Title of Document : VALIDATING A THEORY - BASED MODEL OF L 2 READING COMPREHENSION : RELATIVE CONTRIBUTIONS OF CONTENT - SPECIFIC SCHEMATIC KNOWLEDGE AND L 2 VOCABULARY KNOWLEDGE TO COMPREHENDING A SCIENCE TEXT

Title of Document: VALIDATING A THEORY-BASED MODEL OF L2 READING COMPREHENSION: RELATIVE CONTRIBUTIONS OF CONTENT-SPECIFIC SCHEMATIC KNOWLEDGE AND L2 VOCABULARY KNOWLEDGE TO COMPREHENDING A SCIENCE TEXT Eunjou Oh, Ph.D., 2010 Directed By: Professor Peter Afflerbach, Department of Curriculum and Instruction Professor Robert Mislevy, Department of Measurement, Statistics, and Evaluation The study explores the question of what explains second language (L2) reading comprehension by proposing a comprehensive theory building on the Construction Integration (CI) model of reading comprehension (Kintsch, 1998) and conducting an experimental study within this theoretical framework. The proposed theory maintains that the construction of a textbase is a function of L2 proficiency and the construction of a situation model is a function of first language (L1) reading competence. The effect of two different types of intervention tapping into each representation system (textbase and situation model) is experimentally tested; vocabulary knowledge, conducive to building textbase, and content-specific schematic knowledge, facilitative to building situation model. Two different measures of reading comprehension for both L1 and L2 reading comprehension are used to analyze how different cognitive processes are involved in L2 reading comprehension. Thirty two 9 grade Korean students were given a vocabulary acquisition activity and a content-specific schematic knowledge acquisition activity between a pretest and a posttest on science texts. The findings suggest that the ability to form macropropositions, as measured by a recall task, is a route through which L1 reading competence emerges. Thus, it is an influential factor for L2 reading comprehension. Different patterns in the role of L1 reading competence and L2 proficiency in different treatment conditions provide evidence for a reader constructing a textbase as a function of L2 proficiency and a reader constructing a situation model as a function of L1 reading competence. Three latent variables of textbase, situation model, and L2 reading comprehension were entered in LISREL to conduct structural equation modeling; the indicators of the textbase include the scores of vocabulary knowledge and the scores of listening comprehension (LC) and reading comprehension (RC) in an L2 proficiency measure; the indicators of the situation model include the scores of L1 reading competence and the scores of schematic knowledge; and the indicators of L2 reading comprehension include the scores of the pretests and the posttests. The fit indices of various Structural Equation Modeling (SEM) models of a given text demonstrate the viability of the comprehensive theory of L2 reading comprehension. VALIDATING A THEORY-BASED MODEL OF L2 READING COMPREHENSION: RELATIVE CONTRIBUTIONS OF CONTENT-SPECIFIC SCHEMATIC KNOWLEDGE AND L2 VOCABULARY KNOWLEDGE TO COMPREHENDING A SCIENCE TEXT.