Some election predictions nail outcome, others miss badly
For political prognosticators, their day of reckoning has arrived. Some are gloating about accurate predictions of the presidential election; others have their tails between their legs.
The New York Times’s Nate Silver, a celebrity in the world of predictive modeling, is looking smart. On his blog, FiveThirtyEight (named for the total number of electoral votes), Silver correctly called the winner in 49 states and the District of Columbia. In Florida, the 50th state, President Obama holds a narrow lead over Mitt Romney but has not yet been declared the winner. Silver forecast an Obama win.
Silver is a stat geek who developed the Player Empirical Comparison and Optimization Test Algorithm (PECOTA) and sold the evaluation model to Baseball Prospectus. He prides himself on dispassionate numerical analysis.
In an interview on the eve of Election Day, Silver explained his methodology to Comedy Central’s Stephen Colbert .
“Go and look at the polls and take an average and add up the states and see who has 270 electoral votes,” Silver said. “It’s not really that complicated, but people treat it like it’s Galileo or something.”
Indeed, the polls, taken together, were highly accurate in the presidential election. The final state-by-state polling averages compiled by the website RealClearPolitics suggested Obama would beat Romney, 303 electoral votes to 235. With the result in Florida still not finalized, Obama leads, 303-206.
A Romney win in the Sunshine State would mean the polling averages were perfect. An Obama win would mean the polls were 50-for-51.
Other election pickers missed badly. Republican strategist Karl Rove, a Fox News contributor, predicted an electoral college win for Romney, 285 to 253. Rove projected Romney victories in the battleground states of Virginia, New Hampshire, Ohio, Iowa and Colorado -- all of which the GOP nominee lost. Rove also said Romney would take Florida, where Obama appears to have won.
David Paleologos, director of the Suffolk University Political Research Center in Boston, surprised election observers last month when he said in an interview on Fox News that “in places like North Carolina, Virginia, and Florida, we’ve already painted those red. We’re not polling any of those states again.”
Romney won North Carolina by 2 points but lost Virginia by 3 and appears to have lost Florida, also.
The website PolicyMic posted the electoral college predictions of two of its top pundits, conservative Jesse Merkel and liberal Matthew Rozsa. Rozsa’s picks lined up with the RealClearPolitics polling averages exactly, and he projected an Obama win, 303 to 235.
Merkel, citing Republican enthusiasm, picked Romney in a landslide, 325 to 213, incorrectly projecting GOP wins in the same states Rove did, plus Pennsylvania, Wisconsin, and Minnesota.
Yale University economist Ray Fair forecast a Romney win, based not on polling but on economic data. Fair’s model -- which includes such factors as the per capita growth rate of the gross domestic product and inflation -- is typically very accurate. Before Tuesday, it had accounted for all but two presidential election winners since 1916.Callum Borchers can be reached at email@example.com. Follow him on Twitter @callumborchers.