Collecting data from migrants in Ghana: lessons learned using respondent-driven sampling

Background: Policymakers and program implementers require high-quality data on migrants and migration in low- and middle-income countries (LMIC); however, a shortage of high-quality data exists in these settings. Sampling migrant populations requires better techniques. Respondent-driven sampling (RDS) may be one such solution. Objective: Using Ghana as a case study, the objectives of this paper are to: 1) assess RDS recruitment productivity, network size, and ties of internal migrants; 2) test for homophily; and 3) detail the successes of and challenges to implementing RDS in Ghana and how these lessons can be applied to migrant populations in other LMIC settings. Methods: This study used RDS to sample 625 rural–urban female migrants working as market porters (kayayei) in Accra, Ghana. Results: This study generated the most comprehensive data set on kayayei to date. Network size increases as participants become more educated and migrate more often to Accra. Ethnic group membership is an important determinant of recruitment, with certain groups preferring to recruit from within. Employing members of the kayayei population to collect data built crucial trust. Conclusions: Whilst RDS is not a one-size-fits-all solution for sampling hard-to-reach migrants in LMIC, it can be a powerful tool to uncover and to recruit hard-to-reach migrant populations. In countries with multiple ethnolinguistic groups, recruiting a migrant population with greater ethnolinguistic overlap may facilitate quicker equilibrium. Contribution: This study expands the evidence base on use of RDS among migrant populations in LMIC and provides lessons learned to assist other researchers implementing RDS in LMIC settings.