Transition in Older Parent–Adult Child Relations in U.S. Chinese Immigrant Families

Abstract

Background and Objectives

The family fundamentally underpins the immigration and acculturation processes. But most existing research on acculturation focuses on individual instead of family experience. Guided by Berry’s acculturation theory and Intergenerational Solidarity Theory, this study examined continuity and changes in parent–adult child relations of older Chinese immigrants over a 2-year period, and their implications for older adults’ depression and quality of life (QoL).

Research Design and Methods

Participants included 2,605 older Chinese immigrants from the Population Study of Chinese Elderly in Chicago (PINE). Latent Transition Analysis (LTA) was used to identify transitions in multidimensional parent–child relations over time. Negative binomial and logistic regressions were used to examine the influence of family transitions on depression and QoL, respectively.

Results

LTA revealed five types of family relations: traditional, modified traditional, coresiding-unobligated, independent, and detached. Over 40% of the respondents shifted to a different relation type, with more families classified as modified traditional or independent over time. Transitioning into modified traditional relations or out of detached relations was associated with fewer depressive symptoms and better QoL at the follow-up. Transitioning into independent relation was associated with more depressive symptoms over time.

Discussion and Implications

Parent–child relations among Chinese older immigrants demonstrate significant complexity, including both heterogeneity and fluidity. Better well-being of these older adults seems to stem from the optimal combination of retaining the supportive heritage culture and embracing the host society’s instrumental cultural elements. Services to this population need to include the family context in assessment and interventions.

Immigration is a life-changing event which often requires years of adaptation (Elder & Johnson, 2003). Along the adaptation process, immigrants often face elevated stress and cumulative mental health risks due to factors such as language barriers, homeland estrangement, and social isolation (Alegría, Álvarez, & DiMarzio, 2017). Adaption upon immigration also involves the entire family. Most studies on family adaptation of immigrants have focused on immigrant families with children, adolescents, or college students (Glick, 2010). But the topic is important for older adults for several reasons. First, older immigrants’ lives are often centered around their families for cultural and practical reasons, making family relationships crucial to their well-being (Treas & Batalova, 2009). Second, acculturation may be particularly challenging for older immigrants because of their longer exposure to the heritage culture and fewer opportunities for acculturation than younger immigrants (Batalova, 2012). Third, family relations are the key to the question of who cares for senior citizens. Disruptive family relations may mean greater needs of formal support for the aging immigrant populations. Understanding how older immigrants achieve successful adaptation in the family setting is critical to enhance the well-being of this rapidly increasing population.

Among older immigrants, much attention has been devoted to examining their depression (Diwan, Jonnalagadda, & Balaswamy, 2004; Guo & Stensland, 2018). As a leading cause of morbidity in later life, depression is associated with decreased physical, cognitive and social functioning, greater self-neglect, and increased mobility (Fiske, Wetherell, & Gatz, 2009). However, focusing on depression alone is not sufficient to understand how older immigrants sustain or generate positive affect under stressful situations (Diwan et al., 2004). Positive evaluation of life, captured by high rating of quality of life (QoL) highlights strengths of older immigrants and sheds light on the way through which this vulnerable population remains resilient in the host society (Mui & Shibusawa, 2008). Thus, when examining successful adaption of older immigrants, it is important to examine both the absence of negative outcomes such as depression and the high end of QoL.

Recognizing the vital role of the family for older immigrants and lack of studies on their family transitions, this study aims to answer questions about aging experience of immigrants in the family context: What transitions occur in older parent–adult child relationships in immigrant families? What factors contribute to the transitions? How are these family transitions related to older immigrants’ depression and QoL?

To address these questions, this study used data from the Population Study of Chinese Elderly in Chicago (i.e., PINE), the largest population-based survey on Chinese older adults in the United States. Chinese older immigrants are the largest and fastest growing Asian American group who face greater mental health risks than the general older population (Chang, Beck, Simon, & Dong, 2014). Chinese immigrants are characterized by a strong sense of family identity and interdependence among family members (Guo, Xu, Liu, Mao, & Chi, 2016). Whereas national data are not available, in the PINE baseline, about 85% the respondents migrated for family reasons; more than 60% migrated at 50 years or older; and about a third lived with their children. Given their sheer size, relatively high mental health risks, and traditional and interdependent family relations, older Chinese immigrants present an interesting group for examinations of family adaptation and individual well-being in the immigration context.

