Thus, two networks are the empirical basis for this research. The first network covers only respondents and includes 221 nodes; the second one covers both respondents and non-respondents and consists of 587 nodes. The following analysis will focus on the network with 221 nodes, while the respondent-non-respondents network with 587 nodes will be used for the check of robustness of results. This robustness check is very important inasmuch as the network with 221 nodes does not contain a significant part of information. According to the research strategy applied by Wimmer and Lewis (2010), we also use the multiple regression quadratic assignment procedure, thereby ensuring that our results are not an artifact of ERGM methodology.