New blog post: minimal important differences for the social sciences - part one of two chrishanretty.co.uk/posts/mids_p...

The sequel is now up. Minimal important differences part two. Now features a fable! chrishanretty.co.uk/posts/mids_p...

it's kind of interesting how "smallest effect size of interest", relatively popular in psychology, doesn't seem to have taken off at all in political science

Ah, familiar territory for trials where we do power calcs and think about minimum detectable effect size (MDES). But the "detectable" and "substantively important" are often not the same thing / no one has thought about it.