[mrbungie]

Power is not just about sample size, but also (expected/previously informed by some other evidence) effect size. You can't make that conclusion without that.

[iamacyborg]

For sure, but you’d need one hell of a good cta to be getting a sufficient effect size to warrant small samples.