The scenarios: 1-Nate Silver is right

Byline: | Category: 2012 | Posted at: Thursday, 1 November 2012

(With five days left before Election Day,I  foresee see five possible scenarios.  Each day, we’re going to explore one of them.  This post is the first installment of the series.)

There are two schools of thought in polling the presidential race:  one holds that states don’t stray too far from their partisan leanings, so the national poll numbers are a good predictor of electoral outcomes; the other says that the electoral vote determines presidents, therefore, we should analyze state poll numbers in the states most likely to decide the race.  The divergence between the state and national polls this year is so stark that they cannot both be right.

Nate Silver of is of the second school of thought.  He has looked at the 2008 and 2010 races and concluded that the polling of the states yielded a better predictor of the final outcomes.  Silver, therefore, heavily weights state polling in a complex mathematical model that he runs on his site.  He discounts some political conventional wisdoms, such as the idea that there is a sticking point around 47 or 48 percent that serves as an upper boundary to an incumbent’s support, and the notion that last-minute deciders are more likely to break to the challenger. 

Under this scenario, the race ends:  Obama 50.5%, Romney 48.5%, other 1%.  Nearly every swing state breaks Obama’s way.  Colorado and Virginia do so just barely, with only a few tens of thousands of votes separating the contestants.  Iowa, New Hampshire, and Wisconsin choose the incumbent by margins of three to five points.  Ohio provides the 270th electoral vote by slightly less than a hundred thousand votes.  Nevada continues its big shift leftward.  And despite the GOP’s quadrennial sisyphean efforts, Michigan, Minnesota, and Pennsylvania were never very close.  Only North Carolina and Florida go to Romney, and not by very much. 

Exit polls indicate that, while voters are most concerned about jobs and the nation’s economic future, they view Mitt Romney as too extreme and out of touch.  Republican recriminations against Romney begin the very next day:  how could he have possibly sat on hundreds of millions of campaign dollars all summer long and let Barack Obama define him that way?  How could a Republican from Massachusetts possibly be extreme?  How could he have been so cautious and inept?

Downstream, Obama’s comfortable margin of victory is devastating to Republican hopes in the Senate.  The GOP picks up a senator in only North Dakota, while losing in Massachusetts and Maine.  With just 10 seats to defend in 2012, while Democrats had 23 on the line, this is a crippling defeat for the GOP.  The only bright spot in Republicanville next Tuesday is in the lower House.  They manage to keep it, but with a smaller margin of between 225 and 230 seats.

The only changes between the 2012 map under the Nate Silver scenario and the one from 2008 are Florida, Indiana, North Carolina, and Nebraska’s lone electoral vote in the state’s 2nd congressional district.  With near-record high participation from blacks, hispanics, and youth, as well as continued strong support from women, Barack Obama overcomes higher angry-white-male turnout and falls only two-and-a-half points from his totals four years ago to become the first incumbent in modern history to decrease in popular support from his first election and still win a second term.

The final electoral tally is 303-235 in the favor of Obama.


See all the scenarios:

Scenario 1:  Nate Silver is right

Scenario 2:  RCP is right

Scenario 3:  Rasmussen is right

Scenario 4:  Gallup tracking poll is right

Scenario 5:  Gallup electorate poll is right

And the prediction is . . .

Share this post:

12 Responses to “The scenarios: 1-Nate Silver is right”

  1. JR Says:

    Better to get the laughable scenarios out of the way and focus on more serious ones as election day gets closer.

  2. Greg Q Says:

    Do these results occur before or after Santa Claus comes down the chimney and gives all the “good” (read “Democrat”) little boys and girls all the presents they want?

  3. Lurking Observer Says:

    To be fair to Silver, there’s a couple of potential factors here:

    1. State level polling, like state-level economic policies, may be able to get at realities w/ higher fidelity. Polling by folks in-state may well get at questions and issues that would be ignored at the national level.

    2. There could be a last minute October surprise. Never forget that the MSM is firmly in the tank for Obama. Who knows what might be unveiled on Sunday? (Recalling that this is what they did to George W. Bush, w/ the drunk driving charge, which is one reason Silver thinks undecideds don’t always break for the incumbent.)

    3. Along these lines, the jobs number report due out tomorrow might be upbeat (accurately or not, seasonally adjusted or not), persuading some that things are getting better.

    That’s all in addition to minor things like voter fraud (such as in VA), or voting machines oddly recording “Romney” as “Obama.” In a close race, all of this could tip the balance.

    (Which doesn’t get at Silver’s bizarre 75% chance of an Obama win argument.)

  4. Conrad Says:

    I realize that questioning Nate Silver is tantamount to waging a War on Math, but I still think common sense should play a role. And what common sense tells me is that Obama’s overall level of support against his GOP challenger has dropped significantly more than 2.4% from 2008 to 2012. Thus, given that Obama got 52.9% of the vote in 2008, he is not going get 50.5% this time. Not when:

    (a) He’s no longer the candidate of “hope and change”;
    (b) The economy sucks, thanks largely to Obama;
    (c) National debt has ballooned;
    (d) The public is deeply antagonistic toward his one big “accomplishment”;
    (e) America’s desire to elect its first black president is no longer an inducement to vote for BHO;
    (f) He has, it seems, lost substanital support among EVERY SINGLE voting group you can name;
    (g) his opponent in 2012 is not an insufficiently-funded, stale, aging has-been with strong ties to Bush’s unpopular war in Iraq;
    (h) his opponent in 2012 knows how to run a campaign; and
    (i) a lot of other things that I don’t even want to go into right now because the list would simply get too long.

    I don’t know how much all of those factors (a) through (i) are worth, but I’m pretty certain they add up to more than 2.4%. Ten percentage points seems reasonable, but let’s be conservative and call it five.

  5. physics geek Says:

    Nate Silver could be correct, but would it be this year’s version or the one from 2009?

  6. Trent Telenko Says:

    43 months of 8(+)% unemployment followed by a couple of months at 7.5(+)% due to lowered work force participation spells lowered voter turn out by the most affected by long term unemployment, AKA Blacks, Hispanics and Young Voters.

    The Gallup and Pew reports showing lower 2012 voter participation compared to 2004 & 2008 are right in line with that demographic reality.

    Any voting model that does not include that is an exercise in delusion.

    It’s Still the Economy, Stupid!

  7. Parsing the data « Law of Markets Says:

    […] And not this: […]

  8. blahblahblah Says:

    I did a little investigation into the Nate Silver states…

    Take, for instance – Iowa:
    48.6 D, 45.9 R – his real polling average
    48.8 D, 45.8 R – his “adjusted” polling average (because polling is biased towards the GOP, obviously)
    50.9 D, 48.0 R – his “projected vote”

    So he is giving Obama a +2.3 over the “actual” polling average and a +2.1 over his “adjusted polling” average. (blatantly ignoring that no incumbent President has ever picked up undecideds at the polls outside Bush gaining about 1% of them in 2004 after his 9/11 response and his tax cuts just taking place)

    2008 polling versus actual?
    53.9 D, 44.4 R – Actual (+9.5 D)
    54.0 D, 38.7 R – RCP average (+15.3 D)*
    The polls:
    54.0 D, 37.0 R – Des Moines Register (+17 D)
    55.0 D, 40.0 R – Survey USA (+15 D)
    53.0 D, 39.0 R – Research 2000 (+14 D)
    51.0 D, 40.0 R – NBC/Mason-Dixon (+11 D)
    54.0 D, 39.0 R – Quad City Times/R2000 (+15 D)
    52.0 D, 39.0 R – Big 10 Battleground (+13 D)

    52.0 D, 44.0 R – Rasmussen (+8 D) ~ the poll Silver calls “by far the most biased” was by far the most accurate of all the polls; a perfect match when you factor for undecideds having a 2.9 to 0.4 enthusiasm gap (Rasmussen claimed +7D identification enthusiasm, which 2.9/0.4 is 7.25)

    (*side-note – imagine that, apparently the GOP had ALL the enthusiasm in 2008 – undecideds apparently went exclusively McCain without Obama picking up a single one of them…)

    …so, RCP average was 5.7% off on the Republican side of things (no bias there though, right?), while Obama stayed about where he was polling. And when the polling average has 48.6 D right now (with D turnout expectations somehow even matching 2008), its only logical we would just add an additional 2.3 points because, well, Nate Silver just wants it to be that way I guess…

    Now we get to the other major issue in all of it, Iowa Registration changes:
    -11.1% Democrat (roughly 70K voters)
    +5.8 Republican (roughly 40K voters)

    Yet despite these dramatic changes in Registration, which amounts to a 110K registration swing towards Romney in a state Obama won by just 140K, the actual polls which make up Silvers “average” are predicting the same 2008 turnout … why? And with those polls already showing the favorable 2008 Democrat turnout, Silver would increase that average even further by 2.3 because…?

    The only way Silver is correct is if:
    1) a now non-existent Democrat base still turns out in numbers matching 2008 to account for Romney winning Independents
    2) Obama wins Independents at an even higher rate than 2008 to account for the lost Dem voters and enthusiasm

    Nate Silver is 81% confident one of those two things will happen though!

  9. Irony Abounds Says:

    BlahBlahBlah: You are kind of cherry-picking Iowa. In Ohio, the RCP average was about two points too low, with Rasmussen missing by 4.5 points. Likewise, RCP averages for PA, VA, FL and CO all were too low for Obama, with Rasmussen generally being one of the worst offenders in missing O’s votes.

    In any event, I do know one thing, Nate Silver knows a bit more about numbers of all kinds than you do.

  10. Cornfed Says:

    Nate Silver is not a pollster. Nor is he political operative. This has little to do with his mathematical prowess – it has to do with the fact that his model is no better than the info he puts into it (i.e. the state polls).

    To blindly put faith in polls that run so counter to conventional wisdom is foolish. It requires one to dismiss the notions that Democrat enthusiasm is down, that GOP enthusiasm is up, that Party ID is close to even nationally, that GOTV efforts are vastly better. Most of these are clearly true. So, tell us, Irony, is all this wrong? If not, then explain polls that show a D+6, 8 or 10 sample.

  11. blahblahblah Says:

    actually, not really… The RCP average on Ohio was only noticeably off because they didn’t (for some odd reason) include the PPP poll.

    But that is how things should have been anyway – polls with any undecideds remaining should have under-polled Obama – the +7D enthusiasm and +8D on Independents indicates as much. For every 1 vote they underrepresented for McCain, there should have been 8-9 votes for Obama. Outside of the Rasmussen polls, which line up with that nearly perfectly each time, name the 2008 polls you see showing Undecideds swing in that ratio? (and any poll which didnt see something close to a 7D/1R undecided split was almost certainly over-sampling Democrats to begin with)

    Now, that still doesnt explain why Silver would give an incumbent extra percentage points when they never gain them, why he would “adjust” the already-bias polls further towards D turnout despite all evidence showing a much, much, much lower D turnout in 2012, and why he would give Obama an even higher percentage of remaining Undecideds after already adjusting the polls even further towards Democrats anyway.

    Also, he should probably explain things like this (we dont have many of these, but we do have a couple battle-grounds):

    Party Reg
    Dem 34.6%
    Rep 36.6%
    None/Oth 28.8%
    Thats the early-voting breakdown in Colorado

    Party Reg
    Dem 42.0%
    Rep 46.9%
    Other 11.2%
    Thats the early-voting breakdown in Pennsylvania

    Republicans, even in fairly Blue States, consistently win the Election-Day turnout. Yet Republicans are already up in early voting in 2 states which Democrats insist Republicans wont win. Silver has these at 70% Obama and 98% Obama, respectively. Has he bothered to pay attention to the actual voters?

    The Huffington Post sure has:

    Lastly, Silver knows something about numbers but absolutely nothing about Politics and Political Numbers outside of when Obama is giving him his internals (Republicans take only 45 House seats*? Seriously? I think only Pelosi said fewer then that…)

    *To be fair, he did adjust this to 52 right before the election, and he did say “The real FiveThirtyEight forecast, however, has a margin of error of plus-or-minus 29 seats.” while also reassuring Democrats with this little nugget:
    “Now suppose that the forecast understates Democratic support by 2 points. There are 8 seats in which we project the Republican candidate to win by a margin of between 1 and 2 points; now these would also be wiped off the board. Now the Republican gains would be reduced to just 38 seats — and the Democrats would hold the House, 218-217!”

    …plus or minus 29 seats and a 2% enthusiasm gap is the difference between Republicans picking up 63 seats and Democrats holding the House? Wow, thats some fancy number work right there, I tell you…

  12. killrobotkill Says:

    “Better to get the laughable scenarios out of the way and focus on more serious ones as election day gets closer.”