Learning by explaining: the effects of software agents as learning partners

Abstract

Although prior research as shown that generating explanations encourages students to learn new content with deeper understanding and to monitor their own comprehension more effectively, helping students learn how to explain properly remains a significant challenge. This study investigated the use of software agents as learning partners in an activity where students generated explanations about river ecosystem concepts. The results of the experiment demonstrated that software agents can have a positive impact as learning partners in a virtual world environment. It was found that the agents encouraged the use of explanation resources designed to help students generate more effective explanations. Students working with the agents generated deeper explanations than students who did not interact with an agent. A summary report of student explanations produced for the participating teacher was perceived to be a useful source of feedback.