RJI research team finds where Twitter, politics intersect
The team helps to develop tools for processing “big data.”
Feb. 26, 2013
Associate professor of communication Mitchell McKinney’s biggest surprise in his latest research came before he even had findings — he was frustrated by the fickleness of his data-gathering tools.
“My first surprise was the vendors we were using to help us capture tweets, and the Twitter system itself crashed on the night of the first (presidential) debate,” McKinney, who is part of a team at the Reynolds Journalism Institute studying how Twitter enhances political engagement, said. “I was just amazed that the social media folks were not able to anticipate that.”
McKinney, along with assistant professor of communication Brian Houston, have spent close to a year working with a team of graduate students and professionals to study Twitter conversations during the presidential debates. They hope their research will provide insight as to how social media fits in with media coverage of major events.
“We’d like to see Twitter as this sort of electronic town hall, but that might not be what it is,” Houston said. “It’s really quite difficult to make sense of the conversation beyond looking at ‘Here’s what one person said, here’s what another said, that’s the end of the story.’”
Having collected and analyzed data from three debates and conducted a survey to gauge how newspaper readers use social media, Houston, McKinney and members of their research team updated the institute on their progress Monday.
The team found that different parts of the country were excited by different parts of the debate. While nationwide, the most tweeted about moment of the second presidential debate was a comment Mitt Romney made about "binders full of women," Twitter users in Dallas mostly tweeted about jobs and taxes.
The researchers initially hypothesized that people who communicated with others would find watching the debates a more positive experience and would watch more debates.
In a survey of 716 newspaper subscribers, the team found that those who frequently talked and read about politics, and those who watched the debates with others enjoyed the debates more, but they did not find any correlation between social media use and enjoyment of the debate.
“One interesting thing about this data is that the Twitter-use variable kind of came out of nowhere,” graduate research assistant Josh Hawthorne said. “It didn’t have a significant influence on whether a person enjoyed the debate more, but they tended to watch more debates if they did.”
Houston said although the researchers have completed their tweet collection, they have more work to do. After running into problems with data collection during the debates, the team hopes to develop tools to help journalists and researchers use “big data” in their reporting and research. The team also plans to continue working with news organizations to help them better use social media.
“We want to develop tools for researchers like ourselves and journalists can use to better incorporate into this data into their reporting, their research and their understanding of what’s going on, Houston said. “There really aren’t a lot of good easy tools for journalists and academics to use. So our hope is to develop some of these.”
In the future, the researchers will take their work to a politics and social media symposium in Washington, D.C. Houston said the team also plans to bring their work together with a separate national survey that will serve as a Twitter experiment.
“I’m still sort of amazed at all the different lines and threads that have been explored, and that we’ve been pursuing and chasing," Houston said. “A lot of them are very practical, a lot of them are more social-scientific and academic, and several of them were a nice intersection to those two approaches, which is why this has been a really fantastic project."