If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Population, Family and Reproductive Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MarylandGender, Sexual and Reproductive Health, CESP Centre for Research in Epidemiology and Population Health U1018, Inserm, Villejuif, France
This study was undertaken among 10- to 14-year-old girls and boys in disadvantaged areas of Shanghai, China; Cuenca, Ecuador; Flanders, Belgium; and Denpasar and Semarang, Indonesia. It aimed to assess whether gender norms are related to depressive symptomatology, and to examine whether sex differences in depressive symptoms can be explained by differences in gender norm perceptions.
We examined the distributions of depressive symptoms and two gender norms scales, gender stereotypical traits (GST), and sexual double standard (SDS), across sites and by sex. We next assessed crude and adjusted associations between each of the gender norms scales and depressive symptoms. Finally, we conducted path analysis to examine the mediating role of gender perceptions in sex differences in depressive symptoms.
Girls reported more depressive symptoms than boys in all sites except Denpasar. SDS perceptions were more unequal among girls in most sites, while GST perceptions were more unequal among boys in all sites except Semarang. Gender-equal SDS and GST perceptions were associated with fewer depressive symptoms, while unequal perceptions were related to more symptoms. Gendered perceptions about traits and relationships appeared to partially mediate relationships between sex and depressive symptoms in Shanghai, Cuenca, and Semarang.
Unequal gender norm perceptions were linked to poor mental health among boys and girls, suggesting that gender norms may play a role in psychological wellbeing for adolescents of both sexes. Gender norm perceptions appear to play a role in mental health sex disparities we observed across sites.
Unequal gender perceptions are related to more symptoms of depression among disadvantaged young adolescents in five urban settings. Girls report more depressive symptoms than boys; and gender norms mediate those sex differences in some sites. Programs that address unequal gender norms may play a role in improving the mental health of adolescents.
Over the past decade, concern about young people’s mental health has increased globally [
] have posited that unequal gender norms interfere with both adolescent girls’ and boys’ sense of the control over their lives and futures. The authors also highlighted the dearth of evidence about the relationships between gendered experiences and adolescent mental health, a gap that is particularly pronounced for early adolescence (10–14 years old).
Studies that have assessed the associations between aspects of gendered experiences and poor mental health among adolescents in low- and middle-income countries tend to focus on girls. Beattie et al. [
] highlighted the gendered correlates of distress among 13- to 14-year-old girls in Karnataka, India, including experiences of violence, early childbearing and marriage, and limited autonomy. Satyanarayana et al. [
] linked multiple forms of gender disadvantage (including discrimination and gendered barriers to personal growth) to distress among older adolescent girls living in urban slum areas of Bangalore, India. A relationship between gender-unequal peer norms and poor self-esteem has been documented among adolescent refugee girls in Ethiopia [
]. However, the exclusion of boys from such studies assumes that unequal gender dynamics constrain girls’ choices and promote their social isolation without testing whether such norms benefit or harm the mental health of boys.
Adolescence is also the period during which a sex divide in depression is first observed; and by age 16, depression is nearly twice as common among girls than boys [
]. While developmental, biological, and contextual factors have been proposed to help explain this sex difference, researchers are still exploring the extent to which this disparity is shaped by social factors such as gendered expectations. Several gendered determinants of this gap have been hypothesized, including women and girls’ marginalization, disproportionate exposures to adversity and violence, and distress associated with caretaking and household responsibilities [
Given rising concern about adolescent mental health and efforts to bolster gender equality, this study contributes to a growing body of literature that explores how adolescents’ perceptions of gender norms relate to depressive symptoms in adolescence among both boys and girls [
]. We focus specifically on the critical early adolescent period, at the time of both intensifying gender expectations and the emerging sex divide in depression. The present study aims to explore: (1) the relationships between perceptions of unequal gender norms and the presence of depressive symptoms among adolescents ages 10–14 living in low-income communities in five cities in four countries across the world, (2) compare these associations among adolescent boys and girls, and (3) assess whether perceptions of gender unequal norms mediate relationships between sex and depressive symptomatology that reflect a disproportionate mental health burden among girls.
