Talk:Regression to the mean

From RationalWiki
Jump to: navigation, search

"When Aunt Jane's acne gets better..."[edit]

"When Aunt Jane's acne gets better after rubbing mint leaves on her face, that's "anecdotal evidence" based almost entirely on regression to the mean."

This is a simple misunderstanding of regression to the mean.

Regression to the mean can only occur when there are multiple cases measured, to provide a "mean". In the above case, we do not know if "getting better as much as Aunt Jane got better" is regression "to the mean" or movement "away from the mean".

Regression to the mean does not "cause" anything. Aunt Jane's acne got better. Something caused it to get better. Every effect has a cause.

The suggestion that getting better is "anecdotal evidence" suffers a similar misunderstanding. Aunt Jane's experience is a case study, not a clinical study. "Cases" are anecdotes. Every individual case, even an individual case in a clinical study, is an individual story, an anecdote.

It makes no difference if the case is in a clinical study, or Aunt Jane's bathroom. Every individual benefit or loss is an anecdote. Statistics come from collections of anecdotes. — Unsigned, by: Tracychess / talk / contribs