Because We're Still Exhausted "Outliers"
Mark Blumenthal | November 7, 2008
Topics: Outliers Feature
ABC News, Associated Press, Ron Brownstein, Andrew Gellman, The New York Times. Politico, Bill Schneider, The Wall Street Journal, and The Washington Post all review the exit poll results (h/t: The Page, Sullivan).
Jennifer Agiesta reports exit poll results showing 20% of voters were cell-phone only and finds that weighting by age "mostly accounts for the lack of cellphones in many polls."
[Update: AP's has more detail on the exit poll results on cell-phone-only voters].
Gallup reports on reactions to Obama's election, looks back at his winning coalition and relaunches the Gallup Daily.
Obama pollster Joel Benenson shares observations with Time (via Smith).
Gary Langer says lists of the most accurate polls are "an exercise in vacuity of the highest order."
SurveyUSA is posting state-by-state pollster report cards.
The New York Times reviews the Bradley Effect that wasn't.
AP sees no Bradley Effect either.
Fox News reviews how the pollsters did.
The Las Vegas Sun examines the Obama Effect" in Nevada.
The Anchorage Daily News asks why polls missed the mark in Alaska (via OpenLeft).
PPP defends itself from charges of bias.
Carl Bialik reviews turnout estimates.
Karl Rove still doesn't trust exit polls.
Kidzui's kids voted like the adults.
And yours truly comments on how pollsters did on NPR's All Things Considered and WNYC's Brian Lehrer Show.
By Mark Blumenthal | November 7, 2008 4:53 PM | Permalink | Comments (9) | TrackBacks (0)
Comments
How about that Nebraska electoral vote split? http://www.omaha.com/index.php?u_page=2835&u_sid=10481441
I found a wonderful article in MSNBC :
http://www.msnbc.msn.com/id/27582147/
"Young voters not essential to Obama triumph. Analysis shatters exit-polling myth, but shows black voters were vital."
Ana Maria Arumi, who directs the exit poll desk for NBC, MSNBC and Telemundo, has done the calculations based on the exit poll data, State by State.
She shows what happens with Millennials, Latinos, etc ...
Only Blacks were essential to Obama's Presidency and Electoral Votes, not Youth or Latinos.
I continue doing Analysis and gathering articles of Youth, Blacks, Latinos, etc ... for Politics :
Vicente Duque
Vicente,
Arumi's point is technically accurate, but not very informative. The fact is that no Democratic Presidential candidate has won the white vote since LBJ, so it's not surprising that taking away African American voters would deprive Obama of his victory.
More generally, I find statements that a certain voter group provided a candidate's "margin of victory" to be problematic. Despite the overwhelming support Obama received from African Americans, they still made up less than a quarter of Obama's winning coalition; white voters made up 60%.
The African American votes can only give Obama the winning margin if he has the large base of white (and Latino) voters to begin with.
Focusing excessively on target demographic groups is antithetical to Obama's actual strategy, which was to build a broad, winning coalition across diverse groups.
Well said 4xfloor.
To 4xfloor and BitJam :
Thanks for your interest and comments. This election gives us material for many months or years of data analysis.
This is material for predicting the political future, based on demographic trends, and other trends like : urban behaviour, industrial change, service economy, technologies like internet, professionalism or graduate and postgraduate ideology.
It has been alleged recently that the Same Sex marriage propositions, and the abortion wars are not finished, but represent future issues and material for future politics.
I am happy to see that someone here in pollster has touched one of these hot topics ( the same sex marriage ).
Thanks again
Vicente Duque
I'm not sure it's understood that wagering is still available on Intrade on the three unresolved senate races. That's the beauty of man-to-man speculation as opposed to fixed odds.
Many progressive sites believe Franken will prevail. I understand that logic but in this case I agree with the Intrade consensus that Coleman is the proper favorite. There isn't a stash of uncounted votes from a specific favorable area, similar to what bailed out Gregoire in the 2004 Washington recount. It's a 3-way race and we're counting on one specific candidate to benefit by more than 200 net among an undetermined number of ballots that may or may not contain an attempted vote.
