Friday, October 19, 2012

Gallup's sensitive likely voter model

Gallup has been getting walloped, as of late, by critics who say Mitt Romney's 7% lead in its tracking poll is too big, and indeed, that lead is 6% outside the national average and highly weird.

Gallup's editor-in-chief, Frank Newport, tells Ezra Klein that, yes, indeed, Romney's massive margins in the South might be responsible for that big number.

But more interesting is this little statement from Newport, reflecting on how the likely voter model functions.

"It would be very sensitive to changes in enthusiasm. The Denver debate clearly had an impact on Romney’s people. I think your insight is correct there. Whether we see a dulling of that over the next several days is what I want to see.”

The unfortunate thing about polling in this epoch between the second and third debate is that by the time polls are accurately reflecting the second debate, the third (on Monday) will already be upon us.

Once that's completely digested a few days later, it will already be the weekend of the 27th -- also known as about 1.5 weeks until the election.

So that means we won't have a clear picture of where the race stands, post-debates, until the hour is nigh for the final votes to be cast.