Learning From Explanations: Diagrams Can “Inhibit” the Self-Explanation Effect

Contents

Research shows that one can learn from self-explanations (Chi et al. 1989), but the influence of diagrams on this type of learning has not been assessed. This paper examines the role of spatial localization on the self-explanation effect. It is hypothesized that adjacency may blur feature discrimination, leading to inappropriate self-explanations. Ten subjects with naive conceptions about motion in a curved path were asked to think aloud while studying a chapter and a worked-out example. They were divided into Low-Benefit (LB) and High-Benefit (HB) learners, using a post-hoc median split on post-test measures (explanation, isomorphic, and transfer tests). Analyses of verbal protocols and drawings show that the LB learners self-explained without learning. They used diagrammatic representations extensively, relying on their features to make sense of the example lines, whereas the HB learners processed the text conceptually. Contrary to previous emphasis on the benefits of localization of information in a diagram (e.g., Larkin and Simon 1987), spatial localization inhibited the self-explanation effect because access to adjacent features propagated inaccurate comprehension via familiar diagrammatic knowledge. This research shows that learners in the process of acquiring new conceptual knowledge are not necessarily helped by local diagrammatic features because they can use them indiscriminately. It also suggests that the self-explanation effect may be tied to highly constrained learning situations.