer?

The question has come up before. In 1966, Prof. James S. Coleman published a Congressionally mandated study on why schoolchildren in minority neighborhoods performed at far lower levels than children in white areas.

To the surprise of many, his landmark study concluded that although the quality of schools in minority neighborhoods mattered, the main cause of the achievement gap was in the backgrounds and resources of families.

For years, education researchers have argued over his findings. Conservatives used them to say that the quality of schools did not matter, so why bother offering more than the bare necessities? Others, including some educators, used them essentially to write off children who were harder to educate.

The No Child Left Behind law, enacted in 2002, took a stand on this issue. The law, one instance in which President Bush and Congressional Democrats worked together, rests on the premise that schools make the crucial difference. It holds a school alone responsible if the students -'" whatever social, economic, physical or intellectual handicaps they bring to their classrooms -'" fail to make sufficient progress every year.

Yet a growing body of research suggests that while schools can make a difference for individual students, the fabric of children's lives outside of school can either nurture, or choke, what progress poor children do make academically.

At Johns Hopkins University, two sociologists, Doris Entwisle and Karl Alexander, collected a trove of data on Baltimore schoolchildren who began first grade in 1982. They found that contrary to expectations, children in poverty did largely make a year of progress for each year in school.

But poor children started out behind their peers, and the problems compounded when school ended for the summer. Then, middle-class children would read books, attend camp and return to school in September more advanced than when they left. But poorer children tended to stagnate. "The long summer break is especially hard for disadvantaged children," Professor Alexander said. "Some school is good, and more is better."

"Family really is important, and it's very hard for schools to offset or compensate fully for family disadvantage," he said.

In Chicago, a court order to empty public housing projects, which dispersed families and children into the suburbs, led to a rise in children's academic achievement.

"The evidence is pretty clear that the better their housing, the better kids do on tests," said Jack Jennings, president of the Center on Education Policy, a nonpartisan group.

In his 2004 book, "Class and Schools: Using Social, Economic and Educational Reform to Close the Black-White Achievement Gap," Richard Rothstein, a former writer of this column, argues that reforms aimed at education alone are doomed to come up short, unless they are tied to changes in economic and social policies to lessen the gaps children face outside the classroom.

A lack of affordable housing makes poorer children more transient, and so more prone to switch schools midyear, losing progress. Higher rates of lead poisoning, asthma and inadequate pediatric care also fuel low achievement, along with something as basic as the lack of eyeglasses. Even the way middle- and lower-class parents read to their children is different, he writes, making learning more fun and creative for wealthier children.

"I would never say public schools can't do better," Mr. Rothstein said. "I'd say they can't do much better," unless lawmakers address the social ills caused by poverty.

FOR many children in America, public schools are not lacking. A 2001 international reading test put Americans ninth out of students in 35 nations. But only students in Sweden, the Netherlands and England had scores more than marginally higher than the United States average.

More important, the average score of the 58 percent of American students attending schools that were not predominantly poor surpassed that of Sweden, the top-scoring nation.

But for the 42 percent of American students attending the poor schools that are the principal target of No Child Left Behind, the inequities remain. Blacks and Latinos lag more than two years behind white students in math on national assessment tests. In reading, which is more influenced by family background, blacks and Latinos fall three years behind whites.

Yet these gaps have shrunk considerably since 1992, when blacks were 3.5 years behind whites in math. Since 1973, when the federal government began collecting these scores, black 9-year-olds have gained roughly 3.5 grade levels in math, narrowing the achievement gap, even though white scores were also rising at the same time. The cause of these improvements is unclear, although some are most likely related to state efforts, especially in the last 15 years.

A $100 million school voucher bill sponsored by Republicans gives vouchers a prominent place in next year's debate over renewing No Child Left Behind. But other voices are likely to call for a sense of responsibility for improving children's academic success that does not begin and end at the schoolhouse door.

"It can't just be a burden on the schools to do away with social inequality," said Mr. Jennings, of the education policy center. "It has to be a burden on all of us."

Published Monday, Aug. 28, 2006