California alone might make a huge difference. All those left leaning liberals who don't see their vote making a difference in the presidential race have got a lot of Propositions to fight for or against this time.

Another potential consequence of this is that if McCain does manage to eke out an Electoral College win, he will probably still lose the popular vote.

That will be fun.

I live in one of the only liberal enclaves of NH and there is definitely going to be major turnout here for Senator O. Not sure how strong the "because he's black" factor will be. For me, I'm voting for him because he's a Democrat, and for the most part, liberal. In any case, there are many volunteers roaming the streets and they are all Obama friendly. I just wish they would devote their time to other parts of the state. Of course, I'm a lazy asshole because I never volunteer!

BTW, I wonder how long it will be before our latent racist troll surfaces and calls liberals out for voting for Obama because he is black.

"On a predictive level, my sense is that Obama's popular vote margin is going to be quite a bit higher than his electoral vote tally."

That might have been the single dumbest thing I've read on a political blog all year. (Although your saying that Kerry would've lost to McCain is close.) What does that even mean? Say Obama wins by ten points, which is a little more than any poll predicts. If he does, he'll win with at least 340 electoral votes and probably more. That's a 63-37 margin. It's impossible for Obama to come away with a bigger popular vote margin than an electoral vote margin. The electoral vote always magnifies the difference in the popular vote. Take 1996. Clinton wins by 8.51 points in the popular vote, and forty in the electoral. 88 - Bush wins by 8 points in the popular, 59 in the electoral. 59. 1984 - Reagan wins by 18 points in the popular, 95 points in the electoral. 1976 - Reagan wins by 10 points in the popular and 82 in the electoral. Any time a candidate wins by a big margin - anything over six points - he swamps the electoral college.

Asher... that's just not a given, it's been true, but there are bunch of ways it could turn out differently.

Looking at the national+state polls it's not likely to, but you suggest it is impossible, which it is not.

Asher, it's not written well, but I'd read it this way: "Obama will outperform in the popular vote more than he will in the electoral vote."

At least, the next few sentences are a decent argument for that conclusion.

Asher beat me to it.

"Any time a candidate wins by a big margin - anything over six points - he swamps the electoral college."

Yup.

No, it's a given. In a winner take-all system, which the Electoral College is, margins get magnified. Say Obama wins by ten points, which is about the most he could win by. If he wins that big, he wins a ton of states; he gets every state that's 10% or less more Republican than the national average. That's a lot. He probably swings Indiana, maybe Missouri, Georgia, and even Arizona. Now he's got about 380 electoral votes. For him to underperform his popular vote margin, he'd have to win with just 300 electoral votes or so, which would require him to lose states that he's actually winning, like Florida or North Carolina.

No, it's a given. In a winner take-all system, which the Electoral College is, margins get magnified.

Tell that to Al Gore!

ABC has him up by 11 nationally.

The other day you wrote that he'd win by 8.

So...does he beat the odds on popular vote and win by more than 11? I doubt that.

I'll be happy with an 8 point popular vote win. But I'd like a blowout in the EC please.

Asher, that scenario seems to presume the ten-point popular vote margin is evenly distributed -- which is unlikely to occur.

It's possible to win the popular vote, even by ten points or more, and lose in the electoral college, simply by having the popular votes poorly distributed: piled high in the states that are won, but without a plurality in enough states (or the right states) to get 270 electoral votes.

Ezra:

I disagree with you about the electoral outcomes. I agree regarding popular vote maybe higher than expected, but what you wrote is numerically not possilbe.

Re popular vote and more importantly the composition of that vote, I think one early indicator is the early vote. I mentioned or tried to mention this on the Nate Silver's site. He seems good with the theorectical and first hand accounts, but not so good with how to take new additional data not in his model to integrate it effectively.

The reality is that the early voting demographics are very telling in just how different the composition of this election cycle already differs compared to prior election cycles.

I suppose polls are easier in that one can always point to them and plausibly have pre-made arguments, but for me, the more interesting known unknowns are the early voting.

NC, GA, FL, NV, IN and mulitple other places are showing simply stunning numbers. Did you know that NC allows registration and voting in early voting on the same day? If I am understanding the numbers correctly there were 100,000 additional voters added to the rolls in early voting - or 2 percent more new voters.

There was indeed more African Americans in NC. By contrast look at place like Oregon where the conservative enclaves seem to be underperforming compared to 2004. This is all anecdotal because I can get no one to really go into how these turn out models are all based on prior voting patterns, but what happens when the voting patterns are happening as prior cycles? Here's the ultimate known unknown- if this is an reallignment election the pollsters, including Nate, will not be able to fully predicted what will happen by the close of this election of Tues.

However, there is no way to win the popular vote by 10 and lose the electoral vote since one is linked to the other. It is only in close races where such thingscan be delinked. Its simple math.

