10/28-30,08; 1,201 LV 2.9%
Mode: Live Telephone Interviews
Obama 50, McCain 43
Well, call me Nostradamus, but having written yesterday morning "my instincts say that we can probably expect a rise in Obama's support over the next few days", the graph of the weighted averages for the 3-day trackers has obliged rather promptly:
I should add that my prediction was not based on a belief that the real-world support levels would change; rather the reverse: support for Obama has been extraordinarily steady and whenever the trackers hit the bottom of that range they rebound. Its the trackers, not the levels of support, that change :-)
Comparing this weighted average with the weighted average for all yesterday's polls:
3-day trackers: O 50.4/ M 44.4/ Other & Undecided 5.2
All national polls: O 49.9 / M 43.5/ Other & Undecided 6.6
Which suggests the conclusion that remaining Undecideds will break for McCain about 2:1, and therefore we would project Obama to win 51-52% of the national vote based on yesterday's trackers.
Posted on October 31, 2008 6:19 AM
Good morning to everyone --
Is there any sense out there that a post-infomercial bounce is in the works for Obama? Or perhaps it was watched largely by those who already support him?
I also have a question for those who spend (too much?) time checking polls? Where does one get reliable information on what exit polls are showing in Florida's early voting? Thanks.
Posted on October 31, 2008 6:21 AM
Good morning from sunny Scotland. The only question that concerns me is the rate of refusals in polls. Do we know if it is high? Consistent? Biased? The BBC Question Time show from DC last night discussed this -- consensus was there is no obvious motivation for 'refusers' to be disproportionately voters unwilling to admit they are not supporting Obama, since there are various 'respectable' reasons for favouring McCain. As for the Idaho poll yesterday..no offence, but should that read WHYdaho?
Posted on October 31, 2008 6:27 AM
on MSNBC they said, New Mexico is a tie...
Posted on October 31, 2008 6:40 AM
I do not think there are any exit polls out there for early voting..However you can look at these early voting numbers. Some states give out voter demographics and party identification..please be aware though that a Democrat having voted does not mean a vote for Obama and a Republican having voted does not mean a vote for McCain.
Posted on October 31, 2008 6:41 AM
Thank you MArk for your comments about the "comments" - we must always try to disagree without being disagreeable!!
And I'd like to also make a claim for the Nostradamus Prize 2008. I said last week I thought the polls would narrow a little - it is very unlikely to be an Obama +10 victory in what is in fact a non-incumbent election - and that the polls would generally come together. It's been great to see Kos and Ras dovetailing.
If Obama is going to win by +6 (which looks very likely when the polls are averaged) we are likely to see some Obama +2 outliers (even maybe a tie in there somewhere) and some +10 outliers. The media will inevitably pick up on the former to create a bit of excitement as evidence of a late McCain surge and it will ensure that we have a little tension as we watch on election night.
Posted on October 31, 2008 6:43 AM
Yes, an uptick for Obama. A combination perhaps of good days for McCain dropping out and the early effects of the informercial. I wouldn't be surprised to see today and tomorrow show further modest upticks for Obama with the remaining days up to Tuesday showing something of a move to McCain.
But I might be wrong. I have no special source of information.
Posted on October 31, 2008 6:45 AM
@Terranus, you said "on MSNBC they said, New Mexico is a tie..."
It is probably another Mason-Dixon outlier for this morning :-)
4 days until voting. National trackers and polls (which should have been ignored last week anyway) show 5-6 point race and all seem to be converging on that point as was noted 3 weeks ago, 2 weeks ago and again last week.
State polls are very steady especially where it counts Kerry states plus IA and NM are all solid BO - generally in double digits margins.
Leading in high single digits in CO, NV and VA - either or puts him over 270.
Leading by about 4 in OH, FL, and NC - none are necessary.
Early voting (tradittionally a Republican strength) taking place in OH, CO, NC, VA, and FL indicate an energized Dem base and historic turnout. I beleive the Gov of CO said that if they get all the absenee ballots that were requested back, turnout could be 90+%. Extraordinary.
GOTV is key - so phone bank and knock on doors....4 days left.
Posted on October 31, 2008 6:46 AM
There is some information from news reports and from talk with in the true republican right that all independents should refuse to answer polls or to answer untruthfully. This is worrisome regarding states with traditional social conservatives. I have some knowledge of this topic since I am an independent and have friends and co-workers on both ends of the political spectrum. BTW I live in the reddest area of the Red state TX. Here an independent means I am often labeled a member of the radical left. At least my Obama yard sign is still intact...
