Update: Zogby has a poll out for FL-22 which makes the race there look like a Dem pick up. I don’t buy the wave thing automatically, but I would believe Kerry leaning districts dumping GOP incumbents (which is the case in all those really tough races). RCP shows this is one of those districts. end update

We have to very conflicted polls out today to ponder as we head into the final hours of the 2006 election. The first poll we see shows a huge shift towards the Republicans in the usually left leaning Washington Post/ABC News poll for the generic ballot question (who would you vote for – Rep or Dem?). Then we have Rasmussen out today noting the number of self identified Republicans in his electronic survey (he uses automated telephone polling, I believe using keypad responses).

In the final full month before Election 2006, the number of people identifying themselves as Republicans has fallen to its lowest level since we began reporting this measure of partisan trends in January 2004. As a result, Democrats have their biggest net advantage of the past two campaign cycles.

H/T to Ace for the Rasmussen post. So,how could these polls both be measuring the same electorate? Well, they can be measuring the same electorate, but they may be measuring two different things. The theory goes that Democrats are much more likely to respond to telephone pollsters than Republicans. And as the pollsters have moved from the personal touch (live people on the end of the phone), the Republicans fall off in greater numbers. This can be seen in the fact that automated polling by folks like Rasmussen and RT Strategies/CD all show a recurring and regular bias to the dems.

Now I am actually surprised Scott Rasmussen would go out and claim the drop in Republican responses is actually Republicans leaving the party without addressing the issue of poll responses. I guess once you have had a good record with a methodology you fall into the trap of believing it will stay good forever. But we all know that is not the truth. Look at the what digital cameras did to film to understand what the internet has done to communications (that includes media, mail and phones). We will know in a few days if the electorate shifted to the Dems or away from the pollsters. But the data I see doesn’t make sense with a big Democrat shift, especially in the turnout we are seeing. I may be wrong and I may be too Polyannish here. I plan to do one more round of where the House and Senate are tomorrow, but I think I will stick with my view the polls are off by at least 5% due to erroneous turnout models which are totally predicated on who responds to the pollsters.

Addendum: Folks, I would like you to look at another question in that surprising Washington Post/ABC News poll. In this graphic representation of the results just out last night, the post has numbers on George Bush at the top left figure. The question is whether you approve or disapprove of how W is handling the presidency. The graphic illustrates the liberal wishful thinking in the media. If someone like, say, Michelle Malkin is not happy with W on illegal immigration, the Dubia Ports and a host of other issues (yes, I know this example is a bit of a stretch 😉 ), then she would respond “somewhat dissaprove”. As would most conservatives right now. But that doesn’t mean Michelle is already to get out and elect Pelosi speaker! So the true anti-Bush, anti-Republican vote is the 46% of the democrats who always opposed Bush and ‘strongly’ dissaprove. If I was looking towards voter trends, I would say Bush as 52% support and 46% opposition. And that seems very close to the 2004 results (which it is). More indications there is not a huge shift to the Dems. People are ‘somewhat’ disappointed are not going to siwtcht parties.