This proposal seeks support for an interdisciplinary team to collect and analyze data to estimate the impacts on participants through midlife, and the impacts on their children, of the two most influential early childhood education programs: the HighScope Perry Preschool Program (PPP) and the Carolina Abecedarian Project (ABC). Both programs targeted disadvantaged, predominantly African-American children. Evidence from these programs is the cornerstone of wide-scale initiatives for pre-kindergarten programs around the world. Both programs were evaluated by RCTs and have, by far, the longest follow-ups of any experimentally evaluated early childhood programs (PPP through age 40; ABC through age 34). These programs are influential because they have long-term follow-ups that show persistent beneficial effects through young adulthood. We are in the final stages of completing a NIH-supported follow-up survey of PPP at age 50. We seek support to conduct a similar follow-up for ABC through age 45. The proposed study will collect data in the format used in the Health and Retirement Study (HRS), which is also being used in collecting the data for the PPP at 50 study. Both studies will include measurements of (i) physical and mental health, including metabolic and cardiac risk factors, and depression; (ii) income and employment; (iii) cognitive skills and executive functioning; (iv) noncognitive skills, including grit and perseverance among other personality measures; and (v) education and health outcomes of participants? children. For both interventions, we will have self-reported and administrative data on earnings, employment, use of social services, criminal activity, and self- reported and direct assessments of health and cognition. We also seek support to analyze the PPP and ABC data using the same methods. Our analyses of PPP and ABC will account for (a) their small samples; (b) departures from initial randomization protocols; (c) the effects of testing multiple hypotheses on true p-values; (d) non-random attrition; and (e) contamination bias that arises when control groups use alternative forms of early childcare. We will analyze both data sets using a variety of different longitudinal approaches. We will (i) conduct dynamic mediation analyses to explain the sources of treatment effects; (ii) relate experimental estimates to the non- experimental estimates of the processes of skill formation for comparable populations; (iii) use the experimental data in conjunction with non-experimental data to assess the validity of commonly used short-term indicators utilized to predict long-term outcomes in more recently launched interventions lacking long-term follow-up. In addition, we will combine experimental data and comparable non-experimental data to conduct rate of return and cost-benefit analyses that account for the full array of benefits and costs of the interventions, projecting the out-of-sample benefits for treatments and controls. All data sets will be made publicly available at the completion of the project.
Project Relevance: Current evidence from PPP and ABC shapes understanding of the effectiveness of early childhood programs. The proposed study will examine a broad range of impacts of both programs through midlife facilitating understanding of the mechanisms through which ECE influences a variety of midlife outcomes.
