X hits on this document





20 / 86

10 Emilyzen Ignacio and esenia Mejia

were often completed with an Indonesian migrant advocate interpreting, facilitating and fielding questions about the survey.

3.2 Informational interviews

Snowball sampling was also used to obtain informational interviews with government stakeholders in Hong Kong, including interviews at the Hong Kong Labor Department (LD), the Philippine Overseas Labor Office (POLO), and the Consulate General of the Republic of Indonesia. Likewise, interviews with migrant worker service providers and CSOs were obtained through snowball sampling and Internet research.

3.3 Limitations

Reliance on snowball sampling and the desire to increase the sample size by conducting migrant interviews on non-typical, weekend rest days, resulted in a disproportionate amount (27 percent, n=162) of surveys conducted in shelters. Migrants temporarily residing in shelters have had problems with their employers, and as a result, their experiences may differ significantly from those currently employed. Nevertheless, their stories and experiences provide valuable information about critical work conditions in Hong Kong. Other sources of bias include: selection, non-response, and response biases.

Selection bias: Because most of the survey participants were introduced to the authors by NGO and union leaders, a large subset of migrants that are not affiliated or involved with organizations are not included in the sampling.

Non-response bias: Some surveys were completed by the migrants themselves and had some blank responses. Without follow-up from the research team, these non-responses could potentially be excluding valuable information for analysis.

Response bias: The use of interpreters for group surveys, particularly NGO service providers or union leaders, may have influenced migrant responses, especially when the migrants were receiving services. Pressure from peers, especially in group settings and acknowledgment of the research team’s organizational affiliations may have also affected participant responses.

Document info
Document views428
Page views428
Page last viewedSun Jan 22 00:18:20 UTC 2017