Abstract
Crime is, and has always been, a disproportionately male phenomenon. In recent years, the importance of understanding how masculinity is associated with maladjusted behaviors has risen notably. We aimed to explore the association between conformity to masculine gender norms, with sexuality as a domain of masculinity, and deviant behavior, as well as the association between negative and positive childhood experiences and deviant behavior. Two hundred and thirty-one Portuguese 18- and 19-year-olds completed an online questionnaire that included a sociodemographic questionnaire and four scales (Adverse Childhood Experiences Scale, Benevolent Childhood Experiences Scale, Conformity to Masculine Norms, and Deviant Behaviour Variety Scale). The adverse childhood experiences and conformity to masculine norms were positively associated with deviant behavior, whereas benevolent childhood experiences showed a weak negative association. Notably, conformity to masculine norms predicted deviance at a level comparable to adverse childhood experiences. These findings highlight the importance of understanding deviant behavior among young men as embedded within broader processes of gender socialization, masculinity performance, and sexuality, with implications for gender-sensitive prevention efforts targeting male youth.
Public significance statement
Our findings have practical implications not only for parents, in regard to how boys are being raised and socialized, but also for justice systems that, traditionally, have not acknowledged gender norms in their crime prevention approaches.
Introduction
Men, especially young men, are much more likely than women to commit criminal offenses (Campaniello and Gavrilova, 2018; Hoeve et al., 2009; Messerschmidt, 2018). This is one of the most consistent findings in criminology (McFarlane, 2013). Some authors have even argued that this is the most robust finding in all criminological literature (Bartusch and Matsueda, 1996). This is supported by data from official criminal justice systems worldwide and self-reports—which also consistently reveal that men report significantly higher levels of criminal behavior (Cusson, 2006). Between 2008 and 2018, 19 in 20 prisoners in the European Union were men (Eurostat, 2020). While in Portugal, in 2020, 86% of youths placed in juvenile detention facilities (formally known as Centros Educativos) were male (Direção-Geral de Reinserção e Serviços Prisionais, 2021), and, in adulthood, the proportion is even higher, with 94% of convicted and imprisoned persons being male (Direção-Geral de Reinserção e Serviços Prisionais, 2020). This disparity between men and women is known as the gender gap (Smith and Paternoster, 1987).
Gender and crime: Sociological factors
In attempts to explain this gender gap, some research has explored biological correlates of criminal behavior. A popular explanation, based on biological factors, claims a positive relationship between levels of testosterone and violent or aggressive behavior (Mims, 2007). A meta-analysis based on 45 independent studies (N = 9760) published between the early 1970s and late 1990s did find a positive, but weak, relationship between testosterone levels and violent behavior in humans (Book et al., 2001). However, a reanalysis of the same meta-analysis done by Archer and colleagues (2005) concluded a lower mean weighted correlation than initially reported (r = .08 vs r = .14). A systematic review of 27 studies done by Duke and colleagues (2014) found no evidence supporting the notion that testosterone levels were significantly associated with mood or behavioral changes in adolescent males. Instead, the literature has pointed to an association between testosterone and behaviors involved in obtaining and maintaining high social status in males, rather than aggressive or antisocial behavior, per se (Josephs et al., 2006). Dreher and colleagues studied the effects of testosterone in healthy young males by injecting testosterone enanthate or a placebo in a double-blind, between-subjects, randomized design (n = 40) and concluded that testosterone can reinforce aggressive responses to provocation, but it can also trigger prosocial behavior that is appropriate for increasing status. Evidence suggests that associations between testosterone and aggression are weak and heavily context-dependent, and the explanatory power of biological factors alone appears limited.
If testosterone-related effects on aggression are contingent on context and what is socially defined as high status, it becomes necessary to examine how such definitions of status are themselves shaped by gendered social norms.
Connell’s (1995) conceptualization of masculinity provides a foundational framework for understanding gender as a socially constructed and relational practice rather than a fixed set of traits. Within this framework, masculinity is organized hierarchically, with hegemonic masculinity representing culturally dominant and socially legitimized forms of manhood that privilege traits such as heterosexuality, dominance, emotional control, and authority over women and subordinated forms of masculinity. These masculine ideals are not universally attainable but instead function as normative benchmarks against which men are evaluated. Crucially, men who lack access to hegemonic or dominant non-hegemonic masculinities may seek alternative ways to demonstrate masculine status.
Within this framework, masculinity can be understood as a set of culturally sanctioned norms. According to Thompson and Pleck (1986), these norms are based on three primary components: toughness, antifemininity, and status. Toughness relates to the idea that men must be physically strong, emotionally callous, and behave aggressively. Antifemininity refers to the idea that, to be masculine, one must reject qualities conventionally considered feminine, such as emotionality or helping behavior. Status relates to the notion that men need to work toward achieving power, agency (e.g. in social or financial matters), and the respect of others.
Vandello and colleagues’ (2008) theory of precarious manhood further conceptualizes masculinity as a socially constructed status that, to be maintained, must be actively pursued and continuously validated, rather than passively acquired through biological maturation. Across five studies, the authors demonstrate how manhood is widely perceived as both elusive and tenuous. That is, men are seen as having to earn their gender status through social proof and that this status is easily lost through failures to meet socially valued expectations of masculinity. Their findings show that threats to masculinity elicit heightened anxiety and physically aggressive thoughts among men, whereas comparable threats have no such effects among women. Together, these studies illustrate how masculinity is maintained through public demonstrations of toughness, dominance, and risk-taking, and why conformity to masculine norms (CMNI) may be associated with behavior, such as aggression or deviance, that may function to restore threatened masculine status.
Building on Connell’s framework, Messerschmidt’s (2018) work conceptualizes crime and deviance as potential sources of masculinity in contexts where access to hegemonic masculine roles is constrained. Rather than viewing deviance as inevitable, Messerschmidt’s (2018) perspective emphasizes that criminal behavior may emerge as a strategy through which men attempt to assert power, dominance, heterosexuality, or status when other socially legitimate ways are not available. Drawing on developmental case analyses, Messerschmidt (2004) further demonstrates how these masculinity strategies take shape during adolescence in the context of early life constraint and limited access to valued masculine roles.
Sexuality constitutes a central dimension through which masculinities are constructed and regulated. Within hegemonic and dominant non-hegemonic masculinities, heterosexuality has historically functioned as a key marker of masculine legitimacy, while homosexuality has been positioned as subordinate or stigmatized (Connell, 1995; Connell and Messerschmidt, 2005). As a result, sexual identity and behavior play an important role in the policing and performance of masculinity, particularly during adolescence. Recent literature has indicated that norms surrounding homosexuality may be shifting among younger cohorts, allowing for greater variability in masculine expression (Anderson and McCormack, 2018). Situating sexuality within masculinity theory is therefore essential for interpreting its relationship with childhood experiences, CMNI, and deviant behavior. In this study, sexuality is therefore examined not as an independent outcome, but as a key component through which masculine norms are enacted, contested, and possibly associated with deviant behavior.
Childhood factors and crime
Masculinity is learned and negotiated through early socialization processes, making childhood a critical context for the development of gendered behavior. Adverse childhood experiences (ACEs) and benevolent childhood experiences (BCEs) may shape boys’ access to emotional support, models of behavior, and, ultimately, strategies for achieving masculine status. This position situates childhood experiences as integral to understanding the relationship between masculinity and deviant behavior.
Childhood maltreatment has long been associated with antisocial behavior. Braga and colleagues (2018) conducted a meta-analysis of longitudinal studies involving 20,946 participants, demonstrating that children and adolescents who had suffered maltreatment were nearly two times more likely to exhibit antisocial behavior than their non-maltreated peers. Another meta-analysis on the association between parenting and delinquency by Hoeve and colleagues (2009) found that rejection, hostility, and neglect had the most robust link to delinquency.
In 1998, Felitti and colleagues proposed the ACE concept, which encompasses a host of environmental factors, such as childhood abuse, household dysfunction, family mental health, domestic violence, and family criminal behavior. Leban and Gibson (2019) studied these ACEs in connection with delinquency and substance abuse, with a focus on gender differences. The authors found a significant relationship between ACEs and delinquency for boys, especially.
Because most of the research on delinquent behavior focuses on risk factors, which are undoubtedly vital in understanding any phenomenon and predicting it, the protective or promotive factors tend to be left out of the question (Farrington et al., 2012). Moreover, it is harder to change or reduce risk factors themselves; hence, it is relevant to explore and promote protective factors as a complementary or alternative approach to offending risk reduction (Simões et al., 2008).
