Innovation
Methods, ideas, or system contributions that make the work stand out.
AI Tutor
Assistance Dilemma
Impasse Trajectories
Prompt Specificity
Learning Analytics
Examining Variation in How Guided AI Tutors Resolve Student Impasses
🤖 AI Summary
This study addresses the challenge that AI tutors often impede student learning during impasses by prematurely providing answers or posing excessive questions. Leveraging large-scale authentic dialogue logs and statistical modeling, this work employs large language models to simulate diverse tutoring strategies, systematically analyzing AI behavioral differences when responding to conceptual errors, uncertainty, and help-seeking. The findings reveal that as impasses persist, direct error correction outperforms repeated questioning, with each additional impasse turn reducing the likelihood of student recovery by 12.7%. Accordingly, this paper proposes a real-time, tiered scaffolding mechanism based on impasse depth and type, empirically validating the necessity of dynamically adjusting tutoring strategies to optimize learning outcomes.
📝 Abstract
When a student is stuck, a tutor faces the assistance dilemma: help given too early can hinder productive struggle, while help withheld too long leaves the student in a frustrating, persistent impasse (i.e., wheel spinning). Generative AI tutors increasingly use guardrails restricting answer-giving, yet little is known about how such tutors behave once an impasse persists. We analyze 20,462 student turns from 1,260 authentic sessions with a guided LLM chemistry tutor, identifying 6,630 impasse turns of three major types: conceptual errors, expressed uncertainty, or help-seeking. We then used these impasses to simulate three tutoring conditions to study variation in AI tutor guidance through impasses: baseline, no-direct-answer, and guided tutor. For a sample of 150 impasses, prompt specificity changed pedagogy: a baseline tutor provided the answer directly in 50.7% of responses, a no-direct-answer tutor asked a follow-up question every time, and the guided tutor responded in a wide variety of ways depending on the context. We then analyzed impasse trajectories in authentic interactions, finding that each additional impasse turn lowered the odds of next-turn recovery by 12.7% (AOR = 0.873, p < .001), and early dropouts were caught in recursive concept elicitation before reaching execution. The benefit of questioning decayed as impasses persisted (scripted question x depth AOR = 0.78; follow-up x depth AOR = 0.83), whereas addressing the student's error grew more beneficial (AOR = 1.14); after a failed scripted question, repeating it was followed by recovery in 28.1% of cases, compared with 39.8% when the tutor addressed the error instead. For learning analytics, these findings identify impasse depth and type as observable, turn-level dialogue signals that analytics can use to trigger graduated, state-sensitive assistance in real time.
🔎 Similar Papers
No similar papers found.
