26 papers, honestly described
Every key paper, what it found, and its limitations. Filter by question or search by author, journal or topic. Every title links to the original.
Interviews with thousands of Americans showed same-sex attraction was far more common than assumed, and that sexuality varies along a spectrum.
Limits: Samples were not random, so exact percentages are unreliable. Its lasting contribution was showing diversity, not precise rates.
Expert clinicians shown blind personality tests from 30 gay and 30 straight men couldn't tell them apart. Gay men were as psychologically healthy.
Limits: Small sample of non-patient men from one city, but it was the first rigorous challenge to the illness model.
Identical twins of gay men were much more likely to be gay than fraternal twins or adopted brothers, pointing to a genetic influence.
Limits: Volunteers were recruited through ads, which likely inflated concordance. Later population studies found smaller but still real genetic effects.
A small brain region (INAH3) was on average smaller in gay men than in straight men.
Limits: Only 41 post-mortem brains, many affected by AIDS. Important as a first clue; not proof on its own.
Reported a link between a region called Xq28 and male homosexuality in families with several gay brothers.
Limits: Replication was mixed. A larger 2015 study (Sanders et al.) supported it, but the 2019 GWAS did not find it. Treat as unsettled.
Each older brother increased the odds of a man being gay. This fraternal birth order effect is one of the most replicated findings in the field.
Limits: Explains only a portion of gay men. It is a population pattern, not a rule for individuals.
Introduced the minority stress model: higher rates of mental health problems among LGB people are explained by stigma and discrimination, not orientation itself.
Limits: A review and theoretical model; many later studies have since tested and supported it.
Only older brothers born to the same mother mattered, even if raised apart. Stepbrothers or adopted brothers raised together had no effect. Points to the womb, not upbringing.
Limits: Observational data; the biological mechanism was proposed later.
Brain symmetry patterns of gay men resembled straight women's, and lesbians' resembled straight men's, in areas unlikely to be shaped by learning.
Limits: 90 participants. Brain differences can't by themselves say what caused them.
Young people facing high family rejection were about 8.4 times more likely to report suicide attempts than those facing little or none.
Limits: Cross-sectional US data, so it shows strong association rather than proven cause.
A systematic review found no reliable evidence that efforts to change sexual orientation work, and evidence of harm.
Limits: Most change-effort studies it reviewed were themselves low quality, which is part of its conclusion.
About 3,800 twin pairs. Genes explained roughly a third of variation in men and a fifth in women; shared family environment explained almost none.
Limits: Measures behaviour rather than identity, and same-sex behaviour was rare, so estimates have wide uncertainty.
Teenagers raised by lesbian mothers since birth were rated higher in social and academic competence and lower in behaviour problems than a normative sample.
Limits: Volunteer families, mostly well-off and white, so not representative.
LGB youth in more supportive social environments had a significantly lower risk of suicide attempts.
Limits: One US state's data; county-level environment measures.
In 409 pairs of gay brothers, found linkage to regions on chromosome 8 and Xq28.
Limits: Linkage studies are coarse; the 2019 GWAS found different specific loci.
A therapy targeting minority stress (ESTEEM) reduced depression and alcohol use.
Limits: Small pilot trial of 63 men; larger trials have followed.
A major review concluding there is considerably more evidence for non-social causes of sexual orientation than social ones, and no good evidence it can be changed by therapy or upbringing.
Limits: A review, so it depends on the studies it summarises, which it critiques openly.
After US states legalised same-sex marriage, high-school suicide attempts fell by about 7%, with larger drops among sexual minority students.
Limits: Strong design for observational data, though it can't rule out every other change.
Music therapy added to usual care probably reduces depressive symptoms and anxiety in the short term.
Limits: Low to moderate certainty; studies varied in quality.
Mothers of gay sons, especially those with older brothers, had higher antibody levels to a male-specific brain protein, a possible mechanism for the older-brother effect.
Limits: Small sample; needs replication.
477,522 people. No single gay gene; many variants with small effects. Common variants together explained 8 to 25% of variation.
Limits: Measured ever having a same-sex partner, mostly in people of European ancestry. Authors warned against using it to predict anyone's orientation.
Proposes same-sex behaviour may have been part of animal life from early in evolution, rather than an anomaly needing special explanation.
Limits: A theoretical paper that calls for more data.
Young adults whose parents tried to change their orientation had higher depression and suicide attempt rates and lower education and income.
Limits: Retrospective self-reports from one US sample.
Trans adults exposed to conversion efforts had higher odds of lifetime suicide attempts, especially if exposed before age 10.
Limits: Cross-sectional survey data.
Sexual minority adults who went through change efforts had nearly twice the odds of suicidal thoughts and higher odds of attempts, even after accounting for childhood adversity.
Limits: Observational; US national sample.
Analysis of 164 Indian news reports: family non-acceptance and forced separation were the most common narratives; joint couple suicides were frequent.
Limits: Based on news coverage, so it reflects what gets reported, not all cases.
How to read this library. One study never settles a question. Small or early studies (like LeVay 1991 or Hamer 1993) were important but had limits, and some findings didn't fully replicate. The conclusions on this site rest on what many independent studies, across countries and decades, keep finding.