RESEARCH AND PRACTICE

Giving to Others and the Association Between Stress and

Mortality
Michael J. Poulin, PhD, Stephanie L. Brown, PhD, Amanda J. Dillard, PhD, and Dylan M. Smith, PhD

In a seminal review published more than 20

Objectives. We sought to test the hypothesis that providing help to others
years ago, House et al. described the strong
predicts a reduced association between stress and mortality.
association between social connections and
Methods. We examined data from participants (n = 846) in a study in the
physical health.1 The researchers concluded Detroit, Michigan, area. Participants completed baseline interviews that
that socially isolated people, compared with assessed past-year stressful events and whether the participant had provided
those with strong social ties, were at substan- tangible assistance to friends or family members. Participant mortality and time
tially increased risk of mortality and morbidity. to death was monitored for 5 years by way of newspaper obituaries and monthly
In fact, the magnitude of the association be- state death-record tapes.
tween social isolation and mortality was com- Results. When we adjusted for age, baseline health and functioning, and key
parable to that for high blood pressure, smoking, psychosocial variables, Cox proportional hazard models for mortality revealed
and sedentary lifestyle, even after statistical a significant interaction between helping behavior and stressful events (hazard
ratio [HR] = 0.58; P < .05; 95% confidence interval [CI] = 0.35, 0.98). Specifically,
controls for other known risk factors such as
stress did not predict mortality risk among individuals who provided help to
baseline health. Despite the robustness of this
others in the past year (HR = 0.96; 95% CI = 0.79, 1.18), but stress did predict
effect, it remains unclear what aspects of the
mortality among those who did not provide help to others (HR = 1.30; P < .05;
social environment inuence physical health 95% CI = 1.05, 1.62).
outcomes. Conclusions. Helping others predicted reduced mortality specifically by
One hypothesized link between social con- buffering the association between stress and mortality. (Am J Public Health.
nections and health is that the social support 2013;103:16491655. doi:10.2105/AJPH.2012.300876)
people receive from their network of friends
and loved ones may buffer against the detri-
mental physical consequences of psychosocial
stress.2,3 Indeed, stressful life events have helpers health, even when there is statistical system, including oxytocin, prolactin, and en-
long been established to be a predictor of control for plausible confounds such as base- dogenous opioids, have known stress-reducing
increased mortality risk.4,5 However, the social line physical health and functioning or re- effects.20,25---29
support hypothesis has not been consistently ceiving support from others. For example, Laboratory and eld studies provide pre-
supported in empirical studies. Although some volunteering predicts increased self-rated liminary evidence consistent with the predic-
empirical studies suggest health benets of health and longevity.9---13 In a similar way, tion that providing help or support may act as
received social supportand at least 1 indicates providing aid to a relationship partner pre- a physiological stress buffer. Communicating
that these benets accrue via stress buffering6 dicts reduced morbidity and mortality.14---16 affection to a relationship partner has been
a meta-analytic review concluded that the Given the robust associations between sup- shown to predict reduced perceived stress,
overall relationship between receiving support port provision and health, it is possible that lowered baseline cortisol levels,30,31 and faster
and health outcomes may not be considered support provision may have stress-buffering recovery from peak cortisol levels following lab
signicant or generalizable.7(p352) This may effects even if the receipt of support does not. stressors.32 In a similar way, experimentally ma-
be why House, in 2001, concluded that after To date, health research has not explicitly nipulated helping predicts reduced cardiovascular
nearly 2 decades of empirical work, very little tested this hypothesis; however, providing help reactivity to and faster cardiovascular recovery
is known about how social connectivity, as to others has psychological and physiological from laboratory stressors (written communication
opposed to isolation, translates into physical correlates that may buffer against stress. For from Stephanie L. Brown, October 3, 2008). In
health outcomes.8 example, helping leads to improved mood,17,18 addition, eld studies indicate that engaging in
The failure of the social support hypothesis which itself may act as a stress buffer.19 In helping behavior may buffer the effects of stress-
to account for the links between social con- addition, caring for loved ones, in particular, related constructs on health-related outcomes. For
nectedness and health8 has prompted research may draw on the functioning of neural and example, engaging in helping behavior versus
on whether health may be associated with the hormonal mechanisms that support parenting not doing so predicts lessened associations be-
other side of social interactionsnamely, the behavior20---22that is, the caregiving behav- tween grief and subsequent depression,33 nan-
provision of help and support to others. Pro- ioral system.23,24 Several hormones and neu- cial difculties and mortality,34 and functional
viding help to others appears to promote the rochemicals associated with the caregiving limitations and mortality.35

