National Tracking Polls vs. Full National Polls

A few days ago, I asked, "where are the national polls?" Many have pointed out that the national polls and state polls seem to be saying different things and I suggested that one possible explanation for the difference is that we weren't seeing many full national polls at all. Instead, we were awash in national tracking polls. And I questioned whether there is some difference between these full national polls and the tracking polls that could explain the disconnect between state polls and national polls.

Since then, several full national polls have come out. Let's compare these full national polls with the national tracking polls to see if they tell us something different. I've listed the last 20 national polls (full or tracking) below with the most recent dates in the field on top.

The differences here are not quite as large as they might appear at first ... but there does seem to be a little bit of a difference. If we just take a straight average of the polls on each side of the ledger, Romney leads by an average of 1.1 among the tracking polls and Obama leads by an average of 1.0 on the full national polls. But this is mostly because of a few outliers on each side. Let's throw out the biggest outlier on each side (Gallup among the national tracking polls and National Journal among the full national polls). What do we get then? Among the tracking polls, Romney still leads by 0.6 points on average and, among the full national polls, Obama leads by 0.2 points. It is a 0.8% difference.

That's not a lot but it is something. More notable than that to me is that Obama does not trail in any full national poll. It does seem to me there's something different happening with the tracking polls. If we look at just the full national polls, the split between the national vote and the state polling is just a little less large.

You probably saw that Nate Silver discussed the differences between the state polling and the national polls in a post earlier today. In that piece, he does not consider the issue I'm raising here and expresses some rationale for why he throws his lot in with the state polling. Whether what I'm talking about is a part of this story or not, Silver's argument is a good one and is good reason to think the state polling is more to be believed. But this piece may get us a small distance towards reconciling the two.

Romney's Pennsylvania Gambit ...

... is funny.

I woke up this morning to this nonsense about Romney trying to "expand the map" in Pennsylvania. It seems Romney Romney's super PAC has some extra money to burn so they're burning it in Pennsylvania with a media buys of about $2.6 million. The Obama campaign has responded by spending about $600,000.

In defending the move, the Romney campaign has said they think Pennsylvania is in play and Politico (et tu Politico???) reports that "some polls have shown tightening there." Oy.

I have written more posts railing about how Pennsylvania is not in play than I can count. Here's one. Here's another. You can search the blog for more.

I guess it would be helpful if there were some data that could help us determine whether Pennsylvania is in play:



Oops.

It turns out there's been exactly one poll since February that has showed Romney in the lead. That poll was done by ... the Republican State Committee of Pennsylvania. Interesting. Rasmussen polled Pennsylvania just a week ago and has Obama over 50% and 5 points ahead.

Perhaps you know all this but you worry because Obama spent that $600,000 in Pennsylvania. Believe me folks. The Obama campaign is not sweating Pennsylvania. They can't literally spend every dollar in Ohio. So why not spend a fraction of what the other side is spending to get a little bit of a counterargument on the air.

Let me leave you with this thought. In 2008, Obama beat McCain by just over 10 points in Pennsylvania. The whole nation has shifted a bit back towards the Republicans since then. Has Pennsylvania shifted by 10 points? No. Will $2.6 million in last-minute spending on Romney make it shift 10 points? No.

Romney has one hope for winning Pennsylvania and it is not $2.6 million in ads. Romney's only hope in Pennsylvania is ... systemic polling failure.

Rasmussen, Systemic Polling Failure, and "Truth"

There's quite a brouhaha spinning around the interwebs with conservatives complaining that Nate Silver is cooking the books in his polling model because he's a liberal shill. Here's the argument in its original full form. It is pure drivel. Silver himself had this brilliant retort today ...

But, anyway, the whole thing got me to thinking about the state of the polls and the state of the race. Conservatives are reading Rasmussen and Gravis and Gallup and celebrating what is sure to be a big victory for Romney next Tuesday. Liberals are reading ... well, everything else and are sweating the national vote polls that have it close (actually, if you take Gallup and Rasmussen out of the naional poll averages, Obama is very slightly ahead) but feeling pretty comfortable about the state-level polls that show Obama leading in more than enough states to win a majority in the Electoral College.

What is the state of the race and whose polls are right? The truth is we don't really know for sure which set of polls is right. But, if I had to bet, I'd say Rasmussen and Gallup are wrong (for different reasons). Let me lay out the case for each side and explain why I think Rasmussen has it wrong but it is important to acknowledge that ...

Rasmussen could be right. Rasmussen has been identified as a pollster that has a "Republican lean" but this is really a mislabeling of what they are doing. It suggests that the pollster is just putting his thumb on the scale and cooking the books. That's not what Rasmussen is doing. At least I'm pretty sure they're not. What Rasmussen does is to use "dynamic party weighting." They weight their sample of voters so that they get to a designated number of Republicans and Democrats in their sample. The party weighting is "dynamic" because Rasmussen comes up with the relative weights of Republicans and Democrats based on some recent previous period of polling. Effectively, Rasmussen is making an educated guess about what the partisan makeup of the electorate will be (nationally or in a state) and then they make sure their sample looks like that.

So is it possible Rasmussen is right? Absolutely, it is possible. If their dynamic weights are accurate, then their polls will be accurate. And the flipside is also true. If their weights are wrong, they'll be wrong.

So what's the case against Rasmussen? A couple of days ago, Nate Cohn wrote that, if the polls stay where they are up to Election Day, Mitt Romney's only hope will be "systemic polling failure." This means that the vast majority of polls must have done something (question wording, poor sampling, etc.) that led them all to the wrong answer. Nate Cohn is right about this.

Systemic polling failure does happen. To understand why it is unlikely, we have to understand what other pollsters are doing differently than Rasmussen to get such systematically different results. First, other pollsters are not weighting by party ID at all. To most pollsters, party ID is something to be discovered in a poll, not something to use to weight the sample. This is because party ID is a fluid construct. People tend to change their response to a party ID question based on how they feel about the competing presidential candidates or based on how enthused they are to vote for their preferred candidate. Other pollsters simply ask respondents who they are planning to vote for and then use likely voter screens (of different varieties) to determine who is likely to vote and who is a non-voter.

It is entirely possible that this methodology will lead to a sample that is not reflective of the population as a whole. One reason is simply statistical variation. Maybe you just happened to get a disproportionate number of Romney supporters in your sample. This happens but we know the precise likelihood it will happen and this problem is corrected for by the various polling averages out there. Only a very small number of polls will be outliers for this reason and, when they are averaged into the rest, they will not have a substantial effect on the overall average number. A second problem is some kind of problem with the way we are sampling. One problem along these lines is the undersampling of certain kinds of populations like cell phone users or Hispanic voters who do not speak English, either of which may not be called by pollsters for different reasons. Some pollsters do call cell phones and some call Spanish-only voters. Here's Stan Greenberg talking about why undersampling of cell phone voters is causing Obama voters to be under-represented in the public polls we see:



Other pollsters who are good at what they do, use some kind of weighting mechanism to try to get a representative sample without calling them. This is increasingly problematic but is done well by some pollsters (like PPP in my view).

The point is that these pollsters are either getting a truly representative sample or they are using demographic weighting (not partisan ID weighting) to make sure they don't have a problem with their sample. The fact that so many pollsters are coming to basically the same conclusions (for instance, that Barack Obama has a small lead in Ohio) is evidence that they are likely right.

Could Ohio (and other states) still shift? Yes, but it is getting very late.

Could Rasmussen be right and the other polls wrong? Yes. I don't think so but it could happen and, if it did, it would be exactly the kind of "systemic polling failure" Nate Cohn was talking about.

We'll know next Wednesday morning.

Sundays Are Not Always Lazy

Sunday has been a very active day for polls coming out and it has been a very good polling day for the President. There was an Ohio poll that came out late Saturday night that had Obama and Romney tied at 49 and that can be considered bad. But everything else has been either good or trending in the right direction for Obama today.

A PPP poll just came out giving Obama a 4-point lead (51-47) in Ohio. That's not only a good number for the President, it is from a pollster that had him up by just 1 last week. PPP found that 36% of Ohio respondents have already voted and Obama leads 63-36 among them. Democrats are heavily unified in Ohio and there is no enthusiasm gap between Democrats and Republicans. These are all tough, tough numbers for Romney if they are accurate (that's the big "if"). Similarly, PPP had Obama down by 1 in New Hampshire last week and today, they have him up by 2 (49-47). Rasmussen and Gallup each moved a point in the President's direction in their national trackers. Finally, Gravis Marketing has a poll out tonight showing Obama up by 1 in Ohio. This might seem like bad news ... except that Gravis has about as strong a Republican lean as anybody (see Simon Jackman's recent discussion of house effects). Add that house effect in to Gravis and their poll looks a heck of a lot like PPP's.

