Journal of Gerontology: SOCIAL SCIENCES Copyright 2007 by The Gerontological Society of America

2007, Vol. 62B, No. 1, S60S68

Continuous Participation in Voluntary Groups as a

Protective Factor for the Psychological Well-Being
of Adults Who Develop Functional Limitations:
Evidence From the National Survey of
Families and Households
Emily A. Greenfield and Nadine F. Marks

Department of Human Development and Family Studies, University of Wisconsin-Madison.

Objectives. Although previous studies have indicated that declining functional health is associated with individuals
poorer psychological well-being, few studies have examined factors that can protect adults from the loss of well-being
following functional decline. Guided by continuity theory, this study investigated the extent to which continuous
participation in voluntary groups (recreational, religious, and civic) buffers individuals against the harmful psychological
effects of developing functional limitations.

Methods. Longitudinal data came from 4,646 respondents aged 35 to 92 in the National Survey of Families and
Households (19871993) who reported having no functional limitations at Time 1.

Results. Multivariate models controlling for sociodemographic factors, as well as psychological well-being at Time 1,
indicated that developing functional limitations over a 5-year period was associated with greater increases in depressive
symptoms and lower levels of personal growth. Increases in depressive symptoms, however, were less severe among men
who were continuously involved in recreational groups than among men who were not continuously involved in
recreational groups. Additionally, the association between developing functional limitations and lower levels of personal
growth did not hold for men or women who continuously participated in religious groups.

Discussion. Findings suggest that continuous participation in certain types of voluntary groups can moderate the
problematic effects of developing functional limitations on psychological well-being.

F UNCTIONAL limitationsor having health conditions

that interfere with ones ability to complete daily tasks of
livinghave profound implications for the quality of individ-
uals lives, as well as for social systems that must provide care
for people with disabilities (Lawton, 1991). Given the chal-
lenges that functional health impairments pose for both indi-
viduals and society, much research has focused on identifying
the causes of functional limitations so as to inform efforts to
lower rates of impairment (e.g., Kono, Kai, Sakato, & Rubenstein,
2004; McCusker, Kakuma, & Abrahamowicz, 2002). Although
promoting functional health at the individual and population
levels is important, the fact remains that an increasing number of
adults in the United States have functional limitations
(American Association of Retired Persons, 1998). Advancing
the longevity of morbidity hypothesis, scholars have suggested
that it is becoming increasingly common for people with
functional limitations to survive more years living with
disability (Zarit & Zarit, 1998). As a rising number of adults
acquire and maintain functional limitations, scholarly efforts to
more fully and systematically understand the consequences of
living with functional limitations become increasingly critical.
The aim of this study was in part to address this need by
investigating variation in individuals psychological well-being
following their developing functional limitations. Guided by
continuity theory and using prospective, longitudinal data from
the National Survey of Families and Households (19871993),
this study examined continuous participation in voluntary
groups (recreational, religious, and civic) as a protective factor
against the deleterious psychological consequences of having
developed functional limitations.
Theoretical Background
Continuity theory, a well-developed social gerontological
theory, provides insight into how ongoing voluntary group
participation might help individuals to avoid the negative
psychological effects of developing functional limitations.
Continuity theory advances the assumption that as people
age, they are likely to experience significant life changes, such
as widowhood and retirement (Atchley, 1989, 1993). The
theory posits that when facing such changes, individuals are
predisposed to think of themselves and their lives as having
coherence and consistency. Continuity theory postulates that
people strive to achieve a sense of stability both in terms of
internal continuity (i.e., maintaining a coherent sense of who
one is) and external continuity (i.e., maintaining a coherent
sense of ones social world). Although continuity theory rec-
ognizes that striving towards continuity might not be adaptive
for well-being under all circumstances, the theory posits that
continuity is usually an effective means by which people can
develop and adapt as a result of change (Atchley, 1989).

S60
 VOLUNTARY GROUP PARTICIPATION
Continuity theory explicates mental and behavioral strategies
that can foster an individuals sense of continuity. For example,
interpreting new information about ones self in a way that
validates ones existing self-concept can help a person maintain
a sense of internal continuity, and soliciting the company of
familiar persons can enhance a persons sense of external con-
tinuity. Continuity theorists have discussed a variety of life
domains in which people might employ continuity-maintaining
strategies, including the domain of activities (Atchley, 1993).
Researchers have identified participation in voluntary groups,
in particular, as a promising source of activity-related
continuity in that activities occurring within overt social
environmentssuch as within the context of a stable group
infrastructure or within well-defined role relationships can more
readily provide individuals with a sense of stability over time
(Atchley, 1989).
In sum, continuity theory suggests that as people experience
significant life changes, individuals engage in mental and be-
havioral strategies that help them to maintain a sense of stability
in terms of their self and social worlds. As such, continuity
theory suggests that as adults make the transition to a state of
impaired functional health (i.e., a significant life change),
continuous participation in voluntary groups (i.e., an activity-
related source of continuity) can help them to maintain higher
levels of well-being (i.e., adapt more beneficially to this change).
Empirical Background
A substantial body of evidence suggests that loss of physical
health is a risk factor for individuals poor psychological well-
being (Okun, 1984). Moreover, several studies drawing on
longitudinal data from large population samples have indicated
that, overall, individuals greater levels of physical disability
are associated with their poorer mental health over time
(Kunzmann, Little, & Smith, 2000; Taylor & Lynch, 2004).
Nevertheless, findings from other studies have indicated that
there is substantial variation in psychological well-being among
people with physical health problems (Brief, Butcher, & George,
1993) and that individuals perceptions of their health-related
quality of life and overall well-being do not entirely overlap
(Spiro & Bosse, 2000). These findings suggest that there are
likely factors that can protect aspects of adults well-being when
they are faced with the challenges of impaired functional health.
Scholars have long discussed that social integration and
productive activities can promote optimal states of well-being,
particularly among individuals facing age-related changes
(Rowe & Kahn, 1997). Building on this idea, researchers
have focused increasingly on the physical and mental health
effects of participation in voluntary groups, or groups that are
(a) outside of ones private network of friends and families, (b)
not mandated by the government or the market, and (c) formed
as result of shared interests among a collective of individuals.
Previous empirical studies have documented linkages between
voluntary group participation and higher levels of well-being.
For example, previous research has indicated that formal
religious participation is associated with better physical and
mental health (see Ellison & Levin, 1998, for a review),
participation in activities within formal social groups is pre-
dictive of adults greater happiness over time (Menec, 2003),
and formal volunteering is linked with fewer depressive
S61
symptoms and better self-reported health and functioning
(Morrow-Howell, Hinterlong, Rozario, & Tang, 2003).
Although this growing body of research demonstrates the
potential benefits of voluntary group participation for individuals
well-being, understanding of the linkages between voluntary
group participation and enhanced well-being is far from complete.
Although scholars have posited that voluntary group participation
is particularly salient for the well-being of more vulnerable adults
(e.g., Silverstein & Parker, 2002), there has been little in-
vestigation of whether voluntary group participation can protect
individuals who are specifically at risk for poor mental health.
Additionally, few studies regarding the associations between
voluntary group participation and psychological well-being have
simultaneously examined participation in multiple types of
voluntary groups, as well as multiple dimensions of psychological
well-being. Given previous findings that different psychosocial
processes might lead to participation in distinct types of voluntary
groups (Janoski & Wilson, 1995), and that the effects of voluntary
group participation might be particularly salient in terms of certain
aspects of psychological well-being (Greenfield & Marks, 2004),
it is important for additional research to simultaneously examine
a range of types of voluntary group participation as well as
multiple mental health outcomes. Also, very few studies on the
linkages between voluntary group participation and psycholog-
ical well-being have considered the dynamic nature of such
participation (i.e., that participants in voluntary groups differ with
respect to how long they have been involved in such groups).
Continuity theory suggests conditions under which more long-
term participation in voluntary groups might be particularly
advantageous for individuals psychological well-being. Finally,
although scholars have discussed the extent to which voluntary
group participation differs for men and women (e.g., Cutler &
Hendricks, 2000), few studies have explicitly examined whether
the potential psychological consequences of participation in
diverse types of voluntary groups vary by gender.
This study aimed to address these gaps in the literature on
voluntary group participation and psychological well-being, as
well as gaps in research on variation in the effects of functional
limitations on psychological well-being, by examining contin-
uous participation in voluntary groups as a protective factor
against the negative psychological effects of having developed
functional limitations for men and women.
Hypotheses
Guided by continuity theory and previous empirical work, we
formulated a conceptual model that suggests the risk-buffering
effect of continuous participation in voluntary groups for adults
psychological well-being in the face of developing functional
limitations (Figure 1). We derived three hypotheses (H) from
this model and formulated two research questions (RQ).
H1: Adults who develop functional limitations over a 5-year
period will report poorer psychological well-being (greater
increases in depressive symptoms and lower levels of
personal growth) in contrast to adults who do not develop
any functional limitations.
H2: Adults who maintain participation in voluntary groups
(recreational, religious, and civic) over a 5-year period will
experience better psychological well-being (greater de-
clines in depressive symptoms and higher levels of personal
S62 GREENFIELD AND MARKS
growth) in contrast to adults who do not maintain
continuous participation in voluntary groups over a 5-
year period.
H3: Adults who develop functional limitations but who also
maintain participation in voluntary groups over a 5-year
period will report better psychological well-being in
contrast to adults who develop functional limitations but
who do not maintain continuous participation in voluntary
groups over a 5-year period.
RQ1: Do associations between continuous participation in
voluntary groups and psychological well-being differ for
men and women?
RQ2: Does the extent to which continuous participation in
voluntary groups moderates the associations between
functional limitations and psychological well-being
differ for men and women?
METHODS
Data
This study used data from the first and second waves of the
National Survey of Families and Households, a national prob-
ability sample with an oversampling of African Americans,
Puerto Ricans, Mexican Americans, single-parent families,
families with stepchildren, cohabiting couples, and recently
married persons (Sweet & Bumpass, 1996; Sweet, Bumpass, &
Call, 1988). At Time 1 (T1; 19871988), the study included
interviews with 13,007 persons (response rate 75%), and at
Time 2 (T2; 19921993) the study included follow-up inter-
views with 10,007 respondents (response rate 82%). The
current studys subsample included all respondents who
completed interviews at T1 and T2 (to allow for longitudinal
analyses), who were at least 35 years old at T1 (to focus
analyses on midlife and older adults), and who reported no
functional limitations at T1 (to allow for a prospective design),
yielding a sample size of 4,646 respondents.
Measures
Depressive symptoms. At T1 and T2 the study included
a 12-item version of the Center for Epidemiologic Studies
Depression scale (Radloff, 1977). This index asked respondents
how many days during the past week they had experienced
various depressive symptoms, such as feeling that everything
they did was an effort and talking less than usual. If respondents
answered at least 6 of the 12 items, they received an average-
item score, with higher scores indicating more depressive symp-
toms. To partly, although not entirely, correct for a positive skew
in the distribution of respondents scores on this scale, we logged
respondents scores after adding a constant of 1. Cronbachs
alpha for depressive symptoms was .93 at both T1 and T2. Table
1 provides correlations and descriptives for all analytic variables.
Personal growth. Because researchers have identified
experiences of personal growth as one dimension of psycho-
logical well-being particularly vulnerable to decline with age
(Ryff, 1995), we deemed it valuable to explore this aspect of
psychological well-being in our study. At T2, the study included
Ryffs three-item Personal Growth index. Ryff created this index
Figure 1. Conceptual model for the risk-buffering effect of
continuous participation in voluntary groups over a 5-year period on
adults psychological well-being.
for large survey use as an additive measure designed to represent
the conceptual breadth of personal growth, which she found in
factor analyzing her 20-item scale (Ryff & Keyes, 1995). This
index asked respondents to report the degree to which they
agreed or disagreed with statements indicating feelings of
personal growth on a six-point continuum (1 strongly
disagree), including For me, life has been a continuous process
of learning, changing, and growth, I gave up trying to make
big improvements or changes in my life a long time ago, and I
think it is important to have new experiences that challenge how
I think about myself and the world. Items were coded and
summed such that higher scores indicated more personal growth.
This additive index correlates highly (r . .70) with its parent
20-item, highly reliable scale (Ryff & Keyes, 1995).
Developed functional limitations. Respondents reported on
functional limitations at T1 and T2 by responding to a set of
items asking: Do you have a physical or mental condition that
limits your ability to: (a) care for personal needs, such as
dressing, eating or going to the bathroom; (b) move about
inside the house; (c) do day-to-day household tasks; (d) climb
a flight of stairs; (e) walk six blocks? To create this studys
prospective, longitudinal design, we included in the sample
only respondents who reported no to all of these items at T1.
To assess whether respondents transitioned into a state of
impaired functional health between T1 and T2, we coded as 1
respondents who reported that a health condition at all limited
their ability to engage in any of the activities at T2. We coded
all other respondents as 0 on this variable.
Continuous participation in voluntary groups. Because the
focus of this study was on voluntary group participation that is
specifically continuous over time, we created dichotomous
measures to assess whether respondents reported participation
in voluntary groups at both T1 and T2 (1 continuous
participation; 0 noncontinuous participation). More specifi-
cally, on the measure of continuous participation in recreational
groups, we coded as 1 respondents who reported at least monthly
participation in sports, hobby or garden groups, or discussion
groups at both T1 and T2. On the measure of continuous
participation in religious groups, we coded as 1 respondents who
 VOLUNTARY GROUP PARTICIPATION S63

