Scribd
Upload
27 views

11.test of Significance

This document discusses different types of statistical tests used for hypothesis testing, including normal tests, t-tests, and chi-square tests. It provides examples of how to apply each test, including calculating test statistics and comparing them to critical values to determine whether to reject or fail to reject the null hypothesis. Sample problems demonstrate how to perform t-tests to compare two population means and a chi-square test to determine if two variables are independent.

Uploaded by

Tasbir Hasan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
27 views

11.test of Significance

This document discusses different types of statistical tests used for hypothesis testing, including normal tests, t-tests, and chi-square tests. It provides examples of how to apply each test, including calculating test statistics and comparing them to critical values to determine whether to reject or fail to reject the null hypothesis. Sample problems demonstrate how to perform t-tests to compare two population means and a chi-square test to determine if two variables are independent.

Uploaded by

Tasbir Hasan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Test of Significance

Test of Significance
Type-I Error: The Error of rejecting 𝐻0 (accepting 𝐻1 ) when it is true is called error of first kind or Type-I
Ex.: The test that shows a patient to have a disease where in fact the patient does not have disease
Type-II Error: The Error of accepting 𝐻0 when it is false(𝐻1 is true) is called error of 2nd kind or Type-II
Ex.: Blood test failing to detect a disease where in fact the patient does not have disease
The general procedure followed in testing hypothesis comprises the following steps :
 Set up a hypothesis
 Set up a suitable significance level
 Determination of a suitable test statistic
 Determine the critical region
 Doing computations
 Making decisions
Test

Some important widely used tests are as follows:


1. Normal test
2. t-test
3. Ӽ2 (chi −square) test
Normal test: Let u be the statistic whose expected value E(u) specified by the null hypothesis and its
𝑢−𝐸(𝑢)
standard error ϭ(u) is either known or can be estimated from the large sample (𝑛˃30) then 𝑑 = ,
σ(𝑢)
which is distributed normally with mean 0 and variance
t-test: If the population variance σ2 is known and the sample is small (n˂30) the appropriate test is
Student’s t-test.
ҧ 0
𝑥−𝜇 𝑥𝑖 −𝑥ҧ 2
The test statistic 𝑡 = 𝑠 with (n-1) d.f where 𝑠 2 = is the sample variance.
𝑛−1
𝑛

Ӽ2 (chi −square) test: Let 𝑂𝑖 be the observed frequencies and 𝐸𝑖 expected frequencies then the test
𝑂𝑖 −𝐸𝑖 2
statistic is 𝜒 2 =σ
𝐸𝑖
If the calculated value of Ӽ2 is less than the tabled value then null hypothesis is accepted.
Chi-Square Test
An investigator obtained the following data on the effectiveness of inoculation as a means of preventing from a disease.
Diseased Non-Diseased Total
Inoculated 5 20 25
No Inoculated 10 15 25
Total 15 35 50
Test whether the inoculation has preventive effect or not (Where at the 5% level of significance, the tabulated chi-squared
value with 1 d.f is 3.841)
Solution: 𝐻0 : Inoculation has no preventive effect (Inoculation and disease are independent). The expected frequencies are
𝑅1 ×𝐶1 25×15 𝑅1 ×𝐶2 25×35 𝑅2 ×𝐶1 25×15 𝑅2 ×𝐶2 25×35
𝐸11 = 𝑁
= 50
= 7.5 , 𝐸12 = 𝑁
= 50
= 17.5 , 𝐸21 = 𝑁
=50
= 7.5 and 𝐸12 = 𝑁
= 50
= 17.5
𝑂𝑖 𝐸𝑖 𝑂𝑖 − 𝐸𝑖 2 𝑂𝑖 − 𝐸𝑖 2 /𝐸𝑖
5 7.5 6.25 0.833
10 7.5 6.25 0.833
20 17.5 6.25 0.357
15 17.5 6.25 0.357
= 2.38

The calculated value of Ӽ2 =2.38 is less than the tabled value=3.841. Therefore the null hypothesis is accepted. Hence
Inoculation has no preventive effect.
t-test
A set of 15 observations and another set of 14 observations in lea strength are measured and the output
shown below: Mean (in kg) Standard Deviation (in kg) Number of tests
Set-1 66.48 2.40 15
Set-2 64.00 2.70 14
Use t-test to find whether two sets of data were drawn from population with same mean.(The tabulated t-
value with 27 df is 2.052)
Solution: Let 𝐻0 : two sets of data were drawn from population with same mean.
Given that 𝑛1 = 15, 𝑛2 = 14 , 𝑥ҧ1 = 66.48 , 𝑥ҧ2 = 64.00 , 𝑠1 = 2.40 and 𝑠2 = 2.70
𝑥ҧ 1 −𝑥ҧ 2 𝑛1 𝑛2 𝑛1 −1 𝑠1 2 + 𝑛2 −1 𝑠2 2
We know that 𝑡 = where 𝑠 = with 𝑛1 + 𝑛2 − 2 df
𝑠 𝑛1 +𝑛2 𝑛1 +𝑛2 −2

𝑛1 −1 𝑠1 2 + 𝑛2 −1 𝑠2 2 15−1 (2.40)2 + 14−1 (2.70)2


So 𝑠 = = = 2.55 with 27 df
𝑛1 +𝑛2 −2 27

𝑥ҧ 1 −𝑥ҧ 2 𝑛1 𝑛2 66.48−64.00 15.14


Hence 𝑡 = = = 2.617 >The tabulated t-value with 27 df is 2.052
𝑠 𝑛1 +𝑛2 2.55 29

Hence the null hypothesis is rejected. So two sets of data were not drawn from population with same mean.
t-test
The average salary of 17 employees in a textile company is 42 thousand taka and sample standard deviation
5 thousand taka. Now test whether the population mean salary is 50 thousand taka.(The tabulated t-value
with 16 df is 2.12)
Solution: Let 𝐻0 : 𝜇0 = 50
Given that 𝑛 = 17, 𝑥ҧ = 42, 𝑠 = 5,𝜇0 = 50
ҧ 0
𝑥−𝜇 42−50
We have, 𝑡 = 𝑠 = 5 = −11.314 with 16 d.f
𝑛 50

∴ 𝑡 = −11.314 < tabulated t-value.


So the null hypothesis is accepted.
Hence the population mean salary is 50 thousand taka.
Homework
1. The average daily return of a Textile company for 20 days of a month is 1.55 with a sample standard
deviation 0.53. At the 5% level of significant, test whether the population average return of the
company is equal to 1.8 or not.(Where at the 5% level of level of significance, the tabulated t-value
with 19 df is 2.093.)
Is there a significance difference in littering behavior between men and women in the given data?
Deposit Litter Total
Females 18 7 25
Males 42 33 75
Total 60 40 100

(Where the tabulated chi-squared value with 1 d.f is 0.16)

You might also like