1. Question

Do four weeks of the Ergen Yuantong practice (Chapter 7) improve the pitch accuracy of amateur singers more than four weeks of the same sung exercises practised with conventional instructions?

2. Why this comparison

The obvious comparison, practice against no practice, would show only that practising helps. The question here is narrower: whether the method’s instructions about attention (decoupling, turning the hearing inward, starting from the inner note) add anything beyond the singing itself. So both groups sing the same exercises for the same time each day; only the instructions differ.

3. Hypotheses and criterion

4. Participants

• Amateur singers aged 18 to 60 who have sung at least once a week over the past year.
• Pure-tone screening: thresholds no worse than 20 dB HL at 0.5, 1, 2 and 4 kHz in both ears.
• Excluded: professional singers, anyone who has already learned the Ergen Yuantong method, anyone reporting a current hearing or voice disorder, and anyone who has had voice surgery.
• Ethics: before recruitment, the protocol will be submitted to an independent research ethics committee; recruitment will not begin without approval, and the committee’s name and approval number will be added to the OSF registration and to the next version of this book. All participants give written informed consent and may withdraw at any time.

5. Groups and intervention

Participants are randomised by computer into two groups, 1:1, in blocks so that group sizes stay balanced.

The guiding recordings for both groups are made by the same voice, have the same length, and are described to participants in equally positive terms, to weaken the explanation “one group expected more”.

6. Measures

Primary outcome: pitch accuracy, measured once before and once after. Participants hear a reference tone and sing the same note on /a/, 24 target notes in total across their comfortable range, in a fixed random order. The fundamental frequency is extracted from the recordings with Praat; for each note the median of the stable portion (from 300 ms after onset to 300 ms before the end) is taken and converted to an error in cents with the formula in Chapter 2; the outcome is the median of the 24 absolute errors.

Secondary outcomes:

• onset accuracy: the median absolute error of the first 150 ms of each note;
• self-rated singing effort (0–10) after the test;
• inner reference: a pitch chosen at the pre-test, hummed from memory at the post-test, error in cents;
• if equipment is available: surface electromyography of the suprahyoid muscles during the test, and contralateral suppression of otoacoustic emissions. These are exploratory.

Blinding: the person extracting pitch and computing errors works from coded files and does not know which group a participant is in.

7. Sample size

from statistics import NormalDist
import math

def n_per_group(d, alpha=0.05, power=0.80):
    z = NormalDist().inv_cdf
    za, zb = z(1 - alpha / 2), z(power)
    n = 2 * ((za + zb) / d) ** 2 + za ** 2 / 4   # normal approximation + Guenther's correction
    return math.ceil(n)

for d in (0.5, 0.4, 0.3):
    n = n_per_group(d)
    print(f"d = {d}: {n} per group, {math.ceil(n / 0.85)} to recruit (15% drop-out)")
# d = 0.5: 64 per group, 76 to recruit (15% drop-out)
# d = 0.4: 100 per group, 118 to recruit (15% drop-out)
# d = 0.3: 176 per group, 208 to recruit (15% drop-out)

8. Analysis plan

• Primary analysis by intention to treat: everyone randomised is included. Missing post-test values are handled by multiple imputation using the pre-test value and group; a completers-only analysis is reported alongside.
• Adherence (minutes practised) is reported for both groups; a per-protocol analysis of participants who completed at least 80% of sessions is labelled secondary.
• No post-hoc subgroup analyses; any extra analysis is labelled exploratory.
• Data are published as CSV with identifying information removed, together with the analysis code and the Praat settings.

9. What would count as failure

The method fails this test if the practice group’s adjusted pitch error is not at least 5 cents lower than the control group’s with $p < 0.05$. It also counts against the method if the secondary outcomes move in favour of the control group. A failure would not show that the scripture is wrong or that listening does not matter; it would show that these instructions, given in this way for four weeks, do not make amateur singers more accurate than ordinary instructions do. That is a useful thing to know, and it will be published as prominently as a success.

10. Commitment to publish

Whether or not the results support the hypothesis, they will be:

1. published on Zenodo as a new version of this book with a new DOI;
2. released on 5dlibrary.com as raw data and code;
3. stated plainly in this book, including the possible conclusion that the Ergen Yuantong method does not improve pitch accuracy.

Summary

• The protocol is public before any data are collected, and the criterion is fixed in advance.
• Both groups sing the same exercises for the same time; only the instructions differ, so the test isolates the method’s instructions about attention.
• The primary outcome is pitch accuracy measured objectively in cents; onset accuracy, effort and inner-reference error are secondary.
• About 64 per group are needed (76 allowing for drop-out); results will be published whether positive or negative.