2008 Two Bad Polls for Obama
posted by April 18 at 14:39 PM
onGallup says his lead is shrinking, and more Democrats than Republicans think Obama’s a Muslim.
« A Very Bad Poll For Clinton | And One Very Bad Poll for Dino »
posted by April 18 at 14:39 PM
onGallup says his lead is shrinking, and more Democrats than Republicans think Obama’s a Muslim.
Comments
Barack Hussein Osama? Of course he's a fucking Muslim.
The national polls of Democrats - How does this matter? Isn't this why we have a primary, where people vote and all? Please tell me this isn't going to be something to coerce superdelegates with. Most of these people either voted already or passed up their chance to vote already. What gives?
Match ups with McCain I understand. But this? Somebody explain, please.
Sometimes I suspect that some small percentage of people in these polls actually don't recognize his name (there's always some of them) and, when polled on whether he's a Muslim, just guess based on the name. (And some others actually believe it, of course.)
Once again, a poll of "likely voters".
Meaningless.
This is the year of the Blue Tidal Wave. The "unlikely" voters. Many don't have land lines and don't answer surveys, FWIW.
Actually, that's a very positive sign if more Ds than Rs think he's a Muslim. The nomination is a nearly foregone conclusion.
How could this poll be so much different than the previous one?
Not the previous Gallup poll, but the Newsweek poll posted by Ryan.
Maybe something to do with the phrasing of the question? I can't find the actual wording used. Maybe one asked "who do you like better" and the other "who do you want to be the nominee" or something like that.
If a random sample of people, in a number of 512 or larger, is selected for a poll, we say it has a 1 in 20 chance of being totally wrong. As in totally - it could say McCain is winning 80 percent to Obama 20 percent, for example.
That's typical.
Now add in factors for the non-randomness - only land lines phoned, only people with phones called, not everyone votes the way they say ... and you get an even larger number of incorrect polls, @6.
Then add in all the "unlikely" voters this year and you realize all your polling methods are pretty useless - normally we count 3s and 4s, not 1s and 2s ... but this year most voters are people who rarely if ever vote ...
@6, the Newsweek poll is most likely wrong. There is no way anyone has such a large lead. I support Obama, and don't believe the other one.
Yay, research.
The Newsweek question was:
I can't find the Gallup wording. Anyone?
#8, the two polls were conducted at the same time, and almost the same sample size, and the same methodology. What you say about land lines and likely voters is true, but the two polls would at least have the same 'wrongness factor' when conducted under near-identical circumstances.
Margin of error is based on 95% confidence, but that doesn't mean everything outside the margin of error is equally likely. The polls are just too different. You could base this on three standard deviations instead of two, which is 99.3% confidence, and the margin of error still would not account for the difference in the two polls.
90% of statistics are bullshit.
@12 - I think you mean "90 percent of bullshitters use statistics to 'prove' their point".
The other 10 percent use wikipedia.
So the Chicago pastor who was the subject of all the hubbub a few weeks ago was actually an imam in a mosque?
It'll please me to see Obama finally edge past Hill and get the nom, but by now it's a pyrrhic victory and I'm pretty resigned to the idea of "President McCain." The democrats lost it again, before it even got started. Fucking amazing.
-
We obviously deverve the fucking government that we have (will probably have)!
Comments Closed
In order to combat spam, we are no longer accepting comments on this post (or any post more than 14 days old).