13 September 2012

SUSA CA poll / WAA and Crain's IL polls - interesting side data that pollsters write

SUSA, which one of the two best pollsters of 2008, is back with a vengeance and polling all over the place.

SUSA just put out a poll of California, showing Obama up +22 over Romney. That is in and of itself no surprise, though the internals caused my eyebrows to go up: in this poll Romney is getting 11% of the Black vote and 42% of the Asian vote, but Obama is getting 51% of the white vote. This is definitely unusual, but then again, California can be an unusual state. The +22 lead is almost identical to the last SUSA of the state, from 01 June 2012, which had Obama at +21. In that poll, Romney got 13% of the Black vote and 27% of the Asian vote - so the noted shift in the Asian vote is indeed interesting. And worth following in the next weeks.

This SUSA is also almost identical to the findings of the last USC/LA Times poll from 23 August 2012, showing Obama up on Romney by +23 (+22.90). In 2008, Obama won California by +24.03%, and his raw vote margin in CA was larger than Bush's national raw vote margin was for the entire nation in 2004.

So, the current polling of the largest state in our Union shows no appreciable erosion in support for the President. Indeed, the internals show that he has yet room to grow.


"Of those Californians voting "for" Romney, 48% tell SurveyUSA that their vote is really a vote "against Barack Obama" more than it is a vote "for" Mitt Romney. By contrast, of those Californians voting "for" Barack Obama, just 18% say that theirs is a vote "against Mitt Romney.""

That's right, 1/2 of voters who say they will vote for Romney are not voting for him, they are voting AGAINST the President. This means that Romney's support is not as solid as it looks on this survey.

As far as I know, this is the first SUSA survey to ask this question, and it would be interesting to see how it plays out in other states as well.

So, why I am bringing this up?

First, not every single poll deserves so much recognition and it is still the composite average that is much more important, but the polling out of both New York and California show clearly that a GOP wave is no where to be seen, not even in the slightest.

Second, think of the adage "a rising tide lifts all boats". Were Romney really gaining nationally, then in core DEM states that Obama would still win, he would be winning by decidedly lesser margins, and yet, that is not the case in CA, NY or for that matter, in IL. 


In IL, a rabidly Conservative pollster, We Ask America, was unable to get Obama under +17 in his home state. Obama won IL in 2008 by +25.11%, the largest winning percentage for any candidate in the Land of Lincoln since 1936. +17 sounds a lot less than +25, and it would be erosion, were this the case with all pollsters of IL. But just less than 2 weeks ago, in the middle of the Republican convention, Crain's / Ipsos showed Obama ahead in IL by +26, exactly in line with his margin from 2008.

So, how did WAA title its article about a +17 landslide poll for Obama in IL?

"Despite Dem Optimism, Obama Is In Tight Race"

And from the body of the text:

" The poll of 1,382 likely Illinois voters was taken Sept. 5 by We Ask America. It had Obama at 54 percent to Mitt Romney’s 37 percent. Another 3.33 percent said they’d be voting for an unnamed third party candidate and 6 percent were undecided.

That’s way below where Obama was four years ago, when he won Illinois with 62 percent of the vote. 

If you look at the 17-point spread between the two candidates, it’s a blowout, although not as big as his gigantic 25-point victory back in 2008. If you look at where Obama’s numbers are right now (less than four points above 50 percent), you might consider that this race could have the potential to tighten up quite a bit, at least compared to 2008. The spread is generally how polls are judged in the end, so if you’re a Republican you probably shouldn’t get your hopes up too much."

Compared to the Crain's / IPSOS text:

"Obama widens home-state lead over Romney"

And the text body:

"As Mitt Romney prepares to formally accept the Republican presidential nomination tomorrow night, he appears to be falling further behind incumbent Barack Obama in the president's home state of Illinois.
That trend could mean trouble for other Republicans down the ballot this November.

Mr. Obama now leads Mr. Romney by a ratio approaching 2-to-1, 55 percent to 29 percent, according to the latest Crain's/Ipsos Illinois poll. A total of 16 percent of the 600 voting-age residents surveyed said they're undecided or do not intend to vote.

The lead extends to every major region of the state, but is widest in Chicago, where the president is ahead by an extraordinary 71 percent to 16 percent, with 13 percent undecided/not voting.

The margin is 15 percentage points in Chicago's suburbs, and 12 points downstate. But in both cases, the president is short of the 50 percent mark, with about one respondent in six yet to decide or planning not to vote."


So, looking at the text body of what a pollster writes also tells us alot about whether is he being non-partisan or very partisan. The two examples from Illinois are very crass examples of two very, very different pollsters.

Compare this to Dean Debnam at PPP, who wrote in the last PPP poll of Missouri:

"One thing this whole episode hasn't done is soften voters' opinions of Claire McCaskill. Only 40% of voters approve of the job she's doing to 55% who disapprove. Even though 
32% of Republicans have an unfavorable opinion of Akin, only 11% of them actually say 
that they're going to vote for McCaskill.

Missouri continues to look like a tough sell for Barack Obama. We find him trailing 
Romney 53-41, little change from the 52-42 we found on our poll there this week."

No sugar coating it there.

We have now approached the point where the topline poll values are not necessarily enough - the internals also play a role, especially in battleground states.

No comments:

Post a Comment

Constructive comments and critique are always welcome. Please keep it polite and respectful.