And the Other Winner Is…

by Joey deVilla on November 7, 2012

…Nate Silver, who used data and did the math to accurately call the election. It all sounds very Star Trek-y: using a mathematical model that took in poll data, weighting each data source according to how accurate it’s been in the past, and factoring in other conditions that will affect the result. The talking heads in the media who cover the politics beat will have a harder time dismissing “cold equations” in favour of “warm human intuition” — as if mathematics wasn’t as equally human as “that gut feeling”.

The recent high-profile understandable-by-Joe-Average examples of how “doing the math” can pay off in real life — first Moneyball, now this — might spark a short-term interest in statistics and math in general. I’m not sure if we’ll ever see a CNN anchor explaining the differences between mean, median and mode to an audience that has trouble calculating a tip (I’d see it as a good start), but perhaps we might see more news features on how a little math can help improve your life, such as a piece on The Quantified Self, visualizing how much sugar is in that soft drink or the incredible power of compound interest. We don’t face threats that capture the American imagination the way that Sputnik did in 1957 and started the STEM (Science/Technology/Engineering/Mathematics) boom, but there are such threats to which greater STEM knowledge can lead us to solutions. A success story like Silver’s election prediction is just the kind of thing we need.

Macleans columnist Colby Cosh, one of my more seemingly unlikely internet friends, is one of Silver’s more reasonable critics. He’s taken some drubbing from Silver’s true believers (he refers to them as “Silverbacks”), but I think he makes a good point. While Silver’s method is arguably more “scientific” than a pundit’s gut instinct, he hasn’t done one thing that all good scientists and mathematicians do: he hasn’t shown his work. While we know the broad strokes of his model, the precise details are secret and thus the model is not falsifiable (a science term that’s often confused by laypeople — it means that if the model is false, there is a way to prove that it is so; this is a cornerstone of the scientific method). In a math exam, you don’t get full marks if you don’t show your work, and for this reason, Silver shouldn’t get full marks either (a decent mark, yes, but not a perfect score). There’s also the matter that like any predictive model, it won’t be correct all time, as Colby points out in this recent piece. Pure, blind faith in Silver’s model, just because it’s tagged with the impressive labels of “scientific” or “mathematical” is no different from believing the predictions of a psychic or a deck of tarot cards — you’re simply following a different mysterious force that you don’t really understand.

I’ll close with this: from one guy who juggles symbols for living to another — congratulations, Nate Silver!

Leave a Comment

{ 1 trackback }

Previous post:

Next post: