Tertiary students of elementary statistics often have difficulty developing a conceptual understanding of the subject. This article proposes a didactic method that stimulates students to self-explain the study material. This stimulus to self-explanation is directed, because the instructor decides which particular links and relationships between concepts the student should explore. The method is composed of two major steps: having the students identify the propositions that underlie the statistical theory, and then having them manipulate these propositions by constructing arguments showing a given statement to be either true or false. A practical way of implementing the method in a large-scale introductory statistics course is briefly described.