Several studies have reported that gender roles explain variance in anxiety and depression beyond that of biological sex (19–21). Gender is a multidimensional concept that includes sociocultural roles and identities (i.e., men, women, nonbinary people) that are different from one’s biological sex (i.e. sex assigned at birth: male, female, intersex) that is otherwise determined by genes, hormone levels, and gonads (22, 23). While growing up and socializing, children will endorse sex-specific personality traits or gender roles by performing behaviors that are stereotypically classified as “feminine” (e.g. affective concerns) and “masculine” (e.g. cognitive focus on action) (24). Several questionnaires, like the Personal Attributes Questionnaire (PAQ) (25), have aimed to assess one’s femininity and masculinity by analyzing personality traits and classifying their descriptions as “feminine” or communal and “masculine” or agentic factors.

In the 1970s, Bem developed the Bem Sex Role Inventory (BSRI) that assesses femininity and masculinity on two orthogonal but still complementary continuums (26). Two approaches to gender roles have been developed. First, the categorical method that classifies people as masculine, feminine, androgynous (when both of their feminine and masculine scores are high), or undifferentiated (when both of their scores are low). Second, a continuous method that considers one’s sex and mixture of feminine and masculine continuums to obtain diverse gender role profiles. For instance, people can be sex-typed (when their gender role profile matches their birth-assigned sex), cross-typed (when their gender role profile is opposed to their birth-assigned sex), or mix-typed (when their gender role profile is not mainly feminine nor masculine). As gender roles develop across the lifespan and in the face of stressful life events, they may modulate the way mental health problems manifest themselves (27). Indeed, gender roles overlap with personality traits that can synergistically influence coping strategies (e.g. emotional regulation, self-esteem, adjustment) that are collectively associated with anxiety, depression, and even suicidal behaviors (28).

Whether it has been measured with the PAQ or the BSRI, higher masculinity has been consistently associated with fewer reported symptoms of depression and anxiety (20, 21, 29–32). However, femininity has been less consistently associated with mental health. Some studies have shown that higher femininity is related to more general distress (33), anxiety and depressive symptoms (21, 34). On the other hand, several studies have not shown a significant effect of femininity on mental health (30–32, 35–37), and some even found mixed results where femininity was related to fewer depressive symptoms (29, 38). A meta-analysis by Whitley in 1985 concluded that higher masculinity was associated with less depression and greater adjustment, whereas femininity had no consistent relationship with depression (37).

The apparent heterogeneity of results linking feminine gender roles to depression may be due to different ways that gender roles operate among men and women (20). Some studies did not find significant differences between the way that gender roles affect sex differences in mental health (36). By contrast, several studies have highlighted mixed associations between gender roles and mental health. For example, a study by Gibson et al. reported that higher femininity was associated with fewer depressive symptoms in college educated men, while masculinity had no effect only for women. This highlights the need to consider other demographic and socio-cultural factors such as education level in the study of gender roles (38).

Some studies have investigated the effect of gender roles only in one sex or specific race/ethnic groups. Among females, higher femininity has been associated with more anxiety and depression (39, 40), while less masculinity has been related to more depressive symptoms (30). However, for Black women, higher femininity has been associated with less depressive symptoms (41), which underlines again the inconsistency of findings among females, as well as the importance of considering race/ethnicity and intersecting sociodemographics more broadly in gender role research (42).

It has been proposed that higher masculinity decreases self-harm only in males (43). However, some sociocultural profiles related to certain masculine norms such as self-reliance, difficulty expressing emotions and reduced help-seeking behavior may be ‘maladaptive’ with respect to men’s mental health, particularly in developed Western societies where men experience decreasing social and economic role opportunities (real or perceived) since the 1960s (44, 45). Indeed, higher male-to-female suicide ratios in more common in highly developed, gender egalitarian nations (46). In addition, a study of men hospitalized following stressful life events found that adherence to traditional male gender roles mediated previous suicide attempt status in men (47). Beyond extreme masculinity and femininity alone, it appears that the extent that one’s gender role concords with one’s birth-assigned sex is also a unique predictor of suicide risk. Indeed, cross-typed people (i.e., people with a gender role or personality typology opposed to their birth-assigned sex) are considered to be at higher risk of suicidal symptoms than sex-typed people (i.e., people with a gender role or personality typology in accordance with their birth-assigned sex) (48). These results emphasize that even if gender roles are better predictors of psychological health than other factors like sex (19–21, 49) or sexual orientation (48, 50), it is still important to take birth-assigned sex into account to better understand how gender roles influence well-being as gender roles and personality operate differently for males and females (20).

The apparent heterogeneity in the literature could be due to several methodological issues including the sample sizes and the way gender roles are conceptualized, operationalized and assessed. Indeed, even if gender roles seem to be consistent across generations, Adams and Sherer suggested in 1985 that classic inventories of gender role like PAQ or BSRI measure constructs like assertiveness and self-efficacy better than gender roles (51). Besides, even if Bem had the original intention to measure gender roles without a negative desirability bias, a study by Grimmell and Sterne (1992) reported that the Bem feminine scale of the BSRI contains both positively and negatively valued traits (52). That is, classic measures of gender roles can suffer from a social-desirability bias that affects the way participants evaluate their stereotyped personality traits. This valuation of their own traits may confound mental health measurement. For example, one group showed that only negatively evaluated aspects of femininity of the Australian Sex Role Scale were associated with suicidality (53). More recent studies indicate that most people nowadays disagree with the “masculine” and “feminine” classifications of some characteristics described by the BSRI, suggesting the need of an alternative way to evaluate gender roles which does not suffer from a time period bias (54).

In 1979, Storm created a shortened questionnaire that directly asks participants how they feel about themselves regarding their masculinity and femininity (55) with a simple measure of gender role identity. In this manner, participants fall somewhere on two dimensions comprising their self-perception and perception of others’ perception regarding their masculinity and femininity. Participants are asked to evaluate how they feel, act, and appear to others as more or less feminine and masculine without evoking concepts that can bias their answer regarding both gender role and mental health.

Several studies have underlined the importance of other factors that influence the relationship between gender roles and mental health beyond the questionnaire used. For instance, Sandfort et al. (2021) showed in a US national sample that gender conformity is more present among people with less social status (e.g., age, race/ethnicity, education, income) (56). Gender non-conformity can in turn negatively affect mental health. Taken together, the effects of gender roles on mental health differ according to student/worker status (29), education level (38), race/ethnicity (57), or self-esteem (21, 31). Age also seems to play a key role that influences the effect of gender roles on mental health. For instance, Hunt et al. only succeeded in detecting a significant effect of gender roles on suicidal thoughts among early middle-age participants, but not in early adulthood or late middle-age (58).

Beyond age, sex and race/ethnicity, personality traits are another facet of psychosocial functioning that, like gender roles, can influence mental health and behaviors (59–61). The classic NEO-Five factor inventory (62) breaks down personality into five traits: openness, conscientiousness, extroversion, agreeableness, and neuroticism. The link between these personality traits and mental health has been investigated by numerous studies. Overall, it appears that neuroticism, which represents our vulnerability to experience negative affects including anger and emotional instability, is the only trait that strongly and consistently predicts higher anxiety and depressive symptoms for both sexes (63–68). In contract to neuroticism, extroversion, which is the trait that depict enjoyment for activities involving social interactions, seems to correlate with lower anxiety and depression (63, 65).

In addition, different factors are associated with gender roles, like perceived stress (28) or rumination (69), mediate the relationship between neuroticism and anxiety and depressive symptoms (66, 67). Wupperman and Neumann found that masculinity was associated with lower neuroticism (69), highlighting the close relationship between gender roles, mental health and personality traits. It is no surprise then that personality traits also affect suicidality. Indeed, neuroticism, just like for anxiety and depression, strongly and consistently predicts higher suicidal thoughts (17, 70–74) and behaviors (75) regardless of sex (71).

The link between the other traits and suicidality is less clear. The most consistent evidence suggests that lower levels of extroversion (17, 71, 72, 74) and conscientiousness (70, 71, 74) predicts higher rates of suicidal thoughts. A recent review by Szücs highlighted the importance of age in this relationship between personality traits and suicidality. Especially, elderly people who committed suicide displayed lower levels of openness than younger victims (76). As a whole, personality traits can influence the relationship between gender roles and mental health, as they do for sex and depression (77), since they greatly influence how people feel and act, as well as their psychological health.

To date, gender roles have been associated with mental health outcomes, but results are not consistent, especially regarding the manner in which masculinity and femininity are contextualized and operationalized. Indeed, these mixed results may be due to the methods used. Specifically, some studies have not disaggregated analysis by sex as recommended for rigor and reproducibility in health research (78), and gender roles have mostly been measured by questionnaires that used concepts that may bias participants’ answers. In addition, personality traits should be considered in research on gender roles since they may synergize associations with mental health outcomes.

The present study aims to investigate how self-identified gender roles relate to anxiety, depressive and suicidal symptoms in a well-phenotyped sample drawn from the general population. In sex-specific analyses, we also assess whether personality traits moderate the aforementioned associations. A priori covariates like age, race and ethnicity are also accounted for.

In analyses split by sex, we hypothesize that self-reported femininity will be associated with poorer mental health, as evidenced by higher reported anxiety, depression, and suicidal symptoms. By contrast, masculinity will be associated with lower reported anxiety, depression and suicidal thoughts and behavior. Also, in agreement with the literature, we hypothesize that gender role identity’s effect on mental health will be different in males and females. Specifically, we hypothesize that “cross-typed” profiles (i.e., feminine males and masculine females) will present poorer mental health, while “sex-typed” people will report fewer symptoms of anxiety, depression, and suicide. However, since there is insufficient literature on the role of personality traits as a moderator of associations between gender roles and mental health, we have not explicitly hypothesized directionality for this exploratory hypothesis.

This analysis used data from the Nathan Kline Institute Rockland Sample (NKI-RS) (79). The larger NKI-RS project is a comprehensive community sample of participants studied across the lifespan. Efforts have been concentrated on recruitment strategies to avoid over-representation of any portion of the Rockland County community. Recruitment flyers were posted at schools, shopping malls, community centers, and various other locations in Rockland County in New York State. The resulting sample of more than 1000 participants was recruited between 2012 and 2016. As race, ethnicity and economic demographics of Rockland County are similar to those of the United States, the NKI-RS is generalizable to the broader U.S. population. The larger study also includes the collection of rich phenotyping as well as advanced neuroimaging data. For this study, we restricted our analyses to self-reported behavioral data among adults ages 18 and above.

Participants under 18 years old were excluded (n = 155) as the Trait Anxiety Standard Score was only validated and collected for people above 18 years old. Participants with missing data were also excluded (n = 427). In the current analysis, 741 participants (487 females and 254 males) were therefore included from the NKI-RS.

Institutional Review Board (IRB) approval was obtained for the original project at the Nathan Kline Institute (Phase I #226781 and Phase II #239708) and at Montclair State University (Phase I #000983A and Phase II #000983B). Written informed consent was obtained for all study participants. An additional IRB approval for our analyses of these data was not required. Instead, we reviewed and signed the NKI-RS Data Usage Agreement.

Data used in this analysis are part of larger dataset mentioned above. For further information, please refer to the NKI-Rockland sample description by K. B. Nooner et al. (79). All participants took part in a 2-day experiment composed of biological data measurements, interviews, questionnaires, magnetic resonance imaging, various cognitive tasks, and behavioral measures. Demographic and gender role identity data were collected at the arrival on the first day (8:15 AM), while all the other data used in this analysis were collected during the one-hour questionnaire phase starting at 1:45 PM the same first day.

Statistical analyses were conducted using IBM Statistical Package for the Social Sciences (SPSS) (version 26). Considering the large sample size of the NKI-RS and that sex is an important factor for every variable, all analyses were split by sex as recommended by Clayton and Tannenbaum for improving rigor and reproducibility (78). The sample sizes and degrees of freedom for each statistical model differ slightly since not all measures were collected from all participants at the same time.

Preliminary analyses highlighted that self-assessed gender roles correlated with all personality traits except agreeableness in males, and all mental health measures except suicidality in females. Anxiety symptoms, depressive symptoms, and suicidal thoughts and behaviors were also inter-correlated, consistent with the existing literature (86). Contrary to previous studies (11, 12), our analyses did not find a difference in depression scores between males and females. Surprisingly, males from our sample reported higher neuroticism levels than females, which does not align with existing literature showing the opposite pattern (87, 88). However, as expected, gender role identity scores were substantially different between males and females.

Consistent with our hypothesis, people with a “cross-typed” or “incongruent” gender role profile (i.e., with a gender role identity opposite to their birth-assigned sex) reported poorer mental health, which is consistent with previous work (48). This is also consistent with congruence models of gender roles that state that gender role alignment with birth-assigned sex is less distressing (37). However, this effect turned out to be much stronger in males, where femininity was highly associated with more anxiety symptoms, depressive symptoms, and suicidality. Masculinity in females, however, was less strongly but still significantly associated with anxiety and depressive symptoms. These results are consistent with existing literature where femininity in males have been associated with poorer mental health (33, 89). Notwithstanding, these findings must expanded and replicated using approaches that assess multidimensional gender role profiles.

In contrast to males, our findings only partially align with findings regarding gender roles and mental health in females. In assessing gender roles categorically, undifferentiated females (i.e., females with low levels of masculinity and femininity) usually demonstrate higher anxiety and depression. These observations have been attributed to the low levels of masculinity which would otherwise have a protective effect (29, 37). Other research reported that higher femininity in females was associated with more anxiety (21) and depressive symptoms (39, 40). Our results suggest that it is masculinity, and not femininity, which is associated with higher levels of anxiety and depressive symptoms in females. This being said, our analyses did not allow us to explore other gender configurations (e.g., androgyny).

Our results and the method we used strongly suggest that several forces predict how gender role identity impacts mental health. The main force may be related to the congruence between one’s gender role identity and birth-assigned sex. Having masculine gender role identity among females and feminine gender role identity for males is linked to poorer mental health. This “gendered societal burden” may be strengthened by society’s perception which promotes sex-typed gender roles (89, 90), resulting in more stress-related issues in gender “incongruent” or non-conforming people (91). For instance, feminine gay men tend to experience anti-effeminacy prejudice (90, 92). In addition, being feminine is more likely to be associated with being perceived as gay in men than in women (93). That is, feminine men can suffer from homophobic discrimination regardless of their sexuality. In other research, homophobic discrimination impacts both anxiety and depressive symptoms above and beyond other psychological stressors. This effect is stronger among men compared to women (94). Unfortunately, sexual orientation was not ascertained in this study. Taken together, those societal issues relate to the concept of gender relations and can partially explain differences observed between males and females regarding gender roles and their relation on mental health as an elusive form of gendered societal burden related to expectations of masculine and feminine expressions. We encourage future analyses to further explore the role of sexual orientation and gender identity relative to gender role in mental health research.

Another force that can influence the relationship between gender roles and mental health is the nature of what characterizes personality traits as stereotypically “feminine” or “masculine”. Indeed, traits are classified as feminine and masculine based on their prevalence in each sex within a given society at a particular moment in history (25, 26). However, anxiety and depressive symptoms tend to be more prevalent in females than in males (11, 12). Overall, boys and men experience more restrictive prescriptive stereotypes than girls and women across life, and therefore variations in gender expression are more likely to be punished (95). Then, it makes sense to think that characteristics judged as “feminine” should be associated with poorer mental health. Some literature shows that this is the case (40). In this sense, this “gendered traits burden” based on gendered characteristics would not be based on birth-assigned sex per se. Instead, this could be associated with a protective effect in masculine people; however, this will need to be assessed in future studies that nuance different gender models (e.g., congruence, androgyny, gender-typed) as articulated elsewhere (37).

These two forces combined (i.e., gendered societal burden & gendered traits burden) may explain why feminine men in our study report poorer mental health, as well as explain why gender roles impact women’s mental health less consistently. Studies using validated questionnaires with gendered items like the BSRI and the PAQ might focus on the gendered traits burden only, whereas studies like ours may reflect a combination of both forces, especially as they relate to “cross-typed” gender role profiles. This hypothesis needs more investigation to be clearly assessed, but it underlines the importance of diversity in the methods used to assess gender roles to fully understand their relationship with mental health.

Personality traits have been strongly associated with mental health (63) independent of gender roles, but findings are inconsistent (75, 76, 96). In accordance with existing literature and our hypothesis, neuroticism appeared to be the most important personality trait predictive of mental health (63, 65). Indeed, neuroticism is the strongest factor predictive of both anxiety and suicidal thoughts and behaviors. Neuroticism is also the only significant trait to positively predict depressive symptoms for both sexes. In addition, our results suggest that extroversion is the second most important trait to predict anxiety in both males and females. Specifically, the more a person is extroverted, the less they report anxiety. A recent study also showed the best mental health profiles among androgynous individuals with high scores in both masculinity and femininity, as well as high scores in extraversion, openness to experience, emotional stability, agreeableness, and conscientiousness (97). Note, however, that results in our study are driven by feminine males and must be replicated before making any conclusions.

These results are in accordance with some studies (65). However, the present study did not find an effect of extroversion on depressive symptoms, whereas this trait has been more consistently negatively associated with depression in the literature (63, 98). Otherwise, openness was also significatively associated with suicidal thoughts and behaviors, but only in males. This result highlights that different factors affect suicidality among males and females, and suggest that personality traits, in relation to gender roles and birth-assigned sex, may in part account for the “gender paradox” (i.e. higher women suicide attempts versus higher men suicide completion (5)).

As personality traits are linked to the way people feel, act, and appear to others (62), they may impact the relationship between gender role identity and mental health. Moderation analyses revealed that some personality traits moderate this relationship but only among males, where the effect of gender role identity on mental health is much stronger than in females. Specifically, neuroticism, which is strongly associated with a poorer mental health (65), moderates the effect of the gender role identity on suicidality. Interestingly, it appears that only highly neurotic males display the deleterious effect of cross-typed profile on suicidality. Indeed, males with low levels of neuroticism did not report suicidal thoughts or behaviors, neither when they were highly feminine or highly masculine. Conversely, masculine males reported low suicidality regardless of whether they were neurotic or not. Only feminine neurotic males reported significantly higher suicidal thoughts and behaviors.

Our findings highlight the protective aspect of masculinity against suicidal ideation and self-injury in males that have been consistently reported in the literature (53, 58). Furthermore, they bring a new perspective to the “gender paradox” in mental health. In summary, our results suggest that it is important to consider the core relationships among gender roles, personality traits, and birth-assigned sex collectively when endeavoring to explain gender differences in suicidality. This underlines the fact that it is not only women who display more suicidal thoughts and behaviors, but also men with specific characteristics (i.e., feminine and neurotic). One caveat is that our sample was composed of living people, which limits us from making conclusions about gender roles in relation to suicide completion that is an essential endpoint of the “gender paradox” in suicide. From this perspective, specific gender roles and behaviors may indeed influence some suicide-related characteristics, like the use of lethal means that have been proposed to explain why men complete more suicide than women (99, 100). Further research is required to fully understand the impact of specific gender roles in relation to suicide completion to explain higher prevalence in men.

On the other hand, our analyses revealed that extroversion moderates the relation between gender role identity and anxiety. A similar moderation effect was detected for masculinity alone. It appears that extroverted people (i.e., with high levels of extroversion) have low anxiety scores, regardless of whether they are feminine or masculine. Gender role identity differences in anxiety were only observed in males with low levels of extroversion. Taken together, only low-extroverted feminine males reported significantly higher levels of anxiety. Extroversion is a trait that depicts our relationship to others the most by “initiating social contacts” (101). In this sense, by expressing their femininity, more extroverted feminine males may be able to experience anxiety levels more akin to masculine males, thwarting the deleterious effect of their cross-typed profile. Conversely, feminine characteristics of males that are less extroverted may exacerbate their anxiety by concealing a part of them that they perceive society reprimands and censors. This hypothesis needs to be more thoroughly investigated in future studies in cross-cultural contexts.

Similar to our first moderation analysis described above, these results put forward the protective effect of masculinity against anxiety which has been consistently observed in the literature (21, 29, 36, 102). However, our results underline the importance of considering personality traits to better understand the circumstances in which this protective effect manifests itself in diverse contexts. Interestingly, personality traits that moderate the relationship between gender role profiles and mental health are not the same depending on the psychiatric symptom or construct being studied. Considering that anxiety is an important risk factor for suicidal behaviors (10, 103), our study highlights that restricting research to gender roles and mental health without taking into account the influence of different personality traits might miss important connections with mental health.

The present study has limitations worth discussing. In contrast to previous studies that assessed gender roles using inventories of gendered personality traits (e.g., the Bex Sex Role Inventory, Personal Attributes Questionnaire), participants directly assessed their gender roles broadly. Other gender role instruments were also originally constructed to assess adaptive androgyny defined as high masculinity and high femininity. As gender roles can depend on society and how people are socialized, the concepts of what is deemed feminine/masculine may differ from one culture to another. Even if recruiting people from the same geographic location (i.e., Rockland town) might reduce confounding, further studies should be conducted in areas where conceptualizations of masculinity and femininity are different to assess if these results generalize to other cultures and environments.

Furthermore, the present study only considered gender role identity as a continuum. As such, results are limited to cross-typed and sex-typed profiles (i.e., feminine and masculine people), without looking at people presenting high or low levels of both masculinity and femininity (i.e., androgynous and undifferentiated people). To address this limitation, we conducted analyses with both continuous and categorical approaches (data not shown). Unfortunately, categorical approaches did not provide more information than the continuous approach we reported, so we did not include these in our report. This absence of significant differences may be due to the measurement methods. Indeed, gender roles can be conceptualized along a single continuum when participants self-report (e.g., being more masculine will decrease their femininity), resulting in a limitation of the number of androgenous and undifferentiated profiles. This is in contrast to Bem’s theory that masculinity and femininity are two independent continuums; however, this assumption has nothing to do with how people self-report with their own personal assumptions of their gender role identity.

Another limitation is that the present study considered suicidality as a continuum from no suicidal thoughts or behaviors to the presence of both. This decision was made with analytic consideration for the low rate of suicidal behaviors in the studied sample from general population. Suicide scores were then analyzed as a continuous variable, preventing the dissociation between suicidal thoughts only and suicidal behaviors. However, literature has shown that risk factors for suicidal ideations differ from those for suicide attempt (104). In addition, the current study cannot provide information for the relationship between gender roles, personality traits, and suicide completion. Finally, even if the present study emphasizes the effect of gender role identity and personality traits on mental health, it does not assess which factors cause or influence the other. Another issue is potential confounding related to neuroticism that is associated with anxiety and depression (105) while also interacting with gendered behaviors (106). Further longitudinal studies and mediation analysis need to be conducted to address causal pathways, especially with designs assessing lifespan development.