Prevalence Estimation for Hard-to-Reach Populations: Using the Network Scale-up Method for Estimating Child Trafficking Prevalence in a Low-Resourced Country

Estimating the prevalence human trafficking victims, including children, continues to pose a challenge for researchers across many disciplines. This challenge is exacerbated by the fact that human trafficking victims are hard-to-reach. Hard-to-reach populations are characterized by the difficulty in survey sampling by using standard probability methods. Typically, a sampling frame for the target population is not available, and its members are rare or stigmatized in the larger population so that it is prohibitively expensive to contact them through the available frames. Hard-to-reach populations in the US and elsewhere are under-served by current sampling methodologies mainly due to the lack of practical alternatives to address these methodological difficulties. In this workshop, we begin with an overview of the various methods of estimating the prevalence of hard-to-reach populations with application to human trafficking research. Our presentation is based on our work through the African Programming and Research Initiative to End Slavery (APRIES). APRIES is a consortium of researchers and policy advocates within the Center on Human Trafficking Research & Outreach. In collaboration with local researchers, we conducted household surveys with n=3,000 by nesting the Bayesian network scale-up method (NSUM) to estimate the prevalence of child trafficking and child labor in Sierra Leone. NSUM uses information about respondents' networks to produce prevalence estimates. It has been used in public health to measure hard-to-reach populations, including men who have sex with men, sex workers, and heroin users. A major advantage of NSUM is its allowance for the estimation of the size of target populations without interviewing members of the target population, which is particularly useful when estimating human trafficking prevalence. NSUM assumes that people's social networks are roughly representative of the local population. Thus, it is possible to determine the prevalence of a characteristic in the population by knowing the average prevalence of the characteristic in respondents' networks. To produce this estimate, researchers need to determine how many people are in each respondent's network (i.e., how many people the respondent "knows". We encountered and attempted to resolve several methodological and statistical challenges that will be presented in the workshop and that are applicable to others studying hidden populations in the US and abroad. Among these are the operational definitions of child trafficking; designing a representative sample; training enumerators and safely collecting data in the midst of a pandemic; using the correct network questions, addressing the various assumptions and their violations, and finally, presenting the findings to key stakeholders and a wider audience. We also present and explain the varying prevalence estimates from the household surveys and from the NSUM approach. Our goal is to display transparency in reporting findings that have huge implications for policy and programming among vulnerable populations. The US Department of State, Office to Combat and Monitor Trafficking in Persons funded the project.