Modifying Methods for Hidden Population Estimates

Successive sampling population size estimation (SS-PSE) methods used to estimate the size of hidden populations rely on the assumption that the underlying social network of the hidden population is fully connected. A new study, led by Laura J. Gamble, PhD, a Westat senior statistician, “Estimating the Size of Clustered Hidden Populations”, and published in the Journal of Survey Statistics and Methodology, addresses this problem by modifying an existing method for estimating the size of hidden populations whose social networks consist of disjoint clusters.

Gamble and the team of researchers propose two modification methods: a theoretically straightforward extension of SS-PSE that relies on additional prior information, and an extension of the Bayesian SS-PSE model with new parameters that allow for clustered estimation without requiring any additional prior information. After providing justification, they demonstrate the performance of both methods via simulations and application to displaced persons data.

Hidden population size estimation is greatly important to aid organizations and public health agencies because it helps assess the size of vulnerable and underserved populations in need of humanitarian assistance. This research extends the practical application of SS-PSE to populations with strong geographical, social, or linguistic divisions as well as populations that are spread over multiple cities. ~Laura Gamble