What does The Fewest Questions conclude, in numbers, about “The Information Ceiling of Correlated Questions: Why the Sixth Question Is Nearly Wasted”?

Answer: Chapter 2 carries the main argument: at a shared-bias correlation of 0.5 the sixth question adds only 0.0476 of the first question's information and five questions already reach 0.8333 of the ceiling; once dropout is counted, the best length is two to five questions. The project blueprint stated the "sixth question" as a fact; the book corrects it to a model consequence that depends on the correlation.

Scope of the numbers: Marked boxes separate published evidence, runnable Python calculations (19 programs), classical sources and the author's hypotheses.

中文版

Data table

The numbers below are printed verbatim by the chapter's Python program and can be reproduced by running it. printed values that match the answer: 0.0476, 0.8333

n rho0.00 rho=0.20 rho=0.35 rho=0.50 rho=0.70
11.0000 1.0000 1.0000 1.0000 1.0000
22.0000 1.6667 1.4815 1.3333 1.1765
33.0000 2.1429 1.7647 1.5000 1.2500
44.0000 2.5000 1.9512 1.6000 1.2903
55.0000 2.7778 2.0833 1.6667 1.3158
66.0000 3.0000 2.1818 1.7143 1.3333
77.0000 3.1818 2.2581 1.7500 1.3462
88.0000 3.3333 2.3188 1.7778 1.3559
99.0000 3.4615 2.3684 1.8000 1.3636
1010.0000 3.5714 2.4096 1.8182 1.3699
ceil none5.0000 2.8571 2.0000 1.4286
gain of item 6 / item 1:
1.00000.2222 0.0985 0.0476 0.0175
first item adding < 0.05 of item 1:
>6017 9 6 4
share of ceiling reached by 5 items:
-0.5556 0.7292 0.8333 0.9211

Key formula

\[I_n = \frac{n}{1 + (n-1)\rho}.\]
Show the full program
# Chapter 2. p21: correlated questions hit an information ceiling.
# Exact formula, all parameters illustrative.
# 模型:Y_i = theta + u + e_i,u 为各题共享的干扰(锚定、作答风格),
# Var(u) = rho,Var(e_i) = 1 - rho,单题总误差方差为 1。
# n 题合计信息 I_n = n / (1 + (n - 1) rho),上限 1 / rho。
# 第 k 题的信息增量 g_k = I_k - I_(k-1),与第一题信息 1 相比。
RHOS = (0.0, 0.2, 0.35, 0.5, 0.7)


def info(n, rho):
    return n / (1.0 + (n - 1) * rho)


def first_small(rho, cut=0.05, kmax=60):
    for k in range(2, kmax + 1):
        if info(k, rho) - info(k - 1, rho) < cut:
            return k
    return None


print("n    " + "".join("rho=%-6.2f" % r for r in RHOS))
for n in range(1, 11):
    print("%-4d " % n + "".join("%-10.4f" % info(n, r) for r in RHOS))
print("ceil " + "".join("%-10s" % ("none" if r == 0 else "%.4f" % (1 / r))
                        for r in RHOS))
print("gain of item 6 / item 1:")
print("     " + "".join("%-10.4f" % (info(6, r) - info(5, r)) for r in RHOS))
print("first item adding < 0.05 of item 1:")
print("     " + "".join("%-10s" % (first_small(r) or ">60") for r in RHOS))
print("share of ceiling reached by 5 items:")
print("     " + "".join("%-10s" % ("-" if r == 0 else
                                   "%.4f" % (info(5, r) * r)) for r in RHOS))

Single source and citation

Fan, Y. (2026). Fisher Information Matrices and the Cramér-Rao Bound: the Fewest Questions Needed to Estimate a Budget or a Birth Time (Version 1.0.0), Chapter 2: The Information Ceiling of Correlated Questions: Why the Sixth Question Is Nearly Wasted. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-235 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026235,
  author    = {Fan, Yafu},
  title     = {Fisher Information Matrices and the Cramér-Rao Bound: the Fewest Questions Needed to Estimate a Budget or a Birth Time},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book