Posted on October 31, 2008 6:49 AM
I just cannot see anything that will drag Obama below 50%. McCain's firing blanks, while it would appear that the Obama campaign will have 8 million volunteers out on ED.
Posted on October 31, 2008 6:51 AM
New AZU polls out today
AZ M- 51 O- 40
NV O- 48 M- 44
NM tie 46-46
I am not sure I buy New Mexcio, but since I know nothing about AZU's polling and did not see sample breakdown or party id#'s, I cannot give my (highly valued) opinion- LOL!. I hope Mark Blumenthal and Nate Silver will give some insight on this poll(New Mexico in particular) before the day is out.
Posted on October 31, 2008 6:54 AM
Just saw a Rev. Wright ad on CNN! It plays on the association argument to claim that Obama is too risky.
Posted on October 31, 2008 6:56 AM
Joe on MSNBC has gone to poll picking. It is an ASU Poll and the first time they have polled the state. The same pollster said that Obama is up 4 in NV and down by 11 in Arizona. There is no way it is tied in NM however it makes for good television.
Posted on October 31, 2008 6:58 AM
tick tick tick ...
Posted on October 31, 2008 6:59 AM
NM tied doesn't seem immediately credible. Last poll was Rass at +10.
Internals will be interesting.
Posted on October 31, 2008 7:03 AM
that NM poll scares me a little cause Obama was not very active there and there was compared to other states very litlle polling cause everyone thought it was a done deal. I just hope the Ras poll from 10/28 at +10 is accurate. Wouldnt hurt though to increase their push in NM a littl more.
Best about DKos is that yesterday Obama polled at 52-44
New Mexico being tied of the AZU poll must be an outlier because no one barring Zogby has shown the race there to be even within 5 points in the last 6 weeks.
PPP came out with a poll last night with a large sample (1537 LV's) that was in the field Oct 28-30, which shows Obama leading McCain 58-41.
Posted on October 31, 2008 7:04 AM
The good news about the Daily Kos poll is that after polling at plus 5 and 6 the past several days Obama polled at plus 8 yesterday.
Posted on October 31, 2008 7:06 AM
I see Obama maintiaining a lead up to election day and as you say keeping at or about 50%. However there is still the hope that supporters of the trailing candidate always have, i.e. that this is going to be one of those famous times when most pollsters end up with egg on their face.
Of course if McCain beats the poll predictions it will be becaue of a 'late surge' the 'Bradley Effect' 'Racism' or anything but mistakes by the pollsters.
Posted on October 31, 2008 7:07 AM
@politicalmomma2: There have been subsamples of pre-election telephone polls, but they are too small to be robust. Based on registration (last I saw), early voting in FL looks good for Obama, absentee voting looks good for McCain, and all this is hard to interpret anyway.
Any infomercial bounce will be hard to measure. (The R2K/DKos tracker obtained +8 yesterday, but there is a huge margin of error on that one-day margin.)
@James: It's often pointed out that in the primaries, Obama actually outperformed the pre-election polls -- although of course the Democratic primary electorate isn't the general electorate. (And of course there's the study that shows black candidates matching, on average, their pre-election poll results from the mid-90s on.) If there's a problem with refusals, my hunch is that it would be more with people who are unwilling to talk to interviewers at all than with people who (consciously or unconsciously) misrepresent their preferences. When I poked around back in the 90s, it seemed that reluctant or intermittent respondents (best available proxy for nonrespondents) tended to be more likely to express racist, anti-Semitic, and generally xenophobic views. My guess is that this phenomenon still exists, and induces some bias, but it is generally balanced or overbalanced by other biases. Obama may underperform in Appalachia -- there are hints of this in the primaries.
Posted on October 31, 2008 7:12 AM
This might make you fell better about NM. Like I said Joe is cherry picking!
Posted on October 31, 2008 7:13 AM
Politico/Insider Advantage Polls
641 LV's; MOE 3.7%
Obama - 48%
McCain - 48%
508 LV's; MOE 4%
Obama - 53%
McCain - 39%
814 LV's; MOE 3.4%
McCain - 50%
Obama - 47%
St. Louis County
LV's 858; MOE 3.3%
Obama - 53%
McCain - 38%
Posted on October 31, 2008 7:16 AM
Please do not tell me any person in this country watches Morning Joe? That hack gets Republican talking points each morning and throws that crap out like it is researched news.
You might as well turn your TV to another channel and watch Kevin Trudeau pimp you his new book in an infomercial.
Posted on October 31, 2008 7:18 AM
Does New Mexico report its early voting totals? Because that PPP poll says that 56 percent of the respondents had already voted. That number seems way high to me, but it should be confirmable.
