Pollsters Love Asking "What If?"
By Kathy Frankovic, CBS News director of surveys
George Gallup, who did so much to develop and popularize polling from the 1930's until his death in 1984, certainly loved his job. He once said that pollsters have the best job on earth. As he put it, "We can try out any idea in the world!"
He certainly got to try out many ideas. As early as 1958, he started asking whether Americans would support an African-American candidate for president. He invented many of the questions that have become familiar: Do you approve or disapprove of the way the President is handling his job? What do you think is the most important problem facing the country? And so on.
His enthusiasm for explaining how the polling mechanism could be used for democracy made him the discipline's greatest salesman, both in the U.S. and abroad. But sometimes trying out those ideas may put us just outside of the real world and into what might be called Pollsterland.
It's not just that we can ask questions about Iraq, or health care, or immigration, or baseball's Barry Bonds, all of which are aspects of current reality. But we also can ask questions about an invented reality, such as: What would happen if . . .?
The most famous "if" question is about what someone might do "if the presidential election were held today." Today, that means asking voters about what they will be doing nearly 500 days in the future. But at least the election is a real event, with a real scheduled date. Who the candidates will be, of course, is a different matter. All the questions about how someone might vote "if" Republican candidate X ran against Democratic candidate Y are hypothetical; they can't (and don't) predict what public opinion would be even on the day after X and Y (whomever they are) become their party's nominees.
These days pollsters are looking ahead to the start of the nomination process; after all, the first caucuses in Iowa are closer — only a little over 200 days away. There is no national primary. So while some pollsters ask respondents in national surveys whom they would vote for in their home state's event, others ask simply which candidate people want nominated.
Pollsters control the choices. With eight active Democratic and 10 active Republican candidates, the full list is long to begin with. Some pollsters, including CBS News, limit the list. But others make it even longer. Fred Thompson, the actor and former Republican Senator from Tennessee, has been on many pollsters' lists since March. Thompson is likely to announce very soon. Former House Speaker Newt Gingrich also is included on some lists. Gingrich has expressed interest, but he still seems a long way from announcing. On the Democratic side, former Vice President Al Gore often appears on candidate horserace lists, even though he has said frequently that he is not a candidate.
Those additions make long lists even longer. Adding identifying information, like a candidate's current or most recent job title, makes them longer yet. To ask about the full Democratic list (with Gore) takes 67 words, while the full Republican is 83 words long!
Test runs with hypothetical candidates happen most often when there's a significant level of dissatisfaction with the actual choices. This year, however, there isn't much dissatisfaction on the Democratic side. In the May CBS News/New York Times Poll, 63 percent of registered Democratic primary voters said they were satisfied with their already announced candidates. On the Republican side, however, most primary voters are looking for other choices: only 38 percent of registered Republican primary voters said they were satisfied.
There have been other elections where dissatisfaction with actual candidates was high, and a candidate who's not running becomes part of the polling landscape. Remember Ross Perot? Pollsters starting asking about him in March 1992, just as the rumors began about his possible candidacy. But they did more than ask about Perot as a third party candidate. In June, for the last primaries, pollsters ran a hypothetical election. Both of the two (then-existing) exit polls included a question that pitted Perot against the Democratic and Republican candidates who were actually on the ballot.
Perot won the Republican primaries that day – at least in Pollsterland – and ran even with Bill Clinton in the Democratic ones. And the voters really were dissatisfied. At the time, a Times Mirror Center poll found only 6 percent of Americans who were very satisfied with the presidential candidates. Only 31 percent said they were even fairly satisfied. Sixty-one percent were not.
But pollsters' search for someone else can happen even on Election Day, when it can't matter. In polls conducted right before the 1996 election, barely half of registered voters were satisfied with the choices. So pollsters gave them another. According to the VNS exit poll, if Colin Powell had replaced Bob Dole as the Republican candidate in the 1996 presidential election, he would have defeated Bill Clinton by 12 points. If not necessarily in the real world, at least in Pollsterland!
By Kathy Frankovic