The final update of the poll average of seven daily trackers brings some firsts:
1. This is the first time in the months long history of the tracking poll average that Barack Obama has exceeded 51 percent. He broke through 50 percent many times, especially in the last month, but never exceeded 50.7 percent until today.
2. The number of undecided or third party voters has dipped below 5 percent for the first time ever. If John McCain converts all of them to his side he will still be behind.
3. The lead is at 7 for the first time in November (ha ha).
Significantly, while the number of undecideds has fallen in this average from 7.5 percent recently to 4.8 percent today Obama's support has actually moved upwards. McCain moved up a bit last week but as the graph above shows he appears to have hit a 44 percent ceiling. That's also the number the final Gallup poll estimates he'll finish at. Based on these last six days it appears the latest deciders may be breaking for Obama.
What do you think the final percentage gap will be?
Obama has inched up in the tracking polls for the third straight day, and he now leads by 6% according to Pollster.com with less than 77 hours before the polls close in all swing states. (Iowa is the last "swing state" to close, if you can call Iowa a swing state, anyway. Part of Montana closes at 11:00 p.m. eastern). So, McCain has to gain a full 1% on Obama every thirteen hours between now and then in order to win. That means that McCain has to flip about 100,000 voters every hour, or three every second, in order to tie Obama. Best of all, he can only target 70% of the electorate for these votes, because 30% is already spoken for.
You guys are poll obsessed! Anyway, I have a new theory: the more random flucuation seen in a given poll the more comments it will generate. But in order to test this we'll need to know which polls are stable and which are less stable. One approach is to compute the standard deviation of a tracking polls released numbers (which are typically based on 3-day rolling averages but ignore that issue for now). Here are the results for the 7 daily tracking polls over the last few weeks (which excludes Battleground):
Here's the Halloween edition of the tracking poll average, now with seven daily trackers: Zogby, R2Kos, Rasmussen, Hotline, Gallup LVE (using expanded to offset more conservative LV models in other trackers), TIPP and ABC. See Quick Hits for individual poll results. Bottom line: the gentle "tightening" appears to be over.
Really, even over the 12 days shown here, things have been very stable. Barack Obama's support only varied within one point from 49.7 to 50.7 while John McCain's support fluctuated a bit more, between 42.5 and 44.2. Over this period the average Obama lead has been 6.9 percent and today it is 6.7 percent.
Things have narrowed in Pennsylvania but Obama leads there and in most of the battleground states by significantly more than he does nationally.
This will probably cause me to lose my subscriber account to Rasmussen, but there is a major discrepancy in the internals of the Rasmussen tracking poll today that I have to point out. Simply put, no matter how I work the number, Obama has a larger lead than the 51%-46% published on their website today.
Here is the problem:
The partisan weighting targets for the Rasmussen tracking poll are currently 40.0% Democratic, 32.8% Republican, and 27.2% Other.
According to the subscriber page on Rasmussen's website, the partisan internals for the Rasmussen tracking poll are currently as follows:
Democrats: Obama 88%--11% McCain
Republicans: McCain 87%-10% Obama
Other: Obama 50%-44% McCain
If these partisan weights are combined with these partisan internals, the result is Obama 52.1%--44.9%, or 52%-45% instead of the current 51%-46%. McCain. Even if all the partisan weighting and partisan internal numbers were rounded as favorably to Obama as possible on the subsriber page, Obama still leads 51.54%--45.43%, which is still 52%-45%.
Unless I am missing something, Rasmussen reported today's topline numbers to be closer than they really were. Their own partisan internals say the campaign is 52%-45%, not 51%-46%. As far as I can see, either they are cooking their numbers, or they are feeding their subscribers false information. Given that I have paid $80 for this service so far this year, either way, I am really irritated.
Well, Obama's national lead has been stable at 7% for a month now. The national campaign is not tightening, and we are just seeing statistical noise.