(I know I've seen a site that compiles early voting numbers from across the country, but I can't find the links to it now ...)
Posted on October 31, 2008 7:20 AM
There are so many pollsters, with so many different models, I simply do not believe that Obama's support is soft enough for a last-minute "surge" to hurt him. Claims to the contrary are simply not credible.
Interesting and very detailed reply to McCain's pollster here:
I'm so pissed off. I put my usual early morning show in the radio, Spanish radio here in Miami, and what did I hear? Nasty ads saying that Obama doesn't support the Hispanic community because he is black, yuki!!!!
I mean tell me if it is race-baiting? They want to put Hispanics against Obama just because he is AA. That's just horrible, horrible. I am not going to watch TV, except MSNBC, until after the election. In the 11th hour the Republicans are launching the most disgusting, racist ads.
Posted on October 31, 2008 7:21 AM
Do not normally post, but this is one area that I have some knowledge.
Expect volatility in NM in the last few days. Since TX is pretty safely Rep state, Both parties are organizing west Texans to travel into East NM for neighborhood canvassing and GOTV pushes.
Posted on October 31, 2008 7:24 AM
Another interesting nugget about the Fox News Poll
Last week Fox News Poll had a margin of O+9.
Karl Rove wrote an op-ed piece yesterday on WSJ wherein he said:
If Mr. McCain is down by 3%, his task is doable, if difficult. If he's down by 9%, his task is essentially impossible.
Yesterday's Fox News Poll the result was O+3.
I leave the rest to your imagination :-)
Posted on October 31, 2008 7:26 AM
Posted on October 31, 2008 7:29 AM
That ASU poll was contact over 4 weeks...Hmmm (cross post, waiting for linky).
Posted on October 31, 2008 7:30 AM
summary from ppp:
Obama 55 McCain 42
New Mexico Results
Barack Obama 58
John McCain 41
Barack Obama 57
John McCain 42
Barack Obama 54
John McCain 44
Barack Obama 57
John McCain 41
West Virginia Results
John McCain 55
Barack Obama 42
Minnesota Senate: Franken leads
Al Franken 45
Norm Coleman 40
Dean Barkley 14
Mark Udall: Moving On Up
Mark Udall 56
Bob Schaffer 41
Merkley running away
Jeff Merkley 51
Gordon Smith 43
Dave Brownlow 4
Posted on October 31, 2008 7:31 AM
@saywhat90 & PostModernProphet regarding comments in the previous thread:
Yeah, I felt the exact same way you both do for a while. But lately I've actually been pleasantly surprised and feel at peace with the state of the race right now. In order for McCain to win, that would mean EVERY SINGLE NATIONAL and TONS OF STATE poll numbers for the past month would have to be wrong. Stop and think about that for a minute. I don't think the voting machines will be that big of a problem, or else the GOP in Nevada wouldn't be throwing a hissy fit right now. They are screwed out west and they know it. There's a reason McCain has embarked on this Pennsylvania or bust strategy, and a reason why he has started running robo calls in his home state. New Mexico, Nevada and very likely Colorado are gone (still gotta get out and vote, though, folks). The only way McCain can make up for it is by taking Pennsylvania. The only thing I worry about in Pennsylvania is that they don't have early voting, and there may be massive lines in Philly and people could possibly become tired of waiting in line for hours. Hopefully the Obama team does a good job of making sure every last dem in Philly and the rest of the state gets out, is willing to wait in line and vote. Judging by the enthusiasm advantage dems have, I am confident they can handle standing in line for 4 hours in order to change the course of the next 4 years.
Don't forget, there's always North Carolina, Ohio, Florida, Virginia, Missouri, Indiana, Georgia, etc. as back-ups even if things don't go according to plan in PA. These are back up plans that Kerry did not have 4 years ago. I am confident that Obama will carry several of those states, along with PA and the western states. I think us dems have a tendency to worry too much (for good reason, I might add), but the polls were never this good for Kerry or Gore this late in the game-- or even Bush, for that matter. Obama has the advantage of (A) Being a better candidate, in general, and having ran one of the most brilliant campaigns in history, and (B) Running against the record of quite possibly the worst President in American history.
(B) is an advantage that no poll can ever overstate.
Posted on October 31, 2008 7:32 AM
Yes, but four of the trackers also said O+3 on Wednesday. O+3 is the most popular figure this week. One final push by Mccain...
Posted on October 31, 2008 7:33 AM
The early voting numbers are available at this site:
I think Bernalilo county (33% of the state) has early voting and the latest numbers say 55% of the number that voted in 2004 have voted already. Another point to note it some respondents who have got absentee ballots and are ready to mail them or have already mailed them might say they have voted. Those votes may not show up on the SOS website since they have not received them yet.
