Statistical Inference: Estimation

STATISTICAL INFERENCE : ESTIMATION
Objectives:
 Define a confidence and significance level

 Differentiate point estimates from interval estimates
 Solve for confidence interval estimation of the population mean,
population proportion and the ratio of two population variances.
What this module is all about
In the previous module, we learned the concepts classified under descriptive

statistics. In using this concepts, important information were obtained about the
sample under study. The measures of centrality give information about the typical
values of estimates.
However, if you want to know whether one group is significantly different

from another group in terms of some aspects, then. Descriptive statistics can no
longer be used in this purpose. This requires knowledge in inferential statistics.
Inferential Statistics
Techniques that allow us to make inferences about a population based on

data that we gather from a sample
Study results will vary from sample to sample strictly due to random chance
(i.e., sampling error)
Inferential statistics allow us to determine how likely it is to obtain a set of

results from a single sample
This is also known as testing for “statistical significance”

A. Confidence and Significance levels
Confidence level is the degree of assurance that a particular statistical

statement is correct, under specified conditions.
Significance level is the degree of uncertainty about the statistical

statement under the same conditions used to determine the confidence level.
Significance level are symbolized by α ( 0 ≤ α ≤ 1 ) and confidence levels by

1–α.
Confidence level + Significance level = 1
( 1–α) + α = 1
Statistical Estimation
Statistical estimation procedures use sample data to obtain the “best”

possible estimates of the corresponding population parameters. It furnish a
quantitative measure of the probable error involved in the estimation.
There are two types of estimates: point estimates and interval estimates.
Point Estimates, (p) use sample data to calculate a single “best” value which
estimates a population parameter. To be more precise, we use the sample mean as
a point of estimate of population mean and the sample proportion as a point
estimate of p because these statistics are unbiased and efficient estimators of their
respective parameters.
Interval Estimates gives a range of values that can be expected to include

the correct value a certain specified percentage of the time.
Confidence Interval
We use confidence of intervals to estimate a range of possible values of a

parameter, rather than a single value. When you use a confidence interval instead
of a point estimator, you lose a degree of precision but you gain a large degree of
confidence.
The general form of a confidence interval with a ( 1 – α ) confidence level is
P (lower limit ≤ parameter ≤ upper limit ) = 1 – α

where :
lower limit = point estimator - error of estimate
upper limit = point estimator + error estimate
The confidence intervals we will be dealing with will usually be 90%, 95%, or 99%
confidence intervals. Confidence intervals can be set up for any degree of
confidence.
Interval Estimation Procedures Confidence Interval
We use confidence of intervals to estimate a range of possible values of a

parameter, rather than a single value. When you use a confidence interval instead
of a point estimator, you lose a degree of precision but you gain a large degree of
confidence.
The general form of a confidence interval with a ( 1 – α ) confidence level is
P (lower limit ≤ parameter ≤ upper limit ) = 1 – α
where :
lower limit = point estimator - error of estimate
upper limit = point estimator + error estimate
The confidence intervals we will be dealing with will usually be 90%, 95%, or 99%
confidence intervals. Confidence intervals can be set up for any degree of
confidence.
B.
There are several steps to be followed in constructing an interval estimate.
 Decide on the risk you are willing to take of being wrong. This is also
called probability of error or alpha ( α ). The most commonly used value of
alpha in most researches is 0.05 or 5%. Using a probability of error (alpha)
of 0.05 means that a researcher is willing to be wrong only 5% of the time.
In other words, the probability of being right is 95%. This is also called
confidence level. So if alpha (α) or probability of error is 5% or 0.05, the
confidence level is 95% or 0.95.
 Picture the sampling distribution and divide the probability if error alpha into
upper and lower tail if the distribution. As shown in the previous diagram .
If α = 0.05, then α / 2 = 0.025
 Find the z - score that marks the beginnings of the lined areas. From the
previous module , we compute the z - score to find the area under the normal
curve. Here we do the reverse. We find for the Z - score. To do this use the
“Table of Areas of Normal Curve”
Locate 0.025 . the answer is Z = 1.96. Since the curve is symmetrical in both
upper and lower tails. Z = ± 1.96.
Now using different levels of confidence, verify the value of Z using the
“Table of Area of Normal Curve” for the give confidence levels.
Confidence Level Alpha (α ) α /2 Z – score
90% 0.10 0.05 ± 1.96
99 % 0.01 0.005 ± 2.58
C. Interval Estimation Procedures for Sample Means
In an interval estimate based on sample means, the formula used :
Confidence Interval = X ± Z [ ∂ / √ N ]
Where:
X = mean
∂ = standard deviation
N = no. of sample
Z = value in the table depending upon the

