HOW CAN POOR children be helped to do better in later life? The question has, correctly, dogged Head Start and other early childhood development programs from the beginning. One early study (Westinghouse-Ohio University, 1969) suggested that the results imparted by Head Start did not last. But now a new study, done by Irving Lazar and his associates at Cornell University, presents important follow-up evidence.At least for some intensive experimental projects, benefits last for a number of years -- in some cases throughout high school.

Standard measures, like IQ scores, have shown the now predictable pattern of an early spurt followed by continuous slippage so that in a few years no difference can be found between those who have participated and those who have not. But by another test, the children from the experimental projects succeed more often. Compared with similar kids in the same schools who have not been in the special preschool programs, they have not been put at the same rate inot special classes for the retarded or for children with behavior disorders. They were generally promoted on time, while the control group children were not.

These findings make clear that some positive effects of good early childhood projects have held up over a period of years. Two of the projects were part of Head Start; all were intensive, research-oriented, and received the concentrated attention of those in charge. What is needed now is not just for a the children from a dozen carefully controlled projects to be proceeding acceptably in school. This should happen for every child who could benefit.

Head Start, like other human service programs, has been bedeviled for years by the difficulty of spreading to all localities the intensity, concern and intellectual quality found in particular projects that work well. The curriculum and the daily schedule, the number of children and the teacher training are often the same, but the results don't measure up. The problem is that little systematic analysis has been applied to sorting out the personal, organizational and neighborhood characteristics of successful projects.

This is the area in which research should now be concentrated. We need the ability to translate local success to a national scale in order to give all poor children as good a head start as some projects already are managing to impart.