In a blinded study with 50 learners, TACTutor receives the highest overall mean rating (5.54/7) and is rated above its backbone across all four teaching-behavior dimensions (+0.39 overall)¶

Claim · All claims
Evidence · 1 study (2 entries) · design r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? In a blinded, within-participant Prolific study (n=50, 200 conversations), TACTutor achieved the highest mean score (5.54), within 0.09 points of both proprietary tutors, and was rated above its backbone across all four dimensions (+0.39 overall). → Dongjie Yang 2026 q2 i? The gains over the backbone were concentrated in encouragement (+0.58), guiding (+0.40), and scaffolding (+0.38), with self-correction improving more modestly (+0.20). → Dongjie Yang 2026 (2)

Evidence¶

Dongjie Yang 2026¶

Dongjie Yang, Siyan Lin, Leixian Shen, Rui Sheng, Huamin Qu, and Zixin Chen. (2026). TACT: Taxonomy-Aligned Post-Training for Pedagogically Adaptive English Tutoring. https://arxiv.org/abs/2608.03952 doi:10.48550/arxiv.2608.03952

q2 · i? · design · r2

Blinded, within-participant interactive evaluation with 50 real-world learners via Prolific; each participant held four conversations with anonymous tutors, yielding 200 conversations, with 1-7 ratings on guiding, scaffolding, encouragement, and self-correction support. The article reports the "highest mean score(5.54)"; no significance test is printed.

"TACTutor achieved the highest mean score(5.54), within 0.09 points of both proprietary tutors, and was rated above its backbone across all four dimensions (+0.39overall)."

Dongjie Yang 2026 (2)¶

Dongjie Yang, Siyan Lin, Leixian Shen, Rui Sheng, Huamin Qu, and Zixin Chen. (2026). TACT: Taxonomy-Aligned Post-Training for Pedagogically Adaptive English Tutoring. https://arxiv.org/abs/2608.03952

q2 · i? · design · r2

Dimension-level descriptive rating differences from the same blinded learner study. The article reports gains "concentrated inencouragement(+0.58),guiding(+0.40), and scaffolding(+0.38)"; self-correction improved more modestly (+0.20). No significance test is printed.

"The gains were not uniform, but con- centrated inencouragement(+0.58),guiding(+0.40), and scaffolding(+0.38). This pattern suggests that taxonomy- aligned post-training changes the tutor's interactional pro- file toward more supportive and process-oriented responses, rather than producing only a general improvement in con- versational quality."

Discussion¶

Related Claims¶

• Taxonomy-guided post-training raises TACTutor's TACT_Overall on TACTBench from 0.629 (backbone) to 0.832, the highest among evaluated models — related
• Post-training improves TACTutor's auxiliary guardrail flags on TACTBench: accept rate rises from 0.603 to 0.872 and leak-or-overhelp falls from 0.346 to 0.026 — related
• TACTutor shows no degradation relative to its Qwen backbone on external educational benchmarks, with scores rising on MRBench, TutorBench, LongTutor, and DeepTutor — related
• DeepTutor improves overall first-person interactive tutoring quality by 10.76% over a Naive Tutor baseline on TutorBench — related