See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/236916634

Benefits of Having Friends in Older Ages: Differential Effects of Informal

Social Activities on Well-Being in Middle-Aged and Older Adults

Article in The Journals of Gerontology Series B Psychological Sciences and Social Sciences · May 2013
DOI: 10.1093/geronb/gbt029 · Source: PubMed

CITATIONS READS

294 2,815

3 authors:

Oliver Huxhold Martina Gabrian (née Miche)

Deutsches Zentrum für Altersfragen Universität Heidelberg
88 PUBLICATIONS 3,486 CITATIONS 50 PUBLICATIONS 2,324 CITATIONS

SEE PROFILE SEE PROFILE

Benjamin Schüz
Universität Bremen
223 PUBLICATIONS 7,151 CITATIONS

SEE PROFILE

All content following this page was uploaded by Benjamin Schüz on 30 December 2015.

The user has requested enhancement of the downloaded file.

Huxhold, O., Miche, M., & Schüz, B. (2013). Benefits of having friends in older ages: differential effects of informal social activities on well-being in middle-aged and older adults. Journals of
Gerontology, Series B: Psychological Sciences and Social Sciences, 69(3), 366–375, doi:10.1093/geronb/gbt029. Advance Access publication May 16, 2013

Benefits of Having Friends in Older Ages: Differential

Effects of Informal Social Activities on Well-Being
in Middle-Aged and Older Adults
Oliver Huxhold,1 Martina Miche,1,2 and Benjamin Schüz3

1
German Centre of Gerontology, Berlin, Germany.
2
Department of Psychological Aging Research, Institute of Psychology, Heidelberg University, Germany.
3
School of Psychology, University of Tasmania, Australia.

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

Objectives. It has been considered a fact that informal social activities promote well-being in old age, irrespective of
whether they are performed with friends or family members. Fundamental differences in the relationship quality between
family members (obligatory) and friends (voluntary), however, suggest differential effects on well-being. Further, age-
related changes in networks suggest age-differential effects of social activities on well-being, as older adults cease
emotionally detrimental relationships.

Method. Longitudinal representative national survey study with middle-aged (n = 2,830) and older adults (n = 2,032).
Age-differential effects of activities with family members and friends on changes in life satisfaction, positive affect (PA),
and negative affect (NA) were examined in latent change score models.

Results. In the middle-aged group, activities with friends and families increased PA and life satisfaction and were
unrelated to NA. In the older age group, family activities increased both PA and NA and were unrelated to changes in life
satisfaction, but activities with friends increased PA and life satisfaction and decreased NA.

Discussion. Social activities differentially affect different facets of well-being. These associations change with age.
In older adults, the effects of social activities with friends may become more important and may act as a buffer against
negative effects of aging.

Key Words: Adult development—Social integration—Subjective well-being.

S ocial connections are important for individual well-

being (Pavot & Diener, 1993). People with closer
social relations tend to be more optimistic about their
themselves as agentic during their performance (Herzog,
Franks, Markus, & Holmberg, 1998).
However, activity engagement requires self-regulatory and
future, and they are healthier (Antonucci, 2001; Cohen & emotional effort (Hess, Emery, & Neupert, 2012; Hess &
Janicki-Deverts, 2009). Social relations also affect subjec- Ennis, 2012; Sonnentag & Jelden, 2009). If this effort
tive well-being (SWB) by providing the individual with the exceeds individual resources, social activities could be
opportunity for participation (Berkman, Glass, Brissette, & detrimental for well-being and health (Segerstrom, 2008;
Seeman, 2000) and activities in a social context. Windsor, Anstey, & Rodgers, 2008). Furthermore, a key
Participating in social activities has been noted as partic- aspect of social activities relates to the context of the par-
ularly important for older people and has been coined a key ticular social relation they are performed within. It has been
factor in successful aging (Rowe & Kahn, 1997). A number demonstrated that social relation may not only have positive,
of studies support the idea that the level of activity engage- but also have negative aspects (Newsom, Rook, Nishishiba,
ment in general improves SWB and health in older adults Sorkin, & Mahan, 2005; Rook, 1984). Therefore, this study
(Adams, Leibbrandt, & Moon, 2011; Chen & Fu, 2008; will focus on the effects of the type of social relations (i.e.,
Fernández-Ballesteros, Zamarrón, & Ruíz, 2001; Menec, family members vs. friends) with which social activities are
2003). Informal social activities have been identified as performed with indicators of individual well-being.
being the domain of activities that has the strongest asso-
ciation with SWB (Adams et al., 2011; Litwin & Shiovitz- Differential Effects of Social Interactions With Family and
Ezra, 2006). Recently, Huxhold, Fiori, and Windsor (2013) Friends
have demonstrated that informal social activities in older Social relationships within the family are the main source
adults (65 and older) are associated with a number of key of support and social integration in later life and therefore
aspects of successful aging over and above network struc- central to older adults’ SWB (Antonucci, Jackson, & Biggs,
ture and social support potential. Engaging in social activi- 2007; Katz, 2009; Koropeckyj-Cox, 2002; Lowenstein,
ties is particularly beneficial for older adults if they perceive 2007; Silverstein & Bengtson, 1994). In a related vein,

