ETA, this is an awfully small poll — Just 400 people as opposed to the sample Rasmussen takes ( more than double this ).
The Boston Globe had Coakley up 15...bout ten days before she got smoked.....not the best polling group