a) Ipsos-Reid has already had its internals analyzed and was found to be oversampling Dems by a whopping 10%. That alone would give victory to EVERY "VULNERABLE" REPUBLICAN, PLUS KEAN, PLUS STEELE and would make Kennedy and Harris quite competitive.
b) A month ago a pollster pleaded guilty to manipulating data vis-a-vis labeling "Democrats" as "Republicans" to achieve pre-set quota numbers. Surprise, surprise.
The problem is, on election day, the Republicans who do vote really are Republicans, and they vote in higher numbers than the Dems.
Point 1: Ipsos-Reid was a national poll. It applies in no way to analyzing the accuracy of state polls
Point 2: You keep referring to this one poll, not the at least 150 state Senate polls released in the last two months.
Point 3: "Oversampling" Democrats by 10 percent still wouldn't give Santorum a victory.
b) A month ago a pollster pleaded guilty to manipulating data vis-a-vis labeling "Democrats" as "Republicans" to achieve pre-set quota numbers.
Point 4: Said "pollster" is DataUSA or Viewpoint USA. I've never seen a single poll ever released by that corporation. Never seen any here, on pollingreport, on RCP, on Cook, or on any blogs. Nowhere. Besides that, Viewpoint USA was not a media or mainstream pollster -- they specialized in INTERNAL polling, which I don't happen to put a lot of stock in anyway. The ironic thing is that me, the "poll guy," doesn't believe internal polls. On the other hand, you, the tinfoil hat media-conspiracy guy, believe only what campaign workers (and Zogby Interactive -- as long as he shows Blackwell losing by less than 10) tell you while completely denying all independent and nonpartisan evidence to the contrary.