Science of Sex, Sexuality and Gender Identity

Science can describe how sex traits develop, how sexual orientation and gender identity are measured, which patterns recur across families and populations, and what health outcomes follow different forms of care. It cannot reduce a person to one gene, brain scan or diagnosis, nor can it by itself decide moral, legal or political questions.

Confusion begins when several distinct concepts are treated as though they mean the same thing. Chromosomes are not sexual orientation. Gender expression is not a diagnosis. An intersex variation is not automatically a transgender identity. A person may also experience distress related to sex characteristics or social treatment without every part of that distress having the same cause.

Research in this area draws on genetics, endocrinology, developmental psychology, neuroscience, epidemiology, sociology and clinical medicine. Each discipline sees part of the picture. Their findings often support an account involving several interacting influences, but multifactorial does not mean voluntary. Nor does biological influence mean that one test can reveal a person’s identity or predict an individual life.

The responsible position is neither that science has answered everything nor that evidence has nothing useful to say. It is to ask what was measured, which population was studied, how long participants were followed and whether the design can establish causation rather than association.

What do sex, sexual orientation and gender identity mean?

Sex, sexual orientation and gender identity describe different aspects of a person. Sex concerns a cluster of anatomical and physiological traits. Sexual orientation concerns attraction, identity and behaviour. Gender identity concerns a person’s internal understanding of gender, while gender expression concerns presentation. Gender dysphoria refers to clinically significant distress, not to an identity by itself.

These distinctions matter because studies can produce misleading conclusions when they measure one concept and claim to have measured another. A survey of same-sex behaviour, for example, does not automatically describe everyone who identifies as gay, lesbian or bisexual. A study of clothing or childhood interests does not directly measure adult gender identity.

Sex category is also not identical to any single sex trait. Birth classification is usually based on external anatomy, but chromosomes, gonads, internal reproductive structures, hormones and secondary characteristics may become relevant in different medical or research settings. The National Academies therefore describes sex as a construct based on a cluster of sex traits rather than one universal marker.

How can biology, psychology and society interact without making identity a choice?

Biological development, psychological experience and social surroundings can influence one another without turning sexual orientation or gender identity into a deliberate selection. Human traits commonly emerge through interacting causes. What people call an experience, when they disclose it and how safely they express it may be socially influenced even when the underlying feelings are not chosen.

Biological explanations can include genetic variation, prenatal development, hormones and developmental processes that unfold without conscious control. Psychological explanations concern perception, self-understanding, memory, emotion and the development of a stable or changing sense of self. Social explanations include family expectations, available language, cultural categories, stigma, acceptance and opportunities for relationships or expression.

These levels answer different questions. Biology may help explain why certain patterns arise. Psychology may describe how a person experiences and interprets them. Sociology can show why the same experience is hidden in one setting and openly named in another.

The word environment also causes confusion in twin research. In statistical models, environmental influence means influence not attributed to measured genetic differences. It can include prenatal conditions, individual experiences, random developmental variation, measurement error and factors researchers have not identified. It does not necessarily mean parenting, persuasion or peer pressure.

Social conditions can strongly affect behaviour and reported identity. Laws may discourage disclosure. A culture may lack familiar words for bisexual or nonbinary experience. None of this proves that attraction or gender identity was created by a conscious decision. The scientific question of causation is therefore separate from the everyday question of choice.

Is sexual orientation genetic, and what does research on gay and bisexual people show?

Sexual orientation is influenced partly by biology, including genetic factors, but there is no single gay gene and no genetic test that can determine an individual’s orientation. Twin, family and population studies support a contribution from inherited and developmental factors, while also showing substantial variation that cannot be explained by genes alone.

Family studies can show that same-sex orientation or behaviour clusters to some degree among relatives. They cannot by themselves separate genes from shared prenatal conditions, family environments, cultural influences or patterns of reporting.

Twin studies make a more focused comparison. Identical twins share nearly all their inherited DNA, while fraternal twins share less on average. Greater similarity among identical twins supports genetic influence. Yet identical twins are often different in sexual orientation, which rules out simple genetic determination. Results also depend on recruitment, sample size, whether researchers measure attraction, identity or behaviour, and how nonresponse is handled. Population registries reduce some selection problems but do not remove every source of bias.

A 2019 genome-wide association study analysed data from 477,522 people and identified several genetic variants associated with reported same-sex sexual behaviour. Each had a very small effect. The findings indicated a highly polygenic pattern, involving many variants rather than one decisive gene, and the resulting genetic scores could not meaningfully predict an individual’s sexual behaviour. The study measured behaviour rather than the whole of sexual orientation and drew heavily from people of European ancestry.

Developmental research examines when attraction appears, its stability, childhood patterns, prenatal correlates and differences between groups. It finds regularities, but no developmental sign is a diagnostic test. Population surveys further show that attraction, behaviour and identity do not perfectly overlap and that reported patterns can change for some people over time.

Taken together, the evidence supports probabilistic influences rather than a single universal pathway. It does not support the claim that parents can manufacture homosexuality through ordinary upbringing. It also cannot explain one person’s orientation from a family tree, childhood memory or laboratory result. For a fuller discussion of voluntary choice, conversion efforts and the psychiatric history of homosexuality, see whether being gay is a choice.

What causes gender identity, including transgender identity?

Science has not identified one cause of cisgender, transgender or nonbinary identity. Researchers study genetic influence, prenatal development, hormones, cognition, social experience, family and peer relationships, and patterns of brain structure or function. The evidence supports developmental complexity, but present research cannot reconstruct the cause of an individual person’s gender identity.

Twin research asks whether identical twins are more similar in gender identity or gendered behaviour than fraternal twins. A 2025 systematic review found 16 eligible studies and reported evidence for genetic and environmental contributions. Estimates varied greatly because studies used different measures, ages, samples and definitions. The reviewers concluded that the true size of these contributions remains unclear and that evidence concerning prenatal hormone exposure is inconsistent.

Research involving variations in sex development can offer clues about prenatal hormones and later behaviour or identity. Such conditions are highly diverse, however, and findings from one diagnosis cannot simply be applied to all transgender people. A factor that changes probabilities within a group is not a rule for every individual.

Neuroscience studies sometimes find average differences between groups. These studies face several problems: samples are often small; results do not always replicate; hormone treatment may affect the brain; and researchers cannot always tell whether a difference preceded identity development or followed years of experience. There is no accepted brain scan for determining whether someone is transgender.

Developmental psychology examines how children learn gender categories, understand themselves and respond to social expectations. Sociology studies the language and cultural settings through which identities become visible. These approaches can explain why expression, disclosure and available labels vary across place and time. They do not establish that a persistent identity was copied from friends or selected for social advantage.

Heterogeneity is central. People may recognise gender incongruence in early childhood, around puberty or in adulthood. Some experience severe physical dysphoria; others are more troubled by social treatment. Some identify within a male-female binary, while others do not. A single causal story is unlikely to describe all of them.

What are intersex variations, and what do they show about biological sex?

Intersex variations are congenital differences affecting chromosomes, gonads, hormones, internal reproductive structures or external anatomy. They show that sex traits can sometimes develop in combinations outside typical patterns. They do not show that human sex lacks biological organization; rather, they show why no single trait answers every medical, social or classificatory question.

Examples include congenital adrenal hyperplasia, androgen insensitivity, 5-alpha-reductase deficiency, ovotesticular differences and some sex-chromosome variations. These diagnoses have different causes, health implications and patterns of development. Some are apparent at birth, while others are discovered during puberty, fertility investigation or unrelated medical testing.

Most people have sex traits that follow one of two broad reproductive developmental pathways, with chromosomes, gonads, anatomy and hormones largely aligned. Variations are uncommon relative to these typical patterns, although prevalence estimates differ sharply depending on which conditions are included. It is therefore better to name the population and definition than to repeat one universal intersex percentage.

The existence of exceptions does not make categories useless. It makes them context-dependent. A doctor considering pregnancy needs information about reproductive anatomy. Cancer screening may depend on which organs are present. Endocrine treatment requires hormone and gonadal information. A population survey may need sex assigned at birth and current gender identity. One box cannot perform all these tasks.

Terminology is personal as well as clinical. Some people use intersex as an identity, some prefer differences or variations of sex development, and others use the name of their diagnosis. Historical medical practices, including secrecy and non-consensual interventions intended to make bodies appear more typical, have contributed to justified mistrust. That history calls for informed, patient-centred care, not for pretending that the underlying anatomy is scientifically meaningless.

Is gender identity a mental disorder?

Gender identity is not, by itself, a mental disorder. The DSM-5-TR diagnosis of gender dysphoria concerns clinically significant distress or impairment associated with gender incongruence. The World Health Organization’s ICD-11 places gender incongruence in a sexual-health chapter rather than its chapter on mental and behavioural disorders.

