Thursday, June 26, 2008

Which Candidate has a Base Problem?

That title is not meant to be read rhetorically or sarcastically. But the correct answer is "both of them". The thing, however, is that they are somewhat opposite problems.

The below is data compiled from Rasmussen Reports, consisting of over 7,000 likely voter interviews conducted within the past week. What I'm looking at is solely perceptions of the candidate among voters within his own party.



The topline numbers are not very different from one another. Obama's favorables among Democrats are 82:17, and McCain's among Republicans are 84:15. However, that conceals a lot of information about the strength of those perceptions.

A greater number of Democrats' -- about 8 percent -- have a very unfavorable view of Obama. These 8 percent are your PUMAs -- people that will probably not vote for Obama under any circumstances. Only 4 percent of Republicans feel that way about John McCain.

Obama would be thrilled, of course, if he could actually get his defection rate down to 8 percent: John Kerry lost 11 percent of Democrats to George W. Bush; Al Gore lost 11 percent to Bush and 2 to Nader; Bill Clinton lost 10 percent to Bob Dole and 5 percent to Ross Perot. In reality, Obama will probably lose almost all of the "very unfavorables" and perhaps half of the "somewhat unfavorables", which would produce a defection rate of 12-13 percent (not all of those necessarily to McCain). McCain's defection rate, by that calculus, would be 9-10 percent (not all of those necessarily to Obama).

But look, by contrast, at the enthusiasm gap between the two candidates. 56 percent of Democrats have a very favorable view of Barack Obama, while just 34 percent of Republicans have a very favorable view of John McCain. The thing that's a little bit scary for McCain is that this is after a likely voter screen has been applied, and so even after you get done filtering out those Republicans around the margins who weren't planning to vote in the first place, many of the remaining ones are still doing so for McCain somewhat grudgingly.

The good news for McCain is that if the election is close, the vast majority of these people should still wind up voting for him. That's what turnout operations are all about, and the GOP generally runs a pretty good one. Besides, 52 percent of Republicans have a very unfavorable view of Obama, as compared to 33 percent of Democrats who feel that way about McCain.

But if the election doesn't look like it's going to be close, there could be a snowball effect in which Republican turnout is quite low. If that is the case, the map could turn out to be very, very blue, and Republicans might lose a couple more Senate seats than are generally thought to be in play -- somewhere like Idaho, for instance, could be interesting -- and perhaps an extra dozen or half-dozen House seats on top of that.

47 comments

Patrick Noonan said...

Interesting data, even if it just helps quantify and validate what has already been a widely-held (and often-repeated) point of view among the public and the pundits.

It does beg the question, however, of the level of enthusiasm among Independents, the missing piece of the pie. What data are there to show not just likely leanings but level of support among that key group?

Anonymous said...

I think implying that 8% of Democrats are "PUMAs" is a bad idea - those 8% are probably mostly just regular old southern racists who wouldn't have supported Hillary either. If 8% of Ds were actually Clinton deadenders that would be a huge deal.

Anonymous said...

Obama is going to win big

many republican are angry & will not turn out for Mccain

Anonymous said...

any new polls coming out today ?

Chris said...

This (unlike the comments @ 8:41 & 8:43, which have no evidence for them) is a good post. Balanced, and you are actually pointing out both candidates' strengths and weaknesses.

The analysis also suggests an Obama-Clinton ticket wins. It would be interesting to know if the 17% of Dems with unfavorable views are former Clinton supporters (don't know if it is in the poll). With her on the ticket, you probably get only 3-4% of Dems not supporting Obama.

On the other hand, while I think the dislike of Hillary Clinton on the right waned once she morphed into a more populist style on the trail (e.g., Pat Buchanan), the enthusiasm may pick up on the GOP side. But it might anyway, once Obama is portrayed as a left winger who will do all sorts of who knows what the Republicans will claim.

But I'd like more than %s within party. Since Dems have more #s this year, just getting Dems solidified wins the election--even if Republicans are also solidified because their party i.d. numbers are so low.

lilnev said...

Can you show the converse data, R's opinions of Obama and D's of McCain? You cited the "very unfavorable" number, but I'd like to see the full breakdown. And, for independents, too, please.

Juris said...

I agree that not all of the 8% are likely to be PUMAs (or that it's probably misleading to characterize them as such -- keep in mind the race issue, for example).

Your analysis doesn't take into account the relative sizes of the "D" and "R" identifier pools, nor the preferences of those in the "I" category and their evaluations of the candidates. It could well be that Obama doesn't need to close the deal with as large a percentage of the D's as McCain needs to do with the R's.

Jay L said...

Nate, your back of the envelope estimate of a 12% to 13% defection rate ignores a key factor--differential rates of turnout among the four groups. Turnout among Democrats who have a strongly favorable view of Obama will almost assuredly be significantly higher than turnout among Democrats who have a somehwat unfavorable view, for example. When you factor in (a) differential turnout and (b) the impact of a unified convention in which Hillary lavishes praise on Obama in prime time, I think you have to conclude that a 10% or 11% defection rate is a lot more likely than 12% or 13%.

VitoF said...

"But if the election doesn't look like it's going to be close, there could be a snowball effect in which Republican turnout is quite low". -- Nate

Perhaps you just answered yesterday's post about why McCain is fighting back hard against bad polls. :)

p smith said...

New Quinnipiac poll has Obama up in Minnesota (54-37), up in Wisconsin (52-39), up in Michigan (48-42) and up in Colorado (49-44).

Given that the sample size for each is about 1,500, this is terrible data for McCain.

jay l said...

In my previous comment I should have clarified my comment on a unified convention with Hillary lavishing praise on Obama. Specifically, I think it will drive down Obama's unfavorables a bit. Bottom line, I would guess that slightly lower unfavorables by November and differential turnout will result in a defection rate of 10% or 11%.

Anonymous said...

PUMAs are "regular old southern racists".

Mark said...

this is terrible data for McCain

And it's great data for Nate. The 538 projections for CO and MI are each only 1/2 point different from the new Quinnipiac polls. The WI gap is a bit larger at 4.3 points, but only the 17 point Obama lead in MN was unexpected by the 538 model, which had Obama's lead there at 11.

Nice job, Nate.

Anonymous said...

What percentage of likely voters are registered for each party?

Can someone tell me why Obama has 52% very unfavorable on the g.o.p. side? That just seems like a massive number and I don't see what's provoking more than a "not my cup of tea" shrug in so many.

Spike said...

Sorry - can someone post the decoded acronym for PUMA? Hadn't seen it before.

Alex said...

Spike,

It's Party Unity My Ass, a group (I think?) founded by David Bower. Basically PUMAs are Hillary supporters who refuse to vote for Obama.

I would take the very-unfavorables from Obama and subtract out the very unfavorables for Kerry or Gore to get a better idea of how bad the problem is this year as compared to previous years. We know that a lot of people identify themselves as Democrats but do not vote for Democratic presidential candidates.

Juris said...

Anon: re party registration, a lot of states don't register voters by party (e.g., Michigan). So you can't use that indicator consistently.

Anonymous said...
This post has been removed by a blog administrator.
Anonymous said...
This post has been removed by a blog administrator.
Anonymous said...
This post has been removed by a blog administrator.
namelessfaceless said...