RESEARCH DOSSIER
Gender Identity Development and Gender Diversity
Abstract. Gender identity development concerns the formation, interpretation, expression, and possible change of a person's experienced gender across developmental time. This critical research monograph examines conceptual foundations, cognitive-developmental and social-cognitive theories, observational studies of transgender and cisgender children, longitudinal stability and change, nonbinary experiences, minority stress, family and school environments, neurodiversity, cultural variation, and challenges in measurement and causal inference. It integrates original prospective cohorts, behavioral studies, population surveillance, and relevant reviews while distinguishing identity development from clinical intervention effectiveness. Evidence supports meaningful heterogeneity in pathways and the importance of specifying who was sampled, how identity was measured, and when follow-up occurred. The methodology is a critical narrative scientific synthesis, with no registered systematic search, original participant data, randomized intervention trial, or pooled new meta-analysis.
Keywords: gender identity development; gender diversity; gender self-concept; transgender children; nonbinary youth; gender schemas; social transition; longitudinal cohorts; minority stress; gender identity measurement; developmental psychology; childhood; adolescence; social support; causal inference
Gender Identity as a Developmental Research Object
Gender identity refers to an individual's experienced understanding of gender and its relation to available categories, language, relationships, and social institutions. Developmental psychology asks how this understanding emerges, is expressed, becomes meaningful, and sometimes changes across time. The observable material includes spontaneous self-description, repeated choices about names and pronouns, explanations of belonging, attitudes toward gender categories, interactions with peers, and accounts of bodily and social experience. Each indicator captures only part of an internal construct. A child may express a consistent identity while lacking access to an adult vocabulary; an adolescent may change a public label without undergoing an equivalent change in private experience. The research problem therefore requires separating latent experience from its available social expressions Martin et al. (2002) deMayo et al. (2025).
An adequate study specifies at least four dimensions: the person's self-identification, the significance they attach to that identification, how it is expressed to other people, and how these dimensions develop under particular conditions. Gender diversity describes variation within and across populations in these dimensions, including transgender, cisgender, nonbinary, and other locally meaningful identities. These are empirical descriptors that require clear inclusion criteria; they cannot substitute for a uniform developmental pathway. Social transition typically names changes in social presentation, name, pronouns, or participation, while medical interventions are separate events with distinct eligibility, timing, mechanisms, and evidence questions Olson et al. (2022) Coleman et al. (2022).
This monograph evaluates classic developmental theory, experiments on categorization, observational comparisons, longitudinal trajectories, and studies of relational environments. It asks what counts as an informative identity measure, why early category acquisition does not settle later identity, how prospective data distinguish stability from change, what selection processes shape research samples, and which environmental relationships are supported by causal or correlational evidence. Clinical outcomes enter the account where they clarify well-being, distress, or study populations, but efficacy of interventions requires a separate clinical evidence synthesis. The resulting method is a critical narrative review of traceable scholarship. It reports neither a preregistered systematic search nor a newly estimated population effect Martin et al. (2002) Olson et al. (2022) Taylor et al. (2024a).
Sex, Gender Identity, Expression, and Sexual Orientation
Sex-related biological characteristics, gender identity, gender expression, social role, and sexual orientation are analytically distinct constructs. Sex-related variables can refer to chromosomes, gonadal and endocrine processes, anatomy, or an administrative sex designation, depending on the research question. Gender expression includes behavior, clothing, voice, appearance, and other socially interpreted signals. Gender identity refers to experienced gender and self-categorization; sexual orientation describes patterns of attraction, whose development merits its own investigation. A questionnaire that asks only an administrative sex marker cannot answer a question about a nonbinary adolescent's identity, just as clothing preferences alone cannot reliably classify identity [gulgöz2022] deMayo et al. (2025).
The distinction also clarifies everyday language. Gender nonconformity denotes deviation from locally expected patterns of expression or roles, and it occurs among people with many identities. A person may reject conventional gender roles while describing a stable cisgender identity, or conform to conventional patterns while describing a transgender or nonbinary identity. Researchers who use gender nonconformity as a proxy for being transgender introduce construct misclassification. This risk becomes greater across societies and historical periods because the content of gender conventions changes. The interpretation of findings therefore depends on how instruments translate participants' own categories into analytic variables Bem (1981) Bussey and Bandura (1999) Xu et al. (2024).
Measurement choices shape apparent prevalence, developmental stability, and relationships with mental health. A one-question survey with limited response options may place some participants in an inaccurate category, while a two-step measure that separately asks sex assigned at birth and current identity can distinguish some groups more effectively. Open-ended descriptions provide access to terms unfamiliar to investigators but require transparent coding and confidentiality safeguards. Continuous dimensions of felt gender can complement categorical options, although a continuous score should not be treated as a universal substitute for a person's chosen label. Research reports should explicitly describe wording, age suitability, missing responses, and options for participants who decline classification [gulgöz2022].
Historical Development of Gender Psychology
Historical research on gender development emerged from multiple traditions in child psychology, social learning, personality theory, feminist scholarship, and empirical research on categorization. Earlier investigations frequently organized children according to binary administrative groupings and measured the acquisition of culturally familiar roles. This produced important observations about stereotypes, attention, and reinforcement but embedded the assumption that gender development consisted primarily in learning the expectations attached to two categories. Later scholarship increased attention to the distinction between learning social conventions and developing an experienced identity. The historical record should therefore be read as a series of changing research questions and measurement practices rather than a linear progression toward a single comprehensive theory Bem (1981) Bussey and Bandura (1999).
Bem's gender schema theory proposed that children use gender-linked schemas to organize social information, including information relevant to themselves. The approach explains selective attention, recall, and interpretation of culturally salient categories; it does not demonstrate that any person's identity can be deduced from a particular stereotype. Bussey and Bandura's social cognitive account subsequently integrated modeling, reinforcement, self-regulation, and reciprocal interactions between individuals and environments. Martin, Ruble, and Szkrybalo reviewed cognitive theories and specified how categorization, self-socialization, and cognitive development contribute to the emergence of gendered behavior. These contributions remain foundational because they suggest testable mechanisms, while their original measures and population assumptions require contemporary scrutiny Bem (1981) Bussey and Bandura (1999) Martin et al. (2002).
Research including socially transitioned transgender children changed the empirical field by directly measuring participants whose identities differed from sex assigned at birth rather than defining them only through their degree of gender-typed behavior. Prospective studies then allowed investigators to track identity and social experience across multiple waves. The methodological advance was substantial: longitudinal evidence could distinguish a stable childhood identity from a transient response at a single visit, while still leaving generalizability and causal explanation open. The research history therefore connects conceptual revision with better recruitment, consent, and identity-sensitive instruments, as illustrated by cognitive experiments and the Trans Youth Project Olson et al. (2015) [gulgöz2019] Olson et al. (2022).
