Longitudinal associations between school environment and mental health from childhood through early adulthood

Longitudinal associations between school environment and mental health from childhood through early adulthood | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Longitudinal associations between school environment and mental health from childhood through early adulthood Kaili Rimfeld, Rebecca Ferdinand, Anna Suarez, Agnieszka Musial, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8626819/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Children spend a significant part of their lives at school. However, the long-term effects of the school environment on mental health are still not well understood, especially using genetically sensitive designs. Here, we examine the associations between perceived school environment and mental health from childhood to emerging adulthood, and the genetic and environmental factors that underlie these relationships. Using data from over 6,500 participants aged 7–21 from the Twins Early Development Study (TEDS), we found consistent, moderate associations between perceived school environment and mental health (average r ≈ .19). School environment cumulatively explained mental health problems, explaining 26–56% of the variance both contemporaneously and over time. These associations remained substantial after adjusting for genetic predisposition using psychiatric polygenic scores, family socioeconomic status and earlier mental health problems, although the effect sizes were smaller (6–30% of variance explained). Twin analyses showed that not only was psychopathology highly heritable (~ 61%), but also how children experience school was partly due to genetics (~ 46%). The association between perceived school environment and mental health was largely accounted for by shared genetic influences (~ 70%), supporting the role of gene–environment correlation in mental health outcomes. We show that perceived school environments are significantly associated with mental health across development, even after accounting for genetic predisposition, SES and earlier mental health problems. Using a genetically sensitive, longitudinal approach, this research provides a conservative yet clearer estimate of how school environments might influence mental health outcomes over time, because participants completed their schooling more than a decade ago, when reported youth mental health problems were lower, and school environments were likely less pressured. Our findings emphasise the importance of understanding the school environment as a potential setting for prevention and support and underscore the need for further research to identify modifiable factors that could improve children’s wellbeing. Biological sciences/Genetics Biological sciences/Psychology Mental health school environment cumulative risk gene-environment correlation developmental psychopathology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction The rates of reported mental health problems have been rising globally. According to the latest World Mental Health report, approximately one in seven people worldwide lives with a mental disorder ( 1 ). In the UK, mental health issues accounted for 7% of all ill health based on the disability-adjusted life years (DALY) measure in 2019 ( 2 ). This burden is not only personal but also societal, with recent estimates placing the annual cost of mental ill health in the UK at around £300 billion in 2022, more than double earlier estimates ( 3 , 4 ). In England, among 17–19-year-olds, the prevalence of a probable mental disorder increased from 10% (2017) to 23% (2023) ( 5 ). This observed rise in the prevalence of mental disorders among children and adolescents has generated significant concerns. Although it has been speculated that some of this rise may be due to awareness efforts leading to some individuals interpreting milder forms of distress as mental health problems, it is also likely that this trend partly reflects heightened awareness and reduced stigma, appropriately leading to more frequent help-seeking by children and their guardians ( 6 , 7 ). Substantial evidence shows that mental health problems often begin early, typically emerging by adolescence ( 8 , 9 ). Furthermore, studies indicate that the earlier symptoms of mental disorders appear, the more severe the outcomes tend to be ( 10 ). According to NHS Digital’s Mental Health of Children and Young People in England 2020 survey, one in six children aged 5 to 16 showed identifiable mental health problems ( 11 ). Since common mental disorders are classified as extremes of quantitative traits within a population ( 12 ), many children with poor mental health may not meet clinical criteria for a formal diagnosis despite their real suffering, and therefore, are not reflected in these concerning statistics. Population-based estimates indicate that two in five children score above the threshold for hyperactivity, conduct problems, or emotional difficulties ( 13 ). However, UK mental health services are struggling, and even children at the extremes of mental health dimensions are not receiving adequate help: only one in four children in need of mental health services in 2020 were able to access support ( 11 ). The situation is even more critical for outpatient services, with 81% of providers failing to meet current demands ( 14 ). Early intervention and prevention are thus essential. Nevertheless, effective, evidence-based interventions remain elusive, and universal approaches, such as school-focused mindfulness training, have not been effective and may, in some cases, be harmful ( 15 – 17 ). Although many factors influence children's and young people's psychological wellbeing, increased academic demands may have significantly worsened mental health risks ( 18 – 20 ). Children are increasingly concerned about schoolwork; the OECD reports that approximately 59% of children are concerned about the tests they face, and around 66% worry about receiving poor grades ( 21 ). Childline , a free counselling service for children in the UK, reported that one of the main concerns among children was stress and anxiety related to schoolwork and exam performance ( 22 , 23 ). These educational pressures and associated mental health issues may have long-term negative effects on children and young people ( 24 ), their families ( 25 ), and society ( 26 ). Importantly, school environments may not influence all children the same way; those from disadvantaged socioeconomic backgrounds may be especially vulnerable, with socioeconomic status consistently associated with both academic and mental health outcomes ( 27 – 29 ). However, a recent meta-analysis revealed that while academic pressures are linked to mental health in cross-sectional studies, there is no longitudinal evidence connecting mental health and academic pressures ( 30 ). This evidence highlights the urgent need to understand the underlying factors behind the increase in mental health problems ( 31 ). Identifying preventable risk factors for poor mental health in childhood and adolescence is essential for developing effective interventions that could make a difference. Researchers have long sought these modifiable risk factors within children's homes or neighbourhood environments ( 31 – 34 ). However, the influence of school environments on mental health has received far less attention, even though children spend over 15,000 hours of their lives in full-time education ( 35 ). Schools are often viewed as promising settings for early identification and support. Yet, young people and families frequently report barriers to accessing help, and the quality of support in schools varies greatly ( 36 ). Moreover, school life can itself be a source of stress, with research showing marked variation in school practices and limited evidence on which approaches best promote mental wellbeing ( 16 , 37 , 38 ). Although meta-analytic and systematic reviews link selected aspects of the school environment, such as school climate and connectedness, to internalising problems ( 39 – 41 ), these studies typically focus on individual elements of the school environment in isolation. According to the Good Childhood Report by The Children's Society, which has tracked children's well-being in the UK for many years, children are increasingly reporting unhappiness with school and schoolwork (12% of children aged 10–15), and this trend is more pronounced among older children. Among the various aspects of school life, children were most satisfied with a sense of safety and their relationships with peers, whereas they were least satisfied with the amount of schoolwork and the level of attention they received from their teachers ( 42 ). However, although there is cross-sectional evidence for links between the school environment and mental health, longitudinal research remains limited, and most existing studies focus on narrow facets of the school experience rather than its broader, cumulative effects ( 20 ). To address this gap, a more comprehensive understanding of perceived school environment across the developmental spectrum is needed. Even less is known about the links between school environment and mental health using genetically sensitive designs. Yet, we know from decades of twin studies that individual differences in mental health outcomes are partly explained by genetic factors, with heritability estimates ranging between 30 and 80% ( 43 , 44 ). Furthermore, both objective (e.g. neighbourhood characteristics) and subjective (e.g. perceived noise at home) environmental experiences have been shown to be partly heritable, albeit with modest heritability estimates ( 45 – 49 ). This means that genetic differences not only influence mental health outcomes but could also shape how children perceive and interact with their environments. The home and school environment may be particularly susceptible to these genetic influences ( 45 ). Importantly, even within the same settings, children may experience their environments differently, partly due to their genetic predispositions. For example, when growing up in the same family (home environment) or attending the same school (school environment), children often report these environmental experiences differently ( 50 ). Furthermore, as children grow older, their environments become increasingly dynamic and self-directed, with children selecting, modifying, and creating their environments in ways influenced by their genetically driven behaviours and traits ( 44 ). Therefore, it is important to consider genetic factors when studying the associations between environmental factors (school environment) and life outcomes (mental health). Here, we use longitudinal data on self-, parent- and teacher-reported educational environment and mental health, capitalising on the rich data collected from the Twins Early Development Study (TEDS) over two decades. The differential resemblance between identical and fraternal twins enables investigation of the aetiology of associations between perceived school environment and mental health outcomes during childhood and early adulthood. It allows us to test whether the association between school environment and mental health arises from shared genetic factors, shared environments, or unique individual experiences. In addition, it is now possible to calculate genome-wide polygenic scores (PGS) that leverage summary statistics from genome-wide association (GWA) studies to aggregate the effects of single DNA variants into a single index to predict individual-specific propensities for mental health ( 51 ). While twin analyses quantify the overall genetic and environmental contributions, PGS analyses allow us to partially adjust for measured genetic liability when estimating the association between perceived school environment and mental health. We will use PGS to control for genetic confounding when studying the associations between educational environment and mental health. Additionally, we will control for baseline mental health outcomes (mental health prior to starting school) and family socioeconomic status. In our final step, we further control for mental health outcomes in previous measurements, e.g., for age 21 outcomes, by controlling for mental health at age 16. Our research plan was preregistered in the Open Science Framework (OSF; https://osf.io/3c5rk/ ). Methods Participants Participants come from the Twins Early Development Study (TEDS), which comprises 16,810 twin pairs born in England and Wales between 1994 and 1996, who have been assessed in multiple waves across their development from approximately 18 months to the present. Although there has been some attrition, more than 10,000 twin pairs remain actively involved in the study. The demographic characteristics of TEDS participants and their families are reasonably comparable to those of the population in England and Wales for this birth cohort ( 52 ). Here, we use data from a sample of twins with available data on school environment and mental health, N = 6,500. The sample size varies across measures (See Supplementary Table S1 -S2). Written informed consent was obtained from parents prior to data collection and from TEDS participants themselves, once they were over the age of 18. King's College London's Ethics Committee approved the project for the Institute of Psychiatry, Psychology and Neuroscience PNM/09/10–104. Measures The TEDS cohort provides rich data on educational experiences and emotional and behavioural problems from childhood to early adulthood. In the present study, we utilise the data collected at ages 7, 9, 16 and 21. Mental health. Emotional and behavioural problems were evaluated using a battery of questionnaires completed by the parents, teachers, and the twins themselves (Fig. 1(a), Appendix S1(a)). We further calculated a p-factor, i.e., a general factor for psychopathology, for each age group combining all raters, using confirmatory factor analyses (CFA). Here, we were not interested in the structure of mental health problems across childhood, but used CFA as a data reduction technique to derive a general factor of psychopathology, or p-factor ( 51 , 52 ), across raters (parent, teacher, and twin reports) to index mental health problems at every age. Factor loadings and model fit statistics are presented in Supplementary Table S3. As a sensitivity analysis, we also computed separate externalising and internalising factors and only the self-reported factor of internalising problems (see Supplementary Tables S4-S6 for factor loadings and model fit statistics). Perceived school environment. The survey included multiple measures of the school environment, such as classroom environment, homework, and relationships with teachers and peers, which were rated by parents, teachers, and the twins themselves (Fig. 1(b), Appendix S1(b)). Educational experiences were collected across compulsory education (age 7–16), but the measures differ across data collection waves. To evaluate the overall effects of the school environment, we calculated poly-environmental scores (PES) at each age using penalised regression elastic net regularisation, with a hold-out sample to test prediction accuracy, following the procedure of Gidziela et al. (2023) ( 55 ). This method enables us to overcome the problems of multicollinearity and overfitting ( 55 – 57 ). For details on the construction of PES, see Appendix S2. Socioeconomic status (SES). Family SES measures were collected at the first contact when the twins were about 18 months old. The composite measure was computed as a mean of the mother’s and father’s qualification level, the mean of the occupational status, and the mother's age at first birth. SES at first contact was used, as this has the largest sample size. SES is highly stable across development: SES at first contact correlates .77 with SES at age 7 and .57 with SES at age 9, when family income was added to the composite ( 58 ). Figure 1. Measures used across ages and raters: (a) Mental health; (b) Perceived school environment (see Appendix S1 for full details) (a) (b) Genotyping Genotyping was performed using one of two DNA microarrays (Affymetrix GeneChip 6.0 or Illumina HumanOmniExpressExome chips) according to standard protocols. DNA samples were collected from 12,500 individuals in the TEDS sample. Following rigorous quality control procedures, 10,346 samples were available for genomic analysis. Of these, 7,026 were unrelated individuals, and 3,320 to additional dizygotic co-twins. The study genotyped 7,289 individuals using Illumina arrays and 3,057 individuals using Affymetrix arrays. Further information on genotyping and quality control is available elsewhere ( 59 , 60 ). Polygenic risk scores Genome-wide polygenic scores (PGS) were calculated to estimate genetic vulnerabilities for psychiatric problems using LDpred1 ( 61 ). A detailed description of this method and analytic strategies used for the TEDS sample is presented in earlier work ( 62 , 63 ). To construct the PGS, we used the GWA summary statistics for 27 adult psychiatric disorders: ADHD ( 64 ), anxiety ( 65 ), anxious feeling / worrier ( 65 ), ASD ( 66 ), bipolar disorder ( 67 ), cross-disorder ( 68 ), daytime sleepiness ( 69 ), depression ( 70 ), depressive symptoms ( 71 ), excessive daytime sleepiness ( 72 ), externalising problems ( 73 ), fed-up feeling ( 74 ), guilty feeling ( 74 ), insomnia ( 75 ), miserableness ( 74 ), mood disorder spectrum ( 76 ), mood swings ( 74 ), nervous feelings ( 74 ), neuroticism ( 77 ), obsessive-compulsive disorder ( 78 ), posttraumatic stress disorder ( 79 ), risk-taking behaviour ( 68 ), risk tolerance ( 80 ), schizophrenia ( 81 ), sensitivity / hurt feelings ( 74 ), sleep disorders ( 82 ), and wellbeing ( 83 ). We used all available PGS at TEDS related to mental health problems. Statistical analyses The analyses were performed in R version 4.0 ( 84 ). Descriptive statistics and data preparation. We conducted univariate analyses of variance (ANOVA) to explore phenotypic sex differences across measures. Because significant sex differences emerged, the variables were adjusted for sex and age using linear regression. Sex- and age-adjusted standardised variables were used in all downstream analyses. General factor of psychopathology (p-factor) . We calculated a p-factor, i.e., a general factor for psychopathology, for each age group, using confirmatory factor analyses (CFA; See Measures). Phenotypic correlations . Pearson correlations were used to describe the phenotypic correlations between educational experiences and mental health during compulsory education. Here, we used the p-factor (see Measures) as an indicator of childhood mental health problems. Contemporaneous and longitudinal models to predict mental health outcomes (p-factor). We next sought to examine how well the combined perceived school environmental measures explain mental health outcomes across different developmental stages, both at the time when these perceptions were measured and over time. We used multiple regression and prediction modelling (elastic net) to determine the variance explained by school environment in mental health contemporaneously (Model 1) and over time (Model 2). In the subsequent models (Models 3–6), we adopted a conservative approach and controlled for potential confounders: genetics, baseline mental health problems (age 7), and family SES. We also included mental health at the previous age of measurement. However, we acknowledge that this may overcontrol for the effects of the perceived school environment on mental health, particularly as we had already controlled for genetic predisposition to mental health and mental health problems at age 7. In Model 3, we controlled for potential genetic confounding in predictions using 27 adult psychiatric polygenic scores (PGS); in Model 4, we additionally controlled for mental health at the start of compulsory education (p-factor at age 7). In Model 5, we added control for family socioeconomic status (SES). Finally, in fully adjusted Model 6, we also controlled for mental health (p-factor) from the previous measurement (e.g. at age 16, we also controlled for mental health at age 9) Twin analyses. We investigated the aetiology of perceived school environment, measured by the poly-environmental score (PES; see Measures), and mental health, measured by the general factor of psychopathology (p-factor; see Methods), using the univariate twin method. We also examined the aetiology of the relationships between PES and p-factor using bivariate Cholesky decomposition. We employed a twin model to estimate the proportions of individual differences in PES and