Recall that: Principles of Statistics for Admin.

(15060105)
We use point estimate to estimate the population parameter using a one value called: (Sample Statistic).
Now, we use an interval to estimate the population parameter. This interval is called: (Confidence interval).

Confidence intervals= point estimate ∓ margin of error

To construct our confidence intervals: starts by determine the level of significance ( 𝜶 ).

Then, we say that: we trust (insure) that our confidence interval includes the value of the true parameter
by the percent 𝟏 − 𝜶 𝟏𝟎𝟎%.

If we assume ( 𝜶 = 𝟎. 𝟎𝟓 ), then we insure that our confidence interval includes the value of the true parameter
by the percent 𝟏 − 𝟎. 𝟎𝟓 𝟏𝟎𝟎% = 95%. Thus: here 95% is the confidence level.

And the value 0.95 is the confidence coefficient.

Note: Assuming the confidence interval (Lowe, Upper) = (L,U). The length of this interval increases as
confidence level increases.
Note: The Confidence Interval Range is a twice of (margin of error).
* Confidence intervals for population mean (𝝁) when:
1.The sample size is small (𝒏 < 𝟑𝟎) Or large (𝒏 ≥ 𝟑𝟎) and population variance is known. (using Z-tables)

𝝈
𝒙 ∓ margin of error : 𝒙 ∓ 𝒁𝜶 , 𝒁𝜶𝟐 value providing an area of α/2 in the upper tail of the Z distribution
𝟐 𝒏
2.The sample size is large (𝒏 ≥ 𝟑𝟎) and population variance is Unknown. (using Z-tables)

𝒙 ∓ margin of error : 𝑺
𝒙 ∓ 𝒁𝜶 ,
𝟐 𝒏
3.The sample size is small (𝒏 < 𝟑𝟎) and population variance is Unknown. (using t-tables)

𝑺 𝒕𝜶 value providing an area of α/2 in the upper tail of the t distribution

𝒙 ∓ margin of error : 𝒙 ∓ 𝒕𝜶 , 𝟐
𝟐 𝒏 with (n − 1) degrees of freedom

* Confidence intervals for population proportion (𝒑) :

𝒑(𝟏 − 𝒑)
𝒑 ∓ margin of error : 𝒑 ∓ 𝒁𝜶
𝟐 𝒏
Example:
Assuming a sample of 16 students, if the sample mean of student marks is 75.
Assuming the population variance is 64
Obtain a 95% confidence interval of the population mean (𝜇):
𝝈
𝒙 ∓ 𝒁𝜶
𝟐 𝒏
𝟔𝟒
7𝟓 ∓ 𝒁𝟎.𝟎𝟓
𝟐
𝟏𝟔
𝟖
𝟕𝟓 ∓ 𝒁𝟎.𝟎𝟐𝟓 , 𝒁𝟎.𝟎𝟐𝟓 = 𝟏. 𝟗𝟔
𝟒
7𝟓 ∓ (𝟏. 𝟗𝟔)(𝟐)
7𝟓 ∓ 𝟑. 𝟗𝟐 , margin of error = 3.92
We trust by 95% that 𝜇 ∈ (71.08 , 78.92)

95% confidence interval of 𝜇 is (71.08 , 78.92) → 𝟕𝟏. 𝟎𝟖 < 𝝁 < 𝟕𝟖. 𝟗𝟐

Example:
Assuming a sample of 100 students, if the sample mean of student marks is 75.
Assuming the sample variance is 64
Obtain a 95% confidence interval of the population mean:
𝑺
𝒙 ∓ 𝒁𝜶
𝟐 𝒏
𝟔𝟒
7𝟓 ∓ 𝒁𝟎.𝟎𝟓
𝟐
𝟏𝟎𝟎
𝟖
𝟕𝟓 ∓ 𝒁𝟎.𝟎𝟐𝟓 , 𝒁𝟎.𝟎𝟐𝟓 = 𝟏. 𝟗𝟔
𝟏𝟎
7𝟓 ∓ (𝟏. 𝟗𝟔)(𝟎. 𝟖)
7𝟓 ∓ 𝟏. 𝟓𝟔𝟖
We trust by 95% that 𝜇 ∈ (73.432 , 76.568)
95% confidence interval of 𝜇 is (73.432 , 76.568) , margin of error = 1.568
Note:
𝑺
Assuming (1 − 𝜶)100% confidence interval of the population mean (𝝁): 𝒙 ∓ 𝒁 𝜶
𝟐 𝒏

𝑺
Then, the margin of error is 𝒁𝜶 ,
𝟐 𝒏

margin of error increases as (n) decreases.

margin of error increases as (S) increases. And/or 𝒁𝜶 increases.

