In Nate Silver We Trust
On this past Tuesday, Barack Obama was elected to a second term as President of the United States. While he won the presidency, statistician and New York Times blogger Nate Silver won the prediction game that night. His forecasts weeks before the election claimed that Obama would win with almost 80% confidence, and then come election night, he correctly predicted the outcome of every single state. Last election in 2008, he correctly predicted 49 of the 50 states’ results, as well as all 35 Senate elections. Conservative, liberal, and independent media outlets are all proclaiming his genius, and his incredibly accurate predictions have called into question the common practice of “political pundits” trying to predict election results from gut feelings and perceived “success” of a political campaign. Nate Silver was able to so accurately predict the results for the election by employing one of the most powerful theorems in statistics. Bayes Theorem, and its use in Bayesian statistics, is defined as a methodology for assigning probabilities to future events given what you have already observed. He used millions of data points to try to model the way the election would turn out, the assumptions he would use in his models, and the number of different ways he could run each of his models. His blog, FiveThirtyEight, received millions of hits before, during, and after Election Day, everyone marveling at his ability to predict the future. Bayesian statistics helped him run many simulations of the election using the information he already knew, the information from past elections, and assumptions about this election.
We use Bayesian statistics in our class in the analysis of how people should react in an information cascade game. The same methodology that Nate Silver employed in his statistical models of the election apply to an information cascade. The key is distinguishing known and unknown information, and using what you know to help understand the probability of events happening in the future.
Referenced for information: http://today.duke.edu/2012/11/silver,
http://www.businessinsider.com/bayess-theorem-nate-silver-2012-9
http://today.msnbc.msn.com/id/49748540/ns/today-books/#.UJ0tk-Oe92s