Hjalmar Alexander Bang Carlsen
Øster Farimagsgade 5
1353 København K
I hold an associate professorship in mixed digital methods at SODAS, University of Copenhagen.
Most of my teaching is in the Social Data Science master's degree program.
My research and teaching focus is on mixed methods strategies for collecting and analyzing digital data. My substantive research is in political and civic participation on social media.
Currently, I have three ongoing research projects:
1. Social Media enabled Informal Volunteering in Times of Crisis (SoMeVolunteer) Velux Core Group Program. (Joint PI w Jonas Toubøl)
SoMeVolunteer investigates informal support across three crises(refugee support in 2015, COVID-19 support, and refugee support 2022) organized on social media. It focuses on what we call deserving practices and how these influence the mobilization of volunteers, the distribution of support within social media groups, and the relations to other societal actors. It draws on a comprehensive mixed-methods dataset of surveys, interviews, newspaper articles, and social media data for each crisis mobilization.
2. Public and political participation on Social Media and its consequences. (Joint PI w. Snorre Ralund)
This project draws on a unique large-scale Facebook dataset of public participation on Danish Facebook. It investigates civil society actors' political communication, citizens participation, and how these influence one another.
One strand of the research focuses on gender inequality in political participation on Facebook. Both in regards to the unequal amount of participation, the kind of participation, and the causes of these differences. This part of the project is founded by the UCHP Data+ initiative.
3. AInterviewer - using LLMs to collect qualitative interview data at scale(pilot project). Funded with seeds money SODAS and the DataLab.
This methods project seeks to investigate the potential of using generative large language models(LLM) to ask open-ended questions and follow-up questions similar to the qualitative interview. It has three work areas. 1) Conceptual development of the AInterviewer which seeks to translate state of the interview methodology into a computational LLM setting. Addressing both ethical and epistemological questions around the LLM conducting interviews. 2) Implementation and test development, working to implement an AInterviewer that is ethically responsible, GDPR compliant, and produces high-quality questions. 3) Comparative assessment of data quality of AInterview in comparison with the open-ended interview and the human interview.