Expanded Composite Primary Season Exit Poll

by: Chris Bowers

Tue Mar 04, 2008 at 14:43

Dreaminonempty has greatly expanded upon my work from the weekend, providing us with the following composite exit poll for the nomination campaign to date. Here is the remarkable data:

Democratic Nomination Campaign Composite Exit Poll
Demographic Electorate % Clinton % Obama %
Women 58% 50% 46%
Men 42% 41% 53%
Race / Ethnicity
White 63% 50% 42%
Black 19% 15% 83%
Latino 13% 63% 35%
Other 5% 61% 34%
Gender And R / E
White Women 35% 58% 37%
White Men 27% 44% 49%
Black Women 12% 17% 81%
Black Men 8% 12% 86%
Latino Women 7% 69% 30%
Latino Men 6% 58% 40%
18-29 14% 37% 60%
30-44 25% 40% 56%
45-59 32% 47% 48%
60% 27% 55% 39%
Annual Income
<$15K 7% 50% 46%
$15-29K 15% 50% 45%
$30-49K 19% 48% 47%
$50-74K 21% 44% 49%
$75-99K 14% 45% 50%
$100-149K 14% 42% 53%
$150-199K 6% 43% 55%
$200K+ 6% 48% 49%
Less than HS 5% 62% 34%
High School 19% 53% 41%
Some College 28% 46% 49%
College 25% 42% 53%
Post Graduate 23% 42% 54%
Democrat 77% 49% 47%
Republican 3% 29% 64%
Independent 19% 36% 56%
Liberal 48% 46% 50%
Moderate 39% 47% 48%
Conservative 12% 43% 48%
Protestant 21% 48% 46%
Catholic 28% 59% 37%
Other Christian 19% 35% 61%
Jewish 5% 56% 42%
Other 7% 39% 57%
None 14% 40% 47%
Marital Status
Married 55% 48% 46%
Not Married 38% 45% 50%
Married Women 32% 52% 43%
Married Men 26% 42% 52%
Single Women 25% 48% 48%
Single Men 15% 38% 57%
City 500K+ 17% 46% 51%
City 50-500K 18% 43% 53%
Town 10-50K 5% 42% 49%
Suburbs 45% 48% 47%
Rural 9% 51% 38%

These numbers suggest that an exhaustive, microtargeted poll of the electorate would prove extremely useful for both campaigns. If a campaign knew all of this information about any given primary voter, it could probably determine, with about 80-90% accuracy and without even a single voter contact, the candidate toward which that primary voter leans. This would prove extremely useful for making maximizing the effectiveness of voter contacts, and would be similar to what Republicans started doing in 2002-2004.

Also, how does this electorate compare to the 2004 primary electorate? Ron Brownstein sums it up:

The party is growing younger, more affluent, more liberal, and more heavily tilted toward women, Latinos, and African-Americans.

I dig that. After years of being told how we need to bend over backward to appeal to white, middle-class, conservative men, the party has grown larger by expanding its support within groups where it already held an advantage: liberals, women, Latinos, and African-Americans. Further, our newfound strength among youth voters also points to a possible generational shift in politics. Yet Further, our growth among affluent voters indicates changing values connected to the rise of the Creative Class. All your demographic trends are belong to us.

Oh, and if this data isn't enough to tide you over until the returns tonight, check out fladem's analysis of how late deciders have broken during the 2008 Democratic nomination campaign. The results might surprise you.  

Chris Bowers :: Expanded Composite Primary Season Exit Poll

Tags: , , (All Tags)
Print Friendly View Send As Email

I think your conclusions are very much on point (4.00 / 2)
and I'm very hopeful for the future of the party.  I've also advocated for increased outreach to people who rarely, if ever, vote, but would be sympathetic to progressive policy.  On the whole, it is generally easier and cheaper to register a new voter or turn out an irregular voter who is within your base (without giving up an inch on policy) than to try to swing moderate and conservative independents, which is harder, more costly, and could require significant policy compromise.  

For too long, Democratic leadership forgot that a vote is a vote, whether it comes from a white, middle class conservative male or someone more demographically suited to the Democratic base.  Obama, I think, has done a lot to help with this strategy, to which I think he is owed much credit.

good stuff (0.00 / 0)
thanks very much

Awesome (0.00 / 0)
Can anyone point me to some consolidated turnout data?  

"Don't hate the media, become the media" -Jello Biafra

Wow (0.00 / 0)
I don't know how Clinton is managing to lose this election winning both Married and Not Married voters.  :o)

ha! (0.00 / 0)
Yes, that would be a trick. Perhaps threesomes?


[ Parent ]
lol (0.00 / 0)
I guess with Romney out of the race they have to go somewhere...

[ Parent ]
Clinton isn't winning both (0.00 / 0)
Clinton gets 2% more votes from married people than does Obama.  Obama get 5% more votes from non-married people than does Clinton.

A more important, and easily understood category is men vs. women--this includes ALL votes, there are no exceptions.  Clinton gets 50% of the women vote while Obama gets 46%.  Women constitute 58% of all voters.  Clinton gets 41% of the men vote while Obama gets 53%.  Men constitute 42% of all voters.  The sum of (50x58)+(41x42) is how many people voted for Clinton, and equally true, the sum of (58x46)+(41x53) is how many people voted for Obama.

