Can Self-Explanation Improve Metacomprehension Accuracy for Illustrated Text?

Previous work has suggested that adding illustrations to expository science texts can decrease metacomprehension accuracy. Not only are students unable to take advantage of the illustrations for informing their self-assessments, it seems they can actually be hurt by them. One hypothesis to account for these findings is that illustrations increase the salience of cues not based in representations of the text. In turn, this may influence readers to base their judgments on these cues rather than more valid cues based in the quality of their situation models. To test this hypothesis the current study investigated whether providing students with a self-explanation instruction would help to improve their metacomprehension accuracy for illustrated text. Self-explanation instructions were used because they have been shown to assist in mental model construction and inference generation, which also leads to increased access to cues based in readers’ situation models. Benefits of self-explanation were seen in conditions where participants read texts paired with conceptual images, but these same benefits were not found when texts were paired with decorative images or no images. An analysis of the cues participants reported using to make their metacognitive judgments indicated that participants were not using cues based in their situation model representations of the text in the no-image and decorative images conditions. Instead, participants in these conditions reported using cues such as their surface memory for the text, characteristics of the text, or information about the reader such as their interest or prior knowledge in the topics to make their judgments rather than their comprehension of the text. Future studies will investigate if further benefits in metacomprehension accuracy can be found when students are instructed how best to use the images during self-explanation.