Six Pollsters Can't Be Wrong! Bradley Effect Not Likely, Says Polling Panel
CNN Political Director Sam Feist moderated a really interesting panel at yesterday's Time-CNN "Politics '08" summit called "The Power of Polling: Daily Snapshots, Influencing Policy, Influencing the Election," featuring pollsters Evans Witt (principal and CEO, Princeton Survey Research Associates), Brad Kates (President and COO, Opinion Dynamics, and Fox News pollster), Bill Schneider (Senior Political Analyst, CNN), Stanley B. Greenberg (CEO, Greenberg Quinlan Rosner),
David Iannelli (Managing Director, Public Strategies Inc.) and Clifford Alexander Young (Senior Vice President, Ipsos Public Affairs). Wake up, that was the only boring part.
I videotaped one of the most non-boring parts: When Feist asked all the pollsters if they thought there would be a Bradley Effect in this election (if you don't know what that is, go here. And get out from under that rock). Interestingly, they all said no — no Bradley Effect (though Ianelli waffled a bit at the end there). The important takeaway here is not that there is no racism, as Schneider pointed out, but that they think the results won't reflect hidden racism. Schneider said that racism has been "cooked into the process" — that those people who won't vote for Barack Obama on the basis of race are largely declared; furthermore, pointed out Greenberg, the whole Jeremiah Wright imbroglio provided an easy handle for would-be Bradley-ers to latch onto (he called it a "gift" to pollsters for that reason, in the sense that it allowed for people to justify their vote, and thus declare it, providing that much more data for the pollsters to mine).
Also interesting: People who have cellphones instead of landlines are a whopping 20% more likely to vote Obama; people who live alone are more likely to vote Democrat than their married counterparts, across age groups; Sam Feist will be 40 on Jan. 21st — the day after Inauguration Day. When he presumably will finally have time for a nap. Clip of Feist's 'Poll of Pollsters' below; my live-Twitter notes below that; register your approval rating below that, in comments.
- Stan Greenberg: 3 things we're doing in polls: Weighting (a) age (b) percentage african amer vote (c) hispanic vote.
- Talking about the "Bradley-Wilder" effect makes me think of Van Wilder. Ryan Reynolds is Canadian.
- Ianelli: Turnout issues are going to overwhelm the effect. Youth turnout in key states etc.
- Sam Feist is jovial and funny. Younger than I recalled.
- Wow - all the pollsters agree that white male in Ohio w/landline diff from same sample without landline, but with cellphone.
- Cellphone-only vs. landline only: 20 pct more likely to vote for Obama.
- People who live alone more likely to vote dem than married counterparts, across age groups.
- Ianelli: If you have a landline, yr looking at an issue of household formation.
- Question: What are you worried that you've MISSED here? Kates: The undecideds. Greenberg: I'm worried that I'm not worried.
- Greenberg remembers back in the primary when Obama "couldn't close the deal." Less so now.
- Scneider: Saying there's no Bradley effect is not saying there's no racism. Bradley is hidden racism; he thinks racism is already in polls.
- Racism in the polls is why the race is so close. It's "cooked into the process."
- Greenberg: Rev Wright was a "gift" to pollsters - gave racists something to latch onto in declaring. Again: racism, unhidden.
- New voter/likely voter question: Will the actually show up to pull the lever?
- Question about exit polling. Greenberg: They're "so awful, so wrong."
- Feist: He won't have access to any polling data 'til at least 5 (by then at least two waves of polling will be in). Quarantine before too.
- Oregon has no exit polls - there's no exit (Oregon votes by mail.)
- Someone just asked Feist about fivethirtyeight.com; am eavesdropping, he doesn't seem as impressed as most.