S.P.

MANDALI’S
SIR PARASHURAMBHAU COLLEGE (AUTONOMOUS), PUNE- 30
Subject: Statistics
ST36252: Testing of Hypothesis
Paper Code: ST36252 Class: T.Y. B.Sc.

(Assignment)
Q.1. Attempt all of the following. Justify your answers.
i) Let 𝑥1 , 𝑥2 , … , 𝑥𝑛 be a random sample of size 𝑛 from population with continuous and

strictly increasing cumulative distribution function 𝐹(. ) with an unknown median 𝑀.

To test 𝐻0 : 𝑀 = 10 against 𝐻1 : 𝑀 > 10 at level 𝛼. Let the Statistic T denote the

number of observations larger than 10. Let 𝑡𝑜 be the observed value of test statistic T.

Consider the test which rejects 𝐻0 if 𝑇 ≥ 𝑐, then the p-value of the test is:

𝑛! 𝑛!
A) ∑10
𝑖=0 𝑖 !(𝑛−𝑖)! (0.5)
𝑛
C) ∑𝑛𝑖=𝑡0 (0.5)𝑛
𝑖 !(𝑛−𝑖)!

𝑡0 𝑛! 𝑛!
B) ∑𝑖=0 (0.5)𝑛 D) ∑𝑛𝑖=10 (0.5)𝑛
𝑖 !(𝑛−𝑖)! 𝑖 !(𝑛−𝑖)!

ii) In testing of hypothesis problem, which of the following statements is true?

A) The sum of the probability of Type I error and the probability of Type II error
should be 1.

B) The probability of Type II error is always less than the probability of Type I
error.

C) Type II error occurs if the test accepts the null hypothesis when the null
hypothesis is actually false.

D) Type II error occurs if the test rejects the null hypothesis when the null
hypothesis is actually false.

iii) Let 𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 be the random sample of size 4 from a Bernoulli distribution with
parameter 𝜃, 0 < 𝜃 < 1. To test 𝐻0 : 𝜃 = 0.25 against 𝐻1 : 𝜃 > 0.25 suppose Ho is
rejected if and only if ∑4𝑖=1 𝑥𝑖 > 2. If 𝛼 is the probability of Type I error and 𝛾(𝜃) is
the power function of the test, then the value of 16𝛼 + 7𝛾(0.5) is equal to ______
A) 3.03 B) 3 C) 2.06 D) 2.77

1
iv) Let 𝑋1 , 𝑋2 , … , 𝑋𝑛 be a sequence of identically and independently distributed random
variables each having probability density function 𝑓(. ) and median 𝜃. We want to test
𝐻0 : 𝜃 = 𝜃0 against 𝐻1 : 𝜃 > 𝜃0 . Consider the test that rejects Ho if 𝑆 > 𝑐 for some 𝑐
depending upon the size of the test, where 𝑆 is cardinality of the set {𝑖, 𝑋𝑖 > 𝜃0 , 𝑖 =
1,2, … , 𝑛}. Then which of the following statement is true?

A) Under 𝐻0 , the distribution of 𝑆 depends on 𝑓(. )

B) Under 𝐻1 , the distribution of 𝑆 does depends on 𝑓(. )
C) The power function depends on 𝜃.
D) The power function does not depends on 𝜃.

v) Let X be a discrete random variable with probability mass function 𝑃 ∈ {𝑃0 , 𝑃1 }, where
X 7 8 9 10
𝑃1 (𝑋) 0.69 0.10 0.16 0.05
𝑃0 (𝑋) 0.90 0.05 0.04 0.01

To test 𝐻0 : 𝑃 = 𝑃0 against 𝐻1 : 𝑃 = 𝑃1 , the power of the most powerful test of size 0.05
based on single observation X is ____

A) 0.20 B) 0.21 C) 0.31 D) 0.10

vi) Let 𝑥1 , 𝑥2 , … , 𝑥𝑛 (𝑛 ≥ 2) be a sequence of identically and independently distributed

observations from 𝑁(𝜇, 1), where 𝜇 is unknown. Consider the problem of testing
𝐻0 : 𝜇 ∈ [1,2] against 𝐻1 : 𝜇 < 1𝑜𝑟 𝜇 > 2. Which of the following statements is true?

