This study investigates the persistent academic impacts of the Head Start program, a federal government-funded early childhood intervention, using data from the Early Childhood Longitudinal Study-Kindergarten Cohort (ECLS-K). Bayesian Additive Regression Trees (BARTs) are the primary methodology used, and average, conditional, and individual-level treatment impacts on children’s mathematics achievement are estimated. BART estimates a negative Average Treatment Effect (ATE) of ？1.5421 with increasingly larger adverse effects for children with higher Socioeconomic Status (SES), suggesting diminishing marginal returns. This finding demonstrates the strength of BART to detect nonlinear moderation patterns that are evasive to conventional models. It also implies that Head Start and other preschool interventions will yield greater policy returns when targeted at low-SES children, in order to enable more efficient and fair distribution of public funds. For comparison, Causal Forest estimates a larger ATE (？2.4340) and determines SES to be the overarching moderator, while Propensity Score Matching offers a conservative estimate (？1.2606) without considering effect heterogeneity. These findings underscore the utility of BART in estimating subtle, SES-varying effects of Head Start, and suggest the potential value of more targeted intervention strategies guided by adaptive causal inference.
References
[1]
Holland, P.W. (1986) Statistics and Causal Inference. Journal of the American Statistical Association, 81, 945-960. https://doi.org/10.1080/01621459.1986.10478354
[2]
Rosenbaum, P.R. and Rubin, D.B. (1983) The Central Role of the Propensity Score in Observational Studies for Causal Effects. Biometrika, 70, 41-55. https://doi.org/10.1093/biomet/70.1.41
[3]
Robins, J.M. and Rotnitzky, A. (1995) Semiparametric Efficiency in Multivariate Regression Models with Missing Data. Journal of the American Statistical Association, 90, 122-129. https://doi.org/10.1080/01621459.1995.10476494
[4]
Hill, J.L. (2011) Bayesian Nonparametric Modeling for Causal Inference. Journal of Computational and Graphical Statistics, 20, 217-240. https://doi.org/10.1198/jcgs.2010.08162
[5]
Chipman, H.A., George, E.I. and McCulloch, R.E. (2010) BART: Bayesian Additive Regression Trees. The Annals of Applied Statistics, 4, 266-298. https://doi.org/10.1214/09-aoas285
[6]
Michelmore, K. and Dynarski, S. (2017) The Gap within the Gap: Using Longitudinal Data to Understand Income Differences in Student Achievement. AERA Open, 3, 1-18.
[7]
Hahn, P.R., Murray, J.S. and Carvalho, C.M. (2020) Bayesian Regression Tree Models for Causal Inference: Regularization, Confounding, and Heterogeneous Effects (with Discussion). Bayesian Analysis, 15, 965-1056. https://doi.org/10.1214/19-ba1195
[8]
Sparapani, R.A., Logan, B.R., McCulloch, R.E. and Laud, P.W. (2016) Nonparametric Survival Analysis Using Bayesian Additive Regression Trees (BART). Statistics in Medicine, 35, 2741-2753. https://doi.org/10.1002/sim.6893
[9]
Kapelner, A. and Bleich, J. (2016) bartMachine: Machine Learning with Bayesian Additive Regression Trees. Journal of Statistical Software, 70, 1-40. https://doi.org/10.18637/jss.v070.i04
[10]
Linero, A.R. (2018) Bayesian Regression Trees for High-Dimensional Prediction and Variable Selection. Journal of the American Statistical Association, 113, 626-636. https://doi.org/10.1080/01621459.2016.1264957
[11]
Taddy, M., Gardner, M., Chen, L. and Draper, D. (2016) A Nonparametric Bayesian Analysis of Heterogenous Treatment Effects in Digital Experimentation. Journal of Business & Economic Statistics, 34, 661-672. https://doi.org/10.1080/07350015.2016.1172013
[12]
Wager, S. and Athey, S. (2018) Estimation and Inference of Heterogeneous Treatment Effects Using Random Forests. Journal of the American Statistical Association, 113, 1228-1242. https://doi.org/10.1080/01621459.2017.1319839
[13]
Denton, K. and West, J. (2002) Children’s Reading and Mathematics Achievement in Kindergarten and First Grade. NCES 2002-125. U.S. Department of Education and National Center for Education Statistics.
[14]
West, J., Denton, K. and Germino Hausken, E. (2000) America’s Kindergartners: Findings from the Early Childhood Longitudinal Study, Kindergarten Class of 1998-99. NCES 2000-070. U.S. Department of Education and National Center for Education Statistics.
[15]
Najarian, M., Sorongon, A.G., McManus, J.K., Walston, J.T. and Hagedorn, M.C. (2018) Early Childhood Longitudinal Study, Kindergarten Class of 2010-11 (ECLS-K:2011), User’s Manual for the ECLS-K:2011 Kindergarten-Fifth Grade Data File and Electronic Codebook (NCES 2018-032). U.S. Department of Education and National Center for Education Statistics.
[16]
Tourangeau, K., Nord, C., Lê, T., Sorongon, A. and Najarian, M. (2015) Early Childhood Longitudinal Study, Kindergarten Class of 2010-11 (ECLS-K:2011), User’s Manual for the ECLS-K:2011 Kindergarten-Fourth Grade Data File and Electronic Codebook (NCES 2015-073). U.S. Department of Education and National Center for Education Statistics.
[17]
Athey, S., Bayati, M., Doudchenko, N., Imbens, G. and Khosravi, K. (2020) EconML: A Python Package for ML-Based Heterogeneous Treatment Effect Estimation. Microsoft Research. https://github.com/microsoft/EconML
[18]
Heckman, J.J. and Karapakula, G. (2019) Intergenerational and Intragenerational Externalities of the Perry Preschool Project. National Bureau of Economic Research. https://www.nber.org/papers/w25889
[19]
Bailey, D., Duncan, G.J., Odgers, C.L. and Yu, W. (2017) Persistence and Fadeout in the Impacts of Child and Adolescent Interventions. Journal of Research on Educational Effectiveness, 10, 7-39. https://doi.org/10.1080/19345747.2016.1232459
[20]
Heckman, J.J. (2006) Skill Formation and the Economics of Investing in Disadvantaged Children. Science, 312, 1900-1902. https://doi.org/10.1126/science.1128898
[21]
Kline, P. and Walters, C. (2015) Evaluating Public Programs with Close Substitutes: The Case of Head Start. National Bureau of Economic Research. Working Paper Series. https://www.nber.org/papers/w21658
