The Effects of Feedback in Supporting Learning by Teaching in a Teachable Agent Environment

The idea that teaching others is a powerful way to learn is both intuitively compelling, and one that has garnered support in the research literature. The present study investigates aspects of the “learning by teaching” process that contribute to enhanced learning outcomes for students. We developed a computer-based teachable “agent” that students explicitly teach using concept maps. Results indicate that providing students with opportunities to quiz their agent decreases the amount of irrelevant information and increases the proportion of causal information in students’ maps, whereas having opportunities to query their agent increases the interconnectedness of concepts in students’ maps. The results point to the importance of including various forms of feedback in designing teachable agent environments that promote learning. Introduction The idea that teaching others is a powerful way to learn is both intuitively compelling, and one that has garnered support in the research literature. For example, Bargh and Schul (1980) found that people who prepared to teach others to take a quiz on a passage learned the passage better that those who prepared to take the quiz themselves. The literature on tutoring suggests a similar conclusion in that tutors have been shown to benefit as much from tutoring as their tutees (Graesser, Person, & Magliano, 1995; Chi, Siler, Jeong, Yamauchi, & Hausmann, 2001). Biswas and colleagues (Biswas, Schwartz, Bransford, & TAG-V, 2001) report that students preparing to teach made statements about how the responsibility to teach forced them to gain deeper understanding of the materials; other students focused on the importance of having a clear conceptual organization of the materials. Reflection on these studies and others lead us to conjecture that the creation of a computer-based system, where students can assume the role of “teacher,” may provide an effective and motivating environment for learning. We designed an environment that lets students explicitly teach a computer agent. Once taught, the agent reasons with its knowledge and answers questions. Students observe the effects of their teaching by analyzing these responses and by getting additional feedback from a teaching expert. Although the research literature (e.g., Bargh & Schul, 1980) suggests that learning benefits accrue from preparing to teach, it is not clear if and how other aspects of the learning by teaching process contribute to enhanced outcomes. In addition to preparatory activities, teachers provide explanations and demonstrations during teaching and receive questions and feedback from students. These activities also seem significant from the standpoint of their cognitive consequences. For example, we might expect that teachers’ knowledge structures would become better organized and differentiated through the process of communicating key ideas and relationships to students and reflecting on students’ questions and feedback (Chi, et al., 2001). The purpose of the present study was to examine the learning benefits of different activities associated with “learning by teaching” in our teachable agent environment. This was done by constructing computer-based agents that students could teach domain knowledge. In particular, we created an agent environment called Betty’s Brain which can operate in three modes: (i) the TEACH mode, where students impart knowledge to the agent Betty by means of a dynamic concept map interface, and access content materials as needed to learn information for teaching, (ii) the QUERY mode, where students ask Betty questions (using question templates) which she answers by reasoning with information that the student has taught her, and (iii) the QUIZ mode, where students evaluate how well they have taught Betty by observing her performance on a quiz. At times, an expert teacher agent intervenes to make suggestions that may help Betty (and the student) correct her answers. In the present study we examined the effects of the interactive features of the teachable agent environment that emulate the feedback that instructors receive from students during teaching. All students had the opportunity to TEACH their agent, and we manipulated whether students could QUERY Betty and observe her QUIZ performance following their teaching efforts. Crossing these variables created four versions of the teachable agent environment: 1. TEACH Only version (No QUERY or QUIZ), 2. QUERY version, 3. QUIZ version and 4. FULL version (QUERY & QUIZ). We hypothesized that having opportunities to query and/or quiz Betty would positively, but differentially, impact students’ learning. The query feature helps students debug their own thinking and reasoning in the problem domain. If Betty answers questions in unexpected ways, students know that they need to add to or modify their concept maps. In addition, and perhaps more important, when Betty explains her answers, she makes explicit the process of reasoning across links in a concept map (i.e., infer the effect of one concept on another through a chain of relations). Therefore, we might expect that students who use the QUERY versions of the software would create maps containing more inter-linked concepts. With respect to the quiz condition, we expected that students would become better at identifying important concepts and links to include in their maps because they could map backward from the quiz questions. We also expected that overall they would produce more accurate concept maps because they had access to feedback on Betty’s quiz performance.