Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Sep 1.
Published in final edited form as: Soc Sci Med. 2021 Jul 27;285:114274. doi: 10.1016/j.socscimed.2021.114274

Volunteering and Health: The Role of Social Network Change

Noah J Webster 1, Kristine J Ajrouch 1,3, Toni C Antonucci 1,2
PMCID: PMC8416937  NIHMSID: NIHMS1732754  PMID: 34390978

Abstract

Objectives:

Levels of volunteering may differentially influence multiple dimensions of health among older adults. Further, increasing evidence indicates social networks represent critical bridging and bonding contexts for the volunteering-health link. This study examines two research questions: 1) does volunteering level (low, moderate, high) influence physical and mental health in the same ways? and 2) does social network change moderate this link?

Methods:

Data come from Waves 1 and 2 of the longitudinal Social Relations Study (n=556) collected in 1992 and 2005 and the sub-sample of adults age 50 and older at Wave 2. Regression analyses predicting self-rated health and depressive symptoms were conducted to examine main effects of volunteering and moderating effects of social network change.

Results:

Volunteering at a moderate level (101–300 hours/year) was associated with fewer depressive symptoms compared to those not volunteering. Social network change moderated the association between volunteering and self-rated health. Among those reporting a decrease in the proportion of non-family in their network (decrease in bridging), a moderate level of volunteering was associated with better self-rated health.

Conclusion:

While moderate levels of volunteering are associated with better mental health, the link to physical health is only present in the context of decreasing network bridging. Identifying specific circumstances under which volunteering is beneficial is critical for developing interventions to promote health for all, including those in mid and later life.

Keywords: volunteering, physical health, mental health, social networks, older adults

INTRODUCTION

The consequences of volunteering for both physical and mental health overwhelmingly document beneficial effects (Andersen et al., 2014; Jenkinson et al., 2013; Kim, Whillans, Lee, Chen, & VanderWeele, 2020; Li & Ferraro, 2005; Morrow-Howell, Hinterlong, Rozario, & Tang, 2003; Musick, Herzog, & House, 1999; Thoits & Hewitt, 2001). The United Nations defines volunteering as activity not undertaken for financial gain, engaged in of one’s own free will, that benefits others (United Nations, 2001). It is especially important to understand the volunteering-health link among older adults given role changes often experienced at this time in the life course (e.g., retirement, children leaving home). Role changes can result in both more time and greater need (e.g., due to increased risk of social isolation and activity decline) for volunteering, which may amplify the physical and mental health benefits. Although positive effects of volunteering are widely acknowledged, volunteering may not benefit all older people equally (Martinson & Minkler, 2006; Musick, Herzog, & House, 1999). It is unclear whether similar levels of volunteering are equally beneficial for both physical and mental health (Anderson et al., 2014; Jenkinson et al., 2013). Recent studies have also suggested varying health benefits of volunteering across social contexts (e.g., marital status; Carr, Kail, Matz-Costa, & Savit, 2018) and research has found differences in rates of volunteering by race, class and gender (Musick & Wilson, 2007). The role that social networks play in influencing the association between volunteering and multiple dimensions of health, however, has received minimal attention.

Social networks signify critical contexts for volunteering. The social ties one has to others are multidimensional, including both bridging (i.e., connections facilitated by diversity of social ties as indicated by network size and composition) and bonding (i.e., connections facilitated through frequent contact and geographic proximity) elements. Of noteworthy significance is that networks may change, i.e. grow or diminish in size, become more or less diverse in composition, and promote or discourage face-to-face contact over time. Recently, Ajrouch and colleagues (2014) found that change in social networks influence the likelihood and frequency of volunteering. We build on those findings in this study to examine whether volunteering at varying levels interacts with bridging and bonding elements of social network change to influence physical and mental health.

Volunteering-Health Link

The benefits of volunteering extend to multiple dimensions of health (see Anderson et al., 2014 for a review). Though the relationship between volunteering and health may not always be linear, and indeed may be bidirectional (Thoits & Hewitt, 2001; Luoh & Herzog, 2002), evidence suggests that even small amounts of volunteering can lead to positive health outcomes (Li & Ferraro, 2005; Morrow-Howell et al., 2003; Luoh & Herzog, 2002; Kim et al., 2020). Yet, the effects of volunteering on physical and mental health may vary by intensity (Jenkinson et al., 2013). There appears to be a point of diminishing returns given inconsistent findings regarding the positive effects of volunteering at higher intensity levels (Li & Ferraro, 2005; Morrow-Howell et al., 2003; Windsor, Anstey, & Rodgers, 2008; Carr, Kail, & Rowe, 2018; Carr et al., 2018). Next, we briefly review relevant pathways to ground our hypotheses for how different levels of volunteering may be associated with both physical and mental health.

In terms of physical health, volunteering has been linked with higher levels of physical activity, which can lead to better outcomes through cardiovascular mechanisms (Anderson et al., 2014; Barron, Tan, Yu, Song, McGill, & Fried, 2009; Fried et al., 2004; Jenkinson et al., 2013; Kim et al., 2020). In terms of mental health, volunteering can lead to development of psychological resources (e.g., self-esteem, self-efficacy) that are helpful in managing stress (Lin, Ye, & Ensel, 1999) and can promote positive mood and affect (Kim et al., 2020) as well as reduce negative affect (Musick & Wilson, 2003). There is growing consensus that engaging in moderate levels of volunteering (e.g., ~100 hours/year or 2–3 hours/week) is needed to experience mental health benefits (Anderson et al., 2014; Carr et al., 2018; Kim et al., 2020). However, the evidence is less definitive regarding what threshold is necessary to experience physical health benefits. Increasing evidence suggests a similar (moderate) threshold as with mental health (Anderson et al., 2014; Burr, Tavares, & Mutchler, 2011; Carr, Kail, & Rowe, 2018). Additional research is needed to examine dimensions of health concurrently within the same sample to test whether the same (or different) levels of volunteering are associated with both physical and mental health outcomes. Further, recent studies (Carr et al, 2018; Carr, Kail, & Rowe, 2018; Kim et al., 2020) have been restricted to comparing the differential health impact of lower (<100 hours/year) and higher (100+ hours/year) levels of volunteering due to small sample sizes in the higher category. We build upon this recent research by examining the higher category as two distinct groups, those volunteering at a moderate (101–300 hours/year) and higher (300+ hours/year) level.

