Gender Orientation and the Cost of Caring for Others

Abstract

The “cost-of-caring” thesis asserts that observed gender differences in psychological distress are largely a consequence of women’s greater emotional investment in the lives of their loved ones. Research on this topic has supported this thesis by showing that network events result in higher levels of depressive symptoms for women compared to men. However, other evidence challenges this claim. In light of these divergent findings, this paper elaborates this topic in three ways. First, susceptibility to network events is assessed in terms of two dimensions of psychological distress, depressive symptomatology and problem drinking. Second, within-gender analyses are conducted to examine the possibility that masculine and feminine personality traits condition the relationship between network events and psychological distress. Finally, this paper assesses age variation in the cost of caring. Data collected from 1,393 respondents ages 18 to 55 who participated in a Toronto-based community study are employed to address these issues. Findings reveal that the cost of caring for others extends to women and men, and that gender orientation modifies the relationship between network events and psychological distress. These results underscore the need to critically assess the social factors that differentiate risk and well-being for men and women.

Keywords

network events gender orientation masculinity femininity

Is the stress experienced by loved ones more detrimental to the mental health of women relative to men? When assessed in terms of depressive symptoms, most prior research has confirmed this association in support of the “cost-of-caring” hypothesis: Women respond with greater psychological distress following the occurrence of network events (which are defined as stressful life events experienced by one’s significant others) (Kessler and MacLeod 1984; Kessler, McLeod, and Wethington 1985; Wethington, McLeod, and Kessler 1987; R. Turner and Avison 1989; Roxburgh 2005). Gender role theory (Gove and Tudor 1973) and gender identity theory (B. P. Dohrenwend and Dohrenwend 1976) have provided compelling explanations for this disparity.

However, subsequent research suggests that the conclusions drawn from prior research may be premature. There are at least two strands of research that suggest this possibility. First, prior research has made clear that gender differences in responses to stress are disorder specific (Aneshensel, Rutter, and Lachenbruch 1991; Conger et al. 1993; Horwitz, White, and Howell-White 1996; Simon 1998; Aneshensel 2002; Simon, 2002; Van Gundy 2002). Thus, although overall rates of disorder are roughly equivalent, women are more likely to experience more symptoms of anxiety and depression compared to men, while men tend to respond to stress in terms of behavioral and substance-use problems. Thus, when assessments of the consequences of network events are limited to symptoms of depression, the deck is likely stacked in favor of observing female vulnerability. Equally important, conclusions drawn from these assessments may underestimate the effects of stress exposure for men (Simon 2014).

Second, previous research on gender identity, defined by Horwitz and White (1987:159) as “the way that individuals relate to masculine and feminine qualities of behavior,” raises the possibility that there is intragender variation in vulnerability to stressful events for both women and men. For example, risk for psychological distress has been shown to vary across certain masculine and feminine personality traits and by level of interpersonal dependency (R. Turner and Turner 1999; Barrett and White 2002). Variations in these gender-role orientations may explain or moderate the relationship between exposure to network events and psychological distress.

In addition, little is known about the how vulnerability to network events may vary by age. Separate bodies of research have provided competing theories about the relationship between age and exposure to network events. One posits that risk is greatest in late adolescence and young adulthood. A second perspective suggests that middle adulthood is the stage in the life course when investment in others is greatest. However, to date, there is little empirical evidence in support of either perspective.

Based on this more recent information, this paper has three objectives. First, I assess age distributions in the relationship between network events and psychological distress. Second, following the lead of Aneshensel and others, this paper revisits the question of whether gender differences in vulnerability to network events are outcome specific. Specifically, I evaluate the effects of network events across two outcomes, depressive symptoms and alcohol dependence problems. Third, I expand upon previous research on this topic by assessing the mediating and moderating effects of gender-role orientation in the relationship between network events and psychological distress.

Background

Gender and the Cost of Caring

The cost-of-caring hypothesis is based on evidence demonstrating that gendered roles and identities place women at greater risk for psychological distress. Wethington and colleagues (1987:144-45) define the cost of caring as the tendency for women to be “more affected emotionally than men not only by their own stressful experiences, but also by the stressful experiences of people they care about. Women’s roles obligate them to respond to the needs of others.” Prior research also suggests gender contrasts with respect to differences in the levels of exposure to stress as well as gender differences in vulnerability to stressful events and circumstance (for a recent review and interpretation of the literature, see Rosenfield and Mouzon 2013). Women’s social roles result in their exposure to a greater number of network events compared to men. Consistent with the differential vulnerability explanation, women and men differ in the type and amount of stress exposure that they experience as well as in the coping strategies that they employ in response to stressful circumstance (B. S. Dohrenwend 1973; Kessler 1979; Radloff and Rae 1981; Maciejewski, Prigerson, and Mazure 2001; Rosenfield and Mouzon 2013).

The cost-of-caring thesis (Kessler and MacLeod 1984; Kessler et al. 1985; Wethington et al. 1987) was developed as a synthesis of stress-exposure and gender-role vulnerability theories. Collectively, this research has suggested that the greater demand from others for female nurturance underlies women’s heightened vulnerability to stress exposure. In sum, the cost-of-caring argument posits that women experience greater exposure and vulnerability to network events. Women’s greater emotional and instrumental involvement in network events, in turn, is protective for men because it shields them from such obligations. Prior research tends to support this theory. Kessler and McLeod’s (1984) initial study on this topic confirmed that female vulnerability to life events was largely a consequence of adversity experienced by their significant others and that greater female vulnerability explained a significant part of the overall relationship between gender and depression. In addition, later studies (R. Turner and Avison 1989; Aneshensel et al. 1991; Gore, Aseltine, and Colten 1993) found that men were more responsive than women to negative life events but only when events occurred to self.

Gender Orientation and the Cost of Caring

Gendered Roles

Gender-role theory posits that women are at greater risk for psychological distress because of the nature and status of their roles relative to men. Women’s disadvantage also arises from the imposition of gender stratification at home and work (Gore and Mangione 1983; Ross and Huber 1985; Lennon and Rosenfield 1992; Pugliesi 1995; Ferree 2010). Women also typically occupy social roles that require them to be the providers of unpaid labor, emotional support, empathy, and nurturing (Ross and Mirowsky 1989; Almeida and Kessler 1998; R. Turner and Turner 1999; Schieman and Turner 2001; Ferree 2010). Compared to men, women are more likely to be care providers and to spend more hours providing care (Wethington et al. 1987; Lennon and Rosenfield 1992; Umberson et al. 1996).¹ Women also tend to nurture larger social networks (H. Turner 1994; Rosenfield and Mouzon 2013) and to receive and donate more social support than men (H. Turner 1994; R. Turner and Turner 1999; however, for contrasting findings, see Belle 1982).

Moreover, gender differences in intimate affiliations have been shown to expose women to higher levels of negative social interactions (Rosenfield 1992; H. Turner 1994; Ridgeway 2011) and psychological burden in marital, parental, and care-provider roles (Gove 1972; Gove and Tudor 1973; Conger et al. 1993; Lennon and Rosenfield 1992; Umberson et al. 1996; Simon 1997; Reczek and Umberson 2012), especially when these responsibilities compete with time spent in the paid labor force (Roxburgh 2004; Simon 2014).

Gendered Identities

Research on gender and identity is also congruent with the cost-of-caring theory. Gender identity refers to individual’s self-referenced orientation with respect to traditional masculine and feminine behaviors and personal attributes (Horwitz and White 1987). Thus, women who tend to orient toward feminine traits should be at greater risk in response to network events because of their greater emotional investment in the lives of their significant others. By extension, this pattern of associations should also apply to men who rate high in femininity.

