Combining Computational Modeling, Theory, and Data: Steps Toward a Metamodel Framework for the Study of Learning

• In Event: Poster Session 6
  In Poster Session: Learning Sciences SIG Poster Session

Abstract

Learning sciences research can benefit from an integration of computational modeling of learning theory with quantitative and qualitative data. This paper describes a Meta-Model Framework (MMF) for this integration. An agent-based model (ABM) of learning was developed that was based on Analogical Encoding Theory (AET) to inform the design of research on learning and conceptual change. In particular, the AET ABM indicated there would be a non-linear “tipping point” for conceptual learning about complex systems and climate knowledge based on contrasting and comparing different scientific models versus learning with a single scientific model. The empirical findings from a classroom-based study were consistent with this ABM. The implications of the MMF for research on learning more generally are explored.

Authors

• James A. Levin, University of California - San Diego

• Michael J. Jacobson, The University of Sydney

• Lina Markauskaite, The University of Sydney