February 11, 2009

voodoo baloney

You may have noticed recent media coverage of the charge that functional brain imaging in social neuroscience is beset by "voodoo correlations".

Without addressing the questions about statistics & etiquette that have been raised, let me just say that I am very grateful to Ed Vul for providing on his website a marvelous (he calls it "charming") 1950 paper (reporting on a talk given in 1949) that brilliantly and succinctly states the problem. The problem was the same, but the context was different: Sixty years ago, the context was not neuroimaging but rather personality tests, and the phrase wasn't "voodoo" but rather "baloney."

Edward Cureton's 1950 paper, entitled "Validity, Reliability, and Baloney," is only three pages long, and well worth the read.

No comments:

Post a Comment


Blog Archive