Archive | Statistics RSS feed for this section

Correlation is not causation (nudge, nudge, wink, wink)

4 Jul

There is no end to the media presenting correlations as causation. I was struck by this piece in Monitor on Psychology a publication of the American Psychological Society. I think I see a pattern here:

  1. Point out some correlation
  2.  Say that the researchers “controlled for other variables” (always a dubious claim)
  3.  Quickly acknowledge that correlation is not causation
  4. Never report the magnitude of the effect
  5. Write the rest of the article as if causation has been established

 

%d bloggers like this: