Features

In this epic campaign year, when polls have come under extra scrutiny, Langer is one of the nation's most prominent, and ardent, champions of professional survey research. Sure, he tells people, get annoyed by the lousy polls, the ones with "squirrelly" methodology or inaccurate samples--for instance, a poll of Americans in "hurricane zones" that included people living far from any coast. Look with skepticism on all polls, he says, especially the ones done by groups with self-serving motives. The hurricane survey, which found Americans alarmingly unprepared, was sponsored in part by a manufacturer of storm protection products. Needless to say, ABC didn't use the poll.

"At ABC News we try like hell to produce good research. But we burn just as many calories trying to kill bad research," Langer says. So yes, it's crucial to look critically at all survey research--but it's equally crucial that we never stop looking. He firmly believes that no one can be a responsible participant in democracy without paying attention to polls. For one thing, many of the statistics that shape our lives--about health, crime, the economy--come from polls. For another, understanding the world means understanding not just what people do but what they think. "Public opinion has been an essential element of public discourse since time immemorial," Langer says. "Like it or hate it, we strive to know it, and for good reason: It matters."

The Natural
Langer, whose mother is a professor of literacy, never planned to become a de facto professor of numeracy. He wanted to be a reporter, and he chose UNH because the journalism program offered full-time reporting internships. "From the beginning, Gary was a natural," recalls journalism professor Andrew Merton '67. Then as now, Langer tolerated no nonsense. If he was interviewing someone who was trying to snow him, Langer would say quietly, "Mr. Smith, I'm confused." If the snow job continued, Mr. Smith could expect to see his evasion in print.

When Langer graduated in 1980, Jon Kellogg '70, then AP bureau chief in Concord, N.H., hired him as a reporter. "The kid" quickly impressed longtime staffers with his ability and tenacity. "He had a dry sense of humor and a quick wit," says Kellogg. (Those qualities are still evident in some Langer headlines, like "No Huckaboom" on candidate Mike Huckabee's numbers.)

Langer had been accepted at Columbia's graduate school of journalism but chose to stay with the AP, in Concord and later New York. He started working on a small weekly AP poll, studying polling the way a reporter studies any beat--by reading, interviewing experts, taking a course. He had long noticed that even some good reporters were "data-hungry but math-phobic." The more Langer learned about polling, the more he saw this aversion to numbers as not just lazy but dangerous. "I launched what could be fairly called a crusade to change things," he says.

He moved to ABC in 1990 as senior polling analyst and became polling director in 1998. Despite his passion for his work, sometimes--especially in an election year--he'll stop to remind his news-obsessed colleagues: Some of this data won't matter to "normal people who have actual lives." He's careful to maintain an actual life himself, much of it happening not far from ABC's offices on the Upper West Side. Langer and his wife and two daughters share a two-story co-op with that rare Manhattan commodity, a private, tree-shaded backyard. When he's home, he's often building something--a small waterfall, a potting shed-turned-playhouse, an old-fashioned dollhouse with crown molding and working chandeliers. His method of commuting to work is old-fashioned, too. "Mommy," he heard a little girl say as he whizzed by one morning, "there goes an old man on a scooter!"

Fie on conventional wisdom
Three things never to say to Gary Langer:

• Polls are no substitute for good reporting.
• That poll doesn't represent my opinion because nobody asked me.
• But everyone I know thinks ...

Better still, go ahead. He's heard it all before, and his answers provide a short course in the basics of opinion research. Polls aren't a substitute for reporting, Langer says; good polling is reporting. A reporter tries hard to choose representative sources and analyze the resulting information fairly. A pollster goes further, randomly selecting large numbers of sources with absolutely no idea what information they'll provide. (By "large numbers" he does mean large--for example, the 79,281 voters interviewed by a consortium of news organizations during exit polls in 68 presidential primaries this year.)

As for "nobody asked me," the point isn't whether you were called for a particular survey (as Langer says, it's a big country with a lot of phones); the point is whether your chances of being called were the same as everyone else's. He explains with a standard pollster joke: If you don't believe in random sampling, the next time your doctor tests your blood, have him take it all.

If the opinions measured by surveys look different from the opinions of the people you know--well, that's the point. You know only a tiny portion of the population; polls give the other portions equal weight. Polling "enables us to give voice to those who lack it, to assess attitudes independently," Langer says. Independence means that the Poobah of Polling pooh-poohs conventional wisdom. When Ronald Reagan died, everyone's first impulse was to reminisce about his great popularity as president. But the huge database Langer's unit has accumulated allowed him not just to question the prevailing view but to back up his doubts with facts. Among the 11 postwar presidents, it turns out, Reagan's average approval rating ranked fifth, tied with Bill Clinton's. Analyzing data begins with attitude, and Attitude Number One is to clear the mind of assumptions. Or as Langer puts it: "You don't sing to the data; you let the data sing to you."

 Easy to print version