Even if the campaign were tightening, Obama would still have a comfortable national lead. According to polling conducted over the weekend during the tracking poll "tightening," Obama reaches 264 electoral votes in states where he leads by 9.5% or more, passes 277 in Virginia where he leads by 8.0%, and hits 286 in states where he leads by 7.3% or more. So, he is actually doing even better in the Electoral College, where 270 votes are needed to win, than he is doing in national polls where he leads by 7%.
Sound too good to be true? It isn't. I provide complete details in the extended entry.
Yesterday my tracking poll average showed Barack Obama's lead growing to 7.2 after eight straight days around 6.5 points. A small improvement but interesting after such a stable period. Today we have reports in from Zogby, R2Kos, Rasmussen and Hotline (see Quick Hits) and they are up, on average, about 1.2 points from yesterday. IF Gallup and IBD/TIPP, which will report in a few hours, are unchanged, the chart will look like this:
No doubt I've jinxed it now. Let me have it in the comments. But I'll update this post later in the day with the actual numbers.
Today's numbers are mostly based on interviews conducted Sunday, Monday and Tuesday which means, if the Colin Powell endorsement mattered, today is the day we would really see it, with a hint of it showing yesterday. Now, something else could be the cause or, if Gallup and/or TIPP end up narrowing it will be just a one day blip we saw yesterday.
Since I'm interested in trends over time the Battleground poll is not included because it doesn't poll or publish on the weekends. And what a weird poll. About nine days ago Obama was up 13. Yesterday he was up 1 (one!) and today he appears to have rebounded. The truly weird thing is this poll is based on 4 and sometimes 5 days averaged. How can you get such swings with a 4 or 5 day rolling average?
ABC's tracking poll debuted this week and may be added to this average as we move forward.
Update: Jinxed! Gallup and TIPP both show narrowing. Overall the lead today is about the same as the lead yesterday with Obama up to 49.6 and McCain up to 42.3. Still, two straight days above seven after eight straight days below (averaging about 6.5).
* = If Obama only wins Nevada among these seven, that would force a 269-269 tie, which would almost certainly result in a victory in the House of Representatives, given the large Democratic lead there
(Dark Blue (277): Obama +7.0% or more
Lean Blue (24): Obama +2.6%-+6.9%
White / Toss-up (66): Obama +2.5% to McCain +2.5%
Lean Red (30): McCain +2.6%-+6.9%
Dark Red (141): McCain +7.0% or more)
Complete Tracking poll average
Org
Obama
McCain
ABC
53%
44%
Gallup*
51.5%
43.0%
GWU
48%
47%
Hotline
47%
41%
Rasmussen
50%
46%
R2000
50%
42%
TIPP
46.9%
40.9%
Zogby
50.3%
42.4%
Mean
49.6%
43.3%
* = Gallup is an average of their expanded and traditional likely voter models
I notice two things about these numbers. First, Obama is well ahead, with only 14 days left. Second, the number of states where he holds a statistically significant lead has shrunk to its lowest point since October 1st, at least in my reckoning. While that isn't the most heartwarming trend, with only 14 days left, there is good reason to be thankful that Obama built such a large advantage earlier in the month.
Update: New polls from Colorado, Nevada, North Carolina and West Virginia have been added.
A few days ago, I proclaimed that I expected the national polling margin to drop to 4-5%, and then stabilize. So far, my prediction is a little off, as the now eight tracking polls show a slightly larger, and growing, margin for Obama:
Complete Tracking poll average
Org
Obama
McCain
ABC
53%
44%
Gallup*
51%
44%
GWU
49%
45%
Hotline
47%
42%
Rasmussen
50%
46%
R2000
50%
42%
TIPP
46.7%
41.4%
Zogby
49.8%
44.5%
Mean
49.6%
43.6%
* = Gallup is an average of their expanded and traditional likely voter models
So, right now Obama's margin appears to be about 6.0%, a number with which every polling aggregation site comes with a couple tenths of a percent. So, at least for now, it turns out I was wrong about how much Obama's lead would drop. That's cool.
