Two Examples of Epistemic Network Analysis as a Method for Modeling Facets of Moral Development

• In Event: 1-146 - Employing Quantitative Ethnography and Epistemic Network Analysis to Model Emotions and Discourse in Moral Development

Integrative Statement

This paper explains the theory behind Epistemic Network Analysis (ENA) and two studies that demonstrate ENA’s ability to model networks of constructs related to different facets of moral development. ENA was initially developed to measure connections between elements of epistemic frames (Shaffer, 2009) but has subsequently been used to model phenomena in different domains such as connection-making in collaborative discourse (Siebert-Evenstone et al., 2016; Suthers & Desiato, 2012). In these two examples we used ENA to investigate 1) transactive discourse patterns of middle school students discussing historical and present day dilemmas and 2) constellations of emotions related to witnessing peer aggression.
A theoretical foundation of ENA is that connections between elements of discourse and subjective experience are important, not simply the presence or absence of a single trait or construct that can be coded and counted in isolation. As a result, ENA models connections between constructs (Shaffer, 2017). In our first example we examined connections between statements that were on, or off-topic along with statements with various levels of transactivity; Elicitation, Externalization, Representational, and Operational (Berkowitz & Gibbs, 1983). We considered the following hypothesis: Small groups of students will exhibit different transactive discourse patterns while engaging in scripted vs unscripted discussions about moral dilemmas. Using ENA we created and compared transactive discourse networks for groups A and B in Rounds 1&2 (scripted) vs Round 3 (unscripted).
We found that groups A and B both exhibited statistically significantly (p < 0.05) different discourse patterns between the scripted and unscripted rounds. Figure 1 provides an example of these differences from group A. We can see in Figure 2 that what is driving that statistically significant difference is that students made more connections between representational and operational speech acts in Round 3. Sequences of engaging and working with the ideas and positions of others are indicative of higher quality discourse, than sequences where students simply state their own positions and ideas.
Our second example used ENA to map out the constellations of emotions adolescents associated with different responses to witnessing peer aggression – consoling the victim, intervening aggressively, attempting a peaceful resolution, or ignoring the situation. Using ENA we showed adolescents reported almost identical patterns of emotions for consoling the victim and peacemaking. In addition we used ENA to statistically discriminate emotional patterns of the two aforementioned responses and attempts to amplify anger and enact third party revenge (all p<0.01).
In both examples researchers used ENA networks themselves to interpret the statistical results along with coordinated qualitative analyses of the underlying data. This presentation will explain the theoretical background of ENA, how it was implemented in these two examples and discuss how it can be used to address qualitative and quantitative research goals in moral development research. While grounded in examples, the presentation will also explain the larger Quantiative Ethnographic (QE) methodological and theoretical context of which ENA is a part. ENA, and QE more broadly, offer powerful methods for researching various aspects of moral development including connections associated with moral emotions and discourse.

Author

• Brendan Eagan, University of Wisconsin
  Presenting Author