Scribd
Upload
15 views6 pages

F-Test for Variance Comparison in Statistics

qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq

Uploaded by

nikhilrnikhilr97
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
15 views6 pages

F-Test for Variance Comparison in Statistics

qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq

Uploaded by

nikhilrnikhilr97
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

CHI-SQUARE DISTRIBUTION

BANGALORE INSTITUTE OF TECHNOLOGY


Department of Mathematics
Mathematics for Computer Science (BCS301)

Module 4: Statistical Inference - II

F Test or Variance Ratio Test:


F Test is used to find whether the two independent estimates of population
variance differ significantly or whether the two samples may be regarded as drawn
from the normal population having same variance.
To calculate the test of significance for two samples of size N1 and N 2 drawn
from normal population having same variance we calculate the ratio F defined as
S 12 L arg er estimateof var iance
F 2
 with degree of freedom (df)  1  N1  1 and
S 2 Smaller estimateof var iance
 2  N2 1

where  1 = df with larger variance


 2 = df with smaller variance

S 
2 x  X  2

N1  1
1

S 
2 y  Y 
2

N2 1
2

Note:
1. If Fcal  Ftab then hypothesis is accepted.
2. For two samples of x and y variate, Mean of each variate is given by
x y
X and Y  
N1 N2

3. Variance of a sample s   2 x  X 2

.
N 1

DEPARTMENT OF MATHEMATICS , B.I.T 1


CHI-SQUARE DISTRIBUTION

4. Variance of a population    2 x  X 2

.
N

Problems:
1. In a sample of 8 observations, the sum of squared deviations of items
from the mean was 84.4. In another sample of 10 observations, the value
was found to be 102.6. Test whether the difference is significant at 5%
level.
You are given that at 5% level, critical value of F for  1  7 and  2  9
degree of freedom is 3.29 and for  1  8 and  2  10 its value is 3.07

Solution: H 0 = There is no significant difference between the variance of the two


population. i.e.,  12   22 .

H1 = There is significant difference between the variance of the two population.

Given the sum of squared deviations from the mean is 84.4   x  X 2  84.4

and the sum of squared deviations from another mean is 102.6   y  Y 2  102.6

N1  8 N 2  10

S 
2 x  X 
2


84.4
 12.06 S 
2 y  Y 
2


102.6
 11.4
N1  1 N2 1
x y
7 9

Since S 2x  l arg er estimate  12.06  S 12 and S 2y  smaller estimate  11.4  S 22

S 12
 1.06 1
L arg er estimateof var iance 12.6
F - Test for testing H 0 is F  2  
S 2 Smaller estimateof var iance 11.4

Given that for  1  7 and  2  9 degree of freedom is 3.29 , Ftab  3.29 2

From 1 and 2 Fcal  Ftab

 H 0 is accepted.

The other data for  1  8 and  2  10 its value is 3.07 cannot be considered

DEPARTMENT OF MATHEMATICS , B.I.T 2


CHI-SQUARE DISTRIBUTION

2. An instructor has two classes, A and B, in a particular subject. Class A


has 16 students while class B has 25 students. On the same examination,
although there was no significant difference in mean grades, class A had
a standard deviation of 9 while class B had a standard deviation of 12.
Can we conclude at the 0.05 level of significance that the variability of
class B is greater than that of A?

Solution: H 0 = There is no significant difference between the variability of the


class B and class A. i.e.,  12   22 .

H1 = There is significant difference between the variability between two classes

Given for 16 students, standard deviation is 9  s A  9  s 2 A  81

Given for 25 students, standard deviation is 12  s B  12  s 2 B  144


Since S 2B  l arg er estimate  144  S 12 and S 2A  smaller estimate  81  S 2
2

2
 1.77 1
S L arg er estimateof var iance 144
F - Test for testing H 0 is F  1
2
 
S 2 Smaller estimateof var iance 81

Given that for  1  24 and  2  15 degree of freedom is 2.29 , Ftab  2.29 2

From 1 and 2 Fcal  Ftab

 H 0 is accepted.

Conclusion: There is no significant difference between the variability of the class B


and class A.

3. Two random samples were drawn from two populations and their
values are:
A: 66 67 75 76 82 84 88 90 92 -- --
B 64 66 74 78 82 85 87 92 93 95 97
Test whether the two populations have the same variance at the 5% level
of significance (F=3.36) at 5% level for  1  10 and  2  8
Solution: H 0 = There is no difference between the variance of the two population.

DEPARTMENT OF MATHEMATICS , B.I.T 3


CHI-SQUARE DISTRIBUTION

i.e.,  12   22 .

H1 = There is difference between the variance of the two population.

A: B
x x X x  X  2
y y Y y  Y  2

66 -14 196 64 -19 361


67 -13 169 66 -17 289
75 -5 25 74 -9 81
76 -4 16 78 -5 25
82 2 4 82 -1 1
84 4 16 85 2 4
88 8 64 87 4 16
90 10 100 92 9 81
92 12 144 93 10 100
95 12 144
97 14 196
 x  720  x  X   y  913  y  Y 
2 2
 720  1298

N1  9 N 2  11

X 
 x  720  80 Y 
 y  913  83
N1 9 N2 11

S 
2 x  X 2


734
 91.75 S 
2 y  Y  2


1298
 129.8
N1  1 N2 1
x y
8 10

Since S 2y  l arg er estimate  129.8  S 12 and S 2x  smallerestimate  91.75  S 22

S 12
 1.415 1
L arg er estimateof var iance 129.8
F - Test for testing H 0 is F  2
 
S 2 Smaller estimateof var iance 91.75

Given that for  1  10 and  2  8 , Ftab  3.36 2

From 1 and 2 Fcal  Ftab

 H 0 is accepted.

DEPARTMENT OF MATHEMATICS , B.I.T 4


CHI-SQUARE DISTRIBUTION

Conclusion: The two populations have the same variance.


4. The following data present yields in quintals of common ten
subdivisions of equal area of two agricultural plots:
Plot 1 : 6.2 5.7 6.5 6.0 6.3 5.8 5.7 6.0 6.0 5.8
Plot 2 : 5.6 5.9 5.6 5.7 5.8 5.7 6.0 5.5 5.7 5.5
Test whether two samples taken from two populations have the same
variance (5% point of F for  1  9 and  2  9 is 3.18)
Solution: H 0 = There is no difference between the variance of the two population.
i.e.,  12   22 .

H1 = There is difference between the variance of the two population.

Plot 1 : Plot 2 :
x x X x  X  2
y y Y y  Y  2

6.2 0.2 0.04 5.6 -0.1 0.01


5.7 -0.3 0.09 5.9 0.2 0.04
6.5 0.5 0.25 5.6 -0.1 0.01
6.0 0 0 5.7 0 0
6.3 0.3 0.09 5.8 0.1 0.01
5.8 -0.2 0.04 5.7 0 0
5.7 -0.3 0.09 6.0 0.3 0.03
6.0 0 0 5.5 -0.2 0.04
6.0 0 0 5.7 0 0
5.8 -0.2 0.04 5.5 -0.2 0.04
 x  60  x  X   y  57  y  Y 
2 2
 0.640  0.24

N1  10 N 2  10

X 
 x  60  6 Y 
 y  57  5.7
N1 10 N2 10

S 
2 x  X 
2
0.640
 0.071 S 2 y  Y 
2


0.24
 0.027
N1  1 N2 1
x y
9 9

Since S 2x  l arg er estimate  0.071  S 12 and S 2y  smaller estimate  0.027  S 22

DEPARTMENT OF MATHEMATICS , B.I.T 5


CHI-SQUARE DISTRIBUTION

S 12
 2.63 1
L arg er estimateof var iance 0.071
F - Test for testing H 0 is F  2
 
S 2 Smaller estimateof var iance 0.027

Given that for  1  9 and  2  9 , Ftab  3.18 2

From 1 and 2 Fcal  Ftab

 H 0 is accepted.

Conclusion: The two populations have the same variance.

DEPARTMENT OF MATHEMATICS , B.I.T 6

Common questions

Powered by AI

The primary purpose of conducting an F-Test in statistical analysis is to determine whether two independent estimates of population variance differ significantly. It tests if the samples may be drawn from a normal population with the same variance .

Failing to reject the null hypothesis in an F-Test comparing two classes' variability implies that the observed differences in variability are not statistically significant. It suggests that any variability between the classes could be due to random sampling error rather than systematic differences, impacting decisions on resource distribution or teaching methods .

Degrees of freedom affect the calculation of an F-Test by influencing the critical value against which the calculated F statistic is compared. The degrees of freedom for the samples are derived from the sample sizes and are essential in determining the validity of the F-Test. They are calculated as N1-1 for the sample with the larger variance and N2-1 for the sample with the smaller variance .

If the calculated F statistic is less than the tabulated F value in an F-Test, the null hypothesis is accepted, indicating that there is no significant difference in variances between the two populations being tested .

Acceptance of the null hypothesis in testing yield variance between two plots suggests that the yield variance is similar across plots, impacting agricultural policymaking by supporting uniform policy approaches. Resources and interventions can be standardized across plots without concern for increased variability, optimizing efficiency .

Using the sum of squared deviations in calculating variance for the F-Test is relevant because it provides a measure of the data's dispersion around the mean. This sum helps in determining the variance, which is critical in comparing variances between two samples and ultimately testing if they are significantly different .

An F-Test can be used to determine if the variability between two classes significantly differs even when their mean grades are the same. The null hypothesis would state that there is no significant difference between the variability of the two classes, implying equal population variances .

If an F-Test concludes that two agricultural plots have the same variance, it implies that the variability in yield from the plots is statistically similar, suggesting that resource allocation and management practices do not significantly affect the variability between these plots. This may guide decisions in uniform resource distribution .

The level of significance, such as 5%, is important in statistical tests like the F-Test because it defines the probability threshold for rejecting the null hypothesis. It helps determine the likelihood that the observed variance differences occurred by chance, guiding the conclusions about statistical significance .

The F-Test leverages variance in hypothesis testing by comparing the ratio of two sample variances. This comparison helps determine if the observed variances are significantly different, informing decisions about whether population parameters, such as variance, can be considered equal or distinct. Such information is critical for researchers to infer population characteristics accurately .

You might also like