Author Topic: FDR in 1936 and 2012  (Read 5918 times)

MCWAY

  • Getbig V
  • *****
  • Posts: 19227
  • Getbig!
Re: FDR in 1936 and 2012
« Reply #50 on: October 06, 2012, 07:11:50 PM »
Yes, that is funny as well. Though I think the "creaming" was just Romney being a better actor, which is something that probably doesn't indicate much about how he would perform once in office.

1. OK. Let's throw out the results from all organizations that oversample Dems -- if indeed that's what's happening -- by only looking at the results from an organization that definitely doesn't do that, since it oversamples Republicans instead: Rasmussen.

I've noticed that you focus on national polls in your posts; that would be fine if this country were a democracy. As it happens, the candidate who wins does so by winning electoral votes, not the popular vote. We can use Rasmussen's Ohio polls as a metric for Willard's chances since no Republican has won the White House without winning Ohio.

2. Rasmussen's latest poll -- which accounts for any potential effects of the first debate -- indicates Obama is ahead by 1% in Ohio, with a margin of error of 4.5%. It is a common misperception that since the difference between the candidates is within the margin of error, the candidates are 'essentially tied.' But this isn't true: given the margin of error and difference between the candidates, we can infer the probability that the candidate the poll says is ahead really is ahead. This table captures such probabilities:



Since our particular margin of error is 4.5%, there is a 59% chance Obama really is ahead and would win Ohio -- and thus the White House -- if the election were held today. The conclusion that Willard is in poor shape is doubly reinforced by three more facts: Rasmussen uses likely voters and not registered voters, the probability above is as high as it is for an organization that oversamples Republicans, meaning the real probability is higher, and market data -- which as I've said repeatedly is more accurate than polling data -- indicates a 64% chance of an Obama victory.

3. So that's what happens when we remove any potential effects of Dem oversampling. Thanks.

I wasn't just referring to national polls. The state-by-state polls are just as guilty of oversampling Dems, some by double digits. But, don't take my word for it:

After Wednesday night’s smashing debate victory for Romney, we may expect the national and swing state polls to change in the Republican’s direction. But not by as much as they should. These polls are biased in favor of Obama and here’s the data to prove it:

From noted Republican pollster John McLaughlin comes a clear and convincing exposé of the bias of media polls in the swing states of Florida, Ohio, and Virginia.

McLaughlin reviewed exit polls in each state for the past four elections. From this data about who actually voted, he found that the party divisions manifest on election day have little to do with the samples upon which the media is basing its polling. And, coincidentally, it is always the Republican vote that tends to be undercounted.

In Florida, for example, McLaughlin finds that the average of the last four elections produced a turnout of 37% Democrats and 38% Republicans. But here is the partisan distribution of the most recent Florida media polls:

9-26: CBS/NY Times = 36% Dem / 27% Rep

9-23: Wash Post = 35% Dem / 29% Rep

So the media polls reflect a 9 point and six point Democratic edge even though the actual experience of the past four elections has been a 1 point Republican advantage.

Things are no better in Ohio. Here, McLaughlin finds a 2 point Democratic edge in the past four elections (38% Dem, 36% Rep). But the media polls show vastly more Democrats and fewer Republicans in their samples:

9-26: CBS/NY Times = 35% Dem / 26% Rep

9-23: Wash Post = 35% Dem / 27% Rep

9-11: NBC/Wall St Journal = 38% Dem / 28% Rep

Once again, the actual exit poll-measured vote in Ohio shows a 2 point Democratic edge, but the polls reflect Democratic advantages of 9 points, 8 points, and 10 points respectively.

In Virginia, it’s the same story. The last four elections have a combined 1 point Republican edge, 37-36. But the media polls show a big pro-Democratic bias:

10/2: Roanoke College = Democrat 36% / Republican 27%

9/17: CBS/NYTimes = Democrat 35% / Republican 26%

9/16: Washington Post = Democrat 35% / Republican 24%

9/11: NBC/Wall St Journal = Democrat 31% / Republican 26%

So instead of showing a 1 point Republican edge, these media poll samples show Democratic advantages of 9,9,11, and 5.

The correct conclusion to draw from all these polls is that Romney is comfortably ahead in Virginia and Florida while he holds a slight lead in Ohio. And, remember these polls are all pre-debate!

Also, bear in mind that the undecided vote in all of these polls usually goes against the incumbent.

That’s the real story.


http://www.dickmorris.com/swing-state-polls-are-rigged/

syntaxmachine

  • Getbig IV
  • ****
  • Posts: 2687
Re: FDR in 1936 and 2012
« Reply #51 on: October 07, 2012, 06:38:57 PM »
1. Strictly speaking, we won't know whether such numbers constitute "oversampling" until the election is over and we have the real numbers to compare them with. At best, we can assert that there is some probability that polls have overweighted Dem registration/turnout. This assertion may turn out to be incorrect.

2. Even if Dick is correct, it doesn't change the fact that Willard is in trouble even when we correct for potential Dem oversampling, as I did in the above post. This isn't brain science: I used polling that oversamples Republicans and the result is still a high probability for a second Obama administration.

3. Let's assume that it's true that undecideds tend to 'break toward the challenger' (it isn't obvious that this is so -- what may be happening is simple "regression toward the mean"). I've already established that the race in Ohio will decide the election. And according to Rasmussen, only 1% of likely voters in that state are undecided, with the rest of them split 50/49 in favor of Obama. This means that Willard would not only have to capture 100% of undecideds; he'd also need to poach voters from Obama.