Gender Schemas and Cognitive-Developmental Accounts
Gender schemas are organized expectations linking social categories with perceived properties, activities, and roles. Children can use such structures to predict other people's behavior, remember category-consistent information, or evaluate whether an activity seems appropriate for themselves. These effects concern information processing and social learning, and their size depends on context and task. A child who accurately knows that a culture associates a toy with one category need not endorse the expectation or identify with the category. Studies of schema acquisition therefore require separate measures of descriptive stereotype knowledge, prescriptive approval, personal preference, and felt identity Bem (1981) Martin et al. (2002).
Developmental models often describe interactive processes. When a child begins to label gender categories, the label can become a cue for selective attention; peers and adults supply repeated examples; memory and feedback can reinforce particular expectations. Cognitive development changes the ability to compare self and others, reason about category boundaries, and understand that labels can have multiple meanings. The hypothesized pathways are reciprocal because children actively choose information and settings. A model that explains category learning may therefore explain the social conditions in which identity is articulated without establishing what caused that identity in the first place Bussey and Bandura (1999) Martin et al. (2002).
An important research distinction separates explanatory reach from theoretical aspiration. Schema theory provides predictions about attention and memory that can be examined through controlled experiments; it offers fewer direct tests of long-term self-identification among transgender and nonbinary populations unless the measures are extended. The model can contribute to a multilevel account alongside family relationships, language, public classification, and personal experience. Investigators should specify whether a study demonstrates that categories influence task performance, that behavior aligns with categories, or that identity changes as a result of a mechanism. These claims have different evidentiary requirements and cannot be inferred from one another Bem (1981) [gulgöz2019].
Early Category Recognition and the Development of Self-Description
Developmental researchers distinguish children's ability to classify other people, their use of a label for themselves, and the importance of that label to their self-understanding. These abilities become observable through different tasks and at different points in early childhood. Researchers may ask a child to identify figures, describe activities, sort photographs, or state a personal category. Each task places demands on language, memory, and social knowledge. Performance can therefore change with vocabulary and task design even when the person's underlying experience has not changed to the same degree. The cognitive literature reviewed by Martin, Ruble, and Szkrybalo offers a basis for separating these operations rather than treating all gender-related task performance as a single developmental scale Martin et al. (2002).
Gender categories are embedded in settings that make some distinctions continually visible. School assignments, marketing, family language, and peer play may reinforce the expectation that gender organizes social life. Learning these expectations is an outcome worthy of study; their presence does not entail that every child regards the available categories as accurately describing themselves. Some children employ the familiar label assigned to them, others use an alternative label, and others express uncertainty or a vocabulary that develops over time. Investigations must permit an answer to differ from binary categories if the intended research construct is self-identification rather than conformity to an experimenter's sorting task [gulgöz2022] Bem (1981).
Repeated measures are more informative than a single isolated response. A careful developmental design documents what was asked, whether a child understood the question, who was present, and whether response alternatives changed between waves. Researchers should not treat a child's drawing, favorite play material, or choice of clothes as a diagnostic instrument. These acts can express social affiliations, aesthetic interests, experimentation, or identity, and an interpretation requires convergent data rather than an investigator's assumption. The scientific aim is to understand how a person develops and communicates a self-description under specific conditions Martin et al. (2002) [gulgöz2019].
Gender Cognition in Transgender and Cisgender Children
Research on transgender children provides an important test of theories developed in samples classified entirely by sex assigned at birth. Olson, Key, and Eaton examined self-report and implicit measures in transgender children who had socially transitioned. Their findings showed gender-related cognition aligned with the children's experienced gender on assessed tasks, providing evidence that gender cognition cannot be assumed to follow administrative sex classification. This result supports the relevance of children's expressed identities for psychological measurement. The sample was specifically recruited and included young children, so the findings describe the studied group and should not be converted into a claim about all transgender children or the origins of their identities Olson et al. (2015).
Gülgöz and colleagues compared transgender and cisgender children on multiple domains of gender development, examining how self-categorization, preferences, and perceived group similarity relate to identity. Evidence of similarities in some patterns of developmental organization challenges an assumption that transgender identity necessarily entails an entirely different cognitive developmental process. At the same time, the particular measures cannot establish equivalence across all psychological domains or later lifespan stages. A cross-group similarity on one task offers a finding about that task, while a causal account of why an identity emerged requires additional data and research designs [gulgöz2019].
These studies also reveal the problem of conditioning a sample on social transition. A child recruited because caregivers have supported a public change of pronouns or names belongs to a distinctive social and family context. Children with similar private identifications who have not disclosed them may be absent from the study. Differences between socially transitioned participants and other cohorts can reflect selection, support, disclosure, or developmental characteristics as well as identity. A transparent report therefore states recruitment criteria and separates results from theoretical extrapolation. More diverse and independently recruited cohorts are essential for comparing developmental experiences across social conditions Olson et al. (2015) Olson et al. (2022). Rae and colleagues' investigation of predictors of early-childhood gender transitions further underscores the importance of asking which pretransition measures and contextual observations are available, and how prediction is distinguished from causal explanation Rae et al. (2019).
Beyond a Single Binary Measure
Gender identity can be recorded categorically, dimensionally, narratively, or through repeated social indicators. A categorical response identifies the label a participant wishes to use at that occasion. Dimensional approaches can measure the extent of identification with different gender categories, the importance of gender to self-concept, or the degree of personal certainty. Gülgöz, Edwards, and Olson investigated measurement of gender identity on a continuum, illustrating how broader response formats can capture variation missed by simple forced-choice approaches. The interpretation still depends on what the questions actually ask: a continuum of perceived similarity is conceptually different from intensity of dysphoria, social presentation, or health-related functioning [gulgöz2022].
Nonbinary and other gender-diverse participants may reject the assumption that identity consists of a position halfway between two socially defined poles. A participant may report an identity outside those categories, multiple identities, a changing identity, or no particular attachment to a gender label. A useful instrument should therefore combine a clear research definition with accessible response options and opportunities for free description. Where measurement must be standardized, researchers can report both categorical coding decisions and their effects on results. Disagreement between coding systems is evidence about measurement sensitivity and warrants analysis rather than quiet recoding [gulgöz2022] deMayo et al. (2025).
Measurement invariance is a central methodological problem. The same question may mean something different to a seven-year-old, an older adolescent, and an adult; a category may carry different implications across languages, schools, or geographic regions. Instruments that appear statistically reliable within one sample can still lack validity for comparisons across groups. Interviewer behavior, parental presence, and fear of disclosure can further alter responses. Longitudinal studies need to preserve a meaningful construct across waves even when age-appropriate wording changes. This calls for cognitive interviewing, qualitative calibration, sensitivity analyses, and explicit records of changed items deMayo et al. (2025).
