Hidden Discourses in Twitter Data: Examining the Limitations of Data Collection and Information Dissemination in a Climate of Repression

Abstract

Twitter is an information dissemination tool that has been utilized by news organizations and individuals alike to share relevant links and responses to important current events. It is recognized as having an informational focus amongst users, wherein events subsequently become trending topics. However, distinct language based user groups respond differently due to cultural manners in which the users engage with information as well as environmental factors, including political climate and algorithmic biases. This study analyzes how information relating to lese majeste is being communicated. By resurrecting Scott’s hidden discourse, this thesis displays that within social data passive resistance does exist through media tactics that produce creative methods for online activism in a chaotic language only understood by subgroups. Twitter’s algorithm has created a barrier to understand how data is archived. In this thesis, methodological considerations are taken in account to understand the manner in which information can be approached in a comparative analysis of English and a less-resourced language, Thai. Using a wide range of literature from software studies, anthropology, and new media, this thesis presents findings on lese majeste information behavior of activists in Thai and English language groups and offers a critical perspective on alternative channels of resistance. In this thesis, three key challenges for comparing two language groups using specific keywords are presented. Firstly, less-resourced languages have significantly smaller databases. Next, the key terms used to extract information are sensitive in the Thai political environment. Furthermore, user behavior may direct information flow to one language opposed to another. Additionally, there are algorithmic limitations in detecting keywords in language, resulting in the inability to identify participatory culture in digital spaces. Finally, using a mixed approach to data collection to derive information from these limitations, this thesis presents a methodological approach to identify a culture of information dissemination in a manner where resistance can exist under draconian laws.