Intimate Pathways: Changing Patterns in Close Personal
Relationships Across Time*
E. Mavis Hetherington**
‘This paper presents flings from she Virginia Longitudinal Study of Divorce and Remarriage (VESDR) describing diverse patterns of
inmate relationships and personal adjusiment in marviage and following divorce. Both a conficual, unsatisfying marriage and a
divorce were associated with diminished psychological, social, ane physical wellbeing. However, it was the diversity rather thar the
inevitability of outcomes following divorce that was strike, with most people able to adapt constructively to their new life situation
within 2-3 years following divorce, a minority being defeated by dhe marital breakup, and a substantia group of women being enhanced.
Alowgh both marital conflict and divorce in te family of origin elevated he risk of marital instability in young adult offspring, the
‘effect was greater for divorce, Marriage toa supportv
, well-adjusted partner by youths from divorced families eliminated the difference
in marital instability found for these youths and those fram nondivorced families
in human development. Throughout the life course the qual-
ty of close personal relationships can promote or undermine
‘our psychological and physical health; our security, well-being.
and competencies; and the way we view ourselves—whether we
sce ourselves as Valued and worthwhile or as deficient and un-
‘worthy.
The family plays a unique role in forming and s
intimate relationships; however, there have been notable changes
in the family in the past 50 years. AS marriages are being de
layed, birth rates are decreasing, and maternal employment,
vorce, cohabitation, and births to single mothers are increasing,
the course of intimate relationships is becoming more diverse
and less stable and predictable (Teachman, Tedrow, & Crowder
2000; U.S. Bureau of the Census, 1998).
This paper describes intimate pathways through marriage
and divorce and its aftermath and how some of these pathways
lead to unhappiness and despair and others to fulfillment and
often enhancement. It also will describe how marital transitions
in the parents” generation effect the intimate relations of their
‘offspring when they become young adults and marry, The focus
will be on adult couple relationships and the factors that con-
tribute to their success or failure.
In this paper [address four general questions
1. What types of marriage are at greatest or least risk for
‘marital instability’?
2. What are the responses of men and women to divorve
and how do their adaptive patterns shift over time?
3. What are the major factors that contribute to marital in-
stability?
4. What factors contribute to or protect against the inter
‘generational transmission of divorce?
Virginia Longitudinal Study of Divorce
and Remarriage
Because this paper was originally presented as the 2001 Bur-
gess Award Address, it summarizes a wide array of findings and
[em relationships are the most important guiding forces
Pres he Burgess Ava as the mings of he National Cov
Fmity Relaons in Roches, New York, November 20
Dspanmeat of phuley, Univeriy of Vira 102 Gime Hal, PO. Box
so, Chev, VA 22908 nate mal gina)
Sey Wer lirmen ilo, vers, ody, reemrae
318
does not present the methodological detail customary in a re-
search paper. All findings, unless otherwise noted, were signfi-
cant at p< 05, and publications with more methodological de-
tail are referenced throughout
‘The content is based on data from the Virginia Longitudinal
‘Study of Divorce and Remarriage (VLSDR; Hetherington, 1993;
19992, 1999; Hetherington & Kelly, 2002). The VLSDR was
initially designed as a 2-year study of divorce during which the
families would be seen three times, because 2-3 years was
viewed as the period required for parents and children to adjust
to divoree. However, 2 years after divorce many divoree-related
‘changes were still occurring in our families, and family members
were showing diverse patterns in coping with divorce. To ex-
amine these postdivorce pathways, it became obvious that our
rescutch efforts needed to follow the families for a longer period,
and to enlarge our sample.
‘The original sample of the VLSDR included 144 families,
half nondivorced and half divorced with custodial mothers and
target child 4 years of age. Families were identified and re.
cruited through court records of marriage and divorce and
through schools. Eighty-six percent of families contacted agreed
to participate in the study. The attrition rate in the study was,
modest, with only 9% of families who ever began to participate
in the study eventually dropping out. Dropout rates were slightly
higher for noncustodial fathers (15%) than for custodial parents
and chikiren or for nondivorced parents. Families were White
‘and middle class, with parents slightly above national norms in
education and income. Families were studied at 2 months and 1,
6, L1, and 20 years following divorce, when the target children
were aged 4, 5, 6, 10, 15, and 24 years old successively. Non-
divorced families were studied at equi
sample of 144 was expanded in sueceeding waves to 180 at wave
4, 300 at wave 5, and 450 at wave 6, equally distributed among
nondivorced, divorced, and remarried fa id by gender of
the offspring. In addition, from wave 4 on, when the target chil-
dren were 10, if there were two or more children in the family.
the sibling closest in age 10 the target child also was studied. In
this paper, only data from the target offypring and not the siblings
ave presented. When the offspring or sibling married, eohabited
ore than 6 months, divorced, oF had a child, additional assess-
‘ments focusing on the couple and parent-child relationships were
nade within 6 months. An additional wave of assessments of
family relations and adjustment is being made when the first
child of the offspring turns 4, the age at which the offspring’s
parents were when the study began. Brief annval telephone in
Family Relationsterviews also were conducted, and biannual family status forms
were mailed in by parents and youths.
Tt should be noted that by 2002, when the target offspring
were an average of M years of age, there was considerable
iation in the timing of divorce, repartnering, and rem: be
cause as the study proceeded many couples in both generations
of the sample changed their relationship status through cohabi-
tation, divorce, and remarriage, sometimes several times. The
most extreme case was one of the mothers who had divorced
five times over the course of the study and had had 14 cohabiting
relationships. She described herself as being “just a marrying
kind of gal.” Mothers, fathers, and children continued to be fol-
owed in their new status afier a marital transition
These shifis in marital status presented some problems in
expanding the samples in waves 4, 5, and 6, For example, the
divorced group now contained families who had initially heen
in the nondivoreed group and who had divorced after the chil-
dren were 4 years of age. The additional children in the expand-
ed sample were matched on age at time of divorce with the
transformed sample, yielding a more heterogeneous sample on
that variable than was present in the initial sample. Thus, in wave
4, children in the divorced sample did not all have parents who
had divorced when the child was 4 years old, although all chil-
dren in this wave were 10 yeats of age.
In all analyses, age at the time of & marital transition and
time since the transition were examined. Few effects for age at
the time of divorce on the adjustment of children were obtained:
however, older parents showed more difficulty in adjusting to
divorce than did younger parents. In addition, both parents and
children showed mote distress and problems in adjustment in the
‘early years following divorce. Although shifts in relationships
over the course of the study were problematic in sample expan-
sion in waves 4, 5, and 6, they did permit an examination of the
antecedents and sequelae of marital transitions without relying
‘on retrospective reports for a substantial subset of our couples.
‘Throughout the study, detailed multimeasure, multimethod
assessments were made of parent and offspring personality and
adjustment; relationships outside of and within the family, i
cluding the marital relationship; and (in wave 6) of character
tics of and relationship with the partner in the cohabiting and
marital relationship. Interviews, standardized personality and
marital relationship tests, behavior event checklists involving 10
telephone interviews, and diary records were conducted at each
‘wave, Observations of couple behaviors in family problem-solv-
ing sessions and at the dinner table, as well as standardized tests
and interviews of the partner, were integral parts of the assess-
‘ment of the parents’ marital relationship and the offspring couple
relationship. Each wave of data collection usually involved two
to three sessions with two interviewers in the home and a large
take-home battery of tests 10 be completed between the frst and
second session. At the end of each assessment wave, interview-
er completed a set of ratings of family functioning and the ad-
Justment of family members (more details on measures are avail-
‘able in Hetherington, 1993; 1999a; Hetherington & Elmore, in
press; Hetherington & Kelly, 2002),
Because we were interested in the intergenerational trans-
mission of divorce and relationship instability in our second gen-
eration, we were also concerned about the family history and
family relationships of the partner they selected. For this, we had
to rely on the partners” reports supplemented by their parents’
reports in telephone interviews in about 90% of the eases
For most of the constructs examined here, such as socio-
2003, Vol. 52, No. 4
economicflife course risk and personality problems and positive
adjustment in parents, youths, and spouses: authoritative parent-
ing and negative/coercive parenting; problem-solving skills, neg-
ativity, and positivity in couple interactions; commitment: and
‘mate selection risk, composite multimethod, multi-informant
measures were derived. Scores on each variable were converted
to z scores, and factor analytic techniques and image analyses
were used to derive the composite measures (see Hetherington
& Clingempee!, 1992, for a description of methods of compos-
iting), Cronbach's e for the various composites ranged from .72
to .89, Cluster analyses using the K-Means program of BMDP
(Dixon & Brown, 1979) were used to establish typologies or
profiles of such things as types of relationships or patterns of
‘coping, Cluster analyses were done at each wave of data collee-
tion to permit a migration analyses using x? to show how rela-
onships and adjustment styles change over time:
Marital Typologies
By the time they were 32 years of age, over half of our
target youths had cohabited, 76% had married, and of those, over
‘one quarter were now separated or divorced. By age 32. only
14% of our men and 18% of our women had not been in a
cohabiting or marital relationship. However, by assessing only
relationships of 6 months’ duration, we lost 5% of cohabitors
but ess than 1% of married couples. Relationship quality and
ity in youths was assessed 6 months after a cohabiting
relationship or marriage occurred. In parents it was assessed in
every wave.
‘Through cluster analyses of composites of 12 dimensions of
couple functioning, we identified five configurations of couple
relationships in our parents and the same set of marital typolo-
siies emerged in our 342 youths who had married. We labeled
these as pursuer-distancer, disengaged, operatic, cohesive indi-
viduated, and traditional marriages. The percentages of parents
i youths in the different categories varied, with modern youths
having fewer traditional marriages and more cohesive individu-
ated marriages than their parents had in any wave of data col-
lection. Moreover, in parents the percentage of traditional mar-
riages had decteased and of cohesive individuated marriages had
increased over time. In the following discussion, the percentage
‘of youths in different marital typologies was assessed at 6
‘months after marriage, and the percentage of parents in different
typologies was averaged across all waves of assessment. Thus,
youths were in the early stage of a marriage, and parents were
assessed across time in longer-established marriages,
‘We used x° analyses to examine whether youths in the dif-
fering marital typologies were more likely to have come from
high- or low-conflict intact or divorced families. We also were
interested in the risk of marital instability for our five types of
bility was measured with a modification
cale developed by Booth, Johnson, and
combination of the husbands” and
stability scores. This scale has good demonstrated va-
lidity and reliability in predicting marital breakdown (Booth,
Amato, Johnson, & Edwards, 1993; Wolfinger, 2000). In the pare
ents’ generation of the VLSDR, it predicted divorce with 92%
accuracy. However, itis not possible to estimate how accurately
it will predict for their offspring, because many of the youths
hhad been married for only a short time
Multivariate analyses of variance (MANOVA) were run
with marital type predicting marital instability. The ordering of
319
of a Marital Instability
Edwards (1983) and128
WNEG WITH WPOS—-MNEGNWITHH MOS
a chuster
+ Cluster?
che Cluster3
Figure 1. Marital typologies
CONF SEXSAT MALT WAUT SHARE TRAD.
‘Note: Cluster 1 is a poruerdistancer duster) Guster 2, a di
withdrawal; WROS ~ woman warmtvSupport; MNEG
= sexual satisfction: MAUT = man autonomy: WAUT
the risk of marital instability across generations was similar,
ranging from pursuer-distancer marriages with the greatest and
traditional marriages with the least marital instability
Figures 1 and 2 present the cluster profiles for youths 6
months after marriage (n = 342), (Clusters were divided into
two figures for legibility.)
Cluster 1 was the familiar pursuer-distancer cluster (Gott-
man & Notarious, 2001). It was one of the largest clusters iden-
tified for youths and for parents across all waves. This cluster
was characterized by high nagging, hostile criticism, and con-
tempt on the part of wives; withdrawal and denial by husbands:
frequent intense conflict; and the highest rate of masital insta-
bility of any group in the study. The pursuer-distancer combi
nation can be thought of as a misfit. If both wanted to talk about
their problems or if both wanted to avoid conflict, their marriages
might be more successful. Twenty-five percent of our youths,
slightly more than in their parents’ marriages, were in pursuer-
distancer marriages. Youths from families in which there had
been high interparental conflict whether in a divorced or non-
divorced family, in contrast to those from low-conflict families
‘of origin, were overrepresenied in this cluster.
Cluster 2 was a disengaged cluster, characterized by hus-
bands and wives who led largely parallel lives. They had few
shared interests, activities, or friends. Even their sexual relations
were infrequent and unsatisfying. Disengaged couples rarely
fought, because they expressed litle extreme positive or negative
emotion in their interactions and they withdrew from most dis-
agreements. These were affectively bland unions. Our interviews
indicated that some disengaged marriages are marriages of con-
venience. The couple did not want intimacy or companionship,
but they did want children or the security, status, and services a
marriage can provide, Others were based on a romantic or sexual
attraction, but once the attraction faded, there was litle to hold
the couple together. Disengaged couples had the second highest
320.
jgged closer, and Cluster 3, an oper casiez WNEG = woman negativiy
‘man negativity: MWITH = man wit
voman autonomy; SHARE = shared interests and activities; TRAD
WWITH = woman
‘confit; SEX SAT
al; MPOS ~ mun warmihsuppart: CONF
cational
rate of marital instability as assessed through husbands and
wives’ reports on the Marital Instability Scale, In the parents’
generation, where we followed couples over a long period of
time, we found that pursuer-distancer marriages broke up earlier,
often with considerable acrimony, whereas disengaged marriages
lasted longer and ended not with a bang but with a. whimper.
‘The couple drified along in their separate lives. Then, often in
their late 30s or early 40s, one spouse would look at his or her
life and marriage and think, “Is this as good as it gets?” and
bail out, Sixteen percent of our youths, almost the same as in
their parents’ marriages, were in disengaged marriages, and it
was not related to whether they came from a divorced or non-
divorced family.
ur third cluster involved operatic marriages. Fourteen per-
cent of our youths and an average of 10% of parents across time
were in operatic marriages. Operatic men and women were emo-
tionally volatile sensation seekers. A harmonious, placid envi-
ronment bored operatics, who took everything to extremes, in-
cluding fighting and making love. They liked to funetion at a
level of intense emotional arousal. For operaties, quarreling was
often a trigger for sex, and operatics had the highest level of
sexual satisfaction in our studies. Following the parents’ operatic
marriages across time suggested that at fist the fighting-sex cy-
cle was almost a game, but that a danger lies in wait for operatic
ccouples. Angry people often say hurtful, sometimes unforgivable
things; moreover, anger can easily erupt into violence, Operatic
marriages in the parents’ generation had the third highest divorce
rate, and in youths had the third highest level of reported marital
instability. These marriages usually ended when one partner, zen
crally the man, decided that sex was no longer exciting enough
to balance constant conflict, but clean breaks were fairly uncom-
mon. Often, operatics would reconcile, break up, reconcile, and
break up again, Lingering attachments were common, and op-
eratics often reported less tumultuous relationships (0 be bland
Family Relations