I'm short on time today, so let's keep this on point: John McCain is in deep trouble. In spite of some incremental gains that McCain has made in some of the national tracking polls, the set of state polling that follows is so strong for Obama that he continues to hit record marks in all three of our projection metrics. We are now projecting Obama to win the election 90.5 percent of the time, with an average of 346.8 electoral votes, and a 5.4-point margin in the national popular vote.
There simply isn't any good news in here for John McCain (all right, he's kicking butt in Oklahoma). The only swing state poll that he leads is the SurveyUSA result in North Carolina, but even there, Obama has bounced back from a 20-point deficit in a SuvreyUSA poll taken shortly after the Republican Convention.
Moreover, Obama's position in the electoral vote remains even stronger than his position in the popular vote. We project him to win all of John Kerry's states by at least 6.9 points (New Hampshire remains the weakest link). We also project him to win Iowa by 12.5 points, New Mexico by 7.7, Virginia by 7.3 and Colorado by 6.9. Getting this race back to a tie might not be sufficient for John McCain; he might need to pull ahead by 1 or even 2 points nationally to mitigate Obama's edge in the battleground states.
10.08.2008
Today's Polls, 10/8
@ 3:09 PM
Subscribe to:
Post Comments (Atom)
636 comments
schnikeys. i'm loving that trendline, even if there's a statistical blip today. thanks, Nate!
Above 90% YES!!!!!
McCain's debate performance didn't do him any favors. I expect Obama to maintain a 7-9 point national average after a few days.
Good day to be an American.
Amazing!
I asked this question elsewhere but perhaps it is more appropriate on this thread anyway. Does anyone know how often - if at all - has a candidate been in John McCain's position at this point in the campaign, but gone on to win nevertheless? It might help to put the impressive-looking polls into better perspective, at least for me.
why the heck does OK get so much polling? We KNOW McCain wins OK, sheesh. TX is closer (I think) than OK and it is poll bereft.
Good news for Obama, keep it up!
Yay!!!!!!!!
Nate - are you going to take early voting into consideration? (Think CA Primaries)
oh, and voter registration has hot 90% in MN. And we can register at the polls as well. Massive turnout here.
Single Digits!
Good job on the +40 in Oklahoma repubs! lmao! Too bad Obama is winning all the swing states!
I love that this is coming in factoring the smear campaigns of the McCain-Palin ticket.
Where can McCain retreat now that bald-faced character attacks are failing? Do they just keep it up until November hoping for an erosion over time? I don't see much else.
wow... what will obama's coat tails do for the senate and house races, i wonder...
time to start talking about coat tails more.
What keeps knocking North Dakota back down? It was 21% yesterday and I see nothing today to mitigate it...
In a month, Obama win percentage has gone up from ~40% to ~90%. Amazing.
i don't even know which races are up, but anyone have info on hot governor's races?
If Oklahoma is affecting North Dakota, I want some of what the model is smoking.
I've been hearing a lot (see NPR today, also this article: http://www.thesmokinggun.com/archive/years/2008/1007083aig1.html) about issues with voter registration, and how that might impact voter turnout. For example, one story discussed how potential voters are being threatened with arrest if they show up to vote and have prior issues, such as owing money to the state. Not to mention, all the issues with the voting machine technologies. How much can you account for this?
looks like the smear tactic, just like every other stupid gimmick, has failed.
McCain and the Republicans are weak.
Barack will need an 8 point cushion heading into election day. As David Gergen explained last night, we don't know the outcome of a general election with a black. If he leads by 4 on 11/3, McCain may very well eek one out
Before we count votes that have yet to be cast, let us remember that a certain percentage around 7, if I recall correctly) will say they're voting for Obama but then pull the lever for McCain. Do any of the models you run take that into account?
JonathanCR said...
Amazing!
I asked this question elsewhere but perhaps it is more appropriate on this thread anyway. Does anyone know how often - if at all - has a candidate been in John McCain's position at this point in the campaign, but gone on to win
--
NEVER
Gore was up by 8 once in the Gallup tracker in Oct 2000. That was the closest.
If anyone from any level of the Obama campaign reading this, please make it priority to prevent "voting machine fraud" in this election!!!
Obama 375!!!
Damn I posted wrong link. Here's the right one:
http://www.computerworld.com/action/article.do?command=viewArticleBasic&articleId=9116465&source=NLT_AM&nlid=1
jakam said...
What keeps knocking North Dakota back down? It was 21% yesterday and I see nothing today to mitigate it...
read the FAQ. Nate explained this effect a while back. it is all about the demographics in the model. for example, if you can say that UT is x points behind CO in obama support and CO goes up n points then you can say that if they have similar demographics there will be a certain rise in both states based on what happened in the polled state... he explains that concept better, so go look.
Oklahoma... where the mullet lives on.
I, for one, won't be satisfied until I can drive from Florida to Washington in all blue states.
Great site, robust analysis.
Would you put a "finish line" (Nov 4) on the Super Tracker so we can judge just how steep a curve McCain will need to recover? Thanks and keep up the good work.
Cheers
And if anyone from bumper sticker industry reading this, please make a "THAT ONE for PRESIDENT" sticker. I'll take one.
Awesome polls. :)
But, I have to wonder about states like OK - they are obviously single issue voters, as by voting for McCain they are voting clearly against their wallets.
Before we count votes that have yet to be cast, let us remember that a certain percentage around 7, if I recall correctly) will say they're voting for Obama but then pull the lever for McCain. Do any of the models you run take that into account?
Huh?
If you look at the crosstabs, that SurveyUSA NC poll is actually bad news for McCain.
They have AfrAmer turnout at 20%. AA turnout in NC was 26% in 2004.
Its almost guaranteed to go up this year (Not only are AfrAmer more enthusiastic about Obama, southern white evangelicals are less enthusiastic about McCain).
If you normalize to 29% AA, Obama is doing alright in this poll. He actually is leading by 6.5 points.
myfeatfelloff said...
Barack will need an 8 point cushion heading into election day. As David Gergen explained last night, we don't know the outcome of a general election with a black. If he leads by 4 on 11/3, McCain may very well eek one out
--
That is not true. See the posts here about the Bradley effect. The racist vote is likely already included in the polls. Nate commented specifically on that report and the 6% number is off base anyway, but even if it was true the discount is already taken into account in the polls. Gergen was voicing a common concern, but he was factually wrong about what the Stanford study found
Guys, read up a little before posting about the Bradley effect (lying to pollsters about voting for the black man) like it is some iron-clad given. Many feel that it doesn't exist at all, and even if it does, who knows how much it will be, and whether it will be counteracted by other effects.
myfeetfelloff, you got your name wrong, it should be mybrainfellout. you're a moron.
jeffcassens, you're talking about what is commonly known as the bradley effect. the problem is that it doesn't exist anymore. you're also a moron.
thanks for playing guys.