366 © The Author 2013. Published by Oxford University Press on behalf of The Gerontological Society of America.
All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
Received August 14, 2012; Accepted March 26, 2013
Decision Editor: Bob Knight, PhD
 Differential Effects of Social Activities on Well-Being in Old Age 367

friendship relations are also a significant source of enjoy- relations are a better source of companionship than family
ment and socializing (Pinquart & Sörensen, 2000). Thus, relations, which implies that they may also be more impor-
good relationships with family members and with friends tant for informal social activities.
are important for older adults and have a positive influence
on their SWB. SWB is a crucial indicator of successful
Do the Effects of Informal Social Activities Change
aging (Freund & Baltes, 1998), and numerous studies docu-
With Age?
ment beneficial effects of SWB on health, the occurrence of
On average, the frequency of informal social activities
critical life events, and mortality (Diener, 2012).
decreases with advancing age, which is probably a conse-
However, it has also been proposed that the influences
quence of age-related declines of functional status (Huxhold
of social interactions on SWB differ if they are performed
et al., 2013). This decline might apply even more informal
with family members or with friends (Dupertuis, Aldwin, &
social activities with friends because friendship relations
Bosse, 2001; Rook & Ituarte, 1999). For example, although
require more investment of resources (Roberts & Dunbar,
both family and friend relations provide emotional support,
2011).
they differ with regard to instrumental support, which is
Socioemotional selectivity theory predicts that the lim-

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

more prominent in family relations, whereas companion-
ited future time perspective of older adults is associated
ship, social integration, and reaffirmation of self-worth
with a focus on emotionally meaningful social interactions
are more characteristic of friendship relations (Messeri,
(Carstensen, Isaacowitz & Charles, 1999), which further
Silverstein, & Litwak, 1993). A recent study showed that
underlines such effects. In a related vein, the strength and
the effects of support on well-being depended on the source
vulnerability model assumes that older adults avoid nega-
of support and the quality of the relationship between pro-
tive social interactions in particular (Charles, 2010). Thus,
vider and recipient (Merz & Huxhold, 2010): Instrumental
both perspectives would predict that informal social activi-
support provided by family members increased NA and
ties that have low emotional meaning or could trigger nega-
impaired PA if the relationship quality was low. In con-
tive emotions are likely to be reduced. As a consequence
trast, instrumental support from friends was associated
of this floor effect, the effects of informal social activi-
with higher PA and lower NA regardless of the relationship
ties may become more positive in older ages. In contrast
quality.
to this assumption, empirical studies on this moderating
Social interactions with kin or nonkin can also have
effect of age on the relationship between activities and
additive effects. Dupertuis and colleagues (2001) found
well-being found ambiguous results (Adams et al., 2011).
that older men with high perceived support from family
One reason for these findings might be that family relations
sources alone experienced more depressive symptoms than
are more difficult to cease than voluntary friendship rela-
their peers who received support both from their families
tions (Birditt et al., 2009; Krause & Rook, 2003, Litwak,
and friends. In a related vein, Fiori, Antonucci, and Cortina
1981). Therefore, it is plausible to predict that informal
(2006) have demonstrated that family-focused personal
social activities with friends, but not with family, will be
networks are associated with lower mental health than
more positively associated with SWB in older adults than
networks that include family and friend relationships (i.e.,
in younger adults. Previous studies that have shown that
diverse networks). However, they also found that family-
the association of social support provided by friends with
focused and diverse networks did not differ in terms of their
SWB increases with increasing age suggest this assumed
affects on physical health.
age effect (Li, Fok, & Fung, 2011; Stafford, McMunn,
Social interactions with family members or friends differ
Zaninotto, & Nazroo, 2011), but differential effects of age
with regard to relationship structure. Although family rela-
and relationship structure have not yet been examined.
tions are more or less predetermined, friendship relations
are formed on a voluntary basis and with peers who share
characteristics, cohort experiences, and lifestyles (Chen Aims and Hypotheses
& Fu, 2008). Probably because of this voluntary nature, Although there is a good rationale to assume that the effects
relationships with friends require more maintenance effort of informal social relations on SWB may not be uniformly
than with family relations (Roberts & Dunbar, 2011). They positive (see argumentation earlier), the literature on
are also more based on reciprocity than family relations activities in older adults reports mainly positive associations
(Dupertuis et al., 2001). This structural difference might with SWB. A potential reason for these findings may be that
also explain the finding that the perceived quality of family the relationship between activities and well-being might be
relations tends to be lower than friendship relations, as fam- bidirectional, as high well-being has been shown to increase
ily obligations increase the likelihood of perpetual negative activity engagement (Fredrickson & Branigan, 2005). This
interaction, and unsatisfying family relations are more dif- study therefore uses a methodology that can account for this
ficult to dissolve than unsatisfying relations with friends bidirectional relationship, latent change modeling (McArdle,
(Birditt, Jackey, & Antonucci, 2009; Krause & Rook, 2003). 2009). We examine age-differential effects of family and
Moreover, according to Rook and Ituarte (1999) friendship friend activities on three different facets of SWB, namely
368 Huxhold et al.

PA, NA, and life satisfaction, which goes beyond previous or visiting friends, (b) visiting cultural events, (c) visiting
research. This study tests three hypotheses in particular: sport events, (d) courses or lectures, (e) going for walks, (f)
H1: Older adults perform less informal social activities than sports, (g) artistic work, (h) board games, and (i) regular
middle-aged adults, and this age effect is particularly pro- meetings with a certain group of people. Except for item
nounced in activities with friends. one, participants were asked whether they had performed
H2: Informal activities are on average positively related to this activity together with their partner or other relatives
SWB, but informal activities with friends have a stronger
effect on SWB than activities with family members.
(family activities) or with friends, other people, or in a club
H3: There are age differences in that the association or association (friend activities). The frequency of every
between informal social activities with friends and SWB is activity was measured ranging from one (< once per month)
stronger in older adults than in middle-aged adults. No age to five (daily). Time invested in family or friend activities
differences are expected for informal social activities with was derived by summing these answers over all family and
family members.
friend activities.

Methods Subjective well-being.—Life satisfaction was assessed

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

with the Satisfaction With Life Scale (Pavot & Diener,
Sample 1993) on five items using a 5-point scale from 1 (strongly
Data for this study were drawn from the German Aging disagree) to 5 (strongly agree; Cronbach’s α T1/
Survey (DEAS), an ongoing population-based, representa- T2 = 0.84/0.85). Emotional well-being was assessed with
tive survey of community-dwelling individuals in Germany the Positive and Negative Affect Schedule (Watson, Clark,
between 40 and 85 years with in-home interviews and & Tellegen, 1988). Participants indicated on a 5-point scale
questionnaires. DEAS has a cohort-sequential design and from 1 (never) to 5 (very often), how often they experienced
national probability sampling with three survey waves in an 10 positive and 10 negative affective states in the last month
interval of 6 years (1996, 2002, and 2008; Engstler & Motel- (Cronbach’s α T1/T2 = 0.87/0.99).
Klingebiel, 2010). The base samples of 1996 (N = 4,838)
and 2002 (N = 3,084) were reassessed in subsequent waves. Control variables.—All analyses were controlled for
In this study, we use longitudinal data of individuals from age, sex, partner status at T1 (1 = having a partner, 0 = no
2002 (T1) and 2008 (T2) because these measurement points partner), health status, and socioeconomic status (SES), as
assessed all meaningful control variables. The final sample these factors relate to social interactions and network char-
consisted of N = 2,032 older adults (65 and older, mean acteristics (van Groenou & van Tilburg, 2003). SES was
age 73.73 years, 47.9% female participants, 64% with part- indicated by education assessed by the international stand-
ner) and 2,830 middle-aged adults (40 and older, mean age ard classification of education (ISCED; UNESCO, 1997).
53.30, 49.2% female participants, 86.4% with partner). Of Health was indicated by self-rated health (single item;
these, 665 older adults and 1,327 middle-aged adults took “How would you rate your present state of health?” with
part in T1 and T2. Drop out analyses suggest that follow- answers from 1 (very bad) to 5 (very good) and functional
up participants were on average younger, healthier, better health assessed with the physical functioning subscale of
educated, and had higher incomes than the baseline sample. the SF-36 (Ware & Sherbourne, 1992; Cronbach’s α T1/
Effect sizes of the differences between participants who T2 = 0.94/0.93).
dropped out of the study and those who remained in the
study did not exceed d = 0.5 in both age groups, indicating a
small-to-medium selectivity effect (Lindenberger, Gilberg, Statistical Analyses
Pötter, Little, & Baltes, 1996). We accounted for sample We used latent change score (LCS) models (McArdle,
attrition by using the full information maximum likelihood 2009) to predict changes in SWB between T1 and T2 from
(FIML) procedure in MPlus. social activities at T1. The three facets of SWB (life satis-
faction, PA, and NA) were indicated with two parcels of
items each (means from the overall items). Social activi-
Measures ties (time spent on social activities with family members or
To examine longitudinal associations across domains, friends) were entered into the model as manifest variables.
measures at both time points were T-standardized (M = 50, To evaluate measurement invariance within and across age
SD = 10) using the mean and standard deviation at T1 groups, we performed comparative fit index (CFI) difference
(2002) of the respective age group, all other analyses were tests, which are more apt to detect invariance than the
performed using the raw scores. conventionally used chi-square difference tests especially
with large sample sizes (Cheung & Rensvold, 2002).
Social activities.—Participants were asked to report LCS decomposes the score of a T2 variable into the
their engagement in various social activities in the year T1 level and between T1 and T2. This LCS carries all
preceding the interview on a list of nine items: (a) inviting advantages of latent variables, such as being free from
 Differential Effects of Social Activities on Well-Being in Old Age 369

measurement error, and it can be related to other vari- Results

ables in the model. To examine our main hypotheses,
we regressed change in the different facets of well- Age Group Differences in Informal Social Activities
being on the T1 levels of social activities (β parameters Overall, middle-aged adults performed more social
in the model) and additionally controlled both baseline activities than older adults (Δχ2 = 366.97, df = 4, p < .05).
and change variables for age, sex, partner status, educa- Across both measurement occasions, there were larger dif-
tion, and subjective and functional health (controlling ferences between younger and older adults in activities with
for functional and subjective health at T2 in the model friends than in activities with family members (Δχ2 = 10.45,
changed the path coefficients of interest only marginally df = 2, p < .05). We did not find an age group by time point
but worsened the model fit tremendously. Thus, they were interaction with respect to informal activities with friends
left out of the analysis). Differential effects between time (Δχ2 = 0.16, df = 1, p > .05). In contrast, we found a signifi-
spent on family versus friend social activities were exam- cant interaction of age group by time point with respect to
ined using nested model comparisons. Figure 1 shows a activities with family (Δχ2 = 7.60, df = 1, p < .05). This indi-
generic LCS model (not specific to well-being aspect). We cates that older adults reduced their informal social activi-

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

regressed changes in well-being on changes in social lei- ties between the two time points with family members more
sure activities on but only in order to increase the statisti- than their activities with friends (Figure 2).
cal power to detect directional associations. Associations
between changes are merely cross-sectional. Therefore,
we will refrain from interpreting the associations between Differential Effects of Informal Social Activities on
changes in the discussion. Instead, we focus on the asso- Changes in SWB
ciations of levels of social leisure activities at T1 with Contrasts in CFI statistics suggested measurement invari-
changes in well-being. ance between the measurement points (equal factor loadings

Figure 1. Illustration of a bivariate dual−change score model with two waves of measurement controlled for age, sex, education, self-rated health, and functional
health. A and B = factors A or B; ΔA and ΔB = change in factor A or B; α = correlation at T1; β1A → ΔA = autoregression of factor A (regression of change in
A on interindividual differences in A at T1); β2B → ΔB = autoregression of factor B (regression of change in B on interindividual differences in B at T1); β3A →
ΔB = directional effect (regression of change in B on interindividual differences in A at T1); β4B → ΔA = directional effect (regression of change in A on interindi-
vidual differences in B at T1); β5ΔA → ΔB = regression of change in ΔB on change in ΔA.
370 Huxhold et al.

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

Figure 2. Mean differences in the frequency to engage in informal social activities by age group (i.e., middle−aged adults vs. older adults), by source (i.e., family
members vs. friends), and time point (i.e., T1 vs. T2).

resulted in CFI losses of 0.01 and less; Cheung & Rensvold, and life satisfaction (β = 0.08, Δχ2 = 9.95, df = 1, p < .05)
2002). However, we were not able to set the variances of across 6 years. Levels of activity were unrelated to NA
the changes to be equal across groups. Thus, we could not (β = 0.01, Δχ2 = 0.19, df = 1, p >.05). Contrasts indicated
contrast the regression weights of predictions of changes no significant differences in effects if informal social activi-
between groups directly. ties were performed with friends or with family members
We examined if levels of informal social activities pre- (p > .05 in all contrasts).
dicted changes in the three facets of well-being separately in
both age groups. We contrasted a model with the respective Older age group (65 years and older).—Social activities
regression paths set to zero to a model in which activities positively predicted changes in PA (β = 0.09, Δχ2 = 5.96,
with friends and activities with family members were set to df = 1, p < .05). The regression weights did not differ
predict changes in one facet at equal strength. In the sec- between activities with friends and activities with family
ond step, we contrasted the regression weights of informal (Δχ2= 1.58, df = 1, p > .05). Predicting changes in life
social activities with friends with informal social activities satisfaction from social activities with friends and with
with family members by releasing the equality constraint. family members fixed to equal regression weights improved
These contrasts are possible because interindividual differ- the model fit significantly if contrasted to a model, in which
ences in levels of activities were standardized to T-scores the association between levels of social activities and
within each age group. changes in life satisfaction were set to zero (Δχ2 = 4.67,
The overall model fit of the final LCS model was good df = 1, p < .05). Releasing the equality constraint improved
(χ2 = 809.59, df = 223, root mean square of approxima- the model fit further (βfriends = 0.14, βfamily = 0.02, Δχ2 = 3.86,
tion = 0.032, CFI = 0.98). All latent changes had significant df = 1, p < .05), which indicates that only social activities
variances. In the following, we will report only the effects with friends predicted changes in life satisfaction. In
of levels of informal social activities at T1 on changes in contrast, if social activities from both sources were fixed
SWB (directional effects). All other directional regression to predict changes in NA with equal strength, the increase
weights are depicted in Table 1. Path coefficients reported in model fit was not significant (Δχ2 = 0.01, df = 1, and
in Table 1 and in the following sections represent standard- p >.05). Estimating the regression weights freely resulted
ized beta coefficients from the respective models. in a further significant increase in model fit (Δχ2 = 5.48,
df = 1, p < .05). Examining the regression weights, we found
Middle-aged group (40–64 years).—High levels of infor- that the associations of social activities with friends and of
mal social activities at T1 were associated with increases or social activities with family members nullified each other
maintenance in PA (β = 0.08, Δχ2 = 10.01, df = 1, p < .05) (βfriends = −0.08, βfamily = 0.08, Δχ2 = 5.48, df = 1, p < .05). In
 Table 1. Predictors of 6-Year Changes in Subjective Well-Being (SWB) in Middle-Aged and Older Adults
6-year changes
Middle-aged adults Older adults
Life satisfaction Positive affect Negative affect Family activities Friend activities Life satisfaction Positive affect Negative affect Family activities Friend activities
Intercept 6.33 13.27* 37.39* 21.17* 29.02* 17.03* 18.36* 51.06* 26.50* 23.47*
Residual variance 39.29* 40.49* 52.84* 80.23* 79.05* 43.06* 40.83* 53.83* 66.78* 74.71*
Predictor level at T1
Family activities 0.08* 0.08* 0.01 −0.64* n/a 0.02 0.09* −0.08* −0.61* n/a
Friend activities 0.08* 0.08* 0.01 n/a −0.57* 14* 0.09* 0.08* n/a −0.59*
Life satisfaction −0.57* 0.04 0.00 0.13* −0.04 −0.61* 0.21* −0.11 0.08 −0.07
Positive affect 0.07 −0.57* −0.04 0.00 0.08* −0.17* −0.72* −0.08 0.00 0.04
Negative affect 0.06 0.03 −0.63* 0.06 −0.03 −0.05 −0.03 −0.66* −0.04 −0.01
Age 0.10* 0.03 −0.08* 0.03 −0.04 0.08 −0.12* −0.09 −0.10 −0.08*
Sex 0.01 0.04 0.07* −0.01 0.14* 0.00 −0.05 0.13* −0.04 0.15*
Education 0.04 0.06 −0.02 0.03 0.04 −0.01 0.00 0.10* 0.08* 0.03
Partner status 0.02 −0.07* 0.02 0.15* 0.01 0.09 −0.12* 0.04 0.13* 0.03
Subjective health 0.08 0.05 −0.10* 0.01 0.01 0.07 0.02 −0.14* −0.04 0.06
Functional health 0.05 0.00 0.03 0.02 −0.02 0.11 −0.02 0.05 0.02 0.09
Predictors changes
Family activities 0.10* 0.10* −0.03 n/a −.01(r) 0.13* 0.15* 0.01 n/a .03(r)
Friend activities 0.12* 0.09* 0.00 −.01(r) n/a 0.15* 0.10* −0.02 .03(r) n/a
Life satisfaction n/a .45*(r) −.46*(r) n/a n/a n/a .49*(r) −.42*(r) n/a n/a
Positive affect .45*(r) n/a −.22*(r) n/a n/a .49*(r) n/a −.12*(r) n/a n/a
Negative affect −.46*(r) −.22*(r) n/a n/a n/a −.42*(r) −.12* (r) n/a n/a n/a

