Gender and sexual identity-based inequalities in adolescent wellbeing: findings from the #BeeWell Study

The study addresses systematic, avoidable, and unfair differences in adolescent wellbeing associated with gender and sexual identity. Although prior work documents inequalities by age, socioeconomic status, ethnicity, and binary gender/sex, less is known about disparities specifically affecting gender and sexual minority youth, particularly in early adolescence when gender and sexuality identities are developing and wellbeing trajectories have long-term consequences. The paper clarifies wellbeing frameworks (hedonic, eudaimonic, complete state) and distinguishes sex, gender, gender identity, gender modality (cisgender vs transgender), and sexual identity. Guided by minority stress theory, prior research shows elevated distress among sexual and gender minority adolescents. This study aims to provide a rigorous, nuanced assessment of the nature and magnitude of early adolescent wellbeing inequalities across multiple wellbeing domains, disaggregated by both gender identity and sexual identity, using a large contemporary UK dataset. The primary research question: What is the nature and magnitude of adolescent wellbeing inequalities pertaining to gender and sexual identity?

The review outlines that adolescence involves substantial biological, psychological, and social change with concurrent declines in wellbeing and increased mental health difficulties. National UK data indicate high prevalence of probable mental disorders, but symptom-only approaches may miss variability in broader wellbeing indicators. International trends show declining adolescent life satisfaction, including marked declines in the UK. The paper differentiates wellbeing frameworks: hedonic (life satisfaction; positive/negative affect), eudaimonic (autonomy, purpose, positive relations, self-acceptance, etc.), and complete state (balance of mental wellbeing and distress). Concepts of sex, gender, gender identity, gender modality, and sexual identity are defined to avoid conflation. Minority stress theory predicts poorer outcomes among LGBTQ+ youth due to prejudice, discrimination, concealment, expectations of rejection, and internalized stigma, potentially compounded by adverse family reactions. Prior evidence shows higher depression, anxiety, suicidality, and self-harm among sexual minority and transgender youth. Large international analyses also show girls report worse wellbeing than boys across multiple domains, reflecting sociocultural and biological factors, including restrictive gender norms. Gaps remain: a focus on late adolescence and US samples, limited attention to gender identity (beyond binary), and emphasis on distress over broader wellbeing. This study addresses these gaps using early adolescent UK data and multiple wellbeing frameworks.

Design and sample: Cross-sectional analysis of #BeeWell survey data from Greater Manchester (England). Participants were 37,978 adolescents aged 12–15 (Years 8 and 10) across 165 secondary schools. Measures: Wellbeing indicators spanned three frameworks. Hedonic: life satisfaction (ONS single item, 0–10), positive affect (PANAS-C positive subscale), negative affect (five ‘pure’ negative affect items from Me and My Feelings emotional difficulties subscale). Eudaimonic: autonomy (Basic Psychological Need Satisfaction and Frustration Scales), self-esteem (Rosenberg positive subscale, child version), optimism (EPOCH optimism), positive relationships (CYRM-R friendships/social support). Complete state: symptoms of distress (full Me and My Feelings emotional difficulties subscale), mental wellbeing (Short Warwick-Edinburgh Mental Wellbeing Scale). Internal consistencies for all scales were acceptable (Cronbach’s alpha ≥ 0.70). To minimize construct overlap across frameworks, overlapping items were avoided except where negative affect and distress share source scale components. Measure selection was informed by extensive youth and stakeholder consultation. Covariates: age/year group, free school meal eligibility in last six years (FSM/EverFSM6), special educational needs (SEN), and ethnicity (White reference; Black, Asian, Chinese, Mixed, Other, Unclassified). These were sourced from Local Authority administrative data. Exposures: Gender identity (boy, girl, non-binary, describe in another way, prefer not to say) with boy as reference; sexual identity (heterosexual reference; gay/lesbian; bi/pansexual; describe in another way; prefer not to say). Gender modality (transgender vs cisgender) was analyzed in separate models to avoid multicollinearity with gender identity/sex. Analytic strategy: Preliminary measurement modeling compared unidimensional, bifactor, and correlated factors structures for each framework, selecting correlated factors based on fit, interpretability, and dimensionality indices. Structural correlated factors models were then estimated for each framework. Model A included covariates, gender identity, and sexual identity (multi-categorical predictors dummy-coded with specified references). Model B replaced gender identity with gender modality (transgender vs cisgender). Estimation and handling of data structure: Analyses were conducted in Mplus 8.7. Clustered data (students nested in schools) were handled via TYPE=COMPLEX to adjust standard errors. Items were treated as continuous (all but Me and My Feelings had ≥5 categories and acceptable symmetry). Estimator: Maximum Likelihood with Robust SEs (MLR). Missing data were addressed with Full Information Maximum Likelihood (FIML). Sensitivity analyses used WLSMV with multiple imputation (20 datasets) for hedonic and complete state models; results were highly similar to MLR (negligible path differences). Model fit evaluation used conventional cutoffs (CFI ≥ 0.95; RMSEA ≤ 0.06; SRMR ≤ 0.08). Given the large sample size, emphasis was placed on effect sizes and confidence interval overlap rather than p-values. Descriptive results indicated expected directions and magnitudes of correlations across wellbeing domains (e.g., negative affect and distress inversely correlated with other domains; moderate to strong inter-domain correlations).

- Across all frameworks and domains, gender and sexual identity inequalities in wellbeing substantially exceeded those for covariates (age, FSM, SEN, ethnicity). In almost all cases, the smallest gender/sexual identity inequality was larger than the largest covariate inequality. Instances where covariate effects were comparable typically involved “prefer not to say” categories. - Gender identity inequalities: The largest disparities were between gender diverse adolescents and boys. Order of magnitude typically: non-binary (largest disadvantage), describe in another way, girls, prefer not to say (all vs boys). Transgender adolescents also showed notable disadvantages compared to cisgender peers when gender modality was modeled. - Sexual identity inequalities: Consistent pattern of largest disparities for sexual minority adolescents vs heterosexual peers, with bi/pansexual and gay/lesbian adolescents showing very similar magnitudes of disadvantage. Example (eudaimonic self-esteem): gay/lesbian β ≈ −0.76 SD; bi/pansexual β ≈ −0.74 SD; describe in another way β ≈ −0.27 SD; prefer not to say β ≈ −0.17 SD (all vs heterosexual). - Robustness to gender identity specification: Substituting gender identity with gender modality (Model B) did not materially alter sexual identity disparities. Example (complete state symptoms of distress): gay/lesbian vs heterosexual associated with ≈ +0.91 SD in Model A and ≈ +0.94 SD in Model B. - Variability across frameworks and domains: The largest inequalities were observed for negative affect (hedonic) and symptoms of distress (complete state), often exceeding other domains with non-overlapping CIs for girls, bi/pansexual, and gay/lesbian youth. The smallest inequality was typically in positive relationships (eudaimonic). Within frameworks, disparities in symptoms of distress were substantially greater than in mental wellbeing for girls and sexual minority adolescents. Illustratively, bisexual/pansexual adolescents had life satisfaction ≈ 0.65 SD lower than heterosexual peers. Overall, magnitudes for gender/sexual identity (often >0.5–0.9 SD) far exceeded typical associations seen for covariates or common risk factors (e.g., bullying ≈ +0.2 SD with distress in prior large studies).

The study provides comprehensive evidence that in early adolescence, gender and sexual identity are among the strongest correlates of wellbeing across hedonic, eudaimonic, and complete state frameworks, surpassing traditional socio-demographic inequalities. The findings answer the central research question by documenting large, consistent disparities affecting gender diverse (non-binary or otherwise described) and sexual minority (gay/lesbian, bi/pansexual) adolescents across multiple wellbeing domains, with the greatest disadvantages in negative affect and symptoms of distress and relatively smaller differences in positive relationships. Transgender adolescents also showed broad disadvantages relative to cisgender peers. These results align with minority stress theory: excess stressors such as prejudice, discrimination, concealment, expectations of rejection, and internalized stigma likely contribute to the observed disparities, potentially exacerbated in early adolescence due to heightened social comparison and affective sensitivity. Comparisons with large multi-cohort studies underscore the substantive importance of these inequalities: effect sizes for sexual minority status on distress (>0.9 SD) dwarf many established predictors (e.g., bullying ≈ 0.2 SD). The findings also highlight that the magnitude of gender disparities depends on how wellbeing is defined, with smaller gender gaps in eudaimonic wellbeing than in hedonic and complete state domains. Implications include the need to broaden measurement beyond distress, adopt inclusive, granular assessments of gender and sexual identity, and prioritize prevention and intervention tailored to LGBTQ+ youth. Mechanistic research should expand beyond victimization to a fuller range of minority stress processes and consider intersections of gender and sexual identity.

This study demonstrates substantial early adolescent wellbeing inequalities by gender and sexual identity across multiple frameworks and domains, particularly for youth whose identities transcend binary gender norms and for those identifying as gay/lesbian or bi/pansexual. The work advances the field by jointly modeling diverse wellbeing constructs, including both gender identity and sexual identity, while controlling for key socio-demographics. Future research should: routinely assess inclusive gender and sexual identities; adopt intersectional approaches (e.g., by ethnicity, socioeconomic status); and employ longitudinal designs to examine the development and mechanisms (e.g., minority stress processes) underlying these inequalities. The magnitude and consistency of disparities warrant urgent prioritization of comprehensive prevention and intervention strategies to better meet the needs of gender diverse and sexual minority adolescents.

- Exclusive reliance on self-report may introduce common method variance. - Limited coverage of some eudaimonic domains (e.g., environmental mastery, personal growth) due to existing dataset constraints. - Cross-sectional single-wave data preclude causal inference and analysis of developmental trajectories. - Unbalanced sample reduces precision for minority identity groups (wider CIs). - Generalizability may be limited due to societal and institutional differences affecting LGBTQ+ youth across contexts; sample-school composition differences vs national averages also noted. - Lack of data on gender and sexual expression, which may mediate identity–wellbeing associations. - Coding choices for transgender status (gender modality) may mask within-group heterogeneity; further disaggregation (e.g., trans boys vs cis boys) was not pursued due to power and ethical considerations.