Journal Logo


Gender Differences in Depression

Biological, Affective, Cognitive, and Sociocultural Factors

Hyde, Janet S. PhD; Mezulis, Amy H. PhD

Author Information
doi: 10.1097/HRP.0000000000000230
  • Free


The gender difference in depression is considered to be one of the most robust phenomena in psychiatry and psychology, with more depressed women than men.1 The gender difference in depression can thus be viewed as a serious health disparity. Before efforts to eliminate this disparity can be developed successfully, we must understand the factors that create the gender difference. In this review we use the ABC (affective, biological, cognitive) model of gender differences in depression as a framework to synthesize and organize the evidence-based factors that contribute to the gender difference.1 In addition, this review serves to update research reviewed in the original article on the ABC model. Moreover, the ABC model, as updated in the current article, incorporates several potential influential factors that are typically overlooked in reviews of gender differences in depression, including adolescent experiences of sexual harassment by peers and structural gender inequality. Whenever possible, we report results from meta-analyses, and we report effect sizes rather than significance tests. One reason for the emphasis on effect sizes is that, too often, a finding of a significant gender difference is too quickly translated into a categorical statement (women do this and men do that), whereas the gender difference may actually be small, with substantial overlap in the male and female distributions.2


Salk, Hyde, and Abramson3 conducted a meta-analysis of studies of gender differences in depression based on representative national samples. Measured outcomes in the studies included diagnoses of major depression and self-reports of depressive symptoms, the latter based on a continuous, or dimensional, model of depression. Averaged over all samples from all nations and ages, the odds ratio (OR, female:male) was 1.95—roughly twice as many depressed women as depressed men. The dimensional, self-report measures yielded d = 0.27 for gender differences in symptoms. The gender difference was present as early as age 12, with OR = 2.37 (sufficient data were not available to compute an OR for younger children), and peaked at ages 13–15, with OR = 3.02. The gender gap narrowed in early adulthood and remained stable thereafter, with OR = 1.72 for ages 20–29 and OR = 1.57 for ages 60–69.

In that same meta-analysis, odds ratios varied significantly across nations and were sometimes even reversed; for example, in Sri Lanka OR = .83.4 These cross-national variations provide evidence of the impact of culture in producing and magnifying the gender difference in depression.

Some may question whether the gender difference in depression is a “real” difference or, instead, an artifact of gender-biased responding, with men being less likely to report certain symptoms because those symptoms are regarded as unmasculine.5 Some evidence supports such biases, although the effects appear to be slight. In a community sample of individuals who met diagnostic criteria for major depressive disorder (MDD), women were more likely to report 4 of the 26 symptoms.6 An example is “was often in tears.” The male role forbids crying, whereas the female role allows or even encourages it.7 As a result, men may be less likely to report their crying or, alternatively, may restrain themselves from crying. Nonetheless, only 4 of 26 symptoms showed evidence of gender bias. We conclude that, while a small gender bias may influence the assessment of depression, implying that the estimated size of the gender difference may be somewhat inflated, a clinically meaningful “real” difference remains.

Some efforts to assess gender differences often implicitly rest on the assumption that each gender category is homogeneous, when each category is actually heterogeneous. Intersectionality, which provides a framework for considering this heterogeneity, has become a major approach to analyzing gender differences and, in particular, the intersection of gender and ethnicity in the United States.8 Intersectionality holds that people belong to multiple social categories as a function of gender, race, social class, sexual orientation, nationality, immigration status, and so on, and that these categories are interlocking.8 Consideration of the intersection of gender and ethnicity raises the question of whether the gender difference in depression has the same magnitude in different U.S. ethnic groups. Although ethnicity was not a significant moderator in the meta-analysis by Salk and colleagues,3 the odds ratio varied from 1.74 for African American samples to 2.23 for European Americans. Some epidemiological studies analyze the data for gender differences and for race differences, but not for the two jointly, with the consequence that the analysis is not truly intersectional. For example, in their study of adolescents and young adults, Mojtabai and colleagues9 found 12-month prevalences of major depressive episodes of 11.8% for women and 7.4% for men, and of 11.1% for whites, 6.1% for blacks, and 7.8% for Hispanics. From these findings, it cannot be determined what the prevalence was for black women compared with white women, or for Hispanic men compared with white men, or whether the gender difference is larger in some groups than others.

Patil and colleagues10 reviewed studies that used an intersectional approach to the investigation of risk for depression, including in relation to race/ethnicity and gender. They concluded that gender was the greatest risk factor for depression and that race/ethnicity also had an influence. This intersectional approach is also capable of highlighting stressors—such as experiences of discrimination and immigration status—that are typically not part of stress inventories.


The etiology of the gender difference in depression is complex. We proposed the ABC model to account for gender differences in depression and, in particular, the emergence of the gender difference in adolescence.1

The ABC model is a developmental model (see Figure 1 for a diagram). That is, if we are to understand the 2:1 gender ratio for depression in adulthood, we must understand the factors contributing to its emergence in early adolescence. As noted, the gender difference appears by age 12.3 The model therefore focuses on factors from childhood to early adolescence.

Figure 1
Figure 1:
The ABC model of gender differences in depression (updated from Hyde, Mezulis, & Abramson, 2008).1

The ABC model is a vulnerability-stress model. As noted above, it proposes several categories of vulnerabilities: affective, biological, and cognitive. When a person with one or more of these vulnerabilities encounters negative life events, the person has a higher likelihood of developing depression. Below we review the evidence on the biological, affective, and cognitive factors that contribute to vulnerability to depression, as well as the negative life events (stressors) and sociocultural factors that interact with the vulnerabilities.

The ABC model incorporates the principle of equifinality,11 which refers to multiple pathways to the same outcome—in this case, depression. Multiple factors are involved in the development of depression and gender differences in depression. Some individuals follow one pathway, and others follow another. Therefore, any single pathway will account for only a small percentage of the variance in depression. Following this reasoning, one would not expect a single factor to account for the gender difference.

