Marital Quality and Depressive Symptoms Among Older Hispanic Adults in the United States

Abstract

Objectives

Despite the cultural importance of marriage as a social support system and its well-established link to mental health, older Hispanic adult populations, which are the largest racial and ethnic minoritized groups, remain understudied. The current study examined how positive and negative dimensions of marital quality are associated with depressive symptoms.

Methods

Data from Hispanic adults aged 51 years and older (n = 1,012) were obtained from the 2016 and 2018 Health and Retirement Study waves. The Center for Epidemiological Studies—Depression scale (0–8 symptoms) was modeled as a function of positive and negative marital quality measures (1–4), as well as the relevant covariates.

Results

Results from a negative binomial regression model showed that a 1-unit change in positive and negative marital quality was associated with a 23.61% reduction and a 23.74% increase, respectively, in depressive symptoms. The interaction terms with marital quality and gender, as well as marital quality and religion, were not statistically significant.

Discussion

In the United States, a large percentage of older Hispanic adults are immigrants, and their extended family tends to reside in their countries of origin. As such, older Hispanic adults may have smaller social networks, and marital quality most likely represents a culturally important social support network in later life. Significant associations between depressive symptoms and marital quality among older Hispanic adults should receive more attention in family and public health policy discussions, particularly given the increasing diversity in U.S. society.

Being married is positively associated with mental health and negatively associated with mental health issues such as depressive symptoms (Umberson et al., 2013). In the last few decades, only a handful of studies examined marital quality or the link between marriage and mental health among Hispanic populations in the United States (Bulanda & Brown, 2007; Muruthi et al., 2020; Perez & Cruess, 2014). Most previous studies focus on the White adult populations and do not disaggregate racial and ethnic minoritized groups. On average, Hispanic individuals have lower socioeconomic status (SES) and health insurance coverage (18% uninsured) than the general population (9% uninsured), suggesting sources of structural disadvantage and stress that may negatively affect marital quality and mental health (Branch & Conway, 2021). On the other hand, older Hispanic adults tend to have better health outcomes than the general population despite their lower socioeconomic resources (Hernandez et al., 2022).

Compared to the general older adult population, a few unique characteristics of the older Hispanic populations in the United States include a higher proportion of immigrants, greater socioeconomic disadvantages, poorer mental healthcare access, and higher cultural importance of family relationships (Williams et al., 2009). Yet, Hispanic adults tend to have better mental health outcomes than the general adult population (e.g., 28% vs 37% lifetime depression rate, respectively; Williams et al., 2009). Considering that in older age individuals often have a more limited social network, the quality of marriage can significantly affect the mental health of older Hispanic adults. However, population-level empirical evidence is still yet to be documented. The present study analyzes recent nationally representative data to examine the association between marital quality and depressive symptoms among older Hispanic adults residing in the United States.

Marital Quality and Mental Health Outcomes

Fowers and Owenz (2010) suggest viewing marital quality not as the sole focus of individuals’ lives but rather as one dyadic social relationship among many that contributes to overall life satisfaction. Marital quality is best seen as a multidimensional construct and subjective evaluation of the marital relationship (Fincham & Linfield, 1997) in which both positive and negative dimensions of marital quality are critical components (Fincham & Bradbury, 1987). For older adult populations, marriage is particularly important to health due to the decreasing size of social support networks in later life (Carr et al., 2014; Wrzus et al., 2013). However, simple measures of marital status, such as being married or not being married, do not capture the important nuances of the relationship.

The theoretical pathways between marriage and mental health are applicable to the understanding the role of marital quality among older Hispanic adults. Williams et al. (2009) illustrate three models, including the marital resource model, the marital crisis model, and the selection bias model. The marital resource model suggests that marriage provides social and instrumental support, which, in turn, enhances health outcomes. The marital crisis model suggests that marital quality and relevant conflicts are a source of stress, negatively affecting mental health. Finally, the selection bias model shows that individuals with better mental health and resources to maintain higher-quality marriages are more likely to get married compared to their counterparts. These theoretical propositions emphasize that marriage can be both a beneficial resource and a source of stress. Thus, both positive and negative marital quality, as well as other relevant factors (e.g., demographic and socioeconomic characteristics; see below for an expanded discussion), should be taken into account to examine marital quality and mental health.

Empirical studies have found support for these theoretical perspectives. Wong and Waite (2015) analyzed data from adults aged 57 years and older from the National Social Life, Health, and Aging Project, and reported that a high-quality marriage is protective of mental health. Peek and Markides (2003) examined the data of noninstitutionalized Mexican Americans aged 65 and older from the Hispanic Established Population for the Epidemiologic Study of the Elderly data set and found that older married people are more likely to have lower rates of depression and anxiety than their unmarried counterparts. Moreover, Bulanda et al. (2021) explored the data of married older adults in the Health and Retirement Study (HRS) and suggested that marriage may promote mental health by buffering the adverse effects of psychological stress. Also, based on an analysis of HRS data, including married older adults aged over 50 years old, Choi et al. (2016) found that perceptions of one’s partner being engaged in their relationships and providing social support can promote health in middle and older age. However, the meta-analysis of 128 systematically selected studies showed that a poor-quality marriage may become a source of stress and, in turn, result in less optimal mental health (Robles, 2014). Finally, a study with the HRS data of older married adults aged over 50 years old showed that poor-quality marriage may negate mental health benefits from other social support (e.g., friends; Han et al., 2019).

Recent studies further disaggregate the associations between marriage, marital quality and mental health outcomes. Despite differential measurements—for example, relationship quality, decision-making power balance among couples, spousal support, and spousal strain—marital quality is positively linked with depressive symptoms through loneliness (Marini et al., 2020; Saenz, 2021). The consistent findings about the subpopulations of Mexican adults and Hispanic adults in the United States are extended to older Hispanic adults. Muruthi et al. (2020) showed that positive spousal support was associated with fewer depressive symptoms among older Hispanic adults in the United States. Saenz (2021) and Saenz and Rote (2019) found the power balance between married couples and spousal support was related to lower loneliness and fewer depressive symptoms, whereas spousal strain increases the loneliness and depressive symptoms among older Mexican adults.

In the context of marital quality and mental health, another important finding is the role of health and cognitive conditions. Bookwala and Franks (2005) showed that marital disagreement was associated with greater impacts of physical disability on depressive symptoms among the general adult population. Hsu et al. (2023) reported that positive marital quality was associated with greater loneliness among couples with high marital quality. Warner and Adams (2016) pointed out that the associations between relationship quality, loneliness, and physical health may be different by gender among older adults, due to the differing susceptibility. Although these empirical findings support the marital resource and marital crisis models (Williams et al., 2009), and apply to the Hispanic adult populations, more studies are necessary to document the associations between multiple dimensions of marital quality and mental health among the unique subpopulation of older Hispanic adults in the United States.

Older Hispanic Adults, Marital Quality, and Mental Health

The large subpopulation of Hispanic adults in the United States will necessarily have significant impacts on the nation’s public health (Hummer & Hayward, 2015). As of 2020, Hispanics account for approximately 19.1% of the total U.S. population and form the largest racial and ethnic minoritized population (U.S. Census Bureau, 2021). More than half (51%) of older Hispanic adults living in the United States are married (Federal Interagency Forum on Aging-Related Statistics, 2020). Approximately 14% of the U.S. population are immigrants, but 32% of the Hispanic population are immigrants (Ward & Batalova, 2023). Hispanic immigrants may face a greater risk of mental health issues for two competing reasons. More acculturation to the mainstream culture often results in isolation from their traditional cultural background or social support network, whereas the lack of acculturation results in isolation from the mainstream culture (Thiede et al., 2017a). Both acculturation and the lack thereof may negatively affect immigrants’ mental health. Also, greater acculturation to the U.S. mainstream culture has been associated with marital disruption among Latino couples (Trail et al., 2012).

SES, which is known as one of the social determinants of health (Braveman et al., 2011), is generally lower for Hispanic adults than for other racial and ethnic groups in the United States. Also, compared to their White counterparts, Hispanic couples are more likely to have fewer earners in the marriage and therefore face a lower household income (Thiede et al., 2017b). Hispanic adults tend to have lower levels of postsecondary educational attainment in the form of a bachelor’s degree than their White counterparts (21% vs 42%, respectively; U.S. Census Bureau, 2022).

