Skip to main content

Mathematical Statistics I · Bias, risk, consistency, and point estimation

Invariance moves the MLE but not the UMVUE

Problem

Mathematical Statistics I practice problem

Let X1,…,XnX_1,\ldots,X_n be independent Bernoulli⁡(p)\operatorname{Bernoulli}(p) observations, 0≤p≤10\le p\le1, n≥2n\ge2, and let g:[0,1]→Rg:[0,1]\to\mathbb R be any function, not necessarily injective. Let p^\hat p be a maximizer of the Bernoulli likelihood. Prove that g(p^)g(\hat p) maximizes the induced likelihood for the target τ=g(p)\tau=g(p). State the induced-likelihood definition you use so that the claim remains valid when gg is not one-to-one.

Take g(p)=p(1−p)g(p)=p(1-p). Write the MLE of p(1−p)p(1-p) explicitly in terms of Xˉ\bar X. Then compute Ep[Xˉ(1−Xˉ)]\mathbb E_p[\bar X(1-\bar X)] and produce an unbiased estimator of p(1−p)p(1-p). Explain why invariance does not force that unbiased estimator to equal the MLE, and why a UMVUE need not be a function g(p^)g(\hat p) of the MLE of pp.

Do not treat this as a test of H0:p=p0H_0:p=p_0, and do not Rao–Blackwellize a Poisson zero count.

Hint

Hint

If L(θ^)≥L(θ)L(\hat\theta)\ge L(\theta) for every θ\theta, then the supremum of LL over each gg-fiber is attained at a fiber containing θ^\hat\theta.

Check your work

Work the problem yourself first. Then open it in Training to check your answer and read the full worked solution.

The answer check and full solution for this problem come with ProofAnvil Practice membership ($19 USD monthly). See membership. Or start with the free Mathematical Statistics I sample problem: Try the free sample problem.

More Mathematical Statistics I practice problems