Students rarely open transfer-practice cards immediately after solving; they want such problems organized into delayed, spaced wrong-book review instead¶

Claim · All claims
Evidence · 2 studies · 2 design r2–r3 · q2
Currency · no dated study recorded · checked against its sources 2026-10-20 (automated, 2 of 2 entries) · not reviewed by a person

Subclaims¶

q2 i? Of 93 transfer-card exposures, students opened 8 cards (8.6%) and launched 4 practice questions (4.3%), a timing mismatch with homework completion goals. → transfer-funnel (2) q2 i? Students selectively saved problems to history based on difficulty and wanted automated wrong-book organization for spaced weekend retrieval. → transfer-funnel (2)

Evidence¶

transfer-funnel¶

Harry Feng, Yuan Tian, and Erica Zhao. 2026. From Answer Generators to Reasoning Facilitators: Designing AI Tutors for Mathematical Reasoning in High-Stakes Environments. https://arxiv.org/abs/2607.01692 doi:10.48550/arxiv.2607.01692

q2 · i? · design · r2

Telemetry transfer funnel from the field deployment. The article reports "93 transfer-card exposures" with 8 cards opened and 4 practice questions launched, interpreted as a timing mismatch rather than lack of perceived value.

"We observed 93 transfer-card exposures; students opened 8 cards (8.6% of exposures) and launched 4 practice questions (4.3% of exposures)."

transfer-funnel (2)¶

Harry Feng, Yuan Tian, and Erica Zhao. 2026. From Answer Generators to Reasoning Facilitators: Designing AI Tutors for Mathematical Reasoning in High-Stakes Environments. https://arxiv.org/abs/2607.01692

q2 · i? · design · r3

Interview finding from the same field study: participants P09 and P10 wanted saved problems organized into a knowledge-map wrong-book for weekend spaced retrieval, which telemetry on selective saving corroborated.

"Students like KX (P09) and TYK (P10) still wanted AITutor to organize these problems into a knowledge-map "wrong-book" for delayed retrieval practice during weekend reviews."

Discussion¶

Related Claims¶

• Learners Misjudge Spacing Benefits — the general claim this finding bears on, with the studies behind it
• System reliability and latency acted as a prerequisite bottleneck: 56.4% of started solves completed, with a 32.0-second average latency and 68.5-second p90 — related
• Learners misjudge which learning strategies are effective — related
• Timely feedback on homework increases learning, and computer platforms providing immediate feedback help teachers adapt — related
• Retrieval practice improves long-term retention — related
• Follow-up support shows low initiation but reliable completion: the bottleneck is noticing and composing the question, not the AI's ability to answer — related