Panem et Circenses: The Persuasive Effect of Soft News in Berlusconi’s Italy (with Andrea De Angelis)
How do politicians shape news coverage? and to what extent does news content influence voters? This paper exploits an original dataset of TV news transcripts and a natural experiment in Italy to show how the political control of public television by Silvio Berlusconi significantly affected ordinary news reporting. By applying the unsupervised learning algorithm WORDFISH to the transcriptions of almost 20,000 hours of broadcasting, we are able to document that: newscasts in Italy differ across channel on the concentration of soft vs. hard news; that this is predominantly the result of differential news coverage (agenda setting) and not of how the news are covered (framing); that during his tenure as a prime minister Berlusconi significantly altered public television coverage to mimic his private companies by increasing the soft news concentration and, crucially, that this led to an increase in the voters’ probability of voting for him. These results provide evidence on a previously undocumented mode of political communication and its implications for voting behavior and democratic competence.
Status: Under Review. Download (PDF)
How To Make Voters Read More Hard News? Evidence from a Factorial App Experiment in Italy
This study investigates the impact of internet media technology on voters media diet. I deploy a news aggregator app in which content is randomized among different categories of news, such as soft and hard, to understand how the mix of news impinge on voters demand for news, readership and knowledge. Specifically I test for changes on voters readership of hard news and interest in politics, time spent reading the news and learning from headlines. The results in this study will provide important evidence on the anatomy of media consumption and how voters respond to new technologies.
Status: Pre-Analysis Plan preregistered in EGAP (No. 20180710AB). Draft available soon.
Alternative Media and Political Participation
A number of studies have found that media presence is associated with more political participation. Surprisingly, however, partisan press doesn’t affect political behavior significantly, as studies have shown that neither Republican or Democratic newspapers were able to foster turnout for their party. I argue that this result may be limited to mainstream forms of participation, such as voting, for which the press inevitably does extensive reporting and for which readership is intensively self-selected. Yet, the presence of sympathetic alternative media may instead be crucial to motivate, inform and coordinate marginalized groups to political action. I test this theory by exploiting the historical case of the foundation of the NAACP in the U.S. and look at its relationship with the black press. I test how the pre-existing differential presence of black press newspapers affected the speed and efficacy of diffusion of NAACP branches by matching municipalities with similar potential media markets, for instance with respect to income and percentage of black press population, and by looking at whether a branch was established and if its foundation was faster with respect to areas with, ceteris paribus, no distribution of black press publications. I find that the presence of black press increased the diffusion of NAACP by eighty percent, and that in areas with black press distribution the NAACP was able to found branches earlier. Results from this paper provide important evidence of the various effects that alternative media may have on political participation.
Status: Draft available soon.