A Systematic Review of Adolescent Masculinities and Associations with Internalizing Behavior Problems and Social Support

Abstract

Interest in the connection between masculinities and mental health continues to grow. However, no previous systematic review has explored this association for adolescents. We present the systematic review of 29 articles that explore the connection between adherence to stereotypical male gender role norms (e.g., emotional restriction), attributes (e.g., “ambitious”), and identity (most commonly, gender “typicality”) and internalizing behavior problems and social support. A total of 24,795 adolescent boys (6th‐12th grade) were included in the reviewed studies from 1997–2017. In the quantitative articles (n = 20), associations varied by aspect of masculinity assessed. Specifically, we found that greater endorsement of “masculine” traits (e.g., ambitious, assertive) was generally associated with fewer internalizing behavior problems and greater social support. However, lower gender “typicality” and higher adherence to stereotypical gender role norms were generally associated with more internalizing behavior problems and lower social support. In the qualitative articles (n = 9), the most predominant theme was emotional restriction (i.e., a gender role norm) and consequences for mental health. While research in this area is newer for community psychologists, the connection between masculinities and mental health is directly relevant to the field. Given the focus on individual‐level conceptions of masculinity and mental health found in our review, we describe key future directions for masculinities research in community psychology.

Highlights

• Comprehensive review of masculinities and internalizing problems/social support in adolescence.

• Stereotypical masculine attributes were generally associated with better mental health.

• Adherence to stereotypical male gender role norms was generally associated with worse mental health.

• Lower gender “typicality” was generally associated with worse mental health.

• Results suggest pathways to mental health inequities for men may first emerge in adolescence.

Introduction

Poor mental health among adolescents is a pressing community health problem. In the United States, 22.2% of adolescents experience a mental health disorder, and many adult disorders start during the adolescent period (Merikangas et al., 2010). Beyond specific disorders, a substantial minority of adolescents report poor mental health generally; for example, a national survey of Canadian adolescents found that 35–44% of girls and 27–29% of boys in grades 6–10 report that they experienced a high level of emotional problems (e.g., feeling low) in the past six months (Freeman & Luu, 2010). While a substantial body of literature focuses on individual‐level correlates of mental health problems among adolescents, research in community psychology highlights the importance of considering individuals within their broader context when addressing mental health and well‐being (Trickett, 2009). As it pertains to health outcomes, this approach of understanding people in context can be grounded in the application of a social determinants of health framework, where mental health is seen as intrinsically related to things like poverty (instead of individual income level) and racism (as opposed to individual race/ethnicity; Government of Canada, 2019; World Health Organization, 2019). As disparities in mental health status by gender are well‐documented (e.g., Centers for Disease Control, 2017; Freeman & Luu, 2010; Merikangas et al., 2010), this framework is also used to understand these differences, both between‐ and within‐genders (Lohan, 2007).

For men, a growing body of research now indicates connections between aspects of societal expectations for male‐identified individuals and poor mental health (Addis & Cohane, 2005; Courtenay, 2000; Evans, Frank, Oliffe, & Gregory, 2010; Mankowski & Maton, 2010; Wong, Ho, Wang, & Miller, 2017). In this literature, gender is understood as being socially constructed; that is, men act in the ways they do not because of biological traits, but because of the concepts of masculinity that are adopted from their culture (Courtenay, 2000). Specifically, within a framework of social construction, masculinities are not viewed as the biological essence of an individual, but rather, as a set of practices put into action by individuals interacting with others in their environment (e.g., practices such as how one is expected to react to emotional situations; Addis & Cohane, 2005; West & Zimmerman, 1987). In other words, there is no one way of being “a man,” but rather, multiple ways depending on an individual’s given context (Connell, 2005).

The assumption that multiple versions of masculinity exist also corresponds with the view that some masculinities are more powerful than others (Connell, 2005; Connell & Messerschmidt, 2005). At the top of this masculinity hierarchy is a concept referred to as hegemonic masculinity, or the most socially sanctioned form of masculinity in a given time and place (Connell, 2005; Connell & Messerschmidt, 2005). In many Western countries—including the United States and Canada—this form of masculinity “is associated with being White, heterosexual and middle‐class, and possessing stereotypical masculine traits of assertiveness, dominance, control, physical strength and emotional restraint” (Evans et al., 2010, p. 8). Given its association with social power and control, many men may strive for this form of masculinity, though few, if any, will fully achieve it (Connell & Messerschmidt, 2005; Mankowski & Maton, 2010). However, it is in the striving for the hegemonic ideal that pathways to downstream health inequities likely emerge. For example, because of its focus on power, control, and strength, men may avoid seeking help for medical issues, given the perception that help‐seeking is a sign of weakness or loss of control over the body (Addis & Mahalik, 2003). The restricted emotionality that accompanies this idealized form of masculinity is highly detrimental to mental health in particular (Addis & Cohane, 2005; MacLean, Sweeting, & Hunt, 2010).

Masculinities and Mental Health

While research on masculinities has primarily been conducted outside of community psychology (Bond & Wasco, 2017; Mankowski & Maton, 2010), masculinities are connected to a number of mental health outcomes that are of interest to community psychologists (e.g., depression, suicide, psychological distress), and masculinities represent a critical aspect of understanding lived realities within sociocultural context (Trickett, 1996). Thus, research connecting mental health outcomes and masculinities is of direct applicability to the interests of community psychologists. However, given the substantial research that explores connections between masculinities and mental health, systematically reviewing this literature is an important task. Recent reviews in this area tend to be narrowly focused, making a holistic assessment of connections between gender and health difficult. For example, Wong et al. (2017) reviewed 78 samples of both female‐ and male‐identified individuals (though the majority were male‐identified, 91%), specifically focusing on associations between conformity to dominant masculine norms (as measured by one scale, the Conformity to Masculine Norms Inventory) and negative mental health (e.g., depression), positive mental health (e.g., life satisfaction), and psychological help‐seeking. Among the 19,453 participants in this meta‐analysis, Wong et al. (2017) found that higher conformity to masculine norms was associated with poorer mental health and less help‐seeking. In another comprehensive review, O’Neil (2008) explored the use of the Gender Role Conflict Scale (GRCS) to understand men’s experience of mental health, finding that higher gender role conflict (i.e., “a psychological state in which socialized gender roles have negative consequences for the person or others” (p. 362)) was associated with poorer mental health.

In terms of types of mental health outcomes, past reviews have primarily focused on internalizing behavior problems (i.e., the “propensity to experience distress inwards”; Carragher, Krueger, Eaton, & Slade, 2015, p. 340). Internalizing behavior problems include issues related to distress (depression, anxiety) and fear (phobias, panic disorder), while externalizing behavior problems (i.e., the “tendency to express stress outwards”; p. 340) include drug and alcohol dependence and antisocial behaviors (Carragher et al., 2015). In the Wong et al. (2017) review, outcomes included negative mental health (depression, psychological distress/stress, substance use, body image problems, other psychological problems, negative social functioning) and positive mental health (life satisfaction, self‐esteem, psychological well‐being, social well‐being). Except for substance use, all of these issues are related to internalizing behaviors. Similarly, O’Neil (2008) reviewed literature using the GRCS from 1987 to 2007, finding that most studies had focused on internalizing behavior problems (depression, anxiety, stress). The only externalizing behavior problem included was again substance use and was the focus of far fewer studies than internalizing behavior problems.

