A Study Of The Effects Of Two Reader-based Factors On On-line Causal Inference Generation During L2 Expository Text Comprehension

This dissertation presents a study of the effects of two reader factors: reader's reading ability and working memory span on Chinese EFL learners'on-line causal inference generation during expository text comprehension. With an attempt to reveal the cognitive process of causal inference generation and the effects of factors influencing causal inferencing during the process of expository text comprehension, and based on relevant theories and empirical studies, the study aimed to test the following two hypotheses: for current subjects, first, readers with higher reading ability have faster and more accurate responses to causal inference generations while their lower reading ability counterparts do not; second, readers with larger functional working memory spans have faster and better responses to targets, whereas their low functional working memory span counterparts do not. In order to test these two hypotheses, altogether 51 subjects were selected to participate in a causal inference generation experiment, which was designed and carried out among non-English majors from the College of Foreign Languages of Qingdao University of Science and Technology. The Reading Time(RT)method was adopted, with the psychological research software DMDX applied to this investigation to record the subjects'response times and response results.The current study replicated yet expanded former L1 studies to test the effects of the two reader-based factors on the on-line generation of causal inferences during expository text comprehension.The results proved that the process of causal inference generation during expository text comprehension is affected by both EFL readers'working memory span and their reading ability. Firstly, readers with a high-span working memory can respond faster and make fewer errors than their low-span counterparts. Secondly, readers with a high reading ability can respond faster and make fewer errors than their low reading ability counterparts. The results of this experiment support the contention that readers'working memory span and reader's reading ability are important factors affecting the on-line generation of causal inferences.This thesis consists of five parts. It starts with an introduction, providing the research orientation, objectives and outline of this study. Chapter two contains a literature review and an introduction to relevant theories and empirical studies. Related with the theories, the author focused on the review and discussion of two topics: whether causal inferences can be generated or not during the reading process and factors that can affect the generation of causal inferences. Chapter three is concerned with the experiment with detailed descriptions of methods and procedures for the experiment being presented. Chapter four provides the results of reading comprehension and a general discussion of all the results. A summary of the major findings and implications of the present study, and suggestions for future studies are presented in the last chapter.