This study contributes new understandings of the associations between masculinity and suicidal ideation among adolescent males. Specifically, we found evidence that some dimensions of masculinity were associated with suicidal ideation: notably high conformity to violence and self-reliance among adolescents at 15–18 years was associated with higher odds of reporting suicidal ideation at 17–20 years, and higher conformity to norms related to heterosexuality was associated with lower odds of reporting suicidal ideation. To our knowledge, this is one of the first studies to quantitatively examine the associations between masculinity and suicidal ideation among young males.
The results for violence indicate that each unit increase in conformity to violence was associated with 23% higher odds of reporting suicidal ideation. This is consistent with evidence showing that violent behavior is a risk factor for suicide in adults [17]. While as a whole, the sample did not overwhelmingly endorse violence as a mechanism to solve problems, those who did were more likely to think about suicide, illustrating the risky nature of this dimension of masculinity. There is some evidence that males may be at greater risk of death by suicide because they are socialized to conform to certain masculine norms that foster engagement with painful and provocative life events, resulting in greater ‘acquired capability’ for suicide [26]. It is also known that, compared to females, males are more likely to die by suicide using violent means [61]. Importantly, while hegemonic masculinity values physical strength and toughness, it does not equate with physical violence: violence, however, is sometimes used to demonstrate this physical strength and toughness [8].
The associations observed in this analysis for self-reliance show that each unit increase in self-reliance is associated with 40% increased odds of reporting suicidal ideation. These results are concordant with other work among adults, where self-reliance has been associated with suicidal ideation [49] and mental health problems [43]. The self-reliance items used in this scale tap into affective and behavioral responses to help-seeking. Critically then, these results reveal that some of the young men in this sample reporting suicidal ideation have also reported high conformity to norms that indicate resistance to help-seeking. On face-value, self-reliance can be a positive attribute if it fosters independence, however, the potentially positive effects of self-reliance may be circumscribed if it also inhibits communication and help-seeking in times of distress or crisis. Mental health stigma is known to be a barrier to help-seeking, however recent work highlighted that for young men, this is perhaps more keenly experienced because poor mental health, and help-seeking are both at odds with their internalized masculine norms [39].
Our findings that heterosexual presentation was associated with reduced suicidal ideation (20% lower odds for each unit increase in conformity to heterosexual presentation) was unexpected. The results may evince the protective effect of conforming to socially condoned norms. Hegemonic masculinity is clearly heteronormative. Alignment with dominant masculine norms, and more pointedly, the knowledge that one does not deviate from this heterosexual norm, is likely to confer some level of protection for young males’ mental health. The obverse of the relationship that we observed between high endorsement of heterosexual norms and low suicidal ideation is that low endorsement of heterosexual norms is associated with increased odds of reporting suicidal ideation. Such results do not indicate that being heterosexual is protective, but rather, highlight: firstly, the broader buffering effect of conforming to heterosexual masculine norms; and secondly, the potential to avoid the penalties that arise if deviating from socially accepted norms.
Given that there is a well-established literature documenting the fact that sexual minority young men and adolescents are at elevated risk of suicide and self-harm [20, 51, 55], we conducted sensitivity analysis in which we firstly controlled for sexual orientation, and then restricted our sample to heterosexual-identifying adolescents. The results persisted in both sets of sensitivity analysis. It is widely accepted that heterosexuality is a core component of constructions of hegemonic masculinity [14, 32] and that “to a greater or lesser extent hegemonic masculinity is constructed as a gender position that is as much ‘not gay’ as it is ‘not female’” ([32], p. S113). Certainly, the importance of heterosexuality among adolescents in this sample was evidenced by the high mean score for that subscale, although we note that the large standard deviation suggests that this was not uniformly endorsed. Compared to 18–55 year-olds in the same dataset [49], the adolescents in this sample expressed greater conformity to heterosexual norms. The practice of constructing and affirming masculinity through the assertion of heterosexuality among young males has been observed elsewhere [23]. Froyum’s research revealed the ways that adolescent males (and females) disassociate themselves from homosexuality and other non-heterosexual sexual identities to construct and affirm their heterosexuality [23].
Given that they are situated outside the hetero-normative bounds of hegemonic masculinity, it is not surprising that those young males in our sample not conforming to the hegemonic norm of heterosexual presentation fare less well. This is problematic, not only for the negative impact on young people who are not heterosexual, but also because homophobia, and or the fear of being thought to be gay, can act as a barrier to intimacy among men [14]; something that may impart other negative impacts on them in the future, even if not observed now.
These findings suggesting that certain elements of masculinity may place young men at risk of suicidal ideation have implications for suicide prevention programs among adolescent boys. At a broad level, these results indicate the pervasive power of social norms in defining consensual expectations about what group members do, and should do [11]. Previously, gender norms have been examined in relation to the way they delimit the roles, autonomy and control that women have over their lives. Yet, it is increasingly recognized that gender norms and attitudes may also underpin adverse health behaviors and outcomes in boys and men [34, 35]. Recognizing the challenging and often conflicting messages that young adolescents face regarding masculinity is vital [40]. There is clearly a need to destigmatize mental health, and also foster new understandings of masculinity that incorporate help-seeking into masculine ideals [39].
Jewkes et al. [31] proposed an ecological approach to the transformation of masculinities in adolescents. Such an approach seeks to understand and address the drivers of social norms at all levels- societal, institutional (such as schools), interpersonal and individual - and should seek to understand how different factors or identities might intersect [31]. Jewkes et al. [31] proposed that interventions must move away from one dimensional, homogeneous depictions of masculinity. Drawing on Connell’s work [14, 16], they proposed that interventions should emphasize the heterogeneity of masculinity, avoid stereotypes, focus on similarities between men and women, engage with, and acknowledge fears and vulnerabilities and address homophobia [31]. It is possible that relaxing rigid norms regarding masculinity and encouraging acceptance of more diverse masculinities will deliver benefits beyond suicide and mental health improvements for adolescent boys, and contribute to better health and wellbeing in the wider population [22, 36].
Further work is needed to examine the associations observed in this analysis over time (as more waves of data become available), and across different age groups, to understand whether these associations reflect cohort effects, or developmental stages. It is also crucial to understand how these changes relate to other personal, occupational and well-being indicators, regardless of the drivers. Furthermore, it is possible that specific dimensions of masculinity are both protective and also risk factors at different developmental stages: the dynamic nature of these associations needs to be understood.
Dismantling masculine norms to facilitate help-seeking is vital, however it is also important that once adolescent males do seek help, mental health services are available and appropriate to meet their needs – ideally person-centered approaches that acknowledge the diversity of men, and the diversity of their needs [56].
There are several strengths of this analysis. We used a large sample of Australian male adolescents, which strengthens the basis for statistical inference. We also note the use of a validated measure of masculinity. The original 94-item CMNI had good construct validity, and discriminant validity, and the 22-item instrument has been shown to correlate well with the original scale [60]. Additionally, the CMNI-22 has been widely used and is positively regarded. While in this study, basic psychometric parameters of the CMNI were analyzed (i.e. factor structure, item loadings and internal consistency), it should be acknowledged that an exhaustive evaluation of its psychometric properties has not been conducted on an Australian adolescent sample. We also raise the possibility that the construct validity of some of the subscales was low: in particular, it is possible that the items for heterosexual presentation do not align with the construct they are purported to measure in this population (potentially underpinning the findings for heterosexual presentation).
We note that significant measurement differences across ethnic groups have been observed, with evidence that the scale is more theoretically consistent for White American men compared to Asian Americans [30]. Other evidence also indicates that masculinity may be understood, experienced and expressed differently across different ethnic groups and cultural contexts [27], thus the results may not be generalizable to Indigenous Australians and Australians of ethnic minority backgrounds. Further work using qualitative methods is needed to examine how conformity to masculinity may vary across ethnicity and Indigenous identity in Australian adolescents.
As both masculinity and suicidal ideation were self-reported, dependent measurement error, which can arise when two or more variables are based on self-reported subjective responses from the same respondent [63], may have biased findings.
While there is a precedent for the use of single-item suicidal ideation measures [44, 48], there is some evidence that single-item measures may result in a higher proportion of false positives and false negatives [42]. If this occurred in this study, there is potential that it led to some degree of misclassification bias and potentially spurious findings. A further limitation of single-item measures such as this one is that they inadequately capture differences in the severity or frequency of ideation or attempts [42]. It is impossible to ascertain, for example, whether a person was actively planning to engage in suicidal behavior, whether a suicide behavior was stopped, or whether these were simply fleeting and non-serious thoughts.
Selection bias due to missing data potentially affected these results, although we note that for most variables, there was < 4% missing from the eligible sample, and this is unlikely to have biased results. Because our sample comprised of adolescent males (with no parent-reported information on household income or occupation), common socio-economic confounders were either unavailable/not obtained (household income, parental occupation), or not yet realized (educational attainment). We were therefore unable to include these possible confounders of the relationship between masculinity and suicidal ideation in analytical models. This may have introduced some bias, as there is evidence that constructions of masculinity differ across socio-economic position (SEP) [14], and that more disadvantaged groups are at greater risk of suicide [28]. However, we note that while we were unable to control for individual SEP, we did control for area SEP. Given evidence that for adolescent males, enactments of masculinity [12] and suicidal behaviors [5] are known to vary by neighborhood/area deprivation, we contend that by controlling for area SEP we have captured much of the confounding that may have been induced by individual SEP.
