Citation
Cognitive similarity and understanding over time in satisfied, dissatisfied and improved couples

Material Information

Title:
Cognitive similarity and understanding over time in satisfied, dissatisfied and improved couples
Creator:
Hudson, Jean Ellen McLaughlin, 1945-
Publisher:
[s.n.]
Copyright Date:
1987
Language:
English

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Jean Ellen McLaughlin Hudson. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
17739802 ( OCLC )
AEW8178 ( LTUF )
0030594548 ( ALEPH )

Downloads

This item has the following downloads:


Full Text
COGNITIVE SIMILARITY AND UNDERSTANDING OVER TIME IN SATISFIED, DISSATISFIED
AND IMPROVED COUPLES
By
JEAN ELLEN MCLAUGHLIN HUDSON

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE DOCTOR OF PHILOSOPHY

1987




ACKNOWLEDGMENTS

My deepest love and gratitude is given to my husband, Terry. His encouragement and understanding throughout my graduate school studies and dissertation preparation have made the critical difference in my completion of this project. Also, I would like to express my gratitude and affection to my three daughters, Jean-Marie, Laura and Susan, who are a very important part of this student's life.
Since my professional interest is marriage and family, I felt particularly determined to keep my family as my priority and my studies in perspective. Determined as I might
have been, I know there were times when I failed.
Very special thanks are given to my cochairperson, Dr. Greg Neimeyer, who has been encouraging throughout this project. He has spent much time and energy with this
graduate student and her projects. I value the time I have spent with him and I have learned not only from his theoretical orientation but also from his ability to organize, innovate and constructively criticize.
I would like to thank my chairperson, Dr. Ellen Amatea, whose knowledge and enthusiasm fostered my interest in




working professionally with couples and families. Her
understanding, advice and thoroughness have greatly complemnented this dissertation.
I would also like to thank my supervisory committee members: Dr. Harry Grater, Dr. James Archer, and Dr. Mary Fukuyama. Their support and recommendations have contributed to the success of this project.
In addition, I wish to thank the following people who have been instrumental in the day-to-day activities of data collection and analyses: To Martha Cousar-Davis for her
tremendous help in the collection and organization of the data; to Jackie Twitchell, who helped with proctoring; to April Metzler, for her dedication and contribution of time, energy, and interest throughout this project; to Doug Reese, for sharing his computer and programming expertise; to
Jean-Marie Hudson, my daughter, for her help with computer input of data; to Ces Bibby for her patience, tolerance, and
message-taking during the data collection phase; to Marcia Buchanan for her editorial help in the final days of this dissertation.

iii




TABLE OF CONTENTS
PAGE
ACKNOWLEDGMENTS .................................... ii
LIST OF TABLES ...................................... vi
LIST OF FIGURES ..................................... vii
ABSTRACT ........................................... viii
CHAPTER I. INTRODUCTION ............................ 1
Scope of the Problem .............................. 4
Purpose of the Study .............................. 10
Need for the Study ................................ 11
General Research Questions ........................ 12
Definition of Terms ............................... 12
Organization of the Study ......................... 15
CHAPTER II. REVIEW OF LITERATURE ................... 16
Personal Construct Theory ......................... 16
Implications for Relationship Formation and
Development ..................................... 18
Validation ...................................... 20
Construct Similarity ............................ 23
Functional Similarity ........................... 24
Structural Similarity ........................... 25
Extension ....................................... 27
Application of Kelly's Theory to the Marital Relationship .................................... 31
Summary ......................................... 37
CHAPTER III. METHODS ............................... 39
Research Design ................................... 40
Description of the Sample ......................... 43
Instrumentation ................................... 49
Areas of Change Questionnaire ................... 49
The Marital Awareness Repertory Test ............ 52
Demographic Data Questionnaire .................. 55




PA GE
Procedures.............................................. 56
Time Frame of Study............................... .. 57
Research Hypotheses.................................... 58
Analyses of Data....................................... 59
Limitations............................................. 59
Summary.................................... ......... 62
CHAPTER IV. RESULTS.......................... ........... 64
Analyses of Hypothesis I... .....................64
Analyses of Hypothesis II.................... .... ... 67
Summary................o......... ......... ....71.
CHAPTER V. DISCUSSION .......o.............. o.........73
Implicatios. .. o.................................80
Recommnendations.. ...................................... 84
APPENDICES
A THE MARITAL AWARENESS REPERTORY TEST................88
B DEMOGRAPHIC INFORMATION QUESTIONNAIRE ................94
C INFORMED CONSENT FORM............................... .. 96
REFERENCES........................................ .......97
BIOGRAPHICAL SKETCH-.............. ....... .. ..... ...... 108




LIST OF TABLES

PAGE

TABLE 1.

TABLE TABLE

TABLE 4. TABLE 5.

TABLE TABLE

TABLE 8. TABLE 9.

MEAN SATISFACTION SCORES AND STANDARD DEVIATIONS FOR ALL THREE GROUPS AT T IME I AND TIME II............................
ANOVA USING COUPLE SATISFACTION SCORES ......
MEAN AND STANDARD DEVIATION FOR THE SCORES ON THE AREAS OF CHANGE QUESTIONNAIRE (ACQ) FOR THIS STUDY SAMPLE AND THE NORMATIVE SAMPLE........................................
MEAN AGE, EDUCATION AND LENGTH OF COHABITATION FOR COUPLES SATISFACTION GROUPS ......
MEAN SCORE AND STANDARD DEVIATION FOR FUNCTIONAL SIMILARITY AT TIME I AND TIME II ......................................
ANOVA USING FUNCTIONAL SIMILARITY............
ANOVA USING FUNCTIONALLY INDEPENDENT CONSTRUCTS.... ................................
MEAN SCORES AND STANDARD DEVIATION FOR PREDICTIVE ACCURACY AT TIME I AND TIME II...
ANOVA USING PREDICTIVE ACCURACY..............




LIST OF FIGURES

PAGE
41 70

FIGURE 1. FIGURE 2.

LEVEL OF SATISFACTION ......................
MEAN PREDICTIVE ACCURACY SCORES FOR ALL GROUPS AT TIME I AND TIME II ...........

vii




Abstract of Dissertation Presented to the
Graduate School of the University of Florida
in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy
COGNITIVE SIMILARITY AND UNDERSTANDING OVER TIME
IN SATISFIED, DISSATISFIED AND IMPROVED COUPLES By
Jean Ellen McLaughlin Hudson August 1987
Chairperson: Ellen S. Amatea Co-Chairperson: Greg J. Neimeyer Major Department: Counselor Education
This study tested and compared levels of cognitive similarity and understanding in a sample of 47 couples reporting different levels of relationship satisfaction. It also examined changes in these dimensions over time for stable satisfied (n =21), and stable dissatisfied (n 11),
and couples reporting improvement in satisfaction (n 15)
over a period of 8-10 weeks. Levels of satisfaction were
assessed by use of the Areas of Change Questionnaire. Cognitive similarity and understanding were assessed through the use of the Marital Awareness Repertory Test (Reptest).
Results from this study revealed a tendency for the satisfied and improved couples to share higher levels of cognitive similarity as measured by functional similarity.

viii




Also, couples whose satisfaction level improved over time tended to increase their level of cognitive similarity. In
addition, there was a tendency for satisfied and improved couples to have higher levels of understanding than the dissatisfied couples. Also those couples who improved marital satisfaction over time also had statistically
significant improvements in their level of understanding, while satisfied and dissatisfied couples remained relatively stable.
It was concluded that the findings in this study give some support to both the theoretical and clinical notions of
the importance of cognitive similarity and understanding in the couple relationship. Implications for the findings in this study include use of the variables cognitive similarity
and understanding in marital and family outcome assessment. In addition, this study presents further support for the usefulness of the Reptest with couples. Recommendations for future research are discussed.




CHAPTER I
INTRODUCTION
Marriage in the United States today remains a vital institution affecting most individuals. However, the high rate of marital dissolution and associated problems are challenging both mental health practitioners and researchers to improve the present understanding of the marital relationship.
The current status of marriage and divorce in the United States has been described as problematic by Blumstein
and Schwartz (1983) in their book, American Couples. In
reporting on the relationship of money, work and sex for 300 couples, they state
In 1982, there were 2 percent more marriages than in 1981 and 16 percent more than in 1975. We have reached a new national record.
But this is not necessarily a comforting statistic because the divorce rate is almost three
times what it was in the 1960's. Demographers
project that half of first marriages now taking
place will end in divorce and that nationwide, 41 percent of individuals now of marriageable
age will at some time experience divorce. (p. 33)
It is important to note that these statistics refer only to distressed marriages which result in divorce. In




reality, not all unhappy marriages end in divorce (Landis, 1963). Thus, these divorce statistics should be viewed as the number of recorded dissatisfied relationships and not the number of actual dissatisfied marital relationships.
The problems incurred by unsuccessful relationships have challenged contemporary mental health professionals. Indeed, the growing concern with the rate of marital dissatisfaction is reflected in the recent developments in the area of marriage and family practice and research. Since the 1950s, couple and family issues have been the focus of attention for three main therapeutic approaches: the psychoanalytic, behavioral, and systems. This attention is evident in the rapidly increasing research literature on couples and the number of critical reviews of the literature (Beck, 1975; Greer & D'Zurilla, 1975; Gurman, 1973a, 1973b; Gurman & Kniskern, 1978a, 1978b, 1978c, 1981; Jacobson, 1978; Jacobson & Martin, 1976; Patterson, Weiss, & Hops, 1976; Pinsof, 1981). In addition, there is presently
increased attention in the literature on what treatments are most effective in treating distressed couples and which problems are most amenable to treatment (Gurman & Kniskern, 1981).
The variety of preventative/remedial mental health efforts evident in sources such as Birchler (1979) Gurman and Kniskern (1977), L'Abate (1977, 1981), Jacobson (1978),




Luber (1978) and Weiss (1978) also reflect this growing concern with preventing serious distress in the couple relationship. These programs (e.g. skill training) focus on nonclinical populations with the goal of educating or enriching the marital pair.
It is interesting to note that divorce itself does not prevent most people from continuing their search for a satisfying marital relationship. In fact, according to the U.S. Census Bureau Statistics, 80 percent of the individuals whose first marriages end in divorce remarry. Thus, despite
the incidence of divorce, the importance of marriage in the life of the individual in this society appears to remain relatively high. The continued importance of marriage and the need to improve professional understanding of its nature
is expressed by Newcomb and Bentler (1981) in the following remarks:
Marriage, its cultivation and nourishment,
occupies a major portion of most people's lives and can be an invaluable source of satisfaction and confidence, or a place of pain and unhappiness. A clearer understanding of what can go wrong in that relationship may allow for even
greater possibilities of growth and fulfillment
within marriage. (p. 93)
Such understanding may take the form of discovering what particular factors distinguish satisfied from dissatisf ied couples. These factors could then be applied in the treatment of disturbed marital relationships.




Scope of the Problem
Researchers have been concerned with understanding the marital relationship and the problems which occur within it for several decades. However, understanding the basic
nature of personal relationships in general and couples in particular is problematic. A major hindrance in this
attempt is noted by Hinde (1981): "There is yet no integrated body of knowledge encompassing all relevant disciplines and focusing on interpersonal relationships" (p. 1).
The conceptualization and operationalization of marital satisfaction have varied from one researcher to another. Such variance makes it difficult to compare studies. A
number of different ways of understanding the various forms
of marital satisfaction will be presented in this section, followed by methods chosen to assess and measure these qualities.
Most early studies in marital satisfaction lacked a theoretical framework and usually attempted to relate
demographic, environmental, social and general personality variables to marital satisfaction/dissatisfaction. Among
the first investigations of marital satisfaction were Terman
(1938) and Burgess and Cottrell (1939) who looked at specific personality traits. However, the lack of a conceptual
framework limited understanding of reported empirical findings. By the 1960s, emerging theoretical formulations




allowed for variables which were both better defined and operationalized. For example, Orden and Bradburn (1968)
attempted to specify the two components of marital satisfaction. The first was marital tension (i.e., the number of perceived sources of friction), and the second was marital satisfaction (i.e., number of shared pleasurable activities). In the 1970s, researchers focused more on behavioral measures. For example, Howard and Dawes (1976) looked at the frequency of sexual intercourse, number of disagreements and predicted self ratings of marital satisfaction.
While such studies as those listed above provide incremental information on marital satisfaction, together they remain inadequate to describe marital satisfaction because of problems in methodology. For example, some studies focus
only on 1) one partner's report of satisfaction, 2) individual and not dyadic variables which would be able to more
accurately describe the marital relationship, 3) small and unrepresentative samples, and 4) cross-sectional data. In
addition, the two techniques most frequently used to assess marital satisfaction, self report and behavioral observation, have their own associated problems. These problems
include their subjective nature, the influence of social desirability and the lack of statistical validity and reliability (Newcomb & Bentler, 1981).




Therefore, contemporary researchers in interpersonal relationships have come to a consensus concerning vital avenues for advancing the literature in the area. In particular, they suggest that research should attempt to determine the variables which differentiate between successful and
unsuccessful relationships. In order to accomplish this, suggestions have been made (Duck & Sants, 1983; Delia, 1980;
McCarthy, 1981) regarding the focus for future research in interpersonal relationships. More specifically, these
researchers suggest that more attention be given to conceptualizing research designs that address three aspects of relationships in general.
The first aspect concerns the notion of considering
relationships embedded within their context (Duck & Sants, 1983; McCarthy, 1981). The complexity of analyzing relationships in general is acknowledged by Delia (1980), when he argues that the flaw in predicting successful relationships is the assumption of the same trajectory for all couples. This reflects a statement by Jacobson (1981) that
Vlany attempt to describe the successful marriage must take into account the wide divergence of standards and goals on which couples base their union" (p. 558).




one of the implications of this contextual approach is
attention to the dyadic variables as opposed to individualistic variables. The complexity of the individual personality and the compounding of this complexity in the dyadic relationship, "with its behavioral, cognitive and affective
components" (Hinde, 1981, p. 6), make it difficult to accurately describe successful marriages. In addition, a
successful description should include the view of a relationship as a "process, not a state, made up of several continually interacting components" (Duck & Sants, 1983, p. 28).
The second aspect of relationships which should be accounted for in research is its longitudinal context
(McCarthy, 1981; Huston & Levinger, 1978; Newcomb, 1978; Duck, 1979; Delia, 1980). Most research to date has consisted of cross-sectional studies describing relationships in their initial stages. Data gathered from these studies may be misleading when applied to the long-term relationship (Dickens & Perlman, 1981). Longitudinal research should also be concerned with the applicability of results to real
life relationships; that is, a realistic appraisal of relationships ought to study real life relationships (McCarthy, 1981).
The third aspect of relationships needing further study is understanding the nature and processes of change which




distinguish successful from unsuccessful marriages. Such
research necessitates "describing and accounting for the present state and past or future development of relationships--including their decay, dissolution and repair . .1 (Duck & Miell, 1984, p. 229). The resulting improvement in the understanding of personal relationships will have implications for therapeutic treatment of distressed couples. As Duck and Gilmour (1981) note,
One of the concerns with the field of
personal relationships is to show its practical value as well as theoretical value,
which can be seen most easily in the clinical context, so that it is in the treatment
of personal relationships in disorder that the considerable practical benefits of the
field will be felt most strongly. (p. x)
These three aspects, the need for dyadic, longitudinal, and clinically relevant analyses, are beginning to be addressed by a developing model of close relationships emerging from George Kelly's theory of personal constructs (Kelly, 1955).
Kelly's Personal Construct Theory (PCT) is concerned with links between a person's perception and behavior. The
crucial mediating link in this chain is the interpretation of events and stimuli in our world. Kelly's fundamental
assumption is that individuals act as personal scientists. They try to understand and predict events by validating their own way of intepreting the world, i.e., validating




their own social hypotheses. These hypotheses, called constructs, represent bipolar descriptions which distinguish the ways in which things are viewed as similar and different from others (Bannister & Mair, 1968). For example, an individual may view others along the dimensions honest versus dishonest, happy versus sad, friendly versus unfriendly, et cetera. Such personal constructs are formed by
the processes of integration and differentiation, that is, by perceiving the ways in which events are similar and different from one another. All constructs taken together make up the construct system. It is this construct system which
permits the individual to order and anticipate his/her experience. one's experience includes explaining the behavior of others as well as explaining one's own behavior. In this sense, behavior is a reflection of one's personal construct system.
Kelly emphasized that constructs are not the result of deeper personality processes but actually are the basic
elements of the personality itself. So the individual is not merely an observer of events but an active participant in construing his/her own surroundings. In other words,
events have only the meaning ascribed to them by the individual.
Personal Construct Theory was originally formulated to explain individual psychological development and change and




Kelly's development of the Rep Grid provided a tool to both assess this change and test his theory. This theory and the
Rep Grid were f irst used to diagnose problems and evaluate treatment effects for individuals in therapy (Slater, 1970; Rowe, 1971). Later, this theory was enlarged through the work of Duck (1973a, 1973b, 1975, 1977, 1979) by applying Personal Construct Theory to the study of interpersonal relationship formation and development. More recently, this theory has been applied to study of the marital relationship (Richardson & Weigel, 1971; Neimeyer & Hudson, 1985). Finally, use of Personal Construct Theory and Rep Grid specifically, with couples in the clinical setting, has been
reported in single-case (Bannister & Bott, 1973) and crosssectional group study (Wijesinghe & Wood, 1976; Ryle & Breen, 197 2b) .
Purpose of the Study
The purpose of this study was to examine how Kelly's Personal Construct Theory could be applied to looking at different groups of married and cohabiting couples. More
specifically, the purpose was to determine whether the Personal Construct Theory dimensions of cognitive similarity and understanding could be used to discriminate between satisfied and dissatisfied couples, and whether there is a
relationship between these dimensions and changes in reported satisfaction over time.




Need for the Study
Defining the marital relationship as a process (Duck & Sants, 1983) rather than a state of interaction, requires that researchers adopt new perspectives for investigating it. Several major writers and researchers have recognized the need to look at the relationship, per se, and not just the individuals in the relationship. This involves examining the relationship in terms of dyadic variables, such as cognitive similarity and understanding. In addition, there
is a need to look at the marital relationship over time. Such longitudinal data may provide information regarding
specific kinds of changes which occur in the marital relationship and, as noted by McCarthy (1981), be more applicable to real life relationships. Although there has been
some preliminary research which has looked at dyadic variables (e.g., similarity and understanding) research has neglected to look at changes in the marital relationship over time. In addition, there has been a need to examine what distinguishes satisfied couples from dissatisfied couples in terms of changes in specific variables over time.
Kelly's Personal Construct Theory (1955) is a promising approach to operationalizing these new research perspectives
because it allows one to focus on dyadic variables over time to detect changes in relationships. To date, however,




there has been only limited application of Kelly's theory to
the study of the marital relationship (Richardson & Weigel, 1971; Ryle & Breen, 1972b; Wijesinghe & Wood, 1976; and Neimeyer & Hudson, 1985), and no research has applied this theory to the study of the marital relationship in its longitudinal context. In addition, no research has compared stable satisfied and stable dissatisfied control with couples who report changes in satisfaction. The above needs
for investigating the marital relationship formed the backdrop for research in this study.
General Research Questions
This study is concerned with the following questions:
1. Do couples who are satisfied show greater cognitive
similarity in their construct systems than couples
who are dissatisfied?
2. Do satisfied couples show greater understanding
than the dissatisfied couples?
3. Is there a change in the level of cognitive similarity and understanding over time for couples reporting changes in satisfaction not found in
stable dissatisfied!/satisfied couples?
Definition of Terms
For the purpose of clarity, definitions of the following terms are provided.




1. Choice corollary--The corollary in Kelly's Personal
Construct Theory which describes individuals as
choosing to elaborate their construct systems into
an increasingly complex and comprehensive system
for understanding experience.
2. Cognitive complexity--A variable introduced by
Bieri (1955), which permits assessment of structural similarity. It refers to the degree of
differentiation (i.e., functionally independent
constructs [FIC]) in an individual's construct
system.
3. Cognitive Similarity--Commonality of constructs
between two individuals. The term similarity is
used interchangeably with cognitive similarity when referring to a dependent variable in this
study. The different types of similarity are
a. Construct similarity--Refers to the sharing
of constructs with respect to content.
b. Functional similarity--Refers to the common
application of constructs by two individuals.
c. Structural similarity--Refers to the degree of
commonality in the organization of the constructs between two individuals' systems.
4. Commonality corollary--Kelly's corollary which
states that there are certain constructs which
individuals share.
5. Constructs--Social hypotheses represented as bipolar descriptions which distinguish the ways in which things are viewed as similar and different
from others (Bannister and Mair, 1968).
6. Construct system--All an individual's constructs
taken together.
7. Dissatisfied couple--A man and woman married or cohabiting who report desiring fourteen or more
changes in their relationship. These scores are
within 2 standard deviations of the mean for
distressed couples on the Areas of Change
Questionnaire as reported by Birchler and Webb
(1977).




8. Extension--Derived from the Sociality corollary, it is a concept referring to expansion of one's
construct system. A means of extension is by increasing one's interpersonal understanding
(Neimeyer & Neimeyer, 1982).
9. Improved couple--A man and woman married or cohabiting who report increased satisfaction (i.e.,
at least seven fewer changes) on the Areas of
Change Questionnaire at Time II.
10. Individuality corollary--Kelly's corollary which
states that each individual has his/her own way
of interpreting the world and it is those particular constructs which account for a person's
individuality.
11. Predictive accuracy--The accuracy with which individuals predict one another's construct ratings
on the Marital Awareness Repertory Test.
12. Role relationship--The patterns of activities of
an individual in relation to another.
13. Satisfaction--A term which refers to fulfillment
of needs, expectations or desires of an individual or couple; contentment. It is operationalized in
this study by satisfaction score based on the
Areas of Change Questionnaire.
14. Satisfied couple--A man and woman who are married
or cohabiting who report a desire for zero to thirteen changes in their relationship. These
scores are within 2 standard deviations of the mean
for satisfied couples on the Areas of Change
Questionnaire as reported by Birchler and Webb
(1977).
15. Similarity--Commonality of constructs between two
individuals. Also referred to as cognitive similarity (see Cognitive similarity).
16. Sociality corollary--The corollary in Kelly's
Personal Construct Theory which states that the extent to which an individual is able to assume
the viewpoints of others (i.e., constructs) influences the development of interpersonal
relationships.




17. Stable satisfied/dissatisfied couple--A couple
whose reported dissatisfaction over time remains relatively stable. Whether satisfied or dissatisfied initially for this study, stable satisfied
and stable dissatisfied may be referred to as
satisfied and dissatisfied, respectively.
18. Understanding--A concept which describes a cognitive process by which one individual comprehends
or has a clear perception of the meaning of someone or something. It is operationalized in this
study by predictive accuracy between two partners.
19. Validation--Also called consensual validation,
this term refers to the process by which individuals seek to confirm their social hypotheses.
It is one way in which individuals can elaborate
their construct systems.
Organization of the Study
This dissertation is organized into five chapters. Chapter II contains a review of the related literature. This review includes further discussion of Personal Construct Theory and research reporting application of this theory to describing interpersonal relationships in general and friendships and the marital relationship in particular. Chapter III consists of a discussion on research methodology, data collection, and analyses. Chapter IV includes the report of results of this study. Chapter V presents a
discussion of the results, implications and recommendations based on the findings.




CHAPTER II
REVIEW OF LITERATURE
Kelly has formulated a theory to explain individual psychological development and change in the course of
natural development and/or treatment. In addition, he has developed an instrument for measuring and assessing this change. Recently, Kelly's Personal Construct Theory has
been used to explain close interpersonal relationships such as friendships and marriages. The purpose of the present study is to examine Kelly's theoretical dimensions of cognitive similarity and understanding over time in couples reporting different levels of marital satisfaction.
In this chapter, the original formulations of Kelly's Personal Construct Theory will be presented, including both his fundamental postulate and corollaries. In addition,
research applications of Kelly's theory to interpersonal relationships will be reviewed.
Personal Construct Theory
Kelly (1955) has formally articulated one fundamental postulate and eleven corollaries for his theory. The
central postulate states that "a person's processes are




psychologically channelized by the ways in which he/she anticipates events" (p. 46). In addition, four of the
eleven corollaries are related to the topic of close relationships and will be reviewed. The individuality corollary
states that each individual has his/her own way of interpreting the world and it is those particular constructs which account for a person's individuality. The commonality
corollary states that there are certain constructions which individuals share. The sociality corollary suggests that the extent to which an individual is able to assume the viewpoints of others (i.e., constructs) influences the development of interpersonal relationships. Lastly, the
choice corollary views the individual as choosing to elaborate his/her construct system into an increasingly complex and comprehensive system for understanding experience.
These corollaries suggest that an individual's constructs are keys to understanding his/her particular perceptions and behavior. Furthermore, Kelly (1955) believed that the ability to unlock another's phenomenological world would
greatly enhance the understanding of individuals and interpersonal relationships. In addition, such understanding would have valuable therapeutic application in the treatment of distressed individuals and relationships.
To gain access to an individual's construct system, Kelly devised the Role Construct Repertory Test, more commonly known as the Reptest or Rep Grid. This Rep Grid is




a technique based on an individual's expression of his or her personal role constructs. A role is defined as a
pattern of activity carried out in light of one's understanding of others (Kelly, 1955). As such, the Rep Grid
provides a means for assessing change in an individual, not based solely on statistical methods, but as emphasized by Bannister and Mair (1968) on psychological processes of individuals studied.
Originally designed as a clinical instrument to identify problems for the individual, the Rep Grid's use has been extended to the research on personal relationships
(Duck, 1972; 1973a; l973b; 1975; 1977) In addition to
being used as an assessment tool to evaluate progress in individual and group therapy (Slater, 1970; Watson, 1970; Rowe, 1971; Winter & Trippett, 1977; Ryle, 1980, 1981), it has more recently been used to evaluate change in couples in therapy (Bannister & Bott, 1973; Ryle & Lipshitz, 1975;
Burns, Hunter & Lieberman, 1980) Use of a variety of Rep Grid instruments has been related to an extended Personal Construct Theory, which carries implications for intimate relationships.
Implications for Relationship Formation and Development
Implications for Personal Construct Theory have been drawn most clearly from work regarding the formation of




friendship by Duck (1973a; 1973b; 1975; 1977; 1979) This section discusses the cognitive processes which Duck and others regard as important for the development and maintenance of relationships and the literature relevant to each of these processes. After reviewing this literature, an extension of this work to marital relationships is presented.
This author makes the assumption that the reason individuals seek relationships in the first place is pertinent to understanding interpersonal relationships. One such reason or motivation of individuals is eloquently set forth by H. Kelley (1979) in the following remarks:
The unavoidable consequence of human social
life is a realization of the essentially private
and subjective nature of our experience of the world, coupled with a strong wish to break out of that privacy and establish contact with another mind. Personal relationships hold out to
their members the possibility, though perhaps rarely realized in full, of establishing such
contact. (p. 169)
Another reason is presented by G. Kelly (1955), who viewed individuals as personal scientists who choose alternatives which appear to afford the greatest opportunity for personal
elaboration (choice corollary) There are two aspects of personal elaboration: definition and extension. Definition allows one an awareness of oneself while extension permits




awareness of others which, in turn, expands the self. In
other words, this view of relationship development states that individuals choose relationships based on how well the relationship will enhance one's understanding of oneself and others. In addition, Kelly makes it clear that the development of relationships is not incidental to the process of understanding oneself and others, but essential to an individual's growth and dynamic change.
According to Personal Construct Theory, there are at least two ways in which an individual decides which relationships provide the greatest potential for elaboration: validation and extension. The first way, validation, is derived from the commonality corollary; the second way, extension, is derived from the sociality corollary. Validation
The commonality corollary states that while individuals are unique, some of them share common understandings. This commonality of personal constructs provides a measure of consensual validation for interactants' perceptions. That is, as social hypotheses, personal constructs need to be validated (Duck, 1973a). Validation occurs when an individual obtains evidence of the correctness of his/her
hypotheses. In this sense, validation is an aid in the process of definition.




In drawing on these early insights, Duck (1973a; 1973b)
hypothesized that individuals who share similar constructs would be more attractive to each other than dissimilar others. The reason for the attraction is that similarity provides consensual validation for one's own understanding and also facilitates communication. Empirical support for this view is provided by Byrne's (1969) contention that
research showed that similarity of attitudes does provide validation.
According to Duck (1973a) this validation occurs at varying levels of construing as the relationship develops. Individuals may at first be concerned only with superficial/physical aspects of the person (e.g., whether one is attractive versus unattractive). When the relationship deepens, this concern may focus more on psychological aspects (e.g., ambitious versus lazy) of the person. So at each level of construing, individuals seek validation
through construct similarity and make assumptions regarding deeper levels. When this construct similarity is inadequate at any level, Duck proposes that it acts as a filter, which prevents further relationship development. In other words, this filter hypothesis states that over the course of a relationship, individuals change the focus of their attention from a very general cognitive assessment of one another to




an increasingly specific, differentiating one. Specifically, in developed relationships, individuals will be more concerned with personality attributes versus attitudes because this allows partners' awareness of similarities and differences in their construct system. Further, Duck proposes that a focus on personality variables later in the relationship is a way of distinguishing friends from nonfriends if his filter hypothesis is correct.
As noted by Adams-Webber (1979), Duck successfully
tested his view by assessing the construct similarity between pairs of friends and nominal pairs using the Reptest (which can measure specific personality characteristics) and
the California Personality Inventory (which measures more general dimensions) The findings support Duck's position that friendship pairs were more similar in terms of personal constructs and that the Reptest was significantly more accurate in discriminating friendships from nominal pairs.
In summary, it appears that validation through construct similarity is related to relationship development. Against this general backdrop, several different types of similarity have been examined. In particular, there are three forms of similarity which have been explored. The
first concerns the similarity of constructs themselves. The second concerns the application of those constructs, and the third relates to the structure of the construct system.




Construct Similarity
Construct similarity refers to the sharing of constructs with respect to content, e.g. two individuals use the following constructs: honest versus dishonest, happy versus sad, ambitious versus lazy. As noted by Neimeyer and Neimeyer (1982), construct similarity is the most researched
type of validation, due to the work of Duck (1973a; 1973b; 1975; 1977; 1979) and colleagues (Duck & Allison, 1978; Duck & Spencer, 1972; Lea, 1979) on friendship development.
Basically, the findings in these friendship studies support a positive relationship between construct similarity and successful friendship development. Specifically, higher
levels of construct similarity characterize friends versus nominal pairs (Duck, 1973a) or acquaintances (Lea, 1979). Similar findings have been reported in both cross-sectional (Duck, 1973b) and longitudinal studies (Duck & Spencer,
1972) with a variety of populations (Duck, 1973a; 1975). Further support for the importance of construct similarity in friendship development is its reported absence in deteriorated relationships. Duck and Allison (1978) note that less construct similarity characterizes individuals who choose to separate after living together one year than those who stayed together.




Functional Similarity
Functional similarity refers to the common application of constructs by two individuals. Using the above constructs, a husband and wife view the same people similarly (i.e., they describe the same individuals using the same pole of the construct: Susan is ambitious, Larry is dishonest, et cetera). Functional similarity, like construct
similarity, is viewed as a source of consensual validation and, therefore, attractive to individuals. Initial work in assessing the relationship between functional similarity
(FS) and attraction was done by Neimeyer and Neimeyer (1981) on a group of previously unacquainted individuals. They
hypothesized that if FS provided consensual validation for the application of group members' constructs to one another's, then high FS dyads would show more mutual attraction than low FS dyads. Results supported their hypothesis.
A second study investigated the relationship of FS to friendship development, i.e., whether or not it remained important. The amount of FS of close friends was compared to FS displayed by acquaintance pairs. Results showed that
friends shared greater FS than acquaintances along psychological dimensions. In addition, acquaintances showed greater amounts of FS along physical dimensions and the




least amounts of FS along psychological dimensions. These
findings (i.e., greater FS along psychological dimensions for friends and along physical dimensions for acquaintances) lend further support for Duck's filter hypothesis.
From these studies, it appears that FS is positively related to relationship development. An additional study by Neimeyer and Neimeyer (1982) investigated whether functional dissimilarity would be predictive of loss of attraction over time. The results suggest that initial functional dissimilarity may hinder relationship development while functional dissimilarity later may be less threatening to the partners. Structural Similarity
The last type of similarity discussed here is structural similarity. This refers to the degree of commonality in the organization of the constructs between two individual's systems. One way of assessing structural similarity is by looking at cognitive complexity, a variable introduced by Bieri (1955). Cognitive complexity refers to the "degree of
differentiation in an individual's construct system, i.e.., the relative number of different dimensions of judgment used by a person" (Tripodi & Bieri, 1964, p. 122). Structural
similarity, in terms of cognitive complexity, has been
reported in several studies. Bieri (1955) argued that the greater cognitive complexity of an individual the better




that individual will be able to predict others. He provided empirical support for his argument by finding a significant positive correlation between his subjects' cognitive complexity scores and their ability to predict the responses of acquaintances on a social questionnaire. In addition, Adams-Webber, Schwenker, and Barbeau (1972) reported that one's cognitive complexity is related to the accuracy in making inferences about the personal constructs of others. Some studies (Leventhal, 1957; Sechrest & Jackson, 1961) do not support this hypothesis. Neimeyer, Neimeyer and Landfield (1983) argue that the cognitive structure of the target person whose behavior is being predicted may be an influencing factor in these outcomes. Supportive evidence in favor of target complexity influence is provided by Adams-Webber (1969), who found that individuals were more accurate in predicting the behavior of less cognitively complex target individuals than more cognitively complex target individuals.
Cognitive complexity has also been looked at in relation to friendship formation and maintenance. Duck (1972) assessed cognitive complexity as a means of differentiating friends and nominal pairs in attraction. Neimeyer and
Neimeyer (1983) compared the cognitive complexity of acquaintance pairs. Results showed that structural similarity assessed by cognitive complexity level was important later




(but not earlier) in the relationship; that is, cognitive complexity may be most important in long-term established relationships.
These empirical studies support the importance of validation through the different types of cognitive similarity within the construct systems. This leads to the second way
in which individuals can decide which relationships provide the greatest elaborative potential: extension. Extension
Eventually in the development of a relationship the need for expansion of the construct system becomes more important than that for consensual validation. In fact, it may be that the validation provided through construct similarity affords the individual the degree of psychological security needed to extend himself or herself (Neimeyer, 1981). A means of extension is through increasing interpersonal understanding (Neimeyer & Neimeyer, 1982). Understanding in Personal Construct Theory terms is the ability to subsume the viewpoint of another and/or predict another. Kelly (1955) emphasized in his sociality corollary that this degree of understanding limits the role relation (i.e., patterns of activities of an individual in relation to another). For example, an individual in a relationship plays a role with respect to another person when he/she tries to make sense out of what the other person is doing.




It is critical to the notion of role that the individual attempts to infer the view or outlook of another person. Bannister (1977, p. 25) reiterates that elaboration of one's own understanding of oneself and the world necessarily means developing new constructs and also escaping from old constructs.
It is in the subsuming of another's constructions that one acquires new constructs and adds to his/her understandings of the other. The acquisition of new constructs by the individual implies support for the attractiveness of dissimilarity as suggested by McCarthy and Duck (1976) Goldstein and Rosenfeld (1969), Kerckhoff and Davis (1962) Winch, Ktsanes and Ktsanes (1954); i.e., increasing one's understanding involves seeking out others with differing views. In this sense, dissimilarity may be an important source for increasing interpersonal understanding.
On the face of it, the attractiveness of dissimilarity appears to contradict the need for similarity hypothesis. However, there are several reasons that this is probably not SO. First, the waxing and waning view of similarity
(McCarthy & Duck, 1976) suggests that the need for both simi- larity and dissimilarity in a relationship is* important at different times in the relationship. Second, this attractiveness of dissimilarity may actually be an




increased tolerance of lack of similarity. Third, dissimilarity may itself be restricted to peripheral issues. As Adams-Webber (1979) has noted, "Maslow (1953) suggested that similarity between spouses' personalities may be more in
terms of basic characteristics than superficial ones" (p. 111). In Personal Construct Theory terms, this similarity would refer to commonality along the basic/most important constructs (i.e., core role constructs) but not along
peripheral constructs. In fact, findings by McCarthy and Duck (1976) support this view of attractiveness of dissimilarity on peripheral issues. Indeed, it may be that dissimilarity is only one of many sources for increasing interpersonal understanding.
Empirical evidence supporting a positive relationship between understanding and attraction in friendship is reported in the literature (Neimeyer & Neimeyer, 1982).
Using the Rep Grid, they compared the predicted and actual descriptions of a dyad. This yielded a predictive accuracy
score which reflected the extent to which partners understood each other. Results showed that attraction was
greater for more accurately predicted others throughout the stages of the interaction. In addition, individuals were more attracted to those who more accurately predicted them. This study suggests that not only understanding but the mutuality of understanding may be important in relationship




development. The importance of this mutuality may be dependent upon the type of relationship (i.e., mutuality
would be more important in the close versus the peripheral relationship).
In summary, theoretically and, to some extent empirically, it appears that cognitive similarity and understanding are important to relationship development. However,
whether similarity and understanding are related, and if so, how, has yet to be explained. Kelly deemphasized any relationship between similarity and understanding by stressing the difference between two people holding the same constructions system and two people understanding each other so they can play roles in relation to each other. Runkel
(1956) found that when individuals were similar or perceived
themselves to be so, understanding in the relationship was enhanced. Similar findings were reported by Triandis (1959) and Menges (1969). Smail (1972) found a positive relationship between similarity and understanding in a psychotherapy group. Duck (1973b, 1979) suggests that similarity of construct systems may facilitate understanding. While the relationship between similarity and understanding remains unclear, Neimeyer and Neimeyer (1982) suggest an interactive
relationship in which understanding may qualify attraction initially based on similarity.
Although the relationship between similarity and understanding is unclear, this model of relationships suggests




their importance for relationship development and maintenance. The following section will discuss the implications of the Personal Construct Theory relationship model and the effect of similarity and understanding on the marital relationship.
Application of Kelly's Theory to the Marital Relationship
Only recently have researchers suggested applying this relationship model to the marital relationship (Richardson & Weigel, 1971; Neimeyer & Hudson, 1985).
Marriage from a construct theory perspective is a comprehensive relationship which involves the development of many different subsystems (i.e., social, sexual, emotional, et cetera) at varying levels of intimacy. The marital role is defined as a process of behavioral interactions which occur in the light of one partner's understanding of the other (Kelly, 1955). This process can also be viewed as a temporal interactive one because partners' understanding
change over the course of the relationship (Duck & Sants, 1983). Findings in the friendship studies suggest that like friendships, successful marital relationships will be
characterized by greater validation and extension than unsuccessful ones. To the extent that validation and extension are processes of the role, one would expect that factors affecting these processes and/or the ability to play




one's role (e.g., similarity and understanding) will influence the success of the relationship.
The comprehensive nature of marriage itself would seem to demand greater understanding between the spouses, yet at the same time, make it more difficult to obtain. Specifically, marriage as a multileveled system and changing interactive process offers great elaborative potential, yet also great opportunity for misunderstanding. Recall that elaborative potential is a basis for choosing to form relationships in the first place. Relationships in turn are made up
of individuals who are at the same time unique (individuality corollary), while sharing some common outlooks (commonality corollary). This gap between one's uniqueness and commonality with others may result in misunderstanding as a consequence of relationship formation and development.
Misunderstanding in Personal Construct Theory terms is when two individuals construe and experience differently and each is unable to see the other's point of view. This
jeopardizes the relationship because it prevents the person from playing his/her role. In addition, a multileveled
system may react on different levels to a specific misunderstanding; that is, a conflict on an economic level may affect the sexual level of the couple's life.
Theoretically, it seems important that couples understand one another. In addition, one might expect that the




less discrepancy between partners' understanding of one
another, the more successful the role relationship because each partner is approximating the other's understanding of his/her role. This parallels Delia's (1980) thinking that "relationships . are not simply a matter of one partnerls expectations; the extent to which the other partner satisfies those expectations, and shares them is also
involved" (P. 101) .
Empirical research on the positive effect of understanding in marital relationships has been reported by Christensen and Wallace (1976). They compared maritally
adjusted couples, divorcing couples, and couples in counseling on their ability to predict the rewarding effects of
certain behaviors on their spouse. The maritally adjusted groups were always more accurate than the other two groups in their understanding of their spouse. Similar findings on
understanding in the marital relationship are reported by Laing, Phillipson and Lee (1966) and Dymond (1954).
Harvey and Wells (as cited in Harvey, Wells & Alvarez,
1976) also questioned the accuracy in understanding of distressed couples. They found evidence of both overestimation
and underestimation for males' and females' predictions of each other. Even in relatively long-term relationships, individuals tended to be inaccurate in their perceptions of




how their partners see the world in terms of areas of conflict. In addition to couples differing in their views
about the important sources of conflict, this study suggested that couples were unaware of these differences. The authors concluded that "a lack of understanding of divergence is as critical or more critical to the viability of close relationships than is the actual divergence" (p. 247); that is, partners are less upset by actual disagreement than when they believe their spouses do not understand them. However, attaining an accurate understanding between spouses
may be particularly complex in a relationship in which one partner is psychologically disturbed. The findings of
Wijesinghe and Wood (1976) on 4 married couples in therapy support this view. They suggested that the ill partners may behave in such a way that prediction by the other spouse is more difficult.
In an early formulation of Personal Construct Theory applied to marital relationships, Richardson and Weigel
(1971) emphasized the role of sociality (understanding) and commonality (similarity) as predictors of marital satisfaction. Specifically, they hypothesized that satisfaction was
dependent upon the ability of both partners to correctly grasp each other's outlooks. They also hypothesized that this mutual understanding may be dependent upon a commonality between spouses' construct systems. This facilitative




function of similarity on understanding parallels Duck's (1977) view. Theoretically, it seems that spouses' similarity may have a positive influence, if only to decrease the opportunity for misunderstanding to occur. On the other
hand, dissimilarity between spouses would seem at least
initially to increase the likelihood of misunderstanding. However, Richardson and Weigel (1971) were unable to confirm these hypotheses. Adams-Webber (1979) suggests that had
they elicited the personal constructs f rom each individual couple instead of providing a standard list for all couples,
results might have been more favorable; that is, elicited constructs are more likely to include a couple's basic (core role) constructs than a standard list.
A recent study paralleling the thinking of Richardson and Weigel and extending Personal Construct Theory work on similarity and understanding to the marital relationship is reported by Neimeyer and Hudson (1985). They hypothesized that satisfied couples would share greater validation (as provided by greater similarity in construct systems) and greater understanding (as provided by accuracy of predictions). They compared the similarity and understanding of 10 satisfied couples with 10 dissatisfied couples. Level of satisfaction was determined by the score on Areas of Change Questionnaire (Weiss, Hops & Patterson, 1973). Similarity (i.e., construct and functional similarity) and predictive




accuracy were assessed from responses of spouses on Kelly's Reptest. Briefly, their findings noted greater
construct and functional similarity for satisfied couples than for dissatisfied couples. In addition, satisfied spouses understood each other more accurately and showed
less discrepancy between partner's levels of understanding (i.e., greater mutuality of understanding) than dissatisfied couples.
In another recent study by Neimeyer (1984), the structural similarity of 20 couples was evaluated in terms of their cognitive complexity as measured by functionally independent constructs (FIC) scores (Landfield, 1977). The hypothesis that marital satisfaction was related to the
similarity between husband and wife's level of complexity but not to the absolute level of couple complexity was confirmed. That is, couples who were similar (partner's low
level of discrepancy) to each other in terms of FIC score, no matter what the absolute couple score, were more satisfied than those who were less similar.
The above literature suggests support for this general model of the marital relationship. The following section
will summarize this literature and present ways in which this model will be extended and tested in the present study.




Summary
Successful marriages have been characterized theoretically as mutually elaborative. That is, spouses encourage each other's development by validating and extending their personal construct systems. The empirical evidence reviewed in this section is consistent with this formulation. Specifically, satisfied spouses were more validating and
understanding in their relationships than were dissatisfied partners (Neimeyer & Hudson, 1985). That is, satisfied
couples were more similar in the content and application of
their construct systems than were the dissatisfied couples. This is consistent with a parallel literature in friendship formation (Duck, 1973; Duck & Allison, 1978; Duck & Spencer, 1972; Neimeyer & Neimeyer, 1981). As with friendships (Neimeyer & Neimeyer, 1983), it seems likely that satisfied couples may also organize their construct systems similarly.
Furthermore, each of these types of similarity (i.e., content, functional and structural) validates the construct system. As partners confirm the validity of some of their hypotheses, they become more open to another's differing views and better able to subsume these views. This is
supported by the findings that more satisfied spouses were better able to understand one another. These satisfied
couples also showed more mutual understanding, while dissatisfied spouses were more discrepant in their understandings




of one another. Such lack of mutuality jeopardizes interpersonal elaboration because it limits the partners' ability to contribute to their joint development in a continuous and reciprocal way (Neimeyer & Hudson, 1985).
The premise of this study is that this theoretical model and related empirical literature may have implications for couple treatment. Specifically, couples in conflict may be helped in therapy by increasing their perception of their
similarities and by broadening the understanding of each spouse by helping them to see the relationship from the other's viewpoint. With such increased understanding, one might realistically expect the couple to become more
similar in the content, application and structure of their constructs, as well as become better able to accurately predict the spouse. To detect such changes, this study will
conduct an analysis of the differences between couples who remain stable on reported satisfaction and couples who report changes in satisfaction over time.




CHAPTER III
METHODS
This study was designed to compare levels of cognitive
similarity and understanding in a sample of 47 couples reporting different levels of relationship satisfaction. In addition, it will compare changes in these dimensions over time for stable satisfied couples (n = 21), stable dissatisfied couples (n = 11), and couples reporting improvement in satisfaction (n = 15). Levels of satisfaction will be assessed by using the Areas of Change Questionnaire (Weiss, Hops, & Patterson, 1973). Cognitive similarity and
the measure for understanding will be measured through use of the Marital Awareness Repertory Test (see Appendix A), a
form of the Exchange Grid (Thomas, 1979). Each of these
instruments were administered to all groups on two separate occasions several weeks apart.
This chapter is divided into several sections. In the
first section, the research approach, design and variables of interest are discussed. In the second section, the population and sample of interest in this study and sample selection and recruitment procedures are discussed. In the third section, the instruments used in this study are presented. In the fourth section the procedure involved and




data collection and recording of the data are described. In
the fifth section, the research hypotheses for this study are presented. In the sixth section the methods of data analyses are presented. Limitations of the present study and summary of the foregoing information are presented in the final two sections.
Research Design
This study employed a descriptive design. Data were collected on the variables of cognitive similarity and
understanding over time for three groups of couples reporting different levels of relationship satisfaction. Differences across three groups of couples were examined for each of the two variables at two different times. Using a 3 x 2
factorial design (Figure 1) the first factor is a between groups variable, couple satisfaction. Satisfaction was operationalized in terms of the total number of desired changes in the relationship based on responses of the couple on the Areas of Change Questionnaire (Weiss, Hops & Patterson, 1973). According to this definition, couples
desiring the least number of changes are most satisfied and those couples desiring the most changes are least satisfied.
The three levels of this variable are 1) stable satisfied, 2) stable dissatisfied, and 3) improved. The stable
satisfied classification refers to couples who score in a




SATISFACTION LEVEL TIME I TIME II
Stable Satisfied
Stable Dissatisfied
Improved
FIGURE 1. LEVEL OF SATISFACTION
comparably satisfied range on the Areas of Change Questionnaire (Weiss, Hops, & Patterson, 1973). The stable dissatisfied classification refers to couples who score in the dissatisfied range on the same instrument. The improved
classification reflects couples who change (improve) over time in level of reported satisfaction. For the purposes of
this study, stable satisfied and stable dissatisfied may be referred to as satisfied and dissatisfied, respectively.
The second factor, time, is a within subjects variable.
Time I refers to intake administration of the instruments. Time II refers to the post-test or second administration of the same instruments. The interval between Time I and Time II was 8-10 weeks for all groups.




The variables of interest in the present study included the couples' 1) understanding, and 2) cognitive similarity. The variable understanding will be operationalized in terms of predictive accuracy derived from the Marital Awareness Repertory Test (MART). It refers to the amount of understanding as measured by the accuracy with which individuals predict one another's ratings. The predictive accuracy
score (Landfield, 1977) is calculated by determining the absolute difference between individuals' own ratings and their partner's predictions of their ratings. Lower discrepancy scores indicate better understanding. It was
expected that couples who understood each other better would have greater predictive accuracy.
The variable cognitive similarity was derived from the MART. It was operationalized in terms of functional and structural similarity of partners' construct systems.
Functional similarity (Neimeyer & Neimeyer, 1981) refers to the degree of similarity in applying the constructs to particular others. It was measured by determining the absolute difference between partners' ratings for all six elements (e.g., mother, father, et cetera) for all the six constructs of both the husband and wife. Scores
could range from 36 (no functional similarity) to 0 (in the




case of functional identity) when using only one partner's constructs rating one element. A lower score indicates less discrepancy between partners' ratings and, therefore, greater functional similarity.
Structural similarity refers to the overall organization of the construct system. One method of assessing
structural similarity is estimating the individual's cognitive complexity. Operationally, cognitive complexity was defined as the number of nonidentical ratings or functionally independent constructs (FIC) of a set of stimulus persons along a group of bipolar constructs (e.g., happy versus sad). The greater the number of nonidentical (i.e., unique) ratings between all possible pairs of constructs, the more cognitively complex is the subject (Bieri, Atkins, Briar, Leaman, Miller, & Tripodi, 1966). Possible FIC score ranges between 2 and 12 for each respondent (see Landfield, 1977, for details of scoring procedures). Structural similarity in this study referred to a discrepancy score between partners' levels of cognitive complexity with lower scores reflecting greater structural similarity.
Description of the Sample
The population of interest in this study was all married and cohabiting graduate and undergraduate students living in the university community of Gainesville, Florida.




Participation in this study required that couples be married
or living together for a minimum of 3 months and a maximum of 15 years. Ages of the partners were restricted to individuals between 18 and 40 years. In addition, at least one of the partners was to be a graduate or undergraduate student.
The sample consists of married and cohabiting couples. These couples were drawn from a group within the university community who responded to a newspaper advertisement. The advertisement requested couples who were interested in
participation in research on relationships in exchange for theater tickets.
A total of 50 married or cohabiting couples completed the test forms. Of these, three couples had only one partner who entirely completed all of the forms. Therefore,
47 couples (N = 94 individuals) who completed the Areas of Change Questionnaire, the Marital Awareness Repertory Test, and a demographic questionnaire at both administrations were included in the analyses.
These couples formed 3 separate groups of varying number: 1) stable satisfied group (n = 21), 2) stable dissatisfied group (n = 11), and 3) improved group (n = 15). Couples' group assignment was based on level of couple
satisfaction as assessed by scores on the Areas of Change Questionnaire (ACQ). Couples were rank ordered according to




their ACQ score and a median split was performed on the basis of these scores. The lower half of the split constituted the satisfied group; the upper half constituted the dissatisfied group. Specifically, the satisfied group
wanted 0 to 13 changes in their relationship and the dissatisfied group wanted 14 or more changes. Assignment of a couple to stable satisfied group or stable dissatisfied group indicated that there was little or no change in couple reported satisfaction over time. Assignment of a couple to
the improved group indicated a change in reported satisfaction at Time II administration. These improving couples
reported wanting at least seven fewer changes in their relationship at Time II (i.e., these couples were more
satisfied at Time II) A minimum of seven changes will be used because it reflects the average of the standard deviation for satisfied (5.3) and dissatisfied (9.43) groups by Birchler and Webb (1977). Therefore, these couples will represent a group which changes in a positive direction by at least one standard deviation on the Areas of Change Questionnaire. Mean satisfaction scores and standard deviations for all groups in this study are listed in Table 1.