Not to be inflammatory, but...
As a Christian, I do not see how anyone could support the disgusting tactics now employed by McCain. I previously donated to his campaign. I now will always regret that any of my money went into this xenophobic fear tactics and smearing.
John McCain's campaign is single handedly bringing down our country's level of discourse, and I think, is actively raising the probability of political violence. I am ashamed that his "base" seems to be other evangelical Christians like me.
I wish I could apologize on behalf of my cultural group. Maybe we'll surprise everyone and get even more tepid in our support of McCain because of his immoral attacks, but I wouldn't count on it because of our willingness to be used cynically by politicians grandstanding on the abortion issue.
Diageo/Hotline and Zogby are crap.
1) Diageo has only 300-person per day sample. It's the small sample ever.
2) Zogby has only Obama +2 but look the numbers. They have Obama with the same numbers than McCain with his party and has Obama lead by 9 with independents. It's not possible than with these facts he leads only by 2.
I think Zogby has the same % DEM/REP in his sample.
Interesting proposition bet
Will OU's margin of victory against Texas exceed McCain's OK margin of victory against Obama. Just for context, OU is a 6.5 point favorite while Nate has McCain as a 26 point favorite (too low?) I would set the odds at 8-1 but maybe Nate could the point spread as a starting point for finding the fair odds.
we will find out what the % of the "bradley effect" on election night
Jeffcassans -- if you have evidence for your 7 point number, you need to produce it. Nate has discussed this issue extensively and there is basically no evidence that large numbers of whites will lie about voting for blacks based on elections in the last 10 years.
race will (fingers crossed) not be a decisive factor. i heard someone make the analogy that "when you're drowning you don't stop to look at the color of the hand reaching down to save you"
americans are in a lot of trouble and so far john mccain does not appear to be offering any substantive solutions.
jeffcassens said...
Before we count votes that have yet to be cast, let us remember that a certain percentage around 7, if I recall correctly) will say they're voting for Obama but then pull the lever for McCain. Do any of the models you run take that into account?
______
aka the bradley effect. this has been discussed extensively on this site and disproven. try searching for it on this site if you want to find out more.
Nate, Are you aware of McCain's benzodiazepine and "upper" usage documented on Americablog?
Robert,
write them a letter and ask for your money back!
Oops - messed up my quote above!
TSVictoriaBC said:
"And if anyone from bumper sticker industry reading this, please make a "THAT ONE for PRESIDENT" sticker. I'll take one."
Just print it on your printer and clear tape it on. This is actually easier to remove than a bumper sticker later.
Over 90% is TREMENDOUS NEWS!!! For John McCain!!!
Robert - it's good to know that not all Evangelical voters are wowed by McCain's smear tactics.
[matt w] "NEVER
Gore was up by 8 once in the Gallup tracker in Oct 2000. That was the closest."
Thanks - I thought that was probably the case, but wanted to check! That makes it all the more encouraging.
Obama is doing good, but to totally seal the game, it would be good to see numbers over 50% in Ohio. Once he has Ohio down like he has Pennsylvania down, McCain goes down to a >1% of winning.
How much of a stretch is it from "that one" to the "N" word? Let's hope McCain makes the leap!
The House, The Senate and The White House. Now if only we could get the Supreme Court!
ugh, please give it a rest about the bradley effect.
Even if there were a bradley effect. It will in no way compare to the the new voter turnout effect, black voter turnout effect, the youth voter turnout effect, and the cell phone effect which we KNOW will be factors in this election and ALL will likely be in favor of Barack Obama.
another great day of polling results - for Obama
stop the concern trolling folks about the need for a HUGE Obama lead in the national trackers going into Nov 4th.
first, it is a state by state election, and all the state metrics are setting up perfectly to ensure an Obama victory as Nate explains above with enough states above the margin to clinch @ 270.
next, millions of votes are already being banked for Obama now & over the next 4 weeks in advance of Nov. 4th.
this election is like no other in many ways, including the % of votes that will be cast in advance thru early voting & absentee ballotting increases.
those favor Obama, as do registration, GOTV, and the polls [despite cellphone undercounting] - so voter fraud & supression will be much more difficult.
I predict that election eve exit polls will be shown to be skewed due to the variables listed above & the inability to poll the absentee/early voters...
just look at GA - who thinks the current polling will turn out to be accurate there ?
I would guess with high AA turnout & 3rd party vote splits, the final result will be quite different...
dannylandulf said...
PSST. A new ras poll is dropping tonight that you all might find...whats the word...interesting:
Ras - GA - President
Obama 45%
Mac 43%
is this for real ??
It's unfortunate the Obama lead isn't even higher given the comparative strength of the candidates.
I think the reason the smears are having the opposite effect for McCain is because the swing voters have seen Obama now in the debates and know how ridiculous the smears are. He's not the radical, no-substance, not-your-values guy they were told he was.
The smears only resonate with the partisans who foam at the mouth over any and all democrats.
tsvictoria--
cafepress.com has some "That One" items--can't remember if they have a bumper sticker or not, but I think they do. Go to cafepress and search on "That One."
As far as the 8% so-called Bradley effect, Obama is half white so in his case it's only 4%.
Rich Merritt(.com)
Hi Nate,
I remember your post recently about rcp and bias. And i'm just curious as to if you have any idea why they include the zogby tracking, while still excluding the r2000 tracking, especially considering by your weights are the same for them.
clarkejeffrey
one more thing to point out on the SUSA NC poll, it has Obama winning 75% of dems (also winning independents 51-39). If he gets this many it's likely a huge win. NC always goes red, but it's registered dems voting republican. I have no idea why. Anyway, democrats outnumber republicans by something like 15%.
mattw - I already did. (not that I think it will happen - it was in the primaries)
I actually called 7 or 8 of his campaign offices and tried to convince volunteers they were taking part in something shameful if they supported what McCain was doing right now. Most actually honestly agreed that the attacks are not justified. I may have actually convinced one not to volunteer any more. Additionally, two PAID volunteers actually said that they were fully aware that the ads were not true, but that they didn't care if it helped the R's win. This attitude should turn people off. It is the way towards very bad stuff, historically speaking.
dannylandulf,
Are you Sh**in' me!?
are you a subscriber? when is the public release?
scratch those RAS poll numbers...I just debunked them.
Just a note: The SUSA poll actually offers an alternative data set, noting that they've tied this poll to a predicted 20% African-American "likely voter" turnout. If you increase AA turnout by 100,000 in their model, you end up with an Obama win, a Kagan win for Senate, and a Purdue win for Gov in NC.
20% is probably a bit low even by historic standards. But it should also be noted that there have been 170,000 new AA voter registrations in NC this year - more than enough, with 60% turnout of registrants, to make up the discrepancy in sample size in this poll.
SUSA has not done a great job polling NC this cycle, but I think their "alternative" data set comes a lot closer to an accurate picture than does the published one. I still think it signals a floor, not a ceiling, of Obama support in the state. Looking at SUSA's polling of NC-8 (the Kissell-Hayes race), Obama has a huge lead in a district that Bush won handily in '04. This is true in most regions of the state, and seems to indicate a strong win - not a "squeaker" - in the making.
Nate, what is up with the Hotline poll? It's not only an outlier, but Obama's numbers tanked on Tuesday for apparently no reason.
tsvictoriabc
Ask and you shall receive:
TSVictoriaBC
They have lots of "That One" stuff.
dannylandulf said...
PSST. A new ras poll is dropping tonight that you all might find...whats the word...interesting:
Ras - GA - President
Obama 45%
Mac 43%
Is this poll for real!?
If so, then the floor is lower than McCain can possibly imagine!!!
Dannylandulf:
"PSST. A new ras poll is dropping tonight that you all might find...whats the word...interesting:
Ras - GA - President
Obama 45%
Mac 43%"
**************
Where is your confirmation of this claim. (I hope you aren't bluffing, but you need to back it up for any of us here to trust it).
As the McCain campaign goes increasingly negative/racist/xenophobic and the DOW continues to fall, doesn't the Bradley effect everyone comes here and asks about become more likely to occur? What I mean is, doesn't it at some point become socially unacceptable to be "for McCain" and then people might lie to pollsters?
I can't imagine at that point it would flip the race, but it might make the race somewhat closer.
@Robert:
PAID volunteers
That seems to be an oxymoron.
THAT ONE'08 TSHIRTS
Not sure who's behind the site but it's cute!
http://www.thatone08.com/
I guess my neighborhood must be comprised of that 26-29% of Oklahomans who polled as voting Obama because I've not met one person here who says they're voting McCain. Of course, I don't get out much and likeminded people tend to stick together.
I've been waiting to see some new OK polls for a while, now I am further disgusted to see this state polling even redder. Why do I always live in states that consistently go the other way (whichever way that may be)???
And how come I never get polled???
LOL
Would anyone feel like explaining to me how the electoral vote projection can change so much while the popular vote projection seems to stay pretty constant? I'm sure this has been explained before on this site, but I can't find it.
I understand why the "Win Percentage" changes closer to the election... But the popular vote has seemed to be about 51-47 for a long time.
Authorities raided ACORN's Vegas office yesterday and found evidence of voter registration fraud like registrations in the names of the Dallas Cowboys starting lineup.
The GOP has diebold and the Dems have ACORN.
Interesting that Obama's margin in Gallup is now as big as his margin in Rasmussen, Hotline, and Battleground combined.
My bad, I ran across those number on a right wing site.
Another subscriber debunked them a couple minutes later.
SORRY.
Ras - GA - President
Obama 45%
Mac 43%
A poll-troll with a result favorable to Obama?
That's a first!
that GA poll is a fake
538 Rule:
NO POSTING UNCONFIRMED POLLS
As Dannylandulf confessed... that GA poll is NOT real
EDWARD
here in FL my elderly prejudiced mom was able to rationalize voting for Obama because he is "only half black" and his VP is the fine irsh catholic Biden who she finds very reassuring.
mom has already cast her absentee ballot for Obama - I know because I helped her fill in the form.
she also voted against every GOPeron the ballot here in Volusia County, and against the Defense of Marriage state ammendment that the GOPers are trying to use to help pad their #'s.
I expect many lingering racists will eventually rationalize away their fear to vote for Obama
@patrick:
I, for one, won't be satisfied until I can drive from Florida to Washington in all blue states.
To get there, all we have to do is flip Georgia, Nebraska (all three CD's), and either Utah or Arizona.
Piece of cake. ;)
@patrick
"I, for one, won't be satisfied until I can drive from Florida to Washington in all blue states."
In order of decreasing probability, these are the states you'd need to flip (counting NC, IN and MO as blue.)
Georgia -15%
Arkansas -12%
Texas - 10%
Arizona - 6%
It's something like one in ten-thousand, even counting every blue state as a lock for Obama, but it'd be such a glorious thing to see.
The Real Joe Effect can seen in the direct geometric relationship between the % voter cushion for the Democratic candidate in a tracking poll in a given state on a given day...and the number of spontaneous testicular explosions among internet-capable Republican males who support John McCain.
Someone can calculate the margin for error on all of this and get a PhD.
The tipping point metric has become kind of academic at this point, since close elections are now freakish occurrences under the model.
"As far as the 8% so-called Bradley effect, Obama is half white so in his case it's only 4%. "
LOL
Real Joe said...
that GA poll is a fake
So says one of the most desperately delusional right-wingers to ever grace the internet...
Hey TSVictoriaBC, a shout out from across the water, here. Feel to join us at our Election Returns party here: http://portangelesdowntownhotel.com/
One block from the Coho ferry, rooms start at $45/night, eh? You can call/email from that web link to make a res or want to get some more info.
It's going to be mostly local school teachers and assorted egghead libs. White wine and arugula for all!
The Gore +8 in 2000 was prior to the debates, before the "sigh" and the other petty bullshit stuff the MSM and public harped on back then brought him down.
Obama's numbers have two debates already baked in the cake, and it's pretty inconceivable that last night's debate will help McCain in the polls, so I don't think Gore's blowing an 8-point lead in 2000 has any bearing on today's situation.
Nate,
I've noticed what appears to be a persistent error in the Scenario Analysis table. The lines for "McCain loses OH/MI, wins election" and "McCain loses OH/PA/MI, wins election" appear to be inconsistent. Today we have the following results:
McCain loses OH/MI, wins election - 53 of 6925
McCain loses OH/PA/MI, wins election - 29 of 6906
Clearly, "McCain loses OH/PA/MI" is a subset of "McCain loses OH/MI". These sets are almost the same with the current odds. For today's numbers, it appears that there are only 19 runs where McCain loses OH/MI but wins PA out of the entire 6925 runs where he loses OH/MI. However, McCain gains 24 "wins election" runs when moving from "loses OH/PA/MI" to just "loses OH/MI". These numbers appear to indicate that McCain loses OH/MI but wins PA in 19 runs and that he wins the election in 24 of these runs. How can the denominator be larger than the numerator here? How is McCain adding to his "wins election" total without adding at least as many matching scenarios? Am I missing something here? I've noticed this same issue occurring with these two lines for at least the last week (it stands out since these numbers are so close now).
Another bug - though it is correct today - there has also been a persistent bug where the 4 lines that use "Obama loses OH" as the denominator do not actually show the same number for the denominator. Specifically, the denominator for "Obama wins VA when losing OH" has consistently been one lower than the denominator for "Obama wins FL when losing OH", "Obama wins CO when losing OH", and "Obama loses OH, wins election" (although, as I mentioned, these all match today, so maybe you have fixed this issue).
liam said:
"Would anyone feel like explaining to me how the electoral vote projection can change so much while the popular vote projection seems to stay pretty constant? I'm sure this has been explained before on this site, but I can't find it.
I understand why the "Win Percentage" changes closer to the election... But the popular vote has seemed to be about 51-47 for a long time."
*************
This is because the electoral vote count is mostly sensitive to individual state polls. If even one large state (eg. FL) starts to look solidly in one or the other's bag, it can be counted on to be added into the EV total for far more simulations than not. Obama passed a real tipping point in this respect in the last few weeks, resulting in far far more simulations giving him FL, CO, WI, MI, NH, MN, etc. Many of these probably were swinging wildly between Obama and McCain before this recent movement in all the state polls.
"scratch those RAS poll numbers...I just debunked them."
The best thing to do, then, would be to delete your original comment. Just go back to it and click on the little trash can icon next to it. That way people who start reading the post won't be fooled. Thanks for correcting yourself. In general, it's best to provide a link to a poll or state where you saw it (Pollster, RCP, etc.).
So says one of the most desperately delusional right-wingers to ever grace the internet...
Hey. Real Joe may be a desperately delusional right-winger, but he's our desperately delusional right-winger.
Is there any record of how accurate 538's electoral predictions have been in the past?? Is there any record of how accurate the predictions are at different lengths of time before the election?
Not to beat a dead horse, but imagine if Ickes and Wolfson were running someone elses campaign with something other than a 50 State strategy.
rich merritt said...
As far as the 8% so-called Bradley effect, Obama is half white so in his case it's only 4%.
Ha Ha
you are on weed
steve,
I agree completely. Nobody has ever come back from this big of a deficit this late in the game. Bush v Gore was just the closest.
Last night Jim Carville said he hopes the pollster models are accurate and he warned of the serious consequences to the country if the final polls show Obama ahead by 5 points yet he loses by 1 point in the actual returns.
The current 10-point spread in the 6 daily trackers is mystifying. The pollsters obviously have different weighting formulas to alter their raw data that presumably is rather similar.
About 21% for a filibuster proof 60 seat majority in the senate. Even if it's unlikely, I love the possibility.
Democrats having control of Congress and the White House for the first time in my adult life will let me determine if I am really a Democrat. I'm optimistic.
It's not just the polls.
Obama advantages:
+5 to +8 in polling aggregates like 538
The electoral map favors him.
All the fundamentals: economic, wrong track numbers, incumbant fatigue
Partisan makeup and registration numbers
Money
Much better polling on the issues
McCain advantages:
None
But I suppose he can hope for some kind of massive external event or a huge Bradley effect.
Can someone shed some light on why we're not seeing a larger increase in potential Senate seats? Those projections have remained more or less static through Obama's upward swing.
DCM in FL:
That anecdote about your mom is awesome
the RR POTUS results are not real - but expect one later today or tomorrow.
RR did drop their new GA Senate poll today - Chambliss up by only +6 over the DEM challenger Martin [50-44]. +1 Martin - but Chambliss is still vunerable, especially if the extremely high AA turnout continues in GA & the GOP gets dispirited or splits their tickets.
Obama should benefit from the 3rd party splits in GA, since Barr & Nader should draw up to 5% from McCain mostly.
I expect Obama to also be within 4-6% of Mac but not in the lead in a RR poll of GA. Also, Scottie R pushes the voters to choose between the 2 parties & attempts to discount the 3rd party vote.
for once I agree with Real Joe, no fucking way is that GA poll real
NO WAY
Ellen said...
Is there any record of how accurate 538's electoral predictions have been in the past?? Is there any record of how accurate the predictions are at different lengths of time before the election?
To follow up on that, I would be interested to know what the plan is for evaluating the results of this election.
Found the correct Ras poll for GA:
McCain 54%
Obama 45%
read the FAQ. Nate explained this effect a while back. it is all about the demographics in the model. for example, if you can say that UT is x points behind CO in obama support and CO goes up n points then you can say that if they have similar demographics there will be a certain rise in both states based on what happened in the polled state... he explains that concept better, so go look.
I'm aware of all that. What I'm saying is that ND is no "cousin" of OK.
1. This is fivethirtyeight's first presidential election so there is no track record. Election projection and electoral-vote.com were both pretty accurate in 2004, each off by one or two states.
2. According to a pollster.com article on Gallup polls, in October 2000 Gore was ahead by 10 points early in the month and Bush was ahead 13 points in a late October poll just before the closest election of the modern era in the EC.
Imagine a 60 Seat Senate majority and Scalia having a massive heart attack. Ahhhhhhhh..........
Mathew
I only have intuition for evidence but I think a clear lead in Florida is more posible than Ohio.
I kind of think that Obama's Ohio ceiling is due to racism and Florida ceiling due to youth & inexperience.
Inexperience stops being a drawback when you show talent & your oponent incompitance.
However Obama is still Black.
You would have thought he would have sorted that out by now.
Oh well. If Obama can't paint himself white, he knows how to paint enough of America Blue. Lol.
The gist of Carville's comment wasn't that he hopes the pollsters don't mislead the electorate, it was that if the country blatantly manifests itself as covertly racist, there will be a tremendous amount of social unrest.
But, I have to wonder about states like OK - they are obviously single issue voters, as by voting for McCain they are voting clearly against their wallets.
Having grown up in Norman, I can answer this, and it's essentially what you'd expect. The vast majority of rural Oklahoma, and even an alarming degree of urban Oklahoma, is dedicated to infusing theology into government. They are the "Southern Strategy".
This year, there is the added bonus of unabashed racism. That aspect is considerably less represented in the urban centers, but heavily prevalent in the rural areas.
I expect to see Obama lose by substantially more than Kerry or Gore did.
Truly incredible polls. Kind of nervous about the coming weeks though..
blame, I think your analysis is likely based in reality btw FL and OH.
Obama just hit 75% on intrade.
MARK in VA
if OK is so deeply racist how did they manage to elect JC Watts ?
I mean just because he was a GOPer & ex-OU football star, they voted for a black man ???
rich merritt said...
As far as the 8% so-called Bradley effect, Obama is half white so in his case it's only 4%.
REAL JOE:
Ha Ha
you are on weed
I wish! I'm at work (same effect, tho now that I think about it).
back on topic of this blog, I'm big on the Supreme Court... Obama's got to win if for no other reason. Anyone read Clarence Thomas's book? After Obama wins, Nate should use this site to start an impeachment movement.
the DOW streak continues. :(
Dow was up 100+ points 30 minutes ago but has lost over 300 points since.
Dow -200 points at 4pm EST
If we click on the McCain-Palin ad at the top (tho it's usually the pitbull's puss that we see) does Nate get their money?
