Multiple Ways Qualitative Techniques Inform Unsupervised Computational Analysis of Video Data Applying Computational Grounded Theory

• In Event: New Uses and Settings for Mixed- and Multimethod Research

Abstract

Building from emerging work advocating for strategic combinations of qualitative and computational analytic techniques (e.g., Nelson, 2020), this paper presents four roles that qualitative analysis can play in supporting the development of computational tools for analyzing video data. Typically, mixed methods models combine existing analytic tools from varying methodological paradigms to more fully understand the phenomenon in question (e.g., Greene, 2007). In contrast, we utilize qualitative tools to inform the development of computational analytic tools. This approach aligns with methodological advances that use learning analytics meaningfully in educational settings (e.g., Hur et al., 2020; Paquette et al., 2018); we argue that it is especially valuable for developing computational analytic tools for complex, multi-featured data, including classroom video.

Authors

• Christina Krist, University of Illinois at Urbana-Champaign

• Erika David Parr, Rhodes College

• Elizabeth B. Dyer, The University of Tennessee - Knoxville