The biased polls examined at Patterico today. Thing I noticed is that the flawed results have to do with "predicted" turnout, a number that those with biases can easily manipulate. Left out is the dead on accuracy (again) of the Rasmussen poll correctly calling it 52-46.