A Report Card for the New Election Predictor, Nate Silver
Nate Silver gained fame for correctly predicting all the state results in the 2008 presidential election. He repeated this success in 2012.
Silver is in much the same position today as was Rasmussen in 2008. With an perfect score, the standard deviation is zero.
Silver is just one of more than a dozen very accurate analysts who compiled and “crunched” results from other organizations’ opinion polls. This approach must trust the methods used by those pollsters, at least to be consistent so that the analyst can weigh and adjust those numbers accordingly. On the other hand, using a variety of sources can reduce the overall error if any one round of polling happens to contact a less-than-representative sample.
Recommending Reliable Pollsters for the Future
A statistician would expect that previously reliable pollsters would continue to predict well.
Gallup’s errors were reduced significantly in the 1956 presidential election, when Eisenhower was re-elected. Only the 1992 Clinton victory was over-stated with a margin of more than 3%, which is still within one standard deviation from Gallup’s current accuracy.
Rasmussen was perhaps too cautious this year, but should certainly not be discounted in future elections.
In the future, the gold standard for prescient prognosticators might become the aggregating analysts such as Silver, whose sterling performance relies on the hard work and diligence of reliable pollsters.
Gallup. Romney 49%, Obama 48% in Gallup’s Final Election Survey. (2012). Accessed November 15, 2012.
Politico. 2012 Election Results Map. (2012). Accessed November 15, 2012.
Rasmussen. Election 2012. (2012). Accessed November 15, 2012.