The right and wrong places to look for the cognitive locus of reading disability

A coherent conception of dyslexia has been difficult to arrive at because research findings have continually created logical paradoxes for the psychometric definition of reading disability. One such paradox is that cognitive research has undermined the assumption of specificity in the definition of dyslexia. The assumption of specificity is the underlying premise that the cognitive problems characteristic of the dyslexic child are reasonably specific to the reading task and do not implicate broader domains of cognitive functioning. This paper demonstrates how to develop hypotheses about the cognitive deficits of dyslexic children that do not undermine the assumption of specificity. Phonological awareness is explored as one candidate process. It is argued that in order to avoid the pitfalls surrounding the assumption of specificity three things are important: 1) We must understand the operation of Matthew effects in education (rich-get-richer and poor-get-poorer effects); 2) Strict psychometric criteria must be used to define dyslexia; 3) We must continually be aware that dyslexia does not demarcate a discrete category, separate from other groups of poor readers, but represents the outcome of the application of an arbitrary criterion in a continuous distribution.     

Components of phonological awareness

The factorial structure underlying different types of tasks within the domain of phonological awareness was examined in two studies. Large sample sizes allowed for sensitive differentiation of constructs. In the first study, 128 preschool children without any experience of formal reading instruction were tested with a battery of tasks intended to tap various aspects of phonological awareness: rhyme recognition, syllable counting, initial-phoneme matching, initial-phoneme deletion, phoneme blending, and phoneme counting. Three basic components were extracted in a principal component analysis: a phoneme factor, a syllable factor and a rhyme factor. Cross-tabulations indicated considerable dissociation between performance on phoneme, syllable, and rhyme tasks. The structural relationships were replicated on a much larger sample (n=1509) in the second study. Subjects in this study were one year older and were attending grade 1 thus providing an opportunity to test their reading achievement. Multiple regression analyses demonstrated that the phonemic factor was by far the most potent predictor. However, the rhyming factor made an independent (although small) contribution to explaining the reading variance. Among the phonemic tasks, phoneme identification proved to be the most powerful predictor.