September 2013, Vol 103, No. 9 | American Journal of Public Health Poulin et al. | Peer Reviewed | Research and Practice | 1649
 RESEARCH AND PRACTICE

Research to date has not specically exam- participants over a 5-year period by checking important potential confounds of the hypothe-
ined whether providing help or support to obituaries in 3 Detroit-area newspaper and sized associations between key study variables
others can buffer the associations between monthly death-record tapes obtained from the and mortality. These control variables included
psychosocial stress and physical health out- State of Michigan. If a participant died, the demographic and socioeconomic factors, self-
comes. We sought to do so by using survey surviving spouse was reinterviewed 3 months rated health and health behaviors, mental health
data from the Changing Lives of Older Couples after the deceased spouses death. These pro- variables (i.e., subjective well-being, depression,
(CLOC) study. The CLOC data set included cedures were approved by the institutional and anxiety), personality factors (i.e., Big Five
5-year survival data on a sample of 846 older review board of the University of Michigan, and personality traits,37 self-esteem, perceived con-
adults along with baseline measures of helping, all participants completed provided informed trol), and social support variables (i.e., social
past-year stressful events, and potential con- consent in writing. contact, received instrumental and emotional
founds (e.g., demographic and socioeconomic support).
factors, baseline health and well-being, per- Measures
sonality, and social support receipt), allowing Mortality data. Throughout the study, re- Analyses
for a test of the stress-buffering role of prosocial searchers updated each spouses status as Cox proportional hazard models tested the
behavior. We hypothesized that exposure to either widowed (1) or not widowed (0). When associations of helping and stress with survival
a recent stressful life event would moderate the researchers learned of the death of a partici- over time. Cox models test predictors of mor-
association between helping behavior and mor- pant, they coded the surviving spouse as tality risk over time while requiring minimal
tality and vice versa. That is, we predicted that widowed. When the last surviving spouse of assumptions about underlying distributions, and
helping behavior would most strongly predict a couple died, they recorded this as well. This yield regression parameters that can be anti-
reduced mortality among individuals exposed to variable was used as indicator of the deceased logged and interpreted as hazard ratios (HRs).
signicant stress compared with those not exposed. spouses death. In addition, the researchers Hazard ratios represent the degree of change in
created a gap variable that represented how mortality risk for a unit change in a predictor.
METHODS many months passed between the baseline For the present study, the rst author con-
interview and the follow-up interview of the ducted all analyses with the STCOX module in
The CLOC study was a prospective study of surviving spouse (at 6 months postdeath). We Stata version 11.0 (StataCorp, College Station, TX),
a 2-stage area probability sample. Participants used this variable as an indicator of time of and used robust standard errors to adjust for
were 1532 members of married couples in postbaseline survival (i.e., time to death). within-couple shared variance by using Statas
which the husband was aged at least 65 years, Stressful events. Recent life stress was mea- cluster option. To better estimate the unique
from the Detroit Standard Metropolitan Statistical sured at baseline by asking respondents if association of helping close others with mortality,
Area.36 Of those selected for participation in they had experienced any of the following analyses controlled for several potential con-
the CLOC study, 65% agreed to participate, events in the past 12 months: serious non--- founds, including demographic and socioeconomic
a response rate consistent with response rates life-threatening illness, burglary, job loss, factors, personality, social interactions, self-rated
in other Detroit, Michigan, area studies. nancial difculties, or death of a family health, health behavior, and mental health.
The CLOC study as a whole was primarily member. We used the number of these events
designed not to examine predictors of mortal- (potentially ranging from 0 to 5) as an index RESULTS
ity, but to assess predictors of a surviving of recent stressful events.
spouses experience of widowhood. However, Helping close others. Helping behavior di- Of the subsample of 846 respondents, 134
more than half of the CLOC sample (n = 846) rected toward close others was measured at died over the course of the study. Most re-
consisted of married couples in which both baseline as self-reported engagement in any of spondents (74%) reported having helped
members were focal respondents and, thus, 4 unpaid helping activities directed toward a close other in some way in the past year,
for whom baseline data could be used to friends, neighbors, or relatives who did not live for an average range of 20 to 39 hours. The
predict mortality of each individual. Therefore, with them. The 4 activities were (1) trans- majority of respondents (70%) had experi-
members of this subsample comprised the portation, errands, or shopping; (2) housework; enced no recent stressful life events, whereas
sample for the present study. Exactly 50.0% (3) child care; and (4) other tasks. In addition, 26% had experienced 1 such event and 4%
of this sample was female, and the ethnic respondents reported the total amount of had experienced 2 or 3 events; the latter 2
composition of the sample was 87.7% White, time in the past 12 months they spent in all groups were grouped for the present analyses.
11.7% African American, and less than 1.0% of these activities combined: (0) no help pro- Individuals who helped close others were,
other ethnicity. The mean age of the sample vided; (1) less than 20 hours, (2) 20 to 39 on average, younger, healthier, more likely to
was 71 years (range = 34---93 years). hours, (3) 40 to 79 hours, (4) 80 to 159 hours, be White, of higher socioeconomic status,
All participants completed a face-to-face or (5) 160 hours or more. and higher in social support and social contact
baseline interview, all of which were conducted Control variables. We assessed several other than those who did not do so. Additional
over an 11-month period in 1987 and 1988. variables that were not directly related to the descriptive statistics and correlations among
Subsequently, we monitored mortality of hypotheses of the present study, but were key study variables can be found in Table 1.

