All Hail Twitter as Presidential Predictor


Social media has helped people connect with people, and so it seems fitting that one of its largest social networks is helping pollsters make those connections, too.

The 2012 election is certainly a leap from 2008 when considering the use of social media and its measurement capabilities. Notable are new solutions launched expressly to assess voter sentiment. Backed by two polling firms and Twitter, for example, Twindex issues ratings based on Twitter messages about President Barack Obama and Governor Mitt Romney. Meanwhile, The Washington Post relies on a media tracker, Mention Machine, to monitor candidate mention.

With all this focus on tweets, is Twitter the best data source for predicting polling results? My vote is "Yes." And with a platform on social media history, I can tell you why you should vote the Twitter ticket as a terrific way to canvass for electoral predictions.

First, Twitter has been savvy in addressing its development into a real-time information source. Steady refinement of its uses, coupled with growing acclimation of people's online behavior, provides pollsters and analysts with a wellspring of relevant data. Twitter has struck a balance between providing useful feature updates that people adopt -- countering Facebook's woes with privacy -- and a lack of development -- countering the potential reason for Google+'s initial slow growth.

In a May post, New York Times blogger Nick Bilton noted a key outcome from Twitter's efforts -- Federal Trade Commission endorsement of Twitter's privacy measures. News like this encourages people to participate on Twitter openly, and makes the resulting data that much more valuable to analyze.

Second, already having learned to manage lean online launches and learning quickly from beta periods, organizations have acquired a wealth of material that can improve assumptions applied to online measurement tools development. Tools have improved accuracy regarding assumptions about how people speak in Twitter. Twindex, for example, accounts for social colloquialisms based on past Twitter history.

All of these Twitter-related improvements are leading to outstanding discoveries on what social media influence truly means. At Voxsup, a social media analytics platform vendor whose name stands for "voice of the people," founder Alok Choundary has developed an influence analysis tool based on social media big-data research. Choundary, who is an engineering and business professor at Northwestern University, suggested that Twitter users with a small following can be highly influential because their followers appreciate seemingly honest comments. Having a large following is not a prerequisite to being influential. (Watch Choundary's recent Social Media Week presentation in Chicago here.)

For Twitter measurement, this means pollsters can potentially identify relevant segments that best reflect true sentiment toward a topic. Followers gravitate toward those whom they feel are making genuine expressions, sharing a reply or retweeting the original message.

Finally, Twitter structure for public commentary posted at the occurrence of the event increases the relevancy of the comments and sentimental accuracy about the event. The 140-character format trains people to state their intent quickly, while the widespread use of mobile phones makes that response complimentary to offline events. Newscasts display tweets as an indicator of what is going on. I recall a New York newscast showing an entire TweetDeck screen to highlight hashtag sentiments efficiently.

Obama's first presidential campaign set a standard for galvanizing a political base with social media. It was inevitable that 2012 would further heighten social media's political use. There are nuances to what is being discovered this election, but as eMarketer noted last week, more people are learning about candidates via social media. The unprecedented success of 2008 makes it understandable that political pundits and the experts begin to measure social media political influence. The analytics community should note that Twitter certainly has a strong lead.

Do you trust Twitter presidential predictions? Read Marshall Sponder's Counterpoint, take our quick poll at right, and share your thoughts on the message boards.

Point / Counterpoint, Founder, Zimana

Pierre DeBois is the founder of Zimana, a small business analytics consultancy that reviews data from Web analytics and social media dashboard solutions, then provides recommendations and Web development action that improves marketing strategy and business profitability. He has conducted analysis for various small businesses and has also provided his business and engineering acumen at various corporations such as Ford Motor Co. He writes analytics articles for AllBusiness.com and Pitney Bowes Smart Essentials and contributes business book reviews for Small Business Trends. Pierre looks forward to providing All Analytics readers tips and insights tailored to small businesses as well as new insights from Web analytics practitioners around the world.

Counterpoint: Train for Data Visualization Skills

Chances are you've already got good data visualization experts on staff, even if you don't know it yet.

Customer-Centric Banking Analytics Scares Me

Banks have tons of customer data at their disposal; unfortunately not all will use it scrupulously.


Re: Advanced, but not quite there
  • 10/22/2012 8:08:26 AM
NO RATINGS

Well I'd agree with younger, probably living in a larger urban or college setting.  The 55+ age group is pretty small, current demographics studies show that twitter tends to have more female users (about 40% male, 60% female).  Some other bullet points that I found interesting.

  • More than one quarter of online African-Americans (28 percent) use Twitter. Thirteen percent use it on a typical day, which is nearly double the overall average.
  • Twenty-six percent of Internet users aged 18-29 use Twitter, which is nearly double the rate for those aged 30-49.
  • Residents of urban areas are significantly more likely to use Twitter than their rural counterparts.

Income doesn't seem to play that big a role, the only numbers I've found for income has the highest % of users falling near the national median income.

Re: Advanced, but not quite there
  • 10/19/2012 9:21:50 AM
NO RATINGS

@SaneIT, by Twitter demographics, what would you say: younger, wealthier, probably living in a larger urban setting or college community, more informed?

Re: Advanced, but not quite there
  • 10/19/2012 8:08:59 AM
NO RATINGS

I think mundane depends on who you are.  There are people to take a lot of people who get serious about things that I might call mundane.  I think twitter tracking can be handy if you're looking at how the average twitter user feels about a trending topic.  Where I think it fails for tracking is that you have to keep in mind what your demographics are when it comes to twitter users.

Re: Advanced, but not quite there
  • 10/18/2012 7:24:47 PM
NO RATINGS

That is a great question and perspective Beth.    And I was really trying to be polite about this issue of tweet analysis, as I know there are many who feel it has some merit.   While there may be some, I really don't think the effort is worth the amount of return.   Just today I have heard two references to tweets on the airways, maybe they were speaking of this a couple a days ago just as fervently and I wasn't as aware of it as I am after considering the Point/Counter  debate.  When I listen just long enough to let in sink in that I just heard some reporter reference some tweet regarding the election  -  I think to myself really ?

Am  I alone here ?  I don't think so but it does underscore your question - why bother at all ?   Though it would not be a focus I would actively champion  - I do think it (Tweet analysis ) has some merit in judging customer sentiment. 

I would be more inclined to believe a tweet (analysis) regarding the relationship of the effectiveness of an tweeted coupon for instance a or  some aspect of customer service - feedback on deals etc.   

And I also think you put it very well Beth.  I  would agree that the "inherent nature of tweets and tweeting" make  any conclusion based on it's analysis  tenuous at best.

Re: Advanced, but not quite there
  • 10/18/2012 6:50:27 PM
NO RATINGS

Louis, I have to ask, if, like you say, we focus tweet analysis on "more mundane issues," why bother at all? Or, in other words, does the nature of tweets make their analysis pointless?

Re: Advanced, but not quite there
  • 10/18/2012 5:39:09 PM
NO RATINGS

Hi Pierre ,   You make a interesting and strong argument for the use of tweets to give some sort of indication of how people really feel about a topic ( the election in this case ).  While I understand tools and methods are improving daily, I am just not convinced that tweets have any reliability.   

I am a part of the tweet universe ( by peer pressure more than anything else ) and I don't pay attention to tweets at all whether they come from the President or not - that is my personal stance which if it seems extreme, imagine the perspectives that most come from when tweeting?  

How many times to do we hear someone apologizing for some hastily made tweet ?   I really think we should focus on more mundane issues when we talk about tweet analysis and leave the serious issues to more tradition means. 

And if we are going to amend the tradition in any way then at least do it with a reliable tool - which I do not think tweets or tweeting ever will become.

Advanced, but not quite there
  • 10/17/2012 5:44:44 PM
NO RATINGS

Hi Pierre, you make some compelling arguments here but I'm still pretty iffy on Twitter's value for presidential predictions. I don't think it's used broadly enough or that we have adequate ways of measuring what's being said, by whom, and where. I'll hold out for that changing for 2016 run.

Re: Twitter
  • 10/17/2012 5:42:54 PM
NO RATINGS

Sadly, Alexis, that's so true. Same goes for what color ties these guys wear!

Re: Twitter
  • 10/17/2012 3:08:21 PM
NO RATINGS

@webmetricsguru, @Alexis Yes, they do look like fireworks, especially because they wisely chose a dark background for contrast with the colors. How you package things does make  a difference in getting people's attention, and that certainly applies to data analytics. Show people a stream of numbers, and their eyes glaze over, but if you show them dramatic looking clusters, they find it more attractive. 

Re: Twitter
  • 10/17/2012 3:06:05 PM
NO RATINGS

@Pierre that does stand to reason. Certainly, it is how many social connections work -- people often associate with like-minded inviduals. My online connnections happen to include supporters on both sides, so I try to avoid political discussions. 

Page 1 / 2   >   >>
INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +