Teaching assistants (TAs) in the United States play a prominent role in educating undergraduates. Their influence can make the difference between students continuing in their majors or leaving them. However, most TAs use teacher-centered, transmission models of teaching, i.e., lecturing to disengaged students. Part of the reason for this is that most TAs receive little training on how to teach, and almost no grounded feedback about their teaching behaviors.
In this thesis I describe my work investigating the use of technology to increase feedback and training for TAs. My focus is understanding how their knowledge, skills, and attitudes should drive the design of algorithms for gathering classroom behavioral data and delivering computer-mediated feedback and consultation. My work evaluates a novel framework for investigating how TAs interact with their data, reflect on what it means, and decide what (if anything) to change in their teaching. I examine how initial beliefs can impact their system interactions, how those beliefs change over time, and the resulting implications for designing data-driven training artifacts.
Amy Ogan (HCII) (Co-chair)
John Zimmerman (HCII) (Co-chair)
Ken Koedinger (HCII/Psychology)
Marsha Lovett (Psychology/Eberly Center)