Full text of DOI:10.1093/geronb/56.5.S311

Journal of Gerontology: SOCIAL SCIENCES
Copyright 2001 by The Gerontological Society of America
2001, Vol. 56B, No. 5, S311–S320
Childlessness and the Psychological Well-Being of
Older Persons
Zhenmei Zhang and Mark D. Hayward
Population Research Institute and Department of Sociology, The Pennsylvania State University, University Park.
Objectives. Rapid growth in the size of the childless elderly population has prompted concerns about the negative ef-
fects of childlessness on psychological well-being. This study adds to this line of inquiry by examining the effects of
childlessness on two important dimensions of elderly persons’ psychological well-being: loneliness and depression.
Methods. Using the 1993 Asset and Health Dynamics Among the Oldest Old data set, the authors estimated logistic
and ordinary least squares regression models of psychological well-being for a nationally representative sample of peo-
ple aged 70 and older (N ꢀ 6,517).
Results. Childlessness per se did not significantly increase the prevalence of loneliness and depression at advanced
ages, net of other factors. There also was no statistical evidence for the hypothesis that childlessness increases loneliness
and depression for divorced, widowed, and never married elderly persons. Sex, however, altered how childlessness and
marital status influenced psychological well-being. Divorced, widowed, and never married men who were childless had
significantly higher rates of loneliness compared with women in comparable circumstances; divorced and widowed men
who were childless also had significantly higher rates of depression than divorced and widowed women.
Discussion. The findings suggest that it is important to understand the consequences of childlessness in the context
of marital status and sex.
HE childless elderly population in the United States has
grown quickly in recent years, although surprisingly lit-
for depression and mortality (Cappeliez & Flynn, 1993). Al-
though research has shown that most childless persons often
have support networks consisting of relatives, friends, and
neighbors, those networks are less likely to provide the
long-term commitment and level of instrumental support
provided by children. In other words, the type of lasting
support that children can provide is not as readily available
to many childless persons (Beckman & Houser, 1982; Choi,
1994). The fear and reality of support inadequacy that may
accompany childlessness potentially imposes a heavy bur-
den on childless elders’ psychological well-being (Rubin-
stein, 1987).
The empirical evidence on this issue, however, is incon-
sistent. On the one hand, childless elderly persons score sig-
nificantly lower on objective measures of social support.
For example, childless elderly persons are more likely than
elderly parents to live alone, have fewer close family ties,
and have less social contact (Chapman, 1989; Koropeckyj-
Cox, 1998; McMullin & Marshall, 1996). In addition, child-
less elderly persons are less likely to expect to have any
caregivers in the event of major bouts of sickness (Choi,
1994). Despite these social support deficits, however, the
majority of available empirical evidence suggests that psy-
chological well-being does not differ significantly between
elderly parents and childless elderly persons (e.g., Glenn &
McLanahan, 1981; Koropeckyj-Cox, 1998; McMullin &
Marshall, 1996; Rempel, 1985). Only a few studies have
reported otherwise (Beckman & Houser, 1982; Kandel,
Davies, & Raveis, 1985). In view of the vast array of studies
reporting that social support directly affects psychological
well-being and mediates the effects of undesirable life
T
tle is known about this group’s health, social service utiliza-
tion, and general well-being. In 1990, childless persons ac-
counted for approximately one fifth of the population aged
older than 65 (Himes, 1992). A special report in American
Demographics (“Childless Couples,” 1993) projected that
childless couples will be one of the fastest growing segments
of the elderly population over the next two decades owing to
delayed marriage, infertility, voluntary childlessness, and
high divorce rates. Wachter (1997) estimated that about 25%
of the elderly persons aged 70–85 years in 2030 will not
have a living spouse, living children, or living stepchildren.
Concerns over the psychological well-being of childless
elderly persons have arisen out of the traditional beliefs
about the importance of adult children as sources of emo-
tional, physical, and financial support for elderly parents
(Brubaker, 1990; McMullin & Marshall, 1996). Some re-
searchers have suggested that “the mere knowledge that off-
spring are there to help if they are needed could contribute
substantially to the psychological well-being of the elderly”
(Glenn & McLanahan, 1981, p. 441). In addition, children
as well as grandchildren can provide elderly family mem-
bers with a sense of meaning and security, immortality, suc-
cess, companionship, and affection (Alexander, Rubinstein,
Goodman, & Luborsky, 1992; Rubinstein, 1987). Childless
people are presumed to fare less well at older ages than par-
ents, because childless elderly persons lack a crucial social
support network created by children (Choi, 1994).
It is well established in the literature that social isolation
and low levels of social support are important risk factors
S311

S312
ZHANG AND HAYWARD
events (e.g., Hammen, 1997; Lin, Dean, & Ensel, 1986), the
similarity in the psychological well-being of elderly child-
less persons and parents seems paradoxical.
ported few statistically significant differences between par-
ents and childless elderly persons in terms of life satisfac-
tion. A more recent analysis of data from the 1988 National
Survey of Families and Households showed similarly that
childlessness is not significantly related to loneliness and
depression among American community-dwelling persons
aged 50–84 (Koropeckyj-Cox, 1998). Such findings are
echoed by research based on a Canadian sample (McMullin
& Marshall, 1996).
Although no significant differences in psychological
well-being were found between parents and nonparents at
the population level, recent research has raised the ques-
tion of whether parental status has complex interactive ef-
fects depending on marital status and sex. Koropeckyj-Cox
(1998) argued that marital status should mediate the nega-
tive effects of childlessness on psychological well-being.
Childlessness is likely to have few negative effects for
married couples whose needs are generally provided for
within marriage. Childlessness, however, may have more
negative consequences for the widowed group, who are
likely to have the greatest shortfall of familial support. Ko-
ropeckyj-Cox suggested that no adverse effect of child-
lessness is expected among the divorced because child-
lessness results in fewer stresses and negotiations in the
divorce process. Childlessness is unlikely to be detrimen-
tal to never married persons because of the likelihood of
their having formed a long-term social network that helps
in contingencies.
The empirical results by Koropeckyj-Cox (1998) pro-
vided mixed support for the above-mentioned arguments.
Using data from the National Survey of Households and
Families, a survey limited to community-dwelling adults
aged less than 85, Koropeckyj-Cox found no significant det-
rimental effect of childlessness on psychological well-being
for married and never married adults. Within almost every
ever-married subgroup, childless persons were more likely
to report feelings of loneliness and depression than those
who had children, net of other factors (although these differ-
ences were not statistically significant).
Beckman and Houser (1982) also found support for the
idea that the effect of childlessness is contextualized by
marital status. On the basis of data for 719 older women
aged 60–75, Beckman and Houser observed that childless-
ness had more negative consequences for well-being among
widows than among married women. Childless widows
were more lonely and depressed than widows with children,
net of other factors. Unfortunately, whether this result can
be generalized to the whole population of older women is
unknown, because the sample was not randomly selected.
Another line of research has suggested that sex is an addi-
tional contextualizing factor that must be considered for in-
vestigators to understand the consequences of childlessness
for psychological well-being. Parenthood is often assumed
to be more important for women than for men, with mother-
hood bringing both greater fulfillment and more restrictions
than fatherhood (Alexander et al., 1992; McLanahan & Ad-
ams, 1987; Veroff, Kulka, & Douvan, 1981). Lending sup-
port to this idea is the report by Kandel and colleagues
(1985) that parenthood is negatively associated with the
prevalence of depression among women. Balancing these
In this article we revisit this paradox and contribute to the
current discussion about the psychological well-being of
childless elderly persons in several ways. First, we take ad-
vantage of a recent nationally representative sample of non-
institutionalized people aged 70 and older in 1993, thus
avoiding possible limitations of prior studies (e.g., Alex-
ander et al., 1992; Beckman & Houser, 1982; Callan, 1987;
Connidis & McMullin, 1993; Dugan & Kivett, 1994; Kivett
& Learner, 1980) based on small or nonrepresentative
samples. Most research on childless elderly persons has fo-
cused on young-old persons (e.g., persons aged 70 and
younger) who are relatively healthy. At least one study
showed that childless persons in good health were able to
avoid social isolation by frequent non-kin contact but
faced a high probability of isolation when in poor health
(Bachrach, 1980). Little research has included persons aged
85 and older. Health declines rapidly among this age group,
triggering the need for social support and potentially elevat-
ing the value of children as a source of that support.
Second, we argue that it is important to study the effects
of childlessness on psychological well-being in the context
of marital status and sex. Although adult children are impor-
tant caregivers, children are less likely to provide the kinds
of support that a marital partner does on a daily basis. Thus,
childlessness may have little negative effect on married el-
derly couples but may aggravate the plight of single per-
sons. In addition, men generally are less integrated socially
than women. This gap may be worsened by childlessness
and being unmarried, with negative consequences for men’s
psychological well-being.
As part of this study, we have extended prior research on
the effects of childlessness by distinguishing between com-
pletely childless people (i.e., no biological children or step-
children) and biologically childless people with stepchil-
dren. American family forms are becoming increasingly
diverse, and stepfamilies have become an important family
form. One third of Americans are involved in stepfamily re-
lationships (Larson, 1992), and it is increasingly likely that
biologically childless persons are becoming stepparents.
Little is known, however, about whether stepchildren are
substitutes for biological children in providing social sup-
port to elderly parents and the subsequent effect of stepchil-
dren on elderly parents’ psychological well-being.
Background
Does childlessness matter for the psychological well-
being of elderly persons? The answer is still not conclusive.
On the whole, the empirical literature provides little evi-
dence that children make a difference in the psychological
well-being of elderly persons (Glenn & McLanahan, 1981;
Kivett & Learner, 1980; Koropeckyj-Cox, 1998; McMullin
& Marshall, 1996; Rempel, 1985). Using pooled cross-sec-
tional data of about 1,500 people aged 50 and older from the
1973–78 General Social Survey, Glenn and McLanahan
(1981) found little support for the idea that older parents
were happier and more satisfied with their life than childless
people. In a similar study in Canada, Rempel (1985) re-

CHILDLESSNESS AND WELL-BEING
S313
findings is evidence that parenthood has more negative con-
sequences for women than men because of less marital sup-
port and the stress of balancing work and family (McLana-
han & Adams, 1987).
A second caveat is that adaptation among childless per-
sons may result in social support networks in which child-
less persons are fully integrated and support is provided.
McMullin and Marshall (1996) observed, for example, that
although older childless persons were less integrated in
close family networks, they were equally integrated in close
friend networks compared with parents. Other evidence
suggests that childless persons depend more on siblings,
nieces, and nephews than elderly parents (Choi, 1994). Un-
clear, however, is the question of the substitutability of the
sources of social support and the consequences for psycho-
logical well-being. One study found that a close family
member can mediate the negative effects of stress on well-
being but a close friend does not (McMullin & Marshall,
1996). Chapman (1989) noted that compared with distant
kin and friends, spouses and children provide better and
more intensive care to elderly persons. This finding sug-
gests that close family ties cannot be easily replaced by
other networks and that childless elderly persons are at
greater risk of loneliness and depression (McMullin & Mar-
shall, 1996).
Sex differences in social integration and support favor the
idea that men have much smaller social support networks
outside of the immediate family and may thus benefit more
from having children than women do (Chapman, 1989). Us-
ing data from the 1991 Canadian Survey on Aging, Wu and
Pollard (1998) observed that women were more often in-
volved in support exchange than men. In general, support
exchange directly corresponds with support receipt. Com-
bining this idea with research suggesting that unmarried el-
derly persons have a shortfall in social support, one would
expect that childless elderly men are at greater risk of social
isolation and distress than childless women. Koropeckyj-
Cox (1998) examined this issue but found no evidence for
the idea that the effects of childlessness on loneliness and
depression differed for men and women. However, she did
not examine the possibility that sex differences in the ef-
fects of childlessness are contingent on marital status. Sex
differences in the effects of childlessness may be minimal
for older married couples because of the presence of the
spouse. In the absence of a spouse, however, childlessness
may become more salient. To our knowledge, prior research
has not examined the simultaneous contextualizing influ-
ences of sex and marital status on the effects of childless-
ness for elderly persons’ psychological well-being.
One caveat in the interpretation of the effects of child-
lessness is whether childlessness is voluntary or involun-
tary. One plausible hypothesis is that childlessness reduces
psychological well-being only if it is involuntary. Individu-
als who want to have children but are unable may feel
stressed and perceive that they lack control over their own
lives, leading to depressed mood. Persons who are voluntar-
ily childless, on the other hand, view childlessness almost
by definition as a desired lifestyle, and over the life course
these persons successfully develop their own social support
networks that buffer life stressors and enhance psychologi-
cal well-being (Andrews, Abbey, & Halman, 1991; Callan,
1987; Connidis & McMullin, 1993). Connidis and McMul-
lin (1993) observed, for example, that the psychological
well-being of voluntarily childless persons did not differ
significantly from that of parents who were emotionally
close to their offspring. However, involuntarily childless
persons were less happy and more depressed than emotion-
ally close parents.
METHODS
Data
Our study drew on the first wave (1993) of the Asset and
Health Dynamics Among the Oldest Old (AHEAD). These
data are nationally representative of community-dwelling
persons aged 70 years and older (N ꢀ 7,443). We included
married, divorced, widowed, and never married persons
who provided complete demographic and health informa-
tion. We excluded age-ineligible respondents, because they
did not constitute a representative sample. We also excluded
respondents who were cohabiting or married with absent
spouses (N ꢀ 142) and persons for whom information was
obtained by proxy reports (N ꢀ 774). Another 8 respon-
dents who did not answer key questions, such as health, as-
sets, and demographic information, were also excluded. The
final sample used in our analysis included 6,517 respon-
dents, 4,081 women and 2,436 men.
AHEAD respondents were generally born between 1900
and 1923. Those born in the years 1901–10, that is, persons
aged 83–92 at the time of the survey, had children primarily
during the Great Depression, a period in which childless-
ness increased to historic highs. Figure 1 shows the percent-
age of AHEAD respondents who reported being childless
by age cohort and sex. Childlessness was defined as having
not given birth to or sired any children; individuals may
have had stepchildren, however, through marriage. The age
cohorts 80–84 and 85–89, indeed, had the highest percent-
age of childlessness for women, reaching 18.5% and 25.6%,
respectively.
We view this as an important distinction—particularly
given the rise of voluntary childlessness since the 1960s (De
Jong & Sell, 1977). Prior to the 1960s, however, the avail-
able evidence suggests that the substantial majority of child-
less persons were childless for largely involuntary reasons
(Blake, 1974; De Jong & Sell, 1977). Although we were un-
able to determine the reasons for childlessness in our data,
our respondents’ prime years of family formation occurred
during the Great Depression and the 1940s—two decades
when childlessness more than likely resulted from involun-
tary reasons. Childless persons in our data, therefore, were
presumed to be at greater risk of social isolation and distress
relative to parents.
The restriction of the baseline sample to community-
dwelling persons omitted persons who were institutionalized.
Research has shown that childlessness significantly in-
creases the risk of institutionalization (Freedman, 1996; Pal-
more, 1976). Under the assumption that childless persons
have lower well-being than elderly parents, we anticipated
that the association between childlessness and psychologi-

S314
ZHANG AND HAYWARD
crisis. In this way, social and economic resources function
as a buffer under acute conditions (Cappeliez, 1993).
In this study we focused on self-reported depressive
symptoms. Depressive symptoms are often used in popula-
tion studies as criteria to screen for possible clinical depres-
sion. An advantage of depressive symptoms (hereafter re-
ferred to as depression) rather than clinical diagnosis is that
symptoms are less likely to reflect lower use of hospital ser-
vices by poor and isolated people seeking treatment of psy-
chiatric illness. Depression in the AHEAD was measured
with a modified shorter version of the Center for Epidemio-
logic Studies-Depression (CES-D) scale. Recent research
has shown that use of the short form of the CES-D results in
little loss of information, compared with the original scale
(Soldo, Hurd, Rodgers, & Wallace, 1997). Respondents are
asked to answer yes or no to 10 statements about the feel-
ings they experienced last week. The items include ques-
tions on how often the respondent felt happy, felt sad, en-
joyed life, had restless sleep, felt tired, and felt lonely. High
scores indicate greater levels of depression. For the current
sample, the Cronbach’s alpha was .79.
Figure 1. Percentage of persons who were childless by age group
and sex, Asset and Health Dynamics Among the Oldest Old, 1993 (N ꢀ
6,517). Childlessness was defined as having no biological children.
cal well-being would be attenuated, resulting in conserva-
tive tests of the hypotheses.
Like many depression indexes, the short-version CES-D
has a positively skewed distribution that can produce prob-
lems in meeting the distributional assumptions of ordinary
least squares (Hays et al., 1998; OLS; Sen & Srivastava,
1990). We thus suggest caution in evaluating the OLS re-
gression estimates. Previous research has often used a cutoff
point (roughly the highest 20% of depression scores for the
CES-D scale) to define depressed and not depressed catego-
ries. This categorization is predictive of clinical depression
(Koropeckyj-Cox, 1998; Lin et al., 1986). We used this cri-
terion to dichotomize the depression scores. Scores above 3
were recoded as 1, indicating the presence of depression,
and scores equal to or below 3 were coded as 0, indicating
no depression. In our sample, about 23.7% of respondents
had depression scores higher than 3. In our analysis, we ex-
amined the sensitivity of our results to alternative cutoff
points. We also used the continuous measure of depression
(scored 0–10, with 10 indicating the highest number of depres-
sive symptoms) to assess whether our modeling results were
sensitive to OLS versus logistic regression specifications.
Dependent Variables
Psychological well-being can be conceptualized and
measured in various ways. In this article we focus on two
dimensions of psychological well-being: loneliness and de-
pression. “Loneliness is the subjective negative feelings as-
sociated with perceived social isolation, a lower level of
contact than that desired or the absence of a specific desired
companion” (Wenger, Davies, Shahtahmasebi, & Scott,
1996, p. 333). It has long been associated with old age and
childless elderly persons in particular. Its negative charac-
teristics are also reflected in the association of loneliness
with depression, anxiety, and grief (Dugan & Kivett, 1994).
Loneliness was measured in AHEAD with a single item.
Respondents were asked to answer yes or no to the follow-
ing statement: “(Much of the time during the past week) I
felt lonely.”
Depression, the most prevalent mental disorder of older
adults, is a multifaceted and multicausal illness. The psy-
chosocial model of the etiology of depression suggests that
depression usually results from the interplay of stressful life
events, the resources of the people, and their coping strate-
gies (Chaisson-Stewart, 1985). Elderly people’s susceptibil-
ity to depression is often thought to be related to the nega-
tive aspects of the aging processes—“a progression of
disengagements, separations, and losses at numerous levels,
from the physical to the social” (Cappeliez, 1993, p. 332).
Under such circumstances, the availability of the resources
and the effectiveness of the coping strategy seem particu-
larly important in fending off depression for elderly per-
sons. Previous research has shown that an individual’s
resources including social support, health, income, and edu-
cation enhance psychological well-being in two ways.
These resources reduce exposure to life stress and promote
a sense of self-worth, achievement, and happiness. In this
sense, social and economic resources have a lasting protec-
tive effect over the life course against loneliness and depres-
sion. These resources also mediate the negative effects of
stress in life by promoting a positive coping adaptation to
Independent Variables
The core independent variables in this study were child-
lessness, marital status, and sex. We also included covari-
ates such as age (in years), race (Blacks, Whites, and oth-
ers), physical health (self-rating of health as excellent, 5, to
poor, 1), education (number of years of schooling), and net
household wealth (in dollars) to control for other resources
that elderly persons have to protect them against loneliness
and depression. We adjusted net household wealth because
of its skewed distribution by adding a constant to all house-
holds to eliminate negative wealth and then logged the
value. We included race as a covariate because it is a marker
of social inequality, discrimination, different cultural beliefs
and practices, and small but uncertain biological factors, all
of which can have different consequences on people’s psy-
chological well-being (Jones, 1992).
