Pollsters need a careful revamping of their methods


If it sounds a bit familiar that public opinion pollsters are going back to the drawing boards after missing the call of the 2016 presidential election …

It’s because you’ve heard it before.


Virtually ever “reputable” poll had Hillary Rodham Clinton winning the presidency on Nov. 8. Some had her winning by a fairly comfortable margin. She, of course, didn’t. Donald J. Trump is now preparing to become the next president.

Why is this familiar?

I recall the 2004 election in which President Bush won a second term over Sen. John F. Kerry. The sticking point that year was in Ohio, where exit pollsters had Kerry carrying the Buckeye State. Then the votes started pouring in. Bush won Ohio. He was re-elected. Kerry and his team were stunned. They thought they had Ohio in the bag. Had they won, they would have had just enough electoral votes to defeat the president.

Those dismal exit poll results, along with other misfires around the nation, signaled the end of Voter News Service, the outfit that coordinated all the polling and vote tabulation around the country.

The screw-ups this time were much more severe. Even the once-highly regarded FiveThirtyEight.com poll done by Nate Silver missed by a mile. Silver’s analysis had Clinton with a 71 percent chance of winning on he eve of the election.

Of course, many of the pollsters are trying to cover their backsides. They say they predicted Clinton’s national popular vote percentage, more or less. They missed, though, in several key battleground states where Trump won: Ohio, Michigan, Wisconsin, Pennsylvania, North Carolina, Florida — all states won by Barack Obama in 2008, who won all of them again except for North Carolina in 2012.

Polling has come a long way since the infamous “Dewey beats Truman” headline of 1948. However, as we witnessed during this election season, it still has some distance yet to travel.

Leave a Reply