Improving Monitoring Accuracy Improves Learning From Text

First, the team found that students’ metacomprehension accuracy is increased when students access the situation models that they create during reading. This mental representation allows students to identify what they do and do not understand during and after reading. However, the research of this team elaborates how students do not always access their situation models prior to making metacomprehension judgments. The research completed to address this outcome also discovered that at-risk readers use different (and less effective) tools to calibrate their understanding. Specifically, at-risk readers rely on surface features to make metacomprehension assessments. When at-risk readers were directed to use features other than surface features (through generating keywords at a delay, for example), their metacomprehension accuracy improved.

The second set of research studies addressed what types of instructional manipulations should be used in order to improve metacomprehension accuracy. The major finding from this series of studies is that asking students to generate keywords that summarize the gist of what was just read after a delay is one powerful way to improve metacomprehension accuracy for all readers. In addition, the studies found that other tasks which require students to generate information about a text recently read after a delay can assist metacomprehension accuracy. Self-explanation is another demonstrated technique that improves metacomprehension accuracy. Finally, the use of concept mapping while reading expository texts was effective for at-risk college readers, but did not improve the metacomprehension accuracy ratings of normal college readers.

Finally, the research team has developed a new set of techniques that were not originally proposed, but which hold promise for fostering metacomprehension accuracy. This technique entails giving readers practice with test questions that require inference generation. This then seems to force readers to generate their own situation models. This line of research was continued in their subsequent Cognition and Student Learning award.