Browse By Day
Browse By Time
Browse By Panel
Browse By Session Type
Browse By Topic Area
In Event: Special Poster Session 05 with Continental Breakfast Reception
In Poster Session: PS 05 - Policy Section
English Language Learners (ELLs) comprise 10% of the total K-12 population, with most speaking Spanish (MPI, 2015). National assessments demonstrate significant reading achievement differences not only between White and Hispanic/Latino students but also within the Hispanic population, with Hispanic ELLs significantly underperforming Hispanic non-ELLs (NCES, 2011). Programs intended to serve students with limited English proficiency are typically divided into two categories: bilingual education and English-only instruction. With researchers debating which type of program best serves the academic needs of ELL students, and state policies having the ultimate say on how best to accommodate this student population, this study aims to address whether a state’s educational policy for language of instruction predicts the early reading growth of Spanish-speaking ELL students in the U.S.
To address this aim, a nationally representative sample of Hispanic students was drawn from the Early Childhood Longitudinal Study: Kindergarten Cohort of 2010-2011. Each state’s language of instruction educational policy (LIEP) was collected from legislative laws, state educational handbooks, and state Department of Education websites: (a) English-only (b) Bilingual or (c) an Unspecified policy where both methods were permitted. A child’s language status was coded as (a) English-Only (b) Bilingual (i.e. proficient in English by spring of kindergarten but Spanish home language) and (c) ELL (i.e. not proficient in English, Spanish home language). Reading achievement was assessed at kindergarten entry and again in the spring of kindergarten, first, second grade, and third grades. All analyses are weighted to ensure the final sample is representative of Hispanic children in the U.S. and included a comprehensive set of control variables (see Table 1).
Hierarchical linear models were conducted in Stata 14 with population weights provided by NCES (W1C0) with schools specified as the clustering level for students (see Table 2). Model intercepts were set at the fall of kindergarten, and loadings on slope terms were set at one-year intervals. Unconditional growth curves were specified with linear and quadratic slopes. Parent and child, school, and state level covariates were then included as predictors of intercept, linear slope, and quadratic slope terms. Interactions between a child’s language status and a state’s language policy were added to the final models.
Results from our unconditional model (see Model 1 in Table 2) suggest that there are significant differences in Hispanic students’ initial reading levels, linear growth rates, and in their deceleration rate. Model 2 of Table 2 suggests that at the start of kindergarten, bilingual and ELL students scored lower in reading than their English-speaking peers. Model 3 of Table 2 demonstrates that bilingual students’ growth was significantly greater than English-speaking students’ trajectories. A significant interaction on the linear slope indicates that over time, larger reading gaps emerged between bilingual and English-speaking students in states with unspecified versus English-only policies. Further analyses will involve multiple data imputations to increase sample size and refine the modeling of associations between state’s language policy on Hispanic children’s early reading achievement during elementary school.