Biological Factors

Biological factors that contribute to vulnerability to depression include genetic factors, pubertal hormones, and pubertal timing.

Genetic factors

Using candidate-gene approaches, researchers identified the 5-HTTLPR polymorphism, in interaction with stress, as a risk factor for depression. This polymorphism is a region within the serotonin transporter gene. A meta-analysis found that this gene × stress interaction was significant across 51 studies, with OR = 1.18.12 That odds ratio is not high, but according to the equifinality model, a single factor is not expected to have a large effect. This meta-analysis, however, did not consider interactions with gender. For a model of how 5-HTTLPR polymorphisms are likely to affect brain functioning differently in males than in females, see Perry and colleagues (2017).13 A more detailed consideration of the literature on brain functioning and depression is beyond the scope of this review. Candidate-gene approaches will not be discussed further, because that approach has been criticized for failures to replicate.14 It has now largely been replaced by genome-wide association studies (GWAS).

A meta-analysis of GWAS identified 44 independent and significant loci linked to MDD.15 No genes on the X or Y chromosome appeared among the 44, nor did 5-HTTLPR (located on chromosome 17). This study also found that genetic risk for depression partially overlapped with genetic risk for schizophrenia. A major implication from the study is that genetic risk for depression is not categorical but, instead, is continuous, with numerous genes contributing.

Despite large sample sizes, GWAS often also fail to replicate. One GWAS identified two loci linked to MDD, both on chromosome 10.16 Another GWAS identified two loci linked to self-reports of depressive symptoms, on chromosomes 12 and 18.17 The most recent and largest GWAS identified 102 independent variants, across multiple chromosomes.18

Those GWAS looked just at links between loci and depression—that is, simple main effects, without regard to gender. Most relevant to the current review, another GWAS considered whether there was genetic heterogeneity between males and females in loci linked to depression and found no evidence of heterogeneity; it appears that the same genes are implicated in depression for women and men.19

The GWAS reviewed above did not consider stress or adverse events in the models, and yet stress is at the heart of vulnerability-stress models, as reviewed in more detail below. One GWAS incorporated a genome-wide × environment interaction approach (GWEIS).20 This study failed to find any influential loci, probably because of sample size. Nonetheless, the approach has merit. Another GWAS, of Han Chinese women with recurrent major depression, took into account exposure to adversity and concluded that the genes contributing to depression may differ between those who have or have not experienced major negative life events.21

Pubertal hormones

Pubertal hormones may be involved in the gender difference in depression—specifically estrogen (estradiol, E2), progesterone, and testosterone, as well as the adrenal androgens DHEA and DHEA-S.1 Pubertal hormones are suspect because of the emergence of the gender difference by age 12—that is, during puberty. Estradiol and the other sex steroids have an impact on brain functions such as neural excitability and neural transmission.22 For example, estradiol (E2) can affect synaptic wiring, both acutely and in a sustained manner.22 Moreover, E2 regulates gene expression through estrogen response elements (ERE); E2 regulates all three RNA polymerases, thus influencing transcription.22

Adrenarche, the early phase of pubertal developmental, is characterized by maturation of the adrenal glands and increased production of DHEA and DHEA-S. A systematic review concluded that, although results are somewhat inconsistent across studies, higher levels of DHEA and DHEA-S are associated with more internalizing symptoms in girls.23 Interpretation of these findings is complicated because higher levels of the adrenal androgens reflect earlier pubertal development, which itself may be a risk factor, particularly because early puberty creates a different social environment for girls, as described below. It is possible that adrenarche is a sensitive period in neurobiological development that has implications for affective disorders.23

The later phase of pubertal development is gonadarche, the maturation and enlargement of the gonads, leading to increased production of estradiol, progesterone, and testosterone. Estradiol is involved in the regulation of multiple systems that are implicated in depression, including neural plasticity, the stress axis (hypothalamus, pituitary, adrenal: HPA), and immune function.22

Pubertal timing

Although it has generally been believed that, in regard to internalizing, early puberty is bad for girls (bad in the sense that it increases risk for both internalizing and externalizing disorders) but neutral or even positive for boys, a meta-analysis found that early puberty was associated with internalizing problems for both girls and boys, although the effects were small: d = 0.19 for girls and 0.16 for boys.23 Internalizing disorders are those in which the individual internalizes the problem to the self, and include depressive disorders, anxiety disorders, and trauma-related disorders. In regard to psychological distress, there was a small effect of early puberty for girls; the effect was not significant for boys.

Affective Factors

Affective models of depression posit that temperament is a vulnerability factor for depression.24–26 Temperament is conceptualized as genetically based individual differences in emotional and attentional reactivity and self-regulation that appear early in life and are relatively stable over time.25,26 Longitudinal studies demonstrate that children scoring high on a temperament dimension known as negative emotionality (sometimes termed negative affectivity or “difficult” temperament) are later at increased risk for depression.25,26

Negative emotionality may lead to depression through several pathways.1 First, negative emotionality may have a simple, direct effect on depression; that is, negative emotionality is a risk factor. Second, negative emotionality may contribute to the development of other vulnerability factors such as negative cognitive style (discussed in a later section), which in turn is a more proximal risk factor. Third, negative emotionality may make the individual more reactive to stress, and the hormonal and social changes that occur at puberty may act as stressors that exacerbate negative emotionality, leading to depression.

How can these affective models account for gender differences in depression? One possibility is that girls are higher in negative emotionality from early childhood. A meta-analysis of studies on gender differences in temperament, however, found a trivial gender difference for negative affect, d = −0.06.27 In the absence of a gender difference in negative emotionality, another possible mechanism is a temperament × stress interaction, with girls and women experiencing more stress than boys and men, beginning in early adolescence. Gender differences in the experience of stressors are discussed in a later section.

Several studies have found that temperament predicts depression differently for girls than for boys. Mezulis and colleagues28 examined predictors of different depression-symptom trajectories from infancy into adolescence and found that girls who were high in negative affectivity in infancy typically displayed increases in depression symptoms in adolescence, whereas boys high in negative affectivity in infancy typically displayed initial higher levels of symptoms and then decreases in depression symptoms in adolescence. Many studies on temperament and depression are methodologically flawed because of the use of cross-sectional designs in which temperament is measured in adolescence, concurrently with the measure of depression.29–31 Longitudinal designs, such as the one in the Mezulis study,28 in which temperament is measured in infancy or the preschool years, are necessary to determine whether temperament is indeed a risk factor for depression and contributes to gender differences in depression in adolescence and beyond (see also Forbes et al. [2017]32).

Cognitive Factors

The ABC model posits three categories of cognitive vulnerability to depression: negative cognitive style, objectified body consciousness, and rumination.

Negative cognitive style

Cognitive vulnerability-stress models of depression posit that people with certain cognitive styles are vulnerable to depression when they encounter stressful events. The two most prominent of these models are hopelessness theory and the ruminative response styles theory.33,34 In this section we consider hopelessness theory; rumination is addressed in a later section.

In hopelessness theory, negative cognitive style (also termed attributional style) refers to the tendency to make negative inferences about causality, the self, and consequences in response to negative life events.33 Individuals who have developed a negative cognitive style, when faced with a new stressor, are at increased risk for depression. Prospective tests of this model have supported it, both with adults and adolescents.35–37 A meta-analysis of longitudinal studies found significant support for negative cognitive style predicting later depression, with correlations in the range r = .22 to .30.38

Two questions are relevant in regard to how cognitive vulnerabilities may contribute to the gender difference in depression: (1) Are there gender differences in negative cognitive style? (2) Does negative cognitive style predict depression differently for males than for females? To our knowledge, no meta-analysis of gender differences in negative cognitive style is available. Individual studies have found, for children age 11 and under, either that there are no gender differences in attributional style or that boys have the more negative style.39,40 In adolescence, girls have more negative styles.41 The evidence indicates, though, that the gender difference in depression appears before the gender difference in negative cognitive style, making negative cognitive style an unlikely explanation for gender differences in depression.40 Moreover, according to a meta-analysis, negative cognitive style predicts depression similarly for females (r = .29) and males (r = .27).38

Objectified body consciousness

Research in the psychology of women and gender indicates that traditional views of cognitive vulnerability should be expanded to include negative cognitions about the body. Objectified body consciousness includes two processes that contribute to negative views of one’s body: (1) self-surveillance, a cognitive process in which individuals become observers and critics of their bodies and appearance and of whether they measure up to societal standards; and (2) body shame, an affective component in which individuals feel shame when their bodies do not meet cultural ideals.42 Objectified body consciousness results from internalization of cultural messages about the ideal body and appearance, and is believed to affect females considerably more than males. Longitudinal research indicates that gender differences in self-surveillance are already present by age 11 (d = 0.49) and widen with age (d = 0.64 at age 13; d = 0.79 at age 15).43–45 Because the gender difference in self-surveillance appears so early, it is a likely contributor to the gender difference in depression.

Longitudinal research indicates that self-surveillance at age 11 predicts depressive symptoms at age 13 for girls but not boys.43 Longitudinal research has also found that, as predicted by theory, self-surveillance mediates the effect of media exposure on depressive symptoms.46


Rumination, as described in ruminative response styles theory, refers to repetitive, unproductive thoughts about one’s negative emotions.47 According to this theory, the greater tendency of girls and women to ruminate contributes to gender differences in depression. Rumination has two components: brooding (passive, perseverative thinking about one’s mood that is maladaptive) and reflection (nonjudgmental contemplation about one’s mood that is often adaptive). Consistent with the theory, a meta-analysis found a gender difference in overall rumination, d = 0.24, and brooding specifically, d = 0.19.48 Thus gender differences are found, but they are not large. Consistent with response styles theory, a meta-analysis of cross-sectional studies found that rumination correlated with depressive symptoms, r = 0.46, and that brooding correlated with depressive symptoms, r = 0.61.49 A second meta-analysis reviewed studies comparing clinical samples with a mood disorder with nonclinical control samples; those with mood disorders scored considerably higher than controls on rumination, d = 1.31.49 Because the studies in the meta-analysis are cross-sectional in design, it is impossible to know whether rumination precedes and contributes to depression, or vice versa. Some longitudinal studies have been conducted and show that rumination prospectively predicts depression.36

Co-rumination refers to ruminating with another person. According to a meta-analysis of research on samples ranging in average age from 11 to 21, females score higher than males on co-rumination, d = 0.55.50 Co-rumination correlated with depression, r = .16.

Overall, then, rumination correlates with and predicts depression, and girls and women engage in more rumination than boys and men do. The gender difference is small for rumination, but larger for co-rumination.


Because the ABC model is a vulnerability-stress model, negative life events, or stress, play a major role. It is well established that negative life events are associated with the onset of episodes of depression and with higher depression trajectories over time.51–53

A distinction can be drawn between stress exposure (the occurrence of the negative life event) and the appraisal of the event (subjectively rated on a scale from very negative to very positive). Looking broadly at both major and minor life events, a meta-analysis found that, compared with males, females were exposed to more stress, but the difference was small, d = 0.12.54 Gender differences in the appraisals of the negative events were larger than gender differences in exposure: d = 0.29 for adolescents. The gender difference in appraisals was also larger for major life events (d = 0.37 for adolescents) than for daily stresses.

Methods of measuring negative life events are often imprecise, requiring participants to recall events over long periods of time such as six months, a year, or more. Hankin and colleagues55 improved on typical approaches by using a daily diary method with adolescents. Gender differences were larger using this method; for example, for interpersonal stressors, d = 0.50.

