Skip to main content
NCBI home page
The Journals of Gerontology Series B: Psychological Sciences and Social Sciences logoLink to The Journals of Gerontology Series B: Psychological Sciences and Social Sciences
. 2020 Mar 30;76(4):790–800. doi: 10.1093/geronb/gbaa033

Local Ties in the Social Networks of Older Adults

Erin York Cornwell 1,✉, Alyssa W Goldman 1
Editor: Markus Schafer
PMCID: PMC7955986  PMID: 32227105

Abstract

Objectives

Family members and friends who live nearby may be especially well-positioned to provide social support and companionship for community-residing older adults, but prior research has not examined the distribution and characteristics of local ties in older adults’ networks. We hypothesize that local ties are newer, more frequently accessed, and more embedded in the network, and that social disadvantage and neighborhood conditions structure older adults’ access to local ties.

Methods

We use egocentric network data from 15,137 alters named by 3,735 older adults in Wave 3 of the National Social Life, Health, and Aging Project (NSHAP). We conduct dyadic analysis to compare characteristics of local and nonlocal ties. Logistic regression models estimate how personal and neighborhood characteristics are associated with naming local kin and local non-kin ties.

Results

Nearly half of the older adults named at least one local network tie, and about 60% of these local ties are non-kin. Local ties are newer, frequently accessed, and highly embedded in older adults’ networks. Local kin ties are most common among socially disadvantaged older adults. Local non-kin ties are most common among white older adults and those who live in areas with high levels of collective efficacy, although local non-kin ties are also associated with residence in high-poverty neighborhoods.

Discussion

Local ties may bring unique benefits for community-residing older adults, but their availability is likely structured by residential mobility, neighborhood context, disparities in resources, and support needs. Future research should consider their implications for health and well-being.

Keywords: Family, Neighborhood context, Social network, Social support

Sociological research emphasizes the role of spatial proximity in structuring social networks by providing opportunities to cultivate ties, mobilize social capital, and exchange support through face-to-face interaction within shared spaces (Small & Adler, 2019). Social gerontological research similarly emphasizes that residential proximity to adult children provides an opportunity structure for intergenerational solidarity and exchanges of support (Silverstein, Bengtson, & Lawton, 1997). Indeed, older adults have higher rates of contact with children who live within 25 km, compared to those who live father away (Hank, 2007). And, children and parents who reside within walking distance are more likely to provide a wide range of forms of help and support than those who live farther away (Logan & Spitze, 1994). Nearby adult children are also particularly well-positioned to help older adult cope with new challenges or losses. For example, widowed older adults who have at least one adult child residing within an hour’s drive report lower levels of psychological distress than those whose adult children live farther away (Ha & Carr, 2005), and proximity to an adult child is thought to reduce the likelihood that newly disabled older adults enter a nursing home or require formal care (Choi, Schoeni, Langa, & Heisler, 2015).

The benefits of spatial proximity likely extend beyond relationships with adult children; other family and friends who live nearby may be important sources of companionship and support. Living nearby provides opportunities for face-to-face interaction and a sense of common ground (Ikkink & van Tilburg, 1999). Close relationships with neighbors may be key sources of support and information about local resources, and may promote social integration within the neighborhood (Logan & Spitze, 1994). Having a close friend or family member nearby may be a critical factor continued independent residence, particularly for the growing share of seniors who are aging without a partner or child (Mair, 2019). However, the availability and relevance of local kin and local non-kin ties in older adults’ networks has been understudied in prior research.

We address this gap by examining the prevalence and characteristics of local kin and non-kin ties in older adults’ social networks. We draw from prior research on social integration in later life to develop a set of hypotheses about how relationships with local ties may differ from those that are more geographically distant. Then, we consider how access to local network ties may be structured by individual and neighborhood characteristics. We use new data from the third wave of the National Social Life, Health, and Aging Project (NSHAP), in which respondents indicated whether each of their network ties live within a mile (or about a 20-min walk) from their residence. We conclude by considering the implications of these findings for older adults’ access to resources provided by spatially proximate network ties.

The Value of Local Ties

Classic sociological theories of urbanism suggested that increases in transportation and communication technologies reduce geographic constraints on the formation of close relationships (Fischer, 1982; Logan & Spitze, 1996; Wirth, 1938), and some researchers today argue that the new communication technologies render geographic proximity irrelevant for social network structure (Cairncross, 2001). Today, about 67% of adults age 65+ use the internet (Anderson & Perrin, 2017) and over 40% use social networking sites such as Facebook (Smith & Anderson, 2018). The use of these new communication technologies enhances individuals’ ability to access companionship and support from geographically distant network ties (Hunsaker & Hargittai, 2018; Quan-Haase, Mo, & Wellman, 2017). However, Fischer (2011) suggests that face-to-face interaction, and the importance of local ties, have not declined over the past several decades.

We propose that, despite the advent of new communication technologies, spatially proximate ties are uniquely positioned to provide particular types of support and companionship for community-residing older adults for several reasons. First, network ties that live nearby can provide locally relevant support—information, advice, and resources that can assist older adults with practical, social, and emotional needs. For example, a local network member can provide information about local organizations, events, and activities, which may promote older adults’ social engagement within the broader neighborhood (Logan & Spitze, 1994).

Second, more frequent interaction allows local ties to provide companionship, recognize support needs, and provide just-in-time support. In addition, local ties can also provide practical support that requires physical copresence, such as personal care, transportation, and household help (Hank, 2007). Geographically distant ties could also provide these forms of support, but we argue that local ties are better positioned to do so. In fact, greater face-to-face contact and ready access to support often underlie decisions to relocate to be closer to adult children or other family members (Spring, Ackert, Crowder, & South, 2017).

Finally, geographic proximity may also enhance network members’ relationships with one another. Social networks that have a high level of density, or interconnectedness, have greater potential for network members to coordinate support or assistance during a crisis such as bereavement or illness (Hurlbert, Haines, & Beggs, 2000). Kin ties generally have a high level of embeddedness in the network, due to their relationships with other kin ties. Geographic proximity may be particularly important for interconnectedness of non-kin ties. Non-kin ties that live nearby have greater access to an individual in his or her home during longer stretches of time, affording opportunities to interact with coresidents and visitors.

If local network ties provide unique forms of companionship and support, individual preferences and strategic actions may drive network proximity. Prior research suggests that older adults cultivate particular types of ties to fulfill needs. For example, older adults who have few kin ties may seek friendship ties as an alternative source of social integration (Mair, 2019). And, later-life changes such as bereavement or health decline may lead older adults to narrow focus to their closest and most emotionally rewarding relationships (Charles & Carstensen, 2010; Wellman & Wortley, 1990). If local ties are particularly well positioned to provide companionship, intimacy, and support, then older adults who shift focus to their closest and most rewarding ties may, effectively, focus on their local ties. Network propinquity may also increase if older adults relocate to be closer to family (Silverstein & Angelelli, 1998).

Network Propinquity as Socially Structured

Social network composition is also socially structured. For example, prior research finds that African American and Latino older adults, as well as those with less education, tend to have smaller networks (Ajrouch, Antonucci, & Janevic, 2001; Cornwell, Laumann, & Schumm, 2008) and greater network instability (Cornwell, 2015). Mechanisms that link social disadvantage to restricted network size may be relevant for network proximity. For example, lower levels of education and unemployment limit opportunities to form ties with non-kin (Cornwell, 2015; Small, 2006), and they may also constrain contact with geographically dispersed ties. Older adults of higher socioeconomic status are more likely to spend time participating in organized community activities (Tang, 2008), which may generate more opportunities to form ties with local non-kin.

Other social disadvantages, including exposure to discrimination and institutional racism may restrict residential mobility and lead African Americans to maintain smaller networks of trusted others (Ajrouch et al., 2001), in some cases residing in more racially segregated neighborhoods to be closer to kin (Spring et al., 2017). Families of lower socioeconomic status are also more likely to have lived longer in the same area, which may lead to greater availability of local kin (Jacobs, Broese van Groenou, Aartsen, & Deeg, 2016; Spring et al., 2017).

Residential contexts may also contribute to differences in network propinquity. Although earlier theories on urbanism suggested that urban residence erodes social bonds (Logan & Spitze, 1996; Wirth, 1938), greater population density provides more potential nearby ties. Fischer (1982), for example, shows that urban dwellers have larger networks and more frequent contact with family and friends. The concentration of local institutions in urban areas, including retail establishments, senior centers, and churches, also provides opportunities for forming ties and interacting, especially with non-kin (Small & Adler, 2019; Torres, 2018).

Socioeconomically disadvantaged neighborhoods pose a number of challenges for older adults that may preclude the development of network ties (Cornwell & Behler, 2015). Qualitative research suggests that residents of disadvantaged neighborhoods may cultivate local ties to cope with the lack of local resources (Newman, 2003; Stack, 1974). However, a lack of local organizations and institutions in socioeconomically disadvantaged areas means less opportunities to meet and interact with neighbors (Sampson, 2012), and less residential stability makes it more difficult to form meaningful and long-lasting local ties (Schieman, 2005).

Concentrated poverty and residential instability are also associated with lower levels of social cohesion, trust, and support exchange among neighbors. Together, these contribute to lower levels of collective efficacy, or residents’ capacity to take prosocial action (Sampson, 2012). Neighborhoods with low collective efficacy are unlikely to facilitate the formation of local ties, and may instead precipitate social withdrawal, through associations with disorder, distrust, and fear (Krause, 1993; Ross, Mirowsky, & Pribesh, 2001).

Research Aims and Hypotheses

Shared contexts may enhance network members’ ability to provide companionship and support, making local ties particularly valuable for individuals who are aging independently in the community. However, prior research has not examined the availability of local ties in older adults’ networks, or how local ties differ from those that are more geographically dispersed. To address these gaps, we pursue three research aims.

First, we describe the overall prevalence of local ties in older adults’ social networks. Second, we compare the characteristics of local and nonlocal ties. We hypothesize that local ties are more frequently accessed (H1) and more embedded in older adults’ social networks (H2). Because distance presents hurdles to forming ties, we also hypothesize that local ties are newer relationships than nonlocal ties (H3). We expect these distinctions to be particularly pronounced within non-kin ties, since kin ties often carry normative obligations that encourage high levels of contact and support despite geographic dispersal (Silverstein et al., 1997; Wellman & Wortley, 1990).

Third, we consider how access to local ties is socially structured. We hypothesize that, compared to white older adults and those with greater socioeconomic resources, racial/ethnic minorities and socioeconomically disadvantaged older adults are more likely to have local kin ties and less likely to have local non-kin ties (H4). Characteristics of the local context may also structure access to local ties, particularly with non-kin. We hypothesize that older adults who reside in urban neighborhoods (with greater population density) are more likely to have local non-kin ties (H5). We also expect that older adults who live in areas with higher poverty and greater residential instability are less likely to name local non-kin ties (H6), while those who live in communities with higher levels of collective efficacy are more likely to name local non-kin ties (H7).

Method

We use data from the third wave of the National Social Life, Health, and Aging Project (NSHAP), a nationally representative, longitudinal survey of older Americans. To the best of our knowledge, the NSHAP is the only nationally representative study of older adults in the United States that includes information on network propinquity. The NSHAP has conducted three survey waves (Wave 1 in 2005–2006, Wave 2 in 2010–2011, and Wave 3 in 2015–2016). We restrict our analyses to Wave 3 because it introduced an indicator of network tie proximity. At Wave 3, NSHAP sought to reinterview all surviving Wave 2 respondents, as well as a new cohort born between 1948 and 1965 and their spouses or coresident partners. The survey included in-home interviews and a leave-behind questionnaire (LBQ) that respondents returned by mail. Altogether, 4,777 respondents were interviewed at Wave 3.