TABLE 1. MEAN SATISFACTION SCORES AND STANDARD
DEVIATIONS FOR ALL THREE GROUPS AT
TIME I AND TIME II

SATISFIED a DISSATISFIEDb IMPROVEDc
TIME I 8.04 22.63 19.53
(5.49) (8.63) (10.33)
TIME II 6.90 24.00 7.00
(7.64) (10.37) (8.32)
an = 21
b
cn = 15
Analyses of variance on the satisfaction scores confirmed that there was a significant difference among groups, F (2,44) = 14.36, p d .0001, and a significant interaction

across time, F(2,44) 34.77, p 6

.001 (see Table 2). In

TABLE 2. ANOVA USING COUPLE SATISFACTION SCORES
SOURCE SS df MS F
MEAN 18922.26 1 18922.26 149.86*
GROUP 3625.78 2 1812.89 14.36*
ERROR 5555.61 44 126.26
R 369.40 1 369.40 32.54*
RG 789.21 2 394.60 34.77*
ERROR 499.42 44 11.35
*p g 0.0001




other words, the satisfied group was more satisfied than the dissatisfied group, and the improved group became more satisfied over time.
All couples' satisfaction scores at Time I were within 1.5 standard deviations of the mean for both the satisfied and dissatisfied couples as reported by Birchler and Webb (1977). Mean scores and standard deviations for the satisfied and dissatisfied couples as compared to the couples' scores in Birchler and Webb's normative study are reported in Table 3.
TABLE 3. MEAN AND STANDARD DEVIATION FOR THE SCORES
ON THE AREAS OF CHANGE QUESTIONNAIRE (ACQ)
FOR THIS STUDY SAMPLE AND THE NORMATIVE
SAMPLE
THIS SAMPLE BIRCHLER & WEBBc
V DISSATIS- DISSATISVARIABLE SATISFIED FIED SATISFIED FIED
MEAN SD MEAN SD MEAN SD MEAN SD
ACQ 8.04 5.49 22.63 8.62 6.90 5.3 28.46 9.43
an = 21
b n 11
cn = 50




Results of the demographic questionnaire (see Appendix B) administered to all subjects in the sample indicated that
the mean age was 26. 3 years. Ages across all three groups were comparable within 5 years (stable satisfied, M=28.7; stable dissatisfied, M=24; improved, M=25.4). At least one partner of each couple was a graduate or undergraduate student attending the University of Florida or Santa Fe Community College. Thirty-four couples were married and 13 were cohabiting. Ninety-three participants in the study were Caucasian and one female participant was oriental. only 6 of the 47 couples in this sample had children. Mean
number of years for cohabitation for the total sample was 3.1, and was comparable across the three groups (3.5 years for the Stable Satisfied; 3 years for the Dissatisfied; and 2. 9 years for the improved) These means are presented in Table 4.
TABLE 4. MEAN AGE, EDUCATION AND LENGTH OF COHABITATION
FOR COUPLES SATISFACTION GROUPS
SATISFIED a DISSATISFIED b IMPROVED c
VARIABLE MEAN SD MEAN SD MEAN SD
AGE 28.7 (4.4) 24.0 (4.0) 25.4 (3.5)
EDUCATION 17.3 (2.4) 16.3 (2.0) 16.0 (1.5)
COHABITATION 3.5 (2.6) 3.0 (1.7) 2.9 (1.9)
a n 21
b n 11
c n 15




Instrumentation
Three instruments were used in this study: 1) the
Areas of Change Questionnaire, 2) the Marital Awareness Repertory Test, and 3) Demographic Data Questionnaire. Areas of Change Questionnaire
The Areas of Change Questionnaire (ACQ; Weiss, Hops & Patterson, 1973) is a self-report inventory which assesses marital partners' presenting complaints, their perceptions of one another's complaints, and perceptual accuracy between
actual and perceived complaints. This instrument provides information on 1) desired change (i.e., what types of changes each partner actually wants from the other), 2) perceived change (i.e., what types of changes in oneself are
believed to be desired by the partner) and 3) perceptual accuracy (i.e., the extent to which each spouse correctly identifies the specific items that the partner wants changed and also the direction of that desired change.
Designed as a behaviorally specific measurement of marital complaints, this instrument was developed to identify problem areas for treatment and to evaluate treatment (Weiss, Hops & Patterson, 1973) As noted by Margolin, Talovic, and Weinstein (1983), the Areas of Change Questionnaire is consistent with recommendations for choosing practical clinical measures outlined by Nelson (1981) because it contains a large number of items representing a




broad range of possible complaints and is sufficiently specific that measurement of change is possible. Approximate time for completion of this inventory is 20 minutes.
Psychometric properties of the Areas of Change Questionnaire have been examined in respect to overall change score by Birchler and Webb (1977). They found mean scores on the ACQ to be 28.46 (SD = 9.43) and 6.90 (SD = 5.3) for 50 distressed (dissatisfied) and nondistressed (satisfied) couples, respectively. Weiss et al. (1973) obtained an internal consistency index of .89 for the scale. Also, there appears to be a significant overlap in the dimensions of marital satisfaction measured by the ACQ and Marital Adjustment Scale (MAS) (r = -.70) by Margolin and Wampold (1981) and Weiss et al. (1973). In addition, this instrument has been used in treatment outcome studies and has been shown to be a sensitive index of pretreatment to posttreatment change (Baucom, 1982; Margolin & Weiss, 1978; Weiss et al., 1973).
Further psychometric information about the ACQ is provided in a study by Margolin, Talovic and Weinstein (1983). They state that "the ACQ proves to be a cost efficient yet comprehensive tool for discriminating global levels of marital adjustment, for identifying therapeutic targets, and for measuring changes in marital complaints over time" (p. 930).




Their contrasted group approach provided evidence of validity for four of the five ACQ scores. In addition, 33 of the
34 items discriminated distress versus nondistress for one or both sexes. All three summary scores showed concurrent
validity with the modified Marital Adjustment Scale while the perceptual accuracy measures were relatively weak.
However, less evidence for convergent validity with reports of daily marital events was provided by this study.
The Areas of Change Questionnaire consists of 68 statements which refer to behavior in which relationship
dissatisfaction can arise (e.g., showing affection, money management, recreational activities, work habits, et cetera). On the first 34 statements, each partner is asked to
indicate how much change he or she would like to see in the spouse along a 7-point rating scale (e.g. I want my partner to . spend time with me: much less -3 -2 -1 -0, +1
+2 +3 much more). On the second 34 statements, they are also asked to indicate the changes they think their spouse would like to see in them along those same items. This
instrument assesses how much change each partner seeks in the relationship and how accurately each partner understands the changes desired by the other partner. The total number of changes each partner desires is summed for a couple
score, but it can be subdivided into the number of changes which partners agree and number of changes on which they disagree. The total number of changes which the couple




desires reflects the degree of marital distress (possible range 0 to 68). Higher scores indicate greater
dissatisfaction.
The Marital Awareness Repertory Test
The Marital Awareness RepertoryTest (MART; see Appendix B) is a form of the Exchange Grid (Thomas, 1979). "All forms of the Repertory Grid technique are derivatives of an original proposed by G. A. Kelly (1955) as an integral part of the development of Personal Construct Theory" (Bannister, 1965, p. 977). This repertory test is a form of a sorting test but differs from conventional sorting tests in that there are no standard sorting materials or sorting categories, any standard single form of administration or scoring procedure. This test permits measurement of the relationship between sorting categories (constructs) for the subject, not the correctness of these constructs.
Several facets of the psychometric properties of the Rep Grid have been investigated. Issues of reliability and validity have been addressed by Hunt (1951), Mitsos (1958),
Fjeld and Landfield (1961), Bannister (1965), and Slater (1965). "The assumption underlying all forms of the repertory grid is that the psychological relationship between any two constructs for a given subject is reflected in the
statistical association between them when they are used as sorting categories by the subject" (Bannister, 1965, p.
977).




The occurrence of significant association between constructs is regarded as establishing the internal consistency of a grid (Slater, 1965). Lack of internal consistency is
equated with the notion that an individual's constructs are totally unrelated to one another. The unlikeliness of this makes testing for it a trivial matter (Slater, 1965).
Test-retest reliability is examined in terms of consistency of ratings by subjects along specific constructs. In a
study by Fjeld and Landfield (1961), there was a strong tendency for individuals to rate the same acquaintances the same way when given the same constructs (contingency
coefficient = .80).
Validity of the Rep Grid has been addressed in terms of representativeness of the elements (Mitsos, 1958) and consistency of elicitation of construct labels (Hunt, 1951; Fjeld & Landfield, 1961). Representativeness of the elements is necessary if the test is supposed to indicate subjects' role relationships. That is, if the subject
develops his/ her role in the light of his/her understanding of others, those people used as elements in the grid should be representative of all the people with whom the subject must relate (Kelly, 1955). Mitsos (1958) confirmed his hypothesis that when elements are representative, construct labels elicited by the Rep Grid will be consistent across time (p.E.02).




In another study, Hunt (1951) found consistency of constructs after one week using different but still representative elements (i.e., the two test elements were matched
as to sex, relatives, authority, et cetera). He f ound an average percentage of agreement in constructs of 69 percent. Fjeld and Landfield (1961) report that subjects (N =20) in
their study were asked to retake the Rep Grid after two weeks. Given neither the original elements nor constructs previously elicited, subjects showed high agreement between
constructs on the two tests and also high agreement on the acquaintance used (Pearsonian r for constructs -.80; percentage of agreement on acquaintances (elements) = 72 percent). Other issues of validity and reliability of the Rep Grid have been discussed by Slater (1965) and Bannister (1965).
The Marital Awareness Repertory Test, a 6 x 6 grid, involves examining similarities and differences among
several important interpersonal figures in the individual's life (e.g., mother, father, partner, partner's parents) and then comparing his/her responses to those of his/her partner. more specifically, using the names of both partners and their parents as elements in the grid, each partner identifies how two of three of the above elements are alike
with a word or phrase. Then the individual identifies the opposite of such an identification for him/her. For example, the partner views the husband and father alike




because they are both ambitious. The opposite of this for that individual may be lazy. This ambitious versus lazy is
a construct, and five more such constructs are elicited. Then each element (i.e., the partners and both their parents) are individually rated along each of these constructs using a 7-point Likert-type scale (e.g., ambitious 3 2 1 0 1 2 3 lazy). In addition, partners are asked to predict their spouse's ratings along those same constructs. Partners then exchange grids and, without access to the other's prior ratings, repeat the procedure using the partner's constructs, first completing their own rating and then predicting their partner's rating. Thus, this instrument not only provides information regarding each partner's constructs, but allows identification of major points of disagreement in relation to particular constructs and elements. Estimated time for completion is 45 to 60 minutes. Demographic Data Questionnaire
The Demographic Data Questionnaire (see Appendix B) is
a data sheet developed by the researcher to assess demographic characteristics of the sample to be studied. The use of this questionnaire will allow the researcher to assess sex, race and age of participants, the length of time the couple has been married or cohabiting, and participants' educational levels and number of children. For the purposes




of this study, husband and wife denoted role relationships and not legal status.
Procedures
Each couple in the study was given an informed consent form specifying the nature of the study and requesting
release of information of these materials for confidential use in research (see Appendix C).
Procedures were set up to insure standardization of administration. The researcher trained undergraduate psychology students in administration of these questionnaires to insure proper completion. In addition, these
proctors were advised of the particular order of the tests,
how to answer possible questions of participants and standard method of checking completed tests.
Each participant was given the questionnaires with
written instructions on how to complete the questionnaire. In addition, a proctor was available to answer participants' questions. After testing, participants were urged by the proctor not to compare their ratings as they remember them.
The MART and ACQ were administered to all groups at two
separate times, 8-10 weeks apart. For the MART only, all partners were asked to fill out two each time, one for themselves and one in which they predict how their spouse would respond.




All couples participating in the study were given a couple number code which identified their responses while maintaining their anonymity.
Time Frame of Study
The following time frame for completion of this study was established:
Week 1
Train proctors and advertise for couples.
Week 2
First administration of questionnaires to
respondents of advertisement. If inadequate number of couples respond, will readvertise.
Week 3
Administration continues for new couples.
Begin to score questionnaires from first
administration.
Weeks 4-6
Continue to administer tests to new
couples until N =50. Score all data.
Week 7
Contact couples and set up appointments for
second administration.

Weeks 8-10
Second administration begins. Ensure that this administration does not coincide with final and mid-term examinations.




Research Hypotheses
The following hypotheses were tested in this study: Hypothesis I:
1) There is greater cognitive similarity in terms of
application (FS) and structure (FIC) of the construct systems of stable satisfied couples and improved couples than stable dissatisfied couples as
measured by the MART.
2) There is an interaction between level of satisfaction and time for improved couples such that these
couples show significant improvement in level of
cognitive similarity over time relative to stable
couples.
Hypothesis II:
1) There is greater understanding (PA) as measured
by the MART in stable satisfied couples and improved
couples than stable dissatisfied couples.
2) There is an interaction between level of satisfaction and time for improved couples such that these
couples show significant improvement in level of
understanding over time relative to stable couples.




Analyses of Data
The analyses compared satisfied, dissatisfied, and improved couples. Eight grids (MART) f rom each couple were analyzed to establish differences generated by each couple. A series of 3 x 2 analyses of variance was performed to determine the main effect of level of satisfaction, time and interactions. Independent and dependent sample t-tests were
used in the analyses for the purpose of isolating significant differences.
Limitations
Limitations of this design included the following:
exclusive use of a student population, use of volunteers, problems associated with longitudinal research and procedural difficulties.
The exclusive use of university student population can bring into question the possible generalizability of the
findings to other than student populations. In addition,
the very fact that they are generally an educated population
also limits the findings. Specifically, this population,
because they are educated and attending a university which offers couples counseling may be biased towards seeking earlier treatment. Therefore, dissatisfied couples may
actually be less distressed and/or 'more amenable to change than other populations. While generalizability of results




may be limited, it is expected that the real value of this research will lie in distinguishing the differences among the varying levels of satisfaction over time.
Couples selected for this study were volunteers and may not have resembled the general population of couples.
Therefore, generalizations of the findings of this study to nonvolunteers should be done with caution.
The problems associated with longitudinal research
which may have affected the results in this study included intervening events and attrition. One such event which
might affect outcome in a student population is examinations. Therefore, precautions were taken to prevent second administration of the questionnaires from coinciding with an examination period.
Another problem in this design was the probability of attrition. This was minimized by selecting a relatively short interval between administrations (8-10 weeks). In
addition, all subjects were encouraged after their first administration to return for the second administration.
They were also reminded that they would receive free movie tickets at the second administration.
Procedural problems in this design included variations in test-taking ability and compliance. For example, on the
Areas of Change Questionnaire, distressed couples may be reluctant to self-disclose on the 68-item questionnaire and




may actually minimize the number of changes desired. These couples will still be distressed but appear to be less so based on their responses. In this case, inaccurate self report may have resulted in improper group assignment.
Another procedural difficulty may have been noncompliance with directions for couples to avoid discussion of their responses on the first administration of the questionnaire. The probability that this was a significant problem was reduced because of the complexity of the task and the length of time between administrations. However, if an
individual had a particularly good memory or compared
responses with his/her partner immediately after the testtaking, these could have influenced second administration responses. However, this seemed unlikely.
Finally, use of married couples and cohabiting couples in the same study may present problems, although there is differing opinion on this in the literature. Cole (1976)
and Yllo (1978) point out there has been no difference found
between mnarrieds and cohabitors with regard to contentment and satisfaction. One argument for not comparing these couples is level of commitment may be different (Hinde,
1981; Newcomb, 1981), but other thinking (Montgomery, 1973) suggests that cohabitors need more dyadic commitment to remain intact than do marrieds because "cohabitors have f ewer external unifiers and more external disrupters that act against the relationship than do marrieds" (p. 149).




In summary, there were a few sampling and procedural limitations of this study. This author acknowledges these limitations and believes they were kept to a minimum.
Summary
This study was designed to test and compare changes in couples who report changes in relationship satisfaction with
couples classified as stable satisfied and stable dissatisfied. Classification of couples is based on the couples' satisfaction score on the Areas of Change Questionnaire (Weiss et al., 1973).
The design is a descriptive 3 x 2 factorial design. Factor I is level of couple satisfaction: stable satisfied, stable dissatisfied, and improved. Factor II refers to time: Time I was first administration, Time II was second administration. The variables of interest were 1) understanding and 2) cognitive similarity.
The sample was drawn from a university community population. It consisted of 47 married and cohabiting couples of which at least one partner is a graduate or undergraduate student. These couples formed three distinct groups: 1) stable satisfied, 2) stable dissatisfied, and 3) improved.




Instruments used in this study include the Marital Awareness Repertory Test, the Areas of Change Questionnaire and the Demographic Information Questionnaire. Procedures for collection and evaluation of data were standardized in order to prevent bias between couples or groups.
Limitations of this study included exclusive use of a student population, use of volunteers, problems associated with longitudinal research, and procedural difficulties.
Analysis of data compared the stable satisfied and dissatisfied couples. A 3 x 2 ANOVA was performed to
determine the main effects of level of satisfaction, time and interactions.




CHAPTER IV
RESULTS
The purpose of this study was to test and compare the degree of cognitive similarity and understanding in couples who report changes in relationship satisfaction with couples
who remain satisfied or dissatisfied. To accomplish this, the Areas of Change Questionnaire and the Marital Awareness
Repertory Test were administered to 47 couples at two separate times several weeks apart. Analysis of the collected data involved use of a series of 3 x 2 Analyses of Variance
to determine the effects of levels of satisfaction across time. This chapter presents the results of the data
analyses testing the two hypotheses.
Analyses of Hypothesis I
Two different research hypotheses were postulated for this study. For each of the hypotheses, two different subhypotheses were proposed. The first sub-hypothesis involved mean effects, while the second addressed interaction effects for each variable.
The first prediction stated that there would be greater
cognitive similarity (FS; FIC) in the construct systems of stable satisfied and improved couples than stable dissatisfied couples as measured by the MART.




The second prediction (Hypothesis I, Section 2), was that the improved couples would significantly increase their
level of cognitive similarity at Time II relative to stable (satisfied and dissatisfied) couples.
To test for the accuracy of these predictions, a twoway analysis of variance was performed on the mean scores, along two independent variables (functional similarity and functionally independent constructs). The factors reflected three levels of couple satisfaction (stable satisfied, stable
dissatisfied and improved), and two levels of time (Time I and Time II).
Results from analysis of the Functional Similarity (FS) variable indicated there was no main effect for groups. The difference in the means for groups (M satisfied = 78;
M dissatisfied = 92; M improved = 77) while in the predicted direction, was not statistically significant, F (2, 4 4) = 2.82, p A .07. (Note: Since this discrepancy score reflects a difference between partners' ratings, a lower score denotes greater similarity.)
Results of this ANOVA also indicates that there was a significant main effect for time, F (2,44) = 4.85, p jfi .03. Mean scores and standard deviations for all couple groups at Time I and Time II are presented in Table 5.
Results also show there was no significant interaction between groups and time for the measure FS. The differences




between the means for groups (Msatisfied = 78; Mdissatisfied = 92; Mimproved = 77) and the means for time (MTI = 84; MTII = 78) were in the predicted direction, but these differences were not statistically significant, F(2,44) = 2.53, p .09. The ANOVA data on FS are presented in Table 6.

TABLE 5.

MEAN SCORE AND STANDARD DEVIATION FOR FUNCTIONAL SIMILARITY AT TIME I AND TIME II

SATISFIED a DISSATISFIEDb IMPROVED c TIME I 79 (20) 94 (25) 84 (21)
TIME II 78 (13) 90 (22) 70 (18)
a
n = 20
n = 12
n = 15
TABLE 6. ANOVA USING FUNCTIONAL SIMILARITY
SOURCE SS df MS F
MEAN 597869.42 1 597869.42 1013.61
GROUP 3326.78 2 1663.39 2.82
ERROR 25953.02 44 589.84
R 799.01 1 799.01 4.85b
RG 834.30 2 417.15 2.53c
ERROR 7246.56 44 164.69
ap 0 .07
p .03
p e .09




Results for the 3 x 2 ANOVA on the second measure of cognitive similarity, Functionally Independent Constructs (FIC), showed no significant main effects for group or time, nor was the interaction between the two factors significant. The ANOVA data on FIC are presented in Table 7.
TABLE 7. ANOVA USING FUNCTIONALLY INDEPENDENT CONSTRUCTS
SOURCE SS df MS F
MEAN 604.70533 1 604.70533 148.32
1) GROUP 2.54656 2 1.27328 0.31a
ERROR 179.38961 44 4.07704
R 0.91302 1 0.91302 0.30b
2) RG 4.53849 2 2.26925 0.73c
ERROR 136.16364 44 3.09463
a p4 .0733
b
p-- .5898
cp .4861
Analyses of Hypothesis II
The third prediction (i.e., the first section of Hypothesis II) suggested a greater understanding (predictive accuracy; PA) in the satisfied and improved couples than in the dissatisfied couples, as measured by the MART.
The fourth prediction (the second section of Hypothesis II) expected the improved couples would show significant improvement in level of understanding at Time II relative to the stable couples.




To test the correctness of these hypotheses, a 3 x 2 ANOVA was performed on the mean predictive accuracy scores for each couple group. Results showed that the difference in the means for PA scores for groups (M satisf ied = 156;
Mdiii 178; Mimroe = 152) (see Table 8) was in
the predicted direction. However, differences did not attain statistical significance, F (2,44) = 2.34, p A .10. mean predictive accuracy scores for all groups at Time I and Time II are presented in Table 8.

TABLE 8.

MEAN SCORES AND STANDARD DEVIATION FOR PREDICTIVE ACCURACY AT TIME I AND TIME II

SATISFIED a DISSATISFIED b IMPROVED C
TIME I 157 (31) 182 (42) 170 (47)
TIME 11 155 (31) 175 (39) 136 (31)
an = 20
b n =12
cn = 15
In addition, there was a significant main effect for Time, F (2,44) = 8.73, p :E .005, such that predictive accuracy scores increased (note: lower discrepancy scores indicate better understanding) from Time I [IM = 166.91 (39.92)]




to Time II [M = 153.74 (33.25)]. Results of the ANOVA on predictive accuracy scores are presented in Table 9.
TABLE 9. ANOVA USING PREDICTIVE ACCURACY
SOURCE SS df MS F
MEAN 2317268.59 1 2317268.59 1089.52
GROUP 9974.20 2 4987.10 2.34a
ERROR 93582.06 44 2126.86
R 4384.37 1 4384.37 8.73b
RG 4844.70 2 2422.35 4.82c
ERROR 22105.61 44 502.40
a 10
p .005
c p .013
Results showed a significant groups by time interaction for predictive accuracy, F(2,44) = 4.82, p .013. Mean predictive accuracy scores for all groups at Time I and Time II have been presented in Table 8. The finding of this significant interaction was followed up by dependent samples t-tests to look at differences in the means between Time I and Time II for each group individually.
Results of these t-tests were not significant for differences between Time I and Time II mean predictive accuracy scores for satisfied [M = 157 (31); 155 (31)] and dissatisfied fM = 182 (42); 175 (39)] couples (see Figure 2).




dissatisfied improved
satisfied

TIME II

ADMINISTRATION TIME

FIGURE 2.

MEAN PREDICTIVE ACCURACY SCORES FOR ALL GROUPS AT TIME I AND TIME II

E- 0 HCI2 u u 94 u

190 180 170
160 150 140 130 120

TIME I




However,, there was a significant statistical difference for the improved couples from Time I [M = 170 (47)] to Time II [M =136 (31)], t (14) = 3.17, plE.O05.
Further tests involved independent sample t-tests to check for differences in mean predictive accuracy' scores overall among the groups. There was a significant difference in mean PA score when comparing satisfied and dissatisfied couple groups, t (30) = 1.769, p :: .05 in the predicted direction. There was also a significant difference in mean
PA when comparing dissatisfied and improved couple groups, t(24) = .718, p 4 .05, in the predicted direction.
Summary
The results of this study indicated that there were some differences in the couple groups designated as satisfied, dissatisfied and improved, and some changes across time. Testing Hypothesis I, Section 1, indicated that there was a tendency for satisfied and improved couples to be more
cognitively similar than the dissatisfied couples in terms of functional similarity scores but not cognitive complexity. Hypothesis I, Section 2, testing resulted in support for the prediction that couples whose satisfaction scores improve also improve functional similarity scores. However,
testing the second section of this hypothesis for the FIC measure resulted in no significant interaction.




Testing Hypothesis II, Section 1, resulted in partial support for the author's prediction that satisfied and
improved couples are better able to understand each other (i.e., have better predictive accuracy scores) than are dissatisfied couples. While the observed difference in the means for groups was in the predicted direction, it did not attain statistical significance. Results of testing
Hypothesis II, Section 2, showed, as predicted, a group by time interaction. Follow-up tests showed a significant improvement in understanding (i.e., predictive accuracy) for the improved couples from Time I to Time II, while satisfied
and dissatisfied couples remained relatively stable on this variable.




CHAPTER V
DISCUSSION
The primary objective of this study was to test and compare changes in cognitive similarity (FS, FIG) and
understanding (PA) in couples who report changes in relationship satisfaction with couples who did not change in their levels of satisfaction or dissatisfaction.
The first hypothesis of this study examined the differences in cognitive similarity as measured by functional similarity (FS) and functionally independent constructs
(FIG) score. It was expected that the satisfied and improved couples would share greater FS than the dissatisfied couples. While the differences between groups on the mean FS scores were not statistically significant, they did present a trend in the predicted direction. Large discrepancy scores indicate low levels of FS (see Functional Similarity, Chapter II). The lowest FS levels were found in the dissatisfied group (M =91.95; SD = 23.38). Scores for
the satisfied (M = 78.38; SD = 16.77) and the improved group (M =77.33; SD =19.31) were similar, but this appears to be a function of the improved group changing at Time II. This trend is reflected in both the friendship (Neimeyer & Neimeyer, 1981) and marital (Neimeyer & Hudson, 1985)




personal construct literature. Neimeyer and Neimeyer (1981) found significantly greater functional similarity among friendship pairs than in nominal (unacquainted) pairs along psychological constructs. Neimeyer and Hudson (1985) found a tendency toward greater FS on superordinate (more important) constructs by satisfied than by dissatisfied couples. Another study (Bankikiotes & Neimeyer, 1981) also links the importance of the construct to the functional similarity in attraction.
It had been expected that FS would distinguish the different couple groups. The tendency for FS scores to be in the predicted direction but not significant could be a result of several possible theoretical and methodological problems.
Theoretically, an explanation of the findings in this study could be that there is a relatively fixed level of FS in marital dyads. That is, established relationships may possess a certain amount of FS which acts as a filter in attraction and relationship development. The existence of differences in FS between satisfied and dissatisfied marital relationships may only be detected when evaluating couples along constructs which they rate as very important to them on their individual system but not along constructs which are superficial, as cited in studies by Neimeyer & Neimeyer (1981) and Neimeyer and Hudson (1985).




