Hey look! More skewed polls! Turd Sanford wins by 9

[OCinBuffalo]

http://www.guardian.co.uk/commentisfree/2013/may/08/mark-sanford-win-south-carolina-bad-polling

 

Here we go again. This is going to be fun. Once again we have people who can't weight their demographics properly telling us that this Sanford turd, and he is a turd, was going to lose by 9, 2 weeks before the election. Trouble is: Sanford won by 9. That's a massively skewed poll, and, importantly, this time it comes from both right and left leaning pollsters.

 

Specific reason for PPP failure:

As my friend Mark Blumenthal points out, this first PPP survey had far too many African Americans as a percentage of the electorate. I don't doubt that some white voters, a mostly Republican demographic in South Carolina's first district, were disenchanted with Sanford by allegations that he violated the terms of his divorce, but the difference in the percentage of black voters was too great. Colbert Busch never had a lead of 9pt. One might wonder whether she even had a lead.

Specific reason for RRH failure:

It wasn't a surprise, therefore, that the RRH poll had women as too great a percentage of the electorate, at 60% versus the about 55% it should have been.

Given the 2012 electorate, we would think these assumptions would be valid. But really, this is just more evidence that 2012 was an outlier. Even with turd-boy Sanford, the black and women vote was overstated = an 18 point swing. It wasn't until a few days before that these guys had the D up by 1. Still wrong by 10 pts.

 

The worst part? This time the internals couldn't even get it right. They had the D up by 2-3. So I will say it again: we cannot trust anybody's polls right now, because as you can plainly see, it's more about getting lucky and guessing at the makeup of the electorate, rather than scientifically proving it. Somebody has a lot of work to do(ahem, Gallup) on getting a proper demographic model together, so that polls can be weighted correctly. I strongly suggest that they start by throwing out the 2012 data.

 

This race tells us nothing about what is to come. This is simply what happens when you try to elect Nancy Pelosi 2 in South Carolina. And, a Nancy Pelosi 2 that was an even worse public speaker Nancy Pelosi. (Apparently, yes, that is possible. It's the only truly shocking thing to come of this. )

 

There was never any way Sanford was down by 9, but, if you listened to these polls, and are a marginal R voter, there's a chance you stayed home, and that may also be part of why Sanford only won by 9, and not 13+ too. That is my chief concern regarding all of this. Polls are supposed to give us the score, not create the score.

[Bigfatbillsfan]

http://www.guardian....ina-bad-polling

 

Here we go again. This is going to be fun. Once again we have people who can't weight their demographics properly telling us that this Sanford turd, and he is a turd, was going to lose by 9, 2 weeks before the election. Trouble is: Sanford won by 9. That's a massively skewed poll, and, importantly, this time it comes from both right and left leaning pollsters.

 

Specific reason for PPP failure:

 

Specific reason for RRH failure:

 

Given the 2012 electorate, we would think these assumptions would be valid. But really, this is just more evidence that 2012 was an outlier. Even with turd-boy Sanford, the black and women vote was overstated = an 18 point swing. It wasn't until a few days before that these guys had the D up by 1. Still wrong by 10 pts.

 

The worst part? This time the internals couldn't even get it right. They had the D up by 2-3. So I will say it again: we cannot trust anybody's polls right now, because as you can plainly see, it's more about getting lucky and guessing at the makeup of the electorate, rather than scientifically proving it. Somebody has a lot of work to do(ahem, Gallup) on getting a proper demographic model together, so that polls can be weighted correctly. I strongly suggest that they start by throwing out the 2012 data.

 

This race tells us nothing about what is to come. This is simply what happens when you try to elect Nancy Pelosi 2 in South Carolina. And, a Nancy Pelosi 2 that was an even worse public speaker Nancy Pelosi. (Apparently, yes, that is possible. It's the only truly shocking thing to come of this. )

 

There was never any way Sanford was down by 9, but, if you listened to these polls, and are a marginal R voter, there's a chance you stayed home, and that may also be part of why Sanford only won by 9, and not 13+ too. That is my chief concern regarding all of this. Polls are supposed to give us the score, not create the score.

 

But the polls are obviously not creating the score. If Pelosi II had won and in the aftermath we find that the only reason she did win was because Marginal R voters stayed home believing their vote was meaningless due to the polls showing that Sanford was going to loose we could argue that the polls are creating the score.

 

This is a good example of the how you can't skew a poll to manipulate the outcome of an election.

[OCinBuffalo]

But the polls are obviously not creating the score. If Pelosi II had won and in the aftermath we find that the only reason she did win was because Marginal R voters stayed home believing their vote was meaningless due to the polls showing that Sanford was going to loose we could argue that the polls are creating the score.

 

This is a good example of the how you can't skew a poll to manipulate the outcome of an election.

Nope. We could easily argue that the +9, is only 9 and not 11, because a significant portion of the R electorate, in a special election, whose turnout historically isn't that great to begin with, stayed home because they thought it was "over". There's no way to know, unless you did a poll now that tried to find people who didn't vote and stayed home for that specific reason. Good luck. Nobody will pay for that, because nobody cares now, so it won't get done.

 

Second, Bob Bradley, Washington Post Editor, is on record as specifically and intentionally skewing polls, and he did affect the outcome of Reagan's 1984 election as a result. Reagan would have won all 50 states, but diverted money away from MN because he believed, actually, Nancy believed, Bradley's bogus polls, and that money was moved to CA instead.

 

History is your friend. There is your on-record example of what you say can't happen. I refuse to do your reading for you. Google is also your friend.

 

Just because she didn't win, doesn't mean the polls had 0 effect. Nate Silver has a pretty good write up http://fivethirtyeig...dals-revisited/ and he lays it all on the sex scandal, and that it jives with an earlier study. But, that can't be the only reason.

 

That article begs so many questions:

1. How come the Pelosi extra strength crappiness of the D candidate wasn't even mentioned?

2. Can't help but notice that Silver uses PPP regularly, even with his "house rules", but decided to "caution" us against these polls this time around? Why? Should we expect him to remove these pollsters from his group next time? Why not?

3. How come he never talks about the obvious: over-representation of demographics?

Edited May 10, 2013 by OCinBuffalo

[Doc]

I still can't believe he won. And he beat out Stephen Colbert's sister, no less.

[KD in CA]

I still can't believe he won. And he beat out Stephen Colbert's sister, no less.

 

Imagine that; someone beating the sister of a second rate TV comedian for a seat in Congress. What is happening to our democracy?!?

[IDBillzFan]

I still can't believe he won. And he beat out Stephen Colbert's sister, no less.

 

Her entire campaign was "I have a famous brother, I refuse to debate, Argentina." Romney won that district by 16. By the time they rolled our her mugshot, she would have lost to a turtle.

[Magox]

It really isn't that surprising, Romney won the district by 18 in a heavy black Democratic Turnout year.

[OCinBuffalo]

It really isn't that surprising, Romney won the district by 18 in a heavy black Democratic Turnout year.

 

The outcome? No. The polls that said she was up by 9, with 2 weeks to go?

 

How can we classify those as anything but "surprising"?

 

Edit: But for me? Not surprising either, really. There are either significant methodology problems with out polling process, or, 2012 was an outlier, or worse: both.

Edited May 10, 2013 by OCinBuffalo