Thursday, September 6, 2012

Convention brown-nose ranking 2012

I used the NYT's characteristically awesome infographic about the words most used at the 2012 Democratic and Republican conventions to rank the states in order of how many times they were mentioned by people on stage.

The first number is the total mentions at both conventions per 25,000 words spoken. Within the parentheses, the number on the left is Democratic mentions per 25,000 and the number on the right is Republican mentions.

wisconsin 16 (4-12)
ohio 15 (8-7)
massachusetts 15 (7-8)
north carolina 14 (11-3)
florida 12 (4-8)
colorado 7 (6-1)
virginia 7 (3-4)
michigan 7 (2-5)
montana 7 (4-3)
texas 7 (4-3)
california 6 (3-3)
oklahoma 6 (0-6)
south carolina 6 (1-5)
iowa 5 (3-2)
north dakota 4 (0-4)
nevada 3 (0-3)
indiana 3 (0-3)
illinois 3 (3-0)
delaware 3 (1-2)
new jersey 3 (0-3)
maine 2 (1-1)
nebraska 2 (0-2)
connecticut 2 (2-0)
kansas 2 (2-0)
kentucky 2 (0-2)
new hampshire 2 (0-2)
oregon 1 (1-0)
arkansas 1 (0-1)
hawaii 1 (1-0)
minnesota 1 (1-0)
maryland 1 (1-0)
georgia 1 (1-0)
south dakota 1 (1-0)
utah 1 (0-1)
tennessee 0
west virginia 0
vermont 0
rhode island 0
alabama 0
mississippi 0
louisiana 0
idaho 0
wyoming 0
alaska 0

Lessons learned:
  1. Politicians are unlikely to kiss up to a swing state if they think they're going to win it, maybe because it makes them look desperate.
  2. People really like talking about the state they're from. Therefore a major component of the convention brown-nosing process is actually the speaker selection.
  3. The deep South is just not cool.


  1. When the hell was the deep south EVER cool?

    Also: WTF your captcha system for commenting? It's easily the worst I've ever had to use.

  2. Blogger's system, not mine. Blame this weird, neglected corner of Google.

  3. Actually, don't. I guess it's gotten better lately. Fixed.

  4. Yeah, looks like candidate home states (excepting Illinois and Hawaii, strangely - apparently it's easier to just pretend Obama was formed as an adult now) and the states in which the conventions are held turn out the big winners. Then a couple of major swing states, and also just some random states (Montana? California? North Dakota?).

    I think the deep south thing is a major development, though. Once Obama won without any of the southern states (North Carolina was a bonus, not a key to his strategy), they immediately became irrelevant to presidential campaigns. Romney's path to victory both requires them and can treat them as given (except North Carolina, barely). This is a big part of why Obama can openly support gay marriage - most (but not all) of the people it pisses off are electorally irrelevant, but the people who are on board donate money he needs.

  5. This is really interesting, and thanks for doing this (my housemate will love it), though I am curious as to how you are contextualizing the mentions (if you are doing so) and drawing conclusions. Does simply uttering the name of a state necessarily curry favor with viewers? Do speakers believe that? (That sounds awfully J.L. Austin to me). Are negative mentions counted the same way for this data set ("Man, I hate New Jersey! It's so full of liberals / conservatives") ?

    Could recent events (e.g., Colorado shooting, failed Walker recall) overdetermine the number of utterances of a state's name? Could the presence of a speaker from a certain place metonymically imply the name of a state? Just curious.

  6. Good points as usual, Kevin.

    Audesapere, you're right - I'm basically assuming the sample size addresses most of your concerns. I guess part of being a politician is that you generally avoid spending much time on national television denigrating large groups of people.

    The better way to describe these findings might be "how important are the various states to the political life of the country right now?"

    1. I think I explained what I was saying poorly, and now I think I will explain it even worse!

      I don't think it's that anyone was clearly denigrating a place, I just think that the raw transcripts elide a lot of important data (audience reactions, body language, timing, inflection) that can give useful insight into context and how a term / place can be used to evoke a reaction from the audience, which I think could shed light on how the use of state names could actually reflect a campaign
      (Maybe I'm biased, though: I've done a lot of work with transcripts).
      So, for example, if this makes sense, the mention of, say, a blue state by a conservative candidate with a dramatic pause and laughter by an audience is significantly different than a candidate merely name-dropping a state within the context of a speech. An utterance like "Well, in X state they have Y policy and..." is weighted the same way in the data set as, say, "Well! In X state (say, MA - 3rd most mentioned state - a state roughly unimportant in electoral politics) [ long pause ] they do this [audience reaction]."