In terms of pathways from masculinities to poor mental health, Addis and Mahalik (2003) suggest that help‐seeking may be an important mediator, as help‐seeking is “an important step toward resolving numerous problems in living” (p. 5), but many steps in the help‐seeking process (e.g., seeking support, asking for help) conflict with Western hegemonic masculine standards. Past research—including the reviews mentioned previously—has demonstrated connections between adherence to masculine norms and both formal (e.g., professional counseling) and informal (e.g., friends) help‐seeking (Addis & Mahalik, 2003; Blazina & Watkins, 1996; O’Neil, 2008; Wong et al., 2017). Beyond help‐seeking specifically, social support generally is also a potential mediator between masculinities and mental health problems. For example, Houle, Mishara, and Chagnon (2008) found that both seeking help from a friend/family member and perceived social support (e.g., having people to depend on, having close relationships) mediated the pathway between adherence to male gender role norms and suicidal behavior (such that having support/seeking help was protective against suicidal behavior) in a sample of men with and without a history of suicide attempts.

Assessing Masculinity

Both the Conformity to Masculine Norms Inventory and Gender Role Conflict Scale assess one aspect of gender, particularly around adherence to (and consequences of adherence to) male gender role norms in Western society. Smiler (2004) refers to measures tapping such adherence as the ideology perspective, which sees gender as “related to a broad variety of personality attributes, attitudes, behaviors, and activities (vocation and leisure), as well as abstract properties” (p. 22). Broadly, then, measures of male gender role norm adherence include a mixture of behaviors (e.g., “I do not let it show to my friends when my feelings are hurt”; Gupta et al., 2013), beliefs (e.g., “Guys should always be able to figure out what they should do”; Levant et al., 2012), and attitudes (“I can respect a guy who backs down from a fight”; Chu, Porche, & Tolman, 2005). Such behaviors, beliefs, and attitudes converge into commonly recognized features of stereotypical masculinity, including restricted emotionality, dominance, and power‐seeking (O’Neil, 2008). These features of stereotypical masculinity exist along a continuum, such that many men and boys may endorse a lower level of these behaviors, some may endorse a moderate level, and a smaller group may endorse a high level. In addition, levels of adherence can change over time and in different contexts.

While measures tapping adherence to male gender role norms have been a primary focus within counseling psychology, the psychology of men and masculinities, and the limited literature in community psychology (likely because of their clear connection to social context and culture), gender role socialization is not the only way masculinity is understood and assessed (Mankowski & Manon, 2010; Smiler, 2004). To more fully understand connections between masculinities and mental health, then, other masculinity‐related constructs are important to consider. For example, there is a body of literature exploring the set of “masculine” and “feminine” attributes considered “desirable” for men and women. Masculine attributes included on a common measure used in this area—the Bem Sex Role Inventory—include things like “self‐reliant,” “makes decisions easily,” “independent,” “ambitious,” “assertive,” and “competitive.” Unlike adherence to role norms, attributes are seen as residing within the individual and do not consider social context (Smiler, 2004).

Attribute‐based research arose separately (and before) research on adherence to gender role norms (Smiler, 2004), and thus, attributes are a distinct conceptualization of masculinity. However, expectations of “appropriate” attributes for males likely also intersect with socialization into the masculine role. In terms of overlap with gender role norm adherence scales, Parent, Moradi, Rummell, and Tokar (2011) found that higher endorsement of “masculine” attributes on the Bem Sex Role Inventory had moderate to large correlations with the winning, risk‐taking, violence, and primacy of work sub‐scales of the Conformity to Masculine Norms Inventory, as well as with the total score, in a university sample. Conversely, in a sample of 10‐ to 14‐year‐old boys, Chu et al. (2005) found a negligible correlation between scores on the Bem Sex Role Inventory (masculine attributes) and scores on the Adolescent Masculinity Ideology in Relationships scale (a measure tapping gender role norm adherence).

Related, adolescents are actively developing a sense of identity (Erikson, 1968), including in the domain of gender (Rogers, Scott, & Way, 2015). When exploring identity, both the individual’s evaluation of their own identity group (known as private regard) and the importance of being a member of that group to an individual’s self (known as centrality) can be assessed (Rogers et al., 2015). However, identity can also be explored in terms of what researchers in this area call “gender typicality” (Egan & Perry, 2001), or how well one feels they align (globally) with their socially perceived gender category. For example, Jewell and Brown (2014) assessed peer and self‐report gender typicality, based on how much peers/the individual felt they were “typical” for their gender (e.g., “I feel I am a good example of being a boy”).

Typicality measures are seen as distinct from measures focused on gender role socialization (Thompson, Pleck, & Ferrera, 1992); however, as beliefs about typicality undoubtedly arise in connection with larger cultural expectations for gender roles, there is likely some overlap. For example, a study with 342 college men by Leaper and Van (2008) found a moderate, negative correlation between perceived gender typicality and adherence to masculine gender role norms (assessed using the Male Role Norms Inventory). Similarly, while both perceived attributes and perceived typicality are likely connected (as, e.g., holding a high level of a perceived masculine attribute like “dominant” may be related to one’s self‐assessment of their typicality), these two concepts also appear to be distinct. For example, looking at the association between gender‐typed attributes and typicality, Tate, Bettergarcia, and Brent (2015) found a small, positive correlation between the Bem Sex Role Inventory—Agency (i.e., masculinity) sub‐scale and self‐assessed gender typicality among heterosexual, cisgender adult males, but no association among cisgender GBQIA + males, while Dinella, Fulcher, and Weisgram (2014) found a small, positive correlation between Bem Sex Role Inventory masculine attributes and felt gender typicality among 185 male college students. In their study of 4^th‐ to 8^th‐grade children and youth, Egan and Perry (2001) also found a small, positive correlation between gender typicality and amount of male‐typed activities for boys.

In sum, past research indicates that at least three related, yet distinct, assessments of masculinity can be considered when examining connections with mental health: gender role norm adherence, gender‐typed attributes, and identity (Smiler, 2004; Thompson et al., 1992).

Current Study

In addition to being narrowly focused on one way of assessing masculinity, past reviews have not focused on adolescents specifically (while the most recent Wong et al. (2017) review did include six adolescent samples, results were not separated by age group, and the majority of samples were from adults). Given the developments in gendered identity, beliefs, attitudes, behaviors, and attributes that occur during adolescence (Galambos, 2004), as well as the onset of many mental health problems (Merikangas et al., 2010), a review of connections between aspects of masculinity and mental health in adolescence is warranted. Thus, the purpose of this systematic review is to explore associations between multiple facets of masculinity (role norm adherence, attributes, and identity) and mental health among adolescent boys.

