Welcome to PatsFans.com

POLITICS: On Polling Models, Skewed & Unskewed

Discussion in 'Political Discussion' started by patsfan13, Nov 2, 2012.

  1. patsfan13

    patsfan13 Hall of Fame Poster PatsFans.com Supporter

    Joined:
    Jan 4, 2005
    Messages:
    24,726
    Likes Received:
    73
    Ratings:
    +154 / 7 / -13

    Good article from Baseball Crank (sabre metrics guy) on the mechanics of polls ans the mathematical models that they are based on. Good read.

    Baseball Crank: POLITICS: On Polling Models, Skewed & Unskewed


    This guy understands the game that the media plays with the polls (which I bet many of the reporters don't understand.

    Next Tuesday will show which understanding of the polls is correct which polls were accurate and which were horrible wrong.
  2. patsfan13

    patsfan13 Hall of Fame Poster PatsFans.com Supporter

    Joined:
    Jan 4, 2005
    Messages:
    24,726
    Likes Received:
    73
    Ratings:
    +154 / 7 / -13

    BTW Romney is campaigning in PA this weekend.
  3. Patsfanin Philly

    Patsfanin Philly Rookie

    Joined:
    Jan 16, 2005
    Messages:
    6,717
    Likes Received:
    14
    Ratings:
    +36 / 0 / -0

    #95 Jersey

    Almost literally in my backyard on Sunday. Darryl, Patters I can get you tickets if you'd like.......

    It is interesting to see the campaigns. I saw Jimmy Carter at Boston College back in 1976 and McCain in 2008....

    Either it is a head fake by Romney to get Obama to spend time and resources in a state that is safely blue or Romney thinks he has a shot.....
    Last edited: Nov 2, 2012
  4. Real World

    Real World Moderator Staff Member

    Joined:
    Aug 15, 2006
    Messages:
    26,953
    Likes Received:
    179
    Ratings:
    +402 / 5 / -2

    I love numbers. Math and it's related subjects were always my favorite in school, and especially in college. I'll have to give the article a read.

    BTW, being a sabermetrics follower, and numbers type, to date I have no idea who is going to win this election. There are so many differing pols, and finite analysis, that I honestly can't predict a winner. Sample sizes are all over the place, and in some ways so close, that I don't think anyone can safely say who will win.
  5. Drewski

    Drewski Rookie

    Joined:
    Jul 25, 2007
    Messages:
    3,816
    Likes Received:
    41
    Ratings:
    +87 / 0 / -1

    No Jersey Selected

    Thanks for the link 13. As a student of stats and Econometrics during my college career I found it an interesting read.
  6. patsfan13

    patsfan13 Hall of Fame Poster PatsFans.com Supporter

    Joined:
    Jan 4, 2005
    Messages:
    24,726
    Likes Received:
    73
    Ratings:
    +154 / 7 / -13

    I am always fascinated by the assumptions and underlying assumptions.

    Some other interesting sites and reads for those interested:

    Jay Cost and his reading of the poll internals:

    Morning Jay: Why Romney Is Likely to Win | The Weekly Standard

    This site is analyzing the poll internals and playing with different turnout models. They show the internals of the Ohio polls the raw data on the independents is especially interesting:

    Rebalanced Presidential Election Polls 2012


    His methodology is similar to what I have been doing, mine is simplier and I use a even R-d turnout model, he modifies the turnout model based on the latest Gallup Rasmussen party affiliation numbers.
  7. Patradomous

    Patradomous Rookie

    Joined:
    Nov 20, 2007
    Messages:
    3,047
    Likes Received:
    41
    Ratings:
    +120 / 0 / -9

    #87 Jersey

    that nate silverman scares me
  8. patsfan13

    patsfan13 Hall of Fame Poster PatsFans.com Supporter

    Joined:
    Jan 4, 2005
    Messages:
    24,726
    Likes Received:
    73
    Ratings:
    +154 / 7 / -13



    The rumor is that he had access to the Obama internals in 2010. In any event his credibility is on the line in this election.

