Measuring sexual identity on federal health surveys

Patterns of poorer health have been noted for sexual minorities, specifically for lesbian, gay, bisexual, and transgender people, leading to the establishment of a reduction in health disparities by sexual identity groups as a federal priority in the United States. However, sexual identity is a complex concept to measure as it is a multi-dimensional construct, rooted in social and political contexts, and changes (sometimes multiple times) over the course of individuals' lives. Furthermore, as a subjective phenomenon, what actually constitutes sexual identity varies immensely among individuals. Community ties, cultural and political values, friendship relations as well as sexual behavior and desire have all been identified as central components of sexual identity. Consequently, individuals' sexual identities do not necessarily conform to discrete, objective, and uniformly-defined categories. Our goal, then, is to construct a question that most accurately reflects this construct, produces little response error and is relevant and accessible to respondents. In this paper we present findings from several question evaluation projects, which assessed questions used to measure sexual identity, focusing on how diverse groups of respondents interpret and respond to these questions. Data for these projects include qualitative data from cognitive interviews, as well as quantitative survey data. While findings indicate that sexual identity is complex phenomenon to capture on a survey questionnaire, we provide strategies for improving question validity.