Research and projects
Research and projects
Job Market Paper
Draft here The effect of polarised information on social preferences: an experimental approach.
I investigate the impact of polarized articles on social preferences and the potential channels able to lead the impact. By running an online experiment in Prolific hiring 600 American workers, I find that reading counter-attitudinal articles decreases the level of altruism and increases anti-social behaviors; while reading pro-attitudinal articles increases the level of cooperation and decreases anti-social behaviors. Emotions and perceptions of the in-group are the main channels driving the impact.
Publications
Working Papers
The Impact of Social Capital on Vaccination Campaigns: Results from the French Case Study
I empirically study the effect of social capital on the success of the vaccination campaign in France during the Covid-19 epoque. I use data at department level. By performing a time-series analysis, I find that there is a positive relationship between the level of social capital and the vaccination success. The trend stops only after the enforcement of stricter public policies, like the introduction of the green pass.
Work in progress
First Impressions Matter: The Lasting Impact of Slanted Media on Attitudes with Federico Innocenti and Marco Minozzo (Under review)
We investigate how slanted articles influence attitudes. We implemented a field experiment where university students were randomly assigned to different treatments, reading news articles with either a positive or negative slant on several topics over four weeks. We collected students’ attitudes weekly and in a follow-up survey four weeks later. We find that: a) slanted information significantly shifts attitudes in the intended direction; b) the effect is more substantial for novel topics; c) the first impression about a novel topic shapes attitudes; d) the change in attitudes persists over time; e) different attitudes drive different economic decisions. Our findings highlight that media are more effective when shaping initial attitudes rather than altering existing ones. Thus, our results show the importance of timely information provision, especially in the early stages of public discourse.
Discrimination on information provision: the case of Tiktok with Federico Innocenti and Tsung-Hsien Li
We study how the characteristics of information providers impact the news consumers’ perception of information quality. The characteristics of information providers are crucial in online news consumption because most information providers do not have a reputation, and interactions are primarily one-shot. We focus on the comparison between characteristics that can lead to statistical discrimination— education, for instance—and characteristics from which no inference about information quality should be possible—such as gender and ethnicity. We plan to bring the experiment to the field—on TikTok—and study whether potential discrimination based on information providers’ characteristics can impact the popularity of content and profiles on social media. The National Science and Technology Council (NSTC) of the Republic of China (Taiwan) has awarded us a research grant for this project. The experiment will be launched in 2025 and is expected to be finished in the spring of 2026.
Sentiment Evolution in the media: the case of international conflicts with Federico Innocenti and Paolo Italiani
We use textual analysis to evaluate newspapers’ behaviors over time. We consider all the written articles of a large set of Italian newspapers for the period (01.01.2015 to present) that present specific keywords (namely, Israel and Palestine) to evaluate how their sentiment change over time and differentiate among them with a specific focus on the conflict period. Therefore, the focus in on the supply side and how it varies in the presence of unexpected events. Lastly, we compare our findings with the predictions of an Hotelling model that is used as benchmark in media economics. We are currently collecting the articles. The code to implement the analysis is ready.