We'll see how the rest of the day plays out (more polls to come tonight) ... but it is a good day so far for the President.

Where are the National Polls?

We all see the divergence between the national tracking polls, which generally show Romney with a small lead on average, and the state-level polls, which generally show Obama with small leads in most of the key swing states on average.

How could these two things be true at the same time? There's actually two things going on here:

First, it is entirely possible that the President is winning by narrow margins in swing states like Ohio and Colorado but is losing by more in "solid Romney" states than he is winning in "solid Obama" states. That would translate into a national popular vote deficit but, potentially, an Electoral College victory. Most of the sites that create election-prediction models like Sam Wang's Princeton Election Consortium and Drew Linzer's Votamatic use state-level polls either exclusively or more heavily than national polls because the state-level polls provide more information that is useful. They are generally right to do this in terms of predicting what matters - the Electoral College outcome. But those national tracking polls are still unpleasant to look at ... which brings me to my second point ...

Second, it is really striking to me how few "full" national polls there have been in this campaign, especially in the final weeks of the campaign. There are at least 8 different national "tracking" polls that are released just about every day. But these are not the same as a regular poll. I can't claim perfect expertise on the impact (or not) of this phenomenon but it does seem to me there is some important effect of looking at only tracking polls as opposed to regular "full" national polls. Tracking polls take a smaller sample every night for 3 or 4 or 5 or even 7 (in the case of Gallup) nights. They put these together to create a national polling number. Then, when a new night comes on board, the sample from the oldest night drops out. What you end up with is a "rolling average" of the most recent nights. One potentially problematic aspect of this is that, because each night is a distinct sample and because the pollster wants to keep the demographics in the full sample correct from night to night, you have to weight some very small samples night to night before adding them to the sample. This would seem to introduce some error as opposed to doing a regular 2 or 3-night poll and then weighting the whole sample from there. In effect, you'd need to do less weighting as you'd have stronger sub-samples to work with.

Now, I'm just kind of spit-balling on all that and I could be wrong. But I do think that regardless, it is notable that there have been so few full national polls. Indeed, there has not been one full national poll that I can find since the third debate. I'd be interested to see one of those from a good pollster who does live interviews with cell phones included.

UPDATE: In a post today, Nate Silver speaks to the question of the split between the national tracking polls and the state-level polling. The bottom line is that he says the state-level numbers are suggesting leads for Obama in swing states like Ohio that are statistically very meaningful. Mark Blumenthal makes some similar arguments in Pollster.com

What a Fool Believes

You know that is flat-out the best Michael McDonald song.



Now that you're in a Michael McDonald mood, let's get to his latest work. The "other" Michael McDonald is the foremost expert on early voting and posted this fantastic review of early voting around the country.

There is a lot of discussion of early voting and its impact among pundits and media folks, not to mention pollsters. A lot of the polling, particularly in Iowa, Wisconsin, and Ohio, has found that early-vote respondents are telling pollsters that they are voting very heavily for Obama. This has led some to think the President is running away with the election. It isn't that simple. If you're one of those Obama supporters who thinks early voting is going to singlehandedly crush Romney, McDonald's data is going to disappoint you. But, there is a lot of good news in what McDonald has to say.

His work is worth a full, careful read. But, the summary is this:

Everyone wants to know who is winning. In my commentary below I explain why I conclude Obama has narrow leads in Iowa and Nevada, Romney has a narrow lead in North Carolina, and in the remainder the early vote is not providing a clear direction yet. There is still ample time for conditions to change, but it will be increasingly difficult to do so as more votes pour in.

There are instructive patterns other than the horse race. The patterns of early voting in Florida and Ohio suggest that Obama supporters are successfully overcoming limitations in early voting enacted by those states' Republican governments.

To translate that, there just isn't any clear evidence yet that early voting is helping the Obama campaign in Ohio. BUT, there is clear evidence that the early voting is really hurting Romney in Iowa and Nevada. It is frankly very hard for me to see how Romney wins Nevada at this point and Iowa is looking like a very difficult state for him.

