Tuesday, January 17, 2012

Going down the statistics rabbit hole

It all started with this post:
http://www.johndcook.com/blog/2012/01/13/interpreting-statistics/

which led to this post:
http://wmbriggs.com/blog/?p=5062

which led to this post:
http://wmbriggs.com/blog/?p=4381

which let to this post:
http://wmbriggs.com/blog/?p=4368

And finally to this hilarious, submitted, manuscript:
http://arxiv.org/abs/1201.2590

Reading them in this sequence will explain why the author, William M. Briggs, writes things like:
"When supplicants approach us for wisdom about uncertainty, because of our ardent love of computation we have developed the unfortunate habit of insisting first on the memorization of mathematical incantations, such as figuring chi-squares (by hand), or we require students look to up values in obscure tables in the backs of textbooks. In pursuit of our love, we forget why civilians come to us. Even we statisticians can forget why the math is there. Because of this, short shrift is paid to interpretation and to the limitations of the methods, to what it all means, which is all that civilians care about."
or something like this (and this is only on page 4 of the manuscript):
"This state of affairs is odd because frequentist theory tells us that the p-value is as silent as the tomb about the truth of the theory at hand. Yet when a civilian cocks his ear towards a wee p-value, he hears music. Angels sing."
Where is Brian Ripley when you need him? He has the same confidence of mind and writing style to potentially provide a valid counterpoint to the issue brought up by  William Briggs, and watching these 2 alpha males take a go at each other would be very entertaining and informative.

On another note, I wished ecological manuscripts were sometimes as personal in writing style, although I would not know what to do with a similarly written manuscript as an associated editor. I would definitely smile more when reviewing papers, that is for sure. Now we ecologists seem so boring.

No comments:

Post a Comment