http://www.gallup.com/poll/156914/am...-election.aspx


We already knew that, didn't we....

I'm glad you'll sleep better tonight.

Now I know how he got his avatar.

.

http://news.yahoo.com/obama-widens-l...201203713.html


Yuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuup.

...but, but, but...that is quite a different result than the recent LTUP poll with a very similar question. What gives?

it could have been a true patriot in the whit house instead it will Obama
can you say


We Paulbots told ya so.enjoy your loss GoP.
Mitt is Obama is mitt. we will never again vote for you lesser of 2 evils.

"...with a random sample of 1,033 adults..."

1,033 adults is 'most Americans'?

Originally Posted by QQuestor

it could have been a true patriot in the whit house instead it will Obama
can you say


We Paulbots told ya so.enjoy your loss GoP.
Mitt is Obama is mitt. we will never again vote for you lesser of 2 evils.

We never knew....probably a plant for the Kenyan is American Klan.

Originally Posted by micro_builder

"...with a random sample of 1,033 adults..."

1,033 adults is 'most Americans'?

A random sample of 1,033* adults is a decent sample size, giving a margin of error of something approaching the order of +-3%. at the 95% confidence level. Close enough for jazz and the sort of sample level that lots of pollsters aim for.

*1067 individuals =3% Margin of error@95% confidence level for a US population of ca 300 million.

Originally Posted by leccyflyer

A random sample of 1,033* adults is a decent sample size, giving a margin of error of something approaching the order of +-3%. at the 95% confidence level. Close enough for jazz and the sort of sample level that lots of pollsters aim for.

*1067 individuals =3% Margin of error@95% confidence level for a US population of ca 300 million.

Random is always the key world as for reliability of predictions. Skewed samples makes for skewed results. Which is why multiples are used to determine which samples tend to be biases.

Originally Posted by LcJ

Random is always the key world as for reliability of predictions. Skewed samples makes for skewed results. Which is why multiples are used to determine which samples tend to be biases.

Yes Lyle, and, given that sampling and polling is the stock in trade of the companies and organisations that do it professionally, they know a thing or two about how to conduct such polls.

Lemme guess though, you believe in the results of LTUP polls when they suit, but discount the results of other polls when they do not.

Yup, that pretty much sums it up.

Originally Posted by leccyflyer

A random sample of 1,033* adults is a decent sample size, giving a margin of error of something approaching the order of +-3%. at the 95% confidence level. Close enough for jazz and the sort of sample level that lots of pollsters aim for.

*1067 individuals =3% Margin of error@95% confidence level for a US population of ca 300 million.

Yes, but still 1033 opinions is certainly not the same as 300 million opinions, no matter margins of error or confidence levels. Calling that "most Americans" seems a bit hopefully at best.

I've come to dislike polls, they can be so easily skewed to suit an agenda (not saying this one is, just in general), and they usually seem to be used as the broadest paint brushes I've ever seen.

Originally Posted by micro_builder

yes, but still 1033 opinions is certainly not the same as 300 million opinions, no matter margins of error or confidence levels. Calling that "most americans" seems a bit hopefully at best.

I've come to dislike polls, they can be so easily skewed to suit an agenda (not saying this one is, just in general), and they usually seem to be used as the broadest paint brushes i've ever seen.

+100