The actual amount of difference that exists between the two populations with unknown means (which you sample from) is the issue of

a. critical value

b. significance level

c. effect sized.

d. one- or two-tailed tests