More confusion and frustration, better learning: The impact of erroneous examples
J. Elizabeth Richey, Carnegie Mellon University, United States ; Juan Miguel L. Andres-Bray, University of Pennsylvania, United States ; Michael Mogessie, Carnegie Mellon University, United States ; Richard Scruggs, Juliana M.A.L. Andres, University of Pennsylvania, United States ; Jon R. Star, Harvard University, United States ; Ryan S. Baker, University of Pennsylvania, United States ; Bruce M. McLaren, Carnegie Mellon University, United States
Computers & Education Volume 139, Number 1, ISSN 0360-1315 Publisher: Elsevier Ltd
Prior research suggests students can sometimes learn more effectively by explaining and correcting example problems that have been solved incorrectly, compared to problem-solving practice or studying correct solutions. It remains unclear, however, what role students' affect might play in the process of learning from erroneous examples. Specifically, it may be that students experience greater confusion and frustration while studying erroneous examples, but that their confusion and frustration lead to greater learning. We analyzed student log data from previously published research comparing erroneous example instruction of decimal number mathematics to problem-solving instruction in a computer-based intelligent tutoring system. We created and applied affect detectors for a combination of confusion and frustration (“confrustion”) and compared the role of confrustion across conditions. As predicted, students in the erroneous example condition experienced greater confrustion while working through the instructional materials. However, contrary to predictions, confrustion was negatively correlated with posttest and delayed posttest performance across conditions, though less so for the erroneous example condition. Given that students in the erroneous example condition performed better on the delayed posttest than students in the problem-solving condition, it appears they learned more despite also experiencing greater confrustion rather than because of it. Results suggest that learning from erroneous examples may be an inherently more confusing and frustrating process than traditional problem solving. More generally, this research demonstrates that logging student actions at a step-by-step problem-solving level and analyzing those logs to infer affect can be a powerful way to investigate learning.
Richey, J.E., Andres-Bray, J.M.L., Mogessie, M., Scruggs, R., Andres, J.M.A.L., Star, J.R., Baker, R.S. & McLaren, B.M. (2019). More confusion and frustration, better learning: The impact of erroneous examples. Computers & Education, 139(1), 173-190. Elsevier Ltd.