Contextual changes in associations between marital status and mental health outcomes: An inter- and intra-cohort comparison


Amanda Deeley, University of Toronto

The nature of the relationship between marital statuses and attendant health outcomes experienced by women and men has long been debated by family and health scholars (Bernard, 1973; Gove and Tudor, 1973; Pearlin and Johnson, 1977; Thoits, 1986; Umberson et al., 1996). As divorce rates and common-law partnerships have increased since the 1980s, some scholarship has highlighted the protective effects of the legal institution of marriage over divorce and cohabitation (Waite and Gallagher, 2000). Others have argued that the effects of marital status and marital transitions on well-being vary little by gender (Ross, 1995; Booth and Amato, 1991) or for common-law unions versus legal marriage (Musick and Bumpass, 2012). Still more research has suggested that remaining in unhappy marriages undermines well-being in ways equivalent to divorcing or remaining unmarried (Williams, 2003; Amato and Hohmann-Marriott, 2007) and that the context of the marriage relationship itself influences the outcomes of those who subsequently exit marriage (Wheaton, 1990). My paper examines the strengths of the association between these various marital statuses and mental health outcomes among two cohorts of American men and women two decades apart: in the mid-1990s as rates of divorce and common-law unions were increasing, and a second cohort surveyed in the early 2010s. As the institution of marriage today is arguably viewed as less socially necessary or consequential than in previous generations (Weissman, 2013; Angus Reid Institute, 2018), my project asks: 1) What is the relationship between marital status and key mental health outcomes (depression, anxiety, and self-rated mental health) at these two time points? 2) (How) do these associations differ by birth and age cohort, as well as by gender, socioeconomic status, parental status, or racial/ethnic group? 3) Is there evidence of change in the magnitude of the protective or deleterious effects of different marital statuses on health across two distinct time points/generational cohorts? The literature suggests some evidence that suggests these factors have different effects by gender, racial/ethnic background, and have shifted over time. For example, Margolis and Choi (2020) point out that between 1991 to 2018 in the Canadian context, the economic gap between divorced men and women narrowed considerably. Recent assessments of stability and change in predictors of marital dissolution and common-law unions suggest similar trends in the American context (Kuo and Raley, 2016; Rosenfeld and Roesler, 2023). Because economic inequality has been argued to pose chronic strain for unpartnered/divorced individuals leading to poorer health outcomes than for married individuals, improvements in economic equality might suggest that health outcomes may be more positive for cohabitators, divorced and single/never-marrieds now than in the 1990s. However, there is also support for McLanahan's (2004) "diverging destinies" hypothesis, suggesting that Black Americans and Americans with lower educational attainment are increasingly more likely to be divorced or to never marry (Rosenfeld and Roesler, 2023). This suggests possible increasing exposure to differential health outcomes based on marital status across these groups. To assess my cohort change hypothesis, I pool and compare cross-sectional survey data from the National Survey of Midlife Development in the U.S. (MIDUS). The study was developed to investigate social factors and age-related variations in health and well-being in a representative national probability sample of adults in midlife. 3,487 eligible participants ("cohort 1") between the ages of 25 and 74 completed the initial survey in 1994. A comparable "refresher" cohort ("cohort 2") was asked the same survey questions in 2011. I assess inter- and intra-cohort change in the effects of marriage, common-law unions, divorce, separation, and singlehood, between those aged 25-42 at the time of the first wave, and two age sub-groups in cohort 2, those aged 25-42, as well as to those aged 42-60 in 2011 (i.e., the same age-period cohort as the cohort 1 group) by testing interactions between marital status and cohort. The dependent variables map associations between mental distress (depression, a key mental health indicator for women), anti-social behaviours (alcohol use, a key indicator of mental distress for men), and self-reported mental health. Preliminary analysis suggests that the negative association between separated/divorced status and mental health outcomes such as depression has diminished between these two time points, while cohabitation suggests a strengthened positive association with mental well-being in the refresher cohort, though more analysis of intra-cohort effects is underway.

This paper will be presented at the following session: