Calculategy: Exploring the Impact of Tutorial Dialogue Strategy in Shaping Student Behavior in Effective Tutorial Dialogue for Calculus
Funded by: NSF
PI: Carolyn Rose
Collaborators: Erica Melis at DFKI, Saarbruecken, Germany
Students and Staff: Gahgene Gweon, Regan Carey, Zachary (Sam) Zaiss
Website: http://www.cs.cmu.edu/~cprose/Calc.html
The purpose of this project is to evaluate the instructional value of errors in connection with meta-cognitive behaviors such as self-explanation. In our recent work we have conducted a controlled investigation of collaborative problem solving in which the behavior of a confederate peer learner is manipulated in order to measure the impact of the confederate peer learner.s behavior on the behavior and learning of the student participant. The independent variables in the 2X2 factorial design include level of engagement (Lazy versus Engaged) and accuracy of problem solving contributions (High performing versus Low performing). We did not find evidence that the errors contributed by the confederate peer learners were harmful to student subjects working with them except in the case of students paired with Engaged Low performing peer learners. On the contrary, we found a small but reliable interaction effect in which students paired with Lazy Low performing peer learners derived some benefit from the errors they were exposed to whereas students paired with Engaged Low performing peer learners were harmed by the errors they were exposed to. An understanding of where errors can be used strategically to stimulate cognitive conflict and student learning may enhance the effectiveness of existing well-established approaches to scaffolding in intelligent tutoring systems. Nevertheless, this is an issue that requires much more investigation. Because the majority of the observed learning in this study is explained by correct problem solving, these results do not argue that errors play a large role in student learning relative to correct examples, contrary to what might be predicted by theories such as impasse driven learning. This discrepancy might be explained by a paucity of what is referred to as .high level. explanation and help seeking behaviors found in our corpus of collaborative problem solving interactions. In future work we plan to explore ways of eliciting more high level explanation and help seeking behaviors.