Articles and Analysis


POLL: NBC/Wall Street Journal National

NBC News/Wall Street Journal
(March 24-25, n=800 adults, 700 registered voters; NBC story, video ; WSJ story, results)

Clinton 45%, Obama 45% (Clinton led 47% to 43% on 3/7-10 survey)
Obama 44%, McCain 42%
McCain 46% Clinton 44%

The lead of Jackie Calmes' Wall Street Journal poll story:

The racially charged debate over Barack Obama's relationship with his longtime pastor hasn't much changed his close contest against Hillary Clinton, or hurt him against Republican nominee-in-waiting John McCain, according to a new Wall Street Journal/NBC News poll.

Democratic pollster Peter Hart, who conducts the Journal/NBC polls with Republican pollster Bill McInturff, called the latest poll a "myth-buster" that showed the pastor controversy is "not the beginning of the end for the Obama campaign."

And from Chuck Todd's must read analysis on First Call:

As for the damage [the Wright] controversy did or didn't do to Obama, it's a mixed bag. Yes, Obama saw some of his numbers go down slightly among certain voting groups, most notably Republicans. But he's still much more competitive with independent voters when matched up against John McCain than Hillary Clinton. And he still sports a net-positive personal rating of 49-32, which is down only slightly from two weeks ago when it was 51-28. Again, the biggest shift in those negative numbers were among Republicans.

The survey also includes questions on the Wright controversy and Obama's response. Todd's write up includes this passage, especially relevant to the previous post on defectors:

One thing about these head-to-head matchups: our pollsters found that for the second poll in a row, more than 20% of Clinton and Obama supporters say they would support McCain when he's matched up against the other Democrat. There is clearly some hardening of feelings among some of the most core supporters of both Democrats, though it may be Obama voters, who are more bitter in the long run.

Why? Because among Obama voters, Clinton has a net-negative personal rating (35-43) while Clinton voters have a net-positive view of Obama (50-29). Taken together, this appears to be evidence that Obama, intially, should have the easier time uniting the party than Clinton.



One single poll cannot be a "myth-buster", as the Democrat Peter Hart (who has a conflict of interest because Obama is a Democrat) said.

It should be Pollster.com's responsibility to tel l the reader that the 5-poll average before the Wright scandal had Obama in a much better position than he is now. But they uncritically let Hart talk about this so-called "myth".



Wow, what a Poll Update love-fest. I've never seen you guys celebrate a poll release like this. You've even gone so far as called a four point margin a lead and is it me or did you guys used to be sticklers on that? You seem hell-bent on proving that Obama is the stronger candidate but, while I agree with you, you shouldn't have to lose your objectivity in the process.

Earlier today you strongly dismissed a Gallup Poll that said Obama loses significantly more Democrats to McCain than Clinton does, and tonight you are going so far as to claim that Clinton has lost her lead since the Wright controversy (a lead she hasn't had since January). To me, and I'm sure to other readers, you are looking at the data with a narrative already in mind. And when you do that, you become just another openleft, mydd, or tpm.

And in the future when you quote Chuck Todd, you should probably mention that he is an Obama supporter.



Speaking of Chuck Todd, in 2006 he predicted that if Nancy Pelosi became speaker of the House, then George W. Bush's approval ratings would rise above 50% by July 4th, 2007. Needless to say, Bush's approval was in the low 30's by Todd's ultimatum date. See video:



I don't know why the above comments are so negative - this theme is one that has been repeated often here at pollster. It seems unreasonable to say that the election has changed until you see evidence that it has changed, and so far the polls have not really moved on key issues such as net favorable.

What I don't understand is how Clinton's net negative, reported here, isn't reflected in her overall polling. Shouldn't we view this with a jaundiced eye until we see something else?

Also, I added up only 44 people out of 138 were from Gen-X or Millenial, or 32%. We can expect that these generations will be between 37 and 45% of the electorate, meaning that once again a poll has under-sampled a key Obama constituency.



Can someone analyze the various sample sizes in the study? It doesnt look like they got accurate data on Latinos - and that's one of Clinton's biggest groups. They oversampled blacks to get a better measure on how they felt, but not Latinos. Well, that's about the equivalent of leaving off Obama's white male support.....




By "oversampling" what that means is that they called up more African-Americans so that they could get proper statistics and margin of errors for demographic-specific questions. Then, for the GE and Dem Primary results they weighted the AA responses so they would represent an accurate percentage (15-20% ?) of the population. It's often done in statistics to obtain stats and sig. tests for an "underrepresented" (statistically speaking) group.



thanks for outing these guys. it simply underscores my point about the toxicity of media including pollsters like zogby and i 'others'.

the partisanship of this election has so polarized the media in favor of some candidates over others that it just trickles down in the
words and leanings of their so called reports.

human emotion is awfully strong and filters data with innuendos that have subliminal impact or more.

i would honor your stats far more if you simply came clean and put out a shingle stating your preference. but parading as a fact finding and news body in an attempt to steer our choices is precisely the arrogance and abuse that i firmly believe is creating a backlash and inevitable racial hegemony.

come clean boys, i'll respect you more. but the american public is catching onto your sheep's disguise. that's the polling you are NOT doing!!

i am a visitor. you are a propagator of data.
if you feel strongly about one candidate, announce that, and we can all move along.



Why haven't the pollsters picked up on this?

Clinton and Obama are essentially tied at 45% each, while about 20% of each would vote for McCain if their candidate loses the nomination.

If Democrats reduce the 20% "revenge factor" to 10%, (very likely in my view), that will bring 4.5% back to the Democrats and take 4.5% away from McCain. In other words, what appears to be a tie between the Dem nominee and McCain is really a 10% Democrat advantage.

That makes more sense. A candidate proposing to remain in Iraq can't be ahead when 2/3's of America wants out!

Of course, I could be wrong in that the revenge factor could be heading even higher. Somebody better hrc drop hrc out hrc of hrc the hrc race soon



Erik, pollster.com's McCain vs. Obama graph shows that McCain pulled ahead of Obama immediately after the Wright fiasco. I don't see a difference when it comes to Clinton vs. Obama, but in the GE most polls have Obama in a significantly worse position than he was before the videos were released. You talk about favorable-unfavorable, but ignore GE matchups, and also ignore the fact that Obama's unfavorable numbers in Rasmussen now outweigh his favorable ones. You are in denial.



I don't know why some of you are so upset. The polls are rather straightforward and so is the analysis.

It seems the Clintonistas are always peeved at polls that don't favor their candidate. And now, after Hillary has thrown anything and everything at Obama including the kitchen sink, hot water heater and the sump pump. And still no effect. That has to hurt. So I can understand the frustration - especially when their candidate was caught in an out and out lie backed by video evidence recently. Expect that to destroy what's left of her "favorable" ratings.

The media has finally caught on to what was apparent over a month ago, namely that Hillary can not win and needs to concede - otherwise she risks destroying the party - something she obviously doesn't care about. I think this message, slowly, combined with the Bosnia lie, will push her numbers further down. These polls (along with the PPP California poll) demonstrate this effect.


kingsbridge77, let's leave nonsense like "you are in denial" out of this. There's no reason to be emotional and nasty.

I said that things haven't changed on this race, and I stand by it. If you want to bring in another poll, such as Rasmussen, we can - and it shows that things haven't changed outside of the margin of error since nearly February. Now, if you want to compare that poll to this one, you'll find that it consistently favors Clinton more, but the point remains that it is consistent. Comparing across polls is not a viable exercise.

So whether or not you think I am in "denial" is not the issue here. The point is that for any given methodology there isn't a significant inflection point. As Mark has said here, "Wait a week". We did. It doesn't look like anything changed significantly.

Now, if you want to have a discussion of the different methodologies employed by different pollsters, and which one you think is more valid, have at it. I think people have found a number of flaws in this one that cut both ways. That's always an interesting point to raise.

This poll showed a marked increase in Clinton's unfavorable, which is the only thing new. I don't buy it until I see it reflected anywhere else. The rest is pretty stagnant, and I'm willing to go with that as the real situation unless that change in Clinton is confirmed somewhere else. Rasmussen, for example, hasn't registered any such change.


On second thought, I've changed my mind.

There does appear to be a big inflection point around March 19th in Clinton's net favorable in the Rasmussen data. Both candidates were gradually trending lower, which I didn't think was a big deal given the tone of the campaign. But Clinton took a 5-6 point hit by the methodology Rasmussen uses. I don't know why that happened, but it looks as though this poll may have confirmed that she is taking a harder beating right now.

That means we have a confirmation for the trend we see in this NBC/WSJ poll.



I think very highly of Chuck Todd, but he pushes things too far when he says, "it may be Obama voters, who are more bitter in the long run." In Gallup's findings, Clinton supporters are more "hardened" against Obama than the reverse - i.e. the data cuts in Clinton's favor. I would have hoped this would temper his interpretation.

