CA: Rasmussen: Clinton Leads CA by 5%; Obama Leads by 7%

[Alcon]

John, I think you're projecting the primary to the General election in a wrongheaded way re: Latinos.  Even considering the Latino protest shift, national polling has Obama doing essentially as well among Latinos as John Kerry did.  The GOP has not exactly endeared themselves to the Hispanic voting base in four years.  McCain's moderate stance on immigration will help, but I doubt Obama will perform much worse among Latinos than Kerry.

Taxes were an issue levied against Kerry too, and with little success.  Invariably, Obama's best group that Clinton does not win tend to be affluent, well-educated males.  He's polling better among them than Clinton by a good deal, and better than Kerry too.

I suppose he could collapse, but I'm always skeptical of "once they know what he's really like..." arguments.

[Alcon]

(Stuff about Hispanics)

It's a problem for him, but hardly crippling.  Those are still heavily white states electorally; Colorado was 92% non-Hispanic in 2004, Nevada was 90%, and New Mexico was 68%.  And New Mexican Latinos are oftentimes quite different from Latinos in other states, but that's beside the point.

Clearly, he is polling ahead more often than not in all three of the latter states, so he is evidently doing fine among someone.  Taking polls to project a weakness among a certain group in a state, and then ignoring the actual polls from that state, doesn't make much sense  This is even considering that Hispanics are probably substantially more likely to be Clinton supporters, and thus make protest votes in polls.  So, being up in most polls out of Colorado now hardly spells "trouble."

Moreover, the exit polls are still types of polls.  If Kerry underpolled among Hispanics, and the exit polls were adjusted to his final margin of loss, obviously he overpolled among some other group.  So, if Obama was polling identically to the exit poll results, he'd be doing worse among Hispanics than Kerry did but better among another group (wealthy whites, maybe?)

Obama is doing well among affluent well educated Democrats.  I don't dispute that.  But that is very different from sbane's claim that Obama will do well among rich white people in Orange County.

Have you been paying much attention to the crosstabs on national polls this year?  It's been a while since I had access to subsamples to look at, but Obama is improving among:

1. Post-grads (and to a lesser degree grads in general)
2. Those with household incomes over $100,000

Which in fairness is the same group that Democrats have been improving in since 2000, but I see no indications of reversal.  Maybe with a Clinton candidacy.

I'm not making a "once they really get to know him" argument.  Voters in California already know Obama and McCain pretty well.  And Obama is only ahead by 7% in an exceddingly Democratic year.

We haven't had many polls yet, and it's early, so I'm hesitant to project.  Earlier SUSA polls had him as far up as 27 points (February), and - understandably considering they tend to poll after major news events - it's swung around.  Could California swing toward McCain?  Quite possibly.  He's a good candidate for a state like California.  The rest of your arguments I find questionable.  So, on to those...

And remember, Obama never does as well when people actually vote as the polls say he will.  In California, the RCP average had Obama up by 1.2% on election day.  He lost by 9.6%.  That's means Obama did 10.8% worse when people voted than he did when they answered the polls.  If that happens in November, there are going to be a great many nasty surprises for Demcorats.

You're incorrect on the first count.  Mr. Morden or someone made a nifty chart.  There are states where he overperforms, and ones where he underperforms.  Alabama (hilariously) comes to mind.  The polls averaged out to a tie; he won by 14.  You're thinking leaked exit polls, in which case you'd be right.

SurveyUSA's final poll had Clinton up 10 in California anyway, and the RCP numbers were polluted by: 1) Zogby's hilarious Obama +14 poll; 2) Suffolk, a bad university poll; 3) Rasmussen, who did not handle early voters correctly and admitted it.  No one who was reading the polls with any real attention would have trusted the RCP average.  Besides, Obama did do much better on election-day voters than early voters.  Remember when the results came in?  He flipped some counties that were nearly 3:2 in early voting when election-day results came in.  That phenomenon was certainly in effect in the polling.

What I'm saying is that Demcorat are about to nominate a horrible, horrible candidate.

Since the polls aren't supporting that, that sounds remarkably like a "once they really get to know him" argument to me.

[Alcon]

Ah, Alcon.  Still impervious to evidence, I see.

What did I do to deserve that mean-spirited rhetorical flourish, besides disagreeing with you? 

Lets talk for a moment about the California Democratic primary polls.  Obama in fact did outperform the polls.  Lets exclude Zogby.  Field had Obama by 2.  Suffolk had Obama by 1.  Rasmussen had Obama by 1.  Even if you disregard Zogby, Obama still did much better in the telephone polls than he did in the actual results.

This trend is duplicated in most of the states to vote.  Obama underperformed the RCP average in Arizona, New Jersey, Ohio, Texas, and Pennsylvania.  He underperformed in Massachussetts by 8%.  He underperformed in New Hampshire by 10%.  And of course, he underperformed in California despite your protests to the contrary.

Field also had 18% undecided, and 12% "other."  Obama did underperform in California.  However, as you are (for some reason) ignoring in your analysis, he did better on election day than beforehand.  Suffolk is a university poll.  Considering this, and the flaws in using the RCP average alone, he underperformed less than you were trying to demonstrate.  That may seem rationalized in hindsight, but if you look back at my posts, I imagine I said that all before the primary too.  I may be a crappy political analyst, but objectivity is not generally my weakpoint.

Ignoring the California thing, or the (fairly unsupported, IMHO) argument that Clinton receives most late-breakers in almost every state, let's look at your assertion.  I suspected you were wrong, but I didn't know offhand, so I did the math, and your assertion is indeed incorrect.

First, I took the primary results and compared them to RCP averages.  I excluded traditional caucuses because, obviously, that gives Obama an unfair advantage in my analysis.  The RCP site is a little confusing, so I'm sorry if I missed any polls.

Outside of Super Tuesday, Obama was underestimated in margin in 8 of 11 states.  Clinton was underestimated in three.  In Obama's states, the average underestimation was 7.6 points.  For Clinton, the average was 5.2.  Overall, the error was a 4.1-point under-estimation of Obama's margin.

For Super Tuesday, I did the same analysis.  There were eleven states that weren't caucuses and had enough polls to be included.  Clinton was underestimated in 6 of 11; only 4 of these were by more than one point.  However, the average Obama underestimation was significantly higher than the average Clinton estimation.  Overall, the error was a 2.3-point under-estimation of Obama's margin.

In total, Obama was under-estimated in 14 of 22 states, and the average error was 3.2 points.  You can check my math.  It is correct.

1) He can add Ohio.  This would get him to 279.  This is a problematic strategy because of Obama's deficiency with the white working class vote.
2) He can add Colorado plus either Nevada or New Mexico.  This would get him to 273.  Unless he gets well over 60% of the latino vote, it will be hard for him to do this.

Again, we have plenty of polls out of these states, so you can keep doing subgroup theorizing or actually read the polls.  Unless, for some reason, you think that these polls are mis-handling Hispanics badly, I don't see any rational reason not to look at the polls.

My favorite statistical analysis site, FiveThirtyEight, has them all rigged up for you, with weights and everything.  His conclusion is that Obama has a 60% chance in Colorado, 52% in New Mexico and 48% in Nevada.  Clinton ranks 6%, 42% and 19%, respectively.  You may disagree with his methodology (I think it's solid but you can take a look), but at least that will provide you with a link to the polls that show Obama performing well in the region.

[Alcon]

Also, those rich, affluent white people in Orange and San Diego counties are Republicans, and after all the gaffes, I doubt they are going to be tempted to go out and vote for Obama over the very acceptable McCain.

There are plenty of affluent California voters who are Democrats.  In 2004, the California exit poll had Kerry win those in $100,000+ households by four points.  It's a sub-sample, and exit polls are often wrong, but wealthy people in CA aren't anywhere near as Republican on the whole as the suburbs of San Diego and Orange County you're talking about.  Although, yes, there are some incredibly conservative suburbs in California that many people improperly ignore.

Again, the upper middle class tax issue (if I'm remembering correctly) was an issue Bush tried to levy against Kerry in 2004.  It was one of the few things that didn't seem to stick at all relative to 2000.  There's a reasonable argument that this was solely because of Bush, but that's a discussion for a different day. 

[Alcon]

To be fair (and again, I hate being fair), all indications way back when were not that Obama was going to obtain ahistorical support among Latinos.  They voted against him 2-1 in the primaries (versus 2-1 for him in the general), and McCain historically had been one of the most pro-immigration Republicans.

Few predicted that Obama would do as well as he did among Hispanics back then

Even back then, General election polling was pretty much showing the GOP brand was deathly tainted among Hispanics.  Of course, all the pundits were totally ignoring that, but the signs were definitely there.  I even said as much in this thread.  That's why the Nevada polling at the time was so f'in puzzling.  Demographically, Obama being 5-6 points behind there made no sense.