Heidi Reichert PhD Dissertation Defense

[Tiffany Barnes]

Today, June 9, 10 am - 12 pm

Heidi Reichert PhD Dissertation Defense

zoom: https://ncsu.zoom.us/j/91631546840?pwd=Ymgm5uio20tCbLaJNaFn4PhTa7bOpy.1

Title: Understanding Generative AI and Large Language Models for Teaching and Learning in Secondary Classrooms

Abstract:
Large Language Models (LLMs) and Generative AI (GenAI) have markedly changed the educational landscape, creating both significant opportunities and critical challenges for secondary classrooms (grades 6-12). Despite growing interest in educational AI, most research has focused on higher education, and few studies have centered the perspectives of secondary teachers, who are often most directly responsible for classroom implementation. This dissertation addresses this gap through an educator-centered approach, engaging secondary teachers as learners, designers, and implementers of LLM-based tools across three complementary studies.

First, we developed and evaluated a five-session professional development workshop series introducing secondary teachers to GenAI. Through pre- and post-focus groups and long-term follow-up interviews, we found that structured, active-learning-based PD significantly improved teachers' knowledge and attitudes toward GenAI, with sustained positive impacts on their teaching practices. Second, we engaged a cohort of secondary teachers in participatory design of LLM-powered classroom chatbots. Our analysis revealed that teachers conceptualize AI systems as ``bounded experts'' operating within authority boundaries that preserve teacher professional responsibility and expertise boundaries that recognize teachers' irreplaceable contextual knowledge of their students and institutions. This bounded expertise framework offers design principles for human-centered AI deployment extending beyond education to other high-stakes professional domains. Third, we investigated how secondary computing teachers currently integrate LLM tools in their classrooms through interviews and surveys. We developed and empirically grounded a six-phase implementation framework (Prepare, Adapt Curriculum, Lead, Assess, Synthesize, and Share), revealing that preparation-phase barriers, including administrative restrictions and resource scarcity, represent the primary obstacles to successful integration.

Together, these three studies contribute a theoretically grounded and practically oriented framework for supporting secondary teachers in understanding, shaping, and implementing GenAI in their classrooms. This work offers implications for professional development design, human-centered AI tool development, and institutional policy, while laying the groundwork for future research developing and evaluating educator-informed LLM tools for secondary contexts.

Committee: 

Tiffany Barnes, Chair

Lina Battestilli

Dongkuan Xu

Shiyan Jiang

Tiffany Barnes

Distinguished Professor of Computer Science

NC State University

Tiffany...@gmail.com 

http://eliza.ncsu.edu/