A Meta-Analysis of Early Language and Vocabulary Interventions

• In Event: Wholistic Aspects of Early Childhood: Language, Music, Play and Health

Abstract

Language development is one of the many building blocks children need to function successfully as they grow. Specifically, early vocabulary size is an important predictor of later language skills (Scheffner Hammer et al., 2017; Grossheinrich et al., 2019). Further, toddlers with smaller vocabularies at around 2 years of age continue to score below peers in multiple domains throughout adolescence, indicating a need to understand the efficacy of early language interventions already in use (Rescorla, 2002; 2005; 2009). The most recent large-scale effort to synthesize the impacts of early language interventions on vocabulary was published over a decade ago (Marulis & Neuman, 2010), though more focused meta-analyses have been conducted since (e.g., Flack et al., 2018; Fitton et al., 2018). Here we present a meta-analysis of early language interventions, broadly similar to Marulis and Neuman (2010), but focusing solely on preschool and kindergarten students, and only evaluating standardized vocabulary measures as opposed to author-created measures.
Searching studies published after 2008 when Marulis and Neuman ended their search, we reviewed 65 studies and 105 effect sizes from interventions spanning 13,609 children that focused on vocabulary as one of the main targets. Results indicated a small but significant overall effect size of .22, demonstrating that interventions geared toward young children result in positive gains on standardized language measures. Moderator analyses were conducted to understand the impact of multiple factors, including but not limited to, dosage, intervention administrator, at-risk status, size, and intervention type. Smaller effects were found for published studies than dissertations, and for studies that recruited larger rather than smaller numbers of children. However, other intervention characteristics were not found to impact efficacy, suggesting that interventions may be beneficial for the majority of those enrolled, regardless of general intervention type. To note, most studies enrolled low socio-economic status participants, but only studies with typically developing populations were analyzed. A smaller effect size was found as compared to the previous meta-analysis conducted by Marulis and Neuman, though this may be due to our more stringent inclusion criteria. Including an additional 23 studies with 53 effect sizes reported in Marulis and Neuman (2010) from prior to 2008 that met the current inclusion criteria resulted in a total of 86 studies with 158 effects. We found an overall effect of .24, similar to that found when only analyzing more recently conducted interventions (see Figure 1). In all, the current analysis suggests that early language interventions result in modest, though positive gains in vocabulary.
A qualitative investigation of how interventions have changed in the past decade suggests that more recent studies tend to be larger, longer, and have more whole-class interventions rather than individually administered interventions. In addition, the more recent interventions utilized many different treatment plans, some with multiple components including professional development for classroom teachers, direct instruction for learners, and more general literacy exposure. We suggest that to truly understand which components of these interventions are most beneficial for young learners, a more thorough analysis of intervention characteristics will be critical.

Authors

• Jennifer Weber, University of Colorado Boulder
  Presenting Author

• Eliana Colunga, University of Colorado Boulder