Re NH polling, Freakonomics has a chapter on this (the Weakest Link study)

[AP]

In the chapter about racism, the authors analyzed data from The Weakest Link which showed that contestants do not act rationally and strategically against black, latino and old contestants. With the black contestants, other contestants over-compensate for racism by not voting off black contestants when it would be wise to do so (you have to explain your vote to the audience) -- ie, in the later stages when you want to keep poor performers strong black contestants were not voted off, and in the early stage you want to get rid of people performing poorly weak black contestants were not voted off. However, for old people and latino contestants, other contestants had no problem displaying a bias, even if it wasn't based on reality (people voted off strong performing old people/latinos in early stages, and kept them around in the end, even when they were getting a lot of questions right -- no matter what the facts were, they were perceived as being weak).

The theory is that the civil rights campaign has been successful to the degree that at least people realize that it's not socially acceptable to be racist (and, to a degree, people think that if they're trying to say that they're open minded, even if they don't really understand race, they'll chose a public display of not being a racist as a way to tell the world something about themselves).

This study was meant to explain why black candidates tend to poll higher than eventual vote tallies. Freakonomics argues that you'll tell another human being (the person who calls you from the polling firm) one thing because that's an observed event, but, in the privacy of the polling booth, you'll do another -- and that may be, for some, that you are free to show your racism, but also, for many people, they don't get the social "credit" for the vote if nobody really knows how they're voting, and they weigh other factors.

You can see how this plays out in the caucus situation too. In a caucus all your neighbors see how you vote, so that "observed event" element remains, which wasn't the case in NH.

[Anouka]

1. Except that the CW is that minorities last 2 episodes and are then booted off.

Minorities aren't expected to last long at all on predominantly white shows... yes, minorities are not booted off the first episode, but that's only to keep minority viewers glued to the tube the following week.

It's about ratings and money.

Minorities don't make it all the way to the end on predominantly white shows, and in minority communities are not expected to.

I don't think the same holds true at all for predominantly minority shows. But I don't have the evidence in front of me. Do white people look at predominantly minority reality shows and expect that the white is going to be booted off by the second or third show?

Besides, all minorities are not created equal.

What this has to do with New Hampshire is disturbing and insulting to me :) It puts the blame on minorities, and not whites. It seems to me, at least. But thank you for putting this up.

[AP]

2. I don't understand what you're saying.

On the Weakest Link, you need to increase the pot in the first half (so you want strong people to remain) and you need to win in the second half (so you need to get rid of the strong people).

After each round, everyone votes on who goes, and you say your vote in front of the audience. How people vote says something about their perceptions, since you only find out how many questions were answered correctly AFTER they decide who goes (and it doesn't always match).

The study looked at the difference between who got voted off and how they performed. For blacks, latinos and old people, there was a perceptible difference between candidate strength and whether they got voted off.

Each episode has a new cast of characters, so this isn't about keeping people on over several episodes. And it is the contestants, with a real financial motivation to make an accurate assessment of their fellow contestants who decides who goes.

[AP]

3. Huh???

Still don't understand after the edit.

[slick8790]

4. I actually just finished reading this.

I think that is definitely something to consider with polls with obama. Honestly, i have no idea how true that is, and it should not keep anyone from voting for him, but it's just something else that worries me. Same thing with hillary being a woman. Still...Gobama, GoHillary, and GoEdwards!