Dyslexic Reliance on Statistical Knowledge and Lexical Processing Under Time-Pressure Conditions

The prevalent reliance on sublexical processing leads to slower reading times in adults with dyslexia. This study examined the lexical processing and use of distributional knowledge by adults with dyslexia under time-pressure conditions. Forty-four adults with dyslexia (mean age: 21.4 years) and 51 typical readers (mean age: 20.9 years) performed an orthographic judgement task involving high- and low-frequency irregular words (with either typical or atypical transcription) and regular words. Stimuli were presented in pure lists. Fakes were homophonic only to irregular words. Accuracy scores within the time limit were analyzed using an ANOVA, with group (dyslexic vs controls) as a between-subjects factor and stimulus type (regular, typical irregular, atypical irregular) and word frequency (high vs low) as within-subjects factors. The analysis included correct words and derived fakes, focusing on the frequency effect of the original words. A second ANOVA incorporated the version factor (correct words vs fakes). Variations in the number and order of words and fakes across pages, along with incomplete judgments due to time constraints, mean the version effect findings are suggestive. Participants performed worse when judging low-frequency irregular stimuli, especially those with atypical transcription. Controls showed a transcription typicality effect (better performance on typically transcribed words than atypical ones) and a very large regularity effect (better performance on regular words than irregular ones); this latter effect was maximized by using a pure (blocked) list. Adults with dyslexia showed a smaller regularity effect, significant only when atypical irregular stimuli and regular stimuli were compared. Statistical knowledge allowed them to process irregular words with typical transcription more accurately, achieving an accuracy like regular stimuli. Furthermore, adults with dyslexia showed a larger version effect (with greater difficulty in detecting homophonic fakes) and smaller pure list facilitation compared to typically developing readers. The results indicate that adults with dyslexia have a limited orthographic lexicon. However, they partially compensate for their lexical difficulty by the reliance on statistical-distributional knowledge. Time pressure further compromises dyslexic performance due to their reliance on sublexical reading strategy but does not undermine their use of statistical-distributional knowledge as a compensatory strategy.