Ha, Nate tweeted about this more in amusement than anger. Truly the stuff of parody.
Like one of the comments under the article: “Shorter Dean Chambers: Nate Silver’s math is bad because he’s a girly-man who looks like a castrati. My math is better because I’m so manly I even look like a scrotum.”
Ahaha, now Mittens is looking to flip Wisconsin.
This is kind of funny. Depending on who you ask, Wisconsin either is or isn’t a swing state at all. I guess that that state is his best shot at taking away an Obama stronghold, but I don’t fucking see it happening.
The national polls and the state polls do not agree. If you like, you can just pick one and ignore the other as you prefer to do. But most of us think a more intelligent approach is to attempt integrate both into a model of what is actually going on.
Rasmussen’s tracking poll went up a point today to Romney +4, Gallup’s stayed the same at Romney +5, and the RAND poll has Obama up by about 6.5 points.
There’s a few more I’m missing. But the question is, which polls do you trust?
Like I said earlier, I think it’s quite likely that Romney has opened up a narrow lead in the popular vote, but he is not making any headway in the swing states that matter. Period.
Firefox stops them from opening.
Actually, he does, sort of, by linking to a National Review article that criticizes Nate’s model. But that, of course, is very lazy.
Idle speculation: I’m just wondering if the Nate Silver smear campaign is in preparation for him having a look at that vote-flipping data that’s going around at the moment.
I laughed.
I noted elsewhere that a man who looks like Baron Harkonnen from Lynch’s “Dune” has very little room to talk about attractiveness or masculinity.
Both are not true. If Romney indeed leads in the popular vote, he is probably also leading in the swing states. gallup’s swing state survey confirms this.
So either the polls showing Romney ahead are wrong, or the polls showing Obama ahead are wrong.
But we’re not talking about individual polls- that’s a common mistake. In aggregate, the polls show a very tight race nationally, and an Obama lead in OH, NV, WI, and IA (with Romney leads in NC and FL, and CO, VA, and NH close to toss-up). It’s very possible that it’s nearly 50-50 nationally, with small leads for Obama in those swing states, just like the polls say.
Silver’s model says Obama is up by 1.7. The swing state polls, as you have pointed out, are part of that estimate.
If Romney is actually up by 1, then that’s a 2.7 point swing, which means Romney is probably ahead in Ohio, Colorado, Virginia, etc.
Perhaps. But for one thing, obviously, we don’t know if Romney’s actually ahead by 1. For another, Silver’s national and state estimates have some in common, but aren’t completely linked together. He might just be off a little nationally, but pretty close to accurate on the states (or vice versa).
But we’ll see.
But you are leaving out a bunch of national polls when you say that the average has Romney up 1. RCP doesn’t include: RAND, Google, PPP, UPI/CVoter, or Ipsos/Reuters. The average of all of the daily national polls (on Thursday) was an exact tie.
If that is true (and Nate’s average is wrong) then you have about a 1.5 swing to Romney - Ohio is currently averaging O+2.3.
Friday all national polls averaged to O+0.2.
http://www.realclearpolitics.com/epolls/2012/president/us/general_election_romney_vs_obama-1171.html
Romney leads in six national polls, Obama leads in 3, and one is tied. I think it’s pretty safe to say that Romney leads nationally.
When you take into account the fact that all three polls where Obama leads have small sample sizes(under 1000), it becomes even more clear that Romney is winning the popular vote at this time.
Silver’s model is dependent on the swing state polls being more accurate or at least as accurate. If that proves to be wrong, then Silver’s model is going to need a bit of retooling for 2016.
Nope. RCP doesn’t look at all the polls. If you look at all the polls, the picture is different.
Again, you are leaving out at least 4 daily national polls when you rely on RCP only.
RCP’s selection of polls is solid enough. In 2008, they had a 7.6 point spread in the RCP average the day of the election. The actual spread was 7.2. In 2004, the RCP average had Bush up 1.5. The actual result was Bush +2.4
Average margin of error: .65 percentage points.
All polls are not created equal. Does Silver have a reason to weight those other polls equally to the RCP selections?
I don’t know if he weighs them equally, but at least he includes them in his analysis.