Theoretical Framework

We conceptualize social networks as an important resource, and use the Convoy Model of Social Relations as the guiding framework for the present study (Antonucci, 2001; Antonucci, Ajrouch & Birditt, 2014; Kahn & Antonucci, 1980). Convoys represent an assembly of family and friends who surround the individual, and are available in times of need. Convoy structure encompasses multiple dimensions or characteristics of a social network including size, composition, contact frequency and geographic proximity, providing a unique ability to consider bridging and bonding elements of social networks. Convoys are also thought to be dynamic and lifelong, changing in some ways, but remaining stable in others, across time and situations. The ways in which networks change may signify critical insights into potential avenues of sociability including integration, active engagement, and overall well-being. The systematic study of the multi-dimensional and dynamic nature of social networks has been shown to yield important insights for better understanding the likelihood and frequency of volunteering in later life (Ajrouch, Antonucci, & Webster, 2014). In the present study we extend this framework to better understand the role of social network change in moderating the volunteering-health link.

Specifically, we examine how changes in bridging and bonding elements of social networks moderate the association between volunteering and health. Paik and Navarre-Jackson (2011) describe each element. Bridging elements of social networks promote connections as a result of the increasing diversity of social ties. This includes whether networks grow in size through the development of new relationships, i.e., inclusion over time of new people in one’s social convoy. This can also include changes in network composition, such as changes in the average age of network members and extent to which specific relationship types are represented in the network (e.g., family members, friends). For older adults, an increase in bridging (i.e., diversity in ties) in terms of composition would be driven by inclusion of new network members different from (or discontinuation of ties with those similar to) the individual. This would be indicated by a decrease in the average age of network members and an increase in network proportion of non-family members.

Bonding elements, on the other hand, promote connections as a result of spending time with network members. This may be inferred from increases in geographic proximity (i.e., living closer to) and contact frequency with network members. In testing interactions between changes in bridging and bonding elements with different levels of volunteering, we seek to further specify how the dynamic nature of social networks influences the volunteering-health link. Though previous research shows that volunteering can facilitate access to social resources (Jenkinson et al., 2013; Musick & Wilson, 2003), and through this pathway help reduce social isolation (Mirowski & Ross, 1989) and promote social integration (Moen, Dempster-McClain & Williams, 1992), we conceptualize social networks as a resource representing critical contexts in which volunteering occurs. Social networks change over time (Antonucci, 2001; Kahn & Antonucci, 1980), and so we advance that such an investigation will add greater specificity to recent evidence of the role of social context (Carr et al., 2018) in shaping the volunteering-health link.

The Moderating Role of Social Networks in the Volunteering-Health link

Research suggests social context plays an influential role in determining who volunteers and benefits from volunteering. Social networks, in particular, have received systematic attention as exerting a potential moderating effect on the volunteering-health link among older adults (see Anderson et al., 2014 for a review). For instance, Morrow-Howell and colleagues (2003) found no evidence to support their hypothesis that older adults with lower levels of social integration would experience more positive effects from volunteering on their health. In contrast, more recent research suggests volunteering has a greater health benefit in the context of bridging elements (Carr, et al., 2018; Jang, Tang, Gonzales, Lee, & Morrow-Howell, 2018; Jiang, Hosking, Burns, & Anstey, 2019). For example, Jiang and colleagues (2019) found that the link between volunteering and life satisfaction over a four-year period was stronger among those who lost friends; that is, those who experienced less bridging in their networks over time. In contrast, Carr and colleagues (2018) found that loneliness is attenuated among new widows when volunteering two or more hours per week (i.e., a moderate level). The experience of widowhood sets the stage for a potential increase in bridging as widowhood reflects loss of a family relationship, and in turn, may result in an increasing proportion of non-family members within a widow’s network.

In the context of social network change exemplified by widowhood, volunteering may play a greater role in shaping mental health due to the widowed individual being positioned to rely more on non-family members as well as benefit from the potential of volunteering to facilitate connections with non-family, which can in and of itself have health benefits. In this study we broaden the examination of social networks as contexts critical to informing and influencing the volunteering-health link. We examine whether changes in bridging and bonding elements of social network change shape the extent to which low, moderate, and high levels of volunteering benefits those in later life.

Present Study

By attending to changes in bridging and bonding elements of social network change, we aim to test two competing hypotheses, substitution and amplification. The substitution hypothesis suggests that lower amounts of volunteering can be compensated for by resources found in social networks. It may be that network changes that lead to more bridging in later life, known to be beneficial for health (e.g., larger or more diverse networks), substitute for the absence of volunteering. Therefore, in the context of increasing bridging elements of social networks, we expect those who do not volunteer to have health on par with those who volunteer at low or moderate levels. In contrast, in the context of decreasing bridging elements we expect those volunteering at low or moderate levels to have better health outcomes compared to those not volunteering. Amplification, on the other hand, emphasizes a cumulative advantage position (Dannefer, 2020). Guided by this perspective we argue that those who engage in more volunteering are better able to leverage the advantages that come with increases in bonding elements of social networks (e.g., increases in geographic proximity of and contact frequency with network members). Therefore, in the context of increasing bonding elements of social networks, we expect those not volunteering to have worse health than those who volunteer at low or moderate levels. By contrast, in the context of decreasing bonding elements, we expect there to still be an association between volunteering and health, but not as strong. Due to less consensus regarding the beneficial role of high levels of volunteering, we expect that high levels of volunteering will have no effect on physical and mental health across all social network change contexts. These hypotheses are illustrated in Figure 1.

Fig. 1.

Fig. 1.

The main and interactive effects of volunteering and social network change on health.a

a Solid lines with arrows depict associations examined in the current study; Dashed lines illustrate and acknowledge the: a) potential effect of volunteering on social network change; b) reciprocal effect of health on volunteering; and c) the dynamic nature of volunteering and its bi-directional relaitonship with health outcomes.

Determining whether the benefits of low to moderate volunteering differentially influence physical and mental health in later life across multiple contexts of social network change may indicate why some who volunteer benefit from this activity, while others do not.

METHODS

Sample

Data for this study come from the longitudinal Social Relations Study collected by the Survey Research Center at the University of Michigan. The first wave of this regionally representative sample from the Detroit Metropolitan area (n=1,703) was collected in 1992 from those aged 8–93, with a 72% response rate. Wave 2, collected in 2005 consisted of 1,076 of the original respondents. Three hundred twenty had deceased, and the remainder refused, were lost to follow-up, or were unable to participate resulting in a response rate of 78%. Respondents aged 50 and older at Wave 2 were selected from the larger study to focus on mid and later life. The sample consisted of 556 adults who participated at both waves, ranging in age from 50 to 99. The Social Relations Study was approved by the Institutional Review Board at the University of Michigan.

Measures

Outcome Variables - Health.

