Countdown the Election 2012
Post Reply
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
29-10-2012, 09:32 AM
RE: Countdown the Election 2012
(28-10-2012 10:52 AM)germanyt Wrote:  Obama will not have the same turnout as he did in 2008. That is clear.

That's not at all clear to me if the the length of the early voting lines in MD and DC is any indication. Long lines for early voters in DC, MD

(28-10-2012 10:52 AM)germanyt Wrote:  Romney wins popular, Obama wins electorate, Obama is still president.

Silver's model has that at 5.2% today.

(28-10-2012 09:39 AM)TrulyX Wrote:  Nate Silver is a nerd statistician.

You say that like it's a bad thing. Big Grin

(28-10-2012 09:39 AM)TrulyX Wrote:  I also think, however, it is clear he wants Obama to win, but I don't know if that plays a part in his analysis or not.

That's not clear to me, think he keeps his personal politics pretty close to the vest. In either case, it plays no part in his analysis. He's developed a statistical model that he simply plugs all the day's polls into and then cranks the numbers. The only state his model missed in 2008 was Indiana which went to Obama instead of McCain by 1%. The model also predicted all 35 Senate seats correctly.

I am us and we is me. ... bitches.
Find all posts by this user
Like Post Quote this message in a reply
[+] 1 user Likes GirlyMan's post
31-10-2012, 09:10 PM
RE: Countdown the Election 2012
Lines are long because polling locations are fewer and far between. Republicans have done everything they can to reduce the earling voting numbers that heavily favor Democrats. I saw an article recently that said early voting in Florida was down like 70 percent.

It took me less than 5 minutes to vote in the Republican primary. I left an early voting location because people were waiting over an hour and still hadn't made it halfway through the line.

“Whenever you find yourself on the side of the majority, it's time to pause and reflect.”

-Mark Twain
Find all posts by this user
Like Post Quote this message in a reply
31-10-2012, 09:19 PM (This post was last modified: 31-10-2012 09:32 PM by GirlyMan.)
RE: Countdown the Election 2012
(31-10-2012 09:10 PM)germanyt Wrote:  Lines are long because polling locations are fewer and far between. Republicans have done everything they can to reduce the earling voting numbers that heavily favor Democrats. I saw an article recently that said early voting in Florida was down like 70 percent.

My point is that if early voting lines are long in the People's Republic of Maryland and the Communist District of Columbia where it is a foregone conclusion that they will both go Obama and my vote won't make a difference, it indicates that Democratic voter turnout may not be as weak as you might think. ... Republican attempts at voter suppression notwithstanding.

Silver's model got Obama at close to a 4:1 favorite today.

I am us and we is me. ... bitches.
Find all posts by this user
Like Post Quote this message in a reply
31-10-2012, 10:22 PM
RE: Countdown the Election 2012
(27-10-2012 05:05 PM)Bucky Ball Wrote:  
(27-10-2012 11:56 AM)GirlyMan Wrote:  Nate Silver's got Obama as a 3:1 favorite as of last night.

WTF ?!
http://www.ukprogressive.co.uk/breaking-...20598.html

Meh....any pol who thinks he lost due to malfunctioning vote machines can call for hand recounts. Some states are retarded and use automated voting methods that do not originate with a paper ballot. One example of poor audit trails are a machine that prints a ticker tape of votes as punched into a computer. In my State, it's mostly optical scanned paper ballots that are very easy to count up and audit if ever there is a question of machine tampering--no hanging chads like the Florida fiasco.
Find all posts by this user
Like Post Quote this message in a reply
31-10-2012, 10:38 PM
RE: Countdown the Election 2012
(31-10-2012 09:19 PM)GirlyMan Wrote:  
(31-10-2012 09:10 PM)germanyt Wrote:  Lines are long because polling locations are fewer and far between. Republicans have done everything they can to reduce the earling voting numbers that heavily favor Democrats. I saw an article recently that said early voting in Florida was down like 70 percent.

My point is that if early voting lines are long in the People's Republic of Maryland and the Communist District of Columbia where it is a foregone conclusion that they will both go Obama and my vote won't make a difference, it indicates that Democratic voter turnout may not be as weak as you might think. ... Republican attempts at voter suppression notwithstanding.

Silver's model got Obama at close to a 4:1 favorite today.

I live in a deeply liberal community. All I need to see to realize the Dem vote will be much less than 2008 is the huge reduction in lawn signs, bumper stickers, and all around enthusiasm about voting Obama this time around. Four years ago, the community was all excited about the election. This year, you can hardly find an Obama sign buried among the Halloween decorations. Meanwhile, driving through the deeply conservative areas of the state, I find large numbers of Romney/Ryan signs.

Turnout matters. All the polls, even those showing Obama winning, show GOP voters more motivated to vote this time. All the polls, even most showing Obama winning, show Romney winning self identified independents. The only way this result can happen in polling is if the polls are using turnout models that look like the last presidential election. Problem with adjusting/re-weighting the data to that assumption is that the assumption could be wrong. Based on my anecdotal observations, and the enthusiasm gap favoring GOP that is consistent among all the polls, I do think Romney outperforms the polls significantly. I don't think it will be a landslide either way, but I think Romney has the edge.
Find all posts by this user
Like Post Quote this message in a reply
01-11-2012, 06:15 PM
RE: Countdown the Election 2012
(31-10-2012 10:38 PM)BryanS Wrote:  
(31-10-2012 09:19 PM)GirlyMan Wrote:  My point is that if early voting lines are long in the People's Republic of Maryland and the Communist District of Columbia where it is a foregone conclusion that they will both go Obama and my vote won't make a difference, it indicates that Democratic voter turnout may not be as weak as you might think. ... Republican attempts at voter suppression notwithstanding.

Silver's model got Obama at close to a 4:1 favorite today.

I live in a deeply liberal community. All I need to see to realize the Dem vote will be much less than 2008 is the huge reduction in lawn signs, bumper stickers, and all around enthusiasm about voting Obama this time around. Four years ago, the community was all excited about the election. This year, you can hardly find an Obama sign buried among the Halloween decorations. Meanwhile, driving through the deeply conservative areas of the state, I find large numbers of Romney/Ryan signs.

Turnout matters. All the polls, even those showing Obama winning, show GOP voters more motivated to vote this time. All the polls, even most showing Obama winning, show Romney winning self identified independents. The only way this result can happen in polling is if the polls are using turnout models that look like the last presidential election. Problem with adjusting/re-weighting the data to that assumption is that the assumption could be wrong. Based on my anecdotal observations, and the enthusiasm gap favoring GOP that is consistent among all the polls, I do think Romney outperforms the polls significantly. I don't think it will be a landslide either way, but I think Romney has the edge.

The bookmakers agree with Silver's model and have Obama as a 4:1 favorite as well.

Here's a nice analysis of the differences between the aggregated state polls and the national polls by Silver. Oct. 30: What State Polls Suggest About the National Popular Vote

"Suppose, for example, that you take the consensus forecast in each state. (By 'consensus' I just mean: the average of the different forecasts.) Then you weigh it based on what each state’s share of the overall turnout was in 2008, in order to produce an estimate of the national popular vote.

Do the math, and you’ll find that this implies that Mr. Obama leads nationally by 1.9 percentage points — by no means a safe advantage, but still a better result for him than what the national polls suggest.

What if turnout doesn’t look like it did in 2008? Instead, what if the share of the votes that each state contributed was the same as in 2004, a better Republican year?

That doesn’t help to break the discord between state and national polls, unfortunately. Mr. Obama would lead by two percentage points in the consensus forecast weighing the states by their 2004 turnout.

Or we can weigh the states by their turnout in 2010, a very good Republican year. But that doesn’t help, either: instead, Mr. Obama leads by 2.1 percentage points based on this method.

...

Whether the state polls or the national polls characterize the election correctly could well determine its outcome."

I am us and we is me. ... bitches.
Find all posts by this user
Like Post Quote this message in a reply
01-11-2012, 08:14 PM
RE: Countdown the Election 2012
(01-11-2012 06:15 PM)GirlyMan Wrote:  
(31-10-2012 10:38 PM)BryanS Wrote:  I live in a deeply liberal community. All I need to see to realize the Dem vote will be much less than 2008 is the huge reduction in lawn signs, bumper stickers, and all around enthusiasm about voting Obama this time around. Four years ago, the community was all excited about the election. This year, you can hardly find an Obama sign buried among the Halloween decorations. Meanwhile, driving through the deeply conservative areas of the state, I find large numbers of Romney/Ryan signs.

Turnout matters. All the polls, even those showing Obama winning, show GOP voters more motivated to vote this time. All the polls, even most showing Obama winning, show Romney winning self identified independents. The only way this result can happen in polling is if the polls are using turnout models that look like the last presidential election. Problem with adjusting/re-weighting the data to that assumption is that the assumption could be wrong. Based on my anecdotal observations, and the enthusiasm gap favoring GOP that is consistent among all the polls, I do think Romney outperforms the polls significantly. I don't think it will be a landslide either way, but I think Romney has the edge.

The bookmakers agree with Silver's model and have Obama as a 4:1 favorite as well.

