ClinicalPsychology Rev&, Vol. 13, pp. 29-43, 1993 Printed in the USA. All rights reserved.
Copyright
0272.7358/93 $6.00 + .OO 0 1993 Pergamon Press Ltd.
THE PREDICTION AND PREVENTION MARITAL DISTRESS: AN INTERNATIONAL PERSPECTIVE
OF
Howard /. Markman University
of Denver
Kurt Hahlweg Braunschweig
ABSTRACT.
University
This art&presents the con@tual and empirical ration&for attentionto thepreventtin of
marital distress. Tk like of research+r the United States, its history, and then the results of a German repltiatzimand extension study are presented. The results suggest the possibilitiesjr preventing marital distress through teachingcouples to improve communicationand to handle conflit beforeproblems develop. Implicationsfor future researchare discussed.
The objectives of this article are to review the current and potential contributions of applying the principles of a prevention science to marital and family problems, to review a line of longitudinal research on the prediction of marital distress that sets apart the empirical and theoretical foundations for a prevention science for couples, briefly review a line of prevention a German
research in the United States, and present for the first time the results of
replication
and extension
study.
RATIONALE
FOR PREVENTION
The basic rationale for prevention as opposed to a treatment perspective on mental health problems, in general, is that treatment programs are akin to closing the proverbial barn door after the horse has left. In the marital and family therapy
area in particular,
great
debate has arisen concerning the long-term efficacy of marital and family intervention, and many have argued that perhaps even the best marital therapy programs make dis-
Paper presented at the 25th Association for Advancement of Behavior Therapy conference, New York, November 199 1. Correspondence should be addressed to Howard J. Markman, Center for Marital and Family Studies, University of Denver, Denver, CO 80208. 29
30
H. J. Markman and K. Hahlweg
tressed couples less distressed rather than truly happy (Hahlweg & Markman, 1988). In contrast, the prevention perspective has the goal of starting with happy couples (even though they may be at risk for future distress) and helping them maintain their relatively high levels of functioning. A more specific rationale summarized
elsewhere
for the prevention
(Fincham
& Bradbury,
of marital 1990;
and family distress has been
Markman,
Floyd,
Stanley,
& Sto-
raasli, 1988). Suffice it to say, marital and family distress is a major social problem, and successful prevention efforts can save untold personal, social, and economic costs (Duncan & Markman,
1988).
The Basic Principles of a Prevention A number
of basic principles
Science Applied
of prevention
to Marital
Distress
science have been recently
articulated
by a
NIMH study group (Coie et al., 1991). Most relevant for the current article are two points linking work in the prevention of marital distress to general mental health issues: 1. The goal of a prevention strategies
science is to provide the knowledge base and intervention
to prevent or ameliorate
a wide range of diverse disorders,
including
all
forms of psychopathology. Prevention science focused on risk and protectiue factors for these disorders, both in terms of increasing understanding of these variables and developing and evaluating intervention strategies to address the risk factors. 2. Prevention
science has identified
a relatively small number
of common risk factors that
underlie a wide variety of psychopathology. Marital distress is one of these common or generic risk factors. Specifically, marital distress has been linked to a wide range of both adult disorders (e.g., depression) and child (e.g., conduct) disorders. In addition, a good marriage is seen as one of the common and generic protective factors against a wide range of dysfunctions (Coie et al., 1991). A science of the prevention
of marital distress includes the following basic principles:
1. An identification of risk and protective factors based on up-to-date theory and research is essential in order to target the component of couples’ relationships for preventive intervention. 2. Preventive interventions can be conceptualized as adding to our knowledge about the phenomena of interest in that the best way to understand a phenomena is to try to change it. In the marital and family psychology field, it is very hard to do experiments. Preventive interventions with their longitudinal components provide one of the best opportunities to conduct experiments that enable us to evaluate change. 3. It is important to recognize that marital distress is not a static outcome measure. Couples may be distressed at one time, but clearly not at another. Here we are not only talking about the fact that most couples are nondistressed when they are married but become distressed later on, but that couples may be nondistressed 5 years but may be distressed 10 years into into marriage, or 7 years into marriage, marriage. 4. Similarly, the antecedents of marital distress, the predictor variables, are also not static entities. Thus, if researchers are interested in personality similarities and how that predicts the distress, they should be aware that different measures and personality at times may produce different indicators of similarity because people do change. Further, communication quality at one point in time may be very different than communication quality at another time. For example, over time, stability indicators
Prediction and Prevention
of communication
quality
negative communication
indicate
average
stability
31
coefficients
and .30 for positive communication
of about
(Clements,
.45 for
1992).
5. Finally, not only are predictor variables not necessarily stable, but there may be differences in predictability based on when the predictor variables are assessed. Thus, for example, we have found that in predicting child outcome for marital functioning variables, indicators of quality of marital interaction during the transition to parenthood are substantially better predictors than premarital variables as well as concurrent
marital interaction
variables (Balaguer
& Markman,
1992).
In the next section we describe briefly our up-to-date findings on the prediction of marital distress and divorce that provide the empirical basis for our prevention program.
