Skip to main content

View Diary: Daily Kos Elections Live Digest: 1/29 (263 comments)

Comment Preferences

  •  I would trust exit polls (3+ / 0-)
    Recommended by:
    MichaelNY, wadingo, tommypaine

    More than any Gallup poll.

    "What do you mean "conspiracy"? Does that mean it's someone's imaginings and that the actual polls hovered right around the result?" - petral

    by conspiracy on Wed Jan 29, 2014 at 08:44:42 AM PST

    [ Parent ]

    •  Gallup is excrement till proven otherwise (1+ / 0-)
      Recommended by:
      MichaelNY

      And Gallup is week old excrement in the Latino west.

      /notice garbage poll then move on

      All the problems we face in the United States today can be traced to an unenlightened immigration policy on the part of the American Indian. -- Pat Paulsen

      by tommypaine on Wed Jan 29, 2014 at 01:46:18 PM PST

      [ Parent ]

      •  That's ridiculous. They were off in 2012 but (2+ / 0-)
        Recommended by:
        wwmiv, jncca

        it was in a consistent and predictable manner due to their likely voter model overestimating how white the electorate would be. A Gallup registered voter poll isn't garbage. Sure take it with a grain of salt if you like, but they're not Rasmussen.

        •  That's bizarrely naive (0+ / 0-)

          C'mon, bad methodology/judgment in one thing is a red flag in all things.

          The kid down the street who puts his tongue on a frozen lamppost is not the person to follow as he jumps off the cliff into the quarry.

          All the problems we face in the United States today can be traced to an unenlightened immigration policy on the part of the American Indian. -- Pat Paulsen

          by tommypaine on Wed Jan 29, 2014 at 02:08:51 PM PST

          [ Parent ]

          •  ... (0+ / 0-)

            No, it isn't naive. What you're arguing is bizarrely unscientific and unaware of basic terms like accuracy and precision.

            24 Burkean Post Modern Gay Democrat; NM-2 (Raised), TX-20 (B.A. & M.A. in Political Science), TX-17 (Home); 08/12 PVIs

            by wwmiv on Wed Jan 29, 2014 at 02:12:52 PM PST

            [ Parent ]

            •  You demonstrated previously you don't (0+ / 0-)

              know what those words mean, so you probably should avoid them.

              Still, it is funny to imagine that some people who got wretched food in a restaurant when ordering the enchiladas will go in the same restaurant and naively order the burritos.

              Bad in one thing does not necessarily mean bad in another thing, but it's very naïve to assume that bad in one thing does not relate at all to other things.

              All the problems we face in the United States today can be traced to an unenlightened immigration policy on the part of the American Indian. -- Pat Paulsen

              by tommypaine on Wed Jan 29, 2014 at 02:21:52 PM PST

              [ Parent ]

              •  ... previously demonstrated? (3+ / 0-)

                I'm pretty sure I previously demonstrated exactly what those things mean...

                Accuracy: close to the truth.
                Precision: close to - under unchanging conditions - what you said every other time.

                Gallup may not be the most accurate, but Gallup damn well has precision. Given we now know the truth we can also, at least going forward, be able to assume that they're about as far away from the truth in the same direction away from that truth as they used to be, because they have always had fairly good precision.

                Check yourself before you wreck yourself.

                24 Burkean Post Modern Gay Democrat; NM-2 (Raised), TX-20 (B.A. & M.A. in Political Science), TX-17 (Home); 08/12 PVIs

                by wwmiv on Wed Jan 29, 2014 at 02:24:49 PM PST

                [ Parent ]

                •  First, you should start here: (0+ / 0-)

                  https://en.wikipedia.org/...

                  Then read "because they have always had fairly good precision"

                  Then look up
                  http://www.thefreedictionary.com/...
                  Then ask yourself how you came to conclude "always".

                  Then consider why you think a pollster that averages two points more Republican is more reliable than a pollster that averages the correct result.

                  All the problems we face in the United States today can be traced to an unenlightened immigration policy on the part of the American Indian. -- Pat Paulsen

                  by tommypaine on Wed Jan 29, 2014 at 02:33:16 PM PST

                  [ Parent ]

                  •  First... (0+ / 0-)

                    You should start at the exact link you sent me and then re-read my comment, where I basically gave you the same definitions.

                    Then, you need to go back through my comments and note that I never once said that Gallup was more reliable than other pollsters. I simply backed up Stephen when he said that they aren't total garbage... because, well, they aren't total garbage.

