The Marital Satisfaction Scale: Development of a Measure for Intervention Research

Author(s): Arthur J. Roach, Larry P. Frazier, Sharon R. Bowden

Source: Journal of Marriage and Family, Vol. 43, No. 3 (Aug., 1981), pp. 537-546
Published by: National Council on Family Relations
Stable URL: http://www.jstor.org/stable/351755
Accessed: 14/03/2010 17:55

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=ncfr.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

National Council on Family Relations is collaborating with JSTOR to digitize, preserve and extend access to
Journal of Marriage and Family.

http://www.jstor.org
 The Marital Satisfaction Scale:
Development of a Measure
for Intervention Research*
ARTHUR J. ROACH**
Texas A. & M. University
LARRY P. FRAZIER***
Southern Arizona Mental Health Center
SHARON R. BOWDEN****
University of Arizona Student Counseling Services

This study reviews the development of the Marital Satisfaction Scale (MSS). The
MSS was designed to assess one's level of satisfaction toward his or her own mar-
riage. Marital satisfaction was defined as an attitude of greater or lesserfavorabil-
ity toward one's own marital relationship. Measurement of the perception of one's
marriage by means of an attitude scale was the focus of the effort. Design objec-
tives for developing the measure were to generate a new set of items, to utilize a
single-item style with an easy scoring system, to guard against contamination by
marital conventionalization and social desirability, and to provide items which
could reflect attitudinal change likely to occur as a result of marital intervention.
Research results indicate that the instrument has very high internal consistency,
sufficient test-retest reliability and validity, and a low degree of contamination with
social desirability.

This study is an attempt to develop a reliable
and valid scale designed to measure the de-
gree of favorableness of attitude the respon-
*Research on this paper was conducted from 1975 to
1980. Part of the research contributed by the second
author, Larry P. Frazier, was presented at the 23rd
annual convention of the Southwest Psychological Asso-
ciation, Albuquerque, New Mexico, 1976. Part of the
research was conducted in 1975 and 1976 as part of the
second author's dissertation, and part in 1976 and 1977
as part of the third author's dissertation.
**Department of Educational Psychology, Texas
A & M University, College Station, Texas 77843.
***Southern Arizona Mental Health Center, Tucson,
Arizona 85721.
****University of Arizona Student Counseling
Services, Tucson, Arizona 85721.
dent has toward his or her own marriage. For
purposes of this study, marital satisfaction is
defined as an attitude of greater or lesser fa-
vorability toward one's own marital relation-
ship. This study also attempts to provide an
ship. It is not attempt to assess the status or
quality of the marital relationship. This study
also attempts to provide an administratively
instrument based on a clear and consistent
conception of marital satisfaction.
It is hoped that the instrument will be
capable of reflecting changes in one's level of
marital satisfaction which may occur in
response to various types of helping inter-
ventions. The helping professional typically
seeks to modify behaviors, especially atti-
tudes, by means of relatively brief interven-
tions. There is a need for effective
instruments which can assess changes in the

August 1981 JOURNAL OF MARRIAGE AND THE FAMILY 537

level of attitudes in both pretest and post-
test evaluations. In marriage therapy, in par-
ticular, as well as in developmentally oriented
marriage communication or enhancement
workshops, effectiveness may best be mea-
sured in terms of change in attitude toward
one's own marriage.
NEED FOR THE STUDY
Although there might seem to be a plethora
of adequate instruments available as mea-
sures of marital satisfaction, actually such is
not the case. Straus (1969) surveyed 319
family-related instruments published prior to
1965, of which 34 would be likely to be used
in marriage counseling. Of these 34, only 14
have established reliability and validity, and
of these 14, only four are related conceptually
to marriage satisfaction. These four deal with
marital success (Bernard, 1933), marital
adjustment (Bowerman, 1957; Locke and
Wallace, 1959), and marital happiness
(Terman, 1938). Other surveys by Bonjean et
al. (1967), Lake et al. (1973), Phillips (1973),
Bowerman (1964) and Cromwell et aL (1976)
failed to list other measures dealing with
satisfaction with one's own marriage among
those instruments with acceptable validity
and reliability. Frazier (1976) surveyed the
best known and respected sources of test
information. He listed 30 separate measures
which would possibly be used in marriage
counseling. Only five of these tests reported
adequate data on reliability or validity. How-
ever, none of these measured marital satis-
faction, adjustment, or happiness.
In addition to this paucity of worthwhile
instruments, the related instruments which
do meet the requirement of validity and
reliability fall prey to the attack of conceptual
soundness. Lively (1969) deemed the con-
tinued use of the concepts of marital
happiness, success, and adjustment to be
detrimental to effective research. These
concepts suggest static states, levels of
achievement, or ultimate conditions, none of
which is a realistic conceptualization of the
interrelational dynamics of marriage. More-
over, the meanings of these concepts of
success, adjustment, or happiness, while
related, are dissimilar.
Historically, most studies dealing with the
quality of marriages have in actuality used
inventories of marital adjustment rather than
measures of attitude or satisfaction. The most
538 JOURNAL OF MARRIAGE AND THE FAMILY
notable and widely used of these inventories is
the Locke-Wallace Marital Adjustment In-
ventory (1959). Locke and Wallace put the
capstone on the lengthy effort to develop a
suitable, yet brief, instrument for such a
measure. They defined marital adjustment as
accommodation of a husband and wife to
each other at a given time, which suggests the
measurement of a dyadic process rather than
that of an individual's attitude. Experience
has shown, however, that it simply is not
possible to assess the state of a subject's
marriage from a self-report instrument, since
self-reporting is strongly contaminated with
the individual's perception of the state of his
marriage. One solution to this conceptual
bind is to measure the perception or attitude
toward the marriage relationship, rather than
to attempt to measure the quality of the
relationship or the dyadic adjustment.
Another weakness in traditional measures
of marital adjustment, happiness, success, or
satisfaction has been the problem of marital
conventionalization or social desirability, the
tendency of subjects to distort their responses
in the direction of looking good. Edmonds
(1967) and Crowne and Marlowe (1964) have
signaled the presence of social desirability
contamination and have seen it as a cause for
serious concern.
While the presence of social desirability
contamination in self-report instruments is a
fact, whether or not it should be cause for
concern is a moot question. Hawkins (1966)
demonstrated that social desirability ac-
counted for only a small part of the variance
and did not, therefore, preclude the use of the
Locke-Wallace scale. Murstein and Beck
(1972) did not find that marital convention-
alization was a major contaminating factor in
assessing marital adjustment. They con-
cluded that happily married couples do
exaggerate their spouses' qualities, but that
this overstatement does not affect the
relationship of various perception scores to
marital adjustment.
In spite of this disagreement regarding the
ill effects of social desirability contamination,
it seems appropriate to circumvent the phe-
nomenon by careful construction of items
which do not have a strong or easily apparent
social desirability loading.
Existing measures of marital adjustment or
satisfaction lack various characteristics de-
sired in sound psychological tests. In this
August 1981
respect, the Locke-Wallace scale suffers from
the fact that it contains three different types
of item styles and involves a complicated,
weighted scoring system. From a test
construction standpoint, the avoidance of
such characteristics would enhance the
administrability of a test. There is also a
marked disparity in the weight of the items.
The first item is scored up to 35 points, ap-
proximately one fourth of the total possible
score.
Spanier's Dyadic Adjustment Scale (1976)
was an attempt to start afresh in measuring
marital adjustment. Although he utilized
items from all previous scales, his results
seem primarily to be a revision of the Locke-
Wallace scale and an improvement upon it.
The four factors derived from the Spanier
scale have greater conceptual clarity than
those derived from the Locke-Wallace scale.
Spanier's work was well done, and his instru-
ment deserves greater use than is apparent
from the literature. His instrument, however,
by his own admission, suffers from problems
of direction of wording and halo effects.
Practically, from a test-taking point of view,
there are six different item formats, although
scoring has been simplified relative to the
Locke-Wallace scale. From the point of view
of attitude measurement, Spanier's instru-
ment focuses heavily on estimates of fre-
quency and degrees of difference. Such esti-
mating may involve more cognitive and recall
processes than affective or attitudinal
responses.
Recently Snyder (1979) developed the
Marital Satisfaction Inventory, a multidi-
mensional measure which includes Edmonds'
Marital Conventionalization Scale (1967).
While drawing from the same item pool as
the Locke-Wallace and Spanier instruments,
this instrument avoids one of the difficulties
of previous marriage satisfaction measures by
taking a measure of social desirability directly
into account. By using a profile of scale
scores, the marital conventionalization scale
functions as a measure of truth telling, thus
affecting the clinical interpretation of the
remaining scales. However, the MSI may be
too lengthy for some evaluative research uses.
Also, some items involve historicity and recall
of events and patterns of behavior. Such
factual reporting items do not contain the
element of opinion or projection necessary in
the assessment of attitudes. Because of this,
August 1981 JOURNAL OF MARRIAGE AND THE FAMILY
Snyder's MSI may be measuring something
other than attitude toward one's own mar-
riage. Nevertheless, Snyder's instrument is
based on careful research and holds promise
for the future.
Existing measures of marital quality suffer
from an array of conceptual difficulties, con-
tamination with social desirability, and lack
of sufficient reliability and validity. The use
of behavioral measures has certainly done
much to offset such ills. The fact remains,
however, that in most studies of effectiveness
of marital intervention, present behavioral
approaches are impractical. Rather than
argue for the complete dismissal of these self-
report instruments, it seems more practical to
place greater emphasis on their careful
construction and evaluation and to establish
firmly their reliability and validity.
THE SCALE
Design
The concept of attitude toward one's own
marriage utilized in the present study repre-
sents an attempt to measure a changeable
perception or readiness to respond in the
individual, not an assessment of the state of
the individual's marriage. For purposes of
this study, marital satisfaction is defined as
the perception of one's marriage along a con-
tinuum of greater or lesser favorability at a
given point in time. Satisfaction is by defini-
tion an attitude, which like any perception, is
subject to change over time, and especially in
relation to significant life experiences.
Considering marital satisfaction as an atti-
tude is in contrast to past trends in the liter-
ature which attempted to measure marital
adjustment, success, or happiness as rela-
tively fixed properties or behavior states.
One objective in the development of the
MSS was to develop a single-factor measure
of favorability of attitude toward one's
marriage. The major previously established
measures of marital adjustment have at-
tempted to assess the dyadic relationship,
have drawn from essentially the same item
pool, and have yielded four or more factors.
Especially, these three characteristics are
found in the Locke-Wallace, Spanier, and
Snyder instruments. In each of these instru-
ments, there is one factor which seems related
to global satisfaction with the marriage,
although in some of these instruments, the
539
factor comprises so few items that reliability
of the measure is at risk. An objective of the
present study was to focus on developing an
effective measure of this single factor.
Orden and Bradburn (1968) found sep-
arate clusters of items for marital satisfaction
and marital tension or dissatisfaction and
concluded that marital happiness is a two-
dimensional structure. Their findings make it
clear that a measure of marital satisfaction
alone is not a sufficient predictor of marital
happiness, although it is a necessary one. The
present effort is geared only toward devel-
oping a measure of marital satisfaction,
leaving the adequate measurement of marital
tensions to further research.
Since the purpose of the study was to treat
marital satisfaction as an attitude, Edwards'
(1957) guidelines for constructing Likert-type
attitudinal scales were selected as a design
model. One important design consideration
affected by this selection is scale length.
Increasing the number of items will increase
reliability of the measure. A longer measure
also allows finer discriminations where data
may at least be assumed to be continuous.
Increasing scale length also enables sufficient
sampling of the pool of topics generally
considered to be related to marital satisfac-
tion. On the other hand, the feasibility of
using the scale in connection with typical
marital intervention efforts precludes undue
length. In the development of attitude scales,
the procedure outlined by Edwards is to begin
with a large selection of about 70 items and to
retain only those 20 or so which contribute
most effectively to the performance of the
total instrument.
A major set of design considerations relates
to item characteristics. Attitude scale items
are appropriately comprised of items which
evoke affect or opinion rather than cognition.
When each item in a scale evokes an affective
response toward some aspect of the attitu-
dinal object, the total scale cumulatively and
effectively samples the attitude or affectional
orientation toward the object. Traditional
marital satisfaction measures have included
some items which involve historical recall and
estimation of amounts and frequencies of
behavior. One design criterion for construc-
ting the MSS was to avoid cognitive or recall
items and to require that they deal with
opinion toward some aspect of one's
540 JOURNAL OF MARRIAGE AND THE FAMILY
marriage. An additional requirement was
that each item have the anticipated proba-
bility of evoking divided agreement and
disagreement from subjects in a normal
population. At the same time, following
Edwards' suggestion, items eliciting extreme
favorability and unfavorability would be pre-
ferred to more neutral items.
Other item-characteristic design consid-
erations required that each item be capable of
reflecting change between pretest and post-
test usage. A test item could not reflect
change if the same response were required on
both the pretest and the posttest. (For ex-
ample, an item which asked whether either
partner had experienced an affair, if an-
swered affirmatively on the first usage, could
not be answered differently in a posttest situ-
ation even if the respondent's attitude toward
his/her marriage had changed in a positive
direction.) Only where there is the possibility
that the posttest response can be different
from the pretest response can the test item
detect attitude change between two points in
time.
Another item design criterion was to select
items which prevented contamination with
social desirability or marital conventional-
ization as much as possible. Items should not
tend to evoke unreasonable responses in the
subject. Attributions of perfection to the
spouse or the relationship should be avoided.
An additional requirement was that the
items should be fresh and not drawn from the
traditional item pool used by Locke and
Wallace, Spanier, and Snyder. The intent
here was to emphasize the switch from
marital adjustment and concern with the
quality of the relationship toward the respon-
dent's attitude toward his marital relation-
ship by generating new items.
Yet another design characteristic was to
produce a scale which had a single-item style
throughout and which would have as simple a
scoring scheme as possible. Both these
requirements were met by adopting the
Likert-type attitude scale format.
D)evelopment
In keeping with the desired design charac-
teristics outlined above, items pertaining to
satisfaction with various aspects of marriage
were generated. Content for these items was
suggested by the literature on marriage rela-
August 1981
tionships and by the senior author's experi-
ence in marital therapy and marital enhance-
ment workshops.
Seventy-three suitable items were collected
and formed into an initial version of the scale,
which was originally entitled The Marital
Satisfaction Inventory (MSI). Response
categories were: strongly agree, agree, neu-
tral (undecided), disagree, and strongly
disagree. Items were phrased positively
(favorable toward marriage) and negatively
(unfavorable toward marriage) in approxi-
mately equal proportions. Three experienced
judges agreed completely on the favorability
or unfavorability of the items. These
statements were assembled in random order,
and a set of introductory instructions was
supplied in order to complete the inventory.
Scoring on each item ranged from 1-5, with 5
indicating the most favorable attitude toward
one's marriage and I the least favorable
attitude. For the 73 items, the maximum
possible score was 395 (Roach, 1975). The
development of the scale and the ensuing
pilot study were the work of the senior
author.
STUDIES BASED ON THE INVENTORY
The Pilot Study
An initial administration of the inventory
utilized 88 volunteer subjects, ranging in age
between 20 and 65, with the largest propor-
tion (44 percent) being in the 30-39 age
group. Two thirds of this initial group were
females, and the majority (80 percent) of
subjects were professionals in education.
About half of these subjects (48 percent) were
black. The subjects consisted in part of a
group of Texas school counselors, attending
their annual statewide conference, and
another group of students in counselor
education classes both on and off the campus
of Prairie View A & M University.
Total scores of this initial study group
ranged from 138 to 361. The mean was
282.46, and the standard deviation was
54.81. Eighty percent of the scores were indi-
cative of satisfaction (256 or above). The
results were skewed, with the majority of
scores being high and indicating satisfaction.
Item analysis of the data obtained in this
pilot administration indicated that all but
three items (7, 16, and 43) were correlated
with the total score within the .05 level of
August 1981 JOURNAL OF MARRIAGE AND THE FAMILY
confidence. Of these 70 significantly related
items, all but 10 were related within the .01
level of confidence.
Cronbach's alpha, which is similar to
KR20 as a measure of internal consistency
and which considers all possible split halves,
was .982. This measure provided an initial
indication of very high internal consistency.
A factor analysis of the data was conducted
using the 70 items surviving the item analysis.
Results indicated one dominant factor ac-
counting for 40.33 percent of a trace of 70.00
on a principal axis rotation. This factor had
57 items with loadings above a conservative
criterion of .50. Although 15 other factors
with eigenvalues above 1.0 were extracted
mathematically, only two factors had three
item loadings above the criterion level and
another two had only two item loadings above
criterion.
This initial study of the instrument indi-
cated that the items constituted a single-
factor scale. Of the 70 items subjected to
factor analysis, all were significantly related
to the single factor at the .01 level of confi-
dence, with the exception of four items.
Thus, including the three items failing the
item analysis, only seven of the original 73
items could be considered insufficiently
related to a single factor.
Results of this initial study indicated that,
in general, items correlated well with the total
scale, that there was a very high level of
internal consistency, and that the scale
involved a single factor. These findings justi-
fied further study of the instrument.
The Frazier Study
A subsequent investigation of the 70-item
MSI (omitting items 7, 16 and 43) was con-
ducted by Frazier (1976). He employed a
sample of 309 individuals, including the 88
in the previous study. The majority of the
sample were young white adults in their
twenties, college educated, and married 1-4
years with no children; 19 percent were black
and 8 percent were Mexican-American. The
majority had a family income of less than
$6,000. There were 139 males and 170 fe-
males included in this study.
Subjects were volunteers from a variety of
settings, mostly academic, in southeast and
central Texas. The largest segment consisted
of graduate students and faculty from the
541
Texas A & M University campus and from
the Bryan-College Station area.
Item discrimination was studied on the
basis of item-total score correlation. All items
were significant at or beyond the .01 level of
confidence, and 52 of the items had a dis-
crimination index above .50. The internal-
consistency reliability as calculated by Cron-
bach's alpha formula was .9699, which was
slightly less than in the initial study (.982),
but still quite high.
Subjects who had requested feedback on
their responses to the original testing were
asked to retake the MSI after a period of
three weeks, for the purpose of estimating
test-retest reliability. Twenty subjects re-
sponded a second time. The test-retest coef-
ficient was .76. This coefficient is probably a
conservative estimate, since some of the sub-
jects took as long as seven months to respond
a second time. The standard error of mea-
surement based on internal consistency reli-
ability was 7.01, and based on test-retest reli-
ability, it was 19.49.
To determine concurrent validity of the
MSI, the Marital Adjustment Test (MAT)
(Locke and Wallace, 1959) was used as a
criterion variable and administered at the
same time as the the MSI to the 221 subjects
who were not also in the pilot study. The con-
current validity coefficient was .7851, a rela-
tively high figure for this type of psychological
measure.
In order to assess contamination of MSI
scores with social desirability, 20 subjects
who responded to the retesting for deter-
mining test-retest reliability also responded
to two measures of social desirability, the
Marlowe-Crowne Social Desirability Scale
(SD) (Crowne and Marlowe, 1964) and the
Edmonds Marital Conventionalization Scale
(MC) (Edmonds, 1967). The correlation
between the MSI and the Marlowe-Crowne
SD was .33, which is not significant. This re-
sult suggests that the MSI is low on contami-
nation with social desirability. This compares
suitably with Hawkins' (1966) reported cor-
relation between the Locke-Wallace MAT
and the Marlowe-Crowne SD of .31 for hus-
bands and .37 for wives. In this instance,
Hawkins found that social desirability, while
significantly correlated with the MAT, did
not preclude the use of the MAT, since SD
accounted for only a small portion of the
variance.
542 JOURNAL OF MARRIAGE AND THE FAMILY
The correlation between the MSI and the
Edmonds MC scale was -.13 (sic). However,
this result seems too specious to warrant any
conclusions. The subsample of 20 respon-
dents may have been unique in regard to
marital conventionalization. Further and
careful sampling would be required before
supportable conclusions can be drawn. By
way of comparison, Edmonds (1967) reported
a correlation of .63 between his scale and the
Locke-Wallace MAT.
To check for sex bias in responses, an
analysis of variance was calculated. There
was no sex difference for scores on the MSI (F
ratio = 0.000, df: = 1,307, p = 1.000).
In a factor analysis of the MSI results, only
one factor, which accounted for 36 percent of
the variance, was significant. Although 14
factors with an eigenvalue of 1.0 were ex-
tracted, only this one had item loadings of .5
or better, the criterion used by Kimmel and
van der Veen (1974) in their factor analysis of
the MAT.
The Bowden Study
A third study was conducted by Bowden
(1977) to determine the validity of the MSI.
In order to assess concurrent validity, cri-
terion groups of satisfied and dissatisfied
couples were identified by peer ratings and by
professional marriage counselors. Thirty
married couples (15 satisfied and 15 dissat-
isfied) volunteered and completed the instru-
ments involved. Satisfied couples were mar-
ried for at least one year, and dissatisfied
couples were not divorced for more than one
year. The sample was predominantly 20-39
years of age, white, college educated, with
family incomes beyond $8,000 per year, while
30 percent had incomes of $20,000 or more
per year. Subjects in this study consisted of
couples in therapy with private practitioners
in Austin and College Station, Texas, and of
volunteer couples in the same geographic
location. None of these couples had partici-
pated in the pilot or Frazier studies.
The mean satisfaction score for the satis-
fied group was 306.27 with a standard
deviation of 22.33. The dissatisfied group had
a mean score of 221.30 with a standard devi-
ation of 30.91. The t-test for difference
between means was significant (t = 112.204,
df = 58, p < .0001).
A correlation of MSI scores and scores on
the Marriage Problem Checklist (MPC), an
August 1981
instrument for identifying marital problems
(Roach, 1977), was calculated as a measure
of discriminant validity. Mean scores on the
MPC were 4.10 for the satisfied group and
14.17 for the dissatisfied. Standard devi-
ations were 3.16 and 5.35, respectively. The
correlation coefficient was -.73. Thus, more
satisfied couples reported fewer problems,
and dissatisfied couples reported more
problems.
The Thompson Study
After an instrument is developed and its
reliability and validity are established, its
research practicality must also be demon-
strated. The MSI was designed to measure
change in marital satisfaction as a result of an
intervention such as marital therapy or a
marriage communication workshop. One in-
dication of its proper development should be
that it will function appropriately when used
as intended.
Thompson (1978) used the original 73-item
MSI to study change in the level of marital
satisfaction related to a sex therapy treatment
program. The principle component of the
treatment utilized a weekend sex therapy
workshop format with nine couples. Subjects
in this study consisted of volunteers in Corpus
Christi, Texas. They were predominantly
Anglo-American, well educated and middle
class, and ranged in age from 22 to 50 years
of age.
Thompson reported means and standard
deviations for MSI scores as follows: pretest,
M = 277.39, SD = 32.47; posttest M =
301.44, SD = 30.83; post-posttest M =
295.89, SD = 31.37. MSI scores showed sig-
nificant change between the pretests and
posttests, which were administered three
weeks apart (t = -2.86, df= 17, p<.01).
When a post-posttest was conducted eight
weeks later, there was no significant differ-
ence between MSI scores on the two posttests
(t = .76, df= 17, n.s.). However, there
continued to be a difference between the
pretest and post-posttest measures of marital
satisfaction (t = -2.39, df = 17, p < .05).
These results indicate that the MSI can
reflect the types of attitudinal change likely to
occur in a relatively brief marital interven-
tion, while not reflecting significant change
during a period of nonintervention.
August 1981 JOURNAL OF MARRIAGE AND THE FAMILY
Revised Forms
As part of his reliability study, Frazier
(1976) studied the internal consistency of the
MSI, eliminating 20 items with item-whole
score correlations of less than .50. Cron-
bach's alpha for this short form was .9713,
which was higher than for the total 73-item
instrument. An editorial check of these items
with whole-score correlations of less than .50
indicated that they did indeed fail to fit the
same conceptual pattern of items correlating
with the total score beyond .50. In the revised
form, these items have been eliminated. Four
other items have been eliminated, three
because they were quite similar to others, and
one because it seemed inconsistent with the
definition of marital satisfaction as an indi-
vidual perception. In addition, some of the
items have been slightly edited for greater
clarity and brevity. This shorter form is now
under study to determine its reliability and
validity. However, the preceding research on
which it is based strongly suggests that it is at
least of the same quality as its parent instru-
ment.
Frazier's study (1976) indicated that an
even briefer version of the scale would still
retain a high degree of the scale characteris-
tics. Edwards (1957) suggests that a scale
length of approximately 20 items is adequate.
Research with an even shorter form, Marital
Satisfaction Scale, Form B, involving only the
20 items with the highest item-whole score
correlations is also underway. Since this is the
scale length suggested by Edwards, it is
hoped that this form will have the greatest
practicability as a research measure.
TITLE CHANGE
Snyder's article introducing his Marital
Satisfaction Inventory (1979) appeared in
print while the present report was being pre-
pared. Up to that point, the instrument being
developed in the present study bore the same
title as Snyder's instrument. Although ap-
propriate copyright protection extended back
to 1975, the title of the revised, shorter form
of the scale developed in the present research
has been changed to avoid needless confu-
sion. As a result, the research studies
involved in developing the instrument now
called the Marital Satisfaction Scale have
potentially confusing references to the
543
Marital Satisfaction Inventory in their titles. SUMMARY AND CONCLUSIONS
It should be made clear that these references
are to the parent instrument of the Marital An instrument in the form of a Likert-type
Satisfaction Scale (MSS) and not to Snyder's attitudinal scale was developed for the
instrument. purpose of assessing marital satisfaction.
Items retained in the revised, shorter form, Originally entitled the Marital Satisfaction
now entitled the Marital Satisfaction Scale, Inventory, it was administered to a total of
are presented in Table 1. Coefficients of 369 subjects. Items had very high discrimin-
correlation with the whole-scale scores of the ation, and the instrument proved to be of
original, 73-item scale, which provided the sufficient reliability and of high internal con-
research base for their retention, are also sistency. The original items were found to
included in the table. constitute a single-factor scale which possesses
TABLE 1. WHOLE-SCORE CORRELATIONS OF ITEMS IN THE MARITAL SATISFACTION SCALE
Item Correlationa
1. I know what my spouse expects of me in our marriage. .518
2. My spouse could make things easier for me if he/she cared to. ..582
3. I worry a lot about my marriage. .687
4. If I could start over again, I would marry someone other than my present spouse. .662
5. I can always trust my spouse. .596
6. My life would seem empty without my marriage. .522
7. My marriage is too confining to suit me. .661
8. I feel that I am "in a rut" in my marriage. .714
9. I know where I stand with my spouse. .676
10. My marriage has a bad effect on my health. .633
11. I become upset, angry, or irritable because of things that occur in my marriage. .550
12. I feel competent and fully able to handle my marriage. .649
13. My present marriage is not one I would wish to remain in permanently. .622
14. 1 expect my marriage to give me increasing satisfaction the longer it continues. .559
15. I get discouraged trying to make my marriage work out. .747
16. I consider my marital situation to be as pleasant as it should be. .754
17. My marriage gives me more real personal satisfaction than anything else I do. .638
18. I think my marriage gets more difficult for me each year. .688
19. My spouse gets me badly flustered and jittery. .710
20. My spouse gives me sufficient opportunity to express my opinions. .626
21. 1 have made a success of my marriage so far. .717
22. My spouse regards me as an equal. .571
23. I must look outside my marriage for those things that make life worthwhile and interesting. .630
24. My spouse inspires me to do my best work. .635
25. My marriage has "smothered" my personality. .613
26. The future of my marriage looks promising to me. .816
27. I am really interested in my spouse. .714
28. I get along well with my spouse. .769
29. I am afraid of losing my spouse through divorce. .566
30. My spouse makes unfair demands on my free time. .565
31. My spouse seems unreasonable in his/her dealings with me. .613
32. My marriage helps me toward the goals I have set for myself. .718
33. My spouse is willing to make helpful improvements in our relationship. .688
34. My marriage suffers from disagreement concerning matters of recreation. .511
35. Demonstrations of affection by me and my spouse are mutually acceptable. .606
36. An unhappy sexual relationship is a drawback in my marriage. .522
37. My spouse and I agree on what is right and proper conduct. .513
38. My spouse and I do not share the same philosophy of life. .569
39. My spouse and I enjoy several mutually satisfying outside interests together. .535
40. I sometimes wish I had not married my present spouse. .676
41. My present marriage is definitely unhappy. .572
42. I look forward to sexual activity with my spouse with pleasant anticipation. .605
43. My spouse lacks respect for me. .691
44. I have definite difficulty confiding in my spouse. .644
45. Most of the time my spouse understands the way I feel. .638
46. My spouse does not listen to what I have to say. .608
47. I frequently enjoy pleasant conversations with my spouse. .559
48. I am definitely satisfied with my marriage. .792
aItem score correlations are with whole scores on the original 73-item MSI for 309 subjects in the Frazier study.
All correlations are significant beyond the .01 level of confidence.

544 JOURNAL OF MARRIAGE AND THE FAMILY August 1981

both concurrent and discriminant validity.
These findings indicate that the scale has
merit and deserves further investigation.
A shorter, revised form of the scale was
developed based on research involving the
original, 73-item scale. The resulting 48-item
scale is now entitled the Marital Satisfaction
Scale.
Additional study is now being devoted to
further establishing the reliability and va-
lidity of the instrument with different popu-
lation samples, especially since the validity
studies of the instrument to date have been
with nonrandom samples. Further study of
validity and of possible distortion by social
desirability is also being pursued, and a
number of correlational studies utilizing the
revised instrument are nearing completion.
An even briefer scale would be more de-
sirable and practical in field situations.
Development of an even shorter form (Form
B) is currently underway.
REFERENCES
Bernard, J.
1933 "An instrument for measurement of success in
marriage." Publication of the American Socio-
logical Society 27 (May):94-106.
Bonjean, C. N., R. J. Hill, and S. D. McLemore
1967 Sociological Measurement: An Inventory of
Scales and Indices. San Francisco:Chandler
Publishing Company.
Bowerman, C. E.
1957 "Adjustment in marriage: Overall and in
specific areas." Sociology and Social Research
42 (March-April):257-263.
1964 "Prediction studies." Pp. 215-246 in H. T.
Christensen (Ed.), Handbook of Marriage and
the Family. Chicago: Rand McNally and Com-
pany.
Bowden, S. R.
1977 An Assessment of the Validity of the Marital
Satisfaction Inventory. Doctoral dissertation,
Texas A & M University. Dissertation Ab-
stracts International, 38, 4671 A. (University
Microfilms No. 77-32141, 74.)
Cromwell, R. E., D. H. Olson, and D. G. Fournier
1976 "Tools and techniques for diagnosis and eval-
uation in marital and family therapy." Family
Process 15 (March):1-32.
Crowne, D., and D. Marlowe
1964 The Approval Motive: Studies in Evaluative
Dependence. New York:John Wiley and Sons.
August 1981 JOURNAL OF MARRIAGE AND THE FAMILY
Edmonds, V. H.
1967 "Marital conventionalization: Definition and
measurement." Journal of Marriage and the
Family 29 (November):681-688.
Edwards, A. L.
1957 Techniques of Attitude Scale Construction.
New York:Appleton-Century Crofts.
Frazier, L. P.
1976 An Evaluation of the Marital Satisfaction In-
ventory. Doctoral dissertation, Texas A & M
University. Dissertation Abstracts Interna-
tional, 37, 5062 A. (University Microfilms No.
77-02620, 61.)
Hawkins, J. L.
1966 "The Locke marital adjustment test and social
desirability." Journal of Marriage and the Fam-
ily 28 (May):193-195.
Kimmel, D., and F. van der Veen
1974 "Factors of marital adjustment in Locke's
Marital Adjustment Test." Journal of Mar-
riage and the Family 36 (February):57-63.
Lake, D. G., M. B. Miles, and R. B. Earle
1973 Measuring Human Behavior. New York:
Teachers College Press.
Lively, E. L.
1969 "Toward a concept clarification: The case of
marital interaction." Journal of Marriage and
the Family 31 (February):108-114.
Locke, H. J., and K. M. Wallace
1959 "Short marital-adjustment and prediction tests:
Their reliability and validity." Marriage and
Family Living 21 (August):251-255.
Murstein, B. I., and G. D. Beck
1972 "Person perception, marriage adjustment, and
social desirability." Journal of Consulting and
Clinical Psychology 39 (3):396-403.
Orden, S. R., and N. A. Bradburn
1968 "Dimensions of marriage happiness." Ameri-
can Journal of Sociology 73 (May):715-731.
Phillips, C.
1973 "Some useful tests for marriage counseling."
The Family Coordinator 22 (January):43-53.
Roach, A. J.
1975 "The Marital Satisfaction Inventory." Pp. 45-
50 in L. P. Frazier, An Evaluation of the Mari-
tal Satisfaction Inventory. Doctoral disserta-
tion, Texas A & M University, 1976. Disserta-
tion Abstracts International, 1976, 37, 5062 A.
(University Microfilms No. 77-02620, 61.)
1977 "Marriage Problem Checklist." Pp. 62-63 in
S. R. Bowden, An Assessment of the Validity
of the Marital Satisfaction Inventory. Doctoral
dissertation, Texas A & M University, 1977.
Dissertation Abstracts International, 38, 4671
A. (University Microfilms No. 77-32141, 74.)
Snyder, D. K.
1979 "Multidimensional assessment of marital satis-
faction." Journal of Marriage and the Family
41 (November):813-823.
Spanier, G. B.
1976 "Measuring dyadic adjustment: New scales for
assessing the quality of marriage and similar
dyads." Journal of Marriage and the Family
38 (February):15-38.
545
Straus, M. H.
1969 Family Measurement Techniques. Minneapo-
lis:University of Minnesota Press.
Terman, L. M.
1938 Psychological Factors in Marital Happiness.
New York:McGraw-Hill.
Thompson, R. R.
1978 "Effects of a sex therapy treatment program on
couple relationships." Unpublished master's
thesis, Texas A & I University, Kingsville.

SAVETHESEDATES
November18-21, 1981
Philadelphia,Pa.

Seventh
NASW
National
Professional
Symposium
SocialWork
Practice
ina Turbulent
World
For informationwrite:
Symposium
A National
Association
of Social Workers
|C1•Wl Suite 600
1425 H St., N.W.
Washington, D C. 20005

546 JOURNAL OF MARRIAGE AND THE FAMILY August 1981