JOURNAL OF

AFFECTIVE
ELSEVIER Journal of Affective Disorders 35 (1995) 91-96
DISORDERS

Research report

Parental divorce, self-esteem and depression: an intimate

relationship as a protective factor in young adulthood
U.K. Palosaari *, H.M. Aro
National Public Health Institute, Department of Mental Health, Tampere. Finland
University of Tampere, School of Public Health, Tampere, Finland

Received 4 July 1994; revised 10 April 1995; accepted 10 April 1995

Abstract

This study investigated whether an intimate relationship in young adulthood protects young people from depression in the
presence of risk factors (parental divorce, low self-esteem in adolescence). The data were drawn from a follow-up survey of
adolescents from the age of 16-22 (n = 1656). The prevalence of depression was highest among persons from divorced
families who had reported low self-esteem at the age of 16 and who in young adulthood lacked an intimate relationship. An
intimate relationship was found to protect young adults with the risk factor of earlier low self-esteem from depression
irrespective of family background. The pattern was similar in both sexes.

1. Introduction (1989) reported that girls with a background of

Divorce requires major adjustments from all fam-
ily members. After the initial stress of parental di-
vorce children may have emotional distress, behav-
ioral problems or persistent decrease in their school
performance for several years (Guidupaldi and Perry,
1985; Hetherington et al., 1985; Featherstone et al.,
1992; Goodyer, 1993). It has been suggested that
divorce-related difficulties among girls become visi-
ble especially in adolescence (Kalter et al., 1985;
Wallerstein and Corbin, 1989). Certain problems
stemming from parental divorce may be mediated
via a lowered self-esteem (Parish and Wigle, 1985;
Palosaari et al., in press). Wallerstein and Corbin
* Corresponding author. Address: University of Tampere,
School of Public Health, Box 607, 333101 Tampere 10, Finland.
parental divorce had difficulties in intimate relation-
ships and that some feared being abandoned like
their mothers. The experience of earlier parental
divorce has been reported to be associated with
vulnerability to depression in young adulthood in
both genders (Wallerstein and Blakeslee, 1989; Aro
and Palosaari, 1992).
There is wide individual diversity in responses to
a stressful environment in childhood. In spite of a
risk of a maladaptive life course, many individuals
cope successfully with developmental challenges.
The impact of stressful childhood experiences on
further development is modified by factors or pro-
cesses in the child, family and community, and their
interactions which then affect further development
and life course either protecting from or predisposing
to stress or disorder at another developmental stage.
Subsequent adverse or beneficial circumstances may

016%0327/95/$09.50 0 1995 Elsevier Science B.V. All rights reserved

SSDIOl65-0327(95)00037-2
92 U.K. Pnlosaari, H.M. Are/Journal of Affectiue Disorders 35 (1995) 91-96

intensify or mitigate the influence of early stressful
experiences (Rutter, 1988; Brown et al., 1990). A
protective factor has been defined as a factor which
mitigates or buffers the effects of psychiatric risk
(Rutter, 1985).
The processes of negotiation of life transitions,
both in adolescence and in young adulthood, are of
special interest, because at this point in particular a
persons life course may be redirected into a more
adaptive path (Rutter, 1987; Maugham and Cham-
pion, 1988). Rutter (1987), e.g., has shown that
marriage with a supportive spouse may break the
negative chains of pathological development. Diffi-
culties in commitment and lack of trust in an inti-
mate relationship may be one link mediating nega-
tive long-term impact of parental divorce. On the
other hand, a good relationship may operate as a
protecting factor in the life course of the offspring
from divorced families. The presence of a confiding
relationship, particularly with husband or boyfriend,
reduced appreciably the risk of depressive disorder
among adult women who had experienced parental
loss in childhood (Brown et al., 1986b). One mecha-
nism by which the buffering effect of social support
has been suggested to function is by increasing the
individuals self-esteem and coping skills under stress
(OConnor and Brown, 1984, Rutter, 1987; Brown et
al., 1990).
This paper is one of a series of papers based on
the 6-year follow-up of adolescents. In the present
study (Aro and Palosaari, 1992), young adults with a
history of parental divorce were significantly more
often depressive compared with young people from
nondivorced families. The girls excess risk of de-
pression due to parental divorce was partly mediated
via lack of good self-esteem at the age of 16. Fur-
ther, low self-esteem indicated increased vulnerabil-
ity to subsequent depression among boys and girls,
in general, irrespective of family background (Palo-
saari et al., in press).
In the present paper, we asked whether an inti-
mate relationship in young adulthood protects the
individual from depression in the presence of earlier
risk factors (parental divorce and low self-esteem in
adolescence).
2. Methods
2.1. Study population
The study population included all 9th-grade pupils
attending secondary schools in the spring of 1983 in
Tampere, a Finnish city of 166,000 inhabitants (Aro,
1988). There were altogether 2269 pupils, 97.6% of
whom participated in the study and completed ques-
tionnaires in their classrooms. The mean age of the
pupils was 15.9 years (SD = 0.3 years). The same
young people were re-examined by means of postal
questionnaires in spring 1989 at the age of 22. The
participation rate in the follow-up was 77.4%. No
significant difference was found in parental divorce
between participants and nonparticipants. Before the
first phase of the study, 24% of pupils had experi-
enced parental divorce (girls, n = 210; boys, n =
146). These children were included in the study

Table 1
Adolescents self-esteem at age of 16, proportion of those having an intimate relationship at age of 22 and quality of relationship by family
background and gender (P values based on x2 statistic)
Females Males
Nondivorced Divorced Nondivorced Divorced

(h1 % (n) % P (n) % (n) % p

Self-esteem at age of 16 High (453) 54.3 (881 40.4 ** * (415) 71.8 (117) 75.0 NS
Low (297) 45.7 (130) 59.6 (163) 28.2 (39) 25.0
Intimate relationship at age of 22 Yes (459) 69.2 (155) 69.8 NS (344) 57.9 (94) 57.7 NS
No (204) 30.8 (67) 30.2 (250) 42.1 (69) 42.3
Sum score of quality of relationship (t test)
mean (SD) 10.1 (4.0) 10.6 (4.1) NS 10.0 (3.9) 10.0 (3.6) NS

*** P<O.OOl.
 U.K. Palosaari, H.M. Are/Journal of Affective Disorders 35 (1995) 91-96 93

group and the children living in two-parent families tory. S-BDI score 5 was used as a cut-off point
(girls, n = 675; boys, n = 611) were in the control (Beck et al., 1974; Mattlar et al., 1987). Self-esteem
group. at the age of 16 was measured on a Finnish scale
developed for students. This included 7 assertions
with a 5-point scale resembling those in Rosenbergs
2.2. Variables
measure (Rosenberg, 1965; Aro, 1988). The asser-
tions were: I believe in myself and my prospects; I
Depression was screened by a Finnish modifica- wish I were different from what I am; I suffer from
tion of the short 13-item Becks Depression Inven- feelings of inferiority; I think I have many good
qualities; I feel I lack self-confidence; I am capable
of doing the same as others; I am often dissatisfied
with myself. The theoretical range of the scale was
Females 7-35. The internal reliability of the score was 0.79
among girls and 0.77 among boys measured by
Depresslon % Cronbachs alpha. In the analyses, the scale was
50 dichotomized at about the mean (girls, mean = 17;
Self-esteem
boys, mean = 15). In intimate relationship we in-
0 Hlgh Low
cluded marriage, cohabitation and steady dating. The
quality of intimate relationship was assessed with 6
40 questions concerning closeness and conflicts of the
relationship scored on a 5-point scale: Our relation-
ship is warm, We are close to each other, Our
relationship includes mutual trust, We have many
3a problems, Our relationship is cooling off, We
often quarrel. The sum of the scores was di-
chotomized in the mean (x = 9, SD = 3.4).
The preliminary statistical description was based
2a on cross-tabulation and on logistic regression mod-
elling. Furthermore, we were especially interested in
the multiplicative effects of predictors, i.e., in the
interaction structure of risk and protective factors.
Accordingly, the data were analysed using loglinear
modelling.

Yes No
3. Results
Intimate raiationship Intimate relationship
Non-divorced family Divorced family
Low self-esteem at the age of 16 was more com-
mon among girls from divorced families. Boys from
divorced and nondivorced families did not differ
Relstlonrhlp x Self-esteem x Oeprosslon p = 0.06 from each other. At the age of 22, no difference in
Parental divorce x Dspra8slon pI 0.05
Parental divorce x Self-esteem p = 0.002 the existence of an intimate relationship was found
Ooodner~ of fit of the model p I 0.77 between young people from divorced and nondi-
vorced homes. Further, family background did not
Fig. 1. Prevalences of depression among young females from
nondivorced and divorced families by self-esteem and by presence predict any difference in reported quality of present
of an intimate relationship. relationships (Table 1).
94 U.K. Palosaari, H.M. Are/ Journal of Affective Disorders 35 (1995) 91-96

Males The prevalence of depression was highest among

persons from divorced families who had reported
Depresslon % low self-esteem at the age of 16 and who in young
50
Self-esteem adulthood lacked an intimate relationship. An inti-
mate relationship was found to protect young adults
0 High Low
with the risk factor of earlier low self-esteem from
i depression irrespective of family background. The
association between parental divorce and depression
remained in the model. The pattern was similar
among both sexes. Among girls, there was also an
indirect association between parental divorce and
depression, with girls from divorced families report-
ing more frequently low self-esteem in adolescence.
The results also suggested that good self-esteem
in adolescence reflects resilience against depression
in both genders. The prevalence of depression re-
mained low in all subgroups with good self-esteem.
When the quality of relationship was taken into
account, the association between parental divorce
and depression in the offspring remained significant.
If the relationship was good the prevalence of de-
pression further decreased irrespective of family
background (females, poor relationship 16.7% vs.
good relationship 4.5%, x2 = 25.7, df = 1, P <
Yes NO Yes NO
0.001; males, poor relationship 9.2% vs. good rela-
Intimate relationship intimate relationship
tionship 2.6%, x2 = 8.83, df= 1, P < 0.01). How-
Non-divorced family Divorced family
ever, the proportion of depressive young people
among those who had no intimate relationship
Loglinew model (females, 20%; males, 13%) was bigger than that in
either of the former groups.
p = 0.04
p = 0.02

4. Discussion
Fig. 2. Prevalences of depression among young males from nondi-
vorced and divorced families by self-esteem and by presence of an
intimate relationship. Research showing that social support, including
an intimate relationship, functions as a buffer and
protective factor against depression has been a focus
of methodological consideration, largely due to the
3.1. Model building intertwinedness of risk, protective and outcome-re-
lated factors in the life course (Brown et al., 1986a;
At the age of 16, one of the predictors for subse- Alloway and Bebbington, 1987; Aro et al., 1989).
quent depression was low self-esteem (females, RR Within the methodological complexity, our results
= 2.3, CI 95% 1.6-3.3; males, RR = 2.15, CI 95% also lend support to this hypothesis.
1.2-3.4). In further analyses, parental divorce and Commitment to a partnership is among the devel-
low self-esteem were considered as risk factors for opmental tasks in young adulthood and managing to
depression. We studied whether an intimate relation- form an intimate relationship probably increases
ship protected young adults from depression in the trusting and hopeful attitudes towards the future. A
presence of these risk factors (Fig. 1, Fig. 2). close relationship may compensate for the effects of
 U.K. Palosaari, H.M. Aro / Journal of Affective Disorders 35 (I 995) 91-96
earlier low self-esteem on choices in young adult-
hood. Further, steady dating or an enduring relation-
ship are regarded as a normative part of the transi-
tion to adulthood. The lack of partnership at this age
may be stressful as such. On the other hand, the
results suggested that good self-esteem in adoles-
cence reflects resilience against depression. Those
with good self-esteem in adolescence had low levels
of depression in spite of the lack of an intimate
relationship. An alternative interpretation of our data
is that a low self-esteem in adolescence and lack of
an intimate relationship in young adulthood may
both be indicators of a risk trajectory in develop-
ment. Those who were depressive at the age of 22
may have been depressive already at the age of 16.
Kandel and Davies (1986) suggest that long-term
effects of adolescent depression manifest themselves
in reduced ability in forming an intimate relation-
ship. In the follow-up study, they showed the conti-
nuity of depressive symptoms over time and the
presence of associated long-term difficulties in rela-
tionships, viz., a lack of closeness in relationship
with parents in adolescence and lack of closeness
with their spouses in young adulthood. However, the
protective effect of an intimate relationship on the
proneness to depression among those with low self-
esteem shows that the developmental phase of young
adulthood may be a turning point in such a trajec-
tory.
The protective effect of an intimate relationship
on young adults from divorced families was similar
to the effect on those from nondivorced families.
However, the extra risk due to parental divorce
existed in the presence of an intimate relationship.
Males and females probably have different reasons
for this. Females with a history of parental divorce
were more likely to get divorced themselves and to
report problems and conflicts with spouse than fe-
males from nondivorced families (Aro and Palosaari,
1992) and the discord may bring to their mind
threatening memories of parental divorce (Waller-
stein and Corbin, 1989). Among males from di-
vorced families, failures in job activities may indi-
cate vulnerability to depression. In addition, parental
divorce among both genders may indicate additional
stressful life courses, which are not compensated for
by intimate relationships.
(1) Epidemiologic and population-based methods
95
using crude measures for investigating complex in-
terplay of environmental and developmental pro-
cesses in psychopathology have many limitations. (2)
A depressive person may interpret the quality of the
relationship distorted. The reliability of information
was aimed to improve by defining the intimate rela-
tionship to concern relatively long conditions (mar-
riage, cohabitation or steady dating). (3) Parental
divorce is not a specific risk indicator of depression
due to the great variation in conditions of both
divorced and nondivorced families. This may attenu-
ate the strength of associations. However, the follow-
up design and the focus of the present study on the
key stages of the childs development connect single
variables to the larger context of a persons life
course. The associations of psychosocial stress in
childhood (parental divorce) and developmentally
important factors (self-esteem in adolescence and
commitment to partnership in young adulthood) in
the models contribute to the knowledge of the indi-
viduals susceptibility to depression. In addition, a
relatively large and representative nonclinical sample
gives an opportunity to draw generalizable conclu-
sions.
In accordance with earlier research, our findings
suggest that an intimate relationship in young adult-
hood can act as a buffer against the risk of low
self-esteem, while the lack of it may predispose to
depression in young adulthood.
References
Alloway, R. and Bebbington, P. (1987) The buffer theory of social
support - a review of the literature. Psychol. Med. 17,9 I- 108.
Aro, H. (1988) Parental discord, divorce and adolescent develop-
ment. Eur. Arch. Psychiatry Neurol. Sci. 237, 106-l 11.
Aro, H. and Palosaari, U. (1992) Parental divorce, adolescence
and transition to young adulthood. A follow-up study from the
age of 16 to 22. Am. J. Orthopsychiatry 62, 421-429.
Aro, H., Hanninen, V. and Paronen, 0. (1989) Social support, life
events and psychosomatic symptoms in adolescence. Sot. Sci.
Med. 29, 1051-1056.
Beck, A.T., Ryal, W.Y. and Rickels, K. (1974) Short form of
depression inventory: cross-validation. Psychol. Rep. 34,
1184-1186.
Brown, G.W., Andrews, B., Harris, T., Adler, Z. and Bridge, L.
(1986a) Social support, self-esteem and depression. Psychol.
Med. 16, 813-831.
Brown, CL, Harris, T. and Bifulco, A. (1986b) Long-term effects
of early loss of parent. In: M. Rutter et al. (Eds.), Depression
96 U.K. Palosaari. H.M. Aro / Journal of Afhectiue Disorders 35 (1995191-96
in Young People. Developmental and Clinical Perspectives.
Guilford Press, New York, NY, pp. 3-30.
Brown, G.W., Bifulco, A., Veiel, H.O. and Andrews, B. (1990)
Self-esteem and depression, II. Social correlates of self-esteem.
Sot. Psychiatry Psychiatr. Epidemiol. 25, 225-234.
Goodyer, I.M. (1993) Recent stressful life events: their long term
effects. Eur. Child Adolesc. Psychiatry 2, 1-9.
Guidupaldi, J. and Perry, J. (1985) Divorce and mental health
sequelae for children: a two-year follow-up of a nationwide
sample. J. Am. Acad. Child Psychiatry 24, 531-537.
Featherstone, D.R., Cundick, B.P. and Jensen, L.C. (1992) Differ-
ences in school behaviour and achievement between children
from intact, reconstituted, and single-parent families. Adoles-
cence 27, 1-12.
Hetherington, E.M., Cox, M. and Cox, R. (1985) Long-term
effects of divorce and remarriage on the adjustment of chil-
dren. J. Am. Acad. Child Psychiatry 24, 518-530.
Kalter, N., Riemer, B., Brickman, A. and Chen, J.W. (1985)
Implications of parental divorce for female development. J.
Am. Acad. Child Psychiatry 24, 538-544.
Kandel, D.B. and Davies, M. (1986) Adult sequelae of adolescent
depressive symptoms. Arch. Gen. Psychiatry 43, 255-262.
Mattlar, C.-E., Raitasalo, R., Putkonen, A.-R., Hyyppa, M.T.,
Englund, C., Helenius, H. and Knuts, L.-R. (1987) The preva-
lence of depression in a random sample of Finns, and the
association of depression with various cognitive functions.
Abstract of the Eleventh International Meeting of the Epidemi-
ological Association (No. 629). International Epidemiological
Association, Helsinki, Finland.
Maugham, B. and Champion, L. (19881 Risks and protective
factors in the transition to young adulthood. Paper at European
Science Foundation Workshop; Longitudinal Research and the
Study of Successful Aging. Castle Ringberg, Tegemsee, June
1988.
OConnor, P. and Brown, G.W. (1984) Supportive relationships:
fact or fancy? J. Sot. Pers. Relationships 1, 159-175.
Palosaari, U., Are, H. and Laippala, P. (in press) Adolescents
self-esteem and closeness with parents; vulnerability factors to
later depression. Acta Psychiatr. Stand.
Parish, T.S. and Wigle, SE. (1985) A longitudinal study of the
impact of parental divorce on adolescents evaluation of self
and parents. Adolescence 77, 239-244.
Rosenberg, M. (1965) Society and the Adolescent Self-Image.
Princeton University Press, Princeton, NJ.
Rutter, M. (1985) Resilience in the face of adversity. Protective
factors and resistance to psychiatric disorder. Br. J. Psychiatry
147, 598-611.
Rutter, M. (1987) Psychosocial resilience and protective mecha-
nisms. Am. J. Orthopsychiatry 57, 316-331.
Rutter, M. (1988) Functions and consequences of relationships:
some psychopathological considerations. In: R.A. Hinde and J.
Stevenson-Hinde (Eds.1, Relationships within Families. Ox-
ford University Press, New York, NY.
Wallerstein, J. and Blakeslee, S. (1989) Second Chances: Men,
Women and Children a Decade after Divorce. Who Wins and
Who Loses - and Why. Ticnor and Fields, New York, NY.
Wallerstein, J.S. and Corbin, S.B. (1989) Daughters of divorce:
report from a ten-year follow-up. Am. J. Otthopsychiatry 59,
593-604.