Sampling, Consent, and Researcher–Participant Relationships
Developmental research on gender diversity often relies on volunteers recruited through community networks, clinics, schools, or advocacy settings. Each route offers access to particular experiences while excluding others. Clinic samples select on care seeking and service availability; family volunteer cohorts may overrepresent caregivers with resources and willingness to participate; school surveys can reach broader populations but contain substantial nonresponse and disclosure constraints. A study's recruitment mechanism is thus part of its substantive interpretation. Even when sampling is rigorous, researchers must identify the population from which an inference is warranted rather than implying that a convenient group represents all children Olson et al. (2022) Suarez et al. (2024).
Research with children introduces special responsibilities involving assent, parental permission, privacy, and the risk that publicly shared information can reveal an identity to an unsafe audience. Participants should understand the difference between contributing information and receiving clinical advice or care. Researchers should avoid leading identity questions and ensure that adolescents may describe uncertainty without being forced into categories. Confidentiality plans require special care when the same participant is surveyed at school, at home, and in clinical services. Longitudinal retention strategies must balance scientific continuity with participants' ability to withdraw or change contact arrangements deMayo et al. (2025).
The relation between investigators and participants also affects disclosure. The language used in a study can signal respect, misunderstanding, or assumptions about what outcomes are desirable. Participants who feel pressured may provide answers that maximize perceived acceptance or minimize conflict. Reflexive training and pretested wording reduce some of these risks; reporting attrition by demographic and recruitment group helps interpret remaining biases. Reliable research treats participants as capable narrators of their own experience while checking the validity of observed patterns with independent measures where appropriate [gulgöz2022] Olson et al. (2022).
Distinguishing Observation, Experiment, and Inference
A basic developmental observation records that a participant expresses an identity at a given time. An association may show that a measured environmental condition co-occurs with this expression, and a prospective relation may show that an earlier condition predicts a later outcome. A randomized intervention could identify some causal effects of that intervention, but such designs are often infeasible or unethical for core identity questions. Consequently, developmental research commonly combines repeated observation, theoretical models, and natural variation. Investigators must state which layer of inference their design supports and avoid turning an observed correlation into a mechanism of identity formation Bussey and Bandura (1999) deMayo et al. (2025).
Different methods answer complementary questions. Controlled categorization tasks can identify how children respond to category cues; qualitative interviews can reveal how they interpret labels; repeated cohort assessments can estimate change within sampled individuals; school surveys can describe population patterns under specified sampling conditions. The methods also have different vulnerabilities. A laboratory task can be ecologically narrow, interviews can depend on language and rapport, cohort follow-up can be selective, and anonymous surveys can misclassify identity. No single method supplies a universal measure of internal experience Martin et al. (2002) [gulgöz2022].
Scientific triangulation requires more than collecting many measures. Results must address common constructs, use appropriate comparison groups, and account for plausible confounders. When a survey and an interview disagree, the difference may reflect temporal change, social context, privacy, or an instrument problem. Such disagreement offers a testable question about how identity is articulated rather than evidence that one response is inherently false. Inference becomes stronger when researchers preregister central hypotheses, report alternative operationalizations, retain missing-data information, and share reproducible analysis code when confidentiality permits deMayo et al. (2025) Olson et al. (2022).
Adolescence, Puberty, and the Changing Meaning of Identity
Adolescence brings changes in bodily experience, social evaluation, relationships, autonomy, and the capacity to construct a longer personal narrative. These changes can alter the significance of gender even when the identity label remains stable. Puberty may make sex-related bodily characteristics more salient; peers may create new expectations about expression; growing independence may make earlier private experiences easier to describe. Research needs to separate these processes from changes in the underlying self-identification itself. A study that measures distress, body satisfaction, or peer acceptance at puberty cannot infer a corresponding change in identity without a direct longitudinal identity measure deMayo et al. (2025) Martin et al. (2002).
Timing also matters. Children who disclose a transgender identity before puberty may differ from adolescents first identifying publicly during secondary school in family experience, access to language, and cohort history. The distinction is descriptive rather than a fixed developmental typology. Public recognition can follow private identification by years, and the availability of nonbinary language can alter the form of self-description. Researchers should measure age at reported self-recognition, age at disclosure, age at social transition, and age at first professional contact as separate variables whenever feasible. Collapsing them into an undifferentiated age of onset risks obscuring the developmental process that the study intends to explain Olson et al. (2022) deMayo et al. (2025).
Longitudinal designs must also distinguish calendar time from developmental time. Changing policy, media vocabulary, school environments, and access to online communities can affect who reports an identity in a survey at different dates. A higher observed proportion in one adolescent cohort could reflect actual changes in self-understanding, greater willingness to disclose, different survey items, or population composition. Age, period, and birth-cohort influences are difficult to separate without repeated representative samples or carefully structured longitudinal comparisons. Scientific interpretation therefore joins developmental psychology to demography and social history Suarez et al. (2024) deMayo et al. (2025).
Nonbinary Identities and Developmental Heterogeneity
Nonbinary is an umbrella description used by some people whose experienced gender is not captured by an exclusive identification as a man or woman. Within it are diverse self-descriptions, including identities described as both, neither, fluid, or organized around terms specific to a community or language. Research should allow participants to specify their preferred terms and avoid assuming that every participant uses the umbrella in the same way. Nonbinary identity may be described early in life or during later developmental periods, and public labels can change with increasing access to language. Identity labels and patterns of psychological distress must be measured separately [gulgöz2022] deMayo et al. (2025).
Developmental heterogeneity is not measurement noise to be eliminated by forcing participants into a binary category. The possibility of stability, re-description, or genuine change can be modeled directly through repeated self-reports and within-person trajectories. A researcher may classify a shift from binary transgender to nonbinary as a change in categorical identity; another may classify both as transgender and record continuity at that broader level. Both representations can be legitimate for different questions, provided the decision is disclosed. The same observation should not be described as evidence of complete identity reversal without examining how the participant understands it Olson et al. (2022) deMayo et al. (2025).
Analyses of nonbinary participants also confront sampling and comparison problems. Pooling all gender-diverse youth can conceal meaningful differences in social recognition, stress exposure, identity certainty, or support needs. Creating many small subgroups can, however, produce imprecise estimates and threaten privacy. Hierarchical models, preregistered aggregation rules, and qualitative accounts can help balance these concerns. The goal is to learn which developmental mechanisms are broadly shared and which depend on identity-specific social conditions, while documenting uncertainty where subgroup data are thin deMayo et al. (2025) [gulgöz2022].
