Measurement Concepts

The panel began its task to develop measures that captured aspects of the three distinct concepts—sex, gender, and sexual orientation—by defining the underlying concepts on which the measures are based. To develop these consensus definitions, each panel member submitted a definition for each concept, and the panel then discussed these submissions to identify their common elements and arrive at a consensus understanding of each concept. Although definitions from previously published literature were sometimes used as a starting point for these discussions, those definitions were modified as part of the panel’s discussions: the definitions used in this report reflect the panel’s consensus.

• Sex is a multidimensional construct based on a cluster of anatomical and physiological traits that include external genitalia, secondary sex characteristics, gonads, chromosomes, and hormones.

Sex in Western cultures is typically assigned at birth by medical professionals as either male or female, based solely on visual inspection of external genitalia. Sex traits are often assumed to be unambiguous and correspond to a single sex; however, sex traits may not all correspond to the same sex or with these binary categories, and thus with sex assigned at birth. For example, people who are intersex or have differences in sex development (DSD) have traits that do not correspond to a single sex.² Some sex traits may change or be altered over time, making sex both a complex and temporally fluid concept.

• Gender is a multidimensional construct that links gender identity, which is a core element of a person’s individual identity; gender expression, which is how a person signals their gender to others through their behavior and appearance (such as hair style and clothing); and cultural expectations about social status, characteristics, and behavior that are associated with sex traits.

As noted above, gender and sex are conceptually distinct, but they are often used interchangeably due to normative assumptions that sex is binary and immutable and that sex assigned at birth defines gender. Neither gender identity nor gender expression is defined by sex traits, however, and both can be temporally and contextually fluid. Gender is often conceptualized in Western cultures as man/male and woman/female, although it also includes categories that lie outside this binary, such as nonbinary (an umbrella term for the identities outside of the binary), gender-fluid (does not identify with a fixed gender), or, for Indigenous³ populations, Two-Spirit (see next section). Throughout this report, we focus on measures of gender identity, but it is important to keep in mind that measuring gender identity is one facet of gender. When we discuss the overall concept that encompasses identity, expression, and social and cultural expectations, we use the broader term: gender.

Transgender refers to a person whose gender identity is different from the sex they were assigned at birth; cisgender refers to a person whose gender identity corresponds to the sex they were assigned at birth or is sometimes used to describe someone who is not transgender (Aultman, 2014). This definition of transgender encompasses a wide range of gender minority populations, such as those with nonbinary identities and some people with intersex traits. It is important to note, however, that not all individuals who are classified as transgender under this definition identify themselves as transgender.⁴ For this reason, within this report we make a distinction between transgender experience⁵—when someone currently identifies with a gender identity that is different from their sex assigned at birth—and transgender identity—when someone currently identifies oneself as transgender.

• Sexual orientation is a multidimensional construct encompassing emotional, romantic, and sexual attraction, identity, and behavior.

In Western cultures, sexual orientation is often defined on the basis of the gender(s) of a person’s desired potential or actual sexual or romantic partners relative to the person’s own gender. These dimensions of sexuality—attraction, identity, and behavior—may not correspond to the same sexual orientation. For example, a man may be attracted to other men but exclusively engage in sexual behaviors with women or a woman may be attracted to men but also identify as lesbian.

The most common sexual orientation identity terms are based on a gender binary and include heterosexual, or straight; homosexual, gay, or lesbian; and bisexual. However, there are other less prevalent terms that may eschew the gender binary, such as queer (an umbrella term for belonging to the LGBTQI+ community that can also refer to a nonbinary gender identity, which is also referred to as genderqueer) or pansexual (oriented toward partners of any gender); denote uncertainty (e.g., questioning); or be used only among specific populations (e.g., same gender loving, used in Black communities,⁶ or Two-Spirit, used in Indigenous populations). It is important to note that these terms are only a subset of the ever-evolving sexual orientation terminology that is currently in use and that some of the more common terms (gay, bisexual) are also used to reflect gender diversity in attraction (Callis, 2014; Paz Galupo, Ramirez, and Pulice-Farrow, 2017).

To this point, we have used two different umbrella terms to refer to populations defined by their sexual or gender minority status: “LGBTQI+” and “sexual and gender minority.” Each term has both benefits and drawbacks. LGBTQI+ is a term that is well recognized and the acronym can be shortened to clearly specify which populations are under discussion; however, it is not easily recognizable if the leading letters in the acronym, LG, are excluded from the list, which makes it of limited utility when the focus is gender rather than sexual minorities. Of particular concern, this term conflates gender and sexual orientation, because it does not clearly distinguish between minority populations defined by their sexuality and those defined by their gender. The term “sexual and gender minority” does make this distinction, but it is less well known and less explicit regarding the specific populations that are being discussed. Because the benefits of using one term or the other depend on the context of their use, we use both of these terms throughout the report.

Gender Identity and Sexual Orientation in an Indigenous Context

There are 574 federally recognized American Indian and Alaska Native (AIAN) tribes that are sovereign governments and, as such, have a nation-to-nation relationship with the United States. Two-Spirit is an intertribal term first coined in 1990 that serves as a placeholder for tribally specific gender and sexual orientation identities that are centered in tribal worldviews, practices, and knowledges (see Wilson, 1996). Tribes have their own specific terms for gender statuses (e.g., in Navajo, Nádleehí refers to one who is transformed), and many of these terms go beyond the binary categories of male or female. For these tribal identities that cannot be directly translated or mapped to the Western conception of gender, “Two-Spirit” can be used as the English-language placeholder term. For many tribal nations, gender is not limited to a Western binary construct or expression; in fact, some tribes have as many as five genders. Thus, Two-Spirit often represents a third or fourth gender status that is nonbinary but is linguistically or socially contextualized within a particular tribal nation worldview and cultural understanding, often with certain social, cultural, or ceremonial roles attached to the status (Jacobs, Thomas, and Lang, 1997). Under this holistic view of personhood, Two-Spirit is a placeholder term that captures not only gender identity or sexual orientation identity, but also a social and cultural position that shapes and defines all aspects of one’s life. Because Two-Spirit is a term by and for Indigenous peoples and is culturally anchored with particular meaning and, potentially, social status, it is not appropriate for use by non-Indigenous populations.

Although it is widely recognized and used across many Indigenous communities (Robinson, 2020; Cassels et al., 2010; Walters et al., 2006), the term Two-Spirit is not universally recognized or accepted. Moreover, the term encompasses a large number of heterogeneous identities that may otherwise share little in common. It has been criticized for erasing the specific cultural histories and practices of individual tribes and for evoking (and, in some cases, romanticizing) an ahistorical conception of gender and sexuality that may be more grounded in colonial depictions of Indigenous peoples as deviant than in Indigenous cultural practices (see, e.g., Pember, 2016). Despite these criticisms, the term Two-Spirit is a way to reference Indigenous identities, practices, and traditions in the context of Western data collection practices and ensure that Indigenous sexual and gender minorities are represented and counted (Davis, 2019).

The U.S. government has a trust responsibility to the tribes, derived partly from treaties, executive orders, judicial actions, or legislation, to ensure protection of Indian trust lands and tribal sovereignty, as well as provision of social, medical, and educational services for tribal members. The National Congress of American Indians (NCAI) notes that the collection of accurate, adequate, meaningful data is critical to the health and welfare of tribal nations (Sahota, 2007). In recent years, tribes have exercised sovereign authority over data collection efforts and research in tribal lands and of tribal members.

Currently, the decennial U.S. census and the American Community Survey (ACS) are the two major sources of data on AIAN people and form the main basis of funding for many tribal programs and policies (Connolly and Jacobs, 2020). Tribes have noted that AIAN and other Indigenous populations tend to be undercounted in the ACS as well as in other national surveys due to inadequate data collection and reporting or being collapsed into “other” categories due to small sample sizes. NCAI⁷ reports:

American Indians and Alaska Natives may be described as the “Asterisk Nation” because an asterisk, instead of data point, is often used in data displays when reporting racial and ethnic data due to various data collection and reporting issues, such as small sample size, large margins of error, or other issues related to the validity and statistical significance of data on American Indians and Alaska Natives.

Recognizing that a parallel problem of invisibility often arises in LGBTQI+ data collection methods and research, the panel is highlighting the specific cultural needs of Indigenous populations as part of our evaluation of measures of sex, gender, and sexual orientation.

Importance

LGBTQI+ populations experience differential and inequitable treatment and outcomes in many areas of everyday life, including in health and access to health care services, educational attainment, economic outcomes, and family and social support (NASEM, 2020). The disparities for these populations include higher prevalence of physical and mental health problems, such as HIV and depression; worse self-reported health and health-related quality of life; lower socioeconomic status; and less support from family members and important social institutions, such as schools.

Recent research has sought to describe the origins of these disparities by exploring the multilevel and intersecting factors that influence the well-being of LGBTQI+ populations. These factors include minority stress exposures, including stigma, violence, and discrimination and barriers in access to education, employment, housing, and health care (NASEM, 2020; National Public Radio, Robert Wood Johnson Foundation, and the Harvard T.H. Chan School of Public Health, 2017). At the same time, they are mitigated by factors that help promote and build resilience, such as community and political engagement and strong social relationships (NASEM, 2020). The intensity and effects of these factors can vary across the life course and among different LGBTQI+ populations on the basis of such factors as race, age, and gender. A lack of data on the characteristics, needs, and experiences of LGBTQI+ populations is a major barrier both to better understandings of these disparities and how they are produced and to the development of effective programmatic and policy interventions to address them.

This report lays out a set of recommendations for how best to measure the concepts of sex, gender identity, and sexual orientation in the United States. Everyone has a sexual orientation, a gender identity, and sex traits, and reliable and valid measurement of these constructs are core to understanding population characteristics and outcomes: like race and ethnicity and other demographic characteristics, they are central components of individual identity and experience that shape social relationships and structural opportunities throughout one’s life. Conceptually, sex, gender identity, and sexual orientation are multidimensional, incorporating both social and individual identity components. Social components reflect the interpersonal and societal interaction and recognition aspects of these concepts, while identity refers to an individual’s internal sense of self—whom one understands oneself to be as a person. With respect to gender and sexual orientation, identity is the dimension that is most consistently tied to experiences with material forms of discrimination and health disparities (NASEM, 2020) and is noted explicitly in the application of sex discrimination laws and policies to sexual and gender minority populations.⁸

Measures of identity are the most relevant when the goal of measurement is to enable the identification of sexual and gender minority populations that are most consistently at risk for differential treatment and outcomes (NASEM, 2020). Thus, although the report discusses multiple dimensions of sexual orientation, gender, and sex, its recommendations are primarily focused on identity, including measures that make it possible to ascertain the populations who identify as lesbian, gay, bisexual, or another sexual minority term. This report also discusses recommended measures to identify people with intersex traits—a concept we address in this context because this population motivates the need to rethink current measures of sex and because the discrimination and marginalization that people with intersex traits experience often mirror experiences of discrimination on the basis of sexual orientation or gender identity (NASEM, 2020). This is a complicated task because sexual and gender minority populations (and cultures) are dynamic and extremely diverse, and the recognition and study of them is relatively new.

Developing Consistent Data Collection Practices

The 2020 National Academies report recommended that measures of sexual orientation, gender identity, and intersex status be routinely collected within at least three types of data collection activities:

1.

survey research;

2.

nonsurvey research, including clinical trials, electronic health records, biomedical research, public health surveillance, program evaluations, and assessments of discrimination; and

3.

administrative and program data systems, including intake forms, applications, and civil rights and criminal justice enforcement data.

The report called on the federal government to develop standards to guide the collection of these data throughout the activities of the federal agencies that work and collect data within these domains. It also called on private entities, such as hospitals, to collect these data in a consistent and structured manner that would allow for comparisons across data sources.

This growing recognition of the need to identify LGBTQI+ populations is driving changes to the ways in which demographic data are collected in population surveys and research, health and clinical contexts, and administrative records, both within the United States (Baker, Streed, and Durso, 2021; Keuroghlian, 2021; Streed et al., 2020) and internationally (see, e.g., Statistics Canada, 2021, 2020; Stats NZ, 2021, 2019; United Nations Economic Commission for Europe, 2019). Though the United States has not yet established federal standards for data collection specifically on LGBTQI+ populations, a growing number of federal surveys have introduced measures of sexual orientation and gender identity into their routine data collection efforts (NASEM, 2020; Patterson, Jabson, and Bowen, 2017). At the national level, the Federal Committee on Statistical Methodology (FCSM), an interagency committee dedicated to improving the quality of federal statistics, has released several reports assessing existing sexual orientation and gender identity data collection practices and discussing key implementation issues for the general population (FCSM, 2020; Federal Interagency Working Group on Improving Measurement of Sexual Orientation and Gender Identity in Federal Surveys, 2016a, 2016b, 2016c).

The NIH Sexual and Gender Minority Research Office maintains a comprehensive website⁹ devoted to methods and measurement in sexual and gender minority health research and has helped ensure that the NIH PhenX Toolkit¹⁰ for biomedical research includes standardized measures of sexual orientation and gender identity. Federal regulations already require most electronic health record systems to have the capacity to collect, store, and retrieve structured data on sexual orientation and gender identity,¹¹ and in July 2021 the U.S. Department of Health and Human Services included sexual orientation and gender identity in the U.S. Core Data for Interoperability standards. A number of states have also moved to require the collection of sexual orientation and gender identity data across the activities of their departments of health, aging, and other administrative agencies (State Health Access Data Assistance Center, 2021; San Francisco Human Services Agency, 2020).

