Likelihood inference under latent dyad-independent network models

Extend the exact likelihood correction for multi-wave snowball sampling to latent space models, stochastic block models, graphon models, and other network models in which edges form independently across vertex pairs conditional on latent vertex-level quantities.

Background

The exact derivation in the paper relies on unconditional edge independence in the Erdős–Rényi model. The discussion proposes extending the framework to richer models that preserve dyad independence conditional on latent variables, including latent space, stochastic block, and graphon models. Such models could represent degree heterogeneity, clustering, or community structure that the Erdős–Rényi model cannot capture.

Although the authors expect the likelihood correction to require comparatively modest modification in these settings, they do not provide the extension. A complete treatment therefore remains future work.