Respondent-driven sampling to recruit in-country migrant workers in China: A methodological assessment

Abstract

Aim: Respondent-driven Sampling (RDS) is a new form of chain-referral sampling, which is superior to random sampling and traditional non-probability sampling in hard-to-reach populations. We employed RDS to recruit internal migrant workers in Chengdu, Sichuan Province, China, and examined whether it could be successfully used in this population. Methods: 1,270 migrant workers were recruited into the study. Social demographic information, social network size, and geographic information about participants’ residential locations were collected. RDSAT software and geographic information system (GIS) technology were used to examine whether RDS was successful. Results: The results showed that the sample compositions converged to equilibrium very quickly. Sample representativeness testing results showed that females (t = 3.61, p < 0.001) and people aged 46 years old and above (t = 3.222, p < 0.001) were under-represented. GIS results showed that respondents were concentrated in the vicinity of the third road ring of Chengdu, especially in the areas of 3, 4 and 21, which were close to the investigation site. Conclusions: The results demonstrated that RDS is a robust sampling method in the study of migrant workers. Despite its potential utility, it is also important to recognize and mitigate potential limitations, such as geographic proximity.