Probability and Statistics for Engineers

GENG 200

Tutorial on Chapters 7-8-9

1
 GENG 200: Tutorial on Chapter 7 Exercise 1 (problem 7-20)

Exercise 1 (problem 7-20)

Suppose we have a random sample of size 2n from a population denoted by X, and E(X) = µ and V (X) = σ 2 .
Let
2n n
1 X 1X
X̄1 = Xi , and X̄2 = Xi ,
2n i=1 n i=1
be two estimators of µ.
Which is the better estimator of µ?
Explain your choice.

Solution: !
2n 2n

1 X 1 X 1
E X̄1 = E Xi = E (Xi ) = 2nµ = µ
2n i=1 2n i=1 2n
n
! n

1X 1X 1
E X̄2 = E Xi = E (Xi ) = nµ = µ
n i=1 n i=1 n
This means that X̄1 and X̄2 are unbiased estimators for µ.

σ2 σ2
The variances are V (X̄1 ) = 2n and V (X̄2 ) = n (chapter 5, expression 5-27);

Compare the MSE (only variance in this case since the bias is 0):
M SE(Θ̂1 ) σ 2 /2n n 1
= 2 = =
M SE(Θ̂1 ) σ /n 2n 2
Since both estimators are unbiased, examination of the variances would conclude that X̄1 is the better
estimator with the smaller variance.

Page 2 of 17
 GENG 200: Tutorial on Chapter 7 Exercise 2 (problem 7-21)

Exercise 2 (problem 7-21)

Let X1 , X2 , . . . , X7 denote a random sample from a population having mean µ and variance σ 2 .
Consider the following estimators of µ:
X1 + X2 + · · · + X7
Θ̂1 =
7
2X1 − X6 + X4
Θ̂2 =
2

(a)
Is either estimator unbiased?

Solution:
  1 1
E Θ̂1 = [E(X1 ) + E(X2 ) + · · · + E(X7 )] = 7µ = µ
  1 7 7
1 1
E Θ̂2 = [2E(X1 ) − E(X2 ) + E(X4 )] = [2µ + µ + µ] = 2µ = µ
2 2 2
Both Θ̂1 and Θ̂2 are unbiased estimates of µ since the expected values of these statistics are equivalent to
the true mean, µ.

(b)
Which estimator is best? In what sense is it best?

Solution:
Use here expression 5-27 from chapter 5:
  1 1 σ2
V Θ̂1 = 2 [V (X1 ) + V (X2 ) + · · · + V (X7 )] = 2 7σ 2 =
  71  7
 1
7
6 3σ 2
V Θ̂2 = 2 2 V (X1 ) + (−1) V (X2 ) + V (X4 ) = [4σ + σ 2 + σ 2 ] = σ 2 =
2 2 2
2 4 4 2
Since both estimators are unbiased, the variances can be compared to decide to select the better estimator.
The variance of Θ̂1 is smaller than that of Θ̂2 . Therefore Θ̂1 is the better estimator.

Calculate the relative efficiency of the two estimators.

Solution:
Relative efficiency can be computed as fraction of the variances since since the bias is 0:
M SE(Θ̂1 ) σ 2 /7 2
= 2 /2
=
M SE(Θ̂1 ) 3σ 21
Θ̂1 is is a more efficient estimator.

Page 3 of 17
 GENG 200: Tutorial on Chapter 7 Exercise 3 (problem 7-27)

Exercise 3 (problem 7-27)

Let X1 , X2 , . . . , Xn denote a random sample of size n with from a population mean µ and variance σ 2 .

(a)
Show that X̄ 2 is a biased estimator for µ2 .

Solution:

2
Using E X 2 = V (X) + [E(X)] we will  get:
n
!2 n
! " n
!#2  
n
!2 
2

1 X 1  X X 1 nσ 2 +
X
E X̄ = 2 E Xi = 2 V Xi + E Xi =
2
µ 
n i=1
n i=1 i=1
n i=1
 
2
= n12 nσ 2 + (nµ)
That is:

σ2
E X̄ 2 = + µ2
n
and therefore X̄ 2 is a biased estimator for µ2 .