Network Tie Proximity and Tie Characteristics

Our main interest is the presence of “local” kin and non-kin ties within respondents’ personal (or “egocentric”) social networks. The Wave 3 social network roster was administered during the in-home interview. Respondents were asked to name up to five network members with whom they discussed “important matters” during the prior year. The respondent’s spouse or partner was added to the network if not initially named, so total network size ranges from 0 to 6 members (i.e., “alters”).

Recent research on this name generator has drawn attention to the potential for measurement error at the level of the interviewer (Fischer, 2009) and variation in the substantive importance of discussion topics (Bearman & Parigi, 2004; Small, 2013). However, the NSHAP name generator has been found to be particularly robust to potential sources of measurement error (Paik & Sanchagrin, 2013), and other studies point to its external validity through the observation of expected associations between network size and health and well-being (Liu & Waite, 2014; Schafer & Koltai, 2015; Cornwell & Waite, 2012).

Beginning in Wave 3, respondents were also asked, for each non-coresident alter: “Does this person live in your local area, that is, within a 20-min walk or about 1 mile of your home?” We use responses to this question to categorize non-coresident alters as “local” or “nonlocal.”

Respondents were asked about their relationship with each alter (e.g., spouse, child, friend), which we collapse to “kin” and “non-kin” ties. Frequency of interaction was assessed by asking respondents how often they talk with each alter, including talking on the phone, ranging from 1 “less than once a year” to 8 “every day.” The same question was used to assess the frequency with which each network member speaks with all other alters. For each network member, frequencies of interaction with all other alters were averaged to create an indicator (ranging from 0 to 8) of that member’s embeddedness in the network. This is indicative of the network member’s potential to communicate or coordinate support with other alters.

We use two measures of relationship tenure. Returning respondents were shown a list of the alters named in prior waves and asked to match them, if applicable, with their Wave 3 network members. We use this to identify alters who have been a close confidant for at least 5 years (=1). For network members that had not been named in prior waves, respondents were asked how long they have known the individual (from 1 “less than one year” to 4 “more than 6 years”). The same question was used to assess the length of relationships with all network members named by respondents who were first interviewed at Wave 3.

Sociodemographic and Neighborhood Characteristics

We consider how sociodemographic characteristics are associated with older adults’ inclusion of local confidants in their social networks. Sociodemographic measures include respondent age at Wave 3, gender, race/ethnicity, educational attainment, and employment status. Summary statistics for these and other covariates are presented in Table 1.

Table 1.

Summary Statistics for Respondent-Level Covariates (n = 3,735)

Variable Categories Weighted mean (SD) or proportion
Age 50–64 .37
65–79 .47
80 and older .16
Female .56
Race/Ethnicity Black, non-Hispanic .15
Hispanic .10
White and other .75
Education Less than high school .13
High school degree .24
Some college .35
BA or higher .28
Coresident partner .70
Lives alone .24
Has a living child .92
Currently employed .35
Poor/fair health .22
Residential tenure <6 years .17
6–25 years .43
>25 years .40
Network size (range: 1, 6) 4.05 (1.35)
Characteristics of Respondent’s Local Area
Urbanicity Metropolitan core .81
Suburbs .13
Small town or rural .06
Local poverty (% of tract residents with incomes below poverty) <10% .43
10%–19.99% .27
20%–29.99% .17
30%–39.99% .09
> = 40% .04
Local residential turnover (% of tract residents who moved in past year) <10% .34
10%–19.99% .49
20%–29.99% .12
> = 30% .04
Collective efficacy scale (range: −.02, .04) .01 (.66)
Open in a new tab

We also examine four characteristics of the respondent’s neighborhood. We determine the urbanicity of the respondent’s local area based on the USDA 2010 rural–urban commuting area (RUCA) codes assigned at the tract level. From the 2011–2015 American Community Survey, we identify the percentages of residents in respondents’ tracts that (a) have incomes below the poverty line and (b) moved since the previous year.

NSHAP uses respondent reports of the local context to assess collective efficacy within the local area (Sampson, Raudenbush, & Earls, 1997; York Cornwell & Cagney, 2014). Five questions on the LBQ capture neighborhood social cohesion by assessing whether neighbors get along, can be trusted, share the same values, are close-knit, and help each other. Responses range from 1 “strongly disagree” to 5 “strongly agree.” Three questions on the LBQ assessed neighborhood social ties by asking how often neighbors visit each other, do favors, and exchange personal advice, ranging from 0 “never” to 3 “often.” We standardize and combine these items to create a collective efficacy scale (α = .79).

Covariates

We consider covariates that may be associated with individual characteristics and network propinquity. Based on the Wave 3 household roster, we identify respondents who live alone (=1), and those who have a coresident partner (=1). We use an LBQ question asking how long respondents have lived in their local area (from “less than 1 year” to “more than 25 years”), and we consider whether the respondent has a living child (=1), is currently employed (=1), and reports poor/fair self-rated health (=1).

Analytic Approach

We first compare the characteristics of local and nonlocal network ties named by respondents. These are dyad-level analyses—network ties are the unit of analysis. Because most respondents named multiple ties, we use clustered bootstrapping to assess the statistical significance of differences between local and nonlocal ties.

Second, we use multivariate logistic regression models to predict whether respondents have at least one local kin or non-kin alter in their social network. These analyses are conducted at the respondent level and incorporate person-level weights provided by the NSHAP to adjust for attrition and selection. Standard errors are adjusted for clustering and stratification in the NSHAP sampling design. We implement a Bonferroni correction due to the large number of tests of coefficients and note both the uncorrected and corrected levels of statistical significance.

