Working Papers
"Choosing the Right Words: the Effects of Trump's Communication during the Pandemic"
Abstract: In the first stages of the pandemic, how governments communicated information about Covid-19 impacted on preventive measures taken by people, especially in periods when stay-at-home orders were not implemented. My research focuses on the political communication of former President Donald Trump in the US, in the first period of the onset of the pandemic. On February 28th, Donald Trump stated that Democrats are politicizing the coronavirus, and he referred to it as their “new hoax”. I analyze how these statements affected the online political debate, and how they impacted the propensity to stay at home. I find that Democrats talk relatively more about coronavirus after Trump’s statements. In addition, I find that Trump’s statements decrease the propensity to stay at home by around 4% in Republican counties. Overall, my results point to an effect of Trump’s statements both on the political debate and on individual behaviors.
"Regulation of social media and the evolution of content: a cross-platform analysis"
Abstract: Is self-regulation of social media effective in moderating online content? Or does it shift abusive posts to darker corners of the internet? This paper addresses these questions analyzing the effect of social media regulation from a cross-platform perspective. I exploit an episode of enlargement of Twitter’s regulation against racist hate speech and I investigate whether regulation is effective in curbing this harmful content or whether it shifts abusive content to unregulated platforms. I exploit synthetic difference-in-differences to assess the effects of the new Twitter’s policy, and I construct text classifiers using supervised methods from natural language processing to predict whether a post can be considered targeted hate speech. I find that racist content decrease in percentage both in Twitter and in Parler, but that users that have accounts in both platform increase their racist posts on Parler after Twitter’s policy. Finally, I exploit word embedding to see if the language used toward minorities changed after the regulation, and I find that terms related to some minorities became more offensive in Parler after Twitter’s policy. Overall, my results show a possible substitution effect between platforms, with more abusive content that shifts from regulated to less regulated social media platforms.
"Experience matters? The effect of political experience on municipal policies" (Joint with A. Manello)
"Choosing the Right Words: the Effects of Trump's Communication during the Pandemic"
Abstract: In the first stages of the pandemic, how governments communicated information about Covid-19 impacted on preventive measures taken by people, especially in periods when stay-at-home orders were not implemented. My research focuses on the political communication of former President Donald Trump in the US, in the first period of the onset of the pandemic. On February 28th, Donald Trump stated that Democrats are politicizing the coronavirus, and he referred to it as their “new hoax”. I analyze how these statements affected the online political debate, and how they impacted the propensity to stay at home. I find that Democrats talk relatively more about coronavirus after Trump’s statements. In addition, I find that Trump’s statements decrease the propensity to stay at home by around 4% in Republican counties. Overall, my results point to an effect of Trump’s statements both on the political debate and on individual behaviors.
"Regulation of social media and the evolution of content: a cross-platform analysis"
Abstract: Is self-regulation of social media effective in moderating online content? Or does it shift abusive posts to darker corners of the internet? This paper addresses these questions analyzing the effect of social media regulation from a cross-platform perspective. I exploit an episode of enlargement of Twitter’s regulation against racist hate speech and I investigate whether regulation is effective in curbing this harmful content or whether it shifts abusive content to unregulated platforms. I exploit synthetic difference-in-differences to assess the effects of the new Twitter’s policy, and I construct text classifiers using supervised methods from natural language processing to predict whether a post can be considered targeted hate speech. I find that racist content decrease in percentage both in Twitter and in Parler, but that users that have accounts in both platform increase their racist posts on Parler after Twitter’s policy. Finally, I exploit word embedding to see if the language used toward minorities changed after the regulation, and I find that terms related to some minorities became more offensive in Parler after Twitter’s policy. Overall, my results show a possible substitution effect between platforms, with more abusive content that shifts from regulated to less regulated social media platforms.
"Experience matters? The effect of political experience on municipal policies" (Joint with A. Manello)