Table 1. Correlations and Descriptive Statistics for All Analytic Variables

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
1. Depressive symptoms, T2
2. Personal growth, T2
.20
3. Depressive symptoms, T1 .39
.14
4. Whitea
.09 .00
.05
5. Blacka .06 .02 .04
.81
6. Mexican Americana .05
.05 .03
.39
.09
a
7. Other race/ethnicity .02 .01
.00
.30
.07
.00
8. ,12 years educationa .13
.21 .07
.24 .17 .16 .04
9. 12 years educationa .00
.08 .05 .09
.05
.05
.05
.38
10. 1315 years educationa
.01 .10 .00 .02
.01
.04 .02
.26
.39
11. 16 years educationa
.12 .20
.13 .12
.10
.06 .00
.27
.41
.28
12. Employeda
.06 .20 .01 .01
.00
.01 .01
.28
.04 .12 .19
13. Household income ($1,000s)
.13 .17
.11 .15
.13
.07
.02
.26
.13 .05 .35 .29
14. Child younger than 19 in household, T2 .03 .09 .04
.05
.00 .08 .02
.14
.06 .07 .13 .29 .12
15. Married, T2
.12 .02
.12 .17
.18
.04
.03
.15 .04
.01 .10 .10 .29 .21
16. Age, T1
.06
.20
.09 .05
.02 .00
.05 .30 .02
.15
.16
.58
.25
.52
.19
17. Female, T2 .13 .04 .14
.03 .01 .04 .01 .01 .11
.01
.13
.11
.14 .05
.20 .02
18. Functional limitationsa .26
.18 .12
.08 .08 .02
.00 .18 .04
.07
.15
.26
.20
.14
.11 .26 .10
19. Continuous recreationala
.04 .12
.02 .09
.07
.04
.03
.14
.05 .05 .14 .08 .11 .05 .05
.09
.07
.09
20. Continuous religiousa
.07 .10
.04
.03 .06
.03
.03
.09 .00 .01 .07
.03 .01 .01 .05 .07 .08
.00 .14
21. Continuous civica
.05 .05
.02 .03
.03 .00
.01
.06 .01 .01 .04 .03 .03
.02 .01 .00
.07
.03 .14 .05
M .57 14.75 .57 .78 .16 .04 .02 .21 .36 .20 .22 .59 44,344 .27 .58 50.36 .61 .27 .14 .17 .02
SD .49 2.53 .50 .42 .37 .20 .15 .41 .48 .40 .41 .49 41,693 .44 .49 12.65 .49 .44 .34 .37 .16
Notes: SD standard deviation; T1 Time 1; T2 Time 2. Data are from the National Survey of Families and Households, 19871993; participants at least
35 years of age and without functional limitations at T1 (N 4,646). Analyses used unweighted data.
a
Means for dichotomous variables are reported as proportions.

