Wednesday, February 06, 2008

Demographics of Super Tuesday offer few surprises; Asians finally contribute to an exit poll

++Addition2++New Mexico's contest is officially called the "Presidential Preference Caucus" which explains why the AP, WaPo, Wikipedia, and MSNBC all refer to it as such. But the rules are pretty similar to a standard primary. As Hillary and Obama look to be just about splitting the vote at 13 and 12 delegates awarded, respectively, it doesn't much influence the primary-caucus disparity between the two.

++Addition++In the comments, Argent Paladin (what a cool name, btw) points out, among other insights, the enormous disparity between young white voters and older Hispanic voters in California. Among whites between the ages of 18-29, Obama wins 2-to-1 (63%-32%). Among Hispanics 60+, Hillary dominates almost 4-to-1 (78%-20%). The age trend among whites (younger voters more likely to support Obama, older folks to back Clinton) similarly exists among Hispanics, but it is much less pronounced (and is shifted a good 30 points in Hillary's favor).

In the bit on Jews, I overlooked Florida. In that state, Jewish (58%-26%, Clinton) support did pretty closely mirror white support (53%-23%, Clinton), with the former less supportive of the still-running Edwards than whites as a whole were.


Super Tuesday didn't bring many startling revelations in terms of demographics. However, I was surprised by how overt the bias in favor of Obama and McCain was on ABC and NPR, the two radio broadcasts I toggled back and forth between on Tuesday evening, though. I could feel Aaron Katersky, who was anchoring the coverage, try to cut off a putatively objective Alex Stone as he lectured California's GOP voters for being too restrictionist! NPR's Michel Norris, meanwhile, sounded almost orgasmic in pointing out that Obama had--as expected--won big in Colorado and Idaho!

Blacks overwhelmingly backed Obama, Hispanics favored Clinton (with the anomaly of Connecticut, where, comprising 6% of the total, they apparently preferred Obama by a narrow margin), and the larger the black share of a state's voting population, the more likely whites in that state were to flock to Clinton. Looking at all the contests that have taken place so far, there has been an inverse correlation of .35 (confidence just a hair outside 90%) between the percentage of a state's voters who are black and the amount of support Obama garners among whites in that state.

That's quite rigorous actually, given that obliterations like Illinois are outliers that attenuate the statistical relationship. Further, the real relationship is likely stronger than that, as I computed the results of all contests thus far (including those before Super Tuesday) as though only Hillary and Obama were running--in reality, most of Edwards' (overwhelmingly white) supporters in the southern states would have gone to Hillary if it had been a two horse race at that point. And I gave all the white undecideds in Michigan to Obama (neither he nor Edwards was on the ballot there), so he looked better among whites in that 23% black state that he would've in reality.

In the eyes of whites, Obama is only the Black Candidate when there are lots of blacks rallying behind him. Previously, I underestimated the extent to which this is the case, and am in peril of looking quite foolish in so firmly asserting Clinton would be the Democratic Presidential nominee. She likely faces four consecutive losses before reaching Virginia on the 12th, a state with a racial composition that has served her quite well thus far (a substantial black minority that nonetheless safely represents less than half of the voting population). If it is still 'deadlocked' by March 3rd, Ohio and Texas should be the final knockout punches for Obama, but the media fawning over his smaller state wins prior to those contests may be too much for her to overcome. For those who challenged my prediction, I'll humiliate myself more ostensibly when the time comes for it.

I still suspect she'll win, but I regret not taking sage advice and being a little more prudent. From this point forward, I'll try to stick to informed speculation rather than making bold assertions where uncertainty still remains.

Anyhow, there are a few notable observations to be made regarding the exit polls.

- The Jewish vote is inconsistent. In New York, Hillary beat Obama 2-to-1 among Jews (65%-33%). In New Jersey, her dominance was nearly identical, at 63%-37%. But in Massachusetts (48%-52%, Obama) and California (47%-49%, Obama), they split the Jewish vote, while in Connecticut Obama turned the tables with a nearly 2-to-1 victory, at 61%-38%.

It doesn't mirror the white vote, either. In NY and NJ, Jews were more supportive of Hillary than whites as a whole were. In Massachusetts, Hillary clobbered Obama among whites but narrowly lost among Jews. In Connecticut, she narrowly won among whites, but was clobbered by Obama among Jews. In California, the race was close among both whites and Jews, but Obama had the edge among Jews and Hillary among whites.

- That gallimaufry collectively referred to as "Asian" is solidly in Hillary's camp. In California, the only state in either party with a sizable enough number of Asian voters to adequately report exit polling data on, Hillary outdid Obama by almost 3-to-1 (71%-25%). In New Jersey, extrapolating from the other racial categories, the best estimate for the Asian vote (which comprised 4% of the Democratic total) suggests 59%-41%, in Hillary's favor. And she took American Samoa by a comfortable 57%-43% (the territory sends three delegates to the party's convention; Hillary gets two).

- McCain did better among whites and especially Asians (two-thirds of Asian Republicans backed him--not particularly surprising as Asians tend to shy away from conspicuously Christian politicians) than he did among Hispanics in California. Huh? Yeah, really. He took 66% of the Asian vote, 42% of the white vote, but only 39% of the Hispanic vote. It won't be much of anything the WSJ op/ed board will have to epicycle away (or ignore) though, as Huckabee and Giuliani took double-digit chunks of that vote as well.

- Steve Sailer mentioned Mitt Romney's strong performance in states holding caucuses. I'm looking forward to my state's run at it this Saturday. Although Paul's the guy I've been canvassing for, my little coalition is going to do all we can to get the various non-McCain factions to go after the McCain supporters. If Romney's crew has enough of a presence, though, I'm breaking for him. There's still a sliver of hope for a brokered convention, after all!

But what about the magi of the caci, Obama? He has been racking up lots of lopsided wins 0n the caucus side, coming out ahead in seven of the nine held up to this point. Tallying up delegates won by Hillary and Obama thus far, in each of the two electoral manifestations:

Primaries: Hillary -- 672 , Obama -- 621
Caucuses: Hillary -- 91, Obama -- 157

Are voters more likely to throw in with the 'oppressed' (white women are not as untouchable as black men are) candidate when others are watching them than they would otherwise be? Or are Obama's grassroot supporters simply more enthusiastic (they are younger, after all) and so show up at higher rates than Hillary's do at caucus time (primaries tend to bring out more voters than caucuses do)?

Uniquely, Texas may provide an answer. It holds a primary directly followed by a caucus that those who participated in the primary are eligible (but not mandated) to take part in. There are delegates to be won both from the primary and the caucus. Will the elderly pass on the caucus portion at greater rates than other age groups?


Razib said...

appreciate the analysis. i saw the outlines from the polling data, but you summed it up quite well.

Argent Paladin said...

You missed the importance of gender. In California, for example Obama actually won the male vote, by one point but got crushed in the female vote by nearly 2-1. Still not convinced that gender was the most important issue? Of those who said that gender was not important, Clinton won by 3 points. Those who said that gender was important favored Clinton by an astounding 47 points.
Of course, I found it interesting that Hillary also won by nearly 2 to 1 those who said race was important. I can only assume that those who said that race is important and then voted for the white candidate are either white supremacists or Asians or Latinos.
And when you intersect race and age we find some interesting things. First, under-30 whites LOOOOVE Obama (no surprise if you are on Facebook). They chose Obama at rates nearly two to one over Clinton. But, as much as they love Obama, old Latinos HAAATE Obama, choosing Hillary 81 to 17, which is about the biggest gap of any on the entire exit poll. And, there were almost as many old Latinos voting (5%) as young whites (65).

Fat Knowledge said...

If you had told me that Obama was going to lose California by 10 points, I would have told you that his chances of winning the nomination were 0%. So you weren't the only one who misread this primary. Right now I see it as a 50-50 race.

I still don't get how Obama wins the rural vote, but I find you idea that with no blacks around Obama is no longer the black candidate to be very plausible. I found this blog Rural Votes to have some interesting information (it is pro-Obama, not that there is anything wrong with that :) ).

While Obama take the young voters, surprisingly (well to me) he doesn't get the young Hispanic voters.

I am not sure how Obama won Connecticut and not New Jersey or Massachusetts, but maybe it has to do with the fact that Obama does well with the high income voters and there is a higher % of them in Connecticut.

