Association for Public Policy Analysis & Management

In the United States, preschool participation has risen dramatically from less than 20% in 1965 to nearly 70% in 2013 (Barnett & Yarosz, 2007; National Center for Education Statistics, 2015). Correspondingly, more children today have a parent who attended preschool than ever before. The current study seeks to understand the role of parents’ early experiences in their children’s own school experiences, academic achievement, and well-being and determine whether preschool attendance can benefit both parents and children in the short-term and across generations.    

Chor (2016) demonstrates with experimental Head Start Impact Study (HSIS) data that Head Start induces larger and longer-lasting cognitive and socioemotional benefits for children in “multigenerational Head Start families” in which both mother and child attend the program compared to first-generational program participants, whose gains largely fade out by kindergarten. These findings are suggestive of long-term intergenerational benefits of Head Start participation. However, the ability to make causal inference about the intergenerational impacts of Head Start using HSIS data is limited. Simultaneous two-generation randomization, or an analytic approach mimicking random assignment, is needed to determine the causal effect of multigenerational program participation on children’s educational outcomes.

The current study assesses the intergenerational impacts of parents’ Head Start and other preschool participation. The study uses two nationally representative datasets that allow for: (1) replication and extension of Chor’s previous findings; (2) consideration of additional and longer-term child outcomes including school quality, educational attainment, and adolescent psychological and physical health; and (3) application of a strong multigenerational causal framework not possible with HSIS data. The study first leverages the multigenerational data of the Panel Study of Income Dynamics to estimate marginal structural models using the joint probability of parent-child participation as inverse probability weights. Next, the study draws on rich longitudinal National Longitudinal Survey of Youth data to apply a combination of propensity score matching and fixed effects modeling.   

The study will examine a key early childhood education program for low-income children—Head Start—within the context of multigenerational poverty. Large numbers of children experience multigenerational poverty, which may have more negative consequences for their early development than single-generation poverty (Gans, 2011). The issue of multigenerational Head Start participation is of particular policy relevance, as one-quarter of children enrolled in Head Start in 2002 came from multigenerational Head Start families, or families in which two consecutive generations (mother and child) participated in the program (Chor, 2016). There are even more multigenerational preschool participant families, whose outcomes the study will also investigate. The study will inform researchers, policymakers, and practitioners about strategies to promote the educational opportunities and life chances of children raised in poor and low-income households. An assessment of the impacts of multigenerational Head Start participation can inform program decisions such as how to meet participants’ individual needs and influence policymakers’ assessment of the Head Start program, which may yield intergenerational benefits that have yet to be explored. The study will also contribute to the debate over how to best to allocate resources towards children and/or their parents.