Robin van Rijthoven

123 Compensatory role of verbal learning and consolidation in reading and spelling 5 but not for word and pseudoword decoding. Intercept, Slope, and the GroupxIntercept interaction were all significant predictors of Spelling. A steeper slope and a higher intercept were related to higher spelling scores. The effect of the Intercept is larger for one of both groups. To be more precise, as can be seen in Table 5.2, especially for typically developing children, a higher intercept is related to higher spelling scores. This shows that typically developing children seem to benefit more from their immediate memory span capacity in learning to spell compared to children with dyslexia. In order to check if the found effects remain present when controlling for the severity of the reading and spelling problems (phonological awareness, rapid naming, and verbal working memory) and semantic abilities, these variables were added separately in Step 2 and 3 of each hierarchical regression analyses. The effect of Slope on Pseudoword decoding and the effect of Intercept on Spelling remained significant in Step 2 for each predictive variable. This indicated that despite the severity of the reading and spelling problems or the semantic abilities of children the higher verbal learning rates still predict lower scores in pseudoword decoding and that a higher immediate memory span still predicts higher spelling scores. Step 3 was no longer a significant improvement when adding these variables with the exception of Phonological awareness. When adding Phonological awareness, Step 3 became significant for Word decoding and Pseudoword decoding as well, showing significant effects of Intercept and the GroupxIntercept interaction on Word decoding and no significant effects of Intercept or Slope for Pseudoword decoding. The immediate memory span appeared to be associated with word reading and spelling performances for all children but even more so for typically developing children. The results of the second set of three hierarchical regression analyses, regarding the effect of Trial 5 and Verbal consolidation (see Table 5.4) showed that group was a significant predictor for all three outcome measures (Step 1). Children with dyslexia read and spell fewer words correctly (Step 1). Step 2 and Step 3 were both not leading to a significant improvement of the model. As a result, Trial 5 and Verbal consolidation were both not significantly associated to reading and spelling after controlling for group.