Methodologically, several factors (some derived from
theoretical issues discussed above) may have affected the results of this study. These factors are
1) The limited number of constructs elicited necessarily limits the view of that individual's construct
system and information regarding it.
2) These elicited constructs may include constructs of
varying importance in the individual's construct system. As discussed previously, more important
(superordinate) constructs may differentiate individuals and couples while less important (subordinate)
constructs may not.
The second part of Hypothesis I predicted that improved couples would increase their level of FS and FIC over time, whereas satisfied and dissatisfied couples would remain relatively stable. Partial support for this prediction was
found in the tendency for an interaction between level of satisfaction and time. As the improved group reported increased satisfaction from Time I (M = 19) to Time II (M= 7), the FS also improved from Time I (M =84.33) to Time II (M = 70.33). (Note: a low score reflects greater FS.) Also, the other two couple groups remained relatively stable from Time I to Time II on amount of FS: satisfied couples
T IM = 78.57; T IIM =78.19; dissatisfied couples T IM = 93.81;




T IIM = 90.09. In addition, the improvement of Time II mean FS score for the improved couples resulted in an average overall group score one point lower than the overall group score for the satisfied group. The trend for improved
couples to increase the amount of FS as they become more satisfied is in line with predictions. Indeed, as these
couples improved, they began to resemble the satisfied couples on this dimension.
Hypothesis I was not rejected for the other measure of
cognitive similarity, structural similarity as measured by FIC score. Despite the fact that the means for the couple
groups were in the predicted direction (satisfied M =2.45; dissatisfied M =2.60; improved M = 2.83), the results were not statistically significant. These findings did not replicate Neimeyer's (1984) cross-sectional study of 20
married couples in which marital satisfaction was found to be related to the similarity of husband and wife's levels of cognitive complexity (i.e., a discrepancy score). Since the mean scores and standard deviations for satisfied and dissatisfied couples on the Area of Change Questionnaire in Neimeyer's (1984) study are comparable to the scores in this
study, an obvious explanation for these findings (i.e., the couple groups from these two studies should not be compared
because of very different degrees of satis faction/ dis satisfaction) cannot be made.




A couple's overall level of complexity was also assessed by Neimeyer in his study but no relation between it and marital satisfaction was found. A post-hoc analysis in terms of couples' total FIC score for couples in this study resulted in similar findings.
Hypothesis II, Section 1, of this study focused on the
possible differences in understanding, as measured by the mean predictive accuracy scores for the different couple groups. It was expected that the dissatisfied couples would
show the least understanding of any group. There was a
trend in the predicted direction, F (2,44) 2.34, p :: .10. The dissatisfied couples were least understanding and the improved couples were the most understanding. This latter
finding appears, however, to be a function of improved understanding at Time II.
As previously discussed under FS results section, a significant difference between groups may have been found had the constructs been subdivided into subordinate and
superordinate as in the study of couples by Neimeyer and Hudson (1985) ; that is, predictive accuracy may have been a stronger factor distinguishing the different couple groups had the importance of the constructs been taken into consideration. Theoretically, Wijesinghe and Wood (1976) proposed that the partners understanding of each other, especially




along important constructs, is a critical factor in successful marital relationships. In other words, the findings in
this study may have been influenced by a preponderance of less important or subordinate constructs, which would not be
expected to delineate satisfied versus dissatisfied couples based on past research.
The second part of Hypothesis II focused on the differential effects of level of satisfaction and time on understanding for the couples. It was expected that the satisfied and dissatisfied couples would remain relatively
stable in their understanding, while the improved couples would show improved understanding at Time II. This expectancy was confirmed.
The predictive accuracy of all groups increased from Time I to Time II, with only the improved group's increase being statistically significant ft (14) = 3.17, p ::G. .0051.
In this study, the improved group mean PA score (M
169.66; SD = 46.64) was midway between the mean PA score for the dissatisfied couples' score (M 182.18; SD 41.14) and
the satisfied couples' score (M 156.95; SD 31. 18) at
Time I. So the PA mean score for improved couples was better than the dissatisfied couples score but not as good as the satisfied couples score. However, at Time II, the mean score for the improved group changes (M = 135.73; SD = 31) so that the improved group looks better than the satisfied group (M = 155.28; SD = 31) in accuracy of predictions.




This improvement in the PA score for couples who became
more satisfied over time lends support to the theory that understanding between spouses increases marital satisfaction. It is interesting to note that the satisfied and dissatisfied couples remained stable in their ability to predict.
other studies have addressed the importance of mutual understanding between couples (Neimeyer, 1981; Neimeyer & Hudson, 1985). Since mutual understanding could be readily assessed in this study by obtaining a discrepancy score
between husband's and wife's understanding, this was done. It was predicted that the satisfied and improved couples would share more mutual understanding (i.e., have lower discrepancy scores) than dissatisfied couples. Results of a
3 x 2 ANOVA showed no significant differences among groups. However, the means for the three groups (satisfied =11.69; dissatisfied =15.59; improved = 12.3) were in the predicted direction.
In summary, couples who were satisfied, dissatisfied and improved tended to share different levels of the variables, functional similarity and predictive accuracy. In
addition, there is a significant difference among improved couples from Time I to Time II; i.e., these couples shared both higher levels of FS and PA at Time II when they reported greater marital satisfaction. In general, the raw




scores for these couples tended to fall between the dissatisfied group and satisfied group at Time I but to improve
significantly at Time II to approach or surpass the satisfied couples raw score. These findings lend partial support to the hypotheses in this study.
Implications
The challenge to investigate complex real life interpersonal relationships in terms of dyadic, longitudinal variables has been made by several researchers in the recent past (Delia, 1980; Duck & Miell, 1984; Hinde, 1981). This study has focused on the challenge of examining the established couple relationship over time in terms of the dyadic variables of understanding and cognitive similarity. The findings in the present study provide partial support to the author's hypotheses.
Implications for findings in this study with satisfied,
dissatisfied and improved couples will be found in their application in the clinical arena. In this study, couples who reported increased marital satisfaction over time also became more similar in terms of how they applied their constructs (FS) and in terms of how they understood (PA) their partners along each other's constructs. Provided further studies replicate these findings of increased FS and PA related to increased marital satisfaction, clinical application might include marital and family therapists' use of




these variables in therapy outcome assessment. Specifically, the measurement of these variables of cognitive
similarity and understanding could be a concrete measure of improvement over the course of therapy. So clinical couples who improve with therapy would be expected to increase their levels of cognitive similarity and understanding.
Specific clinical application of these findings would include use of the Reptest not only as an assessment tool to
evaluate progress in therapy (Slater, 1970; Watson, 1970; Rowe, 1971; Winter & Trippett, 1977; and Ryle, 1980, 1981) but also as a therapeutic tool. Its use would be to point out first the accurate predictions and similar ratings and then to point out inaccurate predictions and differences in ratings. As previously suggested by Watson (1970), Neimeyer
(1985), and Neimeyer and Hudson (1985), attending to accuracies and similarities initially in the therapy process
would tend to decrease personal threat and increase consensual validation. Then pointing out inaccurate predictions and differences in ratings to make partners aware of their differences will be more readily accepted.
Conversely, the dissatisfied couples in this study had low levels of predictive accuracy and functional similarity. Such levels of PA and FS can be viewed as greater number of
differences or disagreements in terms of understanding and similarity. While it has been shown that spouses are more




upset when they believe the other misunderstands them than when their spouse disagreed with them (Harvey, Wells, & Alvarez, 1976), dysfunctional relationships commonly exhibit an intolerance of differences (Overturf, 1976). This intolerance of differences sets up an expectation that is generally considered inappropriate in close relationships; that is, irrational beliefs or expectations (such as perfect
harmony) affect the marital relationship (Epstein, 1979). Prevention of disagreements to maintain interspouse harmony has been shown to include making incorrect statements (Ryder & Goodrich, 1966). While beliefs and expectations may
result in outward agreement, literature on interspouse behavior, in general, shows it to be not so agreeable. Specifically, married couples have been shown as less polite, less explicit and more interruptive than strange
couples (Winter, Ferreira & Bowers, 1973). In addition, married individuals treat strangers more gently and generally more nicely than they do their spouses (Ryder, 1968).
Therefore, marital therapy with couples who have low PA
and FS should include understanding the couples' beliefs regarding the consequences of differences in the relationship. Again, unrealistic expectations may be challenged after emphasis of present similarities. In addition,
education of the couple regarding the potential value of disagreement to the extent of encouraging open disagreement




with proper communication skills (Gottman, Notarius, Gonso & Markman, 1980) may facilitate the therapeutic process.
When a couple's low FS and PA are viewed by the therapist to be affected at least in part by the neurotic behavior of one or both spouses, a different therapeutic
technique may be implemented. Kelly (1955) believed that the "idiosyncratic nature of a patient's discriminations may
make it difficult for others to play a meaningful relation to him/her . because roles are defined as interactions based upon an appreciation of an individual's outlook" (p. 97). Paralleling this thinking, Adams-Webber (1979) tentatively explained the finding that neurotic males were better able to predict their spouses than their spouses were them,
by suggesting that these men might act in such a way as to make prediction difficult or impossible. In these cases,
therapy must deal with spouses' role identity and interactions. Techniques suggested in the Personal Construct
Therapy literature include enactment, group psychotherapy, pyramid procedure (Epting & Boger, 1981) and Kelly's (1955) fixed role therapy (Adams-Webber, 1981).
Another implication for the results of this study is further support for the use of the Reptest with couples. The Reptest allows the researcher to look at dyadic variables shared by individuals who, in turn, share a unique relationship while, at the same time, tap commonalities




among couples. In this study, the dyadic variables of
understanding and cognitive similarity were quantified in terms of elicited constructs, thus providing a unique view of each couple relationship. In summary, clinical and
research application for the findings in this study have been presented in this section.
Recommendations
Several articles have been written pointing out the possible problems associated with the study of close relationships (Kelley, 1977; Duck & Miell, 1984; Delia, 1980; Duck & Sants, 1983). Of all the problems involved with the
study of personal relationships, this author believes that Duck and Sants' (1983) viewpoint succinctly describes the complexity of the project in the following statement:
We argue that the existence and range of discrepant viewpoints upon a given relationship itself constitutes evidence about the nature of the relationship rather than being a mere
artifact or irritant to scientists. (p. 35)
That is, relationships are multifaceted by their nature and research on them should begin to realize this. This complexity in the nature of relationships warrants a conservative attitude toward clinical application of positive research findings until replication studies have been completed.




Other issues raised by this study relate to sample selection: the exclusive use of a student population, the small sample size, problems associated with longitudinal
,research and procedural difficulties. Newcomb and Bentler (1980) point out the error in the general results with exclusive use of college students. Students face unique
life dilemmas that may easily and indeterminably alter features of cohabitation. Ideally, then, further studies of
this kind would benefit from utilization of couples from outside the university community. In addition, exclusive use of legally married couples would be recommended.
Further considerations in future studies would be to increase the number of couples involved in the study and to
equalize the numbers in each group. In addition, in order to know whether the findings in this study apply to all couples, research replicating this study with couples with children, couples in therapy and couples at different life stages is recommended.
Another recommendation would be to vary the time
between test administrations and include a time lapse of many months to a period of years. Such testing would be
valuable not only to research in relationship development and satisfaction but also in relationship dissatisfaction and dissolution.




Recommendations for future studies in which the research elicits constructs from the subjects would be concerned with the differential findings regarding superordinate versus subordinate constructs in previous studies as discussed earlier. Therefore, it would be beneficial in the future to: 1) stress the importance of subjects describing
important similarities among the elements; 2) increase the number of elements from which constructs are elicited or elicit two or more constructs for each element, thereby
increasing the number of constructs to be rate; 3) direct the subjects to number the constructs as important versus less important, in order to categorize the constructs as superordinate versus subordinate. However, such changes
would substantially increase the time required to complete the tests. With this in mind, future researchers would do well to compensate their subjects for their time.
In summary, this study investigated the level of understanding and cognitive similarity in satisfied, dissatisfied and improved couples. Results indicated that as couples
improve their levels of satisfaction, levels of understanding and cognitive similarity, marital satisfaction also increases. Implications for these findings included further
support for use of the Reptest in couple studies and possible application to marital therapy. Recommendations were offered for future research.




APPENDIX A
THE MARITAL AWARENESS REPERTORY TEST




Administration 9

1 F1 2 [D REPTEST

.*~ ] I .U *
A A
2~ ji 2

- I ~ Z A J I A L

Note:
self
ratin ags a-predictions of partner black pen ra. per

'El w D

Way Alike

May Different

3 2 1 0 1 2 3
1 2 1 0 1 2 1 ////1 .0
3 2 1 0 1 2 3 / j I U I 3
/I/1!1t 1
//W// Diffren
3///// 2 1 0 ,

Scale
3 most descriptive
2 descriptive
1 snomewhat descriptive
0 not applicable or don't know

F_

V+-

u a 9




89
The Marital Awareness Repertory Test used in this study is presented on the previous page. This test is shown with
the rating scales on which the individual rates himself/ herself along the constructs listed in the way alike/way different columns. In addition, the individual rates the spouse, his/her own parents, and the spouse's parents.




INSTRUCTIONS
This form is called the Reptest. It is designed to
help you determine the kinds of things that are important to
you in your close relationships. The Reptest is a selfexploration exercise. There are no correct answers; what is important is what you value. To complete the Reptest follow the instructions below. Please do not consult your partner
while completing the form since your individual judgments are what is of interest right now.
STEP I:
Notice the numbered columns (1-6) beginning with wife's
mother and ending with husband. These are the people you
will be considering throughout this task. You will also
find six rows, marked A-F.
What you are to do first is to find the three individuals in Row A who have circles under their names. In this
case, circles appear under column #1 (wife's mother) #2 (wife's father), and #3 (wife).
Now ask yourself the question, "how are any two of these individuals similar?" For example, someone might see his mother and his wife both as "talkative". This similarity "talkative" should be written under the "way alike" column in the center of the form. Then ask yourself what is
the opposite of this description and write this in the "way different" column on the far right hand side of the sheet. For example, the opposite of the word "talkative" for someone might be "quiet". "Quiet" should be written in the "way different" column. The dimension "talkative" vs. quiet" is called a construct.
There are many ways in which any two of those people could be considered similar. What we are interested in are the ways which you see as important. There are no right or wrong answers. We are simply interested in your best assessments.
Next, go to Row B and identify the 3 f igures in that row (in this case, #'Is 1, 2, and 6) and again answer the question "in what way are any two of these individuals alike?" Write the way in which you see them as similar in the "way alike" column and whatever you see as the opposite in the "way different" column. Your answers should all be different so that when you have completed Row A through F




you will have 6 different constructs. Complete this procedure now, before going ahead with the instructions.
STEP II:
Now look at each of the constructs and ask yourself which side of each description seems most positive to you and which side seems most negative. Indicate this by
putting a small plus sign in the parentheses which appear next to the description which is positive. Place a minus sign in the parentheses next to the description which is negative. When you have finished, the sides of every
construct should have one positive and one negative sign. If you find it difficult to decide which is more positive or
more negative then assign the positive sign to the description under "way alike".
STEP III:
Now you will notice that to the left of the constructs
there is a grid with the names of individuals above each column. Using a BLACK PEN, please rate each of the individuals along each of the 6 descriptions (constructs) Do this by inserting in the space above the diagonal line in each box the one number which indicates how you view that particular person along each of those dimensions. For
example, if your first scale is "talkative vs. quiet", and you have rated talkative as positive and quiet as negative, then rate the wife's mother towards the -3 on the right if you view her as very quiet. If you see her as "very talkative" you would rate her towards a +3 on the left side of the scale. Remember to include the valence (+ or -) before each number based on your responses in Step II. The higher the number you circle the more you see that person as having the characteristic written on that side of the scale. if
you cannot decide or the scale doesn't seem to apply, then circle the zero in the center. When you finish rating the wife's mother, then continue rating the rest of the individuals (2-6) along those same constructs. It is very
important to assign one and only one number in each rating box above the diagonal line.
STEP IV:
Notice to the left of the "way alike" column there is a small column with a check-mark appearing at the top. What we want you to do is ask yourself which of these constructs you think your partner used in making sense of people. In other words, how many of these constructs will he/she have on his/her grid? Indicate those constructs which you think




Full Text

PAGE 1

COGNITIVE SIMILARITY AND UNDERSTANDING OVER TIME IN SATISFIED, DISSATISFIED AND IMPROVED COUPLES By JEAN ELLEN MCLAUGHLIN HUDSON A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE DOCTOR OF PHILOSOPHY 1987

PAGE 2

ACKNOWLEDGMENTS My deepest love and gratitude is given to my husband, Terry. His encouragement and understanding throughout my graduate school studies and dissertation preparation have made the critical difference in my completion of this project. Also, I would like to express my gratitude and affection to my three daughters, Jean-Marie, Laura and Susan, who are a very important part of this student's life. Since my professional interest is marriage and family, I felt particularly determined to keep my family as my priority and my studies in perspective. Determined as I might have been, I know there were times when I failed. Very special thanks are given to my cochairperson, Dr. Greg Neimeyer, who has been encouraging throughout this project. He has spent much time and energy with this graduate student and her projects. I value the time I have spent with him and I have learned not only from his theoretical orientation but also from his ability to organize, innovate and constructively criticize. I would like to thank my chairperson, Dr. Ellen Amatea, whose knowledge and enthusiasm fostered my interest in

PAGE 3

working professionally with couples and families. Her understanding, advice and thoroughness have greatly complemented this dissertation. I would also like to thank my supervisory committee members: Dr. Harry Grater, Dr. James Archer, and Dr. Mary Fukuyama. Their support and recommendations have contributed to the success of this project. In addition, I wish to thank the following people who have been instrumental in the day-to-day activities of data collection and analyses: To Martha Cousar-Davis for her tremendous help in the collection and organization of the data; to Jackie Twitchell, who helped with proctoring; to April Metzler, for her dedication and contribution of time, energy, and interest throughout this project; to Doug Reese, for sharing his computer and programming expertise; to Jean-Marie Hudson, my daughter, for her help with computer input of data; to Ces Bibby for her patience, tolerance, and message-taking during the data collection phase; to Marcia Buchanan for her editorial help in the final days of this dissertation.

PAGE 4

TABLE OF CONTENTS PAGE ACKNOWLEDGMENTS ii LIST OF TABLES vi LIST OF FIGURES vii ABSTRACT viii CHAPTER I INTRODUCTION 1 Scope of the Problem 4 Purpose of the Study. 10 Need for the Study 11 General Research Questions 12 Definition of Terms 12 Organization of the Study 15 CHAPTER II. REVIEW OF LITERATURE 16 Personal Construct Theory 16 Implications for Relationship Formation and Development 18 Validation 20 Construct Similarity 2 3 Functional Similarity 24 Structural Similarity 25 Extension 27 Application of Kelly's Theory to the Marital Relationship 31 Summary 37 CHAPTER III. METHODS 39 Research Design 40 Description of the Sample 43 Instrumentation 4 9 Areas of Change Questionnaire 49 The Marital Awareness Repertory Test 52 Demographic Data Questionnaire 55

PAGE 5

PAGE Procedures 5 6 Time Frame of Study 57 Research Hypotheses 58 Analyses of Data 59 Limitations 59 Summary 62 CHAPTER IV. RESULTS 64 Analyses of Hypothesis I 64 Analyses of Hypothesis II 67 Summary 71 CHAPTER V. DISCUSSION 7 3 Implications 80 Recommendations 84 APPENDICES A THE MARITAL AWARENESS REPERTORY TEST 88 B DEMOGRAPHIC INFORMATION QUESTIONNAIRE 94 C INFORMED CONSENT FORM 96 REFERENCES 97 BIOGRAPHICAL SKETCH 108

PAGE 6

LIST OF TABLES PAGE TABLE 1. MEAN SATISFACTION SCORES AND STANDARD DEVIATIONS FOR ALL THREE GROUPS AT TIME I AND TIME II 4 6 TABLE 2. ANOVA USING COUPLE SATISFACTION SCORES 46 TABLE 3. MEAN AND STANDARD DEVIATION FOR THE SCORES ON THE AREAS OF CHANGE QUESTIONNAIRE (ACQ) FOR THIS STUDY SAMPLE AND THE NORMATIVE SAMPLE 47 TABLE 4. MEAN AGE, EDUCATION AND LENGTH OF COHABITATION FOR COUPLES SATISFACTION GROUPS 48 TABLE 5. MEAN SCORE AND STANDARD DEVIATION FOR FUNCTIONAL SIMILARITY AT TIME I AND TIME II 6 6 TABLE 6. ANOVA USING FUNCTIONAL SIMILARITY 66 TABLE 7. ANOVA USING FUNCTIONALLY INDEPENDENT CONSTRUCTS 67 TABLE 8. MEAN SCORES AND STANDARD DEVIATION FOR PREDICTIVE ACCURACY AT TIME I AND TIME II... 68 TABLE 9 ANOVA USING PREDICTIVE ACCURACY 69

PAGE 7

LIST OF FIGURES PAGE FIGURE 1. LEVEL OF SATISFACTION 41 FIGURE 2. MEAN PREDICTIVE ACCURACY SCORES FOR ALL GROUPS AT TIME I AND TIME II 70

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COGNITIVE SIMILARITY AND UNDERSTANDING OVER TIME IN SATISFIED, DISSATISFIED AND IMPROVED COUPLES By Jean Ellen McLaughlin Hudson August 1987 Chairperson: Ellen S. Amatea Co-Chairperson: Greg J. Neimeyer Major Department: Counselor Education This study tested and compared levels of cognitive similarity and understanding in a sample of 4 7 couples reporting different levels of relationship satisfaction. It also examined changes in these dimensions over time for stable satisfied (n 21) and stable dissatisfied (n •= 11) and couples reporting improvement in satisfaction (n 15) over a period of 8-10 weeks. Levels of satisfaction were assessed by use of the Areas of Change Questionnaire. Cognitive similarity and understanding were assessed through the use of the Marital Awareness Repertory Test (Reptest) Results from this study revealed a tendency for the satisfied and improved couples to share higher levels of cognitive similarity as measured by functional similarity.

PAGE 9

Also, couples whose satisfaction level improved over time tended to increase their level of cognitive similarity. In addition, there was a tendency for satisfied and improved couples to have higher levels of understanding than the dissatisfied couples. Also those couples who improved marital satisfaction over time also had statistically significant improvements in their level of understanding, while satisfied and dissatisfied couples remained relatively stable. It was concluded that the findings in this study give some support to both the theoretical and clinical notions of the importance of cognitive similarity and understanding in the couple relationship. Implications for the findings in this study include use of the variables cognitive similarity and understanding in marital and family outcome assessment. In addition, this study presents further support for the usefulness of the Reptest with couples. Recommendations for future research are discussed.

PAGE 10

CHAPTER I INTRODUCTION Marriage in the United States today remains a vital institution affecting most individuals. However, the high rate of marital dissolution and associated problems are challenging both mental health practitioners and researchers to improve the present understanding of the marital relationship. The current status of marriage and divorce in the United States has been described as problematic by Blumstein and Schwartz (198 3) in their book, American Couples In reporting on the relationship of money, work and sex for 300 couples, they state In 1982, there were 2 percent more marriages than in 1981 and 16 percent more than in 1975. We have reached a new national record. But this is not necessarily a comforting statistic because the divorce rate is almost three times what it was in the 1960 's. Demographers project that half of first marriages now taking place will end in divorce and that nationwide, 41 percent of individuals now of marriageable age will at some time experience divorce, (p. 33) It is important to note that these statistics refer only to distressed marriages which result in divorce. In

PAGE 11

reality, not all unhappy marriages end in divorce (Landis, 1963). Thus, these divorce statistics should be viewed as the number of recorded dissatisfied relationships and not the number of actual dissatisfied marital relationships. The problems incurred by unsuccessful relationships have challenged contemporary mental health professionals. Indeed, the growing concern with the rate of marital dissatisfaction is reflected in the recent developments in the area of marriage and family practice and research. Since the 1950s, couple and family issues have been the focus of attention for three main therapeutic approaches: the psychoanalytic, behavioral, and systems. This attention is evident in the rapidly increasing research literature on couples and the number of critical reviews of the literature (Beck, 1975; Greer & D'Zurilla, 1975; Gurman, 1973a, 1973b; Gurman & Kniskern, 1978a, 1978b, 1978c, 1981; Jacobson, 1978; Jacobson & Martin, 1976; Patterson, Weiss, & Hops, 1976; Pinsof, 1981). In addition, there is presently increased attention in the literature on what treatments are most effective in treating distressed couples and which problems are most amenable to treatment (Gurman & Kniskern, 1981) The variety of preventative/remedial mental health efforts evident in sources such as Birchler (1979) Gurman and Kniskern (1977), L'Abate (1977, 1981), Jacobson (1978),

PAGE 12

Luber (1978) and Weiss (1978), also reflect this growing concern with preventing serious distress in the couple relationship. These programs (e.g. skill training) focus on nonclinical populations with the goal of educating or enriching the marital pair. It is interesting to note that divorce itself does not prevent most people from continuing their search for a satisfying marital relationship. In fact, according to the U.S. Census Bureau Statistics, 80 percent of the individuals whose first marriages end in divorce remarry. Thus, despite the incidence of divorce, the importance of marriage in the life of the individual in this society appears to remain relatively high. The continued importance of marriage and the need to improve professional understanding of its nature is expressed by Newcomb and Bentler (1981) in the following remarks: Marriage, its cultivation and nourishment, occupies a major portion of most people's lives and can be an invaluable source of satisfaction and confidence, or a place of pain and unhappiness. A clearer understanding of what can go wrong in that relationship may allow for even greater possibilities of growth and fulfillment within marriage, (p. 93) Such understanding may take the form of discovering what particular factors distinguish satisfied from dissatisfied couples. These factors could then be applied in the treatment of disturbed marital relationships.

