Next, we will derive bipartite network of affiliations between respondents and non-respondents, and examine its properties, first separately through bipartite ERG modeling in statnet (Goodreau et al., 2008; Hunter et al., 2008), then jointly with respondents network using valued projection of affiliation networks as a covariate in one-mode ERGMs, and finally we will try to simultaneously model both network as a two-mode network in MPnet software (Wang, Robins, Pattison, & Lazega, 2013).