Cultural and linguistic diversity in reading literacy achievement: a multilevel approach

Current models of reading comprehension consider both reading abilities and reading attitudes and motivation as key factors related to reading (Kamil et al., 2000; Van Diepen et al., 2007). The aim of the current study was to explore the relationship between various school, class and student background characteristics, to the students’ reading literacy achievement in the Netherlands. The influence background variables have on a student’s reading ability may differ between the subgroups of first and second language learners (Van Elsacker, 2002). To explore this potential diversity this study focuses on the variation between the two subgroups of L1 and L2 students. The national data used for this study were selected from the IEA Progress in International Reading Literacy Study 2006 (Mullis, Martin, Kennedy & Foy, 2007). Turkish and Moroccan students were selected for the second language group (N=208). Those students who spoke another language than Dutch were deleted, leaving 3549 students in the first language group. Multiple Imputation was chosen to handle the missing data in the dataset, this procedure was executed using SAS software. The multilevel analyses were conducted with MLWin software, using HOUSE Weight and plausible values (repeated analyses). The variables contributing most to the differences in reading literacy were first selected from a theoretical viewpoint; secondly correlations and reliability of the scales were addressed. Zero model: school, class and student level variance. The antecedent and malleable variables are added in subsequent order. Model 4 is the full model and consists of all antecedent and malleable variables. Model 5: using full model, examining an interaction effect for Language. First preliminary results show that there are strong links between various scales and the reading literacy results; the full model explains 34,7% of the variance in reading literacy. On the ECER, the results of the multilevel analyses and the found differences between the two subgroups of first and second language students will be addressed and an attempt will be made to draw conclusions.