Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Confidence Interval Estimation

Module 2
Study Guide
Prepared By Engr. Emy Barrioquinto
2 Confidence Interval Estimation

Confidence Intervals – An Interval of values computed from sample data


that is likely to cover the true parameter of interest [1]. It tells how
confident you can be that the results from a survey reflect the
population in a certain degree [2].

Let’s say, we want to know the average weight of each apple in a huge
farm. For the lack of resources and/or time, it can be impossible to study
the entire population of apples. A sample of 100 apples are surveyed for
their weight. We desire to know if the resulting average weight taken
from the sample is somehow close to the weight of the population.

Confidence intervals provide an estimate of the unknown population


parameter, which we can estimate with some confidence.

Population Parameter – the numerical value that describes the


population. (μ) Not to be confused with;

Sample Statistic – the numerical value that describes the


sample of the population (x) It is called Point Estimate if the
parameter is a single value (sometimes they are used
interchangeably). In this module, we will use the term Point
Estimate often.

Since we cannot acquire the Population Mean (μ), we can say


that: (x) is a point estimator of the parameter (μ).
Concluding the Confidence Interval

“We are some level of percent confident that the population of interest
lies in the interval confidence interval values.”

“At something % confidence level, the population of interest lies in the


interval, confidence interval values”
Example 1: We are 95% confident that μ lies in the interval (-0.2, 3.1)

Example 2: At 95% confidence level, the mean zinc concentration in the


river lies in the interval, 2.502g/ml < μ < 2.698 g/ml

The usual Confidence levels used are 90%, 95%, and 99%, with 95% being
the famously used. It is solved by (1-α)x100%, where α (level of
significance) is the proportion in the tails of the distribution that is
outside the confidence interval.

The usual Population parameters to use are μ (mean), σ (stdev), and ρ


(proportion).

2.1 Solving the Confidence Intervals

The figure above shows the normal distribution illustrating the level of
confidence. The level of confidence is solved by subtracting 1 to the level
of significance.

Confidence Intervals can be solved using the following formula:

Confidence Interval=Point Estimate ± Margin of Error


Where the point estimate is usually sample mean, and the margin of
error is solved by:

Margin of Error =(Critical Value)(Standard Error)


σ
= Zα/2 ( )
√n
s
= tα/2 ( )
√n
Where the Critical Value is the value found in z or t table, and Standard
Error is the standard deviation of the point estimate [3]

2.1.1 How to get the critical value of Z-distribution


1. Identify the given Confidence Level (CL). (Usually 95%)
2. Get the corresponding z-value of the CL.
a. In excel, use the function =NORMSINV(CL)

b. If the interval is two-tailed, divide the level of


significance (α) into 2, then add to the confidence level.
Example, if the alpha of 95% is 5%, then 0.05/2 = 0.025.
Add this to 95% and you get 97.5%. Find the z-value.

*Important Note: The department requires 4-digit decimal points in


exams, and the problems are solved this way. Know that Excel uses
indefinite decimal points and the final answer solved using the Excel
method can differ.

As much as possible, use the manual method in answering department


exams.
2.1.2 How to get the critical value of t-distribution
1. Identify the Level of Significance (usually 5%)

2. Get the corresponding t-value.


a. In excel, use the function T.INV(α,degrees of freedom)

b. If the problem is two-tailed, use T.INV.2T(α,degrees of


freedom).

Degrees of freedom = n-1

2.1.3 How to know if the test is one-tailed or two-tailed


If the problem translates to “=”, the test is two-tailed.

If the problem translates to > or <, the test is one-tailed.

Example:

We claim that the average weight is 49.5 grams. (two-tailed)

We claim that the average weight is less than 49 grams (one-tailed)

We claim that the average weight is more than 49 grams (one-tailed)


2.1.4 How to get the critical value of Chi-Squared
1. Identify the Level of Significance (usually 5%)

2. Identify the degrees of freedom (n-1)


𝛼
3. For Lower Class Boundary Critical value, get
2
For example, if the level of significance is 0.01, the α of the lower-
class boundary is 0.005. (n=8)

𝛼
4 For Upper Class Boundary Critical value, get 1 -
2
For example, if the level of significance is 0.01, the α of the upper-
class boundary is 0.995

Note: For Excel,


2.2 Confidence Interval on the mean (x̄), variance is known

In this situation, our point estimator is the mean. To use this method, we
will require the following conditions:

• A simple random sample from the population of interest.


• A normally distributed population. (In this method, z-table will
be used.)
• Known population standard deviation/ variance [4]

Step 1: Identify the given values (n, x̄, σ)

Step 2: Identify z-value of the confidence level.

Step 3: Get the Confidence interval using the formula:


σ
Confidence Interval = x̄ ± Zα/2 ( )
√n

Or, if the problem is two-tailed, excel function can be used.

And then add/subtract the resulting margin to the point


estimate.

Step 4: State the claim

“We are some level of percent confident that the population of interest
lies in the interval confidence interval values.”

“At something % confidence level, the population of interest lies in the


interval, confidence interval values”
(1) Sample Problem: The average weight of the apples surveyed
from a sample of measurements taken from 36 apples is found
to be 49 grams. Find the 95% and 99% confidence intervals for
the mean weight of the apples in the farm. Assume that the
population standard deviation is 5.4 grams.

Step1: Identify the given values (n = 36, x̄ = 49, σ=5.4)

Step 2: Identify the z-value of the confidence levels (95% = 1.96, 99% =
2.58)

Step 3: Solve for the confidence intervals.

Or use excel:

Step 4: State the claim:

“At 95 % confidence level, the mean weight of the apples in the farm lies
in the interval, 47.2360 < µ < 50.7640”

Try solving the 99%, the margin of error should be 2.3182.