Heterogeneity in Individual Acculturation Experience

Describing how acculturation occurs to individuals, Berry’s (1997) posits four acculturation “strategies” derived from two bipolar attitudinal dimensions—heritage-culture retention and receiving-culture acquisition. Separation occurs when individuals strive to maintain their original culture without having interaction with others from a culturally different background. Integration occurs when one’s original culture is maintained while also actively participating in the larger social network. Assimilation occurs when individuals do not seek to maintain their cultural identity while interacting heavily with other cultures. Lastly, marginalization occurs when there is little maintenance of the original cultural and little interaction with others due to issues such as exclusion and discrimination.

Berry’s framework has been widely used to study acculturation of individuals. But an individuals’ acculturating experiences are significantly affected by their family members, who play an important role in immigrants’ adaptive outcomes by providing social support, developing a sense of cohesion, and transmitting cultural values (Ward, Fox, Wilson, Stuart, & Kus, 2010). Berry’s framework could potentially speak to the experience of acculturating families. Similar to an acculturating individual, an acculturating family as a unit is also confronted with different sets of norms and cultural expectations regarding family interactions in the host society. The process of acculturation often requires fundamental changes in the functioning of the family (Ward et al., 2010). To fully understand immigrants’ acculturation process and evaluate Berry’s framework, it is important to extend the studies from individual immigrants to the complex and reciprocal family relationships in their families (Berry, Phinney, Sam, & Vedder, 2006).

Intergenerational Solidarity Theory describes family relations as multidimensional, including structural (living arrangement), associational (frequency of contact), functional (support exchange), affectual (cohesion, conflict), and normative solidarity (family norms) (Bengtson & Roberts, 1991). Among them, the normative aspect is most relevant to the concept of cultural retention or assimilation. It helps to regulate other aspects of parent–child relations such as desired geographic distance between generations, frequency of contact, helping behavior and financial exchanges. If Berry’s framework holds true for Chinese immigrants’ family experience, one would expect to observe different patterns of family interactions among these families, such as traditional tight-knit family structure that emphasizes children’s filial obligation and interdependence (separation), family structure that embraces some aspects of Chinese traditional family norms but also individualistic Western family norms (integration), forfeit of traditional Chinese family norms and adoption of relatively independent family norms in the West (assimilation), or loss of traditional tight-knit family structure and also lack of endorsement of Western family norms (marginalization). Recent studies have used typology approaches such as latent class analysis to test Berry’s model and identified distinctive paths of acculturation process among individual immigrants (e.g., Fox, Merz, Solórzano, & Roesch, 2013). These findings are yet to be validated regarding family acculturation process.

Acculturating Families

Most studies on aging immigrant families examined single domains of intergenerational relations and used cross-sectional data. These studies provide evidence of immigrant families’ strong adherence to the heritage culture, in the forms of high prevalence of coresidence, close-knit relations, and frequent support exchanges (Guo et al., 2016; Gurak & Kritz, 2010; Wong, Yoo, & Stewart, 2005). Rooyackers, De Valk, and Merz (2014) reported that family patterns that represent non-Western traditional family relations were more prevalent in immigrant than native families in the Netherlands.

Studies also documented changes or challenges experienced by aging immigrant families. For instance, Chinese older immigrants in the United States were reported to lower their filial expectations in response to children’s struggles in the host society (Lieber, Nihira, & Mink, 2004). From children’s perspective, Chinese immigrants formed relatively nonhierarchical relationships with their aging parents and became more accepting towards the idea of placing parents in senior housing (Hsueh, Hu, & Clarke-Ekong, 2008). Shedding light on the fluid nature of family cultural norms, Sharma and Kemp (2011) observed a continuum of filial support of Indian immigrants in the United States, ranging from traditional, modified, to nontraditional support, illustrating diversity and fluidity of filial practice among immigrants. Similarly, in a latent class analysis, (Guo, Stensland, Li, & Dong, 2019) used the baseline data of the PINE and identified four types of family relations among U.S. Chinese older immigrants, with the tight-knit resembling the most traditional Chinese family structure, and the other three (unobligated ambivalent, detached, commanding conflicted) reflecting variant forms of family adaptation to the Western norms with some conflict or detachment in each type. This study also found that older immigrants’ age, gender, marital status, health status, and level of acculturation were related to their relation types. Together, the above findings suggest that over time, aging immigrant families might be less likely to endorse traditional family structures and more likely to embrace family structures that resemble integrated or assimilated acculturation strategies. Despite these interesting findings, there is an urgent need for longitudinal studies, which are best suited to address issues of adaptation and adjustment.

Adaptation and Psychological Well-being

Acculturation is stress-laden. Mental disorders such as depression could occur “when acculturation experiences overwhelm the individual, creating problems that cannot be controlled or surmounted” (Berry, 1997, p.19). When individuals do not feel successful in acculturating, they may engage in withdrawal from the larger social network (separation) or shedding one’s culture without opportunity to engage in the new culture (marginalization). Severe grief may also occur when there is substantial cultural loss (assimilation). Any of these changes are potential risk factors of immigrants’ depression. Instead, the integration category, sometimes referred to as biculturalism, seems to reflect individuals’ better ability to consolidate competing cultural tenets, and has been found to be associated with lower depression and better QoL (Schwartz, Unger, Zamboanga, & Szapocznik, 2010).

Studies have yet to test this hypothesis in the family context, but longitudinal studies on older adults have reported that over time, departure from cohesive family relations such as infrequent contact, emotionally distant relations, and reduced support elevated depressive symptoms (Liu & Lou, 2017; Stafford, McMunn, Zaninotto, & Nazroo, 2011; Tosi & Grundy, 2019) and reducing QoL (Rafnsson, Shankar, & Steptoe, 2015; Van Dyck, Teychenne, McNaughton, De Bourdeaudhuij, & Salmon, 2015). Cumulative evidence from studies on older Chinese and Korean immigrants further showed that negative family relations such as low support and filial dissatisfaction seem to be more detrimental to older immigrants’ depression than the potential benefit of positive family relations (Guo & Stensland, 2018).

While most of these studies on older immigrants used small cross-sectional convenience samples, the present study used the two-wave panel data from the PINE to address the following questions:

R1. To what extent do parent–adult child relations change among Chinese older immigrants over a 2-year period?

R2. What factors are associated with changes in family relations?

R3. How are these changes related to the depression and QoL of older immigrants?

Design and Methods

Sample

The sample consisted of two-wave panel data of the Population Study of Chinese Elderly in Chicago (PINE), the largest population-based study on Chinese community-living older adults in the United States (Dong, Wong, & Simon, 2014). Between 2011 and 2013 (Wave 1), 3,157 Chinese aged 60+ years were recruited from more than 20 social service agencies, community centers, faith-based organizations, and senior apartments in the greater Chicago area. A follow-up survey was conducted 2 years later between 2013 and 2015 (Wave 2), with 2,709 respondents, or 86% of the original participants. Bilingual interviewers conducted in-home interviews with respondents. The working sample consisted of 2,605 individuals who had at least one child and provided valid information on family relation and depression and QoL at both waves. See Table 1 for sample characteristics.

Table 1.

Descriptive Statistics for Key Variables in the PINE Panel Data (N = 2,605)

Variables	Wave 1	Wave 2
	Mean/% (SD)	Mean/% (SD)
Sample Characteristics
Age	72.87 (8.28)	–
Female	58.03%	–
Married	71.23%	69.57%
Education	8.63	–
Income	1.94	1.14
Good/very good health	38.81%	32.21%
Years in the United States	19.86 (13.05)
Level of acculturation	15.10 (4.84)	15.16 (4.18)
Number of children	3.00 (1.41)	–
Depression	2.57 (4.01)	1.80 (3.70)
Good/very good QoL	50.44	55.29
Life Transitions
Newly widowed		2.23
Having new grandchild(ren)		18.54
Much worse self-rated health		13.79
Onset of new chronic condition(s)		41.31
Parent–Child Relations
Living with children	34.70	32.52
Having weekly contact	85.80	92.48
Receiving instrumental support	40.04	34.05
Caring for grandchildren	26.53	24.80
Feeling close to children	84.34	93.78
Having family conflict	13.86	11.13
Having high filial expectation	37.04	43.55
Receiving high filial piety	42.29	48.18

Note: QoL = Quality of life; SD = Standard deviation.

Measures

Older parent–adult child relations were reported by the respondents to indicate whether they (a) lived with children, (b) had weekly contact with children, (c) received help with Activity of Daily Living (ADL) or Instrumental Activity of Daily Living (IADL) from children, (d) provided childcare to their children, (e) having very close relationship with children, (f) having conflict with children (i.e., perceive children as either demanding or critical), (g) having high filial piety expectation, and (h) receiving high filial piety. The respondents assessed their relationship with each child on the first six items. If the answer was yes for any child, the variable was coded as one. The last two items were asked regarding children in general. To assess filial piety, respondents first reported how much filial piety they expect from children, and then how much they actually received from them on six domains: respect, care, checking on, bring happiness, obedience, and financial support (1 = very little, 2 = rather little, 3 = average, 4 = rather a lot, 5 = very much). Average scores were calculated (α ranged between .84 to.89 across the waves). If the score was four (rather a lot) or five (very much), the variable was recoded as 1 = having high filial expectation, or 1 = receiving high filial piety. The second panel of the Table 1 summarized the frequencies of the eight family relation indicators at both waves.

Depression was measured by Patient Health Questionnaire-9 (PHQ-9, American Psychiatric Association, 1994). Respondents rated how often they experienced nine depressive symptoms (loss of interest, feeling down/depressed/hopeless, changes in sleep, fatigue, changes in appetite, feelings worthless, trouble concentrating, feeling restless or slowed down, suicidal thoughts) in the past 2 weeks (0 = not at all, 1 = several days, 2 = a week or more, 3 = nearly every day). The sum score ranged from 0 to 27, with higher scores indicating more depressive symptoms (α = .82 and .84 at the two waves).

QoL was measured by a question “In general, how would you rate your quality of life?” (1 = poor, 2 = fair, 3 = good, 4 = very good). Given few respondents reported “poor” (6.85% and 9.27% at the two waves) or “very good” QoL (3.2% and 4.51% at the two waves), QoL was further recoded into a dummy variable (1 = good/very good, 0 = poor/fair).

Personal characteristics included age (in years), gender (1 = women, 0 = men), marital status (1 = married, 0 = others), education (in years), self-rated health (1 = good/very good, 0 = poor/fair), years in the United States, level of acculturation, and number of children. Level of acculturation was measured by 12 questions on respondents’ preference for speaking Chinese or English in different settings and preferred ethnicity (Chinese or American) of those they interact (1 = only Chinese to 5 = only English/Americans), capturing language, media use, and ethnic social relations that are relevant to this population (Dong, Bergren, & Chang, 2015). The acculturation sum score ranged from 12 to 60 (α = .91 for both waves), with a higher score indicating a higher level of acculturation.

Life transitions between Wave 1 and Wave 2 were captured by loss of spouse (2.23%), birth of grandchild(ren) (18.54%), reported much worse self-rated health (13.79%), and onset of new chronic condition(s)that included heart disease, stroke, cancer, diabetes, broken/fractured hip, osteoarthritis, thyroid disease, high cholesterol, hypertension (41.31%).

Data Analysis

Latent transition analysis (LTA) was used to assess changes in parent–child relations over time (R1). LTA is an extension of latent class analysis (LCA). Whereas LCA identifies unobserved subgroups (latent classes) within a population based on individuals’ responses to multiple observed variables, LTA uses longitudinal data to address additional questions such as: is there change between latent class over times? If so, how can the change be characterized (Collins & Lanza, 2010)? LTA yields three sets of parameters: (a) latent status (i.e., latent class in LCA) prevalences, which indicate the prevalence of each identified latent status at each time, and (b) item-response probabilities, which indicate the strength of relationship between an observed variable and a latent variable. For instance, a probability of .90 indicates that members in a latent status have a 90% chance of endorsing that particular observed variable. The overall patterns of item-response probabilities are used to interpret and label each latent status. LTA further estimates (c) transition probabilities, which indicate individuals’ probabilities of remaining in a particular latent status versus transitioning to other statuses across times.

The first step of LTA is to determine the number of latent statues. The decision is based on a combination of model fit, parsimony, class membership probabilities, and interpretability of latent statuses (Collins & Lanza, 2010). We relied on likelihood-ratio G² statistics, Akaike’s Information Criterion, and Bayesian Informal Criterion to compare models with different numbers of latent statues. Smaller values of the three statistics generally indicate better model fit. LTA were carried out in SAS using PROC LTA command. Measurement invariance across the time points was assumed (i.e., item-responsibilities are constrained to be equal across time) to help stabilize estimation and improve identification (Collins & Lanza, 2010).

Logistic regressions were carried out to predict changes in family relation types over time (R2), using personal characteristics at Wave 1 and life transition variables between the waves. Negative binomial (for depression) and logistic regressions (for QoL) were carried out to examine the associations between the family relation transitions and the two outcomes at Wave 2 (R3), controlling for Wave 1 personal characteristics, Wave 1 depression/QoL, and between-wave life transition variables. Negative binomial regressions were used to predict depression, because this variable is highly skewed (skewness > 2) and was, thus, treated as a count variable. Among the eligible respondents (N = 2,605), less than 70 (2.7%) had missing values on any variables examined in the study. Thus, listwise deletion was used for each analysis.

Results

Parent–Child Relation Types

Figure 1 shows the goodness-of-fit statistics of LTA models with two to six statuses. Although models with a higher number of statuses yielded overall better model fit, value reduction of all the statistics was more substantial up to the five-status model. We proceeded with the five-status model, which is more conceptually appealing and parsimonious than the six-status model.

Figure 1.

G ², AIC, and BIC for latent transition models of older parent–adult child relations in Chinese immigrant families (PINE Waves 1 and 2; N = 2,605). AIC = Akaike’s Information Criterion; BIC = Bayesian Informal Criterion.

Table 2 summarizes the results of the five-status model. To facilitate interpretation, we highlighted the item-response probabilities that centrally defined each status (i.e., having probabilities that were either much higher or lower than the sample average probabilities). These probabilities also informed the labeling of statuses. The first family relation type was labeled traditional, with the highest likelihood of endorsing all solidarity domains. The second type was named modified traditional, a similar pattern to the traditional except for lack of coresidence and lower likelihoods of support exchanges. The third type was labeled coresiding unobligated, given its very high likelihood of coresidence and contact, but also lack of filial obligations, and the highest likelihood of reporting conflict. The fourth type was labeled independent, with the lowest likelihood of coresidence, upward support from children, and endorsement of filial piety. The last type, named detached, was characterized by the lowest likelihoods of having contact, downward support, and close relationship, and low endorsement of the other relation items.

Table 2.

Five-Latent-Status Model of Older Parent–Adult Child Relationships Among Chinese Immigrants in the PINE Panel Data

	Traditional	Modified traditional	Coresiding unobligated	Independent	Detached
Item-response probabilities of a Yes response a
Living with children	.94	.02	.94 b	.00	.22
Having weekly contact	1.00	.98	.98	.90	.06
Receiving instrumental support	.46	.42	.39	.28	.34
Caring for grandchildren	.43	.18	.40	.19	.11
Feeling close to children	1.00	1.00	.90	.93	.05
Having family conflict	.10	.08	.20	.14	.14
Having high filial expectation	.72	.68	.15	.12	.33
Receiving high filial piety	.88	.88	.09	.06	.25
Latent status prevalences
Wave 1	.16	.25	.18	.30	.11
Wave 2	.19	.31	.14	.33	.03
	Latent Status at Wave 2
	Traditional	Modified traditional	Coresiding unobligated	Independent	Detached
Transition probabilities c Latent Status at Wave 1
Traditional	.68	.11	.17	.04	.00
Modified Traditional	.01	.66	.00	.31	.01
Coresiding Unobligated	.39	.00	.53	.07	.01
Independent	.01	.32	.03	.62	.03
Detached	.12	.29	.10	.36	.13

Note: ^aItem-response probabilities constrained equal across time. ^bItem-response probabilities in bold indicate defining characteristics of each latent class. ^cDiagonal transition probabilities in bold indicate no transition of relationship pattern between waves.

Transitions in Parent–Child Relation Types Over Time

The middle panel of the Table 2 reported prevalences of the five statuses at each wave. The most common and increasingly popular family types were modified traditional and independent, each accounting for about a third of the sample at Wave 2. The least common type was detached, whose prevalence rate dropped from 11% to 3% over time. There was a slight increase in the traditional type (16%–19%), and a slight decrease in the coresiding unobligated type (18%–14%) over the 2 years.

The bottom of the Table 2 summarizes transition probabilities, with the rows corresponding to the earlier time and the columns corresponding to the later time. These probabilities describe the incidence of transitioning to the column latent status, conditional on earlier membership in the row latent status (Collins & Lanza, 2010). The diagonal elements reflect the probability of membership in the same latent status at both waves. According to these numbers, the most stable relation type was traditional, with about two thirds (68%) of such relations remained stable after 2 years. Modified traditional and independent were also relatively stable, with 66% and 62%, respectively, remained stable in 2 years. Slightly more than half (53%) of the coresiding unobligated relation remained stable over time. The detached relation was the most apt to change, with only 13% remaining in this type at Wave 2. Considering these numbers and sample size of each family type at Wave 1, it is calculated that about 43% of the families transitioned into another relation type in 2 years.

The off-diagonal elements present the probability of transitioning from one latent status at Wave 1 to another at Wave 2. It can be seen that 11% and 17% of those who started with traditional relations changed to coresiding unobligated and modified traditional, respectively. For changes occurred to the modified traditional, coresiding unobligated, and independent relations, the most common destinations were independent (31%), traditional (39%), and modified traditional relations (32%), respectively. Changes to the detached relations were more divergent, with 12% changing to traditional, 29% to modified traditional, 10% to coresiding unobligated, and 36% to independent relations, respectively.

Predictors of Parent–Child Relation Transitions

Since it is not feasible to examine all 20 possible transitions involving changes from one family type to another, we focused the analyses to two transitions that yielded the most common family relations at Wave 2 (i.e., modified traditional and independent) and the transition out of the most unstable family type (i.e., detached). The first two analyses used the entire sample, with the two outcomes being 1 = transitioned from any other relation type to modified traditional or independent relations, respectively (0 = no change in relation type over time). The third analysis used a subsample of individuals who were classified as having detached relations at Wave 1 (n = 280), with outcome being 1 = transitioned from detached relation to any other relation type over time (0 = remained in detached relation at Wave 2).

Results in the Table 3 showed that older adults who were older (odds ratio [OR] = 1.03, confidence interval [CI]: 1.01–1.05, p < .001), female (OR = 1.28, CI: 1.00–1.64, p = .049), married (OR = 1.63, CI: 1.22–2.18, p = .001), and less educated (OR = 0.95, CI: 0.92–0.97, p < .001) were more likely to transition into modified traditional relations over time instead of remaining in the same relation type. In addition, those who had more education (OR = 1.04, CI: 1.01–1.07, p = .003) and who reported no new chronic conditions between the waves (OR = 0.74, CI: 0.59–0.93, p = .010) were more likely to transition into the independent relation. Lastly, older adults who were married (OR = 3.15, CI: 1.06–9.38, p = .039), with better health (OR = 2.79, CI: 1.08–7.25, p = .010) and more children (OR = 1.97, CI: 1.23–3.17, p = .049) were more likely to transition out of detached relations instead of remaining in that type of relation.

Table 3.

Summary of Logistic Regression Analysis Predicting the Changes in Parent–Child Relation Types Among Chinese Older Immigrants in the PINE Panel Data

	Into		Into		Out of
	Modified traditional		Independent		Detached
	β	95% CI	OR	95% CI	β	95% CI
Age	1.03***	1.01–1.05	1.00	0.98–1.01	0.97	0.91–1.03
Mothers (ref = fathers)	1.28*	1.00–1.64	.94	0.73–1.19	2.18	0.76–6.25
Married	1.63**	1.22–2.18	.85	0.64–1.12	3.15*	1.06–9.38
Education	0.95***	0.92–0.97	1.04**	1.01–1.07	0.97	0.88–1.07
Good/very good health	1.00	0.80–1.26	1.06	0.85–1.33	2.79*	1.08–7.25
Years in the United States	1.00	0.99–1.01	1.00	0.99–1.01	0.99	0.95–1.03
Level of acculturation	1.02	1.00–1.06	1.01	0.98–1.03	0.99	0.91–1.08
Number of children	1.01	0.93–1.11	1.01	0.92–1.10	1.97**	1.23–3.17
Newly widowed	1.25	0.64–2.47	1.14	0.55–2.38	1.07	0.09–13.37
Having new grandchild(ren)	0.87	0.65–1.17	0.76	0.56–1.04	a
Much worse self-rated health	0.79	0.57–1.11	1.27	0.93–1.74	0.47	0.16–1.41
New chronic condition(s)	1.15	0.92–1.44	0.74*	0.59–0.93	2.25	0.88–5.73
Pseudo r2	.57		.57		.47

Note: CI = Confidence interval; OR = Odds ratio.

^aEffect inestimable due to empty cells of “having new grandchild(ren)” among the “out of detached” group. The variable was thus removed from the analysis to enhance the robustness of the estimation.

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

Parent–Child Relation Transitions and Depression and QoL

The regression analysis results in the Table 4 showed that controlling for baseline characteristics and life transition variables, older adults transitioning into modified traditional relations had less depressive symptoms (β = −0.41, p = .002) and a higher chance of reporting good QoL (OR = 1.29, CI: 1.03–1.64, p = .026) at Wave 2 than those who remained in the same family relation type over time. Instead, transitioning into independent relation was associated with more depressive symptoms over time (β = 0.45, p < .001). Transitioning out of detached relation was associated with less depressive symptoms (β = −1.10, p = .028) and a greater chance of having good QoL (OR = 2.82, CI: 1.23–3.17, p = .046) in 2 years.

Table 4.

Summary of Regression on Wave 2 Depression and Quality of Life Among Chinese Older Immigrants in the PINE Panel Data

	Depression			Quality of life
	β	β	β	OR	95% CI	OR	95% CI	OR	95% CI
Family relation transitions
Into Modified Traditionala	−0.41**			1.29*	1.03–1.64
Into Independenta		0.45***				1.06	0.84–1.33
Out of Detachedb			−1.10*					2.82*	1.02–8.13
Characteristics at Wave 1
Age	0.02**	0.01*	0.02	1.01	1.00–1.02	1.01*	1.00–1.03	1.03	0.99–1.07
Mothers (ref = fathers)	0.19	0.18	0.50	1.28**	1.07–1.53	1.29**	1.08–1.54	0.83	0.48–1.45
Married	−0.13	−0.15	−0.03	0.87	0.70–1.07	0.88	0.72–1.09	0.94	0.46–1.92
Education	0.01	0.00	0.00	1.04**	1.02–1.06	1.04***	1.02–1.06	0.98	0.92–1.04
Good/very good health	−0.43***	−0.44***	−0.92**	1.36**	1.14–1.62	1.36**	1.14–1.63	1.35	0.73–2.51
Years in the United States	0.00	0.00	−0.01	1.00	1.00–1.01	1.00	1.00–1.01	1.00	0.97–1.02
Level of acculturation	−0.03**	−0.03**	0.00	1.03*	1.01–1.05	1.03*	1.01–1.06	1.06	1.00–1.13
Total number of children	−0.03	−0.04	−0.04	1.07	1.00–1.14	1.07	1.00–1.14	0.99	0.81–1.21
Depression	0.13***	0.13***	0.06*
Quality of Life				1.46***	1.23–1.74	1.44***	1.21–1.71	1.110	0.58–2.08
Life translations
Newly widowed	0.23	0.22	0.11	0.66	0.37–1.16	0.66	0.38–1.17	0.26	0.04–1.55
Having new grandchild(ren)	−0.01	−0.01	0.24	1.06	0.86–1.31	1.06	0.86–1.31	0.99	0.48–2.00
Much worse self-rated health	0.74***	0.75***	0.42	0.53***	0.42–0.67	0.53***	0.41–0.67	0.29**	0.13–0.66
New chronic condition(s)	0.44***	0.45***	0.14	0.76***	0.65–0.90	0.77***	0.65–0.90	0.84	0.50–1.41
χ 2/df	0.99	1.01	0.94
Pseudo r2				.75		.75		.75

Note: CI = Confidence intervals; OR = Odds ratio.

^aReference group is no change in family relation type over time. ^bReference group remained in detached relations at Wave 2.

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

Discussion and Implications

The family is a key context for the successful adaptation of older immigrants. But both theories and empirical work have overlooked the divergent ways in which family relations of older immigrants evolve along the acculturation process and the mental health implications of such family transitions. Using longitudinal data from a large sample of Chinese older immigrants in the United States, this study revealed five distinct parent–child relation patterns that best captured family dynamics over a 2-year period. In our earlier study using the baseline data from the same survey, only four types of parent–child relations were identified (Guo, Stensland, Li, & Dong, 2019). In addition, in the present study, over 40% of the families shifted from one relational type to another over time. Together, these findings seem to suggest that even over a relatively short timespan, there are both qualitative (different types emerged) and quantitative changes (different prevalence rate of each type) in immigrant family relations, indicating that family interactions in immigrant families are not only divergent, but also fluid and dynamic.

To some extent, the five parent–child relation types identified mirror the four acculturation strategies/outcomes described by Berry (1997). Arguably, the traditional type reflects the separation status as older immigrants strongly adhere to the original culture and family norms. Modified traditional and coresiding unobligated types represent two forms of integration status, with the former forfeiting the traditional coresiding living arrangement but maintaining traditional family setup in other relational domains, and the latter retaining the traditional living arrangement but with weakened filial obligations. The independent family type may reflect the assimilation status of Berry’s model, as these older immigrants seemingly embrace the relatively independent and individualistic types of family interactions in the United States. The detached relations may represent the marginalization status as family members are essentially unconnected from one another.

Berry’s acculturation framework also helps to contextualize the most common changes in parent–child relations observed in the study, which were shifting into modified traditional or independent relations, suggesting that over time, more older immigrants may achieve the integrated or assimilated family status in the host society. Interesting, there was also a slight increase in the percentage of the most traditional family relation type (i.e., traditional), providing evidence of strong adherence to the traditional family norms among Chinese immigrants. Another important finding of this study is that very few and increasingly less immigrant families fell into the least cohesive and most marginalized relational type (i.e., detached). Shifting away from detached relations may reflect an effort to stop rejecting the traditional familial setup or the Western family norms and beginning to feel more accepted in the host society.

Examining factors contributing to these changes, we found that over time, individuals of older age, women, and married older adults were more likely to transition into modified traditional relations, a highly cohesive family structure without coresidence. The findings are consistent with previous literature which reported that these factors are also associated with closer intergenerational relations (Lin, Bryant, Boldero, & Dow, 2015). Better-educated older immigrants were more likely to transition into independent relations over time, possibly indicating their greater sense of independence and control in the host society. Physical health is another important personal resource that may influence family transitions in the similar way. In our study, older immigrants with better physical health at the baseline and no new chronic condition onset between waves tended to shift into more independent types of family relations (independent and detached). Last but not least, having more children seems to prevent older adults from transitioning into the marginalized family status and having detached relations.

What are the implication of these family transitions on depression and QoL? In our study, an integrated model of family interaction—modified traditional relations (cohesive relations without coresidence)—stood out as having the most optimal well-being outcomes. In contrast, shifting into independent relations that reflect assimilation was related to increased depression longitudinally. Not reported in the study, having traditional family relations at Wave 1 or shifting into it over time was not related to Wave 2 depression or QoL. Taken together, these findings suggest that favorable well-being of older immigrants stems from the optimal combination of retraining the supportive and ego-reinforcing heritage culture (e.g., filial piety) and learning the host society’s instrumental cultural elements (e.g., independent living). Referring to it as “biculturalism paradox,” de Figueiredo (2014) explains that overcoming acculturative stress pushes individuals to seek out and achieve cultural mandates from both cultures; when this is successfully achieved, individuals feel a deep sense of meaning and are able to remove feelings of incompetency, thereby promoting well-being. Intentionally choosing to retain heritage culture and learning new cultural norms likely provides a sense of control and meaning among older immigrants, which in turn, promotes well-being. Summarizing in de Figueiredo’ words “...borrowing values and norms from both the native and the host cultures and applying them judiciously is more conducive to mental health than indiscriminately subscribing to either culture” (p. 176).

Also worth mentioning is that the majority of the respondents in the study (57%) had the same relation type over time, indicating great stability in older immigrants’ relations with their children. The implications of such stability on older immigrants’ well-being vary depending on the type of relations they stayed in. For instance, remaining in detached relations were not only associated with more depressive symptoms but also lower quality life than moving out of that problematic relations. But regardless of the type of parent–child relation an older immigrant has, remaining in such relations was less beneficial than moving into bicultural relations of the modified traditional type. In other words, comparing to either fully endorsing the traditional Chinese family structure or relatively independent family structure in the West, adhering to the traditional cultural norms coupled with independent living arrangement seems to be a happy medium for older immigrants’ well-being.

The study is not without limitations. On average, the respondents had lived in the United States for 20 years. Studying more recent immigrants or following immigrants from the pre- to postmigration stages would better illustrate family transitions at the most intensive encounter at two different cultures. The timespan between the waves was relatively short (2 years). Longer-term observations will yield better insight. Due to the nature of secondary analysis, some family relations measures were aggregate information from all the children in the family, possibly overlooking diversity among parent–child relations and its implications. Measures of children’s acculturation level and English fluency were not available either. Future research needs to further examine how family dynamics vary by children’s gender, marital status, proximity to parents, and their own acculturation status and their implications for older immigrants’ well-being. In addition, based on a sample of Chinese older adults from Chicago, the findings cannot be generalized to Chinese older immigrants in the United States or other immigrant populations. Future research needs to validate the findings using more diverse and representative samples of older immigrants.

Despite these limitations, this study is the first one using population-based, longitudinal data on multifaceted family relations to examine family adaptation of older immigrants, Our findings revealed divergent paths of immigrant families’ adaptation that reflect Berry’s four acculturation strategies/outcomes. Furthermore, the findings show that biculturalism in the sense of retaining one’s native norms and at the same time embracing the instrumental cultural elements in the receiving society, may result in the most optimal mental health of older immigrants.

Findings of this study have important implications for social service agencies serving ethnic minority aging populations. This study demonstrates the dynamic nature of acculturation experience of older immigrants within the family context. The awareness of family transitions may help sensitize gerontological practitioners not only to the older immigrants’ attitude and behaviors related to acculturation but also to the changing family environments in which they are embedded. The findings reveal the protective role of retaining original cultural norms while simultaneously adopting host society’s instrumental cultural elements (i.e., living arrangement) for older immigrants. Social service agencies should develop cultural literacy programs such as validation and facilitation of traditionalism in social services, English language classes, and cross-cultural engagement to improve older immigrants’ ability to understand and participate in the host culture while maintaining their heritage culture. Such programs shall be focused on older immigrants with less education, poorer physical health, and fewer children. Overall, the findings provide valuable insights into unobserved subgroups of older immigrants who face higher mental health risks due to family interaction patterns. The information is helpful to direct comprehensive assessment and tailed intervention to improve the mental health of this rapidly increasing population.

Funding

This work was supported by the National Institute on Aging (grant number R21AG055804).

Conflict of Interest

None reported.