      And which transcribed actions or events evoke roughly the same concepts as those directly and explicitly stated by the speakers? If a chair can stand for Obama, can Christie "stand for" New Jersey and effectively represent a mention (or several) to the audience of the state?

      I am really interested in the data you have here and how you've situated them, but it just gives me more questions. Which I like.


    2. Miranda -

      While I will leave Mike to explain his relationship to the transcripts, I can say that it's unlikely that any of the politicians at either convention said anything either explicitly or implicitly or bodily derogatory about any of the fifty states.

      That's because each convention, as you may know, has delegates representing each state. If Paul Ryan had said, "I've lived within a few miles of Chicago my whole life, and Rahm Emanuel is ruining that city the way Illinois is consistently corrupting American politics." He would not only have risked offending Illinoisans watching at home, but also may have received boos from the hundreds of Illinois delegates in attendance and possibly had funding withdrawn from wealthy Illinoisans. But riots breaking out on the floor of the convention are very bad and to be avoided at all costs.

      It may be possible to point out a specific city as objectionable - San Francisco, Chicago, and the always-evil Hollywood (if I were a consultant for anyone giving a speech, only Hollywood and Washington itself are safe for derisive remarks and/or body language, really) - but states as a whole have to be part of the cheer-leading. Republicans hire Oregonians; Democrats hire Texans. Californians and New Yorkers give money to all the parties.

      In the debates, you can probably get away with at least body language, and maybe some softly spoken scorn directed at cities/states that are politically irrelevant.

      At a random campaign event, so long as you don't go overboard (and make it newsworthy), even the next town over's fair game. Especially if there's an upcoming football game of importance.


    3. Those are good points, but I guess my biggest issue is that context matters a lot for content analysis, and this specific data set weighs, for example, the following (approximately paraphrased) state name utterances equivalently:

      "So Mitt Romney wants to roll back the progress we have achieved under President Obama. In Montana, we have a saying - 'That dog don't hunt!'" Audience: laughter, cheers, applause

      "In my state of Ohio, Republicans are trying to enact onerous and discriminatory voter ID laws!" Audience: boos.

      "Mitt Romney was for Romneycare in Massachusetts before he was against it for the rest of the country!" Audience: laughter

      "I just graduated from the University of Texas ...!" (audience: Cheers)

      What I would be interested in is the terms used to evoke regions or places and how they are used / received by the audience, and the surrounding context of each evocation. Maybe I have a bias that favors audience reception a little too heavily. From my years as a transcriptionist and teaching ethnographic transcription, I generally consider unannotated transcripts as incomplete, to say nothing of isolated / decontextualized data points within them. But that doesn't mean they are not useful.

      Anyway. Sorry. I wasn't trying to argue.

      I'm just genuinely interested in how to interpret this data given the limitations of the transcripts, and I think what Mike did here was really cool.

      Too much coffee, too many commas, not enough sleep.


  7. This comment has been removed by the author.

  8. I think WI and MA are #1 and #3 mostly because they would be the home states of Ryan and Romney. NC and FL might be #4 and #5 because they were the locations that hosted the conventions themselves, but also because both states are competitive this cycle. The others in the top 7 most-mentioned are Ohio, Virginia, Colorado which are all considered strategically important and competitive states this election. If we're looking at Nate Silver's analysis, we can actually see that out of his five highest-ranked "tipping point states" (probability that a state provides the decisive electoral vote) all five actually appear in the seven most name-dropped states on this list (along with MA, Romney's home state, and NC, which hosted the DNC). That seems to probably be significant if we're attempting to perceive any meaning to these numbers.

    On "denigration" of whole states, it might be worthwhile to mention that Republicans in fact would engage in that sort of thing as a long-term pattern against Massachusetts (ironically the home state of their present candidate), this especially against Dukakis but even against Kerry, but I haven't studied the matter enough to be able to say exactly what sort of reaction would have occurred from Republican delegates from MA during conventions in which their state's name was used as a by-word for liberalism run amok. On long-term use of urban areas as symbolic social foils against the geographic areas associated with their own base, I'd actually like to see a long-term analysis of that sort of thing (how often Republicans have used terms like "San Francisco Democrat" or "Chicago politics", how important this was to overall social framing of candidates and geographically important to electoral patterns, and in what sort of context these remarks would occur, etc).



Creative Commons License
This work is licensed under a Creative Commons Attribution-Share Alike 2.5 License.