Journal of Perconality and Social Psychology
1987, Vol. 52, No. 1.27-40
Copyright 1987 by the American Psychological Association, Inc.
Personality and Compatibility: A Prospective Analysis of
Marital Stability and Marital Satisfaction
University of Michigan
James J. Conley
The antecedents of marital stability (divorce or remaining married) and marital satisfaction (within
the group that remains married) were investigated with a panel of 300 couples who were followed
from their engagements in the 1930s until 1980. Twenty-two of the couples broke their engagements;
of the 278 couples who married, 50 got divorced at some time between 1935 and 1980. Personality
characteristics (measured by acquaintance ratings made in the 1930s) were important predictors of
both marital stability and marital satisfaction. The three aspects of personality most strongly related
to marital outcome were the neuroticism of the husband, the neuroticism of the wife, and the im-
pulse control of the husband. In combination, the 17 major antecedent variables were moderately
predictive of a criterion variable composed of both marital stability and marital satisfaction (ft =
.49). The three major aspects of personality accounted for more than half of the predictable variance.
The remaining variance was accounted for by attitudinal, social-environment, and sexual history
There are two major theoretical perspectives on the problem
of marital compatibility (Doherty & Jacobson, 1982). The first
perspective proposes intrapersonal causes of compatibility and
is most popular among psychoanalysts and trait theorists. The
personality characteristics of the two marriage partners make
the match stable and mutually satisfying or unstable and
fraught with discontent. Neuroticism is the personality trait
that has been most often identified as a source of marital insta-
bility (Barry, 1970; Doherty & Jacobson, 1982; Terman & But-
tenweiser, 1935; Zaleski & Galkowska, 1978). The neuroticism
hypothesis is consistent with a common-sense analysis of mari-
tal compatibility. As Terman and Buttenweiser (1935) observed,
"One would hardly expect a man and woman, both highly neu-
rotic, to achieve a very high order of marital happiness" (p.
143). The second perspective concerns interpersonal rather than
intrapersonal sources of compatibility. This perspective is most
popular among behaviorally oriented marriage therapists.
Troubled marriages are seen as dysfunctional behavior ex-
changes that are characterized by high ratios of punishments to
rewards. The distressed spouses are seen as deficient in basic
social skills and more likely to react to and reciprocate the nega-
tive behavior of the partner.
These two perspectives on marital compatibility are not mu-
tually exclusive. In fact, the second (interpersonal) perspective
Various aspects of this study were funded by the National Research
Council, the National Psychiatric Research Fund, and the National In-
stitute on Aging (Grant 02837).
We wish to thank Lisa Danford and other research assistants who
have been affiliated with the study since it began in 1935. Our special
thanks go to the participants in this study. Their cooperation over 5
decades was the single most essential ingredient in its success.
Correspondence concerning this article should be addressed to James
J. Conley, who is now at the Stonington Institute, Swantown Hill Road,
North Stonington, Connecticut 06359.
may be a description of the process by which the first (intraper-
sonal) perspective operates. High levels of neuroticism on the
part of either or both partners could result in the dysfunctional
behavior exchanges noted in troubled couples. Obviously, other
factors (both personal and situational) must enter into the devel-
opment of marital distress, but the neuroticism hypothesis goes
some distance toward integrating the most important concur-
rent and longitudinal findings on marital compatibility. Dis-
tressed couples use more coercive control techniques (Billings,
1979; Gray-Little & Burks, 1983; Margolin &Wampold, 1981)
and make more misinterpretations of basic communications
(Noller, 1980) than do nondistressed couples. These differences
may be attributed to mutually high levels of physiological
arousal during disagreements among the partners. This physio-
logical linkage acts to lock the partners into the expression and
exchange of negative affects (Levenson & Gottman, 1983). Neu-
roticism is primarily physiological overreactivity to stressful en-
vironmental stimuli (Eysenck, 1967). High neuroticism on the
part of one or both partners in a marriage makes it likely that
conflictual physiological linkage will take place and that dys-
functional cognitive and behavioral patterns will be established.
The data from studies that are currently available do not per-
mit a direct test of the neuroticism theory. A direct test would
require physiological measures of emotional reactivity, which
should be administered longitudinally to address the question
of whether neuroticism has a causal role in marital distress.
Physiological measures of emotional reactivity have only re-
cently been used in the study of marital interaction (Levenson
& Gottman, 1983). There are, however, seven longitudinal stud-
ies of marital compatibility. Most of these have included self-
report measures of neuroticism. This is certainly not the ideal
form of measurement for personality characteristics, but the
results of these studies are rather promising for the neuroticism
hypothesis of marital distress.
