Freedom to Tweet or Tweet to Freedom: The Relationship between Freedom Status and Tweets during Elections

Abstract

In this thesis, I conduct an exploratory study of the relationship between a country's freedom and the twitter activity during elections. While there have been many studies of Twitter and elections, there has been no previous research conducted to explore the relationship between a countries' freedom and how Twitter influences elections in that given country. My goal is to identify hypotheses for future work in this area, introduce research designs and to shed light on areas of research where there seems to be little indication of relationships. I explore this space with automated analysis of the tweets' text, election outcomes, freedom ratings for the countries, and sentiment analysis. My results show that there seems to be a weak relationship between the outcome of an election and the sentiment expressed towards a candidate in tweets and that there is no relationship between the freedom in a given country and the sentiment expressed towards the incumbent. I found promising initial results regarding the relationship among content removed from links during an election and freedom status of a country, as well as the correlation between how frequently a candidate is mentioned and the election outcome. In the discussion, I present research questions in areas that are promising for future work.