Notes: r = correlation coefficient; n/a = not applicable. Intercepts cannot be interpreted directly. Path coefficients are reported as standardized β coefficients.
*p < .05.
Differential Effects of Social Activities on Well-Being in Old Age
371

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

372 Huxhold et al.

this age group, activities with family at T1 were associated with friends had taken place before the age of 65. Friends
with increasing NA across time, whereas high levels of who remain in the network in older ages may be selected in
activity with friends at T1 were predictive of decreases in NA. accordance to the emotional needs of the aging individual
(Carstensen, Isaacowitz & Charles, 1999; Charles, 2010)
and thus remain in the network for a longer time. The nature
Discussion of activities with family members could also change with
All three hypotheses were only partly confirmed. As increasing older ages, as the family mostly is the primary
expected, older adults engaged in less informal social source of support for older adults (Antonucci et al., 2007).
activities than middle-aged adults. On average, they also Advancing into old age, social support from the family
engage in more informal activities with their families than may become more important for older adults than informal
with their friends. Surprisingly, with increasing age, older social activities with their family members.
adults seem to reduce their level of informal social activ-
ity with their family members more than the activities with H2: Informal activities are on average positively related to
their friends. SWB, but informal activities with friends have a stronger
effect on SWB than activities with family members.
Our results imply that in middle age, engaging in infor-

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

mal social activities is generally beneficial for SWB. High Our results suggest that in middle age, it is beneficial
levels of activity predicted increasing levels of PA and life to perform leisure activities together with social partners.
satisfaction but were unrelated with NA. In this age group, It can help to maintain or even increase feelings of
it seems not to matter with whom informal social activi- fun, interest, and general satisfaction with one’s life
ties are performed with. In the older age group, however, circumstances. Accordingly, in middle-aged adults, it did
the specific company seems to play a role (albeit not with not matter whether social activities were performed with
respect to changes in PA): Only informal social activities family members or friends. Therefore, we could not find
with friends were associated with increases or maintenance evidence that supports the second part of our hypotheses
of life satisfaction. In addition, informal social activities that activities with friends are more beneficial for SWB than
with friends were associated with reductions in NA, but activities with family members. In older adults, however,
social activities with family members were even associated the picture seems to be different.
with increases in NA in this age group.
H3: There are age differences in that the association
H1: Older adults perform less informal social activities than between informal social activities with friends and SWB is
middle-aged adults, and this age effect is particularly pro- stronger in older adults than in middle-aged adults. No age
nounced in activities with friends. differences are expected for informal social activities with
family members.
Our analyses contrasted the frequency of informal social
activities with friends and with family members between Our analyses revealed that age differences in the asso-
the middle-aged and the older age groups. Older adults ciations between informal social activities and SWB dif-
performed less social activities than their younger counter- fered with regard to the particular facet of SWB under
parts even after controlling for interindividual differences observation.
in health, education, and partner status. In comparison with High levels of social activities were associated with
the middle-aged adults, the older group showed dispropor- increases or maintenance in PA in older adults with similar
tionately less activities with their friends. These findings effects as in the middle-aged group, and it seems to make
are in line with the argumentation that engaging in informal no difference whether the activities were performed with
social activities is resource demanding. Older adults are, friends or family members.
on average, not only less healthy than middle-aged adults Social activities were unrelated to NA in middle-aged
but also have less cognitive and motivational resources that adults. In the older age group, however, informal social
may enable them to get involved in effortful activities (Hess activities with family members were associated with
et al., 2012; Hess & Ennis, 2012). Friendship relations increases or maintenance of NA, but high levels of social
require more investing effort than family relations (Roberts activities with friends were associated with decreases in NA.
& Dunbar, 2011). This may explain why older adults dif- Engaging in informal social activities may require effort
fer from middle-aged adults in particular in informal social and may sometimes be frustrating if conflicts arise with the
activities with friends. However, older adults reduced their social interaction (Hess et al., 2012; Rook, 1984). In older
social activities with friends less than their activities with age, motivational and cognitive resources that may play a
their family members across the 6-year interval. This find- role in shared leisure activities decline (Hess et al., 2012).
ing can neither be explained by a resource-driven perspec- Therefore, engaging in social activities with friends may
tive nor is it likely to be a consequence of participants losing become more demanding. Moreover, it has been shown that
their romantic partner, as the number of persons bereaved conflicts within the family are relatively stable across the
between 2002 and 2008 was very small. One potential life span (Krause & Rook, 2003). According to the strength
explanation could be that the major reductions in activities and vulnerability model (Charles, 2010), older adults are
 Differential Effects of Social Activities on Well-Being in Old Age 373

better than younger adults in navigating their social envi- be that the role of the family changes in older ages, as social
ronment. At the same time, older adults suffer more than support becomes increasingly important for older adults
younger adults if negative social interactions cannot be as a buffer against various negative effects of age-related
avoided. Older adults may have generally less opportunities decline (Bierman & Statland, 2010). The primary provider
to use their social skills in their family relations because of social support in older ages is the family (Antonucci,
these relationships have a strong normative component 2001). This may have consequences on how older adults
(Litwak, 1981). Combining these two arguments, it seems evaluate the relationships with friends and family members:
plausible to assume that familial conflicts that occur within Friends could remain the main source of companionship,
shared leisure activities may particularly affect the SWB of for which it is important to perform activities together,
older adults because they are more vulnerable to negative whereas family members might increasingly be perceived
social interactions. as crucial providers of social support rather than interaction
Our finding is in contrast to previous research argu- partners for informal social activities. Consequently, social
ing that informal social activities are generally beneficial activities with family members may be less important for
for SWB (Adams et al., 2011). In our view, only specific life satisfaction of older adults than of younger adults. The