p-factor explained by genetic, shared environmental, and non-shared environmental factors. Twin methods capitalise on the genetic relatedness between monozygotic (MZ) or identical twins and dizygotic (DZ) or non-identical twins. MZ twins share 100% of their genes, while DZ twins share, on average, 50% of their segregating genes. However, both types of twins share their rearing environment. By comparing the correlations between MZ and DZ twins, it is possible to estimate the relative contributions of genetic, shared environmental, and non-shared environmental effects to individual differences in the trait of interest. Heritability (A), the proportion of variance explained by genetic factors, can be calculated by doubling the difference between MZ and DZ correlations. Shared environmental factors (C), which make children growing up in the same family more similar beyond genetic influences, can be estimated by subtracting heritability from the MZ correlation. The remaining variance is attributed to non-shared environmental factors (E), which include environmental influences that do not contribute to similarities between twins raised together. These are calculated by subtracting MZ correlations from 1. Importantly, the E component also encompasses measurement error. The ACE estimates can be refined more accurately, including 95% confidence intervals, using structural equation modelling ( 44 , 85 , 86 ). The present study employed an OpenMX package in R for all twin analyses ( 87 ). Bivariate twin analyses were used to calculate the proportion of phenotypic association between p-factor and PES explained by genetic, shared-environmental and non-shared environmental factors. Univariate twin analyses can be extended to bivariate genetic analyses to study the aetiology of covariance between traits. The covariance between traits can be decomposed into additive genetic (A), shared environmental (C), and non-shared environmental (E) components by comparing cross-twin cross-trait correlations between MZ and DZ twin pairs. This method also allows for estimating the genetic correlation ( r G), which indicates the extent to which the same genetic factors influence both measures. Shared environmental correlation ( r C) and non-shared environmental correlation ( r E) can be estimated in the same manner ( 86 ). Polygenic score analyses to investigate rGE. PGS provides an index of genetic predisposition to psychiatric problems. We explored how this predisposition is associated with how parents, teachers, and children themselves perceive the school environment, hypothesising that these indicators are not independent. Because adult PGS only explain a small proportion of variance in childhood mental health problems, we used factor analyses to assess how a latent factor of PGS relates to perceived school environment. We calculated a general genomic p factor, using a similar method to the phenotypic p factor, based on 27 adult psychiatric polygenic scores. Then, we used Pearson correlations to describe the phenotypic relationships between perceived school environment and genetic vulnerability to psychiatric disorders. Sensitivity analyses . We conducted a series of sensitivity analyses to assess the robustness, specificity and sensitivity of our findings. First, to examine whether the school environment relates differently to different aspects of mental health, we repeated the analyses using the internalising or externalising factor, and only self-reported internalising factor scores, rather than the p-factor, to assess whether the school environment affects these factors differently. We also ran models using only self-reported internalising scores, as young people may be better placed to report on their own internal states, such as worry or sadness, than parents or teachers. These checks enabled us to determine whether our findings were specific to a single report, thus explained by a reporter bias or reflect a broader pattern. Secondly, because the covariate-adjusted models had significant missing data, we conducted a simpler analysis. We examined the relationship between the perceived school environment during compulsory education and p-factor scores at age 21, approximately five years after completing compulsory education. We repeated this analysis with and without controlling for family SES to determine whether SES or the modelling approach accounted for the associations. Thirdly, because measures of school environment and mental health can overlap or influence each other, it is important to check whether observed associations are genuine or partly reflect the way the data were collected. This is sometimes referred to as a tautological association, where overlapping content contributes to the observed effects ( 88 ). This is especially relevant when both the exposure and outcome rely on self-report, as mental health difficulties may affect how young people perceive and evaluate their school environment. To address this, we conducted sensitivity analyses using teacher-rated mental health outcomes and school environment measures reported by parents and children. We also examined the relationship between teacher-rated school environment and mental health as reported by parents and children. These additional analyses allowed us to assess whether the patterns were consistent across different informants and were not merely due to shared method variance. We applied the Benjamini–Hochberg false discovery rate (FDR) procedure to control for multiple testing across all analyses. All scripts are available on the OSF site ( https://osf.io/3c5rk/ ). Results Descriptive statistics and data preparation. Descriptive statistics (means, standard deviations, skew and kurtosis) were calculated for all perceived school environment and mental health measures for the whole sample and for males and females separately. One twin per pair was randomly selected for all phenotypic analyses to maintain the independence of the data. Analyses of variance (ANOVA) were used to test the significance of sex differences. ANOVA results revealed some gender differences, although on average, gender explained less than 1% of the variance in all measures. The gender differences were larger for some measures; for example, gender differences explained 8% of the variance in prosocial behaviour at age 16. Some variables showed negative and some positive skew (skewness over +/- 1); these variables were transformed using a log transformation. All variables were also corrected for mean sex and age differences, as described in the Methods section. The standardised variables are used in all downstream analyses. Descriptive statistics are presented in Supplementary Table S1 . Descriptive statistics for males and females separately, and the ANOVA results are presented in Supplementary Table S2. Phenotypic correlations . Contemporaneous correlations across all phenotypic measures are presented in Fig. 2 . All perceived school environment measures, regardless of the raters, were associated with mental health reported across raters, as indicated by the general factor of psychopathology (with an average absolute correlation of 0.19). The strongest associations were with teacher-reported classroom environment (r = 0.47) and parent-reported feelings about being accepted in the classroom (r = 0.45), whereas the weakest were for being the first to finish work (r = 0.02) and the first to answer (r = 0.01). These correlations range from positive (protective factors) to negative (harmful factors). Emotional distress (negative affect) and disruptive classroom environment (noise, disorganisation, long waiting times) were consistently linked to higher levels of mental disorders. See Supplementary Table S7 for full statistics, including 95% confidence intervals. Correlations between perceived school experiences and the internalising factor are presented in Supplementary Table S8, with the externalising factor in Supplementary Table S9 and the self-reported internalising factor in Supplementary Table S10. Contemporaneous and longitudinal models to predict mental health outcomes (p-factor). We examined the extent to which perceived school environment explained variance in mental health, both contemporaneously and longitudinally. We employed structural equation modelling (SEM) with full-information maximum likelihood (FIML) as the primary approach, as this method retains participants with partial data and provides robust handling of missing values. Figure 3 shows the results, with full estimates reported in Supplementary Table S11. To assess robustness, we compared these findings with elastic net models, which use cross-validation to limit overfitting (Supplementary Table S12), and with linear regression (LM) using complete cases (Supplementary Table S13). The results were consistent across methods. Variance explained (R²/ΔR²) was very similar in contemporaneous and additive models across SEM, LM, and elastic net methodological approaches. In models that included covariates (Models 4–6), SEM produced estimates consistent with those of LM and elastic net, although the sample sizes were smaller for LM/elastic net due to listwise deletion. Small differences across methods, therefore, likely reflect sample size rather than modelling strategy. We ran a series of models, as presented in Fig. 3 . Model 1 shows contemporaneous variance explained, while Model 2 illustrates longitudinal associations and includes educational perceived school environment from earlier waves (for example, at age 16, it incorporates all perceived school environment collected at ages 7, 9, and 16). We show that perceived educational experiences explain substantial variance in mental health outcomes across development, explaining 26% of variance in mental health outcomes at age 21, years after finishing the compulsory education when these perceived measures were collected. We then sequentially adjusted for potential confounders. First, we added polygenic scores (PGS) for 27 psychiatric traits to account for genetic predisposition to adverse mental health (Model 3). Next, we incorporated baseline mental health at age 7 (Model 4), followed by family socioeconomic status (SES) (Model 5). Finally, we also adjusted for mental health at the previous wave (e.g., age 9 when predicting outcomes at age 16; Model 6). We show that perceived school environment is associated with mental health both contemporarily and over time; these remain substantial and significant even after controlling for possible confounders (explaining 6–30% of variance). Note Model 1 shows contemporaneous variance explained. Model 2 adds longitudinal predictors from earlier waves (for example, at age 16, models also perceived school environment from ages 7 and 9). Model 3 adjusts for polygenic scores for 27 psychiatric traits. Model 4 adds baseline mental health at age 7. Model 5 adds family socioeconomic status. Model 6 further adjusts for mental health at the previous wave Twin analyses. We calculated the poly-environmental score (PES) using the beta values from the elastic net models. Figure 4 shows the aetiology of p-factor at each age as well as the aetiology of school-related PES (see Supplementary Table S14 for full details). As expected, p-factor was highly heritable (average h 2 = 61%), with a small proportion of variance also explained by shared environmental factors (average c 2 = 13%). The remaining variance is attributed to non-shared environmental factors, which also includes the measurement error. Interestingly, the PES also showed substantial heritability (average h²= 46%), with a small proportion of variance also explained by shared environmental factors (average c² = 16%), which was highest at age 7, when the PES included only parent-rated environmental measures. Bivariate analyses revealed that the correlations between the p and PES were primarily explained by genetic factors (averaging 71%), with the remainder of the variance accounted for by shared environmental (averaging 18%) and non-shared environmental factors (averaging 11%), as shown in Fig. 5 (see Supplementary Table S15 for full bivariate model-fitting results). This means that most of the overlap between children’s mental health and their school environment reflects shared genetic influences, with a smaller role for environmental factors. Polygenic score analyses to investigate rGE. The correlations between genomic p and perceived school environment were small in magnitude (average correlation ≈ |0.04|). Although only a handful of associations reached nominal significance and only two survived corrections for multiple testing, this remains noteworthy given that polygenic scores typically account for a small proportion of the variance in childhood psychopathology. These findings provide further evidence of the association between perceived school environment and genetic factors (see Supplementary Table S16). Sensitivity analyses . As a sensitivity analysis, we examined the associations between perceived school environment and internalising, externalising, and self-reported internalising factors separately (see Supplementary Tables S8-S10). The patterns were generally similar, although the strength of the associations varied, with internalising symptoms more strongly linked to emotional aspects of school (e.g., parent-rated acceptance in the class, r = − 0.54) and externalising symptoms more associated with behavioural factors (e.g., teacher-reported homework completion, r = − 0.40). Associations were generally weaker for self-reported internalising, especially for teacher- and parent-rated school environment. Since the p-factor reflects shared variance across internalising and externalising symptoms and combines data from different raters, we use the p-factor as the primary outcome in all other analyses. Missing data are a concern, especially when conducting longitudinal analysis. Data were missing due to attrition, as well as at random, since some data collection waves included only two out of four cohorts. In addition, only a subsample of TEDS participants provided their DNA. We used FIML to handle missing data. As a sensitivity analysis, we tested the associations between school environment and mental health in early adulthood without any intermediate analysis steps (complete cases). Supplementary Table S17 shows these correlations (average absolute correlation 0.12 (range: 0.01–0.31). We also tested whether the correlations change when controlling for family SES, but found only marginal reductions (see Supplementary Table S18). These checks suggest that our main findings are not driven by the modelling strategy, missing data patterns, or family socio-economic status. To check for tautological associations (see Methods), we also examined the associations between teacher-rated mental health outcomes and educational outcomes (parent- and self-rated) (Supplementary Table S19) and parent-rated mental health outcomes and perceived school environment (teacher– and self-rated) (Supplementary Table S20). While these associations were attenuated, they remained significant and substantial. Although we cannot fully rule out tautology, these findings provide added confidence that the observed patterns reflect meaningful associations across different informants. Discussion The impact of school environments on mental health remains understudied, particularly in genetically informed, longitudinal research. Our study addressed this gap by examining the associations between the perceived school environment and mental health from childhood into early adulthood, utilising data from multiple informants, repeated measures, and genetic data. Using over two decades of TEDS data, our findings show that perceived school environment is consistently associated with mental health across compulsory schooling and into early adulthood. The strongest associations were found between the classroom environment and mental health, with weaker links for more peripheral aspects of school life, such as being first to complete homework. Additionally, the long-term relationships were significant, and our analyses indicate that perceived school environment is associated with mental health outcomes in an additive manner, suggesting that these environmental factors may cumulatively elevate the risk of mental health issues. These associations remained substantial and significant even after accounting for earlier mental health problems, genetic confounding through polygenic scores of adult psychiatric disorders, and family socio-economic status. By controlling for these variables, we provide a conservative estimate of the school environment's influence on mental health, as participants in this study attended school more than a decade ago. Since then, educational demands and pressures have increased, with growing concerns about their effects on student wellbeing ( 13 ). It is therefore possible that the associations reported here underestimate the current effects of school environments on young people’s mental health. Our results also showed that gene-environment interplay (rGE) affecting mental health outcomes is widespread in childhood and early adulthood, based on systematic analyses using the twin method and polygenic scores. The perceived environmental measures were significantly heritable, and the association between these perceived school environmental features and mental health was largely explained by genetic factors. This indicates that how pupils, parents and teachers perceive school is partly shaped by children’s inherited dispositions, because genetically influenced behaviours shape peer and teacher relationships, as well as home–school relationships, which in turn influence how adults perceive children’s school environment. By contrast, objective school-level features such as inspection ratings or behaviour policies would be the same for twins in the same school, so a twin model would not be able to capture their heritability. Consistent with rGE, the covariance between perceived school environment and psychopathology was itself substantially explained by genetic factors, indicating that some of the observed association reflects shared genetic influences. We also tested for rGE using a polygenic score approach. Some polygenic scores were associated with school environment, indicating that children’s genetic risks for psychiatric disorders are linked to how they perceive the school environment as well as how parents and teachers perceive children’s educational experiences, although the effect sizes were small. This suggests passive, active, and evocative rGE. In other words, children inherit not only their parents' genes but also their environment (passive rGE), they actively select and shape their environment (active rGE), and they evoke responses from their environment. For example, an inattentive student may elicit specific reactions from teachers, which are associated with their genetic predispositions (evocative rGE) ( 36 ). Children with different genetic predispositions may perceive and experience the same school environment differently, supporting a dual approach that combines whole-school improvements with provision tailored to individual needs, recognising variation in how pupils respond to the same context. Our study design does not allow for causal conclusions, as the methods used were primarily correlational. However, our use of genetically sensitive data and stringent controls enabled us to move closer to causal inference and test whether the observed associations persisted after controlling for major confounders. Although we utilised a large, nationally representative sample and incorporated a range of school environment measures reported by multiple informants across compulsory education, the measures differed at each time point. This variability prevented us from employing longitudinal structural equation modelling, which could have provided stronger evidence for causal directions between perceived school environment and mental health, thereby moving closer to causal inference. We can, however, address the possibility of reverse causation: mental health symptoms at age 21 could not cause the perceived school environment during compulsory education years, especially in adjusted models controlling for mental health symptoms during the school years. We controlled for possible confounders, previous mental health problems, family socioeconomic status, and genetic confounding (polygenic scores), but the associations between school environment and mental health remained significant and substantial. Polygenic scores for adult psychiatric disorders explain a small amount of variance in childhood psychopathology ( 46 , 81 , 82 ). However, this prediction increases linearly as children age, as evident from a smaller drop in prediction when using polygenic scores at age 7 compared to age 21. Nevertheless, we acknowledge that using polygenic scores to control for genetic confounding is limited to the predictive power of these genetic indices ( 91 , 92 ). Our study used subjective measures of educational environment, collected longitudinally from parents, teachers, and twins in the TEDS cohort. Using objective environmental measures is the next step in our research program that aims to establish causal pathways between the educational environment and mental health from childhood to early adulthood. However, a limitation is that participants attended school around 15 years ago, a time when educational pressures were less pronounced, and our measures of the school environment were limited. This likely means our estimates on the associations between school environment and mental health are conservative. Crucially, we were unable to include reports on educational pressures, as these were not collected in TEDS during participants’ school years. Yet, rising educational pressures are a major concern for today’s children, and the longitudinal associations between these pressures and mental health remain neglected in research ( 30 ). Rather than focusing on bullying, which has already been extensively documented as a risk factor for poor mental health ( 93 , 94 ), we examined a broad range of other educational environmental measures. By doing so, our estimates of the association between the school environment and mental health are likely to be conservative, as they do not include the established school-related risk factors. There is increasing interest in the link between school environments and mental health outcomes. For example, during the COVID-19 lockdown, a study found that one-third of children and young people reported an improvement in their mental wellbeing. These improvements were linked to better management of school tasks, reduced pressure and more opportunities for sleep and exercise, factors directly tied to changes in school-related experiences during lockdown ( 95 ). While the study focused on the unique context of the pandemic, these findings highlight the significant influence of school environments on mental health and suggest that certain elements of the school system may be contributing to poor outcomes for some students. Recent government statistics show changes in how families are engaging with the school system. In 2023, the number of children being home-educated in England increased by 20% to an estimated 111,700 children ( 96 ). While 23% of families cited lifestyle or philosophical reasons for this choice, 13% withdrew their children due to dissatisfaction with schools, including inadequate support for special educational needs. These trends highlight potentially important issues within the school environment and suggest that further research is needed ( 96 ). Future research should address these gaps by systematically studying the educational environment using longitudinal, genetically sensitive designs. By understanding the role of these factors in shaping mental health outcomes, we can identify and implement evidence-based changes to educational environments, ultimately supporting better mental health and wellbeing for children and young people. Our study has limitations. We focused only on subjective measures, as previously discussed, and were limited to the measures collected in the TEDS cohort. We used information from teachers, parents and twins; thus, our results may be affected by rater bias. However, different raters were picking up different aspects of the school environment, and although the rater agreement was low (r ~ .2-.3), combining information about perceived school environment from multiple reporters increased prediction of mental health problems both contemporaneously and over time. Using multiple raters may be beneficial for a more comprehensive understanding of children's educational experiences. We also conducted sensitivity analyses using different combinations of raters. Specifically, we examined how parent-rated school environment related to teacher- and child-rated mental health, and how teacher-rated environment related to parent- and child-rated mental health. Although the strength of associations was slightly attenuated in these analyses, the associations remained, increasing confidence in the robustness of our findings. Our study was based on TEDS data, which was a representative sample of the population in England and Wales on recruitment and remains reasonably representative of the population for their birth cohort ( 97 ); however, it does not necessarily reflect the current cohort of schoolchildren in England and Wales. TEDS data primarily consists of white European participants, and the findings may not be applicable to other diverse groups. Using different datasets to replicate our findings is part of our future research program. Another limitation of our study is the issue of missing data, particularly when examining the long-term effects of school environments on mental health outcomes. Neither regression models nor elastic net prediction approaches handle missing data well, which may have affected the robustness of some of our findings. We used FIML in SEM to mitigate missingness issues. Using these different analytical approaches did not affect our results, and all analyses support the view that perceived school environments are important predictors of mental health both during the school years and in later life. Our future systematic research program aims to triangulate between the sample and the study design to establish causal pathways from the educational environment to mental health. In summary, we provide a general overview of the associations between perceived school environment and mental health. In our future research program, we aim to incorporate objective measures of the school environment available through government records, such as the National Pupil Database, which provides quantitative indicators of the school environment (e.g., student-teacher ratio, school resources, absences), as well as Ofsted inspection reports. Risk factors for poor mental health are multifactorial, encompassing biological (e.g., genetic factors), psychological (e.g., experiences and attitudes), and environmental (e.g., socioeconomic disadvantage, school environment) factors. There is an urgent need for research that combines these risk domains to better understand the processes affecting children's mental health. Identifying potentially preventable risk factors of poor mental health in childhood and adolescence is vital to building a solid base for interventions that could make a real difference. Declarations Conflict of Interest The authors declare no conflict of interest. Acknowledgements We gratefully acknowledge the ongoing contributions of the participants in the Twins Early Development Study (TEDS) and their families. TEDS is supported by a programme grant to R.P. from the UK Medical Research Council (Grant Nos. MR/V012878/1 and previously MR/M021475/1). K.R. is supported by a Sir Henry Wellcome Postdoctoral Fellowship and UK Medical Research Council Grant UKRI1503. E.V. is supported by the National Institutes of Health and Care Research (NIHR) University College London Hospitals Biomedical Research Centre (Award: NIHR 203972). MM is supported by a Jacobs Foundation Research Fellowship (#2024-1533-00) and UK Medical Research Council Grant UKRI1506. This research was funded in whole, or in part, by the Wellcome Trust (213514/Z/18/Z). For the purpose of open access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission. References World mental health report: Transforming mental health for all [Internet]. [cited 2022 Dec 11]. Available from: https://www.who.int/publications-detail-redirect/9789240049338 McDaid D, Park AL, Davidson G, John A, Knifton L, Morton A, et al. The economic case for investing in the prevention of mental health conditions in the UK (summary). Ment Health Found. 2022;(February). The economic case for investing in the prevention of mental health conditions in the UK [Internet]. [cited 2022 Dec 11]. Available from: https://www.mentalhealth.org.uk/explore-mental-health/publications/mental-health-problems-cost-uk-economy-least-ps118-billion-year-new-research The economic and social costs of mental ill health: review of methodology and update of calculations [Internet]. Institute for Social and Economic Research (ISER). [cited 2025 Sep 23]. Available from: https://www.iser.essex.ac.uk/research/publications/publication-578405 NHS England Digital [Internet]. [cited 2025 Aug 1]. Mental Health of Children and Young People in England 2022 - wave 3 follow up to the 2017 survey. Available from: https://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2022-follow-up-to-the-2017-survey Deighton J, Lereya ST, Casey P, Patalay P, Humphrey N, Wolpert M. Prevalence of mental health problems in schools: poverty and other risk factors among 28 000 adolescents in England. Br J Psychiatry. 2019;215(3):565–7. NHS England Digital [Internet]. [cited 2024 Dec 29]. Mental Health of Children and Young People in England, 2023 - wave 4 follow up to the 2017 survey. Available from: https://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2023-wave-4-follow-up Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry. 2005;62(6):593–602. Tanner JL. Mental health in emerging adulthood. In: The Oxford handbook of emerging adulthood. New York, NY, US: Oxford University Press; 2016. p. 499–520. (Oxford library of psychology). Otto MW, Pollack MH, Maki KM, Gould RA, Worthington JJ, Smoller JW, et al. Childhood history of anxiety disorders among adults with social phobia: rates, correlates, and comparisons with patients with panic disorder. Depress Anxiety. 2001;14(4):209–13. NDRS [Internet]. [cited 2022 Dec 12]. Mental Health of Children and Young People in England, 2020: Wave 1 follow up to the 2017 survey. Available from: https://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2020-wave-1-follow-up Plomin R, Haworth CMA, Davis OSP. Common disorders are quantitative traits. Nat Rev Genet. 2009;10(12):872–8. Mental health of adolescents [Internet]. [cited 2025 Aug 1]. Available from: https://www.who.int/news-room/fact-sheets/detail/adolescent-mental-health Mental health services: addressing the care deficit [Internet]. [cited 2022 Oct 21]. Available from: https://nhsproviders.org/mental-health-services-addressing-the-care-deficit/the-demand-challenge Marin JM, Allwood M, Ball S, Crane C, Wilde KD, Hinze V, et al. School-based mindfulness training in early adolescence: what works, for whom and how in the MYRIAD trial ? Evid Based Ment Health. 2022;1–8. Mundy J, Moore E, Soan C, Anderson JK, Albajara Saenz A, Baser A, et al. Evaluating the implementation of the Transforming Children and Young People’s Mental Health Provision Green Paper programme: Findings from surveys of schools and colleges and Mental Health Support Teams (2024) [Internet]. London School of Hygiene & Tropical Medicine; 2025 [cited 2025 Jul 15]. Available from: https://researchonline.lshtm.ac.uk/id/eprint/4676423/ Foulkes L, Stringaris A. Do no harm: can school mental health interventions cause iatrogenic harm? BJPsych Bull. 2023;47(5):267–9. Högberg B. Educational stressors and secular trends in school stress and mental health problems in adolescents. Soc Sci Med 1982. 2021;270:113616. Pascoe MC, Hetrick SE, Parker AG. The impact of stress on students in secondary school and higher education. Int J Adolesc Youth. 2020;25(1):104–12. Armitage JM, Collishaw S, Sellers R. Explaining long-term trends in adolescent emotional problems: what we know from population-based studies. Discov Soc Sci Health. 2024;4(1):14. PISA 2015 Results (Volume III): Students’ Well-Being | en | OECD [Internet]. [cited 2023 Apr 19]. Available from: https://www.oecd.org/education/pisa-2015-results-volume-iii-9789264273856-en.htm NSPCC Learning [Internet]. [cited 2022 Dec 13]. Childline annual review. Available from: https://learning.nspcc.org.uk/research-resources/childline-annual-review/ www.basw.co.uk [Internet]. 2015 [cited 2022 Dec 13]. Exam factories? The impact of accountability measures on children and young people. Available from: https://www.basw.co.uk/resources/exam-factories-impact-accountability-measures-children-and-young-people Kelly-Irving M, Lepage B, Dedieu D, Bartley M, Blane D, Grosclaude P, et al. Adverse childhood experiences and premature all-cause mortality. Eur J Epidemiol. 2013;28(9):721–34. Brock-Baca AH, Zundel C, Fox D, Johnson Nagel N. Partnering with Family Advocates to Understand the Impact on Families Caring for a Child with a Serious Mental Health Challenge. J Behav Health Serv Res. 2022;1–18. Prince M, Patel V, Saxena S, Maj M, Maselko J, Phillips MR, et al. No health without mental health. Lancet Lond Engl. 2007;370(9590):859–77. Yang M, Carson C, Creswell C, Violato M. Child mental health and income gradient from early childhood to adolescence: Evidence from the UK. SSM - Popul Health. 2023;24:101534. von Stumm S. Socioeconomic status amplifies the achievement gap throughout compulsory education independent of intelligence. Intelligence. 2017;60:57–62. von Stumm S, Cave SN, Wakeling P. Persistent association between family socioeconomic status and primary school performance in Britain over 95 years. Npj Sci Learn. 2022;7(1):4. Steare T, Gutiérrez Muñoz C, Sullivan A, Lewis G. The association between academic pressure and adolescent mental health problems: A systematic review. J Affect Disord. 2023;339:302–17. Weeland J, Moens MA, Beute F, Assink M, Staaks JPC, Overbeek G. A dose of nature: Two three-level meta-analyses of the beneficial effects of exposure to nature on children’s self-regulation. J Environ Psychol. 2019;65:101326. Baldwin JR, Wang B, Karwatowska L, Schoeler T, Tsaligopoulou A, Munafò MR, et al. Childhood Maltreatment and Mental Health Problems: A Systematic Review and Meta-Analysis of Quasi-Experimental Studies. Am J Psychiatry. 2023;180(2):117–26. Ahmadzadeh YI, Schoeler T, Han M, Pingault JB, Creswell C, McAdams TA. Systematic Review and Meta-analysis of Genetically Informed Research: Associations Between Parent Anxiety and Offspring Internalizing Problems. J Am Acad Child Adolesc Psychiatry. 2021;60(7):823–40. Fowler PJ, Tompsett CJ, Braciszewski JM, Jacques-Tiura AJ, Baltes BB. Community violence: a meta-analysis on the effect of exposure and mental health outcomes of children and adolescents. Dev Psychopathol. 2009;21(1):227–59. Rutter M. Fifteen Thousand Hours: Secondary Schools and Their Effects on Children. Harvard University Press.; 1979. No Wrong Door: bringing services together to meet children’s needs [Internet]. Children’s Commissioner for Wales. [cited 2025 Jul 15]. Available from: https://www.childcomwales.org.uk/publications/no-wrong-door-bringing-services-together-to-meet-childrens-needs/ Anderson M, Werner-Seidler A, King C, Gayed A, Harvey SB, O’Dea B. Mental health training programs for secondary school teachers: A systematic review. Sch Ment Health Multidiscip Res Pract J. 2019;11(3):489–508. Yamaguchi S, Foo JC, Nishida A, Ogawa S, Togo F, Sasaki T. Mental health literacy programs for school teachers: A systematic review and narrative synthesis. Early Interv Psychiatry. 2020;14(1):14–25. Raniti M, Rakesh D, Patton GC, Sawyer SM. The role of school connectedness in the prevention of youth depression and anxiety: a systematic review with youth consultation. BMC Public Health. 2022;22(1):2152. Rose ID, Lesesne CA, Sun J, Johns MM, Zhang X, Hertz M. The Relationship of School Connectedness to Adolescents’ Engagement in Co-Occurring Health Risks: A Meta-Analytic Review. J Sch Nurs Off Publ Natl Assoc Sch Nurses. 2024;40(1):58–73. Aldridge JM, McChesney K, Afari E. Relationships between school climate, bullying and delinquent behaviours. Learn Environ Res. 2018;21(2):153–72. The Good Childhood Report 2022 | The Children’s Society [Internet]. 2022 [cited 2023 Apr 21]. Available from: https://www.childrenssociety.org.uk/information/professionals/resources/good-childhood-report-2022 Polderman TJC, Benyamin B, de Leeuw CA, Sullivan PF, van Bochoven A, Visscher PM, et al. Meta-analysis of the heritability of human traits based on fifty years of twin studies. Nat Genet. 2015;47(7):702–9. Knopik VS, Neiderhiser JM, DeFries JC, Plomin R. Behavioral Genetics. 7th ed. Worth Publishers, New York; 2017. Hanscombe KB, Haworth CMA, Davis OSP, Jaffee SR, Plomin R. The nature (and nurture) of children’s perceptions of family chaos. Learn Individ Differ. 2010;20:549–53. Plomin R, Bergeman CS. The nature of nurture: Genetic influence on ‘environmental’ measures. Behav Brain Sci. 1991;14:373. Plomin R, Reiss D, Hetherington EM, Howe GW. Nature and nurture: Genetic contributions to measures of the family environment. Dev Psychol. 1994;30:32–43. Koellinger P, Harden KP. Using nature to understand nurture. Am J Public Health. 1990;80(6):657–8. Kendler KS, Baker JH. Genetic influences on measures of the environment: a systematic review. Psychol Med. 2007;37:615–26. Daniels D, Dunn J, Furstenberg FF, Plomin R. Environmental differences within the family and adjustment differences within pairs of adolescent siblings. Child Dev. 1985;56(3):764–74. Wray NR, Lee SH, Mehta D, Vinkhuyzen AAE, Dudbridge F, Middeldorp CM. Research Review: Polygenic methods and their application to psychiatric traits. J Child Psychol Psychiatry. 2014;55(10):1068–87. Lockhart C, Bright J, Ahmadzadeh Y, Breen G, Bristow S, Boyd A, et al. Twins Early Development Study (TEDS): A genetically sensitive investigation of mental health outcomes in the mid-twenties. JCPP Adv. 2023;3(2):e12154. Caspi A, Moffitt TE. All for One and One for All: Mental Disorders in One Dimension. Am J Psychiatry. 2018;appi.ajp.2018.1. Allegrini AG, Cheesman R, Rimfeld K, Selzam S, Pingault JB, Eley TC, et al. The p factor: genetic analyses support a general dimension of psychopathology in childhood and adolescence. J Child Psychol Psychiatry. 2020; Gidziela A, Malanchini M, Rimfeld K, McMillan A, Ronald A, Viding E, et al. Explaining the influence of non-shared environment (NSE) on symptoms of behaviour problems from preschool to adulthood: mind the missing NSE gap. J Child Psychol Psychiatry. 2023;64(5):747–57. Gidziela A, Rimfeld K, Malanchini M, Allegrini AG, McMillan A, Selzam S, et al. Using DNA to predict behaviour problems from preschool to adulthood. J Child Psychol Psychiatry. 2022;63(7):781–92. Allegrini AG, Karhunen V, Coleman JRI, Selzam S, Rimfeld K, Stumm S von, et al. Multivariable G-E interplay in the prediction of educational achievement. PLOS Genet. 2020;16(11):e1009153. Hanscombe KB, Trzaskowski M, Haworth CMA, Davis OSP, Dale PS, Plomin R. Socioeconomic Status (SES) and Children’s Intelligence (IQ): In a UK-Representative Sample SES Moderates the Environmental, Not Genetic, Effect on IQ. PLOS ONE. 2012;7(2):e30320. Rimfeld K, Malanchini M, Spargo T, Spickernell G, Selzam S, McMillan A, et al. Twins Early Development Study: A Genetically Sensitive Investigation into Behavioral and Cognitive Development from Infancy to Emerging Adulthood. Twin Res Hum Genet Off J Int Soc Twin Stud. 2019;22(6):508–13. Selzam S, McAdams TA, Coleman JRI, Carnell S, O’Reilly PF, Plomin R, et al. Evidence for gene-environment correlation in child feeding: Links between common genetic variation for BMI in children and parental feeding practices. PLoS Genet. 2018;14(11):e1007757. Vilhjálmsson BJ, Yang J, Finucane HK, Gusev A, Lindström S, Ripke S, et al. Modeling Linkage Disequilibrium Increases Accuracy of Polygenic Risk Scores. Am J Hum Genet. 2015;97(4):576–92. Allegrini AG, Selzam S, Rimfeld K, von Stumm S, Pingault JB, Plomin R. Genomic prediction of cognitive traits in childhood and adolescence. Mol Psychiatry. 2019;24(6):819–27. Selzam S, McAdams TA, Coleman JRI, Carnell S, O’Reilly PF, Plomin R, et al. Evidence for gene-environment correlation in child feeding: Links between common genetic variation for BMI in children and parental feeding practices. PLoS Genet. 2018;14(11):e1007757. Demontis D, Walters GB, Athanasiadis G, Walters R, Therrien K, Nielsen TT, et al. Genome-wide analyses of ADHD identify 27 risk loci, refine the genetic architecture and implicate several cognitive domains. Nat Genet. 2023;55(2):198–208. Purves KL, Coleman JRI, Meier SM, Rayner C, Davis KAS, Cheesman R, et al. A major role for common genetic variation in anxiety disorders. Mol Psychiatry. 2020;25(12):3292–303. Grove J, Ripke S, Als TD, Mattheisen M, Walters RK, Won H, et al. Identification of common genetic risk variants for autism spectrum disorder. Nat Genet. 2019;51(3):431–44. Mullins N, Forstner AJ, O’Connell KS, Coombes B, Coleman JRI, Qiao Z, et al. Genome-wide association study of more than 40,000 bipolar disorder cases provides new insights into the underlying biology. Nat Genet. 2021;53(6):817–29. Strawbridge RJ, Ward J, Cullen B, Tunbridge EM, Hartz S, Bierut L, et al. Genome-wide analysis of self-reported risk-taking behaviour and cross-disorder genetic correlations in the UK Biobank cohort. Transl Psychiatry. 2018;8(1):1–11. Lane JM, Liang J, Vlasac I, Anderson SG, Bechtold DA, Bowden J, et al. Genome-wide association analyses of sleep disturbance traits identify new loci and highlight shared genetics with neuropsychiatric and metabolic traits. Nat Genet. 2017;49(2):274–81. Wray NR, Ripke S, Mattheisen M, Trzaskowski M, Byrne EM, Abdellaoui A, et al. Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet. 2018;50(5):668–81. Howard DM, Adams MJ, Clarke TK, Hafferty JD, Gibson J, Shirali M, et al. Genome-wide meta-analysis of depression identifies 102 independent variants and highlights the importance of the prefrontal brain regions. Nat Neurosci. 2019;22(3):343–52. Wang H, Lane JM, Jones SE, Dashti HS, Ollila HM, Wood AR, et al. Genome-wide association analysis of self-reported daytime sleepiness identifies 42 loci that suggest biological subtypes. Nat Commun. 2019;10(1):3503. Karlsson Linnér R, Mallard TT, Barr PB, Sanchez-Roige S, Madole JW, Driver MN, et al. Multivariate analysis of 1.5 million people identifies genetic associations with traits related to self-regulation and addiction. Nat Neurosci. 2021;24(10):1367–76. Neale lab [Internet]. [cited 2022 Dec 5]. UK Biobank. Available from: http://www.nealelab.is/uk-biobank Jansen PR, Watanabe K, Stringer S, Skene N, Bryois J, Hammerschlag AR, et al. Genome-wide analysis of insomnia in 1,331,010 individuals identifies new risk loci and functional pathways. Nat Genet. 2019;51(3):394–403. Coleman JRI, Gaspar HA, Bryois J, Bipolar Disorder Working Group of the Psychiatric Genomics Consortium, Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium, Breen G. The Genetics of the Mood Disorder Spectrum: Genome-wide Association Analyses of More Than 185,000 Cases and 439,000 Controls. Biol Psychiatry. 2020;88(2):169–84. Luciano M, Hagenaars SP, Davies G, Hill WD, Clarke TK, Shirali M, et al. Association analysis in over 329,000 individuals identifies 116 independent variants influencing neuroticism. Nat Genet. 2018;50(1):6–11. International Obsessive Compulsive Disorder Foundation Genetics Collaborative (IOCDF-GC) and OCD Collaborative Genetics Association Studies (OCGAS). Revealing the complex genetic architecture of obsessive-compulsive disorder using meta-analysis. Mol Psychiatry. 2018;23(5):1181–8. Nievergelt CM, Maihofer AX, Klengel T, Atkinson EG, Chen CY, Choi KW, et al. International meta-analysis of PTSD genome-wide association studies identifies sex- and ancestry-specific genetic risk loci. Nat Commun. 2019;10(1):4558. Karlsson Linnér R, Biroli P, Kong E, Meddens SFW, Wedow R, Fontana MA, et al. Genome-wide association analyses of risk tolerance and risky behaviors in over 1 million individuals identify hundreds of loci and shared genetic influences. Nat Genet. 2019;51(2):245–57. Trubetskoy V, Pardiñas AF, Qi T, Panagiotaropoulou G, Awasthi S, Bigdeli TB, et al. Mapping genomic loci implicates genes and synaptic biology in schizophrenia. Nature. 2022;604(7906):502–8. Yao Y, Jia Y, Wen Y, Cheng B, Cheng S, Liu L, et al. Genome-Wide Association Study and Genetic Correlation Scan Provide Insights into Its Genetic Architecture of Sleep Health Score in the UK Biobank Cohort. Nat Sci Sleep. 2022;14:1–12. Okbay A, Baselmans BML, De Neve JE, Turley P, Nivard MG, Fontana MA, et al. Genetic variants associated with subjective well-being, depressive symptoms, and neuroticism identified through genome-wide analyses. Nat Genet. 2016;48(6):624–33. R Core Team. R: A Language and Environment for Statistical Computing. 2017. Martin NG, Eaves LJ. Stages; the First To Determine the Genetical and Environmental Model. Most. 1977;38:79–95. Rijsdijk FV, Sham PC. Analytic approaches to twin data using structural equation models. Brief Bioinform. 2002;3(2):119–33. Boker S, Neale M, Maes H, Wilde M, Spiegel M, Brick T, et al. OpenMx: an open source extended structural equation modeling framework. Psychometrika. 2011;76:306–17. Berrie L, Arnold KF, Tomova GD, Gilthorpe MS, Tennant PWG. Depicting deterministic variables within directed acyclic graphs: an aid for identifying and interpreting causal effects involving derived variables and compositional data. Am J Epidemiol. 2025;194(2):469–79. Akingbuwa WA, Hammerschlag AR, Allegrini AG, Sallis H, Kuja-Halkola R, Rimfeld K, et al. Multivariate analyses of molecular genetic associations between childhood psychopathology and adult mood disorders and related traits. Am J Med Genet Part B Neuropsychiatr Genet Off Publ Int Soc Psychiatr Genet. 2023;192(1–2):3–12. Akingbuwa WA, Hammerschlag AR, Jami ES, Allegrini AG, Karhunen V, Sallis H, et al. Genetic Associations Between Childhood Psychopathology and Adult Depression and Associated Traits in 42 998 Individuals: A Meta-analysis. JAMA Psychiatry. 2020;77(7):715–28. Allegrini AG, Baldwin JR, Barkhuizen W, Pingault JB. Research Review: A guide to computing and implementing polygenic scores in developmental research. J Child Psychol Psychiatry. 2022;63(10):1111–24. Pingault JB, Allegrini AG, Odigie T, Frach L, Baldwin JR, Rijsdijk F, et al. Research Review: How to interpret associations between polygenic scores, environmental risks, and phenotypes. J Child Psychol Psychiatry. 2022;63(10):1125–39. Arseneault L. Annual Research Review: The persistent and pervasive impact of being bullied in childhood and adolescence: implications for policy and practice. J Child Psychol Psychiatry. 2018;59(4):405–21. Arseneault L. The long-term impact of bullying victimization on mental health. World Psychiatry. 2017;16(1):27–8. Soneson E, Puntis S, Chapman N, Mansfield KL, Jones PB, Fazel M. Happier during lockdown: a descriptive analysis of self-reported wellbeing in 17,000 UK school students during Covid-19 lockdown. Eur Child Adolesc Psychiatry. 2023;32(6):1131–46. Elective home education, Autumn term 2024/25 [Internet]. [cited 2024 Dec 29]. Available from: https://explore-education-statistics.service.gov.uk/find-statistics/elective-home-education/2024-25-autumn-term Twins Early Development Study (TEDS): A genetically sensitive investigation of mental health outcomes in the mid-twenties - Lockhart – 2023 - JCPP Advances - Wiley Online Library [Internet]. [cited 2025 Jan 1]. Available from: https://acamh.onlinelibrary.wiley.com/doi/full/ 10.1002/jcv2.12154 Additional Declarations The authors have declared there is NO conflict of interest to disclose Supplementary Files 20260117SchoolEPSOM.docx Supplementary Material Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8626819","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":578898251,"identity":"0e8ba260-993e-4822-9bac-76aec290691e","order_by":0,"name":"Kaili Rimfeld","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/UlEQVRIie3RsUoDMRzH8V8IpEuut6ZU7BMIV4RbfZW/FM7ltoLc4HAu6dIHcKi+g0vnlINMom+gza5wkzja9Cy45OzokO904fjw/4cAsdg/TIHdGgIkuGFGVoRh94P6SN0RQTDyiSD+JoDZf3mS6CPIaNHUZovXkzPNjfl4KCYC3LVMF0Eylpd+sbnMraDN/bqcaohzxXQZJKfYE5L52zJrknVFu8VyMF2FSep+iE3bJll5MvjsJWN1mGIlmqQud0T6KeHFRnd+SuaJyDYrW0w1l3NFz+Hrq5cr574qusgtd9v3m9kkHSwe2/Z6FiRd2e8DR9+rxGKxWOyYvgFep1GrO2vVEAAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0001-5139-065X","institution":"Social, Genetic and Developmental Psychiatry Centre Institute of Psychiatry, Psychology and Neuroscience, King's College London, UK","correspondingAuthor":true,"prefix":"","firstName":"Kaili","middleName":"","lastName":"Rimfeld","suffix":""},{"id":578898252,"identity":"9adc33b5-bd78-4412-ac06-04f6f93d3003","order_by":1,"name":"Rebecca Ferdinand","email":"","orcid":"https://orcid.org/0009-0008-5491-922X","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Rebecca","middleName":"","lastName":"Ferdinand","suffix":""},{"id":578898253,"identity":"ed9d76eb-04bc-47a7-83d4-ed9b5fcb9b08","order_by":2,"name":"Anna Suarez","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Suarez","suffix":""},{"id":578898254,"identity":"de96b65e-2029-4b7e-a0c9-d9ca37b1ac75","order_by":3,"name":"Agnieszka Musial","email":"","orcid":"https://orcid.org/0000-0002-3400-4638","institution":"Queen Mary University od London","correspondingAuthor":false,"prefix":"","firstName":"Agnieszka","middleName":"","lastName":"Musial","suffix":""},{"id":578898255,"identity":"44ea2469-1980-42b1-bd48-96abb65bbc30","order_by":4,"name":"Jessica Deighton","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jessica","middleName":"","lastName":"Deighton","suffix":""},{"id":578898256,"identity":"953bce9e-550e-48ee-b156-9600ef659a89","order_by":5,"name":"Essi Viding","email":"","orcid":"","institution":"University College London","correspondingAuthor":false,"prefix":"","firstName":"Essi","middleName":"","lastName":"Viding","suffix":""},{"id":578898257,"identity":"d1c93259-a459-4829-b83a-2caf35287550","order_by":6,"name":"Margherita Malanchini","email":"","orcid":"https://orcid.org/0000-0002-7257-6119","institution":"Queen Mary University of London","correspondingAuthor":false,"prefix":"","firstName":"Margherita","middleName":"","lastName":"Malanchini","suffix":""},{"id":578898258,"identity":"710d46e6-f993-4d90-94ce-9b07e5256823","order_by":7,"name":"Dawn Watling","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Dawn","middleName":"","lastName":"Watling","suffix":""},{"id":578898259,"identity":"217e6df9-bf42-46f1-a7cf-c21c4b28e8ea","order_by":8,"name":"Robert Plomin","email":"","orcid":"https://orcid.org/0000-0002-0756-3629","institution":"Institute of Psychiatry, Psychology and Neuroscience, King's College London, London","correspondingAuthor":false,"prefix":"","firstName":"Robert","middleName":"","lastName":"Plomin","suffix":""},{"id":578898260,"identity":"2d992621-f88e-48b5-9669-7b4fa681cf6f","order_by":9,"name":"Kathleen Rastle","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Kathleen","middleName":"","lastName":"Rastle","suffix":""}],"badges":[],"createdAt":"2026-01-17 14:55:48","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8626819/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8626819/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":100951829,"identity":"24ccfb39-307d-4d94-bee0-0ccff63f80f6","added_by":"auto","created_at":"2026-01-23 07:11:19","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1612529,"visible":true,"origin":"","legend":"","description":"","filename":"20260108SchoolEPms.docx","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/2bdb25029215b85926cfdc49.docx"},{"id":100937053,"identity":"37fb50e0-1156-404e-8bb6-8c174a734cf7","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11702,"visible":true,"origin":"","legend":"","description":"","filename":"2026MP000148.json","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/5ecdf4b9bffb703afe40ede8.json"},{"id":100937059,"identity":"e6473a48-f744-403b-bf72-0aa19000510e","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":221351,"visible":true,"origin":"","legend":"","description":"","filename":"20260117SchoolEPSOM.docx","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/086476a2761c7dc22cfe29fb.docx"},{"id":100937067,"identity":"c956a13c-5169-4db7-923c-78dd72509e4d","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":192426,"visible":true,"origin":"","legend":"","description":"","filename":"2026MP0001480enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/6688dcfe24dd312907413baa.xml"},{"id":100937069,"identity":"6a26de75-f27c-499d-aafc-bba4fc256553","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":589273,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/ec53a2589b82a36e3a63db5e.png"},{"id":100937057,"identity":"2c1b002b-ef3e-4bb1-9e9a-a53a6e9e1504","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":391463,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/515dd663a5385950fb2c0d01.png"},{"id":100937063,"identity":"96dd3491-543b-42a6-b98b-f5668e9988b8","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":143021,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/ca32bf72647f9d2b4113173f.png"},{"id":100951263,"identity":"3cb1a943-ea60-4529-aa38-9f1b19e189b3","added_by":"auto","created_at":"2026-01-23 07:10:21","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":76398,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/454df0d3c795ba6ec47a86df.png"},{"id":100937066,"identity":"4c43698b-b60f-4272-a622-d63cb624085e","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":80177,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/6429443d3f8f226f54122270.png"},{"id":100952453,"identity":"f9ac709f-f1ff-416b-b7e2-6b87cb9c5d85","added_by":"auto","created_at":"2026-01-23 07:16:38","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23132,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/6639728ddac268857d78a651.png"},{"id":100937068,"identity":"5225b1d0-de1e-446e-bc6f-89b661530047","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17557,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/0b100c9260b0f1079d13fb43.png"},{"id":100937062,"identity":"e8a1987d-345f-4dc1-996b-baac27e9147f","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16470,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/eacd84cbdd90d7b819cd6991.png"},{"id":100937070,"identity":"4e0ec35d-d482-4e53-a2fb-fb6ccee40936","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"xml","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":189158,"visible":true,"origin":"","legend":"","description":"","filename":"2026MP0001480structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/e1a2a5eefeb29695ad1c64c6.xml"},{"id":100950769,"identity":"175245a3-2d36-4b13-82fb-ce53118b53ff","added_by":"auto","created_at":"2026-01-23 07:09:11","extension":"html","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":207178,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/e2293240571a089e54c11578.html"},{"id":100937052,"identity":"d2b147ec-6e75-486b-979e-f72ca909179e","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":281757,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMeasures used across ages and raters: (a) Mental health; (b) Perceived school environment (see Appendix S1 for full details)\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/75bf2f835c5bfc6b90167c4a.png"},{"id":100951446,"identity":"f2f74ce6-896f-41b2-bcaf-94236dfb8299","added_by":"auto","created_at":"2026-01-23 07:10:36","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":293624,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eContemporaneous correlations between perceived school experiences and mental health at ages 7, 9 and 16.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/a931e60bf21d68848371be67.png"},{"id":100937055,"identity":"da7701b1-f08f-4349-aab1-930d085468d0","added_by":"auto","created_at":"2026-01-23 03:20:28","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":80984,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdjusted R² for mental health explained by educational experiences across six models\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/b76125562d00f4df33e64423.png"},{"id":100951525,"identity":"6620f793-7c74-4f58-bf5b-85c2ff213cf7","added_by":"auto","created_at":"2026-01-23 07:10:45","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":42044,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGenetic (A), shared environmental (C) and non-shared environmental (E) components of variance for the p factor and PES.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/3e2789b3b81fd79d9bfe5ce8.png"},{"id":100951294,"identity":"e55523aa-14df-45ce-bdc9-099e146a2501","added_by":"auto","created_at":"2026-01-23 07:10:26","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":50819,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBivariate genetic analyses illustrating the proportion of phenotypic variance explained by genetic (rA), shared environmental (rC), and non-shared environmental (rE) factors. The length of the bar represents the total phenotypic correlation (rP).\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/a857e07f19dee343465f9d26.png"},{"id":103503949,"identity":"1b93cdd7-e207-4311-8f88-d674cd843c3b","added_by":"auto","created_at":"2026-02-26 13:05:26","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1588980,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/80b00b91-2465-44dd-a187-7694b9caefd8.pdf"},{"id":100951966,"identity":"0cf3de49-baf9-4c22-9a4c-57aaa39a34b5","added_by":"auto","created_at":"2026-01-23 07:11:36","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":221351,"visible":true,"origin":"","legend":"Supplementary Material","description":"","filename":"20260117SchoolEPSOM.docx","url":"https://assets-eu.researchsquare.com/files/rs-8626819/v1/b3012f5e2639a6101e79156d.docx"}],"financialInterests":"The authors have declared there is \u003cb\u003eNO\u003c/b\u003e conflict of interest to disclose","formattedTitle":"Longitudinal associations between school environment and mental health from childhood through early adulthood","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe rates of reported mental health problems have been rising globally. According to the latest World Mental Health report, approximately one in seven people worldwide lives with a mental disorder (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). In the UK, mental health issues accounted for 7% of all ill health based on the disability-adjusted life years (DALY) measure in 2019 (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). This burden is not only personal but also societal, with recent estimates placing the annual cost of mental ill health in the UK at around \u0026pound;300\u0026nbsp;billion in 2022, more than double earlier estimates (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). In England, among 17\u0026ndash;19-year-olds, the prevalence of a probable mental disorder increased from 10% (2017) to 23% (2023) (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). This observed rise in the prevalence of mental disorders among children and adolescents has generated significant concerns. Although it has been speculated that some of this rise may be due to awareness efforts leading to some individuals interpreting milder forms of distress as mental health problems, it is also likely that this trend partly reflects heightened awareness and reduced stigma, appropriately leading to more frequent help-seeking by children and their guardians (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Substantial evidence shows that mental health problems often begin early, typically emerging by adolescence (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Furthermore, studies indicate that the earlier symptoms of mental disorders appear, the more severe the outcomes tend to be (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). According to NHS Digital\u0026rsquo;s Mental Health of Children and Young People in England 2020 survey, one in six children aged 5 to 16 showed identifiable mental health problems (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). Since common mental disorders are classified as extremes of quantitative traits within a population (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e), many children with poor mental health may not meet clinical criteria for a formal diagnosis despite their real suffering, and therefore, are not reflected in these concerning statistics. Population-based estimates indicate that two in five children score above the threshold for hyperactivity, conduct problems, or emotional difficulties (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). However, UK mental health services are struggling, and even children at the extremes of mental health dimensions are not receiving adequate help: only one in four children in need of mental health services in 2020 were able to access support (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). The situation is even more critical for outpatient services, with 81% of providers failing to meet current demands (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Early intervention and prevention are thus essential. Nevertheless, effective, evidence-based interventions remain elusive, and universal approaches, such as school-focused mindfulness training, have not been effective and may, in some cases, be harmful (\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAlthough many factors influence children's and young people's psychological wellbeing, increased academic demands may have significantly worsened mental health risks (\u003cspan additionalcitationids=\"CR19\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e). Children are increasingly concerned about schoolwork; the OECD reports that approximately 59% of children are concerned about the tests they face, and around 66% worry about receiving poor grades (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). \u003cem\u003eChildline\u003c/em\u003e, a free counselling service for children in the UK, reported that one of the main concerns among children was stress and anxiety related to schoolwork and exam performance (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). These educational pressures and associated mental health issues may have long-term negative effects on children and young people (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e), their families (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e), and society (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e). Importantly, school environments may not influence all children the same way; those from disadvantaged socioeconomic backgrounds may be especially vulnerable, with socioeconomic status consistently associated with both academic and mental health outcomes (\u003cspan additionalcitationids=\"CR28\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). However, a recent meta-analysis revealed that while academic pressures are linked to mental health in cross-sectional studies, there is no longitudinal evidence connecting mental health and academic pressures (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). This evidence highlights the urgent need to understand the underlying factors behind the increase in mental health problems (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). Identifying preventable risk factors for poor mental health in childhood and adolescence is essential for developing effective interventions that could make a difference.\u003c/p\u003e \u003cp\u003eResearchers have long sought these modifiable risk factors within children's homes or neighbourhood environments (\u003cspan additionalcitationids=\"CR32 CR33\" citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). However, the influence of school environments on mental health has received far less attention, even though children spend over 15,000 hours of their lives in full-time education (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e). Schools are often viewed as promising settings for early identification and support. Yet, young people and families frequently report barriers to accessing help, and the quality of support in schools varies greatly (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Moreover, school life can itself be a source of stress, with research showing marked variation in school practices and limited evidence on which approaches best promote mental wellbeing (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e). Although meta-analytic and systematic reviews link selected aspects of the school environment, such as school climate and connectedness, to internalising problems (\u003cspan additionalcitationids=\"CR40\" citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e), these studies typically focus on individual elements of the school environment in isolation. According to the Good Childhood Report by The Children's Society, which has tracked children's well-being in the UK for many years, children are increasingly reporting unhappiness with school and schoolwork (12% of children aged 10\u0026ndash;15), and this trend is more pronounced among older children. Among the various aspects of school life, children were most satisfied with a sense of safety and their relationships with peers, whereas they were least satisfied with the amount of schoolwork and the level of attention they received from their teachers (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). However, although there is cross-sectional evidence for links between the school environment and mental health, longitudinal research remains limited, and most existing studies focus on narrow facets of the school experience rather than its broader, cumulative effects (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e). To address this gap, a more comprehensive understanding of perceived school environment across the developmental spectrum is needed.\u003c/p\u003e \u003cp\u003eEven less is known about the links between school environment and mental health using genetically sensitive designs. Yet, we know from decades of twin studies that individual differences in mental health outcomes are partly explained by genetic factors, with heritability estimates ranging between 30 and 80% (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). Furthermore, both objective (e.g. neighbourhood characteristics) and subjective (e.g. perceived noise at home) environmental experiences have been shown to be partly heritable, albeit with modest heritability estimates (\u003cspan additionalcitationids=\"CR46 CR47 CR48\" citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e). This means that genetic differences not only influence mental health outcomes but could also shape how children perceive and interact with their environments. The home and school environment may be particularly susceptible to these genetic influences (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). Importantly, even within the same settings, children may experience their environments differently, partly due to their genetic predispositions. For example, when growing up in the same family (home environment) or attending the same school (school environment), children often report these environmental experiences differently (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e). Furthermore, as children grow older, their environments become increasingly dynamic and self-directed, with children selecting, modifying, and creating their environments in ways influenced by their genetically driven behaviours and traits (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). Therefore, it is important to consider genetic factors when studying the associations between environmental factors (school environment) and life outcomes (mental health).\u003c/p\u003e \u003cp\u003eHere, we use longitudinal data on self-, parent- and teacher-reported educational environment and mental health, capitalising on the rich data collected from the Twins Early Development Study (TEDS) over two decades. The differential resemblance between identical and fraternal twins enables investigation of the aetiology of associations between perceived school environment and mental health outcomes during childhood and early adulthood. It allows us to test whether the association between school environment and mental health arises from shared genetic factors, shared environments, or unique individual experiences. In addition, it is now possible to calculate genome-wide polygenic scores (PGS) that leverage summary statistics from genome-wide association (GWA) studies to aggregate the effects of single DNA variants into a single index to predict individual-specific propensities for mental health (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e). While twin analyses quantify the overall genetic and environmental contributions, PGS analyses allow us to partially adjust for measured genetic liability when estimating the association between perceived school environment and mental health. We will use PGS to control for genetic confounding when studying the associations between educational environment and mental health. Additionally, we will control for baseline mental health outcomes (mental health prior to starting school) and family socioeconomic status. In our final step, we further control for mental health outcomes in previous measurements, e.g., for age 21 outcomes, by controlling for mental health at age 16. Our research plan was preregistered in the Open Science Framework (OSF; \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://osf.io/3c5rk/\u003c/span\u003e\u003cspan address=\"https://osf.io/3c5rk/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eParticipants\u003c/h2\u003e \u003cp\u003eParticipants come from the Twins Early Development Study (TEDS), which comprises 16,810 twin pairs born in England and Wales between 1994 and 1996, who have been assessed in multiple waves across their development from approximately 18 months to the present. Although there has been some attrition, more than 10,000 twin pairs remain actively involved in the study. The demographic characteristics of TEDS participants and their families are reasonably comparable to those of the population in England and Wales for this birth cohort (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e). Here, we use data from a sample of twins with available data on school environment and mental health, N\u0026thinsp;=\u0026thinsp;6,500. The sample size varies across measures (See Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e-S2).\u003c/p\u003e \u003cp\u003e Written informed consent was obtained from parents prior to data collection and from TEDS participants themselves, once they were over the age of 18. King's College London's Ethics Committee approved the project for the Institute of Psychiatry, Psychology and Neuroscience PNM/09/10\u0026ndash;104.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMeasures\u003c/h3\u003e\n\u003cp\u003eThe TEDS cohort provides rich data on educational experiences and emotional and behavioural problems from childhood to early adulthood. In the present study, we utilise the data collected at ages 7, 9, 16 and 21.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMental health.\u003c/b\u003e Emotional and behavioural problems were evaluated using a battery of questionnaires completed by the parents, teachers, and the twins themselves (Fig.\u0026nbsp;1(a), Appendix S1(a)). We further calculated a p-factor, i.e., a general factor for psychopathology, for each age group combining all raters, using confirmatory factor analyses (CFA). Here, we were not interested in the structure of mental health problems across childhood, but used CFA as a data reduction technique to derive a general factor of psychopathology, or p-factor (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e), across raters (parent, teacher, and twin reports) to index mental health problems at every age. Factor loadings and model fit statistics are presented in Supplementary Table S3. As a sensitivity analysis, we also computed separate externalising and internalising factors and only the self-reported factor of internalising problems (see Supplementary Tables S4-S6 for factor loadings and model fit statistics).\u003c/p\u003e \u003cp\u003e \u003cb\u003ePerceived school environment.\u003c/b\u003e The survey included multiple measures of the school environment, such as classroom environment, homework, and relationships with teachers and peers, which were rated by parents, teachers, and the twins themselves (Fig.\u0026nbsp;1(b), Appendix S1(b)). Educational experiences were collected across compulsory education (age 7\u0026ndash;16), but the measures differ across data collection waves. To evaluate the overall effects of the school environment, we calculated poly-environmental scores (PES) at each age using penalised regression elastic net regularisation, with a hold-out sample to test prediction accuracy, following the procedure of Gidziela et al. (2023) (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e). This method enables us to overcome the problems of multicollinearity and overfitting (\u003cspan additionalcitationids=\"CR56\" citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e). For details on the construction of PES, see Appendix S2.\u003c/p\u003e \u003cp\u003e \u003cb\u003eSocioeconomic status (SES).\u003c/b\u003e Family SES measures were collected at the first contact when the twins were about 18 months old. The composite measure was computed as a mean of the mother\u0026rsquo;s and father\u0026rsquo;s qualification level, the mean of the occupational status, and the mother's age at first birth. SES at first contact was used, as this has the largest sample size. SES is highly stable across development: SES at first contact correlates .77 with SES at age 7 and .57 with SES at age 9, when family income was added to the composite (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cb\u003eFigure 1. Measures used across ages and raters: (a) Mental health; (b) Perceived school environment (see Appendix S1 for full details)\u003c/b\u003e \u003c/p\u003e \u003cp\u003e(a)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e(b)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eGenotyping\u003c/h3\u003e\n\u003cp\u003eGenotyping was performed using one of two DNA microarrays (Affymetrix GeneChip 6.0 or Illumina HumanOmniExpressExome chips) according to standard protocols. DNA samples were collected from 12,500 individuals in the TEDS sample. Following rigorous quality control procedures, 10,346 samples were available for genomic analysis. Of these, 7,026 were unrelated individuals, and 3,320 to additional dizygotic co-twins. The study genotyped 7,289 individuals using Illumina arrays and 3,057 individuals using Affymetrix arrays. Further information on genotyping and quality control is available elsewhere (\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e, \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003ePolygenic risk scores\u003c/h3\u003e\n\u003cp\u003eGenome-wide polygenic scores (PGS) were calculated to estimate genetic vulnerabilities for psychiatric problems using LDpred1 (\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e). A detailed description of this method and analytic strategies used for the TEDS sample is presented in earlier work (\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e). To construct the PGS, we used the GWA summary statistics for 27 adult psychiatric disorders: ADHD (\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e), anxiety (\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e), anxious feeling / worrier (\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e), ASD (\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e), bipolar disorder (\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e), cross-disorder (\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e), daytime sleepiness (\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e), depression (\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e), depressive symptoms (\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e), excessive daytime sleepiness (\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e), externalising problems (\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e), fed-up feeling (\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e), guilty feeling (\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e), insomnia (\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e), miserableness (\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e), mood disorder spectrum (\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e), mood swings (\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e), nervous feelings (\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e), neuroticism (\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e), obsessive-compulsive disorder (\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e), posttraumatic stress disorder (\u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e), risk-taking behaviour (\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e), risk tolerance (\u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e80\u003c/span\u003e), schizophrenia (\u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e), sensitivity / hurt feelings (\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e), sleep disorders (\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e), and wellbeing (\u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e83\u003c/span\u003e). We used all available PGS at TEDS related to mental health problems.\u003c/p\u003e\n\u003ch3\u003eStatistical analyses\u003c/h3\u003e\n\u003cp\u003eThe analyses were performed in R version 4.0 (\u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e84\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eDescriptive statistics and data preparation.\u003c/span\u003e We conducted univariate analyses of variance (ANOVA) to explore phenotypic sex differences across measures. Because significant sex differences emerged, the variables were adjusted for sex and age using linear regression. Sex- and age-adjusted standardised variables were used in all downstream analyses.\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eGeneral factor of psychopathology (p-factor)\u003c/span\u003e. We calculated a p-factor, i.e., a general factor for psychopathology, for each age group, using confirmatory factor analyses (CFA; See Measures).\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ePhenotypic correlations\u003c/span\u003e. Pearson correlations were used to describe the phenotypic correlations between educational experiences and mental health during compulsory education. Here, we used the p-factor (see Measures) as an indicator of childhood mental health problems.\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eContemporaneous and longitudinal models to predict mental health outcomes (p-factor).\u003c/span\u003e We next sought to examine how well the combined perceived school environmental measures explain mental health outcomes across different developmental stages, both at the time when these perceptions were measured and over time. We used multiple regression and prediction modelling (elastic net) to determine the variance explained by school environment in mental health contemporaneously (Model 1) and over time (Model 2). In the subsequent models (Models 3\u0026ndash;6), we adopted a conservative approach and controlled for potential confounders: genetics, baseline mental health problems (age 7), and family SES. We also included mental health at the previous age of measurement. However, we acknowledge that this may overcontrol for the effects of the perceived school environment on mental health, particularly as we had already controlled for genetic predisposition to mental health and mental health problems at age 7.\u003c/p\u003e \u003cp\u003eIn Model 3, we controlled for potential genetic confounding in predictions using 27 adult psychiatric polygenic scores (PGS); in Model 4, we additionally controlled for mental health at the start of compulsory education (p-factor at age 7). In Model 5, we added control for family socioeconomic status (SES). Finally, in fully adjusted Model 6, we also controlled for mental health (p-factor) from the previous measurement (e.g. at age 16, we also controlled for mental health at age 9)\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eTwin analyses.\u003c/span\u003e We investigated the aetiology of perceived school environment, measured by the poly-environmental score (PES; see Measures), and mental health, measured by the general factor of psychopathology (p-factor; see Methods), using the univariate twin method. We also examined the aetiology of the relationships between PES and p-factor using bivariate Cholesky decomposition.