---
source_url: https://arxiv.org/abs/2508.14778v1
ingested: 2026-05-07
sha256: 1e643f43af74e3fe1fcdd618c62ea7b367cc7ebf67e5cbdbd5787b361c33cc75
---

Physics education research. Custom Socratic AI chatbot deployed in large-enrollment introductory mechanics course with 150 first-year STEM majors. Full dialogue transcripts logged. Post-interaction surveys: median 4.0/5 for knowledge-based skills, 3.4/5 for overall effectiveness. Question specificity rose from ~10-15% (first turn) to 100% (final turn). Specificity correlated positively with self-reported expected course grade (Pearson r = 0.43). AI-driven Socratic dialogue fosters expert-like reasoning and generates fine-grained analytics for PER, serving as scalable dual-purpose tool for instruction and learning analytics.