Scribd
Upload
52 views1 page

Homework Economic

A random sample of 100 observations from a Bernoulli distribution had 80 successes and 20 failures. An estimator for the odds ratio p/(1-p) was constructed. This estimator approximately follows a normal distribution with mean 4 and variance 1. A consistent estimator θ^ of θ was also described. When n=100, the probability that θ^ is greater than or equal to 1.196θ is approximately 0.025. The standard error of θ^ is θ/√n. The approximate mean and variance of 1/θ^ are 1/θ and 1/(nθ^2) respectively. For large n, 1/θ^ approximately follows a normal distribution with mean 1/θ and variance 1/nθ^

Uploaded by

duc anh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
52 views1 page

Homework Economic

A random sample of 100 observations from a Bernoulli distribution had 80 successes and 20 failures. An estimator for the odds ratio p/(1-p) was constructed. This estimator approximately follows a normal distribution with mean 4 and variance 1. A consistent estimator θ^ of θ was also described. When n=100, the probability that θ^ is greater than or equal to 1.196θ is approximately 0.025. The standard error of θ^ is θ/√n. The approximate mean and variance of 1/θ^ are 1/θ and 1/(nθ^2) respectively. For large n, 1/θ^ approximately follows a normal distribution with mean 1/θ and variance 1/nθ^

Uploaded by

duc anh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

additional problem 2

A random sample from a Bernoulli(p) distribution has 80 successes and 20 failures. Construct a
consistent estimator for the odds ratio p/(1−p). What is the approximate pdf of this estimator?
iid
Solution: If X1 , . . . , Xn ∼ Bernoulli(p), then EX1 = p. Hence, a consistent estimator of p is
p
p̂ := nj=1 Xj /n because p̂ −
P
→ EX1 = p by the WLLN. Hence, by the continuous mapping theorem
d
(CMT), p̂/(1 − p̂) is a consistent estimator of p/(1 − p). From the CLT, we know that n1/2 (p̂ − p) −
→
1/2 d
N(0, p(1−p)). Hence, letting g(p) := p/(1−p), the exact delta method implies that n (g(p̂)−g(p)) − →
d
N(0, p(1−p)[g 0 (p)]2 ). But g 0 (p) = 1/(1−p)2 . Hence, it follows that n1/2 (g(p̂)−g(p)) −
→ N(0, p/(1−p)3 ).
d
Consequently, g(p̂) ≈ N(g(p), p/[n(1−p)3 ]). In the given dataset, we know that n = 100, p̂ = 80/100 =
d d
0.8, and g(p̂) = 0.8/0.2 = 4. Therefore, g(p̂) ≈ N(4, 0.8/[100(1 − 0.8)3 ]) = N(4, 1).

additional problem 5
d
Let θ̂ be a consistent estimator of θ > 0 such that n1/2 (θ̂ − θ) →
− N(0, θ2 ) as n → ∞.
(i) What is Pr(θ̂ ≥ 1.196θ) when n = 100?
(ii) What is the standard error of θ̂?
(iii) Find the approximate mean and variance of 1/θ̂.
(iv) What is the distribution of 1/θ̂?
Solution:
(i) Note that
Pr(θ̂ ≥ 1.196θ) = Pr(n1/2 (θ̂ − θ) ≥ n1/2 (1.196θ − θ))
.
= Pr(N(0, θ2 ) ≥ n1/2 0.196θ)
= Pr(N(0, 1) ≥ n1/2 0.196)
= Pr(N(0, 1) ≥ 1.96) (n = 100)
= 0.025.
√
q
(ii) se(θ̂) := asvar(
[ θ̂)/n = θ̂/ n.
.
(iii) Let g(θ) := 1/θ. By a first order Taylor expansion, g(θ̂) = g(Eθ̂) + g 0 (Eθ̂)(θ̂ − Eθ̂). Hence,
. . d .
Eg(θ̂) = g(Eθ̂) = 1/Eθ̂ = 1/θ because θ̂ ≈ N(θ, θ2 /n). Similarly, var g(θ̂) = [g 0 (Eθ̂)]2 var(θ̂) =
.
var(θ̂)/(Eθ̂)4 = 1/(nθ2 ).
d d
(iv) By the exact delta method, n1/2 (g(θ̂) − g(θ)) −→ N(0, [g 0 (θ)]2 θ2 ) = N(0, 1/θ2 ), i.e.,
1 1 d 1
n1/2 ( − ) − → N(0, 2 ) as n → ∞.
θ̂ θ θ
The practical implication of this result is that, for large enough n,
1 d 1 1
≈ N( , 2 ).
θ̂ θ nθ

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy