Not to be flip here, but the stigma surrounding family and gender identities has a very long and unfortunate history, including in their statistical measurement, so it's important to keep focusing on the proper measurement instruments.
Back in the day, households were "headed" by men, so if a woman listed herself as "household head" in a survey, and there was a "male" household member, the coders changed the answer for tabulation. Back in the day, there were no "unmarried partners." Back in the day, there were no "same sex partners.
Except there were. See, for example, Dan Bouk's Democracy's Data chapter on "Partners."
Or as things began to change, the creation of the POSSLQ, https://en.wikipedia.org/wiki/POSSLQ