Learning Climate Change as a Complex System: Analogical Encoding Meets Agent-Based Models

• In Event: Developing Evidence of Understanding Complex Systems

Abstract

This paper reports on the second year findings from a four-year program of research in Australia exploring ways in which students could learn key conceptual aspects of complex systems of relevance to deeply understanding climate systems and change. Goldstone and Wilensky (2008) have proposed that the representational affordances of agent-based models (ABMs) may activate the visuospatial systems of students to “perceptually ground” learning of complex systems ideas in ways that are impossible to do with equation-based representations.
It is hypothesized that ABMs provide appropriate representations of climate and complex systems knowledge, but the type of learning activities involving the models will also be a critical factor. Earlier research by the first author (Jacobson, Kapur, So, & Lee, 2011) demonstrated that contrasting and comparing different ABMs can enhance the understanding of complex systems concepts, so this study was designed with a treatment condition based an Analogical Encoding Theory (AET) set of learning activities. AET (Gentner, Loewenstein, & Thompson, 2003) posits that contrasting and comparing cases sharing an underlying principle or concept may help a learner focus on the conceptual similarities rather than on idiosyncratic surface features, which in turn may lead to the abstraction of schema associated with the targeted concept.

Table 1. Climate and complex systems content knowledge covered in the study
Content Areas Topics Day
Climate concepts • Temperature and rainfall
• Carbon cycle and greenhouse gases
• Fossil fuel use, albedo, storm formation 1
2
3
Complexity concepts • Feedback loops
• Dynamic equilibrium
• Emergent properties 1
2
3

The design-based experiment included two groups: experimental analogical encoding (AE) and the comparison non-analogical encoding (NAE). The climate concepts and complex systems content for each of the groups was identical (Table 1). Three NetLogo ABMs were developed for climate and complex systems concepts. The AE group used three additional non-climate models from the NetLogo library that represented the same complex systems concepts, whereas the NAE group only used models about climate and complex systems concepts. Activities were carried out in pairs, where two students collaboratively worked on one computer but individually answered the pre- and post-tests (Table 2).

Table 2. Experimental design sequence
Group Models with learning activities
Day 1 Day 2 Day 3 Day 4 Day 5
AE Pre-test Complex systems model 1 Complex systems model 2 Complex systems model 3 Post-test
Climate and complex systems model 1 Climate and complex systems model 2 Climate and complex systems model 3
NAE Climate and complex systems model 1 Climate and complex systems model 2 Climate and complex systems model 3

It was found that both groups significantly increased their scores on climate concepts by the posttest (essentially replicating the previous year’s study (Jacobson, Markauskaite, & Kelly, 2012), however, only the AE group demonstrated significantly increased knowledge of the complex systems concepts. These findings will be discussed more completely as well as theoretical and practical considerations in the final conference paper.

Authors

• Michael J. Jacobson, The University of Sydney

• Yong Moon Jung

• Polly K. Lai, The University of Sydney

• Paul G. Stokes, Hornsby Girls School