Who would poll on Halloween evening and take it seriously?

[Quixote1818]

:rofl:

On Halloween most people are getting ready to go out or are at the bar by 5:30 pm, accept very conservative types, religious types and old folks. It would seem a piss poor day for polling. Enough to cause at least a 7% shift. That is less than 1 in 33 people. If 3.5 people out of 100 shift from Obama to McCain then you have a tie. I would guess that the number of people out on Halloween or getting ready to go out is around 30% mostly young people and those up to about 50 years old. Any sample done on Halloween would need to be taken with a grain of salt. Basically Obama's core supporters are exactly the type to be out on Halloween and McCain's core supporters are exactly the type to be home.

[John Q. Citizen]

1. I would vote on Halloween evening, if that were an option.

[progressivebydesign]

2. Supposedly he only polled 300 people, which is absurd as well.

Zogby is such a lightweight.

[Quixote1818]

4. That means McCain got about 11 more people than he got before

If 300 were polled Obama normally would have gotten about 150 people and McCain about 129. If 11 people shifted, then Obama only got 139 and McCain got 140. The margin of error would be off the charts.

[FrenchieCat]

3. During the day, parents were either at work, or went to their kid's schools,

to watch them in the Holloween parades.

The only folks home were the old folks.

[justiceischeap]

5. Someone who wants to purposely skew the polls?

:shrug:

[tpi10d]

6. according to Nate Silver @ 538

Most of the Zogby calls were made yesterday evening. A few were made this afternoon. According to Nate, Zogby is a trailing indicator.

As you know, most of the tracking polls already reflect the Thurs. evening sample...and most showed an Obama increase.

You can read more here:
http://www.fivethirtyeight.com/2008/10/trick-or-treat.html