UC Irvine

We investigated the contributions of multiple strands of factors-individual characteristics (struggling reader status, working memory, vocabulary, grammatical knowledge, knowledge-based inference, theory of mind, comprehension monitoring), a text feature (narrative vs. expository genre), and question types (literal and inferential)-to one's performance on discourse comprehension in oral language (listening comprehension), using data from 529 second graders. Results from explanatory item response models revealed that substantial variance in listening comprehension was attributable to differences between items, texts, and children, respectively. Narrative versus expository genre distinctions explained almost all of the variance attributable to text differences. In contrast, literal versus inferential question distinctions did not explain item responses after accounting for text and reading comprehension status. However, there was a moderation between struggling reader status and question type such that struggling readers had a slightly higher (2%) probability of getting inferential questions right compared to typically developing readers, after accounting for individual and text factors. Struggling readers have a lower probability of accurate item responses than typically developing readers, but the difference disappeared once language and cognitive skills (e.g., working memory, vocabulary) were taken into consideration. The effects of text genre and question type on item responses did not differ as a function of children's language and cognitive skills. Overall, these results underscore the importance of considering individual, text, and assessment factors for children's performance in listening comprehension.