Mixing methods to improve sexual behaviour research among socially marginalised populations: insights from a study on female sex workers in Indonesia

Reliable and valid data on sexual behaviour and their pathways among socially marginalised populations that engage in illegal or illicit activities, such as female sex workers, remain challenging to collect. Particularly little is known about FSWs working in rural areas. This paper uses the example of a comparative mixed methods study of rural and urban FSWs in Indonesia, to (1) describe how qualitative and quantitative methods can be combined to overcome research problems accentuated in studies of sexual behaviour within the context of transactional sex, and (2) to discuss the strengths, limitations and implications for practice of such an approach. Mixed methods in this study include a survey of rural and urban FSWs (n=310), in-depth interviews (n=11), key informant interviews (n=5) and ethnographic assessments. The sample of FSWs was drawn up using a novel combination of purposive sampling techniques. We show how the use of mixed methods can enable the collection of context-specific behavioural and socio-demographic data and can serve as a tool for internal validation. In addition, failure to include the hardest to reach and often most vulnerable FSW sub-groups, such as those working in rural settings, can be avoided through the use of combined purposive sampling techniques.