DENVER FAMILY DEVELOPMENT The divorce rate indicates
STUDY: NEW LONGITUDINAL
that one in every two marriages
does include the countless other marriages
FINDINGS
will end in divorce,
and this
that are distressed but stable (National
Center
for Health Statistics, 1989). Despite the enormous social costs of divorce and distress, there is virtually no longitudinal research about the causes of marital distress. The present study was designed to test the hypothesis that premarital couples who are unable to handle differencesin their relationship will be those at highest risk for future divorce and distress. A secondary goal was to provide basic descriptive information on the course of relationship development
in the early years of marriage,
marital and family interactions. Early studies investigating the prediction personality
when the basic foundations
are laid for future
of marital distress focused on compatibility
of
and values, and revealed that these factors were generally not strongly related
to future marital happiness (Gottman, 1979). In contrast, studies of marital interaction have revealed strong associations between communication quality and marital happiness (e.g., Gottman, Markman, & Notarius, 1977). However, because most of these studies were cross-sectional,
inferences
are limited to the concomitants
rather than determinants
of marital distress. Since marital discord is a major risk factor for both adult and child psychopathology, and a major social problem in its own right, longitudinal studies of the evolution of marital discord are a research priority (Coie et al., 1991). While there have been other important longitudinal studies of marriage
that have
assessed the predictive power of marital interaction
Filsinger,
1988; Gottman & Krokoff, 1989), th e current marriage and assessed predistress functioning;
and related constructs
(e.g.,
study differs in that it (a) started before (b) continues through the primary risk
period for divorce, enabling the prediction of divorce as well as marital satisfaction; and (c) assesses communication annually to detect changes in important dimensions of marital interaction. Building
on these previous
efforts and our earlier longitudinal
research
(Markman,
1981), we examined the extent to which couples constructively or destructively handled differences when discussing problems before marriage-predicted divorce and marital distress 6 years later. Guiding our attention to the discussing/arguing process was the following model of the development of marital distress. We start with the assumption that couples entering marriage are faced with two major tasks: (a) handling conflict and disagreements and (b) maintaining and promoting intimacy. We assume that most couples will encounter problems and disagreements and that it is the couple’s ability to handle differences (not the differences themselves) that will be the critical factor in determining future marital success (Storaasli & Markman, 1990). F ur th er, since handling differences often requires
expressing
and receiving
negative
feelings in a constructive
manner,
the
32
H. J. Markman and K. Hahlweg
ability to regulate negative affect is highlighted as an important associate of marital success. In other words, couples must learn how to argue constructively, rather than destructively. Based on recent studies, we expect that there will be differences in how men and women handle conflict in marriage (Gottman & Krokoff, 1989; Markman & Kraft, 1989; Notarius & Pellegrini, 1987). In contrast to conventional wisdom, we believe that men do not have trouble with intimacy and connection in marriage, but do have more difficulty than women handling
conflict and negative emotions.
As noted by Gottman
(1983),
“In a sea
of conflict, men sink, women swim.” While we believe this is due to a complicated set of biological, sociological, and psychological factors (see Markman & Kraft, 1989, for details), it is beyond the scope of this paper to search for the cause of these gender differences. Here we predict that, as unresolved disagreements fester and worsen over time, men and women attempt to cope but use different styles: Women try to establish intimacy and connection by directly expressing psychological pain, while men increasingly withdraw to avoid conflict. These synergistic patterns fuel negative escalation cycles. Consistent with clinical experience, we speculate that men may not express their feelings because of fear of conflict, inadvertently increasing chances of destructive conflict, including verbal and physical abuse. Thus we test the hypothesis that husband withdrawal and wife invalidation will be among the best predictors of divorce assessed in the early stages of the relationship.
and distress
when interaction
is
Method Subjects.
The subjects were 135 couples planning
marriage
for the first time,
recruited
through communitywide publicity to participate in a study on relationship development. The demographic characteristics at Time 1 were as follows: Average time that the partners had known each other was 2Y2 years (range
= 4-48
months);
average age of women was
23 years (range = 18-31); average age of men was 24 years (range = 18-32); average years of education were 15l/2 (range = 12-18); and average personal income level was $10,500
(range
= $5,000-$20,000
or more). While there was some diversity in ethnicity,
most of the couples were white. Furthermore, 40 % of the couples were planning marriage and 60% were formally engaged, 80% of the couples were sexually active, and 39% were living together. Procedures.
There
have been eight procedural
phases: pre-assessment
(PRE),
interven-
tion, postassessment (Post), and 11/2- (FUl), 3- (FU2), 4- (FU3), 5- (FU4), and 6-year (FU5) follow-ups. The 7-year follow-up is almost complete, and the 8-year follow-up started in April 1989. The g-year follow-up started in July 1990. This article presents data up to the 6-year follow-up. Pre-Assessment. Couples participated in two 2-hr research sessions that were scheduled 1 week apart. During these sessions, couples were interviewed, completed a set of questionnaires, and participated in four 15-min videotaped problem-solving interaction tasks. The tasks were a discussion of a vignette from the Inventory of Marital Conflicts (Olson & Ryder, 1970) and a discussion of one of the couple’s top three relationship,problem areas. Each of these tasks was completed twice, once using a procedure to rate each other’s communication and once without this procedure, in a randomly determined order. After the initial sessions, a subset of couples were randomly assigned to a couples’ communication program (see Markman, Floyd, Stanley, & Storaasli, 1988, for details). The interven-
Predicttimand Prevention
tion couples are included in this article,
even though their developmental
somewhat
different
due to the intervention.
population
some couples do take advantage of premarital
tional analyses, comparing reveal similar patterns.
