The role of questioning in creating situation models while reading in a second language: does explaining events in a text matter?

Abstract

The primary purpose of this study was to explore ways in which teachers can increase their second language (L2) learners’ reading comprehension through constructing situation models. The author incorporated theoretical frameworks, including the situation model theory (Kintsch, 1998/2007), the event index model (Zwaan & Radvansky, 1998), and the linguistic threshold hypothesis (Clark, 1980; Aldersen, 1984; Carrell, 1991). As an educational intervention, a set of adjunct questions were asked during reading to elicit readers’ explanations of causality and intentionality to promote coherent comprehension. A total of 117 L2 readers’ ability to make appropriate inferences based on situation models was assessed through both quantitative (experimental-control design) and qualitative (think-aloud) methods. This experimental study examined the effect of explaining to target situation models, while also looking at the relationship between the readers’ inferential ability and their L2 proficiency. In addition to the main effects of the intervention and L2 proficiency, the interaction between the intervention and L2 proficiency were discussed as results of the quantitative analysis. Also discussed was the nature of the L2 readers’ responses to adjunct questions, which were designed to measure the quality of their explanations based on the underlying situations in the text that they were reading. This study extended existing research on situation model-level comprehension to L2 literacy, which has not previously been well studied. This made the study theoretically interesting as well as highly applicable to L2 reading instruction. The main findings of this study were: (1) there was a strong effect of question types (inferential vs. non-inferential questions) with greater accuracy overall for non-inferential questions, (2) there was a suggestive trend of question type interacting with L2 proficiency, such that higher proficiency participants showed little difference between question types, (3) there was no statistically significant main effect of the adjunct question manipulation on accuracy, and (4) there was a trend suggesting an interaction between experimental condition and L2 proficiency, with higher proficiency participants showing a somewhat larger effect of the adjunct question manipulation. Additionally, both quantitative and qualitative data trended in the direction consistent with the linguistic threshold hypothesis.