I wonder too about the impact of caucuses vs. primaries. If the Midwest was all primaries, would he have done nearly as well? I don't think so, but I am not sure how big of a difference it would have made. Kansas going 74-26 Obama over Clinton is crazy. Maybe you have some insight on how that happened. No exit/entrance polls for caucuses so hard to know what is going on.

Good point on Texas being a caucus and the advantage that gives to Obama (I thought he won 8 out of 9 caucuses, am I missing one?). I think that could make a big difference. One other thing I just found out, is that while most states hand out their delegates on a proportional basis, Puerto Rico (the final election) is a winner takes all state with 60 delegates at stake. Obama better get the 'Sí, se puede' signs posted everywhere down there pronto.

Argent Paladin said...

correction: the percentage of under-30 whites was 6%. Also, the fact that women were 55% of the vote only magnified the gender effect. This of course is par for the course in Dem primaries. Old women and Latinos will secure victory for Clinton, with the help of college students without the foresight to register in their new state or get absentee ballots in time.

Argent Paladin said...

Comparing the exit data from CT and NJ confirms your observation. The only real difference between the two is that the Black vs White/Latino vote is more polarized in New Jersey than Connecticut. And, of course, New Jersey has a much higher percentage of African Americans voting.

Whites made up 82% of the electorate in CT and were evenly split 49-48 in favor of Clinton. AAs make up 9% and went for Obama 74-24. Latinos make up 6% and went for Obama 53-43. So a 1-50-10 difference for a total difference of 61.

In NJ, where whites made up 59% of the electorate, the total difference is more than twice as much at 141 (66-31 for whites, 82-14 for African Americans and 68-30 for Latinos).

Thus, it seems that in Connecticut, the small number of Latinos identify with the small number of African Americas, against the large white majority. However, in NJ, where those two groups make up more than a third of the population, they compete rather than identify with one another.
As I said in the first post, it is a white woman-Latino coalition that will put Hillary over the to of the young whites and African American coalition for Obama.

Audacious Epigone said...


Thanks. In the appreciation account, I'm still infinity-billion dollars on the debit side, though.


Thanks. The young white-old Hispanic split is especially interesting, and I've added it to the body of the main post.

In previous posts on exit polling (from Iowa, NH, Nevada, FL, SC) I've made note of the gender advantage among white females that Clinton enjoys. That didn't change, so I didn't point it out again.

Gender made little difference among blacks. That, again, shows the same pattern as previous contests have.

Re: the importance of race, it comprised 17% of the total electorate. Definitely would be interesting to get that broken down by race/ethnicity. Since Steve already pointed that out, I held off.


I wish I had some insight into why Kansas was so lopsided. Sebelius is a popular governor here, especially among Democrats, and she endorsed him. Also, in this socially conservative state where marriage, fidelity, etc is held especially sacred, Bill Clinton doesn't have much of a reputation. But those don't seem to offer a satisfactory explanation.

I have a Nicaraguan friend I'll see early next week who went to one of the caucus locations nearby in support of Clinton. Hopefully she'll be able to fill me in. Many locations were overwhelmed by people wanting to participate, and many voters had to wait outside in the miserable cold (it was in the teens Tuesday night, with howling northwestern winds that ended up dumping half a foot of snow on us) for an hour or more to get in. Maybe Obama's supporters packed the place early, and many Clinton supporters left after facing the raw elements for half an hour. As Clinton supporters tend to be older, they'd probably be less likely to tolerate the elements for long.

And yes, the lack of exit poll data for the caucuses is maddening. Of the Super Tuesday states, only NM provided one. I'm especially interested in the ages of the participants in these whitebread caucuses. I suspect the voters tend to be younger than in primaries, but there's not really any way to tell at this point.

Re: caucus victories, Hillary has won Nevada and New Mexico (by a hair). Obama has taken the rest.

Puerto Ricans have more black ancestry than other Hispanic groups in the US. I wonder how that will play out, assuming the contest is still going strong at that point.

Diana said...

My guess about the Jewish split is that Hillary does better among Jews who are scared of the economy, and Obama does better among Jews who are more secure economically.

Also the Iraq blowback thing. Lack of fear about economic insecurity frees up psychic energy about foreign affairs - Connecticut and Massachusetts Jews are more concerned about how Iraq blowback might lead to people blaming the Jews for the rotten war.

