On Sat, 22 May 2004, Doug Henwood wrote:
> >Tossing out their their entire framework on the basis of divergent
> >results would be mindless empiricism.
>
> Pollsters can be mindless empiricists, which is both their virtue and
> their flaw.
>
> But I think the theory is that nonvoters are passive and disengaged,
> and thoughtlessly absorb the dominant discourse.
As a theory, that makes no sense. They can't assert non-voters are fundamentally different than voters and use that to explain why they vote exactly the same.
It does, however, suggest the real problem with their methodology. A nonvoter becomes a voter because for some reason she suddenly cares more. And people who care, and are better informed, vote differently. The transition from being a habitual nonvoter to being a voter is by definition an attitudinal transformation. So using how they'd vote in their nonvoter state as a proxy for how they'd vote *if* they became voters is about as silly as polling people when they are teenagers and asserting that that's how they'll vote when they grow up. As an aggregate, it can't be accurate, because in between they'll change.
This does however bring up a few interesting tangents.
One, registration is a sine non qua, but by itself, it does nothing (except, paradoxically, lower the turnout rate of registered voters). You have to get people to vote. And to do that, you have to make them care. You have to make them want to join a side. There has to be some reason it matters to them.
Two, there is one small but important area in which this is not true: people whom you can register and can then coerce/induce to vote your way. On the Dem side, union members and poor black churchgoers come to mind. This pool of peer-group-coercable people, who can be counted on to vote your way if you can drag their asses to the polls, is small. But in a tight race, 50,000 such people statewide could matter.
Three, in this arena -- where registration matters all by itself, and you don't have to get people to care, and there isn't any attitudinal transformation, just the power of strong-arm get out the vote campaigns -- the AAPOR polls are probably accurate: there are just as many in these small pools on each side. The Dems consistently do better because they have stronger grassroots organizations. But it's conceivable the Repugs could fight this to a draw with money. Theoretically they could hire a very sophisticated get out the vote machine if they wanted to spend enough. And their potential number of coercable votes equal ours. Even if so far, the Dems have always been better at because it's so much cheaper per vote for us to do.
Bottom line: the idea of increasing the electorate on a massive scale is different in kind from the idea of winning a vote through registration drives. Most people aren't coercable; you'd have to give them a real reason to care to make them voters. It would require a change in both platform, party and means of communication. You can hold registration drives until you're blue in the face and you won't every accomplish it.
The good news is that if someone did figure out how to greatly increase the turnout of voters, then the demographics predict this would favor the Democrats.
But the Democrats would have to be radically different Democrats to bring this about. And just registering people, by itself, will do nothing to bring that change in the party about. Registered non-voters are no threat to anyone. And registered voters who you make vote your way are not only not a threat to the powers in charge, they're a comfort.
Michael