Five-Year Follow-Up of Socially Transitioned Children
Olson and colleagues followed 317 children who had initially socially transitioned and were enrolled in the Trans Youth Project. At the assessment approximately five years after transition, most participants continued to identify with a binary transgender identity, while smaller groups were classified as nonbinary or cisgender. The article reported approximately 94 percent binary transgender, 3.5 percent nonbinary, and 2.5 percent cisgender at the last observation. These figures describe this particular cohort under the study's classification rules, not a population-wide probability for children expressing any gender-related variation. The investigators used recorded pronoun changes as a central classification criterion and openly examined its limitations Olson et al. (2022).
The cohort provides valuable prospective evidence because identity-related information was collected over time after a documented social transition. Yet selection is central to its interpretation. Families volunteered for research, were disproportionately affluent and educated, and had already supported a child's transition. The sample's gender composition and early-transition age profile also differed from many adolescent clinical populations. Researchers cannot use this study to estimate the prevalence or outcomes of all children who experience gender uncertainty, all adolescents seeking clinical services, or those who never disclose their identities. The result directly rebuts a claim that this entire particular cohort rapidly reverted to their earlier classification; it does not establish a universal lifelong trajectory Olson et al. (2022).
Follow-up duration sets another boundary. A child whose identity is stable across five years may develop in several ways during adolescence and adulthood. Retransition can involve multiple pathways, including movement to a nonbinary description or a return to a cisgender description, and these categories should not be treated as identical experiences. The appropriate scientific statement is that many early socially transitioned participants in this cohort maintained a transgender identification over the observed period, with documented minority trajectories. Further waves and independently recruited samples are needed to estimate later transitions and to study the social processes around them Olson et al. (2022) deMayo et al. (2025).
Stability and Change Across Childhood and Adolescence
deMayo, Gallagher, Leshin, and Olson extended the longitudinal discussion by examining stability and change in gender identity and sexual orientation across childhood and adolescence. Their 2025 research monograph emphasizes that developmental trajectories require repeated measures and that gender identity and sexual orientation represent distinct domains. A central interpretive advance is to study change as an observed within-person event rather than infer it indirectly from the distribution of labels at two unrelated ages. Its prospective framework creates a basis for analyzing which aspects of experience remain stable and which evolve with age and social context deMayo et al. (2025).
Repeated observation introduces complex analytical choices. Identity categories can evolve as the available vocabulary changes, so longitudinal coding must decide whether to preserve the original answer verbatim, map answers onto a harmonized scheme, or analyze both. Researchers can estimate within-person persistence at narrow and broad levels of classification; different levels produce different numerical results. They should also distinguish people who change a label, change experienced gender, or change only disclosure context. A trajectory becomes interpretable when the coding rule, assessment schedule, and missing-data handling are explicit. The 2025 monograph makes this broader approach especially salient deMayo et al. (2025).
Attrition is consequential because those who continue participating may differ from those who leave. Families experiencing conflict, relocation, economic hardship, or concern about privacy can be differentially lost to follow-up. Statistical approaches such as inverse-probability weighting or sensitivity bounds may assess some bias, but unobserved reasons for attrition remain. Researchers should report how trajectories vary under plausible assumptions and retain the distinction between observed stability and a causal explanation for stability. The most durable conclusion of longitudinal science is that heterogeneity is measurable and that developmental claims require repeated evidence rather than retrospective stereotypes deMayo et al. (2025) Olson et al. (2022).
Retrospective Narratives, Clinic Samples, and Population Cohorts
Retrospective identity histories can provide rich accounts of meaning, social milestones, and the experience of change. They also depend on memory, current identity, and the circumstances of the interview. People may reinterpret an earlier experience using language learned later, and clinical records may retrospectively organize varied childhood expressions around a current reason for seeking care. Such narratives are legitimate qualitative evidence, but they do not function like repeated prospectively recorded observations. The difference matters when authors claim an age of identity onset or the probability of later persistence deMayo et al. (2025) Olson et al. (2022).
Clinic populations are shaped by referral pathways, eligibility criteria, the availability of specialists, insurance, family circumstances, and local policy. Estimates derived from them should be described as characteristics of service users. A clinic sample may contain more adolescents with distress than a community sample because distress motivates contact; a community volunteer study may contain more families comfortable discussing gender than a school-based probability sample. Comparing their outcomes as if the only difference were identity introduces selection bias. A useful evidence matrix records the sampling frame, entry condition, age distribution, identity instrument, follow-up period, and outcome for each study Tordoff et al. (2022) Olson et al. (2022).
Population surveys add breadth but create their own interpretive challenges. Some respondents may be reluctant to disclose in a classroom or household, some may misunderstand an item, and some categories may be absent from the questionnaire. Changes between survey cycles can reflect changes in wording and response mode. Inference is strongest when repeated surveys maintain comparable measures, report response patterns, and connect their estimates to the target population. Population, cohort, and clinical research are complementary when their different sampling mechanisms remain visible in the scientific synthesis Suarez et al. (2024) deMayo et al. (2025).
Family Relationships, Support, and Developmental Context
Family relationships shape the opportunities through which children communicate experiences and obtain practical, emotional, and social assistance. Investigating family support requires measurement of concrete behaviors: listening, protection from harassment, access to stable relationships, respectful language, and willingness to discuss uncertainty. Support may be correlated with mental health because it reduces hostility and isolation, because youth with better mental health can negotiate relationships more easily, or because shared background factors influence both. These pathways have different causal interpretations. A supportive environment can be studied as a determinant of disclosure and quality of life without assuming that family behavior mechanically determines gender identity Simons et al. (2013) Olson et al. (2016).
Simons and colleagues examined parental support and mental health in a clinic-linked sample of transgender adolescents and reported associations that support the relevance of family context. The cross-sectional design and sample selection limit causal inference, while the study contributes a concrete operationalization of a relationship variable. Subsequent research on socially transitioned children has provided complementary observations of mental health in families recruited through different pathways. Comparisons across such studies should consider baseline distress, income, household acceptance, referral status, and the ages of participants. A mean outcome in one unusually well-supported cohort cannot substitute for an estimate in a broader and more varied population Simons et al. (2013) Olson et al. (2016).
Family processes are dynamic. Parents may change their understanding as a child develops; siblings may offer different responses; and a young person can seek support from extended kin or chosen communities. Longitudinal work should measure who is supportive, when assistance becomes available, and whether the relevant outcome is school participation, self-worth, depressive symptoms, or stability of public expression. Interventions directed at family communication should be evaluated on their actual measured outcomes and adverse possibilities, rather than interpreted as tests of the cause of identity. The scientific picture is relational and context-sensitive Simons et al. (2013) Durwood et al. (2017).