Several studies support this premise. For example, Thomeer and colleagues’ (Thomeer, Umberson, and Pudrovska 2013) recent study of a two samples of married couples found that the wives of depressed men were much more likely to engage in caregiving efforts on behalf of their husbands compared to husbands whose wives were depressed. Moreover, depressed wives enacted emotion work to protect their husbands from the burden of having a depressed wife. This pattern of caregiving was not observed among the depressed husbands in this sample. These findings are consistent with an earlier study showing that women are more likely to internalize psychological schemas that privilege others over self (Rosenfield, Lennon, and White 2005).

Pudrovska’s (2010) study found that the experience of cancer was more depressing for men compared to women and that this difference was explained away when masculine identity was controlled for. In addition, Longest and Thoits (2012) found that men who react to stress in “female typical ways” are at much greater risk for internalizing disorders compared to those men who maintained more traditionally masculine tendencies. These studies suggest that high levels of femininity and low levels of masculinity increase risk for depressive symptoms and that masculinity may operate as a more powerful mediator and/or moderator of distress in response to network events. Alternatively, individuals who rate high in masculinity and low in femininity may have a higher predisposition for externalizing behaviors. I evaluate these possibilities below.

Interpersonal Dependency

Emotional reliance and assertion of autonomy are two subdimensions of interpersonal dependency. They reflect contrasting beliefs and behaviors that are employed to maintain feelings of self-worth. They are also hypothesized to capture masculine and feminine personality traits. Assertion of autonomy is thought be a masculine tendency, and emotional reliance is hypothesized to be an estimate of femininity. Hirschfeld et al. (1977:610) define interpersonal dependency as “a complex of thoughts, beliefs, feelings, and behaviors which revolve around the need to associate closely with, interact with, and rely upon valued other people.” Emotionally reliant persons depend almost exclusively on the love and attention of valued others for the maintenance of positive self-evaluations (Hirschfeld et al. 1976). On the other hand, assertion of autonomy is an important dimension of masculinity because it privileges self over others and reflects one’s tendency to maintain positive feelings of self-regard independent of the appraisals of others.

Evidence that emotional reliance is higher among women than among men and that it is positively associated with depression (R. Turner and Turner 1999) raises the possibility that variations in this personality characteristic may mediate observed female vulnerability to network events or perhaps moderate the relationship between these stressful events and psychological distress.

Intragender Variation

Prior research has demonstrated that masculine and feminine personality traits are distributed along a continuum for both women and men (Ferree 2010; Thomeer et al. 2013). Consistent with this view, Connell and Messerschmidt (2005:836) have observed that masculinity “is not a fixed entity embedded in the body or personality traits of individuals. [Instead, masculinities] . . . can differ according to the gender relations in a particular social setting.” Several studies support this statement. For example, Annandale and Hunt (1990) found that high levels of masculinity were associated with better health outcomes for men and women. Ueno (2010) found that self-identification of attractiveness to the same sex was more distressing for women compared to men, and Chuick and colleagues’ (2009) study revealed that masculine socialization was a deterrent to help-seeking behavior among a sample of men with histories of depression.

However, to date, the exact relationship between masculinity, femininity, and mental health remains unclear. While gender-congruent orientations (i.e., males with high levels of masculinity and females with high levels of high femininity) were generally thought to facilitate optimal well-being, other research suggests that masculinity, and not femininity, is a “central axis on which advantages and disadvantages across some dimensions of mental health accumulate” (Barrett and White 2002:451). As such, this study examines the extent to which stress reactivity differs for those high in femininity compared to those high in masculinity, within and across gender status.

Age and the Cost of Caring

There is contrasting evidence regarding the possible linkages between age, network events, and psychological distress. One body of research points toward increased risk among young adults. This is based on research showing that levels of stress exposure (R. Turner, Wheaton, and Lloyd 1995; Hatch and Dohrenwend 2007), depressive symptoms (R. Turner et al. 1995; Schieman, Van Gundy, and Taylor 2002; Ferraro and Wilkinson 2013; George 2014) and alcohol abuse (Kerr et al., 2009) are highest in young adulthood, especially among the unmarried (Uecker 2012). In addition, the size of social networks is also largest among young adults (Haas, Schaefer, and Kornienko 2010) and decreases in later adulthood (Cartensen, Isaacowitz, and Charles 1999). These findings suggest that young adults are at risk for greater exposure and vulnerability to network events.

A second body of research suggests that middle adulthood is the life stage when network events are most salient. In contrast to adolescents and younger adults, midlife is most often characterized by an outward rather than an inward focus as one’s duties and obligations to others increase (Bush and Simmons 1992; Fleeson 2001; Rossi 2001). Midlife is marked by the transition into leadership roles at work, with family, and in the community (Wethington, Cooper, and Holmes 1997; Umberson 2003). From this perspective, age can be viewed as an analogy for experiential growth that is associated with increases in maturity, responsibility, and emotional commitment to others (Mirowsky and Ross 1992; McFadden 1996; Schieman 2001; Umberson 2003). Moreover, this outward focus likely transcends occupancy in specific roles. For example Rossi (2001:102-103) has observed that most people in middle adulthood “contribute to their communities and nations through their primary ties to children, parents, siblings and friends, through the work they do to earn their way in life, and though involuntary, through the taxes that they pay, which provides needed services to the poor, the sick and disabled, and the elderly.”

While commitment and responsibility to others is emotionally rewarding because it promotes a sense of meaning and purpose, it may be taxing as well. As Rosenberg and McCullough (1981:165) have stated, “[we are bound to others] not only by virtue of our dependence on others but by their dependence on us.” However, the rewards of caring for and about others can be tempered by the emotional costs of seeing loved ones suffer or perform badly. Thus, network events are likely to be most damaging during middle adulthood because this is a life stage when one’s relationships to others tend to be defined by a sense of responsibility, caring, and concern.

Challenges to the Cost-of-caring Hypothesis

There are three important challenges to the claim that the mental health costs of network events are greater for women. First, most prior studies have limited their consideration to depressive symptoms. Research has now demonstrated gender parity with respect to overall rates of disorder and that gender differences in mental health responses to stress are disorder specific (Thoits 1982; Aneshensel et al. 1991; Conger et al. 1993; R. Turner et al. 1995; Horwitz et al. 1996; Simon 1998; Aneshensel 2002; Simon 2002; Van Gundy 2002; Elliot 2013). Men tend to respond to stress by abusing alcohol or drugs, while women are at greater risk for depression and anxiety. Because prior research on the cost of caring has employed depressive symptoms as the outcome of interest, and because women tend to be more depressed than men, the issue of gender differences in vulnerability to network events remains unsettled.

This potential bias was articulated by Aneshensel et al. (1991:176), who remark, “[S]tress research which focuses on a single disorder fails to portray accurately social variation in stress processes and mental health outcomes.” To correct for this limitation, they examined the role of events to self and network events in predicting four outcomes: affective and anxiety symptomatology, substance-use problems, affective and anxiety disorder, and substance-use disorder. Their findings provide only partial support for the cost-of-caring thesis. In contrast to expectation, network events were found to be more strongly related to men’s affective or anxiety disorder compared to women and more strongly related to substance-use disorder among women compared to men (p < .06). Gender × Network Event interaction terms were unrelated to affective and anxiety symptomatology and substance-use problems. At the very least, these findings call into question the assumption that the burden of caring for others is limited to women.

Second, as mentioned above, prior investigations of the cost of caring have conceptualized gender as being the same for men and women. Subsequent studies have confirmed the limitation of this research approach. It is now well established that gender orientations vary within and across gender and across age (Horwitz and White 1987).

Third, the statistical associations reported in prior studies have been relatively weak. For example, R. Turner and Avison (1989) were able to observe statistically significant Gender × Network Event interaction terms only within the physically disabled portion of their sample. No such relationships were observed within their nondisabled comparison sample.

Based on the literature reviewed above, the following hypotheses are tested in the analyses that follow.

Hypothesis 1: Network events will be more detrimental to the mental health of study participants in middle adulthood compared to younger respondents.

Hypothesis 2: The coefficient for the Gender × Network Events interaction term will be positive and significant, indicating that the magnitude of the association between network events and depressive symptoms is stronger among women compared to men.

Hypothesis 3: Masculine personality orientation variables (masculinity and assertion of autonomy) will mediate the interaction term predicted in Hypothesis 2.

Hypothesis 4: In analyses conducted separately for men and women, the coefficients for the network events by feminine orientation variables will be positive and significant, indicating that the magnitude of the association between network events and depressive symptoms is stronger among men and women who are high in femininity compared to those who are not.

Hypothesis 5: The coefficient for Gender × Network Events interaction term will be negative and significant, demonstrating that the magnitude of the association between network events and alcohol problems is stronger for men than for women.²

Hypothesis 6: Feminine personality orientation variables (femininity and emotional reliance) will mediate the interaction term predicted in Hypothesis 5.

Hypothesis 7: In analyses conducted separately for men and women, the coefficients for the network events by masculine orientation variables will be positive and significant, indicating that the magnitude of the association between network events and alcohol problems is stronger among men and women who are high in masculinity compared to those who are not.

Method

Design and Sample

The study design consisted of two waves of interviews with a representative sample of adult residents selected from the six boroughs of metropolitan Toronto, Canada. The second wave of data collection was approximately 12 months following initial interviews, which were conducted from October 1990 to June 1991. Eligible participants included all individuals 18 to 55 years of age who were living in their principal residence and fluent in English. Statistics Canada had conducted a household enumeration of the Toronto area in preparation for a separate study. This allowed development of a nonoverlapping, representative sample of 3,415 residential addresses within the six boroughs of Toronto in proportion to the 1986 population. The Kish ([1965] 1995) procedure was used to ensure the random selection of a single individual from each sampled household, and the resulting unequal probability of selection associated with variations in household size has been corrected for by the application of appropriate weights. This yielded a wave 1 sample size of 1,393 study participants. Comparison of these weighted sample frequencies with population values revealed that women and the more highly educated were slightly overrepresented. A second set of weights were also applied to more completely align the sample with census values. The success rate in interviewing selected subjects was 76 percent at wave 1 and 87 percent at wave 2. The analyses presented below are limited to wave 1 survey responses. The study sample consisted of almost 20 percent are visible minorities, with the largest group (Chinese) accounting for 5.3 percent of the sample.

Measurement

Demographics

Age is measured in years based on time 1 self-reports. In multivariate analyses of the full sample, gender is coded 1 for females and 0 for males. Socioeconomic status is operationalized using a composite measure based on occupational level, educational attainment, and income. Hollingshead’s (1965) categorization of occupational prestige was employed for coding. Occupational prestige scores, years of education, and annual household income were standardized and then summed to achieve equal weighting. Marital status is coded 1 for currently married and 0 for currently separated, divorced, widowed, or never married. Parental status is coded 1 for parents and 0 for others. Number of friends is a continuous measure in response to the following question: “How many friends do you have who live nearby, say within an hour’s drive?”

Life Events/Network Events

Life events were assessed using a 34-item checklist of negative events experienced by respondents and/or their partners, parents, other relatives and friends during the 12 months preceding the interview (Avison and Turner 1988; R. Turner and Avison 1992; R. Turner et al. 1995). Two separate counts of events were computed, one of events occurring to the respondent (events to self) and the other composed of events occurring to the respondent’s significant others (network events).³ Examples of life events include reporting a serious accident or injury, an unwanted pregnancy, experiencing a major financial crisis, and experiencing trouble with the law.

Depressive Symptomatology

The Center for Epidemiologic Studies Depression Scale (CES-D; Radloff 1977) is a 20-item scale that is a widely used and highly reliable index of depressive symptomatology. Examples of the symptoms measured in this scale include “I felt depressed,” “I thought my life had been a failure,” “I had crying spells,” and “I could not get ‘going.’” This measure assesses symptoms occurring during the past two weeks. In this study, the reliability coefficient for the CES-D is .90.

Alcohol Dependence Problems (12 Months)

This measure, developed by Robins et al. (1981) assessed criteria from the Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.; American Psychiatric Association 1987) for alcohol-related problems experienced in the 12 months prior to interview. In the analyses that follow, symptoms of alcohol dependence are assessed by a simple count of eight alcohol-related problems. These problems are as follows: drinking larger amounts than intended, symptoms of withdrawal in the absence of drinking, drinking to overcome withdrawal, unsuccessful attempts to drink less, drinking-related neglect of social activities and inability to fulfill social roles and obligations, problematic time spent devoted to drinking and recovering from the effects of drinking, drinking in physically hazardous situations (such as while driving), and continued drinking despite knowledge of having a persistent or recurring social, psychological, or physical problem that is caused or exacerbated by drinking.

Gender-Role Orientation

Gender-role orientation was measured by four variables: emotional reliance, assertion of autonomy, femininity, and masculinity. Emotional reliance and assertion of autonomy are subdimensions of a larger construct called interpersonal dependency. Assertion of autonomy is hypothesized to measure masculine personality traits, and emotional reliance is thought to be an assessment of feminine tendencies. Preliminary analyses (not shown) confirm this. Emotional reliance is negatively correlated with assertion of autonomy (r = −.101) and masculinity (r = −266) and positively correlated with femininity (r = .237). Assertion of autonomy is positively associated with masculinity (r = .284) and negatively associated with femininity (r = −.204).

The measures of interpersonal dependency were developed by Hirschfeld et al. (1977) to assess the relevance of overdependence on others for psychiatric pathology. Factor analyses (not shown) confirmed the distinctiveness of these two subscales, and internal reliabilities of .66 and .71 were obtained for emotional reliance and assertion of autonomy, respectively. However, because crosscutting agreement tendencies associated with positively and negatively worded questions have been shown to both exaggerate relationships among similarly worded items and suppress relationships across differently worded items (Mirowsky and Ross 1996), I verified the presence of distinct latent constructs by comparing measurement models through confirmatory analysis. This test confirmed that these measures comprise two substantive factors within the construct of interpersonal dependency, controlling for the effects of agreement tendencies. Participants responded on a five-point Likert scale ranging from agree strongly to disagree strongly. Higher scores represent greater levels of each characteristic.

Femininity and Masculinity

Femininity and masculinity were measured by a modified version of the Spence and Helmreich’s (1978) Personal Attributes Questionnaire. Respondents were asked to rate themselves on a series of bipolar items. The items assessing masculinity required respondents to rate themselves on a 1-to-7 scale ranging from 1 = not at all masculine to 7 = very masculine. Examples of the five personal characteristics assessed were aggressiveness, self-persistence, and self-confidence. The femininity items were constructed in an identical fashion ranging from not at all feminine to very feminine. These four items measured self-estimates of tenderheartedness, being excitable in a major crisis, and independence. The internal reliability for each scale was .60.

