Social Isolation, Race, and Gender: Differences between Urban and Rural Older Adults

Background: Social isolation is “the distancing of an individual psychologically or physically, or both, from his or her network of desired or needed relationships with other persons.” (Nicholson, 2016 p. 102).  Older adults who are socially isolated have an increased risk for hospitalization (Giuli, Papa, Mocchegiani, & Marcellini, 2012; Longman, Passey, Singer, & Morgan, 2013).  It negatively impacts quality of life (Hawton et al. 2011) and health (Boulos, Salameh, & Barberger-Gateau, 2017; Golden et al., 2009; Franck, Molyneux, & Parkinson, 2016; Shankar, Hamer, McMunn, & Steptoe, 2013), including all-cause mortality (Pantell, Rehkopf, Jutte, Syme, Balmes, & Adler, 2013; Tanskanen & Anttila, 2016).

This study seeks to use the recently developed Social Isolation Scale (SIS) (Nicholson, Feinn, Casey, & Dixon, n.d.) to measure the relationship between population density and social isolation, along with the relationship between gender and social isolation.  The SIS is a six-item scale that includes three questions that ask respondents to indicate frequency of contact with family, friends, and neighbors and three questions that ask respondents to rate agreement on questions about activities and relationships.

Method: The study used data collected by the AARP Foundation at selected Tax-Aide sites between March 7 and April 30, 2012.  The convenience sample totals 15,535 respondents aged 50 or older across 44 states in the US.  Using self-reported zip codes from that survey, as well as the SIS data, the researchers used a separate database containing zip codes to determine population density. Linear mixed models were used to test if population density was predictive of social isolation.  Linear mixed models were also used to test if there is a difference in social isolation by gender.  The study controlled for demographic variables including ethnicity, marital status, age, and employment status.

Results: After controlling for household income, gender, age, and marital status, there is a significant (p = .041) interaction between population density and ethnicity on social isolation. Follow-up analyses showed there was a significant (p = .005) relationship between population density and isolation in Whites, Hispanics, and others, in that higher population density was associated with greater social isolation.  This association was reversed for African American respondents but was not statistically significant.

Males were significantly more isolated than females on both the objective (Cohen’s d=0.21, p<.001) and subjective (d=0.16, p<.001) measures of isolation, as well as the overall measure of isolation (d=0.18, p<.001). After adjusting for demographic variables the difference between males and females increased for the objective measure (d=0.27, p<.001), the subjective measure (0.20, p<.001), and the overall measure of isolation (d=0.23).

Conclusions/Implications: Based on these findings, interventions to combat social isolation must be targeted by population density, race/ethnicity, and gender.  Further analyses looking at impact of demographics, such as race/ethnicity and gender, are needed to fully understand this phenomenon and appropriately tailor interventions based on race/ethnicity and gender.