Social Network Types, Living Alone, and Depressive Symptoms Among Older Korean Americans

Background and Purpose:

Although similar typologies of social networks have been identified across cultures, little is known about variations in social network patterns by living arrangement in older immigrants. The purposes of this study were to: (1) identify social network types in older Korean Americans; and (2) examine the main effects of demographic characteristics, region, living alone, and social network types and the interaction effects of living alone and social network types on depressive symptoms. We hypothesized that distinct social network types would be identified and that different groups would be differentially associated with depressive symptoms, directly and interactively with living alone.

Methods:

Data were drawn from a survey with older Korean Americans aged 60 or over in five geographic locations (California, New York, Texas, Hawaii, and Florida). Data were collected during 2017−2018, and a total of 2,140 were included in analyses. To identify social network types, latent profile analysis (LPA) was conducted using ten network-related criterion variables: religious meetings/activities, informal social meetings/activities, family (or friends) contact at least one a month, family (or friends) talking about private matters, family (or friends) feel close to ask for help, get help from the Korean community, and participation in community events. The characteristics of the identified groups were then examined using an ANOVA/χ² tests. Lastly, hierarchical multiple regression models of depressive symptoms were estimated to examine the direct and interactive effects of living alone and social network types in the sequential entry of blocks of covariates.

Results:

Based on several model selection criteria, a model with five social network types was identified as best-fit (BIC=56888.48, Entropy=.83, LMR-LRT= p<.001, BLRT= p<.001). The groups were labeled as “diverse” (13.8%, the most enriched type), “moderately diverse” (42.3%, overall enriched network), “family deficit” (22.2%, deficient in family-related networks), “friend deficit” (10%, deficient in friends-related networks), and “restricted” (11.6%, minimally engaged/disengaged network). In bivariate relationships, those who live alone belonged to diverse (9%), moderately diverse (36.4%), family deficit (27.7%), friend deficit (8.7%), and restricted (18.2%) groups. Results from direct effects models suggest that greater levels of depressive symptoms were generally associated with younger age, female, lower perceived financial status, poorer self-rated health, living in New York, and belonging to family deficit or restricted network group (the moderately diverse group was the reference group). A statistically significant interaction effect was found between living alone and membership in the restricted group (B=.08, p<.05), indicating that living alone was associated with significantly elevated depressive symptoms in the restricted group compared to the moderately diverse group.

Conclusions and Implications:

As hypothesized, distinct social network types were identified. Notably, older Korean Americans who lived alone were dispersed in all network types. In line with findings with other groups of older immigrants, belonging to family deficit or restricted groups was associated with poorer mental health. The interaction effect suggests that living alone and restricted group membership aggravate depressive symptoms. Using a typology approach, more attention and interventions should be focused on older immigrants who are both residentially and socially isolated.