Oct 06 2008

Where Will The Polls Head Next?

Published by at 7:46 am under 2008 Elections,All General Discussions

There are some modest indications polls may be heading back toward McCain-Palin this week. Some of this can be seen in the state polls which bounced towards Obama during the debate on the Financial Bailout/Rescue package debate. For example, Colorado is apparently done with its bounce if you look at the current RCP polls:

Poll Date Sample Obama (D) McCain (R) Spread
RCP Average 09/19 – 10/01 47.6 44.6 Obama +3.0
Denver Post 09/29 – 10/01 625 LV 44 44 Tie
FOX News/Rasmussen 09/28 – 09/28 500 LV 49 48 Obama +1
InAdv/PollPosition 09/23 – 09/23 505 LV 50 41 Obama +9
CNN/Time 09/21 – 09/23 794 LV 51 47 Obama +4
Ciruli Assoc. 09/19 – 09/23 501 LV 44 43 Obama +1   




The state was tied, then bounced towards Obama, and is now falling back towards McCain. Remove that 9% fantasy poll and Obama’s lead is under 2% – a statistical tie. I see something similar with RCP’s Virginia polls:

Poll Date Sample Obama (D) McCain (R) Spread
RCP Average 09/27 – 10/01 49.0 46.6 Obama +2.4
Mason-Dixon 09/29 – 10/01 625 RV 45 48 McCain +3
CNN/Time 09/28 – 09/30 684 LV 53 44 Obama +9
InAdv/PollPosition 09/29 – 09/29 436 LV 51 45 Obama +6
FOX News/Rasmussen 09/28 – 09/28 500 LV 50 47 Obama +3
ARG 09/27 – 09/29 600 LV 46 49 McCain +3

Same bounce right back to the state the race was in before the financial crisis hit (and was artfully extended by the Democrats who played the GOP like a fiddle).

I am interested in seeing today’s Gallup Tracking Poll after I noted this comment tagged to yesterday’s poll:

The race has been slightly closer on both of these two individual days (Oct. 3-4) than the previous two days. Obama held particularly large leads over McCain from Oct. 1-2, possibly resulting from Americans’ focus on the Wall Street financial crisis and congressional rescue plan dominating the news at that time. Since then, support for Obama has remained about the same, at the 49% to 50% level, while support for McCain has increased slightly, with an associated decline in the percentage of undecided voters.

The polls may have one more large movement in them. The question is will it be towards McCain if it shows up. I think it will be, but who knows – that is my hopeful speculation. What I do think played into this was moderates telling pollsters they were going to vote Obama as the GOP dragged their feet. Yes, few liked the package, but as with all distasteful medicine many just wanted the ordeal over with. Right or wrong doesn’t matter. The mood was no gambling the economy. The GOP needed to remember that they got one big gamble from America already with Iraq. They were not yet entitled to another blind jump into potential hell.

And I am convinced the Democrats tried to make the economic situation worse to help Obama. Pelosi definitely let the first House vote tank with her ridiculous speech right before the vote and the fact many of her leaders and friends made SURE the vote failed, offsetting the GOP support easily. Remember, the Dems could pass this just about anytime – and they did.

With the vote now passed and Congress adjourned and out of the way it comes down to the final stretch. I still think Obama’s numbers are inflated. And to explain how little they have to be off to make a big difference let’s just use Obama’s 7% lead. First off there is a 2% theoretical margin of error, which means 98% of the time this will be the outcome. But let’s ignore that for a moment and just assume the Obama numbers are only 96% accurate, and 4% are off and in the wrong place. That 4% shift would change Obama’s 7% lead to a 1% deficit. 

As can be seen it doesn’t take a lot of error to be way off in polls. Being 96% accurate is something most polls would love to be, they are not. What happens is they usually are way off until the day before the election, and even then they are not 98% correct. Furthermore, this year has demonstrated pollsters cannot rely on their historical models. That is why we see polls in the same state 15% off from each other.

The polls show me, an engineer, that they are have serious problems measuring the mood of the country. The wild gyrations and obvious outliers indicate a methodology out of control, not focusing in on an answer. 

Anyway, we shall see. I think this race will tighten back down, and it may have already begun to.  We need about 3 more days to see if this is true (weekend polls are notoriously flakey).

Addendum: One more factor seems to be folding into the mix as well, pollster trying to survey over cell phones.

The USA TODAY poll of more than 900 young Americans, taken Sept. 18-28, included interviews by land lines and cellphones, which increasingly are relied on by younger Americans. (Three-fourths of those surveyed have a cellphone; one in five report using social networking sites such as Facebook and MySpace to discuss the election.)

By a crushing 6 to 1, those polled say Obama, 47,understands “the problems of people your age” better than McCain, 72.

When you look at cell phone polls they tilt massively to the left. But we all know the nation is not tilting massively to the left, not even across the younger generations. We conservatives have kids too (possibly more than liberals) and our kids are primarily conservative. Yes, younger people are more idealistic and fantasize of Nirvana, but they also are signing up for military careers by the thousands as well, and these youngsters support McCain-Palin. I think it will be more a wash, not some wild 6 to 1 lopsided race.

Is it simply some kids respond to cell phone polls more than others?

Update: Speaking of bizarre polls, check out this Columbus Dispatch poll showing Obama up 7% in Ohio (if that were true Obama would be up 15% nationally). Intriguing questions surface when one looks at their internals (presented in a very confusing manner).  They show three columns for McCain-Palin, Obama-Biden and Undecided – which supposedly will tally up to a 42-49-8 result, respectively.

But here are some oddities. First off is the Clinton voters. Clinton blasted Obama in Ohio 54-44%. Clearly there are more Clinton voters there than Obama voters. Yet McCain supposedly is taking 18% of Clinton supporters, which equates to almost 10% of all democrat primary voters. And other 15% are undecided. Obama is only pulling 2/3rds (66%) of Clinton supporters in a state she won. Yet he is supposedly ahead 7 points? 

McCain takes 7% of democrat voters, Obama 5% of GOP voters. But 8% of Dems are still undecided, only 5% of Reps. A whopping 15% of independents are undecided (and they decide elections). The poll doesn’t show its party affiliation model, which must be skewed way left. Instead they present this convoluted graphic to try and shore up the conclusions. Remove this poll from the RCP average and the Obama lead is 2.5%, as statistical tie. The other wild poll in the mix is a Quinnipiac poll showing a +8% lead, which is why this state is not a tie right now.

Polls in the same state showing a +1 McCain or a +8 Obama indicates the polls are erratic.

3 responses so far

3 Responses to “Where Will The Polls Head Next?”

  1. Aitch748 says:

    I’m also reading posts on other forums from posters who have been called to participate in some poll or other, and the first question is, “What political party are you affiliated with?” and when the poster answered “Republican,” there was a pause at the other end, then a quick “Thank you for your time” and a hangup.

    I wonder how often that has been happening in these polls.

  2. Mike M. says:


    Have a look at http://wizbangblog.com/content/2008/10/05/the-sordid-business-of-opinion-polls.php

    The polls are really getting odd this year. And I have also read reports of McCain voters outright lying to pollsters.

  3. […] Polling results are good for McCain! […]