Intelligent Dialogue Support for Physics Problem Solving: Some Preliminary Mixed Results
John Connelly and Sandra Katz
Many current intelligent tutoring systems aim to foster learning and skill acquisition in problem solving via the incorporation of dialogue between a student and a computerized tutor. In the present study, a follow-up to the study reported in Katz, Connelly, & Wilson (2005) on the staging of tutorial dialogue across different phases of problem solving, we varied the number and types of dialogues (focus on solution planning vs. subgoal execution) presented during problem solving. We expected students who saw both planning and subgoal dialogues to outperform students who saw only subgoal dialogues as well as students in a no-dialogue control group. Although we found no significant differences between treatment groups on our performance measures, the groups differed in their reliance on other available scaffolding. We also found evidence for the differentially greater benefit of our planning dialogues over our subgoal dialogues.