关于「实战工具与检查器规格:三秒信任台账、合成规则核算器与预注册检验」,《三秒薄切片信任》给出的定量结论是什么?

结论:第八章给出蒸馏小工具「三通道一致性信任合成与短板核算器」、按四件套分组的检查器规格、三秒信任台账与预注册检验:主效应 0.03 每格需要 1,047 人,三因子共 8,376 人,中途偷看会把误报率从 0.0540 推到 0.1940。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;本书只写如何让真实、可兑现的信号在前三秒清晰一致,不写任何伪造资质或假权威的手法,不宣称任何气味的健康功效,是非个人化经营建议。 第一章区分信度与效度,示意参数下三秒判断只达到效度上限的 0.612,与长期评价的一致率为 0.579。

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: 0.03, 1047, 8376

A. stay rate past 3 s, baseline 0.40
effectdeltan per celltotal (8 cells)
main0.031,0478,376
main0.053773,016
two-way DiD0.051,50712,056
two-way DiD0.085894,712
three-way0.082,35518,840
three-way0.121,0478,376
B. trust rating, sd 1
main0.2099792
two-way DiD0.203933,144
three-way0.306985,584
C. same delta: an interaction needs 4.0 times the viewers
查看完整验算程序
# Chapter 8. p82: sample size for a 2 x 2 x 2 test of the three channels.
# Outcome: whether a viewer stays past the first 3 s, baseline p = 0.40.
# With n viewers per cell (8 cells), the variance of an estimated effect
# is V p (1 - p) / n where V = 1/2 for a main effect (4 cells against 4),
# 2 for a two-way interaction written as a difference of differences and 8
# for a three-way one. Two-sided alpha 0.05, power 0.80:
# n = (1.959964 + 0.841621)^2 V p (1 - p) / delta^2. Part B does the same
# for a trust rating with sd 1. Effect sizes are illustrative.
import math

P0 = 0.40
ZZ = (1.959964 + 0.841621) ** 2


def n_cell(v, delta, var):
    return math.ceil(ZZ * v * var / (delta * delta))


if __name__ == "__main__":
    pq = P0 * (1.0 - P0)
    print("A. stay rate past 3 s, baseline {:.2f}".format(P0))
    print("   effect               delta   n per cell   total (8 cells)")
    for name, v, d in (("main", 0.5, 0.03), ("main", 0.5, 0.05),
                       ("two-way DiD", 2.0, 0.05), ("two-way DiD", 2.0, 0.08),
                       ("three-way", 8.0, 0.08), ("three-way", 8.0, 0.12)):
        n = n_cell(v, d, pq)
        print("   {:<18}   {:.2f}    {:>10,}   {:>12,}".format(
            name, d, n, 8 * n))
    print("B. trust rating, sd 1")
    for name, v, d in (("main", 0.5, 0.2), ("two-way DiD", 2.0, 0.2),
                       ("three-way", 8.0, 0.3)):
        n = n_cell(v, d, 1.0)
        print("   {:<18}   {:.2f}    {:>10,}   {:>12,}".format(
            name, d, n, 8 * n))
    ratio = n_cell(2.0, 0.05, pq) / n_cell(0.5, 0.05, pq)
    print("C. same delta: an interaction needs {:.1f} times the viewers"
          .format(ratio))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《三秒薄切片信任捕获:视听嗅三通道同频一致性与瞬间权威建立母典》第 8 章「实战工具与检查器规格:三秒信任台账、合成规则核算器与预注册检验」(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-419 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026419,
  author    = {Fan, Yafu},
  title     = {Three-Second Thin-Slicing Trust: Visual-Auditory-Olfactory Congruence and Instant Authority Signalling},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论