Math 322 Probability and Statistical Methods

Chapter 9 Sample Estimation Problems

Classical Methods of Estimation

Point Estimation
A point estimate of some population parameter θ is a single value θ
of a statistic Θ
.
For example the value x of the statistic X computed from a sample of size n is a point
estimate of the population parameter µ .

 is said to be an unbiased estimator of the parameter θ if

Definition. A statistic Θ
( )
µ Θ = E Θ
 =θ .

Example 1. Show that X is an unbiased estimator of the parameter µ .

We need to show that E ( X ) = µ . That is,
 n 
 ∑ Xi  1 n nµ
E  i =1  = ∑ E ( Xi ) = =µ .
 n  n i =1 n
 
 
Example 2. S is an unbiased estimator of the parameter σ 2 .
2

Variance of a Point Estimator

Definition. If we consider all possible unbiased estimators of some parameter θ , the one with
the smallest variance is called the most efficient estimator of θ .

Example 1. For normal populations one can show that both X and  X are unbiased estimators
of the population mean µ , but the variance of X is smaller than the variance of 
X . Hence the

estimator X is more efficient than X .

Example 2. If X is a binomial random variable, show that

 = X is an unbiased estimator of P .
(a) P
n
Solution. E P ( )
 = E  X  = 1 E ( X ) = 1 .nP = P .
 
n n n

X + n /2
(b) P ' = is a biased estimator of P .
n+ n

1
Sonuc Zorlu Lecture Notes
Math 322 Probability and Statistical Methods

Solution.
 X + n /2
E ( P ') = E   =
1
E X + n /2 ( )
 n+ n  n+ n

=
n+
1
n
{E ( X ) + }
n /2 =
nP + n / 2
n+ n
≠P

Therefore P ' is a biased estimator of P .

Example 3. Let X 1 , X 2 , and X 3 be a random sample of size 3. We consider three estimators of

2 X1 + X 2 + X 3 ˆ X + 2X3 X + X2 + X3
µ given by θˆ1 = , θ2 = 1 and θˆ3 = 1 .
4 3 6
(a) Find out if θˆ1 , θˆ2 and θˆ3 are unbiased or not.
(b) If unbiased, which one is more efficient? Give reasons.

Solution. (a)
 2X + X2 + X3  1
( )
E θˆ1 = E  1
 4  = 4 { E ( 2 X 1 ) + E ( X 2 ) + E ( X 3 )}



2µ + µ + µ
= {2 E ( X ) + E ( X 2 ) + E ( X 3 )} =
1
=µ
4 4
Thus, θˆ1 is an unbiased estimator of µ .

 X + 2X3  1
( )
E θˆ2 = E  1
 3  = 3 { E ( X 1 ) + 2 E ( X 3 )}



µ + 2µ
= =µ
3
Thus, θˆ2 is also an unbiased estimator of µ .

 X + X2 + X3  1
( )
E θˆ3 = E  1
 6  = 6 { E ( X 1 ) + E ( X 2 ) + E ( X 3 )}



µ+µ+µ µ
= = ≠µ
6 2
Thus, θ 3 is not an unbiased estimator of µ .
ˆ

(b) To check the efficiency we need to compute their variances.

 2X + X 2 + X3
( )
Var θˆ1 = Var  1
 4
 1
 = 16 {4Var ( X 1 ) + Var ( X 2 ) + Var ( X 3 )}

=
1
16
{ }
4σ 2 + σ 2 + σ 2 = σ 2
6
16

2
Sonuc Zorlu Lecture Notes
Math 322 Probability and Statistical Methods

 X + 2X3
( )
Var θˆ2 = Var  1
 3
 1
 = 9 {Var ( X 1 ) + 4Var ( X 3 )}

1
{ }
= σ 2 + 4σ 2 = σ 2
9
5
9

( ) ( )
Since Var θˆ1 < Var θˆ2 , θˆ1 is more efficient than θˆ2 .

Interval Estimation

Suppose that the response variable X is normally distributed with unknown mean µ and
unknown variance σ 2 . The goal of interval estimation is to obtain a probability interval for
the true mean ( µ ) based on a random sample of n observations.
Recall that sampling from the population may be considered as an experiment. When n is
large, any particular sample of size n is one of many possible samples of size n that we
could have obtained. It follows that the sample mean ( X ) obtained from any particular
sample of size n is one of many possible sample means we could have obtained from the
same size samples.

According to the central limit theorem:

• Sample means are distributed normally if X is normal or if n is greater than about 30.
• The sample means are centered around the true mean. That is, the expected value of
X ( µ X ) is equal to µ , meaning that X is an unbiased estimate of µ .
σ2
• The variance in the distribution of means is .
n

When σ 2 is unknown, we must substitute the unbiased estimate of this quantity. An unbiased
estimate of σ 2 is given by
2
n n
 n 
∑ (Xi − X )
2
n ∑ X i
2
−  ∑ Xi 
 i =1  .
S 2 = i =1 or S 2 = i =1
n −1 n ( n − 1)

When σ 2 is unknown, the sample means follow a -distribution rather than a -distribution. As
the sample size increase, however, the -distribution converges to a -distribution.

3
Sonuc Zorlu Lecture Notes
Math 322 Probability and Statistical Methods

Estimating the Mean

Confidence Interval for µ ( σ known). If X is the mean of a random sample of size

n from a population (need not be normal) with known variance σ 2 , then a (1 − α )100%
confidence interval for µ is given by
X −e ≤ µ ≤ X +e
σ σ
where e = zα / 2 , zα / 2 is the z -value leaving an area of α / 2 to the right and is the
n n
standard error of X .

Example 1. An electrical firm manufactures light bulbs that have a length of life that is
approximately normally distributed with a standard deviation of 40 hours. If a sample of 30
bulbs has an average life of 780 hours, find a 96% confidence interval for the population mean
µ of all bulbs produced by this firm.

Solution. Given σ = 40hours, n = 30, X = 780hours, 1 − α = 0.96, α / 2 = 0.02 , From table

A.4 z0.02 = 2.05 and the confidence interval for µ is
40 40
X − e < µ < X + e ⇒ 780 − 2.05 < µ < 780 + 2.05
30 30

Thus, the 96% confidence interval is 765.1 < µ < 794.9 .

Theorem. If X is used as an estimate of µ , we can be (1 − α )100% confident that the error

will not exceed a specified amount e when the sample size is
z σ
2

n =  α /2  .
 e 

Example 2. The average zinc concentration recovered from a sample of zinc measurements is
found to be 2.6 grams per milliliter, where the population standard deviation is 0.3. How large a
sample is required if we want to be 95% confident that our estimate of µ is off by less than
0.05?

Solution. Given σ = 0.3, 1 − α = 0.95, α / 2 = 0.025, e = 0.05 , From table A.4 z0.025 = 1.96 and
 zα / 2σ   (1.96 )( 0.3) 
2 2

n=  =  = 138.3 ⇒ n = 138 .

 e   0.05 

Example 3. How large a sample is needed in Example 1, if we wish to be 96% confident that
our sample mean will be within 10 hours of the true mean?

4
Sonuc Zorlu Lecture Notes
Math 322 Probability and Statistical Methods

We are attempting to estimate the mean of a population when the variance σ 2 is unknown. If we
have a random sample chosen from a normal population , then the random variable
X −µ
T=
s/ n
has a student’s t-distribution with degrees of freedom υ = n − 1 .

Confidence Interval for µ ( σ unknown). ). If X and S are the mean and standard
deviation of a random sample of size n from a normal population with unknown variance
σ 2 , then a (1 − α )100% confidence interval for µ is given by
X −e ≤ µ ≤ X +e
S
where e = tα / 2 , tα / 2 is the t -value with υ = n − 1 degrees of freedom, leaving an area of
n
α / 2 to the right.

Example 1. The heights of a random sample of 50 college students showed a mean of 174.5cm
and a standard deviation of 6.9cm.
(a) What can we assert with 98% confidence about the possible size of our error if we
estimate the mean height of all college students?
(b) Construct a 98% Confidence Interval for the mean height of all college students.

Solution. (a) Given S = 6.9cm, n = 50, X = 174.5cm, 1 − α = 0.98, α / 2 = 0.01 , From table A.3
2.423 + 2.390
t0.01,49 = = 2.406 ,the size of the error is computed as
2
S 6.9
e = tα / 2 = 2.406 ≈ 2.35 .
n 50

(b) The confidence interval for µ is

X − e < µ < X + e ⇒ 174.5 − 2.35 < µ < 174.5 + 2.35 .

Thus, the 98% confidence interval is 172.15 < µ < 176.85 .

5
Sonuc Zorlu Lecture Notes
Math 322 Probability and Statistical Methods

Estimating the Variance

Confidence Interval for σ 2 . If S 2 s the variance of a random sample of size n from

normal population, then a (1 − α )100% confidence interval for σ 2 is given by
( n − 1) S 2 ( n − 1) S 2
<σ < 2

χα2 / 2 χ12−α / 2
where χα2 / 2 and χ12−α / 2 are χ 2 -values with υ = n − 1 .

Example 1. The following are the weights in decagrams of 10 packages of grass seed distributed
by a certain company.
46.4 46.1 45.8 47.0 45.9
46.1 45.8 46.9 45.2 46.0

Find a 95% confidence interval for the variance of all such packages of grass seeds, assuming
normal population.

Solution. Given n = 10,1 − α = 0.95, α / 2 = 0.025 and the sample variance is computed as
2
n
 n 
n∑ X −  ∑ X i 
2
10 ( 21273.12 ) − ( 461.2 )
i 2

S =
2 i =1  i =1  = = 0.286 , and from table A.5 we have
n ( n − 1) 10.9
χ 0.025
2
= 19.023 and χ 0.975
2
= 2.7 . Thus, the 95% confidence interval for the variance is

( n − 1) S 2 ( n − 1) S 2 9 ( 0.286 ) 9 ( 0.286 )
<σ < 2
⇒ <σ2 < ⇒ 0.135 < σ 2 < 0.953 .
χα / 2
2
χ 2
1−α / 2 19.023 2.7

Example 2. A random sample of 20 students obtained a mean of X = 72 and a variance of

S 2 = 16 on a college placement tests in maths. Assuming the scores are approximately normally
distributed, construct a 98% confidence interval for the true variance σ 2 .

Solution. Given n = 20, X = 72, S 2 = 16, υ = 19, 1 − α = 0.98,α / 2 = 0.01 . From table A.5,
χ 0.01,19
2
= 36.191 and χ 0.99,19
2
= 7.633 . Therefore the confidence interval for σ 2 is
( n − 1) S 2 ( n − 1) S 2 (19 )16 (19 )16
<σ <2
⇒ <σ2 < ,
χα2 / 2 χ12−α / 2 36.191 7.633

That is, 8.40 < σ 2 < 39.83 .

6
Sonuc Zorlu Lecture Notes
Math 322 Probability and Statistical Methods

Exercises

Exercise 1. An electrical firm manufactures light bulbs that have a length of life which is
approximately normally distributed with a standard deviation of 40 hours. If a sample of 30
bulbs fas an average life of 780 hours find a 96% confidence interval for the population mean µ
of all bulbs produced by this firm.

Exercise 2. How large a sample is needed in Q1 if we wish to be 96% confident that our sample
meanwill be within 10 hours of the true mean?

Exercise 3. The weights of a random sample of 30 college students showed a mean of 74.5 cm
and a standard deviation of 5.9 cm.
(a) What can we assert with 98% confidence about the possible size of our errror if we
estimate the mean weight of all college students?
(b) Construct a 98 % confidence interval for the mean weight of all college students.

Exercise 4. Given a random sample of size 24 from a normal population, find k such that
P ( −2.069 < T < k ) = 0.965 .

Exercise 5. A random sample of size n=12 from a normal population has the mean x = 27.8 and
s 2 = 3.24 . Can we say that the given information supports the claim that the mean of the
population if µ = 28.5 ?

Exercise 6. Find k such that P ( k < T < −1.761) = 0.045 for a random sample of size 15 selected
from a normal population.

Exercise 7. A maker of certain brand of low fat cereal bars claims that their average saturated fat
content is 0.5 gr. In a random sample of 8 cereal bars of this brand the saturated fat content was
0.6,0.7,0.3,0.4,0.5,0.4 and 0.2. Would you agree with the claim? (If the computed t -value falls
between −t0.05 and t0.05 , teh maker is satisfied with his claim)

Exercise 8. The average zinc concentration recovered from a sample of zinc measurements is
found to be 2.6 grams per mililiter, where the population standard deviation is 0.3. How large a
sample is required if we want to be 95% confident that our estimate of µ is off by less than
0.05?

Exercise 9. A random sample of 12 graduates of certain secreterial school typed an average of

79.3 words per minite witha standard deviation of 7.8 words per minute. Assuming a normal
distribution for the number of words typed per minute, find a 95% confidence interval for the
average number of words typed by all graduates of this school.

Exercise 10. A paint manufacturer wants to determine the average drying time of a new wall
paint. In 12 test areas of equal size painted by new product, he measured the drying times
12 12
X 1 , X 2 ,.... X 12 in minutes. He obtained the following data: ∑X
i =1
i = 798 and ∑X
i =1
i
2
= 53524 .

Construct a 95% confidence interval for the true mean µ .

7
Sonuc Zorlu Lecture Notes
Math 322 Probability and Statistical Methods

Exercise 11. The scores of a random sample of 20 students have a variance s 2 = 16 on a college
placement test in mathematics. Assuming the scores are normally distributed, construct a 95%
confidence interval for σ 2 .

Exercise 12. (a) If X is a normal random variable with µ = 20 and σ =4, find P ( X < 13) .
(b) Find P (T13 < 2.650 ) . ( T13 is a t random variable with ν = 13 degrees of freedom)
(c) If in a random sampling problem the sample size is n=11 and σ = 0.8 , find
 n

P  ∑ ( X i − X ) 2 > 3.1136  .
 i =1 

Exercise 13. The error made in measuring the heights of plants by a certain electronic instrument
is accepted to be normally distributed with unknown mean µ and unknown variance σ 2 . Five
measurements made by this instrument are checked later and found to have the following errors
(in mm.s):
-3, 2, 4, 1, -2
Find a 98% confidence interval for the mean error µ .

Exercise 14. A random sample of 25 automobile owners shows that an automobile is driven on
the average 20.000 km per year with a standard deviation of 1000 km. Construct a 99%
confidence Interval for the average number of kilometers an automobile is driven annually.

Exercise 15. The scores on a level determination test given to freshmen for the past five years
are approximately normally distributed with a mean µ = 62 . Construct a 98 % confidence
interval for the population variance σ 2 , if a random sample of 20 students who take this test this
year obtain a value of S 2 = 12 .

Exercise 16. A manufacturer of car batteries claim that his batteries will last on average 3 years
with a variance of 1 year. If 5 of these batteries have lifetimes of 2.9,2.4,1.8,3.5 and 4.2 years,
(a) Construct a 96 % confidence interval for σ 2 .
(b) Decide whether the manufacturer’s claim that σ 2 =1 is valid or not.

 = X + n / 2 where X is a Binomial random variable. Show that P

Exercise 17. Define P  is a
n+ n
Biased Estimator of P .

Exercise 18. (a) The weights (in grams) of certain soft drink bottles (filled) have a standard
deviation of σ = 5 gr.s. A random sample of 16 bottles were tested and gave an average of
X = 290 gr.s. Find a 90 % confidence interval for the population mean weight µ .
(b) If X − µ ≤ 1 , what should be our minimal sample size instead of 16 with 90 %
confidence level?

8
Sonuc Zorlu Lecture Notes