At this year’s Artificial Intelligence in Measurement and Education Conference (AIME-Con), researchers from SC TEACHER presented on the potential for large language models (LLMs) to offer scalable support in analyzing open-ended responses from the SC Teacher Working Conditions Survey. Researchers unpacked a sample study and proposed methodology, exploring how potential outcomes could inform a better understanding of the factors most significantly impacting teacher retention across our state.

What it is:

• LLMs are generative artificial intelligence (GenAI) systems that have the potential to efficiently capture the qualitative richness of teachers’ responses to open-ended questions.
• The SC Teacher Working Conditions Survey, administered biennially by SC TEACHER, measures the experiences of classroom teachers across South Carolina, measuring job satisfaction and intent to stay in the profession as outcomes.

Why it matters:

By utilizing LLMs and applying prompt engineering and rigorous validation against human benchmarks, SC TEACHER researchers seek to transform free-text responses from the SC Teacher Working Conditions Survey into reliable indicators of job resources and demands. From there, these findings can be combined with responses to closed-ended survey items to offer more nuanced, context-specific explanations of teacher working conditions. This new level of nuance will allow school districts to enhance their strategic decision-making and develop more tailored policies aimed at improving teacher retention.

 

About SC TEACHER

Housed in the USC College of Education, SC TEACHER is a research consortium focused on informing statewide education policy and practice through a growing data infrastructure and SC-centric educator surveys. The consortium releases new research each month, shared via a monthly newsletter.

To schedule an interview with P. Ann Byrd, Ph.D., Executive Director of SC TEACHER, or Angela Starrett, Ph.D., Director of Research, contact SCTcomms@mailbox.sc.edu.