washingtonpost.com
Pollsters in the Primary Storms

By Deborah Howell
Sunday, January 20, 2008

Public opinion polls revolutionized public policy and journalism. Polls give policy experts, public officials and journalists quick access to what citizens are thinking.

While some experts think polls are overused and manipulative, the results heavily influence what political candidates say, and journalists glom onto them like manna from heaven.

While The Post did not poll in the New Hampshire primary coverage debacle, polling is a frequent part of its news coverage. Last week alone, it had a poll on D.C. Mayor Adrian M. Fenty's first year in office and a national presidential poll, with its frequent partner, ABC News, that included questions about Iraq, the economy and the president's performance. The Post conducts its own polls on local and regional issues and does polls with the Kaiser Family Foundation and the Harvard School of Public Health.

Scientific polling has been around since the 1930s, when The Post was among the first to subscribe to the Gallup Poll. The paper began its own polling in the early 1980s, soon partnering with ABC News.

Andrew Kohut, president of the Pew Research Center and a polling guru, believes in polling. "If the media didn't have independent polls, they would have to live off the polls the campaigns take and only get what the campaigns want you to see." Polling provides "the public with a more comprehensive and accurate portrayal of public opinion [and] at critical junctures . . . as a reality check for the media when they prematurely and inaccurately pronounce what conventional wisdom is emanating from the American public." Remember the high job approval ratings of President Bill Clinton during the impeachment crisis, when many pundits were pronouncing him politically dead?

Polling has its critics; readers often are wary of results. Richard E. Vatz, a mass communications professor at Towson University, said that "polling is simply too much with us." He contends that readers don't understand the margin of error (often up to three or four percentage points) well enough and that opinion polls "are more effective at creating public opinion than accurately measuring public opinion."

Jon Cohen, The Post's polling director, said that there has been "lots of speculation that polls affect elections, but I've seen little evidence of that or that coverage is better without them. Polls elucidate what's going on more than pundits spouting off. Voters are smart and make decisions by and large in their own interests and don't seem to be overly influenced by poll numbers."

Polls are not summary judgments, Cohen said; they tell news organizations what people are thinking at a point in time. Cohen and polling analyst Jennifer Agiesta work with national and local editors and reporters to pick the times and subjects for polls. "The goal is always to be as useful as possible to the broader coverage," he said. Polls also give Post reporters contacts for voters who may agree to be interviewed for stories. The Post and other major polling organizations select respondents at random using a process called random digit dialing.

Reader Luke Mannello of White Plains, Md., said he finds it hard to believe that a sample of 1,000 "is supposed to extrapolate a conclusion for 220 million adults."

The Post explains its methods on washingtonpost.com, and as to whether a sample of 1,000 people can be valid, it says the answer is: "Yes -- according to statistical and probability theory -- if the right methods are used to choose the sample of people. Stripped of statistics-speak, sampling the population is like testing the temperature of a bowl of soup -- you don't need to eat the whole thing, just stir it up and taste a spoonful or two. Or like taking a blood sample -- no need to drain the patient dry, a syringeful will do."

Readers often ask whether the rising use of cellphones affects polling. Cohen, who at the moment doesn't have a land line at home, said, "Most evidence shows that the lack of cellphones does not bias our estimates of national public opinion." The Post will call cellphones in some future polls. Pew periodically calls them now; in a recent national survey of 1,500 respondents, about 300 were cellphone users, and Kohut said their political attitudes are not much different from those of land-line users.

Another complaint is that more respondents in recent Post polls identify themselves as Democrats than as Republicans. The Post uses age, gender, race and education to adjust its final numbers, all of which can be verified by census data, but it does not use party affiliation or income. Party affiliation "is not a demographic," Cohen said. The Democratic-Republican ratio changes over time, he said, and recently more people have identified themselves as Democrats.

Readers also often ask why The Post doesn't quote this or that poll, usually one that agrees with the candidate a reader supports. Post reporters rely only on polls that follow proven methods and offer new facts or insight.

Bill Foster, a reader in Haverhill, N.H., said he was asked to take part in about 20 polls between Christmas and the Jan. 8 primary but participated in none of them; neither did many of his friends. Does that make a difference? No, pollsters keep calling until they find enough respondents.

In fact, while I was writing this column, I was called by a pollster working on a presidential survey. She was looking for the youngest person in my household. When I told her I was writing a column on polling and that The Post isn't a household, she rightly passed me by.

Deborah Howell can be reached at 202-334-7582 or atombudsman@washpost.com.

View all comments that have been posted about this article.

© 2008 The Washington Post Company