The American Psychiatric Association replaced gender identity disorder with gender dysphoria when DSM-5 appeared in 2013. Its stated purpose was to focus diagnosis on distress rather than treating gender nonconformity itself as pathology. DSM-5-TR, published in 2022, retains separate criteria for children and for adolescents and adults. Not every transgender or gender-diverse person meets them.

ICD-11 took a different approach. It uses gender incongruence and moved the classification away from mental disorders while retaining a health code that can support access to services and insurance. WHO explains that gender-variant behaviour alone is not a basis for diagnosing childhood gender incongruence.

The classifications remain controversial for understandable reasons. A diagnosis can open access to treatment, reimbursement and specialist support. It can also make people feel that their identity has been medicalized or placed under unnecessary professional control. Removing every classification might reduce stigma but could make funded healthcare harder to obtain in systems that require diagnostic codes.

Medicine has a documented history of treating homosexuality and gender variance through moral assumptions disguised as diagnosis. That history warrants scrutiny. It does not follow that every clinical question about distress, co-occurring conditions, consent, medication or surgical risk is prejudice. Respect for identity and rigorous medical assessment can exist together.

The distinction is the key: a person is not a disorder. Distress may still require care, and that care should consider its possible sources, including bodily incongruence, depression, anxiety, trauma, family conflict, discrimination or several factors at once. The linked article on sexual orientation, conversion therapy and DSM history covers the psychiatric history of homosexuality in greater detail.

What are different forms of healthcare trying to do?

Different forms of healthcare have different goals. Psychological support may address distress, relationships or decision-making. Puberty suppression pauses pubertal progression. Hormones promote selected secondary sex characteristics. Surgery alters particular anatomical features. Good evidence assessment must ask whether each intervention achieved its intended outcome in the population that received it.

Mental-health care can treat depression, anxiety, trauma, eating difficulties or other concerns whether or not they are related to gender. It can help a person describe dysphoria, consider options, involve family when appropriate and prepare for social or medical changes. Ethical support should not begin with a predetermined demand that a person must transition or must abandon a transgender identity.

Puberty-suppressing medicines are intended to stop or slow the pubertal changes driven by the gonads. Their immediate endocrine effect can be measured directly. Questions about bone development, growth, later fertility, sexual development and long-term psychological outcomes require different evidence and longer follow-up.

Masculinising or feminising hormones are intended to produce physical changes such as voice change, altered hair growth, breast development or changes in body composition. Evidence that these bodily effects occur is stronger than evidence attributing every later change in mood or quality of life solely to hormones.

Surgery is not one intervention. Chest procedures, facial operations, hysterectomy, gonadectomy and genital reconstruction involve different goals, complication profiles and recovery periods. Studies must therefore be procedure-specific whenever possible.

Age and clinical history matter. Evidence from adults cannot automatically be transferred to adolescents, while findings from carefully assessed specialist patients may not apply to every person who questions their gender. Some people seek several forms of care, some seek one, and others want no medical intervention. Identity alone does not prescribe a treatment plan.

What does evidence say about mental-health support, puberty blockers, hormones and surgery?

The evidence differs by treatment, age group and outcome. Puberty blockers reliably suppress puberty, and hormones reliably produce many expected physical changes. Evidence for psychological and long-term health outcomes is less certain, especially for adolescents. Adult studies generally report beneficial associations, but many remain observational and cannot isolate every cause.

A 2024 systematic review found only ten studies of psychosocial interventions for gender-dysphoric or gender-incongruent children and adolescents. Most were rated low quality. Reported outcomes usually showed benefit or no change, but the designs and varied interventions did not permit strong conclusions about which form of support works best.

Recent adult reviews are more encouraging but still cautious. A 2025 review of 28 comparative adult-outcome studies found moderate- to very-low-certainty evidence that gender-affirming care may improve quality of life, mental health, stigma and healthcare use. Another review found improvements in several patient-reported outcomes after hormones or surgery. Results do not remove the need to study adverse effects, long-term physical health and differences between procedures.

How strong is the evidence, and why do studies disagree?

Evidence strength depends on more than whether a study supports or opposes a treatment. Researchers examine study design, comparison groups, sample size, measurement, follow-up and consistency across studies. A large association can still be biased, while a small or imperfect study may provide useful information when stronger designs are impractical.

Randomized controlled trials are powerful for many treatment questions because random assignment helps balance known and unknown differences between groups. Randomization may be difficult or ethically contested for long-term identity development, surgery or care that participants strongly prefer. Observational studies are therefore common. They can identify patterns and possible benefits or harms, but causal claims require caution.

Confounding occurs when a third factor influences both treatment and outcome. People who obtain care may differ from those who do not in family support, income, baseline distress, health coverage, waiting time or access to experienced clinicians. Statistical adjustment helps only when relevant factors have been measured adequately.

Loss to follow-up matters because the people who remain in a study may not resemble those who leave. Satisfied participants might be more willing to return, making outcomes look better. People doing well might instead stop attending clinics, making the remaining group appear less healthy. Researchers must report who was lost and test how missing outcomes could change the result.

Outcome choice can also alter the public story. A medicine may succeed at suppressing puberty while its effect on depression remains uncertain. Surgery may improve body satisfaction without resolving unrelated anxiety. Bone-density measurements, fractures, fertility, satisfaction, suicide attempts and general quality of life are different outcomes with different follow-up requirements.

Systematic reviews sit high in many evidence hierarchies only when their search methods are sound and the underlying studies are relevant. Combining many weak or dissimilar studies does not automatically create certainty. GRADE-style frameworks therefore assess risk of bias, inconsistency, indirectness, imprecision and publication bias across the whole body of evidence.

What can science settle, and what remains an ethical or policy question?

Science can test factual claims about development, prevalence, treatment effects and risks. It cannot determine a person’s worth, settle a religious doctrine or calculate the correct balance between autonomy and protection. Policy must use evidence, but it also requires ethical judgments about consent, acceptable uncertainty, fairness, age and access to care.

A study may estimate the average benefit of an intervention without deciding who should qualify for it. Evidence that a treatment helps many adults does not automatically establish the same benefit-risk balance for early adolescents. Conversely, uncertainty about long-term outcomes does not prove that an intervention has no benefit or that all patients should be denied individual assessment.

Clinical recommendations also involve values. How much evidence is enough when untreated distress may be serious? How should possible future regret be weighed against regret from delayed care? What safeguards protect young people without turning assessment into indefinite obstruction? Reasonable people may weigh these questions differently even when they agree on the study results.

History adds another warning. Medical labels and claims about nature have been used to defend coercion, conversion efforts and discriminatory laws. Scientific findings did not make those policies morally sound. Readers interested in that legal history can continue with the history of LGBTQ+ criminalization.

The opposite mistake is to treat every request for better evidence as hostility. Patients deserve honest information about benefits, limitations, alternatives and unknowns. They also deserve to be discussed as people rather than political symbols. For readers who want to compare broader historical and scholarly perspectives, books about gender identity and sexuality offers a starting point.

Science works best here when claims remain proportionate to evidence: neither certainty borrowed from ideology nor doubt used as a disguise for contempt.

Frequently Asked Questions

Can a brain scan tell whether someone is transgender?

No accepted brain scan can diagnose a transgender identity or reveal an individual’s gender with clinical reliability. Neuroimaging studies examine average group differences, but results can be affected by sample selection, hormone exposure, life experience and differences between research methods.

Does intersex mean that biological sex has no categories?

No. Intersex variations show that individual sex traits do not always align in typical combinations and that classification can sometimes require careful clinical judgment. They do not erase the two broad reproductive developmental pathways around which human sex is organized.

Are puberty blockers completely reversible?

Stopping treatment allows gonadal puberty to resume in many clinical contexts, but calling the entire intervention completely reversible is too broad. Questions remain about bone development, growth, fertility pathways and psychosexual development, particularly when suppression is followed by gender-affirming hormones.

Does low-certainty evidence mean a treatment does not work?

No. Low certainty means researchers have limited confidence in the estimated size or range of an effect, not that the effect is necessarily absent. Better studies may strengthen the finding, reduce it or reveal differences between patients and outcomes.

Sources and further reading

1. Measuring Sex, Gender Identity, and Sexual Orientation
2. Sexual Orientation, Controversy, and Science
3. Large-scale GWAS Reveals Insights into the Genetic Architecture of Same-sex Sexual Behavior
4. Genetic and Environmental Contributions to Gender Diversity: A Systematic Review of the Twin Literature
5. Gender Incongruence and Transgender Health in the ICD
6. What Is Gender Dysphoria?
7. Interventions to Suppress Puberty in Adolescents Experiencing Gender Dysphoria or Incongruence: A Systematic Review
8. Provision of Gender-affirming Care for Trans and Gender-diverse Adults: A Systematic Review