XXXIII Sunbelt Social Networks Conference

Participants of Sunbelt XXXIII (Hamburg, Germany) talk about the Conference.

XXXIII Sunbelt Social Networks Conference took place in Hamburg (Germany) from 21 to 29 of May 2013. It was the official conference of the International Network for Social Network Analysis (INSNA). INSNA currently has over 1,000 members and more people than ever are interested in attending and presenting their work at Sunbelt conferences. Workshops and conference sessions allowed scholars interested in theory, methods, or applications of social network analysis to share ideas and explore common interests. More than 1000 people attended almost 700 paper presentations and discussed 100 posters. There were three participants from our Lab. 

Diaspora Communities on Social Network Sites in the Former Soviet Union, Alexey Gorgadze and Daniel Alexandrov.

Our research is focused on migration, ethnicity and identity politics in Web 2.0 in the former Soviet Union (fSU). Many young people of Armenian or Georgian origin are living in Russia because their parents moved there or because they went to study in Russian universities. Social network sites provide them with powerful tools to form e-diasporas and engage in identity work and boundary work, forging networked virtual communities with different identity labels. We study diaspora groups on VKontakte (VK), Russian social network service most popular among young Internet users in the fSU - about 195 million accounts by December 2012. Certain open features of VK profiles allow for analysis of geographic mobility and educational background of users, for example, how many graduates of high school in Erevan (Armenia) are now studying in the universities in fSU cities. Another important feature is the ability to make 'friendship' ties between groups and public pages, which allows for network analysis on the level of groups. In particular we interested in groups and public pages that create virtual communities with regional 'supra-ethnic' identities (“We are from the Caucasus” or “United Caucasus”) as many users participate both in ethnic groups narrowly defined ("True Armenians" or "Real Azeri Men") and 'supra-ethnic' groups at the same time. We will present the data on ethnic and geographic composition of different groups, on their activity, and on inter-group ties.

Sociometric Popularity in Academic Context, Vera Titkova, Valeria Ivaniushina and Daniel Alexandrov.

Our study is focused on relation of sociometric popularity and individual academic achievement in different academic contexts, i.e. in classes and schools with different levels of academic culture and educational aspirations. The relation between academic success and popularity is found to be controversial, depending on the ethnic status, gender, and class's academic achievement or academic aspirations (Adler 1992, Rodkin et al 2000, LaFontana & Cillessen 2002, Meijs et al 2010). Schools with different academic culture/values provide different conditions for peer popularity among boys and girls, high and low achievers. Our study is based on the survey of 5,058 students from 270 classes in 98 schools in St. Petersburg. We use two independent methods to analyze our data and test the influence of individual academic achievement, gender, family SES on the sociometric popularity: p * and hierarchical linear modeling. First, we analyze cross-level interaction between individual academic achievement and various indicators of classroom context. Second, we produce the p * models for class networks of three types: low, medium and high academic motivation. We find that in classes with low academic motivation student's academic achievement is negatively related to her/his popularity, and we find the threshold effect of academic context - there is a level of aggregate motivation at which the relation is reversed. We also find that relation between individual academic achievement and popularity is highly gender-specific.

Multilevel Analysis of School Peer Effects, Daniel Alexandrov, Valeria Ivaniushina and Mark Tranmer (University of Manchester, United Kingdom).

Peer group effects is well known and widely studied by sociologists of education. However "peer group" is a very loose concept. Very often for modeling purposes with school survey data peers are understood as school or grade cohort. While school composition influence is important, we believe that immediate student's friends within school are most influential. There is no doubt that schools are not homogeneous but divided into small groups with their own group norms and attitudes. The aim of our projects is to compare peer influence on different levels of social organization. We use data collected in a large school survey in St.-Petersburg in 2010 (104 schools, 7300 students) which has network component eliciting full networks in classes. We identify four levels of social organization and analysis: individual, egonetworks (immediate friends), cliques and schools, and we employ multi-level hierarchical regression analysis to differentiate the effects on different levels. Cliques (defined as tight social group within social network) were identified from complete network data using special Kliquefinder software (Frank, 1995, 1996). We investigate students' proschool/anti-school attitudes on different levels and their relation with educational outcomes. We demonstrated that socio-economic differentiation of Russian schools does not lead to polarization of pro-school/anti-school attitudes in different types of schools; polarization of attitudes emerges and is maintained on the clique level; clique attitudes are significantly related to educational outcomes and cliques are better predictors than egonetworks.