To aggregate positive influences into a brief and effective index and to match the already existing ACEs scale, Narayan and colleagues (2018) created the BCEs scale, which is designed to assess positive early life experiences in individuals with histories of childhood maltreatment and other adversities. Higher levels of BCEs have been found to be associated with lower odds of psychological distress, sociodemographic risk, and parenting stress in homeless parents (Merrick et al., 2019), as well as with a lower severity of post-traumatic stress disorder (PTSD) and depression, and positively associated with prosocial behavior (Zhan et al., 2021). In addition, according to Guarda and Almeida (2020), BCEs were also found to predict empathy in adults.
Objectives and hypotheses
Considering the literature review, the main goal of this study is to explore the role of CMNI in deviant behavior. Second, we aim to explore the association between ACEs, BCEs, and deviant behavior.
As this is an exploratory study, no formal hypotheses were formulated. However, based on the previously highlighted research, four exploratory hypotheses were devised:
Hypothesis 1. The CMNI will be positively associated with deviant behavior.
Hypothesis 2. The ACEs will be positively associated with deviant behavior.
Hypothesis 3. The BCEs will be negatively associated with deviant behavior.
Hypothesis 4. The CMNI, ACEs, and BCEs will prove to be significant predictors of deviant behavior.
In addition, we tested the association between sociodemographic characteristics and the four main variables (CMNI, ACEs, BCEs, and DBVS).
Method
Participants and data collection
Participants were 231 male adolescents from Portugal, of whom 128 (55.4%) were 19 years old, and the remaining 103 (44.6%) were 18 years old. The age range was selected to ensure alignment with the validation criteria of the instruments used, as the Deviant Behaviour Variety Scale (DBVS) is validated for individuals up to 19 years of age (Sanches et al., 2016) and the ACEs scale is designed for individuals 18 and older (Original version: Felitti et al., 1998; adapted by Silva and Maia (2008)).
The study protocol was approved by the Ispa—Instituto Universitário Ethics Committee. Following its approval, the protocol was constructed on the Google Forms platform (www.docs.google.com/forms).
Participants were recruited via paid advertisements on Facebook (www.facebook.com) using Meta’s advertising platform. Advertisements included a brief description of the study and a link to the questionnaire. They were displayed to users based solely on age (18–19 years) and geographic location (Portugal). No targeting based on membership of specific groups, interests, or behavioral characteristics was applied. Recruitment took place over a period of 2 weeks, during which individuals who clicked on the advertisement were directed to an informed consent page prior to accessing the survey itself.
As seen in Table 1, most participants (88.3%) had completed secondary/high school. Regarding their parents’ educational level, there were more fathers with only high school level education than mothers (32.4% vs 25.0%). Mothers had higher degrees (41.1% with a college degree) than fathers (29.2%).
Distribution by educational level of participants and their mothers/fathers.
Concerning occupation, most participants (77.9%) described themselves as full-time students.
Regarding sexual identity, most (70.6%, n = 163) identified as heterosexual (Table 2). Nevertheless, when asked about sexual attraction, approximately only one in two described themselves as exclusively heterosexual, that is only being attracted to individuals of the opposite sex (53.2%, n = 123), as seen in Table 3.
Distribution by sexual identity.
Distribution along the sexuality spectrum.
Instruments
Sociodemographic questionnaire
Participants described their sexual identity from three standard classifications (“heterosexual,” “bisexual,” and “homosexual”). They were also asked to describe their sexual attraction on a 7-point sexuality spectrum scale; one end represented exclusive sexual attraction to women (0), and the other was exclusive to men (6). The middle point denoted equal attraction to both men and women (3). There was also an option for participants who experienced no sexual attraction, and another one for those who were not sure. Participants could submit the questionnaire without answering the sociodemographic questions.
BCEs Scale
The BCEs Scale (original version: Narayan et al., 2018; validated to the Portuguese population by Almeida et al. (2021)) is a 10-item self-report checklist assessing positive relational and environmental experiences. Items capture experiences such as feeling safe with at least one caregiver, having supportive friendships or adults, enjoyment at school, and experiencing predictable home routines. Answers are given dichotomously (yes or no), and positively endorsed items are summed, resulting in a total benevolent experiences score.
ACEs Scale
Like the BCE scale, the ACEs Scale (original version: Felitti et al., 1998; adapted by Silva and Maia (2008)) is a self-report scale but focused on evaluating adverse childhood experiences. There are 17 items (answers are also given dichotomously), and those that are endorsed are summed, resulting in a total ACEs score. The instrument comprises 10 dimensions: emotional abuse, physical abuse, sexual abuse, emotional neglect, physical neglect, exposure to domestic violence, parental separation or divorce, household substance abuse, household mental illness, and incarcerated household member.
Conformity to Masculine Norms Inventory
The Conformity to Masculine Norms Inventory, 22 items (CMNI-22; original version: Mahalik et al., 2003; adapted by Leitão (2015)) is a self-report scale to examine the conformity to masculine gender norms in Western societies. This instrument differs from other measures of gender norms by assessing not only cognitive endorsement of masculine norms (“Men and women should mutually respect each other as equals”) but also self-reported behavioral alignment with these norms (“I tend to share my feelings”). The CMNI-22 has two items per dimension: winning, emotional control, risk-taking, violence, power over women, dominance, playboy, self-reliance, disdain for homosexuals, pursuit of status, and primacy of work. Individuals answer on a 4-point Likert-type scale, ranging from 0 (strongly disagree) to 3 (strongly agree). Cronbach’s alpha was .70 in the male group in the Portuguese adaptation study (Leitão, 2015). In the present study, Cronbach’s alpha was .64.
DBVS
The DBVS (Sanches et al., 2016) is also a self-report measure that comprises 19 items that pertain to a variety of deviant behaviors. Answers are given dichotomously (yes or no), and positively endorsed items are summed, resulting in a total DBVS score. Items are grouped into minor infractions and serious infractions. In the original study, Sanches and colleagues (2016) reported high internal consistency (Cronbach’s α of .83), and in the present study, we obtained a Cronbach’s α of .86.
Data analysis
All statistical analyses were conducted using the software IBM SPSS Statistics (version 27.0) (IBM SPSS, Chicago, IL, USA). Descriptive statistics were used to examine the distribution of responses for each item. Internal reliability was estimated through Cronbach’s alpha. Afterward, correlation coefficients were also calculated to measure the correlation between the sets of data from the variables, namely the Pearson correlation coefficient and Spearman’s rank correlation coefficient as appropriate. Kruskal–Wallis tests were also utilized to examine the distribution of our main variables across sociodemographic variables. All statistical tests were two-tailed, at p < .05. Finally, we executed multiple linear regressions to explore how our main variables could predict deviant behavior.
Results
We applied the Portuguese version of the instruments. The following tables contain English translations of the items, but should not be considered as being validated for English-speaking populations.
BCEs
Every item of the scale had a proportion of yes answers above 60%, as shown in Table 4.
Percentage of positive answers by each BCE item.
ACEs
Regarding adversity, experiences of emotional and physical abuse were the most common ones, as seen in Table 5. Approximately one in three (29.4%; n = 68) reported being struck with enough physical force to leave marks at least once. Similarly, one in three (29.4%; n = 68) also reported having parents with psychological problems or who had gone through suicide attempts.
Percentage of positive answers by at least one item of each ACE dimension.
CMNI
CMNI’s items ranged from 0 (completely disagree) to 3 (completely agree). As can be seen in Table 6, participants scored highest in the pursuit of status, emotional control, and primacy of work dimensions. Conversely, dimensions with the lowest scores were power over women, playboy, and disdain for homosexuals.
Descriptive statistics for the CMNI dimensions. sorted by the median.
DBVS
More than half of participants reported lying to adults (82%) and skipping class (56.3%), both minor infractions. 37.7% had used public transportation without paying for a ticket and 32.5% had stolen something worth less than 5€ (minor infractions). The most common serious infractions were: driving a motorbike or car without a driver’s license (26%), carrying a weapon (21.2%), and stealing something worth between 5 and 50€ (17.7%).
Correlations
In regard to educational level, there was a significant negative correlation between participants’ educational level and deviant behavior (rs = −.220, p < .001), 95% confidence interval (CI): [−.339, −.094]. The mother/female caregiver’s educational level was positively associated with the son’s educational level (rs = .220, p < .001), 95% CI: [.094, .339], while the father/male caregiver was not. The participants’ educational level was also negatively correlated with their CMNI score (r = −.151, p = .022), 95% CI: [−.275, −.022], indicating that the higher the educational degree, the lower the CMNI.
We ran the Kruskal–Wallis test to analyze differences in the DBVS across the four categories of occupational status. A significant difference (H(3) = 22.615, p < .001) was found in the student group, with participants in this category reporting lower levels of deviant behavior when compared to all other occupations.
Concerning sexuality, we also ran Kruskal–Wallis tests to analyze if there were differences in the distribution of deviant behavior, conformity, and early experiences across the three sexual identity categories (heterosexual, bisexual, and homosexual). No significant difference was detected for the BCE (H(2) = .164, p = .921) score. However, significant differences were detected in ACEs (H(2) = 7.877, p = .019), with heterosexual-identifying participants reporting less ACEs than bisexual or homosexual individuals. The CMNI scores were also different between the heterosexual and homosexual groups (H(2) = 8.709, p = .013). Furthermore, significant differences were also detected in the distribution of deviant behaviors across sexual orientation categories. The tests revealed that, when compared with the homosexual group, heterosexuals reported a significantly higher rate of perpetration of serious offenses (H(2) = 14.639, p < .001), but no significant differences were detected for minor infractions (H(2) = 1.529, p = .466).
We found three weak but significant correlations regarding sexual attraction. First, a negative correlation with CMNI (r = −.165, p = .013), 95% CI: [−.288, −.037] was detected, meaning that participants who labeled their sexual attraction closer to the exclusively homosexual end of the spectrum were less likely to report CMNIs. Men who reported higher levels of same-sex attraction also reported lower conformity to the dimensions: power over women (r = −.260, p < .001), 95% CI: [−.380, −.140]; violence (r = −.259, p < .001), 95% CI: [−.380, −.134]; and disdain for homosexuals (r = −.231, p < .001), 95% CI: [−.350, −.105]. There was a positive, significant Spearman correlation between sexual attraction and ACEs (rs = .176, p < .001), 95% CI: [.049, .300], revealing that participants with higher levels of same-sex attraction reported more adverse childhood experiences. Probing the correlation with ACEs, the two significant, but weak, associations were with sexual abuse (rs = .197, p < .001), 95% CI: [.070, .317] and with emotional abuse (rs = .183, p < .001), 95% CI: [.056, .304]. Similarly to our sexual identity findings, a negative (rs = −.216, p < .001), 95% CI: [−.335, −.090] Spearman correlation was detected between sexual attraction and DBVS, but only when looking at the serious infractions category, indicating that participants closer to the exclusively heterosexual end of the spectrum were more likely to report serious offenses.
Regarding our first exploratory hypothesis, a significant correlation between CMNI and deviant behavior was found (r = .302, p < .001), 95% CI: [.180, .415]. The dimensions of the CMNI scale most correlated with deviant behavior were risk-taking (r = .281, p < .001), 95% CI: [.158, .400]; violence (r = .264, p < .001), 95% CI: [.140, .380]; playboy (r = .230, p < .001), 95% CI: [.104, .349]; and power over women (r = .142, p = .031), 95% CI: [.013, .266]. There were no significant correlations between early life experiences and masculine norms assessed with the BCE, ACE, and CMNI (r = .064, p = .332), CI: [−.065, .191] and (r = .033, p = .620), CI: [−.096, .161], respectively.
Our second hypothesis was also confirmed. A significant and positive correlation between ACEs and deviant behavior was also found (r = .341, p < .001), 95% CI: [.222, .450]. The ACE dimensions most associated with deviant behavior were physical neglect (r = .286, p < .001), 95% CI: [.163, .400]; domestic violence (r = .257, p < .001), 95% CI: [.132, .374]; emotional neglect (r = .256, p < .001), 95% CI: [.131, .373]; and physical abuse (r = .226, p < .001), 95% CI: [.100, .345]. As expected, ACEs and BCEs were negatively correlated (r = −.395, p < .001), 95% CI: [−.500, −.280].
Finally, data also supported our third hypothesis. A weak but significant negative correlation between BCEs and deviant behavior was found (r = −.139, p = .035), 95% CI: [−.263, −.010].
Regression model
Regarding our fourth hypothesis, a multiple linear regression analysis was calculated to assess if and how conformity, adverse, and benevolent experiences predicted deviant behavior.
A significant regression equation was found (F(3,227) = 19.091, p < .001), with an adjusted R2 of .191. The ACEs and CMNI were significant predictors of deviant behavior, whereas BCEs were not.
