Sunday, February 17, 2008

On estupidment

Jules Crittenden has an interesting post on the "estupidment" of science, among other fields of endeavor. I've argued the same basic premise for awhile. I'm interested what others might think.

-=DLH=-

Posted by at

1 comment:

chris j pluger said...

To me, this sentence seems spot-on: "On the other hand, what we often see in the news business is the enstupiding inverse, which is the exclusion of broader patterns from analyses which are based on attention-grabbing outlying anomalies."

The examples of housing slump and Iraq body count were appropriate. I thought more of the general trend in reporting (a la CBS evening news) about the "sagging economy" -- a phrase that gets whipped out every time growth doesn't meet expectations (since when is "grow" a synonym with "sag"?) or some other singular anomalous economic indicator gives what's-her-name a chance to gravely shake her head at the Impending Doom About To Befall Us.

But in general, doesn't this article say what all of us has known all along -- that you can make statistics say anything? I mean, throw out the anomalies so your data set looks clean, or else make a new data set out of the anomalies to make the general trends less obvious.

"78.4% of statistics are made up on the spot" -- or, might as well be, for all the real, objective data they represent.

It's like women's clothing sizes. She wears a 6. A six what? At least guys' clothes are measured in real, recognizable units of measure. Like inches, e.g.

February 18, 2008 at 8:39 AM

Post a Comment

Subscribe to: Post Comments (Atom)