Learning from the variation in the effectiveness of Head Start
Economists and policymakers alike are increasingly realizing the vital importance of children’s early development on not only the future success and prosperity of the child, but that of the United States as well. The United States, of course, has a widely implemented, low-cost early-childhood education program, known as Head Start, but studies of the program don’t find it to be nearly effective as the “high-equality” programs such as the Perry Preschool program, a project that provided high-quality, high-cost preschool to a small group of low-income children.
Now, though, a new National Bureau of Economic Research working paper shows that there are large variations in the methods and resources of different Head Start programs as well as differences in the short-term effectiveness of different programs. These variations can help us better understand how to best improve early childhood programs for the broad swath of children in the United States
The new paper, by Christopher Walters of the University of California-Berkeley, uses data from the Head Start Impact study, which randomly assigned students to Head Start centers to measure the effectiveness of the program. The headline result is that Head Start programs barely outperformed the control group of standard childcare centers. Yet Walters finds substantial variation in the outcomes of centers and their methods and means—findings that point to ways the program can be improved upon.
When it comes to the variation of outcomes, Walters finds that the variation of test scores is larger than the usual dispersion found for teachers or schools. He estimates that a move to a Head Start center that is one standard deviation above average in test scores would result in a $3,400 boost in adult-earnings per child.
Why do these differences arise? Walters looks at differences in inputs for each program, specifically inputs that have been credited with making the Perry Preschool program effective. Specifically, these inputs are teacher education, class size, instruction time, the use of the so-called High/Scope curriculum, and home visiting by staffers. Walters finds that Head Start centers with full-day instruction do a better job of boosting cognitive skills (essential test scores), while centers that include home-visiting are very good at boosting non-cognitive (social) skills. Teacher education, class size, and the use of the High/Scope curriculum (a widely cited reason for the success of the Perry program) are not correlated with better outcomes.
Walters’ findings build significantly on other recent research. University of Chicago economist and Nobel laureate James Heckman has extolled the virtues of investing early in children for years. And a study led by Harvard University economist Raj Chetty finds that a high-quality kindergarten teacher has long-term effects on adult outcomes such as earnings and health status. But many of these findings, particularly for those looking at the effects of pre-kindergarten, show the effect of high-quality programs that are hard to replicate or scale-up for a large population.
Clearly, the United States would be better off if all Head Start programs were very effective at improving the adult outcomes of disadvantaged children. But for now, we’ll have to use the variation in outcomes and inputs such as those identified by Walters to better understand how to improve our current efforts.