You are on page 1of 19

Parental Marital Quality

and Family Conflict:


Longitudinal Effects on Adolescents
from Divorcing
and Non-Divorcing Families
Ailsa Burns
Rosemary Dunlop

ABSTRACT. Recent studies by Amato and Booth (1997) and Wallerstein,


Blakeslee and Lewis (2000) have reported that the children of highly-con-
flicted parents who remained together, and the children of low-conflict parents
who separated, were as adults more poorly adjusted than others, and less
successful in themselves forming and maintaining intimate relationships.
These associations were explored in a longitudinal study of 67 adolescents
aged 13-16 when first interviewed, at which time half the families involved
were at the point of divorce. Mothers, fathers and adolescents initially rated
the level of conflict in the family, and parents also rated their marital ad-
justment and their satisfaction with marital conflict resolution (in the case
of the divorce group, at the time just before the decision to separate). The
adolescents also reported whether they got involved in their parent’s dis-
agreements, and completed standardised measures of self-image, anxiety
and depression. Ten years later the now-adult children repeated these
items, along with measures of readiness for intimacy and wariness
about relationships, and some further family conflict items. Family con-
flict at Time 1 (as perceived by the child) predicted self-image and anxi-
ety at Year 1, but not at Year 10. However, the adult children’s current

Ailsa Burns, PhD, and Rosemary Dunlop, PhD, are Honorary Associates, Department
of Psychology, Macquarie University, New South Wales.
Address correspondence to: Ailsa Burns, Department of Psychology, Macquarie Uni-
versity, New South Wales, Australia, 2109 (E-mail: aburns@bunyip.bhs.mq.edu.au).
Journal of Divorce & Remarriage, Vol. 37(1/2) 2002
http://www.haworthpressinc.com/store/product.asp?sku=J087
Ó 2002 by The Haworth Press, Inc. All rights reserved. 57
58 JOURNAL OF DIVORCE & REMARRIAGE

rating of happiness in the family of origin predicted current self-image. As


adults, daughters were more anxious than sons, and daughters who had pre-
viously rated their families as highly conflicted were more depressed than
other sample members. Parent’s marital status was relevant only to wari-
ness about relationships. Children from the divorce group were more wary
overall. Their level of wariness was unaffected by their Year 1 reports of
involvement in parental disagreements, but within the intact group, greater
wariness was associated with greater involvement in parental conflict at
Year 1. These findings indicate a different pattern of long-term outcomes
from those reported by Amato & Booth (1997) and Wallerstein,
Blakeslee and Lewis (2000). [Article copies available for a fee from The
Haworth Document Delivery Service: 1-800-HAWORTH. E-mail address:
<getinfo@haworthpressinc.com> Website: <http://www.HaworthPress.com> © 2002 by
The Haworth Press, Inc. All rights reserved.]

KEYWORDS. Children of divorce, family conflict, parent marriage qual-


ity, adult adjustment, adult relationships

Debate over the long term effects of divorce on children has been re-
newed by recent publications from the longitudinal studies of
Wallerstein (Wallerstein & Kelly, 1980: Wallerstein & Blakeslee, 1989;
Wallerstein, Lewis & Blakeslee, 2000) and Amato and Booth (1997).
Wallerstein reports that 25 years after the parental divorce, the
now-adult children still saw the divorce as the formative event in their
lives, and suffered from poor self-image, chronic anxiety, depression
and a range of other psychological problems. In particular, they found it
difficult to form satisfactory intimate relationships, and their marriage
rate by their 30’s and 40’s was significantly lower than that in the gen-
eral population. Over half those in the study had decided not to have chil-
dren themselves because they believed they knew too little about good
parenting. A particular cause of distress was the fact that their chances of
higher education had been jeopardised by lack of paternal (usually) sup-
port. Wallerstein’s conclusion is that even unhappily married parents
should consider staying together for the sake of the children, except in
cases of extreme conflict, domestic violence, substance abuse and se-
vere pathologies. Her views received major publicity, including the
cover story of Time magazine (6.12.00).
Ailsa Burns and Rosemary Dunlop 59

Wallerstein’s study is famous for the detail of its observation of the


children and the very long follow-up period, and has exerted considerable
influence on policy-makers. However it has been consistently criticised
for its methodological shortcomings, including the subjectivity of the in-
terview procedures used, the smallness of the sample, the self-selection of
troubled families into the sample, and the lack of a control group. These
criticisms cannot be levelled against Amato and Booth’s (1997) longitudi-
nal study of divorced and intact families.Their sample was much larger
(471 children, 89 of them with divorced parents), reasonably representa-
tive of the US population, the pre-divorce socioeconomic status of the di-
vorced and intact families was reasonably similar, and standardised and
structured measures of adjustment and well-being were used. The
time-span of the study was not so great as that of Wallerstein, but still sub-
stantial: parents were first interviewed in 1980 and their 19-40 year old
children twelve years later, in 1992.
How do the findings compare? Amato and Booth (1997) found that the
difference between children from divorced and intact families was very
modest, and partly due to the level of marital conflict that preceded it.
Most of the variance observed between children was due to the parent’s
socioeconomic resources, parental education in particular. However the
research did find an interaction between divorce and the level of pre-di-
vorce parental conflict. Where pre-divorce conflict between parents was
high, children were better off following the divorce, a finding also re-
ported by Jekielek (1998). However when pre-divorce conflict was low,
the children were as adults less happy and had lower self-esteem and psy-
chological well-being. The authors conclude that “with marital dissolu-
tion becoming increasingly socially acceptable, it is likely that people are
leaving marriages at lower thresholds of unhappiness now than in the past.
Unfortunately, these are the very divorces that are most likely to be stress-
ful to children. Consequently, we conclude that the rise in marital disrup-
tion, although beneficial to some children, has in balance, been detrimental
to children” (p. 220). Amato & Booth are thus to some extent in agreement
with Wallerstein’s view that divorce should be avoided except in more ex-
treme situations.
The relative merit of divorce or staying together in cases of marital un-
happiness and conflict is of course a critical issue for parents in unsuccess-
ful marriages, making Amato and Booth’s careful treatment very
welcome. However, it is not a simple matter to nominate the level of con-
flict and/or unhappiness at which divorce is likely to be beneficial or detri-
mental to children. It matters whose perception of marital quality we
accept, as we know that different family members do not always see things
60 JOURNAL OF DIVORCE & REMARRIAGE

the same way (Carter & McGoldrick, 1980). Children, for example, may
believe that their parents rarely disagree if all disagreements occur behind
closed doors. In consequence, they can see the divorce as coming “out of
the blue,” when in fact there has been a long and painful buildup. Dis-
agreements can also be voiced or unvoiced, so that an extreme level of
marital unhappiness–of one or both partners–can occur without overt con-
flict. Overt disagreements can also be more or less successfully resolved,
to the greater or less satisfaction of one or both parents. There is also evi-
dence that as far as children are concerned, it is not so much the degree of
conflict as whether or not they become personally involved that is most
likely to cause distress (Buchanan, Maccoby & Dornbusch, 1991). A
greater understanding of these factors is necessary to clarify the distinc-
tion that Wallerstein et al. (2000) and Amato and Booth (1997) make be-
tween beneficial and detrimental divorce.
Our own longitudinal study can make a contribution here, as we as-
sessed parents and their early adolescent children’s perceptions of family
conflict and parent’s assessment of their own marital quality in intact and
divorcing families at the time of the divorce for the latter group. We also
measured the children’s emotional adjustment and self-concept at this
time, and three and ten years later, when they were aged in their mid-twen-
ties. On this occasion we also took measures of their attitudes to close rela-
tionships, and their satisfaction with their own intimate relationships.
These measures are of particular relevance, as this is the area in which
Wallerstein (Wallerstein & Blakeslee, 1989; Wallerstein et al., 2000)
found that the adult children of divorce had particular difficulties. We are
thus in a position to consider the long-term effects of parent’s marital
quality (as reported by different family members) on children whose par-
ents did or did not divorce.
Specifically, the questions that this paper addresses are

1. Do family members agree in their assessment of family conflict and


parental marital quality?
2. Do assessments of family conflict and parental marital quality pre-
dict adjustment in early adolescence?
3. Do assessments of family conflict and parental marital quality pre-
dict the children’s adjustment and the quality of their own intimate
relationships in early adulthood?
4. Do family conflict and parental marital quality have different effects
on children from divorced and intact homes?
Ailsa Burns and Rosemary Dunlop 61

METHOD

Sample and Procedures

The data come from a study comparing the development of adolescents


whose parents divorced with a comparison group from intact families.
The divorcing families were contacted by letters sent in 1981 by the Fam-
ily Court of Australia (NSW) to all couples who had filed for divorce over
a three month period in the two registries that serve the city of Sydney,
who had been separated for less than two years, and had an adolescent
aged 13-16. Couples were invited to take part if they and their adolescent
were prepared to be interviewed. Nineteen boys and 18 girls were re-
cruited, all of whose families were close to divorce at the time of the inter-
view. Intact families (19 girls and 22 boys) were contacted by letters sent
by class teachers to parents of 13-16 year old students in five high schools
chosen to represent a matched spread of socioeconomic background. The
mean age of the total group was 14.7 years.
At time 1, interviews were conducted with all 41 married mothers and
39 married fathers, 26 divorcing mothers and 11 divorcing fathers who
had custody, and 23 of the non-custodial parents. Married parents re-
sponded to some parts of the interview jointly, but completed the standard
measures separately. Divorcing parents completed the full interview sepa-
rately. On a separate occasion, the adolescent was interviewed by a second
researcher.
In both groups, 46 percent of father’s occupations were in the higher (A
and B) status groups, and 54 percent in the lower (C and D) (Congalton,
1969). A series of analyses found the divorced sample and the intact fam-
ily comparison group to be equivalent on major background variables
such as age of children and parents, parent’s country of birth, length of
marriage and general health, and child’s ordinal position and number of
siblings.
The adolescents were re-contacted in 1985, when aged 16-19, and
again in 1991-92, when aged 23-26, at which time 67 were re-interviewed
(19 sons and 17 daughters from the intact group, and 16 sons and 15
daughters from the divorce group: 84%). Of these, the majority (40) were
single, 17 were married, 1 was divorced and 9 were cohabiting. A majority
of the single (27) were currently in a committed relationship. Preliminary
analyses showed that the smaller sample interviewed in 1991-92 did not
differ significantly from the original sample on any of the standard outcome
measures. However in both divorced and non-divorced families there was
62 JOURNAL OF DIVORCE & REMARRIAGE

a significant drop in retention for those of lower SES (c2 (1) = 9.26, p <
.005).

Measures

Parent Marital Quality Variables

Parents completed the following measures separately:


1. The Spanier Dyadic Adjustment Scale (Spanier, 1976), a widely used
scale that examines satisfaction with current relationships. It has 32 items
providing a total score, and subscores tapping satisfaction, cohesion, con-
sensus and affection. Most items are presented on a 6-point scale (from
“all the time” to “never”). A higher score denotes better dyadic adjust-
ment. When married parents were interviewed, one retired to a separate
room to complete this and other standard measures. Divorced parents
were asked to complete the scale (separately) for the time just before the
decision to separate was made.
2. The Parent Disagreement Scale was created by the authors as a mea-
sure of effectiveness in resolving disagreements with a partner. Ten items
were derived from family therapy sources describing the kinds of dis-
agreements that may occur in marriage or marriage-like relationships.
Items are coded so that a high score indicates less satisfactory conflict res-
olution. Items are: A row seems to clear the air and we feel closer to each
other afterwards*; when we disagree we never seem to solve the problem,
it keeps coming back; we keep disagreements to ourselves and don’t talk
about them to each other; a problem usually disappears because the same
person usually gets his/her way; we each give in a bit in order to reach a
compromise*; someone usually shouts and gets angry; we don’t have dis-
agreements; after a disagreement there is still a lot of tension in the air;
children take sides with one or other parent; we talk problems over quietly
and try to find the best solution* (* notes reverse-scoring). Cronbach’s al-
pha for the parent sample was .64.
3. Parents were also asked the multiple response question: Most cou-
ples have disagreements sometimes. Would you say that you and your
partner quarrel more or less than the average couple (much less, a little
less, a little more, much more)?
Adolescents were asked at Year 1: (1) Most families have quarrels
sometime. Do the members of your family fight much? Would you say
there is (not much, a medium amount, a lot) of fighting in your family? (2)
Do you sometimes feel yourself getting involved in your parent’s dis-
Ailsa Burns and Rosemary Dunlop 63

agreements? (no, yes). (These questions were asked about “family con-
flict” as it was considered unethical to ask adolescents to rate “parent
conflict”). At Year 10, adult children were (3) Do members of your family
of origin fight much nowadays (no/not much, a medium amount, a lot)?
(4) Do you ever feel yourself getting involved in your parent’s disagree-
ments (no, yes)? (5) How happy do you think your family of origin was at
the time of the first interview? (not too happy, pretty happy, very happy)?
(6) Taking all things together, how happy would you say your family of
origin is at present (not too happy, pretty happy, very happy)?

Child Adjustment Variables

1. The Offer Self-Image Questionnaire for Adolescents (OSIQ: Offer,


Ostrov & Howard, 1977), is a self-report questionnaire that explores 11
areas of development including psychological, social, sexual, familial and
coping aspects of the self. Permission was granted by the authors of the
OSIQ in 1991 to make slight alterations to 10 of the 130 questions so that
wording was appropriate for adult respondents. A low score denotes a
more positive self-image. Cronbach’s alpha for the total scale was .92 at
Time 2 and .93 at Time 2. Total scores are used in the present analyses.
2. Depression and Anxiety Scales of the Neuroticism Scale Question-
naire (NSQ). This widely used questionnaire provides a “quantitative
evaluation of neurotic trend”(Scheier & Cattell, 1961). It comprises 40
statements (e.g., “sometimes I let small things get on my nerves too much”
“worrying keeps me awake at night”) answered on a 3-point scale (yes,
in-between, no). A high score denotes greater anxiety and depression. The
NSQ is brief, suitable for use with adults and adolescents, and has well-at-
tested reliability and validity. Scheier and Cattell (1961) report split half
consistency coefficients of .57 for depression and .70 for anxiety.
3. Readiness for Intimacy Scale of the Erikson Psychosocial Inventory
Scale (Rosenthal, Gurney & Moore, 1981). This was developed to
operationalise and test Erikson’s 8-stage theory of psychosocial develop-
ment. The Intimacy scale has 12 items measuring the developmental task
of establishing love and intimacy in early adulthood. Rosenthal et al. re-
port Cronbach’s alpha of .73 and .63 in two samples. A high score denotes
greater readiness for intimacy. Cronbach’s alpha in the present sample
was .69.
4. Wariness scale (Burns & Dunlop, 1998). This 11-item scale was con-
structed by the authors to measure wariness about long-term relationships,
marriage and family life, as described by Wallerstein and Blakeslee
(1989). Items are answered using a 4-point scale from “very like me” to
64 JOURNAL OF DIVORCE & REMARRIAGE

“very unlike me.” A high score indicates greater wariness. Cronbach’s al-
pha in the present sample was .86. Principal components analysis revealed
that one factor accounted for 42% of the variance. All items loaded on this
factor, with 10 having loadings of .59 or more.

Analysis

Correlations between assessments of conflict and marital quality are


presented for the full sample of mothers, fathers and adolescents inter-
viewed at Year 1. In the Manovas of conflict/marital quality variables
against child outcome measures at Year 1 and Year 10, however, only the
results for those offspring present at both the initial and adult follow-up
interviews are used.

RESULTS

1. Do family members agree in their assessments of family conflict


and parental marital quality?

Table 1 presents correlations between father and mother assessments at


Year 1, and child assessments at Years 1 and 10. It shows that the father
and mother assessments made at Year 1 are quite highly inter-correlated,
with the Spanier and the conflict resolution measures more strongly asso-
ciated (range of r = 54 to .70) than was either with the level of conflict
measures (range r = .17 to .48). The father and mother assessments of
level of conflict were however well-correlated (r (60) = .53, p < .01).
Adolescent’s Year 1 assessments of level of conflict agree to some ex-
tent with those of their fathers (r (63) = .33, p < .01) and rather more with
those of their mothers (r (72) = .53, p < .001). Their assessments of their
involvement in parental conflict is not associated with any other of the
conflict measures.
Adolescent’s Year 10 assessments of the current level of family con-
flict show only one significant correlation with a parent measure (father
level of conflict: r (55) = .32, p < .05: Column 9 of Table 1). However their
Year 10 assessments of the family of origin’s level of happiness at Year 1
(Column 11) are quite strongly associated with the parent’s initial self-as-
sessments, indicating that the adult children have some understanding of
their parent’s feelings ten years ago. Their current assessments of happi-
ness in the family of origin (Column 12) are also in tune with the parent’s
initial self-assessments. The retrospective and current assessments are not
Ailsa Burns and Rosemary Dunlop 65

TABLE 1. Correlations Between Parent Measures of Conflict and Marital Satis-


faction at Year 1, and Child Measures of Parental Conflict and Happiness at
1,2
Years 1 and 10

PARENTS 2 3 4 5 6 7 8 9 10 11 12
1. Fa Spanier 2.48*** 2.63*** .70*** 2.50*** 2.52*** 2.05 2.19 2.10 2.08 .38** .51***
(62) (62) (56) (57) (57) (61) (61) (53) (52) (53) (51)
2. Fa Conflict 2.17 2.29* .53** .27* .33** .05 .32* .09 234** .46***
(64) (59) (60) (60) (63) (63) (55) (54) (55) (52)
3. Fa Disagree 2.61*** .02 .56*** .11 .07 .00 .01 2.23 2.38**
(59) (60) (60 (64) (63) (55) (54) (55) (53)
4. Mo Spanier 2.23* 254*** 2.10 2.05 2.11 2.29* .36** .54***
(73) (72) (72) (69) (63) (62) (63) (61)
5. Mo Conflict .23* .53*** .09 .14 .04 2.28* 2.29*
(73) (72) (70) (64) (63) (64) (62)
6. Mo Disagree .18 .09 .06 .08 2.31** 2.48***
(72) (70) (64) (63) (64) (62)
CHILD
7. Year 1 – .12 .19 .27* 2.47*** 2.20
Conflict (74) (65) (64) (65) (63)
8 Y1– 2.06 .31* 2.07 2.02
Involvement (63) (62) (63) (61)
9. Year 10– .25* 2.01 2.03
Conflict (66) (67) (65)
10. Year 10– 2.08 2.28
Involvement (66) (65)
11. Year 10– .23
Family of origin (65)
happy at Y1
12. Year 10–
Family of origin
happy at Y10

1
Higher scores denote higher (Spanier) dyadic adjustment, more conflict, more involvement in conflict, less satisfactory conflict resolution, and
greater family happiness.
2
Different Ns reflect the different numbers of mothers and fathers taking part, and missing values.
*p < .05 **p < .01 *** p < .001

themselves significantly correlated, indicating that the adult children are


distinguishing between the happiness of the family of origin then and
now, but that both assessments reflect to some extent the parent’s original
self-reports.
Adolescent’s self-reports of Year 1 involvement in conflicts (Column
8 of Table 1) is correlated only with its repeat measure at Year 10 (r (62)
= .31, p < .05). However those reporting involvement at Year 10 described
more family conflict at Years 1 and 10. This suggests that as adults, but
not as adolescents, offspring are likely to feel more involved when they
perceive parental disharmony as greater. Involvement at Year 10 was also
associated with mother’s reports of lower marital satisfaction at Year 1.
66 JOURNAL OF DIVORCE & REMARRIAGE

This could result from adult empathy with maternal dissatisfactions that
have persisted over time.

2. Do assessments of family conflict and parental marital quality pre-


dict the children’s adjustment and self-image in early
adolescence?

Table 2 presents overall model values and significant predictors for the
intact and divorce group combined. The adolescent’s assessment of the
level of conflict was the only significant predictor of self-image at Year 1,
with those offspring who perceived greater family conflict having poorer
self-image. Similarly, the child’s assessment of level of conflict was the
sole (positive) predictor of anxiety. There were no significant predictors
of depression.
Greater family conflict and poorer parent marital quality were corre-
lated with poorer self-image and more anxiety and depression. However
only the adolescent rating of the level of conflict was a significant predic-
tor. Neither parent group nor gender was significant as a main effect or in
interaction with the parent marital quality variables.

3. Do assessments of family conflict and parental marital quality pre-


dict the children’s adjustment and the quality of their own
intimate relationships in early adulthood?

Table 3 shows that by adulthood, sons had more positive self-image


than daughters. Of the parent marital quality variables, only the adult
child’s estimate of the family’s current level of happiness predicted (posi-
tively) current self-image. There was a (non-significant) trend for greater
family conflict at Year 1 (as perceived by the adolescent) to be associated
with poorer adult self-image.
Daughters (in both family groups) now reported significantly more
anxiety than sons, but no other predictors reached significance.
Depression was significantly predicted by the interaction between the
children’s gender and their Time 1 assessment of the level of family con-
flict (t (65) = 22.76, p = .008). At the same time, gender as a main effect
almost reached significance (t (65) = 1.94, p = .056). Daughters overall re-
ported somewhat more depression than sons, and daughters from higher con-
flict backgrounds were more depressed than those reporting lower Year 1
conflict (as measured by a test of simple slopes: Aitken & West, 1991).
Sons from higher conflict families were also more depressed than other
Ailsa Burns and Rosemary Dunlop 67

TABLE 2. Regression of Year 1 Conflict and Marital Satisfaction Variables on


Child’s Self-Image, Anxiety and Depression at Year 1
1 2
Outcome Predictors Coeff Overall Model
F df p
Self-Image Ch ild–Conflict level*** .21 7.97** 1,65 .007
Anxiety Child–Conflict level** 1.56 8.28** 1,65 .005
Depression None - - -

1 Higher scores denote a higher level of conflict, poorer self-image and greater anxiety and depression.
2 Higher scores denote a higher level of conflict.
** p < .01 *** p < .001

TABLE 3. Regression of Year 1 Conflict and Marital Satisfaction Variables on


Child’s Self-Image, Anxiety and Depression at Year 10

Outcome1 Predictors2 Coeff Overall Model


F df p
Self-Image Ch–Family of origin happy at T3 2.21*** 10.11 2,65 .0002
Gender 2.23**
Anxiety Gender 2.83* 4.18 1,65 .045
Depression Gender 22.66 4.72 3,65 .005
Ch– Conflict level Y1 2.12
Gender (daughters) 3 Ch - Conflict Level Y1 21.64*
Intimacy Ch– Involvement T1 2.21 3.01 1,64 .088
Wariness Parent group 26.87* 3.06 3,65 .035
Ch–Involvement T1 2.46
Group (intact) 3 Ch - Involvement Y1 6.57**

1 Higher scores denote poorer self-image, more anxiety and depression, greater readiness for intimacy and greater wariness.
2 Higher scores denote greater family happiness, higher level of conflict, involvement in parent conflict, female gender and membership of di-
vorce group.
* p < .05 ** p < .01 *** p < .001

sons, but the effect was not significant. Parent’s marital status did not ex-
ert an effect.
Of the two measures used only at Year 10, readiness for intimacy was
not predicted by any of our measures, although there was a trend for those
reporting involvement at Year 1 to be less ready for intimacy. Wariness
however was predicted both by parent group (t (65) = 22.53, p = .014)
and by the interaction between parent group and adolescent reports of be-
ing involved in parent’s conflict (t (65) = 2.26, p = .027). The divorce
group were more wary overall, but their level of wariness was unaffected
by their Time 1 report of involvement (as measured by a test of simple
slopes). Within the intact group however, greater wariness was predicted
68 JOURNAL OF DIVORCE & REMARRIAGE

by greater involvement at Year 1 (t (65) = 2.70, p = .009). Figure 1 illus-


trates this outcome.
Apart from the association noted above between current parental happi-
ness and child self-image, there were no associations between current par-
ent harmony measures and child outcomes.

4. Does parental marital quality have different effects on children


from divorced and intact and divorce homes?

There was no difference between the intact and divorce groups on


self-image, anxiety, depression at Year 1, nor on these measures and
readiness for intimacy at Year 10. The only significant group difference
was that on wariness about relationships (noted above).

DISCUSSION

The starting point of this analysis was the “cross-over” proposal


(Amato & Booth, 1997; Wallerstein et al., 2000) that it is better for chil-
dren’s adult adjustment if their unhappily married parents continue the
marriage when the level of conflict is only moderate or less; but worse for
children if they continue a high-conflict marriage.
We first considered the question of whose reports of marital conflict
and satisfaction are to be used. We used multiple measures and found that
there was some agreement between parents and children, but the correla-
tions were not especially high. Inspection of agreement patterns in indi-
vidual families showed that in some cases where both parents rated
conflict as high, their child rated it as low or moderate; and in other cases,
children rated conflict as high while parents rated it low or moderate. This
comes as no surprise, as it is well-established that children’s views of their
parent’s attitudes and behaviours vary quite a bit from the parent’s self-de-
scriptions (Cashmore & Goodnow, 1985). It does raise the question how-
ever of whose reports should be used as the marker. As phenomenological
theory (Denzin & Lincoln, 1994) would predict, we found that it was the
children’s own perceptions–of the level of conflict and of their involve-
ment in this–that influenced their adjustment. Parent measures of marital
quality were associated with child outcomes in the predicted direction, but
did not approach significance.
In agreement with previous research (Amato & Booth, 1997; Kelly,)
we found that the degree of conflict during children’s early adolescence
(as perceived by the child) influenced their self-image and anxiety level at
Ailsa Burns and Rosemary Dunlop 69

FIGURE 1. Adult Wariness Level by Adolescent Involvement in Conflict

30

28

Wariness

26 Intact
Divorce

24

22

Not Involved Involved

the time. There were no significant differences between the intact and di-
vorcing group on these measures.
Our findings at Year 10 were more complicated. Wallerstein describes
parental divorce as continuing to distress children 25 years after the event,
and Amato and Booth (1997) found that adult children’s psychological
distress, self-esteem, happiness and life satisfaction were predicted by
parent’s ratings of their own marital quality made 12 years earlier. Our
data did not show these outcomes. Our only predictors of adult self-image
were gender (with sons in both groups having better self-image than
daughters) and current perceptions of the happiness in the family of ori-
gin. Our self-image scores were reasonably consistent over the ten years
of the study (r = .54: Dunlop & Burns, 1995), and at each time of measure-
ment they were significantly associated with current assessments of har-
mony in the family of origin. However adult self-image was not
influenced by past reports of this (whether made by parent or child). This
is a positive outcome. It indicates past dysfunctional aspects of the paren-
tal relationship need not dog the children into the own adult years. At the
70 JOURNAL OF DIVORCE & REMARRIAGE

same time, the significance for adult self-image of the current assessment
of happiness in the family of origin suggests that contemporary (rather
than past) family relationships can be more important than is generally
considered.
Earlier parental marital quality also failed to predict anxiety at Year 10,
although it had done so at Year 1; again suggesting a diminution over time
of the negative influence of the past. The only significant effect came from
the well-known greater anxiety of adult females (Scheier & Cattell, 1961),
which was found in both family groups.
Our measure of depression gave more complex results. In adolescence the
divorce group sons tended to be more depressed, and the divorce group
daughters less depressed than the two intact groups, but by adulthood there
had been a change. Daughters in both groups were now more depressed than
sons, although the effect did not quite reach significance. Gender was how-
ever significant in interaction with past reports of conflict. As adults, daugh-
ters from higher-conflict backgrounds reported significantly more depression
than those from low-conflict backgrounds (regardless of family group). The
same trend was found among adult sons, but did not approach significance.
Adult females are known to be more subject to depression than males
(Scheier & Cattell, 1961). Our results suggest that this tendency can be exac-
erbated by a high-conflict family background, although the effect may not be
apparent until adulthood. This intriguing pattern of changes over time merits
further exploration.
Wallerstein (Wallerstein & Kelly, 1980; Wallerstein & Blakeslee,
1989: Wallerstein, Blakeslee & Lewis, 2000) has consistently emphasised
the anxiety felt by the children of divorce regarding their own chances of
entering a successful long-term relationships. Wallerstein and Blakeslee
refer to a “significant number” of daughters who suffer a “sorrowful sense
that they are unlikely to achieve their goals of an enduring love and mar-
riage” (p. 24) and “an intolerable level of anxiety about betrayal” (p. 62).
Amato & Booth by contrast report that parental divorce has very little ef-
fect on the quality of offspring’s relationships. Our own data distinguishes
between the effects on readiness for intimacy and on wariness about rela-
tionships. Our measure of the former (Rosenthal et al., 1981) is based on
Erikson’s (1950) writings about the early adult developmental task of es-
tablishing intimacy with others, and focuses on a person’s ability to be-
come involved (e.g., “I find it easy to make close friends” “I’m warm and
friendly”), while the latter (Burns & Dunlop, 1998) emphasises wary atti-
tudes (“I worry about getting trapped in a long-term relationship”
“Long-term relationships only succeed if both partners work at them”).
Readiness for intimacy was not associated with any of our predictor vari-
Ailsa Burns and Rosemary Dunlop 71

ables, and is perhaps partly a trait of temperament or personality. Wari-


ness however was higher in the divorce group and among those in the
intact group who reported being involved in their parent’s conflicts. We
conclude that a disruptive family background is likely to influence atti-
tudes to adult relationships, but not the personal qualities involved, and
that it is useful to distinguish between the two.
What support then do our findings then provide for the “cross-over”
proposal of Wallerstein, Blakeslee and Lewis (2000) and Amato and
Booth (1996) that children are better off as adults if their highly-conflicted
parents divorce, but worse off if less-conflicted parents do so? There is
partial support from the findings on wariness about relationships. Adoles-
cents in continuing families who felt involved in parental conflicts were as
adults more wary than those whose parents divorced, suggesting that in
this aspect of adult life, living with conflicted but continuing-married but
parents was more damaging than parental divorce. Divorce group off-
spring who were not involved in parental confict as adolescents were as
adults somewhat more wary than those who were, which is in line with the
notion that low-conflict divorce brings its own problems for children.
However, the difference in wariness between the “involved” and the
“non-involved” in the divorce group was not statistically significant. Sup-
port for the “cross-over” hypothesis is thus weak.
With regard to other aspects of adult adjustment, our findings were that
family conflict, and involvement in that conflict, had long term effects,
but the effects were similar for participants from intact and divorced
homes. The interaction between parent group and the various conflict and
parent marital quality variables were included in all analyses, but none
proved significant. In respect of self-image, anxiety, depression and readi-
ness for intimacy, there was thus no support for the hypothesis that chil-
dren from high-conflict intact backgrounds were worse off than those
from high-conflict divorced backgrounds, or those from low-conflict di-
vorced homes worse off than those from low-conflict intact homes.
Our data base is of course different from that of the studies that found
evidence of a cross-over effect. Wallerstein’s sample of children of di-
vorce (Wallerstein, Blakeslee & Lewis, 2000) is similar in size to our
own, but her case study approach makes comparisons difficult. Her
strength lies in the detailed histories of selected individuals that she pro-
vides, and the many insights that come out of these stories. She has said
that “I knew their history, and they could talk frankly without trying to put
a fine face on everything, which people do. All people do. They all say
they’re fine even if they’re dying, but they didn’t with me. And they lev-
72 JOURNAL OF DIVORCE & REMARRIAGE

eled with me, or if they didn’t I called them on it, because I knew them”
(<Borders.com> 2000). This indicates a style of discourse that most re-
searchers do not engage in, and that renders quantification and compari-
son with a non-divorce group virtually impossible. Our data suggests that
it can result in misleading generalisations.
Amato and Booth’s sample is much larger than our own, and differed in
a number of ways. The children involved were older, and their age range
much greater when first interviewed (19 to 40 years). Their age range at
the time of the parental divorce was also greater (preschool to early twen-
ties), while our own study controlled for recency of divorce and adoles-
cent age at the time when this occurred. In addition, we used different
measures, usually more extensive ones such as the OSIQ; and child as
well as parent reports of conflict at Year 1. And these measures were em-
bedded in lengthy interviews, whereas Amato & Booth used briefer tele-
phone interviews. Given these many differences, we would expect some
divergence in findings.
Even so, the differences are puzzling. Except in the case of wariness, the
interactions between parent marital status and parent conflict did not re-
motely approach significance, and thus cannot be explained as due to the
modest power of a relatively small sample. Nor was our sample unduly
biassed. A random sample of families applying for divorce in Sydney was
approached, and analyses (Dunlop & Burns, 1995) showed that those who
took part were spread across the socioeconomic range, and that there were
few sociodemographic differences between the intact and divorce groups.
What then can explain the substantive differences between the US stud-
ies and our own? Cultural factors may be relevant here. The divorce rate in
Australia is considerably lower than in the USA (2.8 per 1000 population
in 1999 vs 4.7 in the USA in 1990), which could mean that the divorce
population in the two countries is somewhat different, with the Australian
group weighted towards more serious marital complaints. Wallerstein’s
(2000) emphasis on the resentment felt by children of divorce about their
reduced educational chances suggests another cultural effect. While sin-
gle-parent families in both countries are economically disadvantaged, it
may be that different social provisions result in Australian children suffer-
ing fewer educational penalties (Smith Family, 2001).

