LESSONS & LIMITATIONS

Researchers and demographers have surfaced a number of questions and concerns regarding the demographic literature on LGBTQ communities, including LGBTQ communities of color. Some of these critiques include:

The Limitations of U.S. Census Data: As acknowledged by the Williams Institute, using Census data on "same-sex unmarried partners" to draw inferences about the broader lesbian, gay and bisexual population poses a few problems. First, the characteristics of same-sex couples might be different than those of single LGB people (as some studies have found). Second, confidentiality concerns among LGB people means that the data only includes couples who are willing to acknowledge their relationship and be "out" about their sexualities (at least in a survey). Third, the U.S. Census does not pose questions to identify transgender people or people who identify as "queer."

• Read more about this methodology for estimating LGB populations (PDF)
• Learn how Census 2010 will count (and not count) LGBTQ people (PDF)

Question Format: To identify the sexualities of respondents, surveys have typically relied on questions about sexual orientation, sexual behavior or sexual attraction—recognizing that most individuals fall along a sexuality spectrum of how they self-identify, behave and/or feel. Here, the phrasing and formatting of these questions can yield different response rates and findings. In regards to gender identity, many surveys simply do not ask about gender identity and expression in a way that identifies transgender people. Similarly, the phrasing and format of these questions can produce different results. What then—we ask—are the best practices for surveying the full range of LGBTQ communities, including LGBTQ communities of color?

Insufficient Data Sets: As many demographers point out, no nation-wide or state-wide representative data exists for transgender people; this means that accurate, national figures on transgender people are unavailable. Further, many publically available, national data sets produced by government agencies and research organizations—such as the Youth Risk Behavior Surveillance System at the Centers for Disease Control, which monitors health-risk behaviors among youth and young adults—do not include variables for sexuality and gender identity. When data sets leave out LGBTQ people, it means that a researcher's ability to examine LGBTQ individuals in a range of empirical research studies (to study how LGBTQ people are affected across various issues and settings—or in relationship to "race/ethnicity," which allows one to study social phenomena on LGBQT people of color) is largely squandered.

Methodological Barriers: The need for large sample sizes—and the lack of corresponding institutional support—has led to a void in demographic reviews about the full range of LGBTQ communities of color, notably Native American/Two Spirit people and LGBTQ people of color sub-groups such as LGBTQ Southeast Asians, LGBTQ Arab and Iranians, etc. Recognizing the smaller numbers, how can funders support research that studies these communities?

Reductionist Variables: "LGBT" lumps the experiences of lesbians, gay men, bisexuals and transgender people. Further, the racial/ethnic categories popularized by the U.S. Census—American Indian or Alaska Native; Asian; Black or African American; Native Hawaiian or Other Pacific Islander; Hispanic; and White—homogenize the distinct cultural, historical, political and economic histories of diverse people across our country and the globe. Worse, many researchers adapt these categories with little regard for cultural complexity; thus, many demographic reviews remain shallow, rarely addressing intra-group distinctions.

Categorizing & Reporting: When research instruments allow individuals to more accurately describe their identity by including multiple response options (such as a gender identity category that includes "transgender," "genderqueer," and other options, or a Latino/a category that includes "Mexican American," "Puerto Rican" and other options), researchers are stuck with determining how best to convert these responses into manageable and meaningful categories of analysis. Does one include "Mexican American" responses under "Latino/a," which discounts their country of origin for analytical purposes? Does one report all the responses, regardless of the enormity of this task or the density of the final report? Categorizing and reporting identity-based responses—in a way that's comprehensive for the respondent, useful for the researcher and digestible for the reader—creates challenges for relating the full breadth of LGBTQ experiences and identities.

More Research, Better Data Collection: Among researchers of LGBTQ issues, a near consensus exists that the professional literature on LGBTQ communities, especially in relation to LGBTQ people of color and transgender research, has much room to grow. They suggest that growing this body means raising interest among government, philanthropic and academic entities to study the various populations and topics that comprise our diverse communities. And it means addressing the structural barriers that prevent a more comprehensive, accurate library of LGBTQ issues—from limited instrument design, to problematic question formats and taxonomies, to the personal biases about race, sexuality and gender identity that exist among researchers.

Scattered, Decentralized Information: From nonprofit research reports to published journal articles, from university-based studies to formal program evaluations, the large body of theoretical and empirical literature that has been produced on LGBTQ people remains decentralized and, thus, largely out of dialogue with one another. Without the ability to easily compare formal studies, distill useful lessons or findings, or identify what's missing, grantmakers (and others) are left wondering if they've assessed the full scan of existing materials, which makes it difficult to base a policy or program design on an evidence base.

Have a suggestion? Email us at