19:05 February, 21 2011

Daniel Alexandrov tells about the conference:

This was one of the liveliest and nicest conferences I have ever been to for the last 20 years. International Network for Social Network Analysis (INSNA) was created about 30 years ago and has immediately launched a series of annual conferences that are called Sunbelt as they are always held in sunny places. In the past years the conference is held on different continents: once in Europe and then twice in the USA (in California and Florida). The last Sunbelt Conference XXX was held in an Italian town Riva del Garda near a marvelous lake. This year Sunbelt XXXI was organized in a town called Saint-Pete Beach on the shore of the Gulf of Mexico.

Many scholars came to participate. Last year the conference gathered 800 researchers, this time there were more than 700 people. Last year we attended the conference at our own expense without a presentation just to have a look how it is organized. I think it is important to listen to some presentations and to see which scientific trends this community finds normative and suitable for this conference. After that first visit we have understood it well enough, so that for this conference we have prepared 2 papers which were presented in Florida.

I would like to specify that Social Network Analysis (SNA) is more a method (or a complex of methods) than a theory or a subject field.  As these scholars are focused on methods, network analysis community unites people that are really interested in each other. The level of cooperation and co-authorship is much higher here than in other branches of sociology. Long ago Randall Collins has written an article, where he stated that sociology won’t become a high-consensus rapid-discovery science like experimental physics or molecular biology until sociology finds some new methods and research techniques. Standard techniques in mathematics and physics help to develop these sciences quickly.  The humanities do not have such techniques; that is why circulating of scientific capital in this field is slow. Also there is no agreement on which results are good and which are not. But in contrast with the other sociological methods we can call network analysis a “high-consensus rapid-discovery science”, as it is focused on upgrading and refinement of the techniques and the methods are being developed rapidly. The reason for that is vigorous growth of the field and its attractiveness for the researchers (nowadays more and more scholars study networks). Another point is that contemporary IT and software give more opportunities to count fairly complicated cases that are required for network analysis.

The scholars in this field need a broad academic community. First of all, they need help from those investigators who work in practical fields and have experience in community studies, academic cooperation, internet communication and student behavior. Methodologists search for good databases which are not easy to collect to test new methods and techniques of analysis.

The second point is that the methodologists need to be in the community to share methods, develop techniques and software and discuss problems. And those scholars who focus on empirical analysis need theorists and methodologists as well. I have rarely seen such a demand for sophisticated technical details and good empirical results in socio-economic sciences before. One more fact is that the community is fairly friendly towards the newcomers.

The conference itself was organized in a way that seems unusual to us. There were no plenary sessions and no division to more and less important reports. Several sections were held simultaneously and the time for each speech was strictly fixed. It was no more than 20 minutes for each speaker and everybody kept to the schedule. If anybody didn’t come, the time would be left free. It was important as many scholars have preplanned which lectures to visit and switch from one session to another according to their interests. This whole process can be also described as an academic network which has appeared because of rapid development of contemporary humanities.

Sessions are usually organized as a discussion of a certain method or a problem, e.g. Online Social Networks, Academic and Scientific Networks, Mathematical and Statistical Network Models etc. Various sections attract more or less people. Of course, those sections where the leading scholars in network analysis participate are the most popular. This year the most people attended the section called Network Methods, where the leading scholars in this field, such as Linton Freeman, Katherine Faust and Steve Borgatti were participating. It was held in a spacious hall and the lecturers had to speak through a microphone for everybody to hear them. At the same time other sections that were held simultaneously gathered those scholars who were not interested in mathematical methods, but concern themselves with network analysis of social capital and qualitative models in network theory.

We’ve been working there noon till dawn. The lectures started at 8:40 and finished at 17:30 or at 19:00. We were coming early in the morning and were running from one section to another as all the other participants. As there are several colleagues at HSE who analyze blogs, I wanted to share the most recent trends in these studies with them and I’ve attended several lectures on data analysis in the Internet. One of the brightest speeches in this section was the lecture of James Danowski who told us about a new method of analysis of semantic similarities. It was illustrated by content-analysis of online forums in the USA and Pakistan.

Many Russian colleagues of ours have attended the conference. The most of them work all around the world, but I’ve met Alexander Semenov from HSE (Moscow). Our reports were held at the same time that is why I did not manage to listen to him, but it was a pleasure to meet him and see that HSE is well represented at this important international conference. My lecture and the lecture of Valeria Ivanushina were held at different sections. One was about the impact of ethnicity at the choice of friends in students’ networks based on our project about migrant children. Another was showing a network of scientific cooperation where INTAS program was taken as an example.

We have bound contacts with our colleagues from Netherlands, for example, from the University of Groningen which is known worldwide as one of the strongest international centers on methodological development in network analysis. Tom Snijders, who created this scientific school in Groningen, is now Oxford professor and editor-in-chief  of “Social Networks” academic periodical. We’ve met his students a year ago and they keep consulting us since then.

Snijders has invited us to participate in a working group that designs models for multilevel network analysis. Multilevel analysis is understood in a broad sense. On contrary to the plain selection that is understood as atomic and every individual is independent from the others. The networks are organized in a complicated way as they are based on intercommunication of individuals. Intercommunication has different levels: dyadic, triadic, in bigger groups, in cliques, etc. These levels are counted as complementary in between individual and school level that are usually analyzed by multilevel regressions. 

Multilevel Network Modeling Group (http://www.ccsr.ac.uk/mnmg/team/) unites outstanding scholars, such as Stanley Wasserman, Tom Snijders, Noshir Contractor, Emmanuel Lazega and other famous researchers. The group was organized by a statistician from Manchester Mark Tranmer, and it is fairly international. I believe that we were invited to participate because of our successful empirical work in collecting data about students’ networks and about international cooperation. This data is of high quality so that multilevel analysis can be used to examine it. This style of academic communication and cooperation was eminently practical for us. We have participated it special seminars to discuss certain narrow specific issues with professionals in this or that particular field. This work was really intensive, but I found it more useful than just attending the conference. We were asked to describe our sample and our methods of analysis in detail. Now we continue to participate in the work of this group, for example we are sharing our data with the colleagues for them to elaborate their techniques and models using this sample.

That’s how we got involved in an international project where scholars from Britain, Holland, France, the USA and Australia participate. Soon we plan to see the first results of this cooperation and to invite some of the colleagues from this group to St. Petersburg. For example, in April several colleagues from the University of Groningen will come to St. Petersburg to hold a seminar on network analysis and to give public lectures. We plan to develop these methods in St.Petersburg branch of HSE for the scholars to use it in various spheres and to give a chance to our students to participate in Sunbelt Conference as soon as possible.  

by Daniel Alexandrov.