SUMMARY
Ailsa Burns and Rosemary Dunlop 73

The study looked at the predictive power of mother’s, father’s and ado-
lescent children’s reports of family harmony on the adolescent’s adjust-
ment at the time and as adults. The ratings made by different family
members were correlated, but only the adolescent’s ratings predicted their
own self-image and emotional adjustment. Ten years on these measures
were no longer predictive, although daughters from higher conflict fami-
lies reported more depression. The readiness for intimacy of the
now-adult children was unaffected by the earlier reports of conflict and
parental marital quality, but wariness about relationships was higher
among those from divorced homes, and those from continuing marriages
who as adolescents had described themselves as involved in parental con-
flicts. The results offer some support for the proposal that involvement in
parental conflict in continuing families can have negative outcomes for
children; but no support for the proposal that divorce by parents whose
level of overt conflict is low exerts a particularly damaging long-term ef-
fect on their offspring.

REFERENCES
Aiken, Leona S. & Stephen G. West. (1991). Multiple regression: Testing and interpret-
ing interactions. Newbury Park, CA: Sage Publications.
Amato, P. & Booth, A. (1997). A generation at risk. Cambridge, MA: Harvard University
Press.
Borders.com (2000). Interview with Judith Wallerstein, 18.9.2000.
Buchanan, C., Maccoby, E.E., & Dornbusch, S. (1991). Caught between parents: Adoles-
cent’s experence in divorced homes. Child Development, 62, 1008-1022.
Burns, A. & Dunlop, R. (1998). Parental divorce, parent-child relationships, and early
adult relationships: A longitudinal Australian study. Personal Relationships, 5,
393-407.
Carter, B. & McGoldrick, M. (Eds.) (1980). The changing family life cycle: A framework
for family therapy. New York: Gardner Press.
Cashmore, J. & Goodnow, J. (1985). Agreement between generations: A two-process ap-
proach. Child Development, 56, 493-501.
Congalton, A. (1969). Appendix B: Status ranking list of occupations in Australia. In A.
Congalton (Ed.) Status and prestige in Australia. Melbourne: Cheshire.
Denzin, N. & Lincoln, Y. (Eds.) Handbook of qualitative research. Thousand Oaks: Sage.
Dunlop, R. & Burns, A. (1995). The sleeper effect–Myth or reality? Journal of Marriage
and the Family, 57, 375-386.
Dunlop, R. & Burns, A. (1988). “Don’t feel the world is caving in”: Adolescents and di-
vorcing families. Melbourne: Australian Institute of Family Studies.
Erikson, E. (1950). Childhood and society. New York: Norton.
Jekielek, S. (1998). Parental conflict, marital disruption and children’s emotional
well-being. Social Forces, 76, 905-936.
74 JOURNAL OF DIVORCE & REMARRIAGE

Kelly, J. (1993). Current research on children’s post-divorce adjustment: No simple an-


swers. Family and Conciliation Courts Review, 31, 29-49.
Offer, D., Ostrov, E. & Howard, K. (1977). The Offer Self-Image Questionnaire for ado-
lescents: A manual. Chicago: Michael Reese Hopital.
Rosenthal, D., Gurney, R. & Moore, S. (1981). From trust to intimacy: A new inventory
for examining Erikson’s stages of psychosocial development. Journal of Youth and
Adolesence, 10, 525-537.
Scheier, I. & Cattell, R.B. (1961). Handbook for the Neuroticism Scale Questionnaire
“The NSQ.” Illinois: Institute for Personality and Ability Testing.
Smith Family (2001). Children’s educational acheivement in a low-income population.
Melbourne: Smith Family.
Spanier, G. (1976). Measuring dyadic adjustment: New scales for assessing the quality of
marriage and similar dads. Journal of Marriage and the Family, 38, 15-28.
Wallerstein, J. & Kelly, J. (1980). Surviving the breakup: How children and parents cope
with divorce. New York: Basic Books.
Wallerstein, J. & Blakeslee, S. (1989). Second chances. New York: Ticknor & Fields.
Wallerstein, J., Blakeslee, S, & Lewis, J. (2000). The unexpected legacy of divorce. New
York: Hyperion.

You might also like