Refining United States Sexual Minority Adult Population Estimates with the Inclusion of "Something Else" and "Don't Know" Survey Responses

Abstract

Purpose: To refine estimates of the U.S. sexual minority population, we sought to characterize trends in the odds of respondents selecting "something else" or "don't know" when asked about sexual orientation on the National Health Interview Survey and to reclassify those respondents likely to be sexual minority adults. Methods: Logistic regression was conducted to test whether the odds of selecting "something else" or "don't know" increased over time. A previously established analytic approach was used to identify sexual minority adults among these respondents. Results: Between 2013 and 2018, the percentage of respondents selecting "something else" or "don't know" increased 2.7-fold, from 0.54% to 1.44%. Reclassifying respondents with >50% predicted probabilities of being sexual minorities increased sexual minority population estimates by as much as 20.2%. Conclusion: A growing proportion of adults are selecting "something else" or "don't know." Properly classifying these responses yields more accurate sexual minority population estimates.

Keywords: administrative data or large data sets; public policy and advocacy; sexual orientation; survey design or survey methodology.