A comparison of Silver’s prediction and actual Electoral College results. Image: Twitter http://bit.ly/YVSKDV

 

Unless you’ve been living under a rock, you know by now that President Obama has been re-elected for a second term. You may not know, however, about Nate Silver, the man many are calling the election’s other big winner. Silver, a statistician and creator of the popular FiveThirtyEight blog, correctly predicted the Electoral College outcomes in all 50 states for this year’s presidential election (an improvement over his already impressive 2008 performance, when he was right on 49 states). His forecasts for the popular vote and Senatorial results were also surprisingly accurate.

Perhaps more interesting is that unlike other political pundits, Silver based his predictions on mathematical analysis of the poll results. This involved weighting the reliability of poll data, according to recency, sample size, and other factors. The resulting poll values were then adjusted, analyzed with a linear regression, and put through algorithms to determine election day projections. You can read in detail about his methods for the Senate forecast here.

Though his predictions were often ridiculed by other political forecasters, Tuesday’s results validated the reliability of statistical predictive modeling. Silver elaborates on this science in his book “The Signal and the Noise,” where he emphasizes the importance of distinguishing between relevant data and misleading information. The book describes how concepts of probability, regression analysis, and computer modeling can be applied to areas besides politics, including the economy, health, and global warming.  More about his book can be found here.

And though Nate Silver is the one garnering praise and popularity, he believes the focus should be on the modeling itself, and where it could lead us in the future. In an interview with Seth Borenstein of the Associated Press, Silver said:

“This is a victory for the stuff (computer modeling) in politics.  It doesn’t mean we’re going to solve world peace with a computer.  It doesn’t mean we’re going to be able to predict earthquakes…but we can chip away at the margins.”

It’s unclear where Silver’s methods will go from here, but Tuesday’s election can be seen as a resounding statement that we could soon have “political science” down to a science.

Comments:

NO COMMENTS

LEAVE A REPLY