Recent Advances on Estimating Population Size with Link-Tracing Sampling

Social workers are committed to advocating the well-being of hard-to-reach or hidden populations through research and practice. These include populations that use drugs, unhoused people, men who have sex with men, and human trafficking victims. It is challenging for social work researchers to establish links and connections with these populations as they may experience stigmatization and marginalization, which are compounded by structural inequities in our society and across the globe. These populations are also underserved, thus increasing their vulnerability and compounding their trauma. Serving these populations or designing evidence-informed policies to reduce the prevalence of the problem is challenging. While strategies for documenting the attributes of hidden populations through link-tracing/network sampling designs have been well studied, methods for accurately quantifying the number of individuals in the study population are scarce. This shortage of accurate prevalence measurement methods contributes to the lack of understanding of the scale and scope of the problem in social work and social science research. From a research perspective, probability samples from these excluded populations are either prohibitively expensive or impractical to administer as population members may not reside or work in the census enumeration areas - they often are clustered in their hidden social networks. Given these challenges, innovative sampling methods are essential for studying and supporting these communities. Current methods include link-tracing or network sampling designs that leverage social connections between individuals for peer-recruitment purposes. These methods comprise the class of a) snowball, b) chain referral, c) respondent-driven, and d) adaptive web sampling designs. The workshop presents a novel approach to link-tracing sampling that enables researchers to estimate the full population size and network structure of a hard-to-reach population. Participants will gain an understanding of: 1) a peer-recruitment strategy, allowing researchers to establish connections between respondents and thus reveal the underlying network structure; 2) an estimate of the full size of the hard-to-reach population; 3) employing a stratified framework to select the initial sample of respondents and prioritizing factors influencing the immediate accessibility of individuals. Emphasis is placed on capturing diverse demographic profiles and achieving maximum variation. Respondents are added in waves through a model based on covariates including strata assignments, network size, and propensity to accept or decline the invitation for study participation through peer recruitment. The workshop stems from our research in human trafficking prevalence estimation as well as some novel developments and simulations in the US with other at-risk populations. Following successful simulation studies, we have used the link-tracing method effectively to provide accurate results for estimating individuals experiencing human trafficking, including labor and sex trafficking, in various regional and national contexts. Both the theoretical analyses and simulation results demonstrated that our proposed method gives approximately unbiased and highly efficient estimates for the size of the study population. Workshop participants will engage in discussions on innovative methodological approaches for reaching and studying hard-to-reach populations. Additionally, they will participate in hands-on demonstrations illustrating the practical application of a novel link-tracing approach and tools for implementing this methodology in their research.