\u003c/p\u003e \u003cp\u003eWe employed a twin model to estimate the proportions of individual differences in PES and p-factor explained by genetic, shared environmental, and non-shared environmental factors. Twin methods capitalise on the genetic relatedness between monozygotic (MZ) or identical twins and dizygotic (DZ) or non-identical twins. MZ twins share 100% of their genes, while DZ twins share, on average, 50% of their segregating genes. However, both types of twins share their rearing environment. By comparing the correlations between MZ and DZ twins, it is possible to estimate the relative contributions of genetic, shared environmental, and non-shared environmental effects to individual differences in the trait of interest. Heritability (A), the proportion of variance explained by genetic factors, can be calculated by doubling the difference between MZ and DZ correlations. Shared environmental factors (C), which make children growing up in the same family more similar beyond genetic influences, can be estimated by subtracting heritability from the MZ correlation. The remaining variance is attributed to non-shared environmental factors (E), which include environmental influences that do not contribute to similarities between twins raised together. These are calculated by subtracting MZ correlations from 1. Importantly, the E component also encompasses measurement error. The ACE estimates can be refined more accurately, including 95% confidence intervals, using structural equation modelling (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e, \u003cspan citationid=\"CR85\" class=\"CitationRef\"\u003e85\u003c/span\u003e, \u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e86\u003c/span\u003e). The present study employed an OpenMX package in R for all twin analyses (\u003cspan citationid=\"CR87\" class=\"CitationRef\"\u003e87\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBivariate twin analyses were used to calculate the proportion of phenotypic association between p-factor and PES explained by genetic, shared-environmental and non-shared environmental factors. Univariate twin analyses can be extended to bivariate genetic analyses to study the aetiology of covariance between traits. The covariance between traits can be decomposed into additive genetic (A), shared environmental (C), and non-shared environmental (E) components by comparing cross-twin cross-trait correlations between MZ and DZ twin pairs. This method also allows for estimating the genetic correlation (\u003cem\u003er\u003c/em\u003eG), which indicates the extent to which the same genetic factors influence both measures. Shared environmental correlation (\u003cem\u003er\u003c/em\u003eC) and non-shared environmental correlation (\u003cem\u003er\u003c/em\u003eE) can be estimated in the same manner (\u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e86\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ePolygenic score analyses to investigate rGE.\u003c/span\u003e PGS provides an index of genetic predisposition to psychiatric problems. We explored how this predisposition is associated with how parents, teachers, and children themselves perceive the school environment, hypothesising that these indicators are not independent. Because adult PGS only explain a small proportion of variance in childhood mental health problems, we used factor analyses to assess how a latent factor of PGS relates to perceived school environment. We calculated a general genomic p factor, using a similar method to the phenotypic p factor, based on 27 adult psychiatric polygenic scores. Then, we used Pearson correlations to describe the phenotypic relationships between perceived school environment and genetic vulnerability to psychiatric disorders.\u003c/p\u003e \u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eSensitivity analyses\u003c/span\u003e. We conducted a series of sensitivity analyses to assess the robustness, specificity and sensitivity of our findings. First, to examine whether the school environment relates differently to different aspects of mental health, we repeated the analyses using the internalising or externalising factor, and only self-reported internalising factor scores, rather than the p-factor, to assess whether the school environment affects these factors differently. We also ran models using only self-reported internalising scores, as young people may be better placed to report on their own internal states, such as worry or sadness, than parents or teachers. These checks enabled us to determine whether our findings were specific to a single report, thus explained by a reporter bias or reflect a broader pattern.\u003c/p\u003e \u003cp\u003eSecondly, because the covariate-adjusted models had significant missing data, we conducted a simpler analysis. We examined the relationship between the perceived school environment during compulsory education and p-factor scores at age 21, approximately five years after completing compulsory education. We repeated this analysis with and without controlling for family SES to determine whether SES or the modelling approach accounted for the associations.\u003c/p\u003e \u003cp\u003eThirdly, because measures of school environment and mental health can overlap or influence each other, it is important to check whether observed associations are genuine or partly reflect the way the data were collected. This is sometimes referred to as a tautological association, where overlapping content contributes to the observed effects (\u003cspan citationid=\"CR88\" class=\"CitationRef\"\u003e88\u003c/span\u003e). This is especially relevant when both the exposure and outcome rely on self-report, as mental health difficulties may affect how young people perceive and evaluate their school environment. To address this, we conducted sensitivity analyses using teacher-rated mental health outcomes and school environment measures reported by parents and children. We also examined the relationship between teacher-rated school environment and mental health as reported by parents and children. These additional analyses allowed us to assess whether the patterns were consistent across different informants and were not merely due to shared method variance.\u003c/p\u003e \u003cp\u003eWe applied the Benjamini\u0026ndash;Hochberg false discovery rate (FDR) procedure to control for multiple testing across all analyses.\u003c/p\u003e \u003cp\u003eAll scripts are available on the OSF site (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://osf.io/3c5rk/\u003c/span\u003e\u003cspan address=\"https://osf.io/3c5rk/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eDescriptive statistics and data preparation.\u003c/span\u003e Descriptive statistics (means, standard deviations, skew and kurtosis) were calculated for all perceived school environment and mental health measures for the whole sample and for males and females separately. One twin per pair was randomly selected for all phenotypic analyses to maintain the independence of the data. Analyses of variance (ANOVA) were used to test the significance of sex differences. ANOVA results revealed some gender differences, although on average, gender explained less than 1% of the variance in all measures. The gender differences were larger for some measures; for example, gender differences explained 8% of the variance in prosocial behaviour at age 16. Some variables showed negative and some positive skew (skewness over +/- 1); these variables were transformed using a log transformation. All variables were also corrected for mean sex and age differences, as described in the Methods section. The standardised variables are used in all downstream analyses. Descriptive statistics are presented in Supplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e. Descriptive statistics for males and females separately, and the ANOVA results are presented in Supplementary Table S2.\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ePhenotypic correlations\u003c/span\u003e. Contemporaneous correlations across all phenotypic measures are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003e. All perceived school environment measures, regardless of the raters, were associated with mental health reported across raters, as indicated by the general factor of psychopathology (with an average absolute correlation of 0.19). The strongest associations were with teacher-reported classroom environment (r\u0026thinsp;=\u0026thinsp;0.47) and parent-reported feelings about being accepted in the classroom (r\u0026thinsp;=\u0026thinsp;0.45), whereas the weakest were for being the first to finish work (r\u0026thinsp;=\u0026thinsp;0.02) and the first to answer (r\u0026thinsp;=\u0026thinsp;0.01). These correlations range from positive (protective factors) to negative (harmful factors). Emotional distress (negative affect) and disruptive classroom environment (noise, disorganisation, long waiting times) were consistently linked to higher levels of mental disorders. See Supplementary Table S7 for full statistics, including 95% confidence intervals. Correlations between perceived school experiences and the internalising factor are presented in Supplementary Table S8, with the externalising factor in Supplementary Table S9 and the self-reported internalising factor in Supplementary Table S10.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eContemporaneous and longitudinal models to predict mental health outcomes (p-factor).\u003c/span\u003e We examined the extent to which perceived school environment explained variance in mental health, both contemporaneously and longitudinally. We employed structural equation modelling (SEM) with full-information maximum likelihood (FIML) as the primary approach, as this method retains participants with partial data and provides robust handling of missing values. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e shows the results, with full estimates reported in Supplementary Table S11. To assess robustness, we compared these findings with elastic net models, which use cross-validation to limit overfitting (Supplementary Table S12), and with linear regression (LM) using complete cases (Supplementary Table S13). The results were consistent across methods. Variance explained (R\u0026sup2;/ΔR\u0026sup2;) was very similar in contemporaneous and additive models across SEM, LM, and elastic net methodological approaches. In models that included covariates (Models 4\u0026ndash;6), SEM produced estimates consistent with those of LM and elastic net, although the sample sizes were smaller for LM/elastic net due to listwise deletion. Small differences across methods, therefore, likely reflect sample size rather than modelling strategy.\u003c/p\u003e \u003cp\u003eWe ran a series of models, as presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Model 1 shows contemporaneous variance explained, while Model 2 illustrates longitudinal associations and includes educational perceived school environment from earlier waves (for example, at age 16, it incorporates all perceived school environment collected at ages 7, 9, and 16). We show that perceived educational experiences explain substantial variance in mental health outcomes across development, explaining 26% of variance in mental health outcomes at age 21, years after finishing the compulsory education when these perceived measures were collected.\u003c/p\u003e \u003cp\u003eWe then sequentially adjusted for potential confounders. First, we added polygenic scores (PGS) for 27 psychiatric traits to account for genetic predisposition to adverse mental health (Model 3). Next, we incorporated baseline mental health at age 7 (Model 4), followed by family socioeconomic status (SES) (Model 5). Finally, we also adjusted for mental health at the previous wave (e.g., age 9 when predicting outcomes at age 16; Model 6). We show that perceived school environment is associated with mental health both contemporarily and over time; these remain substantial and significant even after controlling for possible confounders (explaining 6\u0026ndash;30% of variance).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eNote\u003c/strong\u003e \u003cp\u003e \u003cem\u003eModel 1 shows contemporaneous variance explained. Model 2 adds longitudinal predictors from earlier waves (for example, at age 16, models also perceived school environment from ages 7 and 9). Model 3 adjusts for polygenic scores for 27 psychiatric traits. Model 4 adds baseline mental health at age 7. Model 5 adds family socioeconomic status. Model 6 further adjusts for mental health at the previous wave\u003c/em\u003e \u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eTwin analyses.\u003c/span\u003e We calculated the poly-environmental score (PES) using the beta values from the elastic net models. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003e shows the aetiology of p-factor at each age as well as the aetiology of school-related PES (see Supplementary Table S14 for full details). As expected, p-factor was highly heritable (average h\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;61%), with a small proportion of variance also explained by shared environmental factors (average c\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;13%). The remaining variance is attributed to non-shared environmental factors, which also includes the measurement error. Interestingly, the PES also showed substantial heritability (average h\u0026sup2;= 46%), with a small proportion of variance also explained by shared environmental factors (average c\u0026sup2; = 16%), which was highest at age 7, when the PES included only parent-rated environmental measures.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eBivariate analyses revealed that the correlations between the p and PES were primarily explained by genetic factors (averaging 71%), with the remainder of the variance accounted for by shared environmental (averaging 18%) and non-shared environmental factors (averaging 11%), as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e5\u003c/span\u003e (see Supplementary Table S15 for full bivariate model-fitting results). This means that most of the overlap between children\u0026rsquo;s mental health and their school environment reflects shared genetic influences, with a smaller role for environmental factors.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ePolygenic score analyses to investigate rGE.\u003c/span\u003e The correlations between genomic p and perceived school environment were small in magnitude (average correlation \u0026asymp; |0.04|). Although only a handful of associations reached nominal significance and only two survived corrections for multiple testing, this remains noteworthy given that polygenic scores typically account for a small proportion of the variance in childhood psychopathology. These findings provide further evidence of the association between perceived school environment and genetic factors (see Supplementary Table S16).\u003c/p\u003e \u003cp\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eSensitivity analyses\u003c/span\u003e. As a sensitivity analysis, we examined the associations between perceived school environment and internalising, externalising, and self-reported internalising factors separately (see Supplementary Tables S8-S10). The patterns were generally similar, although the strength of the associations varied, with internalising symptoms more strongly linked to emotional aspects of school (e.g., parent-rated acceptance in the class, r = \u0026minus;\u0026thinsp;0.54) and externalising symptoms more associated with behavioural factors (e.g., teacher-reported homework completion, r = \u0026minus;\u0026thinsp;0.40). Associations were generally weaker for self-reported internalising, especially for teacher- and parent-rated school environment. Since the p-factor reflects shared variance across internalising and externalising symptoms and combines data from different raters, we use the p-factor as the primary outcome in all other analyses.\u003c/p\u003e \u003cp\u003eMissing data are a concern, especially when conducting longitudinal analysis. Data were missing due to attrition, as well as at random, since some data collection waves included only two out of four cohorts. In addition, only a subsample of TEDS participants provided their DNA. We used FIML to handle missing data. As a sensitivity analysis, we tested the associations between school environment and mental health in early adulthood without any intermediate analysis steps (complete cases). Supplementary Table S17 shows these correlations (average absolute correlation 0.12 (range: 0.01\u0026ndash;0.31). We also tested whether the correlations change when controlling for family SES, but found only marginal reductions (see Supplementary Table S18). These checks suggest that our main findings are not driven by the modelling strategy, missing data patterns, or family socio-economic status.\u003c/p\u003e \u003cp\u003eTo check for tautological associations (see Methods), we also examined the associations between teacher-rated mental health outcomes and educational outcomes (parent- and self-rated) (Supplementary Table S19) and parent-rated mental health outcomes and perceived school environment (teacher\u0026ndash; and self-rated) (Supplementary Table S20). While these associations were attenuated, they remained significant and substantial. Although we cannot fully rule out tautology, these findings provide added confidence that the observed patterns reflect meaningful associations across different informants.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe impact of school environments on mental health remains understudied, particularly in genetically informed, longitudinal research. Our study addressed this gap by examining the associations between the perceived school environment and mental health from childhood into early adulthood, utilising data from multiple informants, repeated measures, and genetic data. Using over two decades of TEDS data, our findings show that perceived school environment is consistently associated with mental health across compulsory schooling and into early adulthood. The strongest associations were found between the classroom environment and mental health, with weaker links for more peripheral aspects of school life, such as being first to complete homework. Additionally, the long-term relationships were significant, and our analyses indicate that perceived school environment is associated with mental health outcomes in an additive manner, suggesting that these environmental factors may cumulatively elevate the risk of mental health issues. These associations remained substantial and significant even after accounting for earlier mental health problems, genetic confounding through polygenic scores of adult psychiatric disorders, and family socio-economic status. By controlling for these variables, we provide a conservative estimate of the school environment's influence on mental health, as participants in this study attended school more than a decade ago. Since then, educational demands and pressures have increased, with growing concerns about their effects on student wellbeing (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). It is therefore possible that the associations reported here underestimate the current effects of school environments on young people\u0026rsquo;s mental health.\u003c/p\u003e \u003cp\u003eOur results also showed that gene-environment interplay (rGE) affecting mental health outcomes is widespread in childhood and early adulthood, based on systematic analyses using the twin method and polygenic scores. The perceived environmental measures were significantly heritable, and the association between these perceived school environmental features and mental health was largely explained by genetic factors. This indicates that how pupils, parents and teachers perceive school is partly shaped by children\u0026rsquo;s inherited dispositions, because genetically influenced behaviours shape peer and teacher relationships, as well as home\u0026ndash;school relationships, which in turn influence how adults perceive children\u0026rsquo;s school environment. By contrast, objective school-level features such as inspection ratings or behaviour policies would be the same for twins in the same school, so a twin model would not be able to capture their heritability. Consistent with rGE, the covariance between perceived school environment and psychopathology was itself substantially explained by genetic factors, indicating that some of the observed association reflects shared genetic influences. We also tested for rGE using a polygenic score approach. Some polygenic scores were associated with school environment, indicating that children\u0026rsquo;s genetic risks for psychiatric disorders are linked to how they perceive the school environment as well as how parents and teachers perceive children\u0026rsquo;s educational experiences, although the effect sizes were small. This suggests passive, active, and evocative rGE. In other words, children inherit not only their parents' genes but also their environment (passive rGE), they actively select and shape their environment (active rGE), and they evoke responses from their environment. For example, an inattentive student may elicit specific reactions from teachers, which are associated with their genetic predispositions (evocative rGE) (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Children with different genetic predispositions may perceive and experience the same school environment differently, supporting a dual approach that combines whole-school improvements with provision tailored to individual needs, recognising variation in how pupils respond to the same context.