Arabic (L1) and English (L2) language and literacy skills in Grade 3 children were examined to determine the predictors of literacy skills in L1 and L2 across Grades 3 and 4. Eighty-two Arabic-speaking children from Kuwait participated in the two-year longitudinal study. Children were followed from Grade 3 (when formal literacy instruction of L2 begins in the state of Kuwait) to the end of Grade 4. A battery of tests was used to measure language and literacy skills in both languages at six-month intervals. This included measures of decoding skills (non-word reading), vocabulary, phonological skills (phonological awareness, rapid automatized naming and phonological memory), orthographic skills (orthographic segmentation, orthographic discrimination and visual memory), morpho-syntactic skills (syntactic awareness and morphological segmentation), and literacy skills (isolated word reading, reading fluency, reading comprehension, comprehension fluency, word spelling, text spelling and writing composition).

Results argued for basic skills to support more complex literacy skills in both reading and writing. For example, the data showed that decoding explained variance in literacy skills in both Arabic and English, and spelling levels were predictive of writing performance in both languages too. Vocabulary, on the other hand, showed less contribution than decoding to literacy in both languages, even when measures of reading comprehension were considered. Phonological awareness influenced literacy mainly via word recognition factors (particularly decoding), whereas rapid naming demonstrated a direct relationship with a number of literacy measures, particularly when fluency was required. While phonological memory was less predictive of literacy than the other measures of phonological skills, it was a significant predictor of non-vowelised Arabic reading comprehension. Although early phonological processing skills were not the predominant predictors of early Arabic literacy, they predicted Arabic literacy in Grade 4, when children begin reading non-vowelised school texts. The data also argued for orthographic skills (which included orthographic discrimination, orthographic segmentation and visual memory) to be better predictors of literacy than phonological skills, particularly for fully vowelised Arabic texts, a finding that may be consistent with views about the complexity of the fully vowelised Arabic script. In addition, measures of morpho-syntactic skills predicted literacy levels in both languages; and these skills were related to vocabulary suggesting an overlap between the development of vocabulary and the processing of morpho-syntactic features within this cohort of students.

Overall, the findings suggest that a Simple View of Reading and Writing could be applied to Arabic in a similar way to English. However, additional factors related to orthographic segmentation and rapid naming are included as important factors in the Arabic model. Furthermore, the relationships between vocabulary and morpho-syntactic skills are taken into account. Given the need for modifications, a model is proposed that includes these additional factors and which is discussed in light of previous research and theories in the field.