probability error alpha.
Example:
Assume that you want to estimate the average performance of government

employees on a given municipality using randomly selected sample of 100 with a
mean performance of 80 and standard deviation of 20.
N = 100 , X = 80 , ∂ = 20 α /2 = 0.05/2
Confidence Interval = 80 ± 1.96 [ 20 / √ 100]
= 80 + 1.96 [ 20/ 10] = 80 - 1.96 [ 20/ 10
= 80 + 1.96 [ 2] = 80 - 1.96 [2]
= 80 + 3.92 = 80 - 3.92
= 83.92 = 76.08
So the estimate or average performance of government employees is between 76 .

08 to 83.92.
Exercise 10
1. For each confidence level , determine the corresponding

Z – score.
Confidence Level Alpha (α) α /2 Z – score
95% 0.05 0.025 ± 1.96
96% _____ _____ ______
99% _____ _____ ______
90% _____ _____ ______
80% _____ _____ ______
2. For sample data below, construct three different interval estimates of

the population mean. For 90%, 95%, 99%
Given; X = 100, ∂ = 100 , N = 500

Statistical Inference: Estimation

Uploaded by

Copyright:

Available Formats

Statistical Inference: Estimation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Statistical Inference: Estimation

Uploaded by

Copyright:

Available Formats

STATISTICAL INFERENCE : ESTIMATION

 Define a confidence and significance level

What this module is all about

In the previous module, we learned the concepts classified under descriptive

However, if you want to know whether one group is significantly different

Techniques that allow us to make inferences about a population based on

Inferential statistics allow us to determine how likely it is to obtain a set of

This is also known as testing for “statistical significance”

Confidence level is the degree of assurance that a particular statistical

Significance level is the degree of uncertainty about the statistical

Significance level are symbolized by α ( 0 ≤ α ≤ 1 ) and confidence levels by

Confidence level + Significance level = 1

Statistical estimation procedures use sample data to obtain the “best”

Interval Estimates gives a range of values that can be expected to include

We use confidence of intervals to estimate a range of possible values of a

The general form of a confidence interval with a ( 1 – α ) confidence level is

P (lower limit ≤ parameter ≤ upper limit ) = 1 – α

lower limit = point estimator - error of estimate

upper limit = point estimator + error estimate

Interval Estimation Procedures Confidence Interval

We use confidence of intervals to estimate a range of possible values of a

The general form of a confidence interval with a ( 1 – α ) confidence level is

P (lower limit ≤ parameter ≤ upper limit ) = 1 – α

lower limit = point estimator - error of estimate

upper limit = point estimator + error estimate

There are several steps to be followed in constructing an interval estimate.

Confidence Level Alpha (α ) α /2 Z – score

90% 0.10 0.05 ± 1.96

99 % 0.01 0.005 ± 2.58

C. Interval Estimation Procedures for Sample Means

In an interval estimate based on sample means, the formula used :

Z = value in the table depending upon the

Assume that you want to estimate the average performance of government

Confidence Interval = 80 ± 1.96 [ 20 / √ 100]

= 80 + 1.96 [ 20/ 10] = 80 - 1.96 [ 20/ 10

= 80 + 1.96 [ 2] = 80 - 1.96 [2]

So the estimate or average performance of government employees is between 76 .

1. For each confidence level , determine the corresponding

Confidence Level Alpha (α) α /2 Z – score

95% 0.05 0.025 ± 1.96

96% _____ _____ ______

99% _____ _____ ______

90% _____ _____ ______

80% _____ _____ ______

2. For sample data below, construct three different interval estimates of

You might also like

96% ___ _ ____

99% ___ _ ____

90% ___ _ ____

80% ___ _ ____