RESEARCH
"The Impact of Political Campaigns on Demand for Partisan News"
Abstract
Political polarization in news consumption has recently gained attention, yet policies to limit it are hard to evaluate. This paper introduces a novel, self-collected dataset on Spanish prime-time TV news—the dominant source of political information in Europe. I identify stories minute-by-minute and match them across newscasts to compare editorial treatment. Combining machine-learning methods with large language models, I classify the partisan slant of each story and use the resulting data to document changes in news coverage during the 2023 presidential campaign. I then match these data to high-frequency audience-meter records and estimate a structural random-coefficients demand model to understand how outlets compete in coverage and tone of news about different political parties. Exogenous shifts in the news landscape that constrain channels’ ability to shape content serve as instruments for slant and allow me to recover demand preferences. I find significant evidence of affective polarization only after the election campaign begins. Given the demand estimates, I back out outlets’ content preferences from a horizontal-differentiation game. Finally, I run counterfactual simulations to assess the effects of adopting the proportional-airtime rules enforced in other European countries but not in Spain. The framework offers a tractable tool for evaluating media policies aimed at content fairness or fighting misinformation.
"Breaking the Echo Chamber: Nonviolent Protest and Police Violence on Twitter"
with Hannes Mueller, Daniel Montolio, and Francesco Slataper.
Abstract
Social media platforms play a crucial role in shaping public discourse and political con- flict, but social media landscape is shaped in turn by the unfolding of the conflict. This paper incorporates this dynamic reciprocity in a microfounded model of online inter- actions, built upon a granular analysis of how social media users responded to events around Catalonia’s independence referendum in 2017. We analyze retweets within and across Catalan and Spanish language groups using linguistic markers from a corpus of 26 million tweets. We define two channels to explain change in retweet behavior: ex- posure—the likelihood of encountering specific content— and retweet rates—the prob- ability of retweeting content once exposed. Remarkably, we can document a dramatic increase in retweet rates between language groups and greater exposure to cross-group content. We contrast this response with the reaction to a terror attack in 2017 and explore the role of political ideology in our findings.
"Online and Offline protest participation: An Empirical Analysis for the 2020 Black Lives Matter Movement"
Abstract
The recent wave of world-wide protests that took place after George Floyd’s killing has sparked attention in the Black Lives Matter movement, specially in terms of online activism. How does offline protesting behavior interact with the underlying online social networks? In this work, I build a classification algorithm to identify individuals who physically participated in the BLM demonstrations across the US. Thanks to this unique dataset, I explore at an individual level their full Twitter activity to better understand the role of influential users, coordination patterns and speech evolution. Through this analysis, I aim to examine assumptions regarding slacktivism, which involves engaging in online activism with minimal effort, in comparison to more traditional forms of protesting. By exploring how social media contributes to the development of traditional activism, we can gain a deeper understanding of the role of social media in protesting behavior.