No way can Obama win Texas, but we're still seeing record turnout, and I suspect it's for the reason you state---when Obama wins, Texas Democrats (usually demoralized and don't bother voting) want to say they voted for him.

Like Asher, I think the opposite will apply: there may well be states that were considered borderline or lean-Rep that tip to Obama.

It doesn't take a huge popular vote margin, combined with GOTV where it matters, to turn a hefty EC win into a landslide. Daddy Bush beat Dukakis 53.4-45.6 in the pop. vote, and came away with over 400 EVs.

You have to assume that the GOTV on both sides is designed to optimise the return in terms of EVs. Without a significant third-party spoiler, it's basically a zero-sum game in which everything outside a four-point spread nationally (52-48) tends to magnify the EC total.

You can also see the RNC investing in places like Montana, which suggests that they appreciate the symbolic value of the electoral map, in spite of how it distorts the importance of those big, sparsely-populated right-angled states in the mountain-west.

It's possible to win the popular vote, even by ten points or more, and lose in the electoral college, simply by having the popular votes poorly distributed: piled high in the states that are won, but without a plurality in enough states (or the right states) to get 270 electoral votes.

It's possible in theory, if you have an incredibly geographically polarized electorate, such that you run up 90/10 margins in places like California and Michigan New England and very narrowly lose the South. Or the converse. But that's never happened, and it's never happened for a very good reason; we're not that polarized. If you win by huge margins anywhere, you're probably going to win by small margins most everywhere else. Take 2004. Besides Massachusetts and DC, the biggest margin that Kerry could amass was an 18-point margin in New York. After that it was 10-point margins or less in all his states. For your scenario to play out, you'd need immense margins in the states you won, and tiny margins in the states you lost. But the spread's never that big. You only see huge margins in states in landslides, not in 10-point wins.

I think the fact that the electoral college and popular vote can have different winners, shows that neither simply magnifies the other. But regardless of what is generally possible, pollster has Obama well over 2 to 1 in the electoral college. Does anyone expect anything close to say 70/30 in the popular vote?

But anecdotally, I'm hearing a lot of folks this year who want to vote for Obama not just because they want to assure his win, but because they want to be able to say they voted for the first black president. They want to be part of this moment, if it indeed manifests.

As a resident of the District of Columbia, my vote couldn't have a lower probability of counting, but this is exactly the factor that will motivate me to take the time to vote.

Let me make this clear- if Obama is winning by the high single digits or even in the 6 to 7 range, there is no way your statement is true. Even with the update. He would had to have won Ohio and PA. The reality of the early vote also says your comment is way off because his turnout is high in states like NC, GA and multiple so called red states- several of which like VA and NC he will probably, I predict, win. The only one here who remotely says anything along this thread that makes sense is Asher. I am not even a numbers guy, by the way, and I can say this will some confidence.

This is rather simpleminded. In reality, if Obama outperforms his polls in the popular vote, he will win in places like Ohio. It's a guarantee. Slightly higher liberal turnout in California just doesn't add up to major changes in the national percentage. Not that they shouldn't make the huge sacrifice of voting: the bigger the victory, the stronger the mandate!

But let's also think about what it would take for Obama to really be seen as outperforming his polls: has has a 6-7 point average lead, but many. many major standard polls (non-trackers) have consistently had him at +10-12. So in a sense, even if he hit that margin, he really wouldn't be outperforming, just beating the average. But he could also match the Mason-Dixons of the polling world too -- ie a 2 or 3 point margin. Would that be underperforming? As regard the average, sure. But if the surprise of the night is that he hits a margin that some pollsters pegged, but that isn't the average, is that over- or underperforming? It seems to me he needs to either lose or win by like 13 or more in order to under- or overperform.

"There's a certain set of left-leaning voters in places like California who don't turn out because their vote doesn't matter."

This is stupid because there are local and state wide races that matter. Actually, as a young political science student, one of my professors said that we had the biggest impact in local races. I hope that people in CA, NY, IL, etc. don't take that lazy attitude.

well the crazies will be out in CA due to Prop 8 and 4 because you k now despite a potential great depression and a never ending war, the most important is gay folks and parental notification. Sad, but true.

I'll second Amanda Marcotte...I live in Texas, I know McCain is going to win our electoral votes, but I still went to the polls last week and cast my early vote for Obama. I want to be able to someday tell my kids, grandkids, etc that I was part of this movement, and I voted for Obama.

Put it another way: it's hard to extrapolate state numbers from national numbers, or the outcome in California from the outcome in Utah, or the outcome in Minnesota from the outcome in Texas, but a combination of national and state polls should give you enough data to guesstimate how much of the lead is likely to be 'overkill' in favourable states. That's certainly what the campaigns are working on.

No way can Obama win Texas, but we're still seeing record turnout, and I suspect it's for the reason you state---when Obama wins, Texas Democrats (usually demoralized and don't bother voting) want to say they voted for him.

