Dialogic Analysis: Using Epistemic Network Analysis to Model Dialogic Interactions

• In Event: 34.028 - Dialogic Interactions in Digital Contexts: Multimodal, Multilevel, and Temporal Analyses

Abstract

Objectives or purposes:
1. Describe theoretical and methodological concerns in examining individual contributions to dialogic conversation.
2. Describe a tool that addresses these concerns.


Perspective:
One methodological challenge in the examination of dialogic activity is modeling, assessing, and understanding the role of an individual participant in a dialogic context (Shaffer, 2017). Dialogic processes are not simply the sum of individual actions: as joint activities unfold, information is added to a common ground: a set of shared knowledge and experiences that exist between people when they interact (Clark, 1996). The contents of the common ground influence subsequent actions and how those actions are interpreted (Dillenbourg, 1999). As a result, the interpretation of any single dialogic move by an individual can only be made taking into account the moves made by others. This can be accomplished with traditional qualitative techniques for small amounts of data. However, the growth of technologies for recording dialogic interactions creates a need for automated and/or quantitative modeling that can address this challenge for larger volumes of data.


Methods:
Epistemic Network Analysis (ENA) is a quantitative ethnographic technique for modeling the structure of connections in data. ENA assumes: (1) that it is possible to systematically identify a set of meaningful features in the data (Codes); (2) that the data have local structure (conversations); and (3) that an important feature of the data is the way that Codes are connected to one another within conversations (Shaffer, 2017). ENA models the connections between Codes by quantifying the co-occurrence of Codes within conversations, producing a weighted network that models the relationships between key Codes at the individual and group level. But critically, the nature of these network models makes it possible to understand how students make connections not just by speaking or writing about two ideas together in their own work, but also by responding to ideas, issues, and concerns raised by other students during the dialogic process.


Data sources:
Data are drawn from online conversations about mental health issues among students at two universities in the United States.


Results:
The paper will present a worked example of our analysis, showing not only differences in the framing and discussion of the problem by students at two different universities, but also differences in the characterization of mental health issues based on whether students used personal anecdotes to communicate their ideas as part of the dialogic process. However, the main purpose of the paper will not be to present these analyses per se, but rather to use these analyses as a way to conceptualize individual contributions to, and stances within, a dialogic processes.


Significance:
ENA is currently being used in a number of studies of student discussions to understand the nature of individual contributions to dialogic activity - and also the impact of dialogic activity on individuals (many of these are underway, but see, e.g., Lund et al., 2017; Siebert-Evenstone et al., 2017). Here we provide a worked example with discussion of theoretical and methodological concerns.

Authors

• David W. Shaffer, University of Wisconsin - Madison

• Gideon Dishon, Ben-Gurion University of the Negev

• Brendan R. Eagan, University of Wisconsin - Madison

• Sara Tabatabai, University of Wisconsin - Madison

• Zachary Swiecki, The University of Wisconsin - Madison

• Peter Levine, Tufts University