An examination of phonological and surface subtypes of readers and their relationship to subtypes of the double deficit hypothesis

grantor: University of TorontoThe traditional dual route model of word recognition (Coltheart, 1978; Coltheart, Curtis, Atkins, & Haller, 1993; Humphreys & Evett, 1985) suggests that there are subtypes of poor readers who have deficits in either phonological processing (called phonological dyslexics) or orthographic processing (called surface dyslexics). In contrast the double deficit hypothesis suggests that there is a subtype of poor readers who have deficits in phonological processing only, naming speed only, or both (Wolf and Bowers, 1999; Wolf, Bowers & Biddle, 2000). Subtypes derived from the double deficit hypothesis are based on the assumption that the naming speed deficit in poor readers is: (a) an independent predictor of reading disability--that is, independent of phonological processing--and cannot be subsumed under a phonological processing deficit and, (b) related to orthographic processing. The present investigation examined the validity and possible correspondence between the two alternative subtype classification scheme in adult poor readers. Based on performance on phonological, orthographic, and naming speed measures, dual route subtypes received a greater degree of validation compared to double deficit subtypes. Furthermore, there was no correspondence between the two subtype conceptualizations. Specifically, the surface subtype based on the dual route model did not show a naming speed deficit, as assessed by rapid automatized naming (RAN). Furthermore, the naming speed subtype based on the double deficit hypothesis did not exhibit problems in orthographic processing, irregular word reading, or have slower reading speed for connected text. Finally, to test the assumptions from which double deficit subtypes are derived, a series of hierarchical regression analyses revealed that RAN did not explain unique variance in word recognition over and above that explained by phonological awareness, and orthographic processing in adult poor readers. This latter finding suggests that RAN is not an independent predictor of reading disability. RAN's failure to track poor reading performance in adult readers is attributed to the fact that RAN shares a considerable amount of variance with phonological awareness, suggesting that RAN is related to phonological awareness rather than orthographic processing. Together these findings undermine the double deficit conceptualization of subtypes, at least for adult poor readers.Ph.D