This analysis included data from five sites of the Global Early Adolescent Study (GEAS): Shanghai, China; Cuenca, Ecuador; Flanders, Belgium; and Denpasar and Semarang, Indonesia. The sites all represent urban poor areas within their respective contexts but vary substantially in their demographic, economic, social, and cultural characteristics, including prevailing gender norms. While all GEAS sites used the same study instruments, the sampling and data collection procedures varied. In Shanghai, Flanders, Semarang, and Denpasar, schools were purposively selected based on the socioeconomic status of the student body. In Cuenca, seven schools were randomly selected after stratification by neighborhood and school type. Within those schools, all eligible students in Shanghai and Flanders, all eligible seventh grade students in Semarang and Denpasar, and a randomly selected group of students (stratified by age and sex) in Cuenca, were invited to enroll [
Initial sample sizes were 1,760 in Shanghai, 704 in Cuenca, 1,013 in Flanders, 1,517 in Semarang, and 1,753 in Denpasar. Cases missing 15% or more of overall survey data were excluded (46 in Shanghai, 90 in Cuenca, 223 in Flanders, 23 in Semarang, and 48 in Denpasar), as were those missing items used to construct the depressive symptom index outcome (10 in Shanghai, 9 in Cuenca, 43 in Flanders, 17 in Semarang, and 39 in Denpasar), or items used to construct the key predictors in the form of two gender norms scales (67 in Shanghai, 50 in Cuenca, 123 in Flanders, 81 in Semarang, and 129 in Denpasar). Our final analytical samples consisted of 1,637 in Shanghai, 555 in Cuenca, 624 in Flanders, 1,396 in Semarang, and 1,537 in Denpasar. We used the K-Nearest Neighbor method to impute any missing data on covariates, based on age and sex (K-values of 36 for Shanghai, 22 for Cuenca, 23 for Flanders, 34 for Semarang, and 36 for Denpasar).
In each of the sites, GEAS surveys were administered using computer-assisted self-interview on tablets. The survey was developed during an extensive formative phase of the GEAS study that involved 14 sites across 5 continents, including Flanders, Cuenca, and Shanghai. The surveys were translated into local languages, revised based on face validity testing, and subsequently pilot tested in all sites. The survey collected information on young people’s sociodemographic background, family and peer environments, perceptions of gender norms and levels of agency, and their health, including mental and sexual health. All study protocols were approved by each site’s ethical review committee and deemed exempt for secondary data collection by the Johns Hopkins School of Public Health’s Institutional Review Board. The study in Shanghai was also approved by the World Health Organization Ethical Review Committee.
Our main outcome of interest was a depressive symptom index, developed within the GEAS from six items. Each item was scored on a five-point Likert scale ranging from “Disagree a lot” to “Agree a lot.” A summed depressive symptom index was constructed ranging from 0 (“Disagree a lot” to all six items) to 24 (“Agree a lot” to all six items). The items are described in more detail in Appendix Table 1 and have been previously reported in Blum et al. [
Our key predictors were two scales that reflected perceptions of gender norms related to romantic relationships and traits. The first was a sexual double standard (SDS) scale, comprised of six items that measure the extent to which adolescents perceived that girls are penalized and boys rewarded for romantic and sexual activity [
]. More details about these items are presented in Appendix Table 1. Each scale was scored from 1 to 5, with greater values indicating more gender-unequal perceptions. Each of the two gender norms scales were subsequently used to create a categorical measure within each site with three levels: the reference category (within one standard deviation of the mean), gender-equal (more than one standard deviation below the mean), and gender-unequal (more than one standard deviation above the mean). These categorical measures were employed in models that examined the relationships between gender norms and depressive symptoms, in an effort to examine the associated mental health consequences of outlying progressive or traditional norm perceptions.
Covariates included age (continuous) and several binary measures. Pubertal onset (prepubertal or pubertal) was captured by sex-specific indicators of pubertal maturation reflecting breast growth and menarche for girls, and voice, genital, and facial hair development for boys. Educational attainment (behind in school, or at or above expected school year for age), number of close friends (none or any close friends), and caregiver closeness (not close or no caregiver reported, or close to caregiver) were assessed. In addition, a binary indicator that described adolescents’ adverse childhood experiences (ACEs) (reporting none or any exposures) was included as a covariate in all sites except for Flanders, where the items were not asked [
We first examined the sociodemographic characteristics of each sample by sex. We then described mean scores of the depressive symptom index and the two gender norm scales in each site overall and by sex to assess any significant differences between boys and girls using Student’s t-tests. Next, we conducted simple and multivariable linear regressions in the overall samples, and after stratification by sex, to evaluate the effect of each of the gender norms scales (SDS and GST) on depressive symptoms in separate models. Both gender norm measures were treated as categorical in these analyses to examine the risk of depressive symptoms associated with more gender-equal and more gender-unequal perceptions relative to the reference category. Subsequently, test of interactions between sex and gender norm perceptions were conducted to identify any differential effects for boys and girls.
