Britt Hume quizzes Fox pollster, he admits "we're not very good" Halloween screwed things up

Special Report with Britt Hume | 11/01/2004

[killjoy]

Britt's answer to that was that summing several polls reduces the margin of error.

Not really. That would assume that all the polls were done the same way with the exact same questions. Something I highly doubt. It has been a long time since I took college statistics, but these polls are mostly meaningless. They give a general indication of trends, but they are not specific enough to offer any real details. Without getting into confidence intervals, these polls can only get down to about 5% accuracy. Someone better versed on statistics can give more reasons why.

[Mannaggia l'America]

I've never been polled, and I probably never will be. We do not answer the phone if there is no name on the caller ID, and I know LOTS of other people who do the same thing.

[Spook86]

You should find them a Realclearpolitics.com. Mason-Dixon released their last round of polls Saturday night, I believe.

[KevinDavis]

I think after the election a lot of polling firms will go out of business.

[SteveAustin]

The Halloween thing could have skewed against the GOP. I'm in a fighting mood tonight, so I will give it a go with an explanation.

A TON of Bush voters are families with small kids. If your house was like ours yesterday with getting three kids in costumes and dragging them around the neighborhood, and passing out candy, you had NO time to answer the phone.

The one Kerry guy with a lawn sign in my neighborhood had his outside lights off and wasn't participating in Halloween....that guy was answering the pollsters calls!

Ok...maybe a reach here, but let's have fun tonight!

Go GW!

[Jose Roberto]

I think this is the end of a beautiful (LOL) relationship between Fox and Opinion Dynamics. Next election cycle Fox will team up with Rasmussen and Zogby with Opinion Dynamics because Reuters will kick old Zogby to the street for such an embarrasing job. Forget this poll, believe Harris Phone poll and Pew.

[lepton]

Are you under-representing people with kids who aren't answering the phone?

Nya nya nya-nya nya!

That's what I've been repeating the last two days. Parents with kids who live at home are one of the most pro-Bush (and even Dole) demographics there is, and polling on Halloween almost garauntees you'd miss them.

[Intolerant in NJ]

Then Britt brought up the more glaring state poll results Fox has. Kerry 49 Bush 43 in Florida vs. Quinnipiac having Bush 51 Kerry 43. Again, he said that those 2 polls says the same thing(!!) but the pollster REALLY looked sheepish after he said it...he should have - he skated over the fact that the Quinnipiac at 8 points was outside the margin of error (+/-4), after he had just spent so much time extolling the wonderfulness of the MOE, while the Fox was well within the MOE of +/-3.7...but then who would expect some college professor to point out something positive about what was going on with Bush......

[Verginius Rufus]

I think the "L" is a "J" (the two pumpkins are supposed to show John Kerry's initials). Sure to sway some undecided voters to vote for Kerry.

[lepton]

Bush's lead of 8 points exceeds 2x MOE. So we can be sure with 95% probability that Bush has a real lead.

Isn't "MOE" two Standard Deviations, which is approximately 95%?

[Cableguy]

It's a bell curve, with 95% of the curve within MOE.

[BobFromNJ]

Bump! As I said on another post WEEKEND (HALLOWEEN) POLLS STINK!! GO BUSH!