Abstract
This study investigated the impact of neighborhood social cohesion on the perceived companionship of nearly 1,500 community-dwelling older adults from the Neighborhood, Organization, Aging and Health project (NOAH), a Chicago-based study of older adult well-being in the neighborhood context. We hypothesized that the relationship between neighborhood-level social cohesion and individual residents’ reports of companionship would be more pronounced among those who lived alone than those who resided with others. Controlling for age, gender, education, race, marital status, length of neighborhood residence, and self-rated health, neighborhood social cohesion predicted companionship among those who lived alone; for a one-unit increase in neighborhood social cohesion, the odds of reporting companionship increased by half. In contrast, social cohesion did not predict the companionship of those who resided with others. The results suggest that older adults who live alone particularly profit from the benefits of socially cohesive neighborhood environments.
According to the U.S. Census, the rate of growth of the older population has surpassed that of the population as a whole (Werner, 2011). At the same time, the prevalence of living alone is on the rise, particularly among the older population (Kreider & Elliott, 2009). Given these concurrent trends, in-depth investigations of the living arrangements of older adults are needed to understand the implications of the changing demographic fabric of the United States. While the ability to live alone may signal independence and good health, concerns remain about social isolation, diminished social networks, and a deteriorated life. Where one lives, particularly at older age, may shape opportunities for social integration and engagement. Theoretical and empirical work has suggested that the salience of the neighborhood environment as a social arena increases with age (Glass & Balfour, 2003; Oh, 2003). To expand upon this literature, the current study examined the impact of neighborhood social cohesion on the perceived companionship of older adults with attention to living arrangements.
Companionship and Living Arrangements in Later Life
Companionship has been defined as social involvement in shared activities, recreational or nonrecreational, which is pursued for the intrinsic goal of satisfaction or enjoyment (Rook, 1987, 1990). Although companionship has not received the attention of concepts such as social support (which is a critical social construct, particularly in the face of challenges or stress) and is often considered an indicator of loneliness (e.g., the lack of companionship is an item in the standard survey of loneliness, the Revised UCLA Loneliness scale; Russell, Peplau, & Cutrona, 1980), it is a unique construct that warrants study. Unlike social support, the aim of companionship is not to solve a problem or provide aid but to experience pleasure (Rook, 1987, 1990). As such, companionship does not bear the same costs as receiving support (Rook, 1990). Furthermore, although the lack of companionate relations informs loneliness, the arm of companionship extends far beyond the buffering of loneliness. Research has indicated that companionate relationships and interactions promote psychological well-being, happiness, self-expression, and social satisfaction (Baldassare, Rosenfield, & Rook, 1984; Krause, 2010; Larson, 1978; Rook, 1987, 1990).
Older adults who live alone generally report that they are content with their social life and do not display higher levels of clinical dementia or depressive disorders; however, they report greater levels of loneliness and depressive symptoms, lower perceived social support, and fewer close relationships than those who live with others (Dean, Kolody, Wood, & Matt, 1992; Henderson, Scott, & Kay, 1986; Victor, Scambler, Bond, & Bowling, 2000; Wenger, Davies, Shahtahmesebi, & Scott, 1996; Yeh & Lo, 2004). However, much of this research on the older adult population that lives alone was conducted prior to (or at the early stages of) the increased trend in one-person households. Furthermore, although social support, loneliness, and depression have been examined as key outcome variables in research on living alone, there is a dearth of research on the related but distinct construct of companionship.
In later life, spouses and friends serve as the primary companions, with spouses dominating this role (Connidis & Davies, 1992). Living arrangements that include additional household members, particularly a spouse, offer the most proximal companions and increase the opportunities to engage in companionate activities. Older adults who live alone do not have this in-home resource and, thus, rely more heavily on friends (Alwin, Converse, & Martin, 1985). Physical proximity becomes an increasingly relevant factor of friendship in older age; thus, many nonrelative friends of the elderly reside nearby (e.g., within walking distance or within the same neighborhood or city; Cantor, 1979; Chown, 1981). Because friends are highly ranked on the companionate roster and neighbors may be considered as friends, the neighborhood social environment may play a key role in fostering or hindering companionship, especially for those residing alone.
