The 2016 U.S. presidential election campaign may set a record for the number of news stories and opinion pieces based on the results of opinion polls. In the first four weeks of the year alone, polling organizations issued an average of seven a day — grist for the mill of the 24-hour news cycle.
“In this highly charged election, it’s no surprise that the news media see every poll like an addict sees a new fix,” two respected political scientists, Norman Ornstein and Alan Abramowitz, wrote in a recent essay in the New York Times under the headline “Stop the Polling Insanity.” According to the pair, many of the polls and their resulting coverage have been “cringeworthy.”
That harsh assessment comes amid what some public opinion researchers see as a crisis in the polling industry, not only in the United States but also in other developed countries. In the past few years, long-established polling institutions suffered spectacular failures in predicting the winners of elections or referendums in Britain, Israel, Greece and Scotland.
In 2012, Mitt Romney was so confident he would triumph over Barack Obama that he ordered elaborate victory celebrations. His confidence was based on internal Republican party polls and surveys from outside companies, including Gallup, one of the oldest and most respected U.S. polling companies. Last October, Gallup announced it would stay away from “horse-race polling” on who is ahead in the 2016 cycle. It will instead focus on what voters think of the candidates’ policies.
Although media critics and political scientists bemoan the prevalence of shallow horse-race stories, there is no sign that editors and pollsters are reining them in. There is no sign either that polls will be more accurate or less confusing.
Take the polls in mid-May that probed the presidential prospects of Donald Trump and Hillary Clinton. The first survey was issued by Fox News and showed Trump leading Clinton for the first time, by three points. This was an unexpected showing and prompted a flood of headlines. In the same week, Rasmussen Reports and ABC News/Washington Post polls also saw Trump ahead, by five and two points, respectively.
Joint polls by NBC News/Wall Street Journal and CBS News/New York Times, however, put Clinton in the lead by three and six points, respectively. RealClearPolitics, an outfit that regularly averages the five most recent polls, concluded that the Republican candidate leads his Democratic rival by 0.2 percent.
Such numbers, more than five months ahead of the elections, bring to mind a jocular remark by Arthur Lupia, a former head of the American National Election Survey, that horse-race polls should carry a label: “For entertainment purposes only.”
What accounts for wide disparities in poll results and for wrong calls on important elections? There are a number of explanations, but the problem really starts with the key question pollsters tend to ask: “Who would you vote for if the elections were held tomorrow?” The answers might produce an accurate snapshot in time but are of little use as a forecast. Between “tomorrow” and Nov. 8 any number of things, from a terrorist attack to stock market gyrations, could change the respondent’s mind.
Pollsters point to two trends that have made accurate surveys more difficult and explain wrong calls. One is the explosive growth of cell phones and the other a sharp drop in the number of people willing to answer survey questions. The response rate, to use the technical term, dropped from 90 percent in the civic-minded 1930s to eight percent in 2014, and even that small number is shrinking.
There is consensus that the best way to yield accurate polls is through live telephone interviews. That has become difficult and expensive. When most Americans had land lines, polling companies made automated “robocalls” to random numbers. They cannot do this with cell phones because autodialing is banned by law. Yet, more than 40 percent of the population now uses only cell phones, up from six percent a decade ago.
Reaching cell phone users by dialing their numbers manually is so expensive and time-consuming that few companies can afford it. To give you an idea of the scale of the task: “To complete a 1,000-person survey, it’s not unusual to have to dial more than 20,000 numbers,” according to Cliff Zukin, a former president of the American Association for Public Opinion Research.
There are around 225 million Americans eligible to vote and to select 1,000 — a typical sample – who are representative of the whole is difficult at the best of times, almost impossible if those polled are landline users who tend to be older, more affluent and more conservative. Minorities, young people and the less affluent make up a large proportion of cell phone-only users.
The alternative to traditional method is Internet polling, a subject of lively debate among experts. Online surveys often produce sharply different results from telephone polls. “We simply have not yet figured out how to draw a representative sample of Internet users,” Zulkin says.