And besides, I have enough bitterness for a dozen people. I think *I* should be overweighted in any sample.

Okay, maybe not.

But this talk of Clinton needing to drop out is just so much chin music. Dropping out is not what Democratic voters want. Period.

Really, these national polls are just chew toys. We all know the numbers that really count are state by state in (maybe) fifteen states, we all know the samples don't anticipate projected (such as it is) voter turnout, and we know Obama's negatives will go up (yes, they will).

That said, here's what I'd like to see:

a chart plotting the positive/negative/dunno findings for Gore, Kerry, Obama and Clinton starting from Jan of '00, '04 and present, using at least five colors, of which no more than one can be green.

Hey, a girl can dream.


Ciccina sed:

> here's what I'd like to see: a chart plotting the positive/negative/dunno findings for Gore, Kerry, Obama and Clinton starting from Jan of '00, '04 and present, using at least five colors, of which no more than one can be green.

I took a quick gander around a few sites, and I think Rasmussen has all that data archived. You'd have to include Bush/Bush/McCain to really make sense of it all, but it's quite do-able. The problem is that you'd have to subscribe to their premium service for $20 to get it.

If I had about 2 hours and $20 to devote to this, I'd do it fer ya. I think it's an excellent suggestion.

Incidentally, I did get the hint that at this point, Kerry/Bush was all tied up - not sure what the net Favorables were.



The difference in our approach is that you focus entirely on favorable/unfavorable numbers, while I focus on voters' candidate preference. High favorables that do not translate into votes are not useful, IMO. More voters prefer McCain over Obama now than did before the March 13 airing of the videos, according to Pollster.com's McCain vs. Obama graph in the front page.


> The difference in our approach is that you focus entirely on favorable/unfavorable numbers, while I focus on voters' candidate preference.

Not entirely, and I do owe an explanation:

Net favorables are less volatile, usually, than the voting preferences. Net favorables are about the candidate, as a person, whereas the match-up is a comparison between two people; the latter is something of a sum of two components. Looking at the net favorables allows us to look at the independent variables before we push them together in the dependent.

Also, net favorables are better for predicting things like turnout. Remember, it's not just about percentages, but who shows up. That's been the big surprise so far, after all.

Now, having said all that, you can predict that in the middle of a slug-fest that everyone's net favorable will go down. Some of that is temporary, as either Clinton or Obama will win the nomination and how the deal goes down may or may not cool the hot blood. For example, I'd have to see way, way more negative net favorables before I'd ever believe that more than 20% of the other Democrat's supporters won't eventually go with the nominee.

So while the net favorables are more useful overall, they aren't perfectly sticky. Yet that's my biggest fear in having this contest go on. Well, that and Rush Limbaugh's "Operation Chaos" getting any cred at all, despite the possibility that they influenced Texas.

In short, I'm watching the net favorables because I'm looking for permanent damage. Both Democrats lose to McCain in the current national polls, but that it's even close compared to a guy with a +8 to +10 is remarkable. The damage so far has been slight, which is good. But there has been damage, and some of it will be lasting. It's up to the two of them to figure a way out of this without bringing each other into the sewer, thus handing the election to McCain.

Remember, one of the two will get the nomination. I'm rather sure it's Obama and despite what 62% of my fellow Democrats think I'd like this to be over with. But it continues, and reality has McCain making apparent gains over both of them the whole time. That is to be expected, but how much of that is temporal? We have to dig a little deeper to have a clue on that. That's all we have is a clue.



@ Eric:

"and reality has McCain making apparent gains over both of them the whole time"

Is this true?

Since McCain became the nominee - or maybe a bit before then, perhaps when it got down to McCain and Huckabee - has the percentage by which he beats Clinton or Obama in a head-to-head changed significantly?



On the front page of pollster.com you'll see the charts for the national race, McCain versus the two Dems seperately. The time scale is very long, so it's hard to tell "significance", but before March 4th (V-Day for McCain) Obama led him slightly and Clinton was only narrowly behind - and now both are solidly behind, Clinton slightly moreso.

I don't know why this would be a surprise, given the negative tone of the campaign. What I find surprising is that the effect hasn't been higher, which is why I think what little we've seen so far is only temporary. Those are only my opinions, of course, but I think the charts accessible on the front show the "V" pattern as a definite fact.



Eric: I realized the same thing after I hit "send" - the charts are there. D'oh! But my sense, looking at them, is that little has changed. One or the other slightly up or down, but nothing significant. Too much "slightly" going on. So I'll have to look again.


Post a comment

Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.