| Description |
This study used Kintsch's Construction Integration (CI) Model as a context; for investigating reading comprehension assessment and its relationship to epistemic beliefs. Specifically, questions tied to levels of representation delineated in the CI model were used to investigate how individual differences may predict reading comprehension. Literal questions that tie to the text;base representation of text;, and inference questions that tie to the situation model representation of text;, were used to investigate these effects. In addition to question type, text; availability was also manipulated in this study. Previous studies have reported that a person's epistemic beliefs may have an effect on his or her ability to comprehend text;. The current study was designed to investigate these findings using additional measures as covariates (i.e., vocabulary ability, background knowledge and working memory), and a test structure that is cognitively demanding (e.g., short answer questions to a variety of passages from three domains with and without text; available). Participants completed individual measures and the reading assessment in a 2-hour session. Although epistemic beliefs did account for a significant portion of variance in the reading assessment, this was very small, especially compared to the predictive validity of background knowledge and vocabulary ability. As predicted by the CI model, question type was related to reading comprehension performance and this interacted with epistemic beliefs. Surprisingly, this effect was found with literal questions rather than inference questions. text; availability did not interact with epistemic beliefs. The results of this study suggest that when predicting reading comprehension, it is essential to use several individual differences variables, and that the relationship of epistemic beliefs with reading comprehension is less definitive than indicated in the research literature. Future work in reading comprehension research should establish other individual difference variables such as reader's interest level and specific strategy use when answering literal and inference questions using expository text;. |
