Evaluating Head Start: A Survey
Have you ever wondered what exactly your tax dollars are doing? Or, perhaps, whether they are being put to the best possible use? The field of public economics studies exactly that: the efficiency of government policies.
The government budget, however, is massive, totaling 6.82 trillion dollars in the U.S. in 2021. This makes it next to impossible to study all of it at once. In this article, we focus on the area of early childhood education (e.g. preschool and other services for children below age 5). This field is important because children are developing rapidly at this time, and the environments they are exposed to can have echoing impacts on the rest of their lives. For instance, consider the Zone of Proximal Development (ZPD), which refers to a “middle ground” of skills that children cannot learn on their own, but can master with help1. The primary government investment in early childhood education is Head Start, a program over fifty years old that mainly targets disadvantaged youth, providing them with both preschool and health services (medical, dental, and nutrition).
History of Head Start
Head Start began in January of 1964, when President Lyndon B. Johnson gave his first State of the Union address, in which he declared, “unconditional war on poverty in America.” He later called for, “special school aid funds as part of our education program.” Johnson's “War on Poverty” was put into motion with the Economic Opportunity Act of 1964, which contained the initial funds for Head Start. Spearheaded by the Office of Economic Opportunity (OEO), in the summer of 1965, over 500,000 children participated in the first summer edition of Head Start. That fall, the first full-year programs were launched, enrolling over 20,000 students (Smith and Bissell 1970).
The development of Head Start was motivated by several factors. First, the establishment and early success of model programs such as the Perry Preschool Project and the Early Training Project encouraged further development of early education programs (like Head Start). The relationship between model programs and Head Start is covered in Harvard researchers Marshall Smith and Joan Bissell’s analysis of one of the first evaluations of Head Start. According to Smith and Bissell, “[e]ach [model program] reported that carefully designed and implemented programs increased the cognitive performance of disadvantaged children”. Second, there was a growing belief that issues within the American education system could be addressed during the years just before elementary school. In particular, many studies at the time found that children of minorities and low-income families entered school already behind their peers. Third, new psychological research in the early 1960s pointed towards the malleability of intelligence, especially at early ages (Hunt 1961, Bloom 1964). Lastly, the program drew political popularity. Certain programs within the OEO might have been politically contentious, but Head Start was liked by most Americans, which helped facilitate its fast beginnings (Smith and Bissell 1970).
Over 200,000 children were enrolled in the fall of 1967. At this point, a total of about two million students had participated in Head Start. Since then, Head Start has received repeated reauthorizations (typically for five year periods) and steady funding increases from Congress. In 1984, Head Start reached nine million total children served, and in 1994, Early Head Start was established. In 2000, total annual enrollment climbed to 900,000 and held steady into the 21st century (Office of Head Start).
Studying Head Start: Early Evaluations
There have been numerous evaluations of Head Start, beginning in 1969 with a joint effort between the Westinghouse Learning Corporation and Ohio University. The report found summer programs did not change cognitive or affective development, while full-year programs had some effect on cognitive development. However, several issues have been identified with the report’s methodology. For instance, the OEO asked for studies of overall effectiveness, not more specific measures, making it difficult to draw precise conclusions. Furthermore, the selection procedure was flawed, and the study did not match Head Start and non-Head Start participants for comparison. As part of the 1998 reauthorization of Head Start, Congress asked for another study of the effectiveness of Head Start, which resulted in the Head Start Impact Study (HSIS), a randomized control trial of about 5,000 individuals. The results of the trial found minor cognitive improvements that “faded out” by kindergarten, and no non-cognitive changes (Kline and Walters 2016).
Studying Head Start: Modern Techniques
Early evaluations focused mainly on IQ and test scores in an attempt to measure cognitive improvement caused by Head Start. A newer strand of literature employs quasi-experimental methods to approach Head Start from a different angle. These studies often broaden their results of interest to include factors such as educational attainment (e.g. high school graduation), annual adult income, and involvement with crime. They often analyze data on individuals provided by the Panel Study of Income Dynamics (PSID) and the National Longitudinal Survey of Youth (NLSY), coupled with data on Head Start's availability or funding amounts (at the county level). Most studies are quasi-experimental and draw from three established strategies to estimate the causal effects of Head Start: the plausibly random rollout of Head Start, sibling comparison (comparing individuals who went to Head Start with their sibling who did not), and policy discontinuity (focusing on the extra funding effort for the poorest 300 counties from the OEO).
Garces et al. (2002) uses a supplement to the PSID specifically designed for this study. The 1995 PSID has a retrospective question on Head Start enrollment. They adopt a sibling comparison design, which shrinks their sample size to approximately 300-1700, depending on the specification. They study four outcomes: high school graduation, college enrollment, earnings in the early twenties, and crime, finding benefits in each category depending on the race of the enrollee. For instance, white enrollees have a 20.3% greater chance of graduating high school, and African-American enrollees are 11.6% less likely to be booked or charged with a crime.
In contrast, Bailey et al. (working paper) uses a much larger sample than other papers (about 22.5 million children) by using the Census/American Community Survey in conjunction with the SSA Numident records. This data also differs from others in that it lacks family characteristics to use as controls. They study the long-run effects of Head Start and find positive effects on educational attainment (2.7% increased chance of high school graduation) and economic self-sufficiency.
Next, Deming (2009) studies the children of mothers who were surveyed in the original NLSY79. These children largely enrolled in Head Start between 1984 and 1990 and participated in the Child NLSY (CNLSY). He follows these participants up to 2004, at which point they are at least 19 years old. Deming uses a sibling-based framework to account for family fixed effects and finds significant benefits to a variety of adult outcomes, including high school graduation (8.6% increased probability) and health (7% less likely to suffer from poor health). Notably (in comparison to other studies), he does not find a significant reduction in criminal activity.
Combining the NLSY79 with data on the availability of Head Start (from the Community Action Programs and Federal Outlay System), Barr and Gibbs (2019) study both effects on the female enrollees into Head Start and their children. This differs from Deming (2009) in that there is no restriction that the children of these female enrollees also enrolled in Head Start. They capture intergenerational effects by capitalizing on the plausibly random rollout of Head Start and supplement this with a second strategy following the policy discontinuity framework. They find a positive impact on the second generation in several outcomes, such as a 13.9% increased chance of graduating high school and a 15% lower chance of criminal activity (arrest, conviction, or probation).
Conclusion
Head Start has been a core component of the American early education program for over fifty years. Funded by federal grants, it provides a variety of services on top of preschool instruction to disadvantaged youth and their families. There have been many attempts to study Head Start and quantify its effectiveness. Early studies were criticized for poor experimental design, and their results are controversial. Newer studies exploit experimental studies (HSIS) or use quasi-experimental methods and various sources of randomization. Although the findings of these studies are mixed, there seems to be a general direction of Head Start providing benefits to children, both in the cognitive and non-cognitive sense.
Written by David Yu, UCLA Undergraduate Economics Student
Endnotes
The theory of the Zone of Proximal Development was first proposed by Vygotsky (1978) and empirically supported by Wood and Middleton (1975). This theory claims that any skill falls into one of three categories: known, unknown, or the ZPD. Children can learn skills in the ZPD if and only if a more skilled individual is there to teach them. For a complete discussion of these ideas, refer to the field of developmental psychology.
References
Bailey, Martha J., Shuqiao Sun, and Brenden Timpe. 2021. "Prep School for Poor Kids: The Long-Run Impacts of Head Start on Human Capital and Economic Self-Sufficiency." American Economic Review, 111 (12): 3963-4001.
Barr, Andrew and Chloe R. Gibbs. “Breaking the Cycle? Intergenerational Effects of an Anti-Poverty Program in Early Childhood.” Conditionally accepted, Journal of Political Economy.
Bloom, Benjamin S. Stability and Change in Human Characteristics. New York: Wiley, 1964. Print.
Deming, David. 2009. "Early Childhood Intervention and Life-Cycle Skill Development: Evidence from Head Start." American Economic Journal: Applied Economics, 1 (3): 111-34.
Garces, Eliana, Duncan Thomas, and Janet Currie. 2002. "Longer-Term Effects of Head Start ." American Economic Review, 92 (4): 999-1012.
Hunt, J. M. (1961). Intelligence and experience. Ronald.
Patrick Kline, Christopher R. Walters, Evaluating Public Programs with Close Substitutes: The Case of Head Start, The Quarterly Journal of Economics, Volume 131, Issue 4, November 2016, Pages 1795–1848, https://doi.org/10.1093/qje/qjw027
Smith, M. S., & Bissell, J. S. (1970). Report analysis: The impact of Head Start. Harvard Educational Review, 40(1), 51–104. https://doi.org/10.17763/haer.40.1.f3h6164m816t2552
Vygotsky, Lev Semenovich, and Michael Cole. Mind in society: Development of higher psychological processes. Harvard university press, 1978.
Wood, D. and Middleton, D. (1975), A Study of Assisted Problem-Solving. British Journal of Psychology, 66: 181-191. https://doi.org/10.1111/j.2044-8295.1975.tb01454.x