Visual processes in word recognition by poor readers: Relative use of visual, phonological and orthographic cues

In two reading level design studies, poor and normal readers, matched for reading age, were presented visual matching tasks on a computer screen. In Experiment 1, letter symbols were used. The strings consisted of uppercase-lowercase congruent letters (e.g., o/O) or uppercase-lowercase incongruent letters (e.g. a/A). Real words and pseudowords were used. Poor readers needed significantly more time to match uppercase-lowercase incongruent pairs, especially when the pairs consisted of pseudowords. Experiment 2 examined if phonological, orthographic, or visual mechanisms were responsible for this effect. Strings of letters, digit strings and abstract figure symbols were used Letter strings included words, pseudowords or nonwords. Poor readers needed more time to match incongruent letter case pairs, consistent with Experiment 1. Poor readers were slower on letter and digit suing matching, but not on the figure symbol matching task. No evidence was found for differential use of orthographic (multiletter) constraints. The combined data on the letter, digit and graphic symbol matching experiments point to inadequate access to grapheme-phoneme associations as the underlying factor accounting for the effects found in Experiment 1. Evidence for poor visual processing as an independent factor accounting for ability differences in reading could not be established.