Australia's census now asks sexual orientation. Should the US census do the same?

When Australians fill out their census on August 11, they'll be asked a question no Australian census has asked before: how they describe their sexual orientation. Everyone 16 and older gets it, the options run straight, gay or lesbian, bisexual, or another term, and answering is optional.

The same form now asks about gender separately, with man, woman, non-binary, or another term, while the older sex question has been rewritten to ask specifically about sex recorded at birth. The Australian Bureau of Statistics developed the wording with an LGBTIQ+ expert advisory committee after public consultation and testing. It nearly didn't happen as the Australian government moved to drop the topic in 2024 and reversed course after the backlash.

The U.S. Census Bureau asks none of this.

What the US actually collects

Neither the decennial census nor the American Community Survey asks anyone's sexual orientation. What they do capture is same-sex couples, through the household relationship question, which distinguishes same-sex from opposite-sex spouses and unmarried partners. So the Bureau can tell you how many same-sex couple households are in a county. It can't tell you how many gay people live there, because anyone single, or partnered but living apart, never shows up.

The Bureau came close to changing that though. It won approval to test sexual orientation and gender identity questions on the ACS starting in August 2024, sending them to roughly 480,000 households, with results expected in early 2026. That work stopped. Executive Order 14168, signed January 20, 2025, directed federal agencies to collect data on sex only as male or female, and the Bureau halted its research into adding gender identity measures to the ACS. Gender identity questions were also dropped from the Household Pulse Survey when it relaunched that April.

According to the Census Project, more than 80 percent of the 360 affected federal data collections were changed without a Federal Register notice or public comment period. That's the same pattern as the June order banning noise infusion and the reversal on race and ethnicity standards, both of which arrived the same way.

The case researchers make for asking

Karinna Saxby, a senior research fellow at the Melbourne Institute of Applied Economic and Social Research, published peer-reviewed work that is probably the strongest argument in the Australian debate. She followed people recorded in the 2011 Australian census for 12 years and compared survival rates.

Saxby found about one in five premature deaths among people in same-sex couples were excess deaths, meaning they wouldn't have occurred at the rates seen in different-sex couples. External causes including suicide accounted for nearly 12 percent of deaths among women in same-sex relationships, against about 2 percent among women in different-sex relationships.

Her argument is that you can't direct health spending or suicide prevention at a population you haven't measured, and that a census reaches everyone rather than the self-selected samples most LGBTQ research relies on.

Here's the part worth noticing for an American reader. Saxby did that study using census data on same-sex relationships, which is exactly what the US already collects. The research she used to justify adding the new question was possible without it. What the new question adds is everyone the couple-based measure misses, which is most single people and anyone not living with a partner.

The arguments against it

Critics of adding the question raise a few things. One is response burden and privacy: a census is mandatory in both countries, and asking a sensitive question of everyone is different from asking it of a survey sample who can more easily decline. Australia handled that by making the question optional and limiting it to people 16 and older, which creates its own problem, since optional questions produce incomplete data and the people who skip them are rarely random.

Second is measurement. Sexual orientation is self-described, terminology shifts between generations, and a fixed set of boxes will fit some people badly. The ABS spent years on cognitive testing for that reason.

Third is that the household form is already crowded, and every question added has to displace something or lengthen a form that agencies are trying to keep short to protect response rates.

Whether the US follows isn't really a technical question at this point. The Bureau had the approval, the funding and 480,000 test households, and the testing stopped for policy reasons rather than methodological ones. Australia will publish its first results from the August count, and those results will be the first national, population-level numbers of their kind from a census anywhere with this design.

Sources

The Australian study and the case for the question come from Karinna Saxby writing in The Conversation. Question wording, eligibility and consultation detail are from the Australian Bureau of Statistics. US testing history and the effect of Executive Order 14168 are from The Census Project. Same-sex couple measurement is documented in the American Community Survey relationship question.