Chapter 7 – Interval Estimation
Objectives are:
• Derive confidence intervals for a single population
mean
• Derive the confidence interval for the difference
between two populations means
• To calculate the CI for one population proportion
• To calculate the difference between two population
proportions
Significance Level
• In order to make inferences about a population, a specified
significance or confidence level must be specified
• The significance level indicates the certainty of the population
parameter falling within the interval
• The significance level is denoted by
• Most common significance levels are:
• 90% (),
• 95%() and
• 99%()
z*–values for Various Confidence Levels
Confidence Level z*-value
80% 1.28
90% 1.645 (by convention)
95% 1.96
98% 2.33
99% 2.58
How to find a critical value?
For example if you want a critical value for 95% CI
(1-a)100%= 95%
95%/100%=0.95
1-0.95= 0.05
Za/2= 0.025
If you check 0.025 in a Z-table, you will find that it is 1.96
The Za/2 of 95% CI is equal to 1.96
Significance Level
• Take note of the sample size when determining thefor
appropriate Table value.
• The sample size will also give you an indication of which
formulae to be used
• If then use the Z-Table ()
• If then use the T-Table with the associated degrees of
freedom ()
• The alpha value is always divided by 2 for interval
estimation
Point and Interval Estimation
Point Estimation
• Approximating a single value to be inferred as the
population parameter
Interval Estimation
• Approximating an interval in which the population
parameter will fall using some measure of
significance/probability
Confidence Interval for the Population Mean
When the population standard deviation is
known or given
[ 𝑋−𝑧𝛼
2
( )
𝜎
√𝑛
; 𝑋+ 𝑧𝛼
( )
𝜎
2 √𝑛
]
Lower Limit Upper Limit
Confidence Interval for the Population Mean
When population standard deviation is unknown and
[ 𝑋−𝑧𝛼
2
( )
𝑠
√𝑛
; 𝑋+ 𝑧𝛼
2
( )
𝑠
√𝑛
] 𝑧 𝛼
2
Population is unknown and
[ 𝑋 −𝑡 𝛼
2
( )
𝑠
√𝑛
; 𝑋+𝑡𝛼
2
( )
𝑠
√𝑛
] 𝑡𝛼
2
; 𝑛 −1
Confidence Interval for the difference between
two population means
When the population variances are unknown and
¿
(√ )
2 2
𝑠 𝑠
1 2 𝑧
[( 𝑋 1 − 𝑋 2) ± 𝑧 𝛼 + ] 𝛼
2
𝑛1 𝑛 2 2
Confidence Interval for the difference between
two population means
The population variances are unknown and either
Sp= is called pulled standard deviation estimator
2 2
2 (𝑛1 −1 ) 𝑠1 +(𝑛2 −1)𝑠2
𝑠𝑝 =𝐼𝑠𝑐𝑎𝑙𝑙𝑒𝑑 𝑃𝑢𝑙𝑙𝑒𝑑𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟
𝑛1 +𝑛2 −2
𝑡𝛼
; +𝑛 2 −2
2
Confidence Interval for population proportions
(√ ( ) ) (√ )
𝑋 𝑋 𝑋 𝑋
1− ( 1− )
𝑋 𝑛 𝑛 𝑋 𝑛 𝑛
[ −𝑧𝛼 ; +𝑧 𝛼 ]
𝑛 2
𝑛 𝑛 2
𝑛
𝑧 𝛼
2
Confidence Interval for the difference between
two population proportions
𝑋1 𝑋2
^1=
𝑝 ^2 =
∧𝑝
𝑛1 𝑛2
^1 − 𝑝
[( 𝑝 ^2)± 𝑧 𝛼
2
(√ ^1 ( 1− 𝑝
𝑝
𝑛1
^1 ) 𝑝
+
^2 ( 1− 𝑝
𝑛2
^2 )
)]
𝑧 𝛼
2
Examples
1. The mean and standard deviation of the quality grade-points of a random sample of 36 students from a large
population are 2.6 and 0.3 respectively.
a) Find 95% and 99% CI for the mean of the entire population and interpret
What do we have?
n=36;
Mean= 2.6
Std= 0.3
Level of confidence = 95% and 99%
Are we going to use a t-distribution or Z-distribution table?
Is population Std known or unknown?
Is n greater than or equal to 30? Or it is less than 30?
Solution for 95% CI
1-0.95=0.05
0.05/2= 0.025
Za/2= 1.96
95% ci
= 2.6 1.96=
Solution for 99% CI
1-0.99=0.01
0.01/2= 0.005
Za/2= 2.58
99% CI=
More examples
2. In a random sample of 500 people eating lunch at a cafeteria on various
Fridays, it was found that 160 preferred seafood. Find a 95% CI for the
actual proportion of people eating seafood on Fridays.
= 0.32
= 1.96
95%= 1-0.95= 0.05/2=0.025
From Z-table= 1.96
=
0.32 1.96
95% CI =
More examples
3. A clothing store analysed the value of purchases made on credit cards by a sample of 25
credit card customers. The sample mean was found to be R 164.45 with a sample standard
deviation of R 38.60. Calculate a 90% CI for the mean and interpret your answer.
(1-a)100%= 1-0.90= 0.1
Za/2= 0.1/2= 0.05
95% CI=
, n-1=
164.45=