Sure I do expect some volatility in NM but nothing close to a tie as suggested in the AZU poll is what I am saying....PPP's sample is huge for a state poll and internals looks very realistic (even though they do not weight by party ID the numbers they got match the state registration data).
Posted on October 31, 2008 7:34 AM
You are cherry picking one day results (most based on a consistently bad polling day/night on Sunday for Obama).
What I am in more interested in is why Sunday was so bad across so many polls...
Posted on October 31, 2008 7:35 AM
@Observer, you said "Yes, but four of the trackers also said O+3 on Wednesday. O+3 is the most popular figure this week."
Which 3 trackers are you talking about....we know the Rasmussen never went below 5 in reality but he somehow cooked up his numbers for Tuesday and Wednesday....Zogby and IBD use an outdated voter model and less said about them the better....GWU/Battleground is steady at +3 for so long that one can suspect they are copy/pasting the same result over and over.
Posted on October 31, 2008 7:40 AM
looking at the daily kos internals, mccain has picked up quite a lot of hispanic support over the past week with obama dropping a few points. any ideas why?
Posted on October 31, 2008 7:41 AM
The only +3 that mattered on Wednesday was Rass, and they pretty well admitted the following day it was an outlier. And:
1. Too much attention paid to spread figures, it's the support figures that count. Spreads are by defn 2x the MOE.
2. Even the Rass outlier still had Obama at 50.
Posted on October 31, 2008 7:42 AM
Of course I am. But Wednesday is not so long ago so it shows that the real national position might indeed be O+3 or thereabouts with a similar number undecided. (And there is another Sunday before the election!)
Best news I have seen for McCain so far is that there is absolutely no evidence of a big surge of young voters in early voting. It might yet happen but it might not.
Some pollsters models must be wrong. Until the counting we won't know which ones.
Posted on October 31, 2008 7:43 AM
In Miami Dade Co., South Florida the turnout has being FANTASTIC:-)
Here are the numbers:
A total of 361,322 people have early voted up to last night, that is 46.3% of the TOTAL people who cast ballot for President in 2004. We still have 3 more days of early voting :-)
Posted on October 31, 2008 7:44 AM
Some pollsters models must be wrong. Until the counting we won't know which ones.
And unfortunately for McCain the +3 trackers (2 of which I might add are spread over 5 days with really small sample sizes) all appear to be 2004 models which will absolutely not happen.
Posted on October 31, 2008 7:47 AM
"looking at the daily kos internals, mccain has picked up quite a lot of hispanic support over the past week with obama dropping a few points. any ideas why?"
Statistical noise. The sample sizes for such breakdowns are too small in a single poll to be meaningful. One needs to look across polls to get a sense of such breakdowns.
You are right some models might be wrong. However, what I think we are seeing in the last 5 years is that an aggreation of polls is the best way to counter this. Thus the pollster, election.princeton.edu, rcp, 538 models taking off. It takes out the noise and biases that may be present in any given poll.
I am no stats genius, but it passes my smell test.
Posted on October 31, 2008 7:48 AM
Bottom line is it's easy to pick up the phone and say who you will vote for but it's another thing to do it. I'm not crazy about 2 polls now showing Obama leading only +4-5 points in Pennsylvania. Now we have a "tied" poll in NM.
Aside from checking in on this website, I've stopped watching anything that has to do with this election on TV. It drives one crazy sometimes seeing the media jump from one rotten story or poll to the next.
Young voters have the ability to have a great impact in this election and so far from what I hear, they are MIA. That is bothersome and one can only hope that they GOTV b/c if they don't, I suspect it's going to be a long night on Nov 4th.
I must admit that I check constantly to insure my Mark Warner(VA)/Obama signs are still intact. I'm slightly concerned by a neighbor (actually his kid) down the street. So concerned, that I plan on moving my wife's car to behind the house on Monday night to prevent an "accidental" deflation of tires before Tuesday morning.
That new New Mexico poll samples only 200 people
Posted on October 31, 2008 7:53 AM
More on the youth vote thing:
I said this yesterday as well..it is not the quantity of the youth vote itself but the margin of the youth vote that will make a difference this time around.
Kerry won the 18-29 demographic 54-45(+9)..In most polls this time we see Obama leading that demographic almost 2:1 and this means a turnaround of 20 points at the least and this turns into about 3.0-3.5% in terms of the vote as this demographic represents about 17-18%(it was 17% in 2004).
Don't you think this margin of > 3.0% is extremely significant?
