Predicting language and literacy outcomes of typically-developing and at-risk English-speaking elementary students in French Immersion programs

Many children around the world are educated in a second language, be they speakers of the majority language of the community in which they are educated or speakers of a minority language. Research has demonstrated the effectiveness of second language immersion for typically-developing students and students who are disadvantaged in the academic environment of schooling due to low levels of academic ability, disadvantaged socio-economic background, or minority ethnic group status (for a review see Genesee, 2006). In contrast, although it is generally thought that the level and kind of first language (L1) ability that children acquire prior to coming to school, and especially in domains related to literacy, are important predictors of success in school, there is little research on individual differences in second language (L2) reading and oral language development of immersion students and, in particular, on students who might be at risk for reading or oral language difficulties (Genesee, 2006).;Study 2 involved an investigation of both oral language and literacy outcomes of children who were either typically developing or at risk for difficulties in the area of oral language or literacy. This study provides threepronged evidence for the distinctiveness of oral language impairment and reading impairment, namely that (1) performance on oral language predictor measures was largely independent of performance on literacy predictor measures, suggesting that each domain constitutes a distinct underlying construct; (2) the most significant predictors of oral language and of literacy difficulties were different -- that is, phonological awareness (PA), phonological recoding, and letter-sound knowledge in L1 in K were significant predictors of risk for reading difficulties in L2 while performance on L1 sentence repetition, PA, and tense marking tests in K were the best predictors of risk for L1 and L2 oral language difficulties; and (3) three distinct subgroups were found in our sample and included children with oral language difficulties, children with literacy difficulties, and children presenting with difficulties in both domains. Even though we began administering predictor measures to the children at a point when they had not yet received literacy instruction, English L1 measures taken at the earliest time point (Fall K) and at the end of that academic year (Spring K) both predicted Grade 1 outcomes to a significant extent, although the latter predictors were more accurate.;Overall, our results suggest that risk for oral language and reading development are distinct; and it is possible to identify, early on, children who are likely to struggle in the domains of oral language or literacy, even if they are schooled in an L2. Our findings further suggest that it is possible to predict the specific domain that is likely to be difficult for them.;We studied 86 English-dominant children in early French Immersion (FI) programs from Fall of Kindergarten (K) to Spring of Grade 1 and tested them at three time points: Fall K, Spring K, Spring Grade 1. Our sample included both typically-developing and at-risk children and our test battery included a broad range of oral language and literacy related predictor and control measures. Consistent with extensive L1 research (e.g., NICHD, 2000), the best K predictors of L2 decoding in Study 1 were knowledge of the alphabetic principle in English and phonological awareness in English. Contrary to what others have found (Jared, Cormier, Levy, & Wade-Woolley, 2006), knowledge of French at K entry was a third significant predictor. The same variables were significant predictors of French reading comprehension, with the added contribution of scores related to oral language skills in L1, suggesting that oral language abilities play an important role in reading comprehension. Furthermore, we were able to test a popular theory of reading, the Simple View of Reading (Gough & Tunmer, 1986), and provided support for its intra- but not cross-linguistic applicability.