When seen through a news report or a computer screen, the impact of current political research can seem very disconnected from what’s really taking place. It can be hard to try to understand the results and implications of politician’s behaviors and opinions without an already-written history book. But with this new age of media presentation comes the new age digging tool of data analysis, which is once again proving to be the key to decoding to today’s political discourse.
And that’s not its only use. Once again, Wes students are proving it possible to not only use online politics for research purposes, but to get your foot in the door of Data in the Real World. In April, John Murchison ‘16, Grace Wong ‘18, and Joli Holmes ’17 attended the Midwest Political Science Association conference in Chicago to present a poster about their research on congressional politics.
I got to speak to Grace, a government and American studies major, and John, a government major, about their experiences both in researching this field and attending the conference. Their story begins how so many for student’s involved in data analysis have: at the QAC’s summer apprenticeship program.
During the summer apprenticeship, John and Joli were working with Professor Dancey (an Assistant Professor in the Department of Government). “They started lab without a research question, just a massive amount of data and an interest in congressional politics,” Grace explained. “When the year came around, they wanted to pull me into their project.” This meeting had led them to work together for the past school year as the Congressional Politics Research Lab. “We call ourselves CPEL for short,” Grace added.
For the bulk of their data, CPEL turned to another new age data source – twitter. “We had a big data set of all these tweets from different congress members. They started digging around in the data, seeing what we could find out (“A party’s messaging strategies, how to quantify these strategies, how to use the data to tell us things about how parties behave,” listed John when I asked for examples) and the questions developed from there.
“Our main question for our poster was motivated by something we read suggesting that the two parties are driven by different factors,” John explained. An example of this party asymmetry would be the Republicans being driven by ideological policies, while the democrats are more concerned with group based consideration. “This plays out in the data by the parties being driven by the different variables that matter in their constituency. “So for democrats, the use of #immigration is significantly predicted by the percentage of foreign born people in their district.”
Here are John, Grace, and Joli presenting their poster, titled “Spreading the Message: Predicting Party Hashtag Use by Members of Congress on Twitter” at the MPSA Conference.
But it wasn’t all easy-going. John explained that one of the issues with gathering data from Twitter is that it’s very big and messy – “So you can’t be sure that you’re going to find the relationships that you expect or that the data will reflect something quantifiable at all.” Overall, they didn’t find a lot of clear patterns of differences between the parties in what determines how many times they use a different hashtags. “However we did find that for certain hashtags the demographic variables were in some cases significant. So it wasn’t that it wasn’t helpful to show the differences between parties.”
These kinds of muddled results are fairly normal – as proven by CPEL’s research still being accepted to one of the top political research associations. “It was really intimidating initially,” admitted Grace when I asked about the atmosphere of the conference. “There was a lot going on at first. And it’s crazy having people come up to look at our research and realize that their research has influenced us.” John added, “For me it wasn’t just that we were meeting people citing in our research. [Being at the conference] helps reinforce our understanding of what we are accomplishing.”
But perhaps even more important than the relevance of the field or their admittance to the conference is the collaborative atmosphere that John and Grace recounted as defining their process. “We work very closely with [Professor Dancey] and he’s a big part of the work, but it’s really collaborative,” John explained. “It’s not that we’re assisting him or he’s advising us.”
As opportunities for students in data analysis expand, of course Wesleyan is trying to stay with the curve. And John, Grace, and Joli’s group is probably the first of many. “We’re kind of the guinea pigs, and I know Manolis (Manolis Kaparakis, Director of the QAC) is hoping that it’ll expand,” said John. Grace echoed her enthusiasm for this expansion to happen. “These labs are the epitome of interdisciplinary learning. And we’re practicing the skill of collaboration. I learned so much from John in way of statistics. And I learned from Joli and how she manages data.” John points out that their collaboration with Professor Dancey is another good example of one of the big strengths that Wesleyan gains from our lack of grad students. John, Grace, and Joli are doing very real publishable research before they graduate, helping them to learn and hone their methods before they are launched into the world of media-controlled presentation.
When I asked them about the future growth of CPEL, Grace smiled. “I think Professor Dancey structured it really strategically, with all of us in different years. We’ll probably grab someone below us, and go to these conferences in future years.” Before then, CPEL is planning on expanding their research into a paper to be presented by the end of the year, before John graduates. After that, the sky is the limit.