This chapter treats four methods for multiple independent network data sets, all actor sets being disjoint and without connections between them, and all satisfying the same model specification. It starts by discussing the idea of a population of networks and the challenges related to group size differences. The first, so-called multigroup method, assumes that all networks are direct replications of each other in the sense that the same model with the same parameter values is used for all networks. The second method is a random-coefficient multilevel model and assumes that the networks obey the same model but with different parameters, following a normal distribution in the population of networks. This is estimated using a Bayesian approach. The third assumes that the networks obey the same model but now with different parameter values that have an arbitrary unknown distribution in the population of networks. This is a kind of meta-analysis. The fourth method models the networks separately and then combines them without assuming that they come from a common population of networks. It uses a combination of independent tests. The chapter ends with a comparison between the model assumptions.