Marital status, gender, and depression: Analysis of the baseline survey of the Korean Longitudinal Study of Ageing (KLoSA)q Soong-Nang Jang a, Ichiro Kawachi a,*, Jiyeun Chang b, Kachung Boo b, Hyun-Gu Shin b, Hyejung Lee b, Sung-il Cho c a Harvard School of Public Health, Boston, MA, United States b Korea Labor Institute, Republic of Korea c Seoul National University, Republic of Korea a r t i c l e i n f o Article history: Available online 12 October 2009 Keywords: Depressive symptoms Marital status Korea Ageing Gender a b s t r a c t Marital status is a robust predictor of health outcomes in Western populations. However, data from Asian cultures remain sparse, and some studies suggest marked gender differences in the health benefits of marriage among Asian populations. We investigated the influence of marital status on depressive symptoms in older adult Koreans. Data were obtained from a sample of adults aged 45 to 85 years (4016 men, 5003 women) who participated in the 2006 cross-sectional baseline survey of the Korean Longi- tudinal Study of Ageing. Depressive symptoms were measured by the 10-item Center for Epidemiological Studies���Depression scale. A multiple regression model was used to examine the association between marital status and depressive symptoms, controlling for socioeconomic status, living arrangement, disability, and number of chronic diseases. In women aged 75 to 85 years, no significant differences were found between women who were married versus those who were widowed, divorced, or separated with regard to depressive symptoms. These findings were driven by increasing depressive symptoms among married women with age. Whereas divorced and widowed men in the sample reported higher rates of depressive symptoms than did married men, the difference between married vs. widowed/divorced women converged as they aged. This pattern of depressive symptoms by gender and life stage may reflect the distinctive influence of the Asian context on relations between men and women, such as traditional gender roles and patriarchal norms for older generation. �� 2009 Elsevier Ltd. All rights reserved. Introduction Numerous empirical studies have provided evidence for the protective effects of marriage on health that is, married individuals are more likely to be healthier than are widowed, divorced, sepa- rated, or never-married individuals (Manzoli, Villari, Pirone, & Boccia, 2007 Turner, Lloyd, & Roszell, 1999 Williams, 2004). Married individuals live longer than do unmarried individuals, and this protective effect cannot be entirely explained by the selection of healthy individuals into marriage (Lillard & Panis, 1996). The benefits of marriage also extend to factors other than survival married individuals reportedly experience lower rates of depres- sion (Afifi, Cox, & Enns, 2006 Holt-Lunstad, Birmingham, & Jones, 2008 Inaba et al., 2005 Weissman et al., 1996 Williams, 2004). The health benefits of marriage may be mediated by a variety of mechanisms, including access to economic resources (economies of scale in household consumption, financial security, as well as the so-called ������marriage premium������ by which married individuals command higher wages in the labor market) (Waite, 1995) the exchange of social support (House, Landis, & Umberson, 1988) and the beneficial influence of spouses on marital partners��� health behaviors (Eng, Kawachi, Fitzmaurice, & Rimm, 2005 Lee et al., 2005). According to the social support theory of marriage, the loss of one���s spouse is often associated with increased risk of depression due to loss of financial, emotional, physical, and instrumental support (Levenson, Carstensen, & Gottman, 1993 Lopata, 1978 Mineau, Smith, & Bean, 2002 Ross, Mirowsky, & Goldstein, 1990 Umberson, Wortman, & Kessler, 1992). In contrast, Nagata, Takat- suka, and Shimizu (2003) suggested that widowhood does not have q This work was supported by a Korea Research Foundation Grant (KRF -2007- 321-B00096) funded by the Korean Government (MOEHRD). Kawachi is supported in part by the MacArthur Network on SES and Health. * Corresponding author. Harvard School of Public Health, Department of Health and Social Behavior, 677 Huntingdon Avenue, Boston, MA 02115, United States. Tel.: ��1 617 432 0235 fax: ��1 617 432 3123. E-mail address: ikawachi@hsph.harvard.edu (I. Kawachi). Contents lists available at ScienceDirect Social Science & Medicine journal homepage: www.elsevier.com/locate/socscimed 0277-9536/$ ��� see front matter �� 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.socscimed.2009.09.007 Social Science & Medicine 69 (2009) 1608���1615

an important effect on the mental health of elders because the normative nature of losing one���s spouse at an advanced age might mitigate the consequences of widowhood on mental health (Nagata et al., 2003). Although many studies have reported similar mental health benefits of marriage among men and women (Arber, 2004 House et al., 1988 Manzoli et al., 2007 Seeman & Crimmins, 2001), several studies have reported gender differences with regard to the mental strain deriving from widowhood. For men, widowhood may lead to a weakening of family contacts, and older widowers might suffer more than older widows do from a lack of social support due to smaller networks of friends and more distant relationships with children and other family members (Davidson, 2000 Wenger, Davies, Shahtahmasebi, & Scott, 1996). Moen (1996) observed that the death of one���s spouse might have a more deleterious emotional impact on men, in that they derive more social and emotional support from marriage. The differential emotional impact of widowhood on men and women may be explained using a life course perspective: widowhood is an anticipated life event for women after a certain age, thus they are more likely than are older men to be prepared and to have a support group of peers and family members. For older women, widowhood may represent freedom from the care-giving role, as well as an opportunity to develop closeness with their children by living with them (Carr, 2004 Lee, DeMaris, Bavin, & Sullivan, 2001). In addition, several studies have emphasized the ability of the widowed to recover from grief and depression. Ferraro (1989) suggested that these factors provide some measure of compensation for widowed older women. Older widows seem to have a high degree of emotional resilience (Mas- tekaasa, 1994). The bulk of this evidence about the associations among marital status, health, and gender is based on Western populations. Fewer studies have been conducted in Asian cultures, where the patterns of associations have differed from the results obtained from Western societies. For example, Ikeda et al. (2007) examined the association between marital status and mortality in a 10-year prospective cohort of 94,062 middle-aged Japanese men and women and found that widowhood and divorce increased the risk of all-cause and cause-specific mortality by 1.5 to 2.0 fold in Japa- nese men, but no survival advantage was observed for married women (Ikeda et al., 2007). A study conducted in Beijing and Shanghai reported that the rate of depressive symptoms among men (50���70 years) without a spouse was 2.6 times higher than rates for married men, but marital status was not significantly associated with depression among women (Pan et al., 2008). These gender differences in the health benefit of marriage in Asian soci- eties may stem from the social and cultural environment, which includes strictly demarcated gender roles deriving from patriarchal and Confucian norms. As a result of industrialization and urbanization during recent decades, Korean families have become more similar to Westernized nuclear families with spouse-centered structures. However, the patriarchal ideology persists in Korean families, especially in older couples (Cha, 2007). In recent decades, increasing life expectancy in Asia has resulted in marriages of increased duration. The period of time following the departure of children has lengthened dramati- cally, as reflected in higher rates of married individuals living together without children in the home. In 2005, 55.4% of Korean aged 65 years and older were married, an increase of 3.4% over this rate in 2000 (Korean National Statistics, 2006). Moreover, the prevalence of losing a spouse declined by 4.1% in this population during this period. These increases in the number of married elders, including increased rates of re-marriage among older individuals (Korean National Statistics, 2006), have reduced the traditional dependence on children, leading to lives that are increasingly focused on the couple in later life. Behind these gradual trends, however, conflict remains with regard to family members��� roles and norms. Women���s labor market participation has grown rapidly in Korea, and consequently, the traditional child bearing and partic- ularly care-giving roles have been filled by older women in the family. In addition, the rapid rise in dual-earner couples has resulted in increased need for childcare arrangements and elder care support, as well as informal care support in Korea. Rapid structural change and traditional ideology can create conflict in older couples because wives may want to adopt modern flexible gender roles, whereas husbands might resist change and find it difficult to accept a new role later in life (Kim & Jun, 1997). Given these conditions, the traditional social support provided by marriage may not apply to Korean older women. Instead, these women might experience high levels of stress and depressive symptoms in later life due to marital conflict and the cumulative impact of gender role expectations. This, in turn,, may result in increased care-giving burdens for their husbands, children, and grandchildren. In the present study, we sought to examine the effect of marriage on the mental health of Korean men and women. We assessed the depressive symptoms among married, divorced/ separated, widowed Koreans across different age groups using a representative sample of middle-aged and older adults drawn from the Korean Longitudinal Study of Ageing (KLoSA). The Korean Longitudinal Study of Ageing (KLoSA), based on a nationally representative sample of Koreans aged 45 years or older, was launched in 2006. Funded by the Ministry of Labor, KLoSA is the first national survey of Koreans that measures depressive symp- toms and is available to the public. KLoSA was also designed to compile information on various aspects of ageing for use as basic statistical data in interdisciplinary research on the social, economic, physical and psychological aspects of ageing. We hypothesized that marital status would have different influences on depressive symptoms depending on gender and age group. We assessed age in 10-year categories to reflect stages in an individual���s life course. For example, men and women between the ages of 45 and 54 are still tied to parental, social, and occupational roles. Men and women between the ages of 55 and 64 often go through the processes of retirement and subsequent reconnection with their spouses. Men and women between 65 and 74 years of age are adapting to new roles as grandparents or caregivers for sick spouses or grand- children. Men and women aged 75 years or older are more likely to experience spousal death, health problems, and diminished levels of functioning. In addition, age cohorts separated by 10-year intervals have experienced different social environments during youth and adulthood because Korea has experienced rapid economic development and modernization during the past 50 years. Individuals currently aged 75 years or older have been immersed since childhood in Confucian attitudes about strictly demarcated gender roles. Thus, married older Korean women are burdened with the role of family caregiver for sick spouses and young grandchildren, substituting for working daughters or daughters-in-law with regard to the latter. The burdens of these roles might lead to depressive symptoms among the oldest Korean married women. Methods Study design and data The Korean Longitudinal Study of Ageing (KLoSA) focused on Koreans aged 45 and older living in households selected by multistage stratified probability sampling (based on geographical area) to be representative of the nation. A total of 10,254 individuals S.-N. Jang et al. / Social Science & Medicine 69 (2009) 1608���1615 1609