Age Trajectories of Depressive Symptoms by Age at Immigration among Older Men and Women of Mexican Descent: The Role of Social Resources

Abstract

Drawing on eight waves of data from the Hispanic Established Population for the Epidemiologic Study of the Elderly, this study uses growth curve models to provide a better understanding of the impact of age at immigration and gender on age trajectories of depressive symptoms among older adults of Mexican descent (aged 65+). The findings reveal that (1) regardless of immigrant status and age at immigration, men have similar age trajectories of depressive symptoms; (2) compared with U.S.-born women, late-life (beyond age 50) immigrant women report more depressive symptoms at age 65, whereas midlife (between ages 20 and 49) immigrant women experience steeper increases in these symptoms with age; (3) controlling for socioeconomic status leads to advantages in mental health at age 65, but steeper age-related increases in depressive symptoms among midlife (between ages 20 and 49) immigrant men and fewer depressive symptoms among late-life (beyond age 50) immigrant women.

Keywords

age at immigration age trajectories depressive symptoms gender immigrant status Mexicans older adults

Depressive symptoms in later life are associated with higher prevalence of disability, morbidity, mortality, and rising health care costs (Blazer 2003). Prior research also demonstrates that the prevalence of depressive symptoms varies over the life course: it tends to decrease from young adulthood to middle age, but it starts rising in later life around age 65 to 70 (Wu, Schimmele, and Chappell 2012). Previous studies suggest that gender, immigrant status, and age at immigration might be related to individuals’ mental health. In particular, prior research provides consistent evidence that women tend to have worse psychological well-being than do men (Djernes 2006). However, the results of previous studies on the importance of immigrant status and age at immigration are mixed. For example, some research indicates that the U.S.-born have more depressive symptoms (Wassertheil-Smoller et al. 2014), whereas other studies show that elevated depressive symptomatology is more prevalent among the foreign born (Gerst et al. 2010). Similarly, greater age at immigration has been linked to fewer (Heilemann, Lee, and Kury 2002) as well as more depressive symptoms (Lum and Vanderaa 2010).

Yet research on the consequences of the interplay among gender, immigrant status, and age at immigration for the age trajectories of depressive symptoms in later life is nonexistent. Specifically, there is a limited understanding of whether and how mental health disparities vary by immigrant status/age at immigration with increasing age among older men and women of Mexican descent. At the same time, a few available studies on physical health suggest that these factors might make a difference in the age patterning of health in later life (Garcia and Reyes 2017; Gubernskaya 2014; Wakabayashi 2010). There is also a lack of research on risk and protective social resources that can be associated with immigrant status/age at immigration variations in the age patterning of depressive symptoms among older men and women. Identifying these resources can facilitate the development of prevention and intervention programs for older adults.

There are several reasons why it is important to investigate these issues among older adults of Mexican descent. Mexican Americans comprise two-thirds of the Hispanics in the United States, the largest ethnic minority group representing 16 percent of the total population (Ennis, Ríos-Vargas, and Albert 2011). Moreover, 35 percent of individuals of Mexican descent in the United States are immigrants (Gonzalez-Barrera and Lopez 2013). Immigration from Mexico is also one of the reasons that the highest rise in the U.S. population age 65 and older is predicted among the Hispanic elderly population, which is expected to increase from its present level of 6 percent to 17 percent by 2050 (Pew Hispanic Center 2008). Furthermore, older adults of Mexican descent, especially women, tend to have relatively high rates of depressive symptoms (Black, Markides, and Miller 1998).

On the basis of prior mostly cross-sectional research in this area, the present study examines differences and similarities by immigrant status/age at immigration not only in the average levels of but also in the rates of change in depressive symptoms with age separately for older men and women of Mexican descent. In particular, using data from eight waves of the Hispanic Established Population for the Epidemiologic Study of the Elderly (H-EPESE), this study draws on three competing theoretical frameworks, such as the cumulative disadvantage, age-as-leveler, and persistent health inequality perspectives, suggesting that immigrant status/age at immigration disparities in depressive symptoms can increase, decrease, or remain stable with age, respectively (Crystal and Shea 2002; Ferraro and Farmer 1996). Moreover, the present study investigates whether and how social resources, including socioeconomic status (education, financial strain), acculturation (language of interview), marital status, living arrangements (co-residence with child and other relatives and nonrelatives), and social integration (church attendance), are responsible for immigrant status/age at immigration variations in age trajectories of depressive symptoms among men and women in this population group.

Background

Immigrant Status, Age at Immigration, and Gender

Immigrant status may be related to higher levels of depressive symptoms in later life because it might be associated with lower levels of social resources (González, Haan, and Hinton 2001). Thus, the double jeopardy theory implies that due to differences in the access to and accumulation of resources, old age in combination with immigrant status, greater age at immigration, and/or being female can create a double disadvantage for health outcomes among foreign-born individuals, individuals who immigrated at greater ages, and/or women (Ferraro and Farmer 1996). Specifically, foreign-born women who immigrated at greater ages might experience health issues in later life due to their multiple disadvantaged social positions. In contrast, the healthy immigrant effect perspective suggests that despite multiple hardships and stressors related to foreign-born status, Hispanic immigrants, including those from Mexico, can have comparable or even better mental health than U.S.-born individuals (Gallo et al. 2009).

Overall, the findings of previous studies on mental health discrepancies by immigrant status in the general population of Mexican descent in the United States are mixed. Some research finds fewer depressive symptoms among immigrants from Mexico than among U.S.-born individuals of Mexican descent (e.g., Alegria, Shrout et al. 2007; Wassertheil-Smoller et al. 2014). Also, one study did not reveal any significant differences in depressive symptoms by immigrant status among persons of Mexican origin (Cuellar, Bastida, and Braccio 2004). At the same time, prior research with a specific focus on older adults of Mexican descent shows that immigrants are more susceptible to depressive symptoms than their U.S.-born counterparts (e.g., Black et al. 1998; Gerst et al. 2010; González et al. 2001).

Among the foreign born, a specific age at immigration might have benefits as well as disadvantages in terms of access to and accumulation of various resources in the country of destination, and as a result, it might have distinct implications for individuals’ health outcomes (Fuligni 2004; Leu et al. 2008). For example, younger age at immigration might be associated with greater exposure to the mainstream U.S. culture and subsequent increased acculturation and weaker ethnic identity, which might be linked to poorer health because of adverse lifestyle changes, weaker family ties, and diminished adherence to ethnic values and beliefs. Alternatively, greater age at immigration might be related to fewer opportunities for accruing protective resources such as higher educational attainment, financial assets, English language proficiency, and extensive and supportive social networks outside of family and ethnic community, due to lower overall incorporation into different social institutions among older individuals than younger ones (Fuligni 2004; Leu et al. 2008).

Furthermore, age at immigration might be related to health selection among immigrants because it can serve as an indicator of reasons for immigration. For example, immigration in midlife may be associated with better health outcomes because during this stage in the life course, individuals tend to immigrate looking for employment opportunities. In contrast, early-life and late-life immigrants move to a different country with their parents or to join their families, respectively, and, as a result, they are less likely to be self-selected on health (Markides et al. 2007). At the same time, it might be harder for midlife immigrants to access and accumulate socioeconomic resources (e.g., greater educational attainment) in the country of destination than for early-life immigrants (Fuligni 2004).

Correspondingly, prior research on the impact of age at immigration for mental health has contradictory results. Some research did not find that age at immigration matters for psychological well-being, measured by a dichotomized indicator of depressive symptomatology (Gerst et al. 2010) and self-reported mental health (Lam, Yip, and Gee 2012). Other studies demonstrate that compared with younger age at immigration or being U.S.-born, greater age at immigration might be related to lower levels of depressive symptoms (Heilemann et al. 2002) and depressive disorders (Alegria, Shrout et al. 2007) or to a greater prevalence of depressive symptoms (Lum and Vanderaa 2010) and poorer self-reported mental health (J. Angel, Buckley, and Sakamoto 2001). In addition, some research on individuals over age 18 indicates, albeit inconsistently, that certain life course stages at the time of immigration might be predictive of better mental health. For instance, Margarita Alegría, Norah Mulvaney-Day, and colleagues (2007) found that depressive disorders were more prevalent among Hispanic immigrants who came to the U.S. before age 13 or at age 35 and older (Alegría, Mulvaney-Day et al. 2007). At the same time, another study on Asian immigrants showed that compared with younger immigrants (i.e., by age 6), those who moved to the United States between ages 7 to 17 might have more depressive symptoms, whereas those who immigrated at age 25 and older might have fewer depressive symptoms (Gong et al. 2011). Discrepancies in findings of previous available studies on the implications of immigrant status and age at immigration for mental health across the life course might be attributed to a combination of factors, including the between-study variations in the age ranges of samples, categorizations of age at immigration, measures of mental health, and focus on specific racial/ethnic groups or the general population. It should be noted, however, that no specific patterns in the impact of immigrant status/age at immigration on psychological well-being emerge on the basis of the differences in the measures of mental health in prior studies.

In addition to immigrant status and age at immigration, gender is another key factor in depressive symptomatology among older adults. Women, including Mexican women, report more depressive symptoms in later life than do men (Black et al. 1998; Djernes 2006; González et al. 2001), which might be related to greater disadvantages in terms of different types of resources among women. Moreover, women tend to perceive similar situations as more distressing and to experience daily stressors related to concerns about family, households, and friends more frequently (Aranda et al. 2001). Unique stressors among Hispanic immigrant women might be linked to discrepancies in gender role expectations between their countries of origin and the United States, and to the pressure to maintain traditional cultural identities (Aranda et al. 2001). Furthermore, the implications of immigrant status and age at immigration for mental health may vary for men and women because they might have different reasons for immigration at different ages and might adjust to the country of destination in distinct ways. For example, immigration in midlife among Mexicans is more likely to be associated with health advantages among men than women (Markides et al. 2007). Specifically, literature demonstrates that migration from Mexico has been dominated by working-age men who move to the United States in search of employment, whereas Mexican women predominately come to the United States to join their spouses (Donato 2010; Durand, Massey, and Zenteno 2001). As a result, these women might have less control over the decision to migrate which might make the process of immigration and adjustment to the new country more stressful for them and might adversely affect their postmigration psychological well-being. Also, among older cohorts of Mexican women, late-life immigration that tends to occur as a result of family reunification might lead to additional responsibilities related to family caregiving and caretaking (Parrado, Flippen, and McQuiston 2005), which might have negative consequences for their mental health.

In the present study, we conduct analyses separately for men and women to examine whether age at immigration makes a difference in age trajectories of depressive symptoms between immigrant men and women and their U.S.-born counterparts. In particular, we differentiate between immigrants who moved to the United States in early life (i.e., by age 19), midlife (i.e., between age 20 and 49), and late life (i.e., beyond age 50). This categorization of age at immigration facilitates comparisons between our findings and the results of prior research on later-life changes in physical and cognitive functioning among older adults of Mexican descent that used the same age-at-immigration groups (Garcia and Reyes 2017; Hill, Angel, and Balistreri 2012; Hill, Angel, Balistreri, and Herrera 2012). Moreover, similar groups by age at immigration were considered in available prior research on the age trajectories of physical health among older adults across different racial/ethnic groups in the United States, including Hispanics (Gubernskaya 2014; Wakabayashi 2010).

Overall, prior research and theoretical perspectives suggest that although early life (by age 19) immigrants are less likely to be self-selected on health (e.g., than midlife immigrants), they are more likely to have access to and accumulate resources throughout the life course. At the same time, midlife (between ages 20 and 49) immigrants are more likely to be self-selected on health than early-life and late-life immigrants, whereas late-life (beyond age 50) immigrants are less likely to be self-selected on health and are more likely to be disadvantaged in terms of resources than other age-at-immigration groups. Thus, we propose the following hypotheses for the levels of depressive symptoms at age 65 (the lowest observed age in H-EPESE).

Hypothesis 1: Compared with their U.S.-born counterparts, Mexican-born men and women who immigrated in early life (by age 19) will have comparable levels of depressive symptoms at age 65.

Hypothesis 2: Compared with their U.S.-born counterparts, midlife (between ages 20 and 49) immigrants will have lower levels of depressive symptoms at age 65.

Hypothesis 3: Compared with their U.S.-born counterparts, Mexican-born men and women who immigrated in late life (beyond age 50) will have greater levels of depressive symptoms at age 65.

Trajectories of Depressive Symptoms in Later Life

To the best of our knowledge, prior research did not examine implications of immigrant status/age at immigration for the age patterning of depressive symptoms in later life, in particular, among older adults of Mexican descent. At the same time, a few available studies on trajectories of self-reported health, physical functioning, and cognition among older adults suggest that age at immigration, although with some variations in specific life course stages at immigration and by gender across studies, might matter not only for the levels of but also for changes in these health dimensions in later life (Garcia and Reyes 2017; Gubernskaya 2014; Hill, Angel, and Balistreri 2012; Hill, Angel, Balistreri, and Herrera 2012; Wakabayashi 2010). Namely, those studies that similar to the present research, investigated these issues among older adults of Mexican descent and used data from H-EPESE indicate that especially midlife (between ages 20 and 49) immigrants might have distinct later-life trajectories of health. For example, drawing on six waves of data from H-EPESE, Terrence D. Hill, Jacqueline L. Angel, and Kelly S. Balistreri (2012) found that immigrants from Mexico who started residing in the United States by age 19 (in early life) or beyond age 50 (in late life) had comparable Wave 1 levels and rates of decrease in cognitive functioning across waves to those of their U.S.-born counterparts. In contrast, midlife (between ages 20 and 49) immigrants had better cognition at Wave 1 and slower rates of cognitive decline over time than U.S.-born persons of Mexican descent (Hill, Angel, and Balistreri 2012). Moreover, using the same data, another study by Hill, Angel, and Balistreri (2012) demonstrated that only midlife immigrant men experienced slower cognitive decline over time (Hill, Angel, Balistreri, and Herrera 2012). Using data from seven waves of H-EPESE, Marc A. Garcia and Adriana M. Reyes (2017) revealed that midlife (between ages 20 and 49) immigrant men and women and late-life (beyond age 50) immigrant men had fewer functional limitations measured by the performance-oriented mobility assessment (POMA) at age 65 but steeper increases in these performance-based functional limitations with age, compared with their U.S.-born counterparts. The latter study did not find any age-at-immigration differences in the age patterning of self-reported limitations with activities of daily living (ADL), however.

Three competing theoretical frameworks, including the cumulative disadvantage, age-as-leveler, and persistent health inequality perspectives (Crystal and Shea 2002; Ferraro and Farmer 1996), help delineate whether immigrant status/age at immigration differences in the age trajectories of depressive symptoms might be characterized by change (i.e., divergence or convergence) or stability in later life. Specifically, the cumulative disadvantage perspective suggests that mental health disparities related to immigrant status and age at immigration can increase with age because advantages and disadvantages in different types of resources accumulate over time (Crystal and Shea 2002). Given that individuals who immigrate at greater ages tend to have lower levels of resources, we hypothesize the following:

Hypothesis 4: Steeper increases in depressive symptoms with age will be more likely among immigrants who moved to the United States at greater ages.

In contrast, the age-as-leveler framework indicates that disparities in depressive symptoms related to immigrant status and age at immigration can narrow with increasing age for a combination of reasons, including higher prevalence of selective mortality at younger ages among the more disadvantaged, coping strategies adopted throughout the life course by the latter individuals, as well as the equalizing consequences of entitlement programs for older adults such as Social Security and Medicare (Ferraro and Farmer 1996). Therefore, we can expect the following:

Hypothesis 5: The differences in depressive symptoms by immigrant status/age at immigration will diminish with increasing age across immigrant status/age-at-immigration groups among men and women of Mexican descent.

At the same time, according to the persistent health inequality perspective, mental health disparities related to immigrant status and age at immigration will not change over time because the effects of discrepancies in resources on health remain stable with age (Ferraro and Farmer 1996). Thus, in contrast to our hypotheses on the basis of the cumulative disadvantage and age-as-leveler perspectives, we can predict the following:

Hypothesis 6: Immigrant status/age-at-immigrant differences in depressive symptoms will stay constant with increasing age for men and women of Mexican descent.

Overall, prior research offers little guidance for potential immigrant status/age at immigration and gender variations in the age trajectories of depressive symptoms in later life. Moreover, potential differences by immigrant status/age at immigration and gender in advantages and disadvantages related to reasons for migration, self-selection on health, and access and accumulation of resources in the country of destination might counterbalance each other. Thus, age-related patterns that are suggested by the three theoretical perspectives discussed above are equally possible across immigrant status/age at immigration groups and among both men and women.

Importance of Social Resources

Such social resources as socioeconomic status, acculturation, marital status, living arrangements, and social integration may be linked to the prevalence of depressive symptoms among older adults. In addition, there is some evidence that these resources might vary by immigrant status and may be responsible, to a certain extent, for mental health disparities in later life (DuBard and Gizlice 2008; W. Kim and Chen 2011; Ladin and Reinhold 2013; Law and Sbarra 2009; Schaan 2013). Yet it remains unclear whether and how these resources might make a difference in the potential variations by immigrant status/age at immigration in the age patterning of depressive symptoms among older men and women of Mexican descent.

Limited socioeconomic resources, such as fewer years of education and greater financial strain, can serve as indicators of chronic stress related to difficulties meeting basic needs and, as a result, to poorer mental health (Kahn and Perlin 2006; Wassertheil-Smoller et al. 2014). Lower levels of socioeconomic resources tend to prevail among foreign-born individuals (Wakabayashi 2010) and have been found to be partially responsible for mental health disparities related to immigrant status (Ladin and Reinhold 2013).

Immigrants from non-English-speaking countries, especially those who immigrated at greater ages, have considerably more limited English skills, compared with native-born Americans (Carliner 2000). Lower English proficiency, however, is associated with higher rates of depressive symptoms throughout the life course (G. Kim et al. 2011) because it can serve as an indicator of lower levels of acculturation that might lead to such disadvantages as greater social isolation, decreased autonomy, self-esteem, and self-efficacy, and inadequate access to and utilization of health care services (DuBard and Gizlice 2008).

Literature also shows that married older adults are at lower risk for elevated depressive symptomatology (González et al. 2001; Schaan 2013) because marriage tends to serve as an important source of different types of support for individuals (Carr and Springer 2010). Previous studies demonstrate that marriage rates are higher among foreign-born Hispanics, including Mexicans, than their U.S.-born counterparts (Stykes, Payne, and Gibbs 2014).

Co-residence with children and other relatives and nonrelatives might be beneficial for mental health in later life because it can facilitate older adults’ access to additional assistance, care, and resources as well as decrease feelings of social isolation and loneliness (Choi and Bohman 2007; W. Kim and Chen 2011). Yet co-residence might also be linked to greater levels of individuals’ stress for a combination of reasons, including greater interpersonal conflict and more responsibilities and demands on older adults because these living arrangements can occur in response to the needs of not only older but also younger generations in the family (Hughes and Waite 2002). Prior research also suggests that compared with the U.S.-born, the foreign born, particularly recent immigrants and those who immigrated to the United States at greater ages, are more likely to live in extended households (Burr, Mutchler, and Gerst-Emerson 2013).

Church attendance, as one of the dimensions of social integration, is important for mental health in later life because regular attendance of religious services might be linked to greater availability of social support and resources and stronger social networks (Law and Sbarra 2009). Findings of prior research on immigrant status differences in church attendance are inconsistent. Overall, immigrants from Mexico seem to be more likely to attend religious services in the United States than immigrants from other regions of the world (Cadge and Ecklund 2006). Nonetheless, frequency of church attendance might decrease (Cadge and Ecklund 2006), increase (Akresh 2011), or stay the same (Alanezi and Sherkat 2008) after immigration.

On the basis of prior research, we anticipate the following:

Hypothesis 7: Variations in socioeconomic resources, acculturation, marital status, living arrangements, and social integration will reduce, at least partially, immigrant status/age at immigration disadvantages in depressive symptoms or lead to comparable or even better mental health among immigrants than U.S.-born individuals.

Data and Method

Data

The present study draws on data from eight waves of the H-EPESE. H-EPESE has an area probability, multistage sample of noninstitutionalized Mexican Americans aged 65 and older residing in five southwestern states—Texas, New Mexico, Colorado, Arizona, and California (Markides et al. 1996). Data for Wave 1 were collected in 1993–1994. The response rate for Wave 1 that included 3,050 individuals was 83 percent (Markides et al. 1996). Subsequent waves were administered at approximately two-year intervals: Wave 2 in 1995–1996 (n = 2,438), Wave 3 in 1998–1999 (n = 1,980), Wave 4 in 2000–2001 (n = 1,682), Wave 5 in 2004–2005 (n = 2,069), Wave 6 in 2007 (n = 1,542), Wave 7 in 2010–2011 (n = 1,078), and Wave 8 in 2012–2013 (n = 744). Wave 5 included interviews with new respondents age 75 and over (n = 902) who were not interviewed in previous waves but who were also reinterviewed during the following waves. We excluded data from proxies, which were obtained when the sampled respondents could not answer questions themselves due to cognitive issues, illness, hospitalization, or temporary absence (Wave 1: n = 316 or 10.36 % Wave 2: n = 271 or 11.12 % Wave 3: n = 267 or 13.48 % Wave 4: n = 204 or 12.13 % Wave 5: n = 327 or 15.80 % Wave 6: n = 307 or 19.91 % Wave 7: n = 180 or 16.70 % and Wave 8: n = 158 or 21.24 percent). Furthermore, we did not include respondents with missing information on depressive symptoms (Wave 1: n = 35 or 1.28 % Wave 2: n = 15 or 0.69 % Wave 3: n = 31 or 1.57 % Wave 4: n = 5 or 0.30 % Wave 5: n = 4 or 0.23 % Wave 6: n = 8 or 0.65 % Wave 7: n = 65 or 7.24 % and Wave 8: n = 12 or 2.05 percent). We used the repeated observations from 3,475 respondents (1,450 men and 2,025 women) to create a data file that consisted of 12,396 person-periods (i.e., the total number of repeated observations from the sample respondents across waves; 4,897 for men and 7,499 for women). Respondents could contribute up to eight waves of data and were included in the current analysis if they had participated in at least one wave of H-EPESE. On average, respondents participated in about three waves of data (M = 3.59, SD = 2.12).

Measures

Immigrant status/age-at-immigration groups and education were treated as time-invariant variables measured only at H-EPESE baseline, which was defined as Wave 1 for original respondents and as Wave 5 for additional respondents. The rest of the variables were measured at each wave. Depressive symptoms at each wave were measured with the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff 1977). H-EPESE contains 20 items of the CES-D scale and includes 16 negative affects (e.g., I felt depressed) and four positive affects (e.g., I felt hopeful about the future), capturing whether respondents felt a certain way in the past week (0 = rarely or none of the time, 3 = most or all of the time). We recoded the responses for the positive affect items and used the total score of the CES-D scale as a continuous measure (Cronbach’s α = .84–.96, depending on the wave).

Immigrant status/age-at-immigration groups

At baseline, respondents reported whether they were born in the United States. Foreign-born respondents specified the age at which they came to the United States to stay. The mean age at immigration for our sample was 33 years (SD = 19.02). We created eight dummy variables (four for each gender) to capture immigrant status and age at immigration to differentiate between the U.S.-born and three age-at-immigration groups for each gender (i.e., those who immigrated in early life [by age 19], midlife [between ages 20 and 49], and late life [beyond age 50]). We conducted our analyses separately for men and women to compare age-at-immigration groups to the U.S.-born within the same gender.

Age, assessed in years at each wave and centered at the lowest observed age of 65, was used as a measure of time in growth curve models. The rest of covariates were mean centered to facilitate the interpretation of findings (Singer and Willett 2003).

Socioeconomic status (SES)

Education at baseline was measured in years from 0 to 20. Financial strain at each wave reflected whether the respondent had difficulty paying monthly bills (1 = none, 4 = a great deal). We were not able to include additional measures of SES (e.g., income and occupation) in the current analysis because information on these measures is not available in each wave of H-EPESE. At the same time, income and occupation can be changing throughout later life (Haan, Zeki Al-Hazzouri, and Aiello 2011). However, prior research suggests that compared with objective (e.g., income) measures, subjective (e.g., financial strain) measures of SES may be more adequate indicators of economic resources and even more critical predictors of health in later life (Hansen, Slagsvold, and Moum 2008).

Acculturation

Interview in Spanish captured whether the survey was conducted in Spanish at each wave (0 = no, 1 = yes). Respondents were given a choice to take the survey in English or Spanish at each wave of H-EPESE. We did not use a more comprehensive indicator of English proficiency because questions that measured respondents’ ability to understand, speak, and read English were asked only at Wave 1 but not at Wave 5 when additional respondents were interviewed. In general, it is challenging to capture acculturation with only one indicator or even a combination of items (Hunt, Schneider, and Comer 2004; Lara et al. 2005). At the same time, English proficiency as well as language preference and use, including language of interview, are the most frequently used measures of acculturation (e.g., for language of interview, see DuBard and Gizlice 2008; Gerst et al. 2010; Gubernskaya 2014). In addition, although these variables might be considered only rough proxies of acculturation, language ability tends to be the key component of acculturation scales because it might serve as an important indicator of individuals’ cultural preferences (Hunt et al. 2004; Lara et al. 2005).

Marital status

Married measured whether respondents were married (0 = no, 1 = yes). We were not able to include additional indicators of marital status because men who immigrated to the United States beyond age 50 had no observations for the never-married group.

Living arrangements

Based on the information from household rosters, co-residence with child and co-residence with others (0 = no, 1 = yes) measured whether at least one child, including stepchildren, and other relatives and nonrelatives, excluding spouses or children, resided with the respondent, respectively.

Social integration

Church attendance measured how frequently the respondent attended religious services at each wave (1 = never or almost never, 5 = more than once a week).

Because panel attrition due to the death was linked to elevated depressive symptomatology for the total sample, the analyses controlled for died during survey (0 = no, 1 = yes). Preliminary analyses (not shown) indicated that the inclusion of this measure did not alter the results. Using mi impute in Stata 14, we imputed missing values on all the covariates. These variables had less than 2 percent of missing values. Other methods of dealing with missing data such as mean substitution and listwise deletion produced similar findings (not shown).

Analytic Strategy

Descriptive statistics by immigrant status/age-at-immigration groups for baseline are presented in Table 1. Mean differences between U.S.-born and age-at-immigration groups within the same gender were examined with t-tests. Zero-order correlations (Table 2) confirmed that none of the correlations among the study variables used in the same growth curve models were above .35.

Table 1.

Means (Standard Deviations) for Study Variables by Immigrant Status and Age-at-immigration Groups for Baseline (N = 3,475; 12,396 Person Periods).

Variables	Range	Men				Women
		Age at immigration				Age at immigration
		U.S.-born	0–19	20–49	50+	U.S.-born	0–19	20–49	50+
Depressive symptoms	0–60	7.86 (8.14)	7.68 (7.72)	7.15 (8.12)	9.38 (8.60)	10.21 (9.40)	10.91 (9.57)	11.34 (10.14)^a	12.55 (11.10)^a
Age	65–108	74.22 (6.43)	77.19 (7.92)^b	74.20 (6.42)	76.78 (7.27)^b	74.25 (6.46)	77.50 (7.53)^a	74.18 (6.77)	76.20 (7.12)^a
Socioeconomic status
Education (in years)	0–20	6.17 (4.34)	4.04 (3.55)^b	3.80 (3.32)^b	2.71 (2.73)^b	5.73 (4.01)	4.43 (3.50)^a	4.21 (3.48)^a	3.05 (3.24)^a
Financial strain	1 = none; 4 = a great deal	2.36 (1.09)	2.63 (0.99)^b	2.79 (0.91)^b	2.94 (0.86)^b	2.51 (1.06)	2.67 (1.03)^a	2.75 (0.93)^a	2.83 (0.92)^a
Acculturation
Interview in Spanish	0 = no; 1 = yes	0.65 (0.48)	0.86 (0.34)^b	0.94 (0.23)^b	0.98 (0.16)^b	0.68 (0.47)	0.80 (0.40)^a	0.95 (0.21)^a	0.96 (0.20)^a
Marital status
Married	0 = no; 1 = yes	0.71 (0.45)	0.71 (0.45)	0.76 (0.43)	0.77 (0.42)	0.42 (0.49)	0.29 (0.45)^a	0.43 (0.50)	0.31 (0.46)^a
Living arrangements
Co-residence with child	0 = no; 1 = yes	0.27 (0.45)	0.28 (0.45)	0.36 (0.48)^b	0.46 (0.50)^b	0.29 (0.45)	0.35 (0.48)	0.32 (0.47)	0.38 (0.49)^a
Co-residence with others	0 = no; 1 = yes	0.18 (0.39)	0.22 (0.41)	0.22 (0.41)	0.29 (0.46)^b	0.24 (0.43)	0.25 (0.43)	0.21 (0.40)	0.39 (0.49)^a
Social integration
Church attendance	1 = never; 5 = more than once a week	2.84 (1.35)	2.83 (1.32)	2.81 (1.35)	2.90 (1.26)	3.25 (1.27)	3.23 (1.30)	3.25 (1.26)	3.07 (1.26)
Died during survey	0 = no; 1 = yes	0.64 (0.48)	0.68 (0.47)	0.52 (0.50)^b	0.64 (0.48)	0.65 (0.48)	0.73 (0.45)^a	0.59 (0.49)^a	0.58 (0.49)^a
n		817	161	352	120	1,150	203	472	200
Person-periods across waves		2,689	481	1,333	394	4,337	705	1,785	672

Note. Baseline is defined as Wave 1 for original respondents or Wave 5 for additional respondents. Mean values are reported for noncentered variables. Means for dichotomous variables can be interpreted as the proportion of the observations coded 1 on a specific variable. t-tests were computed for difference in means.

Statistically significant differences between U.S.-born women and each group of immigrant women at the .05 level.

Statistically significant differences between U.S.-born men and each group of immigrant men at the .05 level.

Table 2.

Correlations among Study Variables at Baseline.

Variables	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
1. U.S.-born men	—
2. Immigrant men (0–19)	−.12***	—
3. Immigrant men (20–49)	−.19***	−.07***	—
4. Immigrant men (50+)	−.10***	−.04*	−.06***	—
5. U.S.-born women	−.39***	−.16***	−.24***	−.13***	—
6. Immigrant women (0–19)	−.14***	−.05**	−.08***	−.05**	−.18***	—
7. Immigrant women (20–49)	−.22***	−.09***	−.13***	−.08***	−.28***	−.10***	—
8. Immigrant women (50+)	−.14***	−.05**	−.08***	−.05**	−.17***	−.06***	−.10**	—
9. Education	.16***	−.05**	−.10***	−.11***	.13***	−.04*	−.08***	−.12***	—
10. Financial strain	−.12***	.01	.07***	.07***	−.05**	.02	.06***	.06***	−.30***	—
11. Interview in Spanish	−.17***	.04**	.13***	.09***	−.16***	.01	.17***	.11***	−.35***	.35***	—
12. Married	.20***	.08***	.15***	.09***	−.17***	−.12***	−.09***	−.11***	.08***	−.02	−.02	—
13. Co-residence with child	−.04**	−.01	.04*	.06***	−.04*	.02	.01	.04*	−.07***	.01	.04*	−.01	—
14. Co-residence with others	−.06***	−.01	−.01	.03	.02	.01	−.02	.09***	−.05**	−.01	.04*	−.16***	.30***	—
15. Church attendance	−.10***	−.04*	−.07***	−.02	.10***	.03	.06**	.01	.07***	−.06***	−.05**	.01	−.07***	−.01	—
16. Died during survey	.09***	.05**	−.04*	.03	−.03	.03	−.07***	−.05**	−.04*	.03	−.01	−.04*	.01	−.01	−.11***	—
17. Depressive symptoms	−.10***	−.04**	−.09***	−.01	.05**	.04*	.08***	.08***	−.10***	.17***	.02	−.10***	−.01	.01	−.07***	.08***	—

p < .05. **p < .01. ***p < .001 (two-tailed tests).

Because we used panel data, we employed growth curve models to assess immigrant status/age-at-immigration differences in the age trajectories of depressive symptoms. The analyses were conducted separately for men and women, with U.S.-born individuals serving as a reference group. The growth curve analytic method estimates a single model with two levels (Singer and Willett 2003). Individual trajectories of change are presented at Level 1. The variance in trajectories of change between individuals is estimated at Level 2. We applied Stata xtmixed procedure with maximum likelihood estimation, which is appropriate for analyzing panel data with attrition. At Level 1, we modeled a linear change trajectory of depressive symptoms of individual i at time t (DS_ti) as a function of age (Age_ti). We did not include a quadratic term for age because likelihood-ratio tests of linear and quadratic models suggested that a linear growth curve provided a better empirical fit to the data. We then included our main independent variables, capturing three age-at-immigration groups n (AIG_nti). We specified U.S.-born individuals as a reference group. To examine whether the age-at-immigration effect on the individual’s depressive symptoms can vary by age, we added the interaction terms between age and each of the three age-at-immigration groups n. The Level 1 equation is presented below:

{DS}_{t i} = π_{0 i} + π_{1 i} {Age}_{t i} + π_{2 n} {AIG}_{n t i} + π_{3 n} {AIG}_{n t i} \times {Age}_{t i} + π_{4} Died + ε_{t i},

where π_0i stands for the intercept and indicates the average number of depressive symptoms at age 65 (the lowest observed age), π_1i denotes the linear age slope and represents the average rate of change in depressive symptoms with each additional year of change, π_2n is the change associated with each age-at-immigration group, π_3n specifies the linear slope for each age-at-immigration group, π₄ is the change associated with panel attrition, and ε_ti is the within-individual error term.

At Level 2, the two submodels of the intercept π_0i and the linear rate of change π_1i examine the between-individual variations in the age trajectory of depressive symptoms:

π_{0 i} = β_{00} + ζ_{0 i}, and

π_{1 i} = β_{10} + ς_{1 i} .

In these submodels, β₀₀ and β₁₀ are the average level of depressive symptoms and the average linear rate of change, respectively. Random error terms of the average intercept and the average linear rate of change are presented with ς_0i and ς_1i, respectively.

The combined model of the Levels 1 and 2 submodels was written as follows:

\begin{matrix} {DS}_{t i} = (β_{00} + β_{10} {Age}_{t i} + π_{2 n} {AIG}_{n t i} + π_{3 n} {AIG}_{n t i} \times {Age}_{t i} + π_{4} Died) \\ + [ε_{t i} + ζ_{0 i} + ζ_{1 i} {Age}_{t i}] . \end{matrix}

This model is presented in Model 1 in Table 3 for men and in Model 1 in Table 4 for women.

Table 3.

Trajectories of Depressive Symptoms by Age-at-immigration Groups: Men (n = 1,450; 4,897 Person Periods).

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Fixed effects
Intercept	5.169***	5.437***	5.179***	5.813***	5.172***	5.234***	5.903***
Linear slope (Age)	0.176***	0.173***	0.176***	0.152***	0.175***	0.156***	0.137***
Age at immigration^a
0–19	0.341	−0.167	0.328	0.261	0.289	0.379	−0.079
20–49	−0.782	−1.390*	−0.798	−0.700	−0.695	−0.752	−0.989
50+	1.542	0.726	1.523	1.645	1.703	1.375	1.079
Age at immigration^a × Age
0–19	−0.029	−0.017	0−.028	−0.022	−0.026	−0.027	−0.010
20–49	0.064	0.077*	0.064	0.062	0.060	0.062	0.068
50+	−0.062	−0.048	−0.062	−0.062	−0.069	−0.048	−0.046
Socioeconomic status
Education		−0.077*					−0.070
Financial strain		0.727***					0.779***
Acculturation
Interview in Spanish			0.052				−0.704*
Marital status
Married				−1.638***			−1.392***
Living arrangements
Co-residence with child					−0.832**		−0.785**
Co-residence with others					0.389		0.106
Social integration
Church attendance						−0.751***	−0.699***
Died during survey	0.551	0.470	0.553	0.557	0.532	0.415	0.296
Random effects
Intercept SD	3.855	3.749	3.855	3.834	3.847	3.758	3.626
Linear slope (Age) SD	0.001	0.001	0.001	0.001	0.001	0.001	0.001
Residual SD	6.671	6.660	6.671	6.650	6.666	6.641	6.609
Log likelihood	−16,760.85	−16,734.40	−16,760.84	−16,743.94	−16,756.03	−16,723.82	−16,678.95
AIC	33,545.70	33,496.79	33,547.67	33,513.88	33,540.06	33,473.64	33,395.90
BIC	33,623.66	33,587.74	33,632.13	33,598.33	33,631.01	33,558.10	33,519.33

Note. AIC = Akaike information criterion; BIC = Bayesian information criterion.

Reference category: U.S.-born.

p < .05. **p < .01. ***p < .001 (two-tailed tests).

Table 4.

Trajectories of Depressive Symptoms by Age-at-immigration Groups: Women (n = 2,025; 7,499 Person Periods).

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Fixed effects
Intercept	7.400***	7.422***	7.268***	7.555***	7.401***	8.032***	7.985***
Linear slope (Age)	0.173***	0.177***	0.176***	0.148***	0.172***	0.130***	0.116***
Age at immigration^a
0–19	0.525	0.271	0.665	0.338	0.524	0.620	0.392
20–49	0.546	0.210	0.820	0.627	0.543	0.491	0.689
50+	2.652**	2.078*	2.917**	2.536**	2.655**	2.357*	2.099*
Age at immigration^a × Age
0–19	−0.006	−0.001	−0.008	0.003	−0.006	−0.007	0.005
20–49	0.091*	0.099*	0.086*	0.085*	0.091*	0.099*	0.093*
50+	−0.004	0.010	−0.006	0.001	−0.004	0.017	0.030
Socioeconomic status
Education		−0.064					−0.076
Financial strain		0.689***					0.752***
Acculturation
Interview in Spanish			−0.879**				−1.561***
Marital status
Married				−1.173***			−1.233***
Living arrangements
Co-residence with child					0.025		−0.081
Co-residence with others					−0.022		−0.094
Social integration
Church attendance						−0.881***	−0.874***
Died during survey	0.487	0.313	0.509	0.434	0.487	0.268	0.063
Random effects
Intercept SD	5.350	5.257	5.335	5.330	5.349	5.215	5.064
Linear slope (Age) SD	0.001	0.001	0.001	0.001	0.001	0.001	0.001
Residual SD	7.580	7.576	7.579	7.575	7.580	7.553	7.543
Log likelihood	−26,827.61	−26,802.84	−26,823.49	−26,819.16	−26,827.60	−26,774.81	−26,730.89
AIC	53,679.22	53,633.69	53,672.99	53,664.33	53,683.21	53,575.62	53,499.78
BIC	53,762.29	53,730.60	53,762.98	53,754.32	53,780.12	53,665.61	53,631.30

Note. AIC = Akaike information criterion; BIC = Bayesian information criterion.

Reference category: U.S.-born.

p < .05. **p < .01. ***p < .001 (two-tailed tests).

The goal of our study was also to examine whether such resources as socioeconomic status, acculturation, marital status, living arrangements, and social integration make a difference in the implications of age at immigration for the age patterning of depressive symptoms. In Tables 3 and 4, Models 2 to 6 added variable(s) for each group of these resources separately. The full models controlling for all the resources simultaneously are presented in Models 7 in Tables 3 and 4.

Results

Descriptive Results

Table 1 shows that there were some statistically significant differences at baseline between the U.S.-born and age-at-immigration groups within the same gender. Compared with their U.S.-born counterparts, midlife (between ages 20 and 49) and late-life (beyond age 50) immigrant women reported more depressive symptoms. Regardless of gender, early-life (by age 19) and late-life (beyond age 50) immigrants were older than their U.S.-born counterparts. Moreover, regardless of gender and age at immigration, the foreign born had lower levels of education, greater financial strain, and greater likelihood of taking surveys in Spanish than the U.S.-born. Compared with U.S.-born women, early-life (by age 19) and late-life (beyond age 50) immigrant women were less likely to be married. Co-residence with child was more prevalent among midlife (between ages 20 and 49) immigrant men and among late-life (beyond age 50) immigrant men and women than among their U.S.-born counterparts. Also, regardless of gender, late-life (beyond age 50) immigrants were more likely to co-reside with others than their U.S.-born counterparts.

Results from Growth Curve Models

Results for men

Table 3 presents the results of growth curve models for older men of Mexican descent. Model 1 contains estimates for the implications of age-at-immigration groups on the initial level and rate of change in depressive symptoms controlling only for attrition due to death during the survey. On average, U.S.-born men of Mexican descent had 5.169 depressive symptoms at age 65 (the lowest observed age in H-EPESE). The significant positive estimate for the linear slope (.176) indicates that they experienced increases in depressive symptoms with age. Immigrant men, regardless of age at immigration, did not have significantly different levels at age 65 and rates of increase in depressive symptoms with age, compared with U.S.-born men.

In Model 2, lower educational attainment and greater financial strain were associated with more depressive symptoms among men. In addition, with the inclusion of these measures of socioeconomic resources, the negative estimate for the level of depressive symptoms and the positive estimate for the rate of change with age for midlife (between ages 20 and 49) immigrant men became statistically significant. That is, if these midlife immigrant men were less disadvantaged in terms of socioeconomic resources, they would have lower levels of depressive symptoms at age 65 than U.S.-born men. However, higher socioeconomic status is not protective against steeper increases in depressive symptoms with age among this group of immigrant men.

In Model 3, taking interviews in Spanish was not predictive of depressive symptoms among men. At the same time, Models 4, 5, and 6 demonstrate that being married, co-residence with child, and more frequent church attendance, respectively, are related to fewer depressive symptoms among men. Yet these measures of resources did not have implications for the association between age-at-immigration groups and trajectories of depressive symptoms observed in Model 1. Model 7 controls for all the resources and shows no considerable differences in the estimates for the age-at-immigration groups, compared with Model 1. Interestingly, the estimate for education was reduced to nonsignificance, whereas taking interviews in Spanish became predictive of fewer depressive symptoms among men in Model 7.

Results for women

Table 4 presents the results of growth curve models for older women of Mexican descent. Model 1 shows that U.S.-born women had, on average, 7.400 depressive symptoms at age 65 and, similar to men, experienced increases in depressive symptoms with age. Early-life (by age 19) and midlife (between ages 20 and 49) immigrant women had similar levels of depressive symptoms at age 65 to those of U.S.-born women. In contrast, late-life (beyond age 50) immigrant women reported significantly higher levels of depressive symptoms at age 65, compared with their U.S.-born counterparts. Early-life (by age 19) and late-life (beyond age 50) immigrant women had a similar rate of increase in depressive symptoms with age to that of U.S.-born women. Midlife (between ages 20 and 49) immigrant women, however, experienced a steeper increase in depressive symptoms with age than U.S.-born women.

Of the socioeconomic status measures added in Model 2, greater financial strain was predictive of more depressive symptoms among women. Moreover, the inclusion of socioeconomic status partly explained higher levels of depressive symptoms among late-life (beyond age 50) immigrant women. Model 3 demonstrates that taking surveys in Spanish was linked to fewer depressive symptoms among women. At the same time, an increase in the estimate for the level among late-life (beyond age 50) immigrant women indicates that they would report slightly more depressive symptoms if they completed interviews in English. In Model 4, being married was related to fewer depressive symptoms among women. In addition, the inclusion of marital status led to a slight decrease in the estimate for the initial level of depressive symptoms among late-life (beyond age 50) immigrant women. Namely, a greater likelihood of being unmarried partly explains more depressive symptoms at age 65 among this group of immigrant women than among U.S.-born women. Contrary to the results for men, co-residence with child did not predict depressive symptoms among women in Model 5. More frequent church attendance was associated with fewer depressive symptoms among women in Model 6. Also, the reduced estimate for the initial level among late-life (beyond age 50) immigrant women indicates that lower frequency of church attendance is partially responsible for more depressive symptoms at age 65 among these foreign-born women compared with their U.S.-born counterparts. Model 7 shows that relative to Model 1, the estimate for the level of depressive symptoms at age 65 among late-life (beyond age 50) immigrant women was lower when all the measures of resources were taken into account.

Again, comparing the estimates across Models 1–6 in Tables 3 and 4 reveals that socioeconomic resources are more likely than other resources to have implications for age trajectories of depressive symptoms among midlife (between age 20 and 49) immigrant men and late-life (beyond age 50) immigrant women. To illustrate the implications of socioeconomic resources more effectively, we created Figure 1, which presents age trajectories of depressive symptoms by immigrant status/age at immigration separately for men and women on the basis of Models 2 (i.e., models controlling for education and financial strain) in Tables 3 and 4, respectively. Age trajectories are specific for each immigrant status/age-at-immigration group and all the measures of covariates were mean centered.

Figure 1.

Age trajectories of depressive symptoms by immigrant status and age at immigration among older men and women of Mexican descent.

Discussion

To the best of our knowledge, this is the first study that investigated the extent to which immigrant status/age at immigration can make a difference in the levels and rates of change in age trajectories of depressive symptoms among older men and women of Mexican descent. In this regard, our study extends previous research on mental health in later life by demonstrating that (1) immigrant status/age at immigration might shape the age patterning of depressive symptoms in later life, (2) the implications of age at immigration for mental health may be different for men and women, and (3) social resources may play a role in the immigrant status/age-at-immigration disparities in mental health among older adults of Mexican descent.

Specifically, the present study underscores that depending on their age at immigration and, to a certain extent, gender, immigrants from Mexico might have similar or different age trajectories of depressive symptoms in later life relative to their U.S.-born counterparts. Namely, compared with U.S.-born individuals, early-life (by age 19) immigrants, regardless of gender, and late-life (beyond age 50) immigrant men did not differ in their initial levels and rates of increase in depressive symptoms with age, whereas midlife (between ages 20 and 49) immigrants, regardless of gender, and late-life immigrant women experienced distinct trajectories.

In particular, the present findings indicate that consistent with the healthy immigrant effect (Gallo et al. 2009) and Hypothesis 1, immigrant men and women who came to the United States in early life had comparable initial levels and age-related increases in depressive symptoms to those of their U.S.-born counterparts. Although they are less likely to be self-selected on health because they tend to immigrate with their parents, early-life immigrants might have more opportunities to have access to and to accumulate resources throughout the life course in the United States than those who immigrate at greater ages (Fuligni 2004; Leu et al. 2008; Markides et al. 2007). These findings are in line with some prior research showing that immigrant status (Cuellar et al. 2004) and age at immigration (Gerst et al. 2010; Lam et al. 2012) might not matter for individuals’ mental health. Moreover, previous studies on trajectories of self-reported health, physical functioning, and cognition among older adults of Mexican descent (Garcia and Reyes 2017; Hill, Angel, and Balistreri 2012; Hill, Angel, Balistreri, and Herrera 2012) as well as across different racial/ethnic groups in the United States (Gubernskaya 2014; Wakabayashi 2010) also suggest that early-life immigrants might have similar levels of and changes in these health outcomes in later life to those of the U.S.-born.

In contrast, the present study shows that compared with their U.S.-born counterparts, midlife immigrant men and women experienced steeper increases in depressive symptoms with age. At the same time, relative to the U.S.-born, midlife immigrant men had fewer depressive symptoms at age 65, whereas midlife immigrant women had comparable initial levels. The findings for the levels at age 65 among midlife immigrants provide support for the healthy immigrant effect (Gallo et al. 2009) and Hypothesis 2, whereas the results for the steeper increases with age are in line with the cumulative disadvantage perspective (Crystal and Shea 2002) and Hypothesis 4. It should be noted, however, that midlife immigrant men had lower levels of depressive symptoms at age 65 only when socioeconomic status was taken into account. The latter finding is consistent with a weaker version of the healthy immigrant effect, according to which immigrant health advantage or less deleterious health outcomes become evident only net of socioeconomic resources (Riosmena, Wong, and Palloni 2013). In addition, this finding suggests that it is essential to address issues related to limited socioeconomic resources among immigrants to improve their psychological well-being, which is in accord with some prior research on the importance of education and financial strain for individuals’ mental health (Kahn and Perlin 2006; Wassertheil-Smoller et al. 2014) and for immigrant status disparities in depressive symptoms (Ladin and Reinhold 2013).

Although midlife immigrants from Mexico, particularly men who usually immigrate in search of employment, are more likely to be self-selected in terms of health, they might accumulate lower levels of resources across the life course than the U.S.-born or early-life immigrants (Fuligni 2004; Leu et al. 2008; Markides et al. 2007), which might precipitate more adverse changes in their mental health in later life. At the same time, midlife immigrant women from Mexico who tend to immigrate to join their spouses might be less likely to experience mental health advantages at age 65 because the process of immigration and adjustment to the new country might be more stressful for them than for their male counterparts who might have more control over the decision to migrate (Donato 2010; Durand et al. 2001). Relatedly, using the same age-at-immigration groups as the present study, prior research on trajectories of physical and cognitive functioning among older adults of Mexican descent also found that compared with the U.S.-born, midlife immigrants might experience advantages in the levels of these health outcomes (Garcia and Reyes 2017; Hill, Angel, and Balistreri 2012; Hill, Angel, Balistreri, and Herrera 2012) and steeper increases in functional limitations with age (Garcia and Reyes 2017). In addition, research on different racial/ethnic groups in the United States that used similar age-at-immigration groups to those in the studies on older adults of Mexican descent demonstrates that immigrants that moved to the United States at age 18 and older might experience faster deterioration in self-reported health and physical functioning than the U.S.-born (Gubernskaya 2014; Wakabayashi 2010).

In the present study, late-life immigration was only related to greater levels of depressive symptoms at age 65 among women, which is in accord with the double jeopardy theory (Ferraro and Farmer 1996) and Hypothesis 3. Moreover, this group of immigrant women experienced mental health disadvantages throughout later life because they had a similar rate of increase in depressive symptoms with age to that of their U.S.-born counterparts, which provides support for the persistent health inequality perspective (Ferraro and Farmer 1996) and Hypothesis 6. At the same time, the findings indicate that greater financial strain partially accounted for their higher initial levels of depressive symptoms. Individuals who immigrate in late life may find it more difficult to get integrated into a new society and to have access to the same resources that become available over time to individuals who immigrate at younger ages (Fuligni 2004; Leu et al. 2008; Markides et al. 2007). Some prior research also demonstrates that older age at immigration can be linked to poorer mental health than younger age at immigration or being U.S.-born (e.g., J. Angel et al. 2001; Lum and Vanderaa 2010). In addition to greater stress related to acculturation and more limited resources, late-life immigration among women might have additional distinctive stressors related to household chores, family caregiving, and child caretaking due to gender differences in expectations and obligations related to traditional family roles that can be particularly true for older cohorts of women of Mexican descent (Parrado et al. 2005). Previous studies on age at immigration and health outcomes among older adults of Mexican descent suggest, however, that late-life immigration may be related to comparable levels of and rates of decrease in cognitive functioning over time among both genders (Hill, Angel, and Balistreri 2012; Hill, Angel, Balistreri, and Herrera 2012) or lower initial levels of, but steeper age-related increases in functional limitations only among men (Garcia and Reyes 2017).

Overall, the present results highlight a need for further research to better understand specific factors responsible for steeper increases in depressive symptoms among midlife immigrants and persistently elevated depressive symptomatology among late-life immigrant women. In particular, future research would benefit by taking into account individuals’ reasons for immigration and immigrants’ experiences associated with adaptation to life in the United States. However, the fact that we did not find disadvantages or steeper increases in depressive symptoms among all the age-at-immigration groups compared with the U.S.-born and our findings on the importance of certain social resources for the age trajectories of mental health (Hypotheses 7) indicate that particularly social connectedness might play an essential role in psychological well-being among older adults of Mexican descent. Specifically, several resources (i.e., interviews in Spanish, being married, church attendance, and co-residence with children) that might serve to a certain extent as indicators of greater social embeddedness were related to fewer depressive symptoms in this population group, which is consistent with greater importance of and reliance on informal networks for support and assistance among individuals of Mexican descent (Gonzales 2007; Russell and Taylor 2009).

In the present study, interviews in Spanish were associated with fewer depressive symptoms among men (only in the full model) and women. Our findings also suggest that immigrant women, regardless of age at immigration, would have reported slightly more depressive symptoms if they had completed interviews in English. Similarly, prior research demonstrates that interviews in Spanish among U.S. Hispanics can be associated with certain health advantages (DuBard and Gizlice 2008), including fewer depressive symptoms (Gerst et al. 2010). Cultural and linguistic differences in perceptions, assessment, and self-reporting of health issues might be responsible for the associations between interviews in Spanish and health outcomes (Lara et al. 2005). In addition, although the usage of Spanish might be an indicator of lower acculturation and, as a result, inadequate access to some socioeconomic resources, it might also be linked to stronger ethnic identity and greater embeddedness in family ties and ethnic communities, which might be beneficial for individuals’ mental health (Torres 2010). At the same time, our results for the association between acculturation and depressive symptoms could have been different if we had been able to include additional or more comprehensive indicators of acculturation, including formal assessment of English language proficiency; individuals’ sense of ethnic identity; knowledge of the country of origin; adherence to ethnic values, beliefs, and traditions; perceptions of discrimination (Finch and Vega 2003; Lara et al. 2005), and the Geriatric Acculturation Rating Scale that evaluates not only English and Spanish use but also such factors as connections to Latin America and dietary practices (González et al. 2001).

The present findings demonstrate that being married was related to fewer depressive symptoms among men and women of Mexican descent. In addition, a lower likelihood of being married, compared with the U.S.-born women, slightly accounted for higher initial levels of depressive symptoms among late-life (beyond age 50) immigrant women. Marriage may be beneficial for mental health among older adults (González et al. 2001; Schaan 2013) because spouses serve as important sources of support and assistance (Carr and Springer 2010).

Co-residence with children in the present study was predictive of depressive symptoms only among men. In multigenerational households, older adults might benefit from easier access to additional support, caregiving, and companionship (Choi and Bohman 2007; W. Kim and Chen 2011). Some prior research also suggests that the implications of living arrangements for older adults’ mental health might vary by gender. For example, living alone might be associated with poorer mental health among men, including Hispanic men (Russell and Taylor 2009), whereas co-residence with children might be less beneficial for women’s psychological well-being because it might lead to more responsibilities and demands among women due to the gendered nature of caregiving and caretaking (Hughes and Waite 2002).

The present findings also demonstrated that more frequent church attendance was beneficial for mental health among both older men and women of Mexican descent. Furthermore, lower frequency of church attendance might be partially responsible for more depressive symptoms at age 65 among late-life (beyond age 50) immigrant women. Regular attendance of religious services may be linked to stronger social networks and greater access to social support (Law and Sbarra 2009). Decreased religious participation among immigrants in the country of destination might be explained by a combination of factors, including the availability of the place of worship, geographic proximity to a religious center, employment opportunities, family responsibilities, demographic composition of the community, individuals’ desire to be with co-ethnics, search for support, and resources (Akresh 2011; Alanezi and Sherkat 2008; Cadge and Ecklund 2006).

Because of data limitations, we were not able to consider several factors that could better explain immigrant status/age-at-immigration differences in depressive symptoms among older adults of Mexican descent. In particular, future research would benefit by considering changes in income, homeownership, assets, and availability of health insurance and variations by immigrant status, age at immigration, and gender in reasons for immigration.

In spite of its limitations, this study provides us with a better understanding of the implications of immigrant status/age at immigration for later-life trajectories of depressive symptoms among older adults of Mexican descent. Specifically, the finding illustrates the importance of distinguishing between different age-at-immigration groups. This study suggests that compared with their U.S.-born counterparts, early-life immigrants might have similar trajectories of depressive symptoms, whereas midlife immigrants, regardless of gender, might be at risk for steeper increases in depressive symptomatology as they get older, despite lower or comparable levels of depressive symptoms at age 65. In contrast, late-life immigrant women might experience persistent disadvantages in psychological well-being throughout later life. Furthermore, insights from this study on risk and protective factors for depressive symptoms in this population group suggest that social programs addressing socioeconomic disadvantages and adverse consequences of being unmarried and facilitating regular religious participation might be beneficial for older adults of Mexican descent.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and /or publicationof this article: The work was supported by grant RO1 AG10939 from the National Institute on Aging and by Joseph Sidney Werlin Sociology Faculty Award at the University of Houston.

Author Biographies

Maria A. Monserud is an assistant professor of sociology at the University of Houston. Her research focuses on family inequality, health disparities, minority aging, and gender.

Kyriakos S. Markides is Gnitzinger Distinguished Professor of Aging at the University of Texas Medical Branch. His research has focused on Hispanic health and aging in the Mexican American population of the Southwestern United States.

References

Akresh

Ilana R.

2011. “Immigrants’ Religious Participation in the United States.” Ethnic and Racial Studies 34(4):643–61.

Alanezi

Fawaz

Sherkat

Darren E.

2008. “The Religious Participation of US Immigrants: Exploring Contextual and Individual Influences.” Social Science Research 37(3):844–55.

Alegría

Margarita

Mulvaney-Day

Norah

Torres

Maria

Polo

Antonio

Cao

Zhun

Canino

Glorisa

. 2007. “Prevalence of Psychiatric Disorders across Latino Subgroups in the United States.” American Journal of Public Health 97(1):68–75.

Alegria

Margarita

Shrout

Patrick E.

Woo

Meghan

Guarnaccia

Peter

Sribney

William

Vila

Doryliz

Polo

Antonio

Cao

Zhun

Mulvaney-Day

Norah

Torresa

Maria

Canino

Glorisa

. 2007. “Understanding Differences in Past Year Psychiatric Disorders for Latinos Living in the US.” Social Science & Medicine 65(2):214–30.

Angel

Jacqueline L.

Buckley

Cynthia J.

Sakamoto

Art

. 2001. “Duration or Disadvantage? Exploring Nativity, Ethnicity, and Health in Midlife.” The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences 56(5):S275–84.

Aranda

Maria P.

Castaneda

Irma

Lee

Pey-Jiuan

Sobel

Eugene

. 2001. “Stress, Social Support, and Coping as Predictors of Depressive Symptoms: Gender Differences among Mexican Americans.” Social Work Research 25(1):37–48.

Black

Sandra A.

Markides

Kyriakos S.

Miller

Todd Q.

1998. “Correlates of Depressive Symptomatology among Older Community Dwelling Mexican Americans: The Hispanic EPESE.” The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences 53(4):S198–208.

Blazer

Dan G.

2003. “Depression in Late Life: Review and Commentary.” The Journals of Gerontology, Series A: Biological Sciences & Medical Sciences 58(3):M249–65.

Burr

Jeffrey A.

Mutchler

Jan E.

Gerst-Emerson

Kerstin

. 2013. “Residential Segregation, Nativity Status, and Hispanic Living Arrangements in Later Life.” Population Research and Policy Review 32:25–45.

10.

Cadge

Wendy

Ecklund

Elaine H.

2006. “Religious Service Attendance among Immigrants: Evidence from the New Immigrant Survey-pilot.” American Behavioral Scientist 49(11):1574–96.

11.

Carliner

Geoffrey

. 2000. “The Language Ability of U.S. Immigrants: Assimilation and Cohort Effects.” International Migration Review 34(1):158–82.

12.

Carr

Deborah

Springer

Kristen W.

2010. “Advances in Families and Health Research in the 21st Century.” Journal of Marriage and Family 72(3):743–61.

13.

Choi

Namkee G.

Bohman

Thomas M.

2007. “Predicting the Changes in Depressive Symptomatology in Later Life: How Much Do Changes in Health Status, Marital and Caregiving Status, Work and Volunteering, and Health-related Behaviors Contribute?” Journal of Aging and Health 19(1):152–77.

14.

Crystal

Stephen

Shea

Dennis G.

2002. “Introduction: Cumulative Advantage, Public Policy and Late Life Inequality.” Annual Review of Gerontology and Geriatrics 22(1):1–13.

15.

Cuellar

Israel

Bastida

Elena

Braccio

Sara Maria

. 2004. “Residency in the United States, Subjective Well-being, and Depression in an Older Mexican-origin Sample.” Journal of Aging and Health 16(4):447–66.

16.

Djernes

Jens K.

2006. “Prevalence and Predictors of Depression in Populations of Elderly: A Review.” Acta Psychiatrica Scandinavica 113(5):372–87.

17.

Donato

Katharine M.

2010. “U.S. Migration from Latin America: Gendered Patterns and Shifts.” International Migration Review 630:78–92.

18.

DuBard

C. Annette

Gizlice

Ziya

. 2008. “Language Spoken and Differences in Health Status, Access to Care, and Receipt of Preventive Services among US Hispanics.” American Journal of Public Health 98(11):2021–28.

19.

Durand

Jorge

Massey

Douglas S.

Zenteno

René M.

2001. “Mexican Immigration to the United States: Continuities and Changes.” Latin American Research Review 36(1):107–27.

20.

Ennis

Sharon R.

Ríos-Vargas

Merarys

Albert

Nora G.

2011. “The Hispanic Population: 2010.” 2010 Census Briefs. U.S. Census Bureau. Retrieved September 26, 2015 (http://www.census.gov/prod/cen2010/briefs/c2010br-04.pdf).

21.

Ferraro

Kenneth F.

Farmer

Melissa M.

1996. “Double Jeopardy, Aging as Leveler, or Persistent Health Inequality? A Longitudinal Analysis of White and Black Americans.” The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences 51(6):S319–28.

22.

Finch

Brian K.

Vega

William A.

2003. “Acculturation Stress, Social Support, and Self-rated Health among Latinos in California.” Journal of Immigrant Health 5(3):109–17.

23.

Fuligni

Andrew J.

2004. “Convergence and Divergence in the Developmental Contexts of Immigrants to the United States.” Pp. 89–98 in Historical Influences on Lives and Aging, edited by Schaie

Elder

New York: Springer.

24.

Gallo

Linda C.

Penedo

Frank J.

de los Monteros

Karla Espinosa

Arguelles

William

. 2009. “Resiliency in the Face of Disadvantage: Do Hispanic Cultural Characteristics Protect Health Outcomes?” Journal of Personality 77(6):1707–46.

25.

Garcia

Marc A.

Reyes

Adriana M.

2017. “Physical Functioning and Disability Trajectories by Age of Migration among Mexican Elders in the United States.” The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences. Published electronically January 4. doi:10.1093/geronb/gbw167.

26.

Gerst

Kerstin

Al-Ghatrif

Majd

Beard

Holly A.

Samper-Ternent

Rafael

Markides

Kyriakos S.

2010. “High Depressive Symptomatology among Older Community-dwelling Mexican Americans: The Impact of Immigration.” Aging & Mental Health 14(3):347–54.

27.

Gong

Fang

Jun

Fujishiro

Kaori

Takeuchi

David T.

2011. “A Life Course Perspective on Migration and Mental Health among Asian Immigrants: The Role of Human Agency.” Social Science & Medicine 73(11):1618–26.

28.

Gonzales

Alicia M.

2007. “Determinants of Parent-child Coresidence among Older Mexican Parents: The Salience of Cultural Values.” Sociological Perspectives 50(4):561–77.

29.

González

Hector M.

Haan

Mary N.

Hinton

Ladson

. 2001. “Acculturation and the Prevalence of Depression in Older Mexican Americans: Baseline Results of the Sacramento Area Latino Study on Aging.” Journal of the American Geriatrics Society 49(7):948–53.

30.

Gonzalez-Barrera

Ana

Lopez

Mark H.

2013. “A Demographic Portrait of Mexican-origin Hispanics in the United States.” Retrieved September 26, 2015 (http://www.pewhispanic.org/2013/05/01/).

31.

Gubernskaya

Zoya

. 2014. “Age at Migration and Self-rated Health Trajectories after Age 50: Understanding the Older Immigrant Health Paradox.” The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences 70(2):279–90.

32.

Haan

Mary N.

Zeki Al-Hazzouri

Adina

Aiello

Allison E.

2011. “Life-span Socioeconomic Trajectory, Nativity, and Cognitive Aging in Mexican Americans: The Sacramento Area Latino Study on Aging.” The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences 66(S1):i102–10.

33.

Hansen

Thomas

Slagsvold

Britt

Moum

Torbjørn

. 2008. “Financial Satisfaction in Old Age: A Satisfaction Paradox or a Result of Accumulated Wealth?” Social Indicators Research 89(2):323–47.

34.

Heilemann

MarySue V.

Lee

Kathryn A.

Kury

Felix S.

2002. “Strengths and Vulnerabilities of Women of Mexican Descent in Relation to Depressive Symptoms.” Nursing Research 51(3):175–82.

35.

Hill

Terrence D.

Angel

Jacqueline L.

Balistreri

Kelly S.

2012. “Does the ‘Healthy Immigrant Effect’ Extend to Cognitive Aging?” Pp. 19–33 in Aging, Health, and Longevity in the Mexican-origin Population, edited by Angel

J. L.

Torres-Gil

Markides

K. S.

New York: Springer.

36.

Hill

Terrence D.

Angel

Jacqueline L.

Balistreri

Kelly S.

Herrera

Angelica P.

2012. “Immigration Status and Cognitive Functioning in Late Life: An Examination of Gender Variations in the Healthy Immigrant Effect.” Social Science & Medicine 75:2076–84.

37.

Hughes

Mary E.

Waite

Linda J.

2002. “Health in Household Context: Living Arrangements and Health in Later Middle Age.” Journal of Health and Social Behavior 43(1):1–21.

38.

Hunt

Linda M.

Schneider

Suzanne

Comer

Brendon

. 2004. “Should ‘Acculturation’ Be a Variable in Health Research? A Critical Review of Research on US Hispanics.” Social Science & Medicine 59:973–86.

39.

Kahn

Joan R.

Perlin

Leonard I.

2006. “Financial Strain over the Life Course and Health among Older Adults.” Journal of Health and Social Behavior 47(1):17–31.

40.

Kim

Giyeon

Worley

Courtney B.

Allen

Rebecca S.

Vinson

Latrice

Crowther

Martha R.

Parmelee

Patricia

Chiriboga

David A.

2011. “Vulnerability of Older Latino and Asian Immigrants with Limited English Proficiency.” Journal of the American Geriatrics Society 59(7):1246–52.

41.

Kim

Wooksoo

Chen

Ya-Ling

. 2011. “The Social Determinants of Depression in Elderly Korean Immigrants in Canada: Does Acculturation Matter?” The International Journal of Aging & Human Development 73(4):283–98.

42.

Ladin

Keren

Reinhold

Steffen

. 2013. “Mental Health of Aging Immigrants and Native-born Men across 11 European Countries.” The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences 68(2):298–309.

43.

Lam

Jonathan

Yip

Tiffany

Gee

Gilbert

. 2012. “The Physical and Mental Health Effects of Age of Immigration, Age, and Perceived Difference in Social Status among First Generation Asian Americans.” Asian American Journal of Psychology 3(1):29–43.

44.

Lara

Marielena

Gamboa

Cristina

Kahramanian

M. Iya

Morales

Leo S.

Bautista

David E. H.

2005. “Acculturation and Latino Health in the United States: A Review of the Literature and Its Sociopolitical Context.” Annual Review of Public Health 26:367–97.

45.

Law

Rita W.

Sbarra

David A.

2009. “The Effects of Church Attendance and Marital Status on the Longitudinal Trajectories of Depressed Mood among Older Adults.” Journal of Aging and Health 21(6):803–23.

46.

Leu

Janxin

Yen

Irene H.

Gansky

Stuart A.

Walton

Emily

Adler

Nancy E.

Takeuchi

David T.

2008. “The Association between Subjective Social Status and Mental Health among Asian Immigrants: Investigating the Influence of Age at Immigration.” Social Science & Medicine 66(5):1152–64.

47.

Lum

Terry Y.

Vanderaa

Julianne P.

2010. “Health Disparities among Immigrant and Non-immigrant Elders: The Association of Acculturation and Education.” Journal of Immigrant and Minority Health 12(5):743–53.

48.

Markides

Kyriakos S.

Eschbach

Karl

Ray

Laura A.

Peek

M. Kristen

. 2007. “Census Disability Rates among Older People by Race/Ethnicity and Type of Hispanic Origin.” Pp. 40–49 in The Health of Aging Hispanics: The Mexican Origin Population, edited by Angel

J. L.

Whitfield

K. E.

New York: Springer.

49.

Markides

Kyriakos S.

Stroup-Benham

Christine A.

Goodwin

James S.

Perkowski

Linda C.

Lichtenstein

Michael J.

Ray

Laura A.

1996. “The Effect of Medical Conditions on the Functional Limitations of Mexican-American Elderly.” Annals of Epidemiology 6(5):386–91.

50.

Parrado

Emilio A.

Flippen

Chenoa A.

McQuiston

Chris

. 2005. “Migration and Relationship Power among Mexican Women.” Demography 42:347–72.

51.

Pew Hispanic Center. 2008. “U.S. Population Projections: 2005–2050.” Retrieved September 26, 2015 (http://pewhispanic.org/files/reports/85.pdf).

52.

Radloff

Lenore S.

1977. “The CES-D Scale: A Self-report Depression Scale for Research in the General Population.” Applied Psychological Measurement 1(3):385–401.

53.

Riosmena

Fernando

Wong

Rebeca

Palloni

Alberton

. 2013. “Migration Selection, Protection, and Acculturation in Health: A Binational Perspective on Older Adults.” Demography 50:1039–64.

54.

Russell

David

Taylor

John

. 2009. “Living Alone and Depressive Symptoms: The Influence of Gender, Physical Disability, and Social Support among Hispanic and Non-Hispanic Older Adults.” The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences 64(1):95–104.

55.

Schaan

Barbara

. 2013. “Widowhood and Depression among Older Europeans—The Role of Gender, Caregiving, Marital Quality, and Regional Context.” The Journals of Gerontology, Series B: Psychological Sciences & Social Sciences 68(3):431–42.

56.

Singer

Judith D.

Willett

John B.

2003. Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence. New York: Oxford University Press.

57.

Stykes

Bart

Payne

Krista K.

Gibbs

Larry

. 2014. “First Marriage Rate in the U.S., 2012.” Retrieved September 26, 2015 (http://www.bgsu.edu/content/dam/BGSU/college-of-arts-and-sciences/NCFMR/documents/FP/FP-14-08-marriage-rate-2012.pdf).

58.

Torres

Lucas

. 2010. “Predicting Levels of Latino Depression: Acculturation, Acculturative Stress, and Coping.” Cultural Diversity & Ethnic Minority Psychology 16(2):256–63.

59.

Wakabayashi

Chizuko

. 2010. “Effects of Immigration and Age on Health of Older People in the United States.” Journal of Applied Gerontology 29(6):697–719.

60.

Wassertheil-Smoller

Sylvia

Arredondo

Elva M.

Cai

JianWen

Castaneda

Sheila F.

Choca

James P.

Gallo

Linda C.

Jung

Molly

LaVange

Lisa M.

Lee-Rey

Elizabeth T.

Mosley

Thomas

Jr. Penedo

Frank J.

Santistaban

Daniel A.

Zee

Phyllis C.

2014. “Depression, Anxiety, Antidepressant Use, and Cardiovascular Disease among Hispanic Men and Women of Different National Backgrounds: Results from the Hispanic Community Health Study/Study of Latinos.” Annals of Epidemiology 24(11):822–30.

61.

Zheng

Schimmele

Christoph M.

Chappell

Neena L.

2012. “Aging and Late-life Depression.” Journal of Aging and Health 24(1):3–28.