Does Zogby have * up 47 to 42 in the latest Ohio poll?

[Richardson08]

MSNBC just had it

[montana500]

1. yes

but the latest scripps polls has Kerry up 4, and so do the internals.

[apnu]

2. bullshit.. I don't believe it. If MSNBC says so, I take the opposite view

[Downtown Hound]

3. Zogby's state polls have always been questionable

and that's putting it mildly.

[lancdem]

5. And if you check back to 2000

they show a lot of big swings from day to day.

[Jersey Devil]

4. Zogby has been contradicting himself

Many of the numbers in his new battleground callup polls totally contradict the numbers in the interactive polls take just a few days earlier.

For instance, in the 10/24 Iowa bg poll he has it 47-45 Bush.
But in his interactive poll of 10/19 he had it 51.1-47.9 Kerry.

Now, do you think Kerry lost 6 points in 4 days or that Bush gained 4 points in the same period in Iowa?

NM is even more ridiculous. His interactive poll on 10/19 had Kerry up almost 10 points, 53.6 - 44.1. Yet on 10/24 he has it as Bush 49-44. What could possibly have occurred to turn around a poll from a 9 point lead to a 5 point loss in 4 days?

Either the interactive polls by Zogby are bullshit or the phone call polls by Zogby are bullshit, or maybe both.

Funny, but just a few weeks ago Zogby posted a long article about how Gallup couldn't be trusted due to the big swings in their polls.

[kurt_cagle]

6. Statistical Noise

There aren't many undecideds yet, and people feel strongly on both sides of the aisle. That means that finding good samples actually becomes harder. I know its expensive, but I think that if this isn't just partisan hackery that Zogby would do good to expand their poll to 2,000 to 3,000 people. WHen you get this much variability in polling its usually a good indication that you're sample is too small and hence not representative.