CHP8 – Practice Questions

Q1) An experiment was carried out to find out the number of hours that
actuarial students spend watching television per week. It was discovered that
for a sample of 10 students, the following times were spent watching
television:
8, 4, 7, 5, 9, 7, 6, 9, 5, 7
Calculate a symmetrical 95% confidence interval for the mean time an actuarial
student spends watching television per week. Write down the assumptions
needed to find this confidence interval.

ANS –
X =6.7 , n=10 , S=1.703

Assuming that the data is taken from a normal distribution, we know that :
X−μ
N (0 , 1)
S/ √ n

( )
X −μ
∴ P −1.96< <1.96 =0.95
S
√n
P (−1.96 ×0.5385< 6.7−μ< 1.96× 0.5385 ) =0.95
P (−1.0555<6.7−μ<1.0555 ) =0.95
P ( 5.6445< μ<7.7555 ) =0.95

Therefore required 95% CI is (5.6445,7.7555)

Q2) A researcher investigating attitudes to Sunday shopping reports that, in a

sample of 8 interviewees, 7 were in favour of more opportunities to shop on
Sunday. Use the binomial distribution to calculate an exact 95% confidence
interval for the underlying proportion in favour of this idea.
ANS-
We have one observation, where 7 people were in favour of shopping on
Sunday. ∴ X=7 ,n=8
To calculate an exact 95% CI :
8 x

∑ b ( r ; n , θ1 )=0.025∧∑ b ( r ; n , θ2 ) =0.025
n= x n=0

Here b ( r ; n ,θ ) represents the probability P ( X=r ) , where X Bin ( n , θ )

Solving the first equation:

8 θ1 ( 1−θ 1 ) +θ 1=0.025 → θ1 [ 8−8θ 1+θ 1 ]=0.025

7 1 8 7

7 8 7
→ θ1 ( 8−7 θ 1 )=0.025 → 7 θ1−8 θ1 + 0.025=0

BY trial∧error method ,using table mode of calculator ,θ 1=0.4735

Solving the second equation :

8 8
1−θ 2=0.025 → θ2=0.975
θ2=0.99684

Therefore the required exact 95 % symmetrical confidence interval for θis

givenby :

( θ1 , θ2 ) ≡(0.4735 , 0.99684)

Q3) An opinion poll of 1,000 voters found that 450 favoured Party P. Calculate
an approximate 99% confidence interval for the proportion of voters who
favour Party P. Comment on the likelihood of more than 50% of the voters
voting for Party P in an election.
ANS –
Let X be the no. of voters that favoured Party P. X Bin ( 1000 , p )
For a binomial distribution the MLE of parameter p is given by
X 450
^p= = =0.45
n 1000

By normal approximation of binomial, (√ X −np

n ^p ( 1− ^p ) ) N ( 0 , 1)

We know that P (−2.5758< Z< 2.5758 )=0.99

(
∴ P −2.5758<
( X −np
)
√n ^p ( 1− ^p ) )
< 2.5758 =0.99

P (−40.5228< 450−1000 p< 40.5228 ) =0.99

P ( 409.4772<1000 p< 490.5228 )=0.99
P ( 0.4095< p<0.4905 )=.99

Since p is not a Random Variable, we will remove the probability symbol, and
hence we get our approximate symmetrical 99% CI for p  (0.4095,0.4905)
Q4) Two inspectors carry out property valuations for an estate agency. Over a
particular week they each go out to similar properties. The table below shows
their valuations (in £000s):
A 102 98 93 86 92 94 89 97

B 86 88 92 95 98 97 94 92 91
(i) (a) Make an informative plot of these figures.
(b) Use the diagrams from part (i)(a) to comment on the possible assumption
of equal variance for the two underlying populations.
(c) With the equal variance assumption of (b), calculate a 95% confidence
interval for this common variance.
(d) Calculate a 95% confidence interval for the mean difference between the
valuations by A and B, and comment briefly on the result.
ANS –

Chart Title
105

100

95

90

85

80

75
0 1 2 3 4 5 6 7 8 9 10

Series 1 Series 2
i)a)

b)The spread is almost the same for both A & B and hence the assumption of
equal variances can be considered. (WE don’t have large no. of data points to
confirm this)
c) We have : n A =8 , n B=9 , X A=93.875 , S A =5.1113 , S 2A=26.124 ,
 2
X B=92.56 , S B=3.94 , S B=15.5278
Therefore the pooled variance is given by :

2
S =
[(n A −1 ) S2A + ( nB −1 ) S 2B ]
=
( 8−1 ) 26.124+ ( 9−1 ) 15.5278
=20.473
P
n A + nB −2 8+9−2

( n−1 ) S2 2 ( n P−2 ) S2P 2 15 S 2P 2

We know that χ n −1 → χ n −2 → χ 15
σ2 σ 2P σ 2P P

( )
2 2
15 S P 2 15 S P
P <σ <
P =0.95
χ 20.025 ,15 χ 20.975 ,15

P ( 307.095
27.49
<σ <
6.262 )
307.095 2
=0.95
P

P ( 11.712< σ 2P< 49.041 ) =0.95

Lets remove the Probability symbol because variance is not a RV.

Then our required symmetrical 95% CI is (11.712,49.041)

d) X A − X B=93.875−92.56=1.315 , Variance isunknown∧equal

n A =8 , n B=9 , S P=4.5247 , t 0.025 ,15=2.131

Therefore the 95% CI is given as : ( X A − X B)± t0.025 , 15 × S P

√ 1 1
+
n1 n 2

 1.315 ± 2.131× 4.5247 × 0.4859→ 1.315 ± 4.685

 (-3.37,6)
ii) The estate agency employing the inspectors decides to test their valuations
by sending
them each to the same set of eight houses, independently and without
knowledge that
the other is going. The resulting valuations (in £000s) follow:

(a) Make an informative plot of these figures.

(b) Calculate a 90% confidence interval for the mean difference between
valuations by A and B, and comment briefly on the result.
ANS –
a) This data is paired, that’s why we will look at the differences between data
points ( Di )
Di=D A −D B={2, 2 ,−9 ,3 , 7 ,−1 , 5 , 5}
b) D ±t 0.05 , n−1 ×
√ S 2D
n √
→1.75 ± t 0.05 ,7 ×
24.893
8
→ 1.75± 1.895 ×1.7638 → 1.75± 3.3424

So the required 90% CI is given by (-1.5924,5.0924)

Q5)

−λ x
f ( X ; λ )=λ e ; x >0 , n=20

i)(a)¿ find MLE of λ , we will differentiate L ( λ ) w . r .t λ∧find the value of λ

at which L ( λ ) ismaximum .
20
L ( λ )=∏ ( λ e−λ x )=λ 20 e−λ∑ x i

i=1

∂ 20
logL ( λ )=20 logλ−∑ x i λ→ ( logL ( λ ) ) = −∑ x i
∂λ λ

{ }
2
∂
2
( logL ( λ ) )= −202
< 0 ; thus , log−likelihood is maximum atfirst derivative=0
∂λ λ
20
→ =174 → ^λ=0.11494
^λ

−1 −1 ^λ 2
CRLB= = = =0.000661

( ) ( )
(b) ∂
2
−20 20
E ( logL ( λ ) ) E 2
∂ λ2 λ

(c) ^λ N ( λ , CRLB ) → ^λ N ( λ , 0.000661 )

√
Therefore a 95% CI for λ is givenby ^λ ± 1.96 0.000661 → 0.11494 ±0.05039
1

Ans – (0.06455,0.16533)
ii) X exp ( λ ) → ∑ X i Ga ( 20 , λ ) → 2nλ X χ 22 n → 40 λ X χ 240

( )
2 2
χ 0.025 ,40 χ 0.975 ,40
therefore the 95 % interval is given by : < λ<
40 X 40 X

(
24.43 59.34
)
 40 × 8.7 < λ< 40 × 8.7 →(0.0702< λ <0.170 51)

This CI is narrower than the one in i)c) because this is calculated based on exact
result whereas the earlier was an approximation.

Q6) Heights of males with classic congenital adrenal hyperplasia (CAH) are
assumed to be normally distributed. Determine the minimum sample size to
ensure that a 95% confidence interval for the mean height has a maximum
width of 10cm, if:
(i) a previous sample had a standard deviation of 8.4 cm
(ii) the population standard deviation is 8.4 cm.
ANS-
Let X= height of males with CAH, X N ( μ , σ 2 )
X has a 95 % confidence interval with maximum widthof 10 cm

i)A previous sample had a standard deviation of 8.4cm, S=8.4

S
Then the 95% CI is given by X ± t 0.025, n−1
√n
t 0.025 , n−1 S
Thus the width is given by 2 ×
√n
8.4
2 ×t 0.025 ,n−1 × < 10→ t 0.025 ,n−1 <0.595238 √ n
√n
Thus the minimum such value of n is 14.
So sample size should be a minimum of 14 for the required 95% CI
ii)population standard deviation is 8.4cm σ =8.4
Then the 95 % CI is given by X ±1.96 σ / √ n
1.96 σ 8.4
Therefore width is given by 2 × <10 3.92 × <10
√n √n
√ n>3 .2 928 →n> 10.84253184 → theminimum value of n should be 11

Q7) 8.7 (i) A sample value of 2 is obtained from a Poisson distribution with
mean μ . Calculate an exact two-sided 90% confidence interval for μ .
(ii) A sample of 30 values from the same Poisson distribution has a mean of 2.
Use these data values to construct an approximate 90% confidence interval for
μ
ANS-
i)X=2 (sample value),
thus ¿ find a 90 % CI , we need μ1 ∧μ 2 such that : P ( μ 1< μ< μ 2 )=0.9

for an exact 90 % two−sided interval , we have the following two equations :

∞ x

∑ p ( r ; μ1 ) =0.05∧∑ p ( r ; μ2 ) =0.05
n= x n=0

Solving the first equation , we get :1−( e−μ + μ1 e−μ

1
1 ) =0.05
1

0.95=( e− μ + μ 1 e−1 μ )
1 1

By trial and error method, the value of μ1 is0.36

( )
2 −μ2
μ2 e
Solving the second equation , we get : e− μ + μ 2 e− μ +
2 2
=0.05
2

By trial and error method, the value of μ2 is6.3

Thus the required 90% CI is (0.36,6.3)
ii)n=30, X =2
For a poisson distribution , ^λ= X
^λ− λ
Thus we have : N ( 0 , 1)

√ ^λ
n

(
So , we have P −1.6449<
2−λ
0.2582 )
<1.6449 =0.9

P (−0.4247<2−λ< 0.4247 )=0.9→ P ( 1.5753< λ<2.4247 )=0.9

Thus , the required interval is(1.5753 , 2.4247)

Q8) 8.8 An office manager wants to analyse the variability in the time taken for
her typists to complete a given task. She has given seven typists the task and
the results are as follows (in minutes):
15, 17.2, 13.7, 11.2, 18, 15.1, 14
The manager wants a 95% confidence interval for the true standard deviation
of time taken of the form (k,∞ ¿Calculate the value of k .
ANS-
n=7, X =14.886 , S=2.269
( n−1 ) S 2 2 30.89 2
We have , 2
χ n−1 → χ6
σ σ2

P ( χ 20.025 , 6< r ) =0.95 → P ( χ 20.025, 6> r ) =5 % → r=12.59

Now we have : P ( 30.89

σ
2 )
<12.59 =0.95 → P ( σ > 1.5664 ) =0.95

Since σ is not a RV , we removethe probability sign∧hence get our 95 % CI of the

form ( k , ∞ ) , wherek =1.5664

Q9) The amounts of individual claims arising under a certain type of general
insurance policy are known from past experience to conform to a lognormal
distribution in which the standard deviation equals 1.8 times the mean. An
actuary has found that the lower and upper limits of a 95% confidence interval
for the mean claim amount are £4,250 and £4,750. Evaluate the lower and
upper limits of a 95% confidence interval for the lognormal parameter μ .
ANS –
 √ [( σ 2
) × (e ] ( )
2 2
σ
μ+ μ+
X logN ( μ , σ ) , Also : −1 ) =1.8 e
2
2 2 σ 2
e

1.8=√ eσ −1 → eσ −1=3.24 → 4.24=e σ → σ 2=1.4446

2 2 2

The 95 % CI for ( e ) is ( 4250,4750 ) → (8.355< μ+ <8.4659 )

2
σ 2
μ+
2 σ
2

(7.6327< μ< 7.7436)

Q10) A general insurance company is debating introducing a new screening

programme to reduce the claim amounts that it needs to pay out. The
programme consists of a much more detailed application form that takes
longer for the new client department to process. The screening is applied to a
test group of clients as a trial whilst other clients continue to fill in the old
application form. It can be assumed that claim payments follow a normal
distribution. The claim payments data for samples of the two groups of clients
are (in £100 per year):
Without screening 24.5 21.7 35.2 15.9 23.7 34.2 29.3 21.1 23.5 28.3
With screening 22.4 21.2 36.3 15.7 21.5 7.3 12.8 21.2 23.9 18.4
(i) (a) Calculate a 95% confidence interval for the difference between the mean
claim amounts.
(b) Comment on your answer. [6]
(ii) (a) Calculate a 95% confidence interval for the ratio of the population
variances.
(b) Hence, comment on the assumption of equal variances required in part (i).
[4]
(iii) Assume that the sample sizes taken from the clients with and without
screening are always equal to keep processing easy. Calculate the minimum
sample size so that the width of a 95% confidence interval for the difference
between mean claim amounts is less than 10, assuming that the samples have
the same variances as in part (i). [3]
ANS –
i)a) X 1 =25.74 , S 1=6.0156 , X 2 =20.07 , S 2=7.6443 , X 1−X 2=5.67
2
S P=47.3114→ S P =6.8783

Thus, required interval is given by : ( X 1−X 2 ) ± t0.025 ,18 × S P ×

√ 1 1
+
n1 n2

→ 5.67 ±2.101 ×6.8783 × 0.4472→ 5.67 ± 6.463→ (−0.793 , 12.133)

b) Here we have assumed equal variances of both populations. The screening

has had a positive effect on the claim amount payable, which has reduced most
of the times.
ii)a) For the ratio of population variances, we use the F distribution.
2
S1

( )
2 2 2 2 2
σ1 S 1 /S 2 σ 1 S 1
F n −1 ,n −1 → P < < × F 0.025 ,9 , 9 =0.95
S 22 1 2
F0.025 , 9 ,9 σ 22 S 22
2
σ2

( )
2 2
0.6193 σ 1 σ1
→P < 2 <0.6193 × 4.026 =0.95 →(0.154< 2 <2.493)
4.026 σ 2 σ2
b)Hence , the assumption of equal variances in i) is correct because 1 is present
2
σ1
in the 95% CI of 2 .
σ2

iii) Since maximum width is 10, we have :

2 ×t 0.025 ,2 n−2 × S P ×
√ 1 1
√
+ <10 t 0.025 ,2 n−2 ×
n n
2
n
<0.727

t 0.025 ,2 n−2< 0.514 × √ n → Least such n is16