As a primary focus in empirical and theoretical literature to date has been the connection between masculinities and internalizing behavior problems, and masculinities and help‐seeking/social support, we chose to focus our review on internalizing behavior problems and social support, so that we could systematically review this body of literature specifically for adolescents. By social support, we mean willingness to seek and ability to access help and/or support from both informal (e.g., friend) and formal (e.g., therapist) sources. We do note that although not the focus of this review, connections between masculinities and externalizing behavior problems, including substance use and aggression, are also important to a holistic understanding of mental health, and a topic worthy of future review. Findings from this review will be used to highlight research gaps in the field, and promising directions for future research in community psychology.

Materials and Methods

Search Strategy

To locate relevant publications, we searched six online databases (MEDLINE, PsycINFO, SocINDEX, ERIC, Social Work Abstracts, and CINAHL) on September 10, 2017. Search terms were combinations of boy* or male* or men* AND masculin* or gender or "role norm*" AND teen* or adolescen* or youth* AND help* or support* or loneliness or connect* or isolation or friend* or peer* or "mental health" or depress* or anxi* or well‐being. We also reviewed the reference lists of eight relevant systematic reviews found during our larger search (i.e., the ancestry approach; Johnson & Eagly, 2000), and hand‐searched three relevant journals that were not picked up as part of the larger database search (American Journal of Men’s Health, Journal of Men’s Studies, and Thymos: Journal of Boyhood Studies).

Inclusion/Exclusion Criteria

Searches were restricted to quantitative and qualitative peer‐reviewed articles published in English between 1997 and 2017 (i.e., in the prior 20 years). We chose this search range as we felt this would offer a thorough snapshot of contemporary masculinities research, including understandings of masculinities as socially constructed (Smiler, 2004). Searches were not restricted by geographic region. To be included, articles had to focus on adolescent boys (age 10–18 years, or those in middle/high school; Steinberg, 2001) and assess a masculinity construct (i.e., role norm adherence, attributes, and/or identity), as well as a mental health outcome (i.e., internalizing behavior problems and/or social support). Articles were excluded if they focused on a college or adult sample. We also excluded articles where gender was included as a covariate in analyses only; when results were not stratified by gender; on gender non‐conformity if exclusively a sexual minority sample (as this seemed like a different goal than the purpose of our study); as well as review, measure development and program evaluation articles (unless the article included baseline associations). We excluded four articles because full text was not available (Figure 1).

Fig. 1.

PRISMA diagram

Review Procedures

Screening was completed using Covidence, an online systematic review management software (www.covidence.org). To determine whether articles should be included, a team of six research assistants (primarily master’s and doctoral‐level students) first did three rounds of title and abstract test screens along with the first author, to ensure consistency in screening decisions. Following these test screens, research assistants reviewed the title and abstract of each of the 7,487 eligible articles in teams of two (Figure 1), to determine whether these articles met inclusion criteria. Screening of titles and abstracts followed the standard Cochrane process (Higgins et al., 2019). Following independent review of their assigned titles and abstracts, the team of two met to discuss decisions and come to consensus on any discrepancies. If the team could not come to consensus, a final decision on inclusion was made by the first author. Results were also reviewed during weekly team meetings. After initial screening of titles and abstracts, 301 articles remained for full‐text review (Figure 1). Full text was again reviewed by a team of two, following the same procedure as for screening of titles and abstracts; 272 articles were excluded because they did not meet inclusion criteria (Figure 1). Following full‐text review, 27 articles (19 quantitative and 8 qualitative) from the database searches remained in the sample. From the ancestry and journal hand search, we included an additional two articles, bringing our final total to 29 articles (20 quantitative and 9 qualitative). Of these articles, one (Gupta et al., 2017) reports on two samples (one from the United States, one from China), for a total of 30 samples in the 29 articles.

Data Abstraction and Quality Assessment

Data from the 29 articles were extracted using a standardized template in Covidence. Extracted information included details on the study (e.g., country; setting; sponsorship information); study methods (e.g., study design, population density, analysis used, study setting, study time period); study population (e.g., inclusion criteria, exclusion criteria, mean age of sample); measures for both masculinity and mental health outcomes; and a description of analysis and outcomes. Following test abstractions with the first author, one doctoral‐level and one undergraduate research assistant (the second and fourth authors, respectively) abstracted information for all articles separately, and then met to come to consensus. Abstractions were all reviewed by the first author. To assess data quality for quantitative articles, we reviewed articles for missingness/attrition for primary outcomes; bias in measurement; sampling frame (representative or non‐representative); response rate (>30%); and other quality issues, using a standard Covidence template. For qualitative articles, we used the RATS checklist (Equator Network, 2016). Following test assessments with the first author, one doctoral‐level and one undergraduate research assistant completed quality assessments for all articles, which were then reviewed and discussed with the first author. For clarity, we present quantitative and qualitative findings separately in the Results section and discuss their overlap in the Conclusions section.

Results

Types of Articles

Of the 29 articles that met our inclusion criteria, 69.0% (n = 20) were quantitative, and 31.0% (n = 9) were qualitative (Table 1). Most studies (82.8%) were published between 2007 and 2017; at least one study has been published every year since 2007. Of the 20 quantitative articles, most (n = 17, 85.0%; Table 1) used non‐representative samples, and the majority (n = 16, 80.0%) were cross‐sectional. Samples were, however, drawn from across the globe; overall, 48.3% of articles used a United States‐based sample, and 37.9% used a sample from somewhere other than North America (South Africa, South Korea, China, India, Malaysia, England, The Netherlands, Norway, Scotland, Sweden). Samples were also fairly racially/ethnically diverse (Table 1). Ages in the 29 included articles ranged from 9 to 20 (6^th‐12^th grade; Table 1).

Table 1.

Summary of included articles (n = 29)

