Disinformation, Language, and Identity workshop 2021
The workshop included two plenary talks by William Dance and Dr. Philip Seargeant and panel sessions, with a total of eight speakers. The talks approached disinformation from a range of disciplines and perspectives including computer science, psychology, linguistics, and politics. The plenary talks and panel sessions were followed by Q&As, and the event was brought to a close by a closing discussion (led by Dr Tereza Spilioti, Cardiff University), where all attendees were invited to participate.
The final discussion focused on some recurring points that emerged across disciplines and various presentations in this half-day event. One of these points was the tension between mainstream and non-mainstream sources in the circulation of information online, where mainstream sources (e.g., mainstream news corporations and broadcasters) are often associated with distrust, while non-mainstream sources (e.g., online forums, twitter threads) can be associated with increased trust, often due to a perception that mainstream sources have hidden agendas. Such perceptions are linked with the kinds of conspiracies that often accompany the spread of disinformation narratives. Participants discussed the role of education and development of digital literacy as potential ways of reducing the risk of exposure to disinformation sources, while acknowledging that this is time-consuming and not everyone may be motivated to educate themselves on the sources that they use.
Furthermore, the participants discussed the cultural differences associated with dominant narratives: there are not universal norms about what is left/right, nor is there a universal acceptance that disinformation is an issue- who decides that it is a problem, and how do we account for these cultural differences? In response to these questions, we acknowledged issues of bias within research and, particularly, in relation to specific approaches to research that may reflect certain academic traditions. White, European perspectives can often dominate research frameworks, but it is important to reflect on potential bias to make sure that everyone is safe, without excluding certain groups by the adoption of primarily western perspectives.
From the perspective of the individual researcher, there was also discussion of the role that personal bias can play in research. For the sake of research replicability and objectivity, it was suggested that it would be helpful to have a strategy or model for identifying disinformation, particularly in regard to coding data that to some may seem ‘extreme’ but to others, might not. Similarly, we acknowledged the necessity for defining and clarifying concepts that are shared across disciplines, as was discussed by William Dance in his plenary talk. Nevertheless, bias and perspective in research are probably aspects that need to be problematised and acknowledged; all knowledge comes from humans who are also social actors and have particular positions. Provided that we, as researchers, acknowledge, understand and critically reflect on any biases in the context of our research, we can still produce meaningful findings and research outcomes.
With respect to research challenges in disinformation research, there was discussion about the need and the challenge of capturing what may not appear or may be ‘not there’ in data. While emphasis in data analysis is on documenting patterns that exist in texts, in the context of disinformation it is equally important to detect what is missing or the lack of specific patterns. For example, it is important to detect campaigns that may have resulted in no response from their targets (such as individual political figures). So, the issue is how to detect or quantify a lack of something. In response to this challenge, combining methods and approaches appears to be an appropriate way forward. Quantitative data from computer science perspectives can provide us with interesting graphs and visual presentations of networking nodes, providing a broad picture of what may be taking place within certain networks. On the other hand, qualitative approaches provide a closer understanding of data, paying attention to the context surrounding the object of study and identifying specific themes and patterns relating to the data set. A mixed method approach would therefore allow for a compromise between the challenges associated with each approach; in qualitative research, you can miss wider pictures, while in quantitative research, the picture provided can sometimes be too wide.
Summary
Emergence of main challenges:
· Bias and objectivity: It is important to acknowledge and critically reflect on the bias that comes with dominantly white, European research perspectives. One way to counter issues of both personal and wider biases is through the understanding and use of clear definitions and replicable models. This should also hopefully aid replicability.
· Challenges of qualitative/quantitative research models: Within both quantitative and qualitative research approaches, there are challenges that can lead to valuable insights into data being missed. To counter these challenges, combining quantitative and qualitative research method should lead to a more comprehensive understanding of data.