AI-assisted feedback drafts do not reduce TAs' time spent per character of feedback¶

Claim · All claims
Evidence · 1 study · causal r2 · q3
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? No statistically significant effect of treatment on time per character of feedback (coeff. = 0.29 s/char, SE = 0.35, p = 0.41); equivalence was not tested. → Mahinpei 2026

Evidence¶

Mahinpei 2026¶

Mahinpei, R., Dean, V., Fong, R., Liu, L. T., & Ribeiro, M. H. (2026). AI Assistance for Discretionary Work: Increasing Feedback Provision in Higher Education. arXiv. https://arxiv.org/abs/2606.03095 doi:10.48550/arxiv.2606.03095

q3 · i? · causal · r2

Randomized field experiment, analysis restricted to submissions that received feedback (N = 664). Within-question regression found "no statistically significant effect" on time per character (p = 0.41); the article interprets this as AI lowering initiation barriers without eliminating review and adaptation effort.

"We find no statistically significant effect of treatment assignment on time spent per character of feedback (coeff. = 0.29 s/char, SE = 0.35, 95% CI [-0.39, 0.97],𝑝= 0.41)."

Discussion¶

Related Claims¶

• AI-assisted feedback does not reduce student-perceived usefulness of rated feedback — related
• AI-assisted feedback drafts increase discretionary feedback provision by TAs by 10.81 percentage points — related
• AI-assisted feedback drafts increase total feedback length by 39.79 characters per submission — related
• Draft usefulness varies by grading context: higher for submissions with mistakes and applied questions — related
• TAs adapt rather than copy AI feedback drafts, with moderate draft–final overlap (ROUGE-L F1 = 0.60) — related
• In the Burgess test about one quarter of total time is spent drawing, and paragraphs unequal in completion time reduce the test's validity — related