(b)
Find the amount of bias in the estimator.

Solution:
σ2
Bias = E X̄ 2 − µ2 =


n

(c)
What happens to the bias as the sample size n increases?

Solution:
Bias decreases as n increases.

Page 4 of 17
 GENG 200: Tutorial on Chapter 7 Exercise 4 (problem 7-40)

Exercise 4 (problem 7-40)

Consider the probability density function
f (x) = c(1 + θx), −1 ≤ x ≤ 1

(a)
Find the value of the constant c.

Solution:
Z 1
θx2 1
    
θ θ
c(1 + θx)dx = c x +  = c 1 + − −1 + = 2c = 1
−1 2 −1 2 2
Therefore constant c should be equal to 0.5.

(b)
(b) What is the moment estimator for θ?

Solution:
First find theZexpectation of X:
1
1 x2 θx3 1
     
1 1 1 θ 1 θ θ
E(X) = x (1 + θx)dx = + = + − − =
2 2 2 3 2 2 3 2 3 3

−1 −1
Now applying method of moments we can set:
n
θ 1X
E(X) = = Xi = X̄
3 n i=1
⇒ θ̂ = 3X̄

(c)
Show that θ̂ = 3X̄ is an unbiased estimator for θ.

Solution: !
n n n
1X 1X 1Xθ 1 nθ
E(θ̂) = E 3 Xi =3 E(Xi ) = 3 =3 =θ
n i=1 n i=1 n i=1 3 n 3
That is E(θ̂) = θ which means that θ̂ = 3X̄ is an unbiased estimator for θ.

(d)
Find the maximum likelihood estimator for θ.

Solution:
Maximum likelihood estimators are not covered in the course.

Page 5 of 17
 GENG 200: Tutorial on Chapter 7 Exercise 5 (problem 7-55)

Exercise 5 (problem 7-55)

A random sample of 36 observations has been drawn from a normal distribution with mean 50 and standard
deviation 12. Find the probability that the sample mean is in the interval 47 ≤ X̄ ≤ 53. Is the assumption
of normality important? Why?

Solution:
From the Central Limit Theorem we know that the sample mean follows normal distribution with mean
√
µ and standard deviation
√ σ/ n. That is:
X̄ ∼ N (50, 12/ 36)

Then the probability that the mean is in the interval 47 ≤ X ≤ 53 can be found as follows:
 
47 − 50 X̄ − 50 53 − 50
PN (50,12/√36) (47 ≤ X̄ ≤ 53) = PN (0,1) √ ≤ √ ≤ √
12/ 36 12/ 36 12/ 36
− 32 ≤ Z ≤ 23 = PN (0,1) (Z ≤ 1.5) − PN (0,1) (Z ≤ −1.5)


= PN (0,1)

= 0.9331928 − 0.0668072 = 0.8663856

The assumption of normality is NOT important, because Central Limit Theorem states that with large
samples (n > 30), X̄ is approximately normally distributed.

Exercise 7 (problem 7-60)

Let X be a random variable with mean µ and variance σ 2 . Given two independent random samples of sizes
n1 and n2 , with sample means X̄1 and X̄2 , show that

X̄ = aX̄1 + (1 − a)X̄2 , 0<a<1

is an unbiased estimator for µ. If X̄1 and X̄2 are independent, find the value of a that minimizes the standard
error of X̄ .

Solution:
Applying the definition of the unbiased estimators we get that
E(X̄) = E(aX̄1 + (1 − a)X̄2 ) = aµ + (1 − a)µ = µ
That is X̄ is an unbiased estimator for µ

Parameter a should be chosen such that the variance of the estimator is minimal.
Find first the variance of X̄:

σ2 σ2
V (X̄) = a2 V (X̄1 ) + (1 − a)2 V (X̄2 ) = a2+ (1 − a)2
n1 n2
Then by solving equation:
σ2 σ2
   
d 1 1 1
V (X̄) = 2a − 2(1 − a) = 2σ 2 a + − =0
da n1 n2 n1 n2 n2
 
1. 1 1 n1 n2 n1
⇒a= + = =
n2 n1 n2 n2 (n1 + n2 ) n1 + n2
n1
a=
n1 + n2

Page 6 of 17
 GENG 200: Tutorial on Chapter 8 Exercise 8

Exercise 8
(a)
For a normal population with known V (X) = σ 2 . What is the confidence level for the interval
√ √
X̄ − 2.49σ/ n ≤ µ ≤ X̄ + 2.49σ/ n

Solution:
Parameter zα/2 = 2.49
Then
P (Z < 2.49) = Φ(2.49) = 1 − α/2
⇒ α/2 = 1 − Φ(2.49)
⇒ confidence level = (1 − α) = 1 − 2(1 − Φ(2.49)) = 2Φ(2.49) − 1
⇒ confidence level = 2 · 0.993 − 1 = 0.987 = 98.7%

(b)
For a normal population with known variance σ 2 . What value of zα/2 in the following equation
√ √
X̄ − zα/2 σ/ n ≤ µ ≤ X̄ + zα/2 σ/ n

gives 98% confidence?

Solution:
Confidence level = (1 − α) = 98% ⇒ α/2 = (1 − 0.98)/2 = 1%

zα/2 is defined as the point above which the area is α/2

(or −zα/2 is defined such that the area below it is α/2).

Therefore Φ(zα/2 ) = 1 − α/2 ⇒ zα/2 = Φ−1 (1 − α/2) = Φ−1 (0.99) = 2.32

Page 7 of 17
 GENG 200: Tutorial on Chapter 8 Exercise 9

Exercise 9
The diameter of holes for a cable harness is known to have a normal distribution with σ = 0.01 inch . A
random sample of size 10 yields an average diameter of 1.5045 inch.

(a)
Find the 99% two-sided confidence interval on the mean hole diameter.

Solution:
First we need to find the parameter zα/2
zα/2 = Φ−1 (1 − α/2) = Φ−1 (1 − (1 − 0.99)/2) = Φ−1 (0.995) = 2.58

Then the confidence interval will be formulated as follows: 0.5

σ σ α/2 α/2
X̄ − zα/2 √ ≤ µ ≤ X̄ + zα/2 √ 0.3
n n
0.01 0.01 0.1
1.5045 − 2.58 √ ≤ µ ≤ 1.5045 − 2.58 √
10 10 −zα/2 zα/2 z
⇒ 1.496 ≤ µ ≤ 1.512 −3.0 −2.0 −1.0 0.0 1.0 2.0 3.0

(b)
What happens to the length of the interval if the sample size is increased by a factor of a ( a ≥ 1)?

Solution:
√
By definition the length of the confidence interval = 2zα/2 σ/ n
If we increase the sample size by a factor a ( a ≥ 1). That is, we set n0 = a · n.
Then the new length will be

σ 1 σ
2zα/2 √ = √ 2zα/2 √
a·n a n
√
The new interval will become smaller and will be equal to the old one divided by the factor of a.

(c)
If the sample size n size is increased by a factor of four, by how much the length of the Confidence Interval
(CI) on µ will be reduced?

Solution:
Here we need to apply the findings from (b) and set the factor to a = 4.
Then we get that if we muptiply n by factor 4, the length of the interval will be divided by 2.

Page 8 of 17
 GENG 200: Tutorial on Chapter 8 Exercise 10 (problem 8-4)

Exercise 10 (problem 8-4)

A confidence interval estimate is desired for the gain in a circuit on a semiconductor device. Assume that
gain is normally distributed with standard deviation σ = 20.

(a)
Find a 95% CI for µ when n = 10 and X̄ = 1000.

Solution:
First we need to find the parameter zα/2
zα/2 = Φ−1 (1 − α/2) = Φ−1 (1 − (1 − 0.95)/2) = Φ−1 (0.975) = 1.96
Then the confidence interval will be formulated as follows:

0.5
σ σ
X̄ − zα/2 √ ≤ µ ≤ X̄ + zα/2 √
n n α/2 0.3 α/2
20 20
1000 − 1.96 √ ≤ µ ≤ 1000 + 1.96 √
10 10 0.1

⇒ 987.6039 ≤ µ ≤ 1012.396 −zα/2 zα/2 z

−3.0 −2.0 −1.0 0.0 1.0 2.0 3.0

(b)
Find a 95% CI for µ when n = 25 and X̄ = 1000.

Solution:
The parameter zα/2 remains the same as in (a) zα/2 = 1.96
Then in the CI formula we just need to correct the parameter n and we get:

20 20
1000 − 1.96 √ ≤ µ ≤ 1000 + 1.96 √
25 25
⇒ 992.16 ≤ µ ≤ 1007.84

c)
Find a 99% CI for µ when n = 10 and X̄ = 1000.

Solution:
Here we need to update the parameter zα/2
zα/2 = Φ−1 (1 − α/2) = Φ−1 (1 − (1 − 0.99)/2) = Φ−1 (0.995) = 2.58
Then the confidence interval will be formulated as follows:

20 20
1000 − 2.58 √ ≤ µ ≤ 1000 + 2.58 √
10 10
⇒ 983.709 ≤ µ ≤ 1016.291

(d)
Find a 99% CI for µ when n = 25 and X̄ = 1000.

Exercise 10 (problem 8-4) continued on next page. . . Page 9 of 17

 GENG 200: Tutorial on Chapter 8 Exercise 10 (problem 8-4)

Solution:
The parameter zα/2 remains the same as in (c) zα/2 = 2.58
The corrected CI (with n = 25) will be:

20 20
1000 − 2.58 √ ≤ µ ≤ 1000 + 2.58 √
25 25
⇒ 989.6967 ≤ µ ≤ 1010.303

(e)
How does the length of the CI’s computed above change with the changes in sample size and confidence
level?

Solution:
When n is larger, the CI will get narrower.
The higher the confidence level, the wider the CI.

Page 10 of 17
 GENG 200: Tutorial on Chapter 8 Exercise 11 (problem 8-8)

Exercise 11 (problem 8-8)

Following are two confidence interval estimates of the mean µ of the cycles to failure of an automotive door
latch mechanism (the test was conducted at an elevated stress level to accelerate the failure).
3124.9 ≤ µ ≤ 3215.7 3110.5 ≤ µ ≤ 3230.1

(a)
What is the value of the sample mean cycles to failure?

Solution:
The sample mean can be found as the middle of the Confidence Interval
X̄1 = 3124.9 + (3215.7 − 3124.9)/2 = 3170.3
X̄2 = 3110.5 + (3230.1 − 3110.5)/2 = 3170.3
The sample means are the same.

(b)
The confidence level for one of these CI’s is 95% and the confidence level for the other is 99%. Both CI’s are
calculated from the same sample data. Which is the 95% CI?
Explain why.

Solution:
Interval (1): was calculated with 95% Confidence because it has a smaller half-length, and therefore a
smaller confidence interval.
The 99% confidence level will make the interval larger.

Page 11 of 17
 GENG 200: Tutorial on Chapter 8 Exercise 12 (problem 8-13)

Exercise 12 (problem 8-13)

A manufacturer produces piston rings for an automobile engine. It is known that ring diameter is normally
distributed with σ = 0.001 millimeters. A random sample of 15 rings has a mean diameter of x̄ = 74.036
millimeters.

(a)
Construct a 99% two-sided confidence interval on the mean piston ring diameter.

