I have 2 sets of independent (unpaired) continuous data and want to show the relative difference between their means (i.e group B mean is 40% of group A's mean) with a 95% confidence interval.

Calculating the % difference and SD for each group is obvious, but the calculation for CI includes the number of samples (I'm running with wiki's equation which I assume is appropriate: https://en.wikipedia.org/wiki/Confid...al#Basic_steps).

Do I use the number of samples in each group (30), or the total combined number of samples (60)?

Thanks, and sorry if this is an obvious answer!