𝟐

Where 𝒁𝜶 increases if confident level ((1 − 𝜶)100%) increases i. e. when ( 𝜶 ) decreases

𝟐
Example:
Assuming a sample of 16 students, if the sample mean of student marks is 75.
Assuming the sample variance is 64
Obtain a 95% confidence interval of the population mean:
𝑺
𝒙 ∓ 𝒕𝜶
𝟐 𝒏
𝟔𝟒
7𝟓 ∓ 𝒕𝟎.𝟎𝟓
𝟐
𝟏𝟔
𝟖
𝟕𝟓 ∓ 𝒕𝟎.𝟎𝟐𝟓 , 𝒕𝟎.𝟎𝟐𝟓 ,𝟏𝟓 𝒅.𝒇. = 𝟐. 𝟏𝟑𝟏
𝟒
7𝟓 ∓ (𝟐. 𝟏𝟑𝟏)(𝟐)
7𝟓 ∓ 𝟒. 𝟐𝟔𝟐
We trust by 95% that 𝜇 ∈ (70.738 , 79.262)
95% confidence interval of 𝜇 is (70.738 , 79.262) , margin of error = 4.262
Example:
Assuming a sample of 30 students, if the proportion of failed student is 0.6.
Obtain a 90% confidence interval of the population proportion:

𝒑(𝟏 − 𝒑)
𝒑 ∓ 𝒁𝜶
𝟐 𝒏
𝟎.𝟔(𝟏−𝟎.𝟔)
0.6∓𝒁𝟎.𝟏𝟎
𝟐
𝟑𝟎

0.6∓𝒁𝟎.𝟎𝟓 𝟎. 𝟎𝟎𝟖 , 𝒁𝟎.𝟎𝟓 = 𝟏. 𝟔𝟒𝟓

0.6∓(𝟏. 𝟔𝟒𝟓) 𝟎. 𝟎𝟎𝟖
0.6∓𝟎. 𝟏𝟒𝟕𝟏𝟑𝟑𝟐𝟕
We trust by 90% that 𝒑 ∈ (0.4528667 , 0.74713327)
90% confidence interval of 𝒑 is (0.4528667 , 0.74713327) , margin of error = 0.14713327
Note:
𝒑(𝟏−𝒑)
Assuming (1 − 𝜶)100% confidence interval of the population proportion (𝑷): 𝒑 ∓ 𝒁 𝜶
𝟐 𝒏

𝒑(𝟏−𝒑)
Then, the margin of error is 𝒁𝜶 ,
𝟐 𝒏

margin of error increases as (n) decreases.

margin of error increases as (𝒑) near 0.5. And/or 𝒁𝜶 increases.

𝟐

Where 𝒁𝜶 increases if confident level ((1 − 𝜶)100%) increases i. e. when ( 𝜶 ) decreases

𝟐
90% confidence interval of 𝒑 is (0.4528667 , 0.74713327)

The confidence interval Range =

= Upper-Lower= 0.74713327 - 0.4528667 = 0.29426657 =(2* margin of error)
 Note: The Confidence Interval Range is a twice of (margin of error).
The confidence interval Range = Upper - Lower = (2* margin of error)

Example:
Assuming 95% confidence interval (C.I.) of the population mean of 100 students
is [62,84], find:
a) margin of error = ( C.I. Range ) / 2 = ( 84 – 62 ) / 2 = 11
b) Sample mean (𝒙 ∓ margin of error )
𝒙 = (Upper - marginal error) = ( 84 - 11) = 73 OR 𝒙 = (Lower + marginal error) = ( 62 + 11) = 73
73 is the sample mean
𝝈 𝝈
c) Population variance 𝒙 ∓ 𝒁𝜶 , the marginal error = 𝒁 𝜶 = 11
𝟐 𝒏 𝟐 𝒏
𝝈 𝝈
𝒁𝜶 = 𝟏. 𝟗𝟔 = 𝟏𝟏 ≫> 𝝈 = 𝟓𝟔. 𝟏𝟐 ≫ 𝒗𝒂𝒓𝒊𝒂𝒏𝒄𝒆 = 𝝈𝟐 = 𝟑𝟏𝟒𝟗. 𝟓𝟗
𝟐 𝒏 𝟏𝟎𝟎
Note:
The most 𝒁 values for Confidence Intervals:
from Normal Table

