The Complex Systems Conceptual Framework of Learning and Methodological Implications

• In Event: 81.026 - WITHDRAWN: Theories of Learning and Education Through the Looking Glass of Complexity

Abstract

Embracing complexity conceptual perspectives for educational research can also have methodological implications. Our respective research activities have generally employed a range of quantitative and qualitative methods (e.g., experimental and quasi-experimental designs, discourse analysis, data mining, observations) and we have also used conceptual aspects of the CSCFL in our research, such as Kapur, Voiklis, and Kinzer (2007) (e.g., sensitivity to initial conditions) and Jacobson et al. (2017) (e.g., system levels, emergence). In this presentation, we propose to extend the applicability of the CSCFL from considerations of theory to another key issue in the field: methodologies for doing research and their suitability for studying learning as an emergent phenomenon. Our reading of the educational research literatures associated with cognitive and situative empirical studies is that researchers in these “camps” typically use different methodologies. In general, we find papers based on cognitive perspectives tend to employ quantitative research methods, whereas situative researchers tend to use qualitative and descriptive methods (i.e., design research), with some use of mixed methods (of course, we recognize there are exceptions).
This is important as there is an issue about whether either quantitative or qualitative methods alone are sufficient for testing explanations for the emergence of learning. Unfortunately, quantitative approaches generally used in education use mathematical modelling tools (e.g., differential equations, statistical models) that are primarily (but not exclusively) linear tools to break a system into its components or parts, study the parts individually, and then add the parts together to form the whole. However, emergent phenomena have important characteristics that are nonlinear and that cannot be analyzed by “adding up the parts” (Jacobson & Kapur, 2012). There also are limitations for using qualitative methods. For example, these approaches generally focus on observable events at a behavioral level, although from the CSCFL perspective, learning events are sustained by complex neural, cognitive, and situated events that are not observable as behaviors or activities, and the behavior-level events are constrained by socio-cultural forces that are also not directly observable.
We proposed to address these inherent limitations with quantitative and qualitative approaches used in educational research with a hybrid approach that incorporates methodologies being used to study other types of complex physical and social systems (Jacobson & Kapur, 2012), such as computer modeling methods such as agent-based and network modeling (ABM) that are increasingly being used in the natural sciences (Holland, 1995; Mitchell, 2009) and social science areas (e.g., Epstein, 2006). The symposium presentation will discuss research employing such hybrid methods (Abrahamson, Blikstein, & Wilensky, 2007; Kapur, Voiklis, & Kinzer, 2008; Levin & Datnow, 2012; Maroulis et al., 2010), and considered challenges and issues for using such approaches.

Authors

• Michael J. Jacobson, The University of Sydney

• Manu Kapur, ETH Zurich

• Peter Reimann, Heidelberg University