In: Statistics and Probability
5. (36 pts) Suppose three samples are selected from three populations as follows: Sample 1: 2, 6 Sample 2: 6, 8, 10 Sample 3: 9, 11, 13, 15 At the α = 0.01 level of significance, determine whether there is evidence of a difference in the population means.
Since there are three samples, we need to use an ANOVA for testing the equality of means.
Null hypothesis:
At least one pair of means are different.
Level of significance:
The given data is presented with the totals for samples and the overall total.
Sample1 | Sample 2 | Sample3 | |
2 | 6 | 9 | |
6 | 8 | 11 | |
10 | 13 | ||
15 | |||
8 | 24 | 48 | 80 |
We have the number of observations in each sample as
The total umber of observations is =2+3+4=9
We calculate the following :
1). Correction factor
2). Total sum of squares:
We make use of the table below for individual squares and add them to get the sum of squares.
Sample 1 | Sample 2 | Sample 3 | |
4 | 36 | 81 | |
36 | 64 | 121 | |
100 | 169 | ||
225 | |||
40 | 200 | 596 | 836 |
3). Sample sum of squares:
where is the total for ith sample.
8 | 24 | 48 | 80 |
64 | 576 | 2304 | 2944 |
32 | 192 | 576 | 800 |
4). Error sum of squares: ESS=TSS-SSS
ESS=124.8889-88.8889
5) Total df=9-1=8
Samles =3-1=2
Error=8-2=6
We shall now construct the ANOVA table:
Source of Variation | SS | df | MS | F | P-value | F crit |
Samples | 88.88889 | 2 | 44.44444 | 7.407407 | 0.023952 | 5.143253 |
Error | 36 | 6 | 6 | |||
Total | 124.8889 | 8 |
Since the F-calculated value is greater than the Critical value, we reject the Null hypothesis. Hence, we conclude that there is enough evidence for difference in population means.