PAGE 13

Scope of the Problem Researchers have been concerned with understanding the marital relationship and the problems which occur within it for several decades. However, understanding the basic nature of personal relationships in general and couples in particular is problematic. A major hindrance in this attempt is noted by Hinde (1981) : "There is yet no integrated body of knowledge encompassing all relevant disciplines and focusing on interpersonal relationships" (p. 1) The conceptualization and operationalization of marital satisfaction have varied from one researcher to another. Such variance makes it difficult to compare studies. A number of different ways of understanding the various forms of marital satisfaction will be presented in this section, followed by methods chosen to assess and measure these qualities. Most early studies in marital satisfaction lacked a theoretical framework and usually attempted to relate demographic, environmental, social and general personality variables to marital satisfaction/dissatisfaction. Among the first investigations of marital satisfaction were Terman (1938) and Burgess and Cottrell (1939) who looked at specific personality traits. However, the lack of a conceptual framework limited understanding of reported empirical findings. By the 1960s, emerging theoretical formulations

PAGE 14

allowed for variables which were both better defined and operationalized. For example, Orden and Bradburn (1968) attempted to specify the two components of marital satisfaction. The first was marital tension (i.e., the number of perceived sources of friction) and the second was marital satisfaction (i.e., number of shared pleasurable activities). In the 1970s, researchers focused more on behavioral measures. For example, Howard and Dawes (1976) looked at the frequency of sexual intercourse, number of disagreements and predicted self ratings of marital satisfaction. While such studies as those listed above provide incremental information on marital satisfaction, together they remain inadequate to describe marital satisfaction because of problems in methodology. For example, some studies focus only on 1) one partner's report of satisfaction, 2) individual and not dyadic variables which would be able to more accurately describe the marital relationship, 3) small and unrepresentative samples, and 4) cross-sectional data. In addition, the two techniques most frequently used to assess marital satisfaction, self report and behavioral observation, have their own associated problems. These problems include their subjective nature, the influence of social desirability and the lack of statistical validity and reliability (Newcomb & Bentler, 1981)

PAGE 15

Therefore, contemporary researchers in interpersonal relationships have come to a consensus concerning vital avenues for advancing the literature in the area. In particular, they suggest that research should attempt to determine the variables which differentiate between successful and unsuccessful relationships. In order to accomplish this, suggestions have been made (Duck & Sants, 198 3; Delia, 1980; McCarthy, 1981) regarding the focus for future research in interpersonal relationships. More specifically, these researchers suggest that more attention be given to conceptualizing research designs that address three aspects of relationships in general. The first aspect concerns the notion of considering relationships embedded within their context (Duck & Sants, 1983; McCarthy, 1981). The complexity of analyzing relationships in general is acknowledged by Delia (1980) when he argues that the flaw in predicting successful relationships is the assumption of the same trajectory for all couples. This reflects a statement by Jacobson (1981) that "any attempt to describe the successful marriage must take into account the wide divergence of standards and goals on which couples base their union" (p. 558)

PAGE 16

One of the implications of this contextual approach is attention to the dyadic variables as opposed to individualistic variables. The complexity of the individual personality and the compounding of this complexity in the dyadic relationship, "with its behavioral, cognitive and affective components" (Hinde, 1981, p. 6) make it difficult to accurately describe successful marriages. In addition, a successful description should include the view of a relationship as a "process, not a state, made up of several continually interacting components" (Duck & Sants, 1983, p. 28) The second aspect of relationships which should be accounted for in research is its longitudinal context (McCarthy, 1981; Huston & Levinger, 1978; Newcomb, 1978; Duck, 1979; Delia, 1980). Most research to date has consisted of cross-sectional studies describing relationships in their initial stages. Data gathered from these studies may be misleading when applied to the long-term relationship (Dickens & Perlman, 1981). Longitudinal research should also be concerned with the applicability of results to real life relationships; that is, a realistic appraisal of relationships ought to study real life relationships (McCarthy, 1981) The third aspect of relationships needing further study is understanding the nature and processes of change which

PAGE 17

8 distinguish successful from unsuccessful marriages. Such research necessitates "describing and accounting for the present state and past or future development of relationships — including their decay, dissolution and repair ..." (Duck & Miell, 1984, p. 229). The resulting improvement in the understanding of personal relationships will have implications for therapeutic treatment of distressed couples. As Duck and Gilmour (1981) note, One of the concerns with the field of personal relationships is to show its practical value as well as theoretical value, which can be seen most easily in the clinical context, so that it is in the treatment of personal relationships in disorder that the considerable practical benefits of the field will be felt most strongly, (p. x) These three aspects, the need for dyadic, longitudinal, and clinically relevant analyses, are beginning to be addressed by a developing model of close relationships emerging from George Kelly's theory of personal constructs (Kelly, 1955). Kelly's Personal Construct Theory (PCT) is concerned with links between a person's perception and behavior. The crucial mediating link in this chain is the interpretation of events and stimuli in our world. Kelly's fundamental assumption is that individuals act as personal scientists. They try to understand and predict events by validating their own way of intepreting the world, i.e., validating

PAGE 18

their own social hypotheses. These hypotheses, called constructs, represent bipolar descriptions which distinguish the ways in which things are viewed as similar and different from others (Bannister & Mair, 1968) For example, an individual may view others along the dimensions honest versus dishonest, happy versus sad, friendly versus unfriendly, et cetera. Such personal constructs are formed by the processes of integration and differentiation, that is, by perceiving the ways in which events are similar and different from one another. All constructs taken together make up the construct system. It is this construct system which permits the individual to order and anticipate his/her experience. One's experience includes explaining the behavior of others as well as explaining one's own behavior. In this sense, behavior is a reflection of one's personal construct system. Kelly emphasized that constructs are not the result of deeper personality processes but actually are the basic elements of the personality itself. So the individual is not merely an observer of events but an active participant in construing his/her own surroundings. In other words, events have only the meaning ascribed to them by the individual. Personal Construct Theory was originally formulated to explain individual psychological development and change and

PAGE 19

10 Kelly's development of the Rep Grid provided a tool to both assess this change and test his theory. This theory and the Rep Grid were first used to diagnose problems and evaluate treatment effects for individuals in therapy (Slater, 1970; Rowe, 1971) Later, this theory was enlarged through the work of Duck (1973a, 1973b, 1975, 1977, 1979), by applying Personal Construct Theory to the study of interpersonal relationship formation and development. More recently, this theory has been applied to study of the marital relationship (Richardson & Weigel, 1971; Neimeyer & Hudson, 1985) Finally, use of Personal Construct Theory and Rep Grid specifically, with couples in the clinical setting, has been reported in single-case (Bannister & Bott, 1973) and crosssectional group study (Wijesinghe & Wood, 197 6; Ryle & Breen, 1972b) Purpose of the Study The purpose of this study was to examine how Kelly's Personal Construct Theory could be applied to looking at different groups of married and cohabiting couples. More specifically, the purpose was to determine whether the Personal Construct Theory dimensions of cognitive similarity and understanding could be used to discriminate between satisfied and dissatisfied couples, and whether there is a relationship between these dimensions and changes in reported satisfaction over time.

PAGE 20

11 Need for the Study Defining the marital relationship as a process (Duck & Sants, 1983) rather than a state of interaction, requires that researchers adopt new perspectives for investigating it. Several major writers and researchers have recognized the need to look at the relationship, per se, and not just the individuals in the relationship. This involves examining the relationship in terms of dyadic variables, such as cognitive similarity and understanding. In addition, there is a need to look at the marital relationship over time. Such longitudinal data may provide information regarding specific kinds of changes which occur in the marital relationship and, as noted by McCarthy (1981) be more applicable to real life relationships. Although there has been some preliminary research which has looked at dyadic variables (e.g., similarity and understanding), research has neglected to look at changes in the marital relationship over time. In addition, there has been a need to examine what distinguishes satisfied couples from dissatisfied couples in terms of changes in specific variables over time. Kelly's Personal Construct Theory (1955) is a promising approach to operationalizing these new research perspectives because it allows one to focus on dyadic variables over time to detect changes in relationships. To date, however,

PAGE 21

12 there has been only limited application of Kelly's theory to the study of the marital relationship (Richardson & Weigel, 1971; Ryle & Breen, 1972b; Wijesinghe & Wood, 1976; and Neimeyer & Hudson, 1985) and no research has applied this theory to the study of the marital relationship in its longitudinal context. In addition, no research has compared stable satisfied and stable dissatisfied control with couples who report changes in satisfaction. The above needs for investigating the marital relationship formed the backdrop for research in this study. General Research Questions This study is concerned with the following questions: 1. Do couples who are satisfied show greater cognitive similarity in their construct systems than couples who are dissatisfied? 2. Do satisfied couples show greater understanding than the dissatisfied couples? 3. Is there a change in the level of cognitive similarity and understanding over time for couples reporting changes in satisfaction not found in stable dissatisfied/ satisfied couples? Definition of Terms For the purpose of clarity, definitions of the following terms are provided.

PAGE 22

13 1. Choice corollary — The corollary in Kelly's Personal Construct Theory which describes individuals as choosing to elaborate their construct systems into an increasingly complex and comprehensive system for understanding experience. 2. Cognitive complexity — A variable introduced by Bieri (1955) which permits assessment of structural similarity. It refers to the degree of differentiation (i.e., functionally independent constructs [FIC] ) in an individual's construct system. 3. Cognitive Similarity — Commonality of constructs between two individuals. The term similarity is used interchangeably with cognitive similarity when referring to a dependent variable in this study. The different types of similarity are a. Construct similarity — Refers to the sharing of constructs with respect to content. b. Functional similarity — Refers to the common application of constructs by two individuals. c. Structural similarity — Refers to the degree of commonality in the organization of the constructs between two individuals' systems. 4. Commonality corollary — Kelly's corollary which states that there are certain constructs which individuals share. 5. Constructs — Social hypotheses represented as bipolar descriptions which distinguish the ways in which things are viewed as similar and different from others (Bannister and Mair, 1968) 6. Construct system — All an individual's constructs taken together. 7 Dissatisfied couple — A man and woman married or cohabiting who report desiring fourteen or more changes in their relationship. These scores are within 2 standard deviations of the mean for distressed couples on the Areas of Change Questionnaire as reported by Birchler and Webb (1977).

PAGE 23

14 8. Extension — Derived from the Sociality corollary, it is a concept referring to expansion of one's construct system. A means of extension is by increasing one's interpersonal understanding (Neimeyer & Neimeyer, 198 2) 9. Improved couple — A man and woman married or cohabiting who report increased satisfaction (i.e., at least seven fewer changes) on the Areas of Change Questionnaire at Time II. 10. Individuality corollary — Kelly's corollary which states that each individual has his/her own way of interpreting the world and it is those particular constructs which account for a person's individuality 11. Predictive accuracy — The accuracy with which individuals predict one another's construct ratings on the Marital Awareness Repertory Test. 12. Role relationship — The patterns of activities of an individual in relation to another. 13. Satisfaction — A term which refers to fulfillment of needs, expectations or desires of an individual or couple; contentment. It is operationalized in this study by satisfaction score based on the Areas of Change Questionnaire. 14. Satisfied couple — A man and woman who are married or cohabiting who report a desire for zero to thirteen changes in their relationship. These scores are within 2 standard deviations of the mean for satisfied couples on the Areas of Change Questionnaire as reported by Birchler and Webb (1977). 15. Similarity — Commonality of constructs between two individuals. Also referred to as cognitive similarity (see Cognitive similarity) 16. Sociality corollary — The corollary in Kelly's Personal Construct Theory which states that the extent to which an individual is able to assume the viewpoints of others (i.e., constructs) influences the development of interpersonal relationships

PAGE 24

15 17. Stable satisfied/dissatisfied couple — A couple whose reported dissatisfaction over time remains relatively stable. Whether satisfied or dissatisfied initially for this study, stable satisfied and stable dissatisfied may be referred to as satisfied and dissatisfied, respectively. 18. Understanding — A concept which describes a cognitive process by which one individual comprehends or has a clear perception of the meaning of someone or something. It is operationalized in this study by predictive accuracy between two partners. 19. Validation — Also called consensual validation, this term refers to the process by which individuals seek to confirm their social hypotheses. It is one way in which individuals can elaborate their construct systems. Organization of the Study This dissertation is organized into five chapters. Chapter II contains a review of the related literature. This review includes further discussion of Personal Construct Theory and research reporting application of this theory to describing interpersonal relationships in general and friendships and the marital relationship in particular. Chapter III consists of a discussion on research methodology, data collection, and analyses. Chapter IV includes the report of results of this study. Chapter V presents a discussion of the results, implications and recommendations based on the findings.

PAGE 25

CHAPTER II REVIEW OF LITERATURE Kelly has formulated a theory to explain individual psychological development and change in the course of natural development and/or treatment. In addition, he has developed an instrument for measuring and assessing this change. Recently, Kelly's Personal Construct Theory has been used to explain close interpersonal relationships such as friendships and marriages. The purpose of the present study is to examine Kelly's theoretical dimensions of cognitive similarity and understanding over time in couples reporting different levels of marital satisfaction. In this chapter, the original formulations of Kelly's Personal Construct Theory will be presented, including both his fundamental postulate and corollaries. In addition, research applications of Kelly's theory to interpersonal relationships will be reviewed. Personal Construct Theory Kelly (1955) has formally articulated one fundamental postulate and eleven corollaries for his theory. The central postulate states that "a person's processes are 16

PAGE 26

17 psychologically channelized by the ways in which he/she anticipates events" (p. 46) In addition, four of the eleven corollaries are related to the topic of close relationships and will be reviewed. The individuality corollary states that each individual has his/her own way of interpreting the world and it is those particular constructs which account for a person's individuality. The commonality corollary states that there are certain constructions which individuals share. The sociality corollary suggests that the extent to which an individual is able to assume the viewpoints of others (i.e., constructs) influences the development of interpersonal relationships. Lastly, the choice corollary views the individual as choosing to elaborate his/her construct system into an increasingly complex and comprehensive system for understanding experience. These corollaries suggest that an individual's constructs are keys to understanding his/her particular perceptions and behavior. Furthermore, Kelly (1955) believed that the ability to unlock another's phenomenological world would greatly enhance the understanding of individuals and interpersonal relationships. In addition, such understanding would have valuable therapeutic application in the treatment of distressed individuals and relationships. To gain access to an individual's construct system, Kelly devised the Role Construct Repertory Test, more commonly known as the Reptest or Rep Grid. This Rep Grid is

PAGE 27

18 a technique based on an individual's expression of his or her personal role constructs. A role is defined as a pattern of activity carried out in light of one's understanding of others (Kelly, 1955) As such, the Rep Grid provides a means for assessing change in an individual, not based solely on statistical methods, but as emphasized by Bannister and Mair (1968) on psychological processes of individuals studied. Originally designed as a clinical instrument to identify problems for the individual, the Rep Grid's use has been extended to the research on personal relationships (Duck, 1972; 1973a; 1973b; 1975; 1977). In addition to being used as an assessment tool to evaluate progress in individual and group therapy (Slater, 1970; Watson, 1970; Rowe, 1971; Winter & Trippett, 1977; Ryle, 1980, 1981), it has more recently been used to evaluate change in couples in therapy (Bannister & Bott, 1973; Ryle & Lipshitz 1975; Burns, Hunter & Lieberman, 1980). Use of a variety of Rep Grid instruments has been related to an extended Personal Construct Theory, which carries implications for intimate relationships Implications for Relationship Formation and Development Implications for Personal Construct Theory have been drawn most clearly from work regarding the formation of

PAGE 28

19 friendship by Duck (1973a; 1973b; 1975; 1977; 1979). This section discusses the cognitive processes which Duck and others regard as important for the development and maintenance of relationships and the literature relevant to each of these processes. After reviewing this literature, an extension of this work to marital relationships is presented. This author makes the assumption that the reason individuals seek relationships in the first place is pertinent to understanding interpersonal relationships. One such reason or motivation of individuals is eloquently set forth by H. Kelley (1979) in the following remarks: The unavoidable consequence of human social life is a realization of the essentially private and subjective nature of our experience of the world, coupled with a strong wish to break out of that privacy and establish contact with another mind. Personal relationships hold out to their members the possibility, though perhaps rarely realized in full, of establishing such contact, (p. 169) Another reason is presented by G. Kelly (1955) who viewed individuals as personal scientists who choose alternatives which appear to afford the greatest opportunity for personal elaboration (choice corollary) There are two aspects of personal elaboration: definition and extension. Definition allows one an awareness of oneself while extension permits

PAGE 29

20 awareness of others which, in turn, expands the self. In other words, this view of relationship development states that individuals choose relationships based on how well the relationship will enhance one's understanding of oneself and others. In addition, Kelly makes it clear that the development of relationships is not incidental to the process of understanding oneself and others, but essential to an individual's growth and dynamic change. According to Personal Construct Theory, there are at least two ways in which an individual decides which relationships provide the greatest potential for elaboration: validation and extension. The first way, validation, is derived from the commonality corollary; the second way, extension, is derived from the sociality corollary. Validation The commonality corollary states that while individuals are unique, some of them share common understandings. This commonality of personal constructs provides a measure of consensual validation for interactants perceptions. That is, as social hypotheses, personal constructs need to be validated (Duck, 1973a) Validation occurs when an individual obtains evidence of the correctness of his/her hypotheses. In this sense, validation is an aid in the process of definition.

PAGE 30

21 In drawing on these early insights, Duck (1973a; 1973b) hypothesized that individuals who share similar constructs would be more attractive to each other than dissimilar others. The reason for the attraction is that similarity provides consensual validation for one's own understanding and also facilitates communication. Empirical support for this view is provided by Byrne's (196 9) contention that research showed that similarity of attitudes does provide validation. According to Duck (1973a) this validation occurs at varying levels of construing as the relationship develops. Individuals may at first be concerned only with superficial/physical aspects of the person (e.g., whether one is attractive versus unattractive) When the relationship deepens, this concern may focus more on psychological aspects (e.g., ambitious versus lazy) of the person. So at each level of construing, individuals seek validation through construct similarity and make assumptions regarding deeper levels. When this construct similarity is inadequate at any level, Duck proposes that it acts as a filter, which prevents further relationship development. In other words, this filter hypothesis states that over the course of a relationship, individuals change the focus of their attention from a very general cognitive assessment of one another to

PAGE 31

22 an increasingly specific, differentiating one. Specifically, in developed relationships, individuals will be more concerned with personality attributes versus attitudes because this allows partners' awareness of similarities and differences in their construct system. Further, Duck proposes that a focus on personality variables later in the relationship is a way of distinguishing friends from nonfriends if his filter hypothesis is correct. As noted by Adams-Webber (1979) Duck successfully tested his view by assessing the construct similarity between pairs of friends and nominal pairs using the Reptest (which can measure specific personality characteristics) and the California Personality Inventory (which measures more general dimensions). The findings support Duck's position that friendship pairs were more similar in terms of personal constructs and that the Reptest was significantly more accurate in discriminating friendships from nominal pairs. In summary, it appears that validation through construct similarity is related to relationship development. Against this general backdrop, several different types of similarity have been examined. In particular, there are three forms of similarity which have been explored. The first concerns the similarity of constructs themselves. The second concerns the application of those constructs, and the third relates to the structure of the construct system.

PAGE 32

23 Construct Similarity Construct similarity refers to the sharing of constructs with respect to content, e.g., two individuals use the following constructs: honest versus dishonest, happy versus sad, ambitious versus lazy. As noted by Neimeyer and Neimeyer (1982) construct similarity is the most researched type of validation, due to the work of Duck (1973a; 1973b; 1975; 1977; 1979) and colleagues (Duck & Allison, 1978; Duck & Spencer, 1972; Lea, 1979) on friendship development. Basically, the findings in these friendship studies support a positive relationship between construct similarity and successful friendship development. Specifically, higher levels of construct similarity characterize friends versus nominal pairs (Duck, 1973a) or acquaintances (Lea, 1979) Similar findings have been reported in both cross-sectional (Duck, 1973b) and longitudinal studies (Duck & Spencer, 1972) with a variety of populations (Duck, 1973a; 1975) Further support for the importance of construct similarity in friendship development is its reported absence in deteriorated relationships. Duck and Allison (1978) note that less construct similarity characterizes individuals who choose to separate after living together one year than those who stayed together.

PAGE 33

24 Functional Similarity Functional similarity refers to the common application of constructs by two individuals. Using the above constructs, a husband and wife view the same people similarly (i.e., they describe the same individuals using the same pole of the construct: Susan is ambitious, Larry is dishonest, et cetera) Functional similarity, like construct similarity, is viewed as a source of consensual validation and, therefore, attractive to individuals. Initial work in assessing the relationship between functional similarity (FS) and attraction was done by Neimeyer and Neimeyer (1981) on a group of previously unacquainted individuals. They hypothesized that if FS provided consensual validation for the application of group members' constructs to one another's, then high FS dyads would show more mutual attraction than low FS dyads. Results supported their hypothesis. A second study investigated the relationship of FS to friendship development, i.e., whether or not it remained important. The amount of FS of close friends was compared to FS displayed by acquaintance pairs. Results showed that friends shared greater FS than acquaintances along psychological dimensions. In addition, acquaintances showed greater amounts of FS along physical dimensions and the

PAGE 34

25 least amounts of FS along psychological dimensions. These findings (i.e., greater FS along psychological dimensions for friends and along physical dimensions for acquaintances) lend further support for Duck's filter hypothesis. From these studies, it appears that FS is positively related to relationship development. An additional study by Neimeyer and Neimeyer (1982) investigated whether functional dissimilarity would be predictive of loss of attraction over time. The results suggest that initial functional dissimilarity may hinder relationship development while functional dissimilarity later may be less threatening to the partners. Structural Similarity The last type of similarity discussed here is structural similarity. This refers to the degree of commonality in the organization of the constructs between two individual's systems. One way of assessing structural similarity is by looking at cognitive complexity, a variable introduced by Bieri (1955) Cognitive complexity refers to the "degree of differentiation in an individual's construct system, i.e.., the relative number of different dimensions of judgment used by a person" (Tripodi & Bieri, 1964, p. 122). Structural similarity, in terms of cognitive complexity, has been reported in several studies. Bieri (1955) argued that the greater cognitive complexity of an individual the better

PAGE 35

26 that individual will be able to predict others. He provided empirical support for his argument by finding a significant positive correlation between his subjects' cognitive complexity scores and their ability to predict the responses of acquaintances on a social questionnaire. In addition, Adams-Webber, Schwenker, and Barbeau (1972) reported that one's cognitive complexity is related to the accuracy in making inferences about the personal constructs of others. Some studies (Leventhal, 1957; Sechrest & Jackson, 1961) do not support this hypothesis. Neimeyer, Neimeyer and Landfield (198 3) argue that the cognitive structure of the target person whose behavior is being predicted may be an influencing factor in these outcomes. Supportive evidence in favor of target complexity influence is provided by Adams-Webber (1969) who found that individuals were more accurate in predicting the behavior of less cognitively complex target individuals than more cognitively complex target individuals. Cognitive complexity has also been looked at in relation to friendship formation and maintenance. Duck (1972) assessed cognitive complexity as a means of differentiating friends and nominal pairs in attraction. Neimeyer and Neimeyer (1983) compared the cognitive complexity of acquaintance pairs. Results showed that structural similarity assessed by cognitive complexity level was important later

PAGE 36

27 (but not earlier) in the relationship? that is, cognitive complexity may be most important in long-term established relationships These empirical studies support the importance of validation through the different types of cognitive similarity within the construct systems. This leads to the second way in which individuals can decide which relationships provide the greatest elaborative potential: extension. Extension Eventually in the development of a relationship the need for expansion of the construct system becomes more important than that for consensual validation. In fact, it may be that the validation provided through construct similarity affords the individual the degree of psychological security needed to extend himself or herself (Neimeyer, 1981) A means of extension is through increasing interpersonal understanding (Neimeyer & Neimeyer, 198 2) Understanding in Personal Construct Theory terms is the ability to subsume the viewpoint of another and/or predict another. Kelly (1955) emphasized in his sociality corollary that this degree of understanding limits the role relation (i.e., patterns of activities of an individual in relation to another) For example, an individual in a relationship plays a role with respect to another person when he/she tries to make sense out of what the other person is doing.