Posted on October 31, 2008 7:54 AM
@platanoman, you said "That new New Mexico poll samples only 200 people"
Is it....HaHa....and the one PPP has out samples 1500+....which one would you trust?
Do you have a link for that AZU poll?
Posted on October 31, 2008 7:57 AM
This was posted in a previous thread:
I haven't been able to find any information about early voting exit polls; however, many of the pollsters have identified in the internals of their poll those individuals that have already voted and who their selection was.
The four were:
Posted on October 31, 2008 8:02 AM
More on the ASU poll....
Among registered voters, McCain leads Obama 50 percent to 42 percent overall (undecided voters are not included in the percentages). The breakdown by state shows strong support for McCain in Arizona and Texas, while Nevada leans toward Obama, and New Mexico is even. The breakdown by state is:
* Arizona – 51 percent for McCain, 40 percent for Obama
* Texas – 50 percent for McCain, 41 percent for Obama
* Nevada – 48 percent for Obama, 44 percent for McCain
* New Mexico – 46 percent for each
The strength of intent to vote shows a slight advantage for McCain; 89 percent of McCain voters say they would “definitely vote” or “had already voted,” compared to 79 percent of Obama voters.
The Arizona State University-Southwest Poll was conducted by telephone Sept. 22 to Oct. 17 among a random sample of 1,208 adult residents in the Southwest United States (293 in Arizona, 190 in Nevada, 525 in Texas and 200 in New Mexico). The results from the full survey have a margin of sampling error of plus or minus 3 percentage points.
Very small samples except in Texas.
Posted on October 31, 2008 8:08 AM
I agree that even at +3 it will be a very good night for Obama. That is 5/6 points better than Kerry and he came very close. It is only if +3 is accurate and McCan mops up the undecided voter and/or the polls as a whole are a little out that McCain will squeeze out a win.
That ASU poll is kinda out-of-date huh?
Posted on October 31, 2008 8:09 AM
@Observer, you said " The four were:
I already commented about the other 3 and let me say something about Gallup now....I think everyone agrees that the turn out model this year even if the does not change as much as expected by the "Expanded" model it will still be better than "Traditional"....When you average those 2 LV models you will see Gallup never went below 4.5.
The bottom line to me is this (look at any national tracker or poll):
McCain's ceiling seems to be 45-46%
Obama's floor seems to be 47-48%
Now tell me who has a better chance of winning this one?
That new New Mexico poll samples only 200 people"
Are you for real? Can't say that I'm surprised.
I think that pretty much goes to show that we should leave the polling to the big boys. PPP has Obama up HUGE in New Mexico, with a much larger sample size, and Uncle Ras has him up by 10.
Really, how can a pollster even publish a poll with only 190 and 200 people responding? I wouldn't feel right publishing a poll with sample sizes that small.
So, enough concern trolling about this obvious New Mexico outlier. I believe Obama will take both states by solid margains. One Nevada GOP official pretty much said McCain has no chance to win Nevada, while some others are already trying to scream voter fraud. That tells you all that you need to know about how things are going out west for McCain and the GOP.
Posted on October 31, 2008 8:10 AM
Re: ignoring pollsters
Ever since the 2000 Florida debacle there has been an unspoken message passed around in conservative circles to either ignore or outright lie to pollsters. Hence a recent report I read or heard about how 'liberal' voters are far more likely to respond to exit pollsters thus exit polls are likely faulty.
Overall this concerns me in regards to the accuracy of polls.
As it stands I do not believe that we can count on even aggregate polls as being a clear indication because:
Models are being based on demographics which may ignore larger turnouts of minority and young voters.
Respondents lying so not to appear 'racist' or "against the flow."
The Bradley-Wilder effect.
The cellphone effect.
My brain just shut down so I'm certain I missed some effects so if you want to add please do.
However, what I am heartened by is the "poll of yard signs". My neighborhood, mostly educators and the like, tends to the left while the neighborhood nearest is heavily right. In 2004 my wife counted a roughly equal number of Kerry and Bush signs. This year we are running about 11 Obama signs (adding one a day recently) to only 2 McCain signs (it was three but someone removed theirs).
As such, I am hedging my bets and am subtracting 5% from Obama on state polls (still shows an Obama win). Overall, I cannot help believe that it will be either very close or a total Obama blowout (350+ EVs).
Posted on October 31, 2008 8:11 AM
The ASU poll is not only out-of-date, it was never a poll worth considering in the first place....imagine the MOE on a poll of 200 voters even if it was conducted over a single day.