Author and year of publication	N	Sampling	Location	% male	Racial/ethnic distribution	Age range
Quantitative (n = 20)
Basterfield et al. (2014)	568	Two public high schools	South Africa	100%	51% White, 19% Black, 19% Indian	15‐18
Choi et al. (2010)	454 b	Four urban high schools	South Korea	100%	n/a	14‐18
Gupta et al. (2013) a	872	Middle schools	United States and China	100%	US: 27.7% White; 20.6% African American; 19.5% Chinese American; 15.2% Dominican; China: not specified	6th grade (US); 7th grade (China) at initial assessment
Jewell and Brown (2014)	84	Health class in a suburban middle school	United States	40.5%	71.4% White, non‐Hispanic	11‐15
Kulis et al. (2010)	151	Community settings in Arizona and North Carolina	United States	39.7%	100% Latinx	13‐18
Lapointe and Marcotte (2000)	339	Urban high schools in Quebec	Quebec, Canada	50.1%	n/a	13‐18
Mahalingam and Balan (2008)	233	Communities in districts with historically reported male‐biased sex ratio	India	100%	n/a	14‐16
Mansor et al. (2014) c	282	6 secondary schools	Malaysia	100%	Malay (100%)	13‐17
Marcotte et al. (1999) d	306	French‐speaking upper middle‐class high school	Quebec, Canada	46.4%	White (French‐speaking)	14‐17
Marcotte et al. (2002) d	547	French‐speaking public elementary and high school	Quebec, Canada	49.0%	White (French‐speaking)	11‐18
Menon (2011) e	357	An urban secondary school	England	50.4%	Primarily White (86%)	11‐13
Rogers et al. (2015) a	183	All male charter urban high school	Midwest United States	100%	African American (100%)	9th grade (initial assessment)
Santos et al. (2013) a	226	Co‐educational public junior high school	Arizona, United States	100%	44% White, 43% Latinx, 7% Black	8th grade (initial assessment)
Sears et al. (2009)	171	Rural high schools	Eastern Canada	100%	Predominately White	14‐18
Shepard et al. (2011)	58	A summer enrichment program at a university‐based center for gifted education	United States	100%	Primarily White (90%)	12‐18
Smith et al. (2018) a , f	5412	26 urban middle schools	California, United States	48.2%	31% Latinx; 20% White; 13% Asian American; 12% African American	Grade 6 (initial assessment)
Steinfeldt and Steinfeldt (2010)	179	Football players enrolled in participating high schools	Midwestern United States	100%	Primarily White (81%)	9th‐12th grade
van Beusekom et al. (2016)	1026	Secondary schools	The Netherlands	50.4%	Dutch/Western ethnicity (92.5%)	11‐16
Wichstrøm (1999) c	10,839	Public junior and senior high schools	Norway	n/a	n/a	12‐20 (grades 7‐12)
Young and Sweeting (2004) c	2194	School‐based sample of adolescents	Scotland	50.8%	n/a	15
Qualitative (n = 9)
Heinrich (2013)	4	High school English class in a small Midwestern town	Midwest United States	100%	Primarily White	11th grade
Mac An Ghaill and Haywood (2012)	28	An urban middle school in northeast England	England	42.9%	n/a	9‐13
Maclean et al. (2010)	90	One primary and one secondary school	Scotland	n/a	White Scottish students	10‐15
Randell et al. (2016)	33	Schools in a mid‐sized suburban town in central Sweden	Sweden	100%	Primarily Swedish	16‐17
Reigeluth and Addis (2016)	30	Boys from several settings (a small university, an all‐boys private school and an urban youth mentoring program)	Northeast United States	100%	50% White, 23% Hispanic, 10% Black, 10% other, 7% Asian/Pacific Islander	14‐19
Reigeluth et al. (2016)	40	Boys in a public high school	Northeast United States	100%	Racially/ethnically diverse (30% White, 37% Latinx, 17% African American, 13% Asian/Asian American)	10th‐12th grade
Spencer (2007)	12	Boys in a mentoring program	Northeast United States	100%	Racially/ethnically diverse (75% non‐White)	12‐16
Way et al. (2014) a	55	Ethnically diverse middle school boys	New York City, United States	100%	20.8% Black; 32.7% White; 23.6% Chinese American; 14.5% Dominican American; 9.1% Puerto Rican	6th grade at baseline
Wisdom et al. (2007)	22	Adolescents at an urban health center with a history of depression	Northwest United States	40.9%	Primarily White, non‐Hispanic	14‐19

^a Longitudinal study.

^b Abstract states n = 464 and Method section states n = 454, so we report n = 454 here.

^c Representative (probability) sample. All other samples non‐representative.

^d These did not appear to be the same sample, and so both are included here.

^e Presents nearly identical results to Menon, Schellhorn, and Lowe (2013). As such, only Menon (2011) is presented in this review.

^f Uses same sample as Smith and Juvonen (2017) but presents more robust outcome analysis and additional outcomes, and so Smith et al. (2018) is presented here.

Summary of Quantitative Findings

In terms of assessing associations with masculinity constructs, the included quantitative articles fell fairly evenly into three broad categories: associations with masculine gender role norm adherence; associations with masculine attributes; and associations with masculine identity (primarily, gender "typicality”). We describe findings for each broad category and associations with mental health below. A summary of mental health outcomes assessed is presented in Table 2, and a summary of findings is presented in Table 3 (full study findings are described in Supplemental Tables S1 and S2).

Table 2.

Summary of quantitative articles (n = 20)

Mental health outcome	Masculinity measure, % (n))
	Adherence	Attributes	Identity	All
Depression	35.0 (7)	30.0 (6)	25.0 (5)	90.0 (18)
Self‐esteem	20.0 (4)	15.0 (3)	20.0 (4)	55.0 (11)
Perceived social support	15.0 (3)	10.0 (2)	5.0 (1)	30.0 (6)
Other internalizing behavior a	5.0 (1)	10.0 (2)	15.0 (3)	30.0 (6)
Anxiety	10.0 (2)	0.0 (0)	15.0 (3)	25.0 (5)
Stress	10.0 (2)	0.0 (0)	0.0 (0)	10.0 (2)
Help‐seeking intentions	10.0 (2)	0.0 (0)	0.0 (0)	10.0 (2)
TOTAL	45.0 (9)	35.0 (7)	30.0 (6)	100.0 (20)

Numbers do not add to 100% because some studies reported multiple outcomes.

^a Other internalizing behaviors included internalized shame, personal strength, general internalizing behaviors, psychological distress, and loneliness. See Tables S1 and S2 for details.

Table 3.

Summary of quantitative associations ^a

	Internalizing behavior problem					Social support
	Depression	Anxiety	Stress	Self‐esteem	Other	Help‐seeking intentions	Perceived support
Adherence (n = 9) b
Basterfield et al. (2014)	No	–	–	–	–	–	–
Choi et al. (2010)	Yes (−)	–	–	Yes (−)	–	–	–
Gupta et al. (2013) c	Yes (−)	–	–	Yes (−)	–	–	Yes (‐)‐
Mahalingam and Balan (2008)	Yes (+)	–	–	–	Yes (+) d	–	–
Mansor et al. (2014)	No	Yes (−)	Yes (−)	–	–	–	–
Santos et al. (2013) c	‐	−	–	Yes (−)	–	–	–
Sears et al. (2009)	Yes (−)	−	–	–	–	Yes (−)	Yes (−)
Shepard et al. (2011)	No	No	No	No	–	–	No
Steinfeldt and Steinfeldt (2010)	−	−	−	−	–	Yes (–)	–
Attributes (n = 7) e
Choi et al. (2010)	Yes (+)	–	–	Yes (+)	–	–	–
Kulis et al. (2010)	–	–	–	–	Yes (−) f	–	–
Lapointe and Marcotte (2000)	No	–	–	–	–	–	No
Marcotte et al. (1999)	Yes (+)	–	–	–	–	–	–
Marcotte et al. (2002)	Yes (+)	–	–	Yes (+)	–	–	–
Wichstrøm (1999)	No	–	–	–	–	–	–
Young and Sweeting (2004) g	Yes (+)	–	–	Yes (+)	Yes (+) h	‐	Yes (+)
Identity (n = 6) i
Jewell and Brown (2014)	No	Yes (−)	–	Yes (‐)	–	–	–
Menon (2011)	Yes (−)	–	–	Yes (−)	–	–	–
Rogers et al. (2015) c	Yes (+)	–	–	Yes (+)	–	–	–
Smith et al. (2018) c	No	Yes (−)	–	–	Yes (−) h	–	–
van Beusekom et al. (2016)	–	Yes (−)	–	–	Yes (−) j	–	–
Young and Sweeting (2004) g	Yes (−)	–	–	No	Yes (−) h	‐	No

^a For full details on study findings, see Tables S1 and S2.

^b For beliefs, “Yes (−)” means higher adherence to stereotypical gender role norms was associated with worse mental health; “Yes (+)” means higher adherence to stereotypical gender role norms was associated with better mental health; “No” means an association was explored but not found; and “‐” means an association was not explored.

^c Longitudinal study.

^d Other behaviors measured were shame and personal strength.

^e For traits, “Yes (−)” means more identification with “masculine” attributes was associated with worse mental health; “Yes (+)” means more identification with “masculine” attributes was associated with better mental health; “No” means an association was explored but not found; and “‐” means an association was not explored.

^f Other behavior measured was general internalizing behaviors, using the Achenbach Youth Self‐Report.

^g This is a longitudinal study, but only data from the age 15 wave are included in the paper; thus, it is considered cross‐sectional for the purposes of this paper.

^h Other behavior measured was loneliness.

ⁱ All identity studies measure gender “typicality” except Rogers et al. (2015). For Rogers et al. (2015), “Yes (+)” means higher gender private regard was associated with better mental health; “‐” means an association was not explored. For typicality, “Yes (‐)” means higher gender “atypicality” was associated with worse mental health; “Yes (+)” means higher gender “atypicality” was associated with better mental health; “No” means an association was explored but not found; and “‐” means an association was not explored.

^j Other behavior measured was psychological distress (generally).

Associations With Gender Role Norm Adherence

Nine (45.0%) of the quantitative articles assessed associations between masculine gender role norm adherence (i.e., adherence to ideologies and norms associated with being male) and mental health; the majority of these articles (n = 7) assessed associations with depression, with others assessing self‐esteem, anxiety, other internalizing behavior problems, stress, perceived social support, and help‐seeking intentions (Table 2). Scales used to assess beliefs included the Traditional Masculinity Ideology Scale, the Gender Role Conflict Scale for Adolescents (GRCS‐A) scale, the Adolescent Masculinity Ideology in Relationships Scale (AMIRS), the Conformity to Masculine Norms Inventory, the Male Role Norms Inventory‐Adolescent scale (MRNI‐A), and the Gender‐Typed Behavior in Relationships Scale (Table S1). All seven articles assessed gender role norm adherence and mental health outcomes at the individual level (Table S1).

Of articles focusing on gender role norm adherence, two did not find an association with any internalizing behavior problems. The first explored depression using a sample of South African youth from public high schools (Basterfield, Reardon, & Govender, 2014), and the second explored depression, anxiety, stress, and self‐esteem with a small sample of male youth from a summer enrichment program within a university‐based center for gifted education (Shepard, Nicpon, Haley, Lind, & Liu, 2011). Five articles found negative associations (i.e., higher adherence to gender role norms emphasizing things like emotional restriction and dominance predicted poorer mental health; Choi, Kim, Hwang, & Heppner, 2010; Gupta et al., 2013; Mansor, Othman, Yasin, Husain, & Yaacob, 2014; Santos, Galligan, Pahlke, & Fabes, 2013; Sears, Graham, & Campbell, 2009). These studies were all school‐based, used established measures of male gender role norm adherence, and were conducted across the globe (South Korea, United States, China, Malaysia, and Canada). Outcomes in these five studies included depression (n = 3), self‐esteem (n = 3), anxiety (n = 1), and stress (n = 1). Finally, one study found a positive association (i.e., higher adherence to gender role norms emphasizing things like emotional restriction and dominance predicted better mental health; Mahalingam & Balan, 2008). This study was conducted in India (Mahalingam & Balan, 2008), looked at depression as an outcome, and drew its study population from a somewhat specialized community‐based sample of youth (Table 1). As noted below, this study was also assessed as at higher risk of bias. A summary of associations by type of internalizing behavior problem is shown in Table 3, and full study details are in Table S1.

In terms of social support seeking, of the four studies that looked at this outcome, one found no association between gender role norm adherence and social support (Shepard et al., 2011), while the other three found negative associations (i.e., higher adherence to gender role norms like emotional restriction and dominance predicted less social support seeking; Gupta et al., 2013; Sears et al., 2009; Steinfeldt & Steinfeldt, 2010; Table 3). Two of these studies (Gupta et al., 2013 and Sears et al., 2009) used the same outcome measure to assess perceived social support (Table S2). Sears et al. (2009) and Steinfeldt and Steinfeldt (2010) assessed help‐seeking intentions and attitudes toward help‐seeking, respectively, and in both cases found that higher adherence to male gender role norms was associated with less potential help‐seeking (Table S2).

Associations With Attributes

Seven (35.0%) of the quantitative articles assessed associations between masculine attributes and mental health. As with gender role norm adherence, the majority of articles focusing on masculine attributes assessed associations with depression (n = 6), but half also assessed associations with self‐esteem (n = 3; Table 2). Additional mental health variables assessed were perceived social support and other internalizing behavior problems (Table 2). All but one article used versions of the Bem Sex Role Inventory to assess masculine (instrumental) attributes (Table S1); the other article used an investigator‐created scale (Kulis, Marsiglia, & Nagoshi, 2010). All seven articles assessed masculine (instrumental) attributes and mental health outcomes at the individual level (Table S1).

Of these articles, four found a positive association (i.e., greater identification with masculine attributes, such as “ambitious” and “assertive,” was associated with better mental health; Choi et al., 2010; Marcotte, Alain, & Gosselin, 1999; Marcotte, Fortin, Potvin, & Papillon, 2002; Young & Sweeting, 2004). These articles were all school‐based and included samples from South Korea, French‐speaking Canada, and Scotland (Table 1). Outcomes in these four articles were depression (n = 4), self‐esteem (n = 3), and loneliness (n = 1). However, two other articles also using school‐based samples to assess masculine attributes found no association with depression (Lapointe & Marcotte, 2000; Wichstrøm, 1999). These samples were drawn from French‐speaking Canada and Norway, and were considered to be at higher and very low risk of bias, respectively (see below). Finally, one included article found a negative association with general internalizing behaviors (i.e., greater identification with masculine attributes was associated with poorer mental health; Kulis et al., 2010; Table 3); however, it is important to note that this is the only article to use an investigator‐created scale, and not the Bem Sex Role Inventory, and was rated as having a high risk of bias.

Two articles assessed associations between masculine attributes and social support. Lapointe and Marcotte (2000) assessed social support seeking and did not find an association in a school‐based French Canadian sample, while Young and Sweeting (2004) found a positive association between greater endorsement of masculine attributes and number of friends (Table S2).

Associations With Identity

Seven (33.3%) of the quantitative articles assessed associations with masculine identity, primarily by measuring gender “typicality” (or, how well one feels they align (globally) with their socially perceived gender category; n = 6). A majority of these articles assessed associations with depression (n = 5), self‐esteem (n = 4), and anxiety (n = 3), and the remainder assessed associations with other internalizing behavior problems (loneliness, general psychological distress). To assess gender typicality, four articles used Egan and Perry’s (2001) measure, one used the Childhood Gender Non‐conformity Scale, and one assessed gender diagnosticity (i.e., the degree of “maleness” an individual possesses, and used in this study to categorize participants as gender atypical, extremely gender‐typed, or gender typical; Young & Sweeting, 2004; Table S1). The remaining identity article from Rogers et al. (2015) assessed gender private regard and gender centrality. Six articles assessed identity and mental health outcomes at the individual level; however, one article (Smith, Schacter, Enders, & Juvonen, 2018) also considered a school‐level measure of gender norm salience, and associations with individual‐level loneliness, depressed mood, and social anxiety (Table S1).

