UC Berkeley

From 2008-09 through 2018-19, economically disadvantaged students in California scored 2.7 grade levels below their peers on academic achievement tests administered in grades 3 through 8, on average. Nationally, academic achievement gaps are best explained by differences in school readiness, as the achievement gap evident at kindergarten entry does not appear to change as students progress through later grades (Reardon and Portilla, 2016). California’s 2010 Kindergarten Readiness Act established the universal Transitional Kindergarten (TK) program to better prepare all students for elementary school. In the 2012-13 school year, California began implementing TK to serve children whose fifth birthday fell in the fall. By 2018-19, over 100,000 children enrolled in TK at a cost to the state of roughly $1 billion.

Chapter 1 of this dissertation examines the overall impact of TK participation on kindergarten English reading assessment performance among all students who entered kindergarten in the study district from 2015-16 through 2018-19. I use a regression discontinuity design to compare children who were eligible for TK to their kindergarten-cohort peers who would have been eligible but for having been born on slightly later dates. To assess whether TK participation ameliorates socioeconomic gaps in school readiness, I estimate the difference in TK impacts on kindergarten reading scores between socioeconomically disadvantaged students and their peers. No prior study of TK has examined this issue. I find that TK participation had a positive effect on midyear reading assessment scores which did not persist into the spring. I find stronger and more persistent effects among students from socioeconomically disadvantaged neighborhoods and Latino students. Despite these results, I am unable to reject the hypothesis that TK affected all students equally.

In Chapter 2, I study variation in kindergarten English reading assessment performance across multiple early learning sequences experienced by students in the same large urban public school district. TK is an academic-oriented program that uses a modified kindergarten curriculum and was explicitly designed to serve as a “bridge” from preschool to kindergarten as part of a multi-year early learning sequence. In a study sample designed to be representative of the state, 77 percent of children who attended TK were first enrolled in center-based preschool (Manship et al., 2017). The district in this study administers two state-funded pre-K programs: Transitional Kindergarten (TK) and the California State Preschool Program (CSPP). CSPP is a targeted program serving socioeconomically disadvantaged three- and four-year-olds. 19 percent of students in the analytical sample attended CSPP before enrolling in TK. I use a regression discontinuity design to interrogate TK’s “bridge” model and examine potential synergies between TK and CSPP. Research on dynamic complementarity between investments in early childhood development suggests that the impacts of early learning investments moderate and are moderated by the impacts of subsequent educational experiences (Aizer and Cunha, 2012; García and Gallegos, 2017; Johnson and Jackson, 2019). I find that TK had strong and persistent effects on CSPP participants. I find no effect of TK participation on the reading assessment performance of students who did not first attend early CSPP. TK impacts on kindergarten reading assessment performance in the spring were 0.447 (p=0.051) standard deviations higher for CSPP participants than they were for other students who attended TK. This study falls short of identifying complementarity between CSPP and TK as I am unable to isolate exogenous variation in CSPP participation. Thus I do not disentangle the influence of CSPP programming on TK effects from that of factors governing selection into the program.

The results of this study suggest that estimates of a given pre-K program’s effects on children’s outcomes depend on the sequence of early educational experiences those children would have had but for enrolling in that pre-K program. In 2018, half of all three-year-olds were enrolled in pre-K (Friedman-Krauss et al., 2019). Future studies of one-year pre-K programs and curricula will be more informative if they are able to characterize children’s counterfactual care sequences.