National Tracking Poll Thread

[ajb]

Current Poll numbers

Obama 47% Romney 43 %

Obama 47 % Santorum 43 %

And yet just yesterday Rasmussen said that Obama led by four in FL, OH, NC and VA, four populous states located right in the middle of the left-right spectrum in America. Hmmm...

[ajb]

Good God. Remember their "traditional"/"expanded" fiasco in 2008?

Yeah, not only in 2008, also in 2010.

Remember when they had about 3 different models to forecast the "generic GOP lead" - one of them even had the GOP winning Congress by about 18 (they actually won by 6).

Gallup is just a crack pollster to me sometimes.

Mind you, then the alternative seems to be Rasmussen, who weights for party ID, which has methodological issues of its own. In the end, I'm glad for a lot of variety in how pollsters do their work.

[ajb]

Just to keep it interesting... here's the latest CNN poll:
Obama 52
Romney 43

http://i2.cdn.turner.com/cnn/2012/images/04/16/rel4a.pdf.

[ajb]

Just to keep it interesting... here's the latest CNN poll:
Obama 52
Romney 43

http://i2.cdn.turner.com/cnn/2012/images/04/16/rel4a.pdf.

Crap poll. CNN hasn't exactly been the gold standard in national polling.

Well, we'll never know if they were right or not, will we? Putting all of the polls together, we're left with a race somewhere between "Romney leads by 2-3 points" and "Obama leads by 10," depending partly on methodologies and assumptions. Seems reasonable enough to me.

[ajb]

various simultaneous polls are pretty likely to be all over the place until November. the obvious thing to do for everyone is shut up and trust Nate Silver

Speaking of which, by this time in 2008 he was doing much more than he is now towards projecting the general election. Looking forward to seeing that happen soon!

[ajb]

Gallup's tracker has Obama up by 7 over Romney today, 49 (+2) to 42 (-2). Clearly, this is going to fluctuate as much as their approval rating poll does (of course, since it's actually the same poll...).

[ajb]

Today's Gallup is Obama 48 (+1), Romney 43 (-1). Guess being stuck with Obamacare is really killing Obama in the polls.

[ajb]

The RAND American Life Panel (9/26):
Obama: 50.46%
Romney: 42.46%

https://mmicdata.rand.org/alp/?page=election

The graph here is pretty eloquent, if you follow the link...

[ajb]

Meanwhile, the erstwhile-gold-standard Gallup tracker has the race steady, at Obama 50-Romney 44.

[ajb]

Meanwhile, the erstwhile-gold-standard Gallup tracker has the race steady, at Obama 50-Romney 44.

Disapproval is up one on Gallup.

Irrelevant to what he was saying.

Kind of just irrelevant, generally. Among other things, I believe that the approval numbers are among adults, and the horserace numbers among RV.

[ajb]

Meanwhile, the erstwhile-gold-standard Gallup tracker has the race steady, at Obama 50-Romney 44.

Disapproval is up one on Gallup.

Irrelevant to what he was saying.

Kind of just irrelevant, generally. Among other things, I believe that the approval numbers are among adults, and the horserace numbers among RV.

Approval ratings do matter ... but only to a point. I've argued all year that if the horse race is really close, then the approvals will matter. But if the challenger is clearly down nationally, and the approvals of the incumbent are down or even in negative terms, then it shows the incumbent can be ahead in spite of that.

But this election could break many established rules of elections... having said that I doubt Obama's approvals will be underwater in many places outside of Ras on election day.

Well, approval ratings do matter, but across all pollsters, not in one daily tracking poll with a small sample, and not really among adults at this point.

[ajb]

Meanwhile, the erstwhile-gold-standard Gallup tracker has the race steady, at Obama 50-Romney 44.

Disapproval is up one on Gallup.

Irrelevant to what he was saying.

I generally have been reporting both as a matter of course, when I have them.

In the interest of thoroughness, then, I'm posting the Rasmussen tracking poll numbers here, for anyone who may have missed them in the Obama Approval rating thread:

without leaners: Obama 48-Romney 46
with leaners: Obama 49-Romney 47

Still watching that 47% blip roll off the numbers.

[ajb]

Still holding on Gallup with Romney 46, Obama 50.

Approval now at 48%, 46% disapprove.

Romney 44, Obama 50, that is.

[ajb]

Still watching that 47% blip roll off the numbers.