(McCain's support among Jews is so transparently clear I need not elucidate it here.)

Audacious Epigone said...


Is the Jewish economic situation that much different in New Jersey and Massachusetts?

None of the exit polling data contains a sizable enough Jewish contingent on the GOP side to offer anything, but it's hardly a mystery. I would be interested in seeing what percentage of Jews backed Huckabee though, out of curiosity more than anything.

Diana said...


"Is the Jewish economic situation that much different in New Jersey and Massachusetts?"

I have no data - just lifelong experience as a New York Jew who moved from "one of the boroughs" to Manhattan, with relatives in NJ and Long Island, and lots of exp. traveling in New England. So these are just impressions. Most Jews in all 4 states are prosperous, but here are some distinctions:

1) NY Jews: mixed bag - small business owners mixed in w/professionals. And many of the "professionals" are public school teachers, not Wall St. moneybags. Professor at a community college, city college, state college, or state university. Home worth $500K to $1M. Most of wealth is in home.

2) NJ Jews: refugees from NY, same demographic.

3) Mass/CT: rich or top professionals. Professors, Wall St. bonus babies, film execs. Harvey Weinstein lives in Westport, CT. Professor at Harvard, Yale, fancy private college, etc.

Home worth $1K +. (Hell, the prop. taxes in Westport are $50K per year.) Several million+ in other assets.

Again, I have no data but these are my lifetime impressions.

"I would be interested in seeing what percentage of Jews backed Huckabee though, out of curiosity more than anything."

Me too. My guess would be 2%. Jewish Repubs went for McCain/Romney, in that order.

Diana said...

In short, among Dems Hillary is the candidate of people who are *scared* of the economic future, and they were reasonably satisfied with the performance of her husband from 1992-2000, and want them both back. He's a huge part of their appeal, and they don't give a stuff about "racially coded remarks."

Geez, it might even increase his appeal!

My big surprise factor is how poorly Romney fared. I thought for sure that the Repubs would recognize a winner and line up behind him. Boy, was I wrong!!

Audacious Epigone said...


The NY explanation certainly seems plausible, and if NJ Jews are cut from the same cloth as NY Jews, as opposed to highly successful Connecticut Jews (and, from your experience, Massachusetts as well), it's pretty satisfying given the limited data.

What is the Jewish take on Mormonism? Essentially the same as it is on other conservative Protestant denominations, or is a distinction made? Would that have effected Romney among Jews more/less than it did among other voters?

Diana said...


My gut feeling is that most Jews think that Mormonism is just bizarro-world nonsense. They may disagree w/fundamentalist Christianity, but they don't think it's a cult from bizarro world. In that sense, aren't they like a lot of other Americans?

That said, if Romney had just run a better campaign, Jewish Repubs would have looked at his record in Mass. and found him acceptable.

As if you needed another wrap up: NY/NJ Jews are middle-middle class to upper-middle class (w/a sprinkling of megabucks types, in a few select suburbs and of course, 5th-Park Avenue), whereas CT/MA Jews range from the solidly upper-middle to upwards...

The difference between a middle-middle class professional depedent on the vagaries of macroeconomic trends, with a $1M home and two/three kids' college education, and a truly wealthy man who commands resources, is vast.

The former may appear to most observers to be nicely set up in life, but the fact is, these guys are scared spitless about the future. They may not BE insecure - but they FEEL insecure. And they feel insecure about their kids, because their kids future is all based on leveraging opportunities.

Fat Knowledge said...

I came across this post that says that Obama does well in states that are less than 5% black and those over 20% black, but not those in between. I think that goes to your premise of when he is and isn't the black candidate.

New Mexico was actually a primary.

Interesting point about Puerto Ricans having more black ancestry. I wonder how that will play out. I wonder how black Hispanics have been voting so far and what percentage of all Hispanics they make up.

Audacious Epigone said...


Ah, the NM vote is called the "Presidential Preference Caucus", which is why WaPo, Wikipedia, the AP, and msnbc all refer to it as a caucus. But the rules seem to be pretty similar to a standard primary, not a caucus.

Thanks for spurring me to the bottom of that. It doesn't do a whole lot in terms of the caucus/primary results between the two, since it looks like Clinton will get 13 and Obama 12.