Stygius is tweaking me about a post I wrote a few weeks back.
I agree with some of his reservations, but I think my general argument works: It seems to be the Rovian strategy to start building from the extreme and then try to add the middle. That strategy is inherently unstable, because as opinion changes, large swathes of people can peel off. That seems to have been what is happening as President Bush's popularity declines. It may also account for a sudden jump in his popularity as he tries to put on the appearance of reasonableness. Then again, that jump might not exist.
Bottom line: building support from the extremes can lead to wide swings in popular support. Wide swings in popular support don't make for good government.
Stygius:
One way I view the problem, is that the situation of the "center" is always historically driven, there is no abstract, independent "center." For instance, the Bolsheviks remained a minority of a minority well into the Russian Revolution, but were still able to out-organize and out-maneuver everyone such that it was they in the end who determined what the orthodox status quo was. Hence, the center is always temporary. So let's caveat before celebrating centrism over all.
Agreed, but my argument doesn't depend on "right," "left," or "center." The fact that there is no independent "center" is what drives the instability. People whose opinons fall toward the center of the hump in that distribution curve may also be swayed toward one side or the other more easily than the fanatics at either end.
The opinions of large groups of people are likely to be represented by a bell curve. That curve might be narrower or wider, or it might be skewed toward one side or the other. But that big hump, somewhere midwayish, remains.
So approval of President Bush probably follows a curve sort of medium-high and medium-wide, like this one, but skewed today toward disapproval according to CNN or more towards the middle according to WaPo.
Approval of stem-cell research by the American public is probably skewed seriously toward approval as opposed to "it's against God's will," I would guess to the extent that there isn't even a skinny little tail at the approval end of the curve.
If you asked a question like whether people think President Bush is the Second Coming or The Beast, with "neither" charted in the middle, you'd probably get a very narrow and high peak. (At least I hope you would!)
Now, all these factors are coming together at one time, and probably one or another prevails in any poll. We haven't seen the effect of the NSA spying question on the polls yet, for example. So yes, it's reductionistic and the temptation is to lump it all together into left, right, and center. Statistics tend to work remarkably well on large numbers, though, so that just might work, just as Rove's strategies have been working.
But the statistics say that those strategies will always have large instabilities.