nuckles87 said:
|
Not actually true.
Multiple polls signficiantly reduce the margin of error.... if they're looking at the same thing and their stnadads are fairly even.
All polls have their own models and often measure different things and also oddly, try and be as next to each other as possible because nobody likes risking being "too far out there."
Which is why composite polling has actually done a pretty shitty job at predicting vote totals.
It's just rarely come into play in elections because elections have been so far apart usually.
The 2008 presidental primary polls come to mind... and hell... the presidential election polls come to mind.
http://www.livescience.com/2183-presidential-polls-wrong.html
http://ns.umich.edu/new/releases/7067
Rasmussen was considered an outlier... yet won first place in accuracy the last couple times around.
Polls are fun.... but not remotely definitive.








