Professional Documents
Culture Documents
Chapter Eight
Statistical Intervals for a Single Sample
Goals of chapter:
After completing this chapter, you should be able to:
Construct and interpret a confidence interval estimate for a single population mean using both the z
and t distributions and single population proportion
Determine the required sample size to estimate a single population mean or single population
proportion within a specified margin of error
8.1.1 Develpoment of the Confidence Interval Estimate and its Basic Properties:
Suppose that a sample of size n selected randomlly from a normal distribution with unknown mean 𝜇 and
known standard deviation 𝜎. From the results of chapter 7 we know that the sample mean is normally
σ
distributed with mean 𝜇 and standard deviation 𝑛. We may standardized the sample mean using that fact
√
that
(𝑥̅ − 𝜇)√𝑛
𝑍=
𝜎
The random variable 𝑍 has a standard normal distribution.
In General
A confidence interval estimate for the population mean 𝜇 is an interval of the form
𝑙≤𝜇≤𝑢
where the end points 𝑙 & 𝑢 are computed from the sample data
There is a probability of 1 − 𝛼 of selecting a sample for which the C.I will contain the true value of 𝜇
The endpoints or bounds 𝑙 & 𝑢 are called the lower and the upper confidence limits and
1 − 𝛼 is called the confidence coefficient
Page |2
(𝑥̅ −𝜇)√𝑛
In our problem, because 𝑍 = has a normal distribution , we may write
σ
(𝑥̅ − 𝜇)√𝑛
𝑃 (−𝑧𝛼 ≤ ≤ 𝑧𝛼 ) = 1 − 𝛼
2 𝜎 2
Manuipulate the quantities inside the brackets, this results in
𝜎 𝜎
𝑃 (𝑥̅ − 𝑧𝛼 ≤ 𝜇 ≤ 𝑥̅ + 𝑧𝛼 ) = 1 − 𝛼
2 √𝑛 2 √𝑛
𝛼
where 𝑧𝛼 is the upper 100 percentage point of the standard normal distribution
2 2
b) What can you do when you don’t know µ? you use x to estimate µ
If x = 362.3 the interval is 362.3 ± 1.96 * 15 / 25 = (356.42, 368.18)
Since 356.42 ≤ µ ≤ 368.18 the interval based on this sample makes a correct statement about µ.
But what about the intervals from other possible samples of size 25?
In practice you only take one sample of Sample # X Lower Limit Upper Limit Contain µ?
size n
In practice you do not know µ so you do 1 362.3 356.42 368.18 Yes
not know if the interval actually contains µ 369.5 363.62 375.38 Yes
2
However you do know that 95% of the
3 360 354.12 365.88 No
intervals formed in this manner will
contain µ 4 362.12 356.24 368 Yes
Thus, based on the one sample, you 5 373.88 368 379.76 Yes
actually selected you can be 95%
confident your interval will contain µ (this is a 95% confidence interval)
Page |3
|𝑥̅ − 𝜇|
will not exceed a specified amount 𝑒
Cases for Confidence Interval for the true population mean (μ):
A (1-) 100% Confidence Interval for μ, when (σ unknown) and the sample size is large
Assumptions
1. Population standard deviation σ is unknown
2. Sample size large n 30
𝑠
𝑥̅ ± 𝑧∝
2 √𝑛
Case III (8.2.1 Confidence Interval on the mean on a Normal distribution, variance unknown):
A (1-) 100% Confidence Interval for μ, when (σ unknown) and the sample size is small
Assumptions
1. Population standard deviation σ is unknown
2. Population is normally distributed
3. Sample size small n 30
𝑠
𝑥̅ ± 𝑡∝,𝑛−1
2 √𝑛
Theorem:
Let 𝑋1 𝑋2 , 𝑋3 , … , 𝑋𝑛 , be a random sample from a normal distribution with unknown mean and
𝑥 2 −𝑛𝑥̅ 2
unknown variance and 𝑠 2 = ∑ , then
𝑛−1
The random variable
(𝑥̅ − 𝜇)√𝑛
𝑇=
s
has a t – student distribution with degrees of freedom 𝑑. 𝑓 = 𝑛 − 1
Page |5
It has bell shape like normal distribution. It has the same properties as the normal distribution,
mean = median = mode. And it is depend on a dgree of feedom (𝑑. 𝑓 = 𝑛 − 1).
As n increases, t – destribution will be z – normal distribution.
dgrees of feedom
Number of observations that are free to vary after sample mean has been calculated
Example:
Use the t – student table to find the followng:
1. t0.05,13 =
2. t0.95,13 =
3. P(T12 < 1.34) =
Example
Assume a sample of size n has been obtained from a normal distribution. Determine the critical value from
a t – distribution when you wish to estimate the population mean in each of the following cases:
1. confidence coefficient = 0.95 and n = 26
2. confidence coefficient = 0.90 and n = 31
3. confidence coefficient = 0.98 and n = 15
4. confidence coefficient = 0.80 and n = 32
5. confidence coefficient = 0.99 and n = 19
Page |6
Exemple:
Ten measurements of impact energy (J) on specimens of A238 steel cut at 60°C are as follows:
64.1 64.7 64.5 64.6 64.5 64.3 64.6 64.8 64.2 64.3
Example:
A sample of 11 circuits from a large normal population has a mean resistance of 2.20 ohms. We know from
past testing that the population standard deviation is 0.35 ohms.
1. Determine a 95% confidence interval for the true mean resistance of the population.
2. In part 1 above, do you need any assumptions, if yes what, if no why.
Page |8
Example:
A machine produces metal rods in an automobile suspension system. A random sample of size 15 rods is
selected and the diameter is measured. The results ( in millimeters) as follows :
|𝑥̅ − 𝜇|
will not exceed a specified amount 𝑒
n
e2 e
Example:
Suppose that it is known that the population standard deviation is 40. if you wish to estimate the population
mean using a 95% confident that the error in estimating the true population mean to be within ± 2.5. what
sample size will be required?
P a g e | 10
Example:
A decision maker is interested in estimating the mean of a population based on a random sample. He wants
the confidence level to be 90% and the total length of the confidence interval on the mean to beat most 0.6.
He does not know what the population standard deviation is, so he has selected the following pilot sample:
Based on the pilot sample, how many more items be sampled so that the decision maker can make the
desired confidence interval estimate?
P a g e | 11
𝑝̂ (1 − 𝑝̂ )
𝑠𝑝̂ = √
𝑛
Then,
A (1-) 100% Confidence Interval for the true proportion is given by:
𝑝̂ (1 − 𝑝̂ )
𝑝̂ ± 𝑧𝛼 √
2 𝑛
where
z is the standard normal value for the level of confidence desired.
𝑥
𝑝̂ = is the sample proportion.
𝑛
n is the sample size.
Note:
Increases in the sample size reduce the width of the confidence interval
Example:
A random sample of 500 people shows that 25 are left-handed.
Example:
In a random sample of 85 automobile engine crankshaft bearings, 10 have a surface finish that is rougher
than the specifications allow.
1. Construct a 95% two-sided confidence interval for the true proportion of bearings in the population that
exceeds the roughness specification.
2. How large a sample is required if we want to be 95% confident that the error in using the sample
proportion to estimating the ture value p is less than 5%?
3. How large must the sample be if we wish to be at least 95% confident that the error in estimating the
true proportion is less than 2% regardless of the true value of P ?