IS/BM-B

25/02/2020

DEPARTMENT OF STATISTICS AND ACTUARIAL SCIENCE

STAT 334: STATISTICAL METHODS II
EXERCISE ONE

1. (a) By given an appropriate intuitive example, explain the meaning of statistical decision.
(b) Explain the concept of hypothesis testing and how it aids in statistical decision.

2. With respect to statistical inference distinguish between the following pairs of terminologies.

(a) Null Hypothesis and Alternate Hypothesis.

(b) Simple Hypothesis and Composite Hypothesis.
(c) Significance probability and Significant Values.
(d) Level of significance and Power of a test.
(e) Type I Error and Type II Error.

3. Explain what is meant by ”critical region” of a test and write the mathematical form of it.

4. It has been asserted that a decision regarding a hypothesis test about a population parameter at
significance level could be reached using a 100(1α)% confidence interval for the parameter I
question. Is this assertion true? Explain your answer.

5. Consider a sample of independent observations x1 , x2 , ..., xn from an exponential distribution with

an unknown parameter θ. It is required to test the null hypothesis H0 : θ = θ0 against H1 : θ < θ0
(where θ0 is a specified value).
1
(a) If the test statistic T = is used;
x̄n
i. What values of the test statistic constitute departures from? Justify your answer and
write the form of the critical region.
ii. Given that θ0 = 2.5, n = 10 and α = 0.05, determine the critical region.
iii. What conclusion will you reach if x̄ = 0.35?
(b) Suppose the test statistic T = X(1) , the minimum order statistic is used instead;
i. What values of the test statistic constitute departures from? Justify your answer and
write the form of the critical region.
ii. Given that θ0 = 2.5, n = 10 and α = 0.05, determine the critical region.
iii. Compute the probability of Type II error and hence the power of the test if θ = 1.5

6. Suppose a random sample of independent observations x1 , x2 , ..., xn is drawn from a population

with the distribution f (x) = βe−β(x−θ) , x > θ. It is required to test the hypothesis H0 : θ = θ0
against H1 : θ > θ0 . The test statistic proposed for the test is T = X(1) , the minimum order
statistic.

(a) What values of T constitute significant departures from H0 ?

(b) Determine the critical region if θ0 = 1.2, n = 12, β = 0.20 and α = 0.01.
(c) Compute the significance probability given that X(1) = 1.8 and hence make the necessary
conclusions.
(d) If θ = 2.10, calculate the power of the test.

STAT 334/Exx1 Page 1 of 4 IS/BM-B/2019

 7. A random sample y1 , y2 , ..., yn is drawn from a population that is uniformly distributed over the
interval [0, θ]. It is required to test the hypothesis H0 : θ = β (where β is a specified value of θ
greater than zero) against the alternative H1 : θ < β.

(a) If the test statistic to be used is T = Y(n) , the maximum order statistic;
i. what value of T constitute departures from the null hypothesis?
ii. hence find the critical region if β = 15.4, , n = 9 and α = 0.05.
iii. Compute the power of the test if β = 10.
∑n
(b) If the test statistic T = − ln yi
i=1
i. what value of T constitute departures from the null hypothesis?
ii. hence find the critical region if β = 15.4, , n = 9 and α = 0.05.
[ ]
∑ n (y )
i
iii. Compute the power of the test if β = 10. Hint : − ln ∼ Γ(n, 1)
i=1
θ

8. A random sample y1 , y2 , ..., yn is drawn from a population that is uniformly distributed over the
interval [2.5, β]. It is required to test the hypothesis H0 : β = 8 (where k is a specified value of β
greater than 2.5) against the alternative H1 : β > 8. The test statistic T = Y(n) , the maximum
order statistic, is to be used.

(a) What values of the test statistic constitute departures from the null hypothesis?
(b) Given that Y(n) = 10.5 and n = 15, calculate the p-value and make necessary conclusions at
α = 0.01
(c) Calculate the power if β = 10.0

9. A random sample from a population normally distributed with mean µ and variance σ 2 is given
below:
92.3 89.4 76.9 65.2 49.1 96.7 69.5 72.8 67.5 52.8 88.5 79.2 72.9 68.7 75.7

(a) Suppose it is known that σ = 8.8 and it is required to test the hypothesis, H0 : µ = 70.5
against H1 : µ ̸= 70.5. Test these hypotheses using the test statistic t(x) = x̄ at α = 0.02
(b) Suppose σ is unknown and the hypotheses in (a) is being tested,how would you proceed?
Conduct the test at α = 0.02. Does it conflict the results in (a)?
(c) Assuming σ is unknown it is required to test the hypothesis that H0 : σ 2 = 100 against
H1 : σ 2 ̸= 100, conduct the test and draw necessary conclusions at α = 0.01.

10. State explain situations where non-parametric procedures are needed in statistical inference.

11. In order to check whether people save time by using a new method in performing a particular task,
the time recorded (in hours) when a standard method and the new method are used by 9 randomly
selected people were observed. The results are as follows:

Person 1 2 3 4 5 6 7 8 9
Standard 9.5 11 9 10.5 8 12.5 4 7.5 7.5
New 15 19 10 17 12 12.5 9 17.5 9.5

(a) Using the sign test and the p-value approach, test the hypothesis that, the median of the time
saved (N ew − Standard) is zero.
(b) Find the critical region of the test you conducted in a above at 5% level of significance.

STAT 334/Exx1 Page 2 of 4 IS/BM-B/2019

 (c) Test the hypothesis that, the 50th percentile of the time saved is greater than zero using the
Wilcoxon Signed Rank test and the p-value method.

12. In order to find out whether the mean length of stay in a particular hospital is more than 5 days,
the hospital’s bio-statistician randomly selected 8 patients and recorded the number of days each of
them spent at the hospital as follows.

6.1 3.2 7.4 5.0 3.1 10.0 8.0 5.0

By means of p – value and using the sign rank test, what conclusion can you draw at 5% level of
significance?
13. (a) Define the power of a test based on a statistic T in testing the hypothesis H0 : θ = θ0 against
H1 : θ = θ1 , where θ0 < θ1
(b) A test of H0 : µ = µ0 against H1 : µ > µ0 is to be carried out for the mean µ of a normal
population with known standard deviation σ. The available data is a random sample of size n
from this population. Given that the level of significance of this test is α, show that the power
function of this test is [ √ ]
n
K(µ) = 1 − Φ Zα − (µ − µ0 )
σ
where Φ denotes the cumulative standard normal distribution function and Zα is the upper α
percentile value of the standard normal distribution.
14. An experimenter has prepared a drug dosage level that she claims will induce sleep for 80% of
people suffering from insomnia. After examining the dosage, we feel that her claims regarding the
effectiveness of the dosage are inflated. In an attempt to disprove her claim, we administer her
prescribed dosage to 20 insomniacs and we observe Y , the number for whom the drug dose induces
sleep. We wish to test the hypothesis H0 : p = 0.8 against the alternative H0 : p < 0.8. Assume the
rejection region {y ≤ 12} is used.
i. In terms of this problem, what is a type I error? Find α
ii. In terms of this problem, what is a type II error? Find β when p = 0.65
15. Suppose that we wish to test the null hypothesis H0 that the proportion of ledger sheets with
errors is equal to 0.05 versus the alternative H1 that the proportion is larger than 0.05, by using
the following scheme, two ledger sheets are selected at random. If both are error free, we reject H0
. If one or more contains an error, we look at a third sheet. If the third sheet is error free, we reject
H0 . In all other cases, we accept H0 .
i. In terms of this problem, what is a type I error?
ii. Find the value of α associated with this test.
iii. In terms of this problem, what is a type II error?
iv. Find β = P (type II error) as function of p.
16. Consider the following two hypothesis about the probability function of the random variable X.
{1 1
+ x , −1 ≤ x ≤ 1
H0 : f (x) = 2 4
0 , otherwise
{1 1
− x , −1 ≤ x ≤ 1
H1 : f (x) = 2 4
0 , otherwise
A single observation of X is made and the test procedure is that if the value of X is less than C,
where −1 ≤ C ≤ 1 , the null hypothesis is accepted, otherwise H1 is accepted.

STAT 334/Exx1 Page 3 of 4 IS/BM-B/2019

 (a) Determine the form of the critical region, based on a single observation of X, such that the
probability of wrongly accepting H0 is 0.05.
(b) Hence find the probability of wrongly accepting H1

17. The label on a certain malaria drug warns that 20% of persons using the drug are likely to develop
an allergic reaction. A medical doctor who has been prescribing the drug for some time suspects
that the proportion of users who develop an allergy is higher than 20%.

(a) In reaching a policy decision concerning the true proportion of allergies developed, explain the
circumstances under which the doctor would be committing;

i. Type I error ii. Type II error

(b) If the true proportion of potential allergies is 25%, how large should a sample be to ensure
that the probability of Type I error is 0.05 and the probability of Type II error is 0.015?
What would be the decision rule?

18. In order to know the age at which student smokers started smoking, 15 randomly selected students
smokers about the age they started smoking. The results is shown below.

21 17 20 19 16 15 17 18 25 24 19 16 22 19 23

(a) Using the sign test and the critical region method, test at 10% significance level the
hypothesis that the third quartile is 20 years.
(b) Use to the Wilcoxon Signed Rank test to test at 10% level of significance whether the median
age is less than 19 years.

19. The hypothesis H0 : p = 0.42 is being tested against H1 : p > 0.42. The test statistics is T = X,
the number of successes and the critical region for the test is {X : X > 5}.

(a) Compute the probability of type I and II error for these values of n {10, 15, 20, 30, 40}
(b) Compute the corresponding power for the various values of n. Comment on the relationship
between the sample size and the power.

20. It is required to test the Hypothesis H0 : µ = 42.30 against H1 : µ < 42.30 for N (µ, 153.76)
population. The test statistic is T = x̄ and the critical region is {x̄ : x̄ ≤ 40.27}

(a) What sample size will ensure that the probability of type I error is 0.05?
(b) What sample size will ensure that the probability of type II error is 0.135 given that
µ = 37.50?
(c) Derive the power function for µ = µ0 , where µ0 < 42.30 as a function of the sample size n and
µ0
(d) Write the power function when µ0 = 35.40 and hence compute the power of the test when
n = {50, 150, 200, 500} and comment on your answers.

STAT 334/Exx1 Page 4 of 4 IS/BM-B/2019