jeudi 29 janvier 2015

Correlation: The devil rides again?

Forty two years ago (!), Birnbaum1 argued that correlational research basically sucks (i.e., correlation is an “instrument of the devil,” especially when used to fit models). Differential psychology, however, is almost exclusively correlational, as we can’t randomly assign people to sex, race, age, personality, IQ, or pretty much anything we want to study.



Today, many believe correlational research basically sucks for two reasons: (1) Correlations don’t imply causation, and (2) squaring correlations usually results in embarrassingly-small amounts of variance explained.



The first argument is cowardly; it allows dismissal of any correlation (especially politically incorrect ones, or those inconsistent with confirmation-biased world views) as worthless, and not worthy of serious consideration. However, it is also trivial that cause implies correlation, and the field is starting to turn the tables on correlation skeptics:



Woodley’s Challenge: "Sure, correlation does not equal causation, but find me just one single instance of a causal relationship where there is no correlation (just one would suffice)2."...



to continue ;):



http://thegeneralfactor.blogspot.com...des-again.html





via International Skeptics Forum http://ift.tt/1ERIDVc

Aucun commentaire:

Enregistrer un commentaire