Monday, August 31, 2009

2012 Presidential Trial Heats (Clarus Research Group): Obama vs. Gingrich/Huckabee/Palin/Romney

Last night we looked at the 2012 Republican primary numbers from the recent Clarus Research Group poll. The same poll that found Mitt Romney doing quite well among Republican respondents in the primary setting also found the former Massachusetts governor matching up against Obama the best among the full national sample.*

...but not by much. The Mike Huckabee/Barack Obama pairing was a near mirror image of the Romney/Obama race (Obama's support against Romney was just one percent shy of the president's against Romney's while the two Republicans' shares remained the same). Before we dig too far in, let's look at the toplines:
Obama: 52%
Gingrich: 34%
Not Sure: 15%

Obama: 48%
Huckabee: 38%
Not Sure: 15%

Obama: 53%
Palin: 34%
Not Sure: 13%

Obama: 47%
Romney: 38%
Not Sure: 15%

Margin of error: +/- 3.1%
Sample: 353 Republican voters
Conducted: August 14-18, 2009
[Click to Enlarge]

For Gingrich, this is the worst showing since the May Public Policy Polling survey (the one that had the most representative sample until the August poll). There really isn't that much to say about the former Speaker. He has been polled among the Big Four throughout 2009, but has yet to measure up to the Huckabee/Palin/Romney group either in the primaries or against the president. That certainly doesn't change here.

[Click to Enlarge]

And while Huckabee does drop off compared to the PPP poll in the field nearly simultaneously, the former Arkansas governor emerges with something to hang his hat on. Earlier we discussed how this poll had Huckabee as the clear choice of Republican women in the primary race which is noteworthy with a woman in the race. But that doesn't translate once a Democrat is added to the equation. Now, we would expect the gender gap to continue to focus female support behind the Democrat and male support with the Republican in any prospective 2012 match up and that trend is not broken here. However, we do see that the gap is smaller for Huckabee (-20) and Romney (-18) than it is for Gingrich (-34) and Palin (-37).**

[Click to Enlarge]

Odd as it may seem, this is the continuing trend in this data as of now. Palin is doing worse among women than with men against Obama and has the highest gender gap (relative to the other Republican candidates). And that is striking. What's worse is that the overall gap between Palin and Obama is now wider than it was prior to her resignation. The three August polls conducted pitting the former Alaska governor against Obama, show the governor under 40% and the president above 50%. That wasn't the case after she announced her resignation, but actually having step down (at least at first glance) seems to be the line of demarcation between these poll trends.

[Click to Enlarge]

Meanwhile, Mitt Romney continues to place, if not the closest to Obama (Huckabee has been consistently closer in the PPP polls.), then second closest. And the former Massachusetts governor has been among (along with Huckabee) the candidates who have been able to minimize the amount of support the president garners in these polls.

The biggest pro and con in all of this 2012 trial heat polling is that one firm (Public Policy Polling) has been doing most of the work. That is good for the sake of comparison across polls, but by the same token is subjecting us to the same house effect over and over again. In other words, we're just missing out on the view from other polling outfits. That's why the Obama/Romney tie in the July Rasmussen poll was so surprising; it followed up the PPP polls that showed Romney close, but struggling to break the 40% barrier. Again, the one thing Romney could hang his hat on there, was that he was keeping the president under 50% in many of those polls. And honestly, that's not a very strong hook and the hat is somewhat heavy.

I still have to caution that this is all extremely early in the 2012 cycle. But the trend among women with Sarah Palin included continues to be a noteworthy statistic across all of these polls.

*Please note that the full national sample size was used in the write up of the Republican primary poll last night. That has been corrected. The Republican portion of the sample was comprised of 353 respondents.

**This figure is calculated as the sum of the deficits among women and men for each candidate against President Obama. For example, in a hypothetical match up between Obama and Minnesota governor Tim Pawlenty, if the president had a 53-37 advantage among women (-16 for Pawlenty), but the governor had a slight advantage among men, say, 44-42 (+2 for Pawlenty), the gap would be -14 (-16+2).

Recent Posts:
Have the Races Changed in New Jersey and Virginia?

2012 GOP Presidential Primary Poll (Clarus Research Group): Romney Jumps

Defense Authorization Bill Amendment Could Affect Primary Timing


Robert said...

Interesting trends. Obama's popularity goes down, but he does better against all comers.

Josh Putnam said...

I wanted to play devil's advocate with your statement, Rob and that made me go check the report again. In actuality, you're comment made me want to compare the partisan breakdown from other recent polls like this and see how Clarus stacked up.

My thinking was that this could just be an overwhelmingly Democratic sample.

Well, that isn't entirely clear. For starters, Clarus doesn't provide the partisan numbers. This is why PPP does such a bang up job: they disclose everything. So I may poke at their polls from time to time, but that is because they show it all (or comparatively more). This Clarus poll isn't up front about the partisan distribution. They make you hunt for it. Even then it isn't clear.

For example, the GOP primary numbers are based on 353 Republican respondents. But there is also a question that was asked only of the 595 Democrats and Independents in the sample. That sums to 968; 35 respondents below the stated size of the sample (1003). That's fine. I'm sure it is nothing more than survey attrition. Some chose not to answer.

But there's more. The report also references a question that was asked of 418 independents. ??? That would leave only 177 Democrats, in which case there is no way the results should show what they do for 2012 (A 2:1 Republican advantage coupled with independents that don't approve of the president does not add up to the 2012 margins the president had over the Republicans).

Of course, that 418 independents is likely derived from the base question before the leaner question is added. But the point is, we don't know. Does that make this a good or bad poll? I don't know, but it is a 2012 poll and I'll look at.

...and point out the flaws as I see them.

Incidentally, Nate Silver had an interesting piece out yesterday that combines both elements of your comment, Rob: presidential approval and the 2012 polls. The bottom line: Obama's popularity will have to go even further down for him to lose and the GOP would probably be best served by going with someone outside of the quartet that has commonly been listed. It is worth a read if you haven't see it already.

Robert said...

Very interesting. If the economy springs back, I think Obama's popularity will also bounce back.