MODULE TWO: MEASURING

UNCERTAINTY; AND DRAWING

CONCLUSIONS ABOUT POPULATIONS
BASED ON SAMPLE DATA
TOPIC 6: CONFIDENCE INTERVALS

Deakin University CRICOS Provider Code: 00113B

+ 2

Learning Objectives
At the completion of this topic, you should be able to:
• calculate estimates and their standard errors
• construct and interpret confidence interval estimates for the mean and
the proportion
• determine the sample size necessary to develop a confidence interval for
the mean or proportion
+Confidence Intervals 3

A confidence interval gives a range of values

•Takes into consideration variation in sample statistics from sample to
sample
•Based on observations from one sample
•Gives information about closeness to unknown population parameters
•Stated in terms of level of confidence
•Can never be 100% confident
+Point Estimates 4
+Point and Interval Estimates 5

A point estimate is the value of a single sample statistic

A confidence interval provides a range of values constructed
around the point estimate
+Estimation Process 6
+Level of Confidence 7

Common confidence levels = 90%, 95%

or 99% (α is therefore 0.1, 0.05 or 0.01)
•Also written (1 - α) = 0.90, 0.95 or 0.99
A relative frequency interpretation
•In the long run, 90%, 95% or 99% of all the
confidence intervals that can be constructed
(in repeated samples) will contain the
unknown true parameter
A specific interval either will contain or
will not contain the true parameter
+Confidence Intervals for Different 8

Samples
+Normal Curve for 95% and 99% 9

Level of Confidence
+Confidence Intervals 10

Confidence
Intervals

Population Population
Mean Proportion

σ Known σ Unknown
+Confidence Interval 11

The general formula for all confidence intervals is:

Point Estimate ± (Critical Value)*(Standard Error)

Represents confidence for which the interval will contain

the unknown population parameter
+Confidence Interval Estimation for 12

the Mean (σ Known)

Assumptions:
•Population standard deviation σ is known
•Population is normally distributed
•If population is not normal, use Central Limit Theorem
Confidence interval estimate σ
X±Z
n
Where: X is the point estimate
Z is the normal distribution critical value for a probability of α/2 in each tail
σ/ n is the standard error
+Confidence Interval (σ Known) 13

Example:
A sample of 11 pizza shops from a large normal population has
a mean pizza cooking time of 2.2 minutes. We know from past
testing that the population standard deviation is 0.35 minutes
Determine a 95% confidence interval for the true mean pizza
cooking time of the population
+Confidence Interval (σ Known) 14

σ
X±Z
n
0.35
= 2.2 ± 1.96
11
= 2.2 ± 0.2068
= 1.9932 ≤ µ ≤ 2.4068

We are 95% confident that the true mean pizza cooking time is
somewhere between 1.9932 and 2.4068 minutes
Note: Although the true mean may or may not be in this
interval, 95% of intervals formed in this manner (in repeated
samples) will contain the true mean
+Confidence Interval Estimation for 15

the Mean (σ Unknown)

If the population standard deviation σ is unknown, we can
substitute the sample standard deviation, S
This introduces extra uncertainty, since S is variable from
sample to sample
So we use the Student t distribution instead of the normal
distribution
•the t value depends on degrees of freedom denoted by sample size minus
1; i.e. (d.f = n - 1)
•d.f are the number of observations that are free to vary after sample mean
has been calculated
+Confidence Interval Estimation for 16

the Mean (σ Unknown)

Confidence Interval Estimate (σ Unknown)
S
X ± t n-1
n
(where t is the critical value of the t distribution with n -1 degrees of
freedom and an area of α/2 in each tail)
+Student’s t Distribution 17

At the beginning of the twentieth century a statistician for

Guinness Breweries in Ireland, William S. Gosset, wanted to
make inferences about the mean when σ was unknown
Because Guinness employees were not permitted to publish
research work under their own names, Gosset adopted the
pseudonym 'Student'. The distribution that he developed is
known as Student's t distribution
•If the random variable X is normally distributed, then the following statistic
has a t distribution with n - 1 degrees of freedom: X −µ
t=
S
N
+Properties of the t Distribution 18

The t distribution appears very similar to the standardised

normal distribution - both distributions are bell shaped
However, the t distribution has more area in the tails and less in
the centre than the standardised normal distribution
The degrees of freedom (n – 1) are directly related to the
sample size n
As the sample size increases, S becomes a better estimate of σ
and the t distribution gradually approaches the standardised
normal distribution until the two are virtually identical
+Properties of the t Distribution 19

(cont)
With a sample size of about 120 or more, there is little
difference between the t values and Z values.
+The Concept of Degrees of Freedom 20

The Degrees of Freedom are the number of observations that

are free to vary after a sample mean has been calculated
Example:
The mean of 3 numbers is calculated as 8. If X1 = 7 and X2 = 8,
then X3 must be 9. In other words, X3 is not free to vary.
For this example, the first two values can be any numbers, but
the third is not free to vary for a given Mean
Therefore, if n = 3, the degrees of freedom will be n – 1 or 2
+The Concept of Degrees of Freedom 21
+The Concept of Degrees of Freedom 22
+Confidence Interval (σ Unknown) 23

Example:
A random sample of 25 people have a S
X ± t n -1
mean age of 50 and sample standard n
deviation of 8 8
50 ± 2.0639
Calculate a 95% confidence interval for μ 25
•Degrees of Freedom are 25-1 = 24 50 ± 3.30224
46.69776 ≤ µ ≤ 53.30224
•t = 2.0639 (from Table E3 - Critical values of t)
Note: Table E3 is Upper Tail only so need to use
α/2
+MS Excel® Template to Calculate 24

Mean Confidence Interval Estimate

+Confidence Interval Estimation for 25

the Proportion
Confidence interval estimate
The Upper and lower confidence limits for the population
proportion are calculated with the formula
p(1− p)
p±Z
n
Where: Z is the standard normal value for the level of confidence desired
p is the sample proportion
n is the sample size
+Confidence Interval Estimation for 26

the Proportion (cont)

Example: p(1 − p)
p±Z
A random sample of 100 people shows that n
0.25(1 − 0.25)
25 are left-handed 0.25 ± 1.96
100
Calculate a 95% confidence interval for the 0.25(1 − 0.25)
0.25 ± 1.96
true proportion of left-handers 100
0.25 ± 0.0849
Answer: 0.1651 ≤ π ≤ 0.3349
We are 95% certain that the true proportion
of all people who are left handed lies
somewhere between 16.51% and 33.49%
+MS Excel® Template to Calculate 27

Proportion Confidence Interval Est.

+Determining Sample Size for the 28

Mean
In order to determine the required sample size for the mean,
we need to know:
•the desired level of confidence (1 - α), which determines the critical Z value
•an acceptable sampling error, e
•the standard deviation, σ (can be estimated with past data or pilot sample)
We rework the Sampling Error (the Margin of Error) of the
Confidence Interval Estimation for the Mean to resolve n
σ σ Z 2
σ 2
X ±Z is denoted e = Z then isolate n n = 2
n n e
+Determining Sample Size for the 29

Mean (cont)
Example:
What sample size is needed to estimate the mean within ± 5 with
90% confidence? If σ = 45
Z 2 σ 2 1.6452 452
n= 2
= 2
= 219.19
e 5
Therefore, the required sample size is n = 220 (always round up)
+MS Excel® Template to Calculate 30

Sample Size for a Mean

+Determining Sample Size for the 31

Proportion
In order to determine the required sample size for the
Proportion, we need to know:
•the desired level of confidence (1 - α), which determines the critical Z value
•an acceptable sampling error, e
•true proportion of ‘successes’, π (can be estimated with past data or pilot
sample, p, or conservatively use π = 0.5)
We rework the Sampling Error (the Margin of Error) of the
Confidence Interval Estimation for the Proportion to resolve n
2
π ±Z
π (1 − π ) is denoted e = Z π (1 − π ) then isolate n n = π (12 − π )
Z
n n e
+Determining Sample Size for the 32

Proportion (cont)
Example:
What sample size is needed to estimate the true proportion of
defective parts in a large population to within ±3%, with 95%
confidence? (Assume a pilot sample yielded 12% defects. i.e p =
0.12)
Z 2 π (1 − π ) 1.96 2 0.12 (1 − 0.12)
n= 2
= 2
= 450.74
e 0.03

Therefore, the required sample size is n = 451 (always round up)

+MS Excel® Template to Calculate 33

Sample Size for a Proportion

+More on Confidence Interval 34

Estimation and Ethical Issues

A confidence interval estimate (reflecting sampling error)
should always be included when reporting a point estimate
The level of confidence should always be reported
The sample size should be disclosed
An interpretation of the confidence interval estimate should
also be provided