Tell me again why we should be trusting the polls we’re being fed? Why are pollsters insisting that 2008 turnout models will be predictive this time? Why are they insisting that trends have moved in the direction of the Democrats when the opposite is true?

Registration advantages and shifts
(Data from Jim Geraghty at the Campaign Spot) 

Iowa

2009: D+ 110,000
2012: R+ 20,000
Shift to R: 90,000

Florida

2008
D: 4,800,890
R: 4,106,743
Advantage: D+694,147

2012
D: 4,627,929
R: 4,173,177
Advantage: D+ 454,752

Nevada

2008
D: 531,317
R: 430,594
Advantage: D+ 100,723

2012
D: 447,881
R: 400,310
Advantage: D+ 47,571

New Mexico

2008
D: 594,229
R: 375,619
Advantage: D+ 218,610

2012
D: 582,656
R: 385,898
Advantage: D+ 196,758

North Carolina

2008
D: 2,866,669
R: 2,002,416
Advantage: D+ 864,253

2012
D: 2,778,535
R: 2,008,609
Advantage: D+ 769,926

Pennsylvania

2008
D: 4,479,513
R: 3,242,046
Advantage: D+ 1,237,467

2012
D: 4,185,377
R: 3,099,371
Advantage: D+ 1,086,006

Colorado

2008
D: 1,056,077
R: 1,065,150
Advantage: R+ 9,073

2012
D: 739,778
R: 837,732
Advantage: R+ 97,954

One of the most fascinating takeaways here is how, in many cases, both parties are losing registered voters. That is a subject of longer-term interest. But for now, the big story is this:

• The Democrats are losing more members than the Republicans. In every one of these states, the Republicans either gained registered voters or lost fewer of them than the Democrats did.

• The shift since 2008 has been towards the GOP and away from the Democrats.

So, how can any poll that uses 2008 turnout as a model be useful for 2012?

The answer is that they cannot. The pollsters are either intentionally misleading you in order to create a bandwagon effect for Obama, or they are incompetently missing obvious trends.