In other words, having a literally flawless performance with undecideds in Ohio is necessary but not sufficient for Willard's goals. And since such a flawless performance is literally impossible -- no candidate ever acquires 100% of the undecided vote -- Willard is still screwed. This isn't due to anything about Willard; as everyone knows, running against an incumbent is difficult, and this year in particular was going to present an intimidating electoral college situation for any challenger. There's a reason the Republican heavy hitters sat this cycle out!

4. With that said, the election will likely grow tighter, which will afford you more opportunities to copy and paste Dick articles (I have no idea why anyone would be a fan of such a man). After 6 November, however, you'll hopefully come to understand what I've been telling you all this time.

MCWAY

  • Getbig V
  • *****
  • Posts: 19227
  • Getbig!
Re: FDR in 1936 and 2012
« Reply #52 on: October 07, 2012, 07:07:56 PM »
1. Strictly speaking, we won't know whether such numbers constitute "oversampling" until the election is over and we have the real numbers to compare them with. At best, we can assert that there is some probability that polls have overweighted Dem registration/turnout. This assertion may turn out to be incorrect.

2. Even if Dick is correct, it doesn't change the fact that Willard is in trouble even when we correct for potential Dem oversampling, as I did in the above post. This isn't brain science: I used polling that oversamples Republicans and the result is still a high probability for a second Obama administration.

3. Let's assume that it's true that undecideds tend to 'break toward the challenger' (it isn't obvious that this is so -- what may be happening is simple "regression toward the mean"). I've already established that the race in Ohio will decide the election. And according to Rasmussen, only 1% of likely voters in that state are undecided, with the rest of them split 50/49 in favor of Obama. This means that Willard would not only have to capture 100% of undecideds; he'd also need to poach voters from Obama.

In other words, having a literally flawless performance with undecideds in Ohio is necessary but not sufficient for Willard's goals. And since such a flawless performance is literally impossible -- no candidate ever acquires 100% of the undecided vote -- Willard is still screwed. This isn't due to anything about Willard; as everyone knows, running against an incumbent is difficult, and this year in particular was going to present an intimidating electoral college situation for any challenger. There's a reason the Republican heavy hitters sat this cycle out!

4. With that said, the election will likely grow tighter, which will afford you more opportunities to copy and paste Dick articles (I have no idea why anyone would be a fan of such a man). After 6 November, however, you'll hopefully come to understand what I've been telling you all this time.

Romney doesn't need to get 100% of the undecided voters to beat Obama. As for the heavy hitters sitting out, that doesn't make much sense. Romney was/is one such heavy hitter. This warped idea that you have that people are somehow afraid of Obama is without merit.

Running against an incumbent with a lousy record hardly strikes fear into anyone. And there's nothing intimidating about the electoral college. There's basically no state that McCain won in 2008 that looks to flip blue. It's all about the swing states. While Romney has to win more of them than Obama does, as the polls suggest, that's hardly a lock for the president at this point.

If he fails to come close to the turnout he had in 2008 (blacks and young voters may be shaky, at this point), he could lose and lose big.

syntaxmachine

  • Getbig IV
  • ****
  • Posts: 2687
Re: FDR in 1936 and 2012
« Reply #53 on: October 13, 2012, 05:42:56 PM »

Romney doesn't need to get 100% of the undecided voters to beat Obama.


My logic pertains to the poll I used an example, viz., the Rasmussen poll indicating a 50-49 lead for Obama in Ohio. This means that only 1% of voters there remain undecided and that by extension Romney will have to win that entire 1% just to tie Obama. In other words, if the election were held today, Romney could not win (given that the poll is accurate and that Ohio is necessary but not sufficient for a Republican victory).


As for the heavy hitters sitting out, that doesn't make much sense. Romney was/is one such heavy hitter.


We'll simply have to agree to disagree on this matter. Given the current status of the GOP,  I for one do not think that a one-term, pro-life governor counts as a 'heavy hitter', especially not when there are more popular politicians on offer (e.g., Christie). Despite a historically vulnerable incumbent, we had a Republican primary with Herman Cain leading the polls for weeks. The best explanation for this fact needs to cite a lack of heavy hitters, or so it seems to me.


This warped idea that you have that people are somehow afraid of Obama is without merit.


It isn't my idea, for I said no such thing. What I did say is that the electoral college situation is 'intimidating' in that it presents challenging prospects for a Republican victory, regardless of the candidate.
 

And there's nothing intimidating about the electoral college. There's basically no state that McCain won in 2008 that looks to flip blue. It's all about the swing states. While Romney has to win more of them than Obama does, as the polls suggest, that's hardly a lock for the president at this point.


1. McCain didn't win very many states, so the fact that none of them are liable to flip blue doesn't affect the proposition that this election cycle marks a challenging race for Republicans.

2. Yes, the swing states will decide the election. The same swing states Obama has had grassroots organizations operative in for several years and where his numbers have been strong for a similar amount of time (this has of course changed in several of them, but the current topic is what the prospects were for a Republican victory at the start of the cycle -- what a Republican deciding to run in the primary would be looking at -- and not the current prospects). Incumbent presidents win 80% of the time as is; the above facts would only reinforce the tendency of a qualified candidate to wait this cycle out.