The Shrug Primary
The best bad poll yet, with a margin of error of only +/- 57 percent.
Thursday, June 14, 2007
No Idea: In the latest Rasmussen robo-poll, pre-release candidate Fred Thompson pulled into a tie with Rudy Giuliani at 24 percent. That makes Thompson the fourth Republican candidate to lead a national, New Hampshire, or Iowa poll in the last three weeks. The race is so fluid that a fifth potential front-runner, Newt Gingrich, is giving up Spanish and switching to French.
Seven months before the Iowa caucuses, polling and news organizations are cranking out an average of four nationwide and two state polls each week. Pollsters and journalists alike warn not to read too much into early polls, then read into them whatever they want.
This week, a CNN/WMUR poll in New Hampshire shed new light on how silly the media's polling obsession has become. The poll's sample size of 304 likely Republican primary voters was sketchy enough. A poll that small has a +/- 5.3 percent margin of error, which means the apparent leader (Romney, with 28 percent) is actually in a statistical dead heat with two who seem far behind (McCain and Giuliani, each at 20 percent), who in turn do not have a statistically significant lead over the man in fourth (Thompson, with 11 percent).
At first glance, the poll results imply that 92 percent support one of the GOP candidates, and only 8 percent "don't know yet" who they'll vote for. But on further inspection, the numbers are closer to the reverse. Of those 304 voters surveyed, a grand total of 18 say they've "definitely decided" to vote for their candidate. That's 6 percent. Given the +/- 5.3 percent margin of error, it's possible that nobody in New Hampshire has made up their minds. Another 37 percent say they're "leaning" to a candidate.
But the overwhelming majority of New Hampshire Republicans give an answer that puts the whole concept of "margin of error" to shame. According to the poll, 57 percent of likely Republican primary voters admit they have "no idea" who they'll for in the primary. With news organizations bombarding us with half a dozen polls a week, it's good to know they've got the margin of error down to just +/- 57 percent.
While its numbers are flimsy and meaningless, the CNN/WMUR poll may be the most revealing portrait of the Republican presidential race so far. If there's one principle that still unites Republicans these days, it's that they have no idea which presidential candidate to support. (The warning holds true for Democrats as well, although the CNN/WMUR poll found that a majority of likely Democratic voters had either decided or were leaning to a candidate.)
Of course, some flinty New Hampshirites go out of their way to remain undecided, so they can get candidates to help out around the house. But there's a big difference between flinty and clueless. The people of New Hampshire have had a front-row seat on this race for the past six months. If only one New Hampshire voter has chosen a candidate for every 10 who have no idea, that doesn't say much for the GOP field—and underscores why the press should forget the polls in favor of worthier obsessions.
The shrugs aren't just coming from New Hampshire. In last month's Des Moines Register poll, 87 percent of Republicans expressing a preference said they might change their minds.
For years, news organizations have offered margin of error as the standard polling disclaimer. In the future, they should routinely report the more revealing percentage of voters who have no earthly idea whom they'll actually vote for.
Bruce Reed, who was President Clinton's domestic policy adviser, is CEO of the Democratic Leadership Council and co-author with Rahm Emanuel of The Plan: Big Ideas for Change in America.E-mail him at email@example.com. Read his disclosure here.
Photograph of Mitt Romney on Slate's Wednesday home page by Eric Thayer/Getty Images.