CHAPTER 5: Estimation

5.1 Point and interval estimation

5.2 Confidence interval
5.3 Confidence interval for the mean in single population
5.4 Confidence Interval for the difference population mean
5.5 Confidence interval for single population variance and ratio of two variance

Chapter 5

Estimation

One aspect of inferential statistics is estimation, which is the process of estimating the
values of an unknown population parameter from a sample. These values are only estimates the
true parameters and are derived from data that collected from samples.
Estimator
An estimator is a sample statistic from sample data that used to estimate
the unknown population parameter.

Estimate
An estimate is a specific value of a sample statistic that assigned to the
unknown population parameter.

Estimator and Estimate

If from a sample data, the mean of body temperature is
true population mean

. Thus, the sample statistic used to estimate the true population

parameter is estimator and the assigned value is estimate. Here,

the estimate is

, then we may estimate the

is called the estimator while

There are two types of estimation, which are (a) point estimate, and (b) confidence interval. As
shown in Figure 5.1 below:

Estimation

Point estimate

Confidence Interval

Figure 5.1 : Estimation

5.1

The Point Estimation

A point estimator, generally designated by the symbol , is the rule or formula that tells
us how to use the observations in a sample to compute a single number (a point) which serves as
a point estimate of the value of

Point Estimate
A point estimate is a specific numerical value that used to
estimate an unknown population parameter

Example 5.1
a) State that how to get the point estimate of the population parameter .
b) Find the point estimate of the sample data below to estimate the population mean, .
3.55

3.61

3.47

3.48

3.80

Solution:
a) We would like to estimate the value of
and obtained the statistic . The value of

with consideration of a random sample

of the statistic , computed from this sample of size n

, is a point estimate of the population parameter , ie

b) From (a), the point estimate of the population mean

is

5.1.1 The Properties of a Point Estimation

Since a point estimator is calculated from a sample, it possesses a sampling distribution.

The sampling distribution of a point estimator completely describes its properties. A good
estimator satisfies the three properties below:

Three Properties of a Good Estimator

1. The estimator

should be unbiased estimator.

That is, the expected value or the mean of the estimates obtained from samples of
a given size is equal to the parameter

being estimated, i.e.

2. The estimator should be consistent.

For the consistent estimator, as sample size increase, the value of the estimator
approaches the value of the parameter estimated.
3. The estimator should be a relatively efficient estimator.
That is, of all the statistics that can be used to estimate a parameter, the relatively
efficient estimator has the smallest variance.

Example 5.2:
If X has the binomial distribution with the parameters n and p, show that the sample proportion,
is an unbiased estimator of p.
Solution:
Since

Hence

it follows that

is an unbiased estimator of p.

5.2

The Confidence Interval

For the purpose of knowing how close the point estimate to the population means or
variance, statisticians prefer to other types of estimate that is called an interval estimate.

Interval estimate
An interval estimate of a parameter is an interval or range of
value that is used to estimate the true parameter.

Example 5.3
If the average age of all students in a college is 22.3 years and the error would be 0.4 year, then
write an interval estimate for the average age.
Solution:
The interval estimate for the average age of all the students in the college might be
years or

years.

A degree of confidence interval (usually a percentage) can be assigned before making

interval estimation. For instance, one may be wish to be 95% confident that the interval contains
the true population mean. If one desires to be more confident, such as 98% or 99% confidence,
then the interval must be larger. These confidence interval estimate would named as confidence
interval.
Confidence Level
The confidence level of an interval estimate of a parameter is the
probability that the interval estimate will contain the true parameter.

Confidence Interval
A confidence interval is a specific interval estimate of a parameter that determined by
using collected data from a sample with the specific confidence level of estimate.

Confidence Interval for Population Mean

Confidence Interval for Sample

Example 5.4
With 95% confidence, state the interval for (a) population mean, and (b) sample mean, where
standard deviation is

and sample size is n.

Solution:
Since we are 95% confident, then with
a) The interval for population mean,

is

b) The interval for a specific selected sample mean, , that falls within the range of
is

Refer to Example 5.4, the value used for the 95% confidence interval, 1.96 is obtained
from statistical table. It is based on the standard normal distribution. Since other confidence
intervals would be used in estimation, thus the symbol

(read z sub alpha over two) is used

in the general formula for confidence intervals. The Greek letter

(alpha) is the total are in

both of the tails of the standard normal distribution curve while

is the area in each one of

tails.

Figure 5.2:

For

the probability of selecting a random variable that will produce an

interval where containing

is

sample, is called a

. The interval

computed from the selected

confidence interval.

Confidence Coefficient
The values

is called the confidence coefficient or the degree of

confidence.

Confidence Limit
The endpoints of interval

and

are called the lower and

upper confidence limits, respectively.