Results

Preliminary descriptive and multivariate analyses (not shown) were conducted to assess the possibility that the relationship between network events and psychological well-being varies by age. The findings were affirmative. Descriptive analyses revealed that 68.6 percent of all alcohol problems and 86.9 percent of the total number of network events were reported by study participants age 30 or younger. CES-D scores were also highest among respondents in this age group. Multivariate analyses confirmed that network events are most damaging for younger study participants. Regression models with the full sample and among respondents age 31 or older showed no significant linkages between network events and the outcomes of interest. Significant associations were observed only when analyses were restricted to the younger portion of the sample. Therefore, hypothesis 1 was not supported, and the analyses that follow are limited to the 539 study participants ages 30 and younger.

Descriptive information for all study variables is presented in Table 1. These findings conform to those of previous investigations. Women reported higher depression scores, while mean scores for alcohol problems were roughly three times greater for men than those reported by women. In addition, women reported higher levels of femininity and emotional reliance compared to men, who scored higher on measures of masculinity and assertion of autonomy.

Table 1.

Mean Values of Variables of Interest, by Gender (Females, n = 309; Males, n = 230).

Variable	Male	Female
Depressive symptoms	12.220***	15.239
Alcohol problems	0.140***	0.050
Events to self	1.149	1.404
Network events	0.840	0.915
Emotional reliance	13.811***	15.193
Assertion of autonomy	9.428*	8.893
Femininity	19.421***	21.621
Masculinity	30.449***	28.411
Age	24.077	24.371
Socioeconomic status	−0.040	0.032
Marital status	0.155***	0.276
Parental status	0.068***	0.187
Friends	4.344*	3.891
Employment status	0.577	0.545

p < .05. **p < .01. ***p < .001.

Table 2 presents the distribution of depressive symptoms and alcohol problems across gender-role orientation and stress variables, by gender. For the most part, depressive symptomatology is distributed in a linear pattern and in the expected direction across the risk and protective variables considered here. The highest mean levels of depression are distributed across the high categories of emotional reliance and femininity, and the lowest depression scores are distributed across the high categories of assertion of autonomy and masculinity. This pattern was not observed with respect to men’s depression scores across the female orientation variables. Here, depressive symptoms were distributed in inverse u-shaped pattern. Also, depressive symptoms are positively associated with network events and life events for women and men. Although the contrasts are not statistically significant, alcohol problems are positively related to both stress-exposure variables for women and men. However, no clear pattern of distributions was observed across the gender orientation variables.

Table 2.

Mean Depression and Alcohol Problems Scores across Gender-role Orientation and Stress Variables, by Gender (Females, n = 309; Males, n = 230).

Variable	Depressive symptoms		Alcohol problems
	Men	Women	Men	Women
Emotional reliance
High reliance	12.81^a	19.41	.11	.05
Medium reliance	13.03	13.99^b	.13^a	.03
Low reliance	11.30	11.09^b	.15^a	.03
Assertion of autonomy
High autonomy	9.04^a	13.00	.11	.07
Medium autonomy	12.36^a	15.49^b	.14^a	.03
Low autonomy	14.54^b	16.50^b	.15^a	.04
Femininity
High femininity	12.12	18.10	.17^a	.06
Medium femininity	13.03	13.16^b	.06	.01
Low femininity	11.78^b	13.10^b	.17^a	.06
Masculinity
High masculinity	10.18	11.97^{b, c}	.16^a	.04
Medium masculinity	11.91	12.73	.11	.03
Low masculinity	15.26^{a, b}	18.81	.14^a	.06
Events to self
2 or more events	14.49	18.51	.25^b	.09
1 event	12.90^b	15.63	.10^{a, b}	.02
0 events	9.50^b	11.32^{b, c}	.05	.02
Network events
2 or more events	16.25	19.90	.25^a	.10
1 event	11.73^a	15.17^b	.16^a	.03
0 events	10.72	13.02^b	.08^b	.03

Note. Significant at p < .05. For variables with high, medium, and low categories, the distribution of scores was divided equally into tertiles.

Significant gender contrast.

Significant contrast with the high category.

Significant contrast with the moderate category.

Table 3 tests hypotheses 2 and 3. Here, depressive symptoms are regressed on statistical controls, stress exposure, and gender orientation. Model 1 demonstrates the independent contributions of events to self and network events in the prediction of depressive symptoms. Model 2 tests hypothesis 2 by adding the multiplicative interaction of female with network events.⁴ The coefficient for that term reveals whether there is a gender difference in the effect of network events on depressive symptomatology. Here, the positive coefficient confirms the cost-of-caring hypothesis as originally conceived. The significant Female × Network Events interaction terms indicate that the harmful effects of network events are more detrimental for female compared to male study respondents. Models 3 through 6 address hypothesis 3, that masculine personality variables will mediate the effects of the interaction term. Models 3 through 6 introduce each gender orientation variable one at a time. As shown, each of these personality traits is significantly associated with psychological distress in the expected pattern. Feminine personality traits (emotional reliance and femininity) increase risk for depression, while assertion of autonomy and masculinity are negatively associated with depressive symptoms. In addition, model 6 reveals that masculinity mediates about one quarter of the magnitude of the Gender × Events to Others interaction term. The coefficient reported in model 2 is reduced 26 percent from 1.387 to 1.028. Thus, hypothesis 3 was partially supported. When all of the countervailing effects of each variable are simultaneously assessed in the final model (7), the interaction term is returned to marginal statistical significance.

Table 3.

CES-D Regressed on Demographic Controls, Life Events, and Personality Traits (N = 539).

Variable	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Female (= 1)	3.222***	3.183***	2.126**	2.884***	2.558**	2.286**	1.171
Age	−0.553***	−0.556***	−0.489***	−0.555***	−0.583***	−0.545***	−0.507***
Socioeconomic status	−0.534*	−0.580**	−0.387†	−0.637**	−0.514*	−0.405†	0.294
Currently married (= 1)	−1.288*	−1.142	−1.492	−0.899	−0.886	−1.403	−1.090
Parent (= 1)	−0.598	−0.838	−0.910	−0.760	−0.863	−0.629	−0.723
Number of friends	−0.326†	−0.329†	−0.366*	−0.306†	−0.325†	−0.272	−0.305†
Currently employed (= 1)	−0.725	−0.746	−0.822	−0.645	−0.637	−0.729	−0.682
Life events to self	1.353***	1.406***	1.362***	1.409***	1.417***	1.362***	1.346***
Network events	1.543***	0.712	0.762	0.590	0.604	0.941†	0.785
Female × Network Events		1.387*	1.516*	1.339†	1.392*	1.028	1.223†
Emotional reliance			0.742***				0.621***
Assertion of autonomy				−0.551***			−0.312*
Femininity					0.277**		0.150
Masculinity						−0.427***	−0.295***
Constant	24.032	24.715	13.267	29.755	20.007	37.183	24.054
R ²	.236	.242	.303	.268	.253	.285	.343

Note. Unstandardized regression coefficients. CES-D = Center for Epidemiologic Studies Depression Scale.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Table 4 addresses hypotheses 4 by assessing the possibility that female gender orientation moderates (amplifies) the positive association between network events and depressive symptoms for women and men. Within-gender analyses are presented separately for women (models 1 through 4) and men (models 5 through 8). Seven of the eight first-order terms presented in models 1 through 8 are statistically significant and replicate the pattern of findings reported above with the full sample. The coefficients in these models show that feminine personality traits increase risk for depressive symptoms, while masculinity is negatively related to depression. Only one of the four interaction terms presented in models 1 through 4 attains marginal statistical significance. The negative regression coefficient shown in model 4 indicates that the harmful effects of network events are lower among women who have high levels of masculinity (compared to women who are low in masculinity). In contrast to prediction, the Network Events × Feminine Orientation interaction terms were not significant. However, consistent with expectation, the positive and significant interaction coefficients observed in models 5 and 7 suggest that the magnitude of the association between network events and depressive symptoms were greater for men with higher levels of feminine orientations. Thus, hypothesis 4 was supported but only among men.

Table 4.

CES-D Regressed on Demographic Controls, Life-Events and Personality Traits, by Gender (Females n=309, Males n=230).

Variable	Women				Men
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Age	−0.503**	−0.670***	−0.682**	−0.740***	−0.449*	−0.442*	−0.463*	0.409*
Socioeconomic status	−0.522†	0.891**	−0.749*	−0.523†	−0.108	−0.201	−0.173	−0.077
Currently married (= 1)	−1.204*	0.258	−0.120	−0.223	−2.456	−1.992	−2.072	−2.237
Parent (= 1)	−0.875	−0.995	−1.392	−0.641	0.385	0.548	0.911	0.971
Number of friends	−0.038	0.067	0.028	0.159	−0.866***	−0.719**	−0.832***	−0.728**
Currently employed (= 1)	−0.538	−0.592	−0.747	−0.666	−0.848	−1.267	−1.438	−1.160
Life events to self	1.932***	2.132***	2.150***	2.120***	0.701*	0.628†	0.569	0.613†
Network events	2.251***	1.082***	1.944***	1.449**	1.688**	1.128*	1.335*	1.530**
Emotional reliance	0.970***				0.436**
Assertion of autonomy		−0.455**				−0.565**
Femininity			0.409**				−0.015
Masculinity				−0.457***				−0.322**
Events × Emotional Reliance	−0.044				0.436*
Events × Assertion of Autonomy		−0.024				0.188
Events × Femininity			−0.057				0.274+
Events × Masculinity				−0.136†				0.156
Constant	8.649	30.818	18.473	41.452	19.095	30.383	26.175	33.964
R ²	.373	.306	.311	.346	.244	.239	.219	.243

Note. Unstandardized regression coefficients. CES-D = Center for Epidemiologic Studies Depression Scale.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Tables 5 and 6 present parallel analyses to those shown in Tables 3 and 4. In Table 5, alcohol problems are regressed on the variables of interest. The statistically significant coefficient for network events in model 1 of Table 5 confirms that network events are positively associated with alcohol problems, net of sociodemographic factors and life events to self. The negative and significant Female × Events to Others interaction term presented in model 2 indicates that the magnitude of the relationship between network events and problem drinking is significantly greater for men compared to women. This finding is consistent with hypothesis 5. Succeeding models introduce each hypothesized mediating gender orientation variable individually (models 3 through 6) and collectively (model 7). However, no mediation effects were observed in these models. Therefore, hypothesis 6 was not supported.

Table 5.

Twelve-month Alcohol Problems Regressed on Demographic Controls, Life Events, and Personality Traits (N = 539).

Variable	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Female (= 1)	−.085***	−.083**	−.082**	−.085***	−.072**	−.083**	−.072**
Age	.000	−.001	.000	−.001	.000	−.001	.000
Socioeconomic status	−.006	−.004	−.005	−.005	−.006	−.004	−.006
Currently married (= 1)	−.038	−.044	−.043	−.043	−.049	−.044	−.048
Parent (= 1)	.002	.011	.011	.011	.012	.011	.012
Number of friends	.002	.002	.002	.002	.002	.002	.002
Currently employed (= 1)	.028	.029	.029	.029	.027	.029	.027
Life events to self	.056***	.054***	.054***	.054***	.053***	.054***	.053***
Network events	.021*	.052**	.052**	.052**	.054**	.052**	.054**
Female × Network Events		−.052*	−.052*	−.052*	−.052*	−.052*	−.052*
Emotional Reliance			−.001				.000
Assertion of autonomy				−.002			−.003
Femininity					−.005†		−.006†
Masculinity						.000	.000
Constant	.022	−.004	.007	.014	.085	.005	.109
R ²	.153	.163	.164	.164	.168	.164	.169

Note. Unstandardized regression coefficients.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Table 6.

Twelve-month Alcohol Problems Regressed on Demographic Controls, Life Events, and Personality Traits, by Gender (Females, n = 309; Males, n = 230).

	Women				Men
Variable	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Age	−.006	−.007	−.006	−.006	.007	.009	.008	.008
Socioeconomic status	−.005	−.005	−.007	−.006	−.006	−.009	−.007	−.006
Currently married (= 1)	−.007	−.008	−.013	−.010	−.149†	−.136†	−.161*	−.148*
Parent (= 1)	−.037	−.042	−.043	−.041	.170†	.149	.170†	.164†
Number of friends	−.006	−.006	−.005	−.005	.006	.007	.004	.005
Currently employed (= 1)	.035	.038	.034	.036	.031	.029	.030	.037
Life events to self	.029**	.028**	.028**	.027**	.082***	.086***	.082***	.083***
Network events	.003	.002	.009	.007	.041†	.040†	.046*	.048*
Emotional reliance	.001				−.002
Assertion of autonomy		.002				−.014†
Femininity			−.004				−.005
Masculinity				.001				−.004
Events × Emotional Reliance	.004				−.004
Events × Assertion of Autonomy		−.004				−.003
Events × Femininity			−.001				.005
Events × Masculinity				.001				−.002
Constant	.163	.162	.235	.149	−.181	−.147	−.150	−.139
R ²	.088	.087	.088	.082	.193	.240	.234	.231

Note. Unstandardized regression coefficients.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Table 6 assesses the extent to which feminine and masculine personality traits condition the relationship between network events and problem drinking. As before (Table 4), within-gender analyses are presented. Analyses are presented separately for women (models 1 through 4) and men (models 5 though 8). In models 1 though 4 of Table 6, life events to self were the only significant coefficients. Importantly, the coefficient for network events was not significant, indicating that for women, network events are not associated with problem drinking. Not surprisingly, none of the interaction terms presented in these equations were statistically significant. Models 5 through 8 present analyses for men. They show that the coefficients for network events range from being marginally significant (models 5 and 6) to being significant at the .05 alpha level (models 7 and 8). Only one of the four interaction terms presented here was marginally statistically significant, suggesting that the magnitude of the relationship between network events and problem drinking is weaker among men who possess high levels of assertion of autonomy. The findings presented here do not support hypothesis 7.

Discussion

There are emotional costs for caring about others. Prior research and day-to-day experiences make this clear. Less clear is the understanding of who is most adversely affected by the suffering of their loved ones and how this upset is expressed. The present paper has attempted to address these important issues. The findings presented here indicate that the cost of caring is contingent upon age and gender orientation and that men and women, for the most part, respond differently to this burden. Despite these differences, the analyses presented here make clear that network events elevate risk for psychological distress for both men and women. The implications of these findings are discussed below.

The distribution of network events, alcohol problems, and depressive symptoms within this sample was unambiguous. Preliminary descriptive analyses revealed that the youngest study participants were at greatest risk and that exposure to network events and reports of alcohol problems declined dramatically after age 30. Depressive symptoms were also highest among those age 25 or younger, but levels of symptoms remained relatively stable at older age categories, and declines in depression were not monotonic. Multivariate analyses confirmed an absence of associations between network events and psychological distress after age 30.

As mentioned, two theoretical perspective have been developed that predict alternative explanations for the relationship between age and the cost of caring. The first explanation, that network events are more harmful at younger ages, is based on evidence that stress exposure and the size and density of social networks are greater during this time in the life course. A second body of research has posited that middle adulthood is the time of life demarcated by the occupancy in social roles that obligate self to others. The findings presented here clearly support the former explanation over the latter.

In the analyses predicting depressive symptoms, masculine and feminine gender role orientations were found to mediate and moderate the relationship between network events and psychological distress. These analyses also reveal a distinct pattern of effects. Overall, they show that masculine orientations decrease risk for women, and feminine orientations increase risk for men.

The findings reveal that network events are damaging to the mental health of men and women, but the way that stress is translated into distress tends to vary by gender. The statistically significant interaction terms presented in Tables 3 and 5 confirm this. Thus, when assessed across separate outcomes, the cost of caring for others was shown to be positively associated with depressive symptoms among women and to increase risk of alcohol problems among men. An important caveat to this pattern of findings is revealed in models 5 through 8 of Table 4, which show that network events are also positively related to depressive symptoms for men.

The findings presented here yield important interpretations about the role of gender and gender orientation in raising and lowering risk for psychological distress. Importantly, they suggest that the advantages associated with masculine orientations are conferred to women and men. The corollary to this is that the mental health disadvantages that accompany feminine identities are harmful for both genders. These findings also confirm what is self evident in our day-to-day lives: Not all men are the same as one another, and not all women are the same. This is consistent with recent studies (Ferree 2010; Longest and Thoits 2012; Reczek and Umberson 2012) that have documented intragender variation across an array of social settings, behaviors, and health outcomes.

In addition, intrastatus variation in stress reactivity has been verified across a number of other social statuses. For example, Keith and Herring (1991) found that skin complexion was a powerful predictor of socioeconomic attainment within a national sample of African Americans. Geronimus’s (1996) weathering hypothesis and more recent studies on telomere length (Needham et al. 2012) have found circumstances where chronological age is only a rough estimate of aging. Importantly, both of these studies were conducted among young study populations. Finally, accurate assessments of the relative effects of socioeconomic status cannot be made in the absence of information on wealth (Conley 1999) and debt (Drentea and Reynolds 2012). These examples validate the need to carefully consider the multidimensionality of a given social status.

Although this study attempted to advance the understanding of the cost of caring, three study limitations merit discussion. First, the data that were analyzed for this report are somewhat dated. As noted earlier, data collection was conducted in 1990 and 1991. Thus, it is reasonable to assume that some period and/or cohort effects reside in the analyses presented here. This is so because women and men now enter the paid labor force, marry, and have children at relatively older ages compared to earlier cohorts (Arnett 2000; Gutman, Pullum-Piñón, and Pullum 2002; Stranger-Ross, Collins, and Stern 2005; Mathews and Hamilton 2009; U.S. Decennial Census 2010; Statistics Canada 2011a; 2011b). These cohort trends may also account for why network events were more harmful for the younger study participants in this sample. At the very least, these findings should be viewed with caution, because the issue of age and vulnerability to network events is far from being resolved. Therefore, analyses from more recently collected data could yield somewhat different results than those presented here.

Second, the analyses presented here say very little about the relative importance of gender roles and gender identities in the prediction of mental health. This is an important shortcoming that the present study was not able to adequately address. Finally, because this study employed a cross-sectional study design, it was not possible to establish the temporal ordering of the independent and dependent variables.

The limitations detailed above offer several avenues for future research. Ideally, this research would be founded on prospective data across a broad age spectrum. Future research should attempt to distinguish the relative salience of gendered roles and gendered identities. Doing so would provide a more nuanced understanding of what it is about gender that raises and lowers risk for psychological distress. Finally, it is crucial that these studies cast a wider net in terms of the outcomes considered as well as the psychosocial processes that make caring for others emotionally costly.

Footnotes

Acknowledgements

I thank David Russell, Joel Andress, and the anonymous reviewers for their comments and advice on this manuscript.

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by a research grant and a National Health Scientist Award from the National Research and Development Program of Health Canada, awarded to R. Jay Turner.

Notes

References

Almeida

David M.

Kessler

Ronald C.

1998. “Everyday Stressors and Gender Differences in Daily Distress.” Journal of Personality and Social Psychology 75:670-80.

American Psychiatric Association. 1987. Diagnostic and Statistical Manual of Mental Disorders, 3rd ed., rev. Washington, DC: Author.

Aneshensel

Carol S.

2002. “Answers and Questions in the Sociology of Mental Health.” Journal of Health and Social Behavior 43:236-46.

Aneshensel

Carol S.

Rutter

Carolyn M.

Lachenbruch

Peter M.

1991. “Social Structure, Stress, and Mental Health: Competing Conceptual and Analytic Models.” American Sociological Review 56:166-78.

Annandale

Ellen

Hunt

Kate

. 1990. “Masculinity, Femininity, and Sex: An Exploration of Their Relative Contribution to Explaining Gender Differences in Health.” Sociology of Health and Illness 12:24-46.

Arnett

J. J.

2000. “Emerging Adulthood: A theory of Development from the Late Teens through the Twenties.” American Psychologist, 55:469-80.

Avison

William R.

Turner

R. Jay

. 1988. “Stressful Life Events and Depressive Symptoms: Disaggregating the Effects of Acute Stressors and Chronic Strains.” Journal of Health and Social Behavior 29:253-64.

Barrett

Anne E.

White

Helene Raskin

. 2002. “Trajectories of Gender Role Orientations in Adolescence and Early Adulthood: A Prospective Study of the Mental Health Effects of Masculinity and Femininity.” Journal of Health and Social Behavior 43:451-68.

Belle

Deborah

. 1982. “The Stress of Caring: Women as Providers of Social Support.” Pp. 496-505 in Handbook of Stress: Theoretical and Clinical Aspects, edited by Goldberger

Breznitz

New York: Free Press.

10.

Bush

Diane Mitsch

Simmons

Roberta G.

1992. “Socialization over the Life Course.” Pp. 133-64 in Social Psychology: Sociological Perspectives, edited by Rosenberg

Turner

R. H.

New Brunswick, NJ: Transaction.

11.

Cartensen

Laura L.

Isaacowitz

Derek M.

Charles

Susan T.

1999. “Taking Time Seriously: A Theory of Socioemotional Selectivity.” American Psychologist 54:165-81.

12.

Chuick

Christopher D.

Greenfeld

Jay M.

Greenberg

Stefanie Teri

Shepard

Samuel J.

Cochran

Sam V.

Haley

James T.

2009. “A Qualitative Investigation of Depression in Men.” Psychology of Men and Masculinity 10:302-32.

13.

Conger

Rand D.

Lorenz

Frederick O.

Elder

Glen H.

Simons

Ronald L.

Xiaojia

. 1993. “Husband and Wife Differences in Response to Undesirable Life Events.” Journal of Health and Social Behavior 34:71-88.

14.

Conley

Dalton

. 1999. Being Black, Living in the Red. Berkeley: University of California Press.

15.

Connell

Robert W.

Messerschmidt

James W.

2005. “Hegemonic Masculinity: Rethinking the Concept.” Gender & Society 19:829-59.

16.

Dohrenwend

Barbara S.

1973. “Social Status and Stressful Life Events.” Journal of Personality and Social Psychology 9:203-13.

17.

Dohrenwend

Bruce P.

Dohrenwend

Barbara S.

1976. “Sex Differences in Psychiatric Disorders.” American Journal of Sociology 81:1447-54.

18.

Drentea

Patricia

Reynolds

John R.

2012. “Neither a Borrower nor a Lender Be: The Relative Importance of Debt and SES for Mental Health among Older Adults.” Journal of Aging and Health. 24:668-90.

19.

