Anova

Question 1:

ANOVA:

The data is divided into two parts, one part contains 8 observations and the other part contains 7 observations, the table below summarizes the data:

group 1

group 2

1

1/33

Anova

30

0

2

30

0

3

20

2/33

Anova

20

4

20

20

5

0

30

3/33

Anova

6

0

30

7

0

30

4/33

Anova

8

0

Total

100

130

230

N

5/33

Anova

8

7

15

mean

12.5

18.57143

The mean for group one data is 12.5 and for group 2 the mean is 18.57, in order to construct the ANOVA table the square of the observations in each group is determined and totals calculated, the table below summarizes the results:

6/33

Anova

group 1

x2

group 2

x2

1

30

900

0

0

7/33

Anova

2

30

900

0

0

3

20

400

20

400

8/33

Anova

4

20

400

20

400

5

0

0

30

900

9/33

Anova

6

0

0

30

900

7

0

0

30

10/33

Anova

900

8

0

0

0

Total

100

11/33

Anova

130

230

N

8

7

15

mean

12.5

12/33

Anova

18.57143

sum X2

2600

3500

6100

Sum x2 /n

13/33

Anova

1250

2414.286

3664.286

From the above grand total of X = 230

Total n = 15

The following values are determined using the data above:

14/33

Anova

value

A

(grand  total of X)2 / total n

= (230*230)/15 3526.667

B

total sum  of X2 for both groups

= 2600 +  3500

6100

C

sum of x2  group 1/n + sum of x2 for group 2/n

15/33

Anova

= 1250 +  2414.286

3664.285714

Using the above values the ANOVA table is determined as follows:

variation source

SS

df

between

C-A

1

within

16/33

Anova

B-C

13

total

B-A

14

The ANOVA table is as follows:

variation source

SS

df

ms

f

17/33

Anova

between

137.6190476

1

137.619

0.734506

within

2435.714286

13

187.3626

total

18/33

Anova

2573.333333

14

Given that the critical value from the F table when alpha = 0.05 = 4.6672, and that the F value calculated is 0.734506, the F critical value > F statistics value calculated, this means that there is no significant difference between the two halves.

Question 2:

Regression analysis:

The table below summarizes calculations undertaken to estimate the model of the form:

Y = a + BX

We use the formula:

19/33

Anova

B = ∑xy/ ∑x2

And a = Y’ – BX’ where Y’ and X’ are the mean values,

independent

dependent

X

20/33

Anova

Y

small y

small x

xy

x2

1

30

14.66667

-7

-102.667

49

21/33

Anova

2

30

14.66667

-6

-88

36

3

20

4.666667

22/33

Anova

-5

-23.3333

25

4

20

4.666667

-4

-18.6667

16

23/33

Anova

5

0

-15.3333

-3

46

9

6

0

-15.3333

-2

24/33

Anova

30.66667

4

7

0

-15.3333

-1

15.33333

1

8

25/33

Anova

0

-15.3333

0

0

0

9

0

-15.3333

1

-15.3333

1

26/33

Anova

10

0

-15.3333

2

-30.6667

4

11

20

4.666667

27/33

Anova

3

14

9

12

20

4.666667

4

18.66667

16

28/33

Anova

13

30

14.66667

5

73.33333

25

14

30

14.66667

6

29/33

Anova

88

36

15

30

14.66667

7

102.6667

49

TOTALS

120

30/33

Anova

230

110

280

MEAN

8

15.3333333

31/33

Anova

The above table shows total sums of x2 and xy where x = X-X’ and y=X-X’ where X and Y’ are the mean values.

∑xy=110 and ∑x2=280

B = 110/ 280

B =0.392857

Given that a = Y’ – BX’,

Then a =15.3333333-(0.392857*8)

a = 12.1904762

The estimated model is therefore:

y = 12.19 + 0.393X

This model means that as we increase X by one unit then Y increases by 0.393 units and that if the value of x was zero than Y value will be 12.19 given that we hold all other factors constant.

32/33

Anova

## REFERENCE:

Mendenhall, W. (2003) Introduction to statistics, Prentice Hall press, New  Jersey

33/33