Model-Based Learning With Productive Failure and Analogical Encoding: Unpacking Learning Dynamics With Contrasting Designs

• In Event: 35.060 - Learning Sciences: STEM Learning With and Through Technological Tools

Abstract

Computer modeling, productive failure, and contrasting cases all support science learning, however, little is known about why and how students actually learn, particularly when these principles are combined. This paper, building on a larger study, explores some of these mechanisms by comparing learning processes of two high-performing groups learning about climate and complex systems under different design conditions. A detailed process analysis, including temporally-sensitive impact coding at three parallel grain sizes, reveal the potential productivity of contrasting case design. The superior performance of the treatment group may have been due to higher student engagement in generating ideas, including making more suggestions, predictions, experimental questions, experimental designs, and explanations. The emergent pattern of successful idea generation is also discussed.

Authors

• Alisha Erica Portolese, The University of Sydney

• Lina Markauskaite, The University of Sydney

• Polly K. Lai, The University of Sydney

• Michael J. Jacobson, The University of Sydney