September 17, 2004
A general thought
By Jim Dallas
Atrios continues the attack on Gallup and NYT/CBS, noting that the 2000 choice of the NYT sample is something like 36-28 Bush over Gore. That would be a really big problem, if the respondents were being honest.
However, I'd add on a healthy +/- 5 to whatever the margin of error is, because it's a proven fact that people lie to pollsters, especially as it regards who they voted for in the last election.
So, looking at these Bush-Gore stats, the CBS/NYT poll may be unrepresentative, but then again, it may not.
I'm bullish on Kerry, but I think it would be a bad idea not to think that we are potentially down 14 points right now.
Coming of age in Texas at the turn of the century, my general advice has to be, "be prepared to lose.... but go down fighting."
Posted by Jim Dallas at September 17, 2004 11:20 PM
| TrackBack
Byron maybe you could ask 'Atrios' how the final results of the Gallup poll of Bush +2 LV and -1 RV are unrepresentative of the general election.
Makes you dream....IF we only had another week...President ALGORE...BZZZZZZZZZZZ.
Wake up! You want to change the polls, get to work. i forgot you've only got Kerry to work with.
Where is Howard Dean when you need him??? Doing a talk show.
Where's Waldo i mean Edwards??? Hadn't glimpsed him lately. Maybe he's in an undisclosed location practicing.
After the point where you confused me with Byron, I totally lost your train of thought.
On Jim's point about people lying to pollsters is the polls taken in 2001. In polls taken on W's job approval ratings in the spring and summer of 2001, the pollsters also asked questions to the effect of "Who did you vote for in the last election" because they wanted context for their "do you think Bush was legitimately elected" questions. Up until 9/11, the polls after the election showed respondents evenly split on who they said they had voted for.
Then 9/11 happened. Pollsters took similar polls. And suddenly 60 - 65 % of the voters were saying that they had voted for Bush over Gore in 2000. Obviously Bush didn't win by that much, so about 15% of the respondents were being less than accurate in their reccollections.
Aladextra
Sorry Jim
My point is, Gallup polled 10/26/2000 a 13 point gap in the 2000 race. The final poll released indicated Bush up 2 likely voters (LV) and down 1 registered voters (RV). How is the final poll not representative of the general election outcome? The next thought is to explain the difference. My memory tells me that Dems made gains in the House and Senate that year. ALGORE worked his butt off in the final two weeks of the campaign to reel in Guv. Bush. He knew how much work he had done and when the results did match he felt he had to contest the election in Florida.
Atrios's inference seems to be to show how far off Gallup was, not how much work ALGORE put in.
The poll was where the race was on 10/26. Just like this poll is indicating where the race is NOW. ALL it means is its time to get to work and change the numbers, get your message out. If America sides with your message, the poll numbers will change.