\u003c/p\u003e \u003cp\u003eOur study design does not allow for causal conclusions, as the methods used were primarily correlational. However, our use of genetically sensitive data and stringent controls enabled us to move closer to causal inference and test whether the observed associations persisted after controlling for major confounders. Although we utilised a large, nationally representative sample and incorporated a range of school environment measures reported by multiple informants across compulsory education, the measures differed at each time point. This variability prevented us from employing longitudinal structural equation modelling, which could have provided stronger evidence for causal directions between perceived school environment and mental health, thereby moving closer to causal inference. We can, however, address the possibility of reverse causation: mental health symptoms at age 21 could not cause the perceived school environment during compulsory education years, especially in adjusted models controlling for mental health symptoms during the school years. We controlled for possible confounders, previous mental health problems, family socioeconomic status, and genetic confounding (polygenic scores), but the associations between school environment and mental health remained significant and substantial. Polygenic scores for adult psychiatric disorders explain a small amount of variance in childhood psychopathology (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e, \u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e, \u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e). However, this prediction increases linearly as children age, as evident from a smaller drop in prediction when using polygenic scores at age 7 compared to age 21. Nevertheless, we acknowledge that using polygenic scores to control for genetic confounding is limited to the predictive power of these genetic indices (\u003cspan citationid=\"CR91\" class=\"CitationRef\"\u003e91\u003c/span\u003e, \u003cspan citationid=\"CR92\" class=\"CitationRef\"\u003e92\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e Our study used subjective measures of educational environment, collected longitudinally from parents, teachers, and twins in the TEDS cohort. Using objective environmental measures is the next step in our research program that aims to establish causal pathways between the educational environment and mental health from childhood to early adulthood. However, a limitation is that participants attended school around 15 years ago, a time when educational pressures were less pronounced, and our measures of the school environment were limited. This likely means our estimates on the associations between school environment and mental health are conservative. Crucially, we were unable to include reports on educational pressures, as these were not collected in TEDS during participants\u0026rsquo; school years. Yet, rising educational pressures are a major concern for today\u0026rsquo;s children, and the longitudinal associations between these pressures and mental health remain neglected in research (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). Rather than focusing on bullying, which has already been extensively documented as a risk factor for poor mental health (\u003cspan citationid=\"CR93\" class=\"CitationRef\"\u003e93\u003c/span\u003e, \u003cspan citationid=\"CR94\" class=\"CitationRef\"\u003e94\u003c/span\u003e), we examined a broad range of other educational environmental measures. By doing so, our estimates of the association between the school environment and mental health are likely to be conservative, as they do not include the established school-related risk factors.\u003c/p\u003e \u003cp\u003eThere is increasing interest in the link between school environments and mental health outcomes. For example, during the COVID-19 lockdown, a study found that one-third of children and young people reported an improvement in their mental wellbeing. These improvements were linked to better management of school tasks, reduced pressure and more opportunities for sleep and exercise, factors directly tied to changes in school-related experiences during lockdown (\u003cspan citationid=\"CR95\" class=\"CitationRef\"\u003e95\u003c/span\u003e). While the study focused on the unique context of the pandemic, these findings highlight the significant influence of school environments on mental health and suggest that certain elements of the school system may be contributing to poor outcomes for some students.\u003c/p\u003e \u003cp\u003eRecent government statistics show changes in how families are engaging with the school system. In 2023, the number of children being home-educated in England increased by 20% to an estimated 111,700 children (\u003cspan citationid=\"CR96\" class=\"CitationRef\"\u003e96\u003c/span\u003e). While 23% of families cited lifestyle or philosophical reasons for this choice, 13% withdrew their children due to dissatisfaction with schools, including inadequate support for special educational needs. These trends highlight potentially important issues within the school environment and suggest that further research is needed (\u003cspan citationid=\"CR96\" class=\"CitationRef\"\u003e96\u003c/span\u003e). Future research should address these gaps by systematically studying the educational environment using longitudinal, genetically sensitive designs. By understanding the role of these factors in shaping mental health outcomes, we can identify and implement evidence-based changes to educational environments, ultimately supporting better mental health and wellbeing for children and young people.\u003c/p\u003e \u003cp\u003eOur study has limitations. We focused only on subjective measures, as previously discussed, and were limited to the measures collected in the TEDS cohort. We used information from teachers, parents and twins; thus, our results may be affected by rater bias. However, different raters were picking up different aspects of the school environment, and although the rater agreement was low (r\u0026thinsp;~\u0026thinsp;.2-.3), combining information about perceived school environment from multiple reporters increased prediction of mental health problems both contemporaneously and over time. Using multiple raters may be beneficial for a more comprehensive understanding of children's educational experiences. We also conducted sensitivity analyses using different combinations of raters. Specifically, we examined how parent-rated school environment related to teacher- and child-rated mental health, and how teacher-rated environment related to parent- and child-rated mental health. Although the strength of associations was slightly attenuated in these analyses, the associations remained, increasing confidence in the robustness of our findings. Our study was based on TEDS data, which was a representative sample of the population in England and Wales on recruitment and remains reasonably representative of the population for their birth cohort (\u003cspan citationid=\"CR97\" class=\"CitationRef\"\u003e97\u003c/span\u003e); however, it does not necessarily reflect the current cohort of schoolchildren in England and Wales. TEDS data primarily consists of white European participants, and the findings may not be applicable to other diverse groups. Using different datasets to replicate our findings is part of our future research program. Another limitation of our study is the issue of missing data, particularly when examining the long-term effects of school environments on mental health outcomes. Neither regression models nor elastic net prediction approaches handle missing data well, which may have affected the robustness of some of our findings. We used FIML in SEM to mitigate missingness issues. Using these different analytical approaches did not affect our results, and all analyses support the view that perceived school environments are important predictors of mental health both during the school years and in later life. Our future systematic research program aims to triangulate between the sample and the study design to establish causal pathways from the educational environment to mental health.\u003c/p\u003e \u003cp\u003eIn summary, we provide a general overview of the associations between perceived school environment and mental health. In our future research program, we aim to incorporate objective measures of the school environment available through government records, such as the National Pupil Database, which provides quantitative indicators of the school environment (e.g., student-teacher ratio, school resources, absences), as well as Ofsted inspection reports. Risk factors for poor mental health are multifactorial, encompassing biological (e.g., genetic factors), psychological (e.g., experiences and attitudes), and environmental (e.g., socioeconomic disadvantage, school environment) factors. There is an urgent need for research that combines these risk domains to better understand the processes affecting children's mental health. Identifying potentially preventable risk factors of poor mental health in childhood and adolescence is vital to building a solid base for interventions that could make a real difference.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eConflict of Interest\u003c/h2\u003e \u003cp\u003eThe authors declare no conflict of interest.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e \u003cp\u003eWe gratefully acknowledge the ongoing contributions of the participants in the Twins Early Development Study (TEDS) and their families. TEDS is supported by a programme grant to R.P. from the UK Medical Research Council (Grant Nos. MR/V012878/1 and previously MR/M021475/1). K.R. is supported by a Sir Henry Wellcome Postdoctoral Fellowship and UK Medical Research Council Grant UKRI1503. E.V. is supported by the National Institutes of Health and Care Research (NIHR) University College London Hospitals Biomedical Research Centre (Award: NIHR 203972). MM is supported by a Jacobs Foundation Research Fellowship (#2024-1533-00) and UK Medical Research Council Grant UKRI1506.\u003c/p\u003e \u003cp\u003eThis research was funded in whole, or in part, by the Wellcome Trust (213514/Z/18/Z). For the purpose of open access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eWorld mental health report: Transforming mental health for all [Internet]. [cited 2022 Dec 11]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/publications-detail-redirect/9789240049338\u003c/span\u003e\u003cspan address=\"https://www.who.int/publications-detail-redirect/9789240049338\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcDaid D, Park AL, Davidson G, John A, Knifton L, Morton A, et al. The economic case for investing in the prevention of mental health conditions in the UK (summary). Ment Health Found. 2022;(February).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThe economic case for investing in the prevention of mental health conditions in the UK [Internet]. [cited 2022 Dec 11]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.mentalhealth.org.uk/explore-mental-health/publications/mental-health-problems-cost-uk-economy-least-ps118-billion-year-new-research\u003c/span\u003e\u003cspan address=\"https://www.mentalhealth.org.uk/explore-mental-health/publications/mental-health-problems-cost-uk-economy-least-ps118-billion-year-new-research\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThe economic and social costs of mental ill health: review of methodology and update of calculations [Internet]. Institute for Social and Economic Research (ISER). [cited 2025 Sep 23]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.iser.essex.ac.uk/research/publications/publication-578405\u003c/span\u003e\u003cspan address=\"https://www.iser.essex.ac.uk/research/publications/publication-578405\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNHS England Digital [Internet]. [cited 2025 Aug 1]. Mental Health of Children and Young People in England 2022 - wave 3 follow up to the 2017 survey. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2022-follow-up-to-the-2017-survey\u003c/span\u003e\u003cspan address=\"https://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2022-follow-up-to-the-2017-survey\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeighton J, Lereya ST, Casey P, Patalay P, Humphrey N, Wolpert M. Prevalence of mental health problems in schools: poverty and other risk factors among 28 000 adolescents in England. Br J Psychiatry. 2019;215(3):565\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNHS England Digital [Internet]. [cited 2024 Dec 29]. Mental Health of Children and Young People in England, 2023 - wave 4 follow up to the 2017 survey. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2023-wave-4-follow-up\u003c/span\u003e\u003cspan address=\"https://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2023-wave-4-follow-up\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry. 2005;62(6):593\u0026ndash;602.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTanner JL. Mental health in emerging adulthood. In: The Oxford handbook of emerging adulthood. New York, NY, US: Oxford University Press; 2016. p. 499\u0026ndash;520. (Oxford library of psychology).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOtto MW, Pollack MH, Maki KM, Gould RA, Worthington JJ, Smoller JW, et al. Childhood history of anxiety disorders among adults with social phobia: rates, correlates, and comparisons with patients with panic disorder. Depress Anxiety. 2001;14(4):209\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNDRS [Internet]. [cited 2022 Dec 12]. Mental Health of Children and Young People in England, 2020: Wave 1 follow up to the 2017 survey. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2020-wave-1-follow-up\u003c/span\u003e\u003cspan address=\"https://digital.nhs.uk/data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2020-wave-1-follow-up\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePlomin R, Haworth CMA, Davis OSP. Common disorders are quantitative traits. Nat Rev Genet. 2009;10(12):872\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMental health of adolescents [Internet]. [cited 2025 Aug 1]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/news-room/fact-sheets/detail/adolescent-mental-health\u003c/span\u003e\u003cspan address=\"https://www.who.int/news-room/fact-sheets/detail/adolescent-mental-health\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMental health services: addressing the care deficit [Internet]. [cited 2022 Oct 21]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://nhsproviders.org/mental-health-services-addressing-the-care-deficit/the-demand-challenge\u003c/span\u003e\u003cspan address=\"https://nhsproviders.org/mental-health-services-addressing-the-care-deficit/the-demand-challenge\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarin JM, Allwood M, Ball S, Crane C, Wilde KD, Hinze V, et al. School-based mindfulness training in early adolescence: what works, for whom and how in the MYRIAD trial ? Evid Based Ment Health. 2022;1\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMundy J, Moore E, Soan C, Anderson JK, Albajara Saenz A, Baser A, et al. Evaluating the implementation of the Transforming Children and Young People\u0026rsquo;s Mental Health Provision Green Paper programme: Findings from surveys of schools and colleges and Mental Health Support Teams (2024) [Internet]. London School of Hygiene \u0026amp; Tropical Medicine; 2025 [cited 2025 Jul 15]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://researchonline.lshtm.ac.uk/id/eprint/4676423/\u003c/span\u003e\u003cspan address=\"https://researchonline.lshtm.ac.uk/id/eprint/4676423/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFoulkes L, Stringaris A. Do no harm: can school mental health interventions cause iatrogenic harm? BJPsych Bull. 2023;47(5):267\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH\u0026ouml;gberg B. Educational stressors and secular trends in school stress and mental health problems in adolescents. Soc Sci Med 1982. 2021;270:113616.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePascoe MC, Hetrick SE, Parker AG. The impact of stress on students in secondary school and higher education. Int J Adolesc Youth. 2020;25(1):104\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArmitage JM, Collishaw S, Sellers R. Explaining long-term trends in adolescent emotional problems: what we know from population-based studies. Discov Soc Sci Health. 2024;4(1):14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePISA 2015 Results (Volume III): Students\u0026rsquo; Well-Being | en | OECD [Internet]. [cited 2023 Apr 19]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.oecd.org/education/pisa-2015-results-volume-iii-9789264273856-en.htm\u003c/span\u003e\u003cspan address=\"https://www.oecd.org/education/pisa-2015-results-volume-iii-9789264273856-en.htm\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNSPCC Learning [Internet]. [cited 2022 Dec 13]. Childline annual review. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://learning.nspcc.org.uk/research-resources/childline-annual-review/\u003c/span\u003e\u003cspan address=\"https://learning.nspcc.org.uk/research-resources/childline-annual-review/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ewww.basw.co.uk [Internet]. 2015 [cited 2022 Dec 13]. Exam factories? The impact of accountability measures on children and young people. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.basw.co.uk/resources/exam-factories-impact-accountability-measures-children-and-young-people\u003c/span\u003e\u003cspan address=\"https://www.basw.co.uk/resources/exam-factories-impact-accountability-measures-children-and-young-people\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKelly-Irving M, Lepage B, Dedieu D, Bartley M, Blane D, Grosclaude P, et al. Adverse childhood experiences and premature all-cause mortality. Eur J Epidemiol. 2013;28(9):721\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrock-Baca AH, Zundel C, Fox D, Johnson Nagel N. Partnering with Family Advocates to Understand the Impact on Families Caring for a Child with a Serious Mental Health Challenge. J Behav Health Serv Res. 2022;1\u0026ndash;18.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrince M, Patel V, Saxena S, Maj M, Maselko J, Phillips MR, et al. No health without mental health. Lancet Lond Engl. 2007;370(9590):859\u0026ndash;77.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang M, Carson C, Creswell C, Violato M. Child mental health and income gradient from early childhood to adolescence: Evidence from the UK. SSM - Popul Health. 2023;24:101534.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evon Stumm S. Socioeconomic status amplifies the achievement gap throughout compulsory education independent of intelligence. Intelligence. 2017;60:57\u0026ndash;62.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evon Stumm S, Cave SN, Wakeling P. Persistent association between family socioeconomic status and primary school performance in Britain over 95 years. Npj Sci Learn. 2022;7(1):4.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSteare T, Guti\u0026eacute;rrez Mu\u0026ntilde;oz C, Sullivan A, Lewis G. The association between academic pressure and adolescent mental health problems: A systematic review. J Affect Disord. 2023;339:302\u0026ndash;17.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeeland J, Moens MA, Beute F, Assink M, Staaks JPC, Overbeek G. A dose of nature: Two three-level meta-analyses of the beneficial effects of exposure to nature on children\u0026rsquo;s self-regulation. J Environ Psychol. 2019;65:101326.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaldwin JR, Wang B, Karwatowska L, Schoeler T, Tsaligopoulou A, Munaf\u0026ograve; MR, et al. Childhood Maltreatment and Mental Health Problems: A Systematic Review and Meta-Analysis of Quasi-Experimental Studies. Am J Psychiatry. 2023;180(2):117\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAhmadzadeh YI, Schoeler T, Han M, Pingault JB, Creswell C, McAdams TA. Systematic Review and Meta-analysis of Genetically Informed Research: Associations Between Parent Anxiety and Offspring Internalizing Problems. J Am Acad Child Adolesc Psychiatry. 2021;60(7):823\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFowler PJ, Tompsett CJ, Braciszewski JM, Jacques-Tiura AJ, Baltes BB. Community violence: a meta-analysis on the effect of exposure and mental health outcomes of children and adolescents. Dev Psychopathol. 2009;21(1):227\u0026ndash;59.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRutter M. Fifteen Thousand Hours: Secondary Schools and Their Effects on Children. Harvard University Press.; 1979.