Adams (1946) reported on the personality-inventory scores
E. LOWELL KELLY AND JAMES J. CONLEY
of the members of 100 married couples. The personality inven-
tories were administered before marriage, and self-report mea-
sures of marital adjustment were taken 2 to 3 years after mar-
riage. Emotional instability and irritability before marriage
were predictive of low marital-adjustment scores. Terman and
Oden (1947) found that a self-report measure of emotional sta-
bility administered to their gifted subjects when they were 7 to
14 years of age was related to marital happiness 18 years later.
The relation was statistically significant for both sexes, and the
combined biserial correlation was .25. In 1940, Terman admin-
istered to his gifted subjects a marital-aptitude test that con-
sisted mainly of neuroticism items but also included material
on social and family background. The concurrent correlation
of marital aptitude with marital happiness was .62, and the
great bulk of this relation was accounted for by the neuroticism
items ("The effects of childhood and family background are
nearly all contained in the personality items themselves," con-
cluded Terman and Oden, 1947, p. 258). The members of cou-
ples that divorced between 1940 and 1946 had marital-aptitude
scores from 1940 that were a full standard deviation lower than
those of the members of couples that remained married. Sears
(1977) related the 1940 marital-aptitude test scores with mari-
tal outcome as of 1972 (when the average age of the subjects
was 62). The marital-aptitude scores of both sexes significantly
differentiated broken marriages from unbroken marriages (the
point-biserial correlation was .28 for the wives and . 12 for the
Burgess and Wallin (1953) reported correlations of .25 for
men and . 18 for women between scores on the Thurstone Neu-
rotic Inventory before marriage and marital-adjustment scores
after 3 to 5 years of marriage. Uhr (1957) analyzed a portion of
the data of the Kelly Longitudinal Study (KLS). Neuroticism
scores on the Bemreuter Personality Inventory taken before
marriage were significantly related to men's, but not women's,
marital happiness 18 years after marriage. Vaillant (1978)
found a strong relation between psychopathology and poor so-
cial adjustment, on the one hand, and low levels of marital ad-
justment on the other hand. Antecedents of poor marital adjust-
ment included poor social adjustment and fearfulness in col-
lege. Bentler and Newcomb (1978) followed 77 newlywed
couples for a period of 4 years. The emotional stability and ob-
jectivity of women and the deliberateness and introversion of
men were predictive of a composite marital-adjustment score
that was administered to couples who were still married and to
those who had divorced. Markman (1979, 1981) followed a
small group of couples from before marriage through the fifth
year of marriage. Unrewarding communication patterns during
the premarital period were predictive of low levels of marital
All seven of the longitudinal studies of marital compatibility
found some predictive role of neuroticism or disturbed commu-
nication patterns. The findings of Bentler and Newcomb (1978)
indicate that an additional predictive factor may be the impulse
control or conscientiousness of the male member of the couple.
This is consistent with nonlongitudinal studies by Terman, But-
tenweiser, Ferguson, Johnson, and Wilson (1938) and Johnson
and Harris (1980) that found that divorced men were character-
ized by more impulsive behavior than were men with stable
marriages. Terman et al. (1938) and Burgess and Wallin (1953)
reported the impulse-related characteristics of irregular em-
ployment, aggressive and domineering social demeanor, and
high levels of premarital sexual activity to be associated with
marital instability in men. Thus the set of probable personality
predictors of negative marital outcome seems likely to include
low impulse control in the husband as well as high neuroticism
in both members of the couple.
Although the existing longitudinal studies indicate that per-
sonality characteristics are predictive of marital compatibility,
there are serious methodological problems both on the side of
the predictors and on the side of the criterion. The two major
problems involving the predictor variables are the use of self-
report measures of personality characteristics and the embed-
ding of personality measures in composite predictive indexes.
Self-report methods of assessing personality characteristics are
a serious problem when the criterion is marital happiness or
marital satisfaction, which invariably are also assessed by way
of self-reports. The correlation of self-reported neuroticism and
marital satisfaction may be attributable to social desirability or
other biases of self-presentation. When personality variables are
analyzed only as part of composite predictive indexes, it is
difficult to determine the contribution of antecedent personal-
ity variables relative to social background and other compo-
nents of the indexes. This is a serious problem in Terman's
Study of the Gifted and also affects most of the analyses of Bur-
gess and Wallin. Terman initiated the practice of constructing
predictive indexes for use in applied settings, especially mar-
riage counseling, but this tends to obscure the interesting theo-
retical problem of whether personality characteristics or social
experience has the greater predictive relation with marital com-
The criterion of marital compatibility is also beset by two
major methodological problems. The first is whether to use
marital stability (whether the marriage continued or resulted in
divorce) or marital satisfaction (typically restricted to those
who remain married) as the measure of marital compatibility.
