关于「实证闭环:同心圆配置器的实战工具与检查器规格」,《同心圆志业与亲密配置》给出的定量结论是什么?

结论:第八章给出周台账、红黄绿规则、配置器小工具与每组 36 人的预注册检验,差值置信区间上界低于 0.05 即推翻核心命题。书中以标注框区分已发表的证据、可运行的 Python 计算(22 段程序)、原典与作者的假说。

数据口径:书中参数为示意;全部计算由书中程序生成;本书不替代心理咨询、婚恋咨询或家庭辅导,也不是个人化建议。 第二章用社会脑层级与衰减动力学算出核心底座定额,示意参数下每周 14.38 小时。

English version

对照数据表

下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.05, 36

independent weeks per arm121
design effect for 12 weeks per person, ICC 0.203.20
person-weeks per arm388; people per arm: 33
simulated power at 33 people per arm0.769
people per arm for simulated power >0.80: 36 (power 0.828)
refuted if the upper 95% bound of the reduction is under 0.05
查看完整验算程序
# Chapter 8. p82: pre-registered test of the floor quota
# Unit: a person-week. Outcome: a breach week (any core tie rated
# under 5). Arms: weeks where the core floor was locked first versus
# weeks where it was not. Planning values (illustrative): breach rate
# 0.30 without the floor, 0.15 with it; two-sided alpha 0.05, power
# 0.80. Repeated weeks inflate variance by 1 + (m - 1) ICC with
# m = 12 weeks per person and ICC 0.20. A seeded simulation checks the
# power of the resulting design with a logistic random intercept.
import math
import random
from statistics import NormalDist

P1, P2, ALPHA, POWER = 0.30, 0.15, 0.05, 0.80
M, ICC = 12, 0.20
Z = NormalDist()
za, zb = Z.inv_cdf(1 - ALPHA / 2), Z.inv_cdf(POWER)
pbar = (P1 + P2) / 2
num = (za * math.sqrt(2 * pbar * (1 - pbar))
       + zb * math.sqrt(P1 * (1 - P1) + P2 * (1 - P2))) ** 2
n = math.ceil(num / (P1 - P2) ** 2)
deff = 1 + (M - 1) * ICC
weeks = math.ceil(n * deff)
people = math.ceil(weeks / M)
print("independent weeks per arm: %d" % n)
print("design effect for 12 weeks per person, ICC 0.20: %.2f" % deff)
print("person-weeks per arm: %d; people per arm: %d" % (weeks, people))

rng = random.Random(42582)
SD = math.sqrt(ICC / (1 - ICC) * math.pi ** 2 / 3)  # latent-scale ICC


def power(npeople, sims):
    hits = 0
    for _ in range(sims):
        arms = []
        for p in (P1, P2):
            base = math.log(p / (1 - p))
            k = 0
            for _ in range(npeople):
                q = 1 / (1 + math.exp(-(base + rng.gauss(0.0, SD))))
                for _ in range(M):
                    if rng.random() < q:
                        k += 1
            arms.append(k / (npeople * M))
        pooled = (arms[0] + arms[1]) / 2
        se = math.sqrt(2 * pooled * (1 - pooled) / (npeople * M) * deff)
        if (arms[0] - arms[1]) / se > za:
            hits += 1
    return hits / sims


print("simulated power at %d people per arm: %.3f"
      % (people, power(people, 2000)))
n2 = people
while True:
    n2 += 3
    pw = power(n2, 2000)
    if pw >= POWER:
        break
print("people per arm for simulated power >= 0.80: %d (power %.3f)"
      % (n2, pw))
print("refuted if the upper 95%% bound of the reduction is under %.2f"
      % 0.05)

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《普世志业大爱与私密关系小爱的相容边界:同心圆预算分配与非零和伦理母典》第 8 章「实证闭环:同心圆配置器的实战工具与检查器规格」(版本 1.0.0). 5D Library. DOI 待发布

DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-429 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026429,
  author    = {Fan, Yafu},
  title     = {Reconciling Universal Mission and Intimate Attachment: Concentric Time-Energy Allocation, Role-Conflict Damping and Non-Zero-Sum Ethics},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论