Aug 18 2006
I am still chuckling over all the fantasies being generated by the left regarding the latest Qinnipiac Poll showing Lieberman clobbering Lamont 53 to 41 among likely voters (previous post here). They have truly come with some amazing ideas on how Lamont is actually winning! The most hillarious claims are the huge gains in Democrat support for Lamont since his squeaker 52-48% win in the primary. Supposedly the poll shows a 65-35 split amongst Democrats only a week later! This has the left all thrilled. What happened?
Nothing. A poll is a statistical estimate based on a small sample. And they can be very far off if the pollster do not have a historic basis to adjust the mathematical weightings in unique circumstances – like third party candidates. Votes, on the other hand, are precise measurements. We all know the liberals still believe Kerry won in 2004 because the estimate (the polls) are supposedly more accurate than the measurement (the ballots). Nothing has changed in liberal la la land. I am well steeped in science and mathematics and engineering, and I know which number is more accurate – it is the measurement from the primary.
My guess is the truth is somewhere in between. It is hard to believe someone who voted against Lamont in the primary is all of a sudden going to ‘see the light’ in the general. But some will fall into the party line – for now. If I am correct, that means the poll is under-estimating Lieberman’s pull from the Democratic base. And unless the pollsters change their models – which they are loath to do since it makes comparisons over time impossible – then this hidden bias will remain in the estimates until the next measurement is taken in November.