Examining Moderators of Vocabulary Acquisition from Kindergarten Through Elementary School Using Local Structural Equation Modeling

Abstract

Does parental involvement or formal education reduce the often found competence gap between children from different socioeconomic backgrounds? To answer this question, we examined the joint and unique effects of parental education, socioeconomic status, and home learning environment on children’s vocabulary competence and growth in early childhood. We used latent growth curve models to distinguish between pre-school vocabulary competence and growth across primary school. Analyses were based on data from the German National Educational Panel Study (NEPS), a large-scale longitudinal study assessing vocabulary competence and family background from Kindergarten to the 3rd grade of elementary school. To scrutinize the moderating effects of socioeconomic status, parental education, and home learning environment on model parameters, we relied on local structural equation modeling. Results revealed a moderate positive effect of parental education on children’s receptive vocabulary competence, which fully explained the effect of socioeconomic status on this language skill. Home learning environment had no effect on children’s competence level. Initially lower performing children showed steeper growth trajectories across school, but the effect was small compared to the initial differences due to parental education. In summary, the results indicate that formal schooling helps to reduce vocabulary competence differences across children from different educational backgrounds, but are not able to fully overcome them.

The development of the rank-order stability of the Big Five across the life span.

“…Next, to test for metric measurement invariance, we estimated LSEMs with added constraints on the factor loadings and compared the fit statistics: First, we examined whether the same construct was measured across the four waves by constraining the factor loadings of the same item to be equal across the four measurement points. Then, we additionally established metric measurement invariance across age by restricting the factor loading of an item to be equal across all ages (for the introduction of invariance constraints across moderator values in LSEM, see Olaru et al, 2020). The model with metric invariance across both time points and ages was used to estimate the rank-order stabilities.…”