Balancing Community and Research Needs in Gender Measurement Decisions

There has been increasing attention to—and debate about—best practices related to gender measurement. We add to this conversation by testing whether an approach of providing participants with an extensive list of gender options and then having them self-select into a more limited group of choices (that cohere with the research questions) could be useful.

In this study of adults (N = 1813), in 2021–2022, we measured gender using a three-part approach: the Gender EXPAND (EXPANsive responsive genDer) approach. Participants were first asked if they identified as transgender. They were then asked their current gender (check all that apply: woman, trans woman, trans feminine, man, trans man, trans masculine, nonbinary or genderqueer, agender, and an option to write in a response). Participants then selected a gender that best fit for them from a limited set of options (transgender, cisgender, nonbinary, unsure).

We evaluated researcher reclassification from the expanded list of gender options compared to participant self-selection from the limited gender categories. We miscategorized 10.5% of participants when reclassifying their gender from the extensive list of options compared to their self-identification as transgender. We miscategorized 11.2% of participants compared to participants’ self-selection as cisgender, transgender, nonbinary, or unsure.

Participants generally responded well to the Gender EXPAND approach and our transparency in our explanations for our questions. This approach should be refined to reduce misclassification, potentially through different reclassification processes or modified response categories.

Measurement of gender has downstream implications for representation in data, which informs policy. Inadequate measurement can lead to inaccurate data and undercounting of gender diverse individuals and further marginalizes transgender and nonbinary people at a time when stigma and anti-transgender legislation is at an all-time high.