DCM-don't discount the OU football star aspect. Seriously. And wasn't he only a Congressman, or did he win statewide office?
"Obama just hit 75% on intrade."
Remember this is another 12 hour post-debate period where there's a contract on "Will Obama's numbers go up". I expect there to be a lot of volatility and inflated numbers until midnight from people trying to win that contract. It should probably stabilize to 70-72 tomorrow.
Here's what I'd like you to address: 346 EV, 90% likelihood of winning... these are impressive numbers, but when you break it down state by state (and I know you do, very carefully), you can say Obama has 264 EV in the bag, but isn't it being a bit facile to assume he'll hold his lead in the other states?
As several people have pointed out, this has all happened very fast, and can un-happen just as fast.
I know he's leading in almost all of those swing states now, but I don't think it would be remarkable if McCain takes Florida and Ohio, and even less remarkable if he takes NC, VA, MO, IN, and NV. Assuming it plays out like this, then it all hinges on CO. The numbers out of there have been good recently, but not as solid as I'd like.
What I'm asking is, while this seems to be resembling a landslide more and more each day, my concern is that it's actually going to be extremely close, and may come down, as the last two votes have, to some questionable behaviors or wonky voting machines in Boulder.
Mark in VA
did you adopt the HUSSEIN surname in support of Barack ?
------------------------------------
Barack 'Hussein' Obama is now up to 75 on INTRADE !!!
Even if Obama took a 20 point lead, Fox News would just report, "Breaking News: "Experts predict 30-point Bradley Effect". Followed shortly by "Is BHO Bill Ayers' Love-Child?" Conservative Radicalist Media at its best.
Here is the new Ga. rasmussen sorry but it`s MC 53 obama 45
One interesting poll coming out is he has Franken leading the MN. senate race 43-37. Obama is leading there 52-45, no surprise.
Seems like Coleman may go down. This seems to confirm the star poll about the senate race.
@Robert, (thanks for the suggestion)
You don't have to "apologize" for the Evangelical cultural group, whatever that means. Instead, you should "energize" the group. Here's the link:
http://my.barackobama.com/page/group/EvangelicalsforObama
This group comes with their "bumper sticker" (professionally made :) )
We all know that "real" evangelicals and Christians can't fall for the Palin type of extremism. If you are to follow THE Prince of Peace, you cannot accept the hate and war mongering bs. That's so obvious.
did you all hear McCain accidentally refer to his fellow "prisoners" today? I'm almost certain this was not a reference to the past, but senility and exhaustion providing a gaffe. I think he meant to say fellow Americans, but instead said prisoners. anyone catch it?
We need to realize how totally absurd 90% is in a hugely partisan environment. These will come back slightly, but it will likely be too late. As an Obama supporter, it gives me hope -- but it also makes me want to lock in as many votes as possible RIGHT NOW.
What´s this Michigan poll?
in PA:
Another McCain-Palin Introducer Declares "Barack Hussein Obama"
this is just really funny
Maybe he should have stuck to saying "my friends"
Does any part of this model account for the increasing numbers of "already voted"? With every passing day the proportion of the electorate that hasn't voted yet shrinks. Some estimates suggest that 1/3 of votes will be cast before election day. Obviously, there's way too much that would go into any approximation about already cast votes -- could be that the most persuaded partisans vote the earliest, or it could be that the laziest least-politically engaged independents just wanna get it out of the way. Hard to call, of course. But it might be worth starting to weight state election results according to past polling in a way that would be reflective of this slow picking off of the electorate. Again, i'm not sure what the best model for that is. If it's just a matter of writing code, maybe Nate could explore a couple of scenarios.
did you all hear McCain accidentally refer to his fellow "prisoners" today? I'm almost certain this was not a reference to the past, but senility and exhaustion providing a gaffe. I think he meant to say fellow Americans, but instead said prisoners. anyone catch it?
How long until he refers to Palin as 'Cindy'?
I believe the studies that show the Bradley/Wilder effect to be a thing of the past. However, my feeling is that this is not because we are less racist than we once were, but rather because there's actually less stigma associated with racism than there was a few short years ago.
I'm getting very nervous about this election. Not some of the polls are tightening and not because I worry about whether I can believe the polling, but because if the McCain/Palin race baiting tactics combined with typical Republican voter suppression efforts actually work to swing this election, coupled with the dismal state of the economy, we are going to have one unbelievable mess on our hands on November 5th.
I don't want to see McCain inaugurated, but what I really don't want to see is riot police clearing Pennsylvania Avenue for his parade.
Has anyone noticed the following:
Background… People talked about the Bradley Effect (B.E.) during the primaries. They claimed it was 3% back then, but then had to admit that it seemed nonexistent or lower (I saw claims by rational people around 1-2%, but these same people knew that what they felt they saw was likely noise. Most people agreed that the dreaded B.E. was low or nonexistent.
(1) Early Presidential: People (R trolls and D worrywarts) claimed the B.E. was 4%
(2) When Obama started rising in the summer, mainly the Republicans on sites started claiming the B.E. could be estimated to be 5%.
(3) As McCain started losing steam after the conventions, same posters talking about 6%.
(4) A week ago I saw more and more 6-7%’s
(5) Today, I am seeing a few people claiming the B.E. is 8%.
Estimates of the B.E. seem to be rising in a nonlinear fashion, related to time!
*******************************
Inspired by Nate, I have created a model that tracks this exponential growth related to time, and now I’m really worried:
A week from now:
B.E. = 10%
2 weeks:
B.E. = 28%
BY ELECTION DAY
Bradley Effect = 78%! - we’re screwed!
darío said...
What´s this Michigan poll?
WHAT ?
Most wanted states for new polling:
KY
TN
SC
None of these states have been polled in weeks and they border battlegrounds like IN, OH, and NC. I wonder where they are now?
JESALEH
but the BIG leads now for Obama do directly corelate to a huge advantage during the early voting & absentee ballots that are already being banked for Obama.
even if the polls later regress toward the mean [which should still favor Obama btw], these banked early votes give Obama a strong structural 'lead' going into Nov 4th.
every day in October & November with a lead should translate into more banked votes for Obama to hedge the bets later.
encourage early voting for all Obama voters
I am thrilled Obama appears to be absolutely indefeatable. McCain, even in his own words, in his own book, is so ill suited for a job in politics-let alone President-it's absolutely ridiculous. Some war hero-he appears to be at least partly responsible for over one hundred deaths of his own fellow airmen in one incident alone-Christ almighty!
A deranged moron and dangerous career criminal text messages me a death and stabbing threat against Bristol Palin and her unborn child-48 hours later he causes an alcohol related three vehicle head on collision near Santa Barbara-see the complete story at my blogs:
www.myspace.com/372390338
CUTTING CONFESSIONS-a love story with high drama and cutting edge comedy like no other-a midwestern man and an abused and artistic Beverly Hills girl are brought together by seemingly near impossible acts of fate-this true story is set in Beverly Hills, Pasadena and the Las Vegas Strip-see the story at these blogs:
www.myspace.com/372390338
www.cuttingconfessionsfilm.blogspot.com
intrade:
Obama @ 77.0
McCain @ 25.0
TSVictoriaBC said:
"And if anyone from bumper sticker industry reading this, please make a "THAT ONE for PRESIDENT" sticker. I'll take one."
http://shop.cafepress.com/that-one?cmp=knc--g--us--pol--elect08--a--default_ad_URL&gclid=CMuBv7C4mJYCFRZjnAodnDaf6w
While it's nice to see Obama projected to win so handily, we can't count those chickens just yet. Click on the link that takes you back to 2004. On this day back then, Kerry was winning big, too, in the projected EV tally. And he was also surging.
How is this year different? And what makes the polling this year so much more "sure-thing"? Or is it? I'd be interested to hear what Nate has to say.
I think Dario was asking why a 2 month old MI poll was added today. Did it just get overlooked until now?
that RR poll of GA is suspect - where are the 3rd party votes ?
at least 5% will vote for Barr, nader & McKinney in GA - and almost all of those will come out of the McCain vote.
RR pushes voters to choose only between the 2 parties - other pollsters who track all 5 parties show a much smaller gap between Mac & Obama...
Here's the link to the MY FELLOW PRISONERS comment by McCain today.
My Fellow Prisoners
At this time Bush lean Kerry for 2-3 points, see RCP 2004.
Yes, Paul, that poll.
Quick thought, Be able to drive from Florida to Washington in Blue States nly is not necesarily completely out of the realms of possibility, if it did come down to a meltdown election. Here is the route - Florida, Georgia, The Carolinas, Virginia, through Maryland and Pennsylvania, Ohio, Indiana, Illinois, Iowa, Minnesota, North Dakota, Montana, Idaho (probably the biggest stretch, but it would only be a tiny bit of red if it didnt turn) ansd onto Washington.
I am still trying to think of a way for McCain to win this thing. Even the Bradley effect wouldn't help him now. (But even if it could there is no historic evidence of an 8% swing caused by the Bradley effect anywhere. I suppose a Bradley effect, tagged onto a 'shy tory' effect might add up to a pretty big number, but 8%??)
I think whats tough for McCain is that Democrats have been leading throughout the year, this isn't JUST a question of the campaigns having an effect (Clinton and Obama both had leadsof around 4% in May/June). This is public dissatisfaction with the GOP and McCain just hasn't been able to be an anti-GOP candidate enough.
I thnk the other thing is that in any historic evidence of it being possible to comeback from so far back, the electorate has been pretty volatile. This year it has been remarkably settled. (Consistent 4% leads at least for Democrats save for GOP bounce moments) Its actually pretty hard to imagine that changing.
Paul, you said:
McCain loses OH/MI, wins election - 53 of 6925
McCain loses OH/PA/MI, wins election - 29 of 6906"
You're right, I can't make any sense of that either, unless these come from separate sets of simulations (which I thought it did not, that the RNG was run once per day).
Nate?
"How is this year different?"
Seriously? You don't know how this year is different than 2004? Well, to start with, the incumbent isn't running, and he's also got some of the lowest approval ratings in history. And perhaps you've noticed that the economy is not doing so well? These are both huge advantages for Obama that Kerry did not have. Finally, the Republican party is in shambles and Democrats have been picking up huge gains in registered voters lately.
MrInsight (sic):
"According to a pollster.com article on Gallup polls, in October 2000 Gore was ahead by 10 points early in the month and Bush was ahead 13 points in a late October poll just before the closest election of the modern era in the EC."
Of course, you cleverly forget to mention any state polling results.
If that's the best example you've got, you might want to change your handle to MrPullStuffOutMyArse.
where the fu** are new state polls ?
LA Mayor Ed Bradley actually won more votes on election day in the CA 1980 race for Governor. But GOP winner Deukmejian had a much better organized *absentee* ballot turnout.
Exit polling and pre-election polling were within the MOE for ballots cast in person, but did not adequately sample mail-in early voters.
So says columnist Dan Walters in the Sacramento Bee Aug 8, 2008. Whatever "Bradley Effect" was present in that election was not about race or ethnicity.
Robert, you might be interested in this:
Petition to Gov. Sarah Palin: Stop Bearing False Witness
http://www.crossleft.org/node/6557
Robert said...
While it's nice to see Obama projected to win so handily, we can't count those chickens just yet. Click on the link that takes you back to 2004. On this day back then, Kerry was winning big, too, in the projected EV tally. And he was also surging.
That just isn't true... as we have seen posted earlier
Jonathan CR -- I think it's dangerous to take historical facts as a guide to any particular election. No Democrat has ever won the White House while losing both Texas and Kentucky, but it looks like both of those states will go for McCain.
I realize the analogy is imperfect, but the point is each election is different.
It doesn't do for Democrats to get comfortable before the election. That's for Republicans.
We need more VA polls.
Actually, electoral-vote.com was predicting a (narrow) Kerry victory on this day in 2004, even though national trackers had Bush ahead.
where are the WV polls?
Kennedy won without Ohio.
No republican has ever won the WH losing OH.
http://www.pollingreport.com/wh04gen.htm
Anyone else notice how Realclearpolitics.com has been dragging their feet on projecting Obama to get more than 270 votes? It's getting a little ridiculous. He seems to be re-revising his weighting of the VA polling numbers in order to avoid projecting an Obama win there that puts him over the top.
Pathetic.
looks like there's a 2nd robert now.
"While it's nice to see Obama projected to win so handily, we can't count" is the new one (to this discussion). I'll demarcate mine with a TX at the beginning.
*************
eve:
Thanks for the link, eve.
RCP 2004 poll averages 1-10 october.
Zogby (1223 LV) 10/9 - 10/11 45% 45% 2% TIE
CBS News (760 LV) 10/9 - 10/11 48% 45% 2% Bush +3
ICR (763 LV) 10/9 - 10/11 48% 43% 2% Bush +5
CNN/USAT/Gallup (793 LV) 10/9 - 10/10 48% 49% 1% Kerry +1
Time (886 LV w/leaners) 10/6 - 10/7 46% 45% 4% Bush +1
GW/Battleground (1250 LV) 10/3 - 10/7 49% 46% 0% Bush +3
AP/Ipsos (944 LV) 10/4 - 10/6 46% 50% 2% Kerry +4
Marist (642 LV) 10/4 - 10/5 49% 46% 1% Bush +3
Fox News (1000 LV) 10/3 - 10/4 47% 45% 1% Bush +2
ICR (762 LV) 10/1 - 10/5 51% 45% 2% Bush +6
ARG (800 LV) 10/2 - 10/4 46% 46% 2% TIE
ABC News/WP (1169 LV) 10/1 - 10/3 51% 46% 1% Bush +5
CBS News/NYT (561 LV) 10/1 - 10/3 47% 47% 1% TIE
Zogby (1036 LV) 10/1 - 10/3 46% 43% 2% Bush +3
Pew Research (801 LV) 10/1 - 10/3 49% 44% 2% Bush +5
CNN/USAT/Gallup (772 LV) 10/1 - 10/3 49% 49% 1% TIE
Nate,
Would you please examine the party weighting in these polls (national and state). There's a lot of talk now on right wing sites that most polls are oversampling democrats grossly.
