Articles and Analysis


Daily Tracker Update

It is probably premature to start talking about a "widening" of Barack Obama's lead over John McCain, although today's national tracking results certainly produce no evidence of any continuing "narrowing." Of the eight national tracking polls out today, five show slight movement in Obama's direction today, only one in McCain's direction and two show no change in the margin.

081031 trackersa.png

While there have been ups and downs, the average Obama lead on the eight daily tracking surveys today (6.3%) is almost exactly what it was (6.4%) on October 20, the day of the first ABC/Washington Post release. The Washington Post's Ben Pershing used a modified version of national trend chart today -- filtered to show just the national trackers -- to make a similar point: "Today's margin [on the trackers] is almost exactly where it was a month ago."

I'm not sure which filter Pershing applied, but the following chart displays the trend line based only on the eight national tracking surveys included in the table above



Steady as she goes...



McCain needs to gain roughly 1.4% per day to have a shot at winning this. Losing a day is terrible news. Obama gaining more than 0.6% on average is beyond terrible.



The polls are just remarkably (sorry) stable. Take the tightening before the conventions away and respective bounces and Obama has consolidated on the back of the economy and social issues which he has always had the advantage.

I'm one who thinks Obama's lead is wider than the average based on my calculation of likely turn out, this time out turn out advantages Obama.

It seemed to me and when its polled its apparent that McCain's age is a bigger disqualification than Obama's race.

McCain is every bit of his 72 years. I'd expect greater width over the weekend.

Makes it hugely important for Obama to get his vote out.



One thing I don't hear enough mention of is the enthusiasm gap between Dem and Rep. Everything I've heard shows close to 30% enthusiasm gap between the two, and I feel that it is probably not apparent in the likely voter models. It's much easier to say you're voting over the phone than it is to actually go out and do it, and I feel as though that 30% enthusiasm gap will show up once the polls open on Tuesday (if it's not already apparent in early voting).



A weighted average of all the state consensus figures from another poll tracking site shows Obama up by 7.9. It reached a peak on Tuesday of 8.6 and has slipped a bit since. 7.9 is essentially where it was last weekend. It was down around 5.8 the prior weekend. So, Obama's still way up.

Aggregating all the state figures gives a tremendous N and is probably quite robust, despite the wide variations in methodology. Hey, if you can aggregate all the polls within a state, why not across states too.


Gary Kilbride:

Definitely looks like a good day for Obama in the national numbers. I was noting that as I started from today's first thread and scrolled upwards.

I'll say it again, comebacks in a general election are for fools.

I still don't understand Plouffe dismissing the national margin. Who does he think he's trying to kid? He said that as recently as yesterday in a TV interview. Then Peter Hart followed and applied some real world perspective to the same question: "Of course they matter."

Plouffe's approach would be like a golfer not understanding he benefits dramatically from hitting a shorter iron to the green. That 300 yard drive cleansed your second shot, genius, by allowing a wedge instead of your buddy trying to shape the 6 iron from around the corner. Likewise, there is no such thing as competing in Indiana and North Carolina and Missouri, etc, without the national mood and margin dumped on the individual states, and amended based on their partisan tendencies.



I can't wait for Mcsame and Palin, I mean unstable and unable to finally lose in 4 days. I don't think they can catch up to Obama and Biden in such a short time. I just wonder if Palin will try at 2012?



Only the losing side talks about polls, as in "don't believe the polls". And I think Plouffe was saying that they're really focused on what's going on in the states, not in the national polls.



When I aggregate Pollster's state trend figures, weighting by the number of U.S. representatives, I get Obama up by 6.4%, comparable to the national trend. The figure itself is suspect, of course, but it was 1.6% four weeks ago, and has steadily climbed since then, with only 2 small one-day drops (I haven't computed it every day, though, and there might be other one-day reverses). What interests me about this is that my state aggregate is rising but the national trend has been pretty much flat over the same period. It will be interesting to see, next Tuesday night, whether Obama has been strengthening or just hanging in there.



@ Gary Killbride

You are quite right about that. The margins on the national polls are actually more valid than the state ones as the sample sizes area always bigger on the national polls. It's counterintuitive, but you higher statistical validity polling 1,000 Chinese than if you polled 700 Hawaians. The margin on the massive chinese population is actually smaller than the one on the relatively tiny Hawaian population.

It's usually at the state level where surprises are seen in elections.


We've been using the Oct-election day version of the tracker graph for some time at Daily Kos (we supply one of the trackers, the Research 2000 poll) but I would recommend putting it on high sensitivity because of the immanent election.

