THE INVESTIGATION OF POPULATION HETEROGENEITY OF PARENTAL INVOLVEMENT IN ENGLISH LANGUAGE LEARNERS WITH FACTOR MIXTURE

Abstract

Parental involvement is an influential factor for child's positive learning outcomes, as well as their social, cognitive, and emotional development. However, few studies have evaluated the effects of parental involvement with English Language Learners (ELL). As ELL students are an underrepresented population that makes up a burgeoning population in the U.S, this student subpopulation appears to benefit significantly from parental involvement, at least in terms of the empirical studies that exist. Nevertheless, empirical studies have found that parents of ELL students are less likely to become involved in their children's academic activities than other parents due to many factors. Thus, the purpose of this study is to classify and examine parental involvement with ELL students through factor mixture modeling. Factor mixture model can accurately differentiate subgroups and allow for the examination of individual differences within each identified class. The identification of different subpopulation membership of parental involvement are of primary interest for this study because they assist researchers in understanding parent subgroups with respect to their level of involvement. The factor mixture model that combines dimensional and categorical structures in one analysis can provide a more flexible framework and advantages over the conventional models. Therefore, this study evaluated whether latent class, factor analysis, and factor mixture model provides good model data fit using 2002 Education Longitudinal data for parental involvement in English Language Learners. The results support a two-class two-factor Factor Mixture Model-3 model that provides the most parsimonious solution for the data. This suggests the factor has different meaning in each class due to the violation of a strong factorial invariance.