Election polls: Why they got it wrong
“The polls blew it,” said Michelle Jamrisko and Terrence Dopp in Bloomberg.com. In the run-up to this week’s presidential election, virtually every major survey of the U.S. electorate gave Hillary Clinton the edge by 2 to 4 percentage points. Actual voters, however, “still possessed the capacity to shock.” In the end, Donald Trump delivered a “commanding performance” that virtually no professional pollsters saw coming, despite a mountain of mathematical models and sophisticated survey methods. The incredible misreading of the national mood, coming on the heels of pollsters’ failure to predict the U.K.’s vote in June to leave the EU or to spot Colombia’s rejection of a peace deal with rebels last month, leaves the credibility of the polling industry in tatters and its practitioners facing questions “on how they got it so wrong.”
“The signs of a polling disaster were all there,” said Steven Shepard in Politico.com. The handful of surveys showing Trump winning—which were mostly conducted by small, partisan consulting firms—were dismissed by major news outlets. No major nonpartisan polls were conducted in the battleground states of Colorado, Nevada, Ohio, and Wisconsin during the last week of the election.
“Taken together, it was a recipe for an epic polling failure.” Obviously, some voters “were apparently sheepish about admitting to a human pollster that they were backing Trump,” said Nathan Bomey in USA Today. The Los Angeles Times/University of Southern California tracking poll, one of the few that consistently pegged Trump as the winner, is conducted online, which may have led respondents to be more truthful.
“This should be the end of industrial political polling as we know it,” said Mark Balnaves in Qz.com. Thirty years ago, pollsters could reliably use landline phone numbers to survey a statistically representative group of people. But the rise of mobile phones has scrambled the equation. Even when voters do have a landline, fewer of them are willing to be interviewed than in the past, clouding the resulting data. We put way too much faith in polls to begin with, said Mona Chalabi in The Guardian (U.K.). Human beings can’t be predicted as easily as the weather. “They can change their minds, they can decide to not share their opinions, or they can flat out lie.” Pollsters and the media didn’t see this coming for a simple reason: “We are obsessed with predicting opinions rather than listening to them.”
■ Over the last week of the presidential campaign, Donald Trump’s aides persuaded him to stop using Twitter to attack his critics. All his subsequent tweets came from an Android phone used by his staff. Los Angeles Times
■ North Carolina eliminated 27 polling places for this election, and cut hours at other sites, producing long lines that required voters to wait hours. Other Republican-controlled states freed of Voting Rights Act restrictions also dramatically cut the number of polling places, including Texas (403 poll closures), Louisiana (103), Alabama (66), and Arizona, which closed 140 of its 200 sites. Slate.com
■American Millennials are pushing global coffee demand to a new record high, sending prices soaring. In the U.S.’s coffeehouse culture, young people now start drinking coffee at the average age of 14.7 years, and those ages 19 to 34 now account for about 44 percent of coffee drinking. Bloomberg.com
■As Baby Boomers age, the number of Americans 65 and older getting cosmetic surgery is rising. In 2015, eyelid surgeries rose 62 percent among seniors, to 39,772 procedures, while face-lifts doubled to 37,632 procedures. The Washington Post
■ About 1.8 million U.S. children were homeschooled in 2012—3.4 percent of the U.S. student population, and double the number since 1999. About 4 in 10 homeschoolers have parents who graduated college. The Washington Post