Learner Characteristics and Feedback in Tutorial Dialogue

Learner Characteristics and Feedback in Tutorial Dialogue Kristy Elizabeth Robert Michael D. Mladen A. James C. a ab ab a a Boyer Phillips Wallis Vouk Lester aDepartment of Computer Science, North Carolina State University bApplied Research Associates, Inc. Raleigh, North Carolina, USA {keboyer, rphilli, mdwallis, vouk, lester}@ncsu.edu ber of systems devised to support a broad range of Abstract conversational phenomena. Systems such as CIRCSIM (Evens and Michael 2006), BEETLE (Zinn Tutorial dialogue has been the subject of in- et al. 2002), the Geometry Explanation Tutor creasing attention in recent years, and it has (Aleven et al. 2003), Why2/Atlas (VanLehn et al. become evident that empirical studies of hu- 2002), ITSpoke (Litman et al. 2006), SCOT (Pon-man-human tutorial dialogue can contribute Barry et al. 2006), ProPL (Lane and VanLehn important insights to the design of computa-2005) and AutoTutor (Graesser et al. 2003) support tional models of dialogue. This paper reports research that has begun to the see the emergence of on a corpus study of human-human tutorial a core set of foundational requirements for mixed-dialogue transpiring in the course of problem-initiative natural language interaction that occurs in solving in a learning environment for intro-the kind of tutorial dialogue investigated here. ductory computer science. Analyses suggest that the choice of corrective tutorial strategy Moreover, recent years have ...