Thanks to Mark and Charles for the ability to do this!



There may be more samples in a given national poll than a given state poll, but in light of the amount of state polling going on, collectively there are far more samples at the state level than the national level. In that sense, there is actually a lot more data behind the simulated national polls using weighted state polling results that the various people have been calculating as opposed to the national polls.

But I don't think that quite renders the national polls useless, because that state polling is not randomly distributed--certain states tend to get polled a lot more than others, and for some the data is quite sparse. So, we don't really have an up-to-date picture in every state at a given time.

Accordingly, I do see a point to national polling, as a sort of helpful gap-filler for states that are not polled frequently. That said, it is nice that the average of the national polls and the simulated national polls using weighted state polls have never varied too much from each other, since that consistency suggests nothing too important is being missed at either level.



the obama informercial seems key. today was the first day that it kicked into the trackers.

it was a masterful part of the strategy. at a time when the obama campaign figured that the squeeze might be on and momentum might shift to the GOP, the ad stopped any slide at a critical time.



Read the headline on Drudge tonight before you count your chickens. A few new ones are coming home to roost. Birds of a Different Feather.

"McCain actually ahead by one on Zogby in (one-day) Friday polling." An outlier? Perhaps.

But I think there are a lot of Republicans like me (though I voted Barr because McCain doesn't need my vote to carry Texas) who have been very publicly quiet because of all the hype, BDS (Bush Rerangement Syndrome) and Groupthink. We have been bullied into stillness, but we still vote in high numbers.



Gary K: Thanks for your well-informed insights. I'm still not sure, however, why anyone would NOT shrug off the national polls. Aren't you outthinking yourself? It's perfectly possible (and apparently true today) that McCain can gain in
1) states he already has, and
2) states he has no realistic chance of winning.
Both make the national margin closer without diminishing in any way Obama's real chances of winning. Let me know if I'm missing something.



If a McCain lead in a 1-day poll from an in-the-tank Republican pollster who chooses to leak his result to an in-the-tank Republican website is the best McCain can get, I'll take that, especially when Zogby weights his poll equally between D and R. In addition, this supposed sample came on the same day when every other tracker showed an increase for Obama. Pathetic...



The issue with Drudge's hyped-up one day results from Zogby is that it's not a one day poll. They likely do not have sampling that could be modeled accurately. For all we know, every voter in the one day could be 65 and over. There's a reason that it's a multi-day poll; it would be wildly inaccurate if it were a one day poll. Hey... wait a minute... that means that the one day results Drudge reported could be... WILDLY INACCURATE.



What puzzles me about the polling being reported is how this race is even as close as it is being claimed to be. Bush has the lowest approval rating of any President since approval was tracked, and John McCain voted with Bush 90% of the time, and has the same policy positions as Bush, so how does Bush have a 23% approval, yet McCain is 40%+ in the polls? Something there just doesn't add up. It's like the World Series, where it always seems to go to the very last game to decide the winner. Are the polls being skewed to make it appear to be a close race? Toss in the fact that McCain chose Palin as his VP candidate, when she thinks the Flintstones is a documentary, and I cannot fathom that Obama is ONLY up by 6-7%.



They are conditioning the public for the big flip on Election Day. They (republican operatives) are setting up plausible scenarios for the sudden 'surge' in the vote tallies that will put McCain up 51 to 49, despite exit polling to the contrary. The AP has floated a couple of polls, one saying that there was only a one point difference between Obama and McCain, and another one this week that suggests that there are 14% undecided...that gives them a point of reference when the skullduggery occurs and the punditry need to explain away the facts in favor for the fairy tale they all want the public to accept. Hope to be proved wrong…but be aware…


None of the above explanations for "why the race is close" seem plausible.

More plausible is

1. 6 points is not close
2. an A-A candidate represents generational change that is difficult to embrace by some



shhhhh. watch crazy bobby knight knock off texas and his legion of detractors. it will be a soothing change and limit any further unnecessary peaks in the body politic's bp. as i've said from my early days as a gamesage, it's all there in the numbers. when you get to be a numbersman as old as i am, you will look at this game...take a cool look at the numbers...and clink a skybrew with our bud Studs (thanks for your life with us Louis...) and wait to see Michelle and kids sharing the light with Barack at this century's BestofAmerica celebration.


The discrepancy between McCain/Palin support and Bush administration approval has one or more of the following explanations:

1) The McCain campaign's efforts to separate itself from Bush among traditional Republican voters have been largely successful.

2) The Bradley effect is already represented in the polls.


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.