Childlessness was defined primarily as a social rather
than a biological variable. Most previous studies have de-

CHILDLESSNESS AND WELL-BEING
S315
Table 1. Sample Characteristics for Persons Aged 70 Years and
Older by Parental Status and Sex, AHEAD, 1993
fined childlessness as having no biological children. This
definition does not reflect the possible acquisition of paren-
tal status via marriage and divorce. In our analytic sample,
about 12% of biologically childless women and 22% of bio-
logically childless men had stepchildren. We regarded step-
children as a potential source of social support. Childless-
ness was defined in our analysis as having neither biological
children nor stepchildren. This status was contrasted with
respondents reporting stepchildren but no biological chil-
dren and respondents reporting biological children (and pos-
sibly stepchildren as well). The last category, biological par-
ent, was designated as the reference category in the analysis.
Unfortunately, the AHEAD does not provide information
about whether respondents had adopted children. We were
thus unable to identify whether biologically childless per-
sons were parents by adoption; thus we classified these indi-
viduals as childless. To the best of our knowledge, extant
data are not available to give us some sense of what percent-
age of childless persons in these birth cohorts acquired pa-
rental status via adoption.
Significance Testa
Variable
Parents Stepparents Childless
I
II
III
Women (n ꢀ 4,081)
n
3,413
2.39
0.27
0.24
77.34
2.97
81
2.19
0.19
0.17
77.44
3.02
587
2.37
0.25
0.26
78.72
2.97
Depression
Depression (1/0)
Loneliness (1/0)
Age (years)
Physical health
–
–
–
–
–
–
–
–
**
–
*
–
–
–
–
–
–
–
Education (years) 10.79
11.93
169,771
11.16
149,051
**
–
Wealth ($)
Marital status (%)
Married
Divorced
Widowed
Never married
Race (%)
Whites
149,497
–
0.37
0.06
0.57
0.005
0.46
0.01
0.53
0
0.23
0.09
0.45
0.23
–
**
–
**
**
**
**
**
**
–
–
**
0.85
0.13
0.02
0.70
0.28
0.01
0.78
0.20
0.02
**
**
–
**
**
–
–
–
–
Blacks
Others
Marital status measured the respondent’s status at the
time of the baseline survey. Marital status was defined as
married, divorced, widowed, and never married, with mar-
ried serving as the reference category.
Men (n ꢀ 2,436)
n
2,069
1.85
0.18
0.15
76.90
3.05
82
285
Depression
Depression (1/0)
Loneliness (1/0)
Age (years)
Physical health
2.20
0.23
0.15
77.29
2.73
2.09
0.23
0.22
77.62
2.95
–
–
–
–
*
–
–
–
–
**
*
–
–
–
–
–
–
–
–
–
Statistical Analysis
Loneliness and depression were defined as binary vari-
ables taking on the values of 1 (lonely/depressed) or 0 (not
lonely/not depressed), and we used a logistic regression ap-
proach to statistically model the two measures of psycho-
logical well-being. In the absence of a well articulated theo-
retical model, we did not attempt to model the possible
interdependence between these two facets of psychological
well-being. Instead we treated each outcome as distinct in
the modeling. As noted previously, we used OLS regression
to analyze the continuous score of depression.
We began our analysis by examining the main effect of
childlessness on psychological well-being, controlling for
marital status, sex, and the other covariates. We then exam-
ined whether childlessness was conditional on marital sta-
tus. Because of sparse data problems, we created a nine-cat-
egory family status measure based on the interaction of the
four marital statuses and three parental statuses. Married
persons with children were treated as the reference cate-
gory. In creating this measure, we combined divorced per-
sons with stepchildren together with divorced persons with
biological children because of the small number of divorced
stepparents. Lastly, to examine how sex contextualized the
effects of family status, we introduced interaction terms be-
tween sex and the family status categories into the model.
Education (years) 11.11
10.83
271,488
10.88
225,592
Wealth ($)
Marital status (%)
Married
Divorced
Widowed
Never married
Race (%)
Whites
242,097
–
0.76
0.05
0.19
0.003
0.72
0.04
0.23
0
0.52
0.06
0.22
0.20
–
–
–
–
**
–
–
**
–
–
**
**
0.87
0.11
0.02
0.87
0.12
0.12
0.83
0.15
0.02
–
–
–
–
–
–
–
–
–
Blacks
Others
Notes: All means and percentages are based on unweighted sample data.
Some percentages do not add up to 100% because of rounding. AHEAD ꢀ As-
set and Health Dynamics Among the Oldest Old.
aThe t statistics for Columns I, II, and III identify covariate differences for
parents and stepparents (I), parents and childless persons (II), and childless per-
sons and stepparents (III).
*p ꢁ. 05; **p ꢁ. 01.
found little statistical evidence that childlessness was asso-
ciated with depression among women and men. Overall, the
results shown in Table 1 suggest that childlessness was
more strongly associated with loneliness than depression,
and the strength of association between childlessness and
loneliness appeared stronger for men than women.
RESULTS
The results shown in Table 1 also indicate that childless
women were not systematically disadvantaged in terms of
the various resources examined in this study. Not surpris-
ingly, marital status was the characteristic most strongly as-
sociated with childlessness. Childless women were less
likely to be currently married or widowed, and they were
more likely to be divorced or never married. The results for
men were similar.
Descriptive Statistics
The results shown in Table 1 indicate that childless men
were more likely to feel lonely (22%) compared with par-
ents (15%). For women, loneliness was also relatively high
among childless persons (26%), although the same was true
among parents (24%). Childlessness appeared to have dif-
ferent consequences for loneliness for men and women. We

S316
ZHANG AND HAYWARD
Table 3. Main Effects of Childlessness on Depression: Logistic
Regression Results for Persons Aged 70 and Older, AHEAD,
Main Effects of Childlessness
Net of resources, how did childlessness influence the psy-
chological well-being of the older persons? The results
shown in Tables 2, 3, and 4 indicate that childless persons
were not more likely to be lonely or depressed compared
with parents and stepparents once marital status was con-
trolled. Note also that stepparents were not significantly dif-
ferent from elderly parents in terms of both aspects of psy-
chological well-being.
Although parental status was not statistically associated
with psychological well-being, marriage appeared to bring
substantial psychological benefits: Married persons had
lower rates of loneliness and depression compared with all
other marital status groups. Psychological well-being was
also significantly better among persons with higher levels of
education, better physical health, and greater wealth—a pat-
tern of results reported consistently in prior studies of psy-
chological well-being.
1993 (N ꢀ 6,517)
Independent Variable Model 1 Model 2 Model 3 Model 4 Model 5
Parents (reference)
Stepparents
Childless
0.847
1.026
0.895
1.003
0.896
0.952
0.917
0.950
0.909
0.933
Sex (male ꢀ 1)
Married (reference)
Divorced
0.633**
0.810** 0.779**
2.233** 2.141** 1.783**
2.113** 1.954** 1.672**
Widowed
Never married
Age
Education
Physical health
Wealth (log)
Whites (reference)
Blacks
1.613** 1.534*
1.419
1.014*
0.949**
0.466**
0.742**
0.870
Others
0.616*
0.765
ROC
0.503
0.552
0.594
0.603
ꢂ2log-likelihood
7140.0
7085.4
6982.7
6973.1
6052.2
Childlessness and Marital Status
Notes: Odds ratios are reported here. Confidence intervals are available on
request. AHEAD ꢀ Asset and Health Dynamics Among the Oldest Old.
*p ꢁ .05; **p ꢁ .01.
Although the main effects of childlessness were not sta-
tistically significant, one of our core arguments is that the
effects of childlessness on psychological well-being are
more likely to arise for some marital status groups than oth-
ers. We did not anticipate low levels of psychological well-
being among married couples because of the presence of
marital partners who provide major social support. We an-
ticipated that childlessness would be more likely to have
negative consequences for single persons because childless-
ness, in addition to the lack of a spouse, narrows the social
support network. To test whether childlessness was condi-
tional on marital status, we next examined the effects of our
family status measure based on the interaction of four mari-
tal statuses and three parental statuses.
sociated with loneliness among married persons, when other
variables were controlled. In contrast to what we expected,
however, we found little evidence for the idea that child-
lessness had negative effects on psychological well-being
among divorced, widowed, and never married persons. Al-
though divorced and widowed childless elderly persons
seemed to have higher odds of loneliness than parents, the
differences were small and the confidence intervals (avail-
able upon request) of the estimates were large. To better il-
lustrate the effects of childlessness among persons who
were not married, we compared whether psychological
The results of the family status model are presented in
Table 5. As expected, childlessness was not statistically as-
Table 2. Main Effects of Childlessness on Loneliness: Logistic
Regression Results for Persons Aged 70 and Older, AHEAD,
1993 (N ꢀ 6,517)
Table 4. Main Effects of Childlessness on Depression: Ordinary
Least Squares Regression Results for Persons Aged 70 and Older,
AHEAD, 1993 (N ꢀ 6,517)
Independent Variable Model 1 Model 2 Model 3 Model 4 Model 5
Independent Variable Model 1 Model 2 Model 3 Model 4
Model 5
Parents (reference)
Parents (reference)
Stepparents
Childless
Sex (male ꢀ 1)
Married (reference)
Divorced
Widowed
Never married
Age
0.744
0.791
0.820
0.809
1.126
1.110
0.855
1.172
1.126
Stepparents
Childless
Sex (male ꢀ 1)
Married (reference)
Divorced
Widowed
Never married
Age
Education
Physical health
Wealth (log)
Whites (reference)
Blacks
0.009
0.092
0.070
0.067
ꢂ0.491**
0.069
0.003
0.094
0.0002 ꢂ0.006
ꢂ0.247** ꢂ0.259**
0.021
1.299** 1.270** 1.125
0.598**
4.814** 4.923** 4.292**
5.376** 5.599** 5.116**
3.371** 3.462** 3.159**
0.856**
0.747**
0.495**
0.805**
0.653**
0.435*
0.575**
0.400**
0.277
0.018**
ꢂ0.053**
ꢂ0.827**
ꢂ0.151*
1.005
Education
0.933**
0.683**
0.747*
Physical health
Wealth (log)
Whites (reference)
Blacks
0.819*
1.035
0.760
ꢂ0.248**
ꢂ0.224
0.230
Others
Others
ROC
0.519
0.567
0.693
0.693
R2
0.0002
0.011
0.026
0.028
ꢂ2log-likelihood
6656.4
6594.6
6048.1
6046.2
5716.0
Notes: Unstandardized ordinary least squares regression coefficients are re-
ported here. Standard errors are available on request. AHEAD ꢀ Asset and
Health Dynamics Among the Oldest Old.
Notes: Odds ratios are reported here. Confidence intervals are available on
request. AHEAD ꢀ Asset and Health Dynamics Among the Oldest Old.
*p ꢁ .05; **p ꢁ .01.
*p ꢁ .05; **p ꢁ .01.

CHILDLESSNESS AND WELL-BEING
S317
Table 5. Conditional Effects of Childlessness on Loneliness and
Depression: Logistic and Ordinary Least Squares Regression
Results for Persons Aged 70 and Older, AHEAD,
1993 (N ꢀ 6,517)
idea that childless persons—even those who were not mar-
ried—were disadvantaged. Being divorced, widowed, or
never married did not place childless persons at greater risk
of loneliness and depression. Note also the similarity in the
results for the logistic regression and OLS models for de-
pression.
Depression
Loneliness,
Odds Ratio
(Logistic),
Odds Ratio
Independent Variable
Depression, b
Childlessness, Sex, and Marital Status
Married parents (reference)
Married stepparents
Married childless
Divorced parents
Divorced childless
Widowed parents
Widowed stepparents
Widowed childless
Never married parents
Never married childless
Sex (male ꢀ 1)
Age
Education
Physical health
Wealth (log)
Whites (reference)
Blacks
We have argued that unmarried men may be particularly
vulnerable to the effects of childlessness. Men in general
have smaller social support networks outside the family.
Unmarried men are thus at the greatest risk of a paucity of
social support from kith and kin, and because of the short-
fall in social support, we expected that unmarried men with-
out children would be at the greatest risk of social isolation
and distress.
The results shown in Table 7 provide statistical evidence
supporting this argument. Divorced and widowed men who
were childless had the highest odds of loneliness compared
with men in other family statuses and women in all family
statuses (Figure 2). Divorced men who were childless also
had the highest odds of depression (Figure 3). Among di-
vorced childless elderly persons, men were approximately
twice as likely to feel lonely compared with women, net of
the covariates included in the full model (Figure 2). Among
widowed childless elderly persons, men were 1.74 times
more likely than women to feel lonely, and never married
childless men were 1.44 times more likely to feel lonely
than never married childless women. All of these effects
were statistically significant. As expected, we found no evi-
dence of sex differences in the effects of childlessness
among persons who were married.
The pattern of results was similar for depression. Di-
vorced and widowed childless men were significantly more
likely to feel depressed than divorced and widowed child-
less women (Table 7). The logistic regression models show
that divorced childless men were about 2.07 times more
likely to be depressed than divorced childless women (Fig-
ure 3), and widowed childless men were almost 1.57 times
more likely to be depressed than widowed childless women.
(Although the differences were statistically significant, we
think some degree of caution is warranted in interpreting the
results given the small cell sizes for childless men who were
divorced [n ꢀ 17], widowed [n ꢀ 63], and never married [n ꢀ
57].) In contrast to the results for loneliness, however, we
did not find evidence that never married childless men were
at greater risk of depression than never married childless
women. Again, we observed that there were no statistically
significant sex differences in the effects of childlessness on
the risk of depression for married people. Results for the
OLS regression were largely in line with those of logistic
regression models, except that sex differences of childless-
ness became statistically insignificant among the divorced
persons.
0.761
1.198
0.895
0.817
0.028
ꢂ0.100
0.490**
0.910**
0.395**
0.257
0.411**
0.574
0.225
ꢂ0.256**
0.018**
ꢂ0.052**
ꢂ0.827**
ꢂ0.153*
4.239**
5.319**
5.110**
4.072**
6.152**
5.538**
3.372**
1.129
1.664**
2.112**
1.649**
1.292
1.629**
2.125
1.218
0.781**
1.014*
0.949**
0.466**
0.741**
1.005
0.934**
0.683**
0.747*
0.810*
1.012
0.864
0.602*
ꢂ0.250**
ꢂ0.237
0.231
Others
R2
ROC
ꢂ2log-likelihood
0.760
5713.3
0.765
6049.0
Note: AHEAD ꢀ Asset and Health Dynamics Among the Oldest Old.
*p ꢁ .05; **p ꢁ .01.
well-being differed according to childlessness within the
marital status categories of divorced, widowed, and never
married persons (Table 6). Regardless of the reason for not
being married, we had no statistical evidence that childless
elderly persons differed from parents in the chances of be-
ing lonely.
If psychological well-being is gauged in terms of depres-
sion, the results shown in Tables 5 and 6 do not support the
Table 6. Comparing Effects of Childlessness on Loneliness and
Depression Among the Unmarried Groups: Logistic and Ordinary
Least Squares Regression Results for Persons Aged 70 and Older,
AHEAD, 1993 (N ꢀ 6,517)
Depression
(Logistic),
Odds Ratio Depression, b
Loneliness,
Odds Ratio
Independent Variable
Divorced childless vs divorced
parents
Widowed stepparents vs
widowed parents
Widowed childless vs widowed
parents
Never married childless vs
never married parents
1.255
0.796
1.203
0.609
1.269
0.783
0.988
0.573
0.420
ꢂ0.138
0.016
ꢂ0.349
Looking at the overall pattern of gender differences in the
effects of childlessness on psychological well-being, our
findings indicate that men usually were advantaged com-
pared with women. Men were less likely to feel lonely and
depressed, compared with women. This pattern is consistent
with prior research. Men’s advantages, however, withered
Notes: Divorced parents, widowed parents, and never married parents were
each entered as the reference category sequentially into the full models shown in
Table 5. The results were the conditional effects of childlessness, controlling for
sex, age, education, physical health, wealth (log), and race. AHEAD ꢀ Asset
and Health Dynamics Among the Oldest Old.
*p ꢁ .05; **p ꢁ .01.

S318
ZHANG AND HAYWARD
Table 7. Conditional Effects of Childlessness on Loneliness and
Depression: Logistic and Ordinary Least Squares Regression
Results for Persons Aged 70 and Older, AHEAD,
1993 (N ꢀ 6,517)
Depression
Loneliness,
Odds Ratio
(Logistic),
Odds Ratio
Independent Variable
Depression, b
Married parents (reference)
Married stepparents
Married childless
Divorced parents
Divorced childless
Widowed parents
Widowed stepparents
Widowed childless
Never married parents
Never married childless
Sex ꢃ Married Stepparents
Sex ꢃ Married Childless
Sex ꢃ Divorced Parents
Sex ꢃ Divorced Childless
Sex ꢃ Widowed Parents
Sex ꢃ Widowed Stepparents
Sex ꢃ Widowed Childless
Sex ꢃ Never Married Parents
Sex ꢃ Never Married Childless
Sex (male ꢀ 1)
0.773
1.355
0.550
0.624
1.186
1.305
1.234*
1.063
1.139
1.478
0.952
ꢂ0.089
ꢂ0.172
0.156
0.577*
0.173*
0.223
0.186
0.023
ꢂ0.001
0.214
0.120
0.769**
1.013
0.676**
ꢂ0.119
0.632*
1.519
2.611**
3.059**
3.216**
3.001**
3.869**
3.711*
2.130**
1.106
2.326
1.696
0.679
2.737**
4.012*
3.353**
1.667
3.175**
2.217
2.627*
0.547**
1.006
0.933**
0.677**
0.731**
2.164*
4.109*
2.343**
1.410
3.116**
2.471
Figure 2. The joint effects of childlessness, marital status, and sex
on loneliness, controlling for age, education, health, household wealth,
and race for persons aged 70 and older in the Asset and Health Dy-
namics Among the Oldest Old, 1993 (N ꢀ 6,517). The reference group
was female married parents. The estimated odds ratio for any other
family status by sex was calculated with the estimated odds ratios in
Table 7. For example, the odds ratio of loneliness for the male child-
less widowed group was 3.869 ꢃ 0.547 ꢃ 3.175 ꢀ 6.719, controlling for
age, education, health, household wealth, and race. Asterisks indicate
significant sex differences in the effects of combination of parental
and marital status on loneliness. *p ꢁ .05; **p ꢁ .01.
1.665
0.510
0.504**
1.015*
0.950**
0.462**
0.732**
ꢂ0.531**
0.019**
ꢂ0.052**
ꢂ0.832**
ꢂ0.161*
Age
Education
Physical health
Wealth (log)
Whites (reference)
Blacks
Others
R2
0.796*
0.998
0.860
0.595*
ꢂ0.260**
ꢂ0.232
0.235
ROC
ꢂ2log-likelihood
0.767
5657.1
0.769
6014.0
mine the causal relationship between the two variables, this
pattern is consistent with the idea that living arrangements
were an outcome of psychological well-being in this sample.
In additional analyses, we considered a more stringent
criterion to define depression (i.e., we used 4 as the cutoff
point), and we re-ran the models described above. The di-
rection of the parameter estimates remained unchanged, and
the magnitude of effects changed very little. The only ex-
ception was among never married childless elderly persons:
men were significantly more likely to feel depressed than
women. This effect was not statistically significant when we
used 3 as the cutoff point.
Note: AHEAD ꢀ Asset and Health Dynamics Among the Oldest Old.
*p ꢁ .05; **p ꢁ .01.
in the absence of children and marriage. Being childless and
unmarried at old age deprived men of their relative advan-
tage in psychological well-being, reversing the traditional
relationship between sex and psychological well-being. The
results support the view that childless and unmarried older
men are a highly vulnerable group.
One factor not considered in our investigation of the ef-
fects of childlessness on psychological well-being was the
living arrangements of elderly persons. In the results shown
previously, we excluded living arrangements as a predictor
variable because of the problems of reverse causality. When
parents become depressed or sick, co-residence with or
moving closer to adult children often occurs so that care can
be provided (Siegel, 1993). In this sense, living arrange-
ments are frequently an outcome rather than a predictor of
psychological well-being. As a partial test of this idea, we
operationalized living arrangements as a series of indicator
variables: living with children, having children nearby but
without resident children, and having neither resident chil-
dren nor children nearby. Having resident children was
strongly associated with high rates of depression (results not
reported), net of other factors. Although we could not deter-
DISCUSSION
Prior research on the effects of childlessness on psycho-
logical well-being has largely focused on young, middle-
aged, and young-old persons. Given the dramatic increase
in childlessness and low rates of marriage in recent decades,
the aging of the American population in the 21st century
will result in greater numbers of future elderly persons who
are both childless and unmarried. Whether these conditions
foretell a growing vulnerable population at greater risk of
loneliness and depression has been the guiding question of
this investigation.
Drawing on a nationally representative sample of persons
aged 70 and older, we examined how marital status and sex
condition the effects of childlessness at older ages. Our re-
sults show that these factors significantly interact to influ-

CHILDLESSNESS AND WELL-BEING
S319
ple in terms of the presence of stepchildren. Our study
shows that a sizable percentage of biologically childless
men and women have stepchildren. Stepparents do not ap-
pear to be substantially different from biological parents in
many aspects, and stepparents’ psychological well-being is
akin to that of biological parents. This suggests that step-
children may be an important source of social support for
older people and that biological ties between parents and
children may be less important than familial ties.
Although the main effects of childlessness are minimal in
this study for older persons, it is premature to conclude that
childlessness has no direct effect on the psychological well-
being of older people. By virtue of the sample design, our
analysis excludes institutionalized older people. At ad-
vanced ages, childless persons are more likely to become in-
stitutionalized (Freedman, 1996; Koropeckyj-Cox, 1998;
Palmore, 1976; Siegel, 1993), and rates of depression are
much higher for persons living in institutions than those liv-
ing in communities. This problem may be compounded by
differential mortality selection. Kobrin and Hendershot (1977)
observed, for example, that age-adjusted mortality rates
are higher for nonparents than for parents. Because our
sample only includes aged persons who have survived to 70
and older, many childless persons may have already died
before 70 years old. Fortunately, the nature of these selec-
tion problems suggests that the results of this study are con-
servative tests.
Ultimately, these issues point to the importance of mov-
ing to a longitudinal design to understand the relationship
between childlessness and psychological well-being—par-
ticularly in the presence of mortality and institutionalization
selection processes. Such a design would also enable re-
searchers to examine the effects of childlessness on the pro-
cess of psychological well-being, that is, the onset of loneli-
ness and depression, the duration of these problems, and the
ultimate outcomes of these problems in terms of institution-
alization and death. Childlessness potentially comes into
play at different points in the process. Childlessness may
have few effects on the onset of psychological problems,
for example, but once persons experience those problems,
childlessness may exact a heavy toll in terms of greater risks
of institutionalization and death. We plan to take up these
issues in our future work.
Figure 3. The joint effects of childlessness, sex, and marital status
on depression, controlling for age, education, health, household
wealth, and race for persons aged 70 and older in the Asset and
Health Dynamics Among the Oldest Old, 1993 (N ꢀ 6,517). The ref-
erence group was female married parents. The estimated odds ratio
of depression for any other family status by sex was calculated with
the estimated odds ratios in Table 7. For example, the odds ratio of
depression for the male childless divorced group was 1.305 ꢃ 0.504 ꢃ
4.109 ꢀ 2.703, controlling for age, education, health, household
wealth, and race. Asterisks indicate significant sex differences in the
effects of combination of parental and marital status on depression.
*p ꢁ .05; **p ꢁ .01.
ence both loneliness and depression. Childless unmarried
men, in particular, have significantly higher rates of loneli-
ness and depression compared with men and women in
other family statuses.
Our findings expand current understanding of the effects
of childlessness on psychological well-being over the life
cycle. Prior research has found little support for the idea that
childlessness has negative consequences for the psychologi-
cal well-being of middle-aged and young-old persons. Con-
sistent with prior research findings, we find that the effects
of childlessness on loneliness and depression remain mini-
mal at advanced ages. Our research also provides little sta-
tistical evidence for the idea that the effects of childlessness
are conditional on marital status. Overall, marital status
rather than parental status is a more salient factor influenc-
ing loneliness and depression in old age.
Acknowledgments
Partial support for this study was provided by grants from the National
Institute on Aging (R01 AG 11758) and the National Institute of Child
Health and Human Development (5 P30 HD28263). We thank Alan Booth,
Valarie King, four anonymous reviewers, and the editor for their helpful
comments.
When we consider whether the interaction effects of mar-
ital status and childlessness on psychological well-being are
gendered, the effects of childlessness become more distinct.
The adverse consequences of childlessness are a character-
istic of unmarried men but not women. Among divorced,
widowed, and never married elderly persons, childless men
are much more likely to feel lonely than childless women.
Divorced and widowed childless men are also more likely to
feel depressed than divorced and widowed childless women.
We were able to take advantage of novel information in
the AHEAD data to differentiate biologically childless peo-
Address correspondence to Zhenmei Zhang, Department of Sociology,
Penn State University, University Park, PA 16802. E-mail: zhenmei@
pop.psu.edu
References
Alexander, B. B., Rubinstein, R. L., Goodman, M., & Luborsky, M.
(1992). A path not taken: A cultural analysis of regrets and childless-
ness in the lives of older women. The Gerontologist, 32, 618–626.
Andrews, F. M., Abbey, A., & Halman, L. J. (1991). Stress from infertility,
marriage factors, and subjective well-being of wives and husbands.
Journal of Health and Social Behavior, 32, 238–253.

S320
ZHANG AND HAYWARD
Bachrach, C. A. (1980). Childlessness and social isolation among the el-
derly. Journal of Marriage and the Family, 42, 627–637.
Beckman, L. J., & Houser, B. B. (1982). The consequences of childless-
ness on the social-psychological well-being of older women. Journal of
Gerontology, 37, 243–250.
Blake, J. (1974). Can we believe recent data on birth expectations in the
United States? Demography, 11, 25–44.
Brubaker, T. H. (1990). Families in later life: A burgeoning research area.
Journal of Marriage and the Family, 52, 960–981.
Callan, V. J. (1987). The personal and marital adjustment of mothers and
of voluntarily and involuntarily childless wives. Journal of Marriage
and the Family, 49, 847–856.
Cappeliez, P. (1993). Depression in elderly persons: Prevalence, predic-
tors, and psychological intervention. In P. Cappeliez & R. J. Flynn
(Eds.), Depression and the social environment (pp. 332–368). Montreal
& Kingston, Canada: McGill-Queen’s University Press.
P. R. Costanzo, & M. E. Oliveri (Eds.), The social psychology of men-
tal health: Basic mechanisms and applications (pp. 199–240). New
York: Guilford Press.
Kandel, D. B., Davies, M., & Raveis, V. H. (1985). The stressfulness of
daily social roles for women: Marital, occupational and household
roles. Journal of Health and Social Behavior, 26, 64–78.
Kivett, V. R., & Learner, R. M. (1980). Perspectives on the childless rural
elderly: A comparative analysis. The Gerontologist, 20, 708–716.
Kobrin, F. E., & Hendershot, G. E. (1977). Do family ties reduce mortal-
ity? Evidence from the United States, 1966–1968. Journal of Marriage
and the Family, 39, 737–745.
Koropeckyj-Cox, T. (1998). Loneliness and depression in middle and old
age: Are the childless more vulnerable? Journal of Gerontology: Social
Sciences, 53B, S302–S312.
Larson, J. (1992). Understanding stepfamilies. American Demographics,
14, 36–40.
Cappeliez, P., & Flynn, R. J. (1993). Implications for assessment, the un-
derstanding of social etiology, and intervention. In P. Cappeliez & R. J.
Flynn (Eds.), Depression and the social environment (pp. 395–405).
Montreal & Kingston, Canada: McGill-Queen’s University Press.
Chaisson-Stewart, G. M. (1985). Depression in the elderly: An interdisci-
plinary approach. New York: Wiley.
Chapman, N. J. (1989). Gender, marital status, and childlessness of older
persons and the availability of informal assistance. In M. D. Petersen &
D. L. White (Eds.), Health care of the elderly: An information source-
book (pp. 277–328). Newbury Park, CA: Sage.
“Childless Couples.” (1993). American Demographics, 15, 34–35.
Choi, N. G. (1994). Patterns and determinants of social service utilization:
Comparison of the childless elderly and elderly parents living with or
apart from their children. The Gerontologist, 34, 353–362.
Connidis, I. A., & McMullin, J. A. (1993). To have or have not: Parental
status and the subjective well-being of older men and women. The Ger-
ontologist, 33, 630–636.
Lin, N., Dean, A., & Ensel, W. M. (1986). Social support, life events, and
depression. New York: Academic Press.
McLanahan, S., & Adams, J. (1987). Parenthood and psychological well-
being. Annual Review of Sociology, 13, 237–257.
McMullin, J. A., & Marshall, V. W. (1996). Family, friends, stress and
well-being: Does childlessness make a difference? Canadian Journal
on Aging, 15, 355–373.
Palmore, E. (1976). Total chance of institutionalization among the aged.
The Gerontologist, 16, 504–507.
Rempel, J. (1985). Childless elderly: What are they missing? Journal of
Marriage and the Family, 47, 343–348.
Rubinstein, R. L. (1987). Childless elderly: theoretical perspectives and
practical concerns. Journal of Cross-Cultural Gerontology, 2, 1–14.
Sen, A. K., & Srivastava, M. S. (1990). Regression analysis: Theory, meth-
ods, and applications. New York: Springer-Verlag.
Siegel, J. S. (1993). A generation of change: A profile of America’s older
population. New York: Russell Sage.
De Jong, G. F., & Sell, R. R. (1977). Changes in childlessness in the
United States: A demographic path analysis. Population Studies, 31,
129–141.
Dugan, E., & Kivett, V. R. (1994). The importance of emotional and social
isolation to loneliness among very old rural adults. The Gerontologist,
34, 340–346.
Freedman, V. A. (1996). Family structure and the risk of nursing home ad-
mission. Journal of Gerontology: Social Sciences, 51B, S61–S69.
Glenn, N. D., & McLanahan, S. (1981). The effects of offspring on the
psychological well-being of older adults. Journal of Marriage and the
Family, 43, 409–421.
Hammen, C. (1997). Depression. Hove, England: Psychology Press.
Hays, J. C., Landerman, L. R., George, L. K., Flint, E. P., Koenig, H. G.,
Land, K. C., & Blazer, D. G. (1998). Social correlates of the dimen-
sions of depression in the elderly. Journal of Gerontology: Psycho-
logical Sciences, 53B, P31–P39.
Soldo, B. J., Hurd, M. D., Rodgers, W. L., & Wallace, R. B. (1997). Asset
and health dynamics among the oldest old: An overview of the AHEAD
study. The Journals of Gerontology: Psychological Sciences and So-
cial Sciences, 52B (Special Issue), 1–20.
Veroff, J., Kulka, R. A., & Douvan, E. A. M. (1981). Mental health in
America: Patterns of help-seeking from 1957 to 1976. New York: Ba-
sic Books.
Wachter, K. W. (1997). Kinship resources for the elderly. Philosophical
Transactions: Biological Sciences, 352, 1811–1817.
Wenger, G. C., Davies, R., Shahtahmasebi, S., & Scott, A. (1996). Social
isolation and loneliness in old age: Review and model refinement. Ag-
ing and Society, 16, 333–358.
Wu, Z., & Pollard, M. S. (1998). Social support among unmarried childless
elderly persons. Journal of Gerontology: Social Sciences, 53B, S324–
S335.
Himes, C. L. (1992). Future caregivers: Projected family structures of older
persons. Journal of Gerontology: Social Sciences, 47, S17–S26.
Jones, J. M. (1992). Understanding the mental health consequences of race:
Contributions of basic social psychological processes. In D. N. Ruble,
Received September 14, 2000
Accepted April 24, 2001