19 West 4th Street, Dept of Politics
New York, NY, 10012
I am a 5th year PhD Candidate in the Department of Politics at NYU and a member of the Social Media and Political Participation (SMaPP) lab. My dissertation studies the political implications of new forms of communication enabled by the internet and social media. This work involves developing innovative methods for performing online behavioral experiments and creating new ways to use text as data.
An urgent example of these changes is the explosion of uncivil discourse online. In contrast to deliberative communication, which is essential to a functional democracy, uncivil political discourse polarizes participants and inhibits learning. My Job Market Paper (invited to revise and resubmit, American Political Science Review) describes an experiment I conducted during the 2016 US Presidential election designed to decrease high-profile poltical incivility. I find that various kinds of moral suasion can cause people to be less incivil, but that these interventions tend to be more effective on subjects who are more inveseted in their online identities. This work builds on methods I developed to combat the online harassment of African Americans. In "Tweetment Effects on the Tweeted" (published in Political Behavior), I used bots (fictitious personae with social media accounts I create and control) to sanction Twitter users engaged in racist harassment. I found that sanctioning can have a substantial, durable effect in reducing racist harassment--but that this sanctioning is most effective when the sanctioner and subject share a relevent social identity. These online behavioral experiments represent an important innovation, allowing researchers to conduct controlled experiments on difficult-to-reach populations at scale and with high levels of ecological validity. I recently conducted (with Jorge Gallego, Juan D. Martinez and Mateo Vasquez) a similar experiment aimed at informing politically interested citizens during the 2016 Colombian Peace Plebiscite.
Modern media technology has also enabled more targeted and personalistic elite political communication. Using an innovative design that matches panel surveys with objective measures of respondents' social media diets, I find (with Patrick Egan, Jonathan Nagler, Jonathan Ronen and Joshua Tucker) that information from traditional media sources does in fact increase factual political knowledge while information from parties increases knowledge of those parties' platforms (invited to revise and resubmit, American Political Science Review). Elite communication on social media takes a slightly different form in non-democratic contexts; in research with the SMaPP lab (conditionally accepted at Political Science Research & Methods), I find that the Venezuelan regime strategically distracted their followers from large-scale protests.
Much of my research has entailed methodological innovation, especially in the use of Text as Data to analyze trends in political communication. In more explicitly methodological work (with Arthur Spirling and Ken Benoit), I have developed an improved measure of textual complexity using crowdsourcing and machine learning to discover the textual features that best capture empirical textual complexity. In a forthcoming book chapter, we demonstrate that the State of the Union address has declined in complexity relative to other corpora of political text.
My research agenda and teaching have been strongly complementary. In addition to one semester of undergraduate Intro to American Politics, I have served as a teaching assistant for four semesters at NYU's Center for Data Science: one semester of (Master's-level) Text as Data, and three semesters of (Master's-level) Intro to Data Science. NYU was the first school in the world to offer a MS in Data Science, and for the past two years, I've been responsible for teaching incoming classes the basics of Data Science using Python.
"Tweetment Effects on the Tweeted: An Experiment to Decrease Online Harassment". Political Behavior, 2017, Volume 39, Issue 3
"The Dumbing Down of the State of the Union? Trends in the Complexity of Political Communication" (with Arthur Spirling and Ken Benoit, prepared for Anxieties of Democracy volume edited by Nolan McCarty and Frances Lee)
"Elites Tweet to Get Feet Off the Streets: Measuring Regime Response to Protest Using Social Media" (with Richard Bonneau, Jonathan Nagler and Joshua Tucker) (Accepted at Political Science Research & Methods)
"Choosing in Groups: Analytical Politics Revisited" (with Michael C. Munger). Cambridge University Press, 2015.
"Experimentally Reducing Partisan Incivility on Twitter" (Revise and Resubmit, American Political Science Review)
"Political Knowledge and Misinformation in the Era of Social Media: Evidence from the 2015 U.K. Election" (with Patrick Egan, Jonathan Nagler, Jonathan Ronen and Joshua Tucker) (Revise and Resubmit, American Political Science Review)
How Norms ChangeThe New Yorker, October 2017
Bots aren’t just service tools—they’re a whole new form of media Quartz, April 2017
Twitter bots can fight racism — if they’re white and popular Vice News, December 2016
Telling People to Be Less Racist Online Works, Sometimes New York Magazine, November 2016
Racist tweeters can be convinced to stop spreading hate—if a white man asks them to Quartz, November 2016
Twitter bots can reduce racist slurs—if people think the bots are white Ars Technica, November 2016
Why Online Allies Matter in Fighting Harassment The Atlantic, November 2016
Stop Playing Defense on Hate Crimes Time, November 2016
Troll hunters: the Twitterbots that fight against online abuse New Scientist, August 2016
Are presidential writings getting dumber? We checked — and were surprised. The Washington Post, April 2017
This researcher programmed bots to fight racism on Twitter. It worked. The Washington Post, November 2016
Social media, #Immigration, and political knowledge in #Ukelection2015 The Washington Post, May 2015
Why Tyrants and Despots Love Social Media Newsweek, July 2015