ICPSR Summer Program - Part II

Now Alexander Kustov is telling about his experience of participating in the ICPSR Summer Program

The ICPSR Summer Program in Quantitative Methods of Social Research “is recognized throughout the world as the preeminent forum for basic and advanced training in the methodologies and technologies of social science research”. Thousands of young researchers visit Ann Arbor every year to take statistical courses provided by world-famous scholars. This year (as well as previous) some of researchers of our Lab participated in the ICPSR Program. We have recently published a report of Evgenia Bystrov and Anna Nemirovskaya on their experience of participating ICPSR-2012. Now Alexander Kustov is sharing his impressions about the event.

The second session of the ICPSR included many courses, which were interesting for me and relevant to my current research. That was why it was extremely hard to choose between them so, although I had some already specified preferences, I decided to spend first days on attending as many lectures as I could. After about a week of struggling, I chose "Advanced Topics in Maximum Likelihood Estimation" and "Complex System in Social Sciences".

Despite its name, the Advanced MLE course was primarily focused on handling temporal data and oftentimes (quite ironically) dealt with various OLS-based estimators. Basically, it contained two separate full-fledged courses. The first part included survival/event history/duration analysis and was taught by Brad Jones from UC Davis (the author of the prominent textbook on event history analysis). The second part held by David Darmofal from the University of South Carolina covered TSCS analysis (time-series cross-sectional) and methods for addressing spatial dependence in data, which I urgently needed in my research on the welfare state. 

The Complex Systems class was very different from other ICPSR courses and, frankly, from any other methodological class I had attended before. First of all, we had five different lecturers with a diverse expertise. In general, the course was based around various computational techniques, which might be employed in social science research. We had some game theory, complexity theory, genetic algorithms, programming, network analysis etc. The central topic of the course was the agent-based modeling (ABM) - a relatively new computational tool, which allows simulating artificial social systems under a given set of rules and thus revealing some patterns of behavior inaccessible by more traditional analytical techniques such as game theory. The course included numerous lab sessions on its implementation in NetLogo software and practical discussion of its applications in social science. The course made me really enthusiastic about the method and I hope to use it in my following project on the role of ethnicity in intra-state wars and insurgencies. 

Apart from the courses listed above, I also occasionally attended Advanced Game Theory, where we thoroughly discussed some models of conflicts, and Causal Inference, where I had a chance to improve my knowledge of such trendy statistical technique as instrumental variables and propensity score matching. Our classes were relatively small (about 20 people) and thus we had quite engaging atmosphere in the courses I took. Overall, I really enjoyed the session, which was probably the best possible complement to the rigorous methodological training I acquired at the University of Mannheim.