A new tool for measuring and understanding individual differences in the component processes of reading comprehension.

We propose a new measure of individual differences in reading comprehension ability that is theoretically motivated, is easy to administer, and that has high predictive power. Participants read 3-sentence paragraphs that describe the relations among a set of real and artificial terms, and then they respond to true-false statements that assess their ability to access and integrate long-term memory knowledge with text information, to make text-based inferences, and to recall text. The components of our task predict performance on a test of global reading comprehension and on a range of specific comprehension tests, each of which draws more heavily on one particular component. Our task is better at predicting reading comprehension than is a typical working memory span task and has the potential for advancing researchers' understanding and measurement of a range of linguistic and cognitive tasks.