Our analytic sample is drawn from the 4,607 Wave 3 respondents age 50 and older. We exclude 709 respondents who did not return the LBQ or respond to questions about neighborhood tenure on the LBQ, an additional 63 respondents who did not name any network alters or had missing information on their geographic proximity, and an additional 100 respondents who had missing data on sociodemographic and other covariates. The final sample size for the regression models is 3,735 respondents. Dyad-level analyses examine 15,137 network alters named by these respondents.

Results

About 19.7% of all network ties named by NSHAP Wave 3 respondents are local, living within a 20-min walk or about one mile of the respondent’s home. About 58.3% of network ties are nonlocal and 22.0% are coresidential.

Almost all of the coresidential ties are kin (94.2%). Most of the nonlocal ties are kin (58.6%), while most of the local ties are non-kin (58.7%). Local non-kin ties are primarily friends (73.4%), but they also include alters identified as neighbors (16.3%), coworkers (3.8%), and others (6.5%).

Geographic Proximity and Tie Characteristics

Table 2 presents tie characteristics, with tests of significance based on clustered bootstrap 95% confidence intervals (see Supplementary Table 1). We find support for our first hypothesis—that local ties are more frequently accessed. Approximately 44.4% of local kin ties are accessed every day, and 81.9% at least several times a week. These proportions are significantly higher than those observed among nonlocal non-kin ties (20.0% and 51.6%, respectively). Rates of interaction are lower for non-kin ties compared to kin ties, but geographic proximity also structures their rates of interaction. Nearly a quarter of local non-kin ties (23.2%) are accessed every day, and fully two thirds (68.6%) at least several times per week. These proportions are significantly higher than those for nonlocal non-kin ties (15.6% and 48.3%, respectively).

Table 2.

Characteristics of Network Ties Named by Respondents in NSHAP Wave 3 (n = 15,137 ties)a

Coresident (n = 3,329; 22.0% of all ties) Local kin (n = 1,234; 8.2% of all ties) Nonlocal kin (n = 5,169; 34.1% of all ties) Local Non-kin (n = 1,750; 11.6% of all ties) Nonlocal non-kin (n = 3,655; 24.1% of all ties)
Proportion or mean (SD) Proportion or mean (SD) Proportion or mean (SD) Proportion or mean (SD) Proportion or mean (SD)
Interaction with respondent
 Every day .967 .444* .200 .232* .156
 Several times a week .026 .375* .316 .454* .327
 Once a week .004 .121* .248 .182 .200
 Once every 2 weeks .002 .036* .131 .070* .148
 Once a month or less .000 .024* .104 .063* .169
Average frequency of interaction with other altersb (range: 0–8) 5.166 4.973* 3.999 3.114* 2.462
Named in prior wavec .916 .695 .709 .332* .435
Length of relationship
 Less than 1 year .021 .005 — .034 .024
 1–3 years .041 .005 .002 .155 .104
 3–6 years .075 .023 .017 .199 .174
 More than 6 years .863 .968 .980 .612* .697
Open in a new tab

Note: aStatistics are unweighted and based on all network ties named by the 3,735 respondents included in the final analytic sample. bCalculated among respondents that named at least two network members at Wave 3 and have nonmissing data on how frequently network members interact with one another (n = 3,525). cCalculated among ties named by respondents who participated in Wave 2 (n = 2,097).

*p < .05 (two-tailed tests indicating statistically significant difference from nonlocal counterpart).

Kin ties are also more interconnected with the network than are non-kin ties, regardless of geographic proximity. But within kin and non-kin ties, geographic proximity matters. On average, local kin interact with other network members about once every two weeks (4.973), while nonlocal kin interact with other network members about once a month (3.999; p < .05). Local non-kin interact with other alters a couple times a year, on average (3.114), which is significantly more often than nonlocal non-kin (2.462; p < .05). This supports our second hypothesis—that local ties are more embedded in respondents’ social networks.

Finally, we consider differences in relationship tenure. Of the kin ties named in Wave 3, about 69.5% of local kin and 70.9% of nonlocal kin were named in a prior wave of the NSHAP. Nearly all kin ties first named at Wave 3 were known for more than 6 years. However, local non-kin tend to be newer relationships than nonlocal non-kin, and are significantly less likely to have been named at a prior wave (33.2%, compared to 43.5% of nonlocal non-kin). Local non-kin ties that were introduced at Wave 3 are significantly less likely to have been known for more than 6 years (61.2%, compared to 69.7% of nonlocal non-kin). These differences support our third hypothesis, in that local ties are newer than nonlocal ties.

Local Kin and Local Non-Kin Ties: Respondent-Level Analysis

Nearly half of respondents (47.5%) named at least one local tie in their social network. About 24% of respondents named at least one local kin tie and about 31% named at least one local non-kin tie. However, as shown in Figure 1, few respondents named more than one local kin or non-kin tie.

Figure 1.

Figure 1.

Open in a new tab

Number and types of local network ties named by respondents at Wave 3 of the NSHAP (n = 3,735 respondents)

Table 3 presents results from logistic regression models predicting whether respondents name at least one local kin tie (Models 1–2) and at least one local non-kin tie (Models 3–4). Before turning to our hypotheses, we observe that the propinquity of non-kin network ties increases with age. Compared to those under the age of 65, the oldest-old have approximately 67% higher odds of naming at least one local non-kin tie (OR = 1.668; p < .001). Women are significantly more likely to name a local kin tie and a local non-kin tie.

Table 3.

