>It is a kind of weird trick

Statisticians hate this guy!

Expand full comment

1. Often one isn't interested in differences of any size but only those that exceed a meaningful size.

2. Too seldom the level of decisions is not explicitly taken into account. For individual decisions a much larger effect size might be necessary to justify the intervention (like to quit smoking) than on a population level (like to make cigarettes more expensive).

Expand full comment