A SPLIT CONGRESS GRAPPLES WITH ENIGMATIC HEAD START

Saturday, August 2, 2003

A sharply divided Congress is currently debating what to do with Head Start, but unfortunately, they haven't begun by admitting that its usefulness is vastly overstated by its advocates.

Strange, isn't it, that a program can operate for 38 years, enroll more than 21 million children, be spending $6.5 billion dollars a year, enjoy third-rail status in Congress -- and the experts are still arguing about whether it does any good?

Well, actually it isn't strange. It's just a federal program, but that's a different column.

Most studies of Head Start have shown that it provides a brief boost in academic performance, but that the effects fade out by third grade. If that conclusion is correct it means every dollar of the billions spent on Head Start since its beginning has been wasted.

Steven Barnett, director of the National Institute for Early Education Research, says the conclusion is not correct, and in a 1993 article in Education Week he explained why. Most studies show, he says, that preschool programs generally, including Head Start, have lasting effects on grade retention, placement in special education and high school graduation. It's primarily results on achievement tests that fade, or appear to.

Longitudinal studies usually compare the academic performance of a group of children who were in the program with an otherwise similar control group of children who were not, and often they are based on tests routinely administered at grade level. That means, Barnett says, that over time the studies gradually exclude from their samples children who are left back or move to special education.

But if children never in Head Start are more likely to be left back, or to move to special education, then over time the comparison groups diverge, with weaker students differentially likely to drop out of the control group.

``Studies without these design flaws,'' Barnett says, ``find persistent effects on achievement test scores'' as well.

He has interesting corroborative evidence from children's ages. ``At first grade, the Head Start and comparison groups are roughly the same age,'' he says. ``At second grade, the Head Start group is significantly younger, and the age gap widens at third grade.''

I can't say whether Barnett is correct, but he has raised a methodological point that needs to be addressed, and furthermore he raised it a long time ago. Yet in testimony before Congress, Marnie Shaul of the General Accounting Office said there has been no definitive federal study of Head Start's effectiveness, though one is expected to be done by 2006.

``The 'fade-out' myth,'' Barnett says, ``has led to another questionable notion, that continued intervention beyond Head Start is needed to maintain its benefits.'' There's little evidence for that, he says, though better education for disadvantaged children might independently lead to better results.

He also debunks that idea that preschool programs can have a lasting effect on IQ. There is an early bump -- about 12 points in some very intensive programs, which would be highly significant if it lasted. But it doesn't, although that was an early hope of Head Start proponents.

Of course, no one should have expected it to last, given that an even more intensive intervention in children's lives -- adoption -- has no lasting effect on IQ either. Adopted children, as adults, resemble their birth parents in IQ just as much as children raised by their birth parents, and resemble their adoptive parents no more than by chance.

But IQ is by no means the only factor in academic success. If one looks at just a narrow slice of the IQ distribution, say from 95 to 105, people will be bunched around some average performance, but the range will be from high school dropouts to people with graduate degrees. A higher IQ slice might have more Ph.D.s and fewer GEDs, and a lower slice the opposite, but achievement within a slice could obviously be higher than it is.

Barnett, a supporter of Head Start, believes it could contribute to that goal if it had enough money to attract teachers with a college degree and training in early childhood education. It would, he estimates, add about $1 billion to the program's cost, and could be phased in over several years so that current teachers could get their degrees.

Maybe so. But it might also be true that if additional money is spent, bigger benefits will come from spending it instead on getting more children into the program.

Those are the issues Congress is wrestling with, because there are no clear answers. And after 38 years, there ought to be.