关于「实证闭环:工作室换气排程、会客室成交率与证伪方案」,《室内二氧化碳与穿堂风》给出的定量结论是什么?

结论:第八章给出工作室、直播间与会客室的实证闭环、蒸馏小工具“换气排程器”(示范直播间开窗时间占 0.243)和 32 人的预注册配对检验,中等效应下把握度为 0.801;成交率检验需要 2754 场会谈,故只作监测。书中以标注框区分已发表的证据、可运行的 Python 计算(21 段程序)、原典与作者的假说。

数据口径:书中的房间、人数、呼出量与风压系数均为示意参数;全部计算由书中程序生成;本书不是医疗建议,不对任何健康结果作出承诺,对经营者的建议为非个人化的经营建议。 书中以标注框区分已发表的证据、可运行的 Python 计算(21 段程序)、原典与作者的假说。

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: 0.801, 32

test 1, n32 paired share significant sig. lower
no CO2 effect0.058 0.034
small, dz 0.30.427 0.001
medium, dz 0.50.801 0.000
Allen-like, dz 1.01.000 0.000
test 2, 120 meetings/arm share significant
no lift0.054
lift 0.050.123
lift 0.1760.802
查看完整验算程序
# Chapter 8. p83: how often the pre-registered tests reach each verdict.
# Test 1: 32 people, each does the same decision task once at about
# 600 ppm and once at 1400 ppm (pure CO2 dosed, order balanced); verdict
# by a two-sided paired test on the mean difference (normal approx).
# Test 2: 120 client meetings per arm, closing rate 0.30 at baseline;
# two-proportion z test. 4000 simulated studies per row, fixed seed.
# True effects are scenarios, not claims.
# 预注册检验的蒙特卡洛:各情景下“支持 / 推翻”的频率;固定种子。
import math
import random

rng = random.Random(20261112)
REPS = 4000


def paired_reject(n, dz):
    d = [rng.gauss(dz, 1.0) for _ in range(n)]
    m = math.fsum(d) / n
    sd = math.sqrt(math.fsum((x - m) ** 2 for x in d) / (n - 1))
    return abs(m / (sd / math.sqrt(n))) > 1.96, m < 0


def prop_reject(n, p1, p2):
    a = len([1 for _ in range(n) if rng.random() < p1])
    b = len([1 for _ in range(n) if rng.random() < p2])
    pb = (a + b) / (2.0 * n)
    se = math.sqrt(2 * pb * (1 - pb) / n)
    return se > 0 and abs(b - a) / n / se > 1.96


if __name__ == "__main__":
    print("test 1, n = 32 paired     share significant  sig. lower")
    for label, dz in (("no CO2 effect", 0.0), ("small, dz 0.3", 0.3),
                      ("medium, dz 0.5", 0.5), ("Allen-like, dz 1.0", 1.0)):
        sig = neg = 0
        for _ in range(REPS):
            s, wrong = paired_reject(32, dz)
            sig += s
            neg += s and wrong
        print("{:22s} {:19.3f} {:11.3f}".format(label, sig / REPS,
                                                neg / REPS))
    print("test 2, 120 meetings/arm  share significant")
    for label, lift in (("no lift", 0.0), ("lift 0.05", 0.05),
                        ("lift 0.176", 0.176)):
        hits = 0
        for _ in range(REPS):
            hits += prop_reject(120, 0.30, 0.30 + lift)
        print("{:22s} {:19.3f}".format(label, hits / REPS))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《室内二氧化碳认知衰减、穿堂风伯努利换气与藏风聚气的通风工程重读母典》第 8 章「实证闭环:工作室换气排程、会客室成交率与证伪方案」(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-194 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026194,
  author    = {Fan, Yafu},
  title     = {Indoor CO2 and Cognitive Decline, Bernoulli Cross-Ventilation and an Engineering Re-reading of 'Storing Wind, Gathering Qi'},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论