Therefore, we calculated a second model, excluding the BCEs score, which also resulted in a significant regression equation (F(2,228) = 28.612, p < .001), with an adjusted R2 of .194. The corresponding coefficients can be seen in Table 7.
Linear regression coefficients.
Discussion
Statistical analyses revealed that the data substantiated our four exploratory hypotheses.
Results showed a statistically significant correlation between CMNI (risk-taking, violence, and playboy) and deviant behavior. These three traits, out of the total 11 dimensions, could perhaps most closely be described as impulsive and lacking in self-control (risky behavior, lashing out in violent ways, and multiple sex partners), reminiscent of Gottfredson and Hirschi’s theoretical approach to crime, published in 1990. Often referred to as the “General Theory of Crime” (GTC), it spotlights self-control as the central variable in explaining crime. Interestingly, according to Schulz’s (2005) analysis of GTC, when self-control is split into six different variables, risk-taking stands out as the most predictive of criminal behavior, aligned with our findings. The power over women dimension also correlated with deviant behavior. The prominence of these dimensions as predictors of deviant behavior is consistent with theoretical conceptions of masculinity as a hierarchical and performative gender system. Within hegemonic and dominant forms of masculinity, risk-taking may function as a culturally sanctioned way of demonstrating toughness, fearlessness, and status, while power over women reflects broader gendered relations of dominance (Connell, 1995; Messerschmidt, 2018). The present findings suggest that engagement in risk-taking and dominance-oriented behaviors may serve as situational strategies for asserting masculine status, particularly among young men navigating social, developmental, and structural constraints.
The ACEs were significantly correlated with deviant behavior. Bellis and colleagues (2016) found that, compared with people with no ACEs, those with 4+ ACEs were 15 times more likely to be perpetrators of violence in the last 12 months and 20 times more likely to have been incarcerated in their lives. The present study found that neglect (physical and emotional) had the strongest correlations with deviant behavior. As frequently reported, child neglect increases the risk of future delinquency because of a lack of parental monitoring and parental rejection (Ryan et al., 2013). Exposure to domestic violence presented the second-strongest correlation with deviant behavior, and physical abuse also showed a significant association with deviant behavior. Physical abuse has long been recognized as a risk factor for delinquency (Bentrup, 2020). However, our findings regarding exposure to domestic violence are of particular relevance. Many countries have criminalized domestic violence, but the exposure of children/young people to domestic violence as a form of child abuse has not yet been regarded as an offense or a crime.
The BCEs were significantly correlated with less deviant behavior, albeit with a very weak correlation. Because this study gathered data from 18- and 19-year-old participants, one possible explanation for the weakness of the correlation may be that its effects are not yet noticeable, since participants are only now entering adulthood. Considering that BCEs have positive long-term effects, it is possible to assume they will play a more decisive role later throughout adulthood (Merrick and Narayan, 2020), which may be better appreciated in a study with a broader age range.
Regarding our regression model, we found that ACEs in conjunction with CMNI explained deviance in 19.4%, with the ACEs contributing slightly more than CMNI. These are encouraging results since our findings indicate that CMNI predicts deviance as substantially as another universally recognized risk factor: ACEs.
When compared to the pilot study that validated the DBVS (Sanches et al., 2016), we encountered a higher prevalence of almost every type of deviant behavior; however, this was expected since we purposefully focused this study on the “peak” of the deviant/criminal activity curve when, in contrast, the original study had participants as young as 14.
Regarding sexuality, namely the sexual attraction item, one notable finding was that one in two participants declared some level of attraction for the same sex, confirming a similar trend found in other Western countries—younger people are much more likely to report same-sex attraction. For example, a 2015 British survey of the general public (using a similar 7-point scale) found that 72% of adults also identified as exclusively heterosexual; however, when looking only at young people’s (18- to 24-year-old) responses, that percentage was 46% (YouGov, 2015). In a survey replication, done in 2020, the percentage of younger people identifying as “100% heterosexual” had dropped to 37% (YouGov, 2020). When comparing these numbers, it is also important to note that our data only reflect the sexual attraction feelings of males, whereas the British surveys include both sexes. This is relevant because virtually in all sexuality surveys, females, across all ages, are more likely than males to report feelings of same-sex attraction (Diamond, 2016).