This means that Clinton gets 46.22% of the total votes cast while Obama gets 48.94%.  The remaining 4.84% goes to an unknown candidate or is uncommitted.

[ Parent ]
How is it changing over time? (0.00 / 0)
I had the (perhaps incorrect) impression that Obama was gaining among white voters gernerally, particularly women, certainly since Edwards dropped out.  Maybe she has stemmed the change, though.

On the marital status, there must be 7% "other"  (partnered?), and they must go for Obama, plus looking at married/unmarried vs the gender breakdown, 1% of the men and 1% of the women are missing.  That could be rounding.

John McCain--He's not who you think he is.

Gaining (0.00 / 0)
No one is really "gaining" anywhere. About 75% of voters went to the polls on Super Tuesday, so "gaining" or "losing" is difficult to day.

As for marriage, perhaps 7% just refused to answer.  

[ Parent ]
Your correction on the unmarrieds (0.00 / 0)
explains part of it, but I'm still intrigued aout the 7%.  Probably partnered, so find it difficult to answer yes or no.  I know I always struggle with the question, depending on whether it is a more or less socio/economic or legal question.  Here I might have said yes.  

John McCain--He's not who you think he is.

[ Parent ]
Evenly balanced (0.00 / 0)
It seems to me that on the whole voters in the Dem primary are evenly balanced between the 2 candidates. Each candidate having a small lead over the other in some segment of the voting population. Hillary was never the "inevitable" candidate and neither is Obama favored overwhelmingly.

Dem primary voters would be happy with either candidate when all is said and done. That's why it does not make sense for the campaigns to go excessively negative and focus on battering the other candidate. The only one that wins in that case is McCain as the Dem leaning voter becomes polarized and is unenthusiastic about the other candidate. A scenario with a decent probability as the primary campaign goes through Puerto Rico and into Denver with no clear winner and becoming more contentious by the day.

This would then naturally focus attention on all the machinations to get to the nominee. Not a pretty sight and a perfect platform for McCain to shore up his base and hammer the Dems, while the corporate media feast on all the discord in the Democratic party. Ultimately repulsing many Dem leaning voters and lowering their enthusiasm in November. If this scenario materializes don't count out McCain!

In the aftermath (0.00 / 0)
of the 2000 election, one of the writers at The New Republic write an article that made the argument that the McGovern coalition had won.

The argument noted that the Gore popular vote victory was less about changing margins within demeographic groups, but rather about the makeup of the demographics themselves.

The example I vividly remember was those with graduate degrees, who made up a substantial portion of all those who voted in 2000.  Gore carried that group by over 20 points.  In 1972, the group was too small to be statistically significant.  

All of this is in support of the data in this post.  

I'd like to see that article (0.00 / 0)
Because I wrote the exact same thing after the 2004 campaign: the McGovern coalition was slowly on the rise. This was slowly happening all along, but many Democrats were simply unwilling to wait for it. Their desire to run away from it turned the party to the right for a while, and we are still recovering from that era.  

[ Parent ]
If memory (4.00 / 1)
serves, it was December 2000.  I think it was written by Judis of "The Emerging Democratic Majority" fame, but I could be wrong.  I don't subscribe any more, but I will try to find it.

I read your article - which I completely agree with. Your work after 2004 was far more developed than the article I remember from the TNR in 2000.  

In my opinion one of the most important narratives to change was the "We must be competitive in the South" narrative, which was forever leading the party right.  

[ Parent ]
I felt some oversights were occurring (0.00 / 0)
Specifically that it's not just older white women favoring Clinton but the statistically important Latinas, too. If she wins Texas, that could prove the difference.

I'd like to see more demographic breakouts of certain categories, too. Like: of that under $30K groups, where do the men and women line up in their candidate preferences and how many under $30K are men or women? Because based on the advantage Clinton shows in that 22% of the population, that's her power center of all the economic classes.

Not only do I wonder as a history buff, but it's where I've lived all my life and I've never felt either Clinton represented my family's interests at all. (Not that Obama has, either, but he at least, will motivate me to vote for the top of the ticket on the chance that he might. With Clinton, I know I'll end up at the Mission, so there's no incentive at all.)

Thus, I'm curious why my income class peers feel so different overall, and whether certain demographics help explain that.

Still something overlooked (0.00 / 0)
It may be true, that "if a campaign knew all of this information about any given primary voter, it could probably determine . . . the candidate toward which that primary voter leans." But what is still missing is behavioral data that can help predict the likelihood of that person voting.

Knowing how recently and frequently a person voted, and what were the nature and number of contacts prior to that voting, are just as or more important in forecasting potential turnout. Why so many pollsters frequently overlook this behavioral data escapes me.

As expatriatedjerseyan says upthread, all this demographic profiling has to be put into the context of what "is generally easier and cheaper to do." And an overly of behavioral data is critical to pinpointing that context.

Save Our Schools! March & National Call to Action, July 28-31, 2011 in Washington, DC: http://www.saveourschoolsmarch...

ooops (0.00 / 0)
"overly" should be overlay

Save Our Schools! March & National Call to Action, July 28-31, 2011 in Washington, DC: http://www.saveourschoolsmarch...

[ Parent ]

Open Left Campaigns



Advanced Search

Powered by: SoapBlox