UNIVERSITY OF CAPE TOWN

DEPARTMENT OF STATISTICAL SCIENCES

STA1000F 2008
Internal Examiners: Mr G. Distiller, Ms K. Leask
Internal reviewer: Prof G. Barr
Time: 90 minutes

SECTION A : MCQ (20 MARKS)

THE NEXT TWO QUESTIONS REQUIRE THE USE OF THE FOLLOW-

ING INFORMATION:
The number of users of an automatic banking machine is distributed such that the average
number of users per 5-minute interval is 1.5.
1. Which of the following probability distribution best describes the number of users
of such a facility? [1]

(a) t-distribution
(b) Normal distribution
(c) Binomial Distribution
(d) Poisson distribution *
(e) Chi-square distribution

2. What is the probability (to three decimal places) of observing 2 or more users in
the next 15 minutes? [2]

(a) 0.112
(b) 0.251
(c) 0.939 *
(d) 0.297
(e) None of the above

USE THE FOLLOWING INFORMATION TO ANSWER THE NEXT

TWO QUESTIONS:
The following density function describes the random variable X:
 x
0<x<5
f (x) = 25
10−x
25
5 ≤ x < 10

3. What is the probability (to two decimal places) that X lies between 1 and 3? [1]

(a) 0.18
(b) 0.16 *
(c) 0.20
 (d) 0.02
(e) None of the above

4. The probability (to two decimal places) that X lies between 4 and 8 is: [2]

(a) 0.50
(b) 0.32
(c) 0.18
(d) 0.42
(e) 0.60 *

5. A biologist knows that about 2% of a certain breed of frog mutate. Given a random
sample of 30 developing frogs, what is the probability that the sample will contain
at least one mutated frog? [1]

(a) 0.306
(b) 0.455 *
(c) 0.545
(d) 0.605
(e) 0.694

6. The time until first failure of a brand of ink jet printers is approximately normally
distributed with a mean of 1500 hours and a standard deviation of 200 hours. What
fraction of printers fail before 1000 hours? [1]

(a) 0.99379
(b) 2.5%
(c) 0.00621 *
(d) 0.66667
(e) 0.33333

7. The p-value obtained from a classical hypothesis test is: [1]

(a) the probability that the null hypothesis is true given the data
(b) the probability that the null hypothesis is false given the data
(c) the probability of observing the data or more extreme values if the null hy-
pothesis is true *
(d) the probability of observing the data or more extreme values if the alternative
hypothesis is true
(e) the probability that the observed data were obtained due to chance

2
 USE THE FOLLOWING INFORMATION TO ANSWER THE NEXT
THREE QUESTIONS:
A social scientist claims that the average adult watches less than 25 hours of tele-
vision per week. He collects data on 257 individuals television viewing habits and
finds that their mean number of hours watching television was 23.4 hours. Assume
the population standard deviation is known to be eight hours, and the significance
level adopted is 1%.

8. What are the null and alternative hypotheses? [1]

(a) H0 : µ = 26, H1 : µ 6= 26
(b) H0 : µ < 26, H1 : µ 6= 26
(c) H0 : µ < 26, H1 : µ = 26
(d) H0 : µ = 26, H1 : µ < 26 *
(e) H0 : µ = 26, H1 : µ ≤ 26

9. What is the value of the test statistic? [2]

(a) -0.20
(b) 0.20
(c) 0.49
(d) -3.21 *
(e) 3.21

10. What is the conclusion based on the information above? [1]

(a) Since z < −z0.01 , we reject H0 and conclude that the social scientist is right.
*
(b) Since z < −z0.01 , we fail to reject H1 and conclude that the social scientist is
right.
(c) Since z > −z0.01 , we fail to reject H0 and conclude that the social scientists
claim can not be proved.
(d) Since z < −z0.01 , we fail to reject H0 and conclude that the social scientists
claim can not be proved.
(e) Since z > −z0.01 , we reject H1 and conclude that the social scientist is right.

