These data are from an eye movement experiment examining the effect of semantic typicality and distance on online anaphor resolution during reading in children, and the extent to which individual differences in reading comprehension and working memory influence the pattern of effects observed in the eye movement data. The .csv data file contains all eye movement measures for each region of text (Regions 1-7 for each eye movement measure; columns in the data file are labelled accordingly), plus measures of decoding efficiency (TOWRE in the data file), reading comprehension (YARC5 in the data file) and working memory (AWMA in the data file) for each participant. Reading is a crucial skill for success in today’s society, the ultimate goal of which is to extract meaning from written text. However, approximately 10 per cent of 7-11 year olds in mainstream schools have specific difficulties with reading comprehension, despite age-appropriate levels of reading accuracy, fluency, and phonological skills, placing them at increased risk of poor educational attainment. Although we know that poor comprehenders’ difficulties lie with sentence and discourse-level processing (rather than word-level processing), critically, we do not know why discourse comprehension fails. Identifying the reason for comprehension failure is essential if the specificity of text comprehension training with this population is to be improved. The proposed research will monitor poor comprehenders’ eye movements as they read discourse in order to assess their moment-to-moment understanding of written language in real time, thus elucidating exactly when and why comprehension fails and pointing the way to appropriate intervention.

Monitoring eye movements during reading. 30 children participated in the experiment, their eye movements monitored as they read 16 short paragraphs that required the resolution of an anaphor. The typicality and distance (from the anaphor) of the antecedent was experimentally manipulated, and we examined reading times and regression probabilities on the anaphor itself (Region 4), the post-anaphor region (Region 5) and the antecedent (Region 2). We analysed the data using Linear Mixed Models.