A particularly salient daily stressor for adolescent girls is victimization from peer sexual harassment. Surveys show that girls (56%, one-year prevalence) are somewhat more likely than boys (40%) to experience harassing behavior.56 Today, sexual harassment includes not only verbal harassment (e.g., unwanted sexual comments) and physical harassment (e.g., sexual touching or groping), but also electronic harassment through text, email, Facebook, or other means. Notably, girls report being more negatively affected by the harassment, as measured by outcomes such as having trouble sleeping and not wanting to go to school.56 Longitudinal research finds that victimization from peer sexual harassment in adolescence prospectively predicts depressive symptoms; evidence also suggests a reciprocal cycle between victimization and depressive symptoms.57

Child sexual abuse, a particularly traumatic life event, is linked to depression and anxiety in adulthood.58 A meta-analytic review of data from around the world indicated that child sexual abuse is 2.4 times more prevalent among girls than boys.59 Higher rates of child sexual abuse for girls are therefore a likely factor contributing to gender differences in depression.

Sexual assault in adulthood is a major stressor that shows even more gender-differentiated prevalence rates. According to a well-sampled national U.S. survey, 18 percent of women and 1.4 percent of men have been raped at some time in their lives, for a 13:1 ratio.60


To this point, this review has considered individual factors (e.g., genetics, temperament, negative cognitive style) and interpersonal factors (e.g., victimization from peer sexual harassment) that may contribute to the gender difference in depression. At the level of analysis of culture, other factors may be involved; we categorize these sociocultural factors into structural gender inequalities and cultural factors.

Structural Gender Inequality

Structural inequality refers to differences in access to power and resources between groups—in this case, men and women.61 It includes forces such as income inequality (the gender gap in wages), lack of education of girls in some parts of the world, women’s underrepresentation in parliaments, and men’s violence against women.61 The hypothesis is that gender inequality, or women’s lack of power, should lead to more depression in women and thereby create larger gender differences in depression.

The influence of structural gender inequality on the gender difference in depression may operate through many different pathways. For example, men’s structural power has permitted men to engage in sexual harassment of women in the workplace, and research has linked experiences of sexual harassment to depression, anxiety, and PTSD.62 As another example, income inequality means that many women cannot afford to leave an abusive heterosexual relationship, and intimate partner violence contributes to depression.63,64

A number of measures of nation-level gender inequality are available, created by organizations such as the United Nations Development Programme and the World Economic Forum.65 These nation-level indicators include variables such as women’s share of parliamentary seats, women’s share of executive positions, prevalence of contraceptive use (an index of women’s reproductive freedom), and the female:male ratio in enrollment in secondary education. The meta-analysis of research on gender differences in depression described earlier found that, contrary to hypotheses, in three of six statistical tests, gender-equity indicators did not predict the magnitude of the gender difference, and in the other three tests, greater equality was positively associated with larger gender differences in depression.3 The explanation for this counterintuitive finding may lie in the somewhat subjective nature of reports of depression symptoms, and in findings that for subjective self-ratings, gender differences tend to be larger in more equitable nations.66 As a possible explanation, in equitable nations there is much more opportunity for cross-gender interactions than in inequitable nations, which tend to be much more gender segregated. For example, in a gender-equitable nation where girls have many interactions with boys, girls may be regularly comparing themselves to boys (or their situations to those of boys) and come to see themselves as higher in depression than boys, who seemingly have less depression.3 To be clear, we are not asserting that gender equality is bad for girls and women—only that different social-comparison processes occur in more- versus less-equitable nations.


Culture refers to the ideas, values, and norms found in human societies, which are transmitted to members of the society. Culture has many components, including religion, technology, art, cultural institutions, and the media. We will focus here on media influences, which have been the subject of much research.

In regard to media, the concern is that television, movies, music, and magazines (traditional media) contain massive amounts of sexual content and, in particular, content that sexually objectifies girls and women.67 Social media operate somewhat differently; adolescents can be exposed to sex-related self-disclosures from others and, in turn, are encouraged to engage in sexual self-disclosure themselves.

Research with both adolescents and adults shows that media consumption is linked to depressive symptoms.68 Much of this research has been done with single-sex samples of girls or women, and finds that media consumption predicts self-objectification and body shame, which in turn predict depressive symptoms. In one study of adolescent girls, time spent on Facebook correlated with depressive symptoms, r = .19.69 A study using two large data sets found that social-media use correlated with depressive symptoms significantly for girls (r = .06) but nonsignificantly for boys (r = .01); by contrast, sports and exercise correlated negatively with depressive symptoms for both girls (r = −.20) and boys (r = −.19).70 A review of research on sexting and its correlates with mental health measures concluded that the results are mixed; some studies find no association between sexting behaviors and depression, whereas other studies find an association of sexting with depression and with outcomes such as contemplated or attempted suicide.71 Nonetheless, the findings about media use and depression are controversial.72 One study based on three large data sets found that digital technology use accounted for only 0.4% of the variance in psychological well-being.73 The outcome measure for this study, though, was well-being, which is not precisely the inverse of depression, leaving the results for depression unknown.

An intersectionality approach implies that the effects of the media on depression should be considered for sexual minority (e.g., lesbian) and ethnic minority (e.g., black) individuals.8,74 For example, a sexual minority adolescent who lives in a small, rural town might be able to combat loneliness by interacting with other LGB youth on a social-networking site. A study of the use of social networking sites and mental health among sexual minority and heterosexual persons between the ages of 18 and 24 found that the respondents used these sites for a variety of reasons, including identity exploration, identity expression, social communication, finding a partner, escaping from loneliness, and LGB identity work.75 In the pooled sample, frequency of using social-networking sites correlated negatively with depression scores; that is, more use was related to lower depression scores. For sexual minority participants, the correlation was also negative, but not significant.

A study of black women measured social media use on Facebook, Instagram, Tumblr, Twitter, blogs, and hashtags.76 Hashtags that had been used by participants included #CarefreeBlackGirl and #BlackGirlMagic. Black-oriented blog use included The Root and For Harriet. The researchers also measured the respondents’ endorsement of the “strong black woman” ideal, believing that it is fostered by the media and can have a negative impact on mental health. The results indicated that endorsement of that ideal was strongly and positively associated with depressive symptoms. Tumblr use and blog use also were positively associated with depressive symptoms, although the effects were not as large.

Overall, then, the media’s effects on depression depend on a number of factors, including individual characteristics such as reasons for using media, ethnicity, and sexual orientation. They also depend on characteristics of the media, including sexualized portrayals of women and portrayals of the strong black woman ideal. For good or ill, the media are part of culture and must be monitored for their potential contributions to depression and gender differences in depression.


The well-known phenomenon that roughly twice as many women as men are depressed has often been translated, even by scientists, into a belief that girls and women become depressed and that boys and men do not. The result is the common and probably well-intentioned methodological practice of single-sex, female-only samples in depression research, especially with adolescents.77,78 However, with an odds ratio of 1.95 for the gender difference in depression, one implication is that, in a given sample, roughly 30%–35% of the depressed people will be men. Even at the maximum OR of 3.02 for 13- to 15-year-olds, roughly 25% of those depressed are boys. The implication is that, consistent with National Institutes of Health policy, both males and females should be included in samples for depression research.79 We need to understand the causes of depression in men just as much as we need to understand the causes of depression in women.

Regarding the genetics of depression and the genetics of gender differences in depression, it seems clear that the recent emphasis in research has been to accumulate the enormous sample sizes necessary for GWAS analyses, but high-quality measurement has often been sacrificed in the process. For example, one GWAS study managed to accumulate data from 180,866 people, but depression was measured by only two self-report items tapping just the past two weeks.17 Poor measurement can be as lethal to detecting important genetic variants as small sample size is. An important next step will be for researchers to move toward stronger measurement of depression in genetic studies.

Other important next steps for GWAS will be to integrate gender into the models, testing whether the same variants confer vulnerability similarly for women and men. In addition, GWAS need to incorporate measures of major stressors and the interaction of genes by stressors, as in the GWEIS approach.20

The findings reviewed here also imply that researchers should use models for gender-depression research that incorporate not only individual vulnerabilities (e.g., genes, temperament, cognitive style) but also highly gender-differentiated stressors such as child sexual abuse and rape, as well as sociocultural factors such as structural gender inequality and media exposure.


Given the tremendous scope of the overall individual burden associated with depression and the disproportionate amount of that burden borne by females, it is imperative that empirically supported research into etiological pathways to the gender difference in depression generate possible clinical pathways for more effective interventions. For both adolescents and adults, cognitive-behavioral (CBT) and interpersonal (IPT) therapy are well-established interventions with evidence of efficacy across multiple trials, conditions, and clinical research teams.80,81

We continue to understand very little, however, about how and for whom such interventions work—the mediators and moderators of effective treatments. In view of the findings reviewed here, such knowledge is particularly salient for the treatment of women. Current clinical and research interest in personalizing interventions suggests that the basic knowledge on gender-specific vulnerability factors and stress exposure reviewed here holds promise for improving individualized interventions. In its Guidelines for Psychological Practice with Girls and Women, the American Psychological Association provides some recommendations for the individualization of evidence-based interventions for treating depression among girls and women.82

We know that some vulnerabilities and stressors may be unique to, or elevated among, girls and women, and that these factors may differentially predict poor treatment response. Rumination has been shown to be particularly resistant to traditional CBT interventions, showing little change following active treatment and predicting greater residual symptoms and relapse.83 Mindfulness meditation techniques, however, may be successfully added to traditional CBT to directly reduce rumination, suggesting that this combination may be an important therapy to consider for depressed females.84,85 Challenging ruminative thinking specifically in response to interpersonal stressors may be another targeted approach for girls and women.82

Given the elevated rates of interpersonal trauma exposure among females and the role of such trauma in predicting depression, it is also important to consider if trauma history may moderate the effects of treatment. For traditional CBT, trauma history and specifically sexual abuse history typically predict a diminished or failed treatment effect.81,86 By contrast, IPT both theoretically and empirically may offer more potential for treating females with a history of interpersonal trauma. Gunlicks-Stoessel and colleagues87 found that the benefits of IPT were strongest for adolescents with greater peer conflict. Kubany and colleagues88 found that supplementing traditional therapies for trauma exposure with assertive communication and self-advocacy skills may maximize effectiveness among women. More research is needed to examine whether the interpersonal compensation effects of IPT or other adjunctive therapies extend to effective intervention with females who have histories of sexual or other interpersonal trauma.

Feminist psychotherapy emerged in the 1960s to address issues related to the effects of systemic discrimination and gender roles on the mental health of women. Central to feminist therapy approaches is an emphasis on developing assertiveness skills, self-esteem, and a sense of personal identity.89 Many of the evidence-based interventions reviewed above, such as challenging negative interpersonal thoughts and building self-advocacy skills, are consistent with feminist therapeutic approaches, although they are not typically presented as such. There remains a dearth of rigorous empirical research examining the effectiveness of feminist therapy for treating depression among women.

In general, research specifically examining gender or gender-linked factors as contributors to the effects of depression treatment is still limited. Methodological limitations remain, including continued reliance on single-sex samples, failure to consider the intersection of gender and race, and limited evaluation of potential mediators and moderators of treatment effects. Nevertheless, the ABC model is consistent theoretically and empirically with recent efforts to personalize interventions, and these interventions are demonstrating the potential to enhance effect sizes for treatments tailored to individuals’ unique vulnerability factors and life experiences.90


Using the ABC model as a framework, this review provides an update on research on factors that contribute to gender differences in depression. It synthesizes research on affective, biological, and cognitive vulnerabilities, together with negative life events and sociocultural factors, in a vulnerability-stress approach.

Many of the effect sizes for the hypothesized influential factors reviewed here are small. For example, among adolescent girls, time spent on Facebook correlates with depressive symptoms, r = .19.69 That is, Facebook time accounts for only 3.6% of the variance in depression. Early puberty is associated with internalizing problems for girls, d = 0.19.23 Longitudinally, negative cognitive style predicts later depression, with correlations in the range r = .22 to .30.38

Do these small effects mean that researchers do not know what causes gender differences in depression? On the contrary, researchers know a great deal about the causes. The small effect sizes are entirely consistent with the ABC model, which invokes the principle of equifinality and asserts that multiple pathways lead to depression and, in particular, to gender differences in depression. One girl may have genetic vulnerability—and, when she encounters intense social stress in the form of peer sexual harassment victimization, she becomes depressed. A woman may have a negative cognitive style—and, when faced with divorce combined with her own low earnings (caused by the wage gap), she becomes depressed. The list could go on. But the point is that different individuals have different vulnerabilities and encounter different negative life events, any combination of which might lead to depression. A single factor, such as social media use, would not be expected to have a large effect by itself because (1) its effect depends on individuals’ vulnerabilities, and (2) many other pathways contribute to differences in developing depression. The implications are that researchers must use complex models that take multiple factors into account, and that clinicians need to explore a wide variety of potential causal factors with their patients.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.


