CA-Field: Cruz +2

[Mr. Morden]

California polls tend to be like the national polls. All last year the Field polls had same leaders as national polls, albeit the percents are lower due to more undecideds.

I disagree with that.  This is the fourth poll of California since August (the other three being another Field poll, USC/LA Times, and USC/LA Times/SurveyMonkey).  Those other three all had Trump leading by between 1 and 6 points.  That's a much narrower lead than he had nationally at the time.  I think it's always been a relatively weak state for him.  Other polls also show Trump comparatively weak in the West, and of course California is a pretty big chunk of the West.

[Mr. Morden]

S*** poll (Carson in 4th place, 41% undecided LOL).

There aren’t 41% undecided in the poll.  With Cruz at 25%, Trump at 23%, and Rubio at 13%, that’s 61% just for those three candidates alone.  How could there still be 41% undecided?

[Mr. Morden]

Full #s:

http://www.field.com/fieldpollonline/subscribers/Rls2521.pdf

Cruz 25%
Trump 23%
Rubio 13%
Carson 9%
Paul 6%
Bush 4%
Christie 3%
Fiorina 3%
Kasich 1%
others/undecided 13%

2nd choice:

Cruz 22%
Rubio 14%
Trump 11%
Christie 8%
Fiorina 8%
Carson 7%
Bush 5%

Who leads among…?
coastal counties: Cruz/Trump tie
inland counties: Cruz
under age 50: Cruz
age 50-64: Trump
age 65+: Trump
strongly conservative: Cruz
not strongly conservative: Cruz/Trump tie
born-again Christian: Cruz
not born-again Christian: Trump

fav/unfav % among GOP primary voters:
Cruz 69/20% for +49%
Rubio 61/26% for +35%
Carson 60/30% for +30%
Christie 52/33% for +19%
Trump 51/45% for +6%
Bush 38/58% for -20%

[Mr. Morden]

Weird: they break down favorability for the top three candidates by race, but they don't give the breakdown for race in the actual first-choice vote (Table 3 and Table 6 on the PDF). Can anyone think of a reason to do that?

The favorability by race is among all California voters, not just GOP primary voters.  The GOP primary presumably has too few minorities to get statistically meaningful racial crosstabs.  I assume they're preparing to release their Democratic primary #s, as well as general election #s, and we'll get more info at that point.

[Mr. Morden]

Thanks for both the response to the post and for putting up the poll's PDF. In terms of minorities being statistically minimal for the GOP, based on my research this seams off. According to http://elections.nytimes.com/2012/results/president/exit-polls it ended up being as follows for California in 2012:

White: 53-45 Romney
Black: 96-3 Obama
Hispanic: 72-27 Obama
Asian: 79-21 Obama

From that we can multiply those numbers by the share of voters:

White: 53*55=2915
Black: 3*8=24
Hispanic: 27*22=594
Asian: 21*11 = 231

And then divide the individual shares from the total and we get (any other group is removed here):

White: 77%
Black: 1%
Hispanic: 16%
Asian: 6%

While the white population is still far larger than any other group, the smaller groups are not really small enough to say that they are statistically insignificant.

The 2008 GOP primary exit poll was:

http://edition.cnn.com/ELECTION/2008/primaries/results/epolls/#CAREP

white 76%
Latino 13%
Asian 6%
black 2%
other 3%

So, OK, sure, you can get a crosstab out of the Latino vote.  But Asian, black, and other?  Each of those is probably too small to get a meaningful crosstab in a poll like this.  The MoE would be huge.  That's not to say that those voters won't matter, but in the polling, those are pretty small crosstabs to get a good MoE on.