Names, Pronouns, Recognition, and School Environments
Names and pronouns are social practices through which gender identification may be communicated and recognized. Research on their use addresses a distinct psychological question: how social acknowledgment relates to self-worth, stress, belonging, and willingness to participate in community settings. Russell and colleagues studied chosen-name use among transgender youth and reported associations with fewer depressive symptoms and suicidal thoughts or behavior. The study's observational design does not establish that changing a name causes those outcomes, and the recruited population limits broad generalization. Nevertheless, it identifies social recognition as an empirically measurable feature of young people's environments Russell et al. (2018) Pollitt et al. (2019).
School climates differ in institutional policies, peer behavior, adult responses to harassment, and perceived safety. Connectedness measures can capture whether students feel accepted by peers and teachers, while victimization measures describe exposure to bullying or violence. National surveillance from the United States has documented disparities in school connectedness, violence exposure, housing instability, and mental health among reported gender-identity groups. Its cross-sectional structure cannot reveal the temporal sequence of stigma, distress, or identity disclosure, but it shows why school conditions belong in a developmental account Suarez et al. (2024).
Researchers should analyze multiple levels of context rather than reduce recognition to one pronoun variable. A student may be recognized by close friends but not at home, or supported by teachers while experiencing harassment elsewhere. A measure of the number of settings in which a chosen name is used is analytically different from a binary report of ever having used it. School-based changes may also affect privacy and disclosure choices. An informative design records the setting, voluntary choice, relevant safety conditions, and contemporaneous outcomes. This approach enables research into how institutions shape participation without attributing every difference in health to a single social behavior Pollitt et al. (2019) Suarez et al. (2024).
Stigma, Minority Stress, and Mental Health
Mental health outcomes among transgender and gender-diverse people reflect multiple determinants, including personal vulnerabilities, socioeconomic conditions, discrimination, access to services, interpersonal support, and experiences of recognition. Gender diversity as an identity classification does not itself establish psychiatric pathology. Studies frequently find higher levels of some symptoms or risk indicators in samples exposed to stigma or exclusion, yet the size and meaning of differences depend on how populations were recruited and how outcomes were measured. Psychological research must separate identity, distress associated with incongruence, general mental disorder symptoms, and the additional burden of social adversity Suarez et al. (2024) Xu et al. (2024).
A minority-stress framework examines whether repeated rejection, anticipated discrimination, concealment pressures, and internalized stigma contribute to psychological distress. Research can test its components using measures of observed events and temporal changes. Cross-sectional associations are informative for describing burden but remain vulnerable to reverse causation, reporting bias, and common underlying variables. Comparisons of supported and unsupported youth similarly need adjustments for family resources, age, and care seeking. Evidence about contextual effects should be described in proportion to its design; it should not imply that every individual's distress has a single social cause Simons et al. (2013) Durwood et al. (2017) Suarez et al. (2024).
An integrated account also recognizes heterogeneity within gender-diverse populations. A nonbinary respondent experiencing repeated invalidation may face a different exposure pattern from a transgender respondent with stable recognition, and both can have additional stressors unrelated to gender. Research should therefore report specific mental health outcomes, survey periods, recruitment environments, and patterns of intersectional disadvantage. The goal is to identify actionable sources of psychological burden while retaining the distinction between associations and experimentally established causal pathways Xu et al. (2024) Suarez et al. (2024).
Autism, Neurodevelopmental Diversity, and Identity Assessment
Studies have reported associations between gender diversity and autistic traits or autism diagnoses in some datasets. Warrier and colleagues assembled large-scale survey evidence showing elevated recorded rates of autism and other neurodevelopmental or psychiatric characteristics among transgender and gender-diverse respondents compared with cisgender groups in their samples. Such research establishes statistical relationships under the study's ascertainment and measurement practices; it does not demonstrate that autism causes gender diversity or that a gender-diverse identity is explained by an autism diagnosis. Both constructs require independent, valid assessment Warrier et al. (2020).
Several mechanisms could contribute to the observed association, including differences in social expectations, willingness to report identity, recruitment through specialized services, measurement overlap, and shared or independent developmental influences. These possibilities cannot be distinguished by a single cross-sectional correlation. A study designed to identify mechanisms would need temporally ordered measures, control of referral pathways, appropriate comparison groups, and a definition of each variable that does not treat gender nonconformity as a symptom of autism. Research that assumes a causal direction in advance risks misclassifying participants and overlooking individual variability Warrier et al. (2020) [gulgöz2022].
Assessment with neurodivergent participants should attend to communication preferences, sensory experiences, literal interpretations of questions, and varied ways of describing identity. Accommodations can improve the accessibility of research and clinical interviews without predetermining the participant's answer. Clinicians and investigators should explore concurrent difficulties on their own terms and distinguish autonomy, decision-making support, and access needs from an inference about the authenticity of someone's identity. The empirical task is to investigate interacting developmental characteristics while protecting the validity and dignity of each independently measured construct Warrier et al. (2020) Coleman et al. (2022).
Cultural Variation, Institutions, and Intersecting Conditions
Gender categories and the institutions governing their expression vary across languages, countries, and historical periods. A label familiar in a contemporary United States school may not map cleanly onto a category in another setting. Researchers should therefore distinguish participants' locally meaningful self-descriptions from categories introduced for statistical comparison. Cross-cultural work requires translation that retains conceptual meaning, not merely lexical substitution. Measurement invariance should be tested before asserting that observed group differences reflect a common psychological variable Bem (1981) deMayo et al. (2025).
Social opportunities are also unequally distributed. Household income, migration conditions, disability, ethnicity, neighborhood resources, and school policy may affect disclosure safety, access to research participation, and mental health. These factors interact with gender identity but must be analyzed with enough data to support subgroup conclusions. A simple average among all gender-diverse respondents can conceal substantially different exposure profiles. Intersectional analysis contributes scientifically when categories correspond to explicit sampling strata and modeled mechanisms rather than functioning as a list of social labels without testable measures Suarez et al. (2024) Warrier et al. (2020).
The historical specificity of a study is particularly important during rapid social change. Diagnostic terminology, public debates, digital communities, and the availability of identity labels can shift between birth cohorts. Researchers comparing cohorts should check whether the same question could have been answered in the same way at each time point and whether changing conditions influence who feels safe responding. Developmental diversity is consequently studied at the intersection of individual time, relational environments, and historical time. A well-designed program follows these levels together and states precisely which populations its findings describe deMayo et al. (2025) Suarez et al. (2024).
Gender Identity, Distress, and Diagnostic Classification
Psychological studies distinguish the development of gender identity from clinically significant distress and the classification of health-related conditions. An individual can describe a gender identity without meeting a diagnostic definition, and symptoms of depression or anxiety require their own assessment regardless of identity. The World Health Organization placed gender incongruence in the ICD-11 chapter on conditions related to sexual health rather than within mental and behavioral disorders. That classification is an institutional decision about health terminology and service context; it is not an empirical proof of one developmental mechanism. Precise scholarship identifies the classification system and edition being discussed rather than interchangeably applying historical diagnostic labels World Health Organization (n.d.).
Clinical samples often contain people seeking help for distress, bodily concerns, or assistance with social and medical decisions. These entry reasons are not equivalent and can vary with age and jurisdiction. Research that begins with a clinic cohort therefore cannot estimate distress among all gender-diverse people in the community. Equally, a community survey without a diagnostic interview cannot estimate the prevalence of a narrowly defined clinical diagnosis. A source-critical review records whether a finding comes from self-reported identity, professional diagnosis, symptom scale, or health service utilization, each of which represents a distinct outcome Taylor et al. (2024a) Coleman et al. (2022).
The developmental research question remains how identity and related experiences unfold in time. A health-related construct becomes relevant when it influences disclosure, self-description, participation, safety, or psychological functioning. The study should avoid treating ordinary exploration as a symptom or interpreting the absence of current distress as proof that all future difficulties have been resolved. Good measurement follows identity and distress separately, considers the social conditions of assessment, and specifies the validity of any clinical tool used. Clinical judgments require qualified review and cannot be inferred from a monograph's theoretical synthesis World Health Organization (n.d.) Hembree et al. (2017).
Evidence on Psychological Well-Being and Social Transition
Research on socially transitioned transgender children has examined anxiety, depression, self-worth, and peer relationships. Olson and colleagues reported mental health findings for a supported sample, while Durwood and colleagues compared aspects of functioning among socially transitioned youth and comparison groups. These studies contribute evidence that well-being is not uniform across gender-diverse populations and that supportive social contexts deserve attention. However, support and social transition were not randomly assigned. Baseline group differences, voluntary family recruitment, and contextual factors restrict causal claims about whether a transition independently changed an outcome Olson et al. (2016) Durwood et al. (2017).
Anxiety and depression scales assess symptoms within a time window; they are not direct measures of gender identity stability. Likewise, observing that children in one supported cohort have relatively favorable outcomes does not establish an intervention's efficacy in every population. Conversely, population-level risk disparities do not demonstrate that any individual will experience poor mental health. Interpretation requires reporting the outcome instrument, comparator, participant age, recruitment pathway, and time between assessments. The distinction protects both scientific accuracy and the possibility of studying adverse experiences alongside positive adaptation Olson et al. (2016) Suarez et al. (2024).
Social transition comprises potentially different changes in name use, pronouns, appearance, and participation. An observational study may treat these as a single exposure although each element has a different social meaning and different potential consequences. Future studies should document component timing, who initiated each change, the extent of the participant's choice, and the response from family and institutions. Prospective designs with repeated well-being measures may help identify temporal relationships; they will still need appropriate controls and sensitivity analysis for unmeasured confounding. A developmental monograph must preserve that inferential distinction Olson et al. (2022) Russell et al. (2018).
Medical Intervention Research as an Adjacent Evidence Domain
Studies of puberty suppression, hormone interventions, and related clinical outcomes address health treatments rather than the initial formation of identity. Such interventions are clinically consequential, but a study of post-treatment symptom scores does not by itself demonstrate what caused a person's identity or whether a developmental trajectory is universal. Tordoff and colleagues investigated mental health outcomes in a prospective observational clinical cohort, and Chen and colleagues examined psychosocial functioning after two years of hormone treatment. These designs provide valuable follow-up data in specific treated populations while lacking randomized untreated comparison groups and remaining susceptible to selection and concurrent changes Tordoff et al. (2022) Chen et al. (2023).
Systematic reviews by Taylor and colleagues assessed the evidence on puberty suppression and masculinizing or feminizing hormones for adolescents experiencing gender dysphoria or incongruence. Their assessments highlight important limitations of available study designs, including confounding, small or selected samples, inconsistent outcomes, and challenges in identifying long-term effects. Clinical evidence assessment therefore concerns specified outcomes, treatment populations, follow-up periods, and harms as well as benefits. Reviews can reach different interpretations when eligibility criteria, risk-of-bias assessment, or clinical weighting differ; transparency about these methods is indispensable Taylor et al. (2024a) Taylor et al. (2024b).
Professional standards and endocrine guidelines provide clinical frameworks, yet their recommendations belong to a distinct normative and medical level from the empirical study of psychological development. A developmental article can describe how clinical care intersects with adolescence without prescribing treatment, asserting broad causal effects from uncontrolled cohorts, or concealing uncertainty in long-term outcomes. Medication indications, regulatory decisions, service eligibility, and clinical practice rules are jurisdiction- and date-sensitive and require separate contemporary verification and medical review before a page offers clinical guidance. This dossier's clinical discussion is methodological and explanatory Hembree et al. (2017) Coleman et al. (2022) Taylor et al. (2024a).
Causal Inference, Confounding, and Evidence Quality
Causal inference requires a specified exposure, comparator, outcome, population, and time horizon. In gender-development research, a variable such as family support may occur partly in response to a child's prior identity expression; treating it as an independent cause without modeling this feedback can distort conclusions. Mental health, disclosure, social environment, and help seeking may all influence one another. Repeated measures allow investigators to establish temporal sequences and test some competing models, but temporal precedence alone does not rule out confounding. A careful synthesis therefore distinguishes association, prediction, plausible mechanism, and direct experimental identification Simons et al. (2013) Olson et al. (2022).
Research quality also depends on who is missing. Volunteer recruitment, selective follow-up, changes in survey definitions, and exclusion of participants who cannot safely disclose create bias that may operate in different directions. Data-rich clinical studies can still have narrow generalizability; broad population surveys can still have uncertain classification. Analysts should report denominators, age distributions, confidence intervals, nonresponse patterns, and sensitivity to alternative definitions whenever original data permit. This monograph draws on published studies and does not reconstruct individual-level datasets, so it treats numerical claims as specific to the original reports rather than independently re-estimated population parameters deMayo et al. (2025) Suarez et al. (2024).
Claims based on one cohort should be checked against independent studies and different methods. If a finding is replicated among school-recruited adolescents, community-recruited families, and specialist clinic users, its generality becomes more plausible, though each group must still be analyzed on its own terms. Disagreement can reveal measurement differences rather than invalidity. For stronger inference, prospective preregistration, accessible coding documentation, multiple identities represented in samples, and transparent handling of unknown or missing responses are preferable to post hoc certainty. Evaluation of evidence is a continuing process, especially as the social conditions under study evolve [gulgöz2022] deMayo et al. (2025).
Research Ethics, Autonomy, and Responsible Interpretation
Research on gender identity involves participants whose information can have immediate social consequences. Ethical study design should therefore include meaningful consent and assent procedures, limits on disclosure, protection of names and contact details, and plans for situations in which parental and adolescent privacy interests differ. Public release of small subgroup tables can risk reidentification, particularly within schools or small communities. The ethical goal is to produce verifiable knowledge while protecting participants' ability to decide how and when they disclose personal information. These responsibilities arise from general research ethics and become especially salient when a study records intimate identity histories deMayo et al. (2025).
Interpretation has ethical consequences as well. A cohort's estimate should not be used as a prediction of what any single child must become, and an association between identity and mental health should not be converted into a presumption of pathology. Researchers should also document adverse outcomes, participant dissatisfaction, and changes in self-description without treating such reports as failures of a preferred developmental story. Conceptual neutrality in measurement means preserving the range of outcomes that the instruments allow, including stable identity, fluidity, re-identification, uncertainty, and refusal to adopt a label Olson et al. (2022) deMayo et al. (2025).
Meaningful scientific disagreement can focus on instruments, thresholds for clinical outcomes, confounding adjustment, and standards for evidence. Such disagreements are resolved through transparent methods and additional study, with clear reporting of established results and limits. The same evidentiary standard should apply to claims of universal stability and universal change, as well as claims of benefit and harm. An authorial synthesis can identify which research questions are answerable now and which require independent clinical expertise. Responsible scholarship benefits from both intellectual openness and reproducible source attribution Taylor et al. (2024a) Taylor et al. (2024b).
Developmental Identity in Digital and Human–Artificial Environments
Digital communication expands the settings in which young people encounter gender categories, exchange personal narratives, and test how an identity description is received. Online spaces may provide terminology, peers, information, or exposure to harassment. The availability of such environments is relevant to gender identity research as a possible change in social opportunity, but access to gender-related content is not by itself evidence that a particular identity was induced by that content. To establish influence, researchers would need temporally ordered measures of exposure, prior self-understanding, peer context, algorithmic selection, and subsequent outcomes. The broader theories of reciprocal social learning and active information seeking help specify these questions Bussey and Bandura (1999) deMayo et al. (2025).
The Artificial Era introduces conversational and recommendation systems into environments of self-description. Human–Artificial interactions can provide language, reflect a user's statements, or influence what information is encountered. These systems are potential social and informational exposures whose effects must be measured rather than attributed in advance. A study could compare how participants interpret responses from a peer, a search engine, and a conversational system, while assessing privacy, autonomy, and reliance on recommendations. Such experimental questions concern the psychological organization of human interaction with artificial systems; they are prospective research directions, not established explanations of gender identity development Bussey and Bandura (1999).
Research ethics require special scrutiny when artificial systems process sensitive identity disclosures. Participants may misunderstand data retention or treat automated responses as equivalent to professional care. Studies should distinguish perceived support from demonstrated health benefits, disclose the system's role, and assess adverse effects as well as positive experiences. The conceptual contribution of this horizon is to place new media and artificial interlocutors within a general ecological account of development. It preserves the distinction between a human participant's identity, the artificial system's functions, and the observed consequences of their interaction deMayo et al. (2025) Suarez et al. (2024).
Conclusion
Gender identity development is a temporal and relational process through which people acquire, interpret, communicate, and sometimes revise an experienced understanding of gender. Research identifies contributions from category learning, social cognition, changing developmental capacities, language, relationships, and institutional conditions. Empirical studies of transgender and cisgender children demonstrate that gender cognition can align with experienced identity and that developmental features can be investigated across differently identified groups. Prospective work documents substantial stability for many participants in specific early socially transitioned cohorts while recording distinct changes and underscoring the limits of extrapolation to every child or later life stage Olson et al. (2015) [gulgöz2019] Olson et al. (2022).
The strongest scientific synthesis is organized around what each study actually measured. Cognitive experiments establish responses to stimuli; longitudinal surveys document identities under stated coding rules; family and school studies describe relationships between social experiences and well-being; clinical treatment cohorts estimate outcomes within selected care pathways. These evidence streams complement one another without collapsing into one universal explanation. Particularly important unresolved questions concern representative long-term trajectories, construct-valid nonbinary measurement, differences across cultures and cohorts, the causal role of social environments, and precise separations among identity, distress, disclosure, and intervention outcomes deMayo et al. (2025) Taylor et al. (2024a).
A durable research program will combine inclusive repeated measures, prospective samples beyond volunteer networks, carefully specified comparison groups, qualitative interpretation of participants' terms, and transparent statistical sensitivity analyses. Studies should report uncertainty and preserve the possibility of multiple legitimate developmental pathways. Psychological knowledge grows when its constructs remain clear, its participants are accurately represented, and its causal claims remain proportionate to the available designs. This dossier establishes that research architecture for continued inquiry into gender identity development and gender diversity within Psychology of Gender and Sexuality and the wider science of development Martin et al. (2002) deMayo et al. (2025).