PAGE 37

28 It is critical to the notion of role that the individual attempts to infer the view or outlook of another person. Bannister (1977, p. 25) reiterates that elaboration of one's own understanding of oneself and the world necessarily means developing new constructs and also escaping from old constructs. It is in the subsuming of another's constructions that one acquires new constructs and adds to his/her understandings of the other. The acquisition of new constructs by the individual implies support for the attractiveness of dissimilarity as suggested by McCarthy and Duck (1976) Goldstein and Rosenfeld (1969) Kerckhoff and Davis (1962) Winch, Ktsanes and Ktsanes (1954); i.e., increasing one's understanding involves seeking out others with differing views. In this sense, dissimilarity may be an important source for increasing interpersonal understanding. On the face of it, the attractiveness of dissimilarity appears to contradict the need for similarity hypothesis. However, there are several reasons that this is probably not so. First, the waxing and waning view of similarity (McCarthy & Duck, 1976) suggests that the need for both similarity and dissimilarity in a relationship is important at different times in the relationship. Second, this attractiveness of dissimilarity may actually be an

PAGE 38

29 increased tolerance of lack of similarity. Third, dissimilarity may itself be restricted to peripheral issues. As Adams-Webber (197 9) has noted, "Maslow (195 3) suggested that similarity between spouses' personalities may be more in terms of basic characteristics than superficial ones" (p. 111). In Personal Construct Theory terms, this similarity would refer to commonality along the basic/most important constructs (i.e., core role constructs) but not along peripheral constructs. In fact, findings by McCarthy and Duck (1976) support this view of attractiveness of dissimilarity on peripheral issues. Indeed, it may be that dissimilarity is only one of many sources for increasing interpersonal understanding. Empirical evidence supporting a positive relationship between understanding and attraction in friendship is reported in the literature (Neimeyer & Neimeyer, 1982) Using the Rep Grid, they compared the predicted and actual descriptions of a dyad. This yielded a predictive accuracy score which reflected the extent to which partners understood each other. Results showed that attraction was greater for more accurately predicted others throughout the stages of the interaction. In addition, individuals were more attracted to those who more accurately predicted them. This study suggests that not only understanding but the mutuality of understanding may be important in relationship

PAGE 39

30 development. The importance of this mutuality may be dependent upon the type of relationship (i.e., mutuality would be more important in the close versus the peripheral relationship) In summary, theoretically and, to some extent empirically, it appears that cognitive similarity and understanding are important to relationship development. However, whether similarity and understanding are related, and if so, how, has yet to be explained. Kelly deemphasized any relationship between similarity and understanding by stressing the difference between two people holding the same constructions system and two people understanding each other so they can play roles in relation to each other. Runkel (1956) found that when individuals were similar or perceived themselves to be so, understanding in the relationship was enhanced. Similar findings were reported by Triandis (1959) and Menges (1969) Smail (1972) found a positive relationship between similarity and understanding in a psychotherapy group. Duck (1973b, 1979) suggests that similarity of construct systems may facilitate understanding. While the relationship between similarity and understanding remains unclear, Neimeyer and Neimeyer (1982) suggest an interactive relationship in which understanding may qualify attraction initially based on similarity. Although the relationship between similarity and understanding is unclear, this model of relationships suggests

PAGE 40

31 their importance for relationship development and maintenance. The following section will discuss the implications of the Personal Construct Theory relationship model and the effect of similarity and understanding on the marital relationship. Application of Kelly's Theory to the Marital Relationship Only recently have researchers suggested applying this relationship model to the marital relationship (Richardson & Weigel, 1971; Neimeyer & Hudson, 1985). Marriage from a construct theory perspective is a comprehensive relationship which involves the development of many different subsystems (i.e., social, sexual, emotional, et cetera) at varying levels of intimacy. The marital role is defined as a process of behavioral interactions which occur in the light of one partner's understanding of the other (Kelly, 1955) This process can also be viewed as a temporal interactive one because partners' understanding change over the course of the relationship (Duck & Sants, 1983) Findings in the friendship studies suggest that like friendships, successful marital relationships will be characterized by greater validation and extension than unsuccessful ones. To the extent that validation and extension are processes of the role, one would expect that factors affecting these processes and/or the ability to play

PAGE 41

32 one's role (e.g., similarity and understanding) will influence the success of the relationship. The comprehensive nature of marriage itself would seem to demand greater understanding between the spouses, yet at the same time, make it more difficult to obtain. Specifically, marriage as a multileveled system and changing interactive process offers great elaborative potential, yet also great opportunity for misunderstanding. Recall that elaborative potential is a basis for choosing to form relationships in the first place. Relationships in turn are made up of individuals who are at the same time unique (individuality corollary) while sharing some common outlooks (commonality corollary). This gap between one's uniqueness and commonality with others may result in misunderstanding as a consequence of relationship formation and development. Misunderstanding in Personal Construct Theory terms is when two individuals construe and experience differently and each is unable to see the other's point of view. This jeopardizes the relationship because it prevents the person from playing his/her role. In addition, a multileveled system may react on different levels to a specific misunderstanding; that is, a conflict on an economic level may affect the sexual level of the couple's life. Theoretically, it seems important that couples understand one another. In addition, one might expect that the

PAGE 42

33 less discrepancy between partners' understanding of one another, the more successful the role relationship because each partner is approximating the other's understanding of his/her role. This parallels Delia's (1980) thinking that "relationships are not simply a matter of one partner's expectations; the extent to which the other partner satisfies those expectations, and shares them is also involved" (p. 101) Empirical research on the positive effect of understanding in marital relationships has been reported by Christensen and Wallace (1976) They compared maritally adjusted couples, divorcing couples, and couples in counseling on their ability to predict the rewarding effects of certain behaviors on their spouse. The maritally adjusted groups were always more accurate than the other two groups in their understanding of their spouse. Similar findings on understanding in the marital relationship are reported by Laing, Phillipson and Lee (1966) and Dymond (1954) Harvey and Wells (as cited in Harvey, Wells & Alvarez, 1976) also questioned the accuracy in understanding of distressed couples. They found evidence of both overestimation and underestimation for males' and females' predictions of each other. Even in relatively long-term relationships, individuals tended to be inaccurate in their perceptions of

PAGE 43

34 how their partners see the world in terms of areas of conflict. In addition to couples differing in their views about the important sources of conflict, this study suggested that couples were unaware of these differences. The authors concluded that "a lack of understanding of divergence is as critical or more critical to the viability of close relationships than is the actual divergence" (p. 247) ; that is, partners are less upset by actual disagreement than when they believe their spouses do not understand them. However, attaining an accurate understanding between spouses may be particularly complex in a relationship in which one partner is psychologically disturbed. The findings of Wijesinghe and Wood (1976) on 4 married couples in therapy support this view. They suggested that the ill partners may behave in such a way that prediction by the other spouse is more difficult. In an early formulation of Personal Construct Theory applied to marital relationships, Richardson and Weigel (1971) emphasized the role of sociality (understanding) and commonality (similarity) as predictors of marital satisfaction. Specifically, they hypothesized that satisfaction was dependent upon the ability of both partners to correctly grasp each other's outlooks. They also hypothesized that this mutual understanding may be dependent upon a commonality between spouses' construct systems. This facilitative

PAGE 44

35 function of similarity on understanding parallels Duck's (1977) view. Theoretically, it seems that spouses' similarity may have a positive influence, if only to decrease the opportunity for misunderstanding to occur. On the other hand, dissimilarity between spouses would seem at least initially to increase the likelihood of misunderstanding. However, Richardson and Weigel (1971) were unable to confirm these hypotheses. Adams-Webber (1979) suggests that had they elicited the personal constructs from each individual couple instead of providing a standard list for all couples, results might have been more favorable; that is, elicited constructs are more likely to include a couple's basic (core role) constructs than a standard list. A recent study paralleling the thinking of Richardson and Weigel and extending Personal Construct Theory work on similarity and understanding to the marital relationship is reported by Neimeyer and Hudson (1985) They hypothesized that satisfied couples would share greater validation (as provided by greater similarity in construct systems) and greater understanding (as provided by accuracy of predictions) They compared the similarity and understanding of 10 satisfied couples with 10 dissatisfied couples. Level of satisfaction was determined by the score on Areas of Change Questionnaire (Weiss, Hops & Patterson, 1973). Similarity (i.e., construct and functional similarity) and predictive

PAGE 45

36 accuracy were assessed from responses of spouses on Kelly's Reptest. Briefly, their findings noted greater construct and functional similarity for satisfied couples than for dissatisfied couples. In addition, satisfied spouses understood each other more accurately and showed less discrepancy between partner's levels of understanding (i.e., greater mutuality of understanding) than dissatisfied couples. In another recent study by Neimeyer (1984) the structural similarity of 20 couples was evaluated in terms of their cognitive complexity as measured by functionally independent constructs (FIC) scores (Landfield, 1977) The hypothesis that marital satisfaction was related to the similarity between husband and wife's level of complexity but not to the absolute level of couple complexity was confirmed. That is, couples who were similar (partner's low level of discrepancy) to each other in terms of FIC score, no matter what the absolute couple score, were more satisfied than those who were less similar. The above literature suggests support for this general model of the marital relationship. The following section will summarize this literature and present ways in which this model will be extended and tested in the present study.

PAGE 46

37 Summary Successful marriages have been characterized theoretically as mutually elaborative. That is, spouses encourage each other's development by validating and extending their personal construct systems. The empirical evidence reviewed in this section is consistent with this formulation. Specifically, satisfied spouses were more validating and understanding in their relationships than were dissatisfied partners (Neimeyer & Hudson, 1985). That is, satisfied couples were more similar in the content and application of their construct systems than were the dissatisfied couples. This is consistent with a parallel literature in friendship formation (Duck, 1973; Duck & Allison, 1978; Duck & Spencer, 1972; Neimeyer & Neimeyer, 1981). As with friendships (Neimeyer & Neimeyer, 1983) it seems likely that satisfied couples may also organize their construct systems similarly. Furthermore, each of these types of similarity (i.e., content, functional and structural) validates the construct system. As partners confirm the validity of some of their hypotheses, they become more open to another's differing views and better able to subsume these views. This is supported by the findings that more satisfied spouses were better able to understand one another. These satisfied couples also showed more mutual understanding, while dissatisfied spouses were more discrepant in their understandings

PAGE 47

38 of one another. Such lack of mutuality jeopardizes interpersonal elaboration because it limits the partners' ability to contribute to their joint development in a continuous and reciprocal way (Neimeyer & Hudson, 1985) The premise of this study is that this theoretical model and related empirical literature may have implications for couple treatment. Specifically, couples in conflict may be helped in therapy by increasing their perception of their similarities and by broadening the understanding of each spouse by helping them to see the relationship from the other's viewpoint. With such increased understanding, one might realistically expect the couple to become more similar in the content, application and structure of their constructs, as well as become better able to accurately predict the spouse. To detect such changes, this study will conduct an analysis of the differences between couples who remain stable on reported satisfaction and couples who report changes in satisfaction over time.

PAGE 48

CHAPTER III METHODS This study was designed to compare levels of cognitive similarity and understanding in a sample of 4 7 couples reporting different levels of relationship satisfaction. In addition, it will compare changes in these dimensions over time for stable satisfied couples (n = 21) stable dissatisfied couples (n = 11) and couples reporting improvement in satisfaction (n 15) Levels of satisfaction will be assessed by using the Areas of Change Questionnaire (Weiss, Hops, & Patterson, 1973). Cognitive similarity and the measure for understanding will be measured through use of the Marital Awareness Repertory Test (see Appendix A) a form of the Exchange Grid (Thomas, 1979). Each of these instruments were administered to all groups on two separate occasions several weeks apart. This chapter is divided into several sections. In the first section, the research approach, design and variables of interest are discussed. In the second section, the population and sample of interest in this study and sample selection and recruitment procedures are discussed. In the third section, the instruments used in this study are presented. In the fourth section the procedure involved and 39

PAGE 49

40 data collection and recording of the data are described. In the fifth section, the research hypotheses for this study are presented. In the sixth section the methods of data analyses are presented. Limitations of the present study and summary of the foregoing information are presented in the final two sections. Research Design This study employed a descriptive design. Data were collected on the variables of cognitive similarity and understanding over time for three groups of couples reporting different levels of relationship satisfaction. Differences across three groups of couples were examined for each of the two variables at two different times. Using a 3 x 2 factorial design (Figure 1) the first factor is a between groups variable, couple satisfaction. Satisfaction was operationalized in terms of the total number of desired changes in the relationship based on responses of the couple on the Areas of Change Questionnaire (Weiss, Hops & Patterson, 1973) According to this definition, couples desiring the least number of changes are most satisfied and those couples desiring the most changes are least satisfied. The three levels of this variable are 1) stable satisfied, 2) stable dissatisfied, and 3) improved. The stable satisfied classification refers to couples who score in a

PAGE 50

41 SATISFACTION LEVEL TIME I TIME II Stable Satisfied Stable Dissatisfied Improved FIGURE 1. LEVEL OF SATISFACTION comparably satisfied range on the Areas of Change Questionnaire (Weiss, Hops, & Patterson, 1973). The stable dissatisfied classification refers to couples who score in the dissatisfied range on the same instrument. The improved classification reflects couples who change (improve) over time in level of reported satisfaction. For the purposes of this study, stable satisfied and stable dissatisfied may be referred to as satisfied and dissatisfied, respectively. The second factor, time, is a within subjects variable. Time I refers to intake administration of the instruments. Time II refers to the post-test or second administration of the same instruments. The interval between Time I and Time II was 8-10 weeks for all groups.

PAGE 51

42 The variables of interest in the present study included the couples' 1) understanding, and 2) cognitive similarity. The variable understanding will be operationalized in terms of predictive accuracy derived from the Marital Awareness Repertory Test (MART) It refers to the amount of understanding as measured by the accuracy with which individuals predict one another's ratings. The predictive accuracy score (Landfield, 1977) is calculated by determining the absolute difference between individuals' own ratings and their partner's predictions of their ratings. Lower discrepancy scores indicate better understanding. It was expected that couples who understood each other better would have greater predictive accuracy. The variable cognitive similarity was derived from the MART. It was operationalized in terms of functional and structural similarity of partners' construct systems. Functional similarity (Neimeyer & Neimeyer, 1981) refers to the degree of similarity in applying the constructs to particular others. It was measured by determining the absolute difference between partners' ratings for all six elements (e.g., mother, father, et cetera) for all the six constructs of both the husband and wife. Scores could range from 36 (no functional similarity) to (in the

PAGE 52

43 case of functional identity) when using only one partner's constructs rating one element. A lower score indicates less discrepancy between partners' ratings and, therefore, greater functional similarity. Structural similarity refers to the overall organization of the construct system. One method of assessing structural similarity is estimating the individual's cognitive complexity. Operationally, cognitive complexity was defined as the number of nonidentical ratings or functionally independent constructs (FIC) of a set of stimulus persons along a group of bipolar constructs (e.g., happy versus sad). The greater the number of nonidentical (i.e., unique) ratings between all possible pairs of constructs, the more cognitively complex is the subject (Bieri, Atkins, Briar, Leaman, Miller, & Tripodi, 1966) Possible FIC score ranges between 2 and 12 for each respondent (see Landfield, 1977, for details of scoring procedures). Structural similarity in this study referred to a discrepancy score between partners' levels of cognitive complexity with lower scores reflecting greater structural similarity. Description of the Sample The population of interest in this study was all married and cohabiting graduate and undergraduate students living in the university community of Gainesville, Florida.

PAGE 53

44 Participation in this study required that couples be married or living together for a minimum of 3 months and a maximum of 15 years. Ages of the partners were restricted to individuals between 18 and 40 years. In addition, at least one of the partners was to be a graduate or undergraduate student. The sample consists of married and cohabiting couples. These couples were drawn from a group within the university community who responded to a newspaper advertisement. The advertisement requested couples who were interested in participation in research on relationships in exchange for theater tickets. A total of 50 married or cohabiting couples completed the test forms. Of these, three couples had only one partner who entirely completed all of the forms. Therefore, 47 couples (N = 94 individuals) who completed the Areas of Change Questionnaire, the Marital Awareness Repertory Test, and a demographic questionnaire at both administrations were included in the analyses. These couples formed 3 separate groups of varying number: 1) stable satisfied group (n = 21) 2) stable dissatisfied group (n = 11) and 3) improved group (n 15) Couples 1 group assignment was based on level of couple satisfaction as assessed by scores on the Areas of Change Questionnaire (ACQ) Couples were rank ordered according to

PAGE 54

45 their ACQ score and a median split was performed on the basis of these scores. The lower half of the split constituted the satisfied group; the upper half constituted the dissatisfied group. Specifically, the satisfied group wanted to 13 changes in their relationship and the dissatisfied group wanted 14 or more changes. Assignment of a couple to stable satisfied group or stable dissatisfied group indicated that there was little or no change in couple reported satisfaction over time. Assignment of a couple to the improved group indicated a change in reported satisfaction at Time II administration. These improving couples reported wanting at least seven fewer changes in their relationship at Time II (i.e., these couples were more satisfied at Time II) A minimum of seven changes will be used because it reflects the average of the standard deviation for satisfied (5.3) and dissatisfied (9.43) groups by Birchler and Webb (1977). Therefore, these couples will represent a group which changes in a positive direction by at least one standard deviation on the Areas of Change Questionnaire. Mean satisfaction scores and standard deviations for all groups in this study are listed in Table 1.

PAGE 55

46 TABLE 1. MEAN SATISFACTION SCORES AND STANDARD DEVIATIONS FOR ALL THREE GROUPS AT TIME I AND TIME II

PAGE 56

47 other words, the satisfied group was more satisfied than the dissatisfied group, and the improved group became more satisfied over time. All couples' satisfaction scores at Time I were within 1.5 standard deviations of the mean for both the satisfied and dissatisfied couples as reported by Birchler and Webb (1977) Mean scores and standard deviations for the satisfied and dissatisfied couples as compared to the couples' scores in Birchler and Webb's normative study are reported in Table 3. TABLE 3. MEAN AND STANDARD DEVIATION FOR THE SCORES ON THE AREAS OF CHANGE QUESTIONNAIRE (ACQ) FOR THIS STUDY SAMPLE AND THE NORMATIVE SAMPLE THIS SAMPLE DISSATISVARIABLE SATISFIED FIED MEAN SD MEAN SD BIRCHLER & WEBB SATISFIED MEAN SD DISSATISFIED MEAN SD ACQ 8.04 5.49 22.63 8.62 6.90 5.3 28.46 9.43 n = 21 n 11 'n = 50

PAGE 57

48 Results of the demographic questionnaire (see Appendix B) administered to all subjects in the sample indicated that the mean age was 26.3 years. Ages across all three groups were comparable within 5 years (stable satisfied, M=28.7; stable dissatisfied, M-24; improved, M=25.4). At least one partner of each couple was a graduate or undergraduate student attending the University of Florida or Santa Fe Community College. Thirty-four couples were married and 13 were cohabiting. Ninety-three participants in the study were Caucasian and one female participant was oriental. Only 6 of the 47 couples in this sample had children. Mean number of years for cohabitation for the total sample was 3.1, and was comparable across the three groups (3.5 years for the Stable Satisfied; 3 years for the Dissatisfied; and 2.9 years for the improved). These means are presented in Table 4. TABLE 4. MEAN AGE, EDUCATION AND LENGTH OF COHABITATION FOR COUPLES SATISFACTION GROUPS VARIABLE SATISFIED MEAN SD DISSATISFIED MEAN SD IMPROVED MEAN SD AGE EDUCATION COHABITATION a n 21 n 11 28.7 (4.4) 17.3 (2.4) 3.5 (2.6) 24.0 (4.0) 16.3 (2.0) 3.0 (1.7) 25.4 (3.5) 16.0 (1.5) 2.9 (1.9) 'n 15

PAGE 58

49 Instrumentation Three instruments were used in this study: 1) the Areas of Change Questionnaire, 2) the Marital Awareness Repertory Test, and 3) Demographic Data Questionnaire. Areas of Change Questionnaire The Areas of Change Questionnaire (ACQ; Weiss, Hops & Patterson, 1973) is a self-report inventory which assesses marital partners' presenting complaints, their perceptions of one another's complaints, and perceptual accuracy between actual and perceived complaints. This instrument provides information on 1) desired change (i.e., what types of changes each partner actually wants from the other) 2) perceived change (i.e., what types of changes in oneself are believed to be desired by the partner) and 3) perceptual accuracy (i.e., the extent to which each spouse correctly identifies the specific items that the partner wants changed and also the direction of that desired change. Designed as a behaviorally specific measurement of marital complaints, this instrument was developed to identify problem areas for treatment and to evaluate treatment (Weiss, Hops & Patterson, 1973). As noted by Margolin, Talovic, and Weinstein (1983) the Areas of Change Questionnaire is consistent with recommendations for choosing practical clinical measures outlined by Nelson (1981) because it contains a large number of items representing a

PAGE 59

50 broad range of possible complaints and is sufficiently specific that measurement of change is possible. Approximate time for completion of this inventory is 20 minutes. Psychometric properties of the Areas of Change Questionnaire have been examined in respect to overall change score by Birchler and Webb (1977) They found mean scores on the ACQ to be 28.46 (SD = 9.43) and 6.90 (SD = 5.3) for 50 distressed (dissatisfied) and nondistressed (satisfied) couples, respectively. Weiss et al. (1973) obtained an internal consistency index of .89 for the scale. Also, there appears to be a significant overlap in the dimensions of marital satisfaction measured by the ACQ and Marital Adjustment Scale (MAS) (r = -.70) by Margolin and Wampold (1981) and Weiss et al. (1973). In addition, this instrument has been used in treatment outcome studies and has been shown to be a sensitive index of pretreatment to posttreatment change (Baucom, 1982; Margolin & Weiss, 1978; Weiss et al. 1973) Further psychometric information about the ACQ is provided in a study by Margolin, Talovic and Weinstein (1983) They state that "the ACQ proves to be a cost efficient yet comprehensive tool for discriminating global levels of marital adjustment, for identifying therapeutic targets, and for measuring changes in marital complaints over time" (p. 930)

PAGE 60

51 Their contrasted group approach provided evidence of validity for four of the five ACQ scores. In addition, 33 of the 34 items discriminated distress versus nondistress for one or both sexes. All three summary scores showed concurrent validity with the modified Marital Adjustment Scale while the perceptual accuracy measures were relatively weak. However, less evidence for convergent validity with reports of daily marital events was provided by this study. The Areas of Change Questionnaire consists of 68 statements which refer to behavior in which relationship dissatisfaction can arise (e.g., showing affection, money management, recreational activities, work habits, et cetera). On the first 34 statements, each partner is asked to indicate how much change he or she would like to see in the spouse along a 7-point rating scale (e.g., I want my partner to spend time with me: much less -3 -2 -1 -0, +1 +2 +3 much more). On the second 34 statements, they are also asked to indicate the changes they think their spouse would like to see in them along those same items. This instrument assesses how much change each partner seeks in the relationship and how accurately each partner understands the changes desired by the other partner. The total number of changes each partner desires is summed for a couple score, but it can be subdivided into the number of changes which partners agree and number of changes on which they disagree. The total number of changes which the couple

PAGE 61

52 desires reflects the degree of marital distress (possible range to 68) Higher scores indicate greater dissatisfaction. The Marital Awareness Repertory Test The Marital Awareness Repertory Test (MART; see Appendix B) is a form of the Exchange Grid (Thomas, 1979). "All forms of the Repertory Grid technique are derivatives of an original proposed by G. A. Kelly (1955) as an integral part of the development of Personal Construct Theory" (Bannister, 1965, p. 977). This repertory test is a form of a sorting test but differs from conventional sorting tests in that there are no standard sorting materials or sorting categories, any standard single form of administration or scoring procedure. This test permits measurement of the relationship between sorting categories (constructs) for the subject, not the correctness of these constructs. Several facets of the psychometric properties of the Rep Grid have been investigated. Issues of reliability and validity have been addressed by Hunt (1951) Mitsos (1958) Fjeld and Landfield (1961) Bannister (1965) and Slater (1965) "The assumption underlying all forms of the repertory grid is that the psychological relationship between any two constructs for a given subject is reflected in the statistical association between them when they are used as sorting categories by the subject" (Bannister, 1965, p. 977)

PAGE 62

53 The occurrence of significant association between constructs is regarded as establishing the internal consistency of a grid (Slater, 1965) Lack of internal consistency is equated with the notion that an individual's constructs are totally unrelated to one another. The unlikeliness of this makes testing for it a trivial matter (Slater, 1965) Test-retest reliability is examined in terms of consistency of ratings by subjects along specific constructs. In a study by Fjeld and Landfield (1961) there was a strong tendency for individuals to rate the same acquaintances the same way when given the same constructs (contingency coefficient .80). Validity of the Rep Grid has been addressed in terms of representativeness of the elements (Mitsos, 1958) and consistency of elicitation of construct labels (Hunt, 1951; Fjeld & Landfield, 1961) Representativeness of the elements is necessary if the test is supposed to indicate subjects' role relationships. That is, if the subject develops his/ her role in the light of his/her understanding of others, those people used as elements in the grid should be representative of all the people with whom the subject must relate (Kelly, 1955) Mitsos (1958) confirmed his hypothesis that when elements are representative, construct labels elicited by the Rep Grid will be consistent across time (pi .02)

PAGE 63

54 In another study, Hunt (1951) found consistency of constructs after one week using different but still representative elements (i.e., the two test elements were matched as to sex, relatives, authority, et cetera) He found an average percentage of agreement in constructs of 69 percent. Fjeld and Landfield (1961) report that subjects (N = 20) in their study were asked to retake the Rep Grid after two weeks. Given neither the original elements nor constructs previously elicited, subjects showed high agreement between constructs on the two tests and also high agreement on the acquaintance used (Pearsonian r for constructs = .80; percentage of agreement on acquaintances (elements) = 72 percent) Other issues of validity and reliability of the Rep Grid have been discussed by Slater (1965) and Bannister (1965) The Marital Awareness Repertory Test, a 6 x 6 grid, involves examining similarities and differences among several important interpersonal figures in the individual's life (e.g., mother, father, partner, partner's parents) and then comparing his/her responses to those of his/her partner. More specifically, using the names of both partners and their parents as elements in the grid, each partner identifies how two of three of the above elements are alike with a word or phrase. Then the individual identifies the opposite of such an identification for him/her. For example, the partner views the husband and father alike

PAGE 64

55 because they are both ambitious. The opposite of this for that individual may be lazy. This ambitious versus lazy is a construct, and five more such constructs are elicited. Then each element (i.e., the partners and both their parents) are individually rated along each of these constructs using a 7-point Likert-type scale (e.g., ambitious 3 2 10 1 2 3 lazy) In addition, partners are asked to predict their spouse's ratings along those same constructs. Partners then exchange grids and, without access to the other's prior ratings, repeat the procedure using the partner's constructs, first completing their own rating and then predicting their partner's rating. Thus, this instrument not only provides information regarding each partner's constructs, but allows identification of major points of disagreement in relation to particular constructs and elements. Estimated time for completion is 45 to 60 minutes. Demographic Data Questionnaire The Demographic Data Questionnaire (see Appendix B) is a data sheet developed by the researcher to assess demographic characteristics of the sample to be studied. The use of this questionnaire will allow the researcher to assess sex, race and age of participants, the length of time the couple has been married or cohabiting, and participants' educational levels and number of children. For the purposes

PAGE 65

56 of this study, husband and wife denoted role relationships and not legal status. Procedures Each couple in the study was given an informed consent form specifying the nature of the study and requesting release of information of these materials for confidential use in research (see Appendix C) Procedures were set up to insure standardization of administration. The researcher trained undergraduate psychology students in administration of these questionnaires to insure proper completion. In addition, these proctors were advised of the particular order of the tests, how to answer possible questions of participants and standard method of checking completed tests. Each participant was given the questionnaires with written instructions on how to complete the questionnaire. In addition, a proctor was available to answer participants' questions. After testing, participants were urged by the proctor not to compare their ratings as they remember them. The MART and ACQ were administered to all groups at two separate times, 8-10 weeks apart. For the MART only, all partners were asked to fill out two each time, one for themselves and one in which they predict how their spouse would respond.

