Paul Hur's Dissertation Defense
Paul Hur will defend his dissertation, "Metacognitive Effects of Confusion During Learning: Continuous, Real-Time Self-Report Methods and Insights."
His doctoral committee includes Professor Michael Twidale (co-chair, co-director of research, member); Assistant Professor Nigel Bosch (co-chair, co-director of research, member); Associate Professor Chad Lane; and Assistant Professor Cynthia D'Angelo.
Abstract
Confusion is a complex yet common student experience with cognitive, motivational, and emotional attributes that can be either beneficial or detrimental to the learning process, depending on how it is managed by the student. Current research methods for understanding student confusion conventionally administer time-based, interval confusion self-report survey prompts alone, or in conjunction with affect data such as automatically analyzing eye gaze, facial expressions, or behavioral log data. In this dissertation, via an in-depth 14-student ethnographic study and a large-scale 280-student online experiment, I evaluate novel, continuous, real-time confusion self-report methods designed to be used while engaged in computer-based learning tasks. I examine how these methods could address the shortcomings of existing confusion detection methods. Notably, continuous, real-time confusion self-report methods enable the capture of granular (i.e., high data collection frequency) and rich (i.e., detailed changes to confusion intensities) grounded student confusion data, which can provide representations of confusion’s dynamic nature. This dissertation presents findings on the usability of these tools, including the differences between the physical and digital formats, as well as the implications for research methods and supporting student metacognitive processes. Ultimately, this dissertation takes a step towards developing confusion measurement methods which can accurately capture and depict the uniqueness of how individual students experience confusion when encountering difficulties during learning.