[Nervous Tick]

The Gallup tracking poll, whose procedure for designating likely voters makes it very susceptible to shifts in the balance of enthusiasm, has been showing Romney ahead by 5 to 7 points.

Fred, I'm soliciting your comments on the highlighted section above. Thanks and FRegards

[Skulllspitter]

It’s true. Not that there’s anything wrong with that. Also, since they don’t weight by party ID, Gallup has more natural variation that forced-allocation polls like Rasmussen.

[FredZarguna]

The Gallup tracking poll, whose procedure for designating likely voters makes it very susceptible to shifts in the balance of enthusiasm, has been showing Romney ahead by 5 to 7 points.

I would never gainsay anything said by Michael Barone on politics, especially at the micro-level. I believe Gallup's LV model is highly tied to interview questions from which they determine likelihood.

In this cycle, it would also correspond to the Party-ID model: Rasmussen actually shows Party ID for Republicans is very close to where it was just after 9/11, which was huge. A FReeper who subscribes to Ras's pay-page claims that Ras is currently coming in under Gallup only because he isn't even using his own LV model, which would be R+1, and is sticking with a very conservative D+3.

[FredZarguna]

If you’re worried about this being volatile (see my earlier reply) Ras’s — and other’s — alternative Party-ID methodology can potentially suffer from the same problem. Republican Party-ID dipped to historic lows as people got tired of the wars in 2006-2008. It spiked hugely after Bush climbed up the debris pile at the WTC site. All of the methods of prediction have pitfalls.