Use impasse depth and type as real-time learning-analytics signals to trigger graduated escalation of tutor support¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The article recommends that learning analytics systems track impasse depth and type in tutoring dialogues in real time to flag students who remain stuck and support graduated escalation. As the Notes for Practice state, analytics can "track impasse depth and type in tutoring dialogues in real time to flag students who remain stuck", moving "from questioning to targeted error feedback and, when needed, worked sub-steps". Impasse depth and type are observable, turn-level dialogue signals derived from codebook-annotated logs.

Design Implications¶

Context¶

Requirements¶

• A dialogue annotation pipeline able to code impasse turns, types (conceptual error, expressed uncertainty, help-seeking), and depth from tutor-student logs

Constraints¶

• The authentic dialogue analyses are observational, so associations between tutor moves and recovery are not causal; recovery captures only next-turn correctness, not longer-term learning

Target Learners¶

• undergraduate chemistry students in introductory general chemistry

Target Learning Goals¶

• early identification of stalled tutoring interactions
• timely escalation of assistance before disengagement

Affordances¶

• Graduated State Sensitive Assistance Llm Tutors

Related Strategies¶

• Hybrid Socratic-to-sub-question tutoring with EEG-triggered fallback and progressively concrete hints

Examples¶

-

Key Sources¶

• Ahtisham, B., Vanacore, K., Napoli, A., Arens, J., Ionova, K., Cohn, C., Salehi, S., & Kizilcec, R. (2026). Examining Variation in How Guided AI Tutors Resolve Student Impasses. arXiv. https://arxiv.org/abs/2609.38346. doi:10.48550/arxiv.2609.38346