Gallup poll may go "a galluping" to strange places...

[The Vorlon]

Gallup, never known as the most stable poll to begin with, is likely to go fairly nuts over the next six days...

The three day samples from Oct 19 to 22 showed a net INCREASE in Obama approval from +4 to +11, but the next three days show a DECREASE from +11 to -3...

"assuming" Job approval and ballot preference correlate at least a little bit, could be a bumpy ride down and then up as these 6 days roll off the 7 day tracking poll...

Yikes!

This should be exciting

[Torie]

What is your speculation as to what is generating these wild oscillations in this one poll but not others, Vorlon?

[Lief 🗽]

Yes, Gallup is terrible.

[Sbane]

What is your speculation as to what is generating these wild oscillations in this one poll but not others, Vorlon?

It's Gallup. But even overall it's pretty hard to gauge where the election is. It would be one thing if everyone was showing a tied race, but some are showing a Romney win by a few points and some are showing Obama up by a few points. Election night should be very interesting.

[Torie]

What is your speculation as to what is generating these wild oscillations in this one poll but not others, Vorlon?

It's Gallup. But even overall it's pretty hard to gauge where the election is. It would be one thing if everyone was showing a tied race, but some are showing a Romney win by a few points and some are showing Obama up by a few points. Election night should be very interesting.

As I kept telling you sbane, the demographic adjustments all the pollsters do except those that use partisan adjustments, are a bit of a Hail Mary pass. So the poll variances do not surprise me. Intra poll trends where the same methodology is used are the only thing really that gets my attention these days, given how close it all is.