Patterns of cross-linguistic transfer of language and reading skills in multilingual contexts - findings from a multi-country study

• In Event: Language of instruction transitions: factors that predict multilingual literacy acquisition and cross-linguistic skills transfer from four countries

Proposal

Language skills are widely recognized factors in explaining literacy outcomes (Ouellete, 2006; Ouellette & Beers, 2010; Alidou et al., 2006). According to Hoover and Tunmer’s (2020) Cognitive Foundations of Reading Acquisition (CFRA), language comprehension is a complex, multi-dimensional skill that, when paired with decoding abilities, explains reading comprehension.


In multilingual environments, these relationships get more complex. Several studies show that L1 reading ability is one of the most significant predictors of L2 reading, often referred to as “transfer” of L1 reading skills to L2 skills (August & Shanahan, 2006; Koda & Reddy, 2008). The Transfer Facilitation Model (TFM) (Koda, 2008) specifies which reading sub-skills are transferred between languages, and how. The model suggests that metalinguistic ability - especially phonological awareness and morphological awareness - are the central resources that are “shared” between languages, which explains the significant relationships between L1 and L2 reading sub-skills.


There is emerging evidence of cross-linguistic resource sharing of metalinguistic skills and transfer of reading skills in younger children (Nakamura, Koda, & Joshi, 2013; Zhang & Koda, 2013). But there are significant gaps in the evidence base in terms of: the directionality of transfer or resource sharing, the specific causal pathways by subkill, evidence from LMICs and across languages, and the “sufficiency” of reading skills needed for effective transfer (Nakamura, et al., 2018; 2019). To design effective multilingual models, it is crucial to understand which skills are shared, which are transferred, and the directionality of transfer in varying contexts. This study seeks to enhance understanding of these transfer relationships by reporting findings from a set of 7 language pairs across 4 countries (Kenya, Rwanda, Mozambique, and Senegal). It seeks to answer the following research questions: How do different language and reading sub-skills predict reading comprehension (within and across languages)? What are the pathways through which these variables predict reading comprehension?


We use a quantitative, multi-case research design and cross-sectional sample of 1,200 students per country in 60 schools in a random, stratified sample. In each country (with the exception of Rwanda, which has a predominantly monolingual local language population), we selected two regions with distinct local languages, sampling 30 schools in each region, to test the transfer relationships. Each student was tested in two languages, a local LOI and a second language to which students are expected to transition. The following constructs and variables (in parentheses) were tested:


L2 reading comprehension
L2 language comprehension
L2 decoding
L1 reading comprehension
L1 language comprehension
L1 decoding


We employ structural equation modeling (SEM) to analyze the relationships between language and literacy sub-skills, testing a model from Proctor, August, Snow, and Barr (2010) that seeks to explain the predictors and pathways for L1 and L2 reading comprehension and the cross-linguistic relationships. Doing so will advance understanding of transfer patterns in contributing to multilingual reading comprehension across a diverse set of contexts.

Author

• Evans Wafula Mahaya, University of Nairobi