A new Rasmussen poll in the state of Nevada has John McCain leading Barack Obama by 3 points. This is a slight improvement for Obama from Rasmussen's previous poll, which had shown McCain ahead by 6. Nevertheless, Nevada remains one of those states where our regression model thinks that Obama's numbers have significant room to improve. Obama has outfundraised McCain by better than 4:1 in the Silver State, and Nevada has a highly secular population, a group that has performed well for Obama in other states.
So what gives? Nevada has historically been an apathetic state politically. It's turnout rate in the 2004 election was among the lowest in the country by any and all measures. And not that the following metric is the end-all, be-all, but when I rank the states from 1 to 50 in terms of the amount of per-capita traffic they contribute to FiveThirtyEight.com, Nevada ranks just 36th (the top three states, FWIW, are Massachusetts, Washington and Oregon; the bottom three are Mississippi, West Virginia, and Oklahoma). If Obama wants to win Nevada, he is probably going to have to rattle the cage a little harder than he might in another state.
In New Hampshire, which is anything but apathetic politically, Rasmussen has Barack Obama increasing his lead to 11 points; last month he led John McCain there by 5. Remember my rule of thumb about New Hampshire: its numbers tends to move about twice as much as the national average. So if Obama is leading by 11 points in New Hampshire, that would imply a 5.5 point lead nationally, which is just about where we have him.
Also two new polls out from SurveyUSA. In California, Barack Obama leads John McCain by 12 points, up from 7 points in their tracking last month. In Iowa, however, he leads by 4 -- down from 9 last month. I would be a little bit cautious about reading too much into either of the Iowa polls released within the past week as Iowans presently have bigger things to worry about.
Finally, while we don't usually focus on national polls, that's where a lot of the action has been today, with no fewer than seven of them released within the past 24 hours.
Newsweek's poll is the attention-getter, showing Barack Obama leading John McCain by 15 points. Is Barack Obama actually ahead by 15 points? Of course not. Newsweek's data tends to be fairly volatile, and we have a whole bunch of polling on both the state and national level that implies that Obama's real margin is closer to 5 points. Nevertheless, he has broken through a barrier of sorts. The last instance I can identify when a Democrat held a 15-point lead over a Republican nominee in any individual November trial heat poll is from November, 1996, when CBS News gave Bill Clinton an 18-point lead over Bob Dole on the eve of the election.
Other national polling, CNBC-style: Gallup Tracker: Obama +2; Rasmussen: Obama +4; USA Today/Gallup: Obama +6; FOX: Obama +4; Ipsos: Obama +7; and Harris Interactive, Obama +11.
6.20.2008
Today's Polls, 6/20
@ 7:15 PM
Subscribe to:
Post Comments (Atom)
102 comments
Looks like New Hampshire is going to be out of reach for McCain.
Only 4 electoral votes. But 4 electoral votes is 4 electoral votes.
""but when I rank the states from 1 to 50 in terms of the amount of per-capita traffic they contribute to FiveThirtyEight.com""
Now that is an observation that begs for the posting of the full 50 state ranking list!
Meh, now McCain will be able to point to the Newsweek poll in a couple months and claim to be staging a huge comeback. I'm surprised they published their ridiculous results.
Nate-
It's kind of sad, but I suspect that I account for at least 90% of your traffic from WA state.
We'll know soon enough- I'm moving to CA in a week. See if the traffic from WA drops off after next Thursday.
For the many readers who don't like the detailed stats that underlie the calculated ~76% chance of an Obama victory based on current data, here's a quickie, non-statistical way to arrive at about the same conclusion: go down the individual states on the left hand side of the blog, and look at the percentage chance of an Obama victory for each state. now sort them from 100% (CD) down, and see how low you have to go before the running total reaches 269 electoral votes. The current answer is 72% (for CO). Put another way, Nate's results currently say that Obma will win if he can just hold the states in which Nate thinks he has a 72% or better chance. The rest of the looking-good-for-Obama states (MI at 68%, MI at 63%, FL, VA, etc.) would just be gravy.
David in CA
Nate, a suggestion for presentation of the EV distribution. Given that you have classified states into 7 categories (SafeDem, LikelyDem, LeanDem, Toss-Up, etc.), and assuming that the toss-up's go in one direction or the other but are within the MOE of your estimates, then could you consider the following?
In addition to your "best estimate" of the EV outcome allocating all state projections that you currently have, could you consider having another EV outcome pie chart in which you do something more like what Rasmussen does? Namely, a pie chart in which the toss-ups are left blank or just put into a "too close to call" category?
A third possibility would be to cut the states into just 5 categories: the Safes, the Likelies, and the "others" (Leans and Tossups).
Then you could do an analysis, or rather write an article say once a week where you discuss which states are in the Tossup or the Lean-Tossup category that week, something that would be different from your other ways of classifying which states are critical.
These alternative distributions of the EV would perhaps be more representative of the current situation as well as the trends.
Your column's title based on this? "Slicing the Pie." And show three versions of the pie charts and tell the story of the race as of that date.
Thanks.
re Aaron.. I'l try to visit twice as much to make up for it!
I can't believe Obama is supporting immunity for the telecoms from their illegal assitance of the NSA. KGB style internal surveillance and the one person that projected himself as wanting to the best for America and its freedoms is helping turn it in to a Stalinist nightmare. I honestly believed Obama would make the right call on these issues, what's next Iran? This is a real stab in the back to Americans. I don't care if the bill was a "compromise." You do not compromise with inalienable rights, G-d damm it! This is very dissapointing.
I didn't see any mention of the InsiderAdvantage GA poll which showed Obama just one point behind McCain (Barr polled at 6% I believe).
Any thoughts on that?
PS. (((Oklahoma)))---my home state
Nate - Could you explain whether, and how, your model would deal with a 15 point lead in the following two (Newsweek poll)scenarios?