The most funny thing about this whole publication to me is this: The results from the full survey have a margin of sampling error of plus or minus 3 percentage points. It is almost as if they were ashamed of publishing the MOE on each individual state's polling result :-)
Posted on October 31, 2008 8:14 AM
I've got a question (I've asked before but never gotten an answer). When Zogby does likely voters, he's basing it on prior voting history. What happens to 18-20 yr olds (no prior national election history)? What about 20-22 (no presidential election history? Do they get chucked? Are they included but weighted to mimic 2004 or 2006 turnout? It's not clear to me what in the heck Zogby is doing here.
Zogby internals Friday October 31
Total: Obama 50.1, McCain 43.1
West: Obama 55.6, McCain 41.2
Holding base: Obama 85.9, McCain 83.8
Independents: Obama 50.0, McCain 35.2
Already voted: Obama 53.8, McCain 41.4
Males: Obama 48.8, McCain 43.7
Females: Obama 51.3, McCain 42.4
White: McCain 51.8, Obama 40.7
Hispanic Obama 70.8, McCain 25.9
AA: Obama 94.8, McCain 2.4
Moderates: Obama 64.2, McCain 28.2
The last two weeks, the differences have ranged from Obama +7.7 on 10-23 to Obama +4.9 on 10-26.
I have listed the internals from 10-26 below so you can see the changes from today:
Zogby internals Sunday October 26
West: Obama 45.9, McCain 46.6
Holding base: Obama 88.2, McCain 88.3
Independents: Obama 49.7, McCain 35.9
Already voted: Obama 57.3, McCain 39.5
Males: Obama 43.7, McCain 47.6
Females: Obama 54.8, McCain 40.8
White: McCain 52.9, Obama 40.9
Hispanic Obama 69.7, McCain 22.9
AA: Obama 90.0, McCain 7.2
Moderates: Obama 60.1, McCain 30.5
I had been pointing out that Zogby's internal West region numbers went against common sense last week, which showed McCain ahead in a region which includes huge Obama states such as California. Between Sunday and today (Friday Oct 31), that West region number has moved Obama +9.7 in the region. The key is obviously not California, but rather Nevada and Colorado (if Zogby has CO in the West and not central - not sure?).
The male vote has reversed itself, now favoring Obama. On the other hand, the female vote has narrowed. Both of these changes could be statistical regression to the mean. That is, statistically, it would not be surprising to see Males move back to a more even position (perhaps Obama +2) and Females to a larger Obama position (perhaps Obama +10) with the overall number at Obama +6.
The only other change which on surface is noticeable has to do with already voted. Again, it makes sense that this change is statistical as well. As more people vote, the difference between Obama and McCain will get closer but reach a resting point The last three days Obama had been at 53.7, 52.3 and 55.4).
Finally, the AA vote is more Obama now as well as the moderate vote, although again these changes may be noise.
Posted on October 31, 2008 8:15 AM
It could also be zogby pulling numbers from out of the air...
Posted on October 31, 2008 8:22 AM
I may be wrong but my understanding is that the likely voter models and prior voting history does not refer to the actual voter but to the demographic. That is, the result is not weighted on the basis of the actual person who has never voted before because they were too young but is weighted on the basis of voting patterns of people of that age in the last election[s]. It would be like saying that if four 18 year olds say they are voting Obama then that only counts as one person voting Obama because in the last election only 1 in 4 18 year olds voted. (My numbers are totally made up, BTW)
As to question above on Hispanics, using Zogby internals, Obama was +47 on Sunday; he is +45 today.
Posted on October 31, 2008 8:23 AM
Yep, Obama has a far better chance to win this. In fact I don't really think it's necessary for Obama supporters to go out and actually vote (!)
Incidentally I now see here and elsewhere talk about people lying to pollsters. Sure it happens but it can happen either way and I don't see it as a factor in current poll results. Of course it will probably be on the top ten list of excuses if the pollsters get it wrong.
Posted on October 31, 2008 8:25 AM
RE: The polls could be wrong
I remember back to 2004, hoping, praying that John Kerry would win - thinking, "the polls could be wrong..."
The AZU poll is essentially a student project. Don't pay any attention to it.
Posted on October 31, 2008 8:26 AM
The thing to remember about yardsigns is:
If there are more for your candidate in your area then your candidate is winning the national vote.
If there are more for the other guy but your guy has more than your candidate had last time then your candidate is winning the national vote.
If the other guy has more yardsigns then it doesn't mean anything.
Posted on October 31, 2008 8:30 AM
Hah, hah, very funny.