1. Hyde JS, Mezulis AH, Abramson LY. The ABCs of depression: integrating affective, biological and cognitive models to explain the emergence of the gender difference in depression. Psychol Rev 2008;115:291–313.
2. Hyde JS. The gender similarities hypothesis. Am Psychol 2005;60:581–92.
3. Salk RH, Hyde JS, Abramson LY. Gender differences in depression in representative national samples: meta-analyses of diagnoses and symptoms. Psychol Bull 2017;143:783–822.
4. Mommersteeg PM, Herr R, Pouwer F, Holt RI, Loerbroks A. The association between diabetes and an episode of depressive symptoms in the 2002 World Health Survey: an analysis of 231,797 individuals from 47 countries. Diabet Med 2013;30:e208–14.
5. Sigmon ST, Pells JJ, et al. Gender differences in self-reports of depression: the response bias hypothesis revisited. Sex Roles 2005;53:401–11.
6. Romans SE, Tyas J, Cohen MM, Silverstone T. Gender differences in the symptoms of major depressive disorder. J Nerv Ment Dis 2007;195:905–11.
7. Motro D, Ellis APJ. Boys don’t cry: gender and reactions to negative performance feedback. J Appl Psychol 2017;102:227–35.
8. Else-Quest NM, Hyde JS. Intersectionality in quantitative psychological research: I. Theoretical and epistemological issues. Psychol Women Q 2016;40:155–70.
9. Mojtabai R, Olfson M, Han B. National trends in the prevalence and treatment of depression in adolescents and young adults. Pediatrics 2016;138:e20161878.
10. Patil PA, Porche MV, Shippen NA, Dallenbach NT, Fortuna LR. Which girls, which boys? The intersectional risk for depression by race and ethnicity, and gender in the U.S. Clin Psychol Rev 2018;66:51–68.
11. Cicchetti D, Rogosch FA. Equifinality and multifinality in developmental psychopathology. Dev Psychopathol 1996;8:597–600.
12. Bleys D, Luyten P, Soenens B, Claes S. Gene-environment interactions between stress and 5-HTTLPR in depression: a meta-analytic update. J Affect Disord 2018;226:339–45.
13. Perry LA, Goldstein-Piekarski AN, Williams LM. Sex differences modulating serotonergic polymorphisms implicated in the mechanistic pathways of risk for depression and related disorders. J Neurosci Res 2017;95:737–62.
14. Border R, Johnson EC, Evans LM, et al. No support for historical candidate gene or candidate gene-by-interaction hypotheses for major depression across multiple large samples. Am J Psychiatry 2019;176:376–97.
15. Wray NR. Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression. Nat Genet 2018;50:668–81.
16. CONVERGE Consortium. Sparse whole-genome sequencing identifies two loci for major depressive disorder. Nature 2015;523:588–91.
17. Okbay A, Baselmans BML, et al. Genetic variants associated with subjective well-being, depressive symptoms, and neuroticism identified through genome-wide analyses. Nature Genet 2016;48:624–33.
18. Howard DM, Adams MJ, et al. Genome-wide meta-analysis of depression identifies 102 independent variants and highlights the importance of the prefrontal brain regions. Nat Neurosci 2019;22:343–52.
19. Trzaskowski M, Mehta D, Peyrot WJ, et al. Quantifying between-cohort and between-sex genetic heterogeneity in major depressive disorder. Am J Med Genet B Neuropsychiatr Genet 2019 Feb 1 [Epub ahead of print].
20. Dunn EC, Wiste A, Radmanesh F, et al. Genome-wide association study (GWAS) and genome-wide by environment interaction study (GWEIS) of depressive symptoms in African American and Hispanic/Latina women. Depress Anxiety 2016;33:265–80.
21. Peterson RE, Cai N, Dahl AW, et al. Molecular genetic analysis subdivided by adversity exposure suggests etiologic heterogeneity in major depression. Am J Psychiatry 2018;175:545–54.
22. Byrne ML, Whittle S, et al. A systematic review of adrenarche as a sensitive period in neurobiological development and mental health. Dev Cogn Neurosci 2017;25:12–28.
23. Ullsperger JM, Nikolas MA. A meta-analytic review of the association between pubertal timing and psychopathology in adolescence: are there sex differences in risk? Psychol Bull 2017;143:903–38.
24. Rothbart MK, Bates J. Temperament. In: Damon W, Lerner RM, eds. Handbook of child psychology, vol. 2: Cognition, perception, and language. New York: Wiley, 1998:105–76.
25. Sherman BJ, Vousura E, Wickramaratne P, Warner V. Temperament and major depression: how does difficult temperament affect frequency, severity, and duration of major depressive episodes among offspring of parents with or without depression? J Affect Disord 2016;200:82–8.
26. Lengua LJ, West SG, Sandler IN. Temperament as a predictor of symptomatology in children: addressing contamination of measures. Child Dev 1998;69:164–81.
27. Else-Quest NM, Hyde JS, Goldsmith HH, Van Hulle C. Gender differences in temperament: a meta-analysis. Psychol Bull 2006;132:33–72.
28. Mezulis AH, Salk R, Hyde JS, Priess-Groben H, Simonson J. Affective, biological, and cognitive predictors of depressive symptom trajectories in adolescence. J Abnorm Child Psychol 2014;42:539–50.
29. Bradley B, Define JA, Guarnaccia D, et al. Emotion dysregulation and negative affect: association with psychiatric symptoms. J Clin Psychiatry 2011;72:685–91.
30. Muris P, Meesters C, Blijlevens P. Self-reported reactive and regulative temperament in early adolescence: relations to internalizing and externalizing problem behaviors and “Big Three” personality factors. J Adolesc 2007;30:1035–49.
31. Verstraeten K, Vasey MW, Raes F, Bijttebier P. Temperament and risk for depressive symptoms in adolescence: mediation by rumination and moderation by effortful control. J Abnorm Child Psychol 2009;37:349–61.
32. Forbes MK, Rapee RM, Cambris A, McMahon CA. Unique associations between childhood temperament characteristics and subsequent psychology symptom trajectories from childhood to early adolescence. J Abnorm Child Psychol 2017;45:1221–33.
33. Abramson LY, Metalsky G, Alloy L. Hopelessness depression: a theory-based subtype of depression. Psychol Rev 1989;96:358–72.
34. Nolen-Hoeksema S. The role of rumination in depressive disorders and mixed anxiety/depressive symptoms. J Abnorm Psychol 2000;109:504–11.
35. Alloy L, Abramson L, Hogan M, et al. The Temple-Wisconsin Cognitive Vulnerability to Depression (CVD) Project: lifetime history of Axis I psychopathology in individuals at high and low cognitive risk for depression. J Abnorm Psychol 2000;109:403–18.
36. Giollabhui NM, Hamilton JL, Nielsen J, et al. Negative cognitive style interacts with negative life events to predict first onset of a major depressive episode in adolescence via hopelessness. J Abnorm Psychol 2018;127:1–11.
37. Hankin BL, Abramson L, Siler M. A prospective test of the hopelessness theory of depression in adolescence. Cognit Ther Res 2001;25:607–32.
38. Huang C. Relation between attributional style and subsequent depressive symptoms: a systematic review and meta-analysis of longitudinal studies. Cognit Ther Res 2015;39:721–35.
39. Abela J. The hopelessness theory of depression: a test of the diathesis-stress and causal mediation components in third and seventh grade children. J Abnorm Child Psychol 2001;29:241–54.
40. Mezulis AH, Hyde JS, Abramson LY. The developmental origins of cognitive vulnerability to depression: temperament, parenting, and negative life events. Dev Psychol 2006;42:1012–25.
41. Hankin BL, Abramson L. Measuring cognitive vulnerability to depression in adolescence: reliability, validity, and gender differences. J Clin Child Adolesc Psychol 2002;31:491–504.
42. McKinley NM, Hyde JS. The Objectified Body Consciousness Scale: development and validation. Psychol Women Q 1996;20:181–215.
43. Grabe S, Hyde JS, Lindberg S. Body objectification and depression in adolescents: the role of gender, shame, and rumination. Psychol Women Q 2007;31:164–75.
44. Lindberg SM, Hyde JS, McKinley NM. A measure of objectified body consciousness for pre-adolescent and adolescent youth. Psychol Women Q 2006;30:65–76.
45. Petersen JL, Hyde JS. Peer sexual harassment and disordered eating in early adolescence. Dev Psychol 2013;49:184–95.
46. Grabe S, Hyde JS. Body objectification, MTV, and psychological outcomes among female adolescents. J Appl Soc Psychol 2009;39:2840–58.
47. Nolen-Hoeksema S. Responses to depression and their effects on the duration of depressive episodes. J Abnorm Psychol 1991;100:569–82.
48. Johnson DP, Whisman MA. Gender differences in rumination: a meta-analysis. Pers Individ Dif 2013;55:367–74.
49. Olatunji BO, Naragon-Gainey K, Wolitzky-Taylor KB. Specificity of rumination in anxiety and depression: a multimodal meta-analysis. Clin Psychol Sci Pract 2013;20:225–57.
50. Spendelow JS, Simonnds LM, Avery RE. The relationship between co-rumination and internalizing problems: a systematic review and meta-analysis. Clin Psychol Psychother 2017;24:512–27.
51. Kendler KS, Karkowski L, Prescott C. Causal relationship between stressful life events and the onset of major depression. Am J Psychiatry 1999;156:837–41.
52. Shore L, Toumbourou JW, Lewis AJ, Kremer P. Review: Longitudinal trajectories of child and adolescent depressive symptoms and their predictors—a systematic review and meta-analysis. Child Adolesc Ment Health 2018;23:107–20.
53. Stroud CB, Davila J, Moyer A. The relationship between stress and depression in first onsets versus recurrences: a meta-analytic review. J Abnorm Psychol 2008;117:206–13.
54. Davis M, Matthews K, Twamley E. Is life more difficult on Mars or Venus? A meta-analytic review of sex differences in major and minor life events. Ann Behav Med 1999;21:83–97.
55. Hankin BL, Mermelstein R, Roesch L. Sex differences in adolescent depression: stress exposure and reactivity models. Child Dev 2007;78:279–95.
56. Hill C. Crossing the line: sexual harassment at school. 2019.
57. Dahlqvist HZ, Landstedt E, Young R, Gädin KG. Dimension of peer sexual harassment victimization and depressive symptoms in adolescence: a longitudinal cross-lagged study in a Swedish sample. J Youth Adolesc 2016;45:858–73.
58. Gallo EAG, Munhoz TN, de Mola CL, Murray J. Gender differences in the effects of childhood maltreatment on adult depression and anxiety: a systematic review and meta-analysis. Child Abuse Negl 2018;79:107–14.
59. Stoltenborgh M, Bakermans-Kranenburg MJ, Alink LRA, van IJzendoorn MH. The prevalence of child maltreatment across the globe: review of a series of meta-analyses. Child Abuse Rev 2015;24:37–50.
60. Black MC, Basile KC, Breiding MJ, et al. The National Intimate Partner and Sexual Violence Survey (NISVS): 2010 summary report. Atlanta, GA: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control.
61. Kunst JR, Fischer R, Sidanius J, Thomsen L. Preferences for group dominance track and mediate the effects of macro-level social inequality and violence across societies. Proc Natl Acad Sci U S A 2017;114:5407–12.
62. Klingensmith K. Military sexual trauma in US veterans: results from the National Health and Resilience in Veterans Study. J Clin Psychiatry 2014;75:e1133.
63. Anderson KL. Who gets out? Gender as structure and the dissolution of violent heterosexual relationships. Gend Soc 2007;21:173–201.
64. Coker AL, Williams CM, Follingstad DR, Jordan CE. Psychological, reproductive and maternal health, behavioral, and economic impact of intimate partner violence. In: White JW, Koss MP, Kazdin AE, eds. Violence against women and children: mapping the terrain. Washington, DC: American Psychological Association, 2011:265–84.
65. Else-Quest NM, Grabe S. The political is personal: measurement and application of nation-level indicators of gender equity in psychological research. Psychol Women Q 2012;36:131–44.
66. Wood W, Eagly AH. Biosocial construction of sex differences and similarities in behavior. Adv Exp Soc Psychol 2012;46:55–123.
67. Collins RL, Strasburger VC, Brown JD, Donnerstein E, Lenhart A, Ward LM. Sexual media and childhood well-being and health. Pediatrics 2017;140(suppl 2):S162–6.
68. Tiggemann M. Mental health risks of self-objectification: a review of the empirical evidence for disordered eating, depressed mood, and sexual dysfunction. In: Calogero RM, Tantleff-Dunn S, Thompson JK, eds. Self-objectification in women: causes, consequences, and counteractions. Washington, DC: American Psychological Association, 2011.
69. Tiggemann M, Slater A. The role of self-objectification in the mental health of early adolescent girls: predictors and consequences. J Pediatr Psychol 2015;40:704–11.
70. Twenge JM, Joiner TE, Rogers ML, Martin GN. Increases in depressive symptoms, suicide-related outcomes, and suicide rates among U.S. adolescents after 2010 and links to increased new media screen time. Clin Psychol Sci 2018;6:3–17.
71. Klettke B, Hallford DH, Mellor DJ. Sexting prevalence and correlates: a systematic literature review. Clin Psychol Rev 2014;34:44–53.
72. Heffer T, Good M, Daly O, MacDonell E, Willoughby T. The longitudinal association between social-media use and depressive symptoms among adolescents and young adults: an empirical reply to Twenge et al. (2018). Clin Psychol Sci 2019;4:462–70.
73. Orben A, Przybylski AK. The association between adolescent well-being and digital technology use. Nat Hum Behav 2019;3:173–82.
74. Cole ER. Intersectionality and research in psychology. Am Psychol 2009;64:170–80.
75. Ceglarek PJD, Ward LM. A tool for help or harm? How associations between social networking use, social support, and mental health differ for sexual minority and heterosexual youth. Comput Hum Behav 2016;65:201–9.
76. Stanton AG, Jerald MC, Ward LM, Avery LR. Social media contributions to Strong Black Woman ideal endorsement and black women’s mental health. Psychol Women Q 2017;41:465–78.
77. Speed BC, Nelson BD, Auerbach RP, Klein DN, Hajcak G. Depression risk and electrocortical reactivity during self-referential emotional processing in 8 to 14 year-old girls. J Abnorm Psychol 2016;125:607–19.
78. Vilgis V, Gelardi KL, Helm JL, et al. Dorsomedial prefrontal activity to sadness predicts later emotion suppression and depression severity in adolescent girls. Child Dev 2018;89:758–72.
79. Clayton JA, Collins FS. NIH to balance sex in cell and animal studies. Nature 2014;509:282–4.
80. Hans E, Hiller W. Effectiveness of and dropout from outpatient cognitive-behavioral therapy for adult unipolar depression: a meta-analysis of nonrandomized effectiveness studies. J Consult Clin Psychol 2013;81:75–88.
81. Weersing VR, Jeffreys M, Minh-Chau TD, Schwartz KTG, Bolano C. Evidence base update of psychosocial treatments for child and adolescent depression. J Clin Child Adolesc Psychol 2017;46:11–43.
82. Mazure CM, Keita GP, Blehar MC. Summit on women and depression: proceedings and recommendations. Washington DC: American Psychological Association, 2002.
83. Riso LP, du Toit PL, Blandino JA, et al. Cognitive aspects of chronic depression. J Abnorm Psychol 2003;112:72–80.
84. Geschwind N, Peeters F, Huibers M, van Os J, Wichers M. Efficacy of mindfulness-based cognitive therapy in relation to prior history of depression: randomized controlled trial. Br J Psychiatry 2012;20:320–5.
85. van Aalderen JR, Donders AR, Giommi F, Spinhoven P, Barendregt HP, Speckens AE. The efficacy of mindfulness-based cognitive therapy in recurrent depressed patients with and without a current depressive episode: a randomized controlled trial. Psychol Med 2012;42:989–1001.
86. Lewis C, Simons A, Nguyen L, et al. Impact of childhood trauma on treatment outcome in the Treatment for Adolescents with Depression Study (TADS). J Am Acad Child Adolesc Psychiatry 2010;49:132–40.
87. Gunlicks-Stoessel M, Mufson L, Jekal A, Turner AB. The impact of perceived interpersonal functioning on treatment for adolescent depression: IPT-A versus treatment as usual in school-based health clinics. J Consult Clin Psychol 2010;78:260–7.
88. Kubany ES, Hill EE, Owens JA. Cognitive trauma therapy for battered women with PTSD: preliminary findings. J Trauma Stress 2003;1:81–91.
89. Brown LS. Feminist therapy. 2nd ed. Washington DC: American Psychological Association, 2018.
90. Shoham V, Insel TR. Rebooting for whom? Portfolios, technology, and personalized intervention. Perspect Psychol Sci 2011;6:478–82.

child sexual abuse; depression; gender differences; genetics; media; negative cognitive style; objectified body consciousness; puberty; rumination; stress

© 2020 President and Fellows of Harvard College