Post hoc power calculations for different test

mariko5797 — Mon, 11 Mar 2024 22:15:51 GMT

A study was powered for a non-inferiority test. However, the PI wants to know if a superiority test is possible now that we've seen the results, i.e., wants to claim one product is superior to the other since the GMT 95% CIs do not overlap.

(1) Is it worth checking what the power would be with our final sample size?

(2) If so, would I use the results I have (geometric mean and log10 SD) rather than theoretical outcomes?

/*sample code*/
proc power;
twosamplemeans test=diff nfrac
meandiff = Actual_MeanDiff
sides = 2
alpha = 0.05
groupstddevs= (Actual_SD1 Actual_SD2)
groupNs= (Actual_N1 Actual_N2)
power = .;
run;

Re: Post hoc power calculations for different test

Mike_N — Wed, 13 Mar 2024 13:43:54 GMT

You should not compute a post-hoc power analysis. A power analysis is only useful prior to the study to estimate a reasonable target enrollment number. Once the study is complete, you can either run the superiority test or not. You will have a better idea if reviewers in your field will accept a superiority test for a trial designed to detect non-inferiority.

topic Post hoc power calculations for different test in Statistical Procedures

Post hoc power calculations for different test

Re: Post hoc power calculations for different test