In: Statistics and Probability
Toothpaste The table shows the costs per ounce (in dollars) for a sample of toothpastes exhibiting very good stain removal, good stain removal, and fair stain removal. At α = 0.05, can you conclude that at least one mean cost per ounce is different from the others?
|
Very Good |
.47 |
.49 |
.41 |
.37 |
.48 |
.51 |
|
Good |
.60 |
.64 |
.58 |
.75 |
.46 |
|
|
Fair |
.34 |
.46 |
.44 |
.60 |
Null Hypothesis, Ho:
There is no significant difference in the means cost per ounce for toothpaste exhibiting very good stain removal, good stain removal, and fair stain removal i.e.

Alternative Hypothesis, Ha:
There exists at least toothpaste for which the mean cost
per ounce is significantly different i.e.
for at least one
Test Statistics: Under Ho,

where k = number of groups = 3 and,
n = number of observations = 15


Calculations:
| Treatment Groups | ni | ![]() |
si |
| Very Good | 6 | 0.455 | 0.0536 |
| Good | 5 | 0.606 | 0.1048 |
| Fair | 4 | 0.46 | 0.1071 |


So,

and ,
Decision:
Since, p-value = 0.02951 <
, so we reject the Null Hypothesis at 5% level of
significance.
Conclusion:
There exists at least toothpaste for which the mean cost
per ounce is significantly different i.e.
for at least one