It is hard to think of a federal program that is admired as much as Head Start. It is also hard to think of one that has fallen so short of its ambitions.

When the program was launched in 1965 as the key-stone of the War on Poverty, it promised to give young children from poor homes a "head start" on their schooling. But over the years, study after study has produced no convincing evidence that it actually does so. The problem isn't that we can't prepare poor children to succeed in school; it's that we've stopped trying. Head Start can't fulfill its promise in its present form. And unless Congress and the executive branch raise Head Start's aspirations, it never will.

From the beginning, Head Start was designed to be comprehensive: It included educational, social, and psychological services; nutrition and health care; programs for families; and a career ladder for paraprofessionals. It began as a summer program for 500,000 children and was expanded to reach 700,000 children in both summer and year-round settings by the late 1960s.

Unfortunately, Head Start's proponents oversold it, promising that it would quickly lift the IQ scores of poor children and break the cycle of poverty. These inflated expectations were dashed in 1969, when a national evaluation by the Westinghouse Learning Corporation and Ohio University found that the program produced no lasting cognitive or behavioral gains. (There have been hundreds of studies of Head Start since, but no comparable evaluations using a nationally representative sample.) Ardent supporters mobilized to save the program, but future growth and funding were reined in. By 1977 enrollment had dropped to about 325,000, with nearly all children in year-round programs.

Although its luster was dimmed, Head Start remained immensely popular. It's not hard to see why: In addition to whatever educational benefits it confers, Head Start also provides a range of important and popular services -- including employment for large numbers of poor parents as teachers and aides. President Bush expanded enrollment in the program from 450,000 in 1989 to 713,000 in 1993. President Clinton further expanded it to 840,000 -- about one-third of eligible children -- and now wants to increase enrollment to 1 million by 2002.

Cognitive Loss

Head Start advocates responded to the early negative evaluations of the program's educational value by quietly abandoning cognitive development as its primary goal. Today, Head Start's chief purpose is not to prepare young children for school, but to provide them with nurturing day care, medical attention, and social services and to provide some of their parents with jobs. These are certainly important purposes, but by themselves they will not narrow the large education gap between poor and middle-class children.

Head Start's institutional place in the federal bureaucracy signifies its emphasis on social services rather than education. It began life in the Office of Economic Opportunity, was moved to the Department of Health, Education, and Welfare, and is now ensconced in the Department of Health and Human Services. Its advocates consistently have opposed any suggestion of transferring the program to the Department of Education, fearing that such a move might narrow the program's scope.

Head Start has no standard curriculum for school readiness and cognitive development; in fact, it has no standard curriculum at all. Its "performance standards" are vague and lack substance. For example, each Head Start center is expected to support "emerging literacy and numeracy development through materials and activities according to the developmental level of each child." But in reality, an open-ended statement of this kind does not set any performance standard. Head Start centers receive no guidance about which skills and knowledge to teach or the best ways to teach them.

Indeed, Head Start prides itself on variability from one center to another. As the General Accounting Office has observed, "fundamental to program philosophy [is] the notion that communities be given considerable latitude to develop their own Head Start programs, an idea that has made variability a defining characteristic of the program." By design, parents must be involved in the development of each center's approach to education (imagine a hospital where patients were expected to decide which treatments were most appropriate). This rule, a remnant of the 1960s ideology of "maximum feasible participation," means that there is no uniform Head Start school-readiness curriculum, no instructional program of proven effectiveness. While local control of public schools is a firmly rooted American tradition, it is odd to see this concept enshrined in a federal program.

Head Start proponents deny that the program de-emphasizes cognitive development, but the low salaries and qualifications of the program's teachers suggest otherwise. As of the 1996-97 school year, the average Head Start teacher earned $17,802 annually. Senior teachers with 13 years of experience earned about $20,000 on average, or about $10,000 less than the starting salary of an inexperienced public school teacher in New York City. With pay that low, Head Start cannot compete for well-qualified teachers. It also suffers high turnover, as the best teachers soon leave for better paying jobs.

Educational qualifications for Head Start teachers are correspondingly low. Teachers must possess at least a child-development associate degree, which can be acquired through on-the-job training. Fewer than 30 percent have a bachelor's degree. As presently organized, Head Start virtually guarantees that the children with the greatest educational needs will have teachers with the lowest qualifications.

Health and social services and parent training certainly are important, but they should not eclipse the need to equip disadvantaged young children with the same school-readiness and learning skills their middle-class peers gain at home and in their preschools. Children enrolled in Head Start deserve well-trained teachers who can prepare them for success in school.

The Preschool Difference

Because there is no Head Start educational curriculum, it is impossible to know what children in the program are learning. And because Head Start makes a point of extreme variability among centers, it is impossible to know whether there is any consistency in what children are taught. This bizarre situation must end.

Federal officials should clearly establish school-readiness as a major goal of Head Start. They should develop a curriculum and set standards for what teachers should know and be able to do and what students are expected to learn. The Head Start classroom should be a cognitively enriched environment where children are exposed regularly to literacy, numeracy, and problem-solving activities. The aim should be to elevate Head Start from day care to preschool, without sacrificing its valuable social and medical services.

There is growing evidence that high-quality preschool makes a big difference in the lives of disadvantaged children. Studies of the Perry Preschool Program in Ypsilanti, Mich., indicate that a well-planned program -- of far higher quality than what is ordinarily offered by Head Start centers -- can have large effects on children's achievement, subsequent grade promotions, placement in regular rather than special education classes, and high school graduation. French preschools -- which are designed as schools, not day care centers-- also have reported positive results. The French programs serve nearly all French children ages 3 to 5. Their teachers and directors are highly qualified and well-paid. The French curriculum is designed to reduce the cognitive gaps between poor and advantaged children; it provides active, carefully planned experiences with language and numbers, as well as arts, crafts, and games. Affluent American parents, eager to give their children a "head start," often send them to preschools with enriched curricula. The Core Knowledge Foundation in Charlottesville, Va., has recently released a set of guidelines for preschool that reflect the best American and French practices.

What's Really Needed

Federal policymakers face a choice about the future of Head Start: Expand it without fundamental change, as the Clinton administration now plans; or change it to emphasize cognitive development and school-readiness. The first option is easier because it requires only more money. But we cannot expect much from expanding a day care program of uneven quality to even more children.

It would be far preferable to redesign the educational component of Head Start while preserving its capacity to deliver nutritional, social, and medical services to children and their families. This would mean recruiting a well-qualified staff of teachers and paying them higher salaries and establishing a national curriculum intended to prepare young children to succeed in school. To be sure, it would cost more, but what we are doing now may be even more expensive down the road. At the very least, when Congress reauthorizes Head Start in the next year or so, it should set up a significant number of high-quality demonstration projects and insist upon careful evaluation of their effects.

In the short run, change will threaten those who see Head Start as a jobs program, a way station for adults leaving the welfare rolls. In the long run, however, a reconfigured Head Start could dramatically improve the lives of the nation's most vulnerable children.

Diane Ravitch is a New York-based senior fellow with the Progressive Policy Institute. Martha Wright, a former intern at the Progressive Policy Institute, assisted the author with research.