Evolution of Friendship and Advice Networks in Blended Data Science Course

We investigated the network dynamics and the influence of structural characteristics on changes in network ties, as well as parameters of agents in the network, such as academic performance, gender and educational program of a student. Our sample consisted of 193 students attending the blended Data Science course at St. Petersburg campus of HSE. The analysis was carried out using the methods of social network analysis (SNA), and the method of constructing stochastic agent-oriented models (SAOMs) included in the RSiena package adopted for R programming language. In the study, we have found that in both friendship and advice networks, students with the same grades were likely to choose those who were similar to them as friends and advisers. In addition, friendship and advice influenced each other positively: friends have chosen friends as advisers, and advisers have chosen advisers as friends. We have also found that students with a high level of academic performance were often asked for help, while themselves they rarely asked for help from other students.