RE: Clinical trials21 Apr 2025 11:55
From a purely statistical point of view, what matters is:
1. Are the results significant? Which we are told they are. The 'p' values being low. Less than 5%.
2. How confident can we be in any claims? This is more pertinent with regard to sample size.
Taking relaxation as an example, let's say the claim is a 12% improvement. ( Not actual figure just for example). With a small sample, this could be say, 12% +/- 9%. With a larger sample this might be 12% +/- 3%. So the margin of error around the %age effect is much greater with smaller samples.
I guess marketing won't go into margins of error!! So they would quote the sample estimate. The 12% in the example above.