In: Statistics and Probability
Suppose the Total Sum of Squares (SST) for a completely
randomzied design with ?=4 treatments and n=12 total measurements
is equal to 480. In each of the following cases, conduct an F-test
of the null hypothesis that the mean responses for the 4 treatments
are the same. Use α=0.025.
(a) The Treatment Sum of Squares (SSTR) is equal to 336 while the
Total Sum of Squares (SST) is equal to 480.
The test statistic is F=
The critical value is F=
The final conclusion is:
A. We can reject the null hypothesis that the mean
responses for the treatments are the same and accept the
alternative hypothesis that at least two treatment means
differ.
B. There is not sufficient evidence to reject the
null hypothesis that the mean responses for the treatments are the
same.
(b) The Treatment Sum of Squares (SSTR) is equal to 48 while the
Total Sum of Squares (SST) is equal to 480.
The test statistic is F=
The critical value is F=
The final conclusion is:
A. There is not sufficient evidence to reject the
null hypothesis that the mean responses for the treatments are the
same.
B. We can reject the null hypothesis that the mean
responses for the treatments are the same and accept the
alternative hypothesis that at least two treatment means
differ.
(c) The Treatment Sum of Squares (SSTR) is equal to 384 while the
Total Sum of Squares (SST) is equal to 480.
The test statistic is F=
The critical value is F=
The final conclusion is:
A. There is not sufficient evidence to reject the
null hypothesis that the mean responses for the treatments are the
same.
B. We can reject the null hypothesis that the mean
responses for the treatments are the same and accept the
alternative hypothesis that at least two treatment means
differ.
Ans:
df(treatments)=4-1=3
df(within)=12-4=8
MS=SS/df
F=MS(treatments)/MS(error)
When F>Fcritical,we reject null hypothesis.
1)
Source | df | SS | MS | F | Fcrit |
treatments | 3 | 336 | 112 | 6.222 | 5.416 |
within | 8 | 144 | 18 | ||
Total | 11 | 480 |
We can reject the null hypothesis that the mean responses for the treatments are the same and accept the alternative hypothesis that at least two treatment means differ.
2)
Source | df | SS | MS | F | Fcrit |
treatments | 3 | 48 | 16 | 0.296 | 5.416 |
within | 8 | 432 | 54 | ||
Total | 11 | 480 |
There is not sufficient evidence to reject the null hypothesis that the mean responses for the treatments are the same.
3)
Source | df | SS | MS | F | Fcrit |
treatments | 3 | 384 | 128 | 10.667 | 5.416 |
within | 8 | 96 | 12 | ||
Total | 11 | 480 |
We can reject the null hypothesis that the mean responses for the treatments are the same and accept the alternative hypothesis that at least two treatment means differ.