What that means is that, while Romney CANNOT win without Ohio, there are paths to victory for Obama that don't include Ohio. Wisconsin is likely going for the President if the polls are to be believed. On cue, the Romney campaign is trying to make a play in Wisconsin ... which is brilliant because, ya know, Mike Dukakis won Wisconsin so why couldn't another Massachusetts Governor do it? Adding Wisconsin, Iowa, and Nevada to the President's "solid" 237 electoral votes means he's at 259 with Florida, Ohio, Virginia, Colorado, and New Hampshire still in play (we'll assume North Carolina is going for Romney for simplicity's sake). At that point, Obama would only need Florida OR Ohio OR Virginia OR the combination of Colorado and New Hampshire. The only state in that list where Obama is not currently ahead in the pollster.com average is Florida.

Polls can be wrong. But, as we get closer to the election, the Romney campaign's hopes are increasingly pinned on that thin thread - hoping the state polls are wrong.

Take us home Michael McDonald ...

Charlie Cook is Kinda Wrong

I like Charlie Cook. He's a very smart guy knows as much about congressional elections as just about anybody. But he's just kinda sorta wrong in his read of what the debates did to the race. Here's part of his analysis:

A strong performance in that first debate would have probably closed the sale for Obama. Instead, his lackluster showing shifted a bunch of voters who had seemed to be drifting gradually in his direction back into neutral, with some reversing course and moving into Romney’s column. ... But this is a horse race, a very close one that can still go either way, and that was not the case before the first debate. The debates—and I would say all three of them—hit a reset button for Romney and put him back into this contest.

The notion that Obama could have "closed the sale" with a strong performance in the first debate is pure nonsense. Similiarly, the notion that "this is a horse race, a very close one that can still go either way, and that was not the case before the first debate" is also nonsense.

Step back and think about the logic of this. The idea Cook is selling is that Obama had a nice lead before the first debate (this is true but it was already starting to recede) and that he could have just held that lead and coasted if he had "a strong performance." This is just fantasy.

The race was already starting to tighten before the first debate. Contrary to what Cook and others think, Obama did not have a disastrous debate performance. If he did, that would be reflected in the polls with his job approval sinking. It isn't. In fact, his job approval has gone steadily up in October. What happened in the first debate is that Romney looked credible after several months of being bashed by Bain Capital ads, discussion of Romney's tax returns, the conventions, and Romney's 47% comments. In the wake of all that, by the end of September, a certain number of voters (a small number but enough to make a difference) who do not care for President Obama were unsure they could support Governor Romney. When these voters saw Romney in the first debate, where he seemed, in a word, presidential, they simply went where they were likely to go in the first place. Polling says they were moving in that direction already. The first debate accelerated that. But it was underway. Some of you may remember that this is exactly the logic I outlined way back in early March when I asked whether Romney would be a credible alternative.

Importantly, there was no possibility that a "strong performance" by the President in that first debate was going to "close the sale." That's just utterly silly.

And let me add just one point to this. You might ask, why does it matter if Charlie Cook has it wrong? What's at stake in this? For starters, Andrew Sullivan, who spent more than a week on the ledge with some insane ideas about what happened in the debate is going to open that window up again and will start threatening to jump. This is sad because I enjoy reading Sullivan's blog and now I'll have to avoid it for another week or so. Second, and more importantly, this nonsensical meme has now infected some of the smarter, more reasonable minds in the journalist class, like Charlie Cook. Whether Obama wins or loses (though it will be worse if he loses), we'll hear this nonsense about how the President looked down too much during the first debate and that's why Romney caught up. It just isn't true.

That's ... Good?

Today was one of those mixed, weird days in the polling. Obama had a very good day yesterday (Wednesday) and I said then that it would interesting to see if he could build on that with another good day on Thursday indicating a nice, little bounce out of the 3rd debate.

It didn't happen ... kinda.

What happened Thursday was sort of confusing. A few of the national tracking polls seemed to move slightly in Romney's direction. ABC/WP was the worst of it (since I think they generally know what they are doing) and that tracker now has Romney ahead 50-47. Rasmussen also has Romney up 50-47. Reuters/Ipsos now has Romney up by 1, 47-46. PPP has moved back to a tie after having Obama up by 1. I could go on. But you get the picture. It seems like Romney had a good polling day in these national tracking polls.

So, Obama getting a nice little bounce out of the 3rd debate is bunk, right? Well maybe. If you look at the state polling numbers today, they are largely good for Obama. PPP released a bunch of polls today and all were good for Obama. They've got Obama up by 4 in Colorado, up by 6 in Wisconsin, up by 2 in Iowa, up by 5 in Virginia, and they've got Obama tied in North Carolina. Those are all good numbers for the President. But that's just one pollster. What else?

NBC/WSJ has a tie in Colorado (which is not a bad result) and Obama up by 3 in Nevada. Fox has Obama down by 2 in Virginia but JZ Analytics has Obama up by 2 there. Grove Insight has Obama up by 2 in Florida and up by 3 in Colorado.

So, these state polls are pretty good overall for Obama ... and the national polls are pretty good overall for Romney. What do we make of it.

I don't know either.

Grab Your Magnifying Glass and Take a Look

It was a very solid polling day for Obama and I think I see little green shoots of a tiny Obama bounce coming off of the 3rd debate. The evidence is limited but it is in enough places that it seems real.

PPP's tracker put Obama ahead 49-48 and PPP's Tom Jensen pointed out:

Four polls came out looking at likely voters in Ohio today. Rasmussen had it tied at 48. The other three had Obama up by 2, 3, and 5 points respectively. The poll that had Obama up by 5 (Time) was conducted on the 22nd and 23rd, the night of the debate and the night after.

Two polls were released today looking at Virginia and both had Obama leading (though both are questionable for different reasons). Two polls were released today looking at Nevada and Obama was leading in both. One was Rasmussen and one was PPP. Obama is leading in Nevada folks.

Finally, the Gallup tracker lurched back towards Obama today. I discussed (and largely dismissed) this earlier. But, in the context of these other polls, it may be a sign of a little bit of movement.

Were there any counter indicators today? To put that another way, what were the worst numbers for Obama in any polls released today? There was a Rasmussen poll that had Obama down 2 in New Hampshire (though there was another poll that had Obama up by 3 there). ABC/WP didn't budge today and Romney leads by 1 nationally in their poll ... and that's about it. Those are the good numbers for Romney today.

Put all that together and mix until you see a "sticky" consistency and what do you get?

It is one day. And it isn't a big move towards Obama. It is very small actually. Most importantly, everything I describe above could VERY easily be statistical noise. We'll know more tomorrow. But it was about as good a day of polling as Obama has had since before the first debate.

Gallup: Is it a Dessert Topping or a Floor Wax?

Is today's Gallup release good news for Obama or just utterly crazy nonsense ... or could it be both???



It's a dessert topping AND a floor wax!!!

All of Gallup's metrics moved in Obama's direction today. He narrowed the gap from 5 to 3 among likely voters. He moved from a 1-point deficit to a 1-point lead among registered voters. And Obama's net job approval went from +6 to a staggering +11. What a day!

There's certainly no bad news in there. But how good is the news? We first have to remember that Gallup's trial heat tracking poll is a 7-day rolling sample while their job approval poll is a 3-day rolling sample. This means that every day, a previous day is dropping out of the sample but in the trial heat numbers it is from 8 days ago while in the job approval numbers, it is from 4 days ago. I mention this because the uptick can easily be explained by a really good day for Romney rolling out of the sample OR by a really good day for Obama on Tuesday (the 23rd). How do we know which it is? We don't know for sure but the fact that both sets of numbers got better suggests that Obama had a good day in Gallup's sample yesterday rather than just 2 different bad days for Obama dropping out of the sample. But the truth is we really don't know in the absence of additional information.

But, hey, this release from Gallup could also be a floor wax! Gallup is so off-the-rails insane with their numbers, it is impossible to say what is going on. For example, let's start with Gallup's numbers on the surface. They've got Romney ahead by 3 among likely voters but Obama has a net job approval rating of +11. Sorry, but no. If Obama's job approval is 53% on Election Day, he wins ... and it isn't close. Part of the problem here is that Gallup has a 7-day rolling sample for their trial heat in the first place. This is fairly absurd. Then, there's the problems with Gallup's likely voter model which has a very, very bad recent history. But forget the likely voter screen. Until today, Gallup had the President trailing among registered voters by as much as 3 points. Virtually every pollster has Obama ahead among registered voters.

So, is Gallup a dessert topping or a floor wax? Is it good news or just insane rantings? It's both!!!