All six articles exploring gender typicality found negative associations with the internalizing behavior problems described above (i.e., higher gender atypicality was associated with poorer mental health; van Beusekom, Baams, Bos, Overbeek, & Sandfort, 2016; Jewell & Brown, 2014; Menon, 2011; Smith & Juvonen, 2017; Smith et al., 2018; Young & Sweeting, 2004; Table 3). All of these articles used school‐based samples; half of these studies were with United States‐based samples, and the other three included youth from England, The Netherlands, and Scotland (Table 1). Examining a different aspect of identity, Rogers et al. (2015) found a positive association between gender private regard and depression and self‐esteem in a school‐based sample of African American youth (i.e., a more positive assessment of the identity group was associated with better mental health). Only one article also assessed associations with perceived social support (Young & Sweeting, 2004), but did not find an association (Table 3).

Article Quality

We used four criteria to assess the quality of quantitative articles: (a) missingness/attrition, (b) measures used, (c) sampling frame, and (d) response rate. We then evaluated risk of bias for each criterion: (a) high, (b) low, or (c) unclear. We also had a category for other quality issues, where the team could highlight general concerns, issues, or questions for consensus. Quality was then determined through consensus between the first, second, and fourth authors. Three (15.0%) studies were found to be at low risk of bias (Mansor et al., 2014; Wichstrøm, 1999; Young & Sweeting, 2004), due to minimal attrition, use of standardized measures, probability sampling, and high response rates. As noted above, nearly all of the studies included in this review used a convenience sample as the sampling frame, which increased the risk of bias. Eight (40.0%) studies were found to be at fairly low risk of bias (van Beusekom et al., 2016; Gupta et al., 2013; Marcotte et al., 2002; Rogers et al., 2015; Santos et al., 2013; Sears et al., 2009; Smith et al., 2018; Steinfeldt & Steinfeldt, 2010). Each of these studies used standardized measures, employed convenience sampling, and reported high response rates, though Sears et al. (2009) and van Beusekom et al. (2016) did not report on missingness/attrition, and missingness/attrition in Santos et al. (2013) was unclear. Nine (45.0%) studies were found to be at higher risk of bias (Basterfield et al., 2014; Choi et al., 2010; Jewell & Brown, 2014; Kulis et al., 2010; Lapointe & Marcotte, 2000; Mahalingam & Balan, 2008; Marcotte et al., 1999; Menon, 2011; Shepard et al., 2011). Though all these studies except Kulis et al. (2010) used standardized measures, most did not report on missingness/attrition or response rates, and all used a convenience sample. Though Jewell and Brown (2014) did report on missingness/attrition, the consent and response rate in this study was quite low (49.1%).

Summary of Qualitative Findings

In terms of assessing associations with masculinity, the nine included qualitative articles focused on the influence of male gender role norm adherence, primarily exploring emotional connection/restriction and mental health (n = 6; 66.7%). Other less prominent themes included policing of masculinity (n = 3; 33.3%) and institutional expectations for masculinity (n = 2; 22.2%). A summary of key findings by article is presented in Table S3.

Emotional Connection

Six (66.7%) of the included qualitative studies discussed emotional connection, including closeness, expression, consequences of, and restriction. Randell, Jerden, Ohman, Starrin, and Flacking (2016) and Reigeluth, Pollastri, Cardemil, and Addis (2016) explored emotional expression related to masculinity and friendships. Reigeluth and Addis (2016) examined the emotional consequences of policing masculinity, and Spencer (2007) examined emotional closeness in healthy male adult–youth relationships. Way et al. (2014) found that the ability to maintain close friendships through adolescence was linked to the ability to resist conforming to stereotypical masculine norms; specifically, emotional connection was directly related to the ability to establish and maintain close friendships, which in turn supported resistance to adopting stereotypical male gender role norms. Finally, Wisdom, Rees, Riley, and Weis (2007) found that adherence to stereotypical gender role norms resulted in emotional restriction and created a barrier to seeking help for depression. In this study, boys were more likely than girls to ignore, repress, or distract themselves from negative emotions and feelings.

Overall, all six articles found that emotional connection, expression, and closeness with friends were affected by adherence to stereotypical male gender role norms. Specifically, boys who adhered more strongly to gender‐normative masculinity were more likely to exhibit emotional restriction than their less adherent peers, and tended to police other boys’ performances of masculinity in order to diminish expressivity and maintain a perceived masculine ideal. Emotional connection was also found to be contextually dependent upon boys’ adherence to or rejection of stereotypical gender role norms.

Policing of Masculinity

One study discussed policing of masculinity specifically (Reigeluth & Addis, 2016), although there were two other studies focused on emotional connection that also highlighted the issue of policing masculinity (MacLean et al., 2010; Reigeluth et al., 2016). The themes of policing masculinity and emotional connection overlapped in the Reigeluth et al. (2016) and Reigeluth and Addis (2016) studies, in which some boys were found to engage in the policing of masculinity among their peers. In these studies, policing of masculinity was used to minimize emotional expressivity and uphold stereotypical gender role norms of stoicism, toughness, and emotional restriction. MacLean et al. (2010) also found that boys demonstrated awareness of and adherence to expectations of gender role norms, including stoicism, strength, control, and independence, such that these norms become strict “rules” for behavior policed by peers. For both physical and psychological symptoms, boys said that they would conceal, downplay, or ignore symptoms in order to better conform to masculine norms in front of their peers.

Institutional Expectations for Masculinity

There were two (22%) studies that focused on institutional modeling and expectations for masculinity (Mac an Ghaill and Haywood, 2012, describe institutional masculinity as gendered expectations within school systems and other institutions that determine acceptable behaviors). Heinrich (2013) focused on the production of masculinity and vulnerability, and found that institutional masculinity reinforced hierarchy, hegemony, and power structures, ultimately privileging some performances of masculinity over others. Emotional connection and expression were consequently restricted by these institutional expectations, resulting in performances of masculinity that adhered to male gender role norms such as stoicism and toughness. Mac an Ghaill and Haywood (2012) examined normative models of masculinity and suicide, finding that assumptions of gendered behaviors developed at the institutional (school) level rather than the individual level; specifically, characteristics and norms of behaviors were determined by institutional assumptions (e.g., around showing fear), rather than individualized traits. Based on their findings, the authors conclude that institutional masculinity results in a “production and regulation of boyhood” (p. 485) and that failure to conform to institutional expectations of boyhood led to anxiety and emotional distress among their participants. The institutional expectations of gendered norms also affected boys’ ability to emotionally disclose, demonstrate emotion, or seek help. Institutional expectations for masculinity thus created structural processes and barriers to help‐seeking, especially in cases where boys were expected to be stoic and independent.

