It is not a choice

How to read this page

In nature

The commonest argument against same-sex attraction is that it is “unnatural”. That is a checkable claim rather than a matter of opinion — and it has been checked for decades.

Same-sex sexual behaviour has been recorded in over 1,500 animal species, distributed across most major clades. That is the opening line of a peer-reviewed paper in Nature Ecology & Evolution from 2019. But the same paper says this, honestly: “most of what we know about the incidence of SSB is the result of opportunistic and haphazard observations.” In other words, 1,500 is an accumulated tally of scattered reports, not a census.

Three numbers that are not the same number

Here is the correction we owe the reader. Bruce Bagemihl’s Biological Exuberance (1999) documented more than 450 species — not 1,500. The 1,500 figure comes later and from elsewhere. Online the two are constantly fused and 1,500 gets attributed to Bagemihl. It is not true, and anyone who checks will find that out in two minutes. Below, each figure sits next to what it actually measures.

The systematic counts

When a whole animal group is analysed systematically rather than tallied, the numbers are smaller — and sturdier. Each dot below is roughly one per cent of the species in that group; the exact ratio is written out in words. In both cases a species counts if it has at least one published report, so these are floors, not ceilings.

Three studies answering the three usual objections

Genes

Care is needed in both directions here. “A gay gene has been found” is exactly as untrue as “there is nothing genetic about it”. The honest answer is duller and stronger: partly heritable, many genes with small effects, no single cause.

In August 2019 Science published the largest study on the subject — a genome-wide association study of almost half a million people.

Ganna et al., Science, 30 August 2019

Participants in the main analysis

477,522 people

Where from

UK Biobank (408,995) and 23andMe (68,527)

Replication in a separate sample

15,142 people

Genome-wide significant loci found

5, of which 3 replicated

Share of variation explained by all tested variants

8–25%

Share explained by polygenic scores

under 1%

Heritability: 32.4% from family data, 8–25% from DNA

Within UK Biobank, using 106,979 pairs related at full-cousin or closer, the authors estimated broad-sense heritability at 32.4% — with a confidence interval running from 10.6 to 54.3. That interval is very wide, and that is itself part of the answer: we know the genetic contribution is substantial, but not exactly how large.

All the measured DNA variants together do not capture the whole of family-based heritability (8–25% versus 32%). That is not odd — across 16 other traits the authors checked in the same sample, the ratio is the same. Family-based heritability typically comes out around three times larger than the part captured directly from DNA.

The rest is “environment”. And here comes the word that gets mistranslated in Bulgaria every time.

Twins — and why “environment” does not mean “upbringing”

In 2010 Archives of Sexual Behavior published a Swedish study of 3,826 monozygotic and dizygotic same-sex twin pairs. The sample is population-based — from a 2005–2006 survey of all adult twins in Sweden aged 20–47. That matters: earlier twin studies worked with volunteers, which skews the result.

The modelling splits the variation three ways: genes, shared environment (the home, the parents, everything siblings have in common) and individual-specific environment (everything not shared — including prenatal biology and personal experience).

Read the “shared environment” row for men again. Zero. The home, the parents, the way children were raised — everything brothers share — explained none of the differences. In women it accounts for around 16–17%. That is the direct answer to “it is bad parenting”. And note: “individual-specific environment” does not mean upbringing. It includes prenatal biology and individual life events, not parenting style.

The authors themselves write that the confidence intervals are wide and that the results should be interpreted cautiously. Their formulation: moderate, primarily genetic familial effects, and moderate to large effects of the non-shared environment — social and biological.

How this study gets twisted

Ganna et al. also found genetic correlations with depression (0.44 in women, 0.33 in men) and with schizophrenia. That result is regularly pulled out of context as “proof” that this is a mental disorder.

Here is what the authors themselves write, verbatim: “We emphasize that the causal processes underlying these genetic correlations are unclear and could be generated by environmental factors relating to prejudice against individuals engaging in same-sex sexual behaviour, among other possibilities.” We publish both, so that this page cannot be quoted by halves.

One more thing worth knowing: before publication the authors held open workshops with LGBT activists.

Where the science stands today

In 2025 Felesina and Zietsch published a synthesis in Trends in Genetics (Zietsch was one of the authors of the 2019 study). Their formulation: thanks to twin studies it has been known for decades that human same-sex sexual behaviour has a substantial heritable component; it is influenced by many common genetic variants, each with tiny but cumulative effects; and the evolutionary explanation for why those variants persist remains uncertain. That is the state of the question — without a single slogan.

Birth order

One effect in this field has been confirmed for decades across many independent samples: the more older brothers a man has, the more often he turns out to be homosexual. It is called the fraternal birth order effect.

In 2018 Ray Blanchard pooled 26 studies spanning 25 years: 30 homosexual and 30 heterosexual male groups, 7,140 and 12,837 men respectively.

• What that number actually is: a ratio of two ratios — homosexual men’s older brothers relative to their other siblings, divided by the same ratio in heterosexual men. It does not mean “47% more likely per older brother”. Every simpler retelling of 1.47 is inaccurate.
• The effect is reliable: significantly above 1.00 in 20 of 30 comparisons, above 1.00 non-significantly in 9, and below 1.00 non-significantly in only one. Its magnitude does not depend on family size.
• The effect applies only to males and explains only a minority of cases. We publish no figure for what share — we could not verify one.

The maternal immune hypothesis

The proposed explanation is prenatal. With each male pregnancy, some mothers develop antibodies against a Y-linked protein important in male brain development, and the chance of that happening rises with each subsequent male gestation.

In 2018 PNAS published a direct test. After statistically controlling for the number of pregnancies, mothers of gay sons — particularly those with older brothers — had significantly higher levels of anti-NLGN4Y antibodies than control samples of women, including mothers of heterosexual sons.

Whatever becomes of this mechanism in the end, the direction is clear and it matters for any parent reading this: the proposed cause operates before birth. It has nothing to do with upbringing, with a circle of friends, with “seduction” or with “recruitment”.

This is an association from a single study with modest samples. The authors write that the results “suggest” a link. We say “suggest” too, not “prove”.

When medicine stopped calling it an illness

This is not one dramatic date, nor one vote. It is a process stretched over some fifteen years, followed by a second process at the level of the World Health Organization. Precisely because it was slow and staged, the explanation “they just changed it for political reasons” does not hold.

1. 1973

   The American Psychiatric Association removes the diagnosis from the DSM

   The diagnosis of “homosexuality” is removed from the second edition of the Diagnostic and Statistical Manual. Source: Drescher, “Out of DSM: Depathologizing Homosexuality”, Behavioral Sciences, 2015 — a peer-reviewed history of exactly this decision.

2. after 1973

   A transitional category appears

   Subsequent DSM-II reprints include “Sexual Orientation Disturbance” — for people “disturbed by it, in conflict with it, or wishing to change their homosexual orientation”.

3. 1980

   DSM-III: “Ego-dystonic Homosexuality”

   The transitional category is renamed but remains — still about the distress, not the orientation.

4. 1987

   And it disappears too

   The term no longer appears in DSM-III-R (1987), nor in DSM-IV (1994), nor in DSM-IV-TR (2000).

5. May 1990

   The World Health Assembly adopts ICD-10

   Resolution WHA43.24. The classification came into effect on 1 January 1993. Those are the dates on WHO’s own page — the widely repeated “in force from 1992” is not supported by it.

6. 2014

   A WHO working group recommends deleting the categories entirely

   A review of everything published since 1990 found little scientific interest in these categories and no clinical utility: they neither help with treatment selection nor provide essential public-health information, and using them “may create unnecessary harm by delaying accurate diagnosis and treatment”. Published in WHO’s own Bulletin.

7. May 2019

   ICD-11 is adopted

   Resolution WHA72.15; in force from 1 January 2022. Neither ICD-11 nor DSM-5 contains a single category that can be applied to a person on the basis of their sexual orientation.

If someone shows you code F66 in a Bulgarian handbook

They will. Bulgaria uses ICD-10, and it does contain section F66 — “psychological and behavioural disorders associated with sexual development and orientation” — with subcategories F66.0, F66.1, F66.2, F66.8 and F66.9. We checked them directly against WHO’s own reference.

Three things are true at once, though. First: ICD-10 itself carries the explicit note that sexual orientation alone is not to be regarded as a disorder. Second: F66.1 (“ego-dystonic sexual orientation”) describes a person whose orientation “is not in doubt, but the individual wishes it were different because of associated psychological and behavioural disorders” — so the code is about the distress around an orientation, not about the orientation. Third: the WHO working group recommended that the whole section be deleted, and in ICD-11 it was.

What this sounds like in plain language

All three sentences are from the same American Psychological Association page (2008) and must be read together. “No consensus about the exact reasons” does not mean “so it is a choice” — the same page says most people experience no sense of choice. We publish all three so the page cannot be quoted by halves.

Can it be “fixed”

This is no longer a matter of curiosity. How common these practices are in Bulgaria we do not know — we found no verified data and will not guess. What the research does show, where it has been measured, is that people rarely go on their own: studies of adolescents describe attempts initiated by a parent.

On whether it works, the American Psychological Association answered as far back as 2008: “To date, there has been no scientifically adequate research to show that therapy aimed at changing sexual orientation (sometimes called reparative or conversion therapy) is safe or effective.”

On whether it harms, the evidence is more plentiful and points one way.

Blosnich et al., American Journal of Public Health, 2020

The strength of this study lies in one detail: the comparison was made after adjusting not only for demographics but for adverse childhood experiences. The harm signal survives accounting for a difficult childhood. Separately, among young people aged 13–24 (Green et al., American Journal of Public Health, 2020) those who went through such practices were more than twice as likely to report a suicide attempt, and Ryan and colleagues (2020) link attempts by parents and being sent to therapists and religious leaders with depression, suicidal thoughts and attempts, less educational attainment and lower income.

Who opposes them

The American Psychological Association’s February 2021 resolution states that “scientific evidence and clinical experience indicate that SOCE put individuals at significant risk of harm”, and lists the bodies holding the same position:

• Pan American Health Organization (2012)
• American Psychiatric Association (2000)
• National Association of Social Workers, USA (2015)
• World Psychiatric Association (2016)
• American Counseling Association (2017)
• American Academy of Child and Adolescent Psychiatry (2018)

The UN position

The report of the UN Independent Expert Victor Madrigal-Borloz of 1 May 2020 (A/HRC/44/53) concludes that these practices are discriminatory in nature and that subjecting people to them is “by their very nature degrading, inhuman and cruel and create a significant risk of torture”.

The report also describes what they rest on: “the medically false pathologization of sexual orientation and gender identity, manifested through interventions that inflict severe pain and suffering and result in psychological and physical damage”. It cites the Pan American Health Organization (2012), which found no medical justification and a severe threat to health and rights, and the World Psychiatric Association (2016), which found there is no sound scientific evidence that innate sexual orientation can be changed.

In a survey cited in the report — 8,000 respondents from 100 countries — of the 940 people who reported having undergone such practices, 98% testified to having suffered damage. The report recommends that states ban the practices across all settings: health care, religious, educational, community and commercial, public and private.

Bans around the world

We have to be especially precise here, because we do not have a verified current list. All we can document is the snapshot in the UN report of 2020: at that time the practice was criminalised in Ecuador and Malta; Canada, Spain and the United States had bans at local level; and Brazil had a professional-body resolution from 1999, extended in 2018. Malta’s law is the Affirmation of Sexual Orientation, Gender Identity and Gender Expression Act, 2016.

The number of countries with bans has grown since, but we could not obtain a verified list as of 2026 — the databases that maintain one were not accessible during our checking. So you will not read a number here. We will add one when we have verified it.

The design of bans differs too: in Malta, adults who freely consent are excluded from the ban; several US bans cover only mental health professionals treating minors, whereas European bans include all providers; and in Aragón and Valencia the practices cannot be validated by consent.

And in Bulgaria

We found no ban on conversion practices in Bulgarian law, and Bulgaria does not appear among the banning jurisdictions in the 2020 UN report. But we could not check this against a current legal source, so we record it as an unconfirmed finding rather than a fact. If you know otherwise, write to us — we will correct the page.

What is verified and concerns Bulgaria: ILGA-Europe’s country record (last modified 12 May 2026) states that one year after Bulgaria’s parliament banned the discussion of LGBTI issues in schools, students and teachers reported that the law has deepened fear and anxiety in the classroom; the amendments prohibit so-called “propaganda” of “non-traditional sexual orientation” in schools.

And one substantive clarification: Bulgarian medicine’s position on classification follows automatically from WHO membership. Bulgaria uses the ICD, in which homosexuality is not a disorder — that is written into ICD-10 itself, and ICD-11 has no category applicable on the basis of sexual orientation at all.

How many people

The numbers differ widely, and that is not a mess — different surveys measure different things. Behaviour (“have you had a same-sex partner”) is one thing, identity (“how do you describe yourself”) another, attraction a third. A census is one thing, a telephone poll another.

So below, every figure sits next to what it measures and the country it applies to.

The rise from 3.5% to 9% — and what it actually means

Gallup has measured this share since 2012, when it was 3.5%. In 2023 it was 7.6% (12,145 interviews, margin of error ±1 point), and on 2025 data it is 9%. That number is regularly used as evidence of a “fashion”, a “contagion” or “recruitment”.

But look at the age breakdown in the chart above: 23% under 30, 10% between 30 and 49, and 3% or less over 50. That is exactly what a country looks like when its older generations stayed in the closet. Gay people over 50 have not evaporated — they are simply less likely to say so to a stranger on the telephone. The rise almost certainly reflects a growing willingness to say so, not a change in people.

One more thing rarely said in Bulgaria: bisexual people are the majority of this population in every wave of the poll. Bisexuality is not a transitional stage and it is not rare — it is the largest group.

For Bulgaria we have no number

And we will not invent one. We found no reliable population survey showing what share of Bulgarians are lesbian, gay or bisexual. Borrowing an American or British figure and presenting it as Bulgarian would be worse than admitting that.

The large European survey by the EU Agency for Fundamental Rights (LGBTIQ Survey III, fieldwork 2023, published May 2024) collected 100,577 responses across 30 countries — but the participants are people who already identify as LGBTIQ and chose to fill it in. It is excellent for one thing and useless for another: it shows what those people experience, but it cannot say how many they are.

From it, on Bulgaria: the country is among those with the highest discrimination rates — alongside Cyprus, both at 50% — and among those with the highest reported rates of physical or sexual attacks over five years. We are not publishing the exact attack figure: in the source we opened it appeared once as 16% and once as 19%, and until we can establish which is right, it is more honest to have no number.

What we still do not know

Claims and evidence

Six sentences Bulgarian gay and lesbian people hear most often, and what the research says against each of them.

What bulgayria covers