Poster #179 - Classroom Composition and Preschoolers’ Receptive German Vocabulary: Unpacking a “Simpson’s Paradox”

• In Event: 3-159 - Poster Session 12
  In Poster Session: PS 12 Section - School Readiness/Childcare

Integrative Statement

Supporting children’s early societal language skills is an objective of early childhood education and care (ECEC). One consideration is whether the composition of an ECEC classroom in terms of the share of dual language learners (DLLs) affects the language skills of individual children in a classroom. Despite interest by parents and policymakers, there is very little research on this link. A US study uncovered a negative connection between the percentage of DLLs in a classroom and DLLs’ English vocabulary (Garcia 2018). Moreover, there is evidence that having peers with more advanced language skills benefits preschoolers’ language skills (Justice, Petscher, Schatschneider, & Mashburn, 2011; Mashburn, Justice, Downer, & Pianta, 2009). However, such connections may differ for single language learners’ (SLLs’) and DLLs’ outcomes as they are exposed to differing amounts of the societal language at home.
In a large sample of German preschoolers (N = 2231), we examined whether there were links between the classroom percentage of DLLs and the German receptive vocabulary of SLLs and DLLs. We tested for differential effects of classroom composition depending on language background of the children (SLL vs. DLL). Of the children, n = 371 were DLLshiExp from families in which German was frequently spoken, and n = 305 DLLsloExp from families in which German was less frequently spoken. Children (age 30 – 80 months) attended 177 classrooms in 95 preschools.
In multilevel regression models, we controlled for children’s gender and age at entry, classroom quality and other classroom variables. German vocabulary was assessed with the German edition of the Peabody Picture Vocabulary Test, 4th edition (Dunn & Dunn, 2015). Teachers reported on children’s age at entry into preschool and the language background of every child in the classroom. From these reports, the classroom percentage of DLLs was computed. Three domains of classroom quality were observed with the Classroom Assessment Scoring System Pre-K (CLASS Pre-K; Pianta, La Paro, & Hamre, 2008).
In models not controlling for individual language background, a higher percentage of DLLs in a classroom predicted lower individual vocabulary scores (illustrated by the bivariate correlations in Figure 1A). However, this connection disappeared when the individual language background (SLL vs. DLLshiExp vs. DLLsloExp) was accounted for (illustrated in Figure 1B). The initially found link between percentage of DLLs and German vocabulary was a case of the so-called Simpson’s Paradox, where associations evident in a total group disappear when examining subgroups separately (or the other way around). It was “caused” by SLLs, DLLshiExp and DLLsloExp attending classrooms with very different average percentages of DLLs.
Our findings suggest that despite much public interest in classroom composition in terms of the percentage of DLLs, it is not highly relevant for SLLs’ or DLLs’ German vocabulary skills. Moreover, our study highlights the importance of accounting for individual language background when analyzing composition effects as failing to do so can result in false conclusions.

Authors

• Jessica Willard, Ruhr-Universität Bochum
  Presenting Author

• Katharina Kohl, Ruhr-Universität Bochum
  Non-Presenting Author

• Alexandru Agache, Ruhr University Bochum
  Non-Presenting Author

• Lilly Bihler, Ruhr-Universität Bochum
  Non-Presenting Author

• Birgit Leyendecker, Ruhr-Universität Bochum
  Non-Presenting Author