We weigh polls differently based on their previous general election accuracy.
We toss polls that have a Margin of Error (MoE) that exceeds +/-5%.
We further take in to consideration the pollsters assumptions of the makeup of the electorate by considering the average PVI for the State polled.
Next we look at momentum of each of the candidates. This is similar to momentum investing in stocks. We look at the polling data and assign a momentum figure based on statistics that suggest whether a candidate has a positive, negative, or neutral momentum that could translate into a higher percentage outcome on Election Day. This is the reason we call our model a “projection” rather than “prediction.” One could argue for either word.
While Nate Silver at FiveThrityEight says he adds a “secret sauce” to come up with his predictions for election outcomes, we too have that additional secret something. Since we are located in the South, we will equate our “special formulations” to the secret formula in Cocoa Cola.
In the Virginia Gubernatorial race to be held on November 5th, 2013 we See normal upward momentum for both the Democrat Terry McAuliffe and Republican Ken Cuccinelli as voters more from undecided to making a final decision before the election. The Libertarian candidate Robert Sarvis has some downward pressure as voters realize that although they are unhappy with the current state of a two party system, have come to the realization that a vote for the Libertarian candidate will ultimately put a Democrat in the Governors Mansion.
Based on pure poll aggregation and weighting, Democrat Terry McAuliffe would win by a seven (7) point margin with 48.7% of the vote. But by judging past Libertarian voting behavior in Virginia, we think that the 9.2% current polling for Sarvis is overstated.
Based on our careful analysis the Politisite Political Projection is as Follows:
Read Rest at Politisite