The 'Polling' Wars

I found that there is an old THREAD from 2004, about 'polls'.....hope I searched correctly.
This should likely be in the '2008' Forum, please move as necessary.

In the run-up to the 'Conventions'.....am seeing more and more political pundits on every 24-hour Cable News channels continue to 'cite' various polls...

Now, I realize that a 24-hour News Channel has a lot of airtime to fill....but these talking heads 'argue' about the polls.....this 'percentage' and that 'percentage'.....I got to thinking (never a good idea!)

'Talking Heads' talking about percentages rarely go deep into HOW the varoius polls were conducted.

WHAT were the 'sampling' criteria? HOW many were sampled? WHEN were they sampled? And, the biggest one....WHO were sampled, and how were they chosen???


I invite comments, thanks.

Ever read the book "How to Lie with Statistics? I generally do not believe any statistics unless I see the raw data and design of experiment. It is ridiculously easy to design a poll to get the answers you want.

Not to mention the usual discrepancy in exit polls... I too am keen to see what comes of this.. polls have always thrown up a few 'flags' in my book.

Nice post

reply to post by crimvelvet


Thanks crimvelvet....never seen that particular book, but it is quite pertinent to this concept. (and I intend to find it...)

I HAVE actually seen some examples of how 'statistics' can be skewed.....depending on how they are 'spun'.....

Hence, my point about these 'polls'.....

Thanks for your post!!