Learning in Multilingual Contexts: Language Policies, Cross-Linguistic Transfer, and Reading Interventions

Abstract

The proportion of children across the world who have access to basic education through formal schooling has never been as high as it is today. Surveys of literacy and numeracy skills in low- and middle-income countries, however, regularly highlight low levels of learning, even for young people with several years of formal schooling. Among the structural issues faced by school systems in low- and middle-income countries experiencing rapid increases in enrollment, linguistic heterogeneity of communities and countries stands out as it poses both a threat and opportunity to quality instruction. This dissertation is organized in three papers and aims to explore: (i) ways in which language in education policies are designed, implemented, and understood; (ii) whether cross-linguistic transfer theories apply in a new context; and (iii) heterogeneity across teachers and schools of the impact in a bilingual literacy intervention. The first paper describes the multi-faceted factors that policymakers take into account when designing language in education policies, and presents a framework for analyzing these policies at the community, regional or state-level. The second paper uses structural equation modeling to test cross-linguistic transfer in a sample of second-grade students in Uganda. Findings suggest children in second grade have developed common decoding skills across two languages, and those staring to learn in their first language can transfer reading comprehension skills to the second. The third paper in this dissertation explores the impact of a literacy intervention in Kiswahili and English in urban public and private nonformal schools in Kenya. The study exploits the random assignment of schools to an intervention to identify whether the impact of an intervention differed across types of teachers and schools. The study finds large average gains in literacy skills in English and Kiswahili and no differences in average impact across teacher characteristics. In contrast, the school characteristics moderate the impact of the intervention on higher-level skills (word reading and comprehension). This work contributes to support the process of designing and implementing language policies in schools focused on inclusion and equity.