Age Differences in Children's Monitoring of Open-Ended and True-False Questions After Reading Texts

• In Event: Advances in Understanding Metacomprehension Accuracy of Expository Text

Abstract

Theory and Objectives: Students’ ability to monitor their text comprehension is a crucial skill that has been considered the basis for regulation of study behaviour and academic performance (e.g., author & colleagues, 2013). Throughout elementary school, children’s evaluations of their performance become more accurate (author & colleague, 2010). However, compared to associative learning tasks, metacognition for text comprehension is less accurate, indicating that distinguishing between what one knows and doesn’t know is difficult when learning from text (Thiede, Griffin, Wiley, & Redford, 2009). There are indications that, when monitoring performance on text questions after text study, the format of these questions may impact the accuracy of Confidence Judgments (CJs). Research using an event-recall task showed that children more accurately monitor the correctness of their answers for open-ended than for yes/no questions (author & colleagues, 2007). When studying texts, it is yet unknown how question type affects children’s monitoring accuracy, and whether age difference in monitoring accuracy of test performance depends on the type of test questions. The aim of the current study is to examine age differences in elementary school children’s ability to monitor their performance on open-ended and true-false questions after reading expository texts.

Methods: Second-graders (N = 140) and 4th-graders (N = 165) read six expository texts. After reading, children completed a test with 12 open-ended questions (e.g. “What happens to a bee’s stinger after stinging a human?”) and 12 statements that had to be indicated as either true or false (e.g. “Bees can survive after stinging.”). After giving each answer, children made a CJ, indicating confidence that the given answer was correct on a 7-point scale ranging from “very unsure” to “very sure”.

Results: Intra-individual Gamma correlations (Nelson, 1984) between scores and CJs were calculated to determine monitoring accuracy (values closer to +1 indicate higher accuracy). For true-false statements, M Gamma is .08 (SD = .57) for 2nd and .06 (SD = .56) for 4th graders, there were no age differences. For open-ended questions, M Gamma is .28 (SD = .66) for 2nd and .47 (SD = .50) for 4th graders, monitoring was more accurate for 4th graders (p = .007). Moreover, monitoring was more accurate for open-ended than true-false questions (p < .001).

Significance: In sum, findings show low monitoring accuracy for true-false tests, and no age related progression in monitoring accuracy for this question type. However, for open-ended questions, CJs are more accurate, and results show an age related progression in monitoring skills. However, even for open-ended questions, monitoring was far from accurate. Inaccurate monitoring has harmful effects on further study regulation and achievement. Our results suggest that educational practitioners should be aware that in a self-regulated learning context, tests with true-false questions may have adverse effects on students’ monitoring of their text comprehension. When aiming to enhance monitoring and regulation of learning in elementary school children, open-ended questions seem to be more suitable.

Authors

• Martina Steiner, University of Bern

• Mariette Van Loon, University of Bern

• Natalie Guggisberg

• Claudia Roebers, University of Bern