The pros and cons of early years programs – where to start!

Smart… Safe… Best… Sure… there are so many ways to 'Start' these days, but once there was only Head Start. Since its inception in 1965 the program has been the most extensively applied and most heavily researched prevention effort in the US. About a million economically deprived children in the country currently receive its support at an annual cost of about $7 billion.

The results of evaluations of Head Start are notably mixed, particularly when compared to those from 'flagship programs' such as Abecedarian, the Chicago Child-Parent Centers and High/Scope Perry Preschool. Meanwhile, claims continue to be made for the benefits of pre-K schooling over early years programs. So where should a policy maker put the money?

In a recent edition of Social Policy Report, a regular publication of the Society for Research on Child Development [1], Jens Ludwig and Deborah Phillips use benefit-cost analysis to argue for Head Start. Their report, The Benefits and Costs of Head Start [2], shows that gains to participants and society have consistently exceeded the costs.

Short-term benefits of Head Start are well established, and are measured mainly in terms of academic skills, literacy, vocabulary and numeracy. Effect sizes of between 0.1 and 0.2 are common in recent robust experimental evaluations. But the equally abundant evidence about long-term gains is more shaky.

Economists Eliana Garces, Duncan Thomas and Janet Currie from the University of California Los Angeles compared children receiving Head Start with siblings who did not. They found significant benefits in terms of high school completion (22 points), college attendance (19 points) and involvement in crime (12 points).

Their study also uncovered significant differences by ethnic group. And, crucially, they detected the impact of funding: children going to better resourced Head Start centers make greater gains. The more you put in the more you get out. So the question is, how much more investment is needed?

Some sense of the answer comes from comparing the impact of Head Start and the High/Scope Perry Preschool project on vocabulary (as measured using the Peabody Picture Vocabulary Test). Head Start shows early benefits of about 0.25 of a standard deviation but these fade within three or four years. They are also less for African American children.

The impact of Perry Pre-school is much greater, about 0.9 of a standard deviation. And the impacts seem to last longer. The total cost per child of Perry Pre-school is roughly twice that of Head Start.

So could resources invested in Head Start be used better elsewhere? The results are not clear cut and readings of the evidence vary. Doug Besharov, Professor at the Maryland School of Public Policy and Senior Scholar at the American Enterprise Institute, for example, described the results of the Head Start national impact study as 'disappointingly small'; while Hirokazu Yoshikawa from New York University finds them 'consistently positive' and 'impressive'.

The difference of opinion leads the authors of “The Benefits and Costs of Head Start” into a discussion about the nuances of different evaluation strategies. Most agree that experimental evaluation is the best way to understand impact of prevention programs on health and development, but what kind of experiment?

Most evaluations of Head Start use 'Intention to Treat' strategies, in which all those randomly allocated to the program group are compared with controls, regardless of whether they actually started the program. By comparing only those who go onto the program with those who do not, what they call 'Treatment on the Treated' effect sizes rise by 0.1 and sometimes more.

These methodological considerations may pass most policy makers and practitioners by, but they may be important in deciding where scarce US prevention dollars are invested.

However, the conclusion of this discussion might be more simply put. First, no prevention program has sufficient 'reach'. Head Start is the most pervasive and well established of early years programs. It reaches just a million among 20 million aged four years or younger, of whom about four million live below the poverty line, drawn at its lowest level..

Second, investment in Head Start and in most other prevention programs is insufficient. One can see why scientists can get drawn into discussions about relative economic merits of pre-K versus Head Start or other prevention programs. Resources are scarce; policy makers have to make difficult decisions about what to buy. But a more rational strategy that invests early for later benefits would make these comparative questions irrelevant.

Finally, the wealth of evidence on Head Start is a reminder of how much there is still to learn. The Ludwig and Phillips paper alone raises important questions about understanding effect size, better comprehending the relationship between short and long term effects, and exploring more the relative merits of different evalation strategies.

See also: Can several thin reeds ever make a strong policy boat? [3], Getting the measure of the 'mileage' of preschool care [4] and Pay now, break out of poverty later [5]

References

Ludwig J and Miller D, "Does Head Start improve children's life chances?" Quarterly Journal of Economics, 122, 1, 159-208.

Puma, M., Bell, S. and colleagues (2005). "Head Start Impact Study: First year findings". Washington, DC: U.S. Department of Health and Human Services.

Graces E, Thomas D and Currie J, "Longer-term effects of Head Start" American Economic Review, 92(4), 999-1012