Epistemic Association Rule Networks: Incorporating Association Rule Mining into the Quantitative Ethnography Toolbox

Since the formalization of Quantitative Ethnography (QE) as a methodology, Epistemic Network Analysis (ENA) has been the most widely used analytical tool in the community. ENA has proven itself highly useful for QE research, particularly in modeling temporal associations between code pairs. However, integrating additional techniques that systematically reveal more complex patterns in data can not only supplement the insights derived from ENA but also broaden the range of research that can be conducted through a QE lens. To that end, this paper proposes Association Rule Mining (ARM) as an additional technique for QE. We introduce a new visualization of the results of ARM, which we term Epistemic Association Rule Networks (EARN), which combines ENA’s visualization strengths with ARM’s ability to identify more complex patterns. Using human-human tutoring transcripts, we illustrate how ARM and EARN can complement ENA by offering insights on directional conditional relationships between groups and pairs of constructs, offering a more nuanced understanding of complex phenomena.