Fabienne Lind is a postdoctoral researcher at the Department of Communication at the University of Vienna. Her main research interests are the advancement of content analysis methodology, comparative research, European media discourses about migration, and knowledge gap research. In her dissertation, she introduces, compares, and evaluates strategies for the automated analysis of text collections in different languages for cross-country research.
Fabienne Lind was part of the Horizon 2020 projects REMINDER and MIRROR and is currently a member of the H2020 project OPTED and the COST Action OPINION.
Hauke Licht is a postdoctoral researcher at the Cologne Center for Comparative Politics, University of Cologne, who received his PhD from the University of Zurich. He develops and applies computational text analysis methods to study political communication, electoral competition, and democratic representation. Taking a comparative perspective, Hauke often analyzes multilingual texts and thus has a strong interest in multilingual text analysis methods.
He also increasingly uses deep learning methods to analyze textual and audio-visual data.
They will teach the course "Going Cross-Lingual: Computational Methods for Multilingual Text Analysis" in Cologne in December 2023.
How did you become interested in your subject?
Fabienne: I became interested in multilingual text analysis when I was involved in a project that aimed to analyze migration reporting in Europe. What truly fascinated me was the challenge of working with texts in different languages and attempting to compare them. Indeed, there were already some approaches to making these comparisons possible through human interpretation and analysis of data.
However, in the realm of automated analysis, which was still relatively new in the field of social sciences at that time, there was a lack of clear guidance and established methods. This presented an exciting and challenging opportunity to explore innovative techniques and contribute to the evolution of the field.
Hauke: Like Fabienne, I also came to the topic as a practitioner. I wanted to study to what extent and under what conditions political parties criticize “the elite” as part of their electoral strategies. But to implement a suitable comparative research design, I needed to measure parties’ emphasis on anti-elite rhetoric across countries and hence languages. I was aware of the option of automated (“machine”) translation to “solve” this problem. In this case, I could just have applied established text-as-data approaches to texts’ English translations. But when I calculated how much it would cost to translate all texts in our dataset using Google Translate, DeepL, or another commercial service, I was shocked.
So I started to look for alternatives across the political, communication, and computer science literature.
What lessons can participants draw from your GESIS course?
Fabienne & Hauke:
One of the core takeaways from this course is the importance of collaboration with case and language experts when undertaking multilingual text analysis projects. For instance, when analyzing social media data from various regions, having a local linguist who understands the nuances of regional slang or dialects can significantly enhance the accuracy of sentiment analysis. Collaboration with experts ensures that the context and cultural nuances of the text are adequately considered, which can make a substantial difference in the project's outcomes.
Moreover, another takeaway is that numerous tools and resources are readily available to facilitate multilingual text analysis projects. We will provide participants with the knowledge and hands-on experience to harness these tools effectively in their projects, streamlining the analysis process and enhancing the quality of their research.
What do you enjoy most about being a social scientist?
Fabienne: Being a social scientist for me is like using a mirror to understand how society changes over time. The mirror isn't perfect, and what it shows is like a constantly moving puzzle.
But I really enjoy the challenge of trying to figure out some of the mysteries.
What do you think is the most exciting recent development in your field?
Hauke: The aspect of research about multilingual text analysis I find most fascinating is its level of interdisciplinarity. Today, social, communication, and political scientists are increasingly aware of the opportunities of multilingual text analysis. Researchers from all across social science seem to realize that analyzing texts across languages can allow comparative insights into communication and political behavior and help understanding how context and culture shape discourse more generally. I believe that this development is in large parts driven by methodological research that expands our theoretical knowledge about multilingual text analysis as well as our methodological toolkit. What’s striking is that while many of these contributions come from social science researchers, they often import and adapt insights and methods developed in (computational) linguistics as well as from neighboring fields within the social sciences.
This makes research on multilingual text analysis in the social sciences extremely interdisciplinary.
We thank Fabienne and Hauke for their exciting insights and look forward to their course.