Syllabic processing in visual word identification by dyslexic children, late readers and normal readers at age 11

Three groups of eleven-year-old children at different reading levels (normal readers, below average readers and dyslexics) were presented with a visual adaptation of the monitoring task by Mehler et al. (1981). The children were required to detect a two or three letter target string at the initial position of a familiar or unfamiliar word. For example, target strings such as "ga" or "gar" had to be detected in the words "garage" and "garcon". The target strings had either a CV or CVC structure which either matched or did not match the first syllable of the word. in normal readers, a syllable compatibility effect for unfamiliar words was observed (detection times were faster when the target corresponded to the first syllable of the word than when it did not). There was also a target length effect for familiar words, with two-letter targets strings responded to fastest. In contrast, for the below average readers the syllable compatibility effect was observed for familiar words, and the target length effect for unfamiliar words. In dyslexic readers, only a target length effect was found at all levels of word familiarity. The results indicated first, that normal readers processed unfamiliar words phonologically, whereas familiar words were processed orthographically. Second, below average readers processed words phonologically irrespective of familiarity, but used syllabic units with familiar words and phonemic units with unfamiliar words. Third, dyslexic readers used a letter-by-letter strategy involving a visuo-orthographic procedure rather than a phonological one. A previous longitudinal study established that the dyslexics (as well as the below average readers, but to a lesser extent) showed a poorer performance on both syllable and phoneme deletion tasks. Thus the letter-by-letter strategy of the dyslexics might be explained by their severe phonological deficits.