a

is economical with your data: you need fewer respondents;

b

makes use of more information in your dataset than a test that is not conservative;

c

allows only for a small chance that you reject the null-hypothesis while it is actually true;

d

makes it easier for you to accept your hypotheses.