To determine which specific group means are significantly different.

To check if the assumption of normality is met.

To increase the sample size of the groups.

To adjust the total sum of squares.