2.3 Confidence Interval on the mean (x̄), variance is unknown

In this situation, our point estimator is the mean. To use this method, we
will require the following conditions:

• A simple random sample from the population of interest.


• A normally distributed population. (In this method, t-table will
be used.)
• Unknown population standard deviation/ variance (In this
method, sample standard deviation will be used, which
introduces extra uncertainty. For that, we will use the t-table) [5]

Step 1: Identify the given values (n, x̄, s)

Step 2: Identify t-value of the Level of Significance and degrees of


freedom.

Step 3: Get the Confidence interval using the formula:


s
Confidence Interval = x̄ ± tα/2 ( )
√n

Or excel function can be used: CONFIDENCE.T (α, s, n)

And then add/subtract the resulting margin to the point


estimate.

Step 4: State the claim

“We are some level of percent confident that the population of interest
lies in the interval confidence interval values.”

“At something % confidence level, the population of interest lies in the


interval, confidence interval values”
(2) Sample Problem: The average weight of the apples surveyed
from a sample of measurements taken from 36 apples is found
to be 48 grams. Find the 95% and 99% confidence intervals for
the mean weight of the apples in the farm. The sample standard
deviation is 5.2

Step1: Identify the given values (n = 36, x̄ = 48, σ=5.2)

Step 2: Identify the t-value of the confidence levels

(95% = 2.0281, 99% = 2.7195)

Step 3: Solve for the confidence intervals.

Or use excel:

And then add/subtract the resulting margin to the point


estimate.
The answers differ because excel uses indefinite decimal points.

Step 4: State the claim:

“At 95 % confidence level, the mean weight of the apples in the farm lies
in the interval, 46.2423 < µ < 49.7577

Try solving the 99%, the margin of error should be 2.3606.


2.4 Confidence Interval on the Variance & Standard Deviation of a
Normal Distribution
In this situation, our point estimator is the variance or standard
deviation. The critical value will come from the chi-squared table [6]

Step 1: Identify the given values (n, σ)

Step 2: Identify the lower and upper boundary Critical Value of Chi-
squared.

Step 3: Get the Confidence interval using the formula:

Confidence Interval:
(𝒏−𝟏)𝒔𝟐 (𝒏−𝟏)𝒔𝟐
𝑿𝟐 𝒐𝒇 𝒍𝒐𝒘𝒆𝒓 𝒄𝒍𝒂𝒔𝒔
≤ σ2 ≤ 𝑿𝟐 𝒐𝒇 𝒖𝒑𝒑𝒆𝒓 𝒄𝒍𝒂𝒔𝒔

Apply square root to final answer for Standard Deviation.

Step 4: State the claim

“We are some level of percent confident that the Standard deviation lies
in the interval confidence interval values.”

“At something % confidence level, the Standard deviation lies in the


interval, confidence interval value.”
(3) Sample Problem: The standard deviation of the apples surveyed
from a sample of measurements taken from 11 apples is found
to be 4.934 grams. Find the 95% and 99% confidence intervals
for the standard deviation of these weights.

Step1: Identify the given values (n = 11, s = 4.934)

Step 2: Identify the Lower and Upper Class Boundary values of Chi-
Squared.

(Lower = 20.48, Upper = 3.25)

Step 3: Solve for the confidence intervals.

Step 4: State the claim:

“At 95 % confidence level, the Standard deviation lies in the interval,


3.4477 ≤ σ ≤8.6548.”
Try solving the 99%, the confidence interval should be 3.1088 ≤ σ
≤10.6163
2.5 Large Sample Confidence for Population Proportion
In this situation, our point estimator is the sample proportion p̂. The
critical value will come from the z-table. [7]

Step 1: Identify the given values (n, σ)

Step 2: Identify z-value of the confidence level.

Step 3: Get the Confidence interval using the formula:

Confidence Interval:

𝐩̂ (𝟏 − 𝐩̂
̂)
𝐩̂ ± 𝒁𝜶⁄ √
𝟐 𝒏

Step 4: State the claim

“We are some level of percent confident that the Standard deviation lies
in the interval confidence interval values.”

“At something % confidence level, the Proportion lies in the interval,


confidence interval value.”
(4) Sample Problem: From a sample of 100 apples, 10 pieces of
apples do not follow the specified weight. Construct a 95% and
99% two-sided confidence interval for p.
𝟏𝟎
Step1: Identify the given values (n = 100, p̂ = )
𝟏𝟎𝟎

Step 2: Identify the z-value of the confidence levels

(95% = 1.96, 99% = 2.58)

Step 3: Solve for the confidence intervals.

Step 4: State the claim:

“At 95 % confidence level, the confidence interval for the proportion of


the apples that are below recommended quality lie between 4.12% ≤ p ≤
15.88 %

Try solving the 99%, the confidence interval should be0.0301 ≤ p ≤0.1699
Sources

[1] Penn State Eberly College of Science, "Estimation and Confidence


Interval".

[2] statisticshowt, "Confidence Interval: How to Find it: The Easy


Way!".

[3] K. I. Cabaña, "Module 4a | Introduction to Confidence Interval


Estimation".

[4] K. I. Cabaña, "Module 4b | Confidence Interval Estimation on the


Mean, Variance is Known," Cebu, 2021.

[5] K. I. Cabaña, "Module 4c | Confidence Interval Estimation on the


Mean, Variance is Unknown," Cebu, 2021.

[6] K. I. Cabaña, "Module 4d | Confidence Interval Estimation on


Variance & Standard Deviatio," Cebu.

[7] K. I. Cabaña, "Module 4e | Large Sample Confidence Interval


Estimation on Population Proportion," Cebu, 2021.

You might also like