Transwoman Population Size

The recent findings of Crissman et al.¹ on trans people in the United States provide much-needed insight on the makeup of the population and help reconcile dilemmas we have encountered in our research with transwomen over many years. In particular, our estimates of the number of transwomen in San Francisco, California (1528 in 2011, for example)² and our characterization of their socioeconomic status (SES)³ have struck us as too low (i.e., our estimates of income and education are lower than in reality). The authors found that 53.8% of trans people nationally are transwomen, and 50% live under 200% of the poverty level. In addition, Flores et al. found that 0.76% of adults in California are trans.⁴ Assuming that these parameters apply to San Francisco, there should be 2977 transwomen in the city (728 059 adults × 0.0076 × 0.538). Crissman et al. and Flores et al. validate our concerns and offer a calibration to correct our projections.

We believe that the discrepancy arises from how transwomen access or are reached for services and are sampled in research. Recruitment methods relying on peer referrals incentivized with cash, such as the respondent-driven sampling used in our surveys,³ tend to lead to recruitment from social networks of lower SES.⁵ Recruitment from publicly funded services may also tend to reach those of lower SES. Population size estimation methods based on such surveys and client registers (e.g., capture–recapture and multiplier methods) will underestimate population sizes because of the positive correlation between the two sources.⁶ They will also overstate the lower SES of the overall transwoman population.

When we reapportion our local estimates to the SES data described by Crissman et al., on the assumption that we undersampled higher SES transwomen, our new estimate for San Francisco would be 3107 transwomen, more than twice the 2011 estimate and very near the number predicted by Flores et al. We conclude that much of the research among transwomen in San Francisco and likely elsewhere, including population size estimates, measures of HIV prevalence, and other health issues, may be overly focused on lower income transwomen. This may also be true for other research on trans people relying on similar methods. Meanwhile, the health needs of higher income transwomen may be less well understood than those of lower income transwomen, and members of this population are deserving of efforts to include them in health research.