Physical health assessed at both Waves 1 and 2, measured how respondents rated their health at the present time on a 5-point scale ranging from very poor=1 to excellent=5. Analyses predicting Wave 2 self-rated health included Wave 1 self-rated health as a covariate. Mental health was measured at both waves using the 20-item Center for Epidemiological Studies Depression (CES-D) scale (Radloff, 1977). Respondents reported the experience of depressive symptoms in the past week on a 4-point scale ranging from 0 (rarely/none of the time) to 3 (most of the time). Item scores were summed to create a total composite score for each wave with higher values indicating experience of more depressive symptoms (α = .88). Analyses predicting Wave 2 depressive symptoms included Wave 1 CES-D as a covariate.

Volunteering was measured only at Wave 2 using a two-part self-report item assessing both the likelihood (i.e., “Do you do any volunteer work?”) and frequency (i.e., “if yes, about how many hours per…(week, month, or year)”) of volunteering. Frequency responses were converted to a metric of hours per year to create a single continuous variable ranging from zero (no volunteering) to the maximum number of hours reported (2,080) per year. We then divided this variable into four categories similar to previous studies (Carr, Kail, & Rowe, 2018; Luoh & Herzog, 2002) to indicate increasing levels of volunteering, including 0=no volunteering (reference category); 1=low level of volunteering (1 to 100 hours/year); 2=moderate level of volunteering (101 to 300 hours/year); and 3) high level of volunteering (300+ hours/year).

Social Network Change.

The hierarchical mapping technique (Antonucci, 1986) was used at both waves to measure five network indicators: size, family and age composition, geographic proximity, and contact frequency. Respondents were shown a diagram containing three concentric circles and asked to nominate people in their lives based on varying levels of closeness and importance. Respondents were then asked questions about up to the first 10 people named in their network age 13 or older. A continuous change score was computed for each of the network indicators by subtracting the Wave 1 value from the one for Wave 2. Positive change scores indicate an increase and negative values a decrease in network characteristics. Two bridging elements were assessed, including: 1) Network Size, which indicates the total number of people the respondent included in her/his network diagram; and 2) Composition, measured as proportion non-family i.e. the percentage of the first 10 people named who are not immediate or extended family members, and average age calculated by averaging the ages of the first 10 network members. Two bonding elements were assessed, including: 1) Proportion proximate, which measured the geographic closeness of social network members, operationalized as the percentage of the first 10 network members living within a one-hour drive of the respondent; and 2) Average contact frequency, where respondents were asked to rate how often they are usually in touch with each of their network members on a 5-point scale ranging from irregularly=1 to everyday=5. We scale proportion non-family and proximate as percentages (0 to 100) when reporting descriptive statistics and proportions (0 to 1) when estimating regression models, which reduces rounding error for small coefficients.

Covariates.

Variables identified in the literature to be significantly associated with volunteering and health were included as covariates in addition to the Wave 1 score for each health outcome. Age (at Wave 2) calculated from birth date. Gender, coded as male=0; female=1. Race/ethnicity was assessed with a self-report item worded as: “Are you white, black, Native American, Asian, Hispanic, Other?” Responses were coded so that white=0 and non-white=1. Education was measured at Wave 2 with a self-report item worded as “What is the highest grade of school or year of college you have completed?” Descriptive statistics revealed 16 respondents with years of education ranging from 1–7 as outliers, and subsequently were recoded to equal 8 years. The final measure used in analyses therefore ranged from ≤8–17+ years. Employment status (at Wave 2): Respondents were asked how many hours they work in an average week. Responses were coded into three categories: not employed=1 (reference category); work part-time (less than 40 hours per week)=2; work full-time (40 or more hours per week)=3. Caregiver (at Wave 2) was measured as “do you currently provide any unpaid care to a parent (in-law), spouse, child, or other person?” (no=0; yes=1). Health Limitation (at Wave 1 and Change): At both waves respondents were asked if they were limited in any way because of their health (no=0; yes=1). A change score was computed by subtracting Wave 1 health limitation status from Wave 2, with positive scores an indication of becoming limited over time, negative values indicating no longer being limited, and zero no change.

Missing data:

We follow recommendations for reporting handling of missing data outlined by Sidi and Harel (2018). Among the 556 respondents age 50 and older at Wave 2 who completed both waves, 37 (6.7%) were missing data on one or more variables, and 27 of these were missing data on a single variable. Regarding variables of interest, the extent of missing data was: volunteering (W2 n=3); depressive symptoms (W1 n=19; W2 n=3: self-rated health (W1&2 n=1); social network (W1 and change n=0–7). We assume data are missing at random and ignorable (i.e., can be accounted for by measured variables and are not conditional on the missing data themselves). Therefore, we used full-information maximum likelihood (FIML) to account for missing data (Allison, 2002). We conducted a sensitivity analysis using list-wise deletion, which produced substantively identical results as the main analyses.

We excluded from the sample 455 respondents who would have been age 50 and older at Wave 2 but did not participate. We assume missing data due to attrition are ignorable, i.e., probability of participating in Wave 2 is not associated with outcomes after accounting for Wave 1 covariates such as age and self-rated health. Use of FIML with these additional respondents is not likely to contribute much information (Allison, 2002, p. 54). Nevertheless, we estimated models using FIML with these respondents included (n=1,011) and found results to be substantively identical to the results when not included (n=556).

Analysis Strategy:

To examine the association between volunteering and health we conducted linear regression analyses predicting each health outcome at Wave 2. First we conducted main effects models, with ten covariates including Wave 1 health, five social network variables at Wave 1, five social network change variables, and three dummy coded variables measuring low, moderate, and high volunteering, with no volunteering as the reference group. Next, to examine the moderating effect of social network change on the volunteering-health link we added interaction (i.e., product) terms into the main effects models. In total, 15 product terms were created, one for each of the five social network change variables by each of the three levels of volunteering. Social network change variables were mean centered prior to the creation of the product terms. Interaction terms were then entered into the model in clusters of three (one for each volunteering level) for each indicator of social network change. Interactions were determined to be significant if the interaction term was statistically significant (p<.05). We focused only on the p-value of specific interaction terms and not the significance of the change in overall model fit (R-square change) given our hypothesis that not all volunteering levels (e.g., high levels) would interact with social network change to predict health (Schad, Vasishth, Hohenstein, & Kliegl, 2020).

RESULTS

We present first a descriptive analysis of the sample, followed by findings that address our hypotheses. Table 1 provides the means and sample distribution of study variables examined from Wave 1 to 2. When variables from both waves were analyzed, we also include a change score as well as a test of whether or not the change over time was significant.

Table 1.

Descriptive Statistics

Wave 1 Wave 2 Δ Wave 1 to 2 Significance

M (SD) Range M (SD) Range M (SD) Range

Covariates
 Age 67.0 (11.8) 50 – 99
 Female (%) 60.1
 Non-white (%) 22.5
 Education (years completed) 13.5 (2.5) 8 – 17
 Employment status (%)
  Not employed/retired 60.2
  Works part-time 13.7
  Works full-time 26.1
 Caregiver (%) 18.3
 Health limitation (%) 15.5 33.1 0.17 (0.50) −1 – 1
Social network characteristics
 Network size 10.93 (5.92) 1 – 35 11.33 (7.03) 0 – 45 0.40 (7.00) −23 – 39 NS
 Average age 45.62 (9.06) 21.25 – 73.25 49.89 (10.10) 17 – 85 4.25 (10.13) −24.29 – 39.66 ***
 Average contact frequency 3.84 (0.58) 1 – 5 3.90 (0.54) 1.86 – 5 0.07 (0.66) −1.83 – 3 *
 Proportion non-family (%) 17.9 19.3 0.01 (0.24) −0.88 – 1 NS
 Proportion proximate (%) 75.9 72.1 −0.04 (0.26) −0.9 – 0.89 ***
Volunteering
 Hours Volunteer per year 88.5 (213.2) 0 – 2,080
 Volunteering level (%)
  None 65.9
  Low (1 to 100 hrs/year) 9.2
  Moderate (101 to 300 hrs/year) 15.7
  High (300+ hrs/year) 9.2
Health
 Self-rated health 3.93 (1.00) 1 – 5 3.71 (1.01) 1 – 5 −0.23 (1.08) −4 – 3 ***
 Depressive symptoms 8.95 (9.02) 0 – 51 7.73 (8.47) 0 – 57 −1.12 (9.86) −43 – 38 *

Notes.

a

Limited in any way because of health?

b

Percentage of network members that live within a 1 hour drive

c

Measured on 5-point scale (1=irregularly to 5=everyday); NS=W1 to 2 difference not significant

*

p<.05

**

p<.01

***

p<.001

The average age of the study sample at Wave 2 was 67.0 years (SD = 11.8). Approximately 60% of the respondents were women, and 23% were non-white. Respondents reported on average completing 13.5 years of education (SD=2.5). One quarter (26%) of the sample reported currently working full-time, 14% part-time, and 60% were not employed at Wave 2. Among the respondents who were not employed 73% reported that they were retired, 14% a homemaker, 8% permanently disabled, 3% unemployed, and 2% in another type of employment situation. At Wave 2, 18% of the sample identified themselves as a caregiver. Respondents who reported being limited by their health more than doubled from Wave 1 (16%) to Wave 2 (33%). Some social network characteristics changed very little over the twelve-year period (network size and proportion of non-family), while others were more dynamic (average age of network, proportion proximate, and average contact frequency). The means, standard deviations, and ranges for each of the five network characteristics at Wave 1, Wave 2 and change scores are presented in Table 1.

On average respondents reported volunteering 88.5 hours per year (SD = 213.2) with a range from 0 hours to as many as 2,080 hours per year. When translated into a monthly and weekly scale this indicates an average of 7.4 hours of volunteering per month or 1.7 hours per week. When hours per year were categorized into four levels just under two-thirds (66%) of the sample reported no volunteering; 9% reported a low level (100 hours a year or less) of volunteering; 16% reported a moderate level (between 101 and 300 hours per year) of volunteering; and 9% reported a high level (more than 300 hours per year) of volunteering. Within the low volunteering group, the average number of hours reported per year was 40.5 (SD = 22.6; Range: 1 – 96). In the moderate volunteering group, the average was 172.0 (SD = 59.9; Range: 104 – 260), and high volunteering group 625.0 (SD = 360.5; Range: 312 – 2080).

On average, the respondents’ rating of their physical health significantly declined from a mean of 3.9 (SD = 1.0) at Wave 1 to 3.7 (SD = 1.0) at Wave 2. By contrast, in terms of mental health, the number of depressive symptoms reported significantly declined from a mean of 9.0 (SD=9.0) at Wave 1 to 7.7 (SD=8.5) at Wave 2.

Are levels of volunteering differentially associated with physical and mental health?

We hypothesized that better health (both physical and mental) would be associated with both low and moderate levels of volunteering compared to those who reported no volunteering. Further we hypothesized that a high level of volunteering would not be associated with either physical or mental health compared to no volunteering.

Contrary to our hypothesis, analyses indicated there was no association between level of volunteering and physical (i.e., self-rated health). Specifically, we found no difference between those volunteering at a low (1 to 100 hours/year) or moderate (101 to 300 hours/year) level when compared to those who reported no volunteering. In support of our hypothesis, a high level of volunteering (300+ hours/year) compared to no volunteering was also not associated with better physical health (see Table 2, Model 1a).

Table 2.

Main and Interactive Effects of Volunteering and Social Network Change on Health

Main effects Main effects models (1) Significant interaction effect model (2)
Moderate volunteering × Proportion non-family change Self-rated health
(a)
Self-rated health
(b)
Depressive symptoms

Covariates b 1 (SE) B (SE) b (SE)

Intercept 1.56** 0.59 11.48* 5.39 1.58** 0.58
Age (W2) 0.00 0.00 −0.02 0.04 0.00 0.00
Female 0.13 0.07 −0.74 0.68 0.14 0.07
Non-white −0.14 0.09 −1.02 0.81 −0.14 0.09
Education (W2) 0.04* 0.02 −0.45** 0.15 0.04* 0.02
Not employed/retired (W2)-reference group - - - - - -
Employed part-time (W2) 0.12 0.11 −2.77** 1.04 0.12 0.11
Employed full-time (W2) 0.02 0.10 −0.22 0.98 0.03 0.10
Caregiver (W2) −0.10 0.09 1.49 0.86 −0.10 0.09
Health limitation (W1) −1.07*** 0.12 2.40* 1.06 −1.06*** 0.12
Health limitation change (W1 to 2) −0.83*** 0.08 2.55*** 0.76 −0.83*** 0.08
Self-rated health (W1) 0.27*** 0.04 - - 0.27*** 0.04
Depressive symptoms (W1) - - 0.33*** 0.04 - -
Social network characteristics (W1)
Network size 0.01 0.01 −0.16* 0.06 0.01 0.01
Proportion non-family −0.44* 0.19 −1.30 1.77 −0.43* 0.19
Average age 0.00 0.01 0.04 0.05 0.00 0.01
Proportion proximate 0.20 0.18 −2.05 1.70 0.21 0.18
Average contact frequency 0.08 0.09 0.80 0.83 0.07 0.09
Social network change (W1 to 2)
Network size change 0.00 0.01 −0.13* 0.05 0.00 0.01
Proportion non-family change 0.10 0.17 0.36 1.62 0.34 0.20
Average age change 0.00 0.00 0.01 0.04 0.00 0.00
Proportion proximate change 0.18 0.17 1.19 1.55 0.19 0.17
Average contact frequency change −0.03 0.07 0.56 0.68 −0.06 0.07
Volunteering level (W2)
None (0 hrs/year)-reference group - - - - - -
Low (1 to 100 hrs/year) 0.24 0.12 0.79 1.15 0.23 0.12
Moderate (101 to 300 hrs/year) 0.15 0.10 −2.14* 0.93 0.16 0.10
High (300+ hrs/year) 0.12 0.13 −1.89 1.18 0.13 0.13
Interaction effect
Low vol. × prop. family change - - - - 0.31 0.62
Moderate vol. × prop. non-family change - - - - −0.70* 0.34
High vol. × prop. family change - - - - 0.72 0.53

Total N (n missing) 556 (16) 556 (35) 556 (16)
R 2 0 39*** 0.24*** 0.39***
1

Presented in the table are the unstandardized regression coefficients (b) and the standard errors (SE)

*

p<.05

**

p<.01

***

p<.001

In terms of mental health, we again compared whether high, moderate, and low levels of volunteering compared to no volunteering were associated with depressive symptoms. As with physical health, analysis showed partial support for our hypotheses. However, in the case of mental health, volunteering at a moderate level (101–300 hours/year) compared to no volunteering was associated with reports of better mental health (i.e., fewer depressive symptoms; beta = −2.14; p<.05). Contrary to our hypothesis and similar to what was found for physical health, volunteering at a low level (1 to 100 hours/year) compared to no volunteering was not associated with mental health. Also, as was the case with physical health, a high level of volunteering compared to no volunteering was not associated with mental health, as hypothesized (see Table 2, Model 1b).

Does social network change moderate the volunteering-physical and mental health link?

We examined whether bridging and bonding elements of social network change influenced the association between volunteering and physical and mental health. We found partial support for our hypotheses. One bridging element of social network change, proportion non-family, did interact with volunteering to influence physical health. Specifically, we found that changes in the network proportion of non-family significantly interacted with volunteering at a moderate level (101 to 300 hours/year) compared to no volunteering to predict physical health (see Table 2, Model 2).

To examine the nature of this interaction we conducted simple slopes analysis using Wald chi-squared tests. This included testing the effect of moderate volunteering (versus none) at one standard deviation below the mean proportion non-family change (decrease of 23%) and one standard deviation above the mean (increase of 26%). See Figure 2. Among respondents who experienced a decrease in their network proportion of non-family (i.e., decrease in bridging), volunteering at a moderate level was associated with better physical health compared to those not volunteering. In contrast, among those who experienced an increase in proportion non-family (i.e., increase in bridging) the association between volunteering at a moderate level (compared to not volunteering) and physical health was not significant. The nature of this interaction indicates a substitution effect as hypothesized for bridging elements of social network change. No bonding elements of social network change were found to interact with volunteering to predict either mental or physical health.

Fig. 2.

Fig. 2.

Interaction effect of network proportion non-family change and moderate (vs. no) volunteering on self-rated health.

Post-hoc analysis:

We replicated all analyses using cross-sectional data (Wave 2 only) to check the robustness of the findings observed longitudinally, and the cross-sectional analyses produced near identical results. One difference was the p-value for the low level of volunteering-self-rated health association was just under 0.05 in the cross-sectional model, whereas it was just over this threshold in the longitudinal model. The size of the coefficient was the same in both models. One potential reason for this difference is the cross-sectional model had increased power due to fewer variables (i.e., no social network and other change variables). Second, in the cross-sectional analysis we did not find a significant interaction effect between network proportion non-family and volunteering on self-rated health. Third, we found a significant interaction effect between network average age and volunteering at a moderate level (compared to none) on self-rated health. Simple slopes analysis indicates that among those with an older average network age (i.e., less bridging), volunteering at a moderate level was associated with better self-rated health compared to those not volunteering. In contrast, among those with a younger average network age (i.e., more bridging), there was no link between volunteering and self-rated health. Even though this cross-sectional finding involves a different network characteristic from that found to be significant in the longitudinal model, there are similarities. First, both are network composition characteristics that indicate bridging. Second, both operate in the same direction (i.e., moderate the volunteering-physical health link) and support our study hypothesis in that increases in or more bridging network resources have a substitution effect on the volunteering-health link.

DISCUSSION

This study extends current research on the importance of social network contexts for understanding the volunteering-health link. To begin, we examined whether levels of volunteering are differentially associated with health, and in doing so considered two dimensions of health, physical and mental health. Additionally, we examined whether changes in bridging and bonding elements of social networks influence this association.

Volunteering-Health Link

While past research has examined whether or not volunteering is associated with health, the current study considered this association in greater detail and across multiple health outcomes. Specifically, we extend recent studies (Carr et al, 2018; Carr, Kail, & Rowe, 2018; Kim et al., 2020), which have distinguished lower (<100 hours/year) and higher (100+ hours/year) levels of volunteering. We examined the higher level as two distinct groups, those volunteering at a moderate (101–300 hours/year) and higher (300+ hours/year) level.

High levels of volunteering:

The mechanisms through which volunteering facilitates better physical and mental health discussed prior provide potential insights into why high levels may have no benefit. In terms of physical health, those able to volunteer at a high level may already be in good physical health, which may facilitate this high level of engagement. As a result, volunteering may do little to further elevate health status. This idea is supported in a previous study which found that volunteers in fair health reported greater improvements in stair climbing speed compared to those in excellent and very good health (Barron et al., 2009). In terms of mental health, volunteering has been shown to buffer cortisol stress reactivity (Han, Kim, & Burr, 2018). Therefore, it is also plausible that when volunteering itself becomes a source of stress, which may occur at higher levels and has been documented across multiple studies (Huynh, Winefield, Xanthopoulou, & Metzer 2012; Scherer, Allen, & Harp, 2016), this accompanying stress may counteract the beneficial (i.e., protective) aspects of the experience.

Moderate volunteering and mental health:

We found that volunteering at a moderate level (101 to 300 hours a year) was associated with fewer depressive symptoms compared to those not volunteering. Further, there was no difference in depressive symptoms between those not volunteering and those volunteering at a low level (100 hours or less in a year). This finding adds to growing consensus in the volunteering literature that engaging in moderate levels is needed to experience mental health benefits because it may indicate actual engagement in the activity, i.e. involvement with other people. This level of engagement would require greater emotional and cognitive energy, and as a result could lead to greater psychological stimulation. Under these circumstances, volunteering may yield mental health benefits that are not present in the case of no or a low level of volunteering.

Links between levels of volunteering and physical health.

Contrary to our hypothesis, there was no association between volunteering and physical health. Though this finding is consistent with findings from a systematic review and meta-analysis focused on all adults conducted by Jenkinson and colleagues (2013), the lack of an association between low volunteering and physical health was especially surprising in our study given our focus on mid and later life. Further, research shows that even small increases in physical activity positively impact health (Lachman et al., 2018). There is less consistency in the literature regarding the association between volunteering level and physical health compared to mental health (Anderson, 2014), suggesting a need to continue identifying contexts under which volunteering matters for physical health, especially in mid and later life. The context of social network change appears to be one such circumstance that impacts the volunteering level-physical health association.

Social Network Change and the Volunteering-Health Link

It appears social network change that promotes bridging over time is a key beneficial context for the effects of volunteering on physical health. Among older adults with increasing proportions of close and important relationships with non-family members, those who volunteer at moderate levels have physical health on par with those who do not. In the context of increasing network diversity, which is often viewed as a resource in later life (Doubova, Perez-Cuevas, Espinosa-Alarcon, & Flores-Hernandez, 2010), volunteering does little to further improve physical health. This could suggest that having diverse network members is sufficient, and volunteering thus does not further improve physical health. It is important to again acknowledge that an increasing proportion of non-family in a social network could result from either the addition of non-family relationships (e.g., friends) or the loss of family relationships. Future research is needed to determine how these two unique situations operate similarly or differently to contextualize the volunteering-health link.

Next, we elaborate theoretical assumptions that may explain the benefits of volunteering for physical health when networks change to reflect less bridging over time, i.e., become more family-centric. In doing so we highlight both the drawbacks as well as potential benefits of experiencing social network change characterized by a growing proportion of family members within one’s social convoy in later life. This network change may result from loss or dissolution of non-family ties (e.g., friends), and is consistent with Jiang and colleague’s (2019) finding of a stronger volunteer-life satisfaction link among those who lost friends. In contrast, this same change can reflect the addition of new family members (e.g., grandchildren).

One explanation for this finding is that physical activity engaged in via volunteering may be particularly beneficial for those who have lost or discontinued interactions with non-family social ties. Older adults with less diverse and predominantly family-focused networks report more functional dependency/limitations (Doubova et al., 2010; Fiori, Antonucci, & Cortina, 2006). Declining network diversity has also been linked to social isolation and negative health outcomes (e.g., mental, physical, and functional; Fuller-Iglesias, Webster, & Antonucci, 2015; Doubova et al., 2010; Fiori, Antonucci, & Akiyama, 2008; Fiori, Antonucci, & Cortina, 2006). Engagement with people outside the family is also linked with physical activity (Cohn-Schwartz, 2020). From this perspective, volunteering plays a substituting role to offset the decline or lack of the other (bridging) resource. This finding is consistent with Anderson and colleagues’ (2014) argument that volunteering plays the biggest role in facilitating positive health outcomes among vulnerable older adults.

A second explanation for the importance of volunteering for physical health in later life when experiencing a decrease in network proportion non-family involves re-framing this network change from a deficit to an alternative resource context. It may be that when social networks undergo a decrease in proportion non-family this change serves as a useful context from which to benefit from a moderate level of volunteering. Indeed, this finding draws unique attention to the importance of family in later life, and allows us to specify conditions under which networks composed of increasing proportions of family incur positive benefits. In particular, this finding requires a critical reassessment of the growing body of literature that shows networks characterized by an increasing proportion of family in later life only indicate deficit, vulnerability, and isolation. This reframing requires an expansion of our original conceptualization of the substituting role of network bridging resources (i.e., increasing proportion non-family) to also include an amplifying role in the specific context of increasing proportion family.

A perspective that considers increasing family-centered networks as a mechanism that amplifies the effects of volunteering on health presents family as an important resource for older adults. An increase in the proportion of family in one’s social network may serve as a key outreach pathway for social engagement activities such as volunteering. Family members may serve as a bridge to volunteering activities. This could occur by way of instrumental support (e.g., suggestion of volunteering opportunities to engage in, ride to volunteer activities) as well as emotional encouragement (e.g., engaging in the activity together, asking about the experience, and encouragement to continue). Further, these individuals likely are in better physical health and thus would require higher (i.e., moderate) levels of volunteering to experience a benefit. These findings support the tenets of the Convoy Model (Antonucci, Ajrouch, & Birditt, 2014; Kahn & Antonucci, 1980), illustrating how attention to the complexity and dynamic nature of social relations yield situational specific health benefits.

Finally, findings from this study demonstrate that not all dimensions of social relations serve as influential contexts in the volunteering-health link, but the bridging element of network composition is. This result further demonstrates the utility of the Convoy Model of Social Relations in clarifying which aspects of social networks influence health outcomes at specific points in the life course. Although some aspects of networks change, these changes we found do not have equal impact on the link between volunteering and health.

Limitations and Future Directions

One potential limitation of this study is the age of the data (i.e., Wave 1 in 1992; Wave 2 in 2005). Many things have changed since 2005 including a rise in age friendly communities which may provide more volunteering opportunities for older adults as well as emphasize importance of social connections for health and longevity. We used the baseline (Wave 1) and first follow up (Wave 2) data as opposed to the more recent data available from the same longitudinal study (Wave 2 collected in 2005; Wave 3 collected in 2015) to maximize the larger sample size and more representative nature of the baseline data. Further, these data are unique in the level of detail they provide regarding social network structure and composition over time. They provide a rare opportunity to more fully understand the impact of social networks on the volunteer-health link with a sufficiently large sample. Future studies are needed with sufficiently large samples to determine if this study’s hypotheses are supported among more recent cohorts of older adults.

A second limitation is the lack of data available on the nature of volunteering activities. Older adults’ volunteer activities vary in multiple ways in addition to the number of hours engaged, including intensity of activity, duration of commitment, extent to which activities involve engagement with others, etc. (Morrow-Howell, 2010). These additional dimensions are likely to influence how volunteering is related to diverse health outcomes as well as the ways in which social networks moderate this link. Future studies are needed that incorporate both detailed data on social networks as well as volunteering activities. Qualitative methods (e.g., ethnography or semi-structured interviews) can also be used to help unpack the diverse range of volunteering experiences and help begin to identify intersections with other themes relevant to volunteering, e.g., facilitators, outcomes, and relevant influential contexts operating at macro- and meso-levels.

