Identifying Emergence in Collaborative Group Discourse

• In Event: Complex Systems Approaches to Studying Motivation and Learning in Social Contexts

Abstract

Objective
This work uses a complex systems approach to gain new insights into the interaction dynamics in collaborative problem-solving groups. We develop an innovative approach that expands the ability of researchers to explore qualitative changes and emergent behavior (defined below) in the collaborative dynamics of groups assigned a problem-solving project.

Theoretical frameworks
This work combines two distinct theoretical frameworks that guide how we conceptualize, and model collaborative dynamics. First, macrocognition characterizes the kinds of talk moves that collaborators use to externalize their cognition, create shared understanding, and coordinate actions to progress a problem-solving task (Fiore, et al., 2010). In particular, the approach of Fiore et al. (2010), is used to code the utterances of elementary students participating in an open-ended engineering design challenge. Second, complex systems theory addresses the relational and learning processes among interconnected individuals and environments that mutually constitute larger learning systems (e.g., Cochran-Smith et al., 2014). Complex systems can have emergent behaviors, behaviors that observable qualitatively but are not able to be logically derived from other observables in the system (Boschetti & Gray, 2013). Emergent behavior is often observed as behavior that is qualitatively different at different points in time. Entropy (Wiltshire et al., 2018) can be used as a marker of these qualitative changes. Although prior work in collaborative dynamics has highlighted the importance of temporal change (Kapur, Volkis, & Kinzer, 2011) and used entropy methods to explore individual changes in interactional patterns (Wiltshire, Butner, & Fiore., 2018), prior work has stopped short of demonstrating the complex nature of such dynamics, failing to include important factors such as inter-level (e.g., individual and group) interactions and failing to explore emergence.

Methods
We use entropy methods along with coding and quantizing (Sandelowski, Volis, & Knafl, 2009) of both individual and group levels of behavior to explore emergence in collaborative groups. In particular, the feedback across levels is used to help identify qualitative transitions in behavior.

Data sources
Transcripts of classroom discourse of small groups of fifth-graders working on an open-ended engineering design project served as the data. These data were coded and quantized in two independent ways. At the individual level, the coding scheme of Fiore, et al. (2010) was applied to all utterances. At the group level, transcripts were segmented into episodes, representing one of five phases of the engineering design cycle (Paul & Beitz, 1996).

Results
Results indicate that qualitative changes in group behavior can be identified from the series of metacognitive codes, and further, that our application of entropy methods can be used to explore the inter-level dynamics of emergent behavior. Surprisingly, we identify not only changes in what might be considered linked dynamics (e.g., individual-level metacognition and group engineering-design-cycle phase), but also changes that occurred along other axes.

Scholarly significance
This work extends prior work on macrocognition by developing a method that allows for work across levels of analysis (e.g., individual and group), highlighting the importance of emergent behavior in collaborative dynamics, and providing a new method for the investigation of collaborative groups.

Authors

• Bernard P. Ricca, University of Colorado - Colorado Springs

• Michelle Jordan, Arizona State University