Odds Ratios from Logistic Regression Models (n = 3,735 Respondents)a

Predicting at least one local kin tie Predicting at least one local non-kin tie
Model 1 Model 2 Model 3 Model 4
OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)
Age
 50–64 (ref.) — — — —
 65–79 1.022 (0.824–1.267) 0.942 (0.728–1.219) 1.143 (0.941–1.387) 0.892 (0.720–1.105)
 80 and older 1.252 (0.934–1.678) 1.080 (0.780–1.497) 1.668*** ,b (1.277–2.179) 1.092 (0.823–1.449)
Female 1.529*** ,b (1.270–1.839) 1.490*** ,b (1.242–1.787) 1.262* (1.074–1.483) 1.138 (0.962–1.348)
Race/Ethnicity
 Black, non-Hispanic 1.502** ,b (1.146–1.970) 1.266 (0.959–1.670) 0.828 (0.652–1.051) 0.741* (0.570–0.963)
 Hispanic 1.266 (0.909–1.764) 1.255 (0.890–1.772) 0.991 (0.747–1.314) 1.050 (0.758–1.453)
 White and other (ref.) — — — —
Education
 Less than HS (ref.) — — — —
 High school 0.845 (0.622–1.149) 0.928 (0.669–1.286) 1.042 (0.729–1.488) 1.112 (0.758–1.629)
 Some college 0.540*** ,b (0.416–0.702) 0.650** (0.498–0.850) 1.128 (0.836–1.521) 1.208 (0.878–1.663)
 BA or higher 0.271*** ,b (0.195–0.376) 0.351*** ,b (0.246–0.500) 0.988 (0.698–1.399) 1.060 (0.714–1.574)
Coresident partner 1.058 (0.751–1.489) 0.891 (0.654–1.215)
Has a living child 1.907** (1.193–3.049) 0.671** (0.511–0.881)
Lives alone 1.496* (1.045–2.141) 1.689** (1.195–2.386)
Residential tenure
 <6 years 0.663* (0.483–0.911) 0.634** (0.493–0.814)
 6–25 years 0.709** (0.565–0.888) 0.972 (0.812–1.163)
 >25 years (ref.) — —
Currently employed 1.098 (0.865–1.392) 0.715** ,b (0.586–0.871)
Poor/fair self-rated health 1.220 (0.987–1.508) 1.257 (0.960–1.648)
Urbanicity
 Metropolitan core (ref.) — —
 Suburbs 1.541** (1.173–2.024) 0.980 (0.707–1.358)
 Small town or rural 1.015 (0.723–1.423) 0.645* (0.456–0.911)
Local poverty rates
 <10% (ref.) — —
 10%–19.99% 0.953 (0.728–1.246) 1.347* (1.061–1.711)
 20%–29.99% 1.440* (1.026–2.020) 1.348 (0.988–1.840)
 30%–39.99% 1.646* (1.069–2.534) 1.126 (0.771–1.644)
 > = 40% 2.043** (1.299–3.211) 1.898* (1.087–3.315)
Local residential turnover
 <10% (ref.)
 10%–19.99% 0.884 (0.676–1.157) 1.100 (0.884–1.369)
 20%–29.99% 0.750 (0.507–1.111) 1.255 (0.870–1.810)
 > = 30% 0.860 (0.445–1.660) 1.579 (0.981–2.542)
Local collective efficacy 0.932 (0.804–1.079) 1.950*** ,b (1.663–2.287)
F (df) 20.50*** (9, 87) 9.07*** (26, 70) 15.17*** (9, 87) 11.06*** (26, 70)
Open in a new tab

Note: aResults are consistent when estimates are adjusted for potential selection bias using a complete-case weighting procedure (Morgan & Todd, 2008). bStatistical significance is robust after applying a Bonferroni correction, which changes the critical threshold to p < .006 in Models 1 and 3, and p < .002 in Models 2 and 4.

*p < .05; **p < .01; ***p < .001 (two-sided tests). Odds ratios (OR) shown with 95% confidence intervals (CI). All models incorporate respondent-level weights that account for the NSHAP survey design and adjust for the probability of selection. Network size is included as a control in each model, but not shown here.

We find partial support for our fourth hypothesis, in that racial minorities and socioeconomically disadvantaged older adults are more likely to have local kin ties. Black respondents have approximately 50% higher odds of including a local kin tie in their networks compared to white respondents (OR = 1.502; p < .01). Latino respondents also may be more likely to name a local kin tie than whites, but the difference is not statistically significant. Model 1 also shows that older adults with lower levels of education are significantly more likely to name a local kin tie. Compared to those who did not complete high school, older adults who attended some college have about 46% lower odds of including local kin in their network (OR = 0.540; p < .001) and those who earned a bachelor’s degree have about 73% lower odds of including local kin (OR = 0.271; p < .001).

We find less support for our hypothesis that racial minorities and socioeconomically disadvantaged older adults are more likely to have local non-kin ties. In fact, we observe that Black respondents are less likely to include a local non-kin tie in their networks. The difference between Black and white respondents achieves statistical significance after the inclusion of covariates in Model 4 (OR = 0.741; p < .05). We do not observe significant differences in the likelihood of including local non-kin ties between Latino and white older adults, or across levels of educational attainment.

Covariates introduced in Models 2 and 4 provide additional insight into how network propinquity is associated with respondent characteristics. Older adults who live alone, and those who have lived longer in the same neighborhood, are more likely to name both local kin ties and local non-kin ties. Current employment is not associated with naming a local kin tie, but older adults who are in the work force are significantly less likely to name a local non-kin tie in their network. In supplemental analyses, we find that employment is primarily responsible for reducing the significance of age in Model 4, suggesting that the oldest-old are more likely to name a local non-kin tie because they are less likely to be in the workforce.

Local Neighborhood Conditions

Models 2 and 4 also incorporate conditions of the local area. Although we did not expect to find significant associations between local conditions and the inclusion of local kin ties, a few results are worth noting. First, suburban residents and those in moderate- and high-poverty tracts are significantly more likely to name local kin ties. Second, supplemental analyses indicate that the inclusion of tract poverty is primarily responsible for reducing the significant difference between Black and white respondents in the likelihood of naming local kin ties.

We focus here how local conditions may be associated with having local non-kin ties. In support of our fifth hypothesis, older adults who reside in urban areas are significantly more likely to include local friends and neighbors in their networks, compared to those who live in small town or rural areas (OR = 0.645; p < .01). Results for localized socioeconomic disadvantage contradict our sixth hypothesis. Older adults who reside in socioeconomically disadvantaged tracts are more likely to have local non-kin ties in their networks. Compared to those in very affluent tracts, older adults in high poverty tracts have nearly 90% higher odds of naming a local friend or neighbor in their network (OR = 1.898; p < .05). Older adults in middle-income tracts are also more likely to include local non-kin in their networks (OR = 1.347; p < .05). We do not find a significant association between residential instability and the likelihood of naming local non-kin ties.

In support of our seventh hypothesis, older adults who live in areas with higher levels of collective efficacy are significantly more likely to include a local friend or neighbor in their network. A one standard deviation increment in collective efficacy is associated with nearly twice the odds of naming a local friend or neighborhood in the network (OR = 1.950; p < .001).

Discussion

Although prior research has documented the value of having nearby adult children, less attention has been devoted to the proximity of confidant networks, including both family and non-kin, and the potential value of local confidants for community-residing older adults. We find that about 20% of older adults’ network ties are local, that is, residing within about a mile of the older adult. About half of the community-residing older adults have at least one local tie, and nearly 60% of these local ties are non-kin. Our results suggest that these local ties are uniquely positioned to provide companionship and support.

Compared to nonlocal ties, local confidants are more frequently accessed and more embedded within the network. Higher levels of interaction with the respondent and with other network members provide greater opportunities for local network members to provide practical assistance and emotional support, and coordinate with other network members to meet older adults’ needs (Hank, 2007). While nonlocal ties may also be able to provide these resources, local ties may be particularly efficient and effective. Local non-kin ties also tend to be newer—they are more recently formed, or have recently become more intimate. Research on network turnover should consider how local non-kin ties may replace or supplement nonlocal ties, perhaps in response to increasing needs for in-person or just-in-time support.

The oldest-old are more likely to name local non-kin ties—is due, in part, to the fact that they are less likely to be employed. Retirement affords greater opportunity for exposure to the local context, at the same time that older adults may seek to replace ties with coworkers. Increasing localization of the network may also reflect age-related changes such as declining health or mobility limitations, or purposive action to cull one’s network to the most emotionally rewarding ties (Charles & Carstensen, 2010). If geographic proximity enhances network ties’ ability to provide companionship and support, then focusing on the most emotionally rewarding ties may effectively increase the presence of local ties in the network.

Our results also indicate that local network ties are socially structured. Socially disadvantaged older adults are more likely to have local kin ties. Black older adults, those with lower levels of education, and those who live in poorer neighborhoods are more likely to have at least one local kin tie. This may reflect lower residential mobility within disadvantaged families (Jacobs et al., 2016) and greater needs for family support within resource-poor areas. Prior research has generally viewed local kin as facilitative of support and integration (Logan & Spitze, 1994; Silverstein et al., 1997), but future work should also consider how the presence of local kin may reinforce spatial inequalities and resource constraints for those in disadvantaged neighborhoods.

Having local non-kin ties is predicted by aspects of both social advantage and disadvantage; although these associations may operate through distinct mechanisms. On one hand, older adults who live in poorer neighborhoods are more likely to name local non-kin ties. This is somewhat unexpected because poorer neighborhoods provide fewer organizational or institutional settings that would promote the formation of close relationships with neighbors (Sampson, 2012). However, qualitative research suggests that residents of poorer neighborhoods may activate exchanges of support or resources in order to satisfy day-to-day needs and subsidize the lack of local institutions (Newman, 2003; Stack, 1974). For older adults in poorer neighborhoods, then, local non-kin ties may reflect adaptation to constraints rather than opportunities.

On the other hand, white older adults and those who live in neighborhoods with higher levels of collective efficacy are more likely to have local non-kin ties. This may reflect differences in opportunities to cultivate close relationships nearby neighbors. For those who want to cultivate close ties with their neighbors, higher rates of neighborly interactions, fewer neighborhood problems, and the presence of local organizations provide fertile ground (Sampson, 2012; Sampson et al., 1997). A related possibility is that having a close relationship with a local friend or neighbor increases individuals’ integration in the neighborhood (Logan & Spitze, 1994) and potentially their perception of collective efficacy. Further research using clustered and longitudinal data could shed light on how neighborhood social context may be a facilitator, or product, of individuals’ local network ties.

This is the first paper, to our knowledge, to examine personal network propinquity within a nationally representative sample of community-residing older adults. Measurement of the network and the neighborhood constrain the scope of our analysis. Because the NSHAP data focus on the confident network, we cannot assess the proximity of other types of network ties, including weaker or elastic ties that may be sources of support (Small, 2013; Torres, 2019). The demarcation of the “local area” as a one-mile radius around respondents’ households standardizes the geographic span of the local (York Cornwell & Cagney, 2014), but it precludes more nuanced considerations such as the distance to nonlocal ties, older adults’ perception of their neighborhood, variations across urban and rural areas, and features like landmarks and roadways that shape daily mobility. Further research should consider these factors, as well as the potentially unique value of ties that are formed and maintained through shared, routine locations that are outside of the residential area (Small & Adler, 2019; Small, 2006; Torres, 2018).

Despite these limitations of the study, our results indicate that local network ties may be particularly valuable sources of support and companionship for community-residing older adults. Prior research in social gerontology has focused on the proximity of adult children as indicative of access to caregiving and assistance, but our findings show that a large share of older adults have close confidents—including but not limited to adult children—who live nearby. In fact, about half of these local ties are non-kin. The inclusion of local ties within older adults’ social networks seems to reflect both resources and needs. We urge additional research to examine the processes through which older adults form and maintain local ties, especially with non-kin, and the implications of network proximity for trajectories of health and well-being in later life.

Funding

This work was supported by the Center for the Study of Inequality at Cornell University and by the National Social Life, Health, and Aging Project (NSHAP), which is funded by the National Institute on Aging and the National Institutes of Health (R01AG043538; R01AG048511; R37AG030481). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Supplementary Material

gbaa033_suppl_supplementary_Appendix_1
Click here for additional data file. (20.4KB, docx)

Acknowledgments

The authors are grateful for comments on earlier versions of this work by Kate Cagney, Benjamin Cornwell, Louise Hawkley, and James Iveniuk.

Author ContributionsE. Y. Cornwell planned the study, supervised the data analysis, and wrote and revised the paper. A. W. Goldman helped to plan the study, conducted data analysis, and contributed to writing and revising the manuscript.

Conflict of Interest

None reported.

References

  1. Ajrouch, K. J., Antonucci, T. C., & Janevic, M. R. (2001). Social networks among blacks and whites: The interaction between race and age. The Journals of Gerontology. Series B: Psychological Sciences and Social Sciences, 56, S112–S118. doi: 10.1093/geronb/56.2.s112 [DOI] [PubMed] [Google Scholar]
  2. Anderson, M., & Perrin, A. (2017). Tech adoption climbs among older adults. Pew Research Center. Retrieved from https://www.pewresearch.org/internet/2017/05/17/tech-adoption-climbs-among-older-adults/ [Google Scholar]
  3. Bearman, P., & Parigi, P. (2004). Cloning headless frogs and other important matters: Conversation topics and network structure. Social Forces, 83(2), 535–557. doi: 10.1353/sof.2005.0001 [DOI] [Google Scholar]
  4. Cairncross, F. (2001). The death of distance: How the communications revolution is changing our lives. Cambridge, MA: Harvard Business Press. [Google Scholar]
  5. Charles, S. T., & Carstensen, L. L. (2010). Social and emotional aging. Annual Review of Psychology, 61, 383–409. doi: 10.1146/annurev.psych.093008.100448 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Choi, H., Schoeni, R. F., Langa, K. M., & Heisler, M. M. (2015). Older adults’ residential proximity to their children: Changes after cardiovascular events. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 70, 995–1004. doi: 10.1093/geronb/gbu076 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cornwell, B. (2015). Social disadvantage and network turnover. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 70, 132–142. doi: 10.1093/geronb/gbu078 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cornwell, B., Laumann, E. O., & Schumm, L. P. (2008). The social connectedness of older adults: A national profile*. American Sociological Review, 73, 185–203. doi: 10.1177/000312240807300201 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Fischer, C. (2011). Still connected: Family and friends in America since 1970. New York, NY: Russell Sage Foundation. [Google Scholar]
  10. Fischer, C. S. (1982). To dwell among friends: Personal networks in town and city. Chicago: The University of Chicago Press. [Google Scholar]
  11. Fischer, C. S. (2009). The 2004 GSS finding of shrunken social networks: An artifact? American Sociological Review, 74(4), 657–669. doi: 10.1177/000312240907400408 [DOI] [Google Scholar]
  12. Ha, J-H., & Carr, D. (2005). The effect of parent-child geographic proximity on widowed parents’ psychological adjustment and social integration. Research on Aging, 27(5), 578–610. doi: 10.1177/0164027505277977 [DOI] [Google Scholar]
  13. Hank, K. (2007). Proximity and contacts between older parents and their children: A european comparison. Journal of Marriage and Family, 69(1), 157–173. doi: 10.1111/j.1741-3737.2006.00351.x [DOI] [Google Scholar]
  14. Hunsaker, A., & Hargittai, E. (2018). A review of Internet use among older adults. New Media & Society, 20(10), 3937–3954. doi: 10.1177/1461444818787348 [DOI] [Google Scholar]
  15. Hurlbert, J. S., Haines, V. A., & Beggs, J. J. (2000). Core networks and tie activation: What kinds of routine networks allocate resources in nonroutine situations? American Journal of Sociology, 65(4), 598–618. doi: 10.2307/2657385 [DOI] [Google Scholar]
  16. Ikkink, K. K., & van Tilburg, T. (1999). Broken ties: Reciprocity and other factors affecting the termination of older adults’ relationships. Social Networks, 21(2), 131–146. doi: 10.1016/S0378-8733(99)00005-2 [DOI] [Google Scholar]
  17. Jacobs, M. T., Broese van Groenou, M. I., Aartsen, M. J., & Deeg, D. J. H. (2016). Diversity in older adults’ care networks: The added value of individual beliefs and social network proximity. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 73(2), gbw012. doi: 10.1093/geronb/gbw012 [DOI] [PubMed] [Google Scholar]
  18. Krause, N. (1993). Neighborhood deterioration and social isolation in later life. International Journal of Aging & Human Development, 36, 9–38. doi: 10.2190/UBR2-JW3W-LJEL-J1Y5 [DOI] [PubMed] [Google Scholar]
  19. Liu, H., & Waite, L. (2014). Bad marriage, broken heart? Age and gender differences in the link between marital quality and cardiovascular risks among older adults. Journal of Health and Social Behavior, 55, 403–423. doi: 10.1177/0022146514556893 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Logan, J. R., & Spitze, G. D. (1994). Family neighbors. American Journal of Sociology. 100, 453–476. doi: 10.1086/230543 [DOI] [Google Scholar]
  21. Logan, J., & Spitze, G. (1996). Family ties: Enduring relations between parents and their grown children. Philadelphia, PA: Temple University Press. [Google Scholar]
  22. Mair, C. A. (2019). Alternatives to aging alone?: “Kinlessness” and the importance of friends across European contexts. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 74, 1416–1428. doi: 10.1093/geronb/gbz029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Morgan, S. L., & Todd, J. J. (2008). A diagnostic routine for the detection of consequential heterogeneity of causal effects. Sociological Methodology, 38, 231–281. doi: 10.1111/j.1467-9531.2008.00204.x [DOI] [Google Scholar]
  24. Newman, K. S. (2003). A different shade of gray: Midlife and beyond in the inner city. New York, NY: The New Press. [Google Scholar]
  25. Paik, A., & Sanchagrin, K. (2013). Social isolation in America: An artifact. American Sociological Review, 78(3), 339–360. doi: 10.1177/0003122413482919 [DOI] [Google Scholar]
  26. Quan-Haase, A., Mo, G. Y., & Wellman, B. (2017). Connected seniors: How older adults in East York exchange social support online and offline. Information, Communication & Society, 20(7), 967–983. doi: 10.1080/1369118X.2017.1305428 [DOI] [Google Scholar]
  27. Ross, C. E., Mirowsky, J., & Pribesh, S. (2001). Powerlessness and the amplification of threat: Neighborhood disadvantage, disorder, and mistrust. American Sociological Review, 66(4), 568–591. doi: 10.2307/3088923 [DOI] [Google Scholar]
  28. Sampson, R. J. (2012). Great American city: Chicago and the enduring neighborhood effect. Chicago, IL: University of Chicago Press. [Google Scholar]
  29. Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science (New York, N.Y.), 277, 918–924. doi: 10.1126/science.277.5328.918 [DOI] [PubMed] [Google Scholar]
  30. Schafer, M. H., & Koltai, J. (2015). Does embeddedness protect? Personal network density and vulnerability to mistreatment among older American adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 70, 597–606. doi: 10.1093/geronb/gbu071 [DOI] [PubMed] [Google Scholar]
  31. Schieman, S. (2005). Residential stability and the social impact of neighborhood disadvantage: A study of gender- and race-contingent effects. Social Forces, 83(3), 1031–1064. [Google Scholar]
  32. Silverstein, M., & Angelelli, J. J. (1998). Older parents’ expectations of moving closer to their children. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 53, S153–S163. doi: 10.1093/geronb/53b.3.s153 [DOI] [PubMed] [Google Scholar]
  33. Silverstein, M., Bengtson, V. L., & Lawton, L. (1997). Intergenerational solidarity and the structure of adult child–parent relationships in American families. American Journal of Sociology, 103(2), 429–460. doi: 10.1086/231213 [DOI] [Google Scholar]
  34. Small, M. L., & Adler, L. (2019). The role of space in the formation of social ties. Annual Review of Sociology, 45(1), 111–132. doi: 10.1146/annurev-soc-073018-022707 [DOI] [Google Scholar]
  35. Small, M. L. (2006). Neighborhood institutions as resource brokers: Childcare centers, interorganizational ties, and resource access among the poor. Social Problems, 53(2), 274–292. doi: 10.1525/sp.2006.53.2.274 [DOI] [Google Scholar]
  36. Small, M. L. (2013). Weak ties and the core discussion network: Why people regularly discuss important matters with unimportant alters. Social Networks, 35(3), 470–483. doi: 10.1016/j.socnet.2013.05.004 [DOI] [Google Scholar]
  37. Smith, A., & Anderson, M. (2018). Social media use in 2018. Pew Research Center. Retrieved from https://www.pewresearch.org/internet/2018/03/01/social-media-use-in-2018/ [Google Scholar]
  38. Spring, A., Ackert, E., Crowder, K., & South, S. J. (2017). Influence of proximity to kin on residential mobility and destination choice: Examining local movers in metropolitan areas. Demography, 54, 1277–1304. doi: 10.1007/s13524-017-0587-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Stack, C. B. (1974). All our kin: Strategies for survival in a black community. New York, NY: Harper and Row. [Google Scholar]
  40. Tang, F. (2008). Socioeconomic disparities in voluntary organization involvement among older adults. Nonprofit and Voluntary Sector Quarterly, 37(1), 57–75. doi: 10.1177/0899764007300409 [DOI] [Google Scholar]
  41. Torres, S. (2018). Aging alone, gossiping together: Older adults’ talk as social glue. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 74(8), 1474–1482. doi: 10.1093/geronb/gby154 [DOI] [PubMed] [Google Scholar]
  42. Torres, S. (2019). On elastic ties: Distance and intimacy in social relationships. Sociological Science, 6, 235–263. doi: 10.15195/v6.a10 [DOI] [Google Scholar]
  43. Wellman, B., & Wortley, S. (1990). Different strokes from different folks: Community ties and social support. American Journal of Sociology. 96, 558–588. doi: 10.1086/229572 [DOI] [Google Scholar]
  44. Wirth, L. (1938). Urbanism as a way of life. American Journal of Sociology, 44(1), 1–24. doi: 10.1086/217913 [DOI] [Google Scholar]
  45. York Cornwell, E., & Behler, R. L. (2015). Urbanism, neighborhood context, and social networks. City & Society (Washington, D.C.), 14, 311–335. doi: 10.1111/cico.12124 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. York Cornwell, E., & Cagney, K. A. (2014). Assessment of neighborhood context in a nationally representative study. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 69 Suppl 2(Suppl_2), S51–63. doi: 10.1093/geronb/gbu052 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. York Cornwell, E., & Waite, L. J. (2012). Social network resources and management of hypertension. Journal of Health and Social Behavior, 53, 215–231. doi: 10.1177/0022146512446832 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gbaa033_suppl_supplementary_Appendix_1
Click here for additional data file. (20.4KB, docx)

Articles from The Journals of Gerontology Series B: Psychological Sciences and Social Sciences are provided here courtesy of Oxford University Press

RESOURCES