Examining an Extended Home Literacy Model: The Mediating Roles of Emergent Literacy Skills and Reading Fluency

ABSTRACT

We examined the developmental relationships between home literacy environment (parent teaching, shared book reading) and emergent literacy skills (phonological awareness, letter knowledge, vocabulary, rapid naming speed) in kindergarten, reading accuracy and fluency in Grade 1, and reading comprehension in Grades 2 and 3 in a sample of Canadian children learning to read English (N = 214). Results from a latent variable model showed that parent teaching predicted letter knowledge and phonological awareness, and shared book reading predicted vocabulary and rapid naming speed after controlling for family socioeconomic status. Moreover, both parent teaching and shared book reading contributed indirectly to reading accuracy and fluency in Grade 1, which then mediated the effects of home literacy environment on reading comprehension in Grades 2 and 3. The results suggest that the effects of home literacy environment on later reading development are distributed via more pathways than previously thought.

Notes

¹ For a similar approach in the context of general academic achievement, see Bornstein, Hahn, and Wolke (2013).

² Reading fluency was assessed in the spring of Grade 1 to allow fluency to become established.

³ This procedure corresponds, in principle, to correcting for the possible bias due to measurement error.

⁴ To examine a potential bias in the model estimates caused by the ordinal-scaled variables of the home literacy activities, we also tested the home literacy environment model using polychoric correlations with the robust weighted least squares estimator implemented in Mplus (Flora & Curran, 2004; Muthén & Muthén, 1998–2017). The result showed that the factor loadings and correlations were very similar to those obtained using the maximum likelihood estimator (∆λs ≤ .07; ∆rs ≤ .04). Therefore, we treated the data from the questionnaire as continuous variables in further analyses (see also Rhemtulla, Brosseau-Liard, & Savalei, 2012).

⁵ Given that parents’ education, but not parents’ occupation, significantly correlated with shared book reading in our sample (see Table 2), we repeated the analyses after removing the two questions about parents’ occupation from the models. The results replicated those that included both parents’ education and occupation, suggesting the robustness of the results (see also Liu, Georgiou, & Manolitsis, 2018; Torppa et al., 2007, for similar findings).

Additional information

Funding

This research was funded by a grant from the Social Sciences and Humanities Research Council of Canada to John R. Kirby and Rauno Parrila.