In the third stage of analysis, we assessed the mediating role of each gender norm construct in the relationship between sex and depressive symptoms using structural equation modeling. We conducted these analyses in sites where we observed significant sex differences in depressive symptoms using Stata’s sem command, which fits two regression models. The first regressed our mediator (gender norms) on the independent variable (sex) adjusting for covariates. Then our outcome (depressive symptoms) was regressed on the independent and mediator variables, with adjustment for covariates. This procedure estimated the total effect, indirect effect, and direct effect of sex on depressive symptoms, as well as direct pathways between sex and gender norms measures, and gender norm measures and depressive symptoms. We then calculated the ratio of the indirect effect to total effect, to estimate the percentage of the total effect mediated in each site. For structural equation models, we employed the gender norms scales as continuous measures (ranging from 1 to 5) to avoid the introduction of bias due to categorization of the mediator. All analyses were stratified by site to account for variation in study contexts and sampling procedures.
Descriptions of each site’s sample are presented in Table 1. Boys were older than girls in all sites except Cuenca. In the two sites in Indonesia, boys were more likely than girls to be behind in school for their age. More girls than boys had begun puberty in all sites except Flanders. ACEs were more common among girls than boys in Cuenca.
Mean depressive symptom scores (on a scale from 0 to 24) are shown in Figure 1. Depressive symptom mean scores were highest in Denpasar (mean: 11.4) and lowest in Flanders (mean: 7.6). Mean depressive symptom scores were significantly higher among girls than boys in all sites, with the exception of Denpasar, where we observed no significant sex difference.
Mean scores of the SDS and GST scales ranged from 1 to 5, and are presented in Figure 2. Perceptions of SDS were most unequal in Cuenca (mean score = 3.6) and most equal in Shanghai and Semarang (mean score = 2.6). In each site but Denpasar, perceptions of SDS were stronger among girls than boys. Perceptions of GST were more unequal in general than those of SDS. GST perceptions were most unequal among adolescents in Denpasar (mean score = 3.9) and most equal in Flanders (mean score = 3.0). Boys’ perceptions of gendered traits were more unequal than girls across the sites except in Semarang.
Associations between gender norms and depressive symptoms
The results of adjusted linear regression models that examined relationships between gender norm perceptions and depressive symptoms are presented in Figure 3 (tables of regression results are presented in Appendix Table 2).
In overall adjusted models, gender-equal SDS perceptions were associated with fewer depressive symptoms, compared to the reference group in all five sites. Meanwhile, adolescents with gender-unequal SDS perceptions had more depressive symptoms compared to the reference group. This finding was significant in all sites but Cuenca. In sex-stratified models, gender-equal perceptions of SDS were significantly related to fewer depressive symptoms among girls in Shanghai, Semarang, and Denpasar, and among boys in Flanders and Semarang. Meanwhile, gender-unequal perceptions were associated with more depressive symptoms only among boys in Shanghai and Flanders, and among both boys and girls in the two Indonesian sites.
Testing for differences in the relationships between SDS perceptions and depressive symptoms by sex in the overall model, findings diverged for girls and boys in Denpasar (p = .026). There, the relationship between gender-equal SDS perceptions and fewer depressive symptoms was stronger among girls than boys (p = .007).
After adjustment, unequal GST perceptions were associated with more symptoms of depression in all sites with the exception of Flanders. Meanwhile, gender-equal perceptions were only significantly linked to fewer depressive symptoms in Shanghai. In sex-stratified models, equal GST perceptions were linked to fewer depressive symptoms among girls in Shanghai, boys in Cuenca, and girls in Semarang. Gender-unequal GST perceptions were associated with more depressive symptoms among both boys and girls in all sites but Flanders.
Tests of interaction terms between sex and GST perceptions were not statistically significant.
Tests of mediation effects
We next explored the role of gender norm perceptions in the observed sex differences in depressive symptoms using a mediation model that was adjusted for covariates (Figure 4). As we observed no significant differences in depressive symptoms in Denpasar by sex, mediation analyses were conducted in Shanghai, Cuenca, Flanders, and Semarang.
The results of the mediation models for SDS are presented in Figure 4A. When we examined each direct path, we observed positive and significant effects of (1) sex on depressive symptoms, (2) sex on SDS, and (3) SDS on depressive symptoms. Together, these direct effects indicate that girls exhibited more depressive symptoms and endorsed more unequal SDS perceptions than boys, and that unequal SDS perceptions were associated with more depressive symptoms. The indirect path through SDS accounted for 13.48% (in Semarang) to 37.23% (in Shanghai) of the total effect of sex on depressive symptoms. No significant mediation effect was observed in Flanders.
The results of the models that examined the mediating effect of GST perceptions on sex differences in depressive symptoms are presented in Figure 4B. In these models, while the direct effects of both sex and GST on depressive symptoms were positive (indicating that girls compared to boys and more unequal GST were each linked to depressive symptoms), the direct effect of sex on GST was negative in most sites (reflecting more unequal GST endorsement among boys than girls). The differing directions of effects in the indirect pathway produced inconsistent mediation, which in turn generated negative proportions of total effect that is indirect in most sites. The mediating effect was significant in Shanghai and Cuenca, where 21.28% and −44.35% of the total effect operated through the indirect pathway, respectively. In Flanders and Semarang, no significant mediation effects were detected.
The present analysis explored the relationships between depressive symptoms and perceptions of gender norms characterizing traits and relationships among young adolescent boys and girls in disadvantaged communities within five urban centers in China, Ecuador, Belgium, and Indonesia. We found that adolescents with unequal gender perceptions about both romantic relationships and traits were at risk for greater depressive symptomology. Conversely, gender-equal perceptions about relationships were associated with fewer reported symptoms of depression, and this relationship was especially pronounced for girls in Denpasar. This study also provides novel evidence that sex divides in depression are partially mediated by gender norm perceptions in at least some settings.
Depressive symptoms were common across settings, with girls reporting greater depression than boys in all sites but Denpasar. These sex differences in depressive symptomatology echo prior findings on depression in this age group [
]. Our results demonstrate that after controlling for other factors, such as pubertal onset and experiences of childhood adversity, by the early adolescent years (ages 10–14), a clear sex disparity has already emerged across most of these settings.
In our analysis, depressive symptomatology was generally related to gender norm perceptions across sites, increasing as young adolescents’ gender perceptions became more unequal. Our study contributes to a small body of literature that has examined the associations between gender norm perceptions and mental health in the early adolescent period. Saeed Ali et al. [
] also demonstrated links between unequal individual- and community-level gender perceptions and poor mental health among young adolescents in low- and middle-income countries settings. Their findings suggest that this association extends beyond personal attitudes, as community-level gender norms also affected adolescents’ mental health. Together, this evidence suggests that unequal gender norms may be detrimental to the mental health of young people.
Our results echo those of Baird, suggesting that gender unequal norms may impact mental health among young adolescents of both sexes. We also found that in most sites, gender norm perceptions were related to both boys’ and girls’ risk of depressive symptoms. These results suggest that restrictive gender norms may constrain young people’s freedom of expression and perceptions of their futures, and that interventions effective in reducing gender unequal norms may have a positive impact on the mental health of adolescents of both sexes. Patterns diverged in Denpasar, Indonesia, where we noted no sex differences in depressive symptoms. Prior research with this cohort in Denpasar has shown that girls have greater voice and decision-making agency than boys, perhaps reflecting a departure from gendered divisions of power that are common across many settings, and which may explain the absence of a gender divide in mental health outcomes in this population [
]. In Denpasar, we also documented a more beneficial effect of gender equal SDS perceptions among girls than boys. In that setting, girls may stand to benefit more from gender-equal perceptions about relationships as such norms function to marginalize women disproportionately. Questions remain as to whether restrictive gender norms lower young people’s perceived control over their lives, how biological factors interact with social vulnerability, and how living in communities with more unequal gender perceptions put young people at increased risks of comorbidities of depression, such as violence.
The sex divide in depression has previously been attributed to both biological and social factors that influence symptomatology, health-seeking behaviors, and symptom management [
]. In our study, we found that SDS perceptions partially mediated the observed sex differences in depressive symptoms in Shanghai, Cuenca, and Semarang. These results suggest that SDS may play a role in the relative mental health disadvantage we see among girls in those sites. While we observed the largest gap between girls’ and boys’ reported mental health in Flanders, SDS perceptions did not appear to be driving this divide. This could be attributed to the smaller sex differential in SDS perceptions observed in Flanders than in other sites. Meanwhile for GST, our findings demonstrate that more unequal trait perceptions among boys compared to girls may be suppressing the mental health sex differences we observe in total effects. The different mediation patterns observed for SDS and GST also highlight the heterogeneity of gender perceptions and their relationships to mental health. Together, these results suggest that the disproportionate burden of depression among girls compared to boys that emerges in early adolescence may be shaped in part by the social norms that govern gender scripts for young people.
While the role of gender norms as a mediator in sex differences in depression among young people has not been widely studied, existing research has linked gender inequality and greater differences by sex in depression. For example, Yu’s [
] cross-national analysis of adult populations found relationships between more equal views about women’s roles and narrowed sex differences in major depression within countries across time. In the U.S., O’Loughlin et al. [
] demonstrated that men'’ greater endorsement of toughness mediated an observed sex difference, with women more likely to seek help for depression. Our findings contribute novel evidence about the interactions of biological and social factors in depression risk during early adolescence, a critical juncture that remains understudied to date. Future research on this topic should explore the mechanisms by which gender norms operate to shape gendered mental health disparities among young people.
This analysis also has several important limitations. First, while our measure of depressive symptoms has been used in prior studies, it has thus far only been validated in limited samples, and cannot be interpreted as diagnostic. The lack of a widely validated measure of depression in this analysis also means we cannot make any comparisons about the prevalence of depression among our sample. Second, our gender norms measures represent adolescents’ perceptions of gender norms, rather than their individual attitudes, and therefore may differ from other research that aims to explore the link between gender attitudes and mental health. This difference, though nuanced, may shape the findings we have observed in settings where adolescents’ attitudes and perceptions of norms may differ. Third, these data are cross-sectional, and therefore causality in the detected associations cannot be inferred. Longitudinal data, which is collected as part of the GEAS, will help to elucidate causal links between gender norms and depression. Fourth, the samples employed in this analysis are not representative, and therefore findings cannot be generalized to young people from different social backgrounds including out of school adolescents, affluent urban populations, or those living in rural settings. Further research should explore these questions using larger representative samples, to examine how perceptions of gender norms contribute to adolescent depressive symptoms according to their social contexts. Relatedly, while the study samples cover geographical representation of three continents, there are wide regional shortcomings of the sites represented in this study. A replication of this analysis in other GEAS sites in sub-Saharan Africa and South America will provide more insights on the ways gender norms inform sex differences in mental health in adolescents according to context. Finally, while we have accounted for additional factors that may shape young adolescents’ risk of depression (including protective factors such as friendship and caregiver relationships, and risks such as ACE exposures), we are unable to capture a full picture of the biological and social factors that shape depression, of which gender is only a singular aspect.
Despite its limitations, among the strengths of this study are its exploration of the relationships between perceptions of two dimensions of gender norms and symptoms of depression among young people living in five geographically and culturally diverse urban contexts, contributing novel cross-cultural data to this question. Employing five distinct samples from three continents, we generally found patterns in the same directions across sites. Future research that explores the mechanisms by which gender unequal norm perceptions relate to adverse mental health will help to elucidate the social factors that shape the emergence of depression in this period, and further inform policies to promote young people’s psychosocial wellbeing.
This study demonstrates that unequal gender norm perceptions are related to more symptoms of depression among young adolescents ages 10–14 across diverse geographies and settings. Results of mediation analyses demonstrated that each gender norm domain (SDS and GST) partially mediated the sex disparities in depressive symptoms we observed in some sites. These findings contribute evidence that suggests a mental health penalty of unequal gender perceptions among both young adolescent boys and girls, as well as the ways gendered perceptions may shape the disproportionate burdens of depression girls face beginning in this period. This study broadens the evidence base that suggests that the mental health of all young people stands to benefit from gender-equal perceptions and norms. Future research, with longitudinal designs more expansive measures of gendered perceptions, can elucidate the relationship between gender expectations and mental health among young people.
This global study was conducted with support from the Bill & Melinda Gates Foundation [ OPP1125119 ], the Oak Foundation [ OCAY-17-649 ], the Packard Foundation [ 2017-66517 ], and the UNDP-UNFPA-UNICEF-WHO-World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), a co-sponsored programme executed by the World Health Organization (WHO). Support for the Global Early Adolescent Study in Indonesia is provided by the Bill & Melinda Gates Foundation [ OPP1178415 ]. In Flanders, the study is supported by the Flemish Ministry of Innovation, Public Investment, Media and Poverty Reduction and the Fund for Scientific Research, Flanders. In Shanghai, the site is supported by the Innovation-oriented Science and Technology Grant from the National Health Commission Key Laboratory of Reproduction Regulation. The study in Cuenca is funded by National Secretary of Science and Technology (SENESCYT) of Ecuador. The authors would also like Jacob Borodovsky for consultation about statistical analysis.
Child and adolescent mental health worldwide: Evidence for action.
Conflicts of interest: None of the authors have potential conflicts of interest to be disclosed.
Disclaimer: Publication of this supplement was supported by the Johns Hopkins Bloomberg School of Public Health Department of Population, Family and Reproductive Health with funding from the Bill and Melinda Gates Foundation.