Both stability and satisfaction are useful measures of marital
compatibility, and ideally, longitudinal studies would use both
of them. This has, unfortunately, not been the case for earlier
studies. Adams, Burgess and Wallin, Uhr, and Markman used
only marital satisfaction and restricted their analyses to intact
marriages. In his analysis of the data of the Terman study of the
gifted, Sears (1977) made the opposite choice, using only mari-
tal stability as the criterion of marital compatibility. Obviously,
either of these two approaches applied in isolation is likely to
underestimate the relation of personality and marital compati-
bility, as personality is likely to operate both on the relative com-
patibility of intact marriages and as a contributing cause of
The second methodological problem in measuring marital
compatibility is the restriction to relatively brief intervals after
marriage. Most of the longitudinal studies examine only the
first 5 or fewer years after marriage. The Vaillant study, Ter-
man's Study of the Gifted, and the KLS are the exceptions in
this regard. These studies gathered data on marital outcome
over 2 or more decades. Most divorces occur after the first 5
years of marriage, and it is quite possible that the causal factors
of instability and dissatisfaction are different in early and in
mature marriages. Interestingly enough, these differences have
PERSONALITY AND COMPATIBILITY 39
Items Derived From "Views About the Ideal Marriage'* Questionnaire
Favors companionate marriage
1. That the husband and wife should have similar intellectual interests, such as scientific, literary, musical, etc.?
2. That the husband and wife should engage in the same outdoor sports (e.g., golf, hiking, swimming, etc.)?
3. That the husband and wife should like the same types of amusements (cards, dancing, theater, etc.)?
4. That husband and wife should be about equally fond of social gatherings?
Favors conventional order
1. That children should be given religious instruction?
2. That the wife should be allowed a definite budget for the household and for her personal expenditures?
3. That husband and wife should each respect the other's religious, political, or ethical convictions and not strive to change them?
4. That children should be held to a strict discipline?
5. That the household affairs should be run in a neat and orderly manner?
Favors equality of partners
1. That the same standard of sexual morality should apply both to husband and wife?
2. That the wife should be kept fully informed of the family finances and of her husband's business?
3. That the father should take an active interest in the discipline and training of the children?
4. That husband and wife should be well-mated sexually?
Opposes premarital sex
1. That young people should be trained never to indulge in "petting" and "spooning"?
2. That husband and wife should not have had sexual intercourse with each other before marriage?
3. That the wife should not have had sexual intercourse with any other man before marriage?
4. That the husband should not have had sexual intercourse with any other woman before marriage?
Favors absolute sexual fidelity
1. That after marriage the wife should be 100 percent faithful to her husband in regard to sex?
2. That after marriage the husband should be 100 percent faithful to his wife in regard to sex?
Stressful Life Events Derived From Part A, 1954-55 Report on Marriage to Research Partner
1. Living with relatives in their home
2. Relatives living in home
3. Death of someone very close and dear to respondent
4. Respondent's ill health or injury
5. Ill health or injury of spouse
6. Ill health or injury of a close relative
7. Respondent's emotional illness
8. Emotional illness of spouse
9. Emotional illness of a close relative
10. Respondent's excessive use of alcohol
11. Excessive use of alcohol of spouse
12. Husband employed, but work not regular
13. Husband lost his job
14. Husband unemployed for several months
15. Husband changed his job once or twice
16. Husband changed his job several times
17. Husband's job required considerable traveling
18. Husband's job required unusual working hours
19. Husband disliked his job
20. Wife worked for pay
21. Family moved from one community to another once or twice
22. Family moved from one community to another several times
23. Husband in military service
24. Inadequate living conditions
25. Serious difficulty in making ends meet financially
(Appendix continues on next page)
40 E. LOWELL KELLY AND JAMES J. CONLEY
Appendix B (continued)
26. Heavily in debt
27. Received financial aid from relatives or friends
28. Received financial aid from social or governmental agencies
29. Gave financial support to relatives or friends
30. Unplanned pregnancy
31. Miscarriage or stillbirth
32. Intentional termination of pregnancy
33. Continued unsuccessful efforts for wife to become pregnant
34. Informed by doctor that wife's pregnancy was unlikely or impossible
35. Wife's menopause (change of life)
36. Respondent had love affair with someone other than spouse
37. Spouse had love affair with someone other than respondent
38. Respondent had sexual relations with someone other than spouse
39. Spouse had sexual relations with someone other than respondent
Received June 22, 1984
Revision received December 31, 1984