reported at least monthly participation in church-affiliated
groups (with the measure specifying that the respondents were
not to consider participation in religious services for this item) at
both T1 and T2. On the measure of continuous participation in
civic groups, we coded as 1 respondents who reported at least
monthly participation in service clubs, fraternal groups, or
political groups at both T1 and T2.
Control variables. Previous studies have demonstrated that
several sociodemographic factorssuch as race/ethnicity,
gender, age, education, income, parental status, marital status,
and employment statusare associated with physical health,
voluntary group participation, and/or psychological well-being
(e.g., Mroczek & Kolarz, 1998; Ryff, 1995). This study
statistically controlled for these factors in all models to reduce
confounding that might have occurred between these variables
and variables of primary substantive interest. We created di-
chotomous variables for gender, employment status, marital
status, and whether respondents had a child younger than 19
years old in the household at T2. We created multicategorical
variables to indicate respondents educational attainment at T2
and race/ethnicity at T1, and we created continuous variables
for respondents age at T1 and household income at T2.
Analytic Sequence
We first examined several sets of bivariate correlations to
provide evidence for the utility of multivariate models (see
Table 1). The modest size of the correlation between the
dependent variables (r
.20), as well as among the types of
continuous voluntary group participation (.05 , r  .14),
indicated that these variables were not empirically redundant
with each other. We also found that the correlations between
the dichotomous measure of functional limitations and the
dichotomous measures of continuous participation in voluntary
groups were small (jrj 
.09). Furthermore, we estimated
binomial logit models that regressed the log odds of continuous
participation in each type of voluntary groups on all control
variables, as well as on a continuous measure of the number of
functional limitations acquired between T1 and T2 (ranging
from 0 to 5). The coefficient for functional limitations did not
achieve statistical significance at a robust level ( p  .05) in any
of these models, further suggesting that continuous participa-
tion in voluntary groups was not significantly associated with
the severity of functional decline.
In additional preliminary analyses, we estimated models that
included three-way interaction terms indicating the product
among developing functional limitations, continuous participa-
tion in recreational or religious groups, and a dichotomous
variable indicating whether respondents were at least 47 years
old (the median age for the sample at T1). (We did not estimate
models with three-way interaction terms involving continuous
participation in civic groups because only 18 respondents
developed functional limitations between T1 and T2 and also
reported continuous participation in civic groups.) None of the
interaction terms achieved statistical significance at a robust
level ( p  .05, two-tailed). Therefore, we subsequently ana-
lyzed data from respondents of all ages together.
We estimated ordinary least squares regression models to
address our hypotheses and research questions. To provide
evidence for the associations between the independent variables
and changes in respondents psychological well-being, we re-
gressed the well-being outcomes at T2 on the control variables,
as well as on respondents scores on depressive symptoms at
T1. (Because the National Survey of Families and Households
did not elicit respondents reports of personal growth at T1, all
models estimated for respondents personal growth included
S64 GREENFIELD AND MARKS

Table 2. Estimated Unstandardized Regression Coefficients for the Effects of Developing Functional Limitations and
Continuous Voluntary Group Participation Over a 5-Year Period on Adults Depressive Symptoms
Variable Model 1 Model 2 Model 3 Model 4 Model 5
Depressive symptoms, T1 .32*** .31*** .31*** .31*** .31***
Race/ethnicity, T1a
Black .01 .01 .01 .01 .01
Mexican American .03 .01 .01 .01 .01
Other race/ethnicity .04 .06 .05 .05 .06
Years of education, T2b
,12 years .08*** .09*** .09*** .09*** .09***
1315 years .01 .01 .01 .01 .01
16 or more years
.02
.01
.01
.01
.01
Employed, T2
.07***
.06**
.05**
.06**
.06**
Household income, T2
.00
.00
.00
.00
.00
Child younger than 19 in household, T2 .01 .01 .01 .01 .01
Married, T2
.05***
.05**
.05**
.05**
.05**
Age, T1
.01***
.01***
.01***
.01***
.01**
Female, T2 .06*** .05*** .08*** .06*** .05***
Developed functional limitations between T1 and T2 .24*** .24*** .30*** .24*** .24***
Continuous participation in voluntary groups
Recreational groups .00 .04 .00 .00
Religious groups
.05**
.05**
.04*
.05**
Civic groups
.14*
.07
.07
.14*
Interactions
Civic groups 3 Female .19* .19*
Recreational groups 3 FL
.24**
Religious groups 3 FL
.03
Civic groups 3 FL
.00
Recreational groups 3 Female
.03
Female 3 FL
.08*
Recreational groups 3 FL 3 Female .23*
Constant .66*** .65*** .64*** .65*** .65*
Valid N 3801 3801 3801 3801 3801
R2 .22 .22 .22 .22 .22
Notes: T1 Time 1; T2 Time 2; FL functional limitations.Data are from the National Survey of Families and Households, 19871993; participants at least
35 years of age and without functional limitations at T1. Analyses used unweighted data.
a
Omitted category is non-Hispanic white.
b
Omitted category is 12 years.
*p  .05; **p  .01; ***p  .001 (two-tailed).

respondents depressive symptoms at T1 as a baseline well-
being control.) Because results based on analyses with the
weighted and unweighted data were comparable, we report
results based on the unweighted data because these analyses
provide estimates with more reliable standard errors (Winship &
Radbill, 1994).
To explore RQ1, we first estimated models that included two-
way interaction terms between gender and continuous partic-
ipation in each type of voluntary group (e.g., Gender 3
Continuous participation in recreational groups). To explore
RQ2, we estimated models that included three-way interaction
terms between gender, continuous participation in religious or
recreational groups, and functional limitations (e.g., Gender 3
Continuous participation in recreational groups 3 Functional
limitations). To examine evidence for H1, we estimated mul-
tivariate models by regressing the well-being outcomes on
the dichotomous variable indicating whether respondents had
developed functional limitations between T1 and T2. To ex-
amine evidence for H2, we added the block of three di-
chotomous variables indicating continuous participation in each
of the voluntary groups to the previous set of models, plus any
two-way interaction terms including gender that had achieved
statistical significance ( p  .05, two-tailed) in previous anal-
yses. To examine evidence for H3, we separately added two-
way interaction terms (i.e., the product between each type of
continuous voluntary group participation and whether respon-
dents had developed functional limitations between T1 and T2)
to a model containing all main effect variables in addition to
relevant three-way interaction terms that had achieved
statistical significance in previous analyses.
RESULTS
Gender, Continuous Voluntary Group Participation,
and Psychological Well-Being
Regarding RQ1, models including two-way interaction terms
between gender and continuous participation in each type of
voluntary group yielded evidence of one statistically significant
interaction. The interaction of Gender 3 Continuous participa-
tion in civic groups achieved statistical significance in terms of
its association with depressive symptoms (b .19, p  .05;
 VOLUNTARY GROUP PARTICIPATION S65

participation in civic groups was associated with lower levels of

Figure 2. Predicted scores of depressive symptoms for men and
women who differed as to whether they continuously participated in
civic groups.
Table 2, Model 2). We used estimates from this model to graph
predicted scores for men and women who differed as to whether
they continuously participated in civic groups across the 5-year
period. The baseline model included respondents at the mean
level on continuous variables and in the zero categories for
dichotomous variables. As Figure 2 demonstrates, continuous
depressive symptoms for men but not for women.
Regarding RQ2, models including three-way interaction terms
among continuous participation in recreational or religious
groups, functional limitations, and gender yielded evidence of
one statistically significant interaction; the three-way interaction
for Gender 3 Functional limitations 3 Continuous participation
in recreational groups achieved statistical significance in the
model for depressive symptoms (b .23, p  .05; Table 2,
Model 3). This three-way interaction term indicates that the
influence of continuous participation in recreational groups on
the association between functional limitations and depressive
symptoms differed for men and women. We further interpret this
result when describing evidence for H3.
Developing Functional Limitations and
Psychological Well-Being
H1 predicted that respondents who developed functional
limitations over a 5-year period would report poorer psycho-
logical well-being than respondents who did not develop
functional limitations. Results from models that evaluated this
hypothesis (refer to Tables 2 and 3, Model 1) indicated that
respondents who developed functional limitations reported
greater increases in depressive symptoms (b .24, p  .001)