They already did, silly.  It was in there on 9/19 to 9/22, then it dropped.  See the trends link on this page:  http://www.rasmussenreports.com/public_content/politics/obama_administration/daily_presidential_tracking_poll

So what's driving Obama's rise in the Rasmussen tracker now? Whatever it is, you must be expecting to see it reflected in Gallup sooner or later.

[ajb]

It is a bit strange to see the constant 50-44 lead with so much movement in the job approval number. You have to wonder if there isn't some huge Obama sample on the 7 day tracker that is holding the lead at 6. Time will tell I guess.

You are probably seeing a strong pro Obama sample in there that was a reaction to the 47% comment.  That is why I was saying wait until Tuesday.  

The week the comment came out, Obama's weekly approval actually went down.  They just were not yet getting the reaction results as of that point (though Rasmussen had them).

I wouldn't read that much into Gallup's approval rating numbers, given that they're of adults, not of RV or of LV. And given that an average of 65 out of every 500 adults they poll every night is not an RV, there's plenty of room for the approval numbers to fluctuate in ways bearing no relationship to the horserace numbers.

[ajb]

The trial heat numbers are RV. The approval numbers are adults:

"Gallup tracks daily the percentage of Americans who approve or disapprove of the job Barack Obama is doing as president. Daily results are based on telephone interviews with approximately 1,500 national adults; Margin of error is ±3 percentage points. "

http://www.gallup.com/poll/113980/Gallup-Daily-Obama-Job-Approval.aspx

[ajb]

What Scott Rasmussen's trying to say I guess is that he will use a D+3 sample in the final week, instead of his current R+2 sample, to save face.

Kind of makes you wonder, right, why he weights for partisan ID, something almost no other pollster does, when he himself thinks that actual turnout in November will be quite different from his weightings.

[ajb]

Just read through the topline results on their page.. they compared pre-debate poll to post-debate... likely voters went from Obama +9 pre-debate to Obama +5 post-debate.. for what it's worth.. only time will tell.

If those numbers are borne out by other polls (and it's too early to know if they will be) that would be both a relatively large shift for a debate to cause, and also not nearly enough on its own to change the election result.

[ajb]

Rand American Life Panel, Oct 2-3-4:

Obama   49.76   49.17   49.87
Romney  44.22   44.83   43.98

[ajb]

Interesting observations this morning on Nate Cohn's twitter feed. It seems that most of the shifts in Rasmussen's numbers in the past month can be accounted for by African-American voters moving back and forth between Obama and undecided. Right now, Ras has Af-Am voters going 85-3 for Obama, with 12% undecided (vs. 3% undecided among all other voters).
Make of that what you will.

https://twitter.com/electionate

[ajb]

In 2008, they consistently had McCain doing better than everyone else, insisting that the race was still close up till the end, when they suddenly allocated 2/3 of the undecideds to Obama, bringing their results to the average of all other pollsters at that point, and hitting the final margin to the nearest tenth of a percent. They had similar results in 2000/2004 as well, with Republicans overperforming relative to other polls till the last minute, when undecideds swung to the Democrat.

http://www.pollster.com/blogs/hitting_a_bullet_with_a_bullet.php?nr=1

[ajb]

The Romney debate bounce continues to fade in Reuters/Ipsos:

Obama 47% (+1)
Romney 45% (nc)

Of course RCP doesn't include this poll in its average, otherwise Obama would pull back into the lead today.

RCP would seriously add to its credibility if it had a "methodology" page for its averages.

[ajb]

Rasmussen LV:

Romney:  50, +1

Obama:  47, +1

Weren't they 50-46 yesterday?

[ajb]

Romney leads by 20 among Indies.

I don't like this ...

A lot of them are really the tea party people that ID as "Independant" that's why Romney is leading with them everywhere but still losing states.

^^^^ This. Hence, ignore the Umengites who post ROMNEY WINNING INDIES BY A MILLION on every poll.

... and who simultaneously complain that there aren't enough Republicans in those same polls, pretending that there's no connection between the two phenomena.

[ajb]

So are those ABC tracking numbers the reason Mitt is undergoing a mini-collapse on Intrade at the moment?

It is likely the ARG Ohio poll.

Or just the unwinding of the unwarranted rise in Romney's Intrade numbers a few days ago. The other betting markets never jumped on that particular bandwagon, so it's only natural that Intrade would revert to the mean at some point.