Posted on 10/16/2004 8:58:03 PM PDT by Jewels1091
This site will show every poll and when it was taken.
(Excerpt) Read more at realclearpolitics.com ...
Thanks for the post.
It is looking good for Bush.
|
Date
|
Bush/
Cheney |
Kerry/
Edwards |
Nader/
Camejo |
Spread
|
|
|
RCP Average
|
10/12 - 10/15
|
48.8%
|
45.4%
|
1.6%
|
Bush +3.4
____________________________ Yes I like that! |
Very interesting!
Kerry now leads Bush 57 percent to 36 percent among those who identify themselves as first-time voters. The number of voters who say they are still open to switching candidates is actually fairly small, but still large enough to determine the popular vote winner: One in ten (11 percent) registered voters are still uncommitted.
Bush has a clear advantage with women, who prefer him 49 percent to 43 percent. Kerry has a slight edge with men, 50 percent to 46 percent. The Democrat also commands 50 percent of the youth vote (with Bush at 41 percent) and 54 percent of the senior vote (to Bush's 39 percent.) But Bush dominates the 30-49 set (56 percent to 39 percent). Voters aged 50 to 64 are split evenly, with 46 percent backing Bush and 48 percent Kerry.
However, I can not find the numbers of those polled, republicans vs. democrats and so forth.
After Late Start, Republican Groups Jump Into the Lead: Since August, 527s Raised 6x vs. Democrats
Averaging polls is a really bad idea. It is as unscietific as you can get. Polls only offer a snapshot of candidate support within their own base. You can get an idea of how they do among "independent" voters, and among different demographic goups or in target areas. But the idea of getting an accurate national or even state wide poll is a myth. The special sauce in polling is in party ID percentages. So why average all of these different polls together? All you do is get an even more confused mush, and you completely spoil even your ability to track trends. Each poll must be viewed isolated, since each poll has its own unique set of factors. They are only good for comparing results over a period of time using a consistent methodology. And those are the respectable ones. Most are only good for creating a news story or influencing public opinion.
I never believed that Bush could easily knock Kerry out in one debate. Kerry gave the best he could, and the polls show people thought he won all three--I disagree, and in the end, it looks like Kerry didn't win any of the debates--he "won" them.
But people considered Kerry, because they believe this IS the most important election of their lives. And they got his message. Unfortunately for him, Kerry forgot that he always did better when he wasn't very visible, he was just ABB.
But now it's getting close to decision time, and voters cannot be so frivolous--it's easy to "give" someone support far away from the election, it's something else to vote for them. I think Kerry's learning that.
Bush won all three debates because he gave Kerry NOTHING to use from them--Bush has harvested a pile of material to use against Kerry, though. And Kerry has no more opportunities to shone--no debates, no convention, zilch.
I still want to see the latest Gallup - if it trends toward Bush - Kerry IS toast ...
"Bush has a clear advantage with women, who prefer him 49 percent to 43 percent. "
This is a major good news for Bush -- in 2000 it was the reverse, more women voted for Gore, by at least 6-8 points, maybe more.
Women realize that Bush is the one who can keep them, their children and families safe, while Kerry is useless for that.
Some of the stats in that Newsweek poll seem suspect. Kerry having an advantage with men is one.
True--one poll results from polling X number of Dems, another Y number of Dems, etc. This would have value if all the polls used the same parameters. But I think it has value if used in the broadest possible terms, which is all national polls are good for anyway. It's a tool to get a feel for the momentum--it's not accurate, yet polls are inaccurate to begin with. I find them interesting but of little consequence. It jsut so happens that the polls have been mathcing my own observations about how things are going.
The last one Gallup has on it's site is from the 12th, I didn't even go in and look at it..but the headline was it's a tie.
Bush has a clear advantage with women, who prefer him 49 percent to 43 percent. Kerry has a slight edge with men, 50 percent to 46 percent.
HUH??? This seems to be OPPOSITE of what it should be?
Perhaps if you are to average a bunch of polls, you should also average the MoE, or just choose something like 4.5%, and then remove any poll that falls outside of it, then reaverage. this time add back any poll which may be within the MoE upon reaveraging. If there are any, then reaverage again. This is for the sake of removing those odd polls which come smothered in special sauce, usually from Newsweek, so they don't tweak you average. Of course this also guarantees you will have a real predictable, median final result, which just goes to show you haw arbitrary and useless all of this poll averaging is in the first place.
It's probably best at that point to just pick the one poll that comes closest to your average after, unbiasing as described above. But then again, why choose the middle most poll? Why not just choose one at random..? You have an equal probability of any one of them being accurate.
Makes sense, although it seems to me Gallup has a track record on its side. I just find Zogby a little weird, the way he futzes with the party levels.
Again, I try to see as many as possible to see if there are any trends. I think it's either that or ignore them ALL.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.