Abstract
Four lines of evidence, proceeding from correlational outcomes to findings from true experiments, are reviewed as preliminary support for the hypothesis that learning is increased when students attempt to construct explanatory answers to questions about to-be-learned content (i.e., answers requiring inferential transformation of questioned material). The lines of evidence are as follows: (a) When students generate explanatory answers to questions as part of learning in a group, better learning occurs for students doing the explaining. (b) Manipulations that increase student generation of explanations to questions during group interaction increase learning. (e) Attempting: to predict the content of upcoming text by responding to prequestions (including explanatory justifications for predicted answers) increases learning of prequestioned content. (d) Attempting to explain the significance of to-be-learned facts increases learning of those facts. Requiring learners to construct explanatory answers about new content probably improves learning because students do not spontaneously attempt to explain to themselves the significance of to-be-learned content. We hypothesize that attempting an explanation induces the learner to relate the new material to prior knowledge. Potential directions for future research are emphasized.