Assuming 95% confidence interval (C.I.) >>>> 𝒁𝟎.𝟎𝟓 = 1.96

𝟐

Assuming 90% confidence interval (C.I.) >>>> 𝒁𝟎.𝟏𝟎 = 1.645

𝟐

Assuming 85% confidence interval (C.I.) >>>> 𝒁𝟎.𝟏𝟓 = 1.44

𝟐

Assuming 80% confidence interval (C.I.) >>>> 𝒁𝟎.𝟐𝟎 = 1.28

𝟐
 Determining the Sample Size
A Case Based on Confidence intervals for population mean (𝝁):
point estimate ∓ margin of error
𝒙 ∓ margin of error
We can use the above formula to make an estimate value of the sample size (𝒏) that
provide a given determined [margin of error]:
𝝈
we use the confidence interval formula: 𝒙 ∓ 𝒁𝜶 .
𝟐 𝒏

𝝈
Then, assuming the [margin of error] is 𝒁𝜶
𝟐 𝒏

Where: 𝒁𝜶 is determined based on how you need a confidence level for your estimator.
𝟐

The value of the population standard deviation (𝝈) is determined using several ways:
The value of the population standard deviation (𝝈) is determined using
several ways:

A common method is to replace the population standard deviation (𝝈) by its

estimator from a sample(use (𝑺)).

If the sample standard deviation (𝑺) is unknown, we can use equation (8.3)
provided we have a preliminary or planning value for (𝑺) .

𝝈 𝑺
margin of error = E = 𝒁𝜶 → 𝑻𝒉𝒆𝒏 𝑬 = 𝒁𝜶
𝟐 𝒏 𝟐 𝒏
In practice, one of the following procedures can be chosen to determine (𝑺) :
1. Use the estimate of the population standard deviation computed from data of previous studies
as the planning value for (𝑺).
2. Use a pilot study to select a preliminary sample. The sample standard deviation from the
preliminary sample can be used as the planning value for (𝑺).
3. Use judgment or a “best guess” for the value of (𝑺). For example, we might begin by
estimating the largest and smallest data values in the population.
The difference between the largest and smallest values provides an estimate of the range for the
data. Finally, the [range divided by 4] is often suggested as a rough approximation of the
standard deviation (𝑺). and thus an acceptable planning value for (𝑺). 𝟐
𝑺
𝒁𝜶/𝟐 𝑺
𝑻𝒉𝒆𝒏 𝒕𝒉𝒆 𝒇𝒐𝒓𝒎𝒖𝒍𝒂 𝑬 = 𝒁𝜶 will be used to estimate (𝒏) at least: 𝒏≥
𝟐
𝟐 𝒏
𝑬
𝒁𝜶/𝟐 𝑺
Ex. if =57.2 Then  We need at least a sample size of 58
𝑬
 𝒁𝜶/𝟐 𝑺 𝟐
Note: 𝒏 ≥
𝑬

Then, the sample size is increases if 𝒁𝜶 increases.

𝟐

Where 𝒁𝜶 increases if confident level ((1 − 𝜶)100%) increases i. e. when ( 𝜶 ) decreases

𝟐
Example:
Suppose that we wanted to estimate the true average number of salmon fish eggs lays with 95%
confidence.
The margin of error we are willing to accept is (0.7).
Suppose we also know that the standard deviation (S) from a previous sample equals to (20).
What is sample size should we use?

𝟐
𝒁𝜶/𝟐 𝑺
𝒏≥
𝑬

𝟐
𝟏. 𝟗𝟔 ∗ 𝟐𝟎 𝟐
𝒏≥ = 𝟓𝟔 = 𝟑𝟏𝟑𝟔 𝐢𝐬 𝐭𝐡𝐞 𝐬𝐚𝐦𝐩𝐥𝐞 𝐬𝐢𝐳𝐞
𝟎. 𝟕

Exercise: solve the above example based on 80% confidence.

 Determining the Sample Size
A Case Based on Confidence intervals for population proportion (𝑷):
point estimate ∓ margin of error
𝐩 ∓ margin of error
We can use the above formula to make an estimate value of the sample size (𝒏) that
provide a given determined [margin of error]:

𝑷(𝟏−𝑷)
we use the confidence interval formula: 𝒑 ∓ 𝒁𝜶 .
𝟐 𝒏

𝑷(𝟏−𝑷)
Then, assuming the [margin of error] is 𝒁𝜶
𝟐 𝒏

Where: 𝒁𝜶 is determined based on how you need a confidence level for your estimator.
𝟐
The value of the population proportion (𝑷) is estimated by a planning value (𝑷∗ ) is determined using several ways:
In practice, one of the following procedures can be chosen to determine a
planning value (𝑷∗ ) for population proportion:
1. Use the sample proportion from a previous sample of the same or similar units.
2. Use a pilot study to select a preliminary sample. The sample proportion from this
sample can be used as the planning value, (𝑷∗ ).
3. Use judgment or a “best guess” for the value of (𝑷∗ ).
4. If none of the preceding alternatives apply, use a planning value of (𝑷∗ = 0.50).

