Kerry Spot    [ jim geraghty reporting ]
[ kerry spot home | archives | email ]

Let me throw out a theory here.

The polls in the last week of October will have a tremendous impact on the election.

If one candidate is ahead outside of the margin of error, the trailing campaign will have an exponentially harder time getting its voters motivated and enthusiastic. The campaign volunteers in every precinct and every swing state will sense they're fighting a losing battle. Every news story about every speech will mention the polls. The candidate who is behind will have to utter those demoralizing words, "The only poll that matters is the one on election day."

I suspect that the number of truly undecided voters is relatively few. I realize I run in political circles, but as I talk to family, friends, and neighbors in different parts of the country, I find that they have encountered very, very few individuals who are definitely voting but who genuinely have no preference for one of the two (or three, for the Nader supporters out there) candidates in this race.

No, this contest is between two highly motivated bases — each suspecting the other of advocating policies that would get innocent Americans killed — with a smattering of late-breaking undecided voters in between.

During the period bounded by the end of the GOP convention and the first presidential debate, John Kerry was behind. Sometimes he trailed by only a few points, sometimes he trailed by double digits. Each day brought more left-of-center columnists giving Kerry's campaign advice on the op-ed pages. Had the election taken place during this cresting wave, Bush would have won in a landslide in both the electoral college and the popular vote.

Then, in the space between the first debate and either the vice-presidential debate or the second presidential debate, Kerry pulled even and seemed positioned to pull ahead. The president dropped the ball by appearing irritable and unfocused in the first debate, and his supporters were angry about it. This was the most important election of their lives, and the president had appeared way too confident that he had it already won.

In fact, some conservatives were ready to ask whether Bush had been campaigning effectively at all.

Why did it take 200-plus veterans, and not the Bush campaign, to put the spotlight on the not-so-gleaming side of Kerry's Vietnam years? Why had the senator's 1971 testimony, stirring anger among so many veterans, been left to a bunch of political outsiders to expose?

Were Bush's ads effective? Or were they just becoming pleasant-sounding background noise against the constant "Bush will enslave your children" barrage from MoveOn.org?

Even Kerry's advocates agreed his style on the stump was awkward at best, and that his views on Iraq changed and spun more frequently than a roulette wheel. Perhaps America could choose a lightweight president if we were in the no obvious threat and booming-economy 1990s. But this was the first post-9/11 election. So why wasn't the race set to be a landslide?

It appears, however, that the president's stronger performance on Friday night — giving birth to the new oddball political catchphrase "Got wood?" — has stopped the slide, and has maybe even built a little momentum for the GOP. Conservatives can look at RealClearPolitics and not cringe.

Zogby has Kerry up by one nationally. Rasmussen has Bush up 49.6 percent to 45.9 percent on Saturday, and 49.5 percent to 45.5 percent Sunday. (By Tuesday, Rasmussen had the race much closer, with Bush at 47.4 percent and Kerry at 45.8 percent. Over two days, Bush lost 2.1 percent and Kerry gained .3 percent?)

The Battleground poll has Bush up 49 to 46, with a job-approval rating at 52 percent.

After a Gallup poll showed Colorado tied, Survey USA has Bush up by eight points. (Interestingly, the poll also finds that a Kerry win in Colorado is almost certain to bring with it a victory for the state's proposed constitutional amendment that would allocate Colorado's electoral votes proportionally. So there are three possibilities on Nov. 2: Bush wins all nine votes; Bush wins five votes; or Bush wins four. I wonder if this analysis would convince Colorado Democrats that this amendment could end up hurting their guy.)

In the electoral college, this analysis remains accurate. Bush has a ways to go before he's out of the woods, but Kerry has little margin for error. State after state that is looking less blue and more red recently is in his "must win" column.

If the numbers remain around this level, then it's just a slight Bush advantage, and Kerry still has a chance of winning the presidency. If Bush builds a stronger lead, say mid-single digits, the enthusiasm on the left will slowly dissipate. (There wasn't that much enthusiasm for Kerry the man, as opposed to passions against Bush, to begin with. Watch Kerry's numbers among African Americans as a key indicator.)

If Kerry pulls even or pulls ahead, then Kerry's fans in the media can follow the prewritten script about their man's mythical "great closer" reputation.

Of course, both sides know this. So each camp has a vested interest in publicizing polls that show their side up. And the pollsters — sworn to objectivity as they are — know that the numbers they roll out between, say, Oct. 24 and Nov. 1 will have an enormous impact on the coverage of each candidate and the public perception of which candidate has momentum.

In short, it's important to check the polls with even greater scrutiny in the coming days. How large is the sample size? Over how many days was the poll conducted? Was it on a weekend? (Whether Democrats do better on weekends is hotly contested, but it seems likely that a certain segment of the voting population is less likely to answer the phone on weekends.) What's the breakdown by party affiliation in the poll? Does this split seem out of whack with what one would assume the electoral breakdown at large to be? (The results from the last election had Democrats at about 35 percent and Republicans at about 31-32 percent; Independents made up the rest. Have the electorate's party-affiliation proportions changed since 2000? Sure, but probably not by more than three or four points.) And did this party breakdown differ greatly from the one in the last poll conducted by this organization? Did the results change greatly from the group's last poll, and does that change seem plausible in light of campaign events?

Good polls act as adrenaline for a campaign. And in this era of sports doping, it's only natural that bad actors on each side would want to give their campaign some opinion-poll steroids. (If journalists are willing to take a Microsoft Word memo and insist that it's a 1972-era artifact, they can over sample certain groups and round down a point or two.)

Those with an agenda will attempt to spin you in the coming weeks. Don't let them.