Polling Truth:
Don't Be Confused by the Tide of Contradictory Voter Surveys. Bush is in Trouble
by Mark Moulitsas
With one week to go, Americans are being inundated by polls. At least 112 have been published for the presidential contest in the last week alone.
While it's tempting to look at the jumble of results and declare polling hopelessly useless, fact is, polls have a great deal to tell us about the state of the race. And not in the way people generally assume:
It's not the head to head
Polls are always reported as though there's a winner, and there's a loser. So a poll showing Bush leading 45-42 is headlined "Bush leads by three", when the reality is that Bush is actually losing.
In US elections, any elected official garnering less than 50% of the vote in polls is considered vulnerable. As Democratic pollster Mark Blumenthal notes:
"Voters typically know incumbents well and have strong opinions about their performance. Challengers are less familiar and invariably fall short on straightforward comparisons of experience and (in the presidential arena) command of foreign policy. Some voters find themselves conflicted - dissatisfied with the incumbent yet also wary of the challenger - and may carry that uncertainty through the final days of the campaign and sometimes right into the voting booth. Among the perpetually conflicted, the attitudes about the incumbent are usually more predictive of these conflicted voters' final decision than their lingering doubts about the challenger. Thus, in the campaign's last hours, we tend to see 'undecided' voters 'break' for the challenger."
Testing this theory, blogger Chris Bowers examined presidential poll results since 1976, and calculated that undecided voters broke for the challenger 86% of the time.
It's a dynamic that clearly weighs on the Bush campaign. Speaking to conservative bloggers at the Republican national convention, Bush's pollster Matthew Dowd said the 50% rule didn't apply to the president:
"Based on the polling data they've aggregated on undecideds in battleground states, the Bush team has compiled the following profile on undecideds: they are overwhelmingly white, tend to be older, go to church often and describe themselves as moderate to conservative. Dowd says they can't find any self-described liberals who remain undecided."
As a result, he predicted that Bush would either split or outright win the vote of the undecided bloc. While not outside of the realm of possibilities, there is nothing indicating that Bush could buck the historical trends embodied in the 50% rule. That they're forced to spin it away hints at their concerns.
There is no national election
While the press is obsessed with horse race national numbers, the fact is that we Americans don't elect our president directly. Rather, we have 51 state elections, including Washington DC. That means that for voters in 35 to 40 states, their votes really don't matter and neither do their responses to pollsters' questions. It is only voters in the small group of "swing states" that essentially elect the US president: Arkansas, Colorado, Florida, Iowa, Minnesota, Nevada, New Hampshire, New Mexico, Ohio, Oregon, Pennsylvania, West Virginia, and Wisconsin (give or take a state).
Different pollsters use different methodologies
No two polls are built alike. Some weigh their results by demographics (race and geography), others weigh by expected turnout and party identification, and yet others don't weigh at all. Some polls attempt to filter out unlikely voters asking questions they think will flag those least motivated to vote. Some polls are partisan and will game their assumptions in order to give their side a boost in the results.
This is critical when trying to sift through seemingly contradictory polling. Polling is nothing more than educated guessing. Some get it right, some get it wrong, and half the time luck is probably involved. That's why it's best to look at polls in the aggregate - easy to do given the sheer number of them - rather than obsess over any one particular poll.
The state of the race
With that brief polling primer, anyone can take a look at the numbers and get a sense for the state of the race. And a strict by-the-numbers calculation shows that Bush is in serious trouble.
In Ohio, Bush numbers range from 43-49%, failing to break 50% in any of the 12 Ohio polls in October. Indeed, there are signs that Bush has essentially abandoned the state, working to build his electoral majority by winning three out of four in Florida, Iowa, Wisconsin, and New Mexico. But October polling in those states also show an incumbent in serious trouble.
In 14 Florida polls, Bush hasn't broken 50% since a SurveyUSA poll conducted between October 1 and October 3. A subsequent SurveyUSA poll now gives Kerry a 50-49 lead in the state. In Iowa, a single poll has him at 51% while six others range between 46% and 49%. Wisconsin is giving Democrats heartburn, but Bush breaks 50% in only one of the nine polls this month. Two independent polls put him as far back as 43%. New Mexico has Bush in the 43-49% range, anemic numbers in a state Gore won by less than 1,000 votes.
Much can happen in one week, and Republicans are doing their part to prevent a fair election. Perhaps the Bush campaign is right and the 50% rule won't apply to them this year. But the Bushies haven't been right about much of anything the past four years, while Democrats are vigorously challenging voter suppression efforts around the country. As of this writing, this is Kerry's election to lose.
· Markos Moulitsas runs the dailykos.com US political blog, and Our Congress, a blog tracking the hottest congressional races |