If Obama wins big and takes a bunch of normally red states with him while narrowing the gap significantly in a state like Texas, I think it's only logical to assume that his campaign would make a run for the state in 2012, assuming he's not fighting for his political life. In fact, I'd say the more red states he takes, the likelier this becomes. The infrastructure will be in place in states like Indiana, North Carolina, Virginia, and Georgia, and while it might need to be refined, it probably wouldn't need as much work as the one in Texas might. (This, of course, assumes that even with a large number of congressional seats and voter files to go along with them, it'll take time to refine the list and expand it.) If the Democrats can make Texas a competitive state in the next few cycles, along with states like Virginia and North Carolina, the Republicans would probably find it impossible to win a national election unless they figured out a way to make inroads in the Northeast.

As I said before, it will be helped along if the margin is a lot smaller than it usually is this year. So I definitely hope those Texas Democrats get out and vote this year.

"Let me make this clear- if Obama is winning by the high single digits or even in the 6 to 7 range, there is no way your statement is true. Even with the update. He would had to have won Ohio and PA. The reality of the early vote also says your comment is way off because his turnout is high in states like NC, GA and multiple so called red states- several of which like VA and NC he will probably, I predict, win. The only one here who remotely says anything along this thread that makes sense is Asher. I am not even a numbers guy, by the way, and I can say this will some confidence."

Thank you. The thing about predicting presidential politics is this. Every state tends to be a certain number of points more or less Republican than the national average. Ohio, 2004 aside, is probably a couple points more Republican than the national average. That is, if the popular vote's a tie, Ohio probably goes Republican by 2. So yeah, if the election's a close one, Obama could lose Ohio. But if he outperforms his polling nationally, like Ezra's saying, Ohio's going to Obama by 5 points. For Obama to lose Ohio while winning nationally, Ohio would have to be a much more Republican-biased state than it is. Now, it isn't so simple as that makes it out to be - demographics and turnout change over time, so you can't always project from what happened four years prior. For instance, NC's historically 7-9 points more Republican than the national average, but this year it's tracking a lot closer to national polling. Black turnout being the main reason. But the thing is, there's no reason, none at all, for Ohio to trend so far away from what happens in the popular vote overall, like Ezra's suggesting it could. Obama will, if anything, overperform the electoral vote expectations people have for him. And about this:

"I think the fact that the electoral college and popular vote can have different winners, shows that neither simply magnifies the other."

If you have a very close race, it's true that you can lose the electoral college and win the popular vote, yeah. But once you're over a five-point margin, you're always going to see big magnification in the electoral college because of the winner-take-all effect.

If that effect is as big as I think it might be, it could pump up Obama's popular vote totals rather significantly.

Ohio isn't Indiana or Kentucky - it has a lot more cities and a lot more people, which is how it gets 20 electoral votes in the first place.

Similar remarks apply to Virginia and Florida, and if Obama carries those three and holds Pennsylvania he's well over 300 EV.

"But anecdotally, I'm hearing a lot of folks this year who want to vote for Obama not just because they want to assure his win, but because they want to be able to say they voted for the first black president. They want to be part of this moment, if it indeed manifests."

I'm offendede by the affirmative action implied in this comment of yours, Ezra.

Obama's not successful solely, or even primarily, because he's black. He's potentially transformative because of his eloquence for one, and his ability to translate his appeal into on-the-ground mobilization.

Democrats have simply not fielded a candidate with his charisma since Bobby Kennedy, and have not fielded a candidate with his organizational abilities since the days of machine politics.

Demoralized Democrats are confident that we will win this election, and win it handily. While this has very much to do with Barack Obama, it has very little to do with his race.

brewmn: I think you're a little quick to react here--I don't think there's anything offensive in Ezra's comment.

Certainly, on my internal list of "reasons to vote for Obama", being able to say I voted for the first black president is on the list. It's not high on the list -- certainly not a decisive factor for me -- but it's there; and I can imagine that this might be the case for lots of people.

I also agree with Ezra that this is likely to be part of the reason for this year's expected-to-be-unusually-high turnout. People who would normally lean Democrat but would not bother to vote may take the extra time to vote to stake their tiny place in history.

I guess you can be offended at me and these hypothetical other voters for acknowledging that race might play some part in our decision of who to vote for (or whether to vote), but I don't think you should be offended by Ezra reporting on the idea.

Thanks for responding, Carl. I'm happy that I am sending a message about racial progress in America by voting for Obama as well.

I just thought the lack of qualifications in Ezra's statement play into the wingnuts' arguments that whites are supporting Obama solely because of racial guilt.

I'm glad Obama's black; I'm more glad that I can finally vote for a Democrat for president who can make the case for our side forcefully and eloquently. We've had too many well-meaning, technically qualified candidates who couldn't inspire.