Talk:Polling
From dKosopedia
I changed the Category - trying to group HOW TO and information info for election activists in categories that will make them useful election season to election season. --SarahLee 17:15, 25 July 2008 (EDT)
Hi Scott!
I had a suggestion with respect to the last half of the third paragraph. "Second, the polling firm then picks a random sample of this universe and polls them, the model provides a calculation of the margin of error. In other words, the mathematical model guarantees that the sample reflects the whole universe and gives an estimate of the maximum possible error created by the sampling method."
Here's what I'd suggest as an alternative:
Second, the polling firm tries to pick the best random sample possible of the universe they specified and they poll this sample. The model provides a calculation of what the error would be if the random sample were perfect, but very often the sample itself is out-of-line. So when you look at "margin of error" estimates, you are looking at what the margin of error would be mathematically, if no other sampling or other errors had crept into the process.
Then when you get to the fifth paragraph, which starts the discussion of weights, I would suggest something like this ...
Almost all polling firms try to bring what they ended up with back into line with "reality" by using weights. If they didn't have enough Hispanic respondents, then they weight them more heavily and down- weight those they have too many of. Weights are well intentioned efforts to correct for sampling problems, but they also can introduce significant biases of their own. etc ...
My firm does a lot of survey research work for entertainment clients, and we have seen all the problems discussed in these pages come up in our work this year. We try to keep the math out of it in our write-ups and presentations, but statistics really is a tough thing to describe in reasonable words! I hope this helps.