If the differences among samples are large, the difference of categories often seems unclearly. Paired t-test is used for the case.
With pair information is such case that
The diagram is an example. It is the case to compare the difference between the average of A and B. In the left diagram, we cannot confirm the difference numerically without pair information testing.
In the right diagram, even if we use pair information, P-value is not small because unevenness ( outlier ) is large.
You may find that one of the data of B is too large. It may be good that we remove the data from testing as an outlier.
NEXT Hypothesis Testing for Diffrence of DispersionTweet