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

analyses, such as in this article, allow to detect the associa- changing role of the family may also be the reason why
tion between high levels of social activity and increases in older adults in our sample decreased their activities with
NA because the LCS approach takes into account recipro- their family members more than their activities with their
cal influences. Adams, Leibbrandt, and Moon (2011), for friends during the 6-year period.
example, argue that high levels of well-being may some-
times be more important for performing activities than the
influence of activities on well-being. We controlled for the Limitations and Strengths
effects of T1 SWB and for the concurrent reciprocal influ- All measures were self-reported and related to poten-
ences between activities and SWB across the 6-year interval tially long recall periods. This might imply that the recall
by regressing changes in SWB on changes in social activi- of activities could be subject to under- or overestimation
ties. Detrimental effects of activities can only be examined based on current affective status. However, our results con-
by controlling for these reciprocal influences. In contrast trolled for baseline differences in SWB. This suggests that
to the negative influence of social activities with the fam- effects on long-term SWB changes are independent of base-
ily, informal social activities with friends were associated line differences. The limitation to two measurement points
with decreases in NA. Silverstein and Parker (2002) found with a time difference of 6 years might mask more com-
that leisure activities buffer against the negative effects plex interacting dynamics between higher SWB and social
of functional impairment, widowhood, and lack of social interactions, in particular with regard to PA and NA, which
support but did not differentiate with whom the activities have been shown to fluctuate over shorter time spans and
were performed with. The socioemotional selectivity theory as a result of various life events (Diener, 2012). We inter-
(Carstensen, Isaacowitz & Charles, 1999) and the strength pret our results as indicating changes in affect dispositions
and vulnerability model (Charles, 2010) predict that older rather than short-term fluctuations, but the possibility that
adults will let go of social relations that are emotionally affect measures change as a result of short-term influences
meaningless or even increase the likelihood of negative needs to be born in mind. Future studies might target these
interactions. Due to the voluntary nature of friendship rela- dynamic interactions using more measurement points in
tions, it might be easier to be selective with friends than with shorter time intervals. Moreover, with the data set at hand,
family members (Krause & Rook, 2003). This expectation we were unable to distinguish between different relation-
of differential selectivity is in line with our finding that age ships within the family. It has been shown, however, that
group differences were on average greater for social activi- relationships with a partner or with children are evaluated
ties with friends than with activities with family members. quite differently and show different developmental trajecto-
In light of this reasoning, it is plausible to assume that activ- ries (Birditt et al., 2009).
ities performed with friends that remain in the social net- Moreover, our measures of informal social activities broad
work of older adults after an emotionally driven selectivity measure of social integration of older adults. Our analyses
process are particularly helpful. Consequently, our findings did not consider the quality of the relationships with friends
add to the results from the study by Silverstein and Parker or with family members. Merz and Huxhold (2010), for
(2002) that only social activities with friends may act as a example, have shown that the negative impact of instrumen-
buffer against the negative consequences of aging and are tal support provided by family members is moderated by the
therefore associated with decreases in NA. relationship quality. Thus, future analyses should investigate
The finding that social activities with family members were if the consequences of performing informal social activities
only beneficial for life satisfaction in middle-aged adults with family members with whom the older adult shares a
but not in the older age group might seem puzzling at first high-quality relationship differ from the effects of performing
sight, and we can only speculate about the reasons. It could social activities in the context of low-quality family relations.
374 Huxhold et al.

Implications and Conclusion Chen, S. Y., & Fu, Y. C. (2008). Leisure participation and enjoyment among
Our study shows that it makes a difference in terms of the elderly: Individual characteristics and sociability. Educational
Gerontology, 34, 871–889. doi:10.1080/03601270802115382
NA and life satisfaction with whom one engages in social Cohen, S., & Janicki-Deverts, D. (2009). Can we improve our physical
activities for older adults but not for middle-aged adults. health by altering our social networks? Perspectives on Psychological
These differential effects might be due to the structural Science, 4, 375–378. doi:10.1111/j.1745-6924.2009.01141.x
and functional differences between family-based and Cohen, G. D., Perlstein, S., Chapline, J., Kelly, J., Firth, K. M., & Simmens,
friend-based social networks that influence the emotion- S. (2006). The impact of professionally conducted cultural programs
on the physical health, mental health, and social functioning of older
ally driven selectivity process in older ages. Informal adults. Gerontologist, 46(6), 726–734. doi:10.1093/geront/46.6.726
social activities with friends are beneficial for all three Diener, E. (2012). New Findings and Future Directions for Subjective
aspects of SWB in older adults. Future research might Well-Being Research. American Psychologist, 67(8), 590–597.
want to examine mediators of these effects, in particular doi:10.1037/a0029541
the role of individual expectations with regard to activity Dupertuis, L. L., Aldwin, C. M., & Bosse, R. (2001). Does the source
of support matter for different health outcomes? Findings from the
outcomes. Our results have strong implications for mental Normative Aging Study. Journal of Aging and Health, 13, 494–510.
health promotion in older adults, as they suggest support- doi:10.1111/j.1745-6924.2009.01141.x

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

ing older adults in building and maintaining friend-based Engstler, H., & Motel-Klingebiel, A. (2010). Datengrundlagen und Methoden
social networks, for example, by encouraging volunteer- des Deutschen Alterssurvey (DEAS). In A. Motel-Klingebiel, S. Wurm,
ing in old age in elder-helping-elder programs (Butler & & C. Tesch-Römer (Eds.), Altern im Wandel: Befunde des Deutschen
Alterssurveys (DEAS) (pp. 34–60). Stuttgart: Kohlhammer.
Eckart, 2007) or programs aiming at increasing informal Fernández-Ballesteros, R., Molina, M. Á., Schettini, R., & Del Rey, Á. L.
social interactions, such as cultural programs (Cohen (2012). Promoting active aging through university programs for older
et al., 2006), university programs (Fernández-Ballesteros, adults. GeroPsych: The Journal of Gerontopsychology and Geriatric
Molina, Schettini, & Del Rey, 2012), or occupational ther- Psychiatry, 25, 145–154. doi:10.1024/1662–9647/a000064
apy programs (Hay et al., 2002). Fernández-Ballesteros, R., Zamarrón, M. D., & Ruíz, M. A. (2001).
The contribution of socio-demographic and psychosocial factors
to life-satisfaction. Ageing and Society, 21, 25–43. doi:10.1017/
Funding
S0144686X01008078
This work was supported by the German Federal Ministry for Family Fiori, K. L., Antonucci, T. C., & Cortina, K. S. (2006). Social network
Affairs, Senior Citizens, Women and Youth (BMFSFJ). typologies and mental health among older adults. The Journals of
Gerontology: Series B: Psychological Sciences and Social Sciences,
Correspondence 61, P25P32. doi:10.1093/geronb/61.1.P25
Correspondence should be addressed to Oliver Huxhold, PhD, German Fredrickson, B. L., & Branigan, C. (2005). Positive emotions broaden
Centre of Gerontology, Manfred-von-Richthofen Strasse, D-12101 Berlin, the scope of attention and thought-action repertoires. Cognition &
Germany. E-mail: oliver.huxhold@dza.de. Emotion, 19, 313–332. doi:10.1080/02699930441000238
Freund, A. M., & Baltes, P. B. (1998). Selection, optimization, and com-
References pensation as strategies of life management: Correlations with sub-
Adams, K. B., Leibbrandt, S., & Moon, H. (2011). A critical review of the jective indicators of successful aging. Psychology and Aging, 13,
literature on social and leisure activity and wellbeing in later life. 531–543. doi:10.1037/0882-7974.13.4.531
Ageing & Society, 31, 683–712. doi:10.1017/s0144686x10001091 Hay, J., LaBree, L., Luo, R., Clark, F., Carlson, M., Mandel, D., . . . Azen, S.
Antonucci, T. C. (2001). Social relations: An examination of social net- P. (2002). Cost-effectiveness of preventive occupational therapy for
works and social support, and sense of control. In J. E. Birren (Ed.), independent-living older adults. Journal of the American Geriatrics
Handbook of the psychology of aging (5th ed., pp. 427–453). San Society, 50, 1381–1388. doi:10.1046/j.1532-5415.2002.50359.x
Diego, CA: Academic Press. Herzog, A. R., Franks, M. M., Markus, H. R., & Holmberg, D. (1998).
Antonucci, T. C., Jackson, J. S., & Biggs, S. (2007). Intergenerational Activities and well-being in older age: Effects of self-concept
relations: Theory, research, and policy. Journal of Social Issues, 63, and educational attainment. Psychology and Aging, 13, 179–185.
679–693. doi:10.1111/j.1540-4560.2007.00530.x doi:10.1037//0882-7974.13.2.179
Berkman, L. F., Glass, T., Brissette, I., & Seeman, T. E. (2000). From Hess, T. M., Emery, L., & Neupert, S. D. (2012). Longitudinal relation-
social integration to health: Durkheim in the new millennium. ships between resources, motivation, and functioning. Journals of
Social Science & Medicine (1982), 51, 843–857. doi:10.1016/ Gerontology Series B-Psychological Sciences and Social Sciences,
S0277-9536(00)00065-4 67, 299–308. doi:10.1093/geronb/gbr100
Birditt, K. S., Jackey, L. M., & Antonucci, T. C. (2009). Longitudinal pat- Hess, T. M., & Ennis, G. E. (2012). Age differences in the effort and
terns of negative relationship quality across adulthood. The Journals costs associated with cognitive activity. Journals of Gerontology
of Gerontology. Series B, Psychological Sciences and Social Series B—Psychological Sciences and Social Sciences, 67, 447–455.
Sciences, 64, 55–64. doi:10.1093/geronb/gbn031 doi:10.1093/geronb/gbr129
Butler, S. S., & Eckart, D. (2007). Civic engagement among older adults Huxhold, O., Fiori, K. L., & Windsor, T. D. (2013). The dynamic interplay
in a rural community: A case study of the senior companion pro- of social network characteristics, subjective well-being and health:
gram. Journal of Community Practice, 15, 77–98. doi:10.1300/ The costs and benefits of socio-emotional selectivity. Psychology
J125v15n03_05 and Aging, 28(1), 3–16. doi:10.1037/a0030170
Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking Katz, R. (2009). Intergenerational family relations and subjective well-
time seriously - A theory of socioemotional selectivity. American being in old age: A cross-national study. European Journal of Aging,
Psychologist, 54(3), 165–181. doi:10.1037//0003-066x.54.3.165 6, 79–90. doi:10.1007/s10433-009-0113-0
Charles, S. T. (2010). Strength and vulnerability integration: A model of Koropeckyj-Cox, T. (2002). Beyond parental status: Psychological well-
emotional well-being across adulthood. Psychological Bulletin, 136, being in middle and old age. Journal of Marriage and the Family, 64,
1068–1091. doi:10.1037/a0021232 957–971. doi:10.1111/j.1741-3737.2002.00957.x
 Differential Effects of Social Activities on Well-Being in Old Age 375

Krause, N., & Rook, K. S. (2003). Negative interaction in late life: Issues Roberts, S. G. B., & Dunbar, R. I. M. (2011). The costs of family and
in the stability and generalizability of conflict across relationships. friends: An 18-month longitudinal study of relationship mainte-
Journal of Gerontology: Psychological Sciences, 58B, 88–99. nance and decay. Evolution and Human Behavior, 32, 186–197.
doi:10.1093/geronb/58.2.P88 doi:10.1016/j.evolhumbehav.2010.08.005
Lindenberger, U., Gilberg, R., Pötter, U., Little, T. D., & Baltes, P. B. (1996). Rook, K. S. (1984). The negative side of social interaction: Impact on psy-
Stichprobenselektivität und Generalisierbarkeit der Ergebnisse in chological well-being. Journal of Personality and Social Psychology,
der Berliner Altersstudie. In K. U. Mayer & P. B. Baltes (Eds.), Die 46, 1097–1108. doi:10.1037/0022-3514.46.5.1097
Berliner Altersstudie (pp. 85–108). Berlin, Germany: Akademi Verlag. Rook, K. S., & Ituarte, P. H. G. (1999). Social control, social support, and
Litwak, E. (1981). The modified extended family, social networks, and companionship in older adults’ family relationships and friendships.
research continuities in aging. New York: Columbia University Personal Relationships, 6, 199–211. doi:10.1111/j.1475–6811.1999.
Center for Social Sciences. tb00187.x
Litwin, H., & Shiovitz-Ezra, S. (2006). The association between activity Rowe, J. W., & Kahn, R. L. (1997). Successful aging. The Gerontologist,
and wellbeing in later life: What really matters? [Article]. Ageing & 37, 433–440. 10.1093/geront/37.4.433
Society, 26, 225–242. doi:10.1017/s0144686x05004538 Segerstrom, S. C. (2008). Social networks and immunosuppression during
Li, T. Y., Fok, H. K., & Fung, H. H. (2011). Is reciprocity always beneficial? stress: Relationship conflict or energy conservation? Brain, Behavior,
Age differences in the association between support balance and life and Immunity, 22, 279–284. doi:10.1016/j.bbi.2007.10.011
satisfaction. Aging & Mental Health, 15(5), 541–547. doi:10.1080/ Silverstein, M., & Bengtson, V. L. (1994). Does intergenerational

Downloaded from http://psychsocgerontology.oxfordjournals.org/ by guest on December 28, 2015

13607863.2010.551340 social support influence the psychological well-being of older
Lowenstein, A. (2007). Solidarity conflict and ambivalence: Testing parents? The contingencies of declining health and widowhood.
two conceptual frameworks and their impact on quality of life Social Science & Medicine (1982), 38, 943–957. doi:10.1016/
for older family members. The Journals of Gerontology Series 0277-9536(94)90427–8
B: Psychological Sciences and Social Sciences, 62, S100–S107. Silverstein, M., & Parker, M. G. (2002). Leisure activities and quality of
doi:10.1093/geronb/62.2.S100 life among the oldest old in Sweden. Research on Aging, 24, 528–
McArdle, J. J. (2009). Latent variable modeling of differences and changes 547. doi:10.1177/0164027502245003
with longitudinal data. Annual Review of Psychology, 60, 577–605. Sonnentag, S., & Jelden, S. (2009). Job stressors and the pursuit of sport
doi:10.1146/annurev.psych.60.110707.163612 activities: A day-level perspective. Journal of Occupational Health
Menec, V. H. (2003). The relation between everyday activities and suc- Psychology, 14, 165–181. doi:10.1037/a0014953
cessful aging: A 6-year longitudinal study. Journals of Gerontology Stafford, M., McMunn, A., Zaninotto, P., & Nazroo, J. (2011). Positive
Series B—Psychological Sciences and Social Sciences, 58, S74–S82. and Negative Exchanges in Social Relationships as Predictors of
doi:10.1093/geronb/58.2.S74 Depression: Evidence From the English Longitudinal Study of
Merz, E. M., & Huxhold, O. (2010). Wellbeing depends on social rela- Aging. Journal of Aging and Health, 23(4), 607–628. doi:10.1177/
tionship characteristics: Comparing different types and providers of 0898264310392992
support to older adults. Ageing & Society, 30, 843–857. doi:10.1017/ UNESCO. (1997). International Standard Classification of Education:
S0144686X10000061 ISCED 1997. Paris, France: Author.
Messeri, P., Silverstein, M., & Litwak, E. (1993). Choosing optimal sup- van Groenou, M. I. B., & van Tilburg, T. (2003). Network size and sup-
port groups: A review and reformulation. Journal of Health and port in old age: Differentials by socio-economic status in childhood
Social Behavior, 34, 122–137. doi:10.2307/2137239 and adulthood. Ageing and Society, 23, 625–645. doi:10.1017/
Newsom, J. T., Rook, K. S., Nishishiba, M., Sorkin, D. H., & Mahan, T. L. S0144686X0300134X
(2005). Understanding the relative importance of positive and nega- Ware, J. E., & Sherbourne, C. D. (1992). The MOS 36-item short-form health
tive social exchanges: Examining specific domains and appraisals. survey (SF-36). I. Conceptual framework and item selection. Medical
The Journals of Gerontology. Series B, Psychological Sciences and Care, 30, 473–483. doi:10.1097/00005650-199206000-00002
Social Sciences, 60, P304–P312. doi:10.1093/geronb/60.6.P304 Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and
Pavot, W., & Diener, E. (1993). Review of the satisfaction with life scale. validation of brief measures of positive and negative affect: The
Psychological Assessment, 5, 164–172. doi:10.10371040-3590.5.2.164 PANAS scales. Journal of Personality and Social Psychology, 54,
Pinquart, M., & Sörensen, S. (2000). Influences of socioeconomic sta- 1063–1070. doi:10.1037//0022-3514.54.6.1063
tus, social network, and competence on subjective well-being in Windsor, T. D., Anstey, K. J., & Rodgers, B. (2008). Volunteering and
later life: A meta-analysis. Psychology and Aging, 15, 187–224. psychological well-being among young-old adults: How much is too
doi:10.1037//0882-7974.15.2.187 much? The Gerontologist, 48, 59–70. doi:10.1093/geront/48.1.59

View publication stats