Table 3. Estimated Unstandardized Regression Coefficients for the Effects of Developing Functional Limitations and
Continuous Voluntary Group Participation Over a 5-Year Period on Adults Personal Growth
Variable Model 1 Model 2 Model 3 Model 4 Model 5
Depressive symptoms, T1
.67***
.67***
.67***
.67***
.66***
Race/ethnicity, T1a
Black .35*** .32*** .32*** .32** .32**
Mexican American
.14
.14
.14
.15
.14
Other race/ethnicity .10 .04 .04 .05 .03
Years of education, T2b
,12 years
.41***
.43***
.43**
.41***
.42***
1315 years .60*** .56*** .56*** .56*** .56***
16 or more years .91*** .79*** .79*** .79*** .79***
Employed, T2 .37*** .31** .31** .31** .31**
Household income, T2
.00** .00** .00** .00** .00**
Child younger than 19 in household, T2
.11
.09
.09
.09
.09
Married, T2
.16*
.21*
.21*
.21*
.21*
Age, T1
.02***
.02***
.02***
.02***
.02***
Female, T1 .50*** .43*** .43*** .43*** .43***
Developed functional limitations between T1 and T2
.44***
.46***
.46***
.55***
.48***
Continuous participation in voluntary groups
Recreational groups .42*** .40*** .42*** .42***
Religious groups .44*** .44*** .30* .44***
Civic groups .46 .36 .37 .15
Interactions
Recreational groups 3 FL .10
Religious groups 3 FL .58*
Civic groups 3 FL 1.28y
Constant 15.46*** 15.51*** 15.51*** 15.54*** 15.51***
Valid N 4152 3747 3747 3747 3747
R2 .13 .14 .14 .14 .14
Note: T1 Time 1; T2 Time 2; FL functional limitations.Data are from the National Survey of Families and Households, 19871993; participants at least
35 years of age and without functional limitations at T1. Analyses used unweighted data.
a
Omitted category is non-Hispanic white.
b
Omitted category is 12 years.
y
p  .10; *p  .05; **p  .01; ***p  .001 (two-tailed).
S66 GREENFIELD AND MARKS
Figure 3. Predicted scores of depressive symptoms for men and
women who differed as to whether they continuously participated in
recreational groups and/or developed functional limitations over a 5-
year period.
and lower levels of personal growth (b
.44, p  .001) in
contrast to respondents who did not develop functional
limitations. Pending interaction results with respect to H3,
these analyses provided evidence in support of H1.
Continuous Participation in Voluntary Groups and
Psychological Well-Being
H2 predicted that respondents who continuously participated
in voluntary groups would report better psychological well-
being than respondents who did not continuously participate.
Findings from models that evaluated this hypothesis (Tables 2
and 3, Model 2) demonstrated that continuous participation in
religious groups was associated with smaller increases in
depressive symptoms (b
.05, p  .01); however, continuous
participation in recreational groups was not associated with
differences in depressive symptoms, and continuous participa-
tion in civic groups was associated with differences in
depressive symptoms among men but not among women. Fur-
thermore, continuous participation in recreational groups was
associated with higher levels of personal growth at T2 (b .42,
p  .001), as was continuous participation in religious groups
(b .44, p  .001), but continuous participation in civic groups
was not associated with levels of personal growth at T2. These
analyses provided partial support for H2.
Continuous Participation in Voluntary Groups
as a Protective Factor
H3 predicted that respondents who developed functional
limitations and who continuously participated in voluntary
groups would report better psychological well-being than
respondents who also developed functional limitations but
who did not continuously participate in voluntary groups. In
addition to further interpreting the association between the
three-way interaction among gender, continuous participation
in recreational groups, and functional limitations and depressive
symptoms reported previously, we also estimated models with
two-way interaction terms between continuous participation in
the remaining types of voluntary groups and functional
limitations in order to evaluate evidence for this hypothesis
(refer to Tables 2 and 3, Models 35). The interaction term
Figure 4. Predicted scores of personal growth for adults who
differed as to whether they continuously participated in religious
groups and/or developed functional limitations over a 5-year period.
between developing functional limitations and continuous
participation in religious groups achieved statistical significance
in models for respondents personal growth (b .58, p  .05;
Table 3, Model 4).
To interpret statistically significant interaction terms, we
computed predicted scores for psychological well-being across
groups of respondents who differed by whether they developed
functional limitations and by whether they continuously
participated in the voluntary group of interest. We used es-
timates from Table 2, Model 3, to interpret the statistically
significant three-way interaction among gender, continuous
participation in recreational groups, and functional limitations
on depressive symptoms, and we used estimates from Table 3,
Model 4, to interpret the statistically significant two-way
interaction between continuous participation in religious groups
and functional limitations on personal growth.
As Figure 3 displays, among men who did not continuously
participate in recreational groups, developing functional
limitations increased predicted levels of depressive symptoms
by two thirds of a standard deviation. Among men who re-
ported continuous participation in these groups, however,
developing functional limitations was associated with a smaller
increase in depressive symptoms over the 5-year period.
Among women, regardless of their continuous participation in
recreational groups, developing functional limitations was
associated with increasing depressive symptoms from T1 to
T2. These results indicate that continuous participation in rec-
reational groups ameliorated the problematic association be-
tween functional limitations and depressive symptoms for men
but not for women.
Figure 4 demonstrates the interactive relationship between
developing functional limitations and continuous participation
in religious groups on respondents feelings of personal growth
across men and women. For respondents who did not con-
tinuously participate in religious groups, developing functional
limitations predicted lower levels of personal growth. Among
respondents who reported continuous participation in religious
groups, however, developing functional limitations was not
associated with lower levels of personal growth.
In sum, these results indicate that continuous participation in
religious groups protected respondents from lower levels of
 VOLUNTARY GROUP PARTICIPATION
personal growth associated with developing functional limi-
tations and that continuous participation in recreational groups
protected men, in particular, from greater increases in de-
pressive symptoms associated with developing functional
limitations. As a whole, these findings provided some evidence
in support of H3.
DISCUSSION
The primary aim of this study was to investigate continuous
participation in voluntary groups as a protective factor against
the negative psychological consequences of developing func-
tional limitations. Results from the current study are congruent
with those of previous studies, which suggest that developing
functional limitations is associated with individuals poorer
psychological well-being over time (Kunzmann et al., 2000;
Taylor & Lynch, 2004). Findings also suggest, however, that
deleterious changes in psychological well-being associated with
developing functional limitations are not uniform across all
adults who experience such limitations. Increases in depressive
symptoms associated with developing functional limitations
were less severe among men who continuously participated in
recreational groups, and developing functional limitations was
not associated with lower levels of personal growth among
respondents who continuously participated in religious groups.
Evidence for these associations held even when statistically
controlling for a variety of sociodemographic factors associated
with psychological well-being, such as respondents education,
race/ethnicity, and gender. Moreover, supplemental analyses
(not shown), in which we separated respondents into three
distinct categories of noncontinuous participation (no partici-
pation at T1 or T2, participation at T1 but not at T2, and
participation at T2 but not at T1), indicated that the pattern of
psychological well-being benefits associated with continuous
participation in the face of functional decline remained in terms
of contrasts with each of the noncontinuous participation
subgroups.
Although these results provide at least some evidence in
support of the beneficial effects of continuous voluntary group
participation on adults psychological well-being, not all types
of voluntary group participation demonstrated independent and/
or interactive effects on both of the psychological well-being
outcomes examined (i.e., depressive symptoms and personal
growth). Additionally, although we observed few gender differ-
ences in the associations between continuous voluntary group
participation and psychological well-being, results indicated
some evidence for the idea that the psychological effects of
voluntary group participation might differ for men and women.
For example, whereas continuous participation in recreational
groups protected men from smaller increases in depressive
symptoms associated with developing functional limitations, we
did not find this protective effect among women. Such complex
patterns of significant and nonsignificant associations suggest
that the ways and degrees to which particular types of continuous
voluntary group participation affect certain aspects of psycho-
logical well-being are likely variable. Additional empirical and
theory-based work is necessary to better understand the sub-
stantive significance of these complex patterns.
The results also indicated that the ways in which continuous
participation in voluntary groups moderates the association
S67
between functional decline and psychological well-being are
similarly complex. Whereas continuous participation in recre-
ational groups served as a moderator for mens psychological
well-being in terms of lessening the severity of the association
between developing functional limitations and increasing
depressive symptoms, continuous participation in religious
groups moderated the association between developing func-
tional limitations and lower levels of personal growth by
eliminating this association altogether for both men and
women. We speculate that because functional decline is a
more potent risk factor for poorer psychological well-being in
terms of more depressive symptoms than lower levels of
personal growth, continuous participation in recreational
groups might not be powerful enough to protect individuals
against experiencing any increases in depressive symptoms
upon developing functional limitations. On the other hand,
because developing functional limitations is a less powerful risk
factor for personal growth, and because continuous participa-
tion in voluntary groups is more robustly associated with this
dimension of well-being, continuous participation in religious
groups can altogether shield individuals from loss of personal
growth following functional decline.
Both types of moderation processes (i.e., the amelioration
and elimination of the association between functional decline
and psychological well-being) are congruent with continuity
theory. Continuity theory suggests that behavioral and mental
strategiessuch as maintaining participation in formal volun-
tary groupscan help individuals to achieve a consistent sense
of self and their social worlds and thereby experience greater
well-being when faced with life changes (Atchley, 1989).
Alternative processes, however, might account for the statisti-
cally significant interactions found in this study. Although this
study draws on the methodological strengths of a longitudinal,
prospective design, it remains possible that respondents main-
tenance of higher levels of psychological well-being in the face
of declining functional health caused them to remain involved
in voluntary groups, rather than their continuous participation
helping them to maintain their psychological well-being.
Additionally, although the associations between the severity
of respondents functional decline and their continuous par-
ticipation in the voluntary groups across T1 and T2 were
modest in size, the potential selection process from risk to
protective factor suggests that continuous participation in
voluntary groups among respondents with functional limita-
tions could indicate a lesser degree of functional decline within
this group. This lesser degree of functional decline might
explain why respondents with continuous participation in
voluntary groups reported better psychological well-being.
Finally, although we investigated voluntary group participation
in terms of three types of groups, the remaining heterogeneity
of types of groups within each of these categories might mask
the implications of particular types of voluntary group
participation for psychological well-being.
Despite these limitations, this studys findings contribute to
a better understanding of the psychosocial implications of
functional limitations and factors comprising processes of risk
and resilience in adulthood. Whereas results from previous
studies have indicated that various types of voluntary group
participation can promote individual well-being within the U.S.
adult population at large (Menec, 2003; Morrow-Howell et al.,
S68 GREENFIELD AND MARKS
2003), the findings from this study demonstrate that voluntary
group participation might also promote adults psychological
well-being by buffering them against losses in psychological
well-being following the onset of functional limitations. Fur-
thermore, the results suggest the advantages of voluntary group
participation that extends across years of an individuals
adulthood, thereby supporting the importance of efforts not
only to engage, but also to maintain, adults participation in
voluntary groups. Future research on the precise mechanisms
through which voluntary group participation promotes partic-
ular aspects of psychological well-being, as well as on factors
that facilitate adults ongoing participation, holds the promise
of contributing to a better understanding of how psychological
well-being can be maintained across adulthoodparticularly in
the face of aging-related challenges.
ACKNOWLEDGMENTS
Support for this research was provided by grants from the National
Institute of Mental Health (MH61083) and the National Institute on Aging
(AG20166 and AG206983). Address correspondence to Emily Greenfield,
University of Wisconsin-Madison, Human Development and Family
Studies, 1430 Linden Drive, Madison, WI 53706. E-mail: eagreenfield@
wisc.edu.
REFERENCES
American Association of Retired Persons. (1998). Aging everywhere.
Washington, DC: Author.
Atchley, R. C. (1989). A continuity theory of normal aging. The
Gerontologist, 29, 183190.
Atchley, R. C. (1993). Continuity theory and the evolution of activity in
later adulthood. In J. Kelly (Ed.), Activity and aging (pp. 516).
London: Sage.
Brief, A. P., Butcher, A. H., & George, J. M. (1993). Integrating bottom-up
and top-down theories of subjective well-being: The case of health.
Journal of Personality and Social Psychology, 64, 646653.
Cutler, S. J., & Hendricks, J. (2000). Age differences in voluntary asso-
ciation memberships: Fact or artifact? Journal of Gerontology: Social
Sciences, 55B, S98S108.
Ellison, C. G., & Levin, J. S. (1998). The religionhealth connection:
Evidence, theory, and future directions. Health Education and
Behavior, 25, 700720.
Greenfield, E. A., & Marks, N. F. (2004). Formal volunteering as
a protective factor for older adults psychological well-being. Journal of
Gerontology: Social Sciences, 59B, S258S264.
Janoski, T., & Wilson, J. (1995). Pathways to voluntarism: Family
socialization and status transmission models. Social Forces, 74, 271292.
Kono, A., Kai, I., Sakato, C., & Rubenstein, L. Z. (2004). Frequency of
going outdoors: A predictor of functional and psychosocial changes
among ambulatory frail elders living at home. Journal of Gerontology:
Medical Sciences, 59A, M275M280.
Kunzmann, U., Little, T. D., & Smith, J. (2000). Is age-related stability of
subjective well-being a paradox? Cross-sectional and longitudinal
evidence from the Berlin Aging Study. Psychology and Aging, 15,
511526.
Lawton, M. P. (1991). Functional status and aging well. Generations, 15,
3135.
McCusker, J., Kakuma, R., & Abrahamowicz, M. (2002). Predictors of
functional decline in hospitalized elderly patients: A systematic review.
Journal of Gerontology: Medical Sciences, 57A, M569M577.
Menec, V. H. (2003). The relation between everyday activities and
successful aging: A six-year longitudinal study. Journal of Gerontol-
ogy: Social Sciences, 58B, S74S82.
Morrow-Howell, N., Hinterlong, J., Rozario, P. A., & Tang, F. (2003).
Effects of volunteering on the well-being of older adults. Journal of
Gerontology: Social Sciences, 58B, S137S145.
Mroczek, D. K., & Kolarz, C. M. (1998). The effect of age on positive and
negative affect: A developmental perspective on happiness. Journal
of Personality and Social Psychology, 75, 13331349.
Okun, M. A. (1984). Health and subjective well-being: A meta-analysis.
International Journal of Aging and Human Development, 19, 111132.
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for
research in the general population. Journal of Applied Psychological
Measures, 1, 385401.
Rowe, J. W., & Kahn, R. L. (1997). Successful aging. The Gerontologist,
37, 433440.
Ryff, C. D. (1995). Psychological well-being in adult life. Current
Directions in Psychological Sciences, 4, 99104.
Ryff, C. D., & Keyes, C. L. M. (1995). The structure of psychological
well-being revisited. Journal of Personality and Social Psychology, 69,
719727.
Silverstein, M., & Parker, M. G. (2002). Leisure activities and quality of life
among the oldest old in Sweden. Research on Aging, 24, 528538.
Spiro, A., & Bosse, R. (2000). Relations between health-related quality of
life and well-being: The gerontologists new clothes? International
Journal of Aging and Human Development, 50, 297318.
Sweet, J. A., & Bumpass, L. L. (1996). The National Survey of Families
and Households waves 1 and 2: Data description and documentation.
Retrieved September 1, 2005, from http://www.ssc.wisc.edu/nsfh/
home.htm
Sweet, J. A., Bumpass, L. L., & Call, V. R. A. (1988). The design and
content of the National Survey of Families and Households (NSFH
Working Paper No. 1). Madison: University of Wisconsin, Center for
Demography and Ecology.
Taylor, M. G., & Lynch, S. M. (2004). Trajectories of impairment, social
support, and depressive symptoms in later life. Journal of Gerontology:
Social Sciences, 59B, 238246.
Winship, C., & Radbill, L. (1994). Sampling weights and regression
analysis. Sociological Methods and Research, 23, 230257.
Zarit, S. H., & Zarit, J. M. (1998). Mental disorders in older adults:
Fundamentals of assessment and treatment. New York : Guilford Press.
Received October 6, 2005
Accepted March 15, 2006
Decision Editor: Charles F. Longino, Jr., PhD