Discussion about this post

User's avatar
Dean's avatar

If I understand correctly, "P= .09" means there is a 91% chance that the effect was not by chance. And if there's a 91% probability that by taking a harmless drug for a few days I can reduce my chance of death by 70%, then how stupid would I be to refuse? The P=.05 cutoff is arbitrary and doesn't take into account risk vs. benefit, right?

Expand full comment
Betsy McDonel Herr, Ph.D.'s avatar

Another great example of undermining a real finding "the effect size" with an underpowered study "statistical significance." Do people ever do power analyses any more before they start a study, it was always required for any Ph.D. candidate.

Expand full comment
294 more comments...

No posts