However, I still think that there is a good reason to expect the final polling margin to be 4%. The theory goes like this: in presidential elections, the average poll margin during the campaign equals the final poll margin at the end of the campaign. So far in this campaign, the all-time average has been Obama by 4%. As such, I predict it as the final margin.
And then there were six. The newest tracker, by IBD/TIPP, published and polled this weekend and therefore it meets the criteria of the other five. It's based on a three day rolling average like the others. It began on Oct. 13 so I've gone back and added to the average to create the graph below. The big takeaway? Extreme stability in the Obama-McCain race over the last 7 days. As I wrote yesterday, Obama's peak in the old five-poll tracking average was Oct. 11 & 12 when he went above eight points which I believe was market related and represented a small decline in McCain's support.
The margin is very stable. Since Oct. 13 it's been 6.6, 6.7, 6.3, 6.3, 6.9, 6.4 and 6.3.
If you look at likely voters instead of registered voters you will find a narrower margin. That doesn't mean the race is narrowing it just means likely voter models usually predict that younger (Democratic leaning) voters will turn out at lower rates than older (Republican leaning) voters. That may or may not be the case this year. African American turnout will be very high and will favor Obama, that much is clear.
The folks over at Redstate are trying their best to ignore the Colin Powell news and the $150 million news and focus on a supposed narrowing of the race. According to this daily tracking poll average that is a thin hope.
* = Can't actually find the number. Going from user comments back on Monday
No matter which way you look at it, McCain has gained ground, really anywhere from 1%-2.5%, The varying movement comes from different organizations using different polls in their averages, and / or from weighting the polls differently in their averages. I tend to prefer Pollster.com, because it using all polls and applies equal weights to them ( I think).
Still, no matter which way you look at it, Obama still holds a strong lead, at least in comparison to the two most recent elections. He appears to lead by about 6% which, according to the partisan leanings of individual states in the 2004 election, would be enough to score 349 electoral votes: all of the Kerry states plus Colorado, Florida, Iowa, Missouri, Nevada, New Mexico, Ohio and Virginia. With none of the Kerry states remotely close right now, one then is prone to look at Iowa and New Mexico, which were the two narrowest Bush victories in 2004. Of the seven pickups listed above, they also happen to have the two widest margins for Obama right now according to Pollster.com and according to my own averages. So, even with a tightened campaign, we reach 264 electoral votes without breaking a sweat. And yes, there have been new state polls during the tightening period to confirm this.
So, which of the five remaining pickups are the best bets to secure victory for Obama? As long as I am allowed two answers that is actually a very easy question: Colorado and Virginia. Obama leads both of these states by 6% or more according to Pollster.com, Real Clear Politics, Electoral-Vote.com, fivethirtyeight.com, and myself. This puts both of these states right at, or possibly even above, the current national trendline. Excepting Iowa and New Mexico, they also currently give Obama noticeably wider leads than any of the other current pickup opportunities.
This is actually very good news, because it now means that Obama has not one, but two "low hanging fruit" paths to victory. If you are worried that Obama's lead is shrinking, my advice is simply to look to Colorado and Virginia. As long as he is ahead in one of those two states, then he is still headed to victory. Right now, he is so far ahead in both states, that there is virtually no chance he would the election. We have also been on real winning streaks in both states recently, picking up the Colorado Governorship (we already had the Virginia Governorship), two--and soon to be all four--U.S. Senate seats, two--and soon to be three or four--U.S. House seats, not to mention three of the four branches of the state legislatures. In other words, these are just not states where Democrats have been losing much at all.
Colorado and Virginia serve as a very strong two-state firewall. And remember, Obama only needs one.
