McCain Pollster: Race Too Close to Call
Bill McInturff, John McCain’s chief pollster, released a memo last night saying the campaign has made “significant progress” in the last week and that “All signs say we are headed to an election that may easily be too close to call by next Tuesday.”
Disclaimer up front: As I’ve noted numerous times before, my wife is Chief Operating Officer of McInturff’s firm, Public Opinion Strategies. For a variety of legal and tactical reasons, neither she nor the other partners of the company are privy to the McCain numbers, which are held to essentially three members of Team McInturff. Given that I neither have the numbers and that I’m often critical of McCain and his campaign strategy, I minimize this connection out of fairness to my wife and POS.
Elizabeth Holmes prints the entire memo in WSJ’s Washington Wire:
1. We are witnessing a significant shift across the battleground states.
The race has moved significantly over the past week, closing to essentially tied on the last two-day roll. These gains are coming from sub-groups it should be possible to sustain over the next week, including:
- Non-college men;
- Rural voters, both men and women;
- Right-to-life voters; and most encouragingly;
- We are beginning to once again get over a 20% chunk of the vote among soft Democrats.
Importantly as well, our long identified target of “Walmart women” — those women without a college degree in households under $60,000 a year in income are also swinging back solidly in our direction.
Finally, in terms of critical improvement, even as this track shows more Republicans voting for us than Democrats supporting Obama, we are witnessing an impressive “pop” with Independent voters.r
As I said during our Sunday briefing, we do substantially more interviews per day than any public poll, but, given the shift we were witnessing, it was my expectation that by Tuesday/Wednesday multiple public polls would show the race closing. A quick glance at Real Clear Politics would indicate this is happening by today, Tuesday, and that’s good!
Indeed, RCP does show the race closing both at the national and the actually important state-by-state level. It also shows Obama with a comfortable lead in Colorado, Ohio, and Virginia — all states that went for Bush in 2000 and 2004 — and slight leads in several other battlegrounds that were Red in those elections. Meanwhile, McCain is leading in precisely zero Blue states.
2. It is not surprising we are witnessing this closing as we are finally having an opportunity to run a campaign that focuses on Senator Obama’s record on taxes and his lack of experience.
We are tracking how much people have seen, read, or heard about a number of thematic elements from both campaigns, including the false charges about Senator McCain’s health care plan, being out of touch on the economy, and the Obama’s campaign charges about Medicare. At the same time, we are testing awareness of “Joe the Plumber,” Senator’s Biden’s quote about his own running mate being so inexperienced it invites being tested by our enemies around the world, and Obama’s proposals that will raise federal spending by a trillion dollars.
This has been the week where “Joe the Plumber” has literally become a household name. An astounding 59% of voters in these battleground states have heard “a lot” about this story, 83% have heard “a lot” or “some” about this episode.
The 59% “a lot” dwarfs the other stories/thematic elements we are tracking this week.
The campaign’s relentless focus has helped strengthen our margins on the issue of taxes and broadened as well to the attribute of handling the economy and jobs.
I do think the “Joe the Plumber” meme, silly as the constant repetition sounded on debate night, is getting traction. Given that it didn’t cause a major swing right away, however, I’m dubious that it’s going to be decisive.
3. Our opponent is being correctly perceived as the most liberal nominee in modern times.
In our tracking, now 59% of battleground voters describe Senator Obama as being a “liberal,” a percentage that is higher than previous Democrat losers Gore/Kerry, and significantly higher than for President Clinton and President Carter.
A majority (54%) of voters profile as saying Senator Obama is more liberal than they see themselves politically.
As Senator Obama’s profile as a “liberal” increases, it has helped further erode his support among key sub-groups.
This is interesting and not something that’s being tracked in media polls. It’s also true, rather than a case of smear tactics working. Then again, given his plans to have one massive government giveaway after another in response to the financial crisis, McCain is the most liberal Republican nominee in my lifetime.
4. Turn-out IS going to go through the roof.
Public Opinion Strategies has been using a 1 to 10 scale to help look at self-described interest in the election since 1993. In 1996, in our last track, 48% of voters described their interest in the election as a “10.” In 2000, the last track was 54% saying “10.” Remarkably, in 2004, our last track had self-described “10s” at 75% of the electorate.