                    Then, I'll ask you to note that a pollster who is consistently about two points more Republican than the truth is, well, what would be considered precise by the very link that you tried to squash me with (and failed).

                    Then, I'll ask that you realize that just because someone might be a bit off (2 points isn't that far off, I might add, and is well within MoE generally) doesn't impugn the rest of their work. That's a pretty high bar that you're setting to judge quality from... a bar that most pollsters would never pass.

                    24 Burkean Post Modern Gay Democrat; NM-2 (Raised), TX-20 (B.A. & M.A. in Political Science), TX-17 (Home); 08/12 PVIs

                    by wwmiv on Wed Jan 29, 2014 at 02:45:21 PM PST

                    [ Parent ]

                    •  The word you are looking for is "consistent" (0+ / 0-)

                      I can't and don't want to go on this endlessly, but the point I made yesterday was:

                      A pollster missing +2D in ten races and +2R in ten other races, is falling in the margin of error, and is averaging the correct result overall in the 20 races.  

                      A pollster that misses +2R all the time has a partisan house effect, and they have an average that is 2% off the correct result.

                      The first pollster is more reliable and their polling average is more accurate.  The second pollster is more consistent, and more precisely wrong.

                      If both pollsters poll a race 10 times, the average of the first pollster will be right on the money, and the average of the second pollster will average to be +2R.  If you know the house effect of the second pollster, both pollsters will give you the exact same information, but there is certainly no advantage to that.

                      The first pollster, with its average result being the exact correct result, would be a reliable pollster.  The second pollster always being +2R would be consistently wrong in its average, and consistently have a house effect, and thus could never be considered "more reliable" than the first pollster even if they are more consistent.

                      All the problems we face in the United States today can be traced to an unenlightened immigration policy on the part of the American Indian. -- Pat Paulsen

                      by tommypaine on Wed Jan 29, 2014 at 03:03:18 PM PST

                      [ Parent ]

                      •  Yet again... (1+ / 0-)
                        Recommended by:
                        Setsuna Mudo
                        The first pollster, with its average result being the exact correct result, would be a reliable pollster.  The second pollster always being +2R would be consistently wrong in its average, and consistently have a house effect, and thus could never be considered "more reliable" than the first pollster even if they are more consistent.
                        I never said that the second pollster would be more reliable. I don't know where you're getting that from. What I said was that having a +2R effect consistently (meaning +2R precision, which is the language you yourself use after debating me on this point to the contrary... for some reason... implicit acknowledgement that I'm correct? at least that's how I'm gonna take it) doesn't make the pollster itself worthless. Certainly it isn't worth as much as the best pollster. Nobody would ever argue that point. Honestly, I'd rank the various combination thusly:

                        1. accuracy and precision (a pollster who hits the true mark within MoE in every poll)
                        2. accuracy but not precision (i.e. fluctuating around the true point, but not hitting it consistently)
                        3. precision but not accuracy (something like Gallup, where they consistently pull results slightly to the right of the "true" point)
                        4. neither accuracy nor precision (a pollster who misses the truth by wide amounts in no consistent pattern with respect to partisanship).

                        24 Burkean Post Modern Gay Democrat; NM-2 (Raised), TX-20 (B.A. & M.A. in Political Science), TX-17 (Home); 08/12 PVIs

                        by wwmiv on Wed Jan 29, 2014 at 04:54:00 PM PST

                        [ Parent ]

        •  In retrospect I should not have said excrement (1+ / 0-)
          Recommended by:
          MichaelNY

          That was basically using an imprecise word that overstates things.

          The point though is Gallup should be viewed with skepticism because their recent history involves very poor methodology choices.  They might be right here.  They might not.  They should have to prove themselves before they are trusted again though.

          All the problems we face in the United States today can be traced to an unenlightened immigration policy on the part of the American Indian. -- Pat Paulsen

          by tommypaine on Wed Jan 29, 2014 at 02:46:12 PM PST

          [ Parent ]

        •  I will note that this: (1+ / 0-)
          Recommended by:
          MichaelNY
          They were off in 2012 but
          it was in a consistent and predictable manner due to their likely voter model overestimating how white the electorate would be.
          ... basically translates to "precise, but not accurate".

          24 Burkean Post Modern Gay Democrat; NM-2 (Raised), TX-20 (B.A. & M.A. in Political Science), TX-17 (Home); 08/12 PVIs

          by wwmiv on Wed Jan 29, 2014 at 02:48:26 PM PST

          [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site