Elliott

Marta

. 2013. “Gender Differences in the Determinants of Distress, Alcohol Misuse, and Related Psychiatric Disorders.” Society and Mental Health 3(2):96-113.

20.

Ferraro

Kenneth F.

Wilkinson

Lindsay R.

2013. “Age, Aging, and Mental Health.” Pp. 183-203 in Handbook of the Sociology of Mental Health, edited by Aneshensel

C. S.

Phelan

J. C.

Bierman

New York: Springer.

21.

Ferree

Myra Marx

. 2010. “Filling the Glass: Gender Perspectives on Family.” Journal of Marriage and Family 72:420-39.

22.

Fleeson

William

. 2001. “Judgments of One’s Overall Contribution to the Welfare of Others: Life-course Trajectories and Predictors.” Pp. 75-96 in Caring and Doing for Others, edited by Rossi

A. S.

Chicago: University of Chicago Press.

23.

George

Linda

. 2014. “Taking Time Seriously: A Call to Action in Mental Health Research.” Journal of Health and Social Behavior 55:251-64.

24.

Geronimus

Arline T.

1996. “Black/White Differences in the Relationship of Maternal Age to Birthweight: A Population-based Test of the Weathering Hypothesis.” Social Science & Medicine 42:589-97.

25.

Gore

Susan

Aseltine

Robert H.

Colten

Mary Ellen.

1993. “Social Structure, Life Stress, and Depressive Symptoms in a High School Aged Population.” Journal of Health and Social Behavior 33:97-113.

26.

Gore

Susan

Mangione

Thomas

. 1983. “Social Roles, Sex Roles, and Psychological Distress: Additive and Interactive Models.” Journal of Health and Social Behavior 24:300-312.

27.

Gove

Walter R.

1972. “The Relationship between Sex Roles, Marital Status, and Mental Illness.” Social Forces 51:34-44.

28.

Gove

Walter R.

Tudor

Jeanette F.

1973. “Adult Sex Roles and Mental Illness.” American Journal of Sociology 78:50-73.

29.

Gutmann

M. P.

Pullum-Piñón

S. M.

Pullum

T. W.

2002. “Three Eras of Young Adulthood: Home Leaving in Twentieth-century America.” Journal of Social History 35:533-76.

30.

Haas

Steven A.

Schaefer

David R.

Kornienko

Olga

. 2010. “Health and the Structure of Adolescent Social Networks.” Journal of Health and Social Behavior 51:424-39.

31.

Hatch

Stephani

Dohrenwend

Bruce P.

2007. “Distribution of Traumatic and Other Stressful Life Events by Race/Ethnicity, Gender, SES, and Age: A Review of the Literature.” American Journal of Community Psychology 40:313-32.

32.

Hirschfeld

Robert M.

Klerman

Gerald L.

Chodoff

Paul

Korchin

Sheldon J.

Barrett

James

. 1976. “Dependency–Self-esteem–Clinical Depression.” Journal of the American Academy of Psychoanalysis 4:373-88.

33.

Hirschfeld

Robert M.

Klerman

Gerald L.

Gough

Harrison G.

Barrett

James

Korchin

Sheldon J.

Chodoff

Paul

. 1977. “A Measure of Interpersonal Dependency.” Journal of Personality Assessment 41:610-18.

34.

Hollingshead

August B.

1965. Two Factor Index of Social Position. New Haven, CT: Yale University.

35.

Horwitz

Allan V.

White

Helene Raskin

. 1987. “Gender Role Orientation and Styles of Pathology among Adolescents.” Journal of Health and Social Behavior 28:158-70.

36.

Horwitz

Allan V.

White

Helene Raskin

Howell-White

Sandra

. 1996. “The Use of Multiple Outcomes in Stress Research: A Case Study of Gender Differences in Responses to Marital Dissolution.” Journal of Health and Social Behavior 37:278-91.

37.

Keith

Verna M.

Herring

Cedric

. 1991. “Skin Tone and Stratification in the Black Community.” American Journal of Sociology 97:760-78.

38.

Kerr

William C.

Greenfield

Thomas K.

Bond

Jason

Rehm

Jürgen

. 2009. “Age-period-cohort Modeling of Alcohol Volume and Heavy Drinking Days on the U.S. Alcohol Surveys: Divergence in Young and Older Trends.” Addiction 104:27-37.

39.

Kessler

Ronald C.

1979. “Stress, Social Status, and Psychological Distress.” Journal of Health and Social Behavior 20:259-72.

40.

Kessler

Ronald C.

McLeod

Jane D.

1984. “Sex Differences in Vulnerability to Undesirable Life Events.” American Sociological Review 49:620-31.

41.

Kessler

Ronald C.

McLeod

Jane D.

Wethington

Elaine

. 1985. “The Cost of Caring: A Perspective on the Relationship between Sex and Psychological Distress.” Pp. 491-506 in Social Support: Theory, Research, and Applications, edited by Sarason

I. G.

Sarason

B. R.

Dordrecht, Netherlands: Martinus Nijhoff.

42.

Kish

Leslie

. [1965]1995. Survey Sampling. New York: Wiley.

43.

Lennon

Mary Clare

Rosenfield

Sarah

. 1992. “Women and Mental Health: The Interaction of Job and Family Conditions.” Journal of Health and Social Behavior 33:316-27.

44.

Longest

Kyle C.

Thoits

Peggy A.

2012. “Gender, the Stress Process, and Health: A Configurational Approach.” Society and Mental Health 2:187-206.

45.

Maciejewski

Paul L.

Prigerson

H. G.

Mazure

C. M.

2001. “Sex Differences in Event-related Risk for Major Depression.” Psychological Medicine 31:593-604.

46.

Mathews

T. J.

Hamilton

B. E.

2009. “Delayed Childbearing: More Women Are Having Their First Child Later in Life.” NCHS data brief (no. 21). Hyattsville, MD: National Center for Health Statistics.

47.

McFadden

S. H.

1996. “Religion, Spirituality, and Aging.” Pp. 349-65 in Handbook of the Psychology of Aging, edited by Birren

J. E.

Schaie

K. Warner

. San Diego: Academic Press.

48.

Mirowsky

John

Ross

Catherine E.

1992. “Age and Depression.” Journal of Health and Social Behavior 33:187-205.

49.

Mirowsky

John

Ross

Catherine E.

1996. “Fundamental Analysis in Research on Well-being: Distress and the Sense of Control.” Gerontologist 35:584-94.

50.

Needham

Belinda L.

Fernandez

Jose R.

Lin

Jue

Epel

Elissa S.

Blackburn

Elizabeth H.

2012. “Socioeconomic Status and Cell Aging in Children.” Social Science & Medicine 74:1948-51.

51.

Pudrovska

Tetyana

. 2010. “Why Is Cancer More Depressing for Men Than Women among Older White Adults?” Social Forces 89:535-58.

52.

Pugliesi

Karen

1995. “Work and Well-being: Gender Differences in the Psychological Consequences of Employment.” Journal of Health and Social Behavior 36:551-71.

53.

Radloff

Lenore S.

1977. “The CES-D Scale: A Self-report Depression Scale for Research in the General Population.” Applied Psychosocial Measurement 1:385-401.

54.

Radloff

Lenore S.

Rae

Donald S.

1981. “Components of the Sex Difference in Depression.” Pp. 77-110 in Research in Community and Mental Health, edited by Simmons

R. G.

Greenwich, CT: JAI Press.

55.

Reczek

Corinne

Umberson

Debra

. 2012. “Gender, Health Behavior, and Intimate Relationships: Lesbian, Gay, and Straight Contexts.” Social Science & Medicine 74:1783-90.

56.

Ridgeway

Cecilia L.

2011. Framed by Gender: How Gender Inequality Persists in the Modern World. New York: Oxford University Press.

57.

Robins

L. N.

Helzer

J. E.

Croughan

J. L.

Ratcliff

K. S.

1981. “National Institute of Mental Health Diagnostic Interview Schedule: Its History, Characteristics, and Validity.” Archives of General Psychiatry 38:381-89.

58.

Rosenberg

Morris

McCullough

B. Clair

. 1981. “Mattering: Inferred Significance and Mental Health Among Adolescents.” Research in Community and Mental Health 2:163-82.

59.

Rosenfield

Sarah

. 1992. “The Costs of Sharing: Wives’ Employment and Husbands’ Mental Health.” Journal of Health and Social Behavior 33:213-25.

60.

Rosenfield

Sarah

Lennon

Mary Clare

White

Helene Raskin

. 2005. “The Self and Mental Health: Self-salience and the Emergence of Internalizing and Externalizing Problems.” Journal of Health and Social Behavior 46:323-40.

61.

Rosenfield

Sarah

Mouzon

Dawne

. 2013. “Gender and Mental Health.” Pp. 277-96 in Handbook of the Sociology of Mental Health, edited by Aneshensel

C. S.

Phelan

J. C.

Bierman

New York: Springer.

62.

Ross

Catherine E.

Huber

Joan

. 1985. “Hardship and Depression.” Journal of Health and Social Behavior 26: 312-27.

63.

Ross

Catherine E.

Mirowsky

John

. 1989. “Explaining the Social Patterns of Depression: Control and Problem Solving—or Support and Talking.” Journal of Health and Social Behavior 30:206-19.

64.

Rossi

Alice S.

2001. “Domains and Dimensions of Social Responsibility: A Sociodemographic Profile.” Pp. 97-134 in Caring and Doing for Others, edited by Rossi

A. S.

Chicago: University of Chicago Press.

65.

Roxburgh

Susan

. 2004. “‘There Just Aren’t Enough Hours in the Day’: The Mental Health Consequences of Time Pressure.” Journal of Health and Social Behavior 45:115-31.

66.

Roxburgh

Susan

. 2005. “Parenting Strains, Distress, and Family Paid Labor.” Journal of Family Issues 26:1062-81.

67.

Schieman

Scott

. 2001. “Age, Education, and the Sense of Control: A Test of the Cumulative Advantage Hypothesis.” Research on Aging 23:153-78.

68.

Schieman

Scott

Turner

Heather A.

2001. “‘When Felling Other People’s Pain Hurts’: The Influence of Psychosocial Resources on the Association between Self-reported Empathy and Depressive Symptoms.” Social Psychology Quarterly 64:376-89.

69.

Schieman

Scott

Van Gundy

Karen

Taylor

John

. 2002. “The Relationship between Age and Depressive Symptoms: A Test of Competing Explanatory and Suppression Influences.” Journal of Aging and Health 14:260-85.

70.

Simon

Robin W.

1997. “The Meanings Individual Attach to Role Identities and Their Implications for Mental Health.” Journal of Health and Social Behavior 38:256-76.

71.

Simon

Robin W.

1998. “Assessing Sex Differences in Vulnerability among Employed Parents: The Importance of Marital Status.” Journal of Health and Social Behavior 39:37-53.

72.

Simon

Robin W.

2002. “Revisiting the Relationship among Gender, Marital Status, and Mental Health.” American Journal of Sociology 107:1065-96.

73.

Simon

Robin W.

2014. “Twenty Years of the Sociology of Mental Health: The Continued Significance of Gender and Marital Status for Emotional Well-being.” Pp. 21-52 in Sociology of Mental Health: Selected Topics from Forty Years, 1970s-2010s, edited by Johnson

R. J.

Turner

R. J.

Link

B. G.

New York: Springer.

74.

Spence

Janet T.

Helmreich

Robert L.

1978. Masculinity and Femininity: Their Psychological Dimensions, Correlates, and Antecedents. Austin: University of Texas Press.

75.

Statistics Canada. 2011a. Canadian vital statistics, marriage database and demography division (population estimates). Ottawa, Canada: Statistics Canada.

76.

Statistics Canada. 2011b. “Fertility: Overview, 2008.” Retrieved September 1, 2014 (http://www.statcan.gc.ca/pub/91-209-x/2011001/article/11513-eng.htm).

77.

Stranger-Ross

Collins

Stern

M. J.

2005. “Falling Far from the Tree: Transitions to Adulthood and the Social History of Twentieth-century America.” Social Science History 29:625-48.

78.

Thoits

Peggy A.

1982. “Life Stress, Social Support and Psychological Vulnerability: Epidemiologic Considerations.” Journal of Community Psychology 10:341-62.

79.

Thomeer

Mieke Beth

Umberson

Debra

Pudrovska

Tetyana

. 2013. “Marital Processes around Depression: A Gendered and Relational Perspective.” Society and Mental Health 3:151-69.

80.

Turner

Heather A.

1994. “Gender and Social Support: Taking the Bad with the Good?” Sex Roles 30:521-41.

81.

Turner

R. Jay

Avison

William R.

1989. “Gender and Depression: Assessing Exposure and Vulnerability to Life Events in a Chronically Strained Population.” Journal of Nervous and Mental Disease 177:443-55.

82.

Turner

R. Jay

Avison

William R.

1992. “Innovations in the Measurement of Life Stress: Crisis Theory and the Significance of Event Resolution.” Journal of Health and Social Behavior 33:36-50.

83.

Turner

R. Jay

Turner

J. Blake

. 1999. “Social Integration and Support.” Pp. 301-20 in Handbook of the Sociology of Mental Health, edited by Aneshensel

C. S.

Phelan

J. C.

New York: Kluwer Academic.

84.

Turner

R. Jay

Wheaton

Blair

Lloyd

Donald A.

1995. “The Epidemiology of Social Stress.” American Sociological Review 60:104-25.

85.

Uecker

Jeremy E.

2012. “Marriage and Mental Health among Young Adults.” Journal of Health and Social Behavior 53:67-83.

86.

Ueno

Koji

. 2010. “Same-sex Experience and Mental Health during the Transition between Adolescence and Young Adulthood.” Sociological Quarterly 51:484-510.

87.

Umberson

Debra

. 2003. Death of a Parent. Cambridge, UK: Cambridge University Press.

88.

Umberson

Debra

Chen

Meichu D.

House

James S.

Hopkins

Kristine

Slaten

Ellen

. 1996. “The Effect of Social Relationships on Psychological Well-being: Are Men and Women Really So Different? American Sociological Review 47:1-16.

89.

U.S. Decennial Census. 2010. American Community Survey. Washington, DC: U.S. Census Bureau.

90.

Van Gundy

Karen

. 2002. “Gender, the Assertion of Autonomy, and the Stress Process in Young Adulthood.” Social Psychology Quarterly 65:346-63.

91.

Wethington

Elaine

Cooper

Hope

Holmes

Carolyne S.

1997. “Turning Points in Midlife.” Pp. 215-31 in Stress and Adversity Over the Life Course, edited by Gotlib

I. H.

Wheaton

Cambridge, UK: Cambridge University Press.

92.

Wethington

Elaine

McLeod

Jane D.

Kessler

Ronald C.

1987. “The Importance of Life Events for Explaining Sex Differences in Psychological Distress.” Pp. 144-56 in Gender and Stress, edited by Barnett

R. C.

Biener

Baruch

G. K.

New York: Free Press.