Metacomprehension accuracy with multiple texts: a comprehensive approach

Abstract

This study investigated metacomprehension accuracy for university students reading multiple texts about the same topic, using a Dutch translation of a German multiple document comprehension instrument. Based on research with single texts, we expected students’ performance estimates to have relatively low accuracy, as well as being associated with several student-related and task-related variables. Moreover, we expected metacomprehension accuracy to be related to students’ ability and confidence. We conducted our study with 401 psychology students from one Dutch university, asking for their exam grades, multiple document comprehension habits, self-concept, and a prediction of their performance. Next, students completed two tasks from the multiple document comprehension instrument and after each task provided performance estimates and confidence ratings, rated perceived difficulty, mental effort, interest, text cohesion and ease of comprehension. Finally, students gave a postdiction of their overall performance. Results showed a significant association between task performance and performance estimates of .33, comparable to what has been found with other tasks. Beyond the variance explained by task performance, students’ estimates were predicted by exam grade, self-concept, and perceived task difficulty, but not by any of the other variables. Postdictions were more accurate than predictions, suggesting that task experience allows students to use more valid cues in their performance estimates. Other than with single texts however, metacomprehension accuracy was unrelated to students’ ability and confidence. These findings can inform efforts to improve university students’ metacomprehension accuracy with multiple texts, benefitting their self-regulation.