Several other countries have similarly introduced or begun to develop recommendations and guidelines for standardized approaches to identifying LGBTQI+ populations. For example, Australia, Canada, and the countries of the United Kingdom have introduced new sex and gender measures on their national census in order to count transgender people. Australia and New Zealand also recently revised their national standards for data collection on sex, gender, and sexual orientation (Australian Bureau of Statistics, 2021; Stats NZ, 2021, 2019), while Canada¹² and the United Kingdom¹³ are currently developing new or revised standards.

While gender identity and sexual orientation are increasingly becoming more common to collect, intersex status is still not routinely assessed in population surveys, research, health care, or administrative settings. In addition, intersex status is often erroneously conflated with or subsumed into discussions of sexual minority and transgender populations. Much of the available research on intersex populations focuses on clinical aspects of specific intersex variations, which hinders the ability of researchers, clinicians, and policy makers to understand the well-being of the intersex population more broadly, in general health care settings and in nonmedical contexts, such as education, housing, or employment (NASEM, 2020). The invisibility of intersex populations reflects, in part, a historical trend that is “largely one of erasure” of sex diversity by medicine and societies (Reis, 2021). This erasure has led to a large deficit in knowledge about intersex people and populations relative to other sexual and gender minoritized groups.

Another major issue in the collection of data on sexual orientation, gender identity, and intersex status is inconsistency in measurement. Without national standards on how to collect, analyze, and report these data, there is increasing heterogeneity in measures deployed across U.S. surveys and in other data collection activities. Some of this heterogeneity may appropriately reflect important cultural attention to specific populations of interest, such as surveys among AIAN communities that specifically include references to Indigenous identities such as Two-Spirit or specific tribal identities (see, e.g., HONOR Project: see Cassels et al., 2010; Walters et al., 2006). However, other forms of heterogeneity reflect a lack of consensus on how to define and measure the constructs of interest. A lack of consistency in data collection measures introduces concerns about data quality; complicates data analysis and reporting; and hinders efforts to advance research and develop effective programs and policies to address the disparate treatment and outcomes in LGBTQI+ populations. It also affects non-LGBTQI+ populations by obscuring essential variation in important characteristics, such as sex and gender, that are routinely used in research, policy, and law.

There are concerns that collection of sexual orientation, gender identity, and intersex status data will expose LGBTQI+ people to harm, given a long history that includes interpersonal and structural violence targeting of LGBTQI+ people (NASEM, 2020; National Public Radio, Robert Wood Johnson Foundation, and Harvard T.H. Chan School of Public Health, 2017), mistreatment by clinicians and other service providers (Peek et al., 2016), and exposure to harmful practices, such as conversion and aversion therapies and other medically unnecessary procedures, including lobotomies and chemical castration of LGBTQI+ adolescents and adults, and genital surgeries on the bodies of intersex children too young to consent. Not collecting these data, however, makes discrimination and mistreatment harder to address: without tracking experiences by sexual orientation, gender identity, and intersex status, it is impossible to identify and rectify patterns of poorer access, treatment, and outcomes for LGBTQI+ people. Routine and standardized measurement over time is also essential for building datasets large enough to permit robust analyses of the needs of groups that face mistreatment and disparities in outcomes at the intersections of multiple axes of identity, such as LGBTQI+ people with disabilities and LGBTQI+ Black, Indigenous, and other people of color (Bi, Cook, and Chin, 2021; Crenshaw, 2017, 1989; Tomlinson and Baruch, 2013).

Based on the utility of sex, gender identity, and sexual orientation data in enumeration of populations and identifying differences between population groups, we conclude that these constructs are important demographic variables that are essential for more fully understanding the broad diversity of people and populations in the United States.

Interpretations and Limitations

The panel was tasked with developing recommended measures of sex, gender identity, and sexual orientation within three broad contexts: surveys and research settings, clinical and medical settings, and administrative settings. Each of these are settings in which respondents are drawn from the general population and none specifically focus on sexual and gender minority populations. Consequently, the recommended measures had to be broad and easily understood by the general population, including those who are not members of sexual and gender minority communities. Moreover, the short time frame allotted to this study, along with the specific needs of the study sponsor, forced the panel to constrain our evaluation in several important ways.

