Low-register pitches are sung more accurately than high-register pitches in Chinese classrooms, reproducing register compression at scale¶

Claim · All claims
Evidence · 1 study · associational r1 · q2 · i0 negligible
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person · recheck by 2031: baselines and conditions change

Subclaims¶

q2 i0 Low-register pitches were sung correctly with probability 0.749 versus 0.703 for high-register pitches, a 4.6-percentage-point gap the article reports as a small standardized effect (Cohen's h ≈ 0.10), stable across robustness checks. → Wei 2026

Evidence¶

Wei 2026¶

Wei, Wang and Dong. (2026). Cognitive and skill acquisition trajectories in school-based music education: evidence from Chinese classrooms. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2026.1905847

q2 · i0 · associational · r1

Descriptive register analysis (Figure 4) of the singKT corpus using tertile register bins at Q1 (16) and Q3 (47). The article reports the register gap as a small standardized effect, "Cohen's h ≈ 0.10", and stable when bin boundaries were redrawn at the 10th and 90th percentiles.

"Low-register pitches (Q1, n = 547,584) were sung correctly with probability 0.749, mid-register pitches (Q1–Q3, n = 1,320,007) with probability 0.720, and high-register pitches (Q3, n = 591,234) with probability 0.703."

Discussion¶

Related Claims¶

• Note-level singing correctness rises with practice repetition in a negatively accelerated curve, gaining about 7 percentage points over 30 repetitions — related
• A one-parameter IRT model outperforms global and per-pitch baselines on AUC, BCE and Brier score but not on threshold accuracy for held-out singing attempts — related
• Latent singing ability varies widely across learners, with sequence-level theta estimates spanning −3.79 to 3.36 logits — related
• Per-pitch BKT attains the highest accuracy, lowest BCE and lowest Brier score of the four model families, but lower AUC than IRT — related