Obama 51
McCain 36
Undecided 13
versus
Obama 57.5
McCain 42.5
Undecided 0
thanks
@6:35. We covered the Georgia poll yesterday.
@6:36. To see what an Obama 15-point map might look like, just add 10 points to our current projections.
That would have him winning all states but Alabama, Arkansas, Idaho, Kentucky, Oklahoma, Utah and Wyoming.
Nate - Sorry - I wasn't clear in my question. I'm asking whether your model is looking at Obama as 51% or as Obama +15. In other words, are you accounting for high undecided percentage? And how, in general, do you deal with the undecided element of poll results?
@6:36. Oops, I totally skimmed through your question and missed the gist of it. Right now, the model allocates undecided voters 50:50, so it wouldn't treat the two results differently.
What combination(s) of state(s) does that really prominent mode in the Electoral Vote Distribution represent?
Damn, I'm in Nevada and I can tell you I visit this site about once an hour. Guess I'm the only one.
AIX NATIONAL POLL
Obama 44
McCain 56
MCCAIN 08
YES WE WILL
THIS SITE IS SO BIASED TOWARD OBAMA! WHAT THE CKFUCKFCKUFKCUFKCUFKUCF
You will winin '08 only if you fix the vote counting the way AIX fixes its polls.
Ummmmm, Nate...
"when I rank the states from 1 to 50 in terms of the amount of per-capita traffic they contribute to FiveThirtyEight.com, Nevada ranks just 36th"
Nevada has the 35th highest population in the United States.
Are you sure you're up to the challenge of handling this whole "statistics" thing?
Only kidding. Well, sort of only kidding. But I do love the site.
Hey penis, don't be a dick. Either give a better way to read the polls of stick it.
What combination(s) of state(s) does that really prominent mode in the Electoral Vote Distribution represent?
After twiddling around with mydd's EV calculator, I concluded that the prominent spike is Georgia and North Carolina, plus the usual suspects (Kerry + IA, NM, CO NV, VA, OH, FL, MO, IN)
Ohhhhhh, per capita. Well aren't I embarrassed.
Hey penis, don't be a dick.
first ballot inductee for the snark hall of fame
SWZ NATIONAL POLL:
Obama 83
McCain 17
MCCAIN 08
YES WITH BUTTS.
""Are you sure you're up to the challenge of handling this whole "statistics" thing?""
Are you up to speed on the "per capita" thing?
"Are you up to speed on the "per capita" thing?"
Are you up to speed on the fact that I already fessed up to missing the "per capita" thing?
Do you have per-county turnout statistics? If you do, I bet you that Clark County is even lower in the country rankings than Nevada is in its state rankings.
the bottom three are Mississippi, West Virginia, and Oklahoma
FWIW... as someone who runs a W.Va. website and studies our own traffic data closely... the majority of West Virginia's population lives along the edges of the state and many ISPs service W.Va. locations from out of state.
Also, as broadband is unavailable is so many places, there are people like me with a satellite Internet connection which renders the geographic location completely meaningless (it reads as where the signal returns from the satellite to ISP, not where my home satellite uplink is located).
So, while I'm visting your site from here in W.Va. (at least!) daily and have linked here several times from West Virginia Blue, a bunch of us reading from W.Va. may be contributing to Maryland, Virginia, Penn., Ohio, and Kentucky state traffic stats more than W.Va. stats.
All that being said... there are some major broadband deployment digital divide issues that a state ordered list of traffic might help highlight.
Where do we find the data for 538 hits per state?
good question obsessed
@ Obsessed - I believe that only Nate can access it.
As an Aussie I would also be interested to see how much foreign traffic you guys get.
countarach - we watched your recent election with great interest.
Australia has a drastically different system with all the different parties. Also odd that "liberal" and "conservative" seem to have the opposite meanings from the US.
NEW OBAMA SONG !
Link
Yes it was a great and historic election for us. the Liberal party means liberal in the sense of neo-liberal. There are still so-called "small-l liberals", which is the meaning you guys have. WE prefer to use Progressive though.
And yes the myriad of other minor parties makes things very interesting.
Michele Obama video released
its all over for Obama !
MCCAIN 08
The EV projection moved, but I didn't notice any states change color. Did I miss something?
It's important to remeber Nevada, especially Las Vegas Metro area with 2/3 of the state's population, is very transient. Few people vote or register to vote here because few people are from here. It was great to be a part of the early primaries, though, and we got record turnouts.
The CBS poll you mentioned on the eve of the election in 1996 was off by 10 points! it had Clinton ahead by 18 and he won by 8 points.
The fact that the Gallup tracking poll has Obama up by 2 points today and the Newsweek poll simultaneously says Obama is now ahead by 15 shows just how bogus polling can be.
One or both of them is so wrong it is pollster malpractice.
By the way, how can Obama still be the projected winner in NV with the last four polls all showing him losing there?
Nate seems to be fielding several other questions, so perhaps someone else with polling expertise can answer this query:
What is the methodological rationale for weighting polling samples by Party ID? The rationale for the demographic weights makes sense, i.e. that women/old people/rural folks are more likely to be at home when the phone rings. But what is the comparable rationale for Party ID adjustments? Is there something inherent in being a Democrat (or Republican) that makes one more likely to respond to a poll?
Another way of asking the same question is: if poll after poll finds that self-identified Democrats are over-represented vis-a-vis previous exit polls or registration figures, at what point does a pollster say to himself, "well, maybe it's just because more people identify as Democrats", rather than chalking it up to some mysterious sampling bias?
Kermit:
Getting a little hyperbolic, aren't we. I don't much care for the agreement, and think it is too much of a cave in to Bush, but to say that this will turn the US into a "Stalinist nightmare reflects a lack of perspective and/or knowledge (or possibl paranoia).
In a "Stalinist" nightmare, the government would intetntially "terminate with extreme prejudice" millions of its citizens (that doesn't inlcude death of troops and people in wartime, or the number hits cose to 25,000,000.)
And there is one good part to this bill, amid the crap. It ends the Bush claim of inherent powers to wiretap. (Not that the SOB wouldn't try it again but now at least it would blatantly be an impeachable offense.)
As a side point, the topic way off topic for this site.
great site
i have a question
how can OBAMA get 343.8 EV ?
0.8 ??
Nevada is a low-info state if there
was ever one. In most Las Vegas neighborhoods, there are tract homes, cookie-cutter apartments and lots of 7-11 stores. That's
not a high-info Obama territory.
Obama should have an edge there, but the state is very libertarian, plus there's a Mormon factor if McCain is the nominee.
It looks like McCain got a Juneteenth bounce in the Gallup tracking poll (any reason why no polling on the 18th--Russert funeral?), but maybe the drilling issue has narrowed the gap. Plus the tracking polls tend to weed out any "Bradley effects."
I think Obama's chances in Nevada are only 40%. And, yes, the Newsweek poll is an outlier.
@ Anonymous 7:26PM:
That is an average result.
Does anyone have an idea how reliable Harris Interactive is? I went to their website, but I couldn't really get a sense of their record.
To anonymous above:
The electoral vote count is the electoral vote mean of 10000 simulations. The way the model works, is that it takes into account polling and demographic data for each state and then assigns a probability that a candidate wins that state. The model then "rolls the dice" for all 50 states and counts the # of electoral votes and determines the winner. This process is repeated 9,999 more times, again flipping a handicapped coin for each state.
Hope this makes things less, and not more confusing.
-Aaron
Interesting note:
Aside from Utah, every single state currently in McCain's column trended towards Obama today.
mikeel,
It was due to the Iowa flooding.
Hmm, let's see. West Virginia has two polls out, one with McCain leading by 8%, and another with McCain leading by 18%. Let's call that a "Toss-Up!" Yeah! We're fivethirtyeight.com! We're retarded!
Holy shit, this site is so biased toward Obama.
1. Today's SurveyUSA poll of California (Obama up 12) has the largest gender gap I have ever heard of. In CA, McCain leads by 9among men and Obama leads by 33 among women. A 42-point gender gap!
2. Interactive online polls by Harris and Zogby are unscientific garbage.
Ah I just saw that Harris is web-based like Zogby. I'm not sure if those surveys will ever be reliable due to volunteer response bias (e.g., a person who signs up to be on the "panel" is probably not representative of the general population).
@SilverBird5000
The reason to weight by Party ID is that it stabilizes inter-poll comparisons. Like race and income, democrats tend to vote certain ways while republicans vote in others (I know, duh). So, if the distribution of party ID in your sample happens to randomly differ from poll to poll (or, more importantly, between the poll and who actually turns up to vote on election day), then your poll results will vary.
So, take SurveyUSA's Iowa poll today and a month ago. Today, Obama leads by +4. A month ago, he led by +9. Now, I haven't done the calculations, but looking at the detailed crosstabs, it seems to me that the variation is mostly or entirely explained by changes in the distribution of Party ID among the respondents.
Today, the poll had 37% R, 38% D, 25% I. A month ago, they sampled 28% R, 45% D, 24% I. So, you can see, this poll had a lot more republicans than the last one. Not surprisingly, McCain did better.
So, which is a better result? Hard to say. Here's where it gets into the guesswork. The Party ID distribution among voters in the Iowa exit poll from 2004 is much closer to the more recent, +4 poll. So, did the exit poll get the distribution of Party ID right? Probably close. But is 2004 a good model for 2008. Probably not. In most states, we've seen much higher democratic turnout this year, and Party ID has become more democratic nationwide.
So, you look at the data hard, then you make a judgment.
Mike H
Part of the reason that the polling average might show a candidate losing, but 538 projects them to win is due to mistaken pollster assumptions on turnout and demographic weighting. 538 attempts to correct for through the multivariable regression.
So the polling average shows NV at McCain +2.8, then adjusting for trend McCain +.9, the regression shows Obama +8.5 and finally the projection of Obama +1.1. Run that simulation 10,000 times and NV comes out with Obama winning 56% of the time.
What this means is that Obama is underperforming in Nevada, and by quite a bit.
Something that I have noticed as the site becomes more popular is that 99% of the complaints about the site are from people who haven't read the methodology.
Judas Priest,
Respectfuly I implied a Stalinist nightmare of internal surveillance, not purges. Although unlawful arrests of Persians do seem within the skope of the foreseeable future. Still, I continue to believe that giving imunity to telecoms for supporting internal surveillance is treacherous and anti American.
@silverbird5000:
For a discussion of this issue consult the website poolster.com archive (these postswere originallt posted on a site mysterypollster.com, which has joined pollster.com.
Here is a link to a number of the comments on this topic:
http://www.mysterypollster.com/main/2004/09/should_pollster.html
It would be intriguing, but tedious for me, to see which states have the greatest degree of difference between the polling average and the regression model. If one was enterprising they could then see if there is in fact a bias of sorts within the regression model.
I have been reading Nate's work in baseball for quite some time, and though he has never figured out to project Ichiro, his accuracy is quite good.
PENIS, the +18 McCain WV poll is ancient and has almost no validity. It will soon disappear from consideration as irrelevant. As for the +8 McCain poll by Rasmussen, that too is over two weeks old. To improve predictive power, Nate does a trend adjustment based on recent trends in polling and employs a demographic regression which helps more accurately to predict the real likelihood of victory for each candidate. Nate is ahead of the polls, but I think will be borne out in the end.
Bedir wrote: "Something that I have noticed as the site becomes more popular is that 99% of the complaints about the site are from people who haven't read the methodology."
What methodology? They not only haven't made any effort to understand it but even if they did they'd reject it because they don't like the results.
Wait til McCain makes an inevitable comeback. They will like the methodology then but won't understand it when somebody says it's regression to the mean.
They understand the difference between propaganda and social science, but this is a crowd that has spent years and years deprecating science without understanding it, and so they have only propaganda and noise-making to rely on.
I'm just hoping that it isn't regression past the mean, ala Mariners 08
(I really do need to get new glasses.) That's pollster.com. My problem is that my computer glasses have a broken frame and my regular glasses aren't all that good for computer work. I have real trouble proofreading, and my data entry skills suck. Appointment next week so I'll have new glasses sometime in July.
Bedir: Good to see you're a BP reader. Actually Nate has two things he's never quite got a handle on: Ichiro and the Braves (at least until Leo Mazzone left). But his projections are, um, "bedir than average."
Agreed about the regression . . . when it happens I hope it's not into quicksand. Despite some of the commentary today, the fact that the GOP 527's appear quiescent at this time doesn't mean anything. The Swift Boaters showed up only around September 1st in 2994. That's two+ months away.
@ penis:
I notice your comments are quite small. Any connection to your name?
Sliughtly more seriously, this site has an extensive discussion of its methodology. Why don't you try reading and understanding that before displaying your shortcomngs, both intellectual and otherwise, here.
Sorry about that date: 2004, not 2994.
The California unemployment rate jumped from 6.2% in April to 6.8% in May, and in L.A. it went from 5.9% in April to 6.7% in May.
Combine that with gas prices that, in southern California are near $5 a gallon, and imminent big contractions in the airlines, and the end of the HELOC-driven (home equity line of credit) consumer spending, and Nevada is soon going to be hurting in a big way.
Foreign toruists drawn by the weak dollar will offset some of this, but not anywhere near all of it.
I don't think the pollsters, analysts, or candidates themselves are truly catching what's happening on the economic front, and just how critical an issue this is. I'm telling you, there's a shitstorm of major proportions and it's about to go crazy.
As for November, you can go back for 60 years, and you'll see that the direction of the unemployment rate in the spring of an election year accurately predicted the result of every presidential election other than 1956.
When unemployment went up, the party in power lost. When it went down, the party in power won. When it stayed the same, the election was tight, and the incumbent lost. Right now, unemployment is 0.3% higher in May than it was in March. The June numbers are going to be a disaster, and McCain's fate will be sealed.
Kermit:
Then the use of the term Stalinist is so hyperbolic that the meaning you intended is submerged in the connotations of that word. The wiretapping Congress is about to legitimize is beyond what is done in countries we generally regarded as free only in that our technology permits greter prying.
I disagree with the calculus that the limitations the new bill imposes are worth it, but without the new bill Bush and his cohorts would feel free to continue without check what they are doing. And Bush would have vetoed any bill that did not have the immunity provision.
It might seem unusual that a retired lawyer (and litigator) would say this, but lawsuits are not the way to handle this issue, especially since there is a great likelihood that the courts would have dismissed them for a number of reasons.
If Nelson Mandela could opt to avoid prosecuting many of the people who committed unspeakable acts in attempting to maintain the racist system in South AFrica, we could be willing to accept this weaker step to get the issue settled.
As i understand from what has been reproted, the new system will at least leave a paper trail behind so that years down the line we can see what has been done in the name of this country.
WV is no different than Kentucky. Kentucky has a big lead for McCain and so should WV. They both were states Hillary won big in and they were both states almost 1/3 of Democrats said they were backing McCain.
Hell 1/2 of the voters in each state do not think Obama is honest. He stands no chance in WV.
Again Romney could help McCain in NV, with the mormon population out there.
Personally I know this site has a lean to the left with its readers, but Romney helps McCain a tad bit in Michigan, Colorado, and Nevada.
Michigan-family is well liked, unpopular Dem governor.
Colorado-Mormons
Nevada-Mormons
Again even if Romney helped McCain out by just 1% in these states, that may be enough to push him to victory in them.
I think I'm going to whine until we get that extra decimal of precision back on the state-by-state Win %.
I only did one daily yesterday so I didn't get the CO and GA polls in, but they're accounted for now.
Rasmussen's NV poll showed an Obama gain of three points from their previous poll, but Nate dropped Obama's chances down by four points (or somewhere between 3.5 and 4.4 points -- how can I tell without the extra precision?)
Survey USA's CA poll showed an Obama gain of four points but no gain in Obama chances (What can you expect? He's up to 96% already.)
Nate reflected the Obama drops in IA and CO and in the Obama gains in GA and NH. But my curiosity is with the movement of the states that weren't polled.
LA is up 5 points, MT(?????) is up 4, both Carolinas are up 4 and Texas is up 4. Texas!
IN is down by two. Go figure.
Your analysis is right, Charles. Where the incumbent party missed badly is that in their continued effort to minimize the consequences of the banking/housing scandals, they decided to use a mild palliative: give everybody $600. That is being eaten up by the increase in gasoline prices (and soon enough, home cooling and heating).
What they coulda-shoulda done but it's too late for the election cycle is to fund a massive public works program -- highways, bridges, waterways, etc., mainly through grants to the states. That would have both shored up state budgets, put a helluva lotta people to work (can't outsource most of that, unless you allow illegal immigrants to do most of the work), and helped to rebuild a badly deteriorating infrastructure.
But that mistake is "water over the damn," so to speak.
@Clark/Judas:
Thanks for your comments/clarifications. I'm making my way through the Mystery Pollster posts, but between those and what Clark said, I'm starting to see the logic- particularly with respect to inter-poll comparisons.
I'm still wary of weighting Party ID like a demographic variable. I guess this puts me in the SurveyUSA camp, rather than the Rasmussen camp - i.e. preferring the less fixed, more fluid conception.
Just a general comment. It is SO nice to be on a site where there is little evidence of manic/emotional anger against statistics. Of course, there can always be a debate but it seems to me that, by and large the folks here can spell and be logical.
Thanks
I went on youtube the other week, and looked at videos from Brian Schweitzer, governor of MT (D)... i think he is just the type of vp choice obama should pick, because of his credentials at understanding the type of low-info voters that are common in places like NV and a few other states.
To those who are complaining about 538's "bias," I have one question:
Which part of Obama leading McCain nationally by anywhere from 4-5 points on average don't you understand? If you think that kind of margin doesn't translate into 300+ EVs, you don't know history prior to 2000.
If McCain were leading by 4-5%, several states that are now tossups would be lean McCain and several states that are now lean Obama would be tossups. That's the way the game works. And I'm sure 538 would reflect that accurately.
Holy shit, reality is so biased toward Obama.
To be fair, though, I can't exactly blame so many people for not understanding the methodology well enough to appreciate all aspects of it. I mean, there hasn't even been a post fully explaining every step taken to make these projections in technical terms, let alone standard mathematical or layman's terms. I was able to understand everything going on before the changes in the past week and now even I'm lost on a couple points.
Also, lol at Obama's mode being over 390 EVs.
Equivalently, lol at McCain's mode being half of what he'll need to even seem competitive.
@Paul Bradford:
Even if Nate put in the 10ths of a percent that you want, it really wouldn't mean anything. The measurement error of those numbers is such that the extra digit is noise.
10 is not the same as 10.0 which is not the same as 10.00 Putting in those extra digits is a claim that you can actually detect that small a difference. If I say a certain line is 10 inches long, I am stating that I can tell the difference betwen 10 and 9 or 11. If I state that the line is 10.0 inches long, I am claiming that I can distinguish between 10.0 and 9.9 or 10.1, and so on.
The sampling and other errors that go into the numbers used to generate the state winning %s make the estimates far too imprecise, I think, to justify adding the third place.
"Trevor Howard
To those who are complaining about 538's "bias," I have one question:
Which part of Obama leading McCain nationally by anywhere from 4-5 points on average don't you understand? If you think that kind of margin doesn't translate into 300+ EVs, you don't know history prior to 2000."
RATS! I was going to say this and then you jumped in ahead of me! LOL!
There are some people who just want to see "pretty maps" with McCain (or Obama) comfortably ahead, and others who expect the map to reflect their "gut instinct" about what will happen in November.
And, if that doesn't happen, then they scream "bias" without the slightest understanding of even what they are complaining about.
A lot of them are Republican trolls who vividly reflect the joke that "reality has a well known liberal bias." They're used to attacking science and supporting a President who "goes with his gut." So, if suddenly reality is showing something different than what they expected or what is reflected in their own narrow circle, it's no stretch for them to attack this site for "librul, librul, librul, bias."
MVRed
(MCCAIN TROLL)
keep dreaming
McCain ain't going to win shi$
OBAMA all the way
RED states will go BLUE
this election will be historic
100 years from now people will be talking about the 2008 election
i can't wait till November!
I'm Hillary & i approve this post
:)
Ok, back to reality...
Any critiques on 270toWin.com? They have a staggering 99% chance of Dems winning the GE, they run 1,000 simulations.
Election Winner
Democrats
98.9%
Republicans
1.1%
Ties
0.0%
Sorry guys,
But I'm one of the people who 'trusts' the polls and not Nate's (new) overemphasis of the national polls. Based on the Newsweek poll, these national polls are completely worthless.
If Nate wants to put his analysis on it, great! But let's also see a map with just the original concept of a few weeks ago...the state surveys weighted properly, with some regional smoothing.
Nevada and West Virginia..ugh.
Matthew H, if the national polls are "completely worthless," what does that say about state polls, which are fewer in number, much less frequent, and come from a larger number of companies with more diverse and less transparent methodologies?
As the cliché goes, "the only poll that counts is on election day." We'll know then which approach to reading the polls worked best. Till then, it's just great to see this site and so many others offer us different approaches!
Charles Pluckhahn, Juris:
You're both half-right about the economy; it's strange how the severity seemed to sneak upon everyone. Sure unemployment and foreclosures are bad, but this is just the tip of the iceberg. Just like the energy debate morphed from environmental to national security concerns, the economy is going to make a marked transition with severe geo-political implications. I'm sure what I'm saying sounds like an exaggeration, but just pick up a British newspaper. Ironically, there's a tipping point for the economy; if the shit really hits the fan, then the experience factor really favors McCain.
Remember when Clinton supporters laughed at the idea Obama could change the map and preached that she would win OH and FL so nothing else matters... Well, it has become clear that the Democrats elected the right person. Obama has looked very strong in OH and PA and FL looks like it is coming around. Clinton would add maybe AR, WV and nothing else to Kerry map. Clinton would make WA, OR, NV, NM, CO way too close or losses and Obama looks like he'll win them all- maybe not NV and the west coast with ease. MN, IA, WI are all very favorable Obama and Clinton would likely do as well too- same with MI.
But VA leaning DEM, and GA, NC, MT, AK, ND, and SD possibly in play or at least close enough to contest in... wow.
McCain will have no chance in competing in all swing states and matching Obama's efforts- if CLinton had won all he'd have to do was focus on OH, MI, FL, and then make a play at the western states that should be democratic.
I love watching the numbers go higher and higher Obama- if this continues election night may be the proudest moment in recent electoral memory
Matthew H and all those complaining about the new methodology: all the unadjusted data is there. If you want to run your own simulations or make up your own pretty map, go right ahead.
Just imagine how different the numbers in Mississippi might be if Hurricane Katrina hadn't happened (or at least, if a competent administration were in office during the time it happened). Would at least be leaning Obama.
Think about that one.
JV - I can't really comment on 270towin.com's simulation because they don't give their data and the method use to derive the probabilities from the polling data. Still, their numbers seem farfetched.
I must say that my irrational "theory" that given all of the biases that reality shows towards Obama, he should be leading by 20 points jibes well with the Newsweek poll.
No, I don't think Obama will win by double-digits, but thanks to the Newsweek poll I now feel justified in believing that Obama will probably win. Or at least that a Democrat can still win the US presidency.
It's a relief.
Nate
Your suggestion that Obama only has a 5 pt national lead is mistaken. Averaging the seven national polls mentioned in your post and comes out Obama 47.7 and McCain 40.7. That is, a 7 poiint Obama lead. Why do you insist that it is only a 5 pt. lead?
What is the traffic from different countries like?
Nate's old method: assumes states that haven't been polled in weeks or months can be lumped in, unadjusted, with fresh polls showing sizable changes.
Nate's new method: uses the best available info to estimate the CURRENT status of all states i.e., an electoral PROJECTION done right.
Complaints about the current projection should explain why the old assumption was better in their view, not just that they "can't believe" Obama is doing so well in state X (...in an election where the incumbent party is in the toilet and the opponent has little enthusiasm from his own party and has so far run a poor campaign.)
McCain fans: if your candidate starts polling better in the right places, the map will change. In fact it will change overall faster to reflect his success under the new method.
Matthew H, if you want to see what the numbers would look like before the trend adjustment change in methodology, Nate has posted the polling average for each state. This should give you an approximation of what the old averages would have been for each state. The 538 Regression is also listed, so if you want, you can take the extra step and average the polling average with the 538 Regression (using correct weighting of course!) to get really close to the old numbers. Admittedly, a lot more work for you, but since the site is free, most readers seem to prefer the new methodology, the new methodology has better projected new polls than the old one, and most importantly, Nate runs the show here and has made his decision, any complaints seem unwarranted.
Anon @1:30, perhaps Nate is looking at more than just the seven national polls released in the last 24 hours. Most pollster averaging sites would include more than just these seven. For example, RCP has Obama +5.5. http://www.realclearpolitics.com/epolls/2008/president/us/general_election_mccain_vs_obama-225.html
The new methodology was actually quite bad at predicting Friday's NV (McCain up 3) and IO (Obama up 4) polls.
To the last anon,
Nice cherry pick, and you can't even do it well. The Friday Nevada poll MOE had the trend adjustment for Nevada being a valid and somewhat likely response.
As for Iowa, I'm of the beleif that the results of polls taken while the state is suffering a centuries worst flood are likely indicative of nothing.
Should a poll taken during Hurricane Katrina be counted?
Anonymous @ 2:30 - Nate does not average only the seven most recent polls, he uses all polling information since February and plots a Loess regression curve to smooth out the data and minimize the effects of noise/outliers. Essentially, this counts previous polls with an exponential rate of decay for the weighting of polls as they age. This gives a more time-stable (and more accurate) prediction. Before criticizing Nate's math, perhaps you should do a little reading about his methods: I can assure you that they are far more sophisticated and robust than an average of daily polls.
Anonymous @ 11:29 - Nate's predictions aren't valid as predictors of any single poll result. They are not intended to be, any more than would an analysis of a baseball player's current batting ability be valid as a predictor of his batting average in a single game. In fact, the whole purpose of Nate's sophisticated regression analysis is to smooth out the noise in the data and give us a far more accurate prediction than any single poll.
richard, my comment on how bad the new methodology was at predicting Friday's NV and IO polls was in response to another mike's comment that the new methodology was supposedly better at predicting the newest polls. I know Nate did not claim that.
Newsweek's polls have frequently varies from the norm. Additionally, the fact that they used registered voters makes it suspect.
Also, as these guys point out two 2004 polls gave Bush a huge lead in 2004 and we know his victory was nowhere near 15 or 16 points.
This is based on personal observation only, but Northern NV looks real good for Obama, there is alot of Republican and Independent support for him, also registered democrats now outnumber Republicans for the first time.
Las Vegas is more complex; elderly, Jewish and Hispanic voters voted for HRC in large numbers in the primary, if Obama wins their support I think he wins NV in the general.
情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣用品,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,情趣,美國aneros,rudeboy,英國rudeboy,英國Rocksoff,德國Fun Factory,Fun Factory,英國甜筒造型按摩座,甜筒造型按摩座,英國Rock Chic ,瑞典 Lelo ,英國Emotional Bliss,英國 E.B,荷蘭 Natural Contours,荷蘭 N C,美國 OhMiBod,美國 OMB,Naughti Nano ,音樂按摩棒,ipod按摩棒,美國 The Screaming O,美國TSO,美國TOPCO,美國Doc Johnson,美國CA Exotic,美國CEN,美國Nasstoy,美國Tonguejoy,英國Je Joue,美國Pipe Dream,美國California Exotic,美國NassToys,美國Vibropod,美國Penthouse,仿真按摩棒,矽膠按摩棒,猛男倒模,真人倒模,仿真倒模,PJUR,Zestra,適趣液,穿戴套具,日本NPG,雙頭龍,FANCARNAL,日本NIPPORI,日本GEL,日本Aqua Style,美國WET,費洛蒙,費洛蒙香水,仿真名器,av女優,打炮,做愛,性愛,口交,吹喇叭,肛交,魔女訓練大師,無線跳蛋,有線跳蛋,震動棒,震動保險套,震動套,TOY-情趣用品,情趣用品網,情趣購物網,成人用品網,情趣用品討論,成人購物網,鎖精套,鎖精環,持久環,持久套,拉珠,逼真按摩棒,名器,超名器,逼真老二,電動自慰,自慰,打手槍,仿真女郎,SM道具,SM,性感內褲,仿真按摩棒,pornograph,hunter系列,h動畫,成人動畫,成人卡通,情色動畫,情色卡通,色情動畫,色情卡通,無修正,禁斷,人妻,極悪調教,姦淫,近親相姦,顏射,盜攝,偷拍,本土自拍,素人自拍,公園露出,街道露出,野外露出,誘姦,迷姦,輪姦,凌辱,痴漢,痴女,素人娘,中出,巨乳,調教,潮吹,av,a片,成人影片,成人影音,線上影片,成人光碟,成人無碼,成人dvd,情色影音,情色影片,情色dvd,情色光碟,航空版,薄碼,色情dvd,色情影音,色情光碟,線上A片,免費A片,A片下載,成人電影,色情電影,TOKYO HOT,SKY ANGEL,一本道,SOD,S1,ALICE JAPAN,皇冠系列,老虎系列,東京熱,亞熱,武士系列,新潮館,情趣用品,情趣,情趣商品,情趣網站,跳蛋,按摩棒,充氣娃娃,自慰套,G點,性感內衣,情趣內衣,角色扮演,生日禮物,生日精品,自慰,打手槍,潮吹,高潮,後庭,情色論譠,影片下載,遊戲下載,手機鈴聲,音樂下載,開獎號碼,統一發票號碼,夜市,統一發票對獎,保險套,做愛,減肥,美容,瘦身,當舖,軟體下載,汽車,機車,手機,來電答鈴,週年慶,美食,徵信社,網頁設計,網站設計,室內設計,靈異照片,同志,聊天室,運動彩券,大樂透,威力彩,搬家公司,除蟲,偷拍,自拍,無名破解,av女優,小說,民宿,大樂透開獎號碼,大樂透中獎號碼,威力彩開獎號碼,討論區,痴漢,懷孕,美女交友,交友,日本av,日本,機票,香水,股市,股市行情, 股市分析,租房子,成人影片,免費影片,醫學美容,免費算命,算命,姓名配對,姓名學,姓名學免費,遊戲,好玩遊戲,好玩遊戲區,線上遊戲,新遊戲,漫畫,線上漫畫,動畫,成人圖片,桌布,桌布下載,電視節目表,線上電視,線上a片,線上掃毒,線上翻譯,購物車,身分證製造機,身分證產生器,手機,二手車,中古車,法拍屋,歌詞,音樂,音樂網,火車,房屋,情趣用品,情趣,情趣商品,情趣網站,跳蛋,按摩棒,充氣娃娃,自慰套, G點,性感內衣,情趣內衣,角色扮演,生日禮物,精品,禮品,自慰,打手槍,潮吹,高潮,後庭,情色論譠,影片下載,遊戲下載,手機鈴聲,音樂下載,開獎號碼,統一發票,夜市,保險套,做愛,減肥,美容,瘦身,當舖,軟體下載,汽車,機車,手機,來電答鈴,週年慶,美食,徵信社,網頁設計,網站設計,室內設計,靈異照片,同志,聊天室,運動彩券,,大樂透,威力彩,搬家公司,除蟲,偷拍,自拍,無名破解, av女優,小說,民宿,大樂透開獎號碼,大樂透中獎號碼,威力彩開獎號碼,討論區,痴漢,懷孕,美女交友,交友,日本av ,日本,機票,香水,股市,股市行情,股市分析,租房子,成人影片,免費影片,醫學美容,免費算命,算命,姓名配對,姓名學,姓名學免費,遊戲,好玩遊戲,好玩遊戲區,線上遊戲,新遊戲,漫畫,線上漫畫,動畫,成人圖片,桌布,桌布下載,電視節目表,線上電視,線上a片,線上a片,線上翻譯,購物車,身分證製造機,身分證產生器,手機,二手車,中古車,法拍屋,歌詞,音樂,音樂網,借錢,房屋,街頭籃球,找工作,旅行社,六合彩,水噹噹,台中隆鼻,果凍隆乳,改運整型,自體脂肪移植,新娘造型,婚禮顧問,下川島,常平,常平,珠海,澳門機票,香港機票,貸款,貸款,信用貸款,宜蘭民宿,花蓮民宿,未婚聯誼,網路購物,婚友,婚友社,未婚聯誼,交友,婚友,婚友社,單身聯誼,未婚聯誼,未婚聯誼, 婚友社,婚友,婚友社,單身聯誼,婚友,未婚聯誼,婚友社,未婚聯誼,單身聯誼,單身聯誼,白蟻,白蟻,除蟲,老鼠,減肥,減肥,在家工作,在家工作,水噹噹,台中隆鼻,果凍隆乳,改運整型,自體脂肪移植,新娘造型,婚禮顧問,下川島,常平,常平,珠海,澳門機票,香港機票,貸款,貸款,信用貸款,宜蘭民宿,花蓮民宿,未婚聯誼,網路購物,婚友,婚友社,未婚聯誼,交友,婚友,婚友社,單身聯誼,未婚聯誼,未婚聯誼, 婚友社,婚友,婚友社,單身聯誼,婚友,未婚聯誼,婚友社,未婚聯誼,單身聯誼,單身聯誼,白蟻,白蟻,除蟲,老鼠,減肥,減肥,在家工作,在家工作,婚友,單身聯誼,未婚聯誼,婚友,交友,交友,婚友社,婚友社,婚友社,大陸新娘,大陸新娘,越南新娘,越南新娘,外籍新娘,外籍新娘,台中坐月子中心,搬家公司,搬家公司,中和搬家,台北搬家,板橋搬家,新店搬家,線上客服,網頁設計,線上客服,網頁設計,植牙,關鍵字,關鍵字,seo,seo,網路排名,自然排序,網路排名軟體,交友,越南新娘,婚友社,外籍新娘,大陸新娘,越南新娘,交友,外籍新娘,視訊聊天,大陸新娘,婚友社,婚友,越南新娘,大陸新娘,越南新娘,視訊交友,外籍新娘,網路排名,網路排名軟體,網站排名優化大師,關鍵字排名大師,網站排名seo大師,關鍵字行銷專家,關鍵字,seo,關鍵字行銷,網頁排序,網頁排名,關鍵字大師,seo大,自然排名,網站排序,網路行銷創業,汽車借款,汽車借錢,汽車貸款,汽車貸款,拉皮,抽脂,近視雷射,隆乳,隆鼻,變性,雙眼皮,眼袋,牙齒,下巴,植牙,人工植牙,植髮,雷射美容,膠原蛋白,皮膚科,醫學美容,玻尿酸,肉毒桿菌,微晶瓷,電波拉皮,脈衝光,關鍵字,關鍵字,seo,seo,網路排名,自然排序,網路排名軟體,英語演講,托福,Toastmaster,汽車借款,汽車借款,汽車借款,汽車貸款,汽車貸款,借錢,借貸,當舖,借款,借貸,借錢,週轉,
^^ nice blog!! thanks a lot! ^^
徵信, 徵信社,徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 外遇, 抓姦, 離婚, 外遇,離婚,
徵信, 外遇, 離婚, 徵信社, 徵信, 外遇, 抓姦, 徵信社, 徵信, 徵信社, 徵信, 外遇, 徵信社, 徵信, 外遇, 抓姦, 徵信社, 征信, 征信, 徵信, 徵信社, 徵信, 徵信社, 征信, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信社, 徵信社, 徵信, 外遇, 抓姦, 徵信, 徵信社, 徵信, 徵信社,
^^ nice blog!! ^@^
徵信, 徵信, 徵信, 徵信社, 徵信社, 徵信社, 感情挽回, 婚姻挽回, 挽回婚姻, 挽回感情, 徵信, 徵信社, 徵信, 徵信, 捉姦, 徵信公司, 通姦, 通姦罪, 抓姦, 抓猴, 捉猴, 捉姦, 監聽, 調查跟蹤, 反跟蹤, 外遇問題, 徵信, 捉姦, 女人徵信, 女子徵信, 外遇問題, 女子徵信, 徵信社, 外遇, 徵信公司, 徵信網, 外遇蒐證, 抓姦, 抓猴, 捉猴, 調查跟蹤, 反跟蹤, 感情挽回, 挽回感情, 婚姻挽回, 挽回婚姻, 外遇沖開, 抓姦, 女子徵信, 外遇蒐證, 外遇, 通姦, 通姦罪, 贍養費, 徵信, 徵信社, 抓姦, 徵信社, 徵信, 徵信公司, 徵信社, 徵信, 徵信公司, 徵信社, 徵信公司, 女人徵信, 外遇
徵信, 徵信網, 徵信社, 徵信網, 外遇, 徵信, 徵信社, 抓姦, 徵信, 女人徵信, 徵信社, 女人徵信社, 外遇, 抓姦, 徵信公司, 徵信社, 徵信社, 徵信社, 徵信社, 徵信社, 徵信社, 女人徵信社, 徵信社, 徵信, 徵信社, 徵信, 女子徵信社, 女子徵信社, 女子徵信社, 女子徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社,
艾葳酒店經紀公司提供專業的酒店經紀, 酒店上班小姐,八大行業,酒店兼職,傳播妹,或者想要到打工兼差、打工,兼差,或者八大行業,酒店兼職,想去酒店上班, 日式酒店,制服酒店,ktv酒店,禮服店,整天穿得水水漂漂的,還是想去制服店當上班小姐,水水們如果想要擁有打工工作、晚上兼差工作、兼差打工、假日兼職、兼職工作、酒店兼差、兼差、打工兼差、日領工作、晚上兼差工作、酒店工作、酒店上班、酒店打工、兼職、兼差、兼差工作、酒店上班等,想了解酒店相關工作和特種行業內容,想兼職工作日領、假日兼職、兼差打工、或晚班兼職想擁有快速賺錢又有保障的工作嗎???又可以現領請找專業又有保障的艾葳酒店經紀公司!
艾葳酒店經紀是合法的公司工作環境高雅時尚，無業績壓力，無脫秀無喝酒壓力，高層次會員制客源，工作輕鬆，可日領、現領。
一般的酒店經紀只會在水水們第一次上班和領薪水時出現而已，對水水們的上班安全一點保障都沒有！艾葳酒店經紀公司的水水們上班時全程媽咪作陪，不需擔心！只提供最優質的酒店上班,酒店上班,酒店打工環境、上班條件給水水們。心動嗎!? 趕快來填寫你的酒店上班履歷表
水水們妳有缺現領、有兼職、缺錢卡奴的煩腦嗎?想到日本留學缺錢嗎?妳是傳播妹??想要擁有高時薪又輕鬆的夜間兼職工作,打工機會和,假日打工,假日兼職賺錢的機會嗎??想實現夢想卻又缺錢沒錢嗎!??
艾葳酒店台北酒店經紀招兵買馬!!徵專業的酒店打工,想要去酒店的水水,想要短期日領,酒店日領,禮服酒店,制服店,酒店經紀,ktv酒店,便服店,酒店工作,禮服店,酒店小姐,酒店經紀人,
等相關服務 幫您快速的實現您的夢想~!!
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店小姐兼職,
便服酒店經紀,
酒店打工經紀,
制服酒店工作,
專業酒店經紀,
合法酒店經紀,
酒店暑假打工,
酒店寒假打工,
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店小姐兼職,
便服酒店工作,
酒店打工經紀,
制服酒店經紀,
專業酒店經紀,
合法酒店經紀,
酒店暑假打工,
酒店寒假打工,
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店小姐兼職,
便服酒店工作,
酒店打工經紀,
制服酒店經紀,
酒店經紀,
菲
梵,
Post a Comment