Modeling Learning Progressions in Epistemic Games with Epistemic Network Analysis
Rupp, A, Choi, Y, Gushta, M, Mislevy, R, Thies, MC, Bagley, E, Nash, P, Hatfield, D, Svarovsky, G, Shaffer DW. (2009). Modeling learning progressions in epistemic games with epistemic network analysis: Principles for data analysis and generation. Paper to be presented at the Learning Progressions in Science conference (LeaPS), Iowa City, IA, USA.
http://epistemicgames.org/eg/wp-content/uploads/leaps-learning-progressions-paper-rupp-et-al-2009-leaps-format1.pdf
Epistemic games have been developed to help players develop domain-specific expertise that characterizes how professionals in a particular domain reason, communicate, and act (Shaffer, 2006; Shaffer & Bagley, 2009). Grounded in a sociocultural and sociocognitive approach to learning, epistemic games are designed to foster situated learning that leads to data structures with high levels of dependencies. As one might expect, traditional measurement models struggle to accommodate such contextual dependencies, especially when data are collected at smaller scales and epistemic network analysis (ENA) has been developed to provide a practically feasible modeling alternative (e.g., Rupp et al., 2009; Shaffer et al., in press). In this paper, we describe a research program that addresses key statistical considerations for modeling data from epistemic games using ENA with an eye toward representing different learning progressions of players within such games. Current approaches for representing learning progressions using ENA are juxtaposed with approaches for simulating such data using particular statistical constraints.