Some narrowing of Barack Obama's lead in the tracking poll average today will no doubt spur numerous theories to explain it. "We expected it to narrow towards election day" or "It's a McCain debate bounce" or "Joe the plumber something" or "the smears are working." Take a look at the graph of the five tracking poll average (does not include the non-daily Battleground or IBD/TIPP which is less than a week old):
Here's my take. The overriding picture is still stability. Obama has been between 49.2 and 50.5 for 12 days now and John McCain between 42.1 and 43.3. In fact, Obama's typical mark in this average has been 49.8 and he's been within 3 tenths of that point for all but a couple of days. Today we are right back to where we started with this 5-tracker on Oct. 7: a lead of about 6.
Remember that the daily reports actually reflect the preceding three days' polling. So the data shown here for Oct. 7 was mostly collected before the market meltdown that started Oct. 6. I believe the stock market collapse of the Oct. 6-10 week created a bounce which has now faded away. It wasn't so much an Obama bounce (although a little, see his numbers on Oct. 11 and 12) as it was a McCain anti-bounce. His numbers fell sharply on the Oct. 9 report (fully reflecting three days of market fear) and stayed down for four more days until starting on Oct. 14 to gradually recover.
There were other events, debates and such, but my take is that the market collapse of Oct. 6-10 hurt McCain and Republicans in general. The market, while still volatile, was better this past week and McCain's numbers have crept back up to where they were 12 days prior.
So I don't expect much more narrowing although Chris Bowers' forecast of a stable 4-5 point Obama lead is reasonable.
Your take?
Update: It appears the TIPP tracking poll, which debuted Oct. 13, will be polling and publishing on weekends. If so, I will be adding to this average. Here's a preview: when you look at Oct. 13 thru Oct. 18 with TIPP in the average, there has been almost no movement, a steady Obama lead of about 6.5 points.
This tightening could be reversed over the next couple of days, as post-debate polling begins to be factored in. However, I don't expect Obama's lead to return to its previous 8% levels. Those numbers, I think, were inflated and simply unrealistic for Obama to hold onto. Most campaigns over the past twenty years have tightened up quite a bit toward the end, as I believe we have seen each of the last five Presidential elections. Further, during the entire campaign, Obama never passed a 6% lead until about a week ago, meaning that his newest supporters were probably the shakiest we have seen the entire campaign. It isn't surprising he is shedding some of them now.
I remain cautiously optimistic, because 6.5% is a pretty solid lead, and probably leads to 349 electoral votes. However, with 19 days left, this recent tightening that we have seen over the last two days should serve as a reminder that the campaign is not over until it is over.
We now have nine days worth of data for the new tracking poll average of Rasmussen, Gallop, Hotline, R2Kos and Zogby. While one can sweat the daily movement in the individual tracking polls, collectively they show stability. Today the lead is at 7 and except for a brief foray up to 8 that's about where it's been for nine days:
Obama has been right around 50 the whole time and McCain at 42/43. Just when I was getting used to this new five-poll average, along comes a new one: IBD/TIPP. Oh how I long for the days of "Galmussen" when all you had to do was average two tracking polls. What to do about IBD? Wish their numbers weren't so squirrely (an Obama lead of 3) but can't exclude them on that account. We'll see if they poll and publish through the weekend which another tracking poll, Battleground, doesn't (that one's at O +8 by the way).
Anyway, CBS poll notwithstanding, I'll believe the national lead is growing when I see it in the tracking poll average.
The important thing is to always look at all polls. Only the collective polling result matters, so don't get hung up on individual polls. Right now, before these polls are factored in, Pollster.com has Obama lead by 8.3%. That is, in my opinion, the absolute most that Obama could possibly win by. I honestly believe there is nowhere to go but down at this point, but that is just my opinion (although it has reaonble basis in recent history).
Unless Battleground starts releasing polls on the weekend, and assuming that everyone will release a poll on Election Day itself, there are 131 tracking poll releases left in this election. Man, I don't know if my stomach can take waiting for that many more tracking poll releases. At least we are in a very good position right now.
Yesterday, I complained about things that will continue to frustrate me after the impending Democratic landslide is complete. Today, I'd like to balance that out with a series of things that I won't miss after the huge Democratic victories this year. In the extended entry, I list a whole bunch of them.
Today's five tracking poll numbers are in and show that Barack Obama has widened his lead to 8.3 points:
You can final the individual poll numbers in Quick Hits. After four days just under 50 percent, Obama made a significant move upward today.
In addition to this measure, the Real Clear Politics average shows a 7.6 Obama lead and the Pollster trend estimates have a spread of 8.5 points in favor of Obama. While the Presidential race isn't over, attention is increasingly turning to House, Senate and other local races. If you have any money left after the market crash and want to get better democrats elected, here is the link.
The Reuters/C-Span/Zogby tracking poll debuted a few days ago and I've decided to add it to my tracking poll average. It fits the criteria: it's daily and based on a 3-day rolling average just like the four we've been using (Rasmussen, Gallup, Hotline and Research 2000/DailyKos). And remember, this is a Zogby phone poll, not their online polls which are notoriously erratic. Zogby's phone polls actually score higher for accuracy than Gallup according to Nate Silver's rankings.
I'm still not adding the GW/Battleground tracking poll because it's not daily, it's based on four days rather than three and has had reporting problems discussed in this Quick Hit. I've also decided not to weight the polls by sample size like they do at TPM. There's a certain logic to that approach but sample size is just one variable affecting a poll's accuracy. You can have a large sample and a bad result if, for example, your sampling procedure is bad or your weighting assumptions are wrong. And some pollsters tend to have a "left" or "right" bias. If you weight by sample size you might inadvertently exagerrate that. So we'll keep it simple and just average the five daily 3-day rolling average tracking polls.
Here's the trend for the last four days in the new tracking poll average:
John McCain's negative attacks may be causing a plateau in Barrack Obama's support but they are coming at the cost of his own support which appears to be doing a slow fade down towards 40 percent. We're only four days in to this new poll average so we'll have to see if that trend continues.
Four-tracker simple mean (no Zogby, no Battleground): Obama +8.0%
Four-tracker weighted mean (from TPM): 7.3%
Five-tracker simple mean (no Zogby): Obama +7.0%
Six-tracker simple mean: Obama +6.5%
Pollster.com national: Obama +6.3%
A very good Obama lead no matter which way you look at it. Personally, I go by Pollster.com. However, no matter which aveage you use, if state polling maintains the same relationship to national polling that it did in 2004, Obama would win the Kerry states plus Colorado, Iowa, Florida, Missouri, Nevada, New Mexico, Ohio, and Virginia, for a total of 349 electoral votes. Interestingly, those are the exact numbers that Electoral-Vote.com is showing right now.
Just when you think the bump or bounce or whatever is over, the lead grows again. That's what makes me think it's not a bounce at all but leaners and undecided voters coalescing around the Democratic ticket. Here's what the tracking poll average (Gallup, Rasmussen, Hotline, R2Kos: see Quick Hits for links) shows for the last 10 days:
Today is the first time Barack Obama has reached or surpassed the 50-point mark, also. The lead is a record 8.25 points, also. McCain is also at his all-time low at 42 points, too.
As I "argued" yesterday, McCain's campaign has the look of a sinking ship. Many countered with caution. We've seen Obama take leads before and then give them up. And many of us bear the emotional scars of 2000 and 2004 and have been unwilling to get excited too early. But:
1. Gore or Kerry never had a lead like this. And certainly they never had a sustained and sizable lead like this.
2. It's October. The post-convention period is exactly when moves like this by one candidate or the other usually take place. Waffling voters are now choosing sides.
I'm not arguing that people should cease their hard work or their donating. Now is the time to go for the kill, so to speak. But I do think, with only a month left, it is time to start redirecting some effort/money to House, Senate and other local races before it is too late. With the rising Democratic tide there is an opportunity to win many races in addition to the Presidential one.
