Investigating Reading Behavior and Inference-making in Advanced L2 Reading Comprehension Assessment Tasks
Author ORCID Identifier
Date of Award
Doctor of Philosophy (PhD)
Applied Linguistics and English as a Second Language
Despite the ubiquity of reading comprehension tasks in English language proficiency tests (or sections of tests), the constructs underlying successful reading comprehension in English as a second/additional language at the advanced academic level are still not completely understood. Part of the reason for this gap in the current state of knowledge comes from how existing models of second language reading neglect higher-order reading skills. Many reading assessments overly target language proficiency skills and assume the transfer of first language literacy skills, leaving unexamined the higher-order skills of language learners who become skilled academic readers in their second or additional language. This study seeks to address the dearth of research on higher-order reading skills in advanced second language reading comprehension by examining the activation of these skills in realistic L2 reading comprehension tasks. A reading comprehension test with three different tasks (MC questions, cloze, and summary) was developed and administered to 102 second language English and multilingual undergraduate and graduate students studying at a university in the US. Eye-movement behavior was recorded during these tasks, and each reading task was followed by a sentence verification task to measure activation of inferencing. Eye-movement behavior and inferencing are compared across the reading tasks, and additionally compared to language proficiency and reading comprehension scores. The tasks each elicited distinct patterns of reading behavior: the cloze task elicited careful local reading, the MC task elicited expeditious linear reading, and the summary task elicited both careful global reading and expeditious strategies. Cloze scores were closely related to language proficiency, but also related to reasoning ability and processing efficiency. MC scores were unrelated to proficiency. They were instead related more to reasoning ability and were predicted by readers’ ability to efficiently process the MC questions. Inferencing ability was only predictive of score in the summary task. Summary scores were additionally influenced by global attention to the text, processing efficiency, reading motivation, and language proficiency. Implications for the use of each task as L2 reading assessment are discussed, as well as implications for the teaching of second language reading.
Tywoniw, Rurik, "Investigating Reading Behavior and Inference-making in Advanced L2 Reading Comprehension Assessment Tasks." Dissertation, Georgia State University, 2020.
File Upload Confirmation