Twitter Poll: GOP Primary -- Week Ending 10/29/2011

Herman Cain maintains a lead. But that lead appears to have softened slightly since last week.

Most of the Cain loss appears to be driven by a Perry bump:

 

But the "Perry bump" may just as likely be a result of negative mentions as of positive mentions. Indeed, it appears to be centered around a number of Perry news items mid-week -- a fairly even split of "negative" and "positive" news, from the perspective of the GOP electorate. This is the first week that I wish I had implemented some sentiment analysis. I'd suspect that the (net) Perry mentions mid-week were an even split (or worse) between "positive" and "negative" mentions. In any event, the bump appears to have no staying power:

A closer look at the mention space shows Ron Paul still "off in his own little world" and Newt Gingrich building a base of conservative support similar to that seen with Cain before his surge. I still expect a Cain stumble will be needed for Newt to exploit this positioning.

As usual, I hope to be able to add some new features for next week's poll. I may be settled enough in my new job for that to be actually possible this week.

Twitter Poll: GOP Primary -- Week Ending 10/22/2011

Herman Cain continues his comfortable lead while Perry falls farther into the middle of the pack:

Average mention share for the week suggests a three-way race between Cain, Romney and Ron Paul:

Like Bachmann, Newt's debate performance lead to a small bump at Cain's expense, but the bump quickly faded.

 Cain and Romney dominate the "mainstream" discussion. Meanwhile Ron Paul is off in his own little world -- drawing significant mention share from conservatives and "independents" but virtually none from progressives and politically-engaged centrists.

Newt has a lot of conservative mention share. I wouldn't rule out a surge, but would expect a Cain stumble first.

Twitter Poll: GOP Primary -- Week Ending 10/15/2011

Herman Cain remains the most mentioned GOP primary candidate this week.

Although Romney did see a debate bump, it faded fast. Cain was consistently most mentioned all week:

Average mention share for the week suggests that the primary is settling into a two candidate contest between Romney and Cain -- with Perry and Ron Paul both slipping into the middle pack of also-rans.

A closer look at the mention space re-enforces the sense of Cain/Perry dominance -- and confirms that both are drawing robust volumes of mentions all across the political spectrum.

As I mentioned last week, I'd like to metricize the purpleness and overlap of the candidates' mentioners. But this week was my first at a new full time role. Stay tuned ... and see y'all next week.

Twitter Poll: GOP Primary -- Week Ending 10/8/2011

Herman Cain has maintained a substantial lead in Twitter mention share this week.

Cain lead comfortably all week, but discussion of Rick Perry's "camp problem" exaggerated both Perry and Cain mention share early in the week. Following the dust-up, Romney re-centered to his typical 20% share and Perry continued his steady decline into the middle of the pack.

On average, mention share analysis this week gives Cain a solid lead, followed by a 3-way race of Romney, Perry and, sometimes, Ron Paul.

Analysis of the mention space suggests that Cain's mention distribution has grown increasingly "purple" -- even showing a large number of "blue" mentions. This seems to be the norm for well-established candidates, which I would suggest Cain has now become:

In the coming weeks, I hope to be able to metricize both the "overall purpleness" of a candidate's mentions and their "supporter overlap" with other candidates. Also stay tuned this fall for a number of text-mining aggregations that have been in the works. And, of course, check back next Sunday for the latest Twitter Insta-polling from Take America Forward.

Twitter GOP Primary Poll -- Week ending October 1, 2011

Since the Florida straw poll Herman Cain has seen a huge boost in Twitter Mention Polling, moving from the center of the pack to clear front-runner status:

For the week of September 26 to October 1, Cain's mention share of 30%+ eclipsed those of his closest competitors -- Willard M Romney and Rick Perry, who each command roughly 15% of this week's mentions:

A look at the daily breakdown shows an uninterrupted Cain lead for the entire week, with Romney and Perry in a back-and-forth battle for second -- occasionally joined by Ron Paul.

Closer inspection of the mention space suggests that Cain dominates the discussion among conservatives, while Romney and Perry are mentioned more broadly on the political spectrum:

Please check back at Take America Forward each Sunday for weekly updates. Also expect an expanding range of analysis and a detailed discussion of Twitter aggregation methodologies.

Back from "Summer Vacation"

Apologies for the long silence. This spring, my wife was transferred from the San Francisco Bay Area to South Jersey (near Philadelphia). Although you can always follow my latest political thoughts in amuse bouche form on Twitter as @TheLoki47, our cross-country move pushed my political blogging to the back seat for the summer.

Since arriving in Philadelphia this July, I've been focused on developing accurate Twitter polling for the upcoming elections. Sunday morning, I will release the first set of weekly results. These results represent my best efforts to aggregate political energy on Twitter.  Posting these early (likely buggy) results, is an effort to share more broadly some insights on campaign buzz and grassroots energy that I've had at my disposal for the past several weeks.

As a preview, below is a visual aggregation of candidate mentions for the past week.

I will explain (and improve) the algorithms over time. For now it suffices to note that these are "active" Tweeps (25+ political tweets this week) with "frequent" mentions (5+ this week) of a political candidate. Colors are a best guess at political lean and node size is indicates a higher volume (of tweets for tweeps; of mentions for candidates) and the red and blue nexuses are the hashtags #p2 and #tcot.

The biggest takeaways are:

1. The Herman Cain phenomenon is real (at least this week).
2. Fading candidates seem to show a high volume of "blue" mentions.
3. Although the Eastern Conference/Western Conference metaphor promoted by Chris Matthews is a convenient first approximation, the competing factions in the GOP are likely far more complex and overlapping than the metaphor suggests .

See y'all Sunday with much more.

Owning It: Killing the "Liberal Bias" Myth

Cenk Uygur joked on MSNBC on Thursday that history has a "liberal bias" -- amending a line made famous in Colbert's coup de grace performance at the 2006 White House Correspondents' Dinner:

Now, I know there are some polls out there saying this man has a 32 percent approval rating. But guys like us, we don't pay attention to the polls. We know that polls are just a collection of statistics that reflect what people are thinking in "reality". And reality has a well-known liberal bias ... Sir, pay no attention to the people who say the glass is half empty, because 32 percent means it's two-thirds empty. There's still some liquid in that glass, is my point. But I wouldn't drink it. The last third is usually backwash.

The hipster in me says: But Monty Python has already flung a cow, so it couldn't possibly be funny again. The political analyst in me says: Now that's how to inoculate an insidious right wing meme. Imagine a world in which the very phrase "liberal bias" has become so cliche that it is a liability.

In short: Lean into it.