Andrew,

Yes, sorry that was an accidental error on my part. Thank you for the clarification.

I may be misunderstanding this, but when we run paired t-tests in NBS with M # of permutations, is that analogous to a paired two-sample for means test in Excel? In running paired t-tests in NBS between group 1 and group 2 to test for differences between before and after, I didn't get a significant p-value (< 0.05) when assessing all of the significant connections identified.

My aim is to test effects on a subject population from before (group 1) and after (group 2), so I need to account for repeated measures. When running a paired two-sample for means test in Excel with group 1 and group 2 using all of the significant connections identified when running a one-sample t-test on group 1 separately and group 2 separately, I get a p-value of 6.40*10^-5, indicating significant changes.

Could you clarify the discrepancy between the 2 results?

Regards,

Anita