That’s how pundits and pollsters who predicted a Hillary Clinton presidency ended up as the first day in President-elect Donald Trump’s America began.
Several experts say the GOP candidate’s seemingly longshot but successful bid for the nation’s highest office should prompt a period of reflection for those who consistently predicted the contest was Clinton’s to lose.
“I think the industry of polling will be introspective now because this was pretty much nationwide,” said Tim Malloy, assistant director of the Quinnipiac University Poll.
On Monday, RealClearPolitics’ polling average gave Clinton a 3.3 percentage-point edge over Trump based on data from 11 organizations such as CBS News, Bloomberg and Monmouth University Polling Institute.
The election’s outcome proved far different, with Trump declaring victory early Wednesday morning after winning key swing states such as Florida, Ohio and Pennsylvania. Clinton maintained a slim lead in the national popular vote as of Wednesday afternoon.
Patrick Basham, director of the self-described politically independent Democracy Institute, said the polling industry is losing credibility and that experts should spend more time focusing on the bigger picture and political climate surrounding elections instead of numbers and candidates. The Democracy Institute was in the minority Monday when it predicted a Trump victory. In June, the group also predicted the United Kingdom’s unexpected vote to leave the European Union.
“The winds were always going to be behind the challenger, and I think that because of the institutional and the financial advantages on the Clinton side and the nature of the Trump candidacy, a lot of people who were looking at the numbers lost that sense,” Basham said.
A Franklin & Marshall poll of 652 likely Pennsylvania voters released Nov. 1 gave Clinton an 11 percentage-point lead. The poll’s margin of error for likely voters was 5.1 percentage points.
Trump won Pennsylvania with nearly 49 percent of the vote, enough to give him slightly more than 2.89 million votes compared to Clinton’s roughly 2.82 million, according to unofficial election results.
G. Terry Madonna, director of Franklin & Marshall’s Center for Politics and Public Affairs, attributed the poll’s inaccuracy, in part, to about half of respondents participating online as opposed to by phone. Madonna said the poll findings released Nov. 1 explained that online respondents tend to be younger, more liberal and better educated.
With online polling in its relative infancy, pollsters need to adjust methodologies to correct differences borne from surveying modes, he said.
“We’re in an experimental mode with polls, and we’re trying to build for the future,” Madonna said. “We can’t just do polls the way that we are.”
Basham partly attributed the institute’s success to the use of accurate voter turnout models. He said the institute predicted Tuesday’s turnout would be about the same for both parties. Other polls gave Democrats a sizable edge, based largely on turnout from President Barrack Obama’s victories in 2012 and 2008.
Basham also said the institute tried to identify “shy” Trump voters by teasing out their political leanings through questions about their attitudes toward the world and other issues.
“The question was, is there 100,000 of them, or is there 10 or 15 million of them?” he said. “You can’t get it perfect, but we were able to tease (answers) out of folks who clearly were planning on voting for Trump but clearly weren’t going to say so.”
The Associated Press contributed.