33
This
reflects
course may be
the fact that in the general
intervention
the results with and without intervention
programs.
Correla-
couples,
generally
Post-Assessment and Follow-Ups 1, 2, 3, 4, and 5. Six to eight weeks after the preassessment
phase, all couples participated
in a postassessment
session (Post), during which
they were readministered the questionnaires and engaged in another series of problemsolving interactions. At approximately yearly intervals, couples participated in follow-up sessions.
Couples
who had moved or who did not want to return to the research
were asked to complete the questionnaire battery at home and return procedures were essentially the same at each follow-up.
Measures. This section describes the self-report and communication
center
it by mail.
measures
The
relevant to
the current article. A full description of the measures for the entire study is presented elsewhere (Markman, Duncan, Storaasli, & Howes, 1987). These inventories were completed independently
by both partners at each assessment
point.
Personal History /?wm. This form assesses demographic including age, ethnic background, orientations, income, education, child care arrangements, birth activity, and contraception
socioeconomic
and miscellaneous
status (SES),
employment
information status, career
occupation, reasons for having or not having children, date of children, religion, therapy experience, sexual
use.
Marital Adjustment Test (MAT, Locke & Wallace, 1959). The MAT is a 15-item
self-
report measure covering such domains of marital functioning as disagreement, communication, leisure-time activities, and regrets about having married their spouse. The MAT is one of the most widely used measures of marital adjustment in the field, and studies have shown it to possess high levels of reliability and validity. A premarital version was used at Time
1 (Markman,
1979).
Marital Agendas Protocol (MAI: Notarius & Venzetti, 1984). The MAP assesses intensity of each of 10 commonly as confidence demonstrated
experienced
Relationship Status. A central (married,
problem areas (e.g.,
money, sex, children),
in being able to solve problems (i.e., relationship efficacy). its reliability and validity (Notarius & Venzetti, 1984).
divorced,
outcome
measure
as well
Research
in the study is relationship
has
stability
separated).
The Couple Interaction Scoring System KISS, Gottman, 1979; Notarius & Markman, 1981). The CISS is a microanalytic behavioral observation system for couples that contains 36 verbal codes and 3 nonverbal codes used to code each unit of interaction (i.e., a grammatical clause). Each unit receives a verbal and a nonverbal code. As suggested by Notarius and Pellegrini (1987), the content and affect codes can be combined to form four summary codes: (problem-solving facilitation, problem-solving inhibition, emotional validation and emotional invalidation). Cohen’s kappas average .91 for content codes and .72 for nonverbal behavior codes (Gottman, 1979; Notarius & Markman, 1981). Generaliza-
34
H. J. Markman and K. Hahlweg
bility studies reveal both content and nonverbal alpha coefficients (Gottman, 1979).
behavior
codes to have high Cronbach
Results Relationship
Status. Of the original
135 couples,
103 married
before marriage (24%). Of the couples who married, have separated, and 1 spouse has died. Attrition.
The actual attrition
rate (i.e.,
subjects
(76%)
14 couples (13%)
and 32 broke up have divorced,
who decided not to participate)
3
is less
than 4% (5 couples). This leaves 83 intact couples who continue to participate in the study, 67 of whom come to the research center and 16 of whom are out of town and complete forms by mail. Compared to the other marital interaction studies where the average sample size is about 25 (Notarius
& Markman,
1989) the current
sample is still relatively
large. Relationship Satisfaction Over Time. Previous studies have indicated that relationship satisfaction is highest before marriage, gradually declines over time, and then rises after children leave home (Spanier,
Lewis, & Cole,
1975). Because
the majority
of studies are
cross-sectional, conclusions are limited. Our results reveal that relationship satisfaction was highest at pre-assessment (MAT M = 122) and declined at Follow-Up 1 (MAT M = 119) and Follow-Up 2 (MAT M = 113), but, in contrast to previous research, stabilized at Follow-Ups
3, 4, and 5 (see Figure 1).
One consistent
finding from the earliest studies to recent times is that husbands
are
Legend
130
m
Male Female
Post FIGURE
( Follow-u 2 1 Follow-u 4 ( Follow-up 6 Follow-up 1 Pollow-up 3 Pallow-up 5 I. Marital
Satisfaction
Over Time.
Prediction and Prevention
35
happier with marriage than wives (Gottman, 1979; Spanier et al., 1975), leading to the statement that marriage may be “better” for men than for women (Stroebe & Stroebe, 1983). As seen in Figure 1, in sharp contrast to previous research, wives reported significantly higher levels of marital satisfaction at every stage than did husbands. Our findings may mirror changes in society that have brought increased options for women and/or reflect the change in sex ratios that started in 1980, which has resulted in more available men than women. F. Pederson (personal communication, April 1990) suggested that the increased competition may be contributing to men “working harder” at marriage. Finally, 30% of our couples maintain or improve levels of marital satisfaction over time, suggesting that declines are not inevitable. In contrast to prevailing beliefs, we suggest that feelings of intimacy
and closeness
negative interactions Communication
do not passively
(Markman
and Marital
& Kraft,
decay over time,
but are actively
eroded
by
1989).
Distress
Overview. In this section we present results on the role of premarital communication quality in the development of divorce and marital distress. The results of univariate and multivariate
analyses of variance are presented with group (distressed
and gender as the independent
variables,
and premarital
indicators
and nondistressed) of relationship
func-
tioning as the dependent variables. Parallel analyses covarying premarital satisfaction yielded the same general pattern of results and so are not presented here. Couples were defined as distressed if one partner scored below 100 on the MAT at FU5 and/or if they were separated or divorced, and as nondistressed if both partners scored above 100 on the MAT at each follow-up point. This definition resulted in 36 distressed and 40 nondistresed couples.
The average MAT scores were 91 and 125 for the distressed
and nondistressed
groups, respectively. In this article we focus on the results of the problem discussion task because it represents the most ecologically valid assessment of couples’ interaction we have available. We compared the premarital interaction of the two groups using the new CBS summary codes. The results are presented in Table 1. As seen in Table 1, there were significant differences
in problem-solving
facilitation
and in problem-solving
inhibition-but
only for
TABLE 1. CISS Predictors of Distress Premarital
Distressed
Nondistressed
F
Communication Variables
(CISS)
Husband
Wife
Husband
Wife
Husband
Wife
Problem-Solving Facilitation
60.6
63.7
70.6
68.0
10.8**
2.5
Problem-Solving
11.7
7.1
5.0
5.0
9.2**
2.7*
Emotional Validation
18.3
19.2
18.8
21.6
Emotional
9.5
10.0
5.6
5.5
Inhibition
Invalidation Note. NS = nonsignificant. *p < .lO. **p < .Ol. ***p
< ,001.
NS
2.0
8.8**
12.0***
36
H. J. Markman
the males.
and
Males who were destined to become
K. Hahlwec
distressed
or divorced had significantly
lower levels of problem-solving facilitation and significantly higher levels of problemsolving inhibition compared to the males who were destined to remain happy. It is noteworthy that only the male interactional behaviors predicted outcome for these two summary categories. These summary categories can be conceptualized as codes that have to do with conflict management.
One tentative conclusion
is that how males handle conflict in a
relationship is more important in terms of predicting the future than how females handle conflict. This fits emerging data on differences between men and women in distressed relationships (Baucom, Notarius, Burnett, & Haefner, 1990; Christensen & Shenk, 1991; Gottman & Krokoff, 1989; Markman & Kraft, 1989). There were significant differences for both males and females for the invalidation summary
codes. Couples
who were destined
higher levels of invalidation
to go on to be distressed
during their premarital
interaction
or divorced
had
than couples who were
destined to remain nondistressed. Contrary to our initial expectations and most theories of marriage, there were no differences, however, on the validation code. Thus, it is not the amount of validation received that seems to be critical for the future but instead, invalidation. When we asked couples in seminars and studies what their major goal was in marrying, they emphasized the importance of having someone who will listen to and understand them-in other words, to validate! When expecting validation, however, for a partner to be actively criticized and put down is extremely arousing and provocative, can generate major arguments and be costly to the relationship over time.
and
In sum, the results show that we can predict marital distress and divorce from examining premarital interaction; and the best predictors of divorce (male withdrawal from conflict, negative affect escalation, conflict) appear to be related to how well couples (especially
men)
regulate
and express
negative
affect when they disagree.
The
reults
suggest that couples who go on to develop marital problems and divorce show patterns of dysfunctional affect regulation while trying to resolve relationship issues. These
findings
echo earlier
findings
from the current
study and other
reports
in
progress indicating that quality of the premarital communication is one of the best predictors of future outcomes for a relationship. Thus it is not the differences between people so much that matters, The Prevention
but how those differences
and Relationship
are handled.
Enhancement
Program (PREP)
Building on the empirical foundation of longitudinal approach to helping couples prevent marital problems
research, we have developed an by teaching them how to handle
differences through effective communication and conflict management. Although there is evidence to support the affectiveness of the PREP program with a U.S. sample (see Markman et al., 1988; Markman, Renick, Floyd, Stanley, & Clements, in press; Renick, Blumberg, & Markman, 1992), there are numerous design problems with this study, including a high rate of couples declining the initial offer of a PREP intervention and the relatively small sample size and relatively homogeneous population in the study. A study was therefore undertaken in Germany in part to address some of the limitations of this initial study.
THE GERMAN PREMARITAL PREVENTION STUDY (EPL: EIN PARTNERSCHAFTLICHES LERNPROCRAMM) The rationale for divorce prevention
programs
in Germany
is similar to that in the United
States. The divorce rate is above 33% and is expected to increase with unification,
since
Predzhhn and Prevention
the divorce rate in East Germany 1980).
The demand
for marital
was considerably
therapy
37
higher
is equally increasing,
(Duss-von
Werdt & Fuchs,
as evidenced
by the long
waiting lists of marital counselling agencies. At the same time, marriage and the family is highly valued, since over 90% of the population do strongly believe in marriage and in family life (Kocher, 1985). In Germany premarital
counselling
is primarily
done by the Catholic
church because
some form of counselling is required before the couple is allowed to marry in the church. The counselling varies from talking 15 minutes to a priest up to participating in a weeklong activity. The content of the programs varies accordingly, and the programs are delivered by leaders/teachers from many different professions, and notably by lay priests. Many of the enrichment programs developed so far focus on current and future religious practices,
highlight
areas of agreement
and disagreement
concerning
intimate
rela-
tionships and expectations for marriage, and enlighten couples concerning current and potential future personality problems. The mode of transmission is often group discussions, lectures by experienced couples, or couples’ discussions without much direct guidance by the workshop leaders or modification of the couples’ communication process. To our knowledge,
none of the programs
has been evaluated longitudinally
with regard to its
effectiveness in preventing marital distress. For the purposes of this article, we will first describe the development of a culturally relevant marital distress prevention program for German couples, the EPL. Next we will present the short-term (pre-post) results of the EPL, when compared to a control group using data from self-report measures and from behavioral observations of couples discussing a conflictual
partnership
problem.
Development of EPL In 1985/86 we (i.e.,
the Munich
group’ in collaboration
with the first author) modified the
PREP in order to satisfy the needs of German couples and to respect the financial restrictions of the typical Catholic prevention program. For example, it was impossible for a routine application of EPL to use more than two trainers for a group of four couples. EPL can be delivered either in six weekly sessions of about 2-hr duration or at a weekend meeting. Group sizes vary from three to live couples with two trainers for each group. The couples meet as a group for the lecture portions of the sessions but meet alone and work with a trainer for all other aspects of each session. At the end of each session, a homework assignment is given. Below, each session is briefly reviewed.
Session 1. The first two sessions deal primarily with communications
skills training.
After
the introduction, a lecture is delivered covering two areas: (a) the philosophy and goals of the program and its theological relevance, and (b) the importance of good communication skills for marriage. The goals are teaching communication and problem-solving skills and helping the couple to apply these newly learned skills to discuss relationship and sexual issues and their theology, using the wedding ceremony
expectations
as a starting point.
‘The Munich group is K. Hahlweg, TU Braunschweig, Institut fur Psychologie, Spielmannstr. 12a, 3300 Braunschweig, Germany; F. Thurmaier, J. Eckert, and J. Engel, Institut fur Forschung und Ausbildung in Kommunikationstherapie e.V., Ruckertstr. 9,800O Munchen 2, Germany; and H. J. Markman, University of Denver, Center for Marital and Family Studies, Department of Psychology, Denver, Colorado 80208, USA.
38
H. J. Markman and K. Hahlweg
Session 2. The second session is devoted to the expression way the partner should respond to such a self-disclosure. couples practice these skills alone. Session 3. In this session a structured live steps: 1. Both partners
problem-solving
of negative feelings and to the During 75% of the time, the
scheme is introduced,
convey their views of the conflict,
disclosing
consisting
their feelings,
and attitudes toward the problem, and give mutual feedback. communication skills learned in the first two sessions.
of
thoughts,
In sum, they use the
2. Each partner describes his needs and wishes regarding the problem. 3. The couple tries to generate specific solutions to the conflict in a brainstorming fashion. 4. Every solution generated is discussed regarding its positive or negative consequences when applied, and the best solutions are selected. 5. The couple plans how to implement the solutions into everyday life. In the next session implementation
is reviewed, and solutions are changed according
to the need
of the couple. Session
4. This session begins with a lecture on the concept of expectations
and the effects
they can have on a relationship. The remainder of the session is spent having the couples discuss some of their expectations regarding their marriage with the help of the trainer and using the communication
skills.
Session 5. This session is devoted to communication
in the sexual domain,
stressing the
fact that ineffective or nonexistent communication about sexual needs and wishes may lead to sexual dissatisfaction. Each couple has to rank order cards with a concept printed on it such as happiness, fantiy, creativity, body contact, orgasm, relaxation, tenderness, love, and reproduction. Couples can also decide to build a house using the cards. Cards at the basement are more important than cards (concepts) further up. The aim is to discuss the meaning of the concepts for the individual and the marriage. Session 6 (When the EPL Is Used With Catholic Couples). The theme of Session 6 is the meaning of Christian marriage for the couple. Using the wedding ceremony (Trauritus) as a starting point, the couple discusses the relevance and for their future relationship.
of Christian
marriage
for each other
Method Overview. The main aim of the study was to evaluate the effectiveness of EPL in a randomized trial, contrasting EPL couples with a control group of couples receiving a conventional Catholic premarital enrichment program. However, contrary to our expectations and despite our efforts, we could not recruit couples and randomize them to the experimental or control condition because couples in the Catholic church book a specific course advertised in a booklet that is published annually. Recruitment of couples using other channels was not successful. Therefore, our trial was quasi-experimental in design. We were able to recruit 77 EPL couples from Munich or from the rural area around Munich. EPL was received weekly by 34%) and 66% received it on a weekend. The control group consisted of two different kinds of couples: 18 couples participating
Prediction and Preuention
in a conventional advertisement
marital
enrichment
program
39
and 14 couples
who responded
to an
in a newspaper and who received no marital preparation.