Actually, I was discussing a similar idea last night as to when the media may call the election. The best response was they will try to wait as long as possible to keep viewers watching. Yet, as always, I'm concerned about the media calling an election early and the affect it can have. I have always thought the media should not be allowed to do ANY reporting until all polls have closed.
Consider the situation in AK. What if the polls are nearly correct and Obma easily wins VA, OH, PA and FL. That is enough for the media to announce him the victor as early as 8PM. So, people waiting around in AK decide "ah, why bother". In that case you could potentially see AK voters somewhat disgusted with Stevens not bothering to vote, and Obama amazingly winning Alaska! Wouldn't that be a kick!
Posted on October 31, 2008 8:32 AM
Re: "Hence a recent report I read or heard about how 'liberal' voters are far more likely to respond to exit pollsters thus exit polls are likely faulty."
I too have heard this for exit polls, but I'm not sure there's any evidence that this holds true for telephone polls.
At least in 2004, there was no systematic biases in the polling averages on RCP; Bush overperformed his polling numbers in FL, for example. I'd be curious if you could identify something that showed an "enthusiasm bias" in poll responders. If anything, it appears the pollsters are having a more difficult time reaching Obama supporters.
Posted on October 31, 2008 8:35 AM
Any one with a early look at Rasmussen? Nate Silver says that Ras showed a good result in the field Thursday, 1 day after infomercial. may be a 1 or 2 point improvement???
Posted on October 31, 2008 8:36 AM
Hey, don't dismiss the POWER of yard signs! They have complete and utter control of the people you know!
Seriously, though, what I find really, really interesting this year is that while there are many small Obama signs in this part of VA and many fewer pro-Rep signs, the Republicans seem to be trying to make up for it by having really BIG signs along the main roads. And to bad for Gilmore(R) in VA, there are quite a few McCain(R)/Warner(D) signs out.
Posted on October 31, 2008 8:38 AM
Yes, you're right. I was supposed to be referring to "exit polls" not polls in general. Beyond the cell phone effect though, I was not aware pollsters were having difficulty reaching Obama supporters.
Posted on October 31, 2008 8:43 AM
As to actual polls, all I can say is that some conservatives I have talked to have outright told me that they have lied to pollsters and that they have "under their breath" spoken to other conservatives about intentionally skewing polls. Not necessarily to lie but more, it seems, to make the polls look bad. One guy I know did a long poll and intentionally answered questions in a contradictory manner to see if he could annoy the pollster. You know "I'm voting for Obama" vs. "Bush is a great president."
Posted on October 31, 2008 8:48 AM
The problem with 3 day trackers is that one off day of polling can cause a perturbation. If you look at the daily's from any pollster and adjust to 5 day trackers, you will see significant smoothing. No pollster is impervious to a bad polling day.
As for some of these more dubious polls (i.e. IBD/TIPP, GWU and FOX), Nate Silver over at 538 has already illustrated the flaws in IBD/TIPP methodology. GWU is a bit of a head scratcher in that it is a poll run by a rep & a dem but what is most perplexing is that it is unchanged for days. Maybe it has a bias but it is showing a remarkably stable period when other polls are bouncing all over the place. Now for FOX, Opinion Research decided to change their party ID weighting from 43D 37R 16I to 41D 39R 16I for the most recent +3% Obama poll. If you use the previous party ID weighting the result would be 50.4%-O 43.6%-M delta +6.6-O.
The good news is that the sites like Pollster, RCP and 538 average these polls to remove some of the noise / manipulation.
Posted on October 31, 2008 8:51 AM
The conventional wisdom re signs (at least here in NH) is that signs on public property or along roadways don't sho much except some effort on the part of volunteer. A yard sign in front of a person's home is an indicator of personal commitment.
Posted on October 31, 2008 8:59 AM
"What I am in more interested in is why Sunday was so bad across so many polls..."
Because all the Obama supporters were out canvassing and not home to participate in polls?
Posted on October 31, 2008 9:05 AM
I agree with zoot on the yard signs at least for NH. I see many signs along the public roadways but these don't mean too much other than effort from the volunteers. I live in a southern NH suburban community which would normally lean heavy McCain and the majority of yard signs are Obama. This by no means would indicate that Obama wins my town but it does show a greater level of enthusiasm.
Posted on October 31, 2008 9:08 AM
Can somebody confirm the ras numbers?
"What I am in more interested in is why Sunday was so bad across so many polls..."
Weekend polling has been shown tightening for some reason which seems to be receed during the week. Who knows, maybe PHGrl is right and the Obama supporters are not home. I know I am not available for pollsters on the weekends until after the 4th. Unless they catch me on my cell phone.
