I’ve been following the current brouhaha about Nate Silver with interest, because it’s fascinating to me to watch people viscerally respond to probabilities. This is, in part, because some of my research deals with quantifying difficult decisions. So I get how people really rebel against this.
I like Silver’s work, though, because it’s relatively transparent. And, I confess, because I like his use of probabilities to describe the world. But Conor Friedersdorf says the best thing I’ve read on this subject (emphasis mine):
I seldom read Silver’s blog at the New York Times, having developed my own method for discerning with 100 percent accuracy the victor in every presidential, senate, house, and gubernatorial race in America. For reasons that are mysterious to me, the public lacks the patience for my “just wait until the votes are counted” method, preferring to obsess for months on end about who is up, who is down, and the precise statistical likelihood of still uncertain events. But if someone interested in that sort of thing asked me where to find it, I’d send them to Silver, both because there’s no one more skilled at what he does and because his work seems to have integrity. What I mean is that when Silver makes an electoral prediction, I trust he believes it. [He] seems extremely interested in excelling at his craft, and uninterested in influencing elections by using his prominence to exaggerate the chances of his favored candidate.
If someone, someday were to describe my work on this blog in this way, I’d be immensely proud.
[Cross-posted at The Incidental Economist]