Working Papers

Clark, L. (2023). Sorting into Schools by Unobserved Disadvantage. Available upon request.

Yoo, P., Domina, T., McEachin, A., Clark, L., Hertenstein, H., Penner, A. (2023). Virtual Charter Students Have Worse Outcomes as Young Adults. Census Working Paper CES 23-32. 

Clark, L., Cole, A., Eng, A., Meiselman, B., Pharris-Ciurej, N., Pierce, K., Voorheis, J. (2023). The Demographics of the Recipients of the First Economic Impact Payment. Census Working Paper CES 23-24.  America Counts story.

Akee, R., Clark, L. (2023). Universal Preschool Lottery Admissions and Its Effects on Long-Run Earnings and Outcomes. Census Working Paper CES 23-09.

Spiegel, M., Clark, L., Domina, T., Radsky, V., Yoo, P., Penner, A. (2022). Measuring School Poverty: An Exercise in Convergent Validity. Census Working Paper CES 22-50.

Radsky, V., Domina, T., Clark, L., Bhaskar, R. (2022). Stigma Free Lunch: School Meals and Student Discipline. Census Working Paper CES 22-23.

Research in progress

Using Administrative Measures of Family Income for School-Aged Children, with Renuka Bhaskar

Race/Ethnicity, Sex, and the Relationship between Educational Attainment and Childhood Family Income, with Jennifer Ortman, Nikolas Pharris-Ciurej, and John Voorheis

Local Variation in the Intergenerational Transmission of Income, with Dan Black, Seth Sanders, Lynne Schofield, and Lowell Taylor

Dissertation Chapters

Disadvantage Beyond Poverty: Adverse Childhood Experiences, School Choice, and Educational Outcomes

Abstract: Adverse home and family circumstances play an important role in shaping children’s educational outcomes, but typical administrative data cannot identify these disadvantages beyond a simple household income proxy (e.g., free or reduced lunch eligibility). By linking human service records to school data, this paper identifies a subset of more severely disadvantaged children: those linked to child welfare investigations prior to kindergarten (one-in-eight children in the city under study). Regardless of which elementary school they attend, these children miss 25% more school days than non-disadvantaged children and have suspension odds that are 57% higher. These gaps are significantly larger than those for low-income children not linked to investigations. Moreover, a multinomial choice model of kindergarten enrollment shows that children linked to child welfare investigations are systematically less likely to enroll in charter or magnet schools over traditional public schools, in contrast to other low-income children who are nearly as likely as their counterparts to enroll in charter or magnet schools. Thus, in aggregate at the school level, a handful of traditional public schools disproportionately enroll the most disadvantaged students. By carefully controlling for these sorting patterns and exploiting quasi-random variation in peer group composition within school-grades over time, I recover causal peer effect estimates, which show that having more disadvantaged peers significantly increases students’ own suspension probabilities.

An Intervention to Reduce Chronic Absenteeism

Abstract: This paper undertakes a quasi-experimental evaluation of a truancy prevention program that was piloted in two urban K-8 schools during the 2013-14 school year. The program offered voluntary and customized services to students deemed at-risk for truancy. Using a triple-differences specification, I find that the program was successful in reducing chronic absenteeism rates among persistently low-income students, but the treatment effect is only marginally significant. These results suggest that encouraging school attendance and providing resources to address attendance barriers can reduce chronic absenteeism for persistently low-income students, though the analysis cannot separate the effects of these two mechanisms. 

Persistent and Wide-Ranging Differences in the Income and Racial Segregation of Children

Abstract: I document the income and racial segregation experienced by children in neighborhoods and schools throughout the nation. Comparing segregation estimates across commuting zones, I show that racial and income segregation exhibit distinct geographical patterns. Racial segregation operates across school district boundaries, while income segregation persists within school districts. Of the demographic variation experienced by children in schools, more than 40% of variation in low-income demographics occurs within school districts, but only 18% of the variation in non-white demographics occurs within school districts. These nuances are relevant to policies that may impact how children of different racial and economic backgrounds access public education, whether directly or through spillover effects.