    He got a good gig flaking/spinning polls for the Obama campaign at the NYT's, probably making 6 figures.
  9. IcyPatriot

    IcyPatriot ------------- PatsFans.com Supporter

    Joined:
    Sep 13, 2004
    Messages:
    38,651
    Likes Received:
    370
    Ratings:
    +790 / 6 / -15

    #87 Jersey


    Math was my minor in college but the funny thing is I have zero interest in polls ... hate them - too much work. I go by the desperation theory ... I see what the campaigns are doing the 2 weeks before election and it tells you everything you need to know - their poll numbers are the best there is - they have to be.
  10. patsfan13

    patsfan13 Hall of Fame Poster PatsFans.com Supporter

    Joined:
    Jan 4, 2005
    Messages:
    24,726
    Likes Received:
    73
    Ratings:
    +154 / 7 / -13


    Spot on.

    The information they have is far different than the top line numbers reported by the press.
  11. Real World

    Real World Moderator Staff Member

    Joined:
    Aug 15, 2006
    Messages:
    26,953
    Likes Received:
    179
    Ratings:
    +402 / 5 / -2


    I believe he has Obama at an 80% probability of winning. If I were a Romney supporter i'd be nervous about his prediction to. The thing with Silver is that his previous prediction was based on Obama's internal polling. He had access to it. I'd venture to guess he has access to it now to. Why wouldn't they give it to him the second time around?

    Silver weights polls, and of the criticisms I've seen of his model, his weighting system will apparently be his undoing, if he's indeed incorrect. We'll likely find out early Wednesday morning I guess.

    What really strikes me this election, is all the different election prognosticators (sp?) on either side, or in the middle even, who are predicting different outcomes. For example Barone predicts Romney, and silver Obama. Both have either long standing reputations of knowing election dynamics, or of predicting an outcome. It confuses me even more when i try to judge for myself what the outcome might be. I honestly have no clue. Almost any outcome wouldn't surprise me, save for a blowout by either side. I could see one side winning 300+ delegates, but by small margins in the states actual vote, but I doubt we see a 7 point win persay across the main board.

    So what does your "Desparation Theory" tell you right now?


    I try to do the same, but I can't figure out if Romney going to PA is cuz he thinks he can win it, and has Ohio in his opinion, or if he thinks Ohio is a tough task, and he needs a PA instead. A PA or a combination of Wisconsin & Iowa among others. I honestly don't know. If I had to bet, I'd say it was the latter. All the surface polls I see show Obama up in most of those states by anywhere from 1-6 points. I say surface polls cuz when the internals of those polls are disected by some pundits, the numbers supposedly look different. Sampling D+9 and getting a 4 point Obama lead renders a specific polls surface number questionable. This is honestly the most unpredictable election I can remember.
  12. Patradomous

    Patradomous Rookie

    Joined:
    Nov 20, 2007
    Messages:
    3,047
    Likes Received:
    41
    Ratings:
    +120 / 0 / -9

    #87 Jersey

    Saw some figures yesterday comparing figures from 2008 and past historical early voting patterns in Ohio.Comparing 2008 to this year is about a 250,000 vote swing for Romney. And that was the margin of victory for Obama in 2008,250k votes.
    Mcain in '08 won the election day vote. Now if this pattern holds up, Ohio shouldn't be close.
    Rience Priebus has been very giddy on what he sees with the early returns thus far.
  13. Real World

    Real World Moderator Staff Member

    Joined:
    Aug 15, 2006
    Messages:
    26,953
    Likes Received:
    179
    Ratings:
    +402 / 5 / -2

    Well, he'd look giddy publically no matter what. He's not going to show anything but a positive confidence regardless of the numbers. The same applies to Obama's team. Happy faces period.
  14. patsfan13

    patsfan13 Hall of Fame Poster PatsFans.com Supporter

    Joined:
    Jan 4, 2005
    Messages:
    24,726
    Likes Received:
    73
    Ratings:
    +154 / 7 / -13

    The Ohio polling show Romney up 11% among indies compared to Obama winning indies by 7% in 2008 and winning the state by 5%.

    An 18 pt swing in the indie vote translates to Romney by 4-6%.


    Susquehanna polling has Romney up 4% in Pa today.
  15. patsfan13

    patsfan13 Hall of Fame Poster PatsFans.com Supporter

    Joined:
    Jan 4, 2005
    Messages:
    24,726
    Likes Received:
    73
    Ratings:
    +154 / 7 / -13

    BTW Gallup may release 1 more poll but Sandra knocked them out of the box this week.

Share This Page

unset ($sidebar_block_show); ?>