Originally Posted By sgtar15:
Hiram. WHy are some polls used more than others?
And how did these polls become known?
What education/training have these people gone thru to become pollsters?
Most of them are statisticians... polling is nothing more than using mathematical formulas to try to glean the overall representation from a sample. Some polls are used more than others because their methodology is determined to be more sound. The hardest part of polling is to develop a set of questions that has as little bias as possible built into it, and utilizing a large enough and diverse enough sample to reduce the margin of error as possible. Once you eliminate the bias, the key is to pick a sample that brings you down to a smaller error rate. Ideally I like sample that are pushing 1,500 or more as I like to see the error rate below 3 points.
Rasmussen is doing rolling averages. In other words, todays number is from polls conducted Friday night, Saturday and Sunday... three very bad days to be polling Republicans. The kicker is I would expect to see Zogby's poll roughly tracking on this as well since they are both rolling and working on three day averages. Instead you see Zoby opening up.
Now, the best I can make of this is Zogby has long been known as a partisan shill and is likely now trending his polls to the reality of the race in order to not come across as blowing it. He gave the media markets what they wanted, a see saw horserace. Now he wants to make sure he is hired in the future. Rasmussen I have always argues downplays Bush to the rate of 3-4 points. If I apply this consistently and take Bush's two points down than we would be looking at Bush up 1 or 2, which would make sense coming off of a weekend. TIPP released a new poll today and shows Bush up 8.
Like I said, I am going to need until Wednesday to really have a good idea as to what is going on with Scott's numbers. Wednesday's sample will have Sunday, Monday and Tuesday nights figured into the sample, including the influence of the clinton campaign swing in PA as well as the NYT's munitions story. If we are trending upward Wednesday, I will consider it confirmation to the theory I have just offered up.
Never forget, even the best polls have outliers that occur... even great samples are flawed from time to time. The DUhers will be crowing, which will be odd because they consider Rasmussen a shill... They will say its really bad news that even Bush's hack is showing a Kerry lead or even Rasmussen wants to be on the winning side... Once the tables turn you will see them attack him again.