A groundbreaking report by the National Academy of Sciences on how to best gather information on gender and sexuality says gender identity, not biological sex, should be the default data collected and reported.

The report, released this month, provides recommendations to the National Institutes of Health for how to measure gender and sexuality throughout different health contexts, including in research, surveys, health care administration and clinical practice. Made up of 27 different branches, the NIH is the nation’s agency for medical research, as well as the largest funder of biomedical research in the world. Its global influence puts it in a unique position to model meaningful data collection on gender and sexual minority populations.

Accurate data collection is essential to understanding disparities across marginalized populations. Integrating these new metrics will help research programs better evaluate the specific health needs of LGBTQ+ populations.

LGBTQ+ communities have rarely been represented in general population surveys. This is the first rigorous study of how or when data on sex, gender or sexuality should be collected. Without a universal standard, data cannot be combined or compared accurately. For instance, one survey may have asked about gender and another about sex, and thus they ended up measuring different things. 

The report, funded by the Department of Health and Human Services, includes suggested language for survey questions, which data should be collected in which context, and extensive research on privacy concerns and how to minimize potential harm while collecting data. Nancy Bates, retired senior methodologist for survey research at the U.S. Census Bureau and co-chair of the committee that wrote the report, told The 19th in email that the nine-member committee was made up of researchers who identified across the gender and sexuality spectrum, specifically including two-spirit (a general term to describe the variety of culture-specific Indigenous genders and sexualities), intersex and gender nonconforming identities.

Overall, the report emphasizes capturing identity over biology and highlights the potential for discrimination if data isn’t collected in an organized, thoughtful way. Because these aspects identify marginalized populations, there is an inherent risk in collecting the data in the first place. To minimize discrimination and misuse, not all data is recommended to be collected in every context.

Many surveys conflate sex and gender, confusing both survey respondents and the analysts trying to draw meaningful conclusions from the data. The report defines sex as a grouping of biological characteristics, while gender encompasses identity and expression. The panel recommends collecting data on sex only in situations where it is explicitly required, like when a doctor is trying to schedule a Pap smear or prostate exam. This ties into the core recommendation to collect only the minimum amount of data necessary.

The report lays out a two-step protocol for classifying gender identity: Respondents are asked about the sex on their birth certificate and their current gender identity. Past research has shown that asking about only gender identity has led to undercounts of the transgender population. For example, maybe one transgender woman would identify her gender as transgender, and another would identify her gender as female. The two-step method ensures both are classified as people with “transgender experience,” which is relevant data in some contexts.

There are also strict privacy protections around how the data is reported, accessed and used. Many LGBTQ+ people have suffered when their identities have been involuntarily disclosed, and professional groups have highlighted the potential dangers of gathering sensitive data on marginalized populations. The authors conclude that population-level data-gathering is unlikely to result in direct harm to individuals.    

The authors note that the most likely harm to come through this method of aggregate data collection is that the insight or benefits provided never make it back to the community in question. They highlight best practices for sharing research findings with the communities participating in a study.

The panel’s recommended language for gender identity includes named options for only male, female and transgender. The limited options were chosen purposefully to prioritize aggregation; given the sample sizes of many surveys, if more specific options were included (e.g. “transgender male-to-female”) there likely would not be enough responses to both give meaningful insight and protect the privacy of respondents.   

A significant part of the report is dedicated to the design of standardized questions about gender, sex and sexuality. The final options for gender include male, female, transgender and two-spirit (a generalized Western term meant to refer to Indigenous identities outside of the gender binary). However, the option for two-spirit on both the sexuality and gender identity questions will appear only if the respondent has selected “American Indian or Alaska Native” as their race. 

The decision to show two-spirit as an option for only Indigenous people is rooted in both cultural competency and data quality. (Bates explained that the committee consulted with Indian Health Services about how to collect data on two-spirit individuals.) As part of the research for this report, a sample of survey questionnaires were evaluated to inform potential language and question design. One of those was the All of Us Survey, an NIH research program gathering diverse health information. All of Us included two-spirit as an option for gender, and an analysis of responses found that over two-thirds of people who identified as two-spirit did not identify as Indigenous. Restricting the choice ensures data on Indigenous populations is being accurately collected.

Another new area of focus detailed in the report is data collection for intersex populations. Little data is available on intersex individuals or people who have been diagnosed with disorder of sex difference (DSD). There has yet to be a federal population survey in the United States that includes questions on intersex status, and the most comprehensive survey of intersex adults was fielded just two years ago.

Intersex, or DSD, is a broad term that includes anyone born with characteristics of both sexes. Because of the varying nature of the designation and how it usually is based on clinical data, many people may not know whether they are intersex, which complicates data collection. Further, many hospitals historically have performed genital surgery on intersex newborns under the auspices of aligning a child to a binary sex — not always with the knowledge or consent of the parents. In this way too, it’s possible for someone to not know they are intersex. California is the only state to have condemned unnecessary genital surgery in children.

The panel recommends the collection of data on intersex traits in clinical settings where sex characteristics are needed to determine course of care. Bates said the committee took into account recommended measures jointly developed by the Fenway Institute and InterAct Advocates For Intersex Youth. However, they note that more research on the phrasing of questions is necessary. Notably, it is recommended that a question about intersex status be separate as opposed to included as an option for sex or gender. This is another example of how the report emphasizes the decoupling of social identity and biological characteristics.

A representative from NIH’s Sexual & Gender Minority Research Office told The 19th that their office is in the process of sharing this information across the organization and the Department of Health and Human Resources. In a public statement shared upon the report’s release, Dr. Lawrence A. Tabak, acting director of the NIH, shared the potential impact on all federal programs. “I hope that this report will facilitate the collection of sex, gender identity, and sexual orientation across the broader federal landscape.”