Thanks, but I am looking for a little better explanation than that.
Any explanations on why it seemed like every pollster and pundit missed Colorado so badly? Did it really break so strongly on election eve? Was there a poll that showed an 8 pt lead for Allard (Wasn't that the final tally)?
As I have been saying since before Tuesday, the Republicans were probably undersampled in most polls. Before taking the poll, the pollster has to make assumptions based on historical data about the proportion of Pubs, Rats, Libertarians, Greens, others etc. are in what area. Then when they pre-qualify the people that take the poll to get the correct quota of each type. If they did not do it this way they would have to take a much much larger sample to get accurate results. The problem, as I see it, is that there are now more people who qualify themselves as Pubs, and fewer that qualify themselves as Rats--and nobody was doing the kind of expensive polling neccesary to show this.
I first suspected this when I noticed a higher then normal difference between the 'registered voter' vs 'likely voter' polls. Usually the 'likely voter' polls favor Pubs over Rats compared to the 'registered voter' polls. But this year the difference seemed larger (no objective measure here...they just seemed higher as I followed the polls). The only plausible explination that occured to me was that the rank and file Rats were not as committed then they used to be--indicating that at least some of them were so much less committed that they no longer qualified to be polled as Rats.