A) Critical region of level ∝ of UMP test is of form 𝑐1 < ∑𝑛𝑖=1 𝑥𝑖 < 𝑐2 , where
𝑐1 & 𝑐2 are such that the test is of level ∝.
B) Critical region of level ∝ of UMP test is of form ∑𝑛𝑖=1 𝑥𝑖 > 𝑐1 𝑜𝑟 ∑𝑛𝑖=1 𝑥𝑖 < 𝑐2 ,
where 𝑐1 & 𝑐2 are such that the test is of level ∝.
C) At any level ∝ (0 <∝ < 1), UMP test does not exists.
D) At any level ∝ (0 <∝ < 1), UMP test exists with power greater than ∝.

vii) Let 𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 be the random sample of size 4 from a population from 𝑈(0, 𝜃),
where 𝜃 > 0 is an unknown parameter. Let 𝑋(4) = max{𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 },
To test 𝐻0 : 𝜃 = 1 against 𝐻1 : 𝜃 = 0.1, consider the critical region that rejects 𝐻0 if
𝑋(4) < 0.3. Let 𝛼 be the probability of Type I error. Then 100 𝛼 equal to ____

viii) Let X be a random sample of size one from a population having 𝑁(0, 𝜃 2 ) distribution.
Where 𝜃 > 0 is unknown parameter. To test To test 𝐻0 : 𝜃 2 = 1 against 𝐻1 : 𝜃 2 = 2
using Most Powerful test of size 0.05. The critical region is given by 𝜆(𝑋) ≥ 𝐶 where
𝐶 ≥ 0 is a constant. The value of C is ______

2
 ix) Let {0,2 } be a realization of random sample of size 2 from 𝐵(𝑛 = 2, 𝑝).
To test 𝐻0 : 𝑝 = 0.5 against 𝐻1 : 𝑝 ≠ 0.5, the observed value of likelihood ratio test
statistic equals ______

x) Let X be a random variable with probability density function

−𝜃𝑥
𝑓(𝑥; 𝜃) = {𝜃𝑒 ; 0 < 𝑥 < ∞ ,𝜃 > 0
0 ; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
To test 𝐻0 : 𝜃 = 1 against 𝐻1 : 𝜃 > 1, the following test is used:
Reject Ho if and only if X > 𝑙𝑜𝑔𝑒 20, then the size of the test is _____

xi) Let 𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 , 𝑥5 be the random sample of size 5 from a population having density
1 −|𝑥−𝜃|
𝑓(𝑥; 𝜃) = {2 𝑒 ; −∞ < 𝑥 < ∞ , −∞ < 𝜃 < ∞
0 ; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
For testing 𝐻0 : 𝜃 = 0 against 𝐻1 : 𝜃 = −1, let ∑5𝑖=1 𝑦𝑖 be the sign test statistic,
Where
1 ; 𝑥𝑖 > 0
𝑦𝑖 = {
0 ; 𝑂𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
The power of the test which rejects Ho if and only if ∑5𝑖=1 𝑦𝑖 ≤ 2 is ______

xii) Let X be a random variable having 𝑁(𝜃, 1) distribution. Consider testing 𝐻0 : 𝜃 =

0 against 𝐻1 : 𝜃 ≠ 0 at 𝛼 = 0.617 level of significance. The power of the likelihood
ratio test at 𝜃 = 1 equals to ______

Q.2. Attempt any all of the following:

i) Suppose that (X, Y) has joint probability mass function
Y
X 0 1
0 𝜃 0.5 − 𝜃
1 0.5 − 𝜃 𝜃
Where 0 ≤ 𝜃 ≤ 0.5 is an unknown parameter. Consider testing
1 1
𝐻0 : 𝜃 = 4 against 𝐻1 : 𝜃 > 3 based on a random sample {(𝑦𝑖 , 𝑥𝑖 ) ; 𝑖 = 1,2,3, … , 𝑛 }

From the above probability mass function. Let 𝑀 be the cardinality of the set
{𝑖; 𝑦𝑖 , = 𝑥𝑖 , 𝑖 = 1,2, … , 𝑛}. If 𝑚 is the observed value of 𝑀, then construct the
likelihood ratio test.
ii) Let X have a p.d.f. of the form:
1 −𝑥/𝜃
𝑓(𝑥; 𝜃) = {𝜃 𝑒 ; 0 < 𝑥 < ∞ ,𝜃 > 0
0 ; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
To test 𝐻0 : 𝜃 = 2 against 𝐻1 : 𝜃 = 1 based on a sample of size 2 and a critical region is
given as 𝑊 = {(𝑥1 , 𝑥2 ): 9.5 ≤ 𝑥1 + 𝑥2 }
Find: i) Power of the test
ii) size of the test.

