Recruiting the Rainbow: Sampling Methodologies for Recruiting Diverse LGBT Populations for Research

Background: Over the past decade, survey research on the health and psychosocial well-being of lesbian, gay, bisexual, and transgender (LGBT) populations has grown tremendously. However, the majority of these studies focus almost exclusively on white LGBT people, relegating the experiences of individuals who are both LGBT and ethnic/racial minority to the margins. In this presentation, we will discuss our experiences with targeted and respondent-driven sampling methodologies in two recent research studies with LGBT adults in which we successfully recruited large numbers of ethnic minority participants. Specifically, we will focus on the strengths and limitations of these strategies and will present considerations and suggestions for future research in this area.

Methods: The first project was a multi-phase examination of minority stress, coping, and health outcomes among diverse LGBT men and women. In order to obtain a sample that was ethnically and racially diverse, we utilized targeted sampling strategies. In the initial phase, we recruited individuals for a qualitative study in Washington State, yielding 46% ethnic minority representation in the sample. Phases 2 and 3 were national web-based surveys in which we modified our targeted sampling strategy to reach participants on a national level; our efforts yielded 30-40% ethnic minority participation. The second project, which is currently in progress, is a comprehensive health survey of ethnically diverse lesbian and bisexual women in Washington State. Our sampling plan initially focused on Respondent Driven Sampling. While we successfully recruited a diverse group of “seed” participants who demonstrated enthusiasm for passing on recruitment “coupons” to their social network members (nominees), a very low percentage of nominees actually contacted the study office to participate. Thus, also we recruited volunteers through targeted recruitment methods. Contrary to our expectations, volunteers were actually more diverse on the dimensions of interest than the RDS participants.

Results: Targeted sampling can yield diverse study samples both nationally and locally when advertising and networking are tailored specifically to communities of color. Respondent Driven sampling shows some promise as a means of reaching individuals who would not otherwise volunteer; however, it did not improve upon Targeted Sampling in terms of inclusion of women of color. Individuals with multiple marginalized identities may need to hear about research studies from multiple sources before they are willing to consider participating.

Conclusions: We will conclude our presentation by looking at our experiences with sampling diverse LGBT populations in social context, examining how historical and contemporary intersecting oppressions contribute to sampling challenges with these populations. This presentation has important significance for social work research and practice. In terms of research, it provides valuable insights for recruitment and inclusion of stigmatized and understudied populations in research. We will offer suggestions for further research in this area and engage session participants in discussion about how best to carry out sampling among these populations. In terms of practice, our presentation will offer service providers with insights into the challenges faced by LGBT people of color and both the strengths and limitations of the current state of knowledge about this group.