Is Nate Silver full of it? A growing legion of Republican-leaning commentators say he is. Silver is author of “The Signal and the Noise: Why So Many Predictions Fail—and Some Don’t.” But he’s more famous as the number-crunching guru who predicted the correct result in 49 of 50 states in the 2008 presidential election.
For a few weeks now, nervous Democrats have found solace in Silver’s election forecast on his FiveThirtyEight blog. Even as Mitt Romney surged to a lead in the Gallup and other national polls, Silver’s analysis of dozens of state polls, demographics and past voting trends listed President Obama as a favorite to win reelection. As of Friday morning, Silver forecasts an 80.9% chance of an Obama victory.
Earlier in the week, when Silver still forecast just a 73.6% chance of an Obama victory, Republican former congressman and MSNBC pundit Joe Scarborough tore into Silver, saying, “Anybody that thinks that this race is anything but a tossup right now is such an ideologue…they’re jokes.”
Indeed, among the pundit class, it’s the conventional wisdom to call the race a toss-up. But then again, if you read Silver’s book you’d probably never take anything a pundit says all that seriously ever again.
In the chapter “Are You Smarter Than a Television Pundit?” Silver relates how, on the eve of the 2008 election, two of the four television pundits on the McLaughlin Group said the race was too close to call, one said John McCain would win, and one predicted an Obama victory.
Obama went on to win the election by nearly 10 million votes. But no one ever called out the McLaughlin pundits on their poor prognosticating. So Silver analyzed 1,000 of the predictions made on the show. The pundits were right exactly half the time—which made their predictions no better than a coin flip. What’s more, none of the pundits on the show was better at making predictions than any other.
Silver goes on to describe how the same qualities that make a good and entertaining TV pundit also tend to blind people to the complexities inherent in analyzing large sets of data. Pundits believe in one “big idea” and their confident expression of that idea is what makes us listen to them.
Silver, by contrast, urges us to “stop and smell the data — slow down, and consider the imperfections in your thinking.” This doesn’t make him a compelling TV personality. Indeed, one mean-spirited, conservative blogger angered at FiveThirtyEight’s forecast sunk so low as to deride Silver for his “soft-sounding voice” and for being “thin and effeminate.”
The soft-spoken statistician emphasizes again and again that Obama’s victory is by no means certain and that his analysis provides a mere probability of a given result. Think of it this way: If a weather forecaster tells you there’s an 80.9% chance of rain, she isn’t guaranteeing it will rain. But she is saying you probably should take an umbrella when you step out the door.