These are the respective rounded odds, for a reference point on theoretical likelihood:
* Alaska -- Stevens slight favorite, roughly 53-47 (big move on Begich today, Stevens dropping from 80% range)
* Georgia -- Chambliss 85%, Martin 15%
* Minnesota -- Coleman 64%, Franken 36%
It will be interesting to watch those Alaska and Minnesota numbers fluctuate wildly on Intrade, once indications are leaked from the resumed count in Alaska, and the Minnesota recount.
Also, I don't think I mentioned the amazing change in the Nevada electorate, from 77% whites in 2004, which matched the national percentage, to 69% this year. That reality was game-set-match before the first vote was counted.
I just wanted to issue a sincere thank you to Pollster.com for all of the hardwork that has gone into covering the longest election in US history. Certainly no small task, but yet more impressive because it was done so well.
Mark, you and the rest of your team have earned a nice vacation to some remote tropical island to take a break for about six months.
Thanks for everything that you've guys done. Pollster.com has made it easy and fun to follow the twist and turns of this election, and I am truly thankful that I get to enjoy the fruits of your hard labor. And, I truly look forward to using Pollster.com as a resource for many election seasons to come.
Thanks again,
Skippy
To Gary Kilbride : thanks for informations and comments on Nevada.
Frank Rich, New York Times gives this Information :
"The only states where the G.O.P. increased its percentage of the presidential vote relative to the Democrats were West Virginia, Tennessee, Louisiana and Arkansas. Even the North Carolina county where Palin expressed her delight at being in the "real America" went for Obama by more than 18 percentage points."
Vicente Duque
Thanks for all the links. That was almost too much, In fact, it was too much. I'm still allocating energy. Even Obama looked tired today, with forced smiles and without the boyish JFK spark I expected in his first press conference. Strangely serious.
NPR could make Charles Manson sound suave. That's not a wisecrack. I have no idea how they manage it but that network is so smooth it almost puts me to sleep and off the road when I listen while driving. Same dynamic while listening to that clip today.
Kudos to Politico for mentioning the white female margin, and wondering if credit doesn't belong to Palin for tilting it in McCain's direction. I've posted the same thing for days on many sites, while the fashionable course is to rip her and ignore that 53-46 gap, which doesn't align with the other major voting blocks in a 6+ Democratic nod.
Nevada was indeed remarkable. I couldn't forecast it and can't fully explain it. None of the political pundits here were forecasting that type of walkaway. There were hints from New Mexico, the jump to 24% liberals in the 2006 exit poll. That's the precise number that always indicates a state is getting away from swing state status. Nevada in 2006 demonstrated increased Hispanic participation by several percent, and 4 Democrats won among the 6 statewide races as opposed to zero in 2002. But Clark County was still under performing and the high profile senate and gov races slanted easily to the GOP, notably the clod Gibbons winning by 4.
The two things I did notice:
* Personal contact from the Obama camp. Previously I'd receive mailings but no visits. This time I had 3 knocks on the door within about 10 days in early to mid October, seemingly destined to continue until I emphasized they needn't worry about my vote.
* The GOP campaigns were fractured and lame. Normally in Nevada the ace attack dog Mike Slanker forms Republican campaign themes and they are devastating, much more impressive that the ones I see nationally. He's hoisted Jon Porter several times, with no ammo to work with. This cycle Slanker left with Ensign to run the GOP senate effort, which didn't work out well and left Nevada less fortified than normal. The GOP attacks seemed forced and came much later in the cycle than typical. Normally they identify a defining theme two months in advance and plant a simple and memorable disqualifying phrase into the local vocabulary. This time I wondered what was going on. The assaults came late, and were mostly a bland rehash of 2006, "Dina Taxes" for Dina Titus, etc. Some of my GOP friends in sportsbooks had instinct that it was pathetically ineffective. On Monday night one of them told me, "Obama and Titus will win here by much higher than you expect."
Posted on November 7, 2008 10:16 PM