1650 | Research and Practice | Peer Reviewed | Poulin et al. American Journal of Public Health | September 2013, Vol 103, No. 9
 RESEARCH AND PRACTICE

Results of a series of Cox regression models

Depression Well-Being
Subjective

are reported in Table 2. Consistent with pre-

-0.24***
1.00
vious research, stressful events predicted in-
creased mortality whereas helping predicted

0.51***
-0.49***
decreased mortality (Table 2, model 1). In

1.00
addition, however, there was a signicant in-
teraction between helping and stressful events

-0.16***
0.19***
Exercise

that remained with signicant adjustment for all

1.00

-0.06
TABLE 1Descriptive Statistics for and Correlations Among Study Variables (n = 846): Detroit-Area Changing Lives of Older Couples Study, 19871993

assessed confounds (Table 2, models 2---6).

To evaluate the nature of this interaction,
Smoking Drinking (No.

0.11**
Month)
in Past

we examined the main effects of stress in 2

1.00

0.02
0.00
0.04
separate models: one in which nonhelping
was coded as 0, revealing the adjusted main
Education, Non-White Satisfaction Functional (Cigarettes
Per Day)

0.10**
18*** effect of stress among nonhelpers, and one in
1.00

0.04
-0.05

-0.01
which helping was coded as 0, revealing
0.26*** the adjusted main effect of stress among

0.18***
-0.29***

-0.25***
With Health Health

helpers. These models indicated that stress was

1.00

0.02
-0.06

not a predictor of mortality for those who

engaged in helping behavior (HR = 0.96; 95%
0.61***

0.26***

0.25***
-0.38***

-0.35***
CI = 0.79, 1.18), but that each additional
1.00

-0.06

-0.01

stressful event (1 vs 0 or 2 vs 1) predicted

a 30% relative increase in mortality risk among
-0.08*

-0.07*
Race

1.00

0.02

0.02
0.01
0.02
-0.02

-0.06

those who did not engage in helping (HR =

1.30; P < .05; 95% CI = 1.05, 1.62). This
0.11***

0.19***
-0.18*** -0.17***

pattern is illustrated in Figure 1, where odds

0.09**

-0.10**
Years

-0.05
1.00

0.05
0.01

0.04

of survival over time are graphed by helping

or not helping close others and levels of
0.46***

0.18***

0.16***
Income

0.09**

stressful events, with adjustment for mean

-0.09*
1.00

0.05

0.02
-0.03

-0.03

levels of control variables.

A follow-up analysis examined whether
Male Gender

0.16***

0.12***
-0.10**

these stress-buffering results depended on de-

-0.08*
1.00

0.02
0.01
0.00

0.05

0.00
-0.04

-0.06

gree of helping in the form of reported amount

of time spent helping (hours in the past year).
0.32***
0.13*** -0.20***

0.07***
2.611.81 -0.24*** 0.21*** -0.14***

0.13*** -0.16***
-0.13***

0.23*** -0.12***
-0.09**

0.12*** -0.09**
Age in
Years

Specically, we created dummy-coded vari-

1.00

-0.11** 0.02
-0.01

-0.03

ables corresponding to low (less than 20 hours)

