This document describes a strategy to:

  • simulate study level analysis by generating network data, implementing different sampling schemes, and implementing different procedures for estimating the size of hidden populations
  • simulate a meta-analysis that uses outputs from individual studies to estimate “meta-estimands”

Methods:

  • We introduce and use an R package hiddenmeta that gathers together various tools for these tasks.
  • The overall simulations are done using the DeclareDesign suite of packages.