The Neighborhood Social Context
Compared to their younger counterparts, older adults are more likely to remain embedded within the neighborhood environment throughout the day and rely on this context as a source of social integration (Glass & Balfour, 2003; Krantz-Kent & Stewart, 2007; Oh, 2003; Robinson & Godbey, 1997). In fact, geographic proximity has been identified as a predictor of who is considered a companion (Connidis & Davies, 1990). Similarly, having social network members who reside in close proximity has been positively associated with older adults’ perceived social connectedness, including companionship (Ashida & Heaney, 2008). Yet, to our knowledge, the impact of the larger neighborhood social environment on companionship has not been examined.
Examinations of the association between neighborhood context and social outcomes among older adults were prevalent in the late 1970s and early 1980s (Carp, 1976; Kasl & Rosenfield, 1980; Lawton, 1976; Lawton, Nahemow, & Yeh, 1980). Although there has been a resurgence in neighborhood-based research, particularly in relation to crime and disadvantage (e.g., Sampson, Raudenbush, & Earls, 1997; Wen, Browning, & Cagney, 2003; Wilson, 1987), few empirical studies have investigated social outcomes such as companionship or friendship. Krause and Thompson have provided a more recent point of departure for research on neighborhood characteristics and indicators of social well-being. Their work on neighborhood physical deterioration identified an increase in social isolation and a decrease in anticipated support with declines in the physical conditions of the neighborhood (Krause, 1993; Thompson & Krause, 1998). Furthermore, those who lived alone were more vulnerable to the negative impact of residential deterioration (Thompson & Krause, 1998). Krause (1993) determined that deteriorated neighborhoods promoted distrust among residents, which was in turn associated with social isolation. Although a general sense of distrust in people (rather than distrust of neighbors specifically) was assessed, this finding has implications for neighborhood-level social cohesion.
Neighborhood social cohesion is defined in terms of neighbors’ mutual trust, solidarity, connectedness, shared values, and support (Kawachi & Berkman, 2000; Leventhal & Brooks-Gunn, 2000; Sampson, Morenoff, & Earls, 1999; Sampson et al., 1997). Such cohesion fosters a host of favorable outcomes, including outdoor activities, access to services and amenities, healthy behaviors, and heightened feelings of safety (Echeverria, Diez-Roux, Shea, Borrell, & Jackson, 2008; Franzini, Caughy, Spears, & Fernandez Esquer, 2005; Kawachi & Berkman, 2000; McNeill, Wyrwich, Brownson, Clark, & Kreuter, 2006; Mendes de Leon et al., 2009; Veenstra, 2000). Furthermore, individuals in socially cohesive neighborhoods are more likely to help and support each other (McNeill et al., 2006), which may be particularly beneficial for older adults.
Importantly, social cohesion is likely to promote face-to-face interaction within the neighborhood. Because companionship requires contact, neighborhoods that are socially cohesive may offer a venue for the development of such relationships. Few empirical investigations of neighborhood social cohesion have examined social outcomes beyond its definition (likely because the definition captures such a rich array of social factors, i.e., mutual trust, solidarity, connectedness, shared values, and support). One exception is Stafford, McMunn, and Volgi’s (2011) work, which found that neighborhood social cohesion significantly predicted friendship quality, with those in cohesive neighborhoods reporting fewer negative aspects of friendships. Although companionship implies a high level of relationship quality, not all friendships are companionate relationships. Therefore, an individual may report friendships that are high in quality and still lack companionship. Furthermore, Stafford et al.’s (2011) study found that social cohesion reduced negative friendship attributes but did not promote positive aspects. This work is quite informative in linking the neighborhood context to individual-level social interactions; however, further work is necessary to affirm the benefits of socially cohesive environments.
Given that social cohesion fosters social interaction, we hypothesize that neighborhood social cohesion will significantly predict perceptions of companionship regardless of living arrangement (living alone and residing with others). Additionally, because those who live alone must pursue human companionship outside of the household, we expect that the relationship will be more pronounced among these older adults.
Method
Participants
Data were drawn from the Neighborhood, Organization, Aging and Health study (NOAH), which surveyed approximately 1,500 adults aged 65 and over residing in 80 selected neighborhood clusters (each including two to three census tracts) in Chicago, Illinois. The sample frame consisted of households with an adult aged 65 years or older. Participants were interviewed via telephone between August 2006 and September 2007. The weighted response rate was 44.3%. This response rate is in keeping with the current trends for telephone surveys. We note that response rates have declined over the past two decades. For example, Centers for Disease Control (CDC) response rates decreased from 71.4% in 1993 to 48.9% in 2000 and response rates to Pew public opinion surveys decreased from 36% in 1997 to only 9% in 2012, without substantial impact on the results (Langer, 2003; Pew Research Center, 2012; for a discussion of reasons for the declines, please see Kempf & Remington, 2007). All participants provided verbal consent, and the study was approved by the institutional review boards of NORC and the Division of Biological Sciences at the University of Chicago. Analyses were conducted using the data from the 1,497 respondents with information on living arrangement (alone vs. with others). Participants were classified as living alone if they reported that zero people, excluding themselves, lived in their household. Those who reported that one or more additional individuals lived in their household were labeled as living with others.
Measures
Neighborhood Social Cohesion
The neighborhood social cohesion items were adapted from the Project on Human Development in Chicago Neighborhoods (PHDCN; Earls & Buka, 1997). Participants rated their agreement (1 = Strongly agree, 4 = Strongly disagree) with the following statements: 1. “This is a close-knit neighborhood,” 2. “People around here are willing to help their neighbors,” 3. “People in this neighborhood generally don’t get along with each other,” and 4. “People in this neighborhood can be trusted.” The items with a positive valence were reverse-coded (i.e., all except Item 3), then the items were summed to create a social cohesion scale (α = .70). Due to the skewness of the distribution (M = 12.4, SD = 2.5), the variable was dichotomized. Because the mean of the scale was 12.4 and the median was 13, those scoring 4–12 on the scale were labeled as 0 (low social cohesion) and those scoring 13–16 on the scale were labeled as 1 (high social cohesion).
Companionship
Companionship was assessed with the question “How often do you feel that you lack companionship?” The respondents answered using the scale 1 = Hardly Ever (or Never), 2 = Some of the time, 3 = Often. For the current analyses, a binary variable was created, with Some of the time and Often = 0 and Hardly Ever (or Never) = 1. This question was one of the items in the 3-item loneliness scale developed by Hughes, Waite, Hawkley, and Cacioppo (2004) for use in telephone surveys. There is conceptual justification for an analysis of the companionship item separately, as perceptions of companionship may contribute something different from reports of loneliness per se.
Covariates
Participant-reported age, gender, highest education level completed, race, marital status (married, separated, divorced, widowed, single), length of neighborhood residence (calculated as proportion of life spent in the neighborhood based on participant-reported age and years and months spent in the neighborhood), and self-rated health were included as covariates. Because the linearity assumption of age was met (verified using the lintrend command in STATA), it was entered as a continuous variable in the model.
Analysis
To identify potential differences in the characteristics of those who lived alone and those who resided with others, Pearson’s Chi-square test for categorical variables and t-tests for continuous variables were conducted. Then, controlling for the participant characteristics of age, gender, race, education, marital status, length of neighborhood residence, and self-rated health, logistic regression was performed, with neighborhood social cohesion predicting companionship. Separate models were run for those who lived alone and those who lived with others to examine the proposed relationship for each group. The analyses were conducted using STATA 11, and statistical significance was defined as p < .05.