First, the panel limited our evaluation to measures that would capture the relevant dimensions of sex, gender identity, and sexual orientation within the general adult population of the United States. Although the panel believes that understanding the process and experience of sexual and gender identity formation and development in childhood is crucial for understanding how the unequal treatment of sexual and gender minorities cumulates throughout the life course, identifying age-appropriate terminology and concepts for each stage of the developmental process was not possible in the time available to the panel. When possible, we indicate the age ranges in which the recommended measures have been tested and when these measures can be used with young populations.

Similarly, the panel would have liked to have been able to offer recommendations that focused on data collection efforts in specific sexual and gender minority populations that are often ignored and do not see themselves reflected in most data collection efforts. We also would have liked to offer recommendations for populations with low English proficiency who might benefit from translations of these questions into their native language. Each of these populations is important in their own right and deserves to be represented in data and policy discussions; however, the timeline of this study precluded this work. We hope that there will be future efforts to address these gaps and ensure that these populations can be included and accurately captured in data collection efforts. Doing so will enrich understanding not only of these specific populations, but also of the diverse array of experiences of sexual and gender minorities in the United States.

The second limitation to the scope of the panel’s efforts was the decision to focus on the identity dimension of sexual orientation. As noted above, the concepts of sex, gender, and sexual orientation are multidimensional, with dimensions that separately capture their social, behavioral, and identity aspects. With respect to gender, the study title, statement of task, and project description for this study specifically focused on measures of gender identity rather than other dimensions of gender; however, this is not the case for sexual orientation.

In comparison with the measurement of sexual orientation identity, which is (relatively) straightforward, the measurement of sexual orientation attraction and behavior is more complex, because both attraction and behavior are multidimensional concepts. Sexual orientation attraction can be conceptualized as encompassing not only the direction or orientation of attraction (the gender[s] to which an individual feels attraction), but also the strength of that attraction, including whether an individual feels attraction at all. Moreover, even for one person, the orientation and strength of sexual orientation attraction may differ depending on whether it is reported based on emotional, romantic, or sexual attraction.

Sexual behaviors are similarly complex, and their relevance for data collection is often context-specific. Depending on the purpose for including such measures, it may be relevant to identify specific sexual activity or activities, the gender(s) or sex traits of a sexual partner or partners, or the frequency with which an individual engages in specific activities. The panel was concerned both that the measurements of sexual orientation attraction and behavior were too complex to be addressed within the study’s time frame and that there would be insufficient guidance in the extant research literature to allow us to make specific recommendations for these dimensions.¹⁴

The identity dimension is most relevant for measuring disparities in treatment and outcomes, and greater effort has been spent on developing and deploying measures of identity than of other dimensions (Patterson, Jabson, and Bowen, 2017). Identity measures are also better suited for identifying members of sexual minority populations. For these reasons, the panel understood its task to include developing recommendations of specific measures that capture the identity dimensions of sexual orientation. Although the panel focused on measures of identity, we acknowledge the importance of other social and behavior dimensions of these concepts and hope that development of measures that capture the complexity of these dimensions continues.

Measures and Information that Informed the Panel’s Work

The panel began its task by evaluating the set of measures of sexual orientation, gender identity, and intersex status that have been included in federally sponsored surveys (see Appendix A). To compose this list, we first listed the measures that were included in the 2020 report on LGBTQI+ well-being (NASEM, 2020), and then supplemented it with information from newer and other non-federally funded surveys that were conducted of the general adult population in the United States. The panel further supplemented this list by considering guidelines on collecting sexual orientation, sex, and gender identity data that were recently issued by several English-speaking countries. Although the social and political contexts within these countries differ, these guidelines and the procedures through which they were developed were informative for the panel’s deliberations.

The panel also held two public information-gathering sessions that provided additional data and contextual information that informed our discussions. The first session convened a panel of federal employees who had used sexual orientation and gender identity data or had implemented such data collection efforts in different data collection contexts, including survey research, electronic health records, human resources management and measurement of discrimination, and assessing workplace effects. The second session was a 1-day workshop that included panels of experts who discussed the measurement of sex and gender for intersex/DSD populations; legal and administrative issues surrounding the collection of sex, gender identity, and sexual orientation; and the measurement of sex, gender identity, and sexual orientation with clinical and medical contexts. Together, these information-gathering sessions provided a rich knowledge base that informed the panel’s discussions and decision making. The agendas for these sessions can be found in Appendix B.