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNo Wrong Door: bringing services together to meet children\u0026rsquo;s needs [Internet]. Children\u0026rsquo;s Commissioner for Wales. [cited 2025 Jul 15]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.childcomwales.org.uk/publications/no-wrong-door-bringing-services-together-to-meet-childrens-needs/\u003c/span\u003e\u003cspan address=\"https://www.childcomwales.org.uk/publications/no-wrong-door-bringing-services-together-to-meet-childrens-needs/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAnderson M, Werner-Seidler A, King C, Gayed A, Harvey SB, O\u0026rsquo;Dea B. Mental health training programs for secondary school teachers: A systematic review. Sch Ment Health Multidiscip Res Pract J. 2019;11(3):489\u0026ndash;508.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYamaguchi S, Foo JC, Nishida A, Ogawa S, Togo F, Sasaki T. Mental health literacy programs for school teachers: A systematic review and narrative synthesis. Early Interv Psychiatry. 2020;14(1):14\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRaniti M, Rakesh D, Patton GC, Sawyer SM. The role of school connectedness in the prevention of youth depression and anxiety: a systematic review with youth consultation. BMC Public Health. 2022;22(1):2152.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRose ID, Lesesne CA, Sun J, Johns MM, Zhang X, Hertz M. The Relationship of School Connectedness to Adolescents\u0026rsquo; Engagement in Co-Occurring Health Risks: A Meta-Analytic Review. J Sch Nurs Off Publ Natl Assoc Sch Nurses. 2024;40(1):58\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAldridge JM, McChesney K, Afari E. Relationships between school climate, bullying and delinquent behaviours. Learn Environ Res. 2018;21(2):153\u0026ndash;72.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThe Good Childhood Report 2022 | The Children\u0026rsquo;s Society [Internet]. 2022 [cited 2023 Apr 21]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.childrenssociety.org.uk/information/professionals/resources/good-childhood-report-2022\u003c/span\u003e\u003cspan address=\"https://www.childrenssociety.org.uk/information/professionals/resources/good-childhood-report-2022\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePolderman TJC, Benyamin B, de Leeuw CA, Sullivan PF, van Bochoven A, Visscher PM, et al. Meta-analysis of the heritability of human traits based on fifty years of twin studies. Nat Genet. 2015;47(7):702\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKnopik VS, Neiderhiser JM, DeFries JC, Plomin R. Behavioral Genetics. 7th ed. Worth Publishers, New York; 2017.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanscombe KB, Haworth CMA, Davis OSP, Jaffee SR, Plomin R. The nature (and nurture) of children\u0026rsquo;s perceptions of family chaos. Learn Individ Differ. 2010;20:549\u0026ndash;53.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePlomin R, Bergeman CS. The nature of nurture: Genetic influence on \u0026lsquo;environmental\u0026rsquo; measures. Behav Brain Sci. 1991;14:373.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePlomin R, Reiss D, Hetherington EM, Howe GW. Nature and nurture: Genetic contributions to measures of the family environment. Dev Psychol. 1994;30:32\u0026ndash;43.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKoellinger P, Harden KP. Using nature to understand nurture. Am J Public Health. 1990;80(6):657\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKendler KS, Baker JH. Genetic influences on measures of the environment: a systematic review. Psychol Med. 2007;37:615\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDaniels D, Dunn J, Furstenberg FF, Plomin R. Environmental differences within the family and adjustment differences within pairs of adolescent siblings. Child Dev. 1985;56(3):764\u0026ndash;74.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWray NR, Lee SH, Mehta D, Vinkhuyzen AAE, Dudbridge F, Middeldorp CM. Research Review: Polygenic methods and their application to psychiatric traits. J Child Psychol Psychiatry. 2014;55(10):1068\u0026ndash;87.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLockhart C, Bright J, Ahmadzadeh Y, Breen G, Bristow S, Boyd A, et al. Twins Early Development Study (TEDS): A genetically sensitive investigation of mental health outcomes in the mid-twenties. JCPP Adv. 2023;3(2):e12154.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCaspi A, Moffitt TE. All for One and One for All: Mental Disorders in One Dimension. Am J Psychiatry. 2018;appi.ajp.2018.1.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllegrini AG, Cheesman R, Rimfeld K, Selzam S, Pingault JB, Eley TC, et al. The p factor: genetic analyses support a general dimension of psychopathology in childhood and adolescence. J Child Psychol Psychiatry. 2020;\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGidziela A, Malanchini M, Rimfeld K, McMillan A, Ronald A, Viding E, et al. Explaining the influence of non-shared environment (NSE) on symptoms of behaviour problems from preschool to adulthood: mind the missing NSE gap. J Child Psychol Psychiatry. 2023;64(5):747\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGidziela A, Rimfeld K, Malanchini M, Allegrini AG, McMillan A, Selzam S, et al. Using DNA to predict behaviour problems from preschool to adulthood. J Child Psychol Psychiatry. 2022;63(7):781\u0026ndash;92.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllegrini AG, Karhunen V, Coleman JRI, Selzam S, Rimfeld K, Stumm S von, et al. Multivariable G-E interplay in the prediction of educational achievement. PLOS Genet. 2020;16(11):e1009153.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanscombe KB, Trzaskowski M, Haworth CMA, Davis OSP, Dale PS, Plomin R. Socioeconomic Status (SES) and Children\u0026rsquo;s Intelligence (IQ): In a UK-Representative Sample SES Moderates the Environmental, Not Genetic, Effect on IQ. PLOS ONE. 2012;7(2):e30320.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRimfeld K, Malanchini M, Spargo T, Spickernell G, Selzam S, McMillan A, et al. Twins Early Development Study: A Genetically Sensitive Investigation into Behavioral and Cognitive Development from Infancy to Emerging Adulthood. Twin Res Hum Genet Off J Int Soc Twin Stud. 2019;22(6):508\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSelzam S, McAdams TA, Coleman JRI, Carnell S, O\u0026rsquo;Reilly PF, Plomin R, et al. Evidence for gene-environment correlation in child feeding: Links between common genetic variation for BMI in children and parental feeding practices. PLoS Genet. 2018;14(11):e1007757.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVilhj\u0026aacute;lmsson BJ, Yang J, Finucane HK, Gusev A, Lindstr\u0026ouml;m S, Ripke S, et al. Modeling Linkage Disequilibrium Increases Accuracy of Polygenic Risk Scores. Am J Hum Genet. 2015;97(4):576\u0026ndash;92.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllegrini AG, Selzam S, Rimfeld K, von Stumm S, Pingault JB, Plomin R. Genomic prediction of cognitive traits in childhood and adolescence. Mol Psychiatry. 2019;24(6):819\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSelzam S, McAdams TA, Coleman JRI, Carnell S, O\u0026rsquo;Reilly PF, Plomin R, et al. Evidence for gene-environment correlation in child feeding: Links between common genetic variation for BMI in children and parental feeding practices. PLoS Genet. 2018;14(11):e1007757.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemontis D, Walters GB, Athanasiadis G, Walters R, Therrien K, Nielsen TT, et al. Genome-wide analyses of ADHD identify 27 risk loci, refine the genetic architecture and implicate several cognitive domains. Nat Genet. 2023;55(2):198\u0026ndash;208.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePurves KL, Coleman JRI, Meier SM, Rayner C, Davis KAS, Cheesman R, et al. A major role for common genetic variation in anxiety disorders. Mol Psychiatry. 2020;25(12):3292\u0026ndash;303.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrove J, Ripke S, Als TD, Mattheisen M, Walters RK, Won H, et al. Identification of common genetic risk variants for autism spectrum disorder. Nat Genet. 2019;51(3):431\u0026ndash;44.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMullins N, Forstner AJ, O\u0026rsquo;Connell KS, Coombes B, Coleman JRI, Qiao Z, et al. Genome-wide association study of more than 40,000 bipolar disorder cases provides new insights into the underlying biology. Nat Genet. 2021;53(6):817\u0026ndash;29.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStrawbridge RJ, Ward J, Cullen B, Tunbridge EM, Hartz S, Bierut L, et al. Genome-wide analysis of self-reported risk-taking behaviour and cross-disorder genetic correlations in the UK Biobank cohort. Transl Psychiatry. 2018;8(1):1\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLane JM, Liang J, Vlasac I, Anderson SG, Bechtold DA, Bowden J, et al. Genome-wide association analyses of sleep disturbance traits identify new loci and highlight shared genetics with neuropsychiatric and metabolic traits. Nat Genet. 2017;49(2):274\u0026ndash;81.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWray NR, Ripke S, Mattheisen M, Trzaskowski M, Byrne EM, Abdellaoui A, et al. Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet. 2018;50(5):668\u0026ndash;81.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoward DM, Adams MJ, Clarke TK, Hafferty JD, Gibson J, Shirali M, et al. Genome-wide meta-analysis of depression identifies 102 independent variants and highlights the importance of the prefrontal brain regions. Nat Neurosci. 2019;22(3):343\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang H, Lane JM, Jones SE, Dashti HS, Ollila HM, Wood AR, et al. Genome-wide association analysis of self-reported daytime sleepiness identifies 42 loci that suggest biological subtypes. Nat Commun. 2019;10(1):3503.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKarlsson Linn\u0026eacute;r R, Mallard TT, Barr PB, Sanchez-Roige S, Madole JW, Driver MN, et al. Multivariate analysis of 1.5 million people identifies genetic associations with traits related to self-regulation and addiction. Nat Neurosci. 2021;24(10):1367\u0026ndash;76.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNeale lab [Internet]. [cited 2022 Dec 5]. UK Biobank. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.nealelab.is/uk-biobank\u003c/span\u003e\u003cspan address=\"http://www.nealelab.is/uk-biobank\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJansen PR, Watanabe K, Stringer S, Skene N, Bryois J, Hammerschlag AR, et al. Genome-wide analysis of insomnia in 1,331,010 individuals identifies new risk loci and functional pathways. Nat Genet. 2019;51(3):394\u0026ndash;403.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eColeman JRI, Gaspar HA, Bryois J, Bipolar Disorder Working Group of the Psychiatric Genomics Consortium, Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium, Breen G. The Genetics of the Mood Disorder Spectrum: Genome-wide Association Analyses of More Than 185,000 Cases and 439,000 Controls. Biol Psychiatry. 2020;88(2):169\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuciano M, Hagenaars SP, Davies G, Hill WD, Clarke TK, Shirali M, et al. Association analysis in over 329,000 individuals identifies 116 independent variants influencing neuroticism. Nat Genet. 2018;50(1):6\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInternational Obsessive Compulsive Disorder Foundation Genetics Collaborative (IOCDF-GC) and OCD Collaborative Genetics Association Studies (OCGAS). Revealing the complex genetic architecture of obsessive-compulsive disorder using meta-analysis. Mol Psychiatry. 2018;23(5):1181\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNievergelt CM, Maihofer AX, Klengel T, Atkinson EG, Chen CY, Choi KW, et al. International meta-analysis of PTSD genome-wide association studies identifies sex- and ancestry-specific genetic risk loci. Nat Commun. 2019;10(1):4558.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKarlsson Linn\u0026eacute;r R, Biroli P, Kong E, Meddens SFW, Wedow R, Fontana MA, et al. Genome-wide association analyses of risk tolerance and risky behaviors in over 1 million individuals identify hundreds of loci and shared genetic influences. Nat Genet. 2019;51(2):245\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrubetskoy V, Pardi\u0026ntilde;as AF, Qi T, Panagiotaropoulou G, Awasthi S, Bigdeli TB, et al. Mapping genomic loci implicates genes and synaptic biology in schizophrenia. Nature. 2022;604(7906):502\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYao Y, Jia Y, Wen Y, Cheng B, Cheng S, Liu L, et al. Genome-Wide Association Study and Genetic Correlation Scan Provide Insights into Its Genetic Architecture of Sleep Health Score in the UK Biobank Cohort. Nat Sci Sleep. 2022;14:1\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOkbay A, Baselmans BML, De Neve JE, Turley P, Nivard MG, Fontana MA, et al. Genetic variants associated with subjective well-being, depressive symptoms, and neuroticism identified through genome-wide analyses. Nat Genet. 2016;48(6):624\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR Core Team. R: A Language and Environment for Statistical Computing. 2017.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartin NG, Eaves LJ. Stages; the First To Determine the Genetical and Environmental Model. Most. 1977;38:79\u0026ndash;95.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRijsdijk FV, Sham PC. Analytic approaches to twin data using structural equation models. Brief Bioinform. 2002;3(2):119\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoker S, Neale M, Maes H, Wilde M, Spiegel M, Brick T, et al. OpenMx: an open source extended structural equation modeling framework. Psychometrika. 2011;76:306\u0026ndash;17.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBerrie L, Arnold KF, Tomova GD, Gilthorpe MS, Tennant PWG. Depicting deterministic variables within directed acyclic graphs: an aid for identifying and interpreting causal effects involving derived variables and compositional data. Am J Epidemiol. 2025;194(2):469\u0026ndash;79.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAkingbuwa WA, Hammerschlag AR, Allegrini AG, Sallis H, Kuja-Halkola R, Rimfeld K, et al. Multivariate analyses of molecular genetic associations between childhood psychopathology and adult mood disorders and related traits. Am J Med Genet Part B Neuropsychiatr Genet Off Publ Int Soc Psychiatr Genet. 2023;192(1\u0026ndash;2):3\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAkingbuwa WA, Hammerschlag AR, Jami ES, Allegrini AG, Karhunen V, Sallis H, et al. Genetic Associations Between Childhood Psychopathology and Adult Depression and Associated Traits in 42 998 Individuals: A Meta-analysis. JAMA Psychiatry. 2020;77(7):715\u0026ndash;28.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllegrini AG, Baldwin JR, Barkhuizen W, Pingault JB. Research Review: A guide to computing and implementing polygenic scores in developmental research. J Child Psychol Psychiatry. 2022;63(10):1111\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePingault JB, Allegrini AG, Odigie T, Frach L, Baldwin JR, Rijsdijk F, et al. Research Review: How to interpret associations between polygenic scores, environmental risks, and phenotypes. J Child Psychol Psychiatry. 2022;63(10):1125\u0026ndash;39.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArseneault L. Annual Research Review: The persistent and pervasive impact of being bullied in childhood and adolescence: implications for policy and practice. J Child Psychol Psychiatry. 2018;59(4):405\u0026ndash;21.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArseneault L. The long-term impact of bullying victimization on mental health. World Psychiatry. 2017;16(1):27\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSoneson E, Puntis S, Chapman N, Mansfield KL, Jones PB, Fazel M. Happier during lockdown: a descriptive analysis of self-reported wellbeing in 17,000 UK school students during Covid-19 lockdown. Eur Child Adolesc Psychiatry. 2023;32(6):1131\u0026ndash;46.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eElective home education, Autumn term 2024/25 [Internet]. [cited 2024 Dec 29]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://explore-education-statistics.service.gov.uk/find-statistics/elective-home-education/2024-25-autumn-term\u003c/span\u003e\u003cspan address=\"https://explore-education-statistics.service.gov.uk/find-statistics/elective-home-education/2024-25-autumn-term\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTwins Early Development Study (TEDS): A genetically sensitive investigation of mental health outcomes in the mid-twenties - Lockhart \u0026ndash;\u0026thinsp;2023 - JCPP Advances - Wiley Online Library [Internet]. [cited 2025 Jan 1]. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://acamh.onlinelibrary.wiley.com/doi/full/\u003c/span\u003e\u003cspan address=\"https://acamh.onlinelibrary.wiley.com/doi/full/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jcv2.12154\u003c/span\u003e\u003cspan address=\"10.1002/jcv2.12154\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Mental health, school environment, cumulative risk, gene-environment correlation, developmental psychopathology","lastPublishedDoi":"10.21203/rs.3.rs-8626819/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8626819/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eChildren spend a significant part of their lives at school. However, the long-term effects of the school environment on mental health are still not well understood, especially using genetically sensitive designs. Here, we examine the associations between perceived school environment and mental health from childhood to emerging adulthood, and the genetic and environmental factors that underlie these relationships.\u003c/p\u003e \u003cp\u003eUsing data from over 6,500 participants aged 7\u0026ndash;21 from the Twins Early Development Study (TEDS), we found consistent, moderate associations between perceived school environment and mental health (average r\u0026thinsp;\u0026asymp;\u0026thinsp;.19). School environment cumulatively explained mental health problems, explaining 26\u0026ndash;56% of the variance both contemporaneously and over time. These associations remained substantial after adjusting for genetic predisposition using psychiatric polygenic scores, family socioeconomic status and earlier mental health problems, although the effect sizes were smaller (6\u0026ndash;30% of variance explained). Twin analyses showed that not only was psychopathology highly heritable (~\u0026thinsp;61%), but also how children experience school was partly due to genetics (~\u0026thinsp;46%). The association between perceived school environment and mental health was largely accounted for by shared genetic influences (~\u0026thinsp;70%), supporting the role of gene\u0026ndash;environment correlation in mental health outcomes.\u003c/p\u003e \u003cp\u003eWe show that perceived school environments are significantly associated with mental health across development, even after accounting for genetic predisposition, SES and earlier mental health problems. Using a genetically sensitive, longitudinal approach, this research provides a conservative yet clearer estimate of how school environments might influence mental health outcomes over time, because participants completed their schooling more than a decade ago, when reported youth mental health problems were lower, and school environments were likely less pressured. Our findings emphasise the importance of understanding the school environment as a potential setting for prevention and support and underscore the need for further research to identify modifiable factors that could improve children\u0026rsquo;s wellbeing.\u003c/p\u003e","manuscriptTitle":"Longitudinal associations between school environment and mental health from childhood through early adulthood","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-23 03:20:23","doi":"10.21203/rs.3.rs-8626819/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"87e7fe78-b976-449b-87a2-b647a469450c","owner":[],"postedDate":"January 23rd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":61595506,"name":"Biological sciences/Genetics"},{"id":61595507,"name":"Biological sciences/Psychology"}],"tags":[],"updatedAt":"2026-02-12T14:51:21+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-23 03:20:23","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8626819","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8626819","identity":"rs-8626819","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}