Originally posted by tinyp3nis
I allways found polls (especially the ones about popularity before elections) so accurate that they make the counting of the votes boring to watch because you already know the results...
I allways found polls (especially the ones about popularity before elections) so accurate that they make the counting of the votes boring to watch because you already know the results...
In Illinois, the election for Gov. and Att. Gen. were both reported to be close, right up to the day of the election, but both ended up reasonable landslides. More people are unwilling to tell the truth to pollsters. (not good for our business, but fortunately we concentrate more on consumer goods and customer satisfaction)
RAH
edited...
And just for clarification. We've really have been talking about completed surveys, not sample. We would guess the response rate to figure out how much sample we would need. (i.e. If we had experienced a 10 % completion rate, or just guessed, we would pull 50000 sample to get 5000 completes.)
Most random phone surveys will not even perform to this level.
Comment