Solution:
First we need to find the parameter zα/2
zα/2 = Φ−1 (1 − α/2) = Φ−1 (1 − (1 − 0.99)/2) = Φ−1 (0.995) = 2.58
Then the confidence interval will be formulated as follows:
σ σ
x̄ − zα/2 √ ≤ µ ≤ x̄ + zα/2 √
n n
0.001 0.001
74.036 − 2.58 √ ≤ µ ≤ 74.036 + 2.58 √
15 15
⇒ 74.03533 ≤ µ ≤ 74.03667

Page 12 of 17
 GENG 200: Tutorial on Chapter 8 Exercise 13 (problem 8-14)

Exercise 13 (problem 8-14)

The life in hours of a 75-watt light bulb is known to be normally distributed with σ = 25 hours. A random
sample of 20 bulbs has a mean life of x̄ = 1014 hours.

(a)
Construct a 95% two-sided confidence interval on the mean life.

Solution:
First we need to find the parameter zα/2
zα/2 = Φ−1 (1 − α/2) = Φ−1 (1 − (1 − 0.95)/2) = Φ−1 (0.975) = 1.96
Then the confidence interval will be formulated as follows:
σ σ
x̄ − zα/2 √ ≤ µ ≤ x̄ + zα/2 √
n n
25 25
1014 − 1.96 √ ≤ µ ≤ 1014 + 1.96 √
20 20
⇒ 1003.043 ≤ µ ≤ 1024.957

Page 13 of 17
 GENG 200: Tutorial on Chapter 8 Exercise 13 (problem 8-16)

Exercise 13 (problem 8-16)

Suppose that in Exercise 8-14 we wanted the error in estimating the mean life from the two-sided confidence
interval to be five hours at 95% confidence. What sample size should be used?

Solution:
The sample size can be estimating as:
z
α/2 σ
2
n=
E
where
zα/2 = Φ−1 (1 − α/2) = Φ−1 (1 − (1 − 0.95)/2) = Φ−1 (0.975) = 1.96,
E = 5 hours
σ = 25.

This will give us:

 2
1.96 · 25 2
= (1.96 · 5) = 96.04
5
Then the sample size should be:
n = 97

Page 14 of 17
 GENG 200: Tutorial on Chapter 8 Exercise 14 (problem 8-15)

Exercise 14 (problem 8-15)

A civil engineer is analyzing the compressive strength of concrete. Compressive strength is normally dis-
tributed with σ 2 = 1000(psi)2 . A random sample of 12 specimens has a mean compressive strength of
x̄ = 3250psi. (a) Construct a 95% two-sided confidence interval on mean compressive strength. (b) Construct
a 99% two-sided confidence interval on mean compressive strength. Compare the width of this confidence
interval with the width of the one found in part (a).

(a)
Construct a 95% two-sided confidence interval on mean compressive strength.

Solution:
First we need to find the parameter zα/2
zα/2 = Φ−1 (1 − α/2) = Φ−1 (1 − (1 − 0.95)/2) = Φ−1 (0.975) = 1.96
Then the confidence interval will be formulated as follows:
σ σ
X̄ − zα/2 √ ≤ µ ≤ X̄ + zα/2 √
n n
√ √
1000 1000
3250 − 1.96 √ ≤ µ ≤ 3250 + 1.96 √
12 12
⇒ 3232.108 ≤ µ ≤ 3267.892

b)
(b) Construct a 99% two-sided confidence interval on mean compressive strength.
Compare the width of this confidence interval with the width of the one found in part(a).

Solution:
Here we need to update the parameter zα/2
zα/2 = Φ−1 (1 − α/2) = Φ−1 (1 − (1 − 0.99)/2) = Φ−1 (0.995) = 2.58
Then the confidence interval will be formulated as follows:
√ √
1000 1000
3250 − 2.58 √ ≤ µ ≤ 3250 + 2.58 √
12 12
⇒ 3226.486 ≤ µ ≤ 3273.514
The 99% confidence level will make the interval larger.

Page 15 of 17
 2B

The heat evolved in calories per gram of a cement mixture is approximately normally distributed. The mean is
thought to be 100, and the standard deviation is 2. You wish to test H0: µ = 100 versus H1: µ ≠ 100 with a sample of
n = 9 specimens.

(a) If the acceptance region is defined as 98.5 ≤ x ≤ 101.5, find the type I error probability α
A:
a = 𝑃 𝑋 < 98.5 OR 𝑋 > 101.5 when 𝜇 = 100 = 𝑃 𝑋 < 98.5 | 𝜇 = 100) + 𝑃(𝑋 > 101.5 | 𝜇 = 100
æ X - 100 98.5 - 100 ö æ X - 100 101.5 - 100 ö
= Pçç < ÷÷ + Pç > ÷
è 2/ 9 2 / 9 ø çè 2 / 9 2 / 9 ÷ø
= P(Z < -2.25) + P(Z > 2.25)
= (P(Z <- 2.25)) + (1 - P(Z £ 2.25))
= 0.01222 + 1 - 0.98778
= 0.01222 + 0.01222 = 0.02444

(b) Find β for the case in which the true mean heat evolved is 103.
A:
b = P(98.5 £ X £ 101.5 | µ = 103)
æ 98.5 - 103 X - 103 101.5 - 103 ö
= Pçç £ £ ÷
è 2/ 9 2/ 9 2 / 9 ÷ø
= P(-6.75 £ Z £ -2.25)
= P(Z £ -2.25) - P(Z £ -6.75)
= 0.01222 - 0 = 0.01222
(c) Find β for the case where the true mean heat evolved is 105. This value of β is smaller than the one found in
part (b). Why?
A:
b = P(98.5 £ X £ 101.5 when µ = 105)
æ 98.5 - 105 X - 105 101.5 - 105 ö
= Pçç £ £ ÷÷
è 2 / 9 2 / 9 2 / 9 ø
= P(-9.75£ Z £ -5.25)
= P(Z £ -5.25) - P(Z £ -9.75)
=0-0
= 0.

The probability of accepting the null hypothesis when it is actually false is smaller in part c since the true
mean, µ = 105, is further from the acceptance region, mean a larger difference exists.

Page 16 of 17
A manufacturer is interested in the output voltage of a power supply used in a PC. Output voltage is assumed to be
normally distributed with standard deviation 0.25 volt, and the manufacturer wishes to test H0: µ = 5 volts against
H1: µ ≠ 5
volts, using n = 16 units.

a) The acceptance region is 4.85 ≤ 𝑥 ≤ 5.15. Find the value of α.

Answer:
a = P(𝑋£ 4.85 | µ = 5) + P(𝑋> 5.15 | µ = 5)
æ X -5 4.85 - 5 ö æ X - 5 5.15 - 5 ö
= Pçç £ ÷÷ + Pçç > ÷÷
è 0.25 / 16 0.25 / 16 ø è 0.25 / 16 0.25 / 16 ø
= P(Z £ -2.4) + P(Z > 2.4)
= P(Z £ -2.4) +(1 - P(Z £ 2.4))
= 2(1 - P(Z £ 2.4))
= 2(1 - 0.99180)
= 2(0.0082)
= 0.0164.

b) Find the power of the test (Power of the test is defined as 1- probability of error type II = 1-𝛽) for
detecting a true mean output voltage of 5.1 volts
Answer:
b = P(4.85 £𝑋£ 5.15 | µ = 5.1)
æ 4.85 - 5.1 X - 5.1 5.15 - 5.1 ö
= Pçç £ £ ÷÷
è 0.25 / 16 0.25 / 16 0.25 / 16 ø
= P(-4 £ Z £ 0.8) = P(Z £ 0.8) - P(Z £ -4)
= 0.78814 - 0 = 0.78814
1 - b = 0.21186

Page 17 of 17