medium (20---80 hours), and high (> 80 hours)
-0.13***
Helping

Others
Close

0.08*
-0.09*
1.00

0.03

-0.06

0.01

-0.05

levels of helping. When entered into a reg-

ression model simultaneously, these variables
0.715.63 -0.16***
Stressful

1.164.27 0.10**
Events

provided information about the unique con-

0.08*
-0.07*

1.741.96 -0.08*
-0.01

-0.05

-0.01

3.101.59 -0.06
1.00

0.02
0.03

0.06

0.01

tribution of each level of helping compared

with no helping. We computed interactions
0.599.41
0450
050
3693

110
017
Range

02
01

01

01

between each of these dummy variables and

stress. None of these interactions attained
signicance; however, the stress-buffering pat-
13.36 (34.33)

Variable centered and standardized.

0.34 (0.56)
0.74 (0.44)
70.85 (6.22)
0.50 (0.50)
5.06 (2.28)
11.68 (2.88)
0.12 (0.33)
0.03 (0.99)

0.03 (0.98)
2.46 (6.95)

0.01 (0.97)
0.01 (0.97)
Subjective well-beinga 0.00 (0.97)
0.04 (0.93)
Mean (SD)

*P < .05; **P < .01; ***P < .001.

tern appeared strongest for low (HR = 0.44;

P = .09) and medium (HR = 0.60; P = .12)
levels of helping. Importantly, the interaction
between high helping levels and stress, though
Helping close others

Smoking (cigarettes
Functional healtha
Satisfaction with

not signicant, was in a stress-buffering, not

Drinking (no. in
Stressful events

Non-White race
Variable

past month)
Male gender

Education, y
Age in years

Depressiona

stress-augmenting, direction (HR = 0.79; P = .56).

per day)

Exercisea
healtha

Anxietya
Income

An additional set of follow-up analyses

tested whether the main effect of helping or
a

the stress-buffering effect pertained only to

September 2013, Vol 103, No. 9 | American Journal of Public Health Poulin et al. | Peer Reviewed | Research and Practice | 1651
 RESEARCH AND PRACTICE

TABLE 2Cox Proportional Hazard Models of Mortality as a Function of the Interaction of Stress and Helping Behavior, With Control
Variables (n = 846): Detroit-Area Changing Lives of Older Couples Study, 19871993

Hazard Ratio (95% CI)

Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Stressful events 1.56*** (1.22, 1.99) 2.09*** (1.49, 2.93) 1.88*** (1.33, 2.64) 1.82** (1.29, 2.56) 1.81** (1.29, 2.54) 1.62** (1.14, 2.32)
Helping activities 0.41*** (0.29, 0.57) 0.54** (0.36, 0.81) 0.66 (0.44, 1.00) 0.68 (0.45, 1.04) 0.70 (0.46, 1.06) 0.94 (0.58, 1.51)
Helping stressful events 0.57* (0.35, 0.92) 0.62* (0.39, 0.99) 0.62* (0.38, 0.99) 0.63 (0.39, 1.00) 0.54* (0.33, 0.90)
Age 1.07*** (1.04, 1.10) 1.07*** (1.04, 1.10) 1.07*** (1.04, 1.10) 1.06*** (1.03, 1.10)
Male gender 1.97** (1.33, 2.91) 1.85** (1.20, 2.85) 1.82** (1.15, 2.89) 2.14** (1.36, 3.36)
Income 0.87** (0.79, 0.96) 0.87** (0.79, 0.96) 0.87** (0.79, 0.96) 0.91 (0.82, 1.01)
Education 1.02 (0.96, 1.08) 1.02 (0.96, 1.09) 1.03 (0.96, 1.10) 1.02 (0.95, 1.09)
Non-White race 1.35 (0.90, 2.04) 1.32 (0.84, 2.07) 1.29 (0.82, 2.01) 1.32 (0.81, 2.15)
Neuroticisma 1.18 (0.93, 1.49) 1.18 (0.93, 1.49) 1.19 (0.93, 1.51)
Extraversiona 0.84 (0.70, 1.02) 0.86 (0.70, 1.05) 0.97 (0.78, 1.21)
Openness to experiencea 1.17 (0.97, 1.40) 1.16 (0.97, 1.39) 1.07 (0.89, 1.29)
Agreeablenessa 0.91 (0.76, 1.10) 0.92 (0.76, 1.11) 0.90 (0.74, 1.10)
Conscientiousnessa 1.04 (0.84, 1.27) 1.04 (0.85, 1.28) 1.11 (0.88, 1.39)
Self-esteema 0.80* (0.64, 0.99) 0.81 (0.65, 1.00) 0.91 (0.72, 1.15)
Internal locus of controla 0.99 (0.81, 1.21) 0.99 (0.81, 1.22) 1.02 (0.82, 1.25)
Social contacta 0.95 (0.79, 1.16) 0.92 (0.75, 1.12)
Received instrumental support 0.95 (0.81, 1.12) 0.96 (0.82, 1.14)
Received emotional supporta 0.95 (0.78, 1.16) 1.05 (0.87, 1.27)
Satisfaction with healtha 0.68*** (0.56, 0.84)
Functional health limitations 1.59*** (1.26, 2.01)
Smoking 1.02 (0.99, 1.04)
Drinking 1.00 (0.99, 1.00)
Exercisea 0.96 (0.78, 1.17)
Depressiona 1.09 (0.84, 1.41)
Well-beinga 1.14 (0.93, 1.41)
Anxietya 0.94 (0.76, 1.16)
a
Variable centered and standardized.
*P < .05; **P < .01; ***P < .001.

subgroups of participants as dened by our ndings, obtained in a prospective study of connectedness have often been described
control variables (demographic and socioeco- mortality in a community sample, go beyond these in terms of feelings of support or belong-
nomic factors, health variables, personality, and past analyses to indicate that the health benets of ing.2,3,40,41 However, our data, along with data
social network characteristics). Specically, helping behavior derive specically from stress- from previous studies,15,42 indicate that help
we tested all 2-way interactions between help- buffering processes. This nding provides impor- given to others is a better predictor of health
ing and each control variable and all 3-way tant guidance for understanding why helping and well-being than are indicators of social
interactions among helping, stress, and each behavior specically may promote health, and engagement or received social support. Al-
control variable; none of these interactions potentially for how social processes in general may though the mechanisms for this phenomenon
was signicant. inuence health. remain unknown, individuals contributions
Benecial social connections are a key re- to their social networks provide them with
DISCUSSION source for health and well-being,39 and our several unique psychosocial benets, including
data help to clarify what kinds of social con- a sense of meaning or mattering,13,43 oppor-
Social connectedness predicts health and nections are benecial, and why. Specically, tunities for generativity,44,45 or improved
longevity.1 Previous research has indicated that our nding that helping behavior serves as social well-being.46 Moreover, several stress-
helping behavior predicts favorable health out- a stress buffer suggests that helping behavior buffering features of helping could distinguish
comes as well, even if that help occurs in a stressful provides unique psychosocial benets that it from other kinds of social interactions, in-
context such as long-term caregiving.18,38 Our promote health. Psychosocial benets of social cluding the emotional state of compassion47

1652 | Research and Practice | Peer Reviewed | Poulin et al. American Journal of Public Health | September 2013, Vol 103, No. 9
 RESEARCH AND PRACTICE

FIGURE 1Product-limit estimator survival probability curves for low (mean 1 SD) or high (mean + 1 SD) numbers of stressful events in the past
year for those who (a) did not help close others and (b) did help close others.

