| Submit your comments on this article |
| Home Front: Politix |
| Poll finds GOP got 34 point bump up in Alaska |
| 2008-09-07 |
| Posted by:tipper |
| #6 This poll may actually understate things, it's of registered Alaska voters for one thing. For another, even before the VP pick Alaska was a red state and was expected to vote McCain in November. Yet this poll has McCain's initial numbers starting as DOWN by 7%. Now up by 27%. |
| Posted by: Minister of funny walks 2008-09-07 12:58 |
| #5 What I think the blogsphere needs is a statstics expert to run down things. Such as how many people are needed for a poll to be stastically viable. And then what "Likely voter" and Registered voter" mean and why each is used to massage outcomes in one direction or another. Then convince a lot of bloggers to link to that article/blog to make sure it rises i google ratings and otherwise gets notice. |
| Posted by: rjschwarz 2008-09-07 12:30 |
| #4 Locally, our (R) congressional candidate was always up to 8 points behind in the polls till election day. Then when the official results are completed, the candidate always won. The last time the difference was 9 points, but after 2 weeks of counting ballots and threaten lawsuits [on method of counting], squeaked by with just over 200 votes.* That's just local, but significant bias in the process. Unfortunately, the candidate tried for a vacated higher office this time around and didn't make it through the party's own primary. The pollsters play a confidence game with the people who pay them. They know what the buyer wants to hear [sound familiar?]. So they play loose with the numbers. Would you go back if they told you, you were toast early in the game? However, to cover their commercial asses, they play 'tighten the numbers' as the day draws closer. It's the proverbial Lucy and Charlie Brown game with kicking the ball. Why do they keep going back. As mentioned they hear what they want to hear. Also the hope that the repeated and constant bombardment of the numbers convinces people to 'join the bandwagon'/back the strongest horse. *The Donk administered Board of Elections intentionally shorted the polling stations in red sub-districts and increased blue sub-districts holding of the mark sense forms election ballots used in the election. It wasn't by a small margin of statistical error either. |
| Posted by: Procopius2k 2008-09-07 11:33 |
| #3 the Anchorage-based polling firm I find this very interesting! The reason being that this is an independent firm that is outside the taint of the bigger national polling firms. There is a consistent pattern among the known polling firms such as CNN, Gallup, Rasmussen etc. etc. (including Real Clear Politic) which ALWAYS show the democrat ahead at this point and yet their numbers always tighten about a two weeks before the election increasing the odds of the republican candidate. And just days before the election we finally see some numbers that actually resemble election results while still tilting slightly towards the democrat. What I love to see is some intelligent blogger do a study of how many points we can expect the polls to favor the liberal candidate at certain points through out the race. This trend has been happening for probably at least eight years now, so there is plenty of data available to do the study. For example, if CNN and Gallup say at this point the dem candidate is up by 2, then when we get to the actual results of the race, how many points should we deduct from that candidate knowing that these polls have been consistently off by a factor of X and will tighten as the election day nears? |
| Posted by: Betty Grating2215 2008-09-07 10:27 |
| #2 Sorry, try this link. |
| Posted by: tipper 2008-09-07 08:20 |
| #1 Link title and article linked do not match. But the article linked is interesting. |
| Posted by: Herman Anginemp6730 2008-09-07 07:54 |