GESIS Training News

Mark your calendars for 26 February to 15 March 2024! Get ready to embark on a journey of learning and innovation. The GESIS Spring Seminar is renowned for providing high-quality training in state-of-the-art techniques in quantitative data analysis taught by leading experts in the field. Each year, we select a topic that reflects the latest developments and innovations in social science research methodology. Next year's topic will be "Recent Developments in Longitudinal Data Analysis".

By attending the Spring Seminar, you can expect the following:

🔍 In-depth knowledge of state-of-the-art quantitative research methods in the social sciences.

💡 Application of cutting-edge quantitative analysis techniques to your own research projects.

🤝 Engaging discussions with interested colleagues and leading experts.

🌐 Networking opportunities in a supportive social environment.

The Spring Seminar is tailored for advanced graduate or PhD students, post-docs, and junior and senior researchers who are eager to enhance their knowledge and skills in state-of-the-art techniques in quantitative data analysis. Each course combines insightful lectures with hands-on exercises, providing you with the opportunity to apply these methods to real data. The courses will be taught in hybrid mode. Don't miss out on this enriching experience! Stay tuned for more information here.

Michał Bojanowski is an assistant professor at the Chair of Quantitative Methods and Information Technology at Kozminski University and a researcher at the COALESCE Lab at the Autonomous University of Barcelona. He holds a PhD in sociology (2012, ICS / Utrecht University), and his research focuses on modeling social network data, especially those collected with non-sociocentric designs, as well as on assembling complex social network datasets from non-obvious sources (such as historical archives) often using technically-advanced procedures. Michał is an R developer with over 20 years of experience in writing packages and providing training in academic and business contexts. He is a member of the Statnet Development Team, maintaining a suite of R packages for statistical network analysis.

He will teach the course "Social Network Analysis with R" at the Fall Seminar in September 2023.

How did you become interested in your subject?

Michał: My interest in network analysis came to be by means of a kind-of a "Bang!" in the early 2000s, two years after I graduated with MA in Sociology from University of Warsaw. I knew very little about networks then, and even less about their relevance for social science in general. By the "Bang!" looking back now, I mean a couple of things came together somewhat unexpectedly. On the one hand, my interest and relatively strong skills in data analysis. On the other hand, a theoretical void or dissatisfaction with social scientific theories. The third element, a catalyst, or an event rather, was me coming across research by Vincent Buskens on trust and social networks. He was using game theory to derive testable sociological hypotheses. I was fascinated because I knew game theory a little bit but treated it as abstract math prescribing how to act rationally in interactive situations, not an approach that can be used to explain observable phenomena. By effort and luck, a year and a half later, Vincent became my PhD supervisor at the ICS at Utrecht University, and I was making a deep dive into the world of social networks, not without the help of Stephanie Rosenkranz and Werner Raub, who complemented the advisory team. Everything was exciting, new, and challenging.

Of all the fascinating instances in which networks can be studied I was always finding network formation questions the most interesting, i.e.: why and how various types of social ties come into existence. I mean here ties in social systems in which, for example, people compete with some set of actors but collaborate with a different set — a situation not unlike between scientists. Or pursue forming relations with others, but yet would like others not to do the same — not unlike adolescents competing for friendship and attention of peers. The sociological question is then what kind of structures of social networks (e.g., in science or school class) we should expect from such lower-level processes. That's fascinating! Even more so if we acknowledge that each tie, when you think of it, represents a relation initiated privately, or even intimately, by a pair of social actors, but when looked at in their totality in a larger population, combine into a large and complex network, which constitutes a social context, thus an antecedent, for the creation of that very tie in the first place.

What lessons can participants draw from your GESIS course?

Michał: I intend to have everybody finish the course feeling empowered and enabled in several ways. Firstly, the participants should feel enabled to perform their own social network analytic project using R by themselves, comprising all the typical stages: import and transformation of network data, descriptive measures giving a sense of structural features, visualization, and perhaps some elements of statistical analysis. Still, a week-long course might not be enough for everybody to get to that level because SNA is a rich paradigm, and the set of tools for SNA in R is big and growing. Thus the secondary but equally important skill is to know how to learn on their own, e.g., how to start using an unfamiliar R function, how to search and read documentation, how to interpret error messages, where to look for help and additional information, and so on. By experience, I think this is an often overlooked aspect of R training, even though R seems unwelcoming to many just because of that. I hope we will disenchant R on that front during the course. Thirdly, I hope the course will broaden participants' social network science imagination in the sense of possible research questions and mechanisms that may hide behind their network data, which they perhaps did not consider in their research.

What do you enjoy most about being a social scientist?

Michał: I think, being a curiosity-driven person, I would say I enjoy the most the breadth of research problems out there, ready to be pursued, and the possibility of jumping between them. I don't think this freedom is achievable in any other profession. Another related aspect I enjoy a lot is interdisciplinarity which enables me to find a common language with, say, statisticians, historians, or lawyers with whom I collaborate on a couple of distinct projects currently.

What do you think is the most exciting recent development in your field?

Michał: There are many exciting things happening currently, even more so because network science attracts researchers from many different fields. I think the most exciting developments are in the statistical models of networks: static, dynamic, multiplex, and so on. The new methods and approaches (such as Exponential-family Graph Models and their extensions or Stochastic Actor-Oriented Models and their extensions) open various new possibilities. One example is the ability to express multiple theoretical ideas, new and classic ones, e.g., about network multiplexity, and rigorously test them against data. Another example is the potential of applying these methods to new datasets that are becoming available in areas beyond core social science, such as history or archaeology.

We thank Michał for his exciting insights and look forward to his course.