11. A difference between the t and the normal distribution is: [1]

(a) The t-distribution is discrete; the normal is continuous

(b) The t-distribution is skew; the normal is symmetric
(c) The t-distribution is valid only for positive values; the normal for all real values
(d) The t-distribution has fatter tails than the normal distribution i.e. a larger %
of the dbn lies in the tails *
(e) None of the above

3
 An analyst examines sample data on two portfolios of common stocks in terms of
the average price-earnings ratio. He wishes to determine if there is a significant
difference between the two portfolios, when the difference is calculated as Portfolio
1 minus Portfolio2.
Portfolio 1 Portfolio 2
n1 = 7 n2 = 5
x1 = 16.60 x2 = 17.62
s21 = 68.74 s22 = 71.95

12. Assuming equal variances, what can you say about the pvalue for this test? [2]

(a) 0.10 < p < 0.20

(b) 0.30 < p < 0.40
(c) 0.60 < p < 0.80
(d) 0.40 < p < 0.60
(e) p > 0.80 *

13. After constructing a confidence interval for a population mean, you believe that the
interval is useless because it is too wide. In order to correct this problem, you need
to: [1]

(a) Increase the population standard deviation

(b) Increase the sample size *
(c) Increase the level of confidence
(d) Decrease the sample size
(e) Increase the sample mean

THE NEXT THREE QUESTIONS REFER TO THE FOLLOWING PROB-

LEM SETTING:
The following information is available for two samples drawn from independent nor-
mally distributed populations.

n1 = 16; s21 = 41.3; n2 = 13; s22 = 35.6

We would like to test if σ12 6= σ22 at a 5% significance level.

14. The correct test statistic to test this hypothesis is: [1]

(a) 1.08
(b) 1.16 *
(c) 1.29
(d) 3.28
(e) none of the above

15. The critical value for this test is: [1]

4
 (a) 2.62
(b) 2.96
(c) 3.18 *
(d) 2.48
(e) 1.96

16. A suitable conclusion is: [1]

(a) Reject H0 and conclude that the variances are not equal.
(b) Fail to reject H0 and conclude that σ12 = σ22 . *
(c) Fail to reject H0 and conclude that σ12 6= σ22
(d) A type II error has been made.
(e) None of the above.

5
 SECTION B: WRITTEN (30 Marks)

1. λ = 4/60 per second.

Pr(X > 11) = e−4/60∗11 = 0.480.

2. PrX < 89) = 0.9 ⇒ z = 1.28

Pr(X < 94) = 0.95 ⇒ z = 1.645
Thus
89 = 1.28σ + µ
and
94 = 1.645σ + µ
And as a result σ = 13.699 and µ = 71.465
R 15 1 1 1 3
3. (a) 12 10 dx = 10 {x|15
12 } = 10 × (15 − 12) = 10 or 30%. (2)
1 3
Or: P r[X > 12] = 25−15 × (15 − 12) = 10
Ra 1
(b) 5 10 dx = 0.8
x a
|
10 5
= 0.8
a−5
10
= 0.8
a = 13◦ C (2)
1
Or: 15−10 × (a − 5) = 0.8
a = 13 C◦
R 15 x 2 152 −52
(c) E[X] = 5 10 dx = x20 |15
5 = 20
= 10◦ C (2)

4. (a) X ∼ B(18, 0.6)

P [X ≥
3] = 1 − P [X < 3]
= 1 − P [2] − P [1]

−0P [0]
18 18 18
1 − ( 2 0.6 0.4 6) − ( 1 0.6 0.4 7) − ( 0 0.6 0.41 8) (3)
2 1 1 1

1 − 0.000026 = 0.999974 or 99.997%.

(b) P [X = 10] = 18

10
10
0.6 0.48 = 0.1734 or 17.34% (1)

5. (a) 4767 ± 1.761 × 3204

√
15
Note: t0.05
14 = 1.761 (1)
4767 ± 1.761 × 827.27(1)
(3310.18; 6223.82)(1)
(b) H0 : µ = 7250; H1 : µ < 7250 (1)
t14 = 4767−7250
√
3204/ 15
= 827.27
−2483
= −3(1)
So p < 0.005 (1) and so there is evidence to reject H0 and conclude that infected
individuals have a lower white blood cell count then the general population.(1)

6. (a) The appropriate pvalue is 0.3078 (1) and so we would conclude that there is
insufficient evidence against H0 and so on average, the daily wage rates in the
2 sectors are equal. (1)
(b) The ttest has to be used when we don’t know the population variances and
have to estimate them with the sample standard deviations (2).