As a previous poster had said in another thread, the right wing sites are absolutely convinced that McCain will win, and he will win comfortably. They back up their work with history and numbers and its difficult to argue with them.
I would really really appreciate your take Nate as I'm very concerned.
LIBERAL JIM
nice post on the debunking of the 'bradley effect'
all pollsters are subject to uncontrolled variables such as lack of control on early voting as well as absentee ballotting plus cellphone undercount, etc.
this year those uncontrolled metrics will be exposed even more - exit polls will be way off IMHO since the absentee & early votes are going up exponentially.
I wish you would do a comparison of FLA polls from 2004 and today and see if that affects your 75% chance of Obama winning the Sunshine State. Because it seems like the big 90% number depends too much on that. And my memory of '04 is that the Florida polls failed, big time. Because, as we know, election-wise, that state is really, really, really f@#ked up.
Here's the link to the "fellow prisoners" footage from HuffPo.
http://www.huffingtonpost.com/2008/10/08/mccain-calls-americans-hi_n_133037.html
I just ordered my pair of Nate Silver pj's! Better than Christmas morning!
Nate - any chance I could set you up with some single friends?!?!? ;)
if you weight in the Real Joe Effect™ i think the Great Blue Road Trip from Miami to Seattle starts approaching the realms of the possible
Sam - yes, of course I know one shouldn't assume that historical precedent is everything. I certainly don't put much stock in all those "No candidate from state
A in age-bracket B running for party C has ever won an election without winning more than D% of the vote from group E in state F" maxims. I just wanted a bit of context - I don't know how volatile the polls can be after this point. If no-one had ever recovered from a position like McCain's before then that wouldn't mean that he couldn't do it, but it would still be worth knowing, if only to consider what special circumstances McCain might have in his favour that candidates in previous races didn't.
Bill (Dad to Deeboo & Jay):
RCP's bias was naked long before Nate called them out on it.
"As a previous poster had said in another thread, the right wing sites are absolutely convinced that McCain will win, and he will win comfortably"
These guys are just grasping at straws. Even Karl Rove does not believe that (www.rove.com/election). The fact is that the number of people identifiying themselves as Democrats has increased substantially in the last few years.
jewsaleh:
you can say Obama has 264 EV in the bag, but isn't it being a bit facile to assume he'll hold his lead in the other states?
By your own scenario, he doesn't have to hold his lead in the other states, he only has to hold his lead in one of them. The idea that it wouldn't be "remarkable" if McCain were to win all of the swing states where he is currently behind is frankly ludicrous.
His lead expanded quickly because we drew closer to the election and people started making up their minds, and they chose Obama. There is no comparable effect that would make this "un-happen." There could be events that would make it happen, but these are people, not just numbers on a page, and there's no basis for arguing that a huge shift back the other way would "just happen."
RWD - deep down in their heart they know they're wrong.
"My fellow prisoners . . ."
I guess that's McCain's new way of letting people know he was a POW. Did you know that? I only learned it just now.
Actually, McCain was trying to say
"My fellow pensioners . . ."
but his tongue jut caused "prisoners" to come out instead.
A real Freudian slip.
"There's a lot of talk now on right wing sites that most polls are oversampling democrats grossly."
Well, of course, what else are they going to go on? The only "evidence" is that party identification was even in 2000 and 2004. It's not now. It's not even close. Rasmussen posts numbers every month taken from 30k interviews; it's at minimum a 5 point advantage and possibly 8-9.
The people you're referring to are basically deluding themselves into thinking that essentially every polling organization except Zogby is horribly wrong. Think for yourself whether that's more likely or whether they're just losing and in denial.
Bill (Dad to Deeboo & Jay) said...
Anyone else notice how Realclearpolitics.com has been dragging their feet on projecting Obama to get more than 270 votes? It's getting a little ridiculous.
Just as ridiculous as CNN and MSNBC (I can't bring myself to look at FOX)refusing to move any state out of their "toss-up" category. They need to keep Obama under 270 so they can pretend this race is close. Just for ratings I assume, or afraid to be called "liberal media" for reporting the facts? Blogs are all that is left reporting actual news and this is why I love this site!
from eve:
Petition to Gov. Sarah Palin: Stop Bearing False Witness
http://www.crossleft.org/node/6557
great!
Emperorwillis, Rasmussen used 39% dems, 33% reps.
That´s normal, the number of democrats has increased in the last few years.
Another think about the Diageo/Hotline tracker. That isn´t a bad tracker but if you compare the numbers you can see thay Hotline takes only 950 LV. Rasmussen takes 3000 LV, 1000 for day; and Gallup takes 2700 RV approx.
The number of people survey is much more less in Hotline than Rass and Gallup.
this has been asked, but not answered (that i am aware of).
are the polls taking into account -in their wording and in their analysis- of the ALREADY CASTED votes? do they weigh more?
May I be the first to say: MCCAIN LANDSLIDE!!1!eleven!
So, what you are saying is McCain still has a chance.....
I would really really appreciate your take Nate as I'm very concerned.
Hi, Mr. Concern Troll!
I don't remember if you're posting under the same name as you did a few weeks ago, but "They back up their work with history and numbers and its difficult to argue with them." is no more convincing now than it was then. If you want to be entertaining, at least take the trouble to come up with something new.
Robert,
I'm not sure what link you are talking about, but if you go to ElectoralVote.com and look that their map for Oct. 8 in 2004 and for Oct. 8 in 2008 you will see some striking differences. First, the 2004 map has many more safe GOP states, notably almost the entire West. Second, Kerry's lead depended on a whole bunch of barely Dem states like OH, PA, NV, ME, IA, and WI. Obama's lead is not only much bigger, but if you exclude the barely Dem states you still get an Obama win. Third, those close states in 2004 were mostly very, very close (1-point or so) and resulted from a single poll result, often from a crappy pollster (Zogby, ARG). Obama's close states have 2-3-point leads and draw from multiple polls, in some cases 4 or 5 of them. They are far less susceptible to outliers and are thus more reliable leads. As with all reports (or in this case, maps) it's important to look beyond just the top line results.
