AI as a Research Partner: Enhancing PER Literature Analysis

Understanding the evolution of a field of study is important for engaging with that field, yet tracking trends in Physics Education Research (PER) is increasingly challenging due to the expanding amount of relevant literature. Traditional literature reviews provide valuable insight but are slow and labor-intensive, limiting researchers’ ability to track how concepts develop over time. To explore this challenge, we develop an automated approach for large-scale literature analysis using modern language models. By applying natural language processing techniques like sentiment analysis, we investigate the potential of machine learning to extract insights from PER publications. We focus on publications from The Physics Teacher, PER Conference Proceedings, American Journal of Physics, and Physical Review PER. While not a replacement for human analysis, this approach may help identify patterns in how ideas emerge, spread, or fade within PER literature.