Speech Processing and Intelligent Assessment in Second Language: Struggling Immigrant Students and Home Language Parallel Instruction.

Abstract

Purpose: In this study, home language (L1), parallel instruction in L1, and grades were the factors examined to identify differentiated writing performance of young second language learners. Method: A total of 102 children, aged between 8-17 years old, (M=13; DP=2,7), learners of Portuguese as a Second Language (L2), belonging to 6 language groups were compared on a narrative essay composition to identify types of writing (in L2 context) and the effects involved. Home language, parallel instruction in L1 and grades were controlled to examine those effects. The written outputs were measured using the score in six components adapted from an English L2 writing assessment (Alberta Education). Results: Analyses of variance showed only significantly differences for home language variable (F(5,58) = 1,769; η2= .132). Covariance analyses revealed impact of parallel instruction for L2 performance improvement F(1,62) = 4,573; p <.05, η2= .069). The groups whose home language had common features as the language of testing revealed lower scores showing no advantage of language overlap as commonly argued. Conclusions: Specific groups of students have more difficulties during writing in L2 and the home language instruction should be prominently integrated as parallel teaching methodology into early literacy instruction of immigrant groups at school.