By Giovanni Russonello – – –
Pre-election polls in numerous countries this year have widely missed their marks, often by underestimating support for candidates on the ideological fringes. The polling failures in countries like Britain, Poland and Israel point to technical issues that could well foreshadow polling problems in the United States, many analysts believe.
“The industry has a collective failure problem,” said John Curtice, the president of the British Polling Council and a professor of politics at the University of Strathclyde in Glasgow. Partly this is the result of changing methodologies. “It’s now a mix of random-digit dialing — that is, telephone polls — and Internet-based polls based on recruited panels,” he said. Both modes present potential problems.
Opinion polls in advance of Britain’s general election in May severely underestimated the number of seats that Prime Minister David Cameron’s Conservative Party would win. After the election, the polling council called for an independent inquiry into what had caused the error. The council plans to release its findings in mid-January, a report that will be closely read by pollsters in Britain and around the globe.
Days after the British election, Polish voters elected Andrzej Duda of the far-right Law and Justice Party as president; no major poll had given him a lead. Poland’s Association of Market and Opinion Research Organizations is investigating what led to the collective flop. Similar slip-ups occurred in the Greek referendum on austerity, the national elections in Israel and Turkey, and elsewhere.
The British and Polish post-mortems are unlikely to yield a satisfactory answer, said Scott Keeter, director of survey research at the Pew Research Center. “It’s not clear that there’s any single, consistent story running through this,” he said.
To some degree, conventional method explanations can help understand what has gone wrong in Europe. Frustration with the political process has recently given rise to more populist and anti-establishment candidates — and survey respondents sometimes shy away from acknowledging support for candidates seen as outside the mainstream. In addition, as people grow more distrustful of the political process, predicting turnout has grown increasingly difficult.
“The people who refuse to tell you what they’re going to do are disproportionately likely to vote Conservative,” Mr. Curtice said.
Most of the flubbed European polls seemed to misconstrue voters’ allegiances within conservative voting blocs. Prime Minister Benjamin Netanyahu’s Likud Party won decisively in March’s legislative election in Israel, despite polls showing a dead heat with the center-left Zionist Union — largely because Mr. Netanyahu was able to draw voters expected to go to other right-wing parties. And in Poland in May, the hard-line Mr. Duda pulled votes away from the center-right incumbent Bronislaw Komorowski, finishing ahead of him in the preliminary election and edging him out in a runoff.
General election polling in the United States typically doesn’t fall victim to factional conflicts, because in most cases only two major candidates and parties are represented on the ballot. But in the Republican nominating contest, where turnout is uncertain and a dozen candidates are vying for conservatives’ hearts, strong parallels abound.
“I think it’s like it is in Israel: It’s a within-the-bloc primary,” said Stanley Greenberg, who runs the polling firm Greenberg Quinlan Rosner. At the moment, Ben Carson is the candidate with the most appeal among evangelical voters, “but it has all the makings of being able to shift to Cruz,” he said, referring to Senator Ted Cruz of Texas. “And I do think it can break late.”
Another issue has been turnout. In the British election, Labour Party supporters proved “markedly less likely to turn out and vote” than they had suggested in polls, Mr. Curtice said. In the United States, the 2014 election saw the lowest voting percentage since World War II, favoring the Republican Party.
Unexpectedly low turnout only compounds creeping methodological issues that are bred by changes in communications technology and causing consternation across the polling world. Random-sample telephone polls have become increasingly fallible, partly because finding cellphone users willing to answer lengthy political surveys is difficult and costly.
After years of rejecting online polls as inherently nonscientific, researchers are beginning to embrace so-called panel surveys, which have produced some of the more accurate results in recent years. These studies rely on a large swath of voters who participate repeatedly over a course of months or years, allowing pollsters to read trends and predict voter likelihood based on respondents’ personal histories.
This could be particularly promising among young people, whose voting habits are notoriously capricious.
In 2016, “the uncertainty will be with the millennials,” Mr. Greenberg said. “Compared to young people at an earlier time, they marry later, they have fewer kids, they’re living on their own, fewer are attending church — the things that are correlated with voting.”
Giovanni Russonello is a member of The Times’s news surveys department.