Studies involving members of “hidden populations” can present unique challenges for researchers.  In many cases, it can be difficult to obtain a statistically representative sample from these populations, because they are difficult to locate, identify, and contact.

In an article published last year in BMC International Health and Human Rights, Sulaiman-Hill and Thompson discussed the challenges faced in recruiting Afghan and Kurdish refugees living in New Zealand and Australia in order to study their mental health, subjective well-being and perspectives on resettlement. In this mixed methods study, the authors used snowball sampling with multiple entry points to identify potential study participants.  Their analysis indicated that making use of multiple entry points was crucial for reducing sampling bias.

In an intriguing comment on this article, Sabin et al. pointed out additional sampling techniques, including time-location sampling (TLS) and respondent driven sampling (RDS), that have been successfully applied for studies of hidden populations.  These ‘socially invisible groups’ include sex workers, injection drug users, men who have sex with men, migrants, and other groups that are difficult to reach.  A key feature of these groups is that the members form social networks, which can be utilized to identify participants for research studies.

Time-location sampling relies on identifying places where members of the target population congregate, and sampling randomly from this population during defined times. One potential limit of this technique is that it has the potential to exclude members of the target population who do not necessarily frequent public areas. A paper recently published online in the Journal of Urban Health by Karon and Wejnert proposes a method of statistical analysis that could help exclude some of these biasing factors. Respondent driven sampling is a form of snowball sampling that relies on mathematical modelling to account for the bias introduced by recruiting study participants through social networks.

For more examples of alternative sampling techniques for reaching hidden populations, check out these recent publications from BMC International Health and Human Rights:

1.       For the study reported in their paper “Forced residential mobility and social support: impacts on psychiatric disorders among Somali migrants”, Bhui, et al. contacted primary care patients with Somali names, and also engaged with local Somali stakeholders to identify appropriate community sites for recruiting participants.

2.       For the study reported in their paper “Compulsory drug detention center experiences among a community-based sample of injection drug users in Bangkok, Thailand”, Csete, et al. recruited injection drug users by word of mouth.