Home literacy environments in early childhood: Exploring survey measurement invariance across SES groups

Home literacy environments in early childhood: Exploring survey measurement invariance across SES groups

First Author: Beth Phillips -- Florida State University
Additional authors/chairs: 
Christopher Lonigan; Sisi Dong; Pamela Burris
Keywords: Home Literacy Environment, Emergent literacy, early childhood (age 4 - 6), Parent-child interactions, Measurement
Abstract / Summary: 

Purpose. Considerable research has identified predictive associations between the home literacy environment (HLE) and children’s early language and literacy development. Furthermore, numerous studies have compared, and found differences in, reported HLE behaviors across families of different socio-economic and cultural backgrounds. However, little research has explicitly investigated the measurement of HLE constructs in family subgroups; therefore, this study evaluated measurement and structural invariance in a large sample of families.

Method. Composed across four source studies, surveys were provided by a sample of 2,766 primary caregivers of children ages 3-5-years-old (M = 54.57 months). The sample is diverse across child gender, race, ethnicity and household income. Respondents completed self-report items on the HLE representing four conceptual components of shared reading frequency, interactive reading quality, language interactions, and code-focused teaching. The invariance of this four-factor model was compared across levels of maternal education (N= 1400 some college and higher, N= 1366 high school diploma or less).

Results. Models were analyzed in MPLUS using the WLSMV estimator to account for the ordinal measurement scales. Sequential tests of invariance across models indicated that whereas configural and weak partial measurement invariance were achieved, strong measurement invariance was not. Evaluation of structural invariance also indicated that some factor intercorrelations differed between the subgroups.

Conclusion. Findings suggest some caution may be required when interpreting differences in the self-reported HLE across families representing distinct socioeconomic strata, as indexed by maternal education. Extensions will explore questions of invariance across other family characteristics, such as income, ethnicity and other cultural distinctions.