Can Kindergarten Teacher Value-Added Sustain the NC Pre-K Boost During Kindergarten?

• In Event: To Pre-K and Beyond: Child Development and Educational Environments Across Pre-K and Elementary School

Abstract

This study drew comparisons between the skill development of children who did or did not participate in North Carolina’s Pre-K program. Emergent literacy skills were assessed at the beginning, middle, and end of kindergarten. Importantly, we examined if a “value-added” measure of kindergarten teacher effectiveness moderated skill differences between program participants and non-participants. In doing so, we considered the interaction between educational environments across pre-K and kindergarten: NC Pre-K participation during pre-K and teacher value-added during kindergarten.

Consistent with previous research (Peisner-Feinberg et al., 2019), we first hypothesized that NC Pre-K participants would outperform non-participants at the beginning of kindergarten. Second, we hypothesized that this difference would diminish by the end of kindergarten—on average—as a former NC Pre-K participants demonstrated lower rates of skill development compared to their non-participating peers during kindergarten (Peisner-Feinberg et al., 2020). The third hypothesis was consistent with the dynamic complementarity hypothesis (Bailey et al., 2020). Specifically, we hypothesized that former NC Pre-K participants would demonstrate comparable rates of skill development as non-participants when enrolled in kindergarten classrooms with higher value-added teachers, which would lead to sustaining the NC Pre-K boost through the end of kindergarten.

Statewide administrative records identified NC Pre-K participants during the 2013-14 program year. Kindergarten teachers administered a series of subtests from the Dynamic Indicators of Basic Early Literacy Skills (DIBELS) Next® to assess children’s decoding, phonological awareness, and phonemic awareness skills at the beginning, middle, and end of the 2014-15 kindergarten year. The composite score of the DIBELS Next subtest scores was calculated and then transformed into the Lexile® scale score for longitudinal analysis (with r = 1.00 between scores). Socio-demographic covariates were drawn from children’s birth records (e.g., maternal education) and kindergarten school records (e.g., gender, race, and economic disadvantage).

Multi-level regression models measured teacher value-added as the classroom-wide average residualized gain in emergent literacy skills (Chetty et al., 2014). Propensity score matching with 1:1 nearest neighbor matching resulted in a matched sample of NC Pre-K participants (n = 8,865) and non-participants (n = 8,865), with standardized mean differences < 0.10 for all covariates and exact matches for kindergarten teacher enrollment and other school record covariates. Multi-level growth curve models examined children’s skill development during kindergarten as well as program differences and moderation of program differences by teacher value-added (Table 1).

Consistent with our first hypothesis, former NC Pre-K participants outperformed non-participants in emergent literacy skills at the beginning of kindergarten (β=0.22). Consistent with our second hypothesis, former NC Pre-K participants demonstrated lower rates of skill development compared to non-participants during kindergarten as the magnitude of the NC Pre-K effect was diminished by the middle of kindergarten (β=0.05) and no longer statistically significant at the end of kindergarten (β=0.02). Kindergarten teacher value-added was positively associated with skill development during kindergarten, but was not found to moderate program differences in skill development—providing no support for dynamic complementarity (Figure 1). Future research should examine the alignment of curriculum and instruction across pre-K and kindergarten to sustain the pre-K boost.

Authors

• Robert C. Carr, University of North Carolina at Chapel Hill
  Presenting Author

• Jade V Marcus Jenkins, University of California - Irvine

• Tyler Watts, Teachers College, Columbia University

• Ellen S Peisner-Feinberg, University of North Carolina at Chapel Hill

• Kenneth A Dodge, Duke University