PAGE 66

57 All couples participating in the study were given a couple number code which identified their responses while maintaining their anonymity. Time Frame of Study The following time frame for completion of this study was established: Week 1 Train proctors and advertise for couples. Week 2 First administration of questionnaires to respondents of advertisement. If inadequate number of couples respond, will readvertise. Week 3 Administration continues for new couples. Begin to score questionnaires from first administration Weeks 4-6 Continue to administer tests to new couples until N = 50. Score all data. Week 7 Contact couples and set up appointments for second administration. Weeks 8-10 Second administration begins. Ensure that this administration does not coincide with final and mid-term examinations.

PAGE 67

58 Research Hypotheses The following hypotheses were tested in this study: Hypothesis I ; 1) There is greater cognitive similarity in terms of application (FS) and structure (FIC) of the construct systems of stable satisfied couples and improved couples than stable dissatisfied couples as measured by the MART. 2) There is an interaction between level of satisfaction and time for improved couples such that these couples show significant improvement in level of cognitive similarity over time relative to stable couples. Hypothesis II : 1) There is greater understanding (PA) as measured by the MART in stable satisfied couples and improved couples than stable dissatisfied couples. 2) There is an interaction between level of satisfaction and time for improved couples such that these couples show significant improvement in level of understanding over time relative to stable couples.

PAGE 68

59 Analyses of Data The analyses compared satisfied, dissatisfied, and improved couples. Eight grids (MART) from each couple were analyzed to establish differences generated by each couple. A series of 3 x 2 analyses of variance was performed to determine the main effect of level of satisfaction, time and interactions. Independent and dependent sample t-tests were used in the analyses for the purpose of isolating significant differences. Limitations Limitations of this design included the following: exclusive use of a student population, use of volunteers, problems associated with longitudinal research and procedural difficulties. The exclusive use of university student population can bring into question the possible generalizability of the findings to other than student populations. In addition, the very fact that they are generally an educated population also limits the findings. Specifically, this population, because they are educated and attending a university which offers couples counseling may be biased towards seeking earlier treatment. Therefore, dissatisfied couples may actually be less distressed and/or more amenable to change than other populations. While generalizability of results

PAGE 69

60 may be limited, it is expected that the real value of this research will lie in distinguishing the differences among the varying levels of satisfaction over time. Couples selected for this study were volunteers and may not have resembled the general population of couples. Therefore, generalizations of the findings of this study to nonvolunteers should be done with caution. The problems associated with longitudinal research which may have affected the results in this study included intervening events and attrition. One such event which might affect outcome in a student population is examinations. Therefore, precautions were taken to prevent second administration of the questionnaires from coinciding with an examination period. Another problem in this design was the probability of attrition. This was minimized by selecting a relatively short interval between administrations (8-10 weeks) In addition, all subjects were encouraged after their first administration to return for the second administration. They were also reminded that they would receive free movie tickets at the second administration. Procedural problems in this design included variations in test-taking ability and compliance. For example, on the Areas of Change Questionnaire, distressed couples may be reluctant to self-disclose on the 68-item questionnaire and

PAGE 70

61 may actually minimize the number of changes desired. These couples will still be distressed but appear to be less so based on their responses. In this case, inaccurate self report may have resulted in improper group assignment. Another procedural difficulty may have been noncompliance with directions for couples to avoid discussion of their responses on the first administration of the questionnaire. The probability that this was a significant problem was reduced because of the complexity of the task and the length of time between administrations. However, if an individual had a particularly good memory or compared responses with his/her partner immediately after the testtaking, these could have influenced second administration responses. However, this seemed unlikely. Finally, use of married couples and cohabiting couples in the same study may present problems, although there is differing opinion on this in the literature. Cole (1976) and Yllo (1978) point out there has been no difference found between marrieds and cohabitors with regard to contentment and satisfaction. One argument for not comparing these couples is level of commitment may be different (Hinde, 1981; Newcomb, 1981) but other thinking (Montgomery, 1973) suggests that cohabitors need more dyadic commitment to remain intact than do marrieds because "cohabitors have fewer external unifiers and more external disrupters that act against the relationship than do marrieds" (p. 149)

PAGE 71

62 In summary, there were a few sampling and procedural limitations of this study. This author acknowledges these limitations and believes they were kept to a minimum. Summary This study was designed to test and compare changes in couples who report changes in relationship satisfaction with couples classified as stable satisfied and stable dissatisfied. Classification of couples is based on the couples' satisfaction score on the Areas of Change Questionnaire (Weiss et al. 1973) The design is a descriptive 3x2 factorial design. Factor I is level of couple satisfaction: stable satisfied, stable dissatisfied, and improved. Factor II refers to time: Time I was first administration, Time II was second administration. The variables of interest were 1) understanding and 2) cognitive similarity. The sample was drawn from a university community population. It consisted of 47 married and cohabiting couples of which at least one partner is a graduate or undergraduate student. These couples formed three distinct groups: 1) stable satisfied, 2) stable dissatisfied, and 3) improved

PAGE 72

63 Instruments used in this study include the Marital Awareness Repertory Test, the Areas of Change Questionnaire and the Demographic Information Questionnaire. Procedures for collection and evaluation of data were standardized in order to prevent bias between couples or groups. Limitations of this study included exclusive use of a student population, use of volunteers, problems associated with longitudinal research, and procedural difficulties. Analysis of data compared the stable satisfied and dissatisfied couples. A 3 x 2 ANOVA was performed to determine the main effects of level of satisfaction, time and interactions.

PAGE 73

CHAPTER IV RESULTS The purpose of this study was to test and compare the degree of cognitive similarity and understanding in couples who report changes in relationship satisfaction with couples who remain satisfied or dissatisfied. To accomplish this, the Areas of Change Questionnaire and the Marital Awareness Repertory Test were administered to 47 couples at two separate times several weeks apart. Analysis of the collected data involved use of a series of 3 x 2 Analyses of Variance to determine the effects of levels of satisfaction across time. This chapter presents the results of the data analyses testing the two hypotheses. Analyses of Hypothesis I Two different research hypotheses were postulated for this study. For each of the hypotheses, two different subhypotheses were proposed. The first sub-hypothesis involved mean effects, while the second addressed interaction effects for each variable. The first prediction stated that there would be greater cognitive similarity (FS; FIC) in the construct systems of stable satisfied and improved couples than stable dissatisfied couples as measured by the MART. 64

PAGE 74

65 The second prediction (Hypothesis I, Section 2), was that the improved couples would significantly increase their level of cognitive similarity at Time II relative to stable (satisfied and dissatisfied) couples. To test for the accuracy of these predictions, a twoway analysis of variance was performed on the mean scores, along two independent variables (functional similarity and functionally independent constructs) The factors reflected three levels of couple satisfaction (stable satisfied, stable dissatisfied and improved) and two levels of time (Time I and Time II) Results from analysis of the Functional Similarity (FS) variable indicated there was no main effect for groups. The difference in the means for groups (M ,...., = 78; M -, • • • -, = 92; M. j 77) while in the predicted dissatisfied improved v direction, was not statistically significant, F. ... •= 2.82, p ^ .07. (Note: Since this discrepancy score reflects a difference between partners' ratings, a lower score denotes greater similarity.) Results of this ANOVA also indicates that there was a significant main effect for time, F._ ... *= 4.85, p — .03. Mean scores and standard deviations for all couple groups at Time I and Time II are presented in Table 5. Results also show there was no significant interaction between groups and time for the measure FS. The differences

PAGE 75

66 between the means for groups (M = 92; M. ^^^ 77) and the m improved = 78) were in the predicted direction, but these differences were not statistically significant, F (2 ... = 2.53, p £: .09. The ANOVA data on FS are presented in Table 6. satisfied ~ dissatisfied 92; improved 77) and the means for time (M TI = 84; M TII TABLE 5. MEAN SCORE AND STANDARD DEVIATION FOR FUNCTIONAL SIMILARITY AT TIME I AND TIME II SATISFIED' DISSATISFIED IMPROVED TIME I TIME II 79 (20) 78 (13) 94 (25) 90 (22) 84 (21) 70 (18) n 20 n = 12 'n 15 TABLE 6. ANOVA USING FUNCTIONAL SIMILARITY SOURCE SS df MS MEAN

PAGE 76

67 Results for the 3x2 ANOVA on the second measure of cognitive similarity, Functionally Independent Constructs (FIC) showed no significant main effects for group or time, nor was the interaction between the two factors significant. The ANOVA data on FIC are presented in Table 7. TABLE 7. ANOVA USING FUNCTIONALLY INDEPENDENT CONSTRUCTS SOURCE

PAGE 77

68 To test the correctness of these hypotheses, a 3 x 2 ANOVA was performed on the mean predictive accuracy scores for each couple group. Results showed that the difference in the means for PA scores for groups (M ^ ,. = 156; ^ c satisfied "dissatisfied = 178; "improved = 152)(see Table 8) was in the predicted direction. However, differences did not attain statistical significance, F,~ ... 2.34, p £ .10. Mean predictive accuracy scores for all groups at Time I and Time II are presented in Table 8. TABLE 8, MEAN SCORES AND STANDARD DEVIATION FOR PREDICTIVE ACCURACY AT TIME I AND TIME II SATISFIED DISSATISFIED IMPROVED TIME I TIME II 157 (31) 155 (31) 182 (42) 175 (39) 170 (47) 136 (31) n = 20 n 12 'n 15 In addition, there was a significant main effect for Time, F._ = 8.73, p — .005, such that predictive accuracy scores increased (note: lower discrepancy scores indicate better understanding) from Time I [M 166.91 (39.92)]

PAGE 78

69 to Time II [M = 153.74 (33.25)]. Results of the ANOVA on predictive accuracy scores are presented in Table 9. TABLE 9. ANOVA USING PREDICTIVE ACCURACY SOURCE

PAGE 79

70 W CO > w H Eh O u u H CO Q W J* U *g 2 D
PAGE 80

71 However, there was a significant statistical difference for the improved couples from Time I [M = 170 (47) ] to Time II [M = 136 (31)], t (14) 3.17, pi. 005. Further tests involved independent sample t-tests to check for differences in mean predictive accuracy scores overall among the groups. There was a significant difference in mean PA score when comparing satisfied and dissatisfied couple groups, t, 3Q = 1.769, p .05 in the predicted direction. There was also a significant difference in mean PA when comparing dissatisfied and improved couple groups, t (24) = 1*718, p ^ .05, in the predicted direction. Summary The results of this study indicated that there were some differences in the couple groups designated as satisfied, dissatisfied and improved, and some changes across time. Testing Hypothesis I, Section 1, indicated that there was a tendency for satisfied and improved couples to be more cognitively similar than the dissatisfied couples in terms of functional similarity scores but not cognitive complexity. Hypothesis I, Section 2, testing resulted in support for the prediction that couples whose satisfaction scores improve also improve functional similarity scores. However, testing the second section of this hypothesis for the FIC measure resulted in no significant interaction.

PAGE 81

72 Testing Hypothesis II, Section 1, resulted in partial support for the author's prediction that satisfied and improved couples are better able to understand each other (i.e., have better predictive accuracy scores) than are dissatisfied couples. While the observed difference in the means for groups was in the predicted direction, it did not attain statistical significance. Results of testing Hypothesis II, Section 2, showed, as predicted, a group by time interaction. Follow-up tests showed a significant improvement in understanding (i.e., predictive accuracy) for the improved couples from Time I to Time II, while satisfied and dissatisfied couples remained relatively stable on this variable.

PAGE 82

CHAPTER V DISCUSSION The primary objective of this study was to test and compare changes in cognitive similarity (FS, FIC) and understanding (PA) in couples who report changes in relationship satisfaction with couples who did not change in their levels of satisfaction or dissatisfaction. The first hypothesis of this study examined the differences in cognitive similarity as measured by functional similarity (FS) and functionally independent constructs (FIC) score. It was expected that the satisfied and improved couples would share greater FS than the dissatisfied couples. While the differences between groups on the mean FS scores were not statistically significant, they did present a trend in the predicted direction. Large discrepancy scores indicate low levels of FS (see Functional Similarity, Chapter II) The lowest FS levels were found in the dissatisfied group (M 91.95; SD = 23.38). Scores for the satisfied (M 78.38; SD = 16.77) and the improved group (M = 77.33; SD = 19.31) were similar, but this appears to be a function of the improved group changing at Time II. This trend is reflected in both the friendship (Neimeyer & Neimeyer, 1981) and marital (Neimeyer & Hudson, 1985) 73

PAGE 83

74 personal construct literature. Neimeyer and Neimeyer (1981) found significantly greater functional similarity among friendship pairs than in nominal (unacquainted) pairs along psychological constructs. Neimeyer and Hudson (1985) found a tendency toward greater FS on superordinate (more important) constructs by satisfied than by dissatisfied couples. Another study (Bankikiotes & Neimeyer, 1981) also links the importance of the construct to the functional similarity in attraction. It had been expected that FS would distinguish the different couple groups. The tendency for FS scores to be in the predicted direction but not significant could be a result of several possible theoretical and methodological problems. Theoretically, an explanation of the findings in this study could be that there is a relatively fixed level of FS in marital dyads. That is, established relationships may possess a certain amount of FS which acts as a filter in attraction and relationship development. The existence of differences in FS between satisfied and dissatisfied marital relationships may only be detected when evaluating couples along constructs which they rate as very important to them on their individual system but not along constructs which are superficial, as cited in studies by Neimeyer & Neimeyer (1981) and Neimeyer and Hudson (1985)

PAGE 84

75 Methodologically, several factors (some derived from theoretical issues discussed above) may have affected the results of this study. These factors are 1) The limited number of constructs elicited necessarily limits the view of that individual's construct system and information regarding it. 2) These elicited constructs may include constructs of varying importance in the individual's construct system. As discussed previously, more important (superordinate) constructs may differentiate individuals and couples while less important (subordinate) constructs may not. The second part of Hypothesis I predicted that improved couples would increase their level of FS and FIC over time, whereas satisfied and dissatisfied couples would remain relatively stable. Partial support for this prediction was found in the tendency for an interaction between level of satisfaction and time. As the improved group reported increased satisfaction from Time I (M 19) to Time II (M = 7), the FS also improved from Time I (M = 84.33) to Time II (M 70.33). (Note: a low score reflects greater FS.) Also, the other two couple groups remained relatively stable from Time I to Time II on amount of FS: satisfied couples TjM 78.57; T^M = 78.19; dissatisfied couples T M 93.81;

PAGE 85

76 T II M = 90.09. In addition, the improvement of Time II mean FS score for the improved couples resulted in an average overall group score one point lower than the overall group score for the satisfied group. The trend for improved couples to increase the amount of FS as they become more satisfied is in line with predictions. Indeed, as these couples improved, they began to resemble the satisfied couples on this dimension. Hypothesis I was not rejected for the other measure of cognitive similarity, structural similarity as measured by FIC score. Despite the fact that the means for the couple groups were in the predicted direction (satisfied M = 2.45; dissatisfied M = 2.60; improved M = 2.83), the results were not statistically significant. These findings did not replicate Neimeyer's (1984) cross-sectional study of 20 married couples in which marital satisfaction was found to be related to the similarity of husband and wife's levels of cognitive complexity (i.e., a discrepancy score). Since the mean scores and standard deviations for satisfied and dissatisfied couples on the Area of Change Questionnaire in Neimeyer's (1984) study are comparable to the scores in this study, an obvious explanation for these findings (i.e. the couple groups from these two studies should not be compared because of very different degrees of satisfaction/dissatisfaction) cannot be made.

PAGE 86

77 A couple's overall level of complexity was also assessed by Neimeyer in his study but no relation between it and marital satisfaction was found. A post-hoc analysis in terms of couples' total FIC score for couples in this study resulted in similar findings. Hypothesis II, Section 1, of this study focused on the possible differences in understanding, as measured by the mean predictive accuracy scores for the different couple groups. It was expected that the dissatisfied couples would show the least understanding of any group. There was a trend in the predicted direction, F,„ ... 2.34, p .10. The dissatisfied couples were least understanding and the improved couples were the most understanding. This latter finding appears, however, to be a function of improved understanding at Time II. As previously discussed under FS results section, a significant difference between groups may have been found had the constructs been subdivided into subordinate and superordinate as in the study of couples by Neimeyer and Hudson (198 5) ; that is, predictive accuracy may have been a stronger factor distinguishing the different couple groups had the importance of the constructs been taken into consideration. Theoretically, Wijesinghe and Wood (1976) proposed that the partners understanding of each other, especially

PAGE 87

78 along important constructs, is a critical factor in successful marital relationships. In other words, the findings in this study may have been influenced by a preponderance of less important or subordinate constructs, which would not be expected to delineate satisfied versus dissatisfied couples based on past research. The second part of Hypothesis II focused on the differential effects of level of satisfaction and time on understanding for the couples. It was expected that the satisfied and dissatisfied couples would remain relatively stable in their understanding, while the improved couples would show improved understanding at Time II. This expectancy was confirmed. The predictive accuracy of all groups increased from Time I to Time II, with only the improved group's increase being statistically significant [t,,,. = 3.17, p £ .005]. In this study, the improved group mean PA score (M = 16 9.66; SD -= 46.64) was midway between the mean PA score for the dissatisfied couples' score (M = 182.18; SD 41.14) and the satisfied couples' score (M 156.95; SD = 31.18) at Time I. So the PA mean score for improved couples was better than the dissatisfied couples score but not as good as the satisfied couples score. However, at Time II, the mean score for the improved group changes (M -= 135.73; SD 31) so that the improved group looks better than the satisfied group (M = 155.28; SD = 31) in accuracy of predictions.

PAGE 88

79 This improvement in the PA score for couples who became more satisfied over time lends support to the theory that understanding between spouses increases marital satisfaction. It is interesting to note that the satisfied and dissatisfied couples remained stable in their ability to predict. Other studies have addressed the importance of mutual understanding between couples (Neimeyer, 1981; Neimeyer & Hudson, 1985) Since mutual understanding could be readily assessed in this study by obtaining a discrepancy score between husband's and wife's understanding, this was done. It was predicted that the satisfied and improved couples would share more mutual understanding (i.e., have lower discrepancy scores) than dissatisfied couples. Results of a 3x2 ANOVA showed no significant differences among groups. However, the means for the three groups (satisfied 11.69; dissatisfied = 15.59; improved = 12.3) were in the predicted direction. In summary, couples who were satisfied, dissatisfied and improved tended to share different levels of the variables, functional similarity and predictive accuracy. In addition, there is a significant difference among improved couples from Time I to Time II; i.e., these couples shared both higher levels of FS and PA at Time II when they reported greater marital satisfaction. In general, the raw

PAGE 89

80 scores for these couples tended to fall between the dissatisfied group and satisfied group at Time I but to improve significantly at Time II to approach or surpass the satisfied couples raw score. These findings lend partial support to the hypotheses in this study. Implications The challenge to investigate complex real life interpersonal relationships in terms of dyadic, longitudinal variables has been made by several researchers in the recent past (Delia, 1980; Duck & Miell, 1984; Hinde, 1981). This study has focused on the challenge of examining the established couple relationship over time in terms of the dyadic variables of understanding and cognitive similarity. The findings in the present study provide partial support to the author's hypotheses. Implications for findings in this study with satisfied, dissatisfied and improved couples will be found in their application in the clinical arena. In this study, couples who reported increased marital satisfaction over time also became more similar in terms of how they applied their constructs (FS) and in terms of how they understood (PA) their partners along each other's constructs. Provided further studies replicate these findings of increased FS and PA related to increased marital satisfaction, clinical application might include marital and family therapists' use of

PAGE 90

81 these variables in therapy outcome assessment. Specifically, the measurement of these variables of cognitive similarity and understanding could be a concrete measure of improvement over the course of therapy. So clinical couples who improve with therapy would be expected to increase their levels of cognitive similarity and understanding. Specific clinical application of these findings would include use of the Reptest not only as an assessment tool to evaluate progress in therapy (Slater, 1970; Watson, 1970; Rowe, 1971; Winter & Trippett, 1977; and Ryle, 1980, 1981) but also as a therapeutic tool. Its use would be to point out first the accurate predictions and similar ratings and then to point out inaccurate predictions and differences in ratings. As previously suggested by Watson (1970) Neimeyer (1985) and Neimeyer and Hudson (1985) attending to accuracies and similarities initially in the therapy process would tend to decrease personal threat and increase consensual validation. Then pointing out inaccurate predictions and differences in ratings to make partners aware of their differences will be more readily accepted. Conversely, the dissatisfied couples in this study had low levels of predictive accuracy and functional similarity. Such levels of PA and FS can be viewed as greater number of differences or disagreements in terms of understanding and similarity. While it has been shown that spouses are more

PAGE 91

82 upset when they believe the other misunderstands them than when their spouse disagreed with them (Harvey, Wells, & Alvarez, 1976), dysfunctional relationships commonly exhibit an intolerance of differences (Overturf 1976) This intolerance of differences sets up an expectation that is generally considered inappropriate in close relationships; that is, irrational beliefs or expectations (such as perfect harmony) affect the marital relationship (Epstein, 1979) Prevention of disagreements to maintain interspouse harmony has been shown to include making incorrect statements (Ryder & Goodrich, 1966) While beliefs and expectations may result in outward agreement, literature on interspouse behavior, in general, shows it to be not so agreeable. Specifically, married couples have been shown as less polite, less explicit and more interruptive than strange couples (Winter, Ferreira & Bowers, 1973). In addition, married individuals treat strangers more gently and generally more nicely than they do their spouses (Ryder, 1968) Therefore, marital therapy with couples who have low PA and FS should include understanding the couples' beliefs regarding the consequences of differences in the relationship. Again, unrealistic expectations may be challenged after emphasis of present similarities. In addition, education of the couple regarding the potential value of disagreement to the extent of encouraging open disagreement

PAGE 92

83 with proper communication skills (Gottman, Notarius, Gonso & Markman, 1980) may facilitate the therapeutic process. When a couple's low FS and PA are viewed by the therapist to be affected at least in part by the neurotic behavior of one or both spouses, a different therapeutic technique may be implemented. Kelly (1955) believed that the "idiosyncratic nature of a patient's discriminations may make it difficult for others to play a meaningful relation to him/her because roles are defined as interactions based upon an appreciation of an individual's outlook" (p. 97) Paralleling this thinking, Adams-Webber (1979) tentatively explained the finding that neurotic males were better able to predict their spouses than their spouses were them, by suggesting that these men might act in such a way as to make prediction difficult or impossible. In these cases, therapy must deal with spouses' role identity and interactions. Techniques suggested in the Personal Construct Therapy literature include enactment, group psychotherapy, pyramid procedure (Epting & Boger, 1981) and Kelly's (1955) fixed role therapy (Adams-Webber, 1981) Another implication for the results of this study is further support for the use of the Reptest with couples. The Reptest allows the researcher to look at dyadic variables shared by individuals who, in turn, share a unique relationship while, at the same time, tap commonalities

PAGE 93

84 among couples. In this study, the dyadic variables of understanding and cognitive similarity were quantified in terms of elicited constructs, thus providing a unique view of each couple relationship. In summary, clinical and research application for the findings in this study have been presented in this section. Recommendations Several articles have been written pointing out the possible problems associated with the study of close relationships (Kelley, 1977; Duck & Miell, 1984; Delia, 1980; Duck & Sants, 1983) Of all the problems involved with the study of personal relationships, this author believes that Duck and Sants' (1983) viewpoint succinctly describes the complexity of the project in the following statement: We argue that the existence and range of discrepant viewpoints upon a given relationship itself constitutes evidence about the nature of the relationship rather than being a mere artifact or irritant to scientists, (p. 35) That is, relationships are multifaceted by their nature and research on them should begin to realize this. This complexity in the nature of relationships warrants a conservative attitude toward clinical application of positive research findings until replication studies have been completed.

PAGE 94

Other issues raised by this study relate to sample selection: the exclusive use of a student population, the small sample size, problems associated with longitudinal research and procedural difficulties. Newcomb and Bentler (1980) point out the error in the general results with exclusive use of college students. Students face unique life dilemmas that may easily and indeterminably alter features of cohabitation. Ideally, then, further studies of this kind would benefit from utilization of couples from outside the university community. In addition, exclusive use of legally married couples would be recommended. Further considerations in future studies would be to increase the number of couples involved in the study and to equalize the numbers in each group. In addition, in order to know whether the findings in this study apply to all couples, research replicating this study with couples with children, couples in therapy and couples at different life stages is recommended. Another recommendation would be to vary the time between test administrations and include a time lapse of many months to a period of years. Such testing would be valuable not only to research in relationship development and satisfaction but also in relationship dissatisfaction and dissolution.

PAGE 95

86 Recommendations for future studies in which the research elicits constructs from the subjects would be concerned with the differential findings regarding superordinate versus subordinate constructs in previous studies as discussed earlier. Therefore, it would be beneficial in the future to: 1) stress the importance of subjects describing important similarities among the elements; 2) increase the number of elements from which constructs are elicited or elicit two or more constructs for each element, thereby increasing the number of constructs to be rate; 3) direct the subjects to number the constructs as important versus less important, in order to categorize the constructs as superordinate versus subordinate. However, such changes would substantially increase the time required to complete the tests. With this in mind, future researchers would do well to compensate their subjects for their time. In summary, this study investigated the level of understanding and cognitive similarity in satisfied, dissatisfied and improved couples. Results indicated that as couples improve their levels of satisfaction, levels of understanding and cognitive similarity, marital satisfaction also increases. Implications for these findings included further support for use of the Reptest in couple studies and possible application to marital therapy. Recommendations were offered for future research.

PAGE 96

APPENDIX A THE MARITAL AWARENESS REPERTORY TEST

PAGE 97

APPENDIX A I'D *D • n a. I" > • u u v •Am Q • o o • -o • a 88

PAGE 98

89 The Marital Awareness Repertory Test used in this study is presented on the previous page. This test is shown with the rating scales on which the individual rates himself/ herself along the constructs listed in the way alike/way different columns. In addition, the individual rates the spouse, his/her own parents, and the spouse's parents.

PAGE 99

90 INSTRUCTIONS This form is called the Reptest. It is designed to help you determine the kinds of things that are important to you in your close relationships. The Reptest is a selfexploration exercise. There are no correct answers; what is important is what you value. To complete the Reptest follow the instructions below. Please do not consult your partner while completing the form since your individual judgments are what is of interest right now. STEP I: Notice the numbered columns (1-6) beginning with wife's mother and ending with husband. These are the people you will be considering throughout this task. You will also find six rows, marked A-F. What you are to do first is to find the three individuals in Row A who have circles under their names. In this case, circles appear under column #1 (wife's mother), #2 (wife's father), and #3 (wife). Now ask yourself the question, "how are any two of these individuals similar ?" For example, someone might see his mother and his wife both as "talkative". This similarity "talkative" should be written under the "way alike" column in the center of the form. Then ask yourself what is the opposite of this description and write this in the "way different" column on the far right hand side of the sheet. For example, the opposite of the word "talkative" for someone might be "quiet". "Quiet" should be written in the "way different" column. The dimension "talkative" vs. "quiet" is called a construct There are many ways in which any two of those people could be considered similar. What we are interested in are the ways which you see as important. There are no right or wrong answers. We are simply interested in your best assessments. Next, go to Row B and identify the 3 figures in that row (in this case, #'s 1, 2, and 6) and again answer the question "in what way are any two of these individuals alike?" Write the way in which you see them as similar in the "way alike" column and whatever you see as the opposite in the "way different" column. Your answers should all be different so that when you have completed Row A through F

PAGE 100

91 you will have 6 different constructs. Complete this procedure now, before going ahead with the instructions. STEP II: Now look at each of the constructs and ask yourself which side of each description seems most positive to you and which side seems most negative. Indicate this by putting a small plus sign in the parentheses which appear next to the description which is positive. Place a minus sign in the parentheses next to the description which is negative. When you have finished, the sides of every construct should have one positive and one negative sign. If you find it difficult to decide which is more positive or more negative then assign the positive sign to the description under "way alike". STEP III: Now you will notice that to the left of the constructs there is a grid with the names of individuals above each column. Using a BLACK PEN, please rate each of the individuals along each of the 6 descriptions (constructs) Do this by inserting in the space above the diagonal line in each box the one number which indicates how you view that particular person along each of those dimensions. For example, if your first scale is "talkative vs. quiet", and you have rated talkative as positive and quiet as negative, then rate the wife's mother towards the -3 on the right if you view her as very quiet. If you see her as "very talkative" you would rate her towards a +3 on the left side of the scale. Remember to include the valence (+ or -) before each number based on your responses in Step II. The higher the number you circle the more you see that person as having the characteristic written on that side of the scale. If you cannot decide or the scale doesn't seem to apply, then circle the zero in the center. When you finish rating the wife's mother, then continue rating the rest of the individuals (2-6) along those same constructs. It is very important to assign one and only one number in each rating box above the diagonal line. STEP IV: Notice to the left of the "way alike" column there is a small column with a check-mark appearing at the top. What we want you to do is ask yourself which of these constructs you think your partner used in making sense of people. In other words, how many of these constructs will he/she have on his/her grid? Indicate those constructs which you think

PAGE 101

92 your partner also used by placing a check-mark next to them in the check-mark column. STEP V: This last section must be completed with a RED PEN only. With the RED PEN rate the same individuals along the same constructs but this time according to how you believe your partner would rate them. This rating should be inserted in the space below the diagonal line. When you finish, each construct should be rated twice; with one number in black and one number in red. Be sure to rate all of the people on all six constructs. STEP VI: When you have completed these ratings do this: 1) on a new Reptest form make one copy of your constructs with positive and negative signs. 2) Now exchange Reptests with your partner so that you have each other's sets of constructs. 3) Using the BLACK PEN again, rate each of the 6 persons along your partner's constructs as you see them. 4) Finally, using the RED PEN, rate each of the 6 people as you think your partner would rate them and place this number below the diagonal line in each box; in other words, guess how they see those same people. When you have completed this step you have finished the Reptest.

PAGE 102

APPENDIX B DEMOGRAPHIC DATA QUESTIONNAIRE

PAGE 103

APPENDIX B COUPLES' DEMOGRAPHIC DATA QUESTIONNAIRE Husband Age_ Race Years of Education Student: Yes No If student, classification Wife Age_ Race Years of Education Student: Yes No If student, classification Married: Yes No Length of cohabitation Number of children 94

PAGE 104

APPENDIX C INFORMED CONSENT FORM

PAGE 105

APPENDIX C University of Florida Department of Psychology Consent Form Subject's Name Subject's Address Project Number Project Title Principal Investigator Date I agree to participate in the research as explained below: The aim of this study is to explore the way in which you view your relationships with others who are important to you. To participate in this research you will need to complete one questionnaire and one structure interview at two different times: one now and one several weeks from now at your convenience. The questionnaire asks you to indicate the things you like and would like to change in your relationship/marriage. The interview involves examining similarities and differences among various persons who are important to you (parents, spouse, brothers/ sisters) Both the questionnaire and the interview are completed individually; separate from your spouse. All information is strictly confidential. Neither your partner no anyone outside this research project will have access to any of your materials. Also, all materials will be number-coded (instead of using your name) to insure confidentiality within legal limits of the law. The above-stated nature and purpose of this research, including discomforts and risks involved (if any) have been explained to me. Furthermore, I understand that this investigation may be used for educational purposes which may include publication. I also understand that I may withdraw my consent at any time without prejudice. Signed I have defined and explained fully this research to the participant whose signature appears above. Signed 96

PAGE 106

REFERENCES Adams-Webber, J.R. (1969) Cognitive complexity and sociality. British Journal of Social and Clinical Psychology 8_, 211-216. Adams-Webber, J.R. (1973). The complexity of the target as a factor in interpersonal judgment. Social Behavioral Personality _1, 35-38. Adams-Webber, J.R. (1979). Personal construct theory; concepts and applications New York: Wiley. Adams-Webber, J.R. (1981) Fixed role therapy. In R. Corsini (Ed.), Innovative psychotherapies (pp. 333343). New York: Wiley. Adams-Webber, J.R., Schwenker, B. & Barbeau, D. (1972). Personal constructs and the perception of individual differences. Canadian Journal of Behavioral Science 4, 218-224. Bannister, D. (1965) The rationale and clinical relevance of repertory grid technique. British Journal of Psychiatry 111 977-982. Bannister, D. (Ed.) (1977). New perspectives in personal construct theory New York: Academic Press. Bannister, D. & Bott, M. (1973). Evaluating the person. In P. Klein (Ed.), New approaches in psychological measurement (pp. 147-153) Chichester, England: Wiley. Bannister, D. & Mair, M.M. (1968). The evaluation of personal constructs London: Academic Press. Baucom, D.H. (1982). A comparison of behavioral contracting and problem-solving/ communication training in behavioral marital therapy. Behavior Therapy 13 162-174. Beck, D.F. (1975) Research findings on the outcomes of marital counseling. Social Casework 56 153-181. Bieri, J. (1955) Cognitive complexity-simplicity and predictive behavior. Journal of Abnormal and Social Psychology 51, 263-268. 97

PAGE 107

98 Bieri, J., Atkins, A., Briar, S. Leamon, R. Miller, H. & Tripodi, T. (1966) Clinical and social judgment New York: Wiley. Birchler, G.R. (1979). Communication skills in married couples. In A. S. Bela & M. Hertson (Eds.), Research and practice in social skill training (pp. 294-312) New York: Plenum Press. Birchler, G.R., & Webb, L.J. (1977). Discriminating interaction on behavior in happy and unhappy marriages. Journal of Consulting and Clinical Psychology 45 494495. Blumstein, P., & Schwartz, P. (1983). American couples New York: William Morrow and Co., Inc. Burgess, E.W. & Cottrell, L.S. (1939). Predicting success or failure in marriage New York: Prentice-Hall. Burns, T. Hunter, M. & Lieberman, S. (1980). A repertory grid study of therapist/couple interaction. Journal of Family Therapy 2, 297-310. Byrne, D. (1969) Attitudes and attraction. In L. Berkowitz (Ed.), Advances in experimental social psychology (pp. 323-329) New York: Academic Press. Christensen, L.K. and Wallace, L. (1976). Perceptual accuracy as a variable in marital adjustment. Journal of Sex and Marital Therapy 2J 2 > 130-136. Cole, CM. (1976). A behavioral analysis of married and living together couples Unpublished doctoral dissertation, University of Houston, Houston, Texas. Delia, J.G. (1980) Some tentative thoughts concerning the study of interpersonal relationships and their development. The Western Journal of Speech Communication 44, 97-103. Dickens, W. J. & Perlman, D. (1981). Friendship over the life cycle. In S.W. Duck & R. Gilmour (Eds.), Personal relationships 2: Developing personal relationships (pp. 187-193). London: Academic Press. Duck, S.W. (1972) Friendship, similarity, and the reptest. Psychological Reports 31 231-234.

PAGE 108

99 Duck, S.W. (1973a) Personal relationships and personal constructs: A study of friendship formation London: Wiley. Duck, S.W. (1973b) Similarity and perceived similarity of personal constructs as influences on friendship choice. British Journal of Social and Clinical Psychology 18 1-6. Duck, S.W. (1975) Personality similarity and friendship choice by adolescents. European Journal of Social Psychology 5_, 351-365. Duck, S.W. (Ed.) (1977). Inquiry, hypothesis and the quest for validation: Personal construct systems in the development of acquaintance. In Theory and practice in interpersonal attraction (pp. 220-231) London: Academic Press. Duck, S. (1979, July). Personal constructs in the development and collapse of personal relationships Paper presented to the Third International Congress of Personal Construct Psychology, Breukelen, Netherlands. Duck, S.W., & Allison, D. (1978). I like you but I can't live with you: A study of lapsed relationships. Social Behavior and Personality 6_, 43-48. Duck, S.W., & Gilmour, R. (eds. ) (1981) Personal relationships: 3 personal relationships in disorder New York: Academic Press. Duck, S., & Miell, D. (1984). Towards a comprehension of friendship development and breakdown. In D. Taffel, I. Fraser, & J. Jaspars (Eds.), The social dimension: European perspectives on social psychology Volume 1. Cambridge: University Press. Duck, S., & Sants, H. (1983). On the origin of the species: Are personal relationships really interpersonal states? Journal of Social and Clinical Psychology 1, 27-41. Duck, S.W., & Spencer, C. (1972). Personal constructs and friendship formation. Journal of Personality and Social Psychology 23 40-45. Dymond, R. (1954) Interpersonal perception and marital happiness. Canadian Journal of Psychology 8, 164-171. Epstein, N. Finnegan, D. & Bythell, D. (1979). Irrational beliefs and perceptions of marital conflict. Journal of Consulting and Clinical Psychology 47 608-610.

PAGE 109

100 Epting, F.R. & Boger, P. A. (1981). Personal construct psychotherapy. In R. Corsini (Ed.), Innovative psychotherapies (pp. 613-625) New York: Wiley. Fjeld, S., & Landfield, A.W. (1961). Personal construct consistency. Psychological Reports 8_, 127-129. Goldstein, J.W. & Rosenfeld, H.M. (1969). Insecurity and preference for persons similar to oneself. Journal of Personality 37 253-268. Gottman, J., Notarius, C., Gonso, J., & Markman, H. (1980). A couple's guide to communication Champaign: Research Press. Greer, S.E., & D'Zurilla, T.J. (1975). Behavioral approaches to marital discord and conflict. Journal of Marriage and Family Counseling _l f 299-315. Gurman, A.S. (197 3a) Marital therapy: Emerging trends in research and practice. Family Process 12, 45-54. Gurman, A.S. (1973b). The effects and af f ectiveness of marital therapy: A review of outcome research. Family Process 12 145-170. Gurman, A.S., & Kniskern, D.P. (1977). Enriching research on marital enrichment programs. Journal of Marriage and Family Counseling 2' 3-10. Gurman, A.S., & Kniskern, D.P. (1978a). Research on marital and family therapy: Progress, perspective and prospect. In S. Garfield and A. Bergin (Eds.), Handbook of psychotherapy and behavior change (pp. 769815) (second ed.). New York: Wiley and Sons. Gurman, A.S., & Kniskern, D.P. (1978b). Behavioral marriage therapy: II. Empirical perspective. Family Process 17 139-148. Gurman, A.S., & Kniskern, D.P. (1978c). Deterioration in marital and family therapy: Empirical, clinical and conceptual issues. Family Process 17 3-20. Gurman, A.S., & Kniskern, D.P. (Eds. ) (1981) Handbook of family therapy New York: Brunner/Mazel Inc.

PAGE 110

101 Harvey, J.H., Wells, G.L. & Alvarez, M.D. (1976). Attribution in the context of conflict and separation in close relationships. In J.H. Harvey, W.J. Ickes, & R.F. Kidd (Eds.), New directions in attribution research (pp. 235-260). Hillsdale, NJ: Lawrence Erlbaum Associates. Heiman, J., LoPiccolo, L. & LoPiccolo, J. (1981). The treatment of sexual dysfunction. In A. Gurman & D. Kniskern (Eds.). Handbook of family therapy (pp. 592627) New York: Brunner/Mazel. Hicks, M.W. & Piatt, M. (1970). Marital happiness and stability: A review of the research of the sixties. Journal of Marriage and the Family 32 553-574. Hinde, R.A. (1981) The bases of a science of interpersonal relationships. In S. Duck & R. Gilmore (eds.), Personal relationships 1: Studying personal relationships New York: Academic Press. Howard, J.W. & Dawes, R.M. (1976). Linear prediction of marital happiness. Personality and Social Psychology Bulletin 2, 478-480. Hunt, D.E. (1951) Studies in role concept repertory: conceptual consistency. Columbus, OH: Unpublished master's thesis, The Ohio State University. Huston, T.L., & Levinger, G. (1978). Interpersonal attraction and relationships. Annual Review of Psychology 2_9, 115-156. Jacobson, N.S. (1978). A review of the research on the effectiveness of marital therapy. In T.J. Paolino & B.S. McCrady (Eds.), Marriage and marital therapy: Psychoanalytic, behavioral and systems theory perspectives (pp. 394-444) New York: Brunner/Mazel Jacobson, N.S. (1981). Behavioral marital therapy. In A. Gurman & D. Kniskern (Eds.), Handbook of family therapy (pp. 556-591) New York: Brunner/Mazel. Jacobson, N.S, & Martin, B. (1976). Behavioral marriage therapy: Current status. Psychological Bulletin 83, 540-556.

PAGE 111

102 Kelley, K. (1977). An application of attribution theory to research methodology for close relationships. In G. Leirnger and H.L. Raush (eds.) Close relationships: Perspective on the meaning of intimacy Amherst: University of Massachusetts Press. Kelley, H.H. (1979) Personal relationships: Their structures and processes New York: Wiley. Kelly, G.A. (1955) The psychology of personal constructs New York: Norton. Kerckhoff, A.C., & Davis, K.E. (1962). Value consensus and need complementarity in mate selection. American Sociological Review 27 295-303. L'Abate, L. (1977) Enrichment: Structured interventions with couples, families and groups Washington: University Press of America. L'Abate, L. (1981) Skill training programs for couples and families. In A. Gurman & D. Kniskern (Eds.), Handbook of family therapy (pp. 631-661) New York: Brunner/ Mazel. Laing, R.D., Phillipson, H. & Lee, A. (1966). Interpersonal perception: A theory and a method of research New York: Springer. Landfield, A.W. (1971) Personal construct systems in psychotherapy Chicago: Rand McNally. Landfield, A.W. (1977). Interpretive man: The enlarged self image. In The Nebraska symposium on motivation 1976 (pp. 110-119) Lincoln: University of Nebraska Press. Landfield, A.W. (1979) Exploring socialization through the interpersonal transaction group. In P. Stringer & D. Bannister (Eds.), Constructs of sociality and individuality (pp. 59-120). New York: Academic Press. Landis, J.T. (1963). Social correlates of divorce or nondivorce among the unhappy married. Marriage and Family Living 25 178-180. Lea, M. (1979) Personality similarity in unreciprocated friendships. British Journal of Social and Clinical Psychology 18 393-394.

PAGE 112

103 Leventhal, H. (1957) Cognitive processes and interpersonal predictions. Journal of Abnormal and Social Psychology 55, 176-180. Levinger, G. & Senn, D.J. (1970). Progress toward permanence in courtship: A test of the Kerckhof f-Davis hypothesis. Sociometry 33 427-443. Luber, R.F. (197 8) Teaching models in marital therapy: A review and research issue. Behavior Modification 2, 77-91. Margolin, G. Talovic, S., & Weinstein, CO. (1983). Areas of change questionnaire: A practical approach to marital assessment. Journal of Consulting and Clinical Psychology 51 920-931. Margolin, G. & Wampold, B.E. (1981). A sequential analysis and accord in distressed and nondistressed marital pairs. Journal of Consulting and Clinical Psychology 49, 554-567. Margolin, G. & Weiss, R.L. (1978). A comparative evaluation of therapeutic components associated with behavioral marital treatment. Journal of Consulting and Clinical Psychology 46 1476-1486. Maslow, A.H. (1953) Love in healthy people. In A. Montague (Ed.), The meaning of love (pp. 37-68). New York: Julian Press. McCarthy, B. (1981). Studying personal relationships. In S.W. Duck & R. Gilmour (Eds.), Personal relationships 1: Studying personal relationships (pp. 23-45). New York: Academic Press. McCarthy, B. & Duck, S.W. (1976). Friendship duration & responses to attitudinal agreement-disagreement. British Journal of Social and Clinical Psychology 15, 377-386. Menges, R.J. (196 9) Student-instructor cognitive compatibility in the large lecture class. Journal of Personality 37 444-459. Mitsos, S. (1958). Representative elements in role construct technique. Journal of Consulting Psychology 2J., 311-313. Montgomery, J. P. (197 3) Commitment and cohabitation cohesion Unpublished manuscript, University of Edmonton, Edmonton, Alberta.

PAGE 113

104 Neimeyer, G.J. (1981, August). Personal construct systems in the development and deterioration of interpersonal relationships Paper presented at the Fourth International Congress on Personal Construct Psychology, Brock University, St. Catherine's, Ontario, Canada. Neimeyer, G.J. (1984) Cognitive complexity and marital satisfaction, Journal of Social and Clinical Psychology 2, 256-263. Neimeyer, G.J. (1985) Personal constructs in the counseling of couples. In F.R. Epting and A.W. Landfield (eds.), Anticipating personal construct theory Lincoln: Nebraska Press. Neimeyer, G. & Hudson, J. (1985). Couples constructs: Personal systems in marital satisfaction. In D. Bannister (Ed.), Perspectives in personal construct theory: Number Three (pp. 225-241) London: Academic Press. Neimeyer, G. & Neimeyer, R. (1981). Functional similarity and interpersonal attraction. Journal of Research in Personality 15, 427-435. Neimeyer, R.A. & Neimeyer, G.J. (1982, July). Interpersonal relationships and personal elaboration: A construct theory model. Paper presented at the International Conference on Personal Relationships, Madison, Wisconsin. Neimeyer, R.A. & Neimeyer, G.J. (1983). Structural similarity in the acquaintance process. Journal of Social and Clinical Psychology 2_, 1-6. Neimeyer, R. Neimeyer, G. & Landfield, A. (1983). Conceptual differentiation, integration and empathic prediction. Journal of Personality 51 185-191. Nelson, R.O. (1981) Realistic dependent measures for clinical use. Journal of Consulting and Clinical Psychology 49 168-182. Newcomb, M.D., & Bentler, P.M. (1981). Martial breakdown. In S.W. Duck, & R. Gilmour (Eds.), Personal relation ships: 3 Personal relationships in disorder (pp. 57-94) New York: Academic Press. Newcomb, T.M. (197 8) The acquaintance process: Looking mainly backward. Journal of Personality and Social Psychology 36, 1075-1083.

PAGE 114

105 Orden, S.R., & Bradburn, N.M. (1968). Dimensions of marriage happiness. American Journal of Sociology / 73 1537-1563. Overturf J. (1976) Marital therapy: Toleration of differentness. Journal of Marriage and Family Counseling 2_, 235-241. Patterson, G.R., Weiss, R.L. & Hops, H. (1976). Training of marital skills. In H. Leitenberg (Ed.), Handbook of behavior modification and behavior therapy (pp. 313351) New York: Prentice-Hall. Pinsof W. (1981) Family therapy process research. In A. Gurman & D. Kniskern (Eds.), Handbook of family therapy (pp. 699-741) New York: Brunner/Mazel. Richardson, F.C., & Weigel, R.G. (1971). Personal construct theory applied to the marriage relationship. Experimental Publication System 10 371-375. Rowe, D. (1971). An examination of a psychiatrist's predictions of a patient's constructs. British Journal of Psychiatry 118 231-234. Runkel, P.J. (1956) Cognitive similarity in formulating communication. Sociometry 19 178-191. Ryder, R.G. (1968) Husband-wife dyads versus married strangers. Family Process 12 233-238. Ryder, R.G., & Goodrich, D.W. (1966). Married couples' responses to disagreement. Family Process 5_, 30-42. Ryle, A. (1980) Some measures of goal attainment in focussed integrated active psychotherapy: A study of 15 cases. British Journal of Psychiatry 137 475-486. Ryle, A. (1981) Dyad grid dilemmas in patients and control subjects. British Journal of Medical Psychology 54, 353-358. Ryle, A., & Breen, D. (1972a). The use of the double dyad grid. British Journal of Medical Psychology 45 375382. Ryle, A., & Breen, D. (1972b). A comparison of adjusted and maladjusted couples using the double dyad grid. British Journal of Medical Psychology 45 375-384.

PAGE 115

106 Ryle, A., & Lipshitz S. (1975). Recording change in marital therapy with the reconstruction grid. British Journal of Medical Psychology 48 39-48. Ryle, A., & Lipshitz, S. (1976). Rep grid elucidation of a difficult conjoint therapy. British Journal of Medical Psychology 49, 281-285. Ryle, A., & Lunghi M. (1970). The dyad grid: A modification of repertory grid technique. British Journal of Psychiatry 117 323-327. Sechrest, L.B, & Jackson, D.N. (1961) Social intelligence and accuracy of interpersonal predictions. Journal of Personality 29 167-181. Slater, P. (1965) The use of the repertory grid technique in the individual case. British Journal of Psychiatry 111 965-975. Slater, P. (1970) Personal questionnaire data treated as a repertory grid. British Journal of Social and Clinical Psychology 9_, 357-370. Smail, D.J. (1972) A grid measure of empathy in a therapeutic group. British Journal of Medical Psychology 45, 165-169. Terman, L.M. (1938) Psychological factors in marital happiness New York: McGraw-Hill. Thomas, L.F. (1979). Construct, reflect and converse: The conversation reconstruction of social realities. In P. Stringer & D. Bannister (Eds.), Constructs of sociality and individuality (pp. 148-179) London: Academic Press. Triandis, H.C. (1959) Cognitive similarity and interpersonal communication in industry. Journal of Applied Psychology 43 321-326. Tripodi, T. & Bieri, J. (1964). Information transmission in clinical judgment as a function of stimulus dimensionality and cognitive complexity. Journal of Personality 32, 119-137. Watson, J. P. (1970) A measure of patient-therapist understanding. British Journal of Psychiatry 117 319-321.

PAGE 116

107 Weiss, R.L. (1978). The conceptualization of marriage from a behavioral perspective. In T.J. Paolino & B.S. McCrady (Eds.), Marriage and marital therapy: Psychoanalytic, behavioral, and systems theory perspectives (pp. 165-239) New York: Brunner/Mazel. Weiss, R.L. Hops, H. & Patterson, G.R. (1973). Framework for conceptualizing marital conflict, technology for altering it, some data for evaluating it. In L.A. Hamerlynch, L.C. Handy, & E.J. Mash (Eds.), Behavior change: Methodology, concepts, and practice (pp. 202247). Champaign, Illinois: Research Press. Wijesinghe, O.B.A., & Wood, R.R. (1976). A repertory grid study of interpersonal perception within a married couple's psychotherapy group. British Journal of Medical Psychology 49 287-293. Winch, R.F., Ktsanes, I., & Ktsanes, V. (1954). The theory of complementary needs in mate selection: An analytic and descriptive study. American Sociological Review 19, 241-249. Winter, D.A., & Trippett, C.J. (1977). Serial change in group psychotherapy. British Journal of Medical Psychology 50 341-348. Winter, W.D. Ferreira, A.J., & Bowers, N. (1973). Decision-making in married and unrelated couples. Family Process 12 83-94. Yllo, K.A. (1978) Nonmarital cohabitation beyond the college campus. Alternative Life Styles Ml), 37-54.

PAGE 117

BIOGRAPHICAL SKETCH Jean Ellen McLaughlin Hudson was born February 5, 1945, in Brockton, Massachusetts. She graduated from Cardinal Spellman High School in Brockton, in 196 3. She attended Massachusetts General Hospital School of Nursing in Boston, Massachusetts, graduating in 1966. After working as a licensed registered nurse in intensive care settings for more than ten years, she returned to college to pursue her interest in psychology. She received the Bachelor of Arts with high honors from the University of Florida, Gainesville, Florida, in August, 1979. In the following year, she began graduate studies in psychology. She entered the Counseling Psychology Program at the University of Florida in 1980 and received a Graduate Council Fellowship for the academic year 1980-1981. The degree of Doctor of Philosophy will be conferred after completion of an internship year, after which she plans to work as a marital and family therapist. 108

PAGE 118

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Ellen S. Amatea, Chairperson Associate Professor of Counselor Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Greg J. Neimeyer, Co-Chairpe Assistant Professor of Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Harry A. Grater, Jr. Professor of Psychology

PAGE 119

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. James Archer, Jr. Associate Professor of Counselor Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^Vv^ %. C\ ^Ui
PAGE 120

UNIVERSITY OF jj|Rfi!Bft| 11111 |l|| y'S2 0828 l 5 ; 342 4


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EEISL3DU4_4XHI30 INGEST_TIME 2017-07-11T21:31:47Z PACKAGE UF00101050_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES