Statistically Speaking

It may be the fault of my Econometrics professor in undergrad who talked about how he enjoyed single malts and parties in his undergraduate days while laughing about "statistical significance." He introduced me to EconoMagic.com, and the glint in his eye as he emphasized the latter half of the site's name coupled with the ease with which I always been able to "massage" the data likely both contributed to my skepticism.

So when I saw this article today, I couldn't help but smile:

Most Science Studies Appear to Be Tainted By Sloppy Analysis

The point of the article is clear from the title - those of us who groan every time someone says "they have proven _____" fill in the blank - are now not only backed by Taleb-like Fooled By Randomness talk, but also an article in the Wall Street Journal about research by a serious scientist.

The irony is of course obvious. Statistically proving that scientists are statistically irresponsible is just kind of funny to me for some reason.

I guess I just feel like more often than not we - scientists, economists, investors, researchers, theorizers - are wrong more often than we are right. Its just really hard to predict the future or come upon anything like the "truth".

That doesn't mean we should give up trying. But when Goldman's Global Alpha gets hammered a few months in a row and Greenspan admits that he saw the Subprime risk but couldn't translate it to a market reaction, it makes one wonder how much we really know about stuff as simple as the capital markets - much less important stuff like science.

Batter up for predicting the next move in the markets...my guess is we have only heard the first of a many shoes (or statistical gaffes) to drop.