Article Quality

We assessed the quality of the qualitative studies using the RATS Checklist (Equator Network, 2016). This checklist assesses the: (a) relevance of the study questions; (b) appropriateness of the qualitative method used; (c) transparency of the sampling, recruitment, roles, and ethical procedures; and (d) soundness of the approach. The majority of the qualitative articles (n = 8, 88.9%) had low to mostly low risk of bias (Mac an Ghaill & Haywood, 2012; Maclean & Sweeting, 2010; Randell et al., 2016; Reigeluth & Addis, 2016; Reigeluth et al., 2016; Spencer, 2007; Way et al., 2014; Wisdom et al., 2007). We considered Randell et al. (2016), Spencer (2007), and Wisdom et al. (2007) to be at low risk of bias as these studies accounted for all of the RATS Checklist criteria and provided clear explanations of the study design, ethical considerations, analysis, and findings. For the others, we consider them to be at mostly low risk of bias as they did not clearly report some of their procedures per RATS guidelines, but were generally transparent (Mac an Ghaill & Haywood, 2012; Maclean & Sweeting., 2010; Reigeluth & Addis, 2016; Reigeluth et al., 2016, Way et al., 2014). We found only one study to have a higher risk of bias. The analysis in Heinrich (2013) was not clearly specified, ethics approval was not discussed in the article and it was unclear how the researcher accounted for their relationship with participants.

Conclusions

In this article, we present the most comprehensive review to date of connections between masculinities and internalizing behavior problems and social support in adolescent boys. Generally, we found that greater identification with stereotypical masculine attributes (e.g., “ambitious”) was associated with better mental health (i.e., fewer internalizing behavior problems, greater social support), while greater adherence to stereotypical male gender role norms, such as restricted emotionality and dominance, as well as lower gender “typicality” (a measure of identity), was associated with poorer mental health (i.e., more internalizing behavior problems, less social support). Qualitative articles exclusively focused on understanding the consequences of emotional restriction (a stereotypical male gender role norm) for mental health. In the included qualitative articles, emotional restriction contributed to boys being less likely to discuss or seek help for mental health problems. However, qualitative articles also demonstrated that when boys were more aware of stereotypical male norms, or had more supportive peer and adult relationships, they felt less of a need to police masculinity or uphold idealized expectations for masculinity, suggesting the need to intervene on interpersonal and institutional settings (in addition to individual‐level beliefs and attitudes) to improve mental health outcomes for adolescent boys.

In terms of masculine gender role norms, in quantitative articles, we primarily found that as adherence to these norms increased, mental well‐being decreased (i.e., higher depression, anxiety and stress, lower self‐esteem and perceived social support); these findings were also supported and expanded upon by included qualitative articles. These findings also align with theory that links adherence to these masculine norms with poor mental health (e.g., Courtenay, 2000; Evans et al., 2010), and indicate that pathways to downstream internalizing behavior problems for men may first emerge in adolescence. This finding also aligns with Wong et al.’s (2017) meta‐analysis, which found that conformity to stereotypical masculine norms had moderate correlations with both poor mental health and less psychological help‐seeking in samples of primarily adult men.

The three studies in our review which demonstrated a positive association/no association between gender role norm adherence and mental health were all rated as at higher risk of bias, which may be one reason for this difference. However, we also note that the one article that found a positive association (i.e., that greater adherence to male gender role norms was associated with lower depression and shame and greater personal strength) was in a community‐based sample, while the articles that found negative associations were all in school‐based samples. The two articles that found no association were in a specialized (a summer enrichment program; Shepard et al., 2011) and a non‐North American (South African youth; Basterfield et al., 2014) sample. Given that masculinities are locally constructed (Ward et al., 2017), it is thus possible that in some contexts, adhering to certain masculine ideologies may be adaptive for mental health (Heller et al., 2015). This highlights the role of domain‐specific contexts in shaping behavior (Trickett, 2009), and the need for future research in community psychology to more deeply explore the role of differential contexts in shaping connections between masculinities and mental health.

Of the six studies that examined identification with gender‐typed attributes as assessed by the Bem Sex Role Inventory, all but two found that endorsement of “masculine” attributes was associated with fewer internalizing behavior problems and better social support. In other words, male adolescents who more strongly identified with attributes such as “ambitious” and “assertive” reported less depression, better self‐esteem, and more perceived social support (the one study that found a negative association did not use a standardized measure and was rated as having high risk of bias). As Gupta et al. (2013) suggest, findings such as these should not be surprising, as qualities that are often perceived as positive (such as being ambitious and assertive) are associated with masculinity (instrumentality) on the Bem Sex Role Inventory. However, although this scale interprets these attributes as masculine, adolescents themselves may no longer view them this way. Specifically, research by Auster and Ohm (2000) compared the validity of the Bem Sex Role Inventory from 1972 to 1999 in college samples, finding that in 1999, only 8 of the original 20 “masculine” attributes were still interpreted as masculine, likely since constructions of masculinity have changed over time (Basterfield et al., 2014). Thus, future research on these attributes should likely be “disentangled” from the concept of masculinity (Spence & Helmreich, 1980, p. 147).

We also note that adherence to gender role norms (e.g., around toughness, stoicism) was generally associated with poorer mental health, while endorsement of “masculine” attributes (e.g., self‐reliance, ambitious, assertive) was generally associated with better mental health. This finding aligns with past research with male adolescents, which did not find a significant correlation between gender role norm adherence and attributes (Chu et al., 2005), and thus, it follows that these two constructs may be differentially related to mental health outcomes. Specifically, as noted by Gupta et al. (2013), masculine attributes typically refer to traits that are viewed positively by individuals and interpersonal others (e.g., in Western culture, it is a good thing to be viewed as an independent, assertive leader who is willing to take risks). Conversely, gender role norms assess behaviors likely to cause internal distress (e.g., hiding emotions, feeling like you have to appear in control at all times). In addition to differential interpretations, these findings may also reflect issues in measurement and conceptualization. Specifically, and as mentioned above, the “masculine” traits presented in gender‐typed attribute studies are likely not the best way to explore associations between masculinities and mental health in contemporary contexts, as respondents themselves may no longer interpret these traits as “masculine” (Basterfield et al., 2014). However, gender role adherence behaviors are still very much interpreted as comprising what it means to be “a man” in many contexts (e.g., Way et al., 2014). These differential interpretations and measurement issues may thus underlie the divergent association of these two constructs with internalizing behavior problems and social support in the articles reviewed in our study.

Finally, all of the articles that assessed gender “typicality” (an aspect of identity) found that boys who reported higher scores on measures of gender atypicality reported more depression, anxiety, loneliness, general psychological distress, and lower self‐esteem than their more gender‐typical peers. This finding likely speaks to the marginalization and violence experienced by gender non‐conforming boys (Brooks, 2000; D’Augelli, Grossman, & Starks, 2006); to make this connection to larger structural marginalization clearer, we encourage future research to take a critical lens on social exclusion of individuals who are not perceived as conforming to the “typical” gender binary, as opposed to a focus which roots causality at the level of the individual. To this end, we encourage future research on this topic to move away from terms that center the problem within individuals (such as gender typical or atypical), and that instead locate the problem within environments that are gender rigid or gender accepting. This structural understanding is also critical when interpreting this finding (i.e., the issue is not helping boys to be more “typical,” but rather creating communities where these boys are accepted for who they are). Only one of the seven gender “typicality” articles assessed any form of social support, and the one article that did investigate this support only looked at number of friends (Young & Sweeting, 2004). Given the adverse mental health outcomes reported by boys who scored high in gender atypicality in our review, additional research exploring social support among this group is critically needed.

Implications for Community Psychology

As Mankowski and Manon (2010) note, the “analysis of men’s gender largely is missing from community psychology’s efforts to understand wellness, oppression and social systems change” (p. 74). Yet, understandings of masculinities and mental health in context are clearly relevant to the work of community psychologists. Thus, community psychology can specifically play a role in addressing gaps identified by this review. First, despite an acknowledgement among researchers that masculinity is socially constructed and context‐dependent (Connell, 2005; Connell & Messerschmidt, 2005; Smiler, 2004), most articles in our review still focused on an individual‐level of analysis (i.e., analysis that explored individual adherence to beliefs or endorsement of attributes, independent of their broader social context). More specifically, while beliefs about masculine role norms, attributes, and identity are of course shaped by the larger culture (Bond & Wasco, 2017), the majority of included quantitative articles did not explore this larger context, focusing instead on how individual‐level beliefs, behaviors, and attitudes were related to individual‐level internalizing behavior problems/social support (this may reflect the discipline of study authors; only three articles were published in journals relevant to community psychology¹; two‐thirds of articles were published in developmental psychology or gender/masculinity journals). Conversely, only three articles in this review included an explicit focus on contexts other than the interpersonal setting (one quantitative—Smith et al., 2018; two qualitative—Heinrich, 2013; Mac an Ghaill & Haywood, 2012), with all demonstrating how the school context shapes expectations for appropriate masculine behavior and presentation, and the detrimental impact of these expectations for mental well‐being. Thus, community psychology research exploring the contextual nature of gendered qualities, and the role of settings in shaping gender norms, expectations, and processes (Bond & Wasco, 2017; Mankowski & Maton, 2010), is warranted. While the call to consider connections between masculinities, context, and mental health is not new (e.g., Connell, 2005), our review demonstrates that these connections have still not been adequately explored, and thus, community psychologists can make a much needed and valuable addition to this literature.

Second, we feel the findings of our review further Smiler’s (2004) point that there is no problematic form of masculinity itself, but rather that problems arise as a result of “overly rigid adoption/adherence to masculine norms [that] can be problematic” (i.e., a lack of behavioral flexibility; p. 20). Since this is the case (i.e., masculinity in and of itself is not something problematic residing within individuals), sociocultural context, and implications of this context for outcomes, is critical to understand. However, despite the importance of cultural and contextual norms and beliefs, the field to date has largely focused on individualistic approaches to understanding masculinities (Smiler, 2004), a finding further bolstered by our review. Thus, we believe community psychology researchers can significantly advance the field by “plac[ing] boys’ identities within institutional contexts in order to explore what gender means and how gender is lived out” (Mac An Ghaill & Haywood, 2012, pp. 485–486), and implications of such living out for mental health. To this end, Bond and Wasco (2017) present a three‐part conceptual model on understanding gendered nature of settings that can guide this work.

Finally, cross‐cultural work suggests that while there may be dimensions of gender‐typed behavior that is similar across cultures (Arciniega, Anderson, Tovar‐Blank, & Tracey, 2008; Way, 2011; Yim & Mahalingam, 2006), the degree to which individuals adhere to masculine expectations and their mental health adjustment may vary depending on the meaning of the behavior within cultures (Gupta et al., 2013). For example, in our review, one of the articles using a school‐based sample of South African youth did not find an association between adherence to male gender role norms and internalizing behavior problems (depression; Basterfield et al., 2014), while another study using a community‐based sample of Indian youth found positive associations (i.e., stronger adherence to male gender role norms was associated with less depression and shame, and greater personal strength; Mahalingam & Balan, 2008). These findings point to the possible influence of local and regional constructions of masculinity on mental health outcomes, and as such, more research exploring the cultural aspects of the impacts of gender‐typed ideologies on mental health is warranted (Connell, 2012). Community psychologists can contribute to this work by exploring cross‐cultural differences in masculinity and mental health (Trickett, 2009), illuminating both the “contexts of [masculine] diversity and the diversity of [masculine] contexts” (Trickett, 1996, p. 218). Further, only one of the included articles (Rogers et al., 2015) specifically explored intersections of race and gender identity. Thus, an explicit focus on intersections between race, gender, and class (and other identity markers) is another critical area for future community psychology research on masculinities in context.

Limitations

Several limitations should be noted. First, while we had originally planned to conduct additional searches for other mental health outcomes (particularly externalizing behavior problems, including substance use and antisocial behavior), the large number of returns on our internalizing behavior problems/social support search meant we were not able to address additional outcomes, due to resource constraints. Thus, a review of connections between masculinities and externalizing behavior problems in adolescence is a needed contribution to the literature. We also did not include personality disorders. These are all important outcomes for future reviews. Second, because of the lack of consistency in reporting of effects, as well as the lower quality of a number of our quantitative articles, we made the decision not to perform a meta‐analysis on these data. Finally, all quantitative studies but three used non‐representative samples, and all but four used a cross‐sectional design. Thus, findings should be interpreted with caution.

Future Directions

Attention to the mental health of men and boys, and the association of masculinities with mental health, continues to grow. Indeed, the recently released American Psychological Association (APA) Guidelines for Psychological Practice with Men and Boys (2018) specifically draw attention to the role of gender role conflict and masculinity ideology in shaping mental health, including through restricting willingness to seek help. Given the importance of the adolescent period for understanding the development of masculinities in context, as well as for promoting mental health, it is critical to continue to explore connections between aspects of masculinity and mental health in this age group. Our review highlights several key gaps in the adolescent literature, including a failure to assess interactions between context, masculinities, and mental health, and areas where community psychology can address these gaps.

In addition to implications for research, our review also has key implications for practice. Specifically, we believe our review highlights the critical need to focus on gender‐transformative health promotion approaches with adolescent boys (Brush & Miller, 2019), in order to help them deconstruct and explore gendered expectations for their attitudes, beliefs, behaviors, attributes, and identity. To be effective, it is critical this work focuses on the larger cultural context that shapes masculine norms, attributes, and identity (e.g., by working to change school culture), and not just individual‐level behavior (Trickett, 2009; Trickett & Rowe, 2012). The APA practice guidelines also specifically highlight the need for “experiential groups to promote friendships and support among boys while helping them critically examine dysfunctional boy codes and restrictive notions of masculinity” (p. 14). The findings of this review demonstrate the critical nature of such programs, and their potential for addressing life course health inequities.

Conflicts of Interest

The authors have no conflicts of interest to declare.

Supporting information

Table S1. Associations between aspects of masculinity and internalizing behavior problems (n = 20).

Table S2. Summary of quantitative findings, social support (n = 6).

Table S3. Summary of qualitative articles (n = 9)