关于「选项数与购买概率:倒 U 型的不确定区间」,《客群分化界面与同价同条款》给出的定量结论是什么?
结论:第五章说明选择过载是有条件的效应:示意合并均值接近零,新情境的预测区间却从 -0.5195 到 +0.6687;筛选加买家自选排序让比较楼盘快 5.69 倍。
数据口径:书中参数为示意,全部计算由书中程序生成,不对应任何真实站点;非法律建议,非个人化经营建议。 第五章说明选择过载是有条件的效应:示意合并均值接近零,新情境的预测区间却从 -0.5195 到 +0.6687;筛选加买家自选排序让比较楼盘快 5.69 倍。
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: -0.5195, 0.6687
| reported | 63 conditions, 50 experiments, N = 5,036 |
| reported | 99 observations, N = 7,202 |
| illustrative studies k | 12 |
| fixed-effect mean +0.0536 | |
| Q | 45.39 on 11 df; I^2 = 0.758; tau^2 = 0.0654 |
| random-effects mean +0.0746 | 95% CI -0.0946 to +0.2438 |
| 95% prediction interval for a new setting -0.5195 to +0.6687 |
查看完整验算程序
# Chapter 5. p51: why a near-zero pooled effect can coexist with large
# effects in single studies. DerSimonian-Laird random-effects pooling of
# 12 ILLUSTRATIVE standardized effects (positive = more options lowered
# buying or satisfaction). They are not the data of any meta-analysis;
# they only mimic the published pattern: a mean near zero and a large
# spread between studies (Scheibehenne et al. 2010 report 63 conditions
# from 50 experiments, N = 5,036; Chernev et al. 2015 report 99
# observations, N = 7,202, and four moderators).
import math
D = [0.77, 0.45, 0.30, 0.12, 0.05, -0.02, -0.08, -0.15, -0.20, -0.31,
0.60, -0.45]
V = [0.040, 0.030, 0.020, 0.010, 0.015, 0.012, 0.020, 0.018, 0.025,
0.030, 0.050, 0.035]
w = [1.0 / v for v in V]
fixed = math.fsum(wi * di for wi, di in zip(w, D)) / math.fsum(w)
q = math.fsum(wi * (di - fixed) ** 2 for wi, di in zip(w, D))
df = len(D) - 1
c = math.fsum(w) - math.fsum(wi * wi for wi in w) / math.fsum(w)
tau2 = max(0.0, (q - df) / c)
ws = [1.0 / (v + tau2) for v in V]
mu = math.fsum(wi * di for wi, di in zip(ws, D)) / math.fsum(ws)
se = math.sqrt(1.0 / math.fsum(ws))
i2 = max(0.0, (q - df) / q)
print("reported: 63 conditions, 50 experiments, N = {:,}".format(5036))
print("reported: 99 observations, N = {:,}".format(7202))
print("illustrative studies k = {}".format(len(D)))
print("fixed-effect mean {:+.4f}".format(fixed))
print("Q = {:.2f} on {} df; I^2 = {:.3f}; tau^2 = {:.4f}".format(
q, df, i2, tau2))
print("random-effects mean {:+.4f} 95% CI {:+.4f} to {:+.4f}".format(
mu, mu - 1.96 * se, mu + 1.96 * se))
pi = 2.201 * math.sqrt(tau2 + se * se) # t(0.975, 10) = 2.201
print("95% prediction interval for a new setting {:+.4f} to {:+.4f}".format(
mu - pi, mu + pi))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《客群认知画像分层、适老与新世代界面分化及千人千面转化交互母典》第 5 章「选项数与购买概率:倒 U 型的不确定区间」(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-363 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026363,
author = {Fan, Yafu},
title = {Demographic Cognitive Segmentation, Elder-to-Gen-Z Interface Divergence and Audience-Adaptive Conversion UX},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}