In the run up to Election Day, most political pundits and pollsters were not projecting victory for Donald Trump in the U.S. presidential race.
Prior to this year, the forecasting project PollyVote accurately predicted the winner in each presidential election dating back to 1992.
One week after the General Election, WUWF called on University of West Florida political science professor Alfred Cuzan, co-founder of PollyVote, to see how it performed this time around.
On the pollyvote.com website, this was the last prediction headline: “Final PollyVote forecast: Clinton will win.”
“I’ll eat my share of crow,” Dr. Cuzan said, as he acknowledged that PollyVote missed the mark for the first time ever on predicting the actual winner. “But, let me just say that we were right and we were wrong.”
PollyVote actually projected that Democrat Hillary Clinton would win the popular vote, projecting she would get 52.5 percent of the two-party vote, meaning all the votes cast for her and Trump, the Republican. Clinton did win the popular vote, but PollyVote’s projection was off by a little over 2 percentage points.
“As far as I’m concerned we did well. We’ve done even better in the past, but we did fine,” said Cuzan. He pointed out that PollyVote’s result was actually better than the election-eve RealClearPolitics average, which was about three points off, and the Huffington Post’s Pollster.com, which was around four points off.
But, where PollyVote went wrong was with the Electoral College vote, which they tracked and projected for the first time in 2016.
“I think, you know, we just got a little too big for our britches,” Cuzan explaining that their success in forecasting the popular vote would translate to similar accuracy in the electoral vote.
“We had roughly the same kind of methods and components that we used in the popular vote, but what can I say, we got it wrong,” Cuzan said. “Everybody got it wrong.”
Cuzan believes one problem they had with predicting the electoral vote is that there’s less polling on the state level as opposed to the national level. And, he says pollsters got it wrong in several of the battleground states in the Midwest.
“And, that’s where basically he (Trump) won, Pennsylvania, Michigan, Wisconsin, Ohio, Iowa; you know he won all those states,” Cuzan said. Trump also came close in Minnesota and Virginia. He won Florida, which was going back and forth. Ultimately, PollyVote predicted Florida would go to Clinton by a narrow margin, showing the state as a very light shade of ‘Democratic’ blue on their Electoral College map.