Assessment
points were pre and post the intervention.
way, and a 3-year follow-up is planned. Assessments and direct observation of the couple’s communication
A il/z year follow-up is under-
included a variety of questionnaires behavior when discussing a relation-
ship problem. Subjects. Partners
had a mean age of 28 years. Eighty percent
were Catholic;
of those,
38% regarded themselves as religious, with 40% regarding themselves as not religious. They had known each other for about 4 years, and 58 % were living together. Eighty-eight percent were certain to marry each other in the 2).
nmrfutum, on average
in 5 months (see Table
Couples in the EPL and the two control groups did not differ in the socioeconomic
relationship
variables,
except that EPL couples had a slightly better education
and
level than
control couples. There were no differences between the two control groups. When compared with the couples in the PREP experiment, our couples were older (28 vs. 24 years) and dating longer (4 vs. 2’/2 years). As in the Markman were generally middle class and white.
Procedures
Procedures.
were matched
study, our couples
as closely as possible to the original Markman
et
al. study as described earlier.
TABLE
2. Pre-Post
KPI
Relative
Frequencies
EPL
for EPL
and Control
Group F
Control
KPI Variables
Pre
Post
Pre
Post
Self-Disclosure
13.7
14.5
11.0
10.1
2.3
3.6
2.0
1.4
NS
NS NS NS NS NS NS ** *
Positive Solution Acceptance
of Partner
Agreement Criticism Negative
Solution
Justification
3.4
10.2
2.6
1.8
13.7
17.8
15.6
16.4
2.2
4.0
1.6
1.1
1.1
0.7
0.8
1.0
2.2
1.3
2.2
0.8
Disagreement
14.2
6.3
12.3
11.8
Problem
34.0
30.2
39.8
43.0
Description
Metacommunication
1.8
2.3
1.1
1.7
Positive Nonverbal
80.4
90.7
79.6
79.7
Negative
11.1
2.9
5.0
2.9
Nonverbal
Pre
Post
*
** *** ****
NS **
NS NS NS NS **** **** NS **** NS
Note. After the training, EPL couples showed highly significant improvements in their communication behavior, specifically an increase in “self-disclosure,” “partner acceptance,” and “positive solution,” and a decrease in “disagreement” when compared to the control group. Control couples emitted significantly more “problem description” than the EPL couples. Furthermore, in the nonverbal domain after the training, EPL couples were significantly more positive than couples in the control group. No sex differences on any variables emerged. NS = nonsignificant. *p < .lO. **p < .05. ***p < .Ol. ****p < .OOl.
40
H. J. Markman and K. Hahlweg
fre-Assessment. Couples participated in a 1.5-2-hr research session. During this session, partners were interviewed, completed a set of questionnaires independently, and were asked to discuss one of the couple’s top three relationship
problem areas.
Postassessment. Six to eight weeks later, all couples participated session, in which the same set of questionnaires were readministered discussion was videotaped.
in a similar research and another problem
In the follow-up sessions underway or to come, the procedures
will be essentially the same. Measures 1. Marital Adjustment 7&t (MAT).
See Method section.
2. Problem inventory. This 18-item inventory assesses each partner’s perceived level of conflict in different areas of the relationship (e.g., money, in-laws, sex, communication). A high relationship.
score
for an individual
indicates
a high level of conflict
in the
3. KPI (Hahlweg et al., 1984). The couples were asked to discuss one of the couple’s three top relationship problem areas. The behavior of both partners during the videotaped interactions
was coded using the KPI.
The aim of the KPI is to assess empirically
the
speaker and listener skills that form the basis of behaviorally oriented communication and problem-solving training. The basic coding unit is a verbal response that is homogenous in content without regard to its duration or syntactical structure. The KPI consists of 11 verbal and 3 nonverbal categories. Reliability has been sufficiently high for the verbal and nonverbal codes (kappa > 0.80). Tapes were coded by two independent raters who were blind to couple status and assessment time. Results Marital Quality. A rather unexpected finding emerged when the couples’ marital quality at pretest were analyzed. Using a cutoff score of 100 on the MAT, 17 % of the EPL couples rated their relationship as low in relationship quality, and in 14% one partner rated it as low, the other as high in relationship
quality. In the control group,
88% of the couples
rated their relationship as high, 12% as mixed in quality. One can assume that the motivation for participating in an EPL group was very different according to the respective relationship quality: Couples with high quality were looking for premarital enrichment, while couples with mixed or low relationship quality may have used the EPL as an aid to decide whether to stay together or to separate. In order to investigate the preventive effects of the EPL and to conduct the closest replication
as possible to the Markman
et al.
study, we will present data only on the 81 couples with high marital quality (53 EPL, 28 control group; mean MAT score = 123.1; SD = 12.0). Th ere were no significant differences between the EPL and the control group couples in the socioeconomic and partnership variables. Again, the two control groups did not differ significantly on any of the variables. analysis.
Therefore,
it seemed appropriate
to collapse both control groups for the further
Pre-Post Results. Marital quality and problem inten+. When analyzing the MAT and problem-intensity scores using analysis of covariance with prescores as covariate, no significant differences between EPL and control couples emerged. KPZ. After the training EPL couples showed higher significant improvements in their
Prediction and Prevention
communication .04),
41
behavior, specifically an increase in self-disclosure
acceptance
of partner
(F(1,
161) = 74.9,
(F(1,
161) = 6.4, p <
p < .OOl), and positive solution
(F(1,
161) = 9.4, p < .003), and a decrease in disagreement (F(1, 161) = 20.3, p < .OOl) and problem description (F(1, 161) = 17.2, p < .OOl) when compared to the control group. Furthermore, in the nonverbal domain after training, EPL couples were significantly more positive (F(1, 161) = 10.9, p < .OOl) than couples in the control group (see Table 2). When looking at the pre-post
results, the EPL seems to be very successful in teaching
the couples the necessary communication
skills.
Discussion While there were no significant differences between the EPL and the control group (both high in marital quality) in the subjective measure at post, highly significant differences were found with regard to couples’ interaction when discussing a relationship issue. In general, the EPL seems to be very effective in teaching couples to communicate better after the training.
These findings stress the importance
data simultaneously. By and large, the results parallel
of using subjective
those of the Markman
study and provide
evidence for cross-national effectiveness of PREP. To conclude, the EPL/PREP seems to be a very efficient program marital distress. mendations training.
Couples
seem to like the EPL very much,
and objective strong
for the prevention
as evidenced
of
by their recom-
to their friends and when looking at couples’ written statements
at the end of
We are eager to see the results of the 3rd-year follow-up.
Conclusions We have attempted to illustrate a model linking prevention research that we feel has the possibilities for making important contributions to a science of the prevention of marital distress. The next step in both programs of research is to (a) modify the content of the prevention program based on the new data revealed about the factors that predict marital distress and on the outcome of the prevention trials, and (b) to deal with issues of selection effects in interpreting
the results of prevention
trials. In both cases it may be that couples
who selected and completed the intervention may be those couples who are destined to have more successful communication and more successful relationships in the short and long run. One way for future research to deal with these selection effects is to recruit couples just for intervention purposes and then to assign to one or more prevention trials and then to a no-intervention control group. We discuss these issues and provide an example of such a study elsewhere (Markman et al., in press; Renick et al., 1992). While we await the long-term results of the two studies and others that are under way, we remain cautiously optimistic about the possibilities of preventing marital and family distress. At the same time, we need to be aware in disseminating these results to professionals and couples that we are not offering a “warranty” against all problems in the future (Weiss, 1991). The PREP/EPL approach can be applied to other more at-risk populations (e.g., couples at risk for depression and couples who are at risk for abuse or are mildly abusive) in the context of a prediction and prevention design to the extent that marital distress is a risk factor for these outcomes. Moreover, we need to assess the indirect and radiating effects of preventive interventions such as PREP/EPL to other variables including child functioning, be done.
work functioning,
and health status. There is clearly a lot of work to
42
H. J. Markman and K. Hahlweg
- Preparation
Acknowledtemmts
of this manuscript
dation grant BNSD-9011848 The project ferenz,
is financed
by the Bavarian
and the Erzdiozose
Forschung
was supported
in part by National
Science
Foun-
to Howard J. Markman. Munchen
und Ausbildung
Ministry
for Social
und Freising.
The
in Kommunikationstherapie,
Affairs,
project
the Deutsche
is conducted
Munchen
Bischofskon-
by the Institut
fur
(V. Eckert and F. Thurmaier).
REFERENCES A.,
Balaguer,
& Markman, H. J.
(1992).
The effectr
ofmaritalcommunicationand con@
managementon child seljkonapt:
A lonsittldinalperspectiue.Manuscript submitted for publication.
Baucom, D., Notarius, C. I., Burnett, marriage.
In F. Fincham
C.,
& Haefner,
& T. Bradberg
(Eds.).
P. (1990).
Gender
differences
and sex role identity
in
The psychology ojmrrzizge=: B&c issues and applications (pp. 150-
17 1). New York: Guilford. Christensen,
A., & Shenk, J. L. (1991).
clinic, and divorcing Clements,
M. (1992). Watt,
Washington,
S., Ramey,
S., Markman,
H. J.,
and procedures. Paper
H. J. (1988).
& W. A. Goldberg
Cambridge
University
in nondistressed,
Long,
presented
B., & West,
S. (1991,
at the National
June).
Prevention
Preuetiion
Conference,
Werdt, J., &Fuchs,
Filsinger,
E. E. (1988). F. D.,
(Eds.),
Intervention
programs
and the transition
to parenthood.
The transitionto parenthood: Current theory and research.Cambridge,
In G. Y. England:
Press.
Duss-van Fincham,
distance
458-463.
D.C.
S. W., & Markman,
Michaels
and psychological
of Denver.
N., Hawkins,
research: Conceptual model oj strut&s Duncan,
conflict,
The input of maritalfunctioning on children’speer relations: An intnactional study. Unpublished
master’s thesis, University Coie, J.,
Communication,
of Consulting Cliniad Psychology, 59,
couples. Journal
Sch&ung in deI Schweiz. Bern,
A. (1980).
Switzerland:
Paul Haupt.
Biosocialperspectiveson thejvni(y. Beverly Hills, CA: Sage.
& Bradbury,
T. N. (1990).
The prycholosy of mm&se: Basic issues and applications. New York:
Guilford. Gottman,
J. M. (1979).
Marital inferaction:Empirical inuestigations.New York: Family Process Press.
Gottman,
J. M. (1983).
The zocialpsychophysiolo~ ojmarrk~e. Unpublished
Gottman,
J. M.,
&
Krokoff,
L. J, (1989).
Marital
interaction
manuscript.
and satisfaction:
A longitudinal
view. Journal oj
Consultingand Clinical Psychology, 57, 47-52. Gottman,
J. M.,
Markman,
H. J,,
& Notarius,
analysis of verbal and nonverbal Hahlweg,
K.,
behavioral
& Markman, techniques
C. I. (1977).
The topography
of marital
conflict:
A sequential
behavior. Journal ojMarriage and the Family, 39, 461-477.
H. J. (1988).
in preventing
The effectiveness
and alleviating
of behavioral
marital
marital
therapy:
Empirical
status of
distress. Journal oj Consulting and Clinical Psychology,
56, 440-447. Hahlweg,
K., Reisner,
L., Kohli,
G., Vollmer,
validity of a new system to analyse
M.,
interpersonal
Schindler,
L., & Revenstorf,
communication
(Eds.), Marital interaction:Analysis and modification(pp. 182-198). Kocher,
R.
(1985).
Demoskopie, Locke,
H. J,,
& Wallace,
D. (1984).
In K. Hahlweg
Development
and
& N. S. Jacobson
New York: Guilford. der Z&t. Unpublished
Einskllungen zur Ehe und Familk im W&l
Allensbach,
(KPI).
manuscript,
Institut
fur
Germany. K. M. (1959).
Short-term
marital
adjustment
and prediction
tests: Their
reliability
and validity. Marrt@e and Family Living, 21, 251-255. Markman,
H. J. (1979).
couples planning Markman,
Application
marriage.
H. J. (1981).
of a behavioral
model of marriage
in predicting
relationship
satisfaction
in
Journal of Consultingand Clinical Psycholo~, 47, 743-749.
Prediction
of marital
distress:
A five-year
follow-up. Journal of Consulting and Clinical
& Howes, P. (1987).
The prediction and prevention of marital Understandingmajor nmfd disorders: The
Psychology, 49, 760-762. Markman, distress:
H. J., Duncan, A longitudinal
S. W., Storaasli, investigation.
R.,
In K. Hahlweg
contributionoffam+ interactionresearch(pp. 266-289). Markman,
H. J.,
longitudinal Markman,
Floyd,
investigation.
H. J., & Kraft,
F., Stanley,
S. M.,
& M. Goldstein,
New York: Family Process Press. R. (1988). The prevention
& Storaasli,
Journal ojConsulting and Clinical Psych&~, S. (1989).
Men and women in marriage:
of marital
distresss:
A
56, 210-217. Dealing
with gender differences
in marital
therapy. Behavior Thmapirt, 12, 51-56. Markman,
H. J,,
distress through
Renick,
M. J.,
communication
Consultingand CliniGalPsychology.
Floyd,
F. J.,
Stanley,
S. M.,
and conflict management
& Clements,
training:
M. (in press). Preventing
marital
A four and five year follow-up. Journal of
Prediction and Prevention
National
Center
for Health
Statistics.
(1989).
Births,
marriage,
43
divorces,
and deaths
1988. Monthly
for October
vital statistics Report, 37(10). Notarius,
C. I.,
Lewis
(Eds.),
Notarius,
& Markman,
H. J. (1981).
The
Couples’
C. I., & Markman,
Notarius,
H. J. (1989).
C. I., & Pellegrini,
standing
Coding
marital
D. P. (1987).
discord.
In K.
Beverly
marital
Differences
Hahlweg
& M.
contributionoffamily interactionresearch(pp. 231-239). Notarius,
C. I., & Venzetti,
arsmmm (pp. D. K.,
209-227). & Ryder,
N. (1984). Beverly
R.
M. J.,
Program
Blumberg,
(PREP):
Hills,
interaction:
The
marital
CA:
Sage.
and Family Liuq,
S. L.,
& Markman,
An empirically
based
between
System. Hills,
In E. Filsinger
& R.
CA: Sage.
A sampling
husbands
Goldstein
(Eds.),
New York: Family
I nventory
G. (1970).
Journal ofMum+
procedure. Renick,
Scoring
117-136).
and discussion
of current
BehavioralAssmmmt, 11, 1- 11.
issues.
Olson,
Interaction
manziqe: New behavzimdapprouha (pp.
Azsersiq
agenda of Marital
and wives:
Process
protocol.
Implications
Um&rstmding mqbr
The
News.
In E. Filsinger
Conflicts
for underdismks:
menfd
(IMC):
(Ed.),
Marital andfamib
An experimental
interaction
32, 433-448. H. J. (1992).
preventive
The
intervention
Prevention program
and
Relationship
for couples.
Enhancement
Farnib Relations, 41,
141-147. Spanier,
G. B., Lewis,
Storaasli,
R.
marriage: Stroebe,
R. A., & Cole,
D.,
& Markman,
A test of family
M.
C. L. (1975).
Marital
adjustment
over the family
life cycle:
The issue of
and
stages
Journal ofMarrarri0p-e and the Family, 37, 263-275.
curvilinearity.
S.,
& Stroebe,
H. J.
development W. (1983).
(1990).
Relationship
problems
in the
premarital
early
of
theory. Journal ofFamily Psychology, 2, 80-98. Who
suffers
more?
Sex differences
in health
risks
at the
annual
of the widowed.
PsychologicalBulletin, 93, 279-301. Weiss,
R.
Association
(1991,
November).
Preuention research: An overview. Paper
for the Advancement
of Behavior
Therapy,
New York.
presented
meeting
of the