« New Polls Show Kerry leads by 9% in NH, 2% in FL | Main | Gallup Vs. Everyone Else »

The "How Can Gallup........" Game (Florida Edition)

Gallup has just released a new Florida poll that has Bush ahead of Kerry by 9 points among RVs (8 points among LVs). This extraordinary result has led DR reader Lawrence Becker to propose another round of the "How Can Gallup....." game.

Remember how to play? It's easy! Just take the latest Gallup outlier and compare it to other publicly available data that seem to contradict it. And let the fun begin!

Here's Becker's contribution:

How can Gallup have Bush up by only 2 among registered voters nationally, yet have Bush up by 9 among registered voters in Florida?

That is really a feat. Letís compare with some other polling organizations that poll both nationally and in Florida. Yesterday Rasmussen (by no means a Democratic firm) had Kerry up by 2 points nationally; in fact, it was Kerryís first lead in Rasmussenís tracking poll since August 23. Rasmussen also had Kerry and Bush tied in Florida. And Zogby had Bush up by 3 in his national tracking poll yesterday. But he had Bush up by just 1.2 points in Florida!

I'm sure you can think of your own. And here's a couple of my own contributions.

How can Gallup have the race dramatically tightening on the national level since its late September poll--from an 11 point Bush lead in their 3-way RV matchup to a mere 2 points today--while they have Bush's lead in Florida nearly doubling over the same period from 5 to 9 points?

How can Gallup have an 8 point (!) Republican party ID advantage in both their LV (44-36) and their RV (43-35) Florida samples, when the Democrats had a 2 point advantage among actual Florida voters (40-38) in the 2000 election? (Thanks to Steve Soto for the Gallup Florida party ID numbers.)

Note: for more on Gallup's outlier state results, see Alan Abramowitz' contribution above.


How much specifics does Gallup give of its RV samples? If they release age, race, income numbers then it might be fun to compare their own sample-to-sample swings in thos categories, and not even worry about the votes. As in, sure Bush is winning, but the electorate just aged 4 years and turned 5% whiter from last week -- now *that's* real news!

How can CNN and USA Today broadcast and publish this garbage?

Galluping the GOTV

I propose a different game. Look at the Gallup difference between BC and KE in a swing state, and commit to bringing that same number of voters to the polls. The way Gallup is manipulating that difference, this game will lead to a KE landslide.
Get angry and then get even with Gallup....GOTV.

I live in south Florida and have been to the polling place for early voting twice and gave up (postponed - I will try to get a place in line at dawn) because there is a 1-plus to 2-plus hour wait. This is just caused by the new touchscreens and not by anybody challenging voters yet. This expected wait is widely known at this point by the south Florida electorate. This dynamic (the uncertainty of being able to vote plus the realization of the time investment plus the perceived likelihood or otherwise of it being counted) is going to have a bigger impact on results than any LV formula could have conceived of. The voters who are hanging in there are a quite committed subset of LVs. For change or for Bush? Who's going to stand in line two hours to throw away a vote on Nader? And God knows what Tuesday is going to be like. Heaven help us all.

Gallup knows something the rest of us don't: they're factoring in Jeb Bush's skulduggery.