3
iii) Let {(𝑦𝑖 , 𝑥𝑖 ) ; 𝑖 = 1,2,3, … , 25, } be the n pairs of observation on (𝑌, 𝑋). Consider the
simple linear regression model 𝑦𝑖 = 𝛽0 + 𝛽1 𝑥1𝑖 + 𝛽2 𝑥2𝑖 + 𝜀𝑖 . 𝛽0 , 𝛽1 & 𝛽2 are unknown
parameters.
0 ; 𝑖𝑓 𝑖 ≠ 𝑗
𝐸(𝜀𝑖 ) = 0, 𝑖 = 1,2, … , 𝑛 and for 𝑖, 𝑗 = 1,2,3, … , 𝑛, 𝑐𝑜𝑣(𝜀𝑖 , 𝜀𝑗 ) = { 2 .
𝜎 ; 𝑖𝑓 𝑖 = 𝑗
Let 𝐹𝑎,𝑚,𝑛 be such that (𝑌 > 𝐹𝑎,𝑚,𝑛 ) = 𝛼 , Where Y is random variable having an F-
distribution with 𝑚 and 𝑛 degrees of freedom. Suppose that testing:
𝐻0 : 𝛽1 = 𝛽2 = 0 against 𝐻1 : 𝐴𝑡 𝑙𝑒𝑎𝑠𝑡 𝑜𝑛𝑒 𝑜𝑓𝛽1 , 𝛽2 𝑖𝑠 𝑛𝑜𝑡 0
𝑅2
Involves computing 𝐹0 = 11 1−𝑅2 and rejecting Ho if computed value of 𝐹0 > 𝐹𝛼,2,22.
Obtain the smallest value of 𝑅 2 that would lead to the rejection of Ho at 𝛼 = 0.05.

iv) Construct Sequential Probability Ratio Test (S.P.R.T.) for testing

𝐻0 : 𝜃 = 𝜃0 against 𝐻1 : 𝜃 = 𝜃1 (> 𝜃0 ) on the basis of a random sample of size 𝑛 from
Pareto distribution with density
𝜃𝑎𝜃
𝑓(𝑥; 𝜃) = { 𝑥 𝜃+1 ; 𝑎 < 𝑥 < ∞ , 𝜃 > 0
0 ; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

v) Let 𝑥1 , 𝑥2 , … , 𝑥𝑛 be a sequence of identically and independently distributed observations

from 𝑁(𝜇, 𝜎 2 ), where 𝜇 is known, 𝜎 2 being unknown. Obtain the S.P.R.T. for testing
𝐻0 : 𝜎 2 = 𝜎02 against 𝐻1 : 𝜎 2 = 𝜎12 (> 𝜎02 )

vi) Let a random sample of size 4 from a population having 𝑁(𝜇, 1) distribution yield a
sample mean 0.16. If the likelihood ratio test of size 0.05 is used to test 𝐻0 : 𝜇 = 0
against 𝐻1 : 𝜇 ≠ 0. Compute the power of the test at the sample mean.
vii) Let 𝑥1 , 𝑥2 , … , 𝑥𝑛 (𝑛 ≥ 2) be identically and independently distributed observations with
𝑝. 𝑚. 𝑓.
𝑝 𝑥 (1 − 𝑝)1−𝑥 ; 𝑥 = 0, 1
𝑃(𝑋 = 𝑥) = {
0 ; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Show that 𝐶 = {(𝑥1 , 𝑥2 , … , 𝑥𝑛 ); ∑𝑛𝑖=1 𝑥𝑖 ≤ 𝐶 } is the best critical region for testing
𝐻0 : 𝑝 = 0.5 against 𝐻1 : 𝑝 = 0.3. Use the Central Limit Theorem(CLT) to find 𝑛 and 𝑐
such that approximately 𝑃(∑𝑛𝑖=1 𝑥𝑖 ≤ 𝐶|𝐻0 ) = 0.10 and 𝑃(∑𝑛𝑖=1 𝑥𝑖 > 𝐶|𝐻1 ) = 0.20

*****