References
- Bem, S. L. (1981). Gender schema theory: A cognitive account of sex typing. Psychological Review, 88(4), 354–364. https://doi.org/10.1037/0033-295X.88.4.354
- Bussey, K., & Bandura, A. (1999). Social cognitive theory of gender development and differentiation. Psychological Review, 106(4), 676–713. https://doi.org/10.1037/0033-295X.106.4.676
- Chen, D., Berona, J., Chan, Y.-M., Ehrensaft, D., Garofalo, R., Hidalgo, M. A., Rosenthal, S. M., Tishelman, A. C., & Olson-Kennedy, J. (2023). Psychosocial functioning in transgender youth after 2 years of hormones. New England Journal of Medicine, 388(3), 240–250. https://doi.org/10.1056/NEJMoa2206297
- Coleman, E., Radix, A. E., Bouman, W. P., et al. (2022). Standards of care for the health of transgender and gender diverse people, version 8. International Journal of Transgender Health, 23(sup1), S1–S259. https://doi.org/10.1080/26895269.2022.2100644
- deMayo, B. E., Gallagher, N. M., Leshin, R. A., & Olson, K. R. (2025). Stability and change in gender identity and sexual orientation across childhood and adolescence. Monographs of the Society for Research in Child Development, 90(1–3), 7–172. https://doi.org/10.1111/mono.12479
- Durwood, L., McLaughlin, K. A., & Olson, K. R. (2017). Mental health and self-worth in socially transitioned transgender youth. Journal of the American Academy of Child & Adolescent Psychiatry, 56(2), 116–123.e2. https://doi.org/10.1016/j.jaac.2016.10.016
- Hembree, W. C., Cohen-Kettenis, P. T., Gooren, L., et al. (2017). Endocrine treatment of gender-dysphoric/gender-incongruent persons: An Endocrine Society clinical practice guideline. Journal of Clinical Endocrinology & Metabolism, 102(11), 3869–3903. https://doi.org/10.1210/jc.2017-01658
- Martin, C. L., Ruble, D. N., & Szkrybalo, J. (2002). Cognitive theories of early gender development. Psychological Bulletin, 128(6), 903–933. https://doi.org/10.1037/0033-2909.128.6.903
- Olson, K. R., Key, A. C., & Eaton, N. R. (2015). Gender cognition in transgender children. Psychological Science, 26(4), 467–474. https://doi.org/10.1177/0956797614568156
- Olson, K. R., Durwood, L., DeMeules, M., & McLaughlin, K. A. (2016). Mental health of transgender children who are supported in their identities. Pediatrics, 137(3), e20153223. https://doi.org/10.1542/peds.2015-3223
- Olson, K. R., Durwood, L., Horton, R., Gallagher, N. M., & Devor, A. (2022). Gender identity 5 years after social transition. Pediatrics, 150(2), e2021056082. https://doi.org/10.1542/peds.2021-056082
- Pollitt, A. M., Ioverno, S., Russell, S. T., Li, G., & Grossman, A. H. (2019). Predictors and mental health benefits of chosen name use among transgender youth. Youth & Society. https://doi.org/10.1177/0044118X19855898
- Rae, J. R., Gülgöz, S., Durwood, L., DeMeules, M., Lowe, R., Lindquist, G., & Olson, K. R. (2019). Predicting early-childhood gender transitions. Psychological Science, 30(5), 669–681. https://doi.org/10.1177/0956797619830649
- Russell, S. T., Pollitt, A. M., Li, G., & Grossman, A. H. (2018). Chosen name use is linked to reduced depressive symptoms, suicidal ideation, and suicidal behavior among transgender youth. Journal of Adolescent Health, 63(4), 503–505. https://doi.org/10.1016/j.jadohealth.2018.02.003
- Simons, L., Schrager, S. M., Clark, L. F., Belzer, M., & Olson, J. (2013). Parental support and mental health among transgender adolescents. Journal of Adolescent Health, 53(6), 791–793. https://doi.org/10.1016/j.jadohealth.2013.07.019
- Suarez, N. A., Trujillo, L., McKinnon, I. I., Mack, K. A., Lyons, B., Robin, L., Carman-McClanahan, M., Pampati, S., Cezair, K. L. R., & Ethier, K. A. (2024). Disparities in school connectedness, unstable housing, experiences of violence, mental health, and suicidal thoughts and behaviors among transgender and cisgender high school students—Youth Risk Behavior Survey, United States, 2023. MMWR Supplements, 73(Suppl-4), 50–58. https://doi.org/10.15585/mmwr.su7304a6
- Taylor, J., Mitchell, A., Hall, R., Heathcote, C., Langton, T., Fraser, L., & Hewitt, C. E. (2024). Interventions to suppress puberty in adolescents experiencing gender dysphoria or incongruence: A systematic review. Archives of Disease in Childhood, 109(Suppl 2), s33–s47. https://doi.org/10.1136/archdischild-2023-326669
- Taylor, J., Mitchell, A., Hall, R., Langton, T., Fraser, L., & Hewitt, C. E. (2024). Masculinising and feminising hormone interventions for adolescents experiencing gender dysphoria or incongruence: A systematic review. Archives of Disease in Childhood, 109(Suppl 2), s48–s56. https://doi.org/10.1136/archdischild-2023-326670
- Tordoff, D. M., Wanta, J. W., Collin, A., Stepney, C., Inwards-Breland, D. J., & Ahrens, K. (2022). Mental health outcomes in transgender and nonbinary youths receiving gender-affirming care. JAMA Network Open, 5(2), e220978. https://doi.org/10.1001/jamanetworkopen.2022.0978
- Warrier, V., Greenberg, D. M., Weir, E., Buckingham, C., Smith, P., Lai, M.-C., Allison, C., & Baron-Cohen, S. (2020). Elevated rates of autism, other neurodevelopmental and psychiatric diagnoses, and autistic traits in transgender and gender-diverse individuals. Nature Communications, 11, 3959. https://doi.org/10.1038/s41467-020-17794-1
- World Health Organization. (n.d.). Gender incongruence and transgender health in the ICD. https://www.who.int/standards/classifications/frequently-asked-questions/gender-incongruence-and-transgender-health-in-the-icd
- Xu, Y., Feng, J., & Rahman, Q. (2024). Gender nonconformity and common mental health problems: A meta-analysis. Clinical Psychology Review, 114, 102500. https://doi.org/10.1016/j.cpr.2024.102500
Social Cognitive Theory, Modeling, and Self-Regulation
Social cognitive theory situates gendered development in reciprocal relations among personal cognition, observed behavior, and social opportunity. Children observe models, anticipate approval or sanctions, develop self-evaluative standards, and select settings in which certain forms of expression are supported. Caregivers, peers, teachers, media, and institutions supply different and sometimes inconsistent models. These interactions can account for variation in gendered preferences and strategies of expression across contexts, but observing a social influence on expression is insufficient to infer that a particular identity was produced by imitation Bussey and Bandura (1999).
The model makes several testable predictions about contexts. The salience of a category may change when teachers divide activities by gender, while the apparent acceptability of an expression may change when a peer group responds differently. Researchers can compare settings, measure exposure to models, and follow the same participants as social incentives vary. However, self-selection confounds exposure: people can seek peers who share an emerging identity, and caregivers may respond to a previously expressed identity. Without appropriate temporal measurement or intervention, an association between social acceptance and identity articulation leaves the direction of causality unresolved. These mechanisms are particularly important for interpreting family-support studies Simons et al. (2013) Russell et al. (2018).
A mature account also grants individuals agency in how they interpret available conventions. The same cultural rule can be followed, resisted, ignored, or strategically used. Social regulation may affect willingness to report an identity publicly and the psychological costs of concealing it, even when internal identification is relatively stable. This distinction makes supportive environments a research topic in their own right: support may predict well-being and disclosure, yet studies need designs that distinguish those outcomes from underlying identity trajectories. Social cognition supplies a necessary level of explanation within a broader developmental science Bussey and Bandura (1999) deMayo et al. (2025).