turey's hsd is the way to go f you have a large number of groups and you want to make all possible comparisons. the downside is that it tends to be somewhat conservative with respect to Type I and Type II errors. on the other hand. the N-K can be a bit too liberal Turkey's Wholly Significant Difference tries to be like Momma Bear, and be "just right." if you have k group, use WSD if you want to do more than k-1 but fewer than kx(k-1)/2 comparison; otherwise, use the N-K