Posted on October 31, 2008 9:13 AM
PPP on their blog had a piece about reaching Obama supporters. Specifically they noted that it generally took more calls per voter to reach Obama supporters as compared to McCain supporters.
I can't find the piece right now, but it's somewhere at:
He suggests being a bit wary of single day polls (a la Rasmussen state polls), because they obviously don't do call-backs. (Although one wonders if Rasmussen's party ID adjustments don't at least partly compensate for such discrepancies).
We'll see who's right on Tuesday, so I'm not putting too much weight on the idea that the trackers and other single day polls are underrepresenting Obama systematically, but I thought it was at least interesting.
Posted on October 31, 2008 9:15 AM
Just found it:
Posted on October 31, 2008 9:16 AM
Conservatives do play games with pollsters. In fact me, my wife and friends all registered as Dems during Rush Limbaugh's "Operation Chaos" with no intention of voting for Obama. Further more conservatives do not get depressed or complacent like Dems because our guy is down in the polls. We are only energized by the pro Obama MSM and their bogus polls. Meaning we will vote reguardless.
Obama followers need that extra emotional support the fake polls provide to them. This also presents a double edge sword for Obama.
Simply because his followers actually believe Obama has the race in the bag so they don't make it to the polls. Perfect exsample is the MIA youth vote.
Posted on October 31, 2008 9:17 AM
Here's another one, Strategic Vision(R) PA - Obama 49 McCain 44. Just as I said yesterday when M-D came out with a 4% spread, the reps are trying to change the narative through polling. PA is the only hope for McCain so let's make it look closer to keep up the enthusiasm. Show me an indepedent pollster that views PA as close and I'll buy it. IMHO SV(R) & M-D don't fit that description.
Posted on October 31, 2008 9:25 AM
@ southern angler
The exact reason why this country is waking up to the neocon side of the republican party and why I have left the party myself. You all think this is a game. It is against the law in some states to do what Rush Limbaugh's "Operation Chaos" did. You are proud of that?
Posted on October 31, 2008 9:32 AM
please visit here on Nov. 4th. at 7 p.m. ET.
And let's find out whether your opinion is right or not.
Posted on October 31, 2008 9:36 AM
and according to the PPP poll in New Mexico and Obama's huge lead there:
"This may begin to sound like a broken record, but he's also banked a huge lead with those who have already filled out their ballots. 56% of poll respondents reported having done that, and within that group Obama is leading 64-36. He is up by a much more modest 50-47 tally with those who have yet to vote."
Posted on October 31, 2008 9:41 AM
It's getting tighter RAS Obama 51 McCain 47
By election day , it will be a dead heat.
I'm in the new home business and a realtor from Deleware brought me a black customer on the day Obama picked Biden as his running mate. This realtor was really excited about Biden. She assumed that I too would share in her excitment because she ask me if I was as excited about Biden as she was. Instead of telling her what I really thought about Obama and Biden I simply lied and told her That I was an undecided independent and that I was still weighing my options. Keep in mind her black clients were standing there listening to our conversation. Anyway I got the deal, but felt kinda dirty about it. I'm sure that I am not the only conservative to play this game.
Posted on October 31, 2008 9:43 AM
Doesn't that simply prove a contention that Reps are far less American than Dems. Essentially what you and others are doing is lying, making some feel that you guys are worthless. How can anyone trust the right?
Posted on October 31, 2008 9:46 AM
Interesting read from PPP. I wonder if anyone is taking this idea and putting it into a model. Namely, make a poll where you TRY to reach either McCain and Obama supporters then weigh a normal poll on that basis.
For a simplistic example, suppose it was discovered that reaching Smith supporters was 20% harder than reaching Jones supporters. Then in the regular poll if Smith and Jones are tied at 50/50 then you weigh the results so that Smith actually leads 55/45.
Posted on October 31, 2008 9:52 AM
The only thing this proves is the conservatives are also conservative with their opinions. We don't feel as if we have to show all of our cards. It's all part of the game. We know how to win, that what really matters, not the polls.
Posted on October 31, 2008 9:53 AM
southern angler: We know how to win ...
Now THAT can't be disputed.
Still, that's the problem. As an optimistic pessimist, I recognize the need to deceive in order to win, but wouldn't it be great if your ideas were solid enough that you wouldn't have to?
Posted on October 31, 2008 10:19 AM
if this double posts, forgive me, I got a strange error
Posted on October 31, 2008 10:21 AM
Comments: (you may use HTML tags for style)
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.
Please email us to report offensive comments.
See our comment policy here. Note that we require commenters to share their email address via Typekey. We will never share your email address with anyone without your explicit permission.
MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR