Social Media Buzz Can Predict Elections
Social media buzz can predict the winners in political elections, although it can’t necessarily forecast the size of their victory, according to a new study by NM Incite.
The NM Incite study examined social media buzz in relation to four races during the 2010 midterm elections, and found that in three out of four races, the candidate who was mentioned most frequently in social media went on to win the election. However, NM Incite also noted that their share of online buzz didn’t necessarily correspond to actual percent of votes, suggesting there is a correlation between social media buzz and electoral outcomes but not necessarily causation.
The NM Incite analysis looked at 50 days of online buzz between September and November, 2010, for four statewide races. Sources of online buzz included blog posts, blog comments, online news sites, video and image sites, message boards and groups, and public posts on Twitter.
In one of the elections examined, California Senator Barbara Boxer had 55% of the online buzz but just 52% of the actual votes cast in her race against Carly Fiorina. Results could vary according to circumstances, including three-way elections. In a three-way race in Florida, Marco Rubio had 40% of the online buzz but 49% of total votes cast. And in the Maryland governor’s race, Marin O’Malley took 55% of the online buzz and 56% of the total votes cast.
In the Ohio governor’s race, on the other hand, Ted Strickland had more online buzz (54%) but fewer votes (47%) than John Kasich.
Social media appears to have a range of predictive possibilities. Last month I wrote about a study from SAS and UN Global Pulse, a UN think tank, which used social media chatter to predict unemployment trends. Titled “Unemployment Through the Lens of Social Media,” the study examined online job-related conversations from blogs, forums and news in the U.S. and Ireland; assigned a quantitative “mood score” based on the tone of the conversations; and correlated the mood scores with changes in the unemployment rate. For example, the researchers found that the volume of conversations in Ireland showing a “confused” mood correlated with an uptick in unemployment three months later. Likewise, conversations about public transportation spiked about a month before unemployment.