You need to understand we are witnessing a day-to-day trend of serious magnitude as self-described “10s” increase in every roll.
Last night, 81% of voters described their interest in this election as a 10! Wow.
Here is the importance of this number: We have watched as turn-out has gone up in the last three presidential elections from roughly 96 million voters in 1996, to 104 million voters in 2000, to a whopping 122 million voters in 2004.
I now believe turn-out will begin to approach levels not seen since other comparable presidential campaigns in 1960 and 1968.
In today’s terms, that could mean breaking the barrier of 130 million voters!
There is simply no model that begins to know or predict the composition of the electorate at this level of turn-out.
My own view … and our own weights in our surveys … reflect a belief that African American turn-out will be at historic levels, there will be a significant boost with voters 18 to 29 years old, yet the overall high level of turn-out will begin to mute the increase in the percentage these sub-groups represent in the overall electorate.
I’ll defer to Bill’s expertise on this one. The conventional wisdom is that high turnout — and Bill’s team acknowledges that it comes from Democratic subgroups — helps the Democrats. We’ve seen several elections, though, where Republicans won despite (?) high turnout.
5. There is more elasticity in this campaign than is imagined.
We have merged all of our interviews over the last three plus weeks to identify undecided and respondents who “refuse to respond” on the ballot question. This can be as high as one out of ten voters, but is generally about eight percent (8%) of the electorate in battleground states.
These voters might generally be non-voters in most cycles. But, in this cycle, 61% describe their interest in the election as a 10. This is higher than the last track among ALL voters in 1996 and 2000.
These voters are older, downscale, more rural, and are certainly economically stressed. They are quite negative about the direction of country and seek change. They voted for Bush over Kerry by a margin of 47% to 24% and this partisan advantage is a critical element to understanding our capacity to “get” these voters.
They have significant hesitations about Senator Obama’s experience and judgment.
Given an Obama TV media barrage we have not witnessed since the last candidate to run without public financing, Richard Nixon in 1972, and the daily drumbeat about Obama’s chances, given their demographics, it is my sense these voters WILL vote in this election and WILL break decisively in our direction.
These undecided/refuse to respond voters breaking decisively against Senator Obama mirrors the pattern of the last two months of the Democrat primary season.
When they do break, I believe they will add a net three plus points to our margins.
This, I think, is the ballgame. So long as Obama stays under 50 percent in the state polls — and he is in Ohio, Florida, Missouri, and North Carolina and is just barely north of there in Colorado, Nevada, and Virginia — the gap is theoretically closeable. Again, however, the conventional wisdom is that the late breaking undecideds either stay home or vote for the challenger. Maybe not, though.
6. I am becoming more and more convinced Senator Obama “gets what he gets in the tracking.”
Typically a Republican candidate trails among African Americans on a survey by a margin of something like 78% to 14%. As a firm, we consistently warn our clients that on Election Day, they will underperform their polling margins with African American voters. If their tracking says 78% – 18%, they should expect to only carry 8% of the African American vote, as the Democrat candidate will typically carry more than 90% of the African American vote.
Senator Obama’s numbers are different than anything we have ever seen before among African Americans.
In most polls, McCain is losing these African American voters by margins like 97% to 1%.
This means when you see Senator Obama’s number in a survey, it already reflects his significant and full support among African American voters.
Functionally, this means the only undecided/refuse to respond voters are white and Latino.
So, in a state like Indiana where he has recently “led” Senator McCain, in most tracks, Senator Obama is at 46% to 47% of the vote.
I am becoming increasingly persuaded it will be very difficult for Senator Obama to perform much above his percentage of the vote in a state. This puts any number of historically red states very much “in play” and MUCH more competitive than is generally believed by the media. But critically, as Obama drops below 50% in other blue states, some of these states may also becoming back in play as well.
This is a variant of the previous point. Very interesting, if true.
Here’s one thing I’m sure of: While Bill McInturff is not only on McCain’s payroll but his longtime friend, he’s not going to put his reputation at stake releasing a memo containing embarrassingly wrong information to give his candidate a PR boost. He makes a very good living by getting the numbers — and, more importantly, the interpretation of said numbers — right and continues to attract clients because he does it time and again. So, contra Ben Smith, this isn’t just some B.S. being released for public consumption.
Further, while I don’t have McInturff’s numbers, I know enough about his methodology that I’m not concerned about Nate Silver‘s objections. While — as the memo acknowledges — predicting turnout in such energized races is more art than science, I can guarantee that McInturff isn’t counting on 2004-level black turnout and that he’s damn well accounting for “young persons with cellphones.”
That said, there are a lot of “Ifs” here and pretty much all of them have to go McCain’s way to turn this thing around. He’s got another six days. And a goodly number of people have, like myself, already voted or will vote before Election Day.
I’ll make my formal state-by-state Electoral College prediction Monday. (I went 51-for-51 in 2004.) My inclination now is that McCain holds all the 2000 and 2004 Red states that are even close. But it’s looking like Obama will peel off Colorado and quite probably Ohio and I don’t see any McCain pickups out there.
via Memeorandum
James,
“But it’s looking like Obama will peel off Colorado and quite probably Ohio”
According to pollster.com state projections this morning, Iowa (53.0-41.3), New Mexico (50.4-43.9) and Virginia (51.8-43.8) all are more in Obama’s camp than Ohio (49.7-44.1). If Obama holds all the Blue states from 2004, that’s all he needs to win, even without Colorado or Ohio.
As you know, James, I’ve discounted the polls for a long time as a predictor of this election’s outcome. However, unless they’re completely unmoored from reality it’s hard for me to see how McCain can pull this one out.
Arguendo let’s assume that the polls are wrong systematically by 7 percentage points. That means that all of the “toss-up” states, e.g. Missouri, go to McCain and most of the “battleground” states do, too. Even that won’t do it for him. He needs to hold every state that Bush took in 2004 (or detach some states that went to Kerry in 2004). That’s looking pretty unlikely at this point.
Interesting stuff. My gut tells me this is likely to be closer than a lot of people currently think.
That said, shouldn’t there be something of a doctrine of “follow the money” when it comes to looking at these sort of claims? If the internal polling is this encouraging, why is the RNC suddenly putting resources into Montana and West Virginia, which should surely be in the bag and largely ignored in favour of all this low hanging fruit that’s allegedly waiting to be plucked by the GOP? I’m happy to be corrected on this, but that sort of move seems to me to smack of a campaign scrambling to avoid a rout, not one that’s ready to launch a counter-offensive.
That’s exactly where I stand.
Every time there’s a story about Joe the Plumber, it’s mentioned that he’ll do better under Obama’s tax plan. Even if the rest of the story is about how Joe doesn’t like Obama, I don’t really see this as a positive for McCain.
Ok, dumb question time: When people say things like, if Obama takes the Kerry states, plus Virginia and Colorado, he’ll win the election, do they mean literally, or just that if it happened, the odds of him losing would be beyond astronomical?
Well, the EC votes in 2004 went 286 Bush, 252 Kerry. VA has 13 electoral votes, Colorado has 9 electoral votes, for a total of 22. If my arithmetic is correct, if the 22 went to Obama (and were subtracted from the 286 in the 2004 Republican column), the final tally would be 272 Obama, 254 McCain.
Do I have that right?
Yeah, if Obama takes all the Kerry states, plus Virginia and Colorado, he wins.
If he takes all the Kerry states plus Ohio (and nothing else) he wins.
If he takes all the Kerry states plus Florida (and nothing else) he wins.
If he takes all the Kerry states plus Nevada, Colorado and New Mexico (and nothing else) he wins.
And now I think about it, I’m forgetting the fact that Iowa appears to be in the bag.
Sam: 252 + 22= 274
either way I expect this to be tighter than is being predicted
Anthony has ably listed some scenarios under which Sen. Obama wins the election. There are dozens of others with varying levels of likelihood or credibility.
The most likely scenario under which McCain wins is that he holds all of the states that Bush won in 2004. That means that he must hold states which multiple polls indicate are in Obama’s camp by nine percentage points or more. Still possible but not particularly likely.
I believe this conventional wisdom may well hold true again. My question, though: who is the challenger in this race? The senator who has been the presumptive winner since Super Tuesday or the senator who has been the presumptive loser since about the same time?
tom…Duh…I’d like to say I fat-fingered the 2 for the 4, but. Anyhoo, I think one hoped-for effect of this poll is to galvanize dispirited McCain voters. However, it just might, if taken seriously, galvanize complacent Obama voters. Talk with you all later; I’m off to early vote to cancel out a McCain vote.
The problem is that for McCain to win, the late breaking decideds really do need to break for him by a very heavy margin. VERY heavy.
I’d just come back to the “follow the money” point. We know the GOP campaign is strapped for cash. On that basis, where they put their money counts, surely? As things stand, doesn’t the way they are distributing their distinctly finite funds – in a manner that looks more like making sure the “lean McCain” states stay that way than an attempt to turn “lean Obama” states red – seem to indicate that the news isn’t going to be good?
Of course, McCain and Palin are putting a lot of effort into PA. If they can flip it into the McCain camp it’ll at least make for a more intersting night’s election viewing.
Be more specific, James. What are the states you’re betting McCain wins? As I interepret you,
you’re betting MCcain holds North Carolina, Missouri, Nevada, Indiana, Georgia, North Dakota, Montana, and Florida. And Arizona.
I assume you’re not talking about Virgnia and New Hampshire, because those states are *not* close. Ditto for Iowa and New Mexico, which were also Bush 2004 states.
I still think you’re not just wrong, but wildly wrong. I’d bet $100 that Obama takes at least two of that list of “long-shots” up there. Your interpretation of this email as anything but McCain propaganda is wishful thinking. We have seen meaningful fall-off from Obama’s peak numbers in *one* state (Florida). Other than that – no movement whatsoever in McCain’s favor since the last debate.
The undecideds from September have already massively broken for Obama, and you think the last fraction of them are going to reverse course overwhelmingly?
PS: while I’d bet $100 on Obama winning two from that list, my estimate is 4-5.
James, I don’t think the following chunk of your argument is as persuasive as you believe it should be:
Most other pollsters (Zogby excluded as he provides the numbers his clients want) depend on their reputation for getting the numbers and interpretation of those numbers correct. Right now Mr. McInturf is projecting an outcome that is significantly different than other pollsters who face the same general set of incentives as you argue that he faces. Getting things right is valuable. However there is a decent chance that his interpretation is wrong for absolutely no nefarious reason beyond he is applying an odd model. The flipside of that is his model is right and everyone else is massively wrong, but given the same incentive structure, I’ll hang out with the cluster of polls and pollsters that are projecting a wider race than the McCain campaign.
Returning to the issue of undecideds breaking for McCain, at TNR they’re quoting Steve Lombardo saying that (presumably assuming the declared Obama vote is correct) they need it to break between 80 and 90 per cent in McCain’s favour in order for it to put McCain over the top. As I understand it, most of the guestimates are that it’s likely to be more like 60-65.
Gallup and others provide numbers for media consumption and debate. That’s far different than providing them for individual paying clients who rely on them for a high stakes campaign.
Here are today’s RCP Battleground numbers:
My hunch is that McCain keeps Florida, Missouri, and North Carolina. I think it’s possible but less likely that he keeps Ohio and Virginia. I don’t see how he keeps Colorado and think Nevada unlikely.
Beyond that, I don’t see any likely Blue State pickups.
I suspect a description of the various polls showing McCain closing is probably correct. I just don’t believe McCain is going to win-he just seems to b e underperforming in too many states Bush won and isn’t really performing well in any of the Kerry states.
I think on election day we still see Obama the winner, and I think he will do well in the electoral vote column, but the popular vote will once again be rather close, or at least a lot closer than what the polls are currently indicating.
to elaborate, agreed.
my own gut feeling is Obama EV 320+ but popular vote will be 52% (or there abouts). MO is hard to call… I am in a blue area of a red county, but the “undecideds” seem to be breaking for Obama.