According to the central limit theorem, we can expect the sampling distribution of
approximately normally distributed with mean

and standard deviation

to be
From

Figure 5.2,
where

As a result, it give

confidence interval of

with

known,

In the interval estimation, the error would be existed in the interval estimate. That error is
called the maximum error of estimate or standard error. It takes into account during the
estimation. If the error is large, it means that the estimation is far from the true parameter. Thus
the estimation is inaccurate. In practice, we want to make the error as small as we could when

estimating the true parameter. Thus, the

confidence interval provides an estimate

of accuracy of point estimate.

From Figure 5.3, as shown below, we could view that if
thus

may estimate

without error. However,

estimate is in error. We are

is the center of the intervals,

will not be exactly equal to . Hence, the point

confident that the error will not exceed

Figure 5.3: Error of Estimating

by

Standard error
Assume that the life of time of the car battery population has a true mean

years and it is

known for others. A quality engineer has make a sampling of the life time of the battery
population and obtained the sample mean

years. Since the sample mean

used to estimate the true mean of the battery population but

. Therefore the error is exist

which is

Standard error and Sample size

a) Error, e = absolute value of the difference between

[In case

and

unknown and sampling is from a normal distribution, s replaces .

b) Sample size, n =

is

From the formula above, we may conclude that there exists an inverse relationship
between standard error and sample size. The standard error would be decrease when the sample
size increasing and conversely.

Statement of Confidence Interval

Probability of selecting a random variable that will produce

1.

an interval containing
2.

Total area in both of the tails of the standard normal

(alpha)

distribution curve, or called level of significance

3. (1-

Confidence coefficient or the degree of confidence

4. Interval

5. Endpoint

Lower confidence limit

6. Endpoint

Upper confidence limit

7. Standard error

Maximum or minimum difference between the point

confidence level

estimate and the true parameter.

5.3

The Confidence Interval for the Mean in Single Population

Case 1: Variance,

known, for any value of sample sizes.

Case 2: Variance,

unknown and sample size

(If is

unknown then replace

Case 3: Variance,

with s)

unknown and sample size

Example 5.4
A study was carried out to estimate the average life of a large shipment of light bulbs. Previous
studies indicated that the standard deviation is known to be 100 hours. A random sample of 50
light bulbs was selected and indicated that the sample average life was 350 hours. Construct a
95% confidence interval estimate of the true average life for light bulbs in this shipment.
Solution:
Write down the information:
Confidence interval 0.95 for

Determine the case

By case 1:

From table,

The conclusion is, we have 95% confident that the average life of light bulbs is between 322.28
and 377.72 hours.

Example 5.5
The brightness of a television picture can be evaluated by measuring the amount of current
required to achieved a particular brightness level. A random sample of 10 tubes indicated a
sample mean of

microamps and a sample standard deviation is

microamps.

Find (in microamps) a 99% confidence interval estimate for mean current required to achieve a
particular brightness level.
Solution:

Example 5.6:
Talking a random sample of 35 individuals waiting to be serviced by the teller, we find that the
mean waiting time was 22.0 min and the standard deviation was 0.8 min. Construct a 90%
confidence interval estimate the mean waiting time for all individuals waiting in the service line.

5.4

The Confidence Interval for the Mean in Difference Population

Case 1: Variance,

and

known, for any value of sample sizes.

Case 2: Variance,

and

unknown and large sample size

Two different situations must be treated.

In Case 2 (a) :

In Case 2 (b) :

Where

Case 3: Variance,

and

unknown and small sample size

Same goes to Case 3, two different situations must be treated.

In Case 2 (a) :

With

and v is degree of freedoms obtained from

In Case 2 (b) :

Where

The degree of freedoms,

Example 5.7
The breaking strength of yarn supplied by two manufactures is being investigated. From the
previous studies, we noted that

and

specimens from each manufacturer results in

. A random sample of 20 test

and

a 90% confidence interval for the difference in the mean breaking strength.

, respectively. Find

Example 5.8:
A study was conducting to compare the starting salaries for university graduates majoring in
education and engineering. A random sample of 50 recent university graduates in each major
was selected and the following information was obtained.
Major

Mean

SD

Education

RM 2000

RM 100

Engineering

RM 2900

RM 150

Construct a 99% confidence interval for the difference in the mean starting salaries for two
majors.

5.5

Confidence Interval for Single Population Variance,

and Ratio of Two Variance,

and

In previous section, confidence intervals were constructed for single mean and difference
mean between two means. In this section, we will construct the confidence interval for
population variances.
There are two types of the confidence intervals for the population variance as shown in
Figure 5.4:
Confidence Interval
Of Population Variances

Single Variance

Ratio of Two Variances

Use Chi-Square distribution

Use F-distribution

Figure 5.4 : Types of Confidence Interval of Population Variance

Formula 1: The Confidence Interval of the Single Variance,

Where

degree of freedom

Formula 2: The Confidence Interval for the Ration of Two Variance,

Where

and

degree of freedom

Example 5.9:
A sample of 20 cigarettes has a standard deviation

Find the 95% confidence interval

for the variance and standard deviation of the nicotine content of cigarettes manufactured.
Solution:

Example 5.10:
The data below was obtained from two normal populations. Find the 98% confidence interval for
.

Solution: