The Impact on Learning of Generating vs. Selecting Descriptions in Analyzing Algebra Example Solutions

Self-explanation of worked examples is an effective strategy for student learning. This paper examines the impact on learning of self-generating explanations of worked examples vs. selecting them from a menu in an intelligent tutoring system. In this study, students describe the structure of algebraic models of real-world problem situations. In one tutor version students select their descriptions from menus and in a second version students type their explanations in their own words. In a third version menu-selection and self-generation are interleaved. In this condition the canonical menu entries may serve to scaffold self-generated descriptions. Students completed the problems fastest with the menu version and the students learned to both explain and generate the target algebraic models equally well in all versions. However, the type-in version led to more successful transfer to describing novel algebraic models of problem situations. The scaffolded version was the least successful of the three.