Within-subject and between-subject conceptions of metacomprehension accuracy

• In two experiments, students watch multimedia science lesson on immune system.
• They make judgments of learning after each of five sections, then take final test.
• Within-subject accuracy is a correlation of section judgment and score within learners.
• Between-subject accuracy is a correlation of mean judgment and score across learners.
• Students show significant metacomprehension accuracy with both conceptualizations.

This study examines two conceptualizations of metacomprehension accuracy, which is recognized as an important feature of individual differences in learning. Within-subject accuracy requires students to judge their learning of parts of a lesson, and take tests for each part. Within-subject accuracy is computed by correlating judgments with performance on all parts for each student. A second conceptualization, between-subject accuracy, is computed by correlating each student's average judgment with average performance across all students. This measure corresponds to how well students judge their learning of the entire lesson compared to other students, and does not require multiple isolated lessons and tests for each learner. In Experiment 1, participants watched multimedia science lessons, completed a summarizing task for each lesson, made a learning judgment, and took transfer and retention tests. In Experiment 2, practice questions replaced the summaries. The study tasks did not consistently improve either type of metacomprehension accuracy. Overall, students generally displayed a statistically significant level of metacomprehension accuracy both on within-subject and between-subject measures of metacomprehension accuracy, yielding support for adding between-subject measures as a complement to the more commonly used within-subject measures.