Dyslexia and Learning English as a Second Language An Experimental Study

Abstract

This dissertation focuses on two groups of grade 3 (9 year old) children in Pakistan: those with a reading deficit, and those who are developmentally unimpaired. The main objective of the study was to establish what the predictors of reading in Urdu might be and whether tasks that have proven useful in screening for reading deficits in other languages might also be useful in Urdu. There is little awareness of dyslexia in Pakistan and there are no tasks available to screen for children with such deficits in public schools. The sample in this study consisted of children studying in both Urdu and English medium schools. In order to design a test battery for screening purposes, some standard tests, such as the Rapid Automatized Naming (RAN) tasks (Denckla & Rudel, 1976), were borrowed in their original form. Certain tests were adapted for content and grade, such as the English 2 Dyslexia Test (Kaasa, Sanne, & Helland, 2004), while others were designed from scratch, such as a RAN task in Urdu, a non-Word Repetition task, and phonological awareness tasks, also in Urdu. This experimental study began with interviews with school teachers whose knowledge of the children’s abilities helped to divide them into 2 groups (a control group, and a reading deficit group). In the first stage of the study, the children were tested on standard RAN and NWR (Non- Word Repetition) tasks, Urdu dictation, and the reading of a short Urdu text. On the basis of their scores, the children were put into more clearly distinguished control and reading deficit groups for further testing. The statistical analyses revealed that RAN and NWR scores were reliable predictors of reading skills in Urdu. These tasks were thus shown to be capable of differentiating between typical children and children with a reading deficit. In the second stage, both groups of children were controlled for non-verbal IQ differences. The results for children in Urdu medium schools were analysed independently of those for children in English medium schools. This was done in order to avoid any confounds based on differences of socio-economic status (SES) or the fact that children in English medium schools are exposed to a great deal of educational extra-curricular activities, unlike children in Urdu medium schools. Furthermore, all children were tested using the English 2 dyslexia test (Kassa et al., 2004). This test was also shown to be capable of differentiating between the L2 performance of typical children as compared to children with a reading deficit on all tasks in the test. The children from English medium schools from both groups performed better than their counterparts in Urdu medium schools. These results demonstrate that increased exposure to a language helps even the reading deficit group to learn that language better. Additionally, the study showed that the typology of the L1 (morphology, orthography, syntax etc.) might be critical in L2 acquisition. In the second stage, the children were also tested on phonological awareness (PA) tasks, but none of the tasks was able to differentiate between the groups. This raises questions about the suitability of traditional PA tasks for the screening of reading deficits in Urdu. This thesis is based on a collection of research papers. The first paper is about the validation of NWR and RAN tasks a s predictors of reading for screening purposes in Urdu. Paper 2 and paper 3 are about the validation of the English 2 dyslexia test (with content and grade adaptation) in a Pakistani context. The original English 2 dyslexia test is intended to test children at grade 6 in Norway, but for this study it was adapted for grade 3. Paper 4 discusses the validity of phonological awareness tasks as r e a d in g deficit screening tools in Urdu.