Effects of Neighborhood-Level Variables on Loneliness and Social Isolation Among Older Adults in New Jersey

Background/ Purpose:

Older adults suffer greater risks of loneliness and social isolation due to loss of partners, friends, social roles, and mobility. In contrast to studies on individual and household-level factors, studies on neighborhood effects on loneliness and social isolation are relatively rare in the literature, and the limited results are often contradictory.

This study contributes to the current literature on neighborhood effects from two perspectives: 1) it compared subjective loneliness with objective social isolation, and examined the effects of neighborhood traits on both concepts; 2) it focused on the effects of neighborhood-level racial/ethnic diversity and income inequality on loneliness and isolation. Racial/ethnic diversity has been shown to have complex effects on health and mental health, but yet no study to date has examined its effects on loneliness and social isolation among older adults. The study also examined neighborhood income inequality, which has been widely studied in the context of research on social cohesion and health, and examined whether neighborhood effects on loneliness and social isolation differed for whites and non-whites.

Methods:

This study used data from a random sample of 3,701 older adults aged 60 and over living in New Jersey, a populous and ethnically diverse US northeastern state. The sample include 61.5% women, 89.0% whites, 6.2% African Americans, and 2.2% Hispanic or Latino. A weight variable was created to match the demographics of the state. The residential addresses of all individual were geocoded and linked to municipal data. The aggregated data at the municipal level are from 2015 American community Survey and multiple state sources.

Loneliness was measured using the 3-item abbreviated UCLA loneliness scale and social isolation was calculated as an aggregated score of 6 social networks items and 3 social participation items. Racial/ethnic diversity was calculated as the Simpson diversity index. Income inequality was measured using Gini index. The study also included individual and community control variables as suggested by the literature.

We used two-level multilevel models to predict loneliness and social isolation for whites and non-whites subgroups. In the models predicting loneliness, the objective social isolation was controlled to estimate only the psychological effects of neighborhood traits on loneliness. 

Results:

The effects of neighborhood-level variables on loneliness and social isolation among older adults differed for whites and non-whites. Among non-whites, a significant U-shape nonlinear relationship was found between diversity and social isolation, and a positive relationship was found between residential stability and social isolation. In contrast, among whites, a marginally significant positive effect was identified for diversity and significant negative effect for percentage of urban population on social isolation. For loneliness, significant positive effect for income inequality and marginally negative significant effect for median household income were identified among non-whites. For whites, significant U-shape nonlinear relationship between diversity and loneliness were identified.

Conclusions/Implications:

Our results reaffirmed the complicated relationships between neighborhood-level variables and important outcomes for older adults. This study provides insights for designing housing policies and community development strategies to improve social integration and wellbeing of older adults.