Social Media, Turnout Rate, and Election (DSI-SRP)
This DSI-SRP fellowship funded Yuqin (George) Yang to work in the laboratory of John Sides in the Political Science department during the summer of 2021. George is a junior with majors in Economics and Political Science.
This project aims to examine how President Trump communicates with his voters and the general public during the 2020 election via twitter, as Trump employs twitter as his primary social media platform to convey his messages. To complete this research, George selects the tweets from August 24, 2020 (the Republican National Convention) to January 6, 2021 (the riots in Washington D.C.). After classifying the intents of the tweets and with the help of text mining packages in R, George finds out that over 50% of the tweets that Trump sent attack someone or something, including but not limited to Joe Biden, Nancy Pelosi, the governor of Georgia, the governor of New York, the news media, the big tech companies, and the voting system. In addition, the result generated by SentiStrength, a sentiment analysis tool that estimates the strength of positive and negative sentiments in short texts, demonstrates that Trump’s tweets during the 2020 Election have an overall neutral sentiment. Yet, we tend to note those aggressive and negative information Trump sent, especially those attack others. Because of that, George employs the OLS regression model to examine whether attack content would draw more attention for Trump’s tweets. After controlling relevant variables such as the number of days, whether the tweets have hashtags, whether the tweets talk about COVID-19 pandemic, whether the tweets address the riots in Washington D.C., whether the tweets discuss the healthcare reforms, etc., George discovers that the attack content has a statistically significant effect on both the number of retweets and favorites that the tweets can receive, showing that the more aggressive attack messages may draw more attention compared to other messages.
In addition to receiving support through a DSI-SRP fellowship, this project was supported and facilitated by the DSI Data Science Team through their regular summer workshops and demo sessions.