𝑷∗ (𝟏−𝑷∗ )
𝑻𝒉𝒆𝒏 𝒕𝒉𝒆 𝒇𝒐𝒓𝒎𝒖𝒍𝒂 𝑬 = 𝒁𝜶 will be used to estimate (𝒏) at least :
𝟐 𝒏

𝟐
𝒁𝜶/𝟐 𝑷∗ 𝟏 − 𝑷∗
𝒏≥
𝑬 𝟐
The largest value of (𝑷∗ ) (𝟏 − 𝑷∗ ) is at the planning value (𝑷∗ = 𝟎. 𝟓)
Example:
Assuming that a random sample of 60 students from PTUK university has been given a
qualified exam, and 48 of these students had success in the exam by levels over than 70.

Construct a 95% confidence interval for the population proportion of students in PTUK who
𝟒𝟖
passed the qualifying exam with graded over than 70. 𝒑 = = 𝟎. 𝟖
𝟔𝟎

𝒑(𝟏 − 𝒑)
𝒑 ∓ 𝒁𝜶
𝟐 𝒏
𝟎.𝟖(𝟏−𝟎.𝟖)
0.8∓𝒁𝟎.𝟎𝟓
𝟐
𝟔𝟎

0.8∓𝒁𝟎.𝟎𝟐𝟓 𝟎. 𝟎𝟎𝟐𝟔𝟔𝟔𝟔𝟕 , 𝒁𝟎.𝟎𝟐𝟓 = 𝟏. 𝟗𝟔

0.8∓ 𝟏. 𝟗𝟔 𝟎. 𝟎𝟓𝟏𝟔𝟑𝟗𝟖
0.8∓𝟎. 𝟏𝟎𝟏𝟐𝟏𝟒
We trust by 95% that 𝒑 ∈ (0.698786035 , 0.901213965)
Example:
Assuming that a random sample of 100 students from PTUK university has been given a
qualified exam, and 78 of these students had success in the exam by levels over than 70.

Construct a 80% confidence interval for the population proportion of students in PTUK who
𝟕𝟖
passed the qualifying exam with graded over than 70. 𝒑 = = 𝟎. 𝟕𝟖
𝟏𝟎𝟎

𝒑(𝟏 − 𝒑)
𝒑 ∓ 𝒁𝜶
𝟐 𝒏
𝟎.𝟕𝟖(𝟏−𝟎.𝟕𝟖)
0.78∓𝒁𝟎.𝟐𝟎
𝟐
𝟏𝟎𝟎

0.78∓𝒁𝟎.𝟏𝟎 𝟎. 𝟎𝟎𝟏𝟕𝟏𝟔 , 𝒁𝟎.𝟎𝟐𝟓 = 𝟏. 𝟗𝟔

0.78∓ 𝟏. 𝟐𝟖 𝟎. 𝟎𝟒𝟏𝟒𝟐𝟓
0.78∓𝟎. 𝟎𝟓𝟑𝟎𝟐
We trust by 80% that 𝒑 ∈ (0.72698, 0.83302)
Example:
Assuming that you want to estimate the population proportion of defect devices. Assuming a
pilot study of 38 devices that contains 5 defected devices. Find the sample size that will be
used to estimate the proportion of defected devices such that we confidence it by 90% with
margin of error 0.206.

Solve:

𝟐
𝒁𝜶/𝟐 𝑷∗ 𝟏 − 𝑷∗
𝒏≥
𝑬 𝟐
𝟐 𝟓 𝟓
𝒁𝟎.𝟏𝟎/𝟐 𝟏 −
𝟑𝟖 𝟑𝟖
𝒏≥
𝟎. 𝟐𝟎𝟔 𝟐
𝟐 𝟓 𝟑𝟑
𝟏. 𝟔𝟒𝟓 𝟑𝟖 𝟑𝟖
𝒏≥
𝟎. 𝟐𝟎𝟔 𝟐

𝒏 ≥ 𝟕. 𝟐𝟖 >>> at least we need 8