The Effects Of Working Memory And Story Structure On Senior High School Students’ On-line Causal Inference

Text comprehension has been one of the hot issues in cognitive psychology study. Thepurpose of readers is to obtain the information from the text and combine it with their priorknowledge and then to construct meaning representation of the text. A considerable number ofinvestigations indicate that the process of text comprehension is constrained by numerousfactors, which can be divided in two major aspects: text-based factors and reader-basedfactors. Some of the major text-based factors are text readability, text title and text structure,and so on. However, language proficiency, background knowledge, reading purpose andworking memory span are important reader-based factors that influence text comprehension.The current study explores the interactive effects of working memory and story structureon the causal inference generation during L2narrative text comprehension by85Chinesesenior high school students. Adopting a2(working memory: high and low)×2(storystructure: sequential and embedded) factorial design, an on-line experiment was conductedwith the help of E-prime software to measure the reaction time and causal inferenceperformance in reading comprehension.The results indicated that:(1) Working memory and story structure affect senior highschool students’ causal inference performance during L2narrative text comprehension.High-span subjects performed better in both sequential structure and embedded structurestories, whereas low-span counterparts failed to do so. Compared with the embedded structure,subjects generated causal inferences more correctly when reading sequential structure story.In addition, there is significant interactive effect on reading comprehension performance ofon-line causal inferences between working memory and story structure. There is significantdifference in performance between high-span and low-span subjects when reading sequentialstructure, namely high-span subjects performed better than low-span counterparts insequential structure. Moreover, there is also significant difference in performance betweenhigh-span and low-span subjects when reading embedded structure. To put it in another way, high-span subjects generated causal inferences more correctly than low-span counterparts inembedded structure. However, for low-span subjects, there is no significant difference onlow-span counterparts’ causal inferences both in sequential structure and embedded structure.In addition, for high-span subjects, there is significant difference in performance betweensequential structure and embedded structure. In other words, high-span subjects performedbetter when reading sequential structure than embedded structure.(2) Working memory andstory structure affect senior high school students’reaction time of causal inferences during L2narrative text comprehension. Both high-span and low-span subjects reacted to thecomprehension questions faster in the sequential structure than in embedded structure. Andthe reaction time for high-span subjects is much shorter than for low-span subjects both insequential structure and embedded structure. However, there is no interactive effect onreaction time of on-line causal inferences between working memory and story structure. Theabove findings show that both working memory and story structure are important factors thataffect L2learners’on-line causal inferences during narrative text comprehension.