关于「防御退出的临界点:扩展平行过程模型的精确三分法」,《恐惧文案的倒U与边界》给出的定量结论是什么?
结论:第二章把扩展平行过程模型写成精确的三分法,证明防御退出的临界点等于受众效能的中位数,混合受众的临界剂量为 1.980。
数据口径:书中参数均为示意;全部计算由书中程序生成;本书不是法律意见,也不构成个人化的商业或理财建议,对任何产品的功效不作判断。 第一章用「先接收、再顺从」的两步模型说明倒 U 型的顶点由注意与防御的速率共同决定,示意参数下顶点在剂量 3.901,并指出牙齿卫生实验的三个数据点恰在一条直线上。
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: 1.980
| dose f | threat T | ignore | adopt | defend | defend/adopt |
|---|---|---|---|---|---|
| 0.0 | 0.0000 | 1.0000 | 0.0000 | 0.0000 | 0.000 |
| 0.5 | 0.1605 | 0.8395 | 0.1554 | 0.0051 | 0.033 |
| 1.0 | 0.2953 | 0.7047 | 0.2490 | 0.0463 | 0.186 |
| 1.5 | 0.4084 | 0.5916 | 0.2728 | 0.1357 | 0.497 |
| 2.0 | 0.5034 | 0.4966 | 0.2485 | 0.2549 | 1.026 |
| 3.0 | 0.6501 | 0.3499 | 0.1528 | 0.4973 | 3.254 |
| 4.0 | 0.7534 | 0.2466 | 0.0753 | 0.6781 | 9.004 |
| 5.0 | 0.8262 | 0.1738 | 0.0328 | 0.7934 | 24.159 |
| 6.0 | 0.8775 | 0.1225 | 0.0133 | 0.8642 | 64.984 |
| 8.0 | 0.9392 | 0.0608 | 0.0019 | 0.9373 | 487.150 |
| 10.0 | 0.9698 | 0.0302 | 0.0003 | 0.9695 | 3800.671 |
| adopt peaks at dose 1.451 with share 0.2730 | |||||
| defend overtakes adopt where T = median E = 0.5, dose 1.980 | |||||
| at dose 10 defenders per adopter: 3800.7 |
核心公式
\[f_c = -\frac{\ln(1 - \text{中位数})}{0.35}.\]
查看完整验算程序
# Chapter 2. p21: the extended parallel process split, exactly.
# Witte (1992): a reader who sees a threat appraises it; if the threat
# matters and the reader believes the advice works and is within reach
# (efficacy), she controls the danger; if threat outruns efficacy, she
# controls the fear instead (denial, avoidance, leaving the page).
# Model: perceived threat T = 1 - exp(-k f), k = 0.35 (illustrative).
# A share T of readers attends; an attender adopts when her efficacy E
# is at least T, otherwise she defends. E ~ Beta(3, 3) (illustrative),
# whose survival function is a binomial tail, so every share is exact.
import math
K = 0.35
def surv(x, a=3, b=3):
"""P(E > x) for E ~ Beta(a, b) with integer a, b."""
if x <= 0:
return 1.0
if x >= 1:
return 0.0
n = a + b - 1
return math.fsum(math.comb(n, j) * x ** j * (1 - x) ** (n - j)
for j in range(a))
def split(f):
t = 1 - math.exp(-K * f)
s = surv(t)
return t, 1 - t, t * s, t * (1 - s)
def main():
print(" dose f threat T ignore adopt defend defend/adopt")
for f in (0, 0.5, 1, 1.5, 2, 3, 4, 5, 6, 8, 10):
t, ign, ad, de = split(f)
ratio = de / ad if ad > 0 else 0.0
print("{:>7.1f} {:>8.4f} {:>6.4f} {:>6.4f} {:>6.4f} {:>12.3f}"
.format(f, t, ign, ad, de, ratio))
best = max((split(g / 1000)[2], g / 1000) for g in range(0, 10001))
print("adopt peaks at dose {:.3f} with share {:.4f}".format(
best[1], best[0]))
fc = -math.log(1 - 0.5) / K
print("defend overtakes adopt where T = median E = 0.5,"
" dose {:.3f}".format(fc))
print("at dose 10 defenders per adopter: {:.1f}".format(
split(10)[3] / split(10)[2]))
if __name__ == "__main__":
main()唯一数据源与引用
范雅福(Fan, Y.). (2026). 《恐惧唤醒倒U型曲线、具体化痛点成像与合规免责高转化文案边界》第 2 章「防御退出的临界点:扩展平行过程模型的精确三分法」(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-143 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026143,
author = {Fan, Yafu},
title = {The Inverted-U of Fear Arousal, Concrete Pain-Point Imagery and the Compliance Boundary of High-Conversion Copy},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}