Understanding climate change as a complex system with agent-based models: A study of contrasting learning designs

This paper reports on a study in which students used agent-based computer models to learn about complex systems concepts of relevance to understanding climate change. Of empirical interest was if varying the Sequencing of Pedagogical Structure (SPS) provided with the computer models would result in differential learning outcomes of the targeted complexity and climate ideas. The experimental Challenge and Guided Learning (CGL) condition used a Low-to-High (LH) SPS sequence based on productive failure (PF) in which ninth grade students demonstrated significant learning gains for the targeted complex systems concepts. The comparison Teach and Challenge (TC) condition was based on a more traditional teaching approach, classified as a High-to-Low (HL) SPS sequence. After discussing the design of the study, the main findings are reported, which found significant learning of ideas such as “greenhouse gases” and “carbon cycle” by both groups on the posttest. However, for the more conceptually challenging complex system ideas, such as “self organization” and “emergent properties,” only the LH CGL group demonstrated a significantly higher performance on the posttest compared to the HL TC comparison condition. These results demonstrate the potential importance of complexity knowledge for understanding climate systems and change. Theoretical and practical implications of these findings are also considered.