Improving students’ self-evaluation of learning for key concepts in textbook materials

Abstract

Why do students have difficulties judging the correctness of information they recall (e.g., definitions of key concepts in textbooks), and how can students improve their judgement accuracy? To answer these questions, we had college students read six expository passages, each including four key terms with definitions. After reading a text, each key term was presented, and participants (a) attempted to recall the corresponding definition and (b) self-scored the correctness of the response (incorrect, partially correct, or entirely correct). Participants were overconfident, with inflated judgements for responses that were objectively incorrect. When participants could inspect correct definitions while judging their responses, judgement accuracy improved. Counterintuitively, however, some overconfidence remained. We discuss implications of these results for theory, education, and the two questions posed above.

Acknowledgements

Preparation of this report was supported by Cognition and Student Learning grant No. R305H050038 from the US Department of Education.

Notes

¹For all analyses involving self-score judgements, we also conducted nonparametric inferential tests (Mann-Whitey Test) to compare self-score judgements from the standard group versus the no standard group. All of these tests supported the same statistical conclusions as obtained from the parametric tests.

²By definition, mean prejudgement recall was 0% for omissions and commissions, and 100% for correct responses. Mean prejudgement recall across partially correct responses was 40% (SE=2) in the standard group and 36% (SE=3) in the no standard group. Note that prejudgement recall for partially correct responses is less than 50% because of the differences between the criterion used to score recall and the criterion used to categorise prejudgement responses. A response was categorised as partially correct if it contained at least one of the idea units from the correct answer, whereas a recall score of half credit was assigned only if the response had 50% or more of the correct idea units. Thus, some responses were categorised as partially correct but received scores of 0 (e.g., when only one of four idea units was recalled). The logic of this categorisation scheme was to keep the commission category “pure”, so that responses in this category contained no objectively correct information. Accordingly, responses that contained any correct information were included in the partial category, even if it did not contain enough to merit a half credit score.