Denis Stukal received has PhD in Politics in 2018 and is currently a Post-Doctoral Fellow at the NYU Social Media and Political Participation (SMaPP) Lab. His work has been published in Big Data and is forthcoming in the Journal of Politics. His research interests lie at the intersection of Political Communication, Political Methodology, and Data Science, in that he develops and applies new computational approaches that help advance our understanding of the novel phenomena in political communication that have emerged with the advent of social media, especially in non-democratic state propaganda.
His dissertation, entitled “(Mis)information in Autocracies: A Computational Approach with Application to Russia”, examines different aspects of the use of misinformation and propaganda in Russia utilizing cutting-edge machine learning and natural language processing methods.
In the first chapter, Denis develops the notion of a decoy political party as a non-ruling party that mimics opposition without actually challenging the regime and applies modern topic modeling methods to publicly available parliamentary speeches to show that, unlike actual opposition parties, the decoys do not aspire to affect agenda setting or question the activities of the dictator. Instead, they focus on promoting populist topics that are close enough to the agenda of the opposition and thus sway away their voters.
Another chapter of his dissertation studies the political use of social media in Russia by focusing on the political use of Twitter bots and reveals the strategies employed by the Russian regime to counteract online opposition activities. This project combines high relevance for public policy development with innovative computational methodologies. Denis introduced a novel predictive model based on an ensemble of supervised learning algorithms that yields highly accurate predictions of bot activity.
Denis also shows that – contrary to the prevalent mass media narrative – not all Twitter bots in Russia are pro-regime. To disentangle pro-regime bots from anti-regime ones, he develops a neural network model that uses textual data to predict the political orientation of Twitter bots and shows that both pro-regime and anti-regime bots maintain a substantial presence on Russian Twitter, although they have different patterns of activity. While anti-regime bots are mostly used to amplify a very limited number of oppositional mass media and politicians’ accounts, pro-regime ones are much more diverse in their activity, thus making it harder to detect and counteract them. All these findings suggests that a systematic analysis of the use of Twitter bots in Russia should not be implemented without distinguishing between automated accounts with different political affiliations.
The final chapter of Denis' dissertation addresses state propaganda on Russian TV by focusing on the strategies used to manipulate economic news. It employs regression discontinuity design and semi-parametric models to show that Russian television does not engage in full-scale censorship of negative economic news. Instead, it uses a different propaganda mechanism that entails selective attribution of events. Specifically, economic achievements are often attributed to President Putin or other Russian officials, whereas bad economic news tends to be associated with foreign governmental decisions or the global or regional economy.
Other projects that Denis is working on include developing software that facilitates research on the political use of social media (in particular, an R package to scrape VK: Rvk).
Denis has extensive experience in teaching introductory and advanced quantitative and data science methods.