Here's a nice analysis of the differences between the aggregated state polls and the national polls by Silver. Oct. 30: What State Polls Suggest About the National Popular Vote

"Suppose, for example, that you take the consensus forecast in each state. (By 'consensus' I just mean: the average of the different forecasts.) Then you weigh it based on what each state’s share of the overall turnout was in 2008, in order to produce an estimate of the national popular vote.

Do the math, and you’ll find that this implies that Mr. Obama leads nationally by 1.9 percentage points — by no means a safe advantage, but still a better result for him than what the national polls suggest.

What if turnout doesn’t look like it did in 2008? Instead, what if the share of the votes that each state contributed was the same as in 2004, a better Republican year?

That doesn’t help to break the discord between state and national polls, unfortunately. Mr. Obama would lead by two percentage points in the consensus forecast weighing the states by their 2004 turnout.

Or we can weigh the states by their turnout in 2010, a very good Republican year. But that doesn’t help, either: instead, Mr. Obama leads by 2.1 percentage points based on this method.

...

Whether the state polls or the national polls characterize the election correctly could well determine its outcome."

The problem with using a consensus forecast is that simply averaging them is not a great way to do things when you have bimodal distribution of poll results, you cannot just average things to get a result. That kind of distribution is an indication of systemic errors in polling methods in general.

Nate Silver doesn't quite average, but he uses some opaque, unrevealed weighting of polls which tends to weight down GOP leaning polls and weight up DEM leaning polls. He is manipulating the result in a way that conveniently agrees with his politics. I only consider Nate Silver's analysis to be one data point, and in no way dispositive.

The way I see it, we won't really know who is right until election day. Nobody has this in the bag. I remain convinced in my state that flipping to Romney this year is possible, and that tells me the ground has changed significantly from 4 years ago. That, and the simple fact that polls almost always when they are wrong are wrong in a way that undercounts Republicans.
Find all posts by this user
Like Post Quote this message in a reply
01-11-2012, 09:04 PM (This post was last modified: 01-11-2012 09:30 PM by GirlyMan.)
RE: Countdown the Election 2012
(01-11-2012 08:14 PM)BryanS Wrote:  
(01-11-2012 06:15 PM)GirlyMan Wrote:  The bookmakers agree with Silver's model and have Obama as a 4:1 favorite as well.

Here's a nice analysis of the differences between the aggregated state polls and the national polls by Silver. Oct. 30: What State Polls Suggest About the National Popular Vote

"Suppose, for example, that you take the consensus forecast in each state. (By 'consensus' I just mean: the average of the different forecasts.) Then you weigh it based on what each state’s share of the overall turnout was in 2008, in order to produce an estimate of the national popular vote.

Do the math, and you’ll find that this implies that Mr. Obama leads nationally by 1.9 percentage points — by no means a safe advantage, but still a better result for him than what the national polls suggest.

What if turnout doesn’t look like it did in 2008? Instead, what if the share of the votes that each state contributed was the same as in 2004, a better Republican year?

That doesn’t help to break the discord between state and national polls, unfortunately. Mr. Obama would lead by two percentage points in the consensus forecast weighing the states by their 2004 turnout.

Or we can weigh the states by their turnout in 2010, a very good Republican year. But that doesn’t help, either: instead, Mr. Obama leads by 2.1 percentage points based on this method.

...

Whether the state polls or the national polls characterize the election correctly could well determine its outcome."

The problem with using a consensus forecast is that simply averaging them is not a great way to do things when you have bimodal distribution of poll results, ...

Here's a nice tight reasoned response to the bimodal survey error distribution hypothesis. Can We Trust the Polls?

(01-11-2012 08:14 PM)BryanS Wrote:  Nate Silver doesn't quite average, but he uses some opaque, unrevealed weighting of polls which tends to weight down GOP leaning polls and weight up DEM leaning polls. He is manipulating the result in a way that conveniently agrees with his politics.

And that would make him an unethical statistician, which I seriously doubt. If I thought this about all of the statistical models using aggregated State polling: FiveThirtyEight, ElectoralVote, Votamatic, HuffPost Pollster, RCP, TPM, and Princeton, which are all effectively in agreement with the bookies, I'd be tempted to pick one of the many bookies who also have Obama as a 4:1 favorite and bet the house on Romney (the one GirlyBoy lives in, not the one I live in, fucker's on his own if I lose). Then I'd own 3 more houses. Big Grin

'Course I don't think that. The bookies are almost always right. Or they cease being bookies quickly.

I am us and we is me. ... bitches.
Find all posts by this user
Like Post Quote this message in a reply
01-11-2012, 10:35 PM
RE: Countdown the Election 2012
Down here in Woof County we are hoping for a strong Democrat showing.Thumbsup
Find all posts by this user
Like Post Quote this message in a reply
Post Reply
Forum Jump: