29 August 2015

Tom Cruise Was Right. Damn!

I'm among those who don't watch daytime TeeVee, save for late afternoon ESPN (though that may stop now that they've canned Olbermann), so I didn't see Tom jump up and down and scream and yell on Oprah's show, and, no I haven't watched the YouTube feed either. Just seen the clips on the news. Doesn't mean all that much to me.

What does matter is Tom's indoctrination against all things psych-. I have some kinship on that point. R-bloggers is awash in posts dealing with marketing and sales schemes built on mind-bending adverts. Bah. Doesn't seem quite fair to manipulate folks into buying worthless widgets through emotional stratagems. Not everybody agrees, since the Big Data crowd make lots of money doing so.

Well, may be it's all a case of an emperor and his new clothes. It's worth noting that many of the methods widely used in quants and stats came the psych- fields.
Now, a painstaking yearslong effort to reproduce 100 studies published in three leading psychology journals has found that more than half of the findings did not hold up when retested.

Could it be that all that commercially driven quant is equally frail? Even supposedly hard science is susceptible:
Stefano Bertuzzi, the executive director of the American Society for Cell Biology, said that the effort was long overdue, given that biology has some of the same publication biases as psychology. "I call it cartoon biology, where there's this pressure to publish cleaner, simpler results that don't tell the entire story, in all its complexity," Dr. Bertuzzi said.

Now, of course, the psycho-adverts crowd can rebut with, "but we get results, and that's all that counts". But are the results because of your work, or just something in the ether? These quants learned their chops in school, doing just these sorts of studies. If the methods and designs are squishy from the outset, may haps they still are when Mr. Market pays $$$ to conduct the self-same protocol to sell Depends?

No comments: