To disentangle various effects on tie formation, we applied exponential random graph modeling (abbreviated as ERGM, and also known as p*, see – Lusher, Koskinen, & Robins, 2012; Goodreau, Kitts, & Morris, 2009; Jackson, 2008; Robins, Pattison, Kalish, & Lusher, 2007; Robins et al., 2007; Snijders, Pattison, Robins, & Handcock, 2006; Holland & Leinhardt, 1981). This framework regards network as a set of random variables, which makes an observed network only a random pick out of large set of possible networks (Robins et al, 2007). ERG framework also respects inter-dependencies among variables, and allows to associate probabilities of different possible networks with their inner structure (Besag, 1974; Frank & Strauss, 1986; Jackson, 2008). Conceptually, this might mean that a network has chances to be observed depending on the presence or absence of particular patterns in its structure. The exact form of these ‘important’ patterns is implied by dependence assumptions, that is a priori believes about how networks ties may depend on each other.