Obama is way, way ahead of where Kerry was. True, he could lose the election. But it's highly unlikely, barring some MAJOR event. Hence, Obama's very high win percentage. It's just really tough to come back from being so far down, especially in an electoral atmosphere that is toxic to the GOP and with a map that heavily favors Obama.
No way that nigger gets elected. No fucking way. You fuckbags will all be wrong. You will see.
A few comments about stuff I have read in this discussion.
At this point, unless Obama shows movement in a state like ND his win percentage is going to fall. A 5% lead 2 weeks out is more predictive than a 5% lead 4 weeks out. If Obama had these numbers the day before the election his win percentage would be over 99%.
The Bradley Effect numbers are directly correlated w/ Obama's lead, because conservatives like to delude themselves into thinking their guy has some chance. They would also like to believe that other people are as racist as they are, but just not as honest about it. Freud called it projection.
Real Joe isn't delusional, a conservative hack who cares more about himself than his country yes, but not delusional. He knows he is on the ass end of what is shaping up to be a historic ass-whooping.
RCP's bias was naked long before Nate called them out on it.
Sure, sure. But the polling numbers are reaching the point where all the supposedly neutral (but actually conservative-tilting) polling evaluators are going to have to start projecting Obama to win, or else discredit themselves.
After Bush took the last two elections from Democrats, isn't it kinda ironic that HE was vital in giving the Dems such a big advantage in Party ID? He really IS the best Democrat Recruiter.
The Gore +8 in 2000 was prior to the debates, before the "sigh" and the other petty bullshit stuff the MSM and public harped on back then brought him down.
Obama's numbers have two debates already baked in the cake, and it's pretty inconceivable that last night's debate will help McCain in the polls, so I don't think Gore's blowing an 8-point lead in 2000 has any bearing on today's situation.
This may not be a fair comparison, since it was a governor's race. But Bobby Jindal led Kathleen Blanco by 10 points with less than a week to go before Louisiana's 2003 gubernatorial election. Blanco ran a series of negative campaign ads in the final week and ended up winning in the end.
Point being: no matter how good the polls look now, it's not over until the votes are actually counted.
McCain's only hope for a landslide will be to eat late night White Castle. They can always use the results for a Palin speech. Very eco-practical.
this has been asked, but not answered (that i am aware of).
are the polls taking into account -in their wording and in their analysis- of the ALREADY CASTED votes? do they weigh more?
I believe Nate mentioned a while back that some pollsters do ask whether you've already voted. It seems like that would be a money-saving measure for them, since they could just put those people in the "definite" categories without taking as much time to tease out how solid their support is.
Personally, I want a bumper sticker saying "This One votes for That One, he does", and a picture of Yoda on one side and Barack on the other. I may have to make my own.
As far as the polls go:
Start here:
http://cgi.cnn.com/2000/ALLPOLITICS/stories/11/01/tracking.poll/index.html
Click on the Resources: View all the results from the CNN/USA Today/Gallup Tracking Poll. You get a really nice chart. Please note that Bush is blue, and Gore is red.
See how the Bush Blue line is equal to or higher than the Red Gore line is from September 22nd on, except one blip where Gore gets a big lead for about 3 days at the beginning of October? That's the outlier that Republicans get so much hope from. Does it look anything at all like the polls today look? No.
So has somebody who has had a one or two day significant lead lost an election later? Sure. Has anybody who has consistently for weeks had a 5+ point lead, including a full week of October, eventually lost? No.
Outliers happen.
This lead for Obama is not an outlier.
@ myfeatfelloff @ 2:29 PM
@ jeffcassens @ 2:29 PM
You both brought up the "Bradley Effect" within the same minute and both use AOL IDs. Coincidence?
But honestly - have you headed over to Pollster.com and used their new interactive charts? Click on "Tools" on the chart and then "filter" and remove "live interviewer phone" (for worries about a person voting for McCain but saying Obama to avoid sounding racist) and "internet" (since they pretty much suck - IMHO) ... and leave only "IVR-Automated Phone" (or "Robocalls" - no live person to hear you say McCain and have you think they perceive you as a "racist").
Live and IVR:
Obama 49.4%
McCain 43.5%
Obama +5.9%
Live Only:
Obama 49.0%
McCain 43.3%
Obama +5.7%
IVR Only:
Obama 49.8%
McCain 45.8%
Obama +4.0%
So, you do see a narrowing of 1.7% between Live and IVR - which might give you the impression of a 1.7% Bradley Effect.
However - looking at the numbers ... Obama has a higher percentage under IVR (49.8%) than Live (49.0%) for a +0.8.
So it's all about methods of each pollster and how the questions are asked and are within a reasonable proximity to each other.
In the end - there is a "Bradley Effect" on the "Bradley Effect" - it's supported more than it the results show.
What happen if Obama is ahead in the popular vote the day before the election by 5 or 6 points and ahead in the EC more than 320 EV and he lose the election?.
I think only racism explain that.
There's about a dozen studies saying that racism is a bigger problem among Democrats than Republicans. Do your research.
TDS - The Stupid Vote
http://onegoodmove.org/1gm/1gmarchive/2008/10/the_stupid_vote.html
RWD,
Thanks for the response. I'm really looking for a scientific or mathematical response to debunk them.
Here's the gist. They say that historically, party gap is no more than 5%. So if history holds, we are looking at Dem+5 on election day in the best case scenario.
Now examine the polls. Almost all of them have sampled too many dems. RAS is the closest at +6, but others are +7 or even +9. State polls are the same way. so what they are doing is "correcting" each poll to a reasonable turnout model which would be Dem+5 nationally and similar historical numbers for each state. They get MAC+2 nationally and MAC winning the EV as well.
I hope Nate reads this, I would like his analysis.
Talking about RCP, has any one else seen this?
They came up with a BS reason for taking PPP out of their averages.
more confirmation that Obama won the debate & won BIG
from POLLSTER:
"Today, we asked the Online100 who they thought came out on top in last night’s debate.
A majority of the Online100, 70%, believe Obama won. 32% of the panel thinks Obama had a decisive victory, while 38% think it was a slight win. One panelist said, “The debate was a surprisingly strong showing by Obama, in a setting that was supposed to favor McCain.”
'In your opinion, who won the second presidential debate? '
[graph]
94% of left leaning panelists think Obama won the debate. The right also chose Obama as the winner, but by a much smaller margin; 46% think it was a win for Obama, 41% think it was a win for McCain. A majority of unaffiliated panel members, 71%, chose Obama as the winner."
====================================
@ http://www.politicshome.com/USA/Landing.aspx?Blog=3660&perma=link#
not even close...
When life is like the Onion:
"Voters appear disinclined to put a Republican in the White House."
--from MSNBC
Post a Comment