A third limitation is the lack of longitudinal data on volunteering. A more complete understanding of the volunteering-health link needs to take into account: a) the reciprocal effects of volunteering and health; and b) the effect of volunteering on social networks, which can in turn directly impact health. As noted prior and shown by others (e.g., Li & Ferraro, 2005) health plays a key role in determining who volunteers in later life. Also, it is likely (as depicted in Figure 1) that volunteering itself promotes bridging and bonding elements of social networks over time. Findings from analysis of data with two or more time points on volunteering, health, and social networks can guide efforts to integrate volunteering policy with more general aging and public health policies. This can help to both increase the volunteer opportunities for older adults as well as the number of older adults able to engage in such activities through health promotion. Additional research is also needed on how more macro-level contexts shape both agency to engage in volunteering as well as who benefits, e.g., economic, political forces, the digital divide, as well as environmental factors (e.g.., differences among those in urban v. rural areas). Doing so can help ensure equity in access to volunteer opportunities, and help create volunteering policy that does not perpetuate, but rather can help reduce existing health disparities.

Conclusions

The present research demonstrates that moderate levels of volunteering are associated with mental health independent of social network change. In contrast, moderate volunteering is only associated with better physical health among those experiencing a decrease in network bridging (i.e., decrease in network proportion of non-family). These findings suggest that bridging elements of social network change may serve as an important resource for integration, embeddedness, and ultimately engagement in various social roles. Further, these data suggest that the increasing presence of family in one’s network signifies a critical context in which volunteering may be especially important for promoting positive physical health. The significance of the family tie dynamic may be an amplifying mechanism to consider as potentially maximizing the health benefits of volunteering. Findings from this and other recent including those in mid and later life.

Highlights.

  • Moderate volunteering among older adults is associated with better mental health.

  • High levels of volunteering are not associated with either physical or mental health.

  • Volunteering benefits physical health only in the context of social network change.

  • As networks change, volunteering can help those who lose diversity in their ties.

  • Increasing presence of family in networks may help maximize volunteering benefits.

Acknowledgements:

A previous version of this paper was presented at the annual scientific meeting of the Gerontological Society of America. The authors thank the Life Course Development Program at the Institute for Social Research for comments made on earlier versions of this paper and Simon Brauer for data analytic assistance.

Declaration of interest: This work was supported by grants from the U.S. National Institutes of Health (R01MH066876 to T.C.A.; R01AG027021 to T.C.A.; and K01AG062754 to N.J.W.)

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

REFERENCES

  1. Ajrouch KJ, Antonucci TC, & Webster NJ (2016). Volunteerism: Social Network Dynamics and Education. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 71(2), 309–319. 10.1093/geronb/gbu166 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Allison PD (2002). Missing Data. Sage Publications. [Google Scholar]
  3. Anderson ND, Damianakis T, Kröger E, Wagner LM, Dawson DR, Binns MA, Bernstein S, Caspi E, & Cook SL (2014). The benefits associated with volunteering among seniors: A critical review and recommendations for future research. Psychological Bulletin, 140(6), 1505–1533. 10.1037/a0037610 [DOI] [PubMed] [Google Scholar]
  4. Antonucci TC (1986). Measuring Social Support Networks: Hierarchical Mapping Technique. Generations: Journal of the American Society on Aging, 10(4), 10–12. JSTOR. [Google Scholar]
  5. Antonucci TC (2001). Social Relations: An Examination of Social Networks, Social Support and Sense of Control. In Birren JE & Schaie KW (Eds.), Handbook of the Psychology of Aging (pp. 427–453). Academic Press. [Google Scholar]
  6. Antonucci TC, Ajrouch KJ, & Birditt KS (2014). The Convoy Model: Explaining Social Relations From a Multidisciplinary Perspective. The Gerontologist, 54(1), 82–92. 10.1093/geront/gnt118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Barron JS, Tan EJ, Yu Q, Song M, McGill S, & Fried LP (2009). Potential for Intensive Volunteering to Promote the Health of Older Adults in Fair Health. Journal of Urban Health, 86(4), 641–653. 10.1007/s11524-009-9353-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Burr JA, Tavares J, & Mutchler JE (2011). Volunteering and hypertension risk in later life. Journal of Aging and Health, 23, 24–51. 10.1177/0898264310388272 [DOI] [PubMed] [Google Scholar]
  9. Carr DC, Kail BL, Matz-Costa C, & Shavit YZ (2018). Does Becoming A Volunteer Attenuate Loneliness Among Recently Widowed Older Adults? The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 73(3), 501–510. 10.1093/geronb/gbx092 [DOI] [PubMed] [Google Scholar]
  10. Carr DC, Kail BL, & Rowe JW (2018). The Relation of Volunteering and Subsequent Changes in Physical Disability in Older Adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 73(3), 511–521. 10.1093/geronb/gbx102 [DOI] [PubMed] [Google Scholar]
  11. Cohn-Schwartz E (2020). Pathways From Social Activities to Cognitive Functioning: The Role of Physical Activity and Mental Health. Innovation in Aging, 4(3). 10.1093/geroni/igaa015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Dannefer D (2020). Systemic and Reflexive: Foundations of Cumulative Dis/Advantage and Life-Course Processes. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 75(6), 1249–1263. 10.1093/geronb/gby118 [DOI] [PubMed] [Google Scholar]
  13. Doubova (Dubova) SV, Pérez-Cuevas R, Espinosa-Alarcón P, & Flores-Hernández S (2010). Social network types and functional dependency in older adults in Mexico. BMC Public Health, 10(1), 104. 10.1186/1471-2458-10-104 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fiori KL, Antonucci TC, & Akiyama H (2008). Profiles of social relations among older adults: A cross-cultural approach. Ageing and Society, 28(2), 203–231. 10.1017/S0144686X07006472 [DOI] [Google Scholar]
  15. Fiori KL, Antonucci TC, & Cortina KS (2006). Social Network Typologies and Mental Health Among Older Adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 61(1), 25–32. 10.1093/geronb/61.1.P25 [DOI] [PubMed] [Google Scholar]
  16. Fried LP, Carlson MC, Freedman M, Frick KD, Glass TA, Hill J, et al. (2004). A social model for health promotion for an aging population: Initial evidence on the Experience Corps model. Journal of Urban Health: Bulletin of the New York Academy of Medicine, 81, 64–78. 10.1093/jurban/jth094 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Fuller-Iglesias HR, Webster NJ, & Antonucci TC (2015). The complex nature of family support across the life span: Implications for psychological well-being. Developmental Psychology, 51(3), 277–288. 10.1037/a0038665 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Han SH, Kim K, & Burr JA (2018). Stress-buffering effects of volunteering on salivary cortisol: Results from a daily diary study. Social Science & Medicine, 201, 120–126. 10.1016/j.socscimed.2018.02.011 [DOI] [PubMed] [Google Scholar]
  19. Huynh J-Y, Winefield AH, Xanthopoulou D, & Metzer JC (2012). Burnout and Connectedness in the Job Demands–Resources Model: Studying Palliative Care Volunteers and Their Families. American Journal of Hospice and Palliative Medicine®, 29(6), 462–475. 10.1177/1049909111430224 [DOI] [PubMed] [Google Scholar]
  20. Jang H, Tang F, Gonzales E, Lee YS, & Morrow-Howell N (2018). Formal volunteering as a protector of health in the context of social losses. Journal of Gerontological Social Work, 61(8), 834–848. 10.1080/01634372.2018.1476945 [DOI] [PubMed] [Google Scholar]
  21. Jenkinson CE, Dickens AP, Jones K, Thompson-Coon J, Taylor RS, Rogers M, Bambra CL, Lang I, & Richards SH (2013). Is volunteering a public health intervention? A systematic review and meta-analysis of the health and survival of volunteers. BMC public health, 13(1), 1–10. 10.1186/1471-2458-13-773 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Jiang D, Hosking D, Burns R, & Anstey KJ (2019). Volunteering benefits life satisfaction over 4 years: The moderating role of social network size. Australian Journal of Psychology, 71(2), 183–192. 10.1111/ajpy.12217 [DOI] [Google Scholar]
  23. Kahn RL, & Antonucci TC (1980). Convoys over the life course: Attachment, roles, and social support. In Baltes PB & Brim OG (Eds.), Life-span development and behavior (pp. 253–286). Academic Press. [Google Scholar]
  24. Kim ES, Whillans AV, Lee MT, Chen Y, & VanderWeele TJ (2020). Volunteering and subsequent health and well-being in older adults: an outcome-wide longitudinal approach. American Journal of Preventive Medicine. Advance online access 10.1016/j.amepre.2020.03.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lachman ME, Lipsitz L, Lubben J, Castaneda-Sceppa C, & Jette AM (2018). When Adults Don’t Exercise: Behavioral Strategies to Increase Physical Activity in Sedentary Middle-Aged and Older Adults. Innovation in Aging, 2(1), 1–12. 10.1093/geroni/igy007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Li Y, & Ferraro KF (2005). Volunteering and Depression in Later Life: Social Benefit or Selection Processes? Journal of Health and Social Behavior, 46(1), 68–84. 10.1177/002214650504600106 [DOI] [PubMed] [Google Scholar]
  27. Lin N, Ye X, & Ensel W (1999). Social support and depressed mood: A structural analysis. Journal of Health and Social Behavior, 40, 344–359. 10.2307/2676330 [DOI] [PubMed] [Google Scholar]
  28. Luoh M-C, & Herzog AR (2002). Individual Consequences of Volunteer and Paid Work in Old Age: Health and Mortality. Journal of Health and Social Behavior, 43(4), 490–509. 10.2307/3090239 [DOI] [PubMed] [Google Scholar]
  29. Martinson M, & Minkler M (2006). Civic Engagement and Older Adults: A Critical Perspective. The Gerontologist, 46(3), 318–324. 10.1093/geront/46.3.318 [DOI] [PubMed] [Google Scholar]
  30. Mirowsky J, & Ross C (1989). Social causes of psychological distress. New York: Aldine de Gruyter. 10.4324/9781315129464 [DOI] [Google Scholar]
  31. Moen P, Dempster-McClain D, & Williams R (1992). Successful aging: A life course perspective on women’s multiple roles and health. American Journal of Sociology, 97, 1612–1638. 10.1086/229941 [DOI] [Google Scholar]
  32. Morrow-Howell N (2010). Volunteering in later life: Research frontiers. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 65(4), 461–469. 10.1093/geronb/gbq024 [DOI] [PubMed] [Google Scholar]
  33. Morrow-Howell N, Hinterlong J, Rozario PA, & Tang F (2003). Effects of Volunteering on the Well-Being of Older Adults. The Journals of Gerontology, Series B: Psychological and Social Sciences, 58(3), S137–S145. 10.1093/geronb/58.3.S137 [DOI] [PubMed] [Google Scholar]
  34. Musick MA, Herzog AR, & House JS (1999). Volunteering and Mortality Among Older Adults: Findings From a National Sample. The Journals of Gerontology, Series B: Psychological and Social Sciences, 54B(3), S173–S180. 10.1093/geronb/54B.3.S173 [DOI] [PubMed] [Google Scholar]
  35. Musick MA, & Wilson J (2003). Volunteering and depression: The role of psychological and social resources in different age groups. Social Science & Medicine, 56, 259–269. 10.1016/s0277-9536(02)00025-4 [DOI] [PubMed] [Google Scholar]
  36. Musick MA, & Wilson J (2007). Volunteers: A social profile. Indiana University Press. 10.1353/sof.0.0309 [DOI] [Google Scholar]
  37. Paik A, & Navarre-Jackson L (2011). Social Networks, Recruitment, and Volunteering: Are Social Capital Effects Conditional on Recruitment? Nonprofit and Voluntary Sector Quarterly, 40(3), 476–496. 10.1177/0899764009354647 [DOI] [Google Scholar]
  38. Radloff LS (1977). The CES-D Scale: A Self-Report Depression Scale for Research in the General Population. Applied Psychological Measurement, 1(3), 385–401. 10.1177/014662167700100306 [DOI] [Google Scholar]
  39. Schad DJ, Vasishth S, Hohenstein S, & Kliegl R (2020). How to capitalize on a priori contrasts in linear (mixed) models: A tutorial. Journal of Memory and Language, 110, 104038. 10.1016/j.jml.2019.104038 [DOI] [Google Scholar]
  40. Scherer LL, Allen JA, & Harp ER (2016). Grin and bear it: An examination of volunteers’ fit with their organization, burnout and spirituality. Burnout Research, 3(1), 1–10. 10.1016/j.burn.2015.10.003 [DOI] [Google Scholar]
  41. Sidi Y, & Harel O (2018). The treatment of incomplete data: Reporting, analysis, reproducibility, and replicability. Social Science & Medicine, 209, 169–173. 10.1016/j.socscimed.2018.05.037 [DOI] [PubMed] [Google Scholar]
  42. Thoits PA, & Hewitt LN (2001). Volunteer Work and Well-Being. Journal of Health and Social Behavior, 42(2), 115–131. JSTOR. 10.2307/3090173 [DOI] [PubMed] [Google Scholar]
  43. United Nations. (2001). Support for volunteering: Report of the Secretary-General. Fifty-sixth session, item 120 of the provisional agenda. https://www.un.org/webcast/events/iyv/a56288.pdf [Google Scholar]
  44. Windsor TD, Anstey KJ, & Rodgers B (2008). Volunteering and Psychological Well-Being Among Young-Old Adults: How Much Is Too Much? The Gerontologist, 48(1), 59–70. 10.1093/geront/48.1.59 [DOI] [PubMed] [Google Scholar]

RESOURCES