Estimating Temporal Exponential Random Graph Models by Bootstrapped Pseudolikelihood

This package vignette is designed as a hands-on tutorial for estimating temporal exponential random graph models (TERGMs) (Desmarais and Cranmer 2010, 2012b; Hanneke et al. 2010) and assessing goodness of fit and predictive performance (Cranmer and Desmarais 2011; Leifeld and Cranmer 2014) using the xergm package (Leifeld et al. 2014) for the statistical computing environment R (R Core Team 2014). The xergm package is compatible with the syntax of statnet (Handcock et al. 2008; Goodreau et al. 2008; Morris et al. 2008) and uses some of its functions, particularly from the ergm package (Hunter et al. 2008) and the network package (Butts 2008). A basic familiarity with ERGMs and their estimation, as in ergm, and network data management, as in statnet, is assumed and thus not treated extensively in this tutorial. Throughout the examples provided below, a dataset collected by Andrea Knecht on the dynamics of adolescent friendship networks in a Dutch school class is used as an illustration (Knecht 2006, 2008; Knecht et al. 2010; Steglich and Knecht 2009). This dataset is delivered with the xergm package and is the classic textbook example for estimating stochastic actor-oriented models (SAOM) using SIENA and RSiena (Ripley et al. 2011; Snijders et al. 2010).