Scribd
Upload
22 views

MAL522Tut4new Inference2

The document contains 21 problems related to statistical inference and hypothesis testing. The problems cover topics such as finding confidence intervals, determining necessary sample sizes, performing tests of significance using asymptotic distributions, and comparing tests to determine critical regions, levels of significance, and power. Tables of common statistical distributions are also provided.

Uploaded by

Shashi Raj
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
22 views

MAL522Tut4new Inference2

The document contains 21 problems related to statistical inference and hypothesis testing. The problems cover topics such as finding confidence intervals, determining necessary sample sizes, performing tests of significance using asymptotic distributions, and comparing tests to determine critical regions, levels of significance, and power. Tables of common statistical distributions are also provided.

Uploaded by

Shashi Raj
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/ 2

Department of Mathematics

MAL 522 (Statistical Inference)


Tutorial Sheet No. 4
1. Let s = 0.68 and x = 3.42 be the standard deviation and the mean respectively for a sample of size n = 16.
Assume that the population is a normal distribution with mean µ and variance σ 2 . Find 90% confidence
interval for µ.
2. In estimating the mean of a normal distribution N (µ, σ 2 ) having known standard deviation σ by using a
confident interval based on a sample of size n, atleast how large should n be in order for the probability to
be 0.99 that the confidence interval for µ be of length not greater than L?
3. Using asymptotic distribution of MLE, find a (1 − α)100% confidence interval for parameter θ of exponential
distribution.
4. Find a shortest length confidence interval for mean µ of N (µ, σ 2 ) population when σ 2 is unknown. Also, find
shortest length confidence interval for σ 2 .
5. Suppose that the variance of the speeds with which men and women can perform a certain task are respectively
σ1 = 12 seconds and σ2 = 15 seconds. If 20 men and 25 women required on the average 29.4 seconds and
32.5 seconds to perform the given task, obtain a 0.95 confidence interval for the difference between the true
average times it takes men and women to perform the task.
6. It is desired to test the hypothesis µ = 0 against the alternative µ > 0 on the basis of a random sample of size
9 from a normal population with variance σ 2 = 1. Show that X > 0.78 is the critical region of size α = 0.01.
7. A sample of size 1 is taken from an exponential pdf f (x, θ) = θ1 e− θ x , x > 0. To test H0 : θ = 1 against
1

H1 : θ > 1, the following test is used. Reject H0 if X > 2, otherwise accept H0 . Find the size, the level of
significance and the power function of the test.
8. Let X1 , X2 be samples from U (θ, θ + 1) distribution. For testing H0 : θ = 0 against H1 : θ > 0 we have two
competing non { randomised tests. {
1, X1 > 0.95 1, X1 + X2 > c
(I) ϕ1 (X1 ) = (II) ϕ2 (X1 , X2 ) =
0, X1 ≤ 0.95 0, X1 + X2 ≤ c

(a) Choose c so that ϕ2 has the same size as ϕ1 .


(b) Find power function of each of the two tests.
9. A sample of size 1 is taken from Poisson distribution with mean λ. To test H0 : λ = 1 against H1 : λ = 2,
consider the nonrandomised test ϕ(X) = 1, if X > 3 and = 0 if X ≤ 3. Find probabilities of type I and type
II errors and the power of the test against λ = 2. If it is required to achieve the size of test equal to 0.05,
how should one modify the test ϕ to a randomised test.
10. An urn contains 10 marbles of which M are white and 10 − M are black. To test that M = 5 against
the alternating hypothesis that M = 6, one draws 3 marbles from the urn without replacement, the null
hypothesis is rejected if the sample contains 2 or 3 white marbles; otherwise it is accepted. Find the size and
the power function of this test.
11. Using a random sample X1 , X2 , . . . , Xn from Bernoulli distribution with parameter ∑p, it is desired to test
n
H0 : p = 0.49 against H1 : p = 0.51. Consider a test that rejects H0 when i=1 Xi is large. Find
approximately the sample size needed in order that the probabilities of errors of both types are approximately
0.01 (Use central limit theorem).
12. A sample of size 1 is taken from pdf f (x, θ) = θ22 (θ − x), 0 < x < θ. Find a MP test of size α for the
hypothesis H0 : θ = θ0 against H1 : θ = θ1 (θ1 < θ0 ).
13. Find Neyman-Pearson size α test of H0 : β = 1 against H1 : β = β1 (> 1) based on a sample of size 1 from
f (x, β) = βxβ−1 , 0 < x < 1.

1
14. Let α be a real, 0 < α < 1 and ϕ∗ be a MP test of size α for testing H0 against H1 . Also let β = EH1 (ϕ∗ (X)) <
1. Show that 1 − ϕ∗ is an MP test for testing H1 against H0 at level 1 − β.
15. Let X1 , X2 , . . . , Xn be a random sample from B(1, p). Find a MP test of size α = 0.0547 of the hypothesis
H0 : p = 0.5 against H1 : p = 0.25. Find the power of the test so obtained. For what α level does there exist
a nonrandomised MP test. Is this test UMP for H0 : p ≤ 0.5 against H1 : p > 0.5?

16. Show that each of the following families has a MLR in a statistic T and find T . (a) N (θ, σ 2 ), σ 2 known. (b)
Poisson distribution with parameter λ. (c) B(n, θ), n known.

17. Find a UMP test for H0 : θ ≤ θ0 against H1 : θ > θ0 based on a sample of size n from pdf/pdf (a)
−θ x
f (x, θ) = e x!θ , x = 0, 1, . . . , θ > 0 (b) f (x, θ) = θ1 x θ −1 , 0 < x < 1, θ > 0.
1

18. Show that f (x, θ) = θ


x2 , x > θ, θ > 0 has MLR in X(1) and obtain a UMP test for H0 : θ ≥ θ0 against
H1 : θ < θ0 .
19. For a sample of size 1 from double exponential distribution with pdf f (x, θ) = 12 exp− | x − θ |, −∞ < θ < ∞,
construct a UMP test of level α = 0.05 for H0 : θ ≤ 0 against H1 : θ > 0.
20. Find a UMP size α test for H0 : θ ≥ θ0 against H1 : θ < θ0 based on a sample of n observations from the
pdf (a) N (θ, 1) (b) Gamma(θ, 1). Verify that UMP test is unbiased.
21. Using a UMP test for H0 : θ = θ0 against H1 : θ < θ0 construct a (1 − α)100% confidence interval for mean
θ of exponential distribution.

Table Values
P( Z is a standard normal distribution ≥ Zα ) = α
P( χ2 r.v. with n degrees of freedom ≥ χ2n,α ) = α
P( t r.v. with n degrees of freedom ≥ tn,α ) = α
P( F r.v. with n1 and n2 degrees of freedom ≥ Fn1 ,n2 ,α ) = α
Z0.025 = 1.96; Z0.05 = 1.645; Z0.0764 = 1.43; Z0.01 = 2.33; Z0.035 = 1.81
χ29,0.05 = 16.917; χ26,0.05 = 12.6; χ25,0.05 = 11.1; χ24,0.05 = 9.48; χ23,0.05 = 7.81; χ22,0.05 = 5.99; χ29,0.95 = 3.334
t8,0.025 = 2.31; t9,0.025 = 2.26; t10,0.025 = 2.22
F9,11,0.025 = 0.1539; F10,15,0.025 = 3.5217; F15,10,0.025 = 3.0602
Note: If above table values are not matched, please leave the answer without numerical.

You might also like