Elektrine lite

← Feed

@AlexGunz@c.im

2023-01-04 00:58 UTC

@uebernerd that post is so long I didn't make it to the end, but perhaps Fielder is confusing regression to the mean with Attenuation of Correlation. RtM requires initial selection bias. AoC occurs whenever you have imperfect measures of two things you are trying to correlate. RtM is relevant to replications, as it is about repeated attempts to estimate how big something is. AoC would suggest most effect sizes are probably slightly bigger than they are measured as - but with this being equally true of originals as of replications.

Replies (0)

No replies.