Traditional conceptions of hegemonic masculinity have regarded heterosexuality as a core component of masculine status, therefore, rendering same-sex attraction as a potential source for stigma and marginalization. However, some research suggests that declining homophobia among younger cohorts may enable greater openness in sexuality without necessarily undermining masculine status (Anderson and McCormack, 2018). From this perspective, the present findings may reflect a shift in masculine norms. Importantly, these dynamics underscore the need to understand sexuality not as an isolated variable, but as a part of a broader process of gender socialization.
Considering sexual attraction’s associations with our variables, we found that it was positively correlated with ACEs, meaning people closer to the same-sex attraction end of the spectrum reported higher levels of adverse childhood experiences. The ACE dimensions that most correlated with sexual attraction were sexual abuse and emotional abuse. Furthermore, our results also showed that heterosexuals reported fewer ACEs than bisexual or homosexual participants. It is well established that members of sexual minorities tend to be at higher risk for several forms of victimization, particularly sexual and emotional abuse (Balsam et al., 2005; Rothman et al., 2011). In our findings, these were the two ACE dimensions most associated with individuals closer to the homosexual end of the spectrum. Conversely, individuals closer to the heterosexual end reported higher CMNI, but only in three specific dimensions: power over women, violence, and disdain for homosexuals.
Data from the sexual attraction and sexual identity items also revealed a connection between these variables and deviant behavior. Individuals closer to the exclusively heterosexual end of the spectrum and heterosexual-identifying individuals were more likely to report serious infractions. However, there was no difference for minor infractions. This is interesting because homosexual individuals were more likely to report higher ACEs, which is corroborated by previous research (e.g. Andersen and Blosnich, 2013) and, as our research demonstrated, ACEs themselves are associated with a higher score of DBVS. These results lead us to question whether ACEs may not be as strong a predictor of criminal behavior for these individuals, as they are for heterosexual individuals. We also found that the closer the individual is to the exclusively heterosexual end of the spectrum, the more likely he is to endorse norms regarding the use of violence. One could, therefore, assume that the higher endorsement of violent action by heterosexuals would translate into committing more serious infractions (i.e. carrying a weapon, damaging property, physically assaulting others, etc.) or, at least, be more likely to report these behaviors. These findings suggest that heterosexuality operates less as an identity category and more as a mechanism through which certain masculine practices are legitimized. The findings align with theoretical accounts that frame deviant and dominance-oriented behaviors as strategies for accomplishing masculinity (Connell, 1995; Messerschmidt, 2004, 2018; Thompson and Pleck, 1986; Vandello and Bosson, 2013).
Limitations
The CMNI instrument was selected because it is the only Portuguese-adapted scale that measures this construct; however, another limitation of this study is that this instrument still requires further psychometric evaluation. Given this aspect, it is also possible that the relatively small correlation that was observed could be tied to low internal/external consistency.
As already discussed, our sexual attraction findings are relevant on their own, just in terms of sexuality demographics. However, they also could have implications on masculinities, particularly dimensions of disdain for homosexuals and antifemininity. The CMNI scale was developed in 2003, which does not make it a particularly old instrument; however, in Western societies, attitudes toward (homo)sexuality have shifted dramatically since 2003. For example, in 2020, The New York Times described how heterosexual-identifying young men purposefully record themselves engaging in homoerotic activities (i.e. cuddling and kissing other boys) to “look cool” and attract female attention. These actions would be tantamount to “social suicide” in 2003. Professor of masculine studies Anderson posited that because today’s young people grew up in a time of declining homophobia, they are now actively rebelling against the “anti-gay” and “anti-feminine” model of previous masculinities (Hawgood, 2020). Considering that disdain for homosexuals is evaluated by these two items, “It would be awful if people thought I was gay” and “I try to avoid being perceived as gay,” and considering current attitudes, it seems crucial to reconsider the measurement of this concept.
Future research could also explore what variables mediate the relationship between masculinity and deviance. Our results indicate a moderate correlation between conforming to masculine norms and crime—which may suggest the presence of other variables explaining the relationship.
More research is also needed to better appreciate which specific dimensions of masculinity most contribute to various types of deviant/criminal behavior. In addition, future studies should also look at how masculinity in other cultural contexts is correlated with crime.
Our results call for an increased awareness of how males are being raised and socialized. Parents should also consider alternative parenting strategies for boys that do not place such importance on masculine notions such as risk-taking, violence, or dominance.
Footnotes
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.