Adapting design patterns across learning environments

Abstract

Pedagogical design patterns offer high-quality solutions to known educational problems. Design patterns are meant to guide teachers' pedagogical decisions to improve student learning in varied learning contexts such as intelligent tutoring systems, traditional classrooms, online learning systems, and so forth. In theory, design patterns are written so that they are applicable to multiple learning contexts, but, in practice, we wonder if pedagogical design patterns intended specifically for one learning environment can be used in other learning environments. We explore this theoretical question for practical reasons. Over the past three years, we applied existing pedagogical design patterns and wrote new pedagogical patterns to enhance student feedback for math problems in an intelligent tutoring system (ITS) called ASSISTments. It was more difficult than expected for two reasons: there are few pedagogical design patterns specifically for ITSs and contextual features for design patterns developed for other learning environments make them either too general or too specific to apply to an ITS. For example, feedback design patterns that involve interpreting learners' misconceptions may be easy for teachers to apply in traditional classroom settings, but difficult for ITSs because current algorithms poorly predict student misconceptions. In this paper, we explore the adaptability of design patterns to different learning environments using guide questions and contextual design pattern features we developed from our experience of adapting design patterns to ITSs.