Neighborhood Social Capital and Achieved Mobility of Older Adults

Abstract

Objective: Evaluate associations of neighborhood social capital and mobility of older adults. Method: A community-based survey (Philadelphia, 2010) assessed mobility (Life-Space Assessment [LSA]; range = 0-104) of older adults (n = 675, census tracts = 256). Social capital was assessed for all adults interviewed from 2002-2010 (n = 13,822, census tracts = 374). Generalized estimating equations adjusted for individual- and neighborhood-level characteristics estimated mean differences and 95% confidence intervals (CIs) in mobility by social capital tertiles. Interactions by self-rated health, living arrangement, and race were tested. Results: Social capital was not associated with mobility after adjustment for other neighborhood characteristics (mean difference for highest versus lowest tertile social capital = 0.79, 95% CI = [−3.3, 4.8]). We observed no significant interactions. In models stratified by race, Black participants had higher mobility in high social capital neighborhoods (mean difference = 7.4, CI = [1.0, 13.7]). Discussion: Social capital may not contribute as much as other neighborhood characteristics to mobility. Interactions between neighborhood and individual-level characteristics should be considered in research on mobility.

Keywords

aging social capital mobility neighborhoods

Introduction

Mobility is the ability of an individual to move about his or her environment. Mobility limitations are difficulties or dependence in movement and affect between one third and one half of adults aged 65 or older (Webber, Porter, & Menec, 2010). While interest in understanding the contribution of neighborhood environment to mobility of older adults is growing, the focus to date has largely been on built environment characteristics (Rosso, Auchincloss, & Michael, 2011) rather than social environment (Beard et al., 2009; Bowling & Stafford, 2007; Fisher, Li, Michael, & Cleveland, 2004; King, 2008; Li, Fisher, & Brownson, 2005; Mendes de Leon et al., 2009; Pollack & von dem Knesebeck, 2004). Individuals reside within social groups, such as neighborhoods, and the characteristics of those groups are important to understanding how individual group members act (Kawachi & Berkman, 2000). Social interactions, including those within one’s neighborhood, can be positive or negative and may act as either facilitators or barriers to mobility.

Social capital is defined as the collective features of social structures that facilitate the actions of their members (Kawachi & Berkman, 2000). It is most often conceptualized as the interpersonal trust and norms of reciprocity that exist within a community (Brehm & Rahn, 1997; Kawachi & Berkman, 2000). Social capital is likely structured by both the characteristics of the individuals living in a community (compositional effects) and by aspects of the communities themselves (contextual effects; Hanibuchi et al., 2012; Mohan & Mohan, 2002). Social capital may influence health-related outcomes by influencing health-related behaviors, by influencing access to services and amenities, and by affecting psychosocial processes such as anxiety and self-esteem (Kawachi & Berkman, 2000). High social capital may represent the ability of a community to work collectively to solve problems (Brehm & Rahn, 1997) and can provide informational channels for exchange of relevant health information (Coleman, 1988). Social capital may also act as a buffer against detrimental neighborhood characteristics such as low socioeconomic status and high crime, increasing the resilience of a neighborhood (De Jesus, Puleo, Shelton, & Emmons, 2010; Poortinga, 2012). Both individual- and neighborhood-level determinants of social capital have been identified, which suggests that the individuals in a community and the community structure itself play a role in building social capital (Cao, Mokhtarian, & Handy, 2010; Hanibuchi et al., 2012; Mohan & Mohan, 2002; Poortinga, 2012).

Social capital is associated with self-rated health and mortality (Kawachi & Berkman, 2000), but associations with mobility are less well studied. Communities with high social capital may be more effective at helping frail elders, may provide greater sense of security for those with impairments, and may provide more opportunities for social engagement for those who may otherwise become socially isolated (Cannuscio, Block, & Kawachi, 2003; Glass & Balfour, 2003). Social relations and social engagement are associated with mobility (Glass, Mendes de Leon, Marottoli, & Berkman, 1999; Mendes de Leon, Glass, & Berkman, 2003; Rosso, Taylor, Tabb, & Michael, 2013) and could represent a pathway through which social capital acts on mobility. Neighborhood social capital may also positively affect mobility of older adults through indirect paths including promotion of healthy behaviors as social norms (Mohnen, Volker, Flap, & Groenewegen, 2012), increased mental well-being (Nyqvist, Forsman, Giuntoli, & Cattan, 2013), and feelings of safety (Pitner, Yu, & Brown, 2011), all of which are associated with mobility (Clark et al., 2009; Stuck et al., 1999).

Social capital is a characteristic of the community but individuals contribute to social capital and ultimately, individuals benefit from the social capital of their community (Kawachi & Berkman, 2000). While social capital benefits even those individuals who do not contribute to its creation (Coleman, 1988; Kawachi & Berkman, 2000), not all individuals in a group have equal access to the social capital that exists within their community. Access to social capital may differ based on specific individual-level characteristics. Ecological models predict that there are multiple levels of influence, including individual and neighborhood environment, on health and health-related behaviors and that influences interact across these levels (Sallis, Owen, & Fisher, 2008).

There are several individual-level characteristics, including living alone, health status, and race, that might influence the association of neighborhood social capital and mobility. Living alone may isolate individuals from the community social structure, thereby decreasing the received benefit from collective social resources (Cannuscio et al., 2003). However, degradation of personal social networks may increase dependence on informal community social structures and may therefore strengthen the association of neighborhood social capital with health outcomes, particularly for those who live alone (Cannuscio et al., 2003; Thompson & Krause, 1998).

When individual competence is low, due to poor health for example, the range of environmental stimuli that can be successfully managed is narrowed (Lawton, 1982). Therefore, the environment may play a stronger role in determining the health of older individuals who have limited capabilities (Clarke & George, 2005; Lawton, 1982). However, individual factors such as poor health may be such strong contributors to mobility in older adults that they may supersede neighborhood-level characteristics (Mendes de Leon et al., 2009).

Individual-level contributors to mobility are known to differ by race (Allman, Baker, Maisiak, Sims, & Roseman, 2004). An individual’s race may also affect the ability to access neighborhood social capital and the quality of the social capital that they have access to (Hutchinson et al., 2009; K. A. Lochner, Kawachi, Brennan, & Buka, 2003; Yen & Syme, 1999). Racial differences in access to social capital may occur due to discrimination and historical segregation (Kawachi & Berkman, 2000). Segregation may lead to a greater concentration of kin and close friends within minority neighborhoods and, therefore, to stronger neighborhood bonds and greater access to neighborhood social capital among minorities (Kim, Subramanian, & Kawachi, 2006; Schieman, 2005). Conversely, racial minorities living within White neighborhoods may find accessing social capital more difficult (Hutchinson et al., 2009).

This study used a cross-sectional, community-based sample of adults aged 65 years and older in Philadelphia to determine the associations of neighborhood social capital with achieved mobility and the influence of individual-level characteristics on this association. Effect modification of the social capital and mobility association was assessed for living alone, self-rated health, and race.

Method

Study Sample

The population-based Household Health Survey (HHS) is conducted biennially in southeastern Pennsylvania by the Public Health Management Corporation (Philadelphia, Pennsylvania). Non-institutionalized individuals aged 18 years and older are recruited by random digit dialing stratified on 54 service areas to maintain geographic representativeness. The overall response rate for the 2010 survey was 24.5% (according to the American Association for Public Opinion Research’s [AAPOR] response rate 3 method, 2011); this was only marginally lower than for other large population-based random digit dial telephone surveys widely used in research (Kempf & Remington, 2007). Participants of the 2010 survey who were aged 65 years and older and who lived in the city of Philadelphia were eligible to participate in the Life-Space Mobility in Older Adults substudy with a recruitment goal of 700 participants. The first 948 eligible participants were invited and 702 (74.1%) agreed to participate. This sample constitutes the primary analytic sample for these analyses.

All participants provided informed consent at administration of the survey. This substudy was approved by the institutional review board of Drexel University (Philadelphia, Pennsylvania).

Neighborhood Characteristics

Census tracts were the smallest geographic units available to geocode participants and were used to define neighborhoods (tract median square miles = 0.23 [25th-75th percentile = 0.15-0.35]). All data used to characterize census tract characteristics, including size (in square miles), total population, percentage of adults over 65 years of age, and median income, were obtained from the 2010 U.S. census and American Community Survey (U.S. Department of Commerce, 2010).

Reported crime incidents were downloaded from the Philadelphia Police Department for the year prior to completion of data collection for the HHS (September 1, 2009 until August 31, 2010) and geocoded using ArcMap 10. Incidents were categorized as crimes against persons (homicides, rapes, robberies, and assaults) or crimes against property (burglaries and thefts; Ball et al., 2010).

To ensure a sufficient sample of individuals for estimating social capital at each census tract, we combined responses to the social capital questionnaire from the 2002-2010 surveys. Participants from the 2002-2008 surveys were geocoded to the 2000 census tracts. To match participants across census years during which some tract boundaries changed substantially, all 2010 participants were also geocoded to the 2000 census tracts. Characteristics of the 2010 census used to describe neighborhood-level characteristics (at the block group level) were spatially aggregated to the 2000 census tract boundaries.

Social Capital

Social capital was averaged for 374 census tracts from individual responses to five questions on the HHS from all participants aged 18 and older interviewed between 2002 and 2010, excluding those who participated in the mobility substudy (n = 13,822). The 8 years of data were used to ensure that we had a sufficient number of individuals contributing to the average social capital measure at each census tract.

The questions assessing social capital (Table 1) represent aspects of reciprocity, civic trust, and group participation (Pollack & von dem Knesebeck, 2004). For social capital items, neighborhood was self-defined by each participant. Each variable was dichotomized as 0 (low) or 1 (high). These scores were then summed across the five questions resulting in a score from 0 (lowest social capital) to 5 (highest social capital; Leader & Michael, 2013). Social capital scores were then averaged across all participants in each census tract and categorized in tertiles based on the distribution for all Philadelphia census tracts (tertile cutoffs: 1.50-2.99, 3.00-3.41, 3.42-5.00).

Table 1.

Items Assessing Social Capital From the Public Health Management Corporation’s Household Health Survey and Scoring Used in These Analyses.

	Responses
Questions	High = 1	Low = 0
How likely people in your neighborhood are willing to help their neighbors with routine activities such as picking up their trash cans or helping to shovel snow	Always or often	Sometimes, rarely, or never
Have people in your neighborhood ever worked together to improve the neighborhood? For example, through a neighborhood watch, creating a community garden, building a community playground, or participating in a block party	Yes	No
I feel that I belong and am a part of my neighborhood	Strongly agree or agree	Strongly disagree or disagree
Most people in my neighborhood can be trusted	Strongly agree or agree	Strongly disagree or disagree
How many local groups or organizations in your neighborhood do you currently participate in such as social, political, religious, school related, or athletic organizations?	1 or more	0

Life-Space Assessment

The Life-Space Assessment (LSA) evaluates the extent of achieved movement of the respondent in the past month. Five levels of movement are assessed: (a) the home, (b) areas immediately outside the home such as yards or driveways, (c) the neighborhood, (d) the town or city beyond their neighborhood, and (e) beyond their town. For each level of achieved mobility, the respondent is then asked how frequently he or she traveled to that area and whether he or she needed assistance from another person or from equipment (Peel et al., 2005). The LSA was modified for the current study by eliminating areas immediately outside the home, as this level lacks relevance to many urban residents. Scores for distance traveled, frequency traveled, and need for assistance were totaled to create a continuous, normally distributed, composite score ranging from 0 to 104, with higher scores indicating greater mobility. This scoring method was the most highly correlated with physical performance and other measures in testing by the LSA developers (Baker, Bodner, & Allman, 2003). The validity and reliability of the LSA have been established in a population of older adults (Baker et al., 2003; Peel et al., 2005). Internal consistency of the modified LSA was good in this sample (α = .77) and confirmatory factor analysis verified that this version of the LSA measures a single factor (Rosso et al., 2013).

Individual Characteristics

Demographics including age, sex, race, education, difficulty with housing costs, home ownership, living arrangement (lives with others or lives alone), and length of residency in the neighborhood were recorded. Income was recorded and coded as being above or below 200% of the federal poverty level. Self-rated health was reported as excellent, very good, good, fair, or poor and was re-categorized as either good (excellent, very good, or good) or poor (fair or poor). Race was self-reported and was categorized as non-Hispanic White, non-Hispanic Black, or other including Hispanic. Due to the small number of other races, stratified analyses included only Whites and Blacks.

Statistical Analysis

All neighborhood variables were either normally distributed or minimally skewed, and no differences were observed between parametric and non-parametric analyses; parametric statistics are reported here. Associations of neighborhood and participant characteristics with tertiles of social capital were calculated by analysis of variance test for continuous measures and the chi-square test for categorical measures.

Regression models estimating the association of tertiles of neighborhood social capital and individual mobility as a continuous outcome used generalized estimating equations (GEE) to account for nesting of individuals within census tracts (Zeger & Liang, 1986). Multicolinearity of neighborhood covariates with social capital was assessed; no variables had a correlation greater than .8 and all were retained. Analyses were adjusted for individual (age, sex, race, education, poverty status, difficulty with housing costs, home ownership, living arrangement, and self-rated health) and neighborhood (median income, total population, area, and median age) level covariates. Stratified results and tests for interaction are presented for several a priori identified potential effect modifiers: self-rated health, living arrangement, and race.

A sensitivity analysis to assess the effects of length of residency in the neighborhood was conducted by removing individuals with less than 5 years of residency (n = 65) and less than 10 years of residency (n = 124). A final analysis was conducted including crime as a covariate to determine whether the association between social capital and mobility was due to crime rates. SAS 9.2 was used for all statistical analyses.

Results

These analyses included 675 individuals with complete LSA and census tract data (96% of substudy participants, 27 with missing mobility data, and one with missing census tract data). Those who had a complete LSA did not differ in age, racial distribution, poverty status, home ownership, or living arrangement compared with the overall Philadelphia population over 65 years of age (p ≥ .05). This sample was more likely to be female (p < .0001), less likely to be married (p = .002), and better educated (p < .0001) than the older Philadelphia population.

Of the 374 residential census tracts in Philadelphia, 256 tracts had study participants resident in them and were included in these analyses. On average, there were 2.6 participants per census tract (range = 1-11). Included tracts were not different from excluded tracts with respect to social capital. However, included tracts were smaller in area and population size and had lower median age and median income of residents than excluded tracts (all ps < .001).

The social capital score had an adequate internal consistency (α = .58). The correlation of each component with the total ranged from .12 to .38. The removal of any one variable from the score did not improve the overall consistency of the scale. A median of 34.5 (range = 5-116) individuals contributed to the social capital score of each census tract. Average tract-level social capital scores (2002 and 2010) were adequately correlated with the social capital scores from 2010 alone (r = .63, p ≤ .001). On average, social capital increased slightly over the study period (mean difference = 0.26, interquartile range = −0.26-0.73). Changes appeared to represent a regression to the mean as the census tract level correlation between the mean change in social capital over time with the baseline (2002) social capital score (r = −.60, p ≤ .001) was stronger than the correlation between 2002 and 2010 social capital scores (r = .22, p ≤ .001).

At the neighborhood level, higher social capital was correlated with higher neighborhood income and age and lower neighborhood population (Table 2). Mobility was highest for those living in neighborhoods with the highest social capital (Table 3). Individuals who were young, old, male, White, better educated, in better health, not in poverty, and owned their homes were more likely to live in neighborhoods with high social capital (Table 3).

Table 2.

Characteristics of Residential Census Tracts (N = 374) in Philadelphia, Pennsylvania, by Tertile of Social Capital.

	Social Capital
	Total (N = 374)	Highest tertile (n = 124)	Middle tertile (n = 125)	Lowest tertile (n = 125)	p value
	M (SD)	M (SD)	M (SD)	M (SD)	p value
Social capital	3.38 (0.63)	3.76 (0.28)	3.20 (0.12)	2.69 (0.25)
Area (mile²)	0.37 (0.40)	0.41 (0.5)	0.37 (0.4)	0.32 (0.3)	.2
Population	10,118 (4,631)	9,186 (5,339)	10,560 (4,170)	10,626 (4,163)	.02
Median income ($)	43,126 (20,063)	57,431 (24,826)	38,566 (12,420)	33,060 (10,386)	≤.001
Median age (years)	36.4 (6.0)	39.4 (6.2)	36.3 (5.6)	33.5 (4.5)	≤.001

Note. Data compiled from Census 2010.

Table 3.

Characteristics of Adults Aged 65 Years and Older by Tertiles of Neighborhood Social Capital.

	Social capital
	Total (N = 675)	Highest tertile (n = 360)	Middle tertile (n = 206)	Lowest tertile (n = 109)	p value
	M (SD)	M (SD)	M (SD)	M (SD)	p value
Mobility score (range = 0-104)	49.6 (24.9)	54.2 (24.0)	43.8 (25.3)	45.5 (24.3)	≤.001
Age	74.5 (6.9)	73.9 (6.8)	75.8 (6.9)	74.0 (7.1)	.005
	N (%)	n (%)	n (%)	n (%)
Sex
Male	487 (27.9)	107 (29.7)	61 (29.6)	20 (18.4)	.05
Female	188 (72.1)	253 (70.3)	145 (70.4)	89 (81.7)
Race
White	348 (52.4)	197 (56.0)	98 (48.0)	53 (49.1)	.05
Black	278 (41.9)	142 (40.3)	92 (45.1)	44 (40.7)
Other	38 (5.7)	13 (3.7)	14 (6.9)	11 (10.2)
Education
<High school	127 (18.9)	50 (14.0)	53 (25.7)	24 (22.2)	≤.001
High school grad	276 (41.1)	137 (38.3)	90 (43.7)	49 (45.4)
>High school	269 (40.0)	171 (47.7)	63 (30.6)	35 (32.4)
Poverty status
Above 200%	368 (54.5)	223 (61.9)	90 (43.7)	55 (50.5)	≤.001
Below 200%	307 (45.5)	137 (38.1)	116 (56.3)	54 (49.5)
Difficulty with housing costs
No	408 (62.2)	226 (64.6)	123 (61.8)	59 (55.1)	.2
Yes	248 (37.8)	124 (35.4)	76 (38.2)	48 (44.9)
Owns home
Yes	498 (76.0)	280 (80.9)	136 (66.7)	82 (78.1)	≤.001
No	157 (24.0)	66 (19.1)	68 (33.3)	23 (21.9)
Lives alone
No	314 (46.8)	177 (49.4)	92 (45.1)	45 (41.3)	.3
Yes	357 (53.2)	181 (50.6)	112 (54.9)	64 (58.7)
Self-rated health
Good	477 (70.8)	273 (75.8)	132 (64.1)	72 (66.7)	.007
Poor	197 (29.2)	87 (24.2)	74 (35.9)	36 (33.3)

Note. Participants (N = 675) in 256 Census Tract Neighborhoods Were From a Population-Based Survey of Philadelphia Residents in 2010.

In unadjusted analyses, those living in neighborhoods in the highest tertile of social capital had greater mobility compared with those in the lowest tertile (mean difference = 8.5; 95% confidence interval [CI] = [3.4, 13.7]). There was no difference between those in the middle and lowest tertiles of neighborhood social capital. The association at the highest tertile was partially attenuated with adjustment for individual-level characteristics and eliminated after adjustment for neighborhood characteristics (mean difference = 0.79; 95% CI = [−3.3, 4.8]; Table 4).

Table 4.

Mean Differences (MD) in Mobility Score (Range = 0-104) and 95% Confidence Interval (CI) by Neighborhood Social Capital and Covariates in a Sample of Older Adults in Philadelphia, PA (2010; n = 675).

	Unadjusted		Adjusted for individual characteristics		Adjusted for individual and neighborhood characteristics
	MD	95% CI	MD	95% CI	MD	95% CI
Neighborhood characteristics
Social capital
Lowest tertile	0 (REF)		0 (REF)		0 (REF)
Middle tertile	−1.96	[−8.16, 4.23]	−0.03	[−4.73, 4.67]	−2.34	[−7.11, 2.44]
Highest tertile	8.54	[−1.26, 9.40]	4.21	[0.50, 7.93]	0.79	[−3.25, 4.83]
Median income (thousands of $)					0.15	[0.03, 0.27]
Total population (thousands of people)					0.12	[−0.30, 0.54]
Area (square miles)					−5.15	[−13.10, 2.80]
Median age (years)					0.27	[−0.002, 0.53]
Individual characteristics
Age (per year)			−0.71	[−0.96, −0.47]	−0.66	[−0.92, −0.39]
Female sex			−4.17	[−8.00, −0.34]	−3.80	[−8.00, 0.39]
Race
White			0 (REF)		0 (REF)
Black			−5.15	[−8.53, −1.78]	−2.49	[−6.67, 1.69]
Other			−5.75	[−13.98, 2.47]	−4.38	[−12.62, 3.87]
Poor self-rated health			−11.81	[−15.81, −7.80]	−11.55	[−15.65, −7.45]
Education
<High school			−8.20	[−13.26, −3.14]	−8.12	[−13.60, −2.64]
High school grad			−5.19	[−9.30, −1.09]	−4.96	[−9.46, −0.47]
>High School			0 (REF)		0 (REF)
Below 200% poverty			−8.40	[−12.17, −4.64]	−7.71	[−11.67, −3.75]
Has difficulty with housing costs			−3.42	[−6.88, 0.05]	−4.33	[−7.95, −0.71]
Owns home			5.05	[0.76, 9.35]	4.82	[0.58, 9.06]

There were no statistically significant interactions between level of neighborhood social capital and self-rated health (highest tertile compared with lowest: beta for interaction [95% CI] = −6.3 [−15.4, 2.8]), living arrangement (highest tertile compared with lowest: beta for interaction [95% CI] = −5.8 [−14.9, 3.4]), or race (highest tertile compared with lowest: beta for interaction [95% CI] = −4.9 [−13.2, 3.5]). For those who were in poor health, living in a neighborhood with the highest social capital compared with those with the lowest was associated with lower mobility (mean difference = −9.3; 95% CI = [−17.6, −0.9]). In contrast, being in a neighborhood with the highest social capital compared with the lowest was not associated with mobility for those in good health. Black participants who lived in neighborhoods with the highest social capital had greater mobility than those living in neighborhoods with the lowest social capital (mean difference = 7.4; 95% CI = [1.0, 13.7]), whereas for Whites, there was no association (mean difference = 0.11; 95% CI = [−0.05, 0.27]). There were no significant associations for mobility and living in neighborhoods with the highest social capital compared with the lowest for either those living with others (mean difference = 4.1; 95% CI = [−2.7, 10.8]) or those living alone (mean difference = −1.7; 95% CI = [−7.8, 4.4]).

Removal of those who lived in the neighborhood for less than either 5 or 10 years did not qualitatively change the results (data not shown). Neighborhood crime rates were significantly but weakly associated with mobility (personal crime: r = −.18, p < .001; property crime: r = −.076, p = .05). Crime was no longer associated with mobility in multivariate analyses. Adjustment for neighborhood crime did not change the observed association between social capital and mobility (data not shown).

Discussion

In this community-based sample of older adults, we observed no overall association between social capital and mobility after adjustment for other neighborhood characteristics. There were also no significant interactions observed between neighborhood social capital and several individual-level characteristics. We had hypothesized that the impacts of neighborhood characteristics are not equivalent for all individuals resident within them (Rosso et al., 2011) based on ecological models of health behaviors which theorize that individual factors and environmental stimuli act together to determine one’s health and well-being (Lawton, 1982; Sallis et al., 2008). However, our interaction results do not support this.

Social capital is not well studied in relation to mobility. Consistent with our results, previous studies that have assessed social capital and walking or physical function, which can be indicators of mobility, have generally not found significant associations. Neighborhood social capital was not associated with walking (Ball et al., 2010; Mendes de Leon et al., 2009) or physical function (Pollack & von dem Knesebeck, 2004) in previous cross-sectional studies of older adults. A longitudinal study of social capital and walking found a significant, positive association at baseline but no effect of social capital with rate of change in walking over a 12-month period in older adults (Li et al., 2005). Our study differed in that it assessed general mobility that encompasses not only walking but also other forms of transportation, including personal vehicles and public transit.

Although we found no association between neighborhood social capital and mobility for the entire study sample, there may have been associations for subgroups by self-reported health and race. While we cannot rule out that these findings were due to chance, the results are intriguing, particularly as no previous studies have assessed these associations by subgroups.

Previous research has suggested that the association of neighborhood social characteristics and health outcomes differs by race (Locher et al., 2005; K. Lochner, Kawachi, & Kennedy, 1999) and that individual-level predictors of life-space can differ by race (Allman et al., 2004). Racial differences in access to social capital may be particularly important in the context of neighborhood racial composition such that Black residents of majority Black neighborhoods may benefit the most from neighborhood social capital (Hutchinson et al., 2009). Differences in neighborhood racial composition may influence social capital and individual access to it through psychological benefits, access to social support, and buffering against negative influences such as discrimination (Hutchinson et al., 2009). The exact nature of these differences is still unclear. This study lacked sufficient sample size to explore three-way effect modification by individual race, neighborhood racial composition, and neighborhood social capital.

The inverse association of social capital and mobility among those in poor health was not consistent with our hypothesis. We expected that poor self-reported health would reflect presence of chronic diseases and other conditions that would influence mobility more strongly than the effects from the neighborhood social environment (Mendes de Leon et al., 2009), resulting in a stronger association between neighborhood social capital and mobility for those in better health. However, prior research suggests that high levels of instrumental support are correlated with greater disability risk as older individuals with declining function rely more heavily on social contacts for help with daily activities (Mendes de Leon, Gold, Glass, Kaplan, & George, 2001). Along these lines, individuals with mobility-limiting health conditions who live in neighborhoods with high social capital may receive greater levels of instrumental support from their neighbors, allowing them to limit mobility within a comfortable range while aging in place. Nearly 90% of people over age 65 want to stay in their home for as long as possible, or age in place (Farber, Shinkle, Lynott, Fox-Grage, & Harrell, 2011). However, the physical and social environment must be accommodating to allow this. Instrumental support that allows older adults to age in place may be less available to those living in neighborhoods with low social capital, forcing individuals in poor health to maintain higher levels of mobility to meet their daily needs.

We found that the association of social capital with mobility was not altered by inclusion of crime rates within the neighborhood. This is consistent with previous research in younger women that showed social capital but not crime rates was associated with walking (Ball et al., 2010). It may be that perceived safety, which is only moderately associated with actual crime rates, is important in determining mobility in older adults (Clarke, Ailshire, & Lantz, 2009; Foster & Giles-Corti, 2008; Piro, Noss, & Claussen, 2006). We did not have data on perceived safety for our study sample.

Several limitations should be noted. First, we did not have access to addresses for participants and were reliant on census tracts to define neighborhoods. Census tracts are frequently used in neighborhood effects research (Diez Roux, 2001; Macintyre, Ellaway, & Cummins, 2002) and are roughly equal in size to people’s perceptions of their neighborhood, but rarely match the boundaries of an individual’s self-defined neighborhood (Coulton, Korbin, Chan, & Su, 2001). Philadelphia census tracts are fairly small and uniform in size. Second, because this was a cross-sectional study, there is a possibility of reverse causation resulting from individuals with declining mobility moving to neighborhoods with higher social capital. However, sensitivity analyses that excluded individuals who lived in their neighborhood for fewer than either 5 or 10 years did not result in qualitatively different results.

We dichotomized responses to the social capital items, which may have led to a loss of variability in the measure. This method was used to limit inadvertent weighting of particular items due to differing number of response options (Leader & Michael, 2013). Loss of variability in our primary exposure variable may have reduced our ability to detect differences in mobility between groups. Furthermore, we combined social capital responses across individuals of all ages. Perception of social capital and response to items, particularly those relevant to child rearing, may differ by age (Galinsky, Cagney, & Browning, 2012). However, none of our social capital items were inherently age-specific. In addition, the aggregation of responses from individuals of all ages is consistent with the theory that social capital is a property of the overall community rather than a subgroup. We also cannot rule out unmeasured confounding or selection bias. However, our sample did not differ substantially from the population of older adults in Philadelphia, so selection bias may not be a likely explanation for our results. Finally, this study was conducted in a large city in the Northeastern United States and may not be directly applicable to other geographic regions.

The strengths of this study include the use of social capital measures from individuals who were not part of the analysis sample. This eliminated the possibility of reverse causation that may have resulted from an individual’s mobility affecting their response to the social capital questionnaire. In addition, the assessment of effect modification allowed us to examine the effects of social capital within subgroups of the population. Evaluation of the associations of neighborhood characteristics with health outcomes within subpopulations is a step toward better understanding the underlying mechanisms by which they act (Clarke & Nieuwenhuijsen, 2009; Rosso et al., 2011).

These results indicate that neighborhood social capital may not contribute to mobility of older adults after consideration of other neighborhood characteristics. Promotion of mobility is important in that loss of mobility leads to increased risk of poor physical and mental health outcomes and loss of independence (Guralnik et al., 1994; Lampinen & Heikkinen, 2003). Although it is theorized that older adults are more dependent on social capital than are younger adults, neighborhood social capital does not appear to contribute to differences in mobility. Future research on neighborhood-level influences on mobility of older adults should consider possible interactions with individual-level characteristics.

Footnotes

Acknowledgements

The authors would like to thank Dr. Amy Auchincloss for thoughtful feedback on this article.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Association of Schools of Public Health and the U.S. Centers for Disease Control and Prevention’s Environmental Health Scholarship Program to A.L.R. Support was also provided by a grant from the National Institute on Aging (AG028254) to Y.L.M., a training grant from the NIA (T32-AG-000181), and funding from Drexel University.

References

Allman

R. M.

Baker

P. S.

Maisiak

R. M.

Sims

R. V.

Roseman

J. M.

(2004). Racial similarities and differences in predictors of mobility change over eighteen months. J Gen Intern Med, 19(11), 1118-1126. doi:JGI30239[pii]10.1111/j.1525-1497.2004.30239.x.

American Association for Public Opinion Research. (2011). Standard definitions: Final dispositions of case codes and outcome rates for surveys. Deerfield, IL: Author.

Baker

P. S.

Bodner

E. V.

Allman

R. M.

(2003). Measuring life-space mobility in community-dwelling older adults. Journal of the American Geriatrics Society, 51, 1610-1614.

Ball

Cleland

V. J.

Timperio

A. F.

Salmon

Giles-Corti

Crawford

D. A.

(2010). Love thy neighbour? Associations of social capital and crime with physical activity amongst women. Social Science & Medicine, 71, 807-814. doi:10.1016/j.socscimed.2010.04.041

Beard

J. R.

Blaney

Cerda

Frye

Lovasi

G. S.

Ompad

. . . Vlahov

(2009). Neighborhood characteristics and disability in older adults. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 64, 252-257. doi:10.1093/geronb/gbn018

Bowling

Stafford

(2007). How do objective and subjective assessments of neighbourhood influence social and physical functioning in older age? Findings from a British survey of ageing. Social Science Medicine, 64, 2533-2549. doi:10.1016/j.socscimed.2007.03.009

Brehm

Rahn

(1997). Individual-level evidence for the causes and consequences of social capital. American Journal of Political Science, 41, 999-1023.

Cannuscio

Block

Kawachi

(2003). Social capital and successful aging: The role of senior housing. Annals of Internal Medicine, 139, 395-399.

Cao

X. Y.

Mokhtarian

P. L.

Handy

(2010). Neighborhood design and the accessibility of the elderly: An empirical analysis in Northern California. International Journal of Sustainable Transportation, 4, 347-371.

10.

Clark

C. R.

Kawachi

Ryan

Ertel

Fay

M. E.

Berkman

L. F.

(2009). Perceived neighborhood safety and incident mobility disability among elders: The hazards of poverty. BMC Public Health, 9, Article 162. doi:10.1186/1471-2458-9-162.

11.

Clarke

Ailshire

J. A.

Lantz

(2009). Urban built environments and trajectories of mobility disability: Findings from a national sample of community-dwelling American adults (1986-2001). Social Science Medicine, 69, 964-970. doi:10.1016/j.socscimed.2009.06.041

12.

Clarke

George

L. K.

(2005). The role of the built environment in the disablement process. American Journal of Public Health, 95, 1933-1939. doi:10.2105/AJPH.2004.054494

13.

Clarke

Nieuwenhuijsen

E. R.

(2009). Environments for healthy ageing: A critical review. Maturitas, 64, 14-19. doi:10.1016/j.maturitas.2009.07.011

14.

Coleman

J. S.

(1988). Social capital in the creation of human capital [Issue title: Supplement: Organizations and institutions: Sociological and economic approaches to the analysis of social structure]. American Journal of Sociology, 94, S95-S120.

15.

Coulton

C. J.

Korbin

Chan

(2001). Mapping residents’ perceptions of neighborhood boundaries: A methodological note. American Journal of Community Psychology, 29, 371-383.

16.

De Jesus

Puleo

Shelton

R. C.

Emmons

K. M.

(2010). Associations between perceived social environment and neighborhood safety: Health implications. Health Place, 16, 1007-1013. doi:10.1016/j.healthplace.2010.06.005

17.

Diez Roux

A. V.

(2001). Investigating neighborhood and area effects on health. American Journal of Public Health, 91, 1783-1789.

18.

Farber

Shinkle

Lynott

Fox-Grage

Harrell

(2011). Aging in place: A state survey of livability policies and practices. Washington, DC: National Conference of State Legislatures and the AARP Public Policy Institute.

19.

Fisher

K. J.

Michael

Cleveland

(2004). Neighborhood-level influences on physical activity among older adults: A multilevel analysis. Journal of Aging and Physical Activity, 12, 45-63.

20.

Foster

Giles-Corti

(2008). The built environment, neighborhood crime and constrained physical activity: An exploration of inconsistent findings. Preventive Medicine, 47, 241-251. doi:10.1016/j.ypmed.2008.03.017

21.

Galinsky

A. M.

Cagney

K. A.

Browning

C. R.

(2012). Is collective efficacy age graded? The development and evaluation of a new measure of collective efficacy for older adults. Journal of Aging Research, 2012, Article 360254. doi:10.1155/2012/360254

22.

Glass

T. A.

Balfour

J. L.

(2003). Neighborhoods, aging, and functional limitations. In Kawachi

Berkman

L. F.

(Eds.), Neighborhoods and health (pp. 303-334). Oxford, UK: Oxford University Press.

23.

Glass

T. A.

Mendes

Leon

C. F.

Marottoli

R. A.

Berkman

L. F.

(1999). Population based study of social and productive activities as predictors of survival among elderly Americans. British Medical Journal, 319, 478-483.

24.

Guralnik

J. M.

Simonsick

E. M.

Ferrucci

Glynn

R. J.

Berkman

L. F.

Blazer

D. G.

. . .Wallace

R. B.

(1994). A short physical performance battery assessing lower extremity function: Association with self-reported disability and prediction of mortality and nursing home admission. Journal of Gerontology, 49, M85-M94.

25.

Hanibuchi

Kondo

Nakaya

Shirai

Hirai

Kawachi

(2012). Does walkable mean sociable? Neighborhood determinants of social capital among older adults in Japan. Health Place, 18, 229-239. doi:10.1016/j.healthplace.2011.09.015

26.

Hutchinson

R. N.

Putt

M. A.

Dean

L. T.

Long

J. A.

Montagnet

C. A.

Armstrong

(2009). Neighborhood racial composition, social capital and black all-cause mortality in Philadelphia. Social Science Medicine, 68, 1859-1865. doi:10.1016/j.socscimed.2009.02.005

27.

Kawachi

Berkman

L. F.

(2000). Social cohesion, social capital, and health. In Berkman

L. F.

Kawachi

(Eds.), Social epidemiology (pp. 174-190). New York, NY: Oxford University Press.

28.

Kempf

A. M.

Remington

P. L.

(2007). New challenges for telephone survey research in the twenty-first century. Annual Review of Public Health, 28, 113-126. doi:10.1146/annurev.publhealth.28.021406.144059

29.

Kim

Subramanian

S. V.

Kawachi

(2006). Bonding versus bridging social capital and their associations with self rated health: A multilevel analysis of 40 US communities. Journal of Epidemiology & Community Health, 60, 116-122. doi:10.1136/jech.2005.038281

30.

King

(2008). Neighborhood and individual factors in activity in older adults: Results from the neighborhood and senior health study. Journal of Aging and Physical Activity, 16, 144-170.

31.

Lampinen

Heikkinen

(2003). Reduced mobility and physical activity as predictors of depressive symptoms among community-dwelling older adults: An eight-year follow-up study. Aging Clinical and Experimental Research, 15, 205-211.

32.

Lawton

M. P.

(1982). Competence, environmental press, and the adaptation of older people. In Lawton

Windley

Byerts

(Eds.), Aging and the environment. New York, NY: Springer.

33.

Leader

A. E.

Michael

Y. L.

(2013). The association between neighborhood social capital and cancer screening. American Journal of Health Behavior, 37, 683-692. doi:10.5993/AJHB.37.5.12

34.

Fisher

Brownson

R. C.

(2005). A multilevel analysis of change in neighborhood walking activity in older adults. Journal of Aging and Physical Activity, 13, 145-159.

35.

Locher

J. L.

Ritchie

C. S.

Roth

D. L.

Baker

P. S.

Bodner

E. V.

Allman

R. M.

(2005). Social isolation, support, and capital and nutritional risk in an older sample: Ethnic and gender differences. Social Science & Medicine, 60, 747-761. doi:10.1016/j.socscimed.2004.06.023

36.

Lochner

Kawachi

Kennedy

B. P.

(1999). Social capital: A guide to its measurement. Health Place, 5, 259-270.

37.

Lochner

K. A.

Kawachi

Brennan

R. T.

Buka

S. L.

(2003). Social capital and neighborhood mortality rates in Chicago. Social Science & Medicine, 56, 1797-1805.

38.

Macintyre

Ellaway

Cummins

(2002). Place effects on health: How can we conceptualise, operationalise, and measure them? Social Science & Medicine, 55, 125-139.

39.

Mendes

Leon

C. F.

Cagney

K. A.

Bienias

J. L.

Barnes

L. L.

Skarupski

K. A.

Scherr

P. A.

Evans

D. A.

(2009). Neighborhood social cohesion and disorder in relation to walking in community-dwelling older adults: A multilevel analysis. Journal of Aging and Health, 21, 155-171. doi:10.1177/0898264308328650

40.

Mendes

Leon

C. F.

Glass

T. A.

Berkman

L. F.

(2003). Social engagement and disability in a community population of older adults: The New Haven EPESE. American Journal of Epidemiology, 157, 633-642.

41.

Mendes

Leon

C. F.

Gold

D. T.

Glass

T. A.

Kaplan

George

L. K.

(2001). Disability as a function of social networks and support in elderly African Americans and Whites: The Duke EPESE 1986-1992. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 56, S179-S190.

42.

Mohan

(2002). Placing social capital. Progress in Human Geography, 26, 191-210. doi:10.1191/0309132502ph364ra

43.

Mohnen

S. M.

Volker

Flap

Groenewegen

P. P.

(2012). Health-related behavior as a mechanism behind the relationship between neighborhood social capital and individual health—A multilevel analysis. BMC Public Health, 12, Article 116. doi:10.1186/1471-2458-12-116

44.

Nyqvist

Forsman

A. K.

Giuntoli

Cattan

(2013). Social capital as a resource for mental well-being in older people: A systematic review. Aging & Mental Health, 17, 394-410. doi:10.1080/13607863.2012.742490

45.

Peel

Sawyer Baker

Roth

D. L.

Brown

C. J.

Brodner

E. V.

Allman

R. M.

(2005). Assessing mobility in older adults: The UAB Study of Aging Life-Space Assessment. Physical Therapy, 85, 1008-1119.

46.

Piro

F. N.

Noss

Claussen

(2006). Physical activity among elderly people in a city population: The influence of neighbourhood level violence and self perceived safety. Journal of Epidemiology and Community Health, 60, 626-632. doi:10.1136/jech.2005.042697

47.

Pitner

R. O.

Brown

(2011). Exploring the dynamics of middle-aged and older adult residents’ perceptions of neighborhood safety. Journals of Gerontology, 54, 511-527. doi:10.1080/01634372.2011.567322

48.

Pollack

C. E.

von dem Knesebeck

(2004). Social capital and health among the aged: Comparisons between the United States and Germany. Health Place, 10, 383-391. doi:10.1016/j.healthplace.2004.08.008

49.

Poortinga

(2012). Community resilience and health: The role of bonding, bridging, and linking aspects of social capital. Health Place, 18, 286-295. doi:10.1016/j.healthplace.2011.09.017

50.

Rosso

A. L.

Auchincloss

A. H.

Michael

Y. L.

(2011). The urban built environment and mobility in older adults: A comprehensive review. Journal of Aging Research, 2011, Article 816106. doi:10.4061/2011/816106

51.

Rosso

A. L.

Taylor

J. A.

Tabb

L. P.

Michael

Y. L.

(2013). Mobility, disability, and social engagement in older adults. Journal of Aging and Health, 25, 617-637. doi:10.1177/0898264313482489

52.

Sallis

J. F.

Owen

Fisher

E. B.

(2008). Ecological models of health behavior. In Glanz

Rimer

B. K.

Viswanath

(Eds.), Health behavior and health education: Theory, research and practice (4th ed., pp. 465-485). San Francisco, CA: Jossey-Bass.

53.

Schieman

(2005). Residential stability and the social impact of neighborhood disadvantage: A study of gender- and race-contingent effects. Social Forces, 83, 1031-1064. doi:10.1353/sof.2005.0045

54.

Stuck

A. E.

Walthert

J. M.

Nikolaus

Bula

C. J.

Hohmann

Beck

J. C.

(1999). Risk factors for functional status decline in community-living elderly people: A systematic literature review. Social Science & Medicine, 48, 445-469. doi:S0277953698003700

55.

Thompson

E. E.

Krause

(1998). Living alone and neighborhood characteristics as predictors of social support in late life. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 53, S354-364.

56.

U.S. Department of Commerce. (2010). U.S. Census Bureau American FactFinder. Retrieved from http://factfinder2.census.gov/main.html

57.

Webber

S. C.

Porter

M. M.

Menec

V. H.

(2010). Mobility in older adults: A comprehensive framework. Gerontologist, 50, 443-450. doi:10.1093/geront/gnq013

58.

Yen

I. H.

Syme

S. L.

(1999). The social environment and health: A discussion of the epidemiologic literature. Annual Review of Public Health, 20, 287-308. doi:10.1146/annurev.publhealth.20.1.287

59.

Zeger

S. L.

Liang

K. Y.

(1986). Longitudinal data analysis for discrete and continuous outcomes. Biometrics, 42, 121-130.