LLM feedback adaptivity to student performance occurs only in frequently produced focus types and only in large or closed models¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims¶

q2 i? The teacher's feedback differs significantly between high- and low-performing students in Task Constraints and Praise, while models that show performance-based adaptivity do so only in the focus type they produce most frequently, and only large or closed models show any such adaptivity. → Norah Almousa 2026

Evidence¶

Norah Almousa 2026¶

Norah Almousa, Shayan Peyghambari Oskoui, Raquel Coelho, Gayle Rogers, Xiang Lorraine Li, Diane Litman. (2026). Evaluating Feedback Focus and Pedagogical Adaptivity in LLM-Generated Feedback on Student Writing. arXiv preprint. https://arxiv.org/abs/2609.28026 doi:10.48550/arxiv.2609.28026

q2 · i? · design · r2

Fisher's Exact Test comparison of feedback focus type distributions between high- and low-performing students (median grade split) in the composition course under zero-shot prompting. The study reports that "all models that show adaptivity based on student performance are large or closed models."

"the teacher's feedback shows significant differences between high- and low-performing students in Task Constraints and Praise. Gemini and Mistral-24B show significant differences in Clarification, LearnLM in Mistakes, and Qwen-70B in Elaboration, while Llama-7B and Mistral-7B show no significant differences across any feedback focus type."

Discussion¶

Related Claims¶

• Teachers adapt feedback focus across draft stages in five types while the best model shows significant changes in only three — related
• LLMs cover most teacher feedback focus types but fail to match teacher feedback focus distributions, with LearnLM closest — related