Results
The characteristics of the sample as a function of living arrangement are presented in Table 1. Approximately 43% of the sample lived alone. Adults who lived alone were older (75.35 [SD = 7.16] vs. 72.73 [SD = 6.30]) and more likely to be female (75.54% vs. 61.39%) than those who lived with others. The group that lived alone included a greater percentage of Caucasians and lower percentage of African Americans and Hispanics than the group that lived with others. As would be expected, the marital status of the two groups differed significantly. Few of those who lived alone were married (1.69% compared to 54.07% of those who lived with others), whereas this group evidenced higher rates of being divorced, widowed, or single. No education differences were found, with a majority of the sample having completed high school or beyond. Those who lived alone and those who lived with others did not significantly differ on length of neighborhood residence, self-rated health, or reported levels of neighborhood social cohesion. However, those who lived with others displayed significantly greater perceived companionship than those who lived alone (χ2 = 47.86, df = 1, p < .001).
Characteristics of Older Adults, by Living Arrangements.
aFrequency of data missing for this variable = 12.
bFrequency of data missing for this variable = 6.
cFrequency of data missing for this variable = 5.
dFrequency of data missing for this variable = 19.
eFrequency of data missing for this variable = 3.
fFrequency of data missing for this variable = 87.
gFrequency of data missing for this variable = 3.
Table 2 presents the logistic regression results by living arrangements. For those who lived alone, the overall model was significant, Likelihood ratio χ2 = 33.62, df = 20, p < .05. Controlling for demographic variables, neighborhood social cohesion significantly predicted companionship among those who lived alone, β = 0.37, SE β = 0.18, p < .05. Furthermore, for a one-unit increase in neighborhood social cohesion, the odds of reporting companionship increased by a factor of 1.45 (95% confidence interval [CI] [1.02, 2.04]). As a model check, the Hosmer-Lemeshow goodness-of-fit test indicated a p value of .91, suggesting good model fit. For those who lived with others, although the overall model was significant, Likelihood ratio χ2 = 76.00, df = 20, p < .001, social cohesion was not a statistically significant predictor of companionship after controlling for demographic variables, β = 0.17, SE β = 0.16, n.s. The covariates in the model also suggested some interesting associations. Among those who lived alone, male gender negatively predicted perceived companionship (β = −0.88, SE β = 0.21, p < .001). Among those who lived with others, higher education, both at the 2-year degree (β = 1.42, SE β = 0.70, p < .05) and post-college level (β = 2.18, SE β = 0.70, p < .01), positively predicted companionship, whereas poor/fair self-rated health (β = −0.61, SE β = 0.18, p < .01) and widowhood (β = −0.68, SE β = 0.21, p < .01) negatively predicted companionship.
Logistic Regression Results for Social Cohesion Predicting Companionship Controlling for Demographics, by Living Arrangements.
β = coefficient; SE β = standard error of the coefficient; χ2 = likelihood ratio chi-square; df = degrees of freedom.
*p < .05. **p < .01. ***p < .001.
Discussion
The study’s first hypothesis, that neighborhood social cohesion predicts perceived companionship for older adults in both living arrangements, was only partially supported. The second hypothesis, that social cohesion is a stronger predictor of companionship among those in one-person households, was supported; neighborhood social cohesion increased perceptions of companionship among those who lived alone.
In this study, those who lived alone reported less companionship than those with additional household members. However, among the older adults in this potentially vulnerable group of solitary residents, those who reported greater neighborhood social cohesion perceived a higher level of companionship. By definition, socially cohesive neighborhoods represent a context of social connectedness, which may facilitate close relationships and shared activities. Such environments are especially advantageous for those who do not reside with potential companions.
The present results are in line with Stafford et al.’s (2011) findings that social cohesion is associated with increased friendship quality, as companionate relationships are typically high in quality. Furthermore, Thompson and Krause (1998) found that solitary residents were more vulnerable to maladaptive neighborhood conditions. The current work expands upon this consideration of living arrangement in the neighborhood context to evidence that those who live alone may particularly profit from the social benefits of positive local environments.
The loneliness and social support literatures have suggested that the quality (as opposed to quantity) of relationships is of paramount importance for individuals’ social life (see Pinquart & Sorensen, 2001; Sarason, Sarason, & Pierce, 1994 for reviews). Neighborhood social cohesion is an indicator of the quality of the social environment (and of the neighborhood more generally, see Macintyre & Ellaway, 2000; Macintyre, Ellaway, & Cummins, 2002). Therefore, the current work suggests that high-quality environments lead to social relationships marked by high quality, particularly for those who may otherwise lack such relationships.
Neighborhood social cohesion did not significantly predict the perceived companionship of individuals residing with others. As mentioned above, household members may provide opportunities for regular companionate activities and spouses are the predominate companions of later life (Connidis & Davies, 1992). This is not surprising, given that the primary reason individuals marry in older age is for a source of companionship (McKain, 1972). In-home relationships may account for such a significant portion of companionship that the effects of the greater social environment may be relatively small.
Although the present study contributes to the nascent literature on companionship in context, several limitations should be noted. First, the current study measured companionship using a single item that indicated respondents’ perceived frequency of a lack of companionship, as defined by the respondent. We posit that this measure is robust; however, it does not provide information concerning specific companions or sources of companionship (e.g., human vs. animal companions) or the quality of companionate relationships or activities. Furthermore, it is possible that participants rated their perceived lack of companionship based on number of companions rather than a desire for more companionship. A question that asks whether participants desire more (or less) companionship would aid in clarifying this response. Second, the neighborhood measurement considered one aspect of context; a further examination of additional characteristics that may impact social interaction and companionship, such as levels of safety, walkability, and resources, should be conducted. Third, living arrangements are marked by greater complexity than the distinction between living alone and living with others; however, the reasons for the current living arrangements and the relationships among household members were not examined. Fourth, advances in technology facilitate long distance relationships and diversity in social networks (Hampton, Sessions, Her, & Rainie, 2009). Such technological impacts were not examined in the current study; however, distance remains a hindrance of relationship survival, satisfaction, and quality and face-to-face contact is the most frequent form of communication in close relationships (Hampton et al., 2009; Sahlstein, 2006). Future research would benefit from a more in-depth consideration of the multifaceted constructs of companionship and living arrangements within a broader conceptualization of the neighborhood context and with attention to the impact of new technology that facilitates social interaction.
The present research suggests that companionship merits greater theoretical and empirical attention. The influence of the neighborhood social process of cohesion on reports of companionship implies that companionship is not driven solely by dyadic connections. The energy of the social environment, from street activity to a nod from a neighbor, may create a sense of embeddedness and belonging that results in companionate benefits. Such environments may foster neighbor interaction, friendship, and companionate activities. The impact of macro-level social interactions—both their influence on individual well-being and the extent to which they shape expectations of companionship—requires further exploration. This work extends to the conceptualization and measurement of companionship. Assessments of companionship are often folded into the construct of loneliness (for notable exceptions, see e.g., Baldassare, Rosenfield, & Rook, 1984; Connidis & Davies, 1990, 1992; Krause, 2010); thus, it is possible that its independent determinants are often masked or generally unexplored. Further analyses of companionship as an independent construct would benefit the field of aging research, particularly that focused on the neighborhood context in which adults grow older.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The NOAH project was funded by a grant from the National Institute on Aging of the National Institutes of Health (Grant No. R01 AG022488-01 to K.A. Cagney). The current study was conducted with a postdoctoral fellowship award from the National Institute on Aging of the National Institutes of Health (Grant No. 2 T32AG000243-16A1 to L. Bromell).