Friday, July 24, 2009

Oops! A 2012 GOP Primary Poll FHQ Missed and Another Rant on the Over-Interpretation of These Polls

Home renovations like the ones FHQ did in mid-May can put a damper on your 2012 poll-watching in a heartbeat. And apparently my blogger-turned-handyman days caused me to miss one of the 2012 GOP primary poll conducted by FOX [pdf] during that period.

Excuses, excuses.

Anyway a hearty thank you to GOP12 via CQ PollTracker via GOP12 for the belated heads up. For the record, here are the particulars:

Huckabee: 20%
Romney: 18%
Gingrich: 14%
Palin: 13%
Giuliani: 12%
Sanford: 4%
Bush: 3%
Jindal: 3%

Margin of Error: +/- 3 points (+/- 6 points among Republicans)
Sample: 900 registered voters (274 Republicans)
Conducted: May 12-13, 2009

I'll skip the analysis and leave it at this: This is the only primary poll thus far that does not have Palin clustered at the top with Mitt Romney and Mike Huckabee; well above everyone else. [And yes, how quaint. Mark Sanford was included -- pre-Argentina.]

Sadly, with Giuliani and Sanford now tacked onto the list of candidates, the key took up too much room and the color scheme Google Docs provided was repetitive and confusing. In sum, that was not really a workable order. The key is now gone from the figure and the names are added nearby the lines or points they correspond to. Most of the color issues were moot once I withheld the "other" line. It matched nearly identically the color given to Jindal's data. The other change is that I've added in the element of time. Everyday is accounted for in the series now so that it doesn't appear as if each poll is equidistant from the next.

Here's the trend updated through today:

[Click to Enlarge]

[If you find anything about the above graph confusing still, please let me know in the comments section.]


Before I close, I did want to mention one other issue with this FOX poll and the poll ABC and the Washington Post released this morning. In each case, we are talking about a 2012 primary question that is based on the responses of less than 300 Republicans (and/or Republican-leaning independents) nationally. When the goal is 1000, less than 300 respondents has the effect of REALLY ramping up the margin of error. In the process, the representativeness of the poll is made all the more questionable for something that is already well in advance of primary season (or even the competitive tail end of the invisible primary for that matter). As I've said recently, I like seeing these numbers and I enjoy seeing the trends, but these things absolutely have to be taken with a grain of salt. And occasionally I like to fold in some discussion of fundraising or organization, but I try to avoid claims like these at all costs. To assert that Huckabee leads this race or that it is beneficial for Romney to "draft" behind Huckabee is patently ridiculous. Given the margins in the polls conducted so far, Romney and Huckabee are tied (with Sarah Palin). Now, it could be that the perception that Huckabee is ahead is helpful to Romney in that "everyone else" is gunning for the former Arkansas governor and not Romney, but still. Let's just watch these numbers come in and not over-interpret them.


Recent Posts:
Presidential Primary Reform Week: The National Association of Secretaries of State's New President

ABC/WaPo Poll: 2012 GOP Primary--Huckabee Back on Top, but...

Presidential Primary Reform Week: The Fair and Representative Presidential Primaries Act of 2009

No comments: