Content area
Grade 1, 2 and 3 children with reading difficulties and normally achieving children in these grades were compared on measures of syntactic processing (syntactic judgement and correction), phonological processing (phoneme segmentation) and verbal working memory. A series of hierarchical regression analyses and commonality analyses indicated that phonological processing was a consistent and unique predictor of reading performance across these stages. On the other hand, syntactic processing was only a unique correlate of pseudoword reading ability in Grade 1 children. These findings contradict the model of Tunmer and Hoover (1992) that views syntactic processing and phonological processing as equally predictive of early reading ability. The finding that syntactic processing and phonological processing share common overlapping variance also supports the phonological processing limitation hypothesis (Shankweiler, Crain, Brady & Macaruso, 1992). Proponents of the phonological processing limitation hypothesis believe that all oral language weaknesses found in children with specific reading difficulties are the result of lower level phonological processing weaknesses. This hypothesis contrasts with the structural lag hypothesis which views the cause of these language as being a delay in the development of more complex linguistic structures.
Subsequently, children with specific language impairments (N = 24) were compared to children with oral language skills in the normal range (N = 48) on the experimental measures. An additional phonological processing measure, pseudoword repetition, was included. Hierarchical regression analyses showed that the phoneme segmentation measure continued to be the unique predictor of reading ability in these groups. However, this phonological processing measure did not correlate with language classification group (SLI versus normal oral language development). On the other hand, syntactic processing and phonological processing as measured by the pseudoword repetition task were both unique statistical predictors of language classification group, while phoneme segmentation ability was not related to language classification group. Therefore, reading ability and language classification group are correlated with different experimental variables in the syntactic, phonological and verbal memory realms in primary school-aged children.