Head Start is currently threatened with a significant funding cut. As part of the political fight over budget deficits, the House of Representatives has proposed significant funding cuts to Head Start for fiscal year 2011 (the current fiscal year) of over $1 billion. This represents about a 15% cut from the previous fiscal year. These funding cuts are estimated to reduce Head Start and Early Head Start by over 200,000 slots.
One argument made for these funding cuts largely relies on the results from the recent Head Start Impact Study. This study used random assignment, so it is worth taking seriously. This random assignment experiment compares effects for Head Start applicants of being offered enrollment in Head Start for a year, versus being denied enrollment. Those denied enrollment may of course access whatever alternative preschool and child care services they choose. In some cases, despite the intent of the experiment, the control group ends up enrolling in Head Start, although at a much lower rate than the treatment group. The control group ended up spending almost as many hours per week in non-paternal care as the treatment group, but with less reliance on Head Start for non-paternal care.
Based on the Head Start Impact Study, the average Head Start center has some test score benefits for the treatment group relative to the control group at the end of the Head Start year. But these test score effects versus the control group are in most cases no longer statistically significant at the end of first grade. The control group is able to catch up with the treatment group by making somewhat faster progress in kindergarten and first grade.
Obviously these estimated results are discouraging for Head Start advocates. Why, despite these results, does it make sense to improve Head Start rather than cut its funding?
First, the overall research findings for Head Start are much more positive than suggested by this one study. There are several research studies using good methodologies that suggest significant long-run benefits of Head Start. A study by Eliana Garces, Duncan Thomas, and Janet Currie suggests that Head Start may have significant effects in increasing educational attainment and reducing crime for some groups. This study uses a good comparison group: siblings in which one or more attended Head Start while one or more did not attend Head Start. This comparison controls for many unobserved family and neighborhood factors that are likely to affect how children develop and their later outcomes as adults.
A study by Jens Ludwig and Douglas Miller also finds evidence that Head Start has long-run effects on educational attainment. This study also uses a good comparison group. The study compares groups of counties. Head Start was started in 1965 as part of the War on Poverty. When Head Start was begun, the Office of Economic Opportunity (OEO) offered extra assistance in writing Head Start grants to the 300 poorest counties. The Ludwig/Miller study is essentially comparing counties that just made the cutoff of being one of the 300 poorest counties in 1965, with counties that just missed being one of the 300 poorest counties. Counties that just made the cutoff of being one of the 300 poorest were much more likely to have more Head Start activity in Head Start’s initial phases, and through the late 1970s.
The Ludwig/Miller study looks at what happens to these counties in later years. In later years, the cohorts that could have been involved with Head Start had higher educational attainment in the counties that just made the cutoff for being one of the 300 poorest, compared to similar cohorts in counties that just missed this cutoff.
A study by Deming finds evidence of long-term Head Start effects despite short-run fading of test score effects. This study also uses as a comparison group siblings who did not attend Head Start. This study found long-run effects of Head Start on educational attainment. However, the test score effects of Head Start faded considerably (for example, two-thirds of these effects disappear by middle school) as students progressed in the K-12 system.
Therefore, Head Start may have long-term effects on important adult outcomes. These effects may occur even with some fading of test score effects. These long-term effects may be due to Head Start effects on soft skills that are not fully captured by effects on test scores.
Second, there is good evidence that at least some Head Start programs have considerably larger effects than those estimated for the average head Start Center in the recent Impact Study. For example, Gormley et al’s estimates for Head Start in Tulsa, Oklahoma suggest effects of Tulsa Head Start on test scores at kindergarten entrance that are over twice the effects in the Head Start Impact Study.
These findings can be reconciled. This reconciliation is speculative, not proven. Head Start may have had large effects relative to available alternatives when first implemented in the 1960s and 1970s. However, since then, state pre-k programs and private preschool programs have expanded significantly in availability and quality. All of these Head Start impact studies are really not comparing Head Start with nothing, but rather with whatever mix of non-paternal care and paternal care are chosen by non-participants. If the mix of Head Start alternatives has improved over time, whereas many Head Start centers have not, then the relative effects of the average Head Start Center may have declined somewhat over time. In addition, it could be that Head Start’s more important effects are on long-term behavior and health, not on test scores.
In my opinion, this evidence argues for improving Head Start rather than reducing its funding. We need to improve Head Start Centers so that we can match the larger short-term test score and other benefits obtained by the better Head Start Centers and pre-k programs. Head Start has too much evidence of long-term effects for significant cutbacks to make sense. Furthermore, preschool education in general has too strong evidence of effects for it to make sense to begin dismantling the largest federally-funded preschool program.
The U.S. Department of Health and Human Services is making significant efforts to improve Head Start. For example, HHS is in the process of developing procedures to identify the lowest performing 25% of all Head Start Centers. The contracts for such Centers will then be required to be rebid.
More far-reaching reforms might allow some states, with rigorous oversight, greater flexibility in coordinating Head Start and other funds to create a more coordinated early childhood education system. Some ideas on how to do this have been developed by Steve Barnett and Ron Haskins.
Other recent arguments for reforming Head Start rather than defunding it include Kathleen McCartney, dean of the Harvard Graduate School of Education. Another recent statement opposing cuts to Head Start comes from Lauren Moore at the blog run by the publication The Future of Children.
One perennial problem in American public policy is that a single negative study seems sufficient to motivate efforts to kill a program rather than reform it. Most educational and social programs at some point may be negatively evaluated versus their alternatives. If the overall logic of the program makes sense, such negative evaluations should be used to motivate reforms rather than kill the program. If any negative evaluation is an excuse to kill a program, why would any program operator want to participate in evaluations? A political climate in which evaluations are only weapons to defund programs, rather than to improve programs, is not likely to lead to continuous improvement in public program quality.