The research team, led by Tetsuji Minami of the National Cancer Center Japan alongside collaborators from several Japanese institutions, broke the 8.6 percent figure into distinct categories: 4.9 percent identified as gay or lesbian, 1.1 percent as bisexual, 1.0 percent as asexual or unsure of their attraction, and 1.6 percent as gender-diverse, meaning their current gender identity did not match the sex assigned to them at birth or fell outside the binary altogether. These estimates align with the 7.0 to 16.8 percent range reported in earlier Japanese studies conducted by private companies, government agencies, and sociologists, but earlier efforts suffered from convenience sampling, inadequate statistical adjustment, and narrow cohorts that limited their generalizability. By contrast, the new study calibrated its web-based sample against the Japanese government’s 2019 Comprehensive Survey of Living Conditions using a statistically demanding technique called inverse probability weighting.

The technical machinery behind the estimates deserves attention because it addresses a well-known weakness of internet surveys. Participants were drawn from a panel of 2.3 million people managed by Rakuten Insight, with stratified multistage sampling by age and sex, and a response rate of 65.7 percent among those invited. The researchers then applied two pre-set algorithms to screen out inattentive respondents, or “satisficers,” who straight-line their answers or fail directed attention checks, removing 9 percent of the raw sample. Sampling weights were predicted from a logistic regression model comparing the web survey respondents with the census-calibrated national survey, adjusting for area of residence, marital status, education, housing tenure, self-rated health, and smoking status. The resulting inverse probability weights allowed the team to produce population-level estimates with 95 percent confidence intervals, computed with a robust variance estimator, treating the findings explicitly as weighted distributions rather than causal effects, in line with modern descriptive epidemiology frameworks.

Measurement of sexual orientation and gender identity followed the internationally recommended “two-step approach,” which asks separately about sex assigned at birth and current gender identity, supplemented by a romantic and sexual attraction item to classify sexual orientation among cisgender respondents. Gender-diverse respondents were deliberately not reclassified by sexual orientation, and the authors caution that the gender-diverse category aggregates heterogeneous subgroups, including transgender, non-binary, and questioning individuals, because the survey instrument could not capture finer self-descriptions. Similarly, respondents who selected neither or unsure on the attraction item were grouped under an inclusive “asexual cisgender or unsure” label, an operationalization the authors urge readers to interpret cautiously since it may blend true asexuality with ongoing questioning. Inattentive respondents who reported anything other than male or female for assigned sex at birth were excluded as quality controls, consistent with Japan’s family register system, and were not intended to exclude minority identities.

Beyond headline prevalence, the study mapped striking demographic and socioeconomic patterns. All sexual and gender minority groups were younger on average than heterosexual respondents, with mean ages of 44.7 years for gay and lesbian participants, 38.3 for bisexual, 40.0 for asexual, and 38.4 for gender-diverse respondents, compared with 48.8 for heterosexual participants, and prevalence was concentrated under age 30. Gay respondents were more often male, while bisexual and asexual respondents were more often female, and gender-diverse respondents were evenly split by assigned sex. Marriage and partnership rates told a more complex story: only 47.7 percent of gay and lesbian respondents and 38.1 percent of gender-diverse respondents were married or partnered, versus 65.1 percent of heterosexual respondents, and a substantial share of gay and lesbian participants reported opposite-sex marriages or partnerships, a pattern the authors link to Confucian family norms emphasizing marriage and procreation, the absence of legal same-sex marriage recognition, and structural stigma.

Socioeconomic disparities emerged with unusual clarity because the national calibration allowed income, education, employment, insurance, and housing comparisons. Lower household equivalent income was more common among minority groups, particularly male sexual minorities and female gender-diverse respondents, and women had lower incomes than men within every category. Educational attainment diverged sharply from Western patterns: whereas studies in the United States and Europe generally find gay and bisexual men more educated than heterosexual men, Japanese gay, lesbian, and bisexual male respondents showed the opposite trend, with heterosexual men most likely to hold degrees. Employment differences also inverted expectations in places, with female sexual minorities more likely to be employed than heterosexual women, plausibly because married Japanese women are often outside the labor force while minority women marry less often. Health insurance coverage was lower among minority respondents of both sexes, and housing tenure varied significantly only among men, with asexual men and lesbian women least likely to own homes.

Health indicators painted a mixed picture that partially contradicts Western literature. Psychological distress, measured with the Kessler-6 scale, was consistently more common among minority respondents of both assigned sexes, echoing minority stress theory, which holds that stigma and marginalization drive mental health disparities. Body mass index patterns were also distinctive: underweight was more common among female minority respondents, while overweight and obesity clustered among male bisexual and asexual respondents and female asexual respondents. Perhaps most surprising, current smoking and habitual drinking showed limited or reversed differences, with minority groups tending to drink less than heterosexual respondents, contrary to consistent findings from England, Canada, and the United States. Substance use other than tobacco and alcohol was higher among minority respondents but remained rare overall. Medical comorbidity did not differ significantly once results were stratified by assigned sex at birth, and self-rated health differences disappeared under stratification, prompting the authors to call for future studies with covariate adjustment and clinical data linkage.

The significance of these findings extends well beyond Japanese borders. Until now, virtually all nationally representative data on sexual and gender minority populations came from Western countries, including the United States, the Netherlands, Sweden, Portugal, Canada, and the United Kingdom, creating a geographic evidence gap the authors argue reflects both research neglect and, in parts of Asia, legal environments where same-sex acts remain criminalized. Japan presents a distinctive hybrid context: same-sex relationships were historically tolerated within certain cultural traditions, cross-national analyses suggest relatively low structural stigma and low non-disclosure, yet Japan remains the only G7 nation without comprehensive anti-discrimination legislation or marriage equality. The study’s documentation of opposite-sex marriages among gay and lesbian respondents, observed at levels between the low Western figures of 3 to 9 percent and the much higher Chinese figures of 33 to 51 percent, offers rare quantitative grounding for understanding how family norms and legal structures shape minority lives in East Asian settings.

The authors draw direct policy implications from their data. They recommend that national administrative surveys routinely incorporate sexual orientation and gender identity items and household modules that recognize diverse partnership structures, enabling ongoing surveillance of socioeconomic disparities and targeted resource allocation. They call for accessible mental health support and anti-discrimination efforts in schools and workplaces to reduce the psychological burden the Kessler-6 results document, and for longitudinal and panel designs that oversample small subgroups and analyze intersections of age and assigned sex. Acknowledging limitations, including self-reported data, potential social desirability bias, residual web-survey selection bias that weighting could reduce but not eliminate, and the coarse aggregation of gender-diverse subgroups, the team nevertheless positions the work as a foundational step. With an estimated 8.6 percent of Japan’s population identifying as a sexual or gender minority, the study transforms a population previously invisible in national statistics into a measurable, monitorable constituency whose health and socioeconomic well-being can now be tracked, compared, and protected.

Subject of Research: National estimation of the proportion and characteristics of sexual and gender minorities in Japan using a nationwide cross-sectional internet survey

Article Title: Estimation of Sexual and Gender Diversity Among People Aged 15 Years and Older in Japan: A Descriptive Epidemiological Study Using a Nationwide Cross-Sectional Internet Survey

Article References: Minami, T., Inoue, M., Matsushima, M., Yoshioka, T., & Tabuchi, T. (2026). Estimation of Sexual and Gender Diversity Among People Aged 15 Years and Older in Japan: A Descriptive Epidemiological Study Using a Nationwide Cross-Sectional Internet Survey. Archives of Sexual Behavior. https://doi.org/10.1007/s10508-026-03515-0

Image Credits: AI Generated

DOI: 10.1007/s10508-026-03515-0

Keywords: sexual orientation, gender identity, sexual and gender minorities, Japan, descriptive epidemiology, national survey, LGBTQ, health disparities, minority stress, inverse probability weighting, bisexual, transgender