The aim of the present study was to examine the effect of refutation texts on generating explanations and explore individual differences in quality of explanations. Seventy-five undergraduates read refutation and non-refutation texts addressing common scientific misconceptions and completed a two-tier post-test including True/False and open-ended questions. Readers' written post-test explanations were coded on accuracy and quality dimensions. The analysis of these explanations showed less circularity and uncertainty descriptors and more accuracy and causal connections in the refutation than the non-refutation condition. Further, three distinct clusters of readers emerged: coherence-building readers, non-coherence building readers, and promiscuous readers. Results demonstrate examining explanation characteristics can provide a useful tool for revealing how readers' knowledge influences learning from texts.
- Knowledge revision