Structural model of reading comprehension ability in English for children with English as second or subsequent language

Purpose: Studies in India have examined the relation of rapid automatized naming (RAN), reading fluency (RF), phonological awareness (PA), spelling ability, and language ability with reading comprehension (RC) in English in children where English is taught for the purpose of education. However, the predictor variables of RC have not been explored. Aim: To propose a structural model for RC in English in children with mother tongue Marathi studying in English medium schools. Materials and Methods: Participants included 80 children with mother tongue Marathi (Grades 1 and 2) in English medium schools. A battery of test for the assessment of language, RAN, PA, RF, spelling and RC in English was administered. Smart PLS 3.0 was used for the analysis of data. Results: The final model proposed had loadings > 0.708, composite reliabilities > 0.7, and AVE of all constructs > 0.5. Discriminant validity of the model was ascertained as the values were less than the heterotraitmonotrait (HTMT) criteria of 0.85 for all components. 86% of variance in RC was explained by language ability (β = 0.409, P < 0.001) and RF (β = 0.453, P < 0.001), when RF was mediated by RAN-letter (β = 0.348, P < 0.001), and PA (β = 0.384, P < 0.001). Conclusion: The structural model proposed in the study is reliable and valid. Language ability and RF have significant influence on RC, and RAN-letter and PA have a significant mediating role in the relationship between RF and RC. Thus, RC in English can be improved if language and RF are developed, and RF can be improved by enhancing RAN-letter speed, and PA skills in these children.