Info 202 F12

With the U.S. election day just two months away, news organizations, blogs, and social networks are abuzz with a constant state of attentiveness to the latest developments in the high-profile races for political power. Political polling is an arena that is now omnipresent in how elections are discussed by the media, as well as in the ways campaigns are conducted. While much of the political coverage in elections has traditionally been based around polling data, the practice of collecting, interpreting and presenting that data has always been, by its very nature, an inexact science. To the amateur political observer such as myself, polling has often been considered intriguing but often unreliable due to its inherent limitations, frequent fluctuations, and, occasionally, biased sources.

Enter Nate Silver and his political analysis and prediction website, FiveThirtyEight. A former baseball statistician, Silver emerged during the 2008 presidential election cycle as an innovator in helping political observers make sense of polls and trends, effectively stripping the noise filter of political spin machines and biased media outlets to create a clearer picture of the state of the race.

FiveThirtyEight’s forecast model, which is based in statistics and data analytics, takes several factors into account when issuing its predictions, one of which is a weighted average of available polling data. Rather than just taking a pollster’s numbers at face value, the model uses three factors to weigh a poll result: recency, sample size and a pollster’s rating; The weight assigned to a poll is “designed to be proportional to the predictive power that it should have in anticipating the results of upcoming elections" (see link above). By combining this information with other factors including demographics, economic trends, and several other layers of analysis, Silver’s model creates an innovative method of organizing, analyzing and interpreting the available information that has thus far proven to be impressively accurate: in the 2008 elections his forecast accurately predicted every Senate result, and called every state except Indiana and Nebraska’s 2nd district in the presidential election.