Chapter 8

Interval
Estimation

1
 Chapter 8: Learning Objectives

Textbook
Learning Objective
Section
• The student must be able to construct and interpret an interval
8.1
estimate of a population mean for known
• The student must be able to construct and interpret an interval
8.2
estimate of a population mean for unknown
• know about the -distribution and how to use it in constructing
8.2
an interval estimate for a population mean
• be able to do interval estimation for the population mean with
8.1, 8.2
the help of Excel
• be able to construct and interpret an interval estimate of a
8.4
population proportion
• be able to do interval estimation for the population proportion
8.4
with the help of Excel

Sections excluded:
• Section 8.3: Determining the Sample Size
• Section 8.5: Practical Advice: Big Data and Interval Estimation (read for interest)
 Interval estimation

Estimates

Point Interval
Estimates Estimates
We are 95% confident that
The mean waiting time at
the mean waiting time at a
a check point is 3 minutes
check point is between 2
and 5 minutes
is a point estimator of
Interval estimate
is a point estimator of

= Point estimate

Margin of Error
: Margin of Error

p: Margin of Error
2
 Confidence interval for

mmmmm…so I am
95% CONFIDENT
that the confidence
interval that I calculate
from a SINGLE
sample will contain
the unknown
population mean

3
 Confidence intervals for

Margin of Error

known unknown

4
 Confidence interval for

1 100% confidence interval

Margin of Error
Point estimator

Confidence Coefficient Standard error

of ̅
1

5
 -values for Confidence Intervals
Most commonly used confidence levels:
Confidence Level ⁄ Margin of error
90% 0.10 0.05 1.645 1.645 ̅
95% 0.05 0.025 1.960 1.96 ̅
99% 0.01 0.005 2.576 2.576 ̅

90% Confidence Level: 0.045

0.05
95% Confidence Level:

6
 Interval estimate of a population mean : σ known

Margin of Error

Confidence Coefficient Level of Significance

z . .

1.96 1.96
7
 Interval estimate of a population mean : σ known
Consider a sample of 16 test marks. Assume
test marks are normally distributed with 10.
Let ̅ sample average test mark 58

Calculate a 95% confidence interval for :

58 1.96 10
16
58 4.9

58 4.9, 58 4.9
53.1, 62.9

We are 95% confident that (the population mean mark for

this test) is between 53.1 and 62.9
8
 Focusing on the Margin of Error

58 1.96 58 4.9

Sampling error = ≤ margin of error

95% of the time the sampling error will be 4.9 or less.

There is a 0.95 probability that the sample mean, ,

will provide a sampling error of 4.9 or less.

9
 Interval estimate of a population mean : σ known
Consider a sample of 16 test marks. Assume
test marks are normally distributed with 10.
Let ̅ sample average test mark 58
Calculate a 99% confidence interval for : level of significance
1– 0.99 0.01
58 2.576 10 ∴
.
0.005
16
58 6.44

58 6.44, 58 6.44
51.56, 64.44

Confidence Level Margin of error

99% 0.01 0.005 2.576 2.576 ̅

We are 99% confident that (the population mean mark for this
test) is between 51.56 and 64.44 10
 Focusing on the Margin of Error

58 2.576 10 58 6.44
16

Sampling error = ≤ margin of error

95% of the time the sampling error will be 6.44 or less.

There is a 0.95 probability that the sample mean, ,

will provide a sampling error of 6.44 or less.

11
 Interval estimate of a population mean : σ known
Consider a sample of 16 test marks. Assume
test marks are normally distributed with 10.
Let ̅ sample average test mark 58

95% confidence interval for : 99% confidence interval for :

53.1, 62.9 51.56, 64.44

Width 9.8 Width 12.88

What can you conclude?

12
 Confidence interval for

Interpretation

% confident that There is a

(the population mean) is probability that the
between the Lower Limit sample mean, , will
and Upper Limit of the provide a sampling error
Confidence Interval. of or less
⁄

13
 Confidence interval for using Excel

( )% CI = Margin of error
NORM.S.INV *

=CONFIDENCE.NORM( , , )

14
 Confidence interval for

Margin of Error

known unknown

15
 Interval estimate of a population mean : unknown

Estimate using

Properties of the -distribution

• Symmetric around 0
• Degrees of freedom 1

See Tables in the textbook

16
 Characteristics of the -distribution
 Symmetric around 0.
 Characterized by the degree of freedom, in this case, – 1.
 As the degrees of freedom increase, the -distribution tends
to the standard normal distribution.
 For “large sample cases” ( 30 the -distribution
approaches the standard normal distribution.

Relationship between the and – distributions

17
 Interval estimate for with unknown

1 100% confidence interval

Margin of Error
Point estimator

Confidence Coefficient Estimator for

the standard
error of
1

⁄ , ⁄, ,

18
 .
.
.

NB: Only the bottom row (∞ degrees of freedom) corresponds to

the standard normal distribution! 19
 Interval estimate of a population mean : unknown – an example

Given: 15, 53.87 and 6.82. Take note: is unknown.

Calculate a 95% confidence interval for
= level of significance
1 – 0.95 0.05

.
∴ 0.025

–1 15 – 1 14

20
 Interval estimate of a population mean : σ unknown – an example
Given: 15, 53.87 and 6.82.
Take note: is unknown. Calculate a 95% confidence interval for

6.82
53.87 2.145
15
53.87 3.78
53.87 3.78, 53.87 3.78
50.09, 57.65
We are 95% confident that is between 50.09 and 57.65.
Obtaining ∝⁄ using Excel: Obtaining the level of significance using Excel:
= T.INV( ⁄ , ) =2 1 T.DIST( ∝⁄ , ))
= T.INV(0.025, 14) =2 1 0.97501
= 2.144787 = 0.04998 0.05
OR OR
= T.INV.2T( , ) = T.DIST.2T( ∝⁄ , )
= T.INV(0.05, 14) = T.DIST.2T(2.145, 14, 2)
= 2.144787 = 0.04998 0.05
Note: The T.DIST.2T function requires a positive value
21
 Focusing on the Margin of Error

53.87 2.145 . = 53.87 3.78

Sampling error = ≤ margin of error

95% of the time the sampling error will be 3.78 or less.

There is a 0.95 probability that the sample mean, ,

will provide a sampling error of 3.78 or less.

22
 Confidence interval for : σ unknown

Interpretation

There is a
% confident that
probability that the sample
(the population mean) is
mean, , will provide a
between the Lower Limit
and Upper Limit of the sampling error of ⁄ or

Confidence Interval. less.

23
 Confidence interval for σ unknown) using EXCEL

% CI = Margin of error

= T.INV.2T( , )*

= CONFIDENCE.T( , , )

24
 Interval estimate of a population proportion
The sampling distribution of is approximately normal for “large” samples
with ̅ and ̅ , i.e. ̅ ~ ,

1
CI: Margin of error ̅ ̅ ̅

Why doesn’t it make sense to use this expression to develop an interval

estimate of a population proportion, ?

To use this expression to develop an interval estimate of a population

proportion, , the value of would have to be known.
But, the value of is what we are trying to estimate, so we simply substitute
the sample proportion in the place of .
Therefore: Margin of Error
Point estimator
Estimate of
the standard
error of ̅ 25
 Interval estimate of a population proportion

Sampling distribution of the sample proportion

̅~ ̅, ̅

1
∴ ̅~ ,

1 100% of all values

Confidence Level Margin of error

90% 0.10 0.05 1.645 1.645 ̅

95% 0.05 0.025 1.960 1.96 ̅
99% 0.01 0.005 2.576 2.576 ̅
26
 Interval estimate of a population proportion : Example
A national survey of 902 women golfers was taken to learn how
female golfers view themselves as being treated at golf courses.
The survey found that 397 if the female golfers felt they were treated
fairly.
Calculate a 95% confidence interval for , the proportion of female
golfers who felt they were treated fairly.

1 Point estimate is:

0.4401
0.4401 1 0.4401 = level of significance
0.4401 1.96
902 1 – 0.95 0.05
0.4401 0.0324
.
0.4401 0.0324, 0.4401 0.0324 ∴ 0.025
0.4077, 0.4725
Confidence Level Margin of error
95% 0.05 0.025 1.960 1.96 ̅

We are 95% confident that the proportion of female golfers who felt they were
being treated fairly, , is between 0.4077 and 0.4725. 27
 Focusing on the Margin of Error

1
2

0.4401 1 0.4401
0.4401 1.96 0.4401 0.0324
902

Sampling error = ≤ margin of error

95% of the time the sampling error will be 0.0324 or less.

There is a 0.95 probability that the sample proportion, ,

will provide a sampling error of 0.0324 or less.

28
 Confidence interval for

Interpretation

% confident that There is a

(the population proportion) probability that the sample
is between the Lower Limit proportion, , will provide a
and Upper Limit of the sampling error of
Confidence Interval.
⁄ or less.

29
 Confidence interval for using EXCEL

( )% CI = Margin of error

= NORM.S.INV *

No Excel function designed explicitly for .

Hint: Consider how the

CONFIDENCE.NORM() function can be
used.
30