4 Introduction to Estimation

1. Point Estimation and Interval Estimation

2. Interval Estimation of Normal Population Mean: Variance Known
3. Determining the Sample Size

Section 4.1 Point Estimation and Interval Estimation

In point estimation we make use of a random variable depending on samples
drawn from a population, called a point estimator, to compute a value that
serves as a point estimate of a population parameter.

Example 1
can be a point estimator of the population mean . The actual value x , the
sample mean obtained from a particular sample, is a point estimate of .

S2 can be a point estimator of the population variance 2. The actual value s2,
the sample variance obtained from a particular sample, is a point estimate of 2.

If the expectation of the point estimator is equal to the population parameter

being estimated, the point estimator is said to be an unbiased estimator of the
population parameter. The value of an unbiased estimator of a population
parameter is an unbiased estimate of the population parameter.

We know from Section 3.3 that

E( X )  μ and E(S2)  2.
Therefore X is an unbiased estimator of  and S2 is an unbiased estimator of
2.

The absolute value of the difference between a point estimate and the
population parameter to be estimated is called the sampling error. The
sampling error is due to the use of a subset of the population instead of the
entire population to collect information.

Example 2
x
|  | is the sampling error when using x to estimate the population mean .
|s  | is the sampling error when using s to estimate the population standard
deviation .
This study source was downloaded by 100000787038097 from CourseHero.com on 12-30-2021 20:55:36 GMT -06:00
STAT0302 ch4 1
https://www.coursehero.com/file/10046339/Lecture-4/
 In interval estimation, a random interval (depending on the sample drawn) is
constructed around a point estimator such that it contains the true value of the
population parameter with probability 1  , where   0. This interval is called
a confidence interval and its two endpoints are called the upper and lower
confidence limits. 1   is the confidence level (or confidence coefficient).

Section 4.2 Interval Estimation of Normal Population Mean: Variance

Known
Assumption
The population follows a normal distribution with mean  and known variance
 2.

Let X be the sample mean of a sample taken from the population. Let n be the
sample size. Then X follows N(,  2  n).

Area Area
/2 /2

Area  1  

 x

Conclusion The 100(1  )% confidence interval for the population mean  is

x  z   
  /2 , x  z /2 
 n n 
where x is the sample mean, n is the sample size and z is the number such
that
P(Z  z)  

for a standard normal random variable Z.

This study source was downloaded by 100000787038097 from CourseHero.com on 12-30-2021 20:55:36 GMT -06:00
STAT0302 ch4 2
https://www.coursehero.com/file/10046339/Lecture-4/
 [Proof] follows N(0,1).
√
)
√ √

 )
√

 )
√

 ( )
√ √
 1- - =1 -

That is, the (random) interval

X z   
  /2 , X  z /2 
 n n 
will contain  with probability 1  .

 The midpoint of the confidence interval is . x

 The length of the confidence interval is 2 , which does not depend on
√
x.
 The value of z can be obtained from the last row of the t distribution table.
(Do not use the standard normal distribution table if the t distribution table is
provided.)

Does there exist a 100% confidence interval? Yes, it is just (, ) and is
absolutely useless.

If we want to reduce the length of a confidence interval, we may

 lower the confidence level, or
 increase the sample size.
This study source was downloaded by 100000787038097 from CourseHero.com on 12-30-2021 20:55:36 GMT -06:00
STAT0302 ch4 3
https://www.coursehero.com/file/10046339/Lecture-4/
 Example 1 A publishing company has just published a new college textbook.
Before the company decides the price at which to sell his book, it wants to know
the average price of all such textbooks in the market. The research department
at the company took a random sample of 10 such textbooks and collected
information on their prices. This information produced a mean price of $48.40.
It is known that the standard deviation of the prices of all such textbooks is
$4.50. Assuming that the population is normal, construct a 90% confidence
interval for the mean price of all such college textbooks.

[Solution] From the given information, n  10, x  48.40,   4.50 and 1  

 90%.
Hence   0.1,   2  0.05,

(1.423) = 2.34
√ √

Since 48.40  2.34  50.74 and 48.40  2.34  46.06, the 90% confidence
interval for the mean price of all such college textbook is ($46.06, $50.74).

Example 2 Suppose that in a large city the annual income of a randomly

selected real estate agent is normally distributed with a standard deviation of
$7,500. A random sample of 16 real estate agents shows a mean annual income
of $52,000. Determine the 99% confidence interval estimate of the mean annual
income of all real estate agents in the city.

[Solution] n  16. x  52000.   7500. 1    99%,   0.01,   2  0.005.

(1875) = 4830
√ √

The 99% confidence interval estimate of the mean annual income of all real
estate agents in the city is
($52000  $4830, $52000  $4830),
that is,
($47170, $56830).

This study source was downloaded by 100000787038097 from CourseHero.com on 12-30-2021 20:55:36 GMT -06:00
STAT0302 ch4 4
https://www.coursehero.com/file/10046339/Lecture-4/
 Section 4.3 Determining the Sample Size

We call the half of the length of a confidence interval the margin of error.
Suppose that we want, at the confidence level of 1  , the margin of error not
to be more than a given number E. Since

the sample size should be greater than or equal to

Example Suppose the bureau of the census of a city wants to estimate the mean
family annual income for all families in the city at the 99% confidence level. It
is known that the standard deviation  for the family annual income is 60
thousand dollars. How large a sample should the bureau select if it wants the
margin of error not to be more than 5 thousand dollars?
[Solution]   1  99%  0.01.

n ( )  (2.576  60  5)2  955.6.

Hence the sample size should be at least 956. (Always round up!)

This study source was downloaded by 100000787038097 from CourseHero.com on 12-30-2021 20:55:36 GMT -06:00
STAT0302 ch4 5
https://www.coursehero.com/file/10046339/Lecture-4/
Powered by TCPDF (www.tcpdf.org)