and the physiology of the caregiving behavioral health and functioning, health behaviors, psy- participate in the CLOC study are unknown,
system.25,27,29 In short, social connections chological well-being, personality traits, and and the sample consisted notably of only
may be benecial to the extent that they social engagement and received social support. married couples. In addition, only a few types
provide individuals with the opportunity to In addition, we found that these possible of stressors were assessed in the CLOC study,
benet others. confounds specically did not function as so it is not clear how our ndings apply to
Our ndings also suggest conditions under moderators of the stress---mortality association, a broader set of stressful experiences. A related
which stressful events may be more or less as did prosocial behavior. Nonetheless, it is issue is the fact that participants in our study
consequential for health and mortality. Expe- possible that unobserved variables account could only report up to 5 stressful events
riencing stressful events signicantly predicted for the associations we found between key in the past year, but it is possible that some
increased mortality over the study period variables and mortality. individuals experienced many more stressors.
among those who had not tangibly helped Second, although there are several aspects of Our nding that the association between stress
others in the past year, but among those who helping that may lead to stress buffering, in- and health was completely eliminated for
had provided help, there was no association cluding increased positive affect, relief of per- those who engaged in helping may not apply
between stress and mortality. In effect, this sonal distress, or activation of the caregiving to individuals exposed to very high levels of
nding suggests that, among individuals who system, assessing these mechanisms was be- stress. Whether similar associations among
do not help others, exposure to a stressful life yond the scope of the present study. Future helping, stress, and behavior will apply to
event is associated with 30% increased mor- research should test potential mediators to nonmarried individuals, those exposed to
tality risk. Given that previous studies of establish the mechanisms by which helping other kinds or greater levels of stress, or other
stress and health have not differentiated those might buffer the effects of psychosocial stress. populations is a ripe topic for future research.
who engage in helping behavior, it is possible Identifying the mechanisms by which helping
that the magnitude of the link between stress buffers stress would have the added benet Conclusions
and health may have been underestimated of helping researchers identify forms of helping Helping behavior, along with other types of
in previous work. Future research should that are more or less likely to have stress- social interaction, is associated with positive
explore this possibility. buffering effects. For example, whereas the health outcomes, including reduced mortality.
present study looked at the provision of The present research indicates that helping
Limitations and Future Directions tangible aid, or instrumental support,2 it is valued others predicts reduced mortality spe-
We recognize that there are noteworthy possible that expressing warmth and caring, cically because it buffers the association
limitations of the current study. First, because or emotional support, would be even more between stress and mortality. To our knowl-
this was a nonexperimental study, it is not benecial.30---32 edge, this is the rst study to nd evidence
possible to claim a causal role for our key Finally, although these results were found for a stress-buffering mechanism for explain-
predictors: stressful events and prosocial be- in a reasonably diverse sample, they may or ing the benecial association of prosocial
havior. We were able to statistically control for may not generalize to all populations. The behavior with mortality. It will be important
many plausible confounds, including baseline characteristics of those who chose not to for research in public health to follow up on

September 2013, Vol 103, No. 9 | American Journal of Public Health Poulin et al. | Peer Reviewed | Research and Practice | 1653
 RESEARCH AND PRACTICE

this nding to determine the causal relation- Human Participant Protection when one into one equals oneness. J Pers Soc Psychol.
All procedures were approved by the institutional 1997;73(3):481---494.
ships between exposure to stress, prosocial
review board of the University of Michigan, and all 18. Poulin MJ, Brown SL, Ubel PA, Smith DM,
behavior, and development of disease, and participants provided informed consent in writing. Jankovic A, Langa KM. Does a helping hand mean
to test the generalizability of these ndings a heavy heart? Helping behavior and well-being
across forms of helping, types of stress, and References among spouse caregivers. Psychol Aging. 2010;25
populations. (1):108---117.
1. House JS, Landis KR, Umberson D. Social relation-
If research continues to nd stress-buffering ships and health. Science. 1988;241(4865):540---545. 19. Folkman S, Moskowitz JT. Positive affect and the
other side of coping. Am Psychol. 2000;55(6):647---
effects of helping, however, it may be wise for 2. Cohen S, Wills TA. Stress, social support and the
654.
buffering hypothesis. Psychol Bull. 1985;98(2):310---
researchers and policymakers to consider 20. Brown SL, Brown RM. Target article: selective
357.
interventions to promote helping, especially investment theory: recasting the functional signi-
3. Cohen S, Pressman S. The stress-buffering hy-
for older adults. As research reveals what cance of close relationships. Psychol Inq. 2006;17(1):
pothesis. In: Anderson N, ed. Encyclopedia of Health
1---29.
types of helping are most benecial, and and Behavior. Thousand Oaks, CA: Sage Publications;
2004. 21. Graziano WG, Habashi MM. Motivational processes
under what conditions, seniors could be underlying both prejudice and helping. Pers Soc Psychol
routinely advised and even incentivized (e.g., 4. Stroebe W, Stroebe MS. Bereavement and Health:
Rev. 2010;14(3):313---331.
The Psychological and Physical Consequences of Partner
through tax deductions) to engage in informal Loss. Cambridge, England: University of Cambridge 22. Shaver PR, Mikulincer M, Shemesh-Iron M. A
helping in their communities, formalized Press; 1987. behavioral systems perspective on prosocial behavior. In:
Mikulincer M, Shaver PR, eds. Prosocial Motives, Emo-
volunteering with a one-on-one emphasis 5. Nielsen NR, Kristensen TS, Schnohr P, Gronbaek M.
tions, and Behavior: The Better Angels of Our Nature.
such as the Foster Grandparents program or Perceived stress and cause-specic mortality among men
Washington, DC: American Psychological Association;
and women: results from a prospective cohort study. 2010.
Experience Corps, or volunteering in general.
Am J Epidemiol. 2008;168(5):481---491.
At-risk populations are frequently advised 23. Bell DC, Richard AJ. Caregiving: the forgotten
6. Rosengren A, Orth-Gomer K, Wedel H, Wilhelmsen element in attachment. Psychol Inq. 2000;11(2):69---83.
to seek support from their social networks. L. Stressful life events, social support, and mortality
in men born in 1933. BMJ. 1993;307(6912):1102--- 24. Bowlby J. Attachment and Loss, Vol 1: Attachment.
A less common message, but one that perhaps
1105. 2nd ed. New York, NY: Basic Books; 1969.
deserves more prominence, is for them to
7. Smith CE, Femengel K, Holcroft C, Gerald K. Meta- 25. Carter CS. Neuroendocrine perspectives on social
support others as well. j attachment and love. Psychoneuroendocrinology. 1998;23
analysis of the associations between social support
(8):779---818.
and health outcomes. Ann Behav Med. 1994;16(4):
352---362. 26. Carter CS, Pournaja-Nazarloo H, Kramer KM, et al.
Oxytocin: behavioral associations and potential as a sali-
About the Authors 8. House JS. Social isolation kills, but how and why?
vary biomarker. Ann N Y Acad Sci. 2007;1098:312---
Michael J. Poulin is with the Department of Psychology, Psychosom Med. 2001;63(2):273---274.
322.
University at Buffalo, Buffalo, NY. Stephanie L. Brown and 9. Morrow-Howell N, Hinterlong J, Rozario PA, Tang
Dylan M. Smith are with the Department of Preventive 27. Numan M. Hypothalamic neural circuits regulating
F. Effects of volunteering on the well-being of older
Medicine, Stony Brook University, Stony Brook, NY. maternal responsiveness toward infants. Behav Cogn
adults. J Gerontol B Psychol Sci Soc Sci. 2003;58(3):
Amanda J. Dillard is with the Department of Psychology, Neurosci Rev. 2006;5(4):163---190.
S137---S145.
Grand Valley State University, Allendale, MI. 28. Rodrigues SM, Saslow LR, Garcia N, John OP,
10. Musick MA, Herzog AR, House JS. Volunteering and
Correspondence should be sent to Michael J. Poulin, Keltner D. Oxytocin receptor genetic variation relates to
mortality among older adults: ndings from a national
University at Buffalo, Department of Psychology, Park empathy and stress reactivity in humans. Proc Natl Acad
sample. J Gerontol B Psychol Sci Soc Sci. 1999;54(3):
Hall 206, Buffalo, NY 14260 (e-mail: mjpoulin@ Sci USA. 2009;106(50):21437---21441.
S173---S180.
buffalo.edu). Reprints can be ordered at http://www. 29. Taylor SE, Klein LC, Lewis BP, Gruenewald TL,
ajph.org by clicking the Reprints link. 11. Shmotkin D, Blumstein T, Modan B. Beyond
Gurung RA, Updegraff JA. Biobehavioral responses to
This article was accepted April 28, 2012. keeping active: concomitants of being a volunteer in
stress in females: tend-and-befriend, not ght-or-ight.
old-old age. Psychol Aging. 2003;18(3):602---607.
Psychol Rev. 2000;107(3):411---429.
12. Oman D, Thoresen CE, McMahon K. Volunteerism
Contributors 30. Floyd K, Hess JA, Miczo LA, Halone KK,
All authors participated in the conceptualization and and mortality among the community-dwelling elderly.
Mikkelson AC, Tusing KJ. Human affection exchange:
design of the analyses reported herein, as well as the J Health Psychol. 1999;4(3):301---316.
VIII. Further evidence of the benets of expressed
interpretation of data, drafting of the article, and 13. Oman D. Does volunteering foster physical health affection. Commun Q. 2005;53(3):285---303.
critical revision of the article for important intellectual and longevity? In: Post SG, ed. Altruism and Health. New
31. Floyd K. Human affection exchange: XII. Affec-
content. M. J. Poulin is fully responsible for all statis- York, NY: Oxford Press; 2007:15---32.
tionate communication is associated with diurnal varia-
tical analyses, had full access to all of the data in the
14. OReilly D, Connolly S, Rosato M, Patterson C. Is tion in salivary free cortisol. West J Commun. 2006;70
study, and takes responsibility for the integrity of the
caring associated with an increased risk of mortality? A (1):47---63.
data and the accuracy of the data analysis.
longitudinal study. Soc Sci Med. 2008;67(8):1282---
32. Floyd K, Mikkelson AC, Tafoya MA, et al. Human
1290.
affection exchange: XIII. Affectionate communication
Acknowledgments 15. Brown SL, Nesse RM, Vinokur AD, Smith DM. accelerates neuroendocrine stress recovery. Health Com-
Funding was provided by the National Institute on Aging Providing social support may be more benecial than mun. 2007;22(2):123---132.
(grants AG15948-01 [ R. M. Nesse], AG610757-01 [C. B. receiving it: results from a prospective study of mortality. 33. Brown SL, Brown RM, House JS, Smith DM. Coping
Wortman], and AG05561-01 [ J. S. House]). Psychol Sci. 2003;14(4):320---327. with spousal loss: potential buffering effects of self-
The authors wish to gratefully acknowledge James S.
16. Brown SL, Smith DM, Schulz R, et al. Caregiving reported helping behavior. Pers Soc Psychol Bull.
House, PhD, Randolph M. Nesse, MD, and Camille B.
behavior is associated with decreased mortality risk. 2008;34(6):849---861.
Wortman, PhD, for collecting the data on which this
research was based and for making the data accessible. Psychol Sci. 2009;20(4):488---494. 34. Krause N. Church-based social support and mor-
The authors also thank Jaymie Meliker, PhD, for helpful 17. Cialdini RB, Brown SL, Lewis BP, Luce C, Neuberg tality. J Gerontol B Psychol Sci Soc Sci. 2006;61(3):S140---
comments on an earlier version of this article. SL. Reinterpreting the empathy---altruism relationship: S146.

1654 | Research and Practice | Peer Reviewed | Poulin et al. American Journal of Public Health | September 2013, Vol 103, No. 9
 RESEARCH AND PRACTICE

35. Okun MA, August KJ, Rook KS, Newsom JT. Does
volunteering moderate the relation between functional
limitations and mortality? Soc Sci Med. 2010;71(9):
1662---1668.
36. Carr D, House JS, Kessler RC, Nesse RM, Sonnega J,
Wortman C. Marital quality and psychological adjust-
ment to widowhood among older adults: a longitudinal
analysis. J Gerontol B Psychol Sci Soc Sci. 2000;55(4):
S197---S207.
37. Costa PT Jr, McCrae RR. Revised NEO Personality
Inventory (NEO-PI-R) and NEO Five-Factor Inventory
(NEO-FFI) Manual. Odessa, FL: Psychological Assess-
ment Resources; 1992.
38. Fredman L, Cauley JA, Hochberg M, Ensrud KE,
Doros G. Mortality associated with caregiving, general
stress, and caregiving-related stress in elderly women:
results of Caregiver-Study of Osteoporotic Fractures.
J Am Geriatr Soc. 2010;58(5):937---943.
39. Phelan JC, Link BG, Tehranifar P. Social conditions
as fundamental causes of health inequalities: theory,
evidence, and policy implications. J Health Soc Behav.
2010;51(Suppl):S28---S40.
40. Baumeister RF, Leary MR. The need to belong:
desire for interpersonal attachments as a fundamental
human motivation. Psychol Bull. 1995;117(3):497---
529.
41. Deci EL, Ryan RM. The what and why of goal
pursuits: human needs and the self-determination of
behavior. Psychol Inq. 2000;11(4):227---268.
42. Schwartz CE, Sprangers MAG, Carey A, Reed G.
Exploring response shift in longitudinal data. Psychol
Health. 2004;19(1):51---69.
43. Rosenberg M, McCullough BC. Mattering: in-
ferred signicance and mental health among ado-
lescents. Res Community Ment Health. 1981;2:
163---182.
44. Erikson E. Identity and the Life Cycle. New York, NY:
International Universities Press; 1959.
45. McAdams D, de St Aubin E. Generativity and Adult
Development: How and Why We Care for the Next
Generation. Washington, DC: American Psychological
Association; 1998.
46. Keyes CLM. Social well-being. Soc Psychol Q.
1998;61(2):121---140.
47. Goetz JL, Keltner D, Simon-Thomas E. Compassion:
an evolutionary analysis and empirical review. Psychol
Bull. 2010;136(3):351---374.

September 2013, Vol 103, No. 9 | American Journal of Public Health Poulin et al. | Peer Reviewed | Research and Practice | 1655