Relatedly, the prevalence of chronic diseases such as diabetes is higher among older Hispanic populations. Although the rates of depression may be relatively lower, only about 22% of Hispanic adults with mental illness receive mental healthcare services, compared to their White counterparts (American Psychiatric Association, 2017). Suboptimal mental healthcare access for Hispanic communities is an urgent public health issue in the United States (Bucay-Harari et al., 2020), and prevention of mental illness is critical for Hispanic adults in later life. Moreover, Pabayo et al. (2022) noted that the lower rates of depression among Hispanic adults may reflect undiagnosed cases from their underutilization of healthcare services. Whereas it has been recommended that services be directed toward expanding social support networks, emphasizing on leveraging the inner strengths of older Hispanics adults to enable them to utilize healthcare services more frequently (Curtin et al., 2019).

Despite SES and healthcare access disadvantages, older Hispanic adults tend to have better overall health outcomes and lower mortality rates. This Hispanic paradox (Markides & Eschbach, 2011) can be partially explained by culturally valued social support networks (e.g., ethnic enclave advantage), along with the healthy migrant effect and return migrant effect (e.g., salmon bias; Hernandez et al., 2022; Valles, 2016). Familismo, which holds a significant place in Hispanic culture, may also help explain this phenomenon and the lower reports of depressive symptoms among Hispanics (Park et al., 2014; Valdivieso-Mora et al., 2016). Hispanic culture places emphasis on familismo (Davila et al., 2011), and the importance of kin relationships. The concept of familismo views marriage as an extension of what is considered the most important social network—family—over other forms of social relationships, such as friends and distant kin (Orengo-Aguayo, 2015).

In addition, the concept of machismo, which illustrates men’s domination, control, and unavailability in a relationship, may result in systematic gender differences in mental health and marital quality among Hispanic couples (Pardo et al., 2013; Valencia, 2021). Indeed, older women tend to report lower levels of marital satisfaction and interaction than older men (Amato & James, 2018; Bulanda, 2011). Relatedly, depressive symptoms may be transmitted within married couples due to relationship distress (Helms et al., 2014; Kouros & Cummings, 2010). Thus, despite familismo, within-marriage gender differences may result in a greater risk of stress and anxiety for women compared to men among Hispanic married couples (Darghouth et al., 2015; Rosenfield & Mouzon, 2013).

The availability of social support networks, which tend to shrink in later life (Wrzus et al., 2013), is also important to consider. Given immigration patterns and histories, Hispanic adults may still have relatives in their countries of origin and have smaller kin networks in the United States. Additionally, older immigrant adults are more likely to face limited English language skills and less acculturation to the mainstream culture in the United States (Juckett, 2013). These factors may restrict older Hispanic adults from expanding their social network outside of their language and cultural groups. Yet, within-Hispanic population differences (e.g., Mexicans vs others) should not be overlooked given differing migration histories. Additionally, according to socioemotional selectivity theory (Carstensen, 2021), adults become more selective in their social networks and activities that are emotionally gratifying as they reach older ages (Wong & Waite, 2015). Thus, their reliance on an immediate social relationships—such as marriage—may increase in later life. As such, marital quality is likely to become incrementally more important to older Hispanic adults from sociocultural and public health standpoints.

At the same time, nonkin social support networks through religion are also important for Hispanic communities. According to a study by Ellison et al. (2013), Catholicism is the most prevalent faith tradition among Hispanic individuals in the United States. Additionally, Orengo-Aguayo (2015) suggests that participating in religious activities is linked to higher levels of marital satisfaction among Hispanic adults. This religious and cultural practice is also viewed as a mental health-promoting mechanism for Hispanic populations in the United States (Caplan, 2019). For example, religion holds deep cultural significance in the Hispanic population, although there seem to be variations across Hispanic subpopulations, arguably due to the differences in the migration histories in the United States (Orengo-Aguayo, 2015).

Other Relevant Factors

Marital quality and depressive symptoms are related to other factors, such as individual and social characteristics. Studies have shown that as individuals age, their relationship quality tends to decrease (Brown et al., 2022). According to socioemotional selectivity theory (Carstensen, 2021), older age is an indicator of increasing focus on a narrower social network and life events, such as retirement and changes in social relationships due to declining health, which may also be linked with mental health. In turn, age is not only associated with relationship quality but also psychological well-being (Pruchno et al., 2009). Also, having living children is correlated with lower marital satisfaction (Brown & Kawamura, 2010). Child care responsibility tends to constrain parents’ time for healthful behaviors, including physically active leisure activities and sleep, and, in turn, negatively affect marital relationship quality and health (Pepin et al., 2018). Moreover, ethnicity is a cultural factor linked to marital quality and mental health. Race and ethnicity are often tied to SES and health behaviors, which collectively affect marital quality and health outcomes (Umberson et al., 2010).

SES, including education, income, assets, and health insurance, is associated with marital quality (Choi & Marks, 2013). Married couples with lower SES tend to have lower marital quality and poorer mental health because of relatively greater life demands (e.g., caregiving responsibilities; financial stress coping skills; healthcare access) than those with greater financial and health resources (Birditt et al., 2012; Choi & Marks, 2013). Lastly, physical health may also play an important role in the relationship between marital quality and depressive symptoms. Bulanda et al. (2016) reported that low-quality marital interactions are associated with poorer health outcomes among older women. Liu et al. (2021) report that positive and negative marital quality measures are related to cognitive decline at a slower rate and a faster rate over the course of 10 years, respectively, among older men but not among older women. Poor marital quality may lead to poorer health behavior, such as smoking and excessive drinking, and harm health and physical and cognitive functions (Umberson et al., 2010).

Research Questions and Hypotheses

Taken together, studies suggest marital quality is linked to mental health above and beyond just marital status in general and that the relationship between marital quality and mental health may differ for Hispanics (Caplan, 2019; Umberson & Thomeer, 2020; K. Williams et al., 2009). Socioemotional selectivity theory (Carstensen, 2021) suggests the importance of marital relationships becomes even greater in later life. The unique characteristics of the Hispanic population, such as a high percentage of immigrants, limited English language skills, relatively lower SES, mental healthcare access disadvantages, as well as familismo, suggest greater reliance on their marriage and more substantial impacts on later-life mental health outcomes. However, there is a paucity of research on marital quality and mental health outcomes among older Hispanic adults in the United States (Carr et al., 2014). Thus, a population-level examination of marital quality and mental health among older Hispanic adults will inform the nation’s public health interests, aiding in better understanding mental health disparities by race and ethnicity through social relationships and cultural differences.

This paper explores the following research questions: How are positive and negative dimensions of marital quality associated with one of the major mental health outcomes—depressive symptoms—among older Hispanic adults living in the United States? Also, are gender and religion associated with depressive symptoms among Hispanic older married adults in the United States? Finally, are the relationships between marital quality and depressive symptoms different by gender and religion? Given the marital resource and marital crisis models (Williams et al., 2009), we hypothesized that (H1) higher positive marital quality is associated with fewer depressive symptoms, and (H2) higher negative marital quality is associated with greater depressive symptoms. We also hypothesized that the relationships between marital quality and depressive symptoms are moderated by gender (H3) and religious affiliation (H4) due to machismo and cultural significance of religion. Specifically, the relationships between depressive symptoms and marital quality are expected to be stronger among women than men, as well as those who reported religion than those who did not.

Method

Data

We obtained the data from the 2016 and 2018 waves of the HRS as well as the RAND HRS, which provide harmonized as well as derived measures across available waves (RAND Center for the Study of Aging, 2022). HRS is an ongoing study to collect nationally representative data from U.S. adults aged 51 years and older every 2 years. In the present study, we focused on married Hispanic adults who responded to the HRS Psychosocial Leave-Behind Questionnaire (LBQ). Given that the LBQ portion of the HRS is administered to a randomly selected half of the total samples every other way. Thus, we pooed the LBQ 50% subsamples from two consecutive HRS data collection waves (2016 and 2018).

Among all Hispanic married samples (n = 1,058), the missing data ranged from one case to up to 17% (n = 179 for marital duration). When accounting for all variables of interest, the missing rate was approximately 25% (n = 268). Of those, the normalized sampling weight variable had 46 missing cases. We decided not to impute these sampling weights, which are constructed based on the HRS complex sampling design (Ofstedal et al., 2011). The patterns of missing values are examined based on the cross-tabulations of all variables of interest. Based on the examination of patterns, the missing at random (MAR) or missing conditional to the variables of interest in the present study is assumed (Enders, 2022). Although MAR cannot be decisively ruled, modern missing data handling, such as multiple imputations, is recommended over a listwise deletion (B. Muthén et al., 2016). After excluding the missing sampling weights, the total analytic sample is 1,012 in 100 imputed data sets without any missing values (see below for the imputation process).

Measures

Dependent variable

Depressive symptoms is assessed with the HRS version of the Center for Epidemiological Studies—Depression (CES-D) scale (Steffick, 2000), which consists of responses to eight questions: whether much of the time over the past week they felt (1) depressed, (2) lonely, (3) sad, (4) everything was an effort, (5) had restless sleep, (6) could not get going, (7) enjoyed life, and (8) were happy. Responses to (7) and (8) were reverse-coded to make them consistent with other CES-D items. The final score ranges from 0 to 8 depressive symptoms.

Independent variables

Positive marital quality is assessed with three survey items about one’s spouse: (1) do they really understand the way you feel about things?; (2) can you rely on them if you have a serious problem?; and (3) can you open up to them if you need to talk about your worries? Negative marital quality is assessed with four survey items, including (1) how often do they [spouse] make too many demands on you?; (2) how much do they criticize you?; (3) how much do they let you down when you are counting on them?; and (4) how much do they get on your nerves? Response categories for these survey items range from 1 (not at all) to 4 (a lot). Responses to these three positives and four negative marital quality items, respectively, were averaged to produce the final positive and negative marital quality scales (1–4).

Covariates

Demographic characteristics include age in years; gender (women vs men); U.S. born (vs non-U.S. born); Mexican (vs non-Mexican Hispanics); and number of children. The characteristics of marriage include marital duration in years; and remarriage (vs first marriage). Socioeconomic characteristics include the total years of formal education, employment status (employed vs not employed), logged household income, and logged household assets. Health characteristics include the number of activities of daily living (ADL) limitations (0–5: bathing, dressing, eating, getting in/out of bed, walking across a room) and instrumental activities of daily living limitations (0–5: using the phone, managing money, taking medications, shopping for groceries, and preparing hot meals); the number of chronic conditions (0–6: high blood pressure, diabetes, cancer, lung disease, heart disease, and stroke) and self-rated health (1–5: poor–excellent). Finally, religion was dichotomized based on whether a respondent reported any religious affiliation (vs no religious affiliation).

Analytical Approach

A survey-weighted descriptive summary was computed. The 2016 and 2018 HRS sampling weights were combined and normalized based on each wave’s corresponding population sizes. The combined sampling, cluster, and stratification weights were applied in all analyses. Given the relatively high rate of missing values and a mixture of both continuous and categorical variables of interest (Jakobsen et al., 2017), Bayesian multiple imputation techniques with the Markov Chain Monte Carlo simulation, and the mean- and variance-adjusted weighted least square estimation in Mplus version 8 (B. Muthén et al., 2016) were employed. The predictor variables were treated as latent continuous response variables in the imputation process. Simply put, each variable with missing values was repeatedly regressed on all other variables to construct the distribution of possible values, and plausible values were randomly selected from the constructed distribution to impute the missing data 100 times. Examination of the trace plot and potential scale reduction factor (PSR = 1.08; about 1 indicated good fit) concluded that the missing data imputation was converged or carried out without major issues. Also, alternative multiple imputation methods with the observed predictor variables were evaluated but the findings including the estimated coefficients and statistical significance were consistent with the latent continuous response variable approach.

To address the research questions, negative binomial (e.g., Poisson-gamma) regression was used to model the count outcome variable—the CES-D score—as the function of the positive and negative marital quality measures and the covariates (Allison, 2012). The negative binomial regression is determined to be the most appropriate model due to the violation of the equidispersion assumption and normal distribution assumption in the alternative models such as Poisson regression and linear regression in our preliminary analysis (DeMaris, 2004).

Let λ_i be the mean count of depressive symptoms for person i, and β₀, β₁, β₂, and β_k be the intercept and coefficients for positive marital quality, negative marital quality, and kth (k > 6) covariates, which includes gender and religions as parts of the research questions, respectively. Additionally, β₃, β₄, β₅, and β₆ are the coefficients of the interaction terms specified to address the research questions. The dispersion parameter and residual term are expressed as σ and ε, respectively (Allison, 2012). The observation/exposure period is assumed to be equal for all respondents. Thus, λ_i can be considered the mean count rather than an incidence rate over varying exposure time. The negative binomial regression analysis with a set of 100 imputed data is conducted as follows. First, one of the imputed data sets is used to explore baseline models. Second, the negative binomial regression was estimated using each of the 100 imputed data sets. Finally, the median is found from the distributions of the estimated regression coefficients. The standard error is also computed from the results of 100 negative binomial regression models and used for significance tests. The statistical significance is evaluated at the Type I error rate of 0.05. The estimated coefficient is exponentiated and interpreted as a factor change (DeMaris, 2004).

Sensitivity analyses with the models with different specifications (e.g., without one of the covariates) showed consistent findings in our final models. We also examined the interaction effects of relevant variables, including positive marital quality, negative marital quality, gender, and religion, which are theoretically relevant to the current study, as a part of the sensitivity analysis. Multicollinearity was checked using the variation inflation factor (between 1.30 and 2.91, <4.0) and is found not to be an issue, except for the interaction terms (DeMaris, 2004). Finally, the Monte Carlo simulation-based power analysis of the negative binomial regression model showed that the statistical power (>0.93) for the key variables of interest met the conventionally accepted level of 0.80 (L. K. Muthén & Muthén, 2002).

Results

The survey-weighted descriptive summary is presented in Table 1. The average CES-D score was 1.8 symptoms on an 8-point scale. Average positive and negative marital quality were 3.4 and 2.0, respectively, on a 4-point scale. There were slightly more men (53%) than women (47%) and more immigrants than those U.S. born (43%). More than half of the respondents (60%) identified as Mexican. The majority (87%) reported a religious affiliation (vs none). Average self-rated health was 2.8 on a 5-point scale.

Results from the survey-weighted negative binomial regression model are presented in Table 2. Results showed that both dimensions of marital quality were significantly related to depressive symptomatology among older Hispanic individuals, controlling for the other variables in the model. A one-unit increase in positive marital quality was associated with 23.51% fewer depressive symptoms (b = −0.27, p < .05). A one-unit increase in negative marital quality, on the other hand, was associated with a 23.74% increase in depressive symptoms (b = 0.21, p < .05). Therefore, we find support for both our first and second hypotheses. While gender was not associated with depressive symptoms among older Hispanic adults, those who reported any religious affiliation had 40.55% fewer depressive symptoms (b = −0.52, p < .05), on average, than their counterparts without any religious affiliation. Although the third hypothesis was not supported, the fourth hypothesis was supported.

The positive and negative marital quality interaction terms showed the estimated coefficients of 0.097 (p = .562) and −0.148 (p = .332). Also, the marital quality and religion interaction terms showed the estimated coefficients of −0.125 (p = .612) and −0.007 (p = .966). The interaction terms including marital quality and gender, as well as marital quality and religion were not statistically significant and, in turn, were not supportive of the hypotheses. Other significant findings are worth noting. Older Hispanic adults who reported higher SES, including employment, income, and assets, were significantly associated with lower levels of depressive symptoms. Finally, two indicators of health were significantly related to depressive symptoms. A one-unit increase in the number of ADL limitations was associated with a 7.25% increase in depressive symptoms (b = 0.07, p < .05), and a one-unit increase in self-rated health was related to 26.14% fewer depressive symptoms (b = −0.30, p < .05).

Discussion

We examined the relationships between two dimensions of marital quality and depressive symptoms among older married Hispanic adults in the United States. Our results showed that positive marital quality was associated with fewer depressive symptoms. This finding is consistent with relevant previous studies of both the general population and Hispanic adults in the United States (Bulanda et al., 2016; Muruthi et al., 2020; Peek & Markides, 2003; Wong & Waite, 2015). Consistent with the marital resource model, emotional and instrumental support from spouses might protect and enhance mental health outcomes. The relationship between positive marital quality and fewer depressive symptoms among older Hispanic adults supports not only the notion of familismo but also possible spillover effects, such as enhanced subjective well-being or reduced loneliness, which might protect mental health (Carr et al., 2014; Marini et al., 2020). Socioemotional selectivity theory (Carstensen, 2021) suggests that positive marital quality may benefit older adults, as they narrow their social networks to concentrate on their most central and meaningful relationships. This increased attention to the marital relationship might enhance the benefits of positive relationships and emotionally gratifying activities with spouses in later life, and familismo, which already preferences the marital relationship, may further elevate these benefits for older Hispanic adults.

Our results showed that negative marital quality was associated with greater depressive symptoms among older married Hispanic individuals. Consistent with the marital crisis model (Williams et al., 2009), poor marital quality might have caused relationship conflicts, stress, and other negative emotions, which could result in greater depressive symptoms. Consistent with a previous study of Mexican adults, negative marital quality such as spousal strain might have increased loneliness and led to depressive symptoms (Saenz & Rote, 2019). In addition, lower marital quality might have offset mental health benefits from other social networks such as friends and relatives, as was the case among older married couples in one study of the general adult population (Han et al., 2019). In other words, negative marital quality might represent a source of chronic stress that negates or diminishes social support benefits. Negative marital quality might be seen as a failure to “do family” (Connidis, 2015) consistent with familismo among older Hispanic adults and, therefore, result in even greater levels of stress for older Hispanic adults. From the standpoint of socioemotional selectivity theory (Carstensen, 2021), negative marital quality might indicate a lack of a meaningful central relationship and arguably contribute to loneliness in later life (Ong et al., 2015; Wang et al., 2018). Our results also point to the importance of not simply examining marital status when examining variation in mental health, which assumes homogeneity for all married individuals. Our results suggest variations in older Hispanic adults’ mental health by marital quality, with married Hispanic older adults in lower-quality marriages reporting more depressive symptoms.

Some other results are worth noting. Interestingly, gender was not significantly associated with depressive symptoms among older Hispanic adults, despite the notion of machismo among Hispanic adults aged between 18 and 64 years old (Valencia, 2021). Instead, older Hispanic men and women reported similar levels of depressive symptoms. This is consistent with one other existing study of older Hispanic American adults, which found no gender differences in depressive symptoms when controlling for other sociodemographic, economic, and health variables, and also no gender differences in the relationship between positive marital support and depressive symptoms (Muruthi et al., 2020). However, there do appear to be gender differences in marital quality among older Hispanic men and women; the bivariate comparison in our follow-up analyses showed significant differences both in positive and negative marital quality between women and men. Future research should further examine the potential moderating role of gender in the relationship between marital quality and mental health from a sociocultural (machismo) standpoint to extend our findings (Warner & Adams, 2016). Results point to the importance of another potential form of social support for older Hispanic adults, namely religion. Religious affiliation was associated with fewer depressive symptoms, on average, which is consistent with some of the existing literature on the protective role of religion. Based on data from 64 Hispanic adults who were church attendees in Maine and New York, Caplan (2019) reported that religious affiliations may indicate additional social support networks through church communities or perceived support from a higher power, and stress coping mechanisms, which collectively counteract depressive symptomatology. Also, religious activities can be emotionally gratifying activities in later life (Carstensen, 2021), particularly given the importance Hispanic culture places on religion. In addition, religion may function as a coping strategy for negative marital quality and associated stress.

Although the Hispanic paradox (Markides & Eschbach, 2011) suggests that Hispanic older adults have more optimal health regardless of lower SES, we found that employment, income, and assets, were significant predictors of depressive symptoms among older married Hispanic adults in the United States. Differences in mental health and marital quality by separate Hispanic groups (e.g., Mexican and Cuban Americans) should be further examined in future research. Another finding was that higher ADL limitations and lower self-rated health were associated with a greater number of depressive symptoms. Having functional limitations and poor self-rated health are linked with lower participation in physical and social activities and, in turn, fewer opportunities for stress coping (Cunningham et al., 2020; Smith et al., 2017). Moreover, functional limitations and poor health may result in social isolation and loneliness, which have significant impacts not only on marital quality but also on mental health outcomes (Latham-Mintus et al., 2022; Leigh-Hunt et al., 2017). Future research needs to examine the roles of gender, health, and functional limitation differences in the relationships between marital quality and mental health (Warner & Adams, 2016). The findings about these statistically significant covariates should be treated as preliminary or exploratory analysis results because these covariates are specifically selected for addressing the research questions with the marital quality indicators. These findings of the covariates are meant to be supplemental information for the main hypothesis tests and for future research.

Limitations

Several limitations should be noted. First, selection bias (Williams et al., 2009), survivor bias, and omitted variable bias cannot be ruled out. Those who got married and stayed married into later life may represent only a subpopulation of the older Hispanic adults. Future research needs to further explore the associations between marital quality and a variety of mental health outcomes over the life course of older Hispanic adults. More detailed information about the country of origin, acculturation, language skills, marital interaction (e.g., spending time more or less with one’s spouse), and specific sources of stress (e.g., racial discrimination) might have improved the performance of the statistical models (Basáñez et al., 2013). For instance, the experience of discrimination can serve as a chronic stressor that can affect both one’s relationship with their spouse and their mental health (Aneshensel & Mitchell, 2014). By the same token, future research should develop a theoretical framework to specifically test the interaction effects of key factors such as SES and physical health (Warner & Adams, 2016). Additionally, building on the cross-sectional analysis of the current study, development of later-life course model of mental health and marital quality, and use of longitudinal data specifically for older Hispanic adults should be the next steps. Also, we did not focus on concordance or discordance between husbands’ and wives’ marital quality. Previous studies showed that depressive symptoms may transmit from husbands to their wives among Hispanic adults (Helms et al., 2014). Future research should utilize dyadic and longitudinal data to further examine the dynamics of within-couple interactions (Kouros & Cummings, 2010), and possible effects on mental health among older Hispanic populations.

Contributions

The present study made three main contributions to the literature. First, the findings moved the socioculturally important study of the relationship between marital relationships and depressive symptoms in later life beyond the crude measure of marital status among older Hispanic adults in the United States. Given the size of and high percentage of immigrants in the Hispanic population, as well as cultural orientation to familismo, marital quality in later life should receive more attention from a public health standpoint. Specifically, our study aids in better understanding how sociocultural characteristics are related to the relationship between marital quality and mental health among the understudied older Hispanic adults in the United States. Second, results for two dimensions of marital quality—positive and negative—are informative for future public health and family policy and practice with older adults. The promotion of positive marital quality and prevention and mitigation of negative marital quality should be simultaneously considered as avenues for potentially improving the mental health of older Hispanic adults. Finally, whereas most previous studies compare older Hispanic adults to the general adult population, the present study provided population-level findings about the within-older Hispanic population differences in depressive symptoms by marital quality, religious affiliation, SES, functional limitation, and health status. Although the focus of the study was on marital quality, other identified factors associated with depressive symptoms are informative to contextualize discussion on mental health promotion and family policy in the diverse U.S. society.

Policy Implications

The current study offered a few preliminary policy implications. The Hispanic population in the United States is currently the largest racial–ethnic minoritized group, and this study shows that sociocultural factors play a significant role in the relationship between marital quality and depressive symptoms among this group. The present study found that both positive and negative marital quality are linked with depressive symptoms. Thus, it is critical to implement multidimensional interventions (e.g., promotion of positive marital quality and prevention of negative marital quality) as well as culturally sensitive approaches, such as mental health checkups and behavioral modifications (e.g., mediation) in older Hispanic individuals’ preferred languages, to ultimately mitigate the onset of depressive symptoms. These insights can be added to the existing marital quality and mental health promotion policy and intervention programs. For example, the Supporting Healthy Marriage (SHM) program, funded by the U.S. Department of Health and Human Services, has been implemented in eight local communities across seven states. The SHM program targeted low-income married couples with a child/ren or expecting a child, and evaluation data show it significantly improved marital conflict coping skills and lowered psychological distress through year-long education interventions. The SHM program was reported to be particularly effective among economically disadvantaged Hispanic couples, though it did not examine older Hispanic married couples (Hsueh et al., 2012). In conclusion, both positive and negative dimensions of marital quality are associated with the depressive symptoms of older Hispanic adults in the United States, even after accounting for demographic, socioeconomic, and health characteristics. Policy interventions to improve marital quality have potentially significant impacts on the mental health of the largest racial–ethnic minoritized population of older adults.

Funding

None.

Conflict of Interest

None.

Data Availability

The data analyzed in this study were from the 2016–2018 Health and Retirement Study. The public use data are available at: https://hrs.isr.umich.edu/about.

Author Contributions

J. M. Nazario-Acevedo designed the study, wrote the introduction, and contributed to the discussion. T. Yamashita contributed to the study’s design, conducted the data analysis, and wrote part of the introduction and methods sections. J. R. Bulanda contributed to the conceptualization of the issues and development of the theoretical framework. J. S. Brown contributed to conceptualizing the issues, provided methodological guidance, and helped revise the paper.

References