Title: Personality structure and mood in female college students
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00098663/00001
 Material Information
Title: Personality structure and mood in female college students
Physical Description: xi, 157 leaves : ; 28 cm.
Language: English
Creator: Tobacyk, Jerome John, 1948-
Copyright Date: 1977
 Subjects
Subject: Mood (Psychology)   ( lcsh )
Personality assessment   ( lcsh )
Psychology thesis Ph. D   ( lcsh )
Dissertations, Academic -- Psychology -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Statement of Responsibility: by Jerome John Tobacyk.
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 154-157.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00098663
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000207367
oclc - 04058376
notis - AAX4163

Downloads

This item has the following downloads:

personalitystruc00toba ( PDF )


Full Text










PERSONALITY STRUCTURE AND MOOD IN
FEMALE COLLEGE STUDENTS









By

JEROME JOHN TOBACYK


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







UNIVERSITY OF FLORIDA
1977














ACKNOWLEDGEMENTS

I wish to thank Franz Epting, my chairman, for serving as a fine

role model. His is a very unique combination of professional competence

and warm human qualities that has enriched my graduate studies. Also,

I owe a great deal to Jim Dixon who has been very helpful, supportive,

and understanding during my graduate studies. He has been a very good

friend. Audrey Schumacher has enriched my graduate training considerably

by gently helping me to keep perspective on my work. Walter Cunningham

has provided me with considerable enthusiasm concerning factor analytic

methods for the study of persons. William Ware has helped me develop

competence in statistics.

I wish to express my thanks to Ada for spending time at the not-so-

creative, but necessary, task of data recording. Yoriko, Jane, and

Robert worked very hard at punching literally thousands of computer cards.

I appreciate the long hours that they spent at this task. Gail, Heather,

and Lisa have my thanks for their outstanding job of rating protocols.

Theirs was a very important contribution to this dissertation. Marty

has my thanks for sharing his knowledge of Erikson with me. His

suggestions helped me considerably in my formulations.

Of course, the most important contribution to this dissertation was

made by my subjects, who showed considerable dedication in systematically

providing me with glimpses of their intimate affective lives.







ii








Many friends have enriched my life through the past years at the

University of Florida. A special thank you goes to my friend Howie,

who, in many ways is responsible for my academic career. Most of all,

I wish to acknowledge my debt to my parents who have always been

understanding and supportive of my goals. They have my love for this

and for ever so much more.



















TABLE OF CONTENTS


ACKNOWLEDGEMENTS. . . . . . . . . .

LIST OF TABLES. . . . . . . . . .

ABSTRACT. . . . . . . . . . . .

CHAPTER
I INTRODUCTION . . . . . . . .

Theoretical Perspectives Toward Mood . .

Phenomenological Perspective . . .
Trait Perspective. . . . . . .
Behavioral Perspective . . . .
Psychodynamic Perspective. . . . .
Relationships Between Mood and Other
Psychological Processes. . . . .
Multidimensional Approach to Mood. . .

Review of Mood and Personality Research. .

Summary of Research Review . . . .
Critique of Previous Research. . . .

Plan of Present Investigation. . . .
Differentiation and Integration. . . .
Relationship Between Differentiation and In
The Psychological Differentiation Construct
The Ego Identity Status Construct. . .
Assessment of Ego Identity Status. . .
Formulation of Personality Patterns. . .
Hypotheses Concerning Mood . . . .

Affective Complexity . . . . .
Mood Level . . . . . . . .
Mood Variability . . . . . .

II METHOD . . . . . . . . . .

Subjects . . . . . . . . .
General Procedure. . . . . . .


Procedure for PSY 201 Subjects .
Procedure for Additional PSY 201
Procedure for PSY 340 Subjects .


Subjects.
. ,


Page
ii

vii

ix


. . . 1

. . . 1

. . . 1
. . . 217
. . . 3
. . . 3

. . . 5
. . . 5









. . . 624
. . . 13
. . . 16

. . . 16
. . . 17
tegration 18
. . . 18
. . . 20
. . . 22
. . . 24
. . . 25

. . . 25
. . . 26
. . . 26

. . . 29

. . . 29
. . . 29

. . . 30


I









CHAPTER


Subjects' Estimation of Accuracy in Following
Instructions . . . . . . . . . .
Completion of Mood Study. . . . . . . .
Data Recording and Verification . . . . . .
Criterion for Inclusion in the Final Sample . . .
Subject Attrition . . . . . . . . .
Number of Days of Mood Rating Employed. . . . .
Estimations of Missing Observations . . . . .
Instrumentation . . . . . . . . .

Instruments for Assessment of Psychological
Differentiation . . . . . . . .
Instruments for Assessment of Identity Achievement.


Instrument for Assessment of Mood .

Measures of Dependent Variables . .

Affective Complexity. . . . .
Mood Level . . . . . .
Mood Variability. . . . .

Construction of Indices . . .
Formation of Groups . . . . .
Statistical Hypotheses. . . .


Affective Complexity. . . . . .
Overall Mood Level . . . . ..
Overall Mood Variability. . . . .
Mood Variability Concerning Academic Work
Intimacy Issues . . . . . .


. . 57
. . 57
. . 57

. . 60


III RESULTS . . . . . . . . . . .


General Strategy for Results. . . .
Relationships Among Dependent Variables .
Affective Complexity. . . . . .
Overall Mood Level . . . . ..
Overall Mood Variability . . . ..
Mood Variability Concerning Academic Work
Intimacy Issues . . . . . .


IV DISCUSSION. . . . . . . . . . .

Relationships Among Mood Variables. . . . .


. 61
. 62
. 63
. 65
. 67

. 71

. 74

. 74


Relationship between Affective Complexity and
MVPC1 . . . . . . . . .
Relationship between MLPC1 and MVPC1. . .
Relationships Among Other Mood Variables. .


. . 41


. . . . 43


Page









CHAPTER


Findings Concerning Affective Complexity. . . .

Methodological Problems of Previous Investigations.
Affective Complexity and PD: A Reformulation . .


Page

76

77
80


Findings Concerning Overall Mood Level. . . .. 83
Findings Concerning Overall Mood Variability. .. . 85

Theoretical Rationale for Findings . . . . 86
Effectiveness of Integration as a Higher
Order Construct . . . . . . . . 87
Differential Implications of IA for Mood
Variability . . . . . . . . . 89
Mood Variability Concerning Intimacy vs. Academic
Work Issues . . . . . . . . . 90
Summary . . . . . . . . . . . 92
Limitations of the Present Investigation. .. .... 94

APPENDICES . . . . . . . . . . . . . 95

APPENDIX
A GENERAL ORIENTATION TO MOOD STUDY . . . . . 97

B DAILY RECORD OF PERSONAL FEELINGS . . . . . 100

C INSTRUCTIONS FOR RFT. . . . . . . . .. 104

D INSTRUCTIONS FOR EFT . . .. . . . . . . 106

E SUMMARY OF MOOD PROJECT . . . . . . . . 109

F INCOMPLETE SENTENCES BLANK. . . . . . . .. 112

G EGO IDENTITY INCOMPLETE SENTENCE BLANK SCORING MANUAL 116

H PERSONAL PREFERENCES FOR COMPLETING SENTENCES . . 146

I PERSONAL FEELING SCALES . . . . . . . 149

BIBLIOGRAPHY . . . . . . . . . . . . 154

BIOGRAPHICAL SKETCH. . . . . . . . . . .. 158















LIST OF TABLES

TABLE Page

1 INTERCORRELATIONS AMONG FIVE MEASURES OF AFFECTIVE
COMPLEXITY. . . . . . . . . . . .. 45

2 DESCRIPTIVE STATISTICS FOR FIVE MEASURES OF AFFECTIVE
COMPLEXITY . . . . . . . . . . 45

3 FACTOR PATTERN MATRIX FOR MLPC1 . . . . . . 48

4 FACTOR PATTERN MATRIX FOR MVPC1 . . . . . . 50

5 FACTOR PATTERN MATRIX FOR MVPC2 . . . . . . 51

6 MEANS AND STANDARD DEVIATIONS FOR SELECTION INSTRUMENTS 53

7 INTERCORRELATION MATRIX FOR SELECTION INSTRUMENTS . . 54

8 DESCRIPTIVE STATISTICS ON ALL VARIABLES FOR GROUPS 1-4
FORMED BY MEDIAN SPLITS AND FOR THE 31 SUBJECT SAMPLE 58

9 DESCRIPTIVE STATISTICS FOR GROUPS 1-4 FORMED BY EXTREME
GROUP SELECTION PROCEDURES. . . . . . . .. 59

10 INTERCORRELATION MATRIX FOR DEPENDENT MEASURES. .. .... 62

11 SUMMARY TABLE FOR MULTIPLE REGRESSION OF AFFECTIVE
COMPLEXITY ON PD AND IA . . . . . . . . 63

12 SUMMARY TABLE OF ANALYSIS OF VARIANCE FOR AFFECTIVE
COMPLEXITY SCORES .. . . . . . . . . . 64

13 MEAN AFFECTIVE COMPLEXITY SCORES FOR GROUPS 1-4,
COMBINED GROUPS 1 AND 4, AND COMBINED GROUPS 2 AND 3 65

14 SUMMARY TABLE FOR MULTIPLE REGRESSION OF MLPC1 ON
PD AND IA . . . . . . . . . . . 66

15 SUMMARY TABLE OF ANALYSIS OF VARIANCE FOR MLPC1 SCORES. 66









TABLE Page
16 MEAN MLPC1 SCORES FOR GROUPS 1-4, COMBINED GROUPS 1
AND 2, AND COMBINED GROUPS 3 AND 4. . . . . .. 67

17 SUMMARY TABLE FOR MULTIPLE REGRESSION OF MVPC1 ON
PD AND IA . . . . . . . . . . . 68

18 SUMMARY TABLE OF ANALYSIS OF VARIANCE FOR MVPC1 SCORES. 69

19 MEAN MVPC1 SCORES FOR GROUPS 1-4 AND ALL COMBINED GROUPS. 70

20 SUMMARY TABLE FOR MULTIPLE REGRESSION OF MVPC2 ON
PD AND IA . . . . . . . . . . . 71

21 SUMMARY TABLE OF ANALYSIS OF VARIANCE FOR MVPC2 SCORES. 72

22 MEAN MVPC2 SCORES FOR GROUPS 1-4, COMBINED GROUPS 1
AND 4, AND COMBINED GROUPS 2 AND 3. . . . . .. 73


v i i i








Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy



PERSONALITY STRUCTURE AND MOOD IN
FEMALE COLLEGE STUDENTS

By

Jerome John Tobacyk

December, 1977

Chairman: Franz Epting
Major Department: Psychology

The present investigation explored the relationship between mood

and two properties of personality structure, differentiation and

effectiveness of integration. Differentiation was operationalized in

terms of Witkin's psychological differentiation construct and was

assessed by an index computed from each subject's rod-and-frame test

and embedded figures test scores. Effectiveness of personality

integration was operationalized in terms of Erikson's identity achieve-

ment construct and was assessed by a modified form of Marcia's Ego

Identity Incomplete Sentences Blank.

Personality patterns were formulated and assessed in terms of

each subject's status on psychological differentiation and identity

achievement. Then, employing Psychological Differentiation Theory as

a framework, hypotheses were derived concerning the relationships

among psychological differentiation, identity achievement, the

personality patterns and mood.

Wessman and Ricks' Personal Feelings Scales were employed as the

mood assessment instrument. Thirty-one female college students rated









their moods three times daily on each of the 16 Personal Feelings Scales

for a minimum of 33 successive days. Measures of three general para-

meters of mood were constructed via factor analysis of mood rating

records. These three parameters were affective complexity, mood level,

and mood variability.

The affective complexity measure was derived from the results of

P-technique factor analysis of each subject's individual mood rating

record. The overall mood level and mood variability measures were

constructed from a principal components analysis of mood rating records

for all subjects. Factor scores reflecting overall mood level and

overall mood variability were computed for each subject and used as

dependent measures.

Three major findings were: (1) greater personality differentiation,

as reflected by scores indicative of greater psychological differentiation,

was associated with lesser affective complexity, (2) neither psychological

differentiation, nor identity achievement, was associated with overall

mood level, and (3) greater effectiveness of personality integration, as

reflected by scores indicative of greater identity achievement, was

associated with lesser overall mood variability.

The employment of the personality pattern approach disclosed two

relationships that would not have been revealed with a univariate

approach. These were: (1) effectiveness of personality integration

appears to have differential implications for overall mood variability

in greater and lesser differentiated persons, and (2) the characteristic

psychosocial concerns of greater and lesser differentiated persons may








crystalize around different issues, depending upon effectiveness of

integration.

Findings were discussed in terms of Witkin's Psychological

Differentiation Theory and Erikson's Psychosocial Developmental

Theory. It was concluded that the higher order personality constructs

of differentiation and effectiveness of integration provide a useful

framework for the conceptualization of mood/personality relationships.














CHAPTER I

INTRODUCTION

Theoretical Perspectives Toward Mood

The experience of mood is a fundamental characteristic of the human

situation. The structure and function of mood can be conceptualized

from several theoretical perspectives. Each perspective illuminates

different facets of mood, resulting in a richer understanding of its

significance for the person. A theoretical analysis of mood will be

conducted from the phenomenological, trait, behavioral, and psycho-

dynamic perspectives.

Phenomenological Perspective

The descriptive phenomenological approach to mood has provided a

rich legacy in penetrating and insightful clinical descriptions of the

vicissitudes of mood. In Fear and Trembling and The Concept of Dread,

Kierkegaard (1954, 1957) developed a graphic, sensitive description of

mood in some of its core manifestations--anxiety, boredom, and despair.

Such descriptions of the subjective experiences of various moods have

provided a deeper understanding of the human situation.

However, mood can also be viewed as a fundamental structure of

human existence. From this categorical phenomenological perspective,

mood is viewed as providing a frame of reference through which human

existence can be understood. In this case, mood is viewed as a mode of

pre-cognitive relatedness to a world of existential concerns. Thus, in





1





2


addition to its subjective psychological expression, mood performs a

revelatory function. Mood discloses aspects of the concrete life-world

as they appear in their existential immediacy. To illustrate this

function of mood, anxiety can disclose man's freedom by bringing him

to an immediate awareness of his future possibilities.

Heidegger (1962) further analyzes mood in terms of this revelatory

or intentional function. For Heidegger, Dasein's "openness to the world"

is co-constituted through the indissoluble unity of two fundamental

structures: mood and understanding. Thus, from the categorical

phenomenological perspective, mood can be viewed as a fundamental

ontological structure. It precedes any cognitive determination of the

character of the world. Mood discloses to man his concrete situation in

the world.

Trait Perspective

Traits can be conceptualized as stable, organized, intra-individual

dispositions which are relatively generalizable across time and situations.

Some (Cattell, 1957, 1965) construe traits as the primary determinants of

behavior. However, employing the trait formulation, it is necessary to

consider the influence of psychological states, situational variables,

and roles, in addition to traits, to predict human behavior. States,

such as moods, are conceptualized as reversible and less temporally

stable processes than traits. From the trait perspective, at least

two types of moods can be construed. First, there may be purely

fluctuant moods which are independent of fixed traits. This type of

mood would not be related to the level of a particular trait. A second

type of mood may be a state of a trait, that is, a trait which is in a

temporary swing concurrent with changes in internal or external stimuli.


i-








Regardless of their origins and nature, moods can significantly influence

human behavior and are, therefore, necessarily incorporated into the

trait formulation.

Behavioral Perspective

The behavioral perspective emphasizes mood as a hypothetical construct

describing behavior. This conception of mood is similar to the notion

of "predisposition." Mood is regarded as having an effect on the

probability of occurrence of certain responses in certain situations.

Conscious mood, as a response to the cues of the hypothetical mood

state, is assumed to supply information to the organism about the current

functioning of the organism. Thus, mood is presumed to be involved in

the self-monitoring and self-regulation of behavior.

Mowlis and Nowlis (1956), leading proponents of the behavioral

perspective toward mood, define mood as "an intervening variable or pre-

dispositional factor that is a source of information or discriminative

stimuli to the organism about the current functioning characteristics

of the organism" (p. 322). Nowlis and Nowlis view conscious mood as

the perceptual/cognitive responses of the organism to this information.

Akin to the behavioral perspective is the view of Ryle (1950), the

linguistic philosopher, who considers mood to be a short-term liability

to act or react in certain general manners. Thus, to summarize the

behavioral perspective, mood is regarded as a factor which influences

the probability of making certain responses in certain situations at

particular times.

Psychodynamic Perspective

The psychodynamic perspective considers mood within a complex

system of personality structure, dynamics, and development. Jacobson







(1957), a representative of this perspective, emphasizes the pervasive

nature of mood through all levels of individual functioning. Dynam-

ically, mood can be viewed as a temporary fixation of a generalized

libidinal discharge process. Once a mood is established, it affects

all other psychological processes of the person. This generalized

influence on all other processes distinguishes mood from other affective

states. According to Jacobson (1957), non-mood emotional states develop

from specific tensions and are related to specific objects. Thus,

emotions can become moods only if they generalize and predominate the

entire sphere of the ego.

A mood can originate from a significant experience "whose discharge

pattern generalizes and lends its qualities to all other discharge

patterns" (Jacobson, 1957, p. 78). For this to occur, the precipitating

experience must be of considerable intensity and, therefore, cannot be

immediately and sufficiently relieved by a focal discharge only. Such

a general influence on all other psychological processes results in a

qualitative change in the experience of self/world.

According to the psychodynamic formulation, moods appear to serve

a useful economic function. Moods allow for a repetitive affective

discharge on a large number and variety of objects. According to

Jacobson (1957), "such a prolonged discharge in small quantities, combined

with reality testing, can liberate psychic energy from fixated positions,

permitting new investments" (p. 81). This gradual discharge process

can protect the ego from intense, sudden affective discharge.

To summarize the psychodynamic perspective, moods influence all

feelings, thoughts, and actions. Thus, mood, providing an indicant of

the current ego state, reflects the current effectiveness of ego

functioning.







Relationships Between Mood and Other Psychological Processes

There is not considerable precision of definition differentiating

mood from other psychological processes, such as character, trait,

style, disposition, temperament, and emotion. This attests not solely

to a lack of conceptual clarity, but also reflects the necessity of

ultimately construing these psychological processes as interpenetrated,

yet somewhat dissociable, processes of the person.

An initial distinction among these processes can be made on a

stability dimension. Processes such as character, trait, style,

disposition, and temperament are generally regarded as reflections of

underlying, intra-individual personality structure, which is relatively

stable across time and situations. Temperament differs from these other

relatively stable processes in that it refers to relatively stable

affective characteristics of the person.

In contrast to these more stable processes, mood and emotion are

state processes which are reversible and less temporally stable.

Although both mood and emotion refer to affective states, mood is

generally characterized by a greater duration and lesser intensity

than emotion. Further, mood is a generalized affective process which

pervades the person's total experience of self/world, while emotion

generally is directed toward a specific object. Thus, mood can be

considered a generalized, reversible affective state of greater

duration than emotion, but of less stability than temperament.

Multidimensional Approach to Mood

A common core of characteristics can be distilled from the pre-

ceding approaches to mood. Mood can be conceptualized as an affective

process of variable duration which significantly affects and reflects








the quality of life of the person. It is viewed as providing a cross

section of the state of the ego, or, more generally, as providing an

indicant of the current state of the person's overall psychological

functioning.

Mood is not an isolated psychological state, but rather a process

which is interpenetratedly woven through the configuration of structural,

dynamic, and developmental processes of the person. Thus, to understand

mood, it appears necessary to investigate it holistically, within the

configuration of psychological processes of the person. However, as the

following review will indicate, research concerning mood and personality

structure has not generally been conducted from a holistic perspective.

Indeed, all reported studies have employed univariate personality con-

structs as predictors of mood parameters. A multivariate approach

employing a pattern of personality constructs as predictors would

appear to more accurately reflect the complex personality configuration

within which the vicissitudes of mood occur.

Review of Mood and Personality Research

Wessman and Ricks (1965) reported a seminal investigation of mood

and personality within the context of a longitudinal personality

assessment project at the Harvard University Psychological Clinic. Two

separate six-week mood studies were reported, one employing 25 females,

the other using 18 males.

The female study was conducted first (Wessman, Ricks, & Tyl, 1960),

but was more fully analyzed and reported by Wessman and Ricks in 1965.

The subjects in the female study were 25 Radcliffe undergraduates, all

from the same dormitory, who volunteered to participate in the mood

study. Of these 25 subjects, 21 responded for a minimum of 30 days








out of the 42. Fourteen of these subjects completed a minimum of 30

successive days of mood rating, as well as a battery of personality

assessment instruments.

The subjects in the male study were 18 Harvard undergraduates who

were being paid for participation in the Harvard Personality Assessment

Project. An attempt was made to select male subjects who represented a

wide range of scores on several unpublished personality assessment

instruments. These instruments were regarded as measures of (1)

alienation, (2) happiness-unhappiness, (3) oral dependency, and (4)

anal-retentiveness. Forty-two successive days of daily mood ratings

were obtained for each subject.

Mood ratings were obtained on the Wessman and Ricks Personal

Feelings Scales (PFS). The PFS is a set of 16 ten point mood rating

scales, each scale representing a rationally-derived bipolar mood dimension.

For the females, an earlier form of the PFS, consisting of 11 scales, was

employed. All subjects were instructed to report their moods once daily,

during the evening, by recording the "highest," "lowest," and "average"

mood level that they had experienced for each of the mood dimensions

during that particular day.

For each subject, measures of mood level, mood variability, and

affective complexity were obtained. Mood level scores were computed

by obtaining the mean of the "average" ratings on the "Elation vs.

Depression" Scale of the PFS for each subject. Mood variability was

similarly measured by computing the standard deviation of the "average"

ratings on the "Elation vs. Depression" Scale for each subject. Wessman

and Ricks (1965) emphasized the "Elation vs. Depression" Scale, one of

the 16 scales of the PFS, because it was considered to reflect the







central and, perhaps, most significant feature of mood (happiness-

unhappiness). A measure of affective complexity was obtained by

recording the number of rotated factors accounting for more than

10 percent common variance, which were extracted from each subject's

30-42 day record of daily mood reports via P-technique factor analysis.

For the males, measures of mood level, mood variability, and

affective complexity were correlated with approximately 360 scores

derived from personality assessment instruments. This information

included scores derived from the Minnesota Multiphasic Personality

Inventory, the 16PF, the Rorschach, life histories, diaries, and

interviews. For the females, a much smaller personality assessment

battery was employed, which included an actual self-ideal self Q-sort

and the Rosenzweig Picture Frustration Test.

Major findings reported by Wessman and Ricks (1965) included:

(1) A relatively wide range of scores was obtained for both males

and females on measures of mood level, mood variability, and affective

complexity, supporting the notion that these variables represent

individual difference dimensions.

(2) Mood level and mood variability were reported to be un-

correlated for both males and females. Affective complexity was

significantly correlated with mood variability for males (r = -.57,

p <.01), with greater affective complexity associated with lesser mood

variability. However, affective complexity was not associated with

mood variability in females.

(3) For both males and females, lower mood level was associated

with a decrease in reported self-esteem, a greater reported actual

self-ideal self discrepancy, and an increase in the expression of self-







derogatory attitudes. From this constellation of findings, which was

reinforced by clinical assessment information derived from interviews,

life histories, and diaries, Wessman and Ricks (1965) concluded that dif-

ferences in mood level appeared to reflect a fundamental psychological

dimension. This dimension, which could be variously construed as

adjustment, ego strength, or psychological health, appeared to assess

the adequacy of the person's current psychological functioning.

(4) For both males and females, the most general personality

characteristic related to mood variability appeared to reflect the

"emotional openness" of the more variable subjects, in contrast to the

"emotional constriction" of the more stable subjects. This characteristic

appeared to be pervasive through various levels of individual functioning,

ranging from fantasy productions to style of interpersonal relations. The

more stable males, generally described as not seeking involvement with

their inner lives or with others, were characterized as relatively

completed and closed personality systems (Wessman & Ricks, 1965). In

contrast, the more variable males, described as more involved with their

inner lives and with others, were characterized as more open, conflicted,

evolving personality systems.

Frank (1967) in his dissertation, reported an investigation of the

relationship between mood and the psychological differentiation construct

(Witkin, Dyk, Faterson, Goodenough, & Karp, 1962). It was hypothesized

that greater psychological differentiation would be associated with

greater affective complexity and with less mood variability. Subjects

were 50 female Columbia University students. The majority of the

sample were graduate student volunteers, with a mean age of 29.3. These

subjects rated their moods nightly for a minimum of 30 successive days on

the Wessman and Ricks PFS.







Level of psychological differentiation was assessed from an index

derived from each subject's scores on the Embedded Figures Test and the

Human Figure Drawings Test. Affective complexity was measured by the

number of factors extracted via P-technique factor analysis from each

subject's 30+ day record of daily mood reports. Frank (1967) initially

employed the same criterion to assess affective complexity as had

Wessman and Ricks (1965). However, this criterion, the number of

rotated factors each accounting for more than 10 percent of the common

variance, resulted in a restriction in the range of affective complexity

scores. In this case, 96 percent of the sample obtained either two or

three factors. As a result, the criterion was changed to the number of

factors necessary to account for 90 percent of the common variance.

This resulted in a more satisfactory range of affective complexity

scores. Mood variability was assessed by computing the standard

deviation of each subject's "average" daily ratings on the "Elation vs.

Depression" Scale. It was reported that greater psychological differen-

tiation was associated with greater affective complexity (r = -.24,

p (.05). However, no relationship was obtained between psychological

differentiation and mood variability.

Gorman and Wessman (1974) reported an investigation concerning the

relationship between cognitive styles and mood. Subjects were 20 male

and 47 female undergraduate students in the first author's abnormal

psychology class at Nassau Community College. A battery of 21 cognitive

style measures was administered to subjects who were also providing

daily mood reports on 14 scales of the Wessman and Ricks PFS for a 28

successive day period.







Measures of mood level, mood variability, and affective complexity

were obtained. Forty-two mean mood level scores were computed for each

subject. This was accomplished by computing the mean for each subject's

"average," "highest," and "lowest" rating for each of the 14 scales.

Similarly, 42 mood variability scores were computed for each subject,

one for each of the 14 Personal Feeling Scales. Affective complexity

scores were computed via principal components P-technique factor analysis

of each subject's four week record of daily mood reports. The following

four indices of affective complexity were employed: (1) the percentage

of total variance accounted for by the first factor, (2) the average

intercorrelation among mood ratings, (3) the number of factors that

accounted for 90 percent of the total variance, and (4) the number of

factors that each accounted for more than 10 percent of the total

variance.

Since few significant differences were obtained between males and

females on either the cognitive stylistic or the mood variables, the

data for the two groups were combined and analyzed together. The

cognitive style scores were intercorrelated and factor analyzed via the

principle axis method and rotated to Varimax criterion. Eight cognitive

stylistic factors were reported. In two further analyses, the set of

(1) mood level scores, and (2) mood variability and affective complexity

scores were extended onto the original cognitive style factor matrix

by using Dwyer's Factor Extension Technique (Dwyer, 1937). Thus,

factor loadings of the mood variables on the cognitive style factors

were obtained without distorting the original cognitive style factor

matrix.








Two major findings relevant to the present investigation were:

(1) Factor VII, labeled "Field Articulation," was inversely associated

with mood variability, with greater field articulation associated with

less mood variability. The marker variables used to define Factor VII

were the Hidden Figures Test and the Hidden Patterns Test. (2) Factor

II, labeled "Sensation Seeking and Openness," was positively associated

with mood level. This bipolar factor loaded positively on the Zuckerman

and the Pearson Sensation Seeking Scales. Factor II loaded negatively

on the Breskin Rigidity Test and the DPI Reoression Scale. This set

of marker variables for Factor II suggests an "open vs. closed to

experience" personality dimension. Further, three out of four affective

complexity measures indicating lower affective complexity were loaded

moderately on Factor II. Thus, greater openness to experience appeared

to be associated with less affective complexity, as well as with higher

mood level.

Schaff (1975), in his dissertation, investigated the relationship

between cognitive complexity and affective complexity, both as construed

within the context of Kelly's Personal Construct Theory (Kelly, 1955),

and mood. It was hypothesized that high affective complexity and

cognitive complexity would be associated with low mood variability and

with high mood level. Forty-six male and 29 female undergraduates

completed the Wessman and Ricks PFS for a minimum of 21 successive days.

Cognitive complexity and affective complexity were measured by

standard Rep Grid procedures (Bieri, 1955; Kelly, 1955). Affective

complexity was conceptualized as the number of independent personal

constructs relating to affective experience which a subject generated.

Cognitive complexity was similarly conceptualized as the number of

independent interpersonal constructs generated.








Forty-eight mood level scores were obtained for each subject by

computing the means of the "high," "low," and "average" ratings for each

of the 16 mood scales across 21 days. Mood variability scores were

computed by obtaining the standard deviation of each of these 48 mood

level scores.

To summarize Schaff's (1975) findings, it was reported that affective

complexity was related to mood variability for males, but not for females.

Cognitive complexity was differentially related to mood variability for

males and for females. Both affective complexity and cognitive complexity

were differentially related to mood level for males and for females.

However, complexity, as conceptualized within Kelly's Personal Construct

Theory (Kelly, 1955), and psychological differentiation, as developed in

Witkin's Psychological Differentiation Theory (Witkin et al., 1962),

appear to represent relatively independent theoretical formulations.

Indeed, cognitive complexity has been reported to be unrelated to the

psychological differentiation construct (Elliott, 1961). Although these

two formulations appear to possess some formal similarities, they may

not reflect clearly comparable dimensions. Thus, Schaff's (1975)

findings may not be easily compared to those of Frank (1967) and Gorman

and Wessman (1974), who employed measures of the psychological differen-

tiation construct in assessing personality structure.

Summary of Research Review

A summary of the preceding research findings indicates three

substantial, though somewhat confusing relationships between mood and

personality structure. First, it appears that the psychological dif-

ferentiation construct (Witkin et al., 1962) is related to mood. Frank

(1967) reported this construct to be related to affective complexity,








but not to mood variability, in his older, female, primarily graduate

student sample. Gorman and Wessman (1974), employing a male and female

undergraduate sample, reported that field articulation, which provides

an indicant of level of psychological differentiation, was associated

with mood variability, but not with affective complexity. One could

speculate that Frank's (1967) sample may have reflected a relatively

narrow range of scores, which resulted in his findings concerning mood

variability. However, besides employing diverse samples, different, and

not interchangeable, measures of psychological differentation were

employed in these studies. Frank (1967), employing the Embedded Figures

Test and the Human Figure Drawings Test, appears to have used more

adequate assessors of psychological differentiation than Gorman and

Wessman (1974). The Hidden Figures Test and Hidden Patterns Test

employed by Gorman and Wessman (1974) may have a relatively indirect

relationship to the psychological differentiation construct. Thus,

although psychological differentiation does appear to possess a sub-

stantive relationship with mood, the exact nature of this relationship

remains unclear.

The second major finding concerns an "open vs. closed to experience"

personality dimension. Both Wessman and Ricks (1965) and Gorman and

Wessman (1974) reported that an "open vs. closed to experience" dimension

was strongly related to mood. Wessman and Ricks (1965) reported that

this dimension was related to mood variability, with greater "openness"

associated with greater mood variability. Gorman and Wessman (1974),

however, reported that their Factor II, "Sensation Seeking and Openness,"

was strongly associated with mood level, with greater "openness"

associated with a higher mood level, as well as with lower affective

complexity.








Gorman and Wessman's (1974) finding appears to be somewhat contrary

to one of Wessman and Ricks' (1965) most compelling findings, which was

a clear relationship between mood level and general effectiveness of

psychological functioning (or adjustment). One possible explanation for

these apparently contradictory findings focuses upon the personality

assessment instruments which were employed to define the "open vs.

closed to experience" dimension. Wessman and Ricks (1965) were able to

employ an extensive amount of assessment information obtained through a

three year longitudinal personality assessment project. These investi-

gators based their formulation of the "open vs. closed to experience"

dimension primarily upon findings from standard clinical assessment

instruments and interviews. Gorman and Wessman (1974) employed a battery

of 21 paper-and-pencil tests of cognitive style. Their "Sensation Seeking

and Openness" Factor was labeled largely on the basis of salient loadings

of sensation seeking, rigidity, and repression scales. Thus, the "open

vs. closed to experience" dimensions reported in these two studies may

be referring to similarly labeled, but substantively different, personality

dimensions.

To speculate, it is conceivable that the sensation-seeking, rigidity,

and repression scales employed by Gorman and Wessman (1974) may be

reflecting the effectiveness of psychological functioning (adjustment).

However, due to the apparently discrepant findings, the relationship

between the "open vs. closed to experience" dimension and mood remains

unclear. Further, Wessman and Ricks' (1965) finding concerning the

relationship between mood level and general effectiveness of psychological

functioning, although strongly supported, must still remain tentative.








A third major finding derived from the research review concerns

findings by Wessman and Ricks (1965) and Schaff (1975) which indicate

that males and females may be characterized by some similarities and

some differences in mood/personality relationships. Thus, it may be

necessary to investigate these two groups separately.

Critique of Previous Research

The research review revealed a need for clearer theoretical bases

for the conceptualization of relationships between mood and personality

structure. Only Frank (1967) and Schaff (1975) employed clearly defined

theoretical frameworks in their investigations, using, respectively, the

Psychological Differentiation Theory (Witkin et al., 1962) and the

Personal Construct Theory (Kelly, 1955).

Further, the main emphasis of the preceding studies involved

univariate conceptions of personality structure. None of these reported

investigations employed a multivariate approach in the conceptualization

or assessment of personality structure. A multivariate approach,

employing theoretically-derived patterns as predictors of mood, would

provide a more holistic assessment of personality structure. Such an

approach would provide greater fidelity to the actual complexity of

personality than a univariate approach.

Plan of Present Investigation

The present investigation will employ a multivariate approach to

personality structure. Two personality constructs, psychological

differentiation (Witkin et al., 1962) and ego identity status (Erikson,

1956, 1963, 1968) will be combined to provide a multivariate assessment

of personality structure. Then, predictions will be derived concerning

the relationship between this multivariate personality structure and








mood. The theoretical rationale for combining these two constructs

and for formulating predictions concerning mood is based upon the

notions of differentiation and integration in Psychological Differen-

tiation Theory (Witkin et al., 1962).

Differentiation and Integration

In Psychological Differentiation Theory (Witkin et al., 1962),

level of psychological differentiation is considered an assessable

property of the person which provides a reflection of the structural

complexity of the psychological system. This notion can be applied to

the personality system as well. In this case, greater differentiation

can be considered an indicant of a more heterogeneous, specialized,

personality system. Conversely, less differentiation can be viewed as

reflecting a relatively homogeneous, unspecialized personality system.

Specialization refers to the degree of separation among various

psychological processes, such as perceptual, cognitive, and emotional

processes. Further, specialization refers to specificity of functioning

within a particular process. In a more differentiated system, more

specific responses are apt to occur to stimuli, rather than diffuse

responses to any of a variety of stimuli, as in a less differentiated

system. Thus, affective processes, such as moods, would be more

adequately channelized in a more differentiated system. Theoretically,

in a less differentiated system, affective processes could more easily

"spill over" and influence other processes. Generally, then, experience

would be more articulated in a more differentiated system and more

global in a less differentiated system. Greater articulation in the

experience of self and world would be a concomitant of greater

differentiation.









Integration refers to the form of relationships among the components

or processes of a psychological system. Thus, integration refers to the

quality of systemic organization. Effectiveness of integration refers

to the degree of harmonious interaction among systemic components or

processes. Further, effectiveness of integration also refers to the

quality of interaction between the system and its environment. Generally,

then, effectiveness of integration provides a reflection of the degree of

organismic adaptation. From a psychological perspective, effectiveness

of integration has been traditionally considered to be reflected by the

effectiveness of psychological functioning. Thus, effectiveness of

integration corresponds to such psychological formulations as adjustment,

psychological health, and ego strength.

Relationship Between Differentiation and Integration

At any level of differentiation, various forms of integration are

possible, both effective and ineffective. A high level of differentiation

clearly does not imply effective integration. Since, theoretically, both

differentiation and integration are significant, but relatively independent,

determinants of psychological functioning, an assessment of both properties

would provide a holistic assessment of the person. In the present investi-

gation, the psychological differentiation construct (Witkin et al., 1962)

will be employed as a reflection of differentiation, and the ego identity

status construct (Erikson, 1956, 1963) will be employed as an indicant

of effectiveness of integration. The rationale for these formulations

follows.

The Psychological Differentiation Construct

The field dependence dimension emerged from research relating

individual differences in performance on a set of perceptual tasks








(the rod-and-frame test, the embedded figures test, the body adjustment

task) to performance on a variety of perceptual and cognitive measures.

The common element in these perceptual tasks appears to involve the

analytic requirement that the subject separate an object from an

organized field in perception. Subjects who have difficulty separating

an item from its context in these perceptual tasks are considered field

dependent, while those who can overcome an embedding perceptual context

are considered field independent.

Because these individual differences were consistently manifest in

a variety of perceptual and cognitive tasks, the field dependence

dimension was considered to provide an indicant of a personality style.

A personality style is conceptualized as a relatively stable, character-

istic mode of individual functioning which is pervasive through a wide

range of psychological areas. This personality style was conceptualized

as a continuum, extending from a relatively "primitive-global" level of

functioning, reflected by field dependence, to a more "analytic-

differentiated" level of functioning, reflected by field independence.

Thus, it was labeled the global-articulated personality style. The field

dependence dimension, then, can be considered a measure of the global-

articulated personality style.

Further research related the global-articulated style to an even

wider range of areas of individual functioning. To summarize these

findings, the global-articulated style appears to consistently pervade

individual's perceptual, intellectual, motivational, emotional,

defensive, and social operations (Witkin et al., 1962; Witkin, Oltman,

Fox, Erlichmann, Hamm, & Ringler, 1973). Employing this pattern of

evidence, it was proposed that this style is an indicant of the even








more general psychological differentiation dimension. A relatively

extensive nomological network supports the theoretical and empirical

utility of the psychological differentiation construct (Witkin et al.,

1973).

Thus, to summarize, the field dependence dimension is considered

to provide a measure of the global-articulated personality style. This

personality style can be viewed as a reflection of the psychological

differentiation dimension. Theoretically, the psychological differen-

tiation dimension can be considered to provide an assessment of level

of differentiation through a wide range of psychological processes of

the person, or, through the entire personality system.

The Ego Identity Status Construct

The ego identity status construct is a formulation which was

developed in Erikson's Psychosocial Developmental Theory (Erikson,

1956, 1963, 1968). According to this theory, the individual, ideally,

progresses successfully through a series of stages in the process of

developing greater psychosocial maturity. Each of these stages is

characterized by the emergence of a particular normative crisis. By

cirsis, Erikson refers to "a turning point," "a decisive encounter

between the person and his environment" (1968, p. 96). The content

area of each crisis crystalizes around the confluence of the psycho-

sexual and psychosocial concerns characteristic of each stage.

Each crisis is a period of both heightened potential and increased

vulnerability. The successful resolution of a crisis can result in the

acquisition of positive ego qualities, or positive identity elements,

which can contribute to the effective functioning of the person.

Conversely, an unsuccessful resolution can result in the acquisition








of negative identity elements, which can contribute to ineffective

functioning or maladjustment.

The ego is the process which is responsible for synthesizing the

psychosexual and psychosocial concerns which emerge into salience during

each crisis. Further, it is the task of the ego to integrate the identity

elements which emerge at each stage with those identity elements already

in existence. The ego can be conceived as "the organizational process by

which the person maintains himself as a coherent personality with a

sameness and continuity both in self and social experience" (Erikson,

1968, p. 73). The various ego processes, such as perceptual, cognitive,

affective, motivational, social, and defensive processes, must be

effectively integrated in order to successfully perform these syntheses.

Erikson (1956, 1963) has especially emphasized the crisis of ego

identity formation, which typically emerges during late adolescence.

Ego identity is a conscious sense of individual uniqueness, "an awareness

of the fact that there is a self-sameness and continuity to the ego's

synthesizing methods and that these methods are effective in safe-

guarding the sameness and continuity of one's meaning for others"

(Erikson, 1959, p. 23). Thus, ego identity is not concerned with

awareness of the mere fact of one's individuality, but emphasizes the

awareness of the style or quality of one's individuality.

Theoretically, the particular manner that each person has experienced

and resolved each of the earlier crises through his entire psychosocial

history has contributed identity elements, both positive and negative.

During the ego identity crisis, these identity elements re-emerge and

must be synthesized into a coherent configuration or ego identity. The

outcome of this identity crisis depends upon the ability of the ego to








synthesize this constellation of identity elements. This constellation

of identity elements includes "significant identifications, libidinal

needs, favored skills, constitutional givens, and available social roles"

(Erikson, 1968, p. 163). Further, the choices and decisions which are

made during this stage, vocational, occupational, and ideological choices,

frequently result in commitments for life. Thus, the manner of resolution

of the ego identity crisis can be viewed as reflecting the effectiveness

of integration of the person's "lifeplan" or "style of life."

Theoretically, then, the assessment of ego identity status would

provide an indication of the effectiveness of individual or ego functioning.

Such an indication would not only provide a reflection of the current

effectiveness of functioning, but would additionally provide a cross-

section of the effectiveness of functioning through the person's entire

psychosocial history. Thus, theoretically, ego identity status can be

considered to provide an assessment of the effectiveness of integration

of the person or personality system.

Assessment of Ego Identity Status

Erikson (1963) proposed four ego identity statuses, or modes of

dealing with the ego identity crisis. These four statuses are: (1)

identity achievement, in which the identity crisis has been experienced

and successfully resolved via a personal commitment to a style of life,

(2) foreclosure, in which crisis has not been experienced, but firm,

often parentally determined commitments to a life style exist, (3)

identity diffusion, in which crisis has not occurred and commitments

have not been established, and (4) moratorium, in which there may be

an active experiencing of the conflict and confusion associated with

decision making concerning commitments to a style of life.








Marcia (1967) proposed an ordering of these ego identity statuses

based upon the apparent proximity of each status to the identity

achievement status. Marcia employed the rationale that:

A moratorium subject, by virtue of his active concerns
with psychosocial issues, is probably closer to identity
achievement than a foreclosure subject, who may be
solidified in a position of close parental identification,
which makes movement more difficult. Diffusion subjects
lack even the appearance of identity achievement found
in foreclosure. (1967, p. 120)

However, when the criterion of most effective current functioning

is employed, the literature does not fully support the notion that

these statuses represent discretely ordered intervals on a continuum.

Generally, however, the identity achievement status is associated with

the most effective performance on the test and behavioral criteria

employed. These criteria included assessments of self-esteem (Marcia,

1967), problem solving ability (Marcia, 1964), moral judgment (Podd,

1969), conformity (Toder & Marcia, 1973), difficulty of college major

(Marcia & Friedman, 1970), grade point average (Waterman & Waterman,

1972), and experienced self-determination (Waterman, Buebel, & Waterman,

1970). Thus, although there is not sufficient discriminant validity to

order the four statuses as suggested by Marcia (1967), it appears that

the identity achievement status is generally associated with the most

effective functioning. The other statuses appear to be associated with

less effective functioning than identity achievement.

Thus, it appears reasonable to construe a single dimension of ego

identity, ranging from a greater approximation to identity achievement,

characterized by more effective functioning, to a lesser approximation

of identity achievement, characterized by less effective functioning.

Indeed, several investigators have successfully employed this approach








in constructing measures of identity achievement and in predicting to

various criteria reflecting effectiveness of functioning (Baker, 1971;

Constantinople, 1969; Rasmussen, 1964; Simmons, 1973a, 1973b). This

approach will be employed in the present investigation. Ego identity

status will be assessed through a measure of the person's approximation

to identity achievement. In summary, based upon the theoretical and

empirical evidence reviewed concerning ego identity, the person's

status on identity achievement will be employed as a measure of the

effectiveness of functioning, from which the effectiveness of

personality integration can be inferred.

Formulation of Personality Patterns

From the perspective of the present investigation, personality is

viewed as the overall psychological organization of the person. Thus,

personality is considered the fundamental configuration within which

other psychological processes, including affective processes such as

mood, are organized. Level of differentiation and effectiveness of

integration are considered assessable properties of this personality

configuration.

An individual's score on the field dependence dimension will be

regarded as an indication of level of psychological differentiation,

from which level of personality differentiation can be inferred.

Similarly, an individual's identity achievement score will be regarded

as a reflection of the effectiveness of psychological functioning, from

which the effectiveness of personality integration can be inferred. Thus,

level of differentiation and effectiveness of integration will be assessed

for each person. Four personality patterns will be conceptualized, based

upon the four possible intra-individual combinations of differentiation








and effectiveness of integration. These four patterns are: (1) high

differentiation/effective integration, (2) low differentiation/effective

integration, (3) low differentiation/less effective integration, and

(4) high differentiation/less effective integration. These four patterns

will be referred to, respectively, as Patterns 1, 2, 3, and 4.

Hypotheses Concerning Mood

Next, a theoretical rationale based upon the concepts of personality

differentiation and integration will be employed as a framework to derive

hypotheses concerning relationships among psychological differentiation,

identity achievement, the four personality patterns,and mood. Relation-

ships will be derived among these personality constructs and three mood

parameters--affective complexity, mood level, and mood variability.

Affective Complexity

The process of experiencing and distinguishing among moods is referred

to as affective complexity, with greater affective complexity associated

with experiencing a wider variety and shading of moods. Theoretically,

greater differentiation is associated with greater specialization and

articulation in the affective domain. Thus, greater differentiation would

be expected to be associated with greater articulation of all affective

processes and, more specifically, with greater affective complexity.

One hypothesis, then, is that greater differentiation, as reflected

by scores indicative of field independence, is associated with greater

affective complexity. Further, the two patterns associated with greater

differentiation (Patterns 1 and 4) are hypothesized to show greater

affective complexity than the two less differentiated patterns (Patterns

2 and 3).








Mood Level

Mood level has been considered to provide a reflection of the

effectiveness of current individual functioning (Wessman & Ricks, 1965).

A higher mood level is considered indicative of greater effectiveness of

functioning than a lower mood level, except, perhaps, at the extreme.

It will be recalled that, theoretically, effectiveness of functioning

can be conceptualized to provide a reflection of the effectiveness of

personality integration. Thus, a higher mood level can be viewed as a

reflection of greater effectiveness of personality integration. Inasmuch

as identity achievement is conceptualized as an indicant of effectiveness

of personality integration, it is expected that level of identity achieve-

ment will be positively related to mood level.

It is hypothesized that higher identity achievement scores will be

associated with a higher reported mood level. Further, the two per-

sonality patterns associated with more effective personality integration

(Patterns 1 and 2) are hypothesized to report a higher mood level than

the other two patterns.

Mood Variability

Theoretically, mood variability can be considered a joint function

of both level of differentiation and effectiveness of integration. A

highly differentiated personality system, characterized by more

specialized psychological processes, could channelize and respond

relatively discretely to affective stimuli. However, a less differentiated

personality system, with less articulated processes of control, could be

more readily "flooded" by an affect or mood, resulting in relatively

diffuse, systemwide reactions. More articulated processes of control

are available in a more differentiated system, allowing greater

specificity of affective response.







Generally, then, greater differentiation would be expected to be

associated with less mood variability than lesser differentiation. One

hypothesis is that greater differentiation, as indicated by more field

independent scores, will be associated with less mood variability.

However, effectiveness of integration would be expected to moderate

the relationship between differentiation and mood variability. It would

be expected that, among the less differentiated persons, the less

effectively integrated persons would be more variable in mood than the

more effectively integrated persons. Conversely, among the more

differentiated persons, the less effectively integrated would be expected

to report less mood variability than the more effectively integrated.

This formulation is based upon findings that ineffective functioning and

defensive operations take different forms depending upon level of

differentiation (Witkin, 1965).

Among less differentiated persons, ineffective functioning char-

acteristically involves emotionally labile reactions, while typical

defensive operations emphasize repression and denial. Ineffective

functioning among more differentiated persons typically involves over-

control and affective constriction. Characteristic defenses among the

more differentiated include intellectualization, rationalization, and

isolation.

Thus, it appears that less effectiveness of integration would

affect mood variability differentially in greater and lesser differen-

tiated persons. Based upon this rationale, Pattern 3 (low differentiation/

less effective integration) is hypothesized to show the greatest mood

variability, followed by Pattern 2 (low differentiation/more effective

integration), Pattern 1 (high differentiation/more effective integration),








and with Pattern 4 (high differentiation/less effective integration)

showing the least mood variability.

Thus, employing differentiation and effectiveness of integration

together, relatively specific predictions can be made concerning mood

variability. From the perspective of the present formulation, the

"open vs. closed to experience" dimension reported by Wessman and Ricks

(1965) to be strongly related to mood variability, may be alternatively

conceptualized as a dimension reflecting affective reactivity to stimuli.

From this perspective, the extremes of the "open vs. closed to experience"

dimension may be viewed as a reflection of ineffectiveness of integration

in, respectively, less differentiated and more differentiated personality

systems.














CHAPTER II

METHOD

Subjects

Thirty-one female University of Florida undergraduate students

comprised the final sample. Subjects were recruited from either the

Introductory Psychology (PSY 201) Subject Pool or from the Experi-

menter's Winter, 1977 Psychology of Personal Growth (PSY 340) class.

Subjects enrolled in PSY 201 received experimental credit for

participation. Subjects from PSY 340 selected participation in the

mood study from among several alternatives in fulfilling a term

project requirement.

Twenty subjects from PSY 201 and 11 subjects from PSY 340

comprised the final sample. For the total 31 subject sample, the

mean age was 19.94 (SD = 2.31). The distribution of subjects by

class was: 10 freshmen, 7 sophomores, 10 juniors, and 4 seniors.

For the PSY 201 subjects alone, mean age was 19.15 (SD = 1.84) and

class distribution was 9 freshmen, 6 sophomores, 4 juniors, and 1

senior. For the PSY 340 subjects, the mean age was 21.36 (SD = 2.49)

and class distribution was 1 freshman, I sophomore, 6 juniors, and

3 seniors.

General Procedure

Although there were only slight variations in procedure between

the PSY 201 subjects and the PSY 340 subjects, the procedures for

these two groups will be presented separately to enhance clarity.








Procedure for PSY 201 Subjects

Approximately 50 students from the PSY 201 subject pool responded

to a posted experimental announcement and attended a preliminary mood

study meeting on 2/2/77. During this meeting, information concerning

the three-part mood study was provided and subjects were recruited for

participation in the full study (see "General Orientation to Mood Study,"

Appendix A). The experimenter attempted to generate a high level of

interest and motivation in subjects by emphasizing the significance of

the mood study and the importance of each subject's contribution to it.

During this preliminary meeting, which comprised Part I of the

mood study, the two identity achievement measures were group administered,

with administration of the Ego Identity Incomplete Sentences Blank

preceding the administration of the Identity Achievement Status Scale.

This order of administration minimized the possibility of item alternatives

provided by the forced-choice Identity Achievement Status Scale from

influencing subjects' responses to the Ego Identity Incomplete Sentences

Blank items. Fifty subjects completed both identity achievement measures

during the preliminary meeting.

Forty-two subjects who attended the preliminary meeting elected to

participate in the full mood study. Each of these subjects was provided

with a modified form of the Wessman and Ricks Personal Feelings Scales,

a one week's supply of modified Daily Records of Personal Feelings,

plus two "emergency" copies of Daily Records. The emergency copies

were to be used if the subject was temporarily unable to obtain a new

supply of Daily Records. Subjects were instructed in the mood rating

procedure (see "Daily Record of Personal Feelings," Appendix B). To

summarize the mood rating procedure, subjects were requested to rate








their moods three times daily--once during the morning (6AM to 12PM),

once during the afternoon (12PM to 6PM), and once during the evening

(6PM to 12AM). Mood rating was accomplished by selecting the item for

each of the 16 Personal Feeling Scales which best described "How I

feel now." The number of the item selected for each scale was recorded

on the Daily Record of Personal Feelings. Following each evening rating,

the subject was also requested to provide information concerning health,

hours of sleep, academic pressure, menstrual and drug use information,

as well as any additional information which she considered important

in understanding her moods.

Mood rating was begun on 2/2/77 and was to be continued for a

minimum of 33 successive days, resulting in a minimum of 100 ratings

for each of the mood scales per subject. The experimenter placed very

strong emphasis on the need for complete and accurate recording of moods.

The experimenter collected completed Daily Records of Personal Feelings

at the end of subjects' PSY 201 classes each Tuesday and Thursday. During

these twice weekly meetings, the experimenter attempted to maintain a

high level of motivation in subjects. This 33 successive day sequence of

mood rating comprised Part II of the mood study.

Part III of the mood study consisted of each subject being individually

tested on the rod-and-frame test and the embedded figures test. Testing

was conducted by the experimenter in sessions of approximately 45 minutes

duration, with test order randomized. Instructions for the rod-and-

frame test and the embedded figures test are provided in Appendices C and

D.

Procedure for Additional PSY 201 Subjects

Six additional PSY 201 subjects were added to the mood study on








2/3/77. These subjects contacted the experimenter on their own

initiative, desiring to participate in the mood study. The procedures

involved in the mood study were individually explained to these subjects

and the identity achievement measures were also individually administered.

Except for these variations, the procedure was the same as that employed

for the previous PSY 201 subjects.

Procedure for PSY 340 Subjects

Except for minor variations, the same procedure was employed for

PSY 340 subjects as with the PSY 201 subjects. Information concerning

participation in the mood study was provided during the first week of

class. The measures of identity achievement were administered during a

class period. Completed Daily Records were collected and new Daily

Records were distributed at the beginning or end of each class period,

Monday through Thursday. Subjects in the PSY 340 sample conducted their

mood rating from 1/26/77 to at least 3/8/77.

Subjects' Estimation of Accuracy in Following Instructions

Upon the termination of the mood study, each subject was requested

to provide an estimate of the proportion of mood ratings that she had

actually completed in compliance with instructions. That is, subjects

were requested to estimate how frequently they had actually rated their

moods three times daily--once in the morning, once in the afternoon,

and once in the evening. Subjects were instructed to consider all

retrospective ratings inaccurate, regardless of how confident they

felt concerning these ratings. Further, if subjects exceeded the limits

of the time intervals for morning, afternoon, or evening ratings by

only minutes, they were to consider these ratings as inaccurate. The

importance of honestly providing this information was strongly








emphasized. Subjects were assured that there were no penalties associated

with reporting a "low" percentage. In fact, since accuracy of the data

was considered crucial to the mood study, an honest report of a low

percentage would be quite valuable. Subjects were provided the option

of responding anonymously, an option which none of the subjects selected.

Completion of Mood Study

Upon termination of the 33+ day study, subjects were provided either

experimental credit or course credit for a term project. Also, an

explanation of the mood study was provided (see "Summary of Mood Project,"

Appendix E).

Data Recording and Verification

The experimenter filed and verified completed Daily Records for the

PSY 201 subjects, while an assistant performed these duties for the PSY

340 sample. This procedure allowed the experimenter to maintain greater

separation between the teacher role and the role of experimenter in which

he was provided access to the intimate affective lives of his student/

subjects. When missing data were discovered, the subject was individually

contacted immediately and asked for an explanation. The subject was re-

informed concerning the need for accurate and complete data and her

cooperation was re-enlisted in attempting to attain this goal. In some

cases, the subject was allowed to complete the missing ratings by

retrospective rating. In the majority of cases, however, these entries

were coded as missing data and later estimated. This recording and

verification procedure resulted in a relatively small proportion of

missing observations.








Criterion for Inclusion in the Final Sample

A minimum self-report of 90 percent accruacy in following mood

rating instructions was the criterion for inclusion of subjects in the

final sample. The mean percentage for accuracy reported for the full

31 subject sample was 93.6 percent, with the PSY 201 subjects reporting

93.6 percent accuracy, while the PSY 340 subjects provided a 93.7

percent rating.

Subject Attrition

Of the 48 subjects from PSY 201 who began the mood project, data

from 20 were retained in the final analyses. Seventeen subjects were

excluded because their self-report of accuracy in following instructions

in mood rating was less than the 90 percent criterion for inclusion.

Three subjects were excluded because they mistakenly terminated mood

rating prior to the 33 day minimum. Eight subjects terminated partici-

pation in the mood study for personal reasons on their own initiative.

Eleven of the 24 PSY 340 subjects who began the mood study were

retained in the final analyses. Two students terminated participation

in the project and completed alternative term projects. Eleven of the

PSY 340 subjects were excluded because their self-reports of accuracy

of mood rating were less than the 90 percent criterion for inclusion.

Number of Days of Mood Rating Employed

The PSY 201 subjects began their mood rating on 2/2/77 and continued

until at least 3/8/77. The PSY 340 subjects completed mood ratings from

1/26/77 through at least 3/8/77. Thus, an equivalent time sample of

mood ratings for each subject from 2/2/77 to 3/8/77 was available. The

time sample extending from 2/3 to 3/8 was employed for analyses, allowing

the first rating for nearly all of the PSY 201 subjects to be eliminated








as training period data. The equivalent time sample used included 99

observations per subject on each of the 16 mood rating scales.

Estimations of Missing Observations

Observations for each subject consisted of 48 mood ratings per day

for 33 successive days. This resulted in 1,584 possible mood ratings

per subject or 49,104 possible mood ratings for all 31 subjects. Of

these 49,104 possible observations, only 309 or approximately 0.6 percent

were missing. Of the 31 subjects, 21 subjects provided records with no

missing observations, while 10 subjects reported some missing observations.

The range of missing observations for these 10 subjects extended from 2

to 92, with a median of 16 missing observations, which was slightly over

1 percent.

Missing observations were estimated by recording the mean score

for the missing variable for each subject. This is a conservative

method for estimating missing observations as the mean score does not

provide a contribution to the variance of the estimated variable.

Instrumentation
Instruments for Assessment of Psychological Differentiation

Rod-and-frame test (RFT). A Marietta 18-10 RFT apparatus was

employed. It consisted of a 43" square frame, mounted on a black shield,

which enclosed a 41" rod. The angular orientation of both the rod and

the frame could be manipulated either singularly or together through

sets of controls mounted behind the shield. A separate set of hand

controls was provided for the subject to adjust the position of the rod.

Both the rod and the frame were covered with luminous paint which was

activated after exposure to a light source. The RFT was administered

in a totally dark room with the subject seated eight feet in front of








the RFT apparatus. Instructions for the administration of the RFT

are provided in Appendix C.

The standard eight trials on Series III (subject seated erect

and frame tilted 280 to the right or left) were administered. Witkin

et al. (1962) have reported that Series III alone can be substituted

for the index computed from Series I, II, and III together without any

loss of validity. The sum of the absolute number of degrees deviation

from true vertical for the eight trials was used as the subject's score.

Generally, a score of less than 180 deviation is considered an indicant

of greater psychological differentiation, while a score of greater than

400 is considered associated with less differentiation.

Considerable evidence supports the validity of the RFT as a measure

of the psychological differentiation construct (Witkin et al., 1962;

Witkin et al., 1973). Concerning the reliability of the RFT, odd-even

reliabilities of .89 or greater have been reported by Witkin et al.

(1954) and by Gardner, Jackson, and Messick (1960). Test-retest

reliabilities of .84 or greater have been reported for males and females

and for time intervals extending through three years (Adevai & McGough,

1968; Bauman, 1951).

Embedded figures test (EFT). The EFT, like the RFT, is an original

member of Witkin's (1954) perceptual battery used in the assessment of

psychological differentiation. In the EFT, the subject is required to

locate a simple geometric figure which is embedded within a more complex

figure. Instructions for the EFT are provided in Appendix D. The

subject's score for each simple figure is the amount of time necessary

for him to locate the simple figure, with a three minute time limit.

The full EFT consists of a successive presentation of 12 embedded







figures. The subject's total score is the time necessary to locate all

12 of the embedded figures. A shorter time is considered an indicant

of greater psychological differentiation.

The EFT has been employed extensively as a measure of psychological

differentiation and considerable evidence supporting its validity has

been accumulated (Witkin et al., 1962; Witkin et al., 1973). Reported

odd-even reliabilities include r = .88 (Loeff, 1961), r -.90 (Linton,

1952), and r = .92 (Longnecker, 1956). Test-retest reliabilities have

been reported to range from .92 after a one week interval (Dana & Goocher,

1959) to .89 after a three year interval (Bauman, 1951).

Instruments for Assessment of Identity Achievement

Ego identity incomplete sentences blank (EI-ISB). A modified 28

item form of Marcia's (1964) original 23 item EI-ISB was employed as a

measure of level of identity achievement (see Appendix F). The original

EI-ISB consisted of 23 items, which were selected according to their

fidelity to five theoretical criteria proposed by Erikson as seminal

to the assessment of ego identity. Briefly, these five criteria are

(1) occurrence of a crisis, (2) occurrence of and degree of commitment,

(3) continuity of self-definition over time, (4) degree of real self-

ideal self-discrepancy, and (5) commitment to an occupation and ideology.

The original EI-ISB appears to possess greater construct validity than

any other ego identity measure, with the exception of interview

techniques (Marcia, 1964, 1967).

Deldin (1976) modified the original EI-ISB by including five

additional items concerning sexual behaviors. This modification was

suggested by Schenkel and Marcia (1972), who reported that ego identity

in females may be more precisely predicted by including information







concerning sexual behaviors. Further, this modification is consistent

with Erikson's contention that a female's potential for reproduction,

with its biological and social implications, must be considered in

assessing ego identity. Thus, a sixth criterion, concerning crisis

and commitment in the sexual sphere, was added to the EI-ISB.

Subjects' responses to each of the 28 items were scored 3, 2, or

1, according to the six theoretical criteria and examples provided in

the EI-ISB Scoring Manual (see Appendix G). The EI-ISB Manual is based

upon the six theoretical criteria for assessing ego identity and upon

response examples taken from studies by Marcia (1964), Deldin (1976),

and the present investigator. A subject's score may range from 28 to

84, with a higher score considered indicative of a higher level of

identity achievement. Deldin (1976) reported the test-retest reliability

of the 28 item EI-ISB to be .77 for a two-week interval.

Training of raters and inter-rater reliability for EI-ISB. Three

female undergraduate raters were trained in scoring responses to the

EI-ISB. It was decided to employ only females as raters because it was

thought that they may be more sensitive to the nuances of ego identity

in women and, therefore, may be more accurate than males in assessing

ego identity in the female sample. Initially, the training of raters

emphasized Erikson's notion of ego identity, and instructions and

examples provided in the EI-ISB Scoring Manual. EI-ISB protocols

obtained from subjects who were not included in the final sample were

used in training the three raters.

The experimenter and three raters met three times for discussion

of rating criteria and for the rating of practice protocols. Ratings

for practice protocols were reviewed item-by-item. Discrepantly scored








items were discussed, clarified, and a final score was re-assigned.

In several cases, the EI-ISB Scoring Manual was amended by listing

additional examples.

A single practice protocol was scored by all three raters prior

to the fourth and fifth rating sessions. The mean intercorrelations

among raters for the protocol scored for Session 4 was .65 and the mean

intercorrelation for the Session 5 protocol was .92. Thus, consistent

scoring of EI-ISB items among the three raters appeared to be developing.

During Sessions 4 and 5, these two practice protocols were reviewed and

rating discrepancies were discussed and resolved.

At the end of Session 5, 15 practice protocols were divided into

three sets of 5 each and one set was provided for each rater to score

at home. During the next two sessions, raters exchanged their scored

protocols for a new set until each rater had completed scoring of all

15 protocols. During Sessions 6 and 7, reviews of practice protocols

were conducted in an attempt to minimize criterion drift.

After each of the three raters had completed scoring the 15 EI-ISB

protocols, Spearman-Brown reliability coefficients were computed (Winer,

1971, p. 286). The results of this procedure indicated that the

reliability of the mean rating of the three raters, of two of the three

raters, and of one rater alone, in estimating each subject's true total

score was, respectively, .93, .90, and .81. Thus, if 2 of 3 raters

scored each of the EI-ISB protocols, the reliability of their mean

rating in estimating each subject's true total score was .90. The

reliability coefficient of .90 for two of the three raters was considered

acceptable. For each rater to have scored all 49 EI-ISB protocols would

have been both uneconomical and tedious.







The final set of 49 EI-ISB protocols was divided into three sets

and each rater scored two of the three sets of protocols. It is noted

that all 49 subjects with accuracy> 75 percent were included in EI-ISB

scoring. Thus, two of the three raters scored each subject's protocol.

For the total set of 1,563 EI-ISB items, there was a total of 3,126

individual ratings. Among these ratings there were only 310, or

approximately 10 percent, disagreements. For the 31 subjects who

comprised the final sample for the total set of 868 items, there was

a total of 1,736 individual ratings. Of these there were only 168,

or approximately 10 percent, disagreements.

Rather than estimating each subject's identity achievement score

by obtaining the mean score of two raters, the experimenter individually

re-scored each of the 310 discrepantly scored items. This procedure

was based upon the rationale that the experimenter could perform expert

judgments when rating the identity achievement construct. Further, in

his decision making the experimenter could employ information already

provided by two raters on each discrepant item in formulating his

judgment. This procedure was thought to provide even more accurate

identity achievement scores than the purely statistical procedure of

relying on the mean score of the two raters for discrepantly scored items.

Identity achievement status (IAS) scale. The IAS Scale, developed

by Simmons (1970), is a 24 item objectively scorable self-report

instrument designed to assess identity achievement status (see Appendix

H). Items on the IAS Scale were derived from Marcia's (1964) EI-ISB.

The IAS Scale provides a single score indicating level of identity

achievement, rather than the nature of the identity resolution. The

IAS Scale has been reported to reliably predict crisis and commitment








ratings derived from Marcia's Interview Technique for the assessment

of ego identity (Simmons, 1970). Further, predicted correlations have

been reported between IAS Scale scores and various measures of

effectiveness of individual functioning, including the Edwards Personal

Preference Survey, the Personal Orientation Inventory, and self-ratings

of adjustment (Simmons, 1973b). Thus, information concerning scale

construction and validity suggest that it is a satisfactory objective

measure of identity achievement for the undergraduate population.

Simmons (1973a) reported test-retest reliability of .73 over a one-

week interval. However, as noted by Simmons (1973b), the IAS is not

considered an interchangeable substitute for more thorough methods for

the assessment of ego identity.

Instrument for Assessment of Mood

Wessman and Ricks' personal feelings scales (PFS). The PFS appear

to be the most widely used instrument for obtaining successive daily

self-reports of mood level. The PFS consist of 16 rationally-derived

10-point mood rating scales. Each scale was designed to assess mood

level on one of 16 bipolar mood dimensions. The Daily Record of Personal

Feelings is the response sheet for the PFS. Instructions on the Daily

Record request the subject to record retrospectively, "before retiring

every night," his daily "highest," "lowest," and "average" mood level

for each of the 16 mood dimensions.

Three modifications of the PFS and Daily Record were made in the

present study (respectively, Appendices I and B). The first modification

involved requesting subjects to rate "how I feel now" on each of the 16

mood scales for three time intervals .daily--morning, afternoon, and

evening. Time intervals were employed, rather than randomly assigning








specific rating times to subjects, in an attempt to make the mood

rating task less demanding and to insure cooperation.

One reason for this modification in mood rating procedure focuses

upon the distinction between "typical" mood ratings and "actual" mood

ratings. According to Nowlis (1965), typical mood ratings, in which

the subject is requested to rate his "typical" or "average" mood, may

be more subject to the fallibilities of memory, defensive operations,

and social desirability response sets than "actual" mood ratings. In

actual ratings, the subject reports "how I feel now," at the moment

of experience.

A second reason for this modification is based upon the review

of earlier studies in which P-technique factor analysis was employed

on a correlation matrix derived from each subject's Daily Record of

Personal Feelings. Each correlation matrix was constructed by

obtaining each subject's "highest," "lowest," and "average" ratings

for each of the 16 mood scales and then intercorrelating them. This

results in a 48 x 48 correlation matrix. The mood ratings from which

the correlations were computed were based upon 42, 30, and 28 successive

days of observation, respectively, for Wessman and Ricks (1965), Frank

(1967), and Gorman and Wessman (1974). Generally, it is not considered

permissible to employ common factor analysis upon a matrix in which

the correlations among variables are based upon less than 90-100

observations. Fewer observations could result in unreliable correlations

among variables, yielding unreliable factors. Such a factor solution

may not accurately reflect the subject's mood structure. Further,

correlations based upon a relatively small number of observations cannot

be assumed to accurately estimate the population parameters. Thus, it








would not be possible to reliably generalize from the sample to a

population of interest. From this perspective, the P-technique factor

analytic findings of the earlier studies may be seriously limited.

In the present investigation, each subject will report her actual

moods on each of the 16 PFS three times daily for 33 successive days.

Thus, the 16 x 16 correlation matrix which will be computed for each

subject will be based upon 99 observations for each of the 16 PFS.

The P-technique factors extracted from these matrices should be more

reliable and generalizability to a population of interest should be

more permissible, than for earlier studies.

The second modification of the PFS, which was also employed by

Frank (1967) and by Gorman and Wessman (1974), involved changing the

numbering of the 10 response categories for each of the 16 scales from

1-10 to 0-9 to facilitate data coding and processing.

Measures of Dependent Variables

Dependent variable measures of affective complexity, mood level,

and mood variability were constructed. P-technique factor analysis was

employed in constructing a measure of affective complexity, while

principal components analysis was used in constructing measures of mood

level and mood variability. Principal components analysis was employed,

rather than common factor analysis, for constructing mood level and

mood variability measures because scores on the PFS were based on data

from only 31 subjects. The specific procedures employed in the con-

struction of these measures are provided below.

Affective Complexity

The measure of affective complexity selected was the percentage of

total variance accounted for by the first unrotated factor extracted via








P-technique factor analysis from the correlation matrix based on each

subject's 33 day (99 observations) record of mood ratings. Initially,

five P-technique derived measures of affective complexity which had

been previously employed (Wessman & Ricks, 1965; Frank, 1967; Gorman

& Wessman, 1974) were obtained. Table 1 provides the intercorrelation

matrix for these five measures. The five measures were: (1) the

percentage of total variance extracted by the first unrotated factor

via P-technique factor analysis from each subject's mood rating record

(PCTVAR), (2) the absolute magnitude of the eigenvalue corresponding

to the first unrotated factor (EIGEN), (3) the number of unrotated

factors each accounting for an eigenvalue of magnitude greater than

1.0 (NBRGTI), (4) the number of unrotated factors necessary to account

for 90 percent total variance of the correlation matrix based on each

subject's mood rating record (NBRGT90), and (5) the number of unrotated

factors necessary to account for 95 percent total variance of each

subject's mood rating correlation matrix (NBRGT95).

Five measures of affective complexity were initially obtained

because little consensus has yet been reached concerning which are

the more appropriate measures. The rationale for employing PCTVAR

and EIGEN is that, in principle factor analysis, the first factor

accounts for the greatest proportion of variance of any factor. Thus,

if the first factor accounts for a relatively large proportion of

variance, then a considerable portion of the subject's mood ratings

are highly intercorrelated in a relatively global fashion. The pro-

portion of variance accounted for by a factor is reflected by the

magnitude of its corresponding eigenvalue. The rationale for employing

NBRGT1, NBRGT90, and NBRGT95 as measures of affective complexity is that








a smaller number of factors would describe a more compact correlation

matrix, therefore reflecting less affective complexity.

TABLE 1

INTERCORRELATIONS AMONG FIVE MEASURES
OF AFFECTIVE COMPLEXITY

EIGEN NBRGT1 NBRGT90 NBRGT95

PCTVAR .97 -.80 -.86 -.80

EIGEN -.72 -.84 -.78

NBRGT1 .71 .63

NBRGT90 .92

All correlations are significant at p <.001.

As indicated in Table 1, the five affective complexity measures were

highly intercorrelated in directions consistent with their theoretical

rationales. This was expected, as these measures are mathematically

related. An inspection of Table 1 revealed that PCTVAR recorded the

greatest magnitude intercorrelations with the other four measures.

Further, as indicated in Table 2, PCTVAR did not show the restriction

in range of NBRGT1, NBRGT90, and NBRGT95. Also, since PCTVAR could be

more readily understood by the general reader than EIGEN, PCTVAR was

selected as the affective complexity measure in the present investigation.

TABLE 2

DESCRIPTIVE STATISTICS FOR FIVE MEASURES
OF AFFECTIVE COMPLEXITY

STANDARD
VARIABLE MEAN DEVIATION RANGE

PCTVAR 70.17 11.22 52.40

EIGEN 9.04 2.11 10.20

NBRGT1 2.19 0.79 3.00

NBRGT90 3.58 0.92 3.00

NBRGT95 4.65 0.88 3.00








To summarize the procedure for constructing the affective com-

plexity measure, each subject recorded 99 observations on each of the

16 PFS. Scores on these scales were intercorrelated over occasions

and factored for each subject. Factoring was conducted by the

principle factor method with iterative estimates of communality. Five

initial affective complexity measures were obtained. After an inspection

of intercorrelations and descriptive statistics for these measures,

PCTVAR was selected as the statistically most satisfactory measure of

affective complexity.

Mood Level

Principal components analysis was employed in the construction of

a measure of mood level. This procedure was used in order to reduce the

information provided by the 16 PFS to a smaller number of independent

dimensions through the elimination of statistically redundant information.

Generally, a relatively small number of principal components can account

for the majority of the information provided by a larger number of

variables.

It was decided to employ principal components analysis to provide

greater conceptual clarity and computational convenience than previous

approaches taken to the operationalization of mood level using the PFS.

For example, Wessman and Ricks (1965) operationalized mood level as a

subject's mean score on PFS Scale 16 (Elation vs. Depression) alone.

This conceptualization of mood level, while intuitively appealing, did

not take advantage of the information provided by the other 15 PFS.

Other investigators (Gorman & Wessman, 1974; Schaff, 1975) have con-

ceptualized mood level in terms of each of the 16 PFS, as if these

scales each reflected an independent affective dimension. This approach







to the operationalization of mood level neglected the dependencies or

redundant information among the 16 PFS. Further, the use of 16

separate mood level scores as dependent variables raises the problem

of incremental Type I error.

Based upon earlier findings concerning the normative factorial

composition of the PFS (Wessman & Ricks, 1965), it was hypothesized

that one dominant principal component would be obtained. It was

expected that this component would be highly positively loaded by each

of the 16 PFS. This component was expected to reflect "overall mood

level." If this finding was obtained, a subject's factor score on

this component would be conceptualized to provide an indicant of her

overall mood level, also referred to as general level of happiness or

"hedonic level" (Wessman & Ricks, 1965).

For each subject, 16 mean mood level scores were computed by

obtaining the average rating for each of the 16 PFS across 99 occasions

of measurement. The 16 mean PFS scores were intercorrelated across

the 31 subjects using the Pearson product-moment statistic. The

resultant 16 x 16 correlation matrix was factored by principal com-

ponents analysis.

The findings of the analysis were consistent with predictions.

Since the 16 PFS were highly intercorrelated, the first principal

component, labeled Mood Level Principal Component One (MLPCI), with an

eigenvalue of 12.36, accounted for 77 percent of the total variance

of the correlation matrix. All other components yielded eigenvalues

less than 1.0 and, therefore, were not included in subsequent analyses.

Mood Level Principal Component One totally dominated the findings. The

factor pattern matrix for MLPCI is provided in Table 3, which lists the

loadings of each of the 16 PFS on MLPC1.








TABLE 3

FACTOR PATTERN MATRIX FOR MLPC1

PFS SCALE LOADING PFS SCALE LOADING

SCALE 1 .93 SCALE 9 .83

SCALE 2 .94 SCALE 10 .90

SCALE 3 .89 SCALE 11 .90

SCALE 4 .91 SCALE 12 .87

SCALE 5 .93 SCALE 13 .88

SCALE 6 .89 SCALE 14 .93

SCALE 7 .90 SCALE 15 .92

SCALE 8 .50 SCALE 16 .94

Lambda (\) = 12.36

As indicated in Table 3, each of the 16 PFS means, with the exception

of Scale 8 (Love and Sex), recorded large magnitude positive loadings,

greater than .80, on MLPC1. This provides a clear substantive inter-

pretation of MLPC1 as reflecting overall mood level. Ten of the 16 PFS

mean scores loaded .90 or greater on this component. Scale 2 (Receptivity

Towards World) and Scale 16 (Elation vs. Depression), both with loadings

of .94, loaded most strongly on MLPC1. This provides further support

towards interpreting this component as overall mood level.

Each subject's factor score was computed on MLPC1. These factor

scores were employed as the dependent variable scores for overall mood

level.

Mood Variability

Principal components analysis was employed in the construction of two

mood variability measures from PFS scores. The standard deviation of

each of the 16 PFS for each subject across 99 occasions was computed.








The standard deviations for the 16 PFS were intercorrelated across the

31 subjects using the Pearson product-moment statistic. The resultant

16 x 16 correlation matrix was factored by principal components analysis.

As with mood level, it was hypothesized that one major component

for mood variability would be obtained, on which all 16 PFS standard

deviations would have high positive loadings. This component was

expected to provide a reflection of the dimensionality of the mood

variability in this sample. Factor scores would then be computed for

this component and employed as the mood variability dependent measure.

A greater magnitude factor score would be associated with greater overall

mood variability, while a smaller factor score would reflect lesser

variability.

The principal components analysis yielded two components with

eigenvalues greater than 1.0, which, together, accounted for 72.1 percent

of the total variance of the correlation matrix of PFS standard deviations.

The first mood variability principal component (MVPCI), with an eigenvalue

of 9.84, accounted for 61.5 percent of the total variance. The factor

pattern matrix for MVPC1, as indicated in Table 4, shows that each of

the 16 PFS standard deviations recorded moderately-high to high positive

loadings on MVPC1, with the exception of Scale 8 (Love and Sex) and Scale

9 (Present Work), which recorded moderate loadings. This is clearly

consistent with the interpretation that MVPC1 reflects a general dimension

of overall mood variability. A factor score was computed for each subject

for MVPC1 and employed as the dependent variable measure of overall mood

variability.








TABLE 4

FACTOR PATTERN MATRIX FOR MVPC1

PFS SCALE LOADING PFS SCALE LOADING

SCALE 1 .82 SCALE 9 .42

SCALE 2 .88 SCALE 10 .66

SCALE 3 .78 SCALE 11 .88

SCALE 4 .71 SCALE 12 .83

SCALE 5 .90 SCALE 13 .86

SCALE 6 .85 SCALE 14 .78

SCALE 7 .85 SCALE 15 .78

SCALE 8 .50 SCALE 16 .87

Lambda () = 9.84

Mood Variability Principal Component Two (MVPC2), with an eigenvalue

of 1.69, accounted for 10.6 percent of the total variance of the cor-

relation matrix of PFS standard deviations. The marker variables for

bipolar MVPC2 were Scale 8 (Love and Sex), with a loading of -0.60, and

Scale 9 (Present Work), with a loading of 0.73. Scale 4 (Personal Freedom

vs. External Constraint) recorded a moderate negative loading of -0.34 on

MVPC2. Scales which recorded moderate positive loadings on this component

were Scale 10 (Thought Processes), Scale 14 (Self-Confidence vs. Feelings

of Inadequacy), and Scale 15 (Energy vs. Fatigue), with loadings of,

respectively, 0.49, 0.36, and 0.34. The factor pattern matrix for

MVPC2 is listed in Table 5.

To speculate, MVPC2 appears to reflect a bipolar mood dimension

upon which less variability in reported satisfaction concerning intimacy

and impulse expression appears associated with greater variability in

reported satisfaction concerning academic work performance. When the








direction of the scale loadings is reflected, greater reported variability

in satisfaction concerning intimacy and impulse expression appears

associated with lesser variability in satisfaction concerning academic

work performance. Although the magnitude of the salient loadings on

MVPC2 appear large enough to support substantive interpretation, the

speculative nature of this interpretation must be emphasized.

TABLE 5

FACTOR PATTERN MATRIX FOR MVPC2

PFS SCALE LOADING PFS SCALE LOADING

SCALE 1 -.18 SCALE 9 .73

SCALE 2 .01 SCALE 10 .49

SCALE 3 .05 SCALE 11 -.15

SCALE 4 -.34 SCALE 12 -.19

SCALE 5 -.17 SCALE 13 -.13

SCALE 6 -.01 SCALE 14 .36

SCALE 7 -.14 SCALE 15 .34

SCALE 8 -.60 SCALE 16 .18

Lambda (A) = 1.69

Relatively large magnitude positive factor scores on MVPC2 would

appear to indicate that the subject is reporting greater mood variability

in areas concerning academic work, concurrent with lesser mood variability

in areas concerning intimacy and impulse expression. Relatively large

magnitude negative factor scores would appear to show that the subject

is reporting greater mood variability in areas concerning intimacy and

impulse expression, concurrent with lesser variability associated with

academic work concerns. Intermediate magnitude factor scores would

appear to indicate relatively similar magnitudes of mood variability








in both intimacy and impulse expression, as well as academic work

areas.

To speculate, from an Eriksonian perspective, a subject's factor

score on MVPC2 may reflect a particular content area which has emerged

into salience during the period of identity formation. From this

perspective, greater mood variability associated with a particular

psychosocial content area may indicate that the person is actively

engaged in conflict or growth in that area. Further, the content area

which becomes the focus of conflict or growth may, perhaps, be related

to the subject's level of psychological differentiation. Both greater

and lesser differentiation are reported to be associated with particular

"assets" and "liabilities" for individual functioning. Greater dif-

ferentiation, associated with well-developed cognitive analytic abilities,

appears to be a definite asset in academic and work performance. Thus,

more differentiated persons would be expected to report less mood

variability in the academic work area, as they possess a set of abilities

conducive to effective functioning in that area. However, in the inter-

personal domain, more differentiated persons have been reported to be

"task oriented, rather than person oriented," "individualistic," "distant,"

and to prefer solitary activities (Witkin & Goodenough, 1976). This set

of interpersonal attributes makes it reasonable to speculate that

effective functioning may be more difficult forthe more differentiated

person when dealing with interpersonal concerns, such as intimacy issues,

than when dealing with academic work issues. Thus, greater differentiation

would be hypothesized to be associated with greater mood variability

concerning intimacy issues than concerning academic work issues.








Conversely, lesser differentiation, although associated with lesser

cognitive analytic abilities, is associated with a constellation of

characteristics which appear to be conducive to the establishment and

maintenance of intimacy. Lesser differentiated persons have been

reported to be "social," to enjoy interpersonal activities, and to be

person-oriented (Arbuthnot, 1968; Baer, 1964; Oltman, Goodenough, Witkin,

Freedman, & Friedman, 1975). Thus, for a lesser differentiated person,

effective functioning may be more difficult when dealing with academic

work performance issues, than when dealing with interpersonal intimacy

issues. Therefore, lesser differentiation would be hypothesized to be

associated with greater reported mood variability concerning academic

work issues than with intimacy issues.

A factor score was computed for each subject for MVPC2 and employed

as the dependent measure of differential mood variability in the academic

work and intimacy domains.

Construction of Indices

Table 6 provides the descriptive statistics for all selection

instruments in the 31 subject sample. The means and standard deviations

for the instruments selected as measures of psychological differentiation

TABLE 6

MEANS AND STANDARD DEVIATIONS FOR
SELECTION INSTRUMENTS

STANDARD
INSTRUMENT MEAN DEVIATION

RFT 32.23 24.78

EFT 622.16 405.15

IAS 13.32 2.63

EI-ISB 55.65 4.10

N = 31







and identity achievement were comparable to those reported in previous

studies (Witkin et al., 1962; Deldin, 1976). None of the distributions

of scores for these instruments demonstrated skewness or kurtosis to

the extent to warrant transformation.

As indicated in Table 7, which provides the intercorrelations among

the selection measures, none of the correlations of the RFT and EFT,

conceptualized as measures of psychological differentiation, with the

IAS and EI-ISB, conceptualized as measures of identity achievement,

exceeded 0.11. The relatively small magnitude of these correlations

provides support for the notion that psychological differentiation

and identity achievement, as assessed by these instruments, are

independent constructs.

TABLE 7

INTERCORRELATION MATRIX FOR
SELECTION INSTRUMENTS

EFT IAS EI-ISB

RFT .60* .02 .01

EFT .11 .09

IAS .11
*p/.001, one-tailed test

As indicated in Table 7, the RFT and EFT correlated 0.60, reflecting

approximately 36 percent shared variance. The magnitude and direction of

this correlation coefficient, which is consistent with previous reported

findings (Witkin et al., 1962), supports the use of these two measures

together, as an index, to assess psychological differentiation.

However, the two instruments selected to assess identity achievement,

the IAS and the EI-ISB, correlated only 0.11. The magnitude of this








correlation coefficient indicated that the IAS and the EI-ISB share

approximately 1 percent variance. Further, an inspection of the

plot of JAS and EI-ISB scores did not provide evidence of a non-linear

relationship between scores on these two instruments. This 1 percent

shared variance appears particularly small considering that the IAS

and the EI-ISB were developed from the same pool of original items--

Marcia's (1964) EI-ISB, as measures of the same construct. Thus, it

was not reasonable to contend that these two instruments may be

assessing different facets of identity achievement. It clearly is not

defensible to combine scores on these two instruments in computing an

identity achievement index.

An inspection of the descriptive statistics and distributional

characteristics of scores for the IAS and for the EI-ISB did not provide

evidence that either of the measures was statistically suspect. Indeed,

the obtained means and standard deviations for the IAS and the EI-ISB,

provided in Table 6, both appear comparable to norms reported in

previous investigations. For the IAS, Simmons (1970) reported a mean

of 12.39 and a standard deviation of 3.18. For the EI-ISB, Deldin (1976)

reported means and standard deviations of, respectively, 55.8 and 4.90.

However, since the EI-ISB appears to possess greater construct

validity than the IAS as a measure of identity achievement (Marcia, 1966;

Breuer, 1974; Deldin, 1976), it was selected as the measure of identity

achievement in the present investigation. Indeed, Simmons (1973b)

recommends that the IAS not be used as a substitute for more thorough

assessments of ego identity unless a short, quick inventory is judged

appropriate.








Formation of Groups

Each subject's score on the psychological differentiation (PD)

index was obtained by computing the mean of her standard scores on

the RFT and on the EFT. Identity achievement (IA) was assessed by

each subject's score on the EI-ISB. The correlation between scores

on the PD index and scores on the IA measure was .05.

Initially, four groups (Groups 1-4) were constructed by median

splits in such a manner as to correspond with Patterns 1-4. Groups

were formed based upon median splits performed on the distribution of

scores for the PD index and the distribution of scores for the IA

measure. Each subject was assigned to one of four groups based upon

her score on each of the two measures, relative to the median score for

the two measures.

Next, the five subjects whose PD and IA scores appeared to most

distinctly represent each of Patterns 1-4 were selected for membership

in each of four extreme groups. Thus, to form Extreme Group 1, which

corresponded to Pattern 1, the five subjects reporting the lowest PD

scores (indicative of the highest level of differentiation), concurrent

with the highest IA score (indicative of the greatest effectiveness of

integration) were selected. The other three groups were similarly

formed. All tests of hypotheses concerning group differences were

conducted with these four extreme groups of five subjects each. Since

the final sample consisted of only 31 subjects, this extreme group

strategy was considered necessary to more clearly demonstrate effects

attributable to the four patterns.

Descriptive statistics for all variables for the four groups formed

by median splits, as well as for the four extreme groups of five subjects








each, are provided in Tables 8 and 9, respectively. Throughout the

remainder of the investigation, the four extreme groups will be referred

to as Groups 1-4 unless otherwise noted.

Statistical Hypotheses
A set of statistical hypotheses are listed for each of the four

dependent variables--affective complexity, overall mood level, overall

mood variability, and mood variability concerning academic work vs.

intimacy issues.

Affective Complexity

(1) It is hypothesized that smaller PD scores (indicative of greater

PD) will be associated with smaller affective complexity scores (indicative

of greater affective complexity). (2) It is hypothesized that the two

theoretically more differentiated groups (Groups 1 and 4) will report

smaller affective complexity scores than the two less differentiated

groups (Groups 2 and 3).

Overall Mood Level

(1) It is hypothesized that higher IA scores (indicative of greater

effectiveness of personality integration) will be associated with higher

scores on MLPC1 (indicative of a higher overall mood level). (2) It

is hypothesized thatthe two theoretically more effectively integrated

groups (Groups 1 and 2) will report a higher MLPC1 score than the two

less effectively integrated groups (Groups 3 and 4).

Overall Mood Variability

(1) It is hypothesized that smaller PD scores (indicative of

greater PD) will be associated with smaller scores on MVPC1 (indicative

of less overall mood variability). (2) It is hypothesized that the two

theoretically more differentiated groups (Groups 1 and 4) will report






























C -- (n f-
S- cz

C\j


0l- ,-
C0) CT)

0)


i-n
uo-

CO o)
00 cc


i-Of






L-u
>-I






C,- -


IL ,"J

CD


LL.




I.J



F- F-
0-













C
Q <
C> ,
)-
LU








--
0 -





:-4
C)

> CD
_ILL
_Il-



SI--7


I- 0


I--"




1-d


>1
(C)



C G) w I-
u On LL- u Q
C CC c: Ca LUJ Q-


LU)
OG ) C 0 0 0C
co .- CD CD CD

CO rN 0 o



c Ci


S-


0 LLj
F-- :


Q,


C -I > "I 0
a- x S i 2


E
0
4-



0
(1)
S-/
u


-o
C3

4-)











C









-j
C-)






















--
CDJ
a3-













0
4-















0
C_)







._>
Q-
O
4-


-0
o'3


C\J




CY)




00


i







59

















m C) o in


r-- Z in o t - ,-
<> C O
<- .. CD
F- LUd





Si in
<- C) I 'D C) C) C) -S-
I- C

O Ln





t- ,"---,--"----I-- '---------------- ,,--
0 -C
m/i iC C0 i n3
cD G-i ^ C 0o) o in o <3- C) ^o i +o
OL l O C)O D W 1 M l
OD LLJ 0) C 0 0 (.0 cL CI" q0 L
: D Ln C) 0 o i in C. o o r. o o Co c
n - c)i i in '- i II


SC) .-




ICL0 0 0 0
:: O 0 0C 0) C\J
Lu CD. C) : CD C) co C C'] Ln Ln u u
Cjo r iC C 0 n C
Go2 iD- Ql *
< F- 2) iLn Co Co 00) r- c0 c) CD CD C
=- 0) C)o C- in Ln in nO 0



















< 0
l- O C co O
-= 2 M C \ C C) C) C J- 0C (\J C\] >- CD L
Fin C CT) cc (D mC C) oYo LC C)

F- C'\j


C-)

-o CD CD C) in C-) cc
0) CD z- O in t 0 r-- CY
) in C r- 0 C; a; C\J CD C) C) C
C-) o) c] C- C') i in C r I i

C) 4-

V)



C.- C/ < r\
:S- in 'c: L- (NJ
u .- I C- CL 4- U)
< 0 0) LL U LL C O -1 Cu) j 0
:|- cU < 3- Qr L) 0- LF- Q- CL Z: CL







smaller MVPC1 scores than the two less differentiated groups (Groups 2

and 3). (3) It is hypothesized that Group 3 (lower differentiation/less

effectively integrated) will report the highest MVPC1 scores, followed

by Group 2 (lower differentiation/more effectively integrated), Group 1

(higher differentiation/more effectively integrated), and with Group 4

(higher differentiation/less effectively integrated) reporting the

smallest MVPC1 scores.

Mood Variability Concerning Academic Work vs. Intimacy Issues

(1) It is hypothesized that smaller PD scores (indicative of greater

PD) will be associated with smaller scores on MVPC2 (indicative of greater

mood variability concerning intimacy issues than academic work issues).

(2) It is hypothesized that the two theoretically more differentiated

groups (Groups 1 and 4) will report a smaller MVPC2 score than the two

less differentiated groups (Groups 2 and 3).














CHAPTER III

RESULTS

General Strategy for Results

In the results section, the relationships among the dependent

variables will be presented first. Then, tests of hypotheses will

be presented for each of the four dependent variables separately. For

each of the four dependent variables, three levels of hypothesis testing

were employed. These were:

(1) The hypotheses concerning PD, IA, and the dependent variable

were tested via a multiple regression analysis performed on the entire

31 subject sample. The multiple regression analysis employed a stepwise

forward selection procedure for the inclusion of predictor variables.

The order of entry of predictor variables into the regression equation

was determined by the maximization of variance criterion. Thus, the

predictor variable which accounted for the greatest proportion of the

criterion variable variance was entered into the regression equation first.

(2) The hypotheses concerning differences in group means were tested

via a fixed effects 2 x 2 factorial analysis of variance (ANOVA). The

20 subjects in Groups 1-4 were classified according to two factors, PD

(greater differentiation/lesser differentiation) and IA (greater

effectiveness of integration/lesser effectiveness of integration). Thus,

the main effect due to PD tested the significance of the mean difference

between the two combined theoretically more differentiated groups (Groups








1 and 4) vs. the two less differentiated groups (Groups 2 and 3). The

main effect due to IA tested the significance of the mean difference

between the two combined theoretically more effectively integrated

groups (Groups 1 and 2) vs. the two less effectively integrated groups

(Groups 3 and 4). The PD x IA interaction tested whether the effect of

the levels of one factor depends upon their combination with the levels

of the other factor.

(3) More refined analysis of differences among means for Groups 1-4

were conducted by inspection of group means.

All hypotheses were tested at a pre-set alpha level of .05.

Relationships Among Dependent Variables

Table 10 provides the intercorrelation matrix among the dependent

variables. The correlation coefficient between MVPC1 and PCTVAR was

statistically significant (r = .48, p <.003). The correlation between

MLPC1 and MVPC1 (r = .25, p<*.09) indicated a trend which did not achieve

significance.

TABLE 10

INTERCORRELATION MATRIX FOR
DEPENDENT MEASURES

MVPC1 MVPC2 PCTVAR

MLPC1 .25* .04 .06

MVPC1 -.01 .48**

MVPC2 -.13
p<.09, one-tailed test

** p .003, one-tailed test









Affective Complexity

It was hypothesized that smaller PD scores would be associated

with smaller affective complexity scores. This hypothesis was not

supported.

To test this hypothesis, a multiple regression analysis using

stepwise forward inclusion of predictor variables was performed.

Affective complexity scores were employed as the criterion variable,

while predictor variables were PD and IA scores.

A summary of the relationships among affective complexity, PD, and

IA scores is provided in Table 11. Psychological differentiation, with

a multiple R of .41, accounted for 17 percent of the affective com-

plexity score variance, and was entered into the regression equation

first. The effect due to PD was statistically significant (F1,29 = 5.84,

p 4.05). The unique contribution of PD, obtained when IA was entered into

the regression first, remained at 17 percent of the affective complexity

score variance. However, the direction of the simple r between PD and

affective complexity scores (r = -.41) indicated that smaller PD scores

were significantly associated with greater affective complexity scores.

TABLE 11

SUMMARY TABLE FOR MULTIPLE REGRESSION OF
AFFECTIVE COMPLEXITY OF PD AND IA

CUMULATIVE CUMULATIVE SIMPLE UNIQUE F FOR
VARIABLE MULTIPLE R R2 r R2 UNIQUE R2

PD 0.41 .17 -0.41 .17 5.84*

IA 0.42 .17 0.08 z.Ol0 1.00
* p <.05








It was hypothesized that the two theoretically more differentiated

groups (Groups 1 and 4) would report smaller affective complexity scores

than the two less differentiated groups (Groups 2 and 3). This hypothesis

was not supported.

To test this hypothesis, a 2 x 2 factorial ANOVA was performed with

affective complexity scores as the dependent variable and with PD and IA

as factors. Prior to conducting the ANOVA, the Bartlett-Box test of

homogeneity of variance among groups was conducted. The obtained F

(F3,CD = 1.01) was not significant, therefore pooled variance estimates

were employed int he ANOVA. As indicated in Table 12, which provides the

summary table for the ANOVA on affective complexity scores, only the main

effect due to PD was significant (F1, 16 = 5.08, p <.05). Thus, the mean

affective complexity score for the two theoretically more differentiated

groups differed significantly from the mean score for the two less dif-

ferentiated groups. However, as indicated in Table 13, the direction of

this difference was opposite from that predicted.

TABLE 12

SUMMARY TABLE OF ANALYSIS OF VARIANCE
FOR AFFECTIVE COMPLEXITY SCORES

SUM OF MEAN
SOURCE SQUARES Df SQUARES F P

PD 648.65 1 648.65 5.08 4.05

IA 11.71 1 11.71 41.00 ns

PDxIA 37.26 1 37.26 <1,.00 ns

WITHIN 2042.38 16 127.65

TOTAL 19

The mean affective complexity scores for each of Groups 1-4, as well

as for the two combined groups, are provided in Table 13.









TABLE 13

MEAN AFFECTIVE COMPLEXITY SCORES FOR GROUPS 1-4,
COMBINED GROUPS 1 AND 4, AND
COMBINED GROUPS 2 AND 3

AFFECTIVE
GROUP COMPLEXITY SCORE n

GROUP 1 78.98% 5
GROUP 2 64.32% 5
GROUP 3 63.12% 5
GROUP 4 77.24% 5

GROUPS 1 AND 4 75.11% 10
GROUPS 2 AND 3 63.72% 10

An inspection of the group means in Table 13 reveals that the effect

due to PD was consistently obtained in each of Groups 1-4. Group 1

recorded the largest affective complexity score, while Group 3 recorded

the smallest score.

Overall Mood Level

It was hypothesized that higher IA scores would be associated

with higher MLPC1 scores. This hypothesis was not supported.

To test this hypothesis, a multiple regression analysis using

step-wise forward inclusion of predictor variables was performed.

The criterion variable was MLPC1, while PD and IA were employed as

predictor variables.

As indicated in Table 14, neither IA, nor PD, nor IA and PD

together, accounted for even 1 percent of the variance of MLPC1

scores.








TABLE 14

SUMMARY TABLE FOR MULTIPLE REGRESSION
OF MLPC1 ON PD AND IA

CUMULATIVE CUMULATIVE SIMPLE UNIQUE F FOR
VARIABLE MULTIPLE R rR2 r R UNIQUE R2

PD 0.05 4.01 0.05 4.01 1.00

IA 0.07 <.01 -0.04 J .01 < 1.00

It was hypothesized that the two theoretically more effectively

integrated groups (Groups 1 and 2) would report a higher MLPC1 score than

the two less effectively integrated groups (Groups 3 and 4). This

hypothesis was not supported.

To test this hypothesis, a 2 x 2 factorial ANOVA was performed with

MLPC1 scores as the dependent variable and with PD and IA as the factors.

Prior to conducting the ANOVA, the Bartlett-Box test for homogeneity of

variances indicated no significant differences among groups (F3,00 = 1.38,

ns). Thus, pooled variance estimates were employed in the ANOVA. As

indicated in Table 15, neither main effects, nor interaction, was

significant.

TABLE 15

SUMMARY TABLE OF ANALYSIS OF VARIANCE
FOR MLPC1 SCORES

SUM OF MEAN
SOURCE SQUARES Df SQUARES F P

PD 1.25 1 1.25 (1.00 ns

IA 0.28 1 0.28 ( 1.00 ns

PDxIA 0.60 1 0.60 /,1.00 ns

WITHIN 23.22 16 1.45

TOTAL 19

The mean MLPC1 scores, in standard score form, for Groups 1-4, as

well as for the two combined groups, are provided in Table 16. As









indicated in Table 16, the MLPC1 group means cluster relatively close

together, with a range of 0.8458 standard deviation units. Thus, none

of the mean MLPC1 score differences among the four groups were signi-

ficant.

TABLE 16

MEAN MLPC1 SCORES FOR GROUPS 1-4, COMBINED
GROUPS 1 AND 2, AND COMBINED
GROUPS 3 AND 4

GROUP MLPC1 SCORE n

GROUP 1 -0.1776 5

GROUP 2 -0.0245 5

GROUP 3 0.5575 5

GROUP 4 -0.2883 5

GROUPS 1 AND 2 -0.1011 10

GROUPS 3 AND 4 0.1346 10

Because the finding that IA and PD scores together did not account

for even 1 percent of the MLPC1 score variance was unexpected, further

analyses were conducted. Inspection of a scatterplot for MLPC1 scores

did not appear to indicate a restriction in range, nor was evidence of

serious skewness or kurtosis revealed. Inspection of scatterplots among

IA, PD, and MLPC1 scores did not appear to indicate a curvilinear com-

ponent. A curvilinear multiple regression of MLPC1 on IA and PD

revealed a negligible curvilinear component accounting for
of the MLPC1 score variance. Further, the residual scores obtained

from the regression of MLPC1 on IA and PD were plotted and inspected.

No strong evidence of heteroskedasticity was obtained.

Overall Mood Variability

It was hypothesized that smaller PD scores would be associated with

smaller MVPC1 scores. This hypothesis was not supported.









To test this hypothesis, a multiple regression analysis employing

stepwise forward inclusion of predictor variables was performed. The

criterion variable was MVPC1, while the predictor variables were PD

and IA.

Identity achievement, with a multiple R of .47, was entered into the

regression equation first. As indicated in Table 17, IA, with a simple

r of -.47, accounted for 22 percent of the MVPC1 score variance. The

unique contribution of IA to MVPC1, obtained when PD was entered into

the regression equation first, was 21 percent. The magnitude of this

unique contribution was significant (F1,29 = 8.01, p (.05). Thus, higher

IA scores were significantly associated with smaller MVPC1 scores.

TABLE 17

SUMMARY TABLE FOR MULTIPLE REGRESSION
OF MVPC1 ON PD AND IA

CUMULATIVE CUMULATIVE SIMPLE UNIQUE F FOR
VARIABLE MULTIPLE R R2 r R2 UNIQUE R2

IA 0.47 .22 -0.47 .21 8.01*

PD 0.51 .26 -0.22 .04 1.60

* p .05

Psychological differentiation, with a simple r of -0.22, accounted for

approximately 4 percent of the MVPCI score variance. The effect due to

PD was not significant (F1,29 = 1.60, ns). Further, it was noted in

Table 17 that the cumulative R2 and the unique R2 values for IA and for

PD are almost identical. Thus, IA and PD are each accounting for

independent variance components of MVPC1 scores.

It was hypothesized that the two theoretically more differentiated

groups (Groups 1 and 4) would report smaller MVPC1 scores than the two








less differentiated groups (Groups 2 and 3). This hypothesis was not

supported.

To test this hypothesis, a 2 x 2 factorial ANOVA was performed with

MVPC1 scores as the dependent variable and with PD and IA as factors.

Prior to conducting the ANOVA, the Bartlett-Box test of homogeneity of

variances indicated no significant differences among groups (F3,W =

1.00, ns), allowing the use of pooled variance estimates in the ANOVA.

As indicated in Table 18, only the main effect due to IA was significant

(F1,16 = 4.50, p <.05). Thus, the mean MVPC1 score of the two theo-

retically more effectively integrated groups differed significantly from

the mean score of the two less effectively integrated groups.

TABLE 18

SUMMARY TABLE OF ANALYSIS OF VARIANCE
FOR MVPC1 SCORES

SUM OF MEAN
SOURCE SQUARES Df SQUARES F P

PD 0.47 1 0.47 / 1.00 ns

IA 4.65 1 4.65 4.50 (.05

PDxIA 0.78 1 0.78 <1.00 ns

WITHIN 16.56 16 1.04

TOTAL 19

Table 19 provides a listing of mean MVPC1 scores for Groups 1-4, for

the two combined groups reflecting greater and lesser differentiation,

and for the two combined groups reflecting greater and lesser effective-

ness of integration. From an inspection of Table 19, it is apparent

that the two more effectively integrated groups reported a significantly

smaller MVPC1 score than the two less effectively integrated groups.








It was hypothesized that Group 3 would report the highest MVPC1

score, followed by Group 2, Group 1, and Group 4 with the smallest

MVPC1 score. This hypothesis was not supported. An inspection of the

means for Groups 1-4, listed in Table 19, reveals that Group 4 reported

the highest MVPC1 score, followed by Group 3, then Group 2, and with

Group 1 reporting the smallest score.

TABLE 19

MEAN MVPC1 SCORES FOR GROUPS 1-4 AND
ALL COMBINED GROUPS

GROUP MVPC1 SCORE n

GROUP 1 -0.3993 5

GROUP 2 -0.3104 5

GROUP 3 0.2586 5

GROUP 4 0.9611 5

GROUPS 1 AND 4 0.2809 10

GROUPS 2 AND 3 -0.0259 10

GROUPS 1 AND 2 -0.3549 10

GROUPS 3 AND 4 0.6099 10

An inspection of the MVPC1 means for Groups 1-4 indicated that

the effect due to IA is relatively consistent for all four groups,

although the effect is more consistent for the theoretically more

effectively integrated groups (Groups 1 and 2), than for the less

effectively integrated groups (Groups 3 and 4). Group 1 reported the

smallest MVPC1 score, while Group 4 recorded the greatest score.








Mood Variability Concerning Academic Work vs. Intimacy Issues

It was hypothesized that smaller PD scores would be associated with

smaller scores on MVPC2. This hypothesis was supported.

To test this hypothesis, a multiple regression analysis employing

stepwise inclusion of predictor variables was performed. Predictor

variables were PD and IA, while MVPC2 was the criterion variable.

As indicated in Table 20, PD, with a multiple R of .33, was entered

into the regression equation first. PD scores accounted for 11 percent

of the variance of MVPC2 scores. This effect approached, but did not

achieve significance at the .05 level (F10,1,29 = 2.89Fobtained, 1,29

3.46(F.05,1,29 = 4.17). However, the simple r between PD and MVPC2 was

significant (r = .33, p/.05, one-tailed test). Thus, smaller PD scores

were significantly associated with smaller MVPC2 scores. As indicated

by the magnitude of the simple r (r = .05) and the unique R2 (R2 = .01),

IA was not related to MVPC2 scores. Even when entered into the regression

equation first, IA accounted for less than 1 percent of the MVPC2 score

variance.

TABLE 20

SUMMARY TABLE FOR MULTIPLE REGRESSION
OF MVPC2 ON PD AND IA

CUMULATIVE CUMULATIVE SIMPLE UNIQUE F FOR
VARIABLE MULTIPLE R R2 r R2 UNIQUE R2

PD 0.33 .11 0.33 .11 3.46*

IA 0.33 .11 0.05 < .01 < 1.00

* p .10

It was hypothesized that the two theoretically more differentiated

groups (Groups 1 and 4) would report a smaller MVPC2 score than the two









less differentiated groups (Groups 2 and 3). This hypothesis received

mixed support.

To test this hypothesis, a 2 x 2 factorial ANOVA was performed

with MVPC2 scores as the dependent variable and with PD and IA as factors.

Prior to conducting the ANOVA, the Bartlett-Box test of homogeneity of

variances was conducted and a non-significant F statistic was obtained

(F3,M C= 2.24, ns), allowing the use of pooled variance estimates in

the ANOVA. As indicated in Table 21, neither main effects nor interaction

was significant for the ANOVA performed on MVPC2 scores.

TABLE 21

SUMMARY TABLE OF ANALYSIS OF VARIANCE
FOR MVPC2 SCORES

SUM OF MEAN
SOURCE SQUARES Df SQUARES F P

PD 1.80 1 1.80 1.57 ns

IA 0.01 1 0.01 4 1.00 ns

PDxIA 1.13 1 1.13 / 1.00 ns

WITHIN 18.29 16 1.14

TOTAL 19

It is noted that, although the Bartlett-Box F statistic was not

significant at the .05 level, it was significant at the .08 level. Thus,

there is considerable variability among the MVPC2 scores within each of

the four groups.

A listing of mean MVPC2 scores, in standard score form, for Groups

1-4, and for the two combined groups, is provided in Table 22. An

inspection of the mean MVPC2 scores for Groups 1-4, as well as for the

two combined groups, reveals that the direction of all group mean scores

is consistent with predictions. Thus, although the magnitude and









direction of all group means are consistent with predictions, the

magnitude of the Bartlett-Box F statistic indicates that there is

considerable variability within each of the groups. Therefore, the

ANOVA did not indicate a significant main effect due to PD.

TABLE 22

MEAN MVPC2 SCORES FOR GROUPS 1-4, COMBINED
GROUPS I AND 4, AND COMBINED
GROUPS 2 AND 3

GROUP MVPC2 SCORE n

GROUP 1 -0.4190 5

GROUP 2 0.6549 5

GROUP 3 0.2153 5

GROUP 4 0.0901 5

GROUPS 1 AND 4 -0.1645 10

GROUPS 2 AND 3 0.4351 10














CHAPTER IV

DISCUSSION
The use of the PD and IA constructs, respectively, as indicants

of personality differentiation and the effectiveness of personality

integration, appeared to provide a productive approach to the investi-

gation of mood/personality relationships. A number of substantial

relationships among PD, IA, and mood variables were obtained. The use

of PD and IA together, in the formation of personality patterns, also

showed some promise and clearly warrants further investigation.

The findings of the present investigation will be discussed

separately for each of the four mood variables--affective complexity,

overall mood level, overall mood variability, and mood variability in

the intimacy vs. academic work domains. Then a summary of findings and

their implications will be presented. First, however, the relationships

obtained among the four mood variables will be discussed.

Relationships Among Mood Variables

Among the four mood variables, only the correlation between

affective complexity and MVPC1 was statistically significant. A trend

was obtained between MLPC1 and MVPC1.

Relationship between Affective Complexity and MVPC1

Smaller affective complexity scores were found to be significantly

associated with smaller overall mood variability scores (r = 0.48,

p<'.003). Thus, greater affective complexity was significantly associated








with less overall mood variability. This finding provided support for

a trend reported by Wessman and Ricks (1965) for their female sample,

in which greater affective complexity was associated with lesser mood

variability.(r = 0.26, ns). The finding that greater affective com-

plexity is associated with lesser mood variability is consistent with

expectations derived from Psychological Differentiation Theory. From

this perspective, affective complexity, which provides a reflection of

the degree to which a person's moods covary together, could be construed

as an indicant of level of differentiation in the affective domain.

Theoretically, then, greater affective complexity would be associated

with greater specialization of affective functioning. Thus, a more

effectively complex person, with more articulated systems of affective

control, could respond relatively discretely to affective stimuli,

showing less overall mood variability. A less effectively complex

person would, perhaps, respond more globally to affective stimuli,

showing greater overall mood variability.

Relationship between MLPC1 and MVPCl

A non-significant trend (r = 0.25, p <.09) was obtained between MLPC1

scores and MVPC1 scores. This trend indicated that higher overall mood

level is related to greater overall mood variability. This finding must,

however, be interpreted quite guardedly, due to the relatively small

sample size, as well as the moderate magnitude of the correlation

coefficient.

However, to speculate, perhaps mood variability, rather than

singularly indicative of intra-psychic conflict, may indicate that one

is more fully "open to experience," and is, therefore, more able to

appreciate the richness of the world.








Relationships Among Other Mood Variables

No other relationships among the mood variables even approached

significance, providing support for the notion that these variables

may represent statistically and, perhaps, experientially, independent

dimensions of affective experience.

Findings Concerning Affective Complexity

Psychological differentiation was significantly associated with

affective complexity, accounting for 17 percent of the affective com-

plexity score variance. However, the direction of the relationship was

clearly in the opposite direction from that predicted. Thus, greater

PD appears to be associated with lesser affective complexity. This

finding was corroborated by the significant F statistic, indicating

that the two theoretically less differentiated groups recorded scores

reflecting greater affective complexity than the two more differentiated

groups. Further, this finding appeared relatively stable for each of

Groups 1-4.

The finding that greater PD is associated with lesser affective

complexity is not consistent with either expectations derived from

Psychological Differentiation Theory, or with findings reported in

previous investigations. It will be recalled that Frank (1967)

reported that greater PD was associated with greater affective complexity,

in his older female sample, although the magnitude of this relationship

was not large (r = -0.24, p .05). Gorman and Wessman (1974) reported

no relationship between their measures of PD and affective complexity

for either females or males.

Before discussing the theoretical implications of the present

findings, several methodological problems concerning the two earlier








investigations (Frank, 1967; Gorman and Wessman, 1974) will be discussed.

Based upon this discussion, it will be contended that the present

investigation provided a more adequate assessment of the relationship

between PD and affective complexity than the previous studies and,

therefore, that the present findings may, perhaps, demonstrate greater

validity.

Methodological Problems of Previous Investigations

Both Frank (1967) and Gorman and Wessman (1974) employed the standard

form of the PFS. In this standard form, the subject is instructed to

record her moods once daily--recording her personal "highest," "average,"

and "lowest" ratings for each of the 16 PFS. Next, the "highest,"
"average," and "lowest" ratings for each of the 16 PFS were inter-

correlated for each subject, resulting in a 48 x 48 correlation matrix

which was factored via P-technique. However, for Frank (1967) and

Gorman and Wessman (.1974), the correlation matrix which was factored

was based on, respectively, 30 and 20 observations for each variable.

This is a relatively small number of observations from which to obtain

a reliable correlation matrix for factoring. In fact, the number of

variables for these matrices exceeded the number of observations.

In the present investigation, each subject reported her actual

moods on each of the 16 PFS three times daily for 33+ successive days.

Thus, the 16 x 16 correlation matrix which was factored was based on

99 observations for each variable. This procedure was thought to provide

a more reliable correlation matrix for factoring than those obtained in

previous studies. Thus, subject's scores on the affective complexity

measure may be more reliable in the present investigation.








As indicated earlier, different investigators did not always employ

the same measure of affective complexity. This, however, may not be a

serious problem, as these measures appear to be formally and mathemati-

cally related. Indeed, in the present investigation, a relatively high

set of intercorrelations was obtained among all five initial measures

of affective complexity. A more serious problem concerns the restriction

in range that previous studies have reported for their affective com-

plexity measures.

Frank (1967) reported a restriction in range employing a measure

based on the number of factors each accounting for greater than 10

percent total variance. Ninety-six percent of his sample obtained

affective complexity scores of either two or three on this measure.

Frank's (1967) other affective complexity measure, the number of factors

necessary to account for 90 percent total variance, fared better. The

mean for this measure was 14.4, with a standard deviation of 3.0.

Gorman and Wessman (1974) employed four measures of affective

complexity, which were reviewed earlier. They reported that a restriction

in range was obtained for scores for each of these measures. Gorman and

Wessman (1974) reported descriptive statistics for two of these measures

which supported their contention. The means and standard deviations of

these two measures were: (1) the number of factors each accounting for

at least 10 percent total variance, mean = 1.89, standard deviation =

0.7, and (2) the number of factors necessary to account for 90 percent

total variance, mean = 12.5, standard deviation = 1.5. In the present

investigation, the measure of affective complexity, PCTVAR, did not

appear to show such a restriction in range.








Although one measure of affective complexity employed by Frank

(1967) did not show a restriction of range, the validity of this measure

as an operationalization of affective complexity may be suspect. This

measure, the number of factors (principal components) necessary to

account for 90 percent of the total variance, was derived from a

principal components analysis of each subject's correlation matrix.

These correlation matrices were based on 30 observations per variable.

Thus, the reliability of the initial correlation matrices may not have

been satisfactory. Further, principal components analysis may include

considerable error variance in the composition of factors, particularly

as the number of extracted factors increases. Generally, only the first

few principal components are retained and interpreted. To speculate,

since a mean of 14 components was obtained for each subject from a matrix

of 48 highly related variables, a considerable number of these components

may be primarily composed of error variance, with no systematic rela-

tionship to the complexity of the person's affective processes.

Another methodological problem concerns the instruments used to

assess PD in previous studies. The Hidden Figures Test and Hidden

Patterns Test employed by Gorman and Wessman (1974) may be relatively

indirectly related to PD. Frank's (1967) Human Figure Drawings Test

and EFT appear to provide a more reliable assessment of PD. In the

present investigation the RFT and EFT were employed and were thought

to provide a more adequate assessment of PD than either of the batteries

employed previously.

To summarize, three methodological problems in previous studies of

PD and affective complexity, which the present investigation attempted

to remedy, were discussed. These were: (1) the possible unreliability








of the correlation matrices from which affective complexity scores were

derived, (2) the restriction in range of affective complexity scores,

and (3) the employment of less satisfactory instruments to assess PD.

Further, the magnitude of the relationship obtained between PD and

affective complexity in the present investigation (17 percent shared

variance) is considerably greater than that obtained previously (5.7

percent shared variance) by Frank (1967). To summarize this pattern of

evidence, it appears reasonable to contend that previous investigations

may not have adequately assessed the relationship between PD and

affective complexity. Further, there appears to be considerable support

for the internal validity of the present investigation.

Affective Complexity and PD: A Reformulation

The finding that greater PD is associated with lesser affective

complexity appears contradictory to expectations derived from Psychological

Differentiation Theory. Psychological Differentiation Theory has provided

an impressive organizational and explanatory scheme for integrating an

extensive body of research concerning field dependence and the global-

articulated personality style. According to this theory, level of

differentiation, as assessed in the perceptual/cognitive domain,

accurately reflects the general level of differentiation throughout the

psychological processes of the person. Thus, field independent scores

on the RFT and EFT, indicative of greater perceptual/cognitive dif-

ferentiation, should also be associated with greater affective com-

plexity, or differentiation in the affective domain.

However, this formulation of Psychological Differentiation Theory

has been primarily based on studies employing college student samples

in academic settings, in which highly developed cognitive-analytic skills,








such as those associated with field independence, are quite generally

adaptive. Thus, over-generalizations could, conceivably have occurred,

from the implications of perceptual/cognitive differentiation in a

specific environment which strongly rewards these skills, to the

general functioning of persons through wider age ranges and in environ-

ments with different adaptive requirements. Such generalizability of

differentiation throughout various psychological domains may, indeed,

be valid, but it certainly has not been adequately tested--particularly

on non-student samples, in non-academic environments. Further, even

within the college student sample, the majority of studies of PD have

appeared to employ performance on perceptual/cognitive tasks as the

criterion variable. Very few studies have investigated the implications

of PD in interpersonal domains, for example. Thus, to summarize, it is

conceivable that the notion of the generalizability of PD through various

psychological domains has not been adequately tested.

Indeed, recent research has indicated that field dependent and

field independent persons may be more accurately construed as each

possessing their own particular adaptive skills, or--to speculate--as

each having their own domains of greater and lesser differentiation.

Thus, field dependent persons, who have been reported to possess a

constellation of characteristics which facilitate effective interpersonal

functioning (Witkin & Goodenough, 1976), may be more differentiated

regarding psychological domains related to interpersonal functioning

than field independent persons, who do not possess these interpersonal

skills. Conversely, field independent persons appear to be more

differentiated in perceptual/cognitive functioning than field dependent

persons. To speculate further, it could be hypothesized that field








dependence may be associated with a set of characteristics which

facilitates interpersonal intimacy and, therefore, affective dis-

closures with others. Such behaviors could, conceivably, result in

greater affective complexity.

To summarize, it appears that the notion that the level of

perceptual/cognitive differentiation accurately reflects the level of

differentiation through the totality of psychological processes of the

person has not been adequately tested. It is conceivable that over-

generalizations from studies concerning perceptual/cognitive functioning

of college student subjects in academic settings may have resulted in a

premature extension of the generalizability of perceptual/cognitive

differentiation. Few studies of the generalizability of perceptual/

cognitive differentiation through different psychological domains in

different samples, and in environments with differing adaptive require-

ments have been conducted. Further, recent evidence indicates that

field dependent and field independent persons might, perhaps, be

construed as differentially differentiated, rather than more or less

differentiated.

Thus, the finding that less differentiated females, according to

RFT and EFT performance, report greater affective complexity may,

indeed, indicate that these persons do possess greater affective com-

plexity than females indicating field independent performance. To

speculate, the Psychological Differentiation Theory itself, with its

notion of generalizability of perceptual/cognitive differentiation

through various domains of psychological functioning, may be

undifferentiated.








Findings Concerning Overall Mood Level

It was found that neither IA scores, nor PD scores, nor IA and

PD scores together, accounted for even 1 percent of the MLPC1 score

variance. Thus, IA and PD appear to be unrelated to overall mood

level. The hypothesis that IA would be related to overall mood level

was clearly not supported.

It is conceivable that selective subject attrition could account

for these findings. Subjects who were currently experiencing adjustment

difficulties would certainly be more likely to terminate participation

in the mood study. Also, subjects who were experiencing adjustment

difficulties may, perhaps, have been less likely to report their moods

as systematically as others, thus increasing the probability that they

would be excluded from the final sample. Indeed, to complete three mood

ratings daily for a period of 33 successive days at a 90 percent plus

level of accuracy in following instructions, could be considered a

reflection of a relatively high level of personal organization,

dedication, and flexibility. Thus, the demands of the mood rating task

may have systematically excluded the less effectively functioning

persons from inclusion in the final sample.

This speculation is supported by evidence reported by Frank (1967),

who interviewed all nine of his mood study dropouts. Of these nine,

six reported that acute emotional distress, exacerbated by the self-

reflective nature of the mood rating task, caused their termination in

the study. Further, it will be recalled that Wessman and Ricks (1965),

who reported a strong relationship between mood level and a battery of

measures related to adjustment, employed a male sample which was

systematically selected to provide a wide range of talent on several








assessment instruments related to adjustment. Indeed, one of these

selection instruments was a measure of the subject's general level of

"happiness-unhappiness."

In order to examine the tenability of this "selective attrition

hypothesis," the MLPC1 scores and EI-ISB scores obtained in the present

investigation were examined. First, a comparison of the MLPCl scores

for the 31 subjects in the final sample with the MLPC1 scores for the

18 subjects who were excluded from the final sample because their mood

rating accuracy was too "low" (between 75 percent and 89 percent) was

conducted. Based upon the combined 49 subject sample, the sample means,

in standard score form, for the 31 subject group and the 18 subject group

were, respectively, -0.0075 and 0.0129. These two sample means are

nearly identical. If the selective attrition hypothesis was valid, it

would be expected that the mean MLPC1 score for the 18 subject group

would be smaller than that of the 31 subject group. Thus, the selective

attrition hypothesis was not supported by an inspection of MLPC1 scores.

Systematic attrition did not appear to have occurred in relation to

overall mood level.

To further test the selective attrition hypothesis, a comparison

of EI-ISB scores was conducted for the 31 subjects in the final sample,

the 18 excluded subjects, and 15 subjects who had either not selected

participation in the mood study or who had terminated participation.

The EI-ISB sample means for these three groups were, respectively,

55.65, 54.18, and 54.78. These three group means are nearly identical.

Thus, it appears that subjects were not systematically selected in

relation to level of IA.








The selective attrition hypothesis was not supported by an

inspection of MLPCl and EI-ISB scores obtained in the present investi-

gation. It could, perhaps, be argued that the MLPC1 data for the 18

subject sample may be unreliable, since up to 25 percent of the

observations for some subjects were missing. Further, it could be

argued that the 49 subjects that the selective attrition hypothesis

was tested on represented only 68 percent of the original 72 subjects

who began the mood project. Thus, the selective attrition hypothesis

may not have been adequately tested for MLPC1. However, no effects due

to selective attrition were obtained with EI-ISB scores as well. To

summarize, there is little support for the selective attrition hypothesis

in the present sample.

An alternative explanation for the finding of no relationship

between IA and overall mood level is, simply, that IA and effectiveness

of functioning as reflected by adjustment are independent constructs,

dissimilarly related to overall mood level. This explanation will be

further developed in the next section, when the findings concerning

overall mood variability will be discussed.

Findings Concerning Overall Mood Variability

Contrary to predictions, PD was not significantly related to

MVPC1 scores. However, IA scores accounted for a significant proportion

(21 percent) of the MVPC1 score variance. Higher IA scores, theoretically

reflecting greater effectiveness of personality integration, were signi-

ficantly associated with less overall mood variability. This finding

was corroborated by a significant F statistic, indicating that the two

theoretically more effectively integrated groups recorded scores

indicative of smaller overall mood variability than the two less








effectively integrated groups. This finding appeared to be relatively

stable for each of Groups 1-4.

Theoretical Rationale for Findings

The finding of this strong relationship between IA and overall mood

variability, although not formally hypothesized, is readily interpretable

from an Eriksonian perspective. Theoretically, as proposed eariler, IA

level can be conceptualized as providing a reflection of the effective-

ness of integration of identity elements into an organized configuration.

Ego identity formation, then, involves a synthesis of identity elements

concerning vocational occupational, ideological, and other psychosocial

issues, into a "style of life." Thus, ego identity formation may result

in a relatively permanent commitment to a particular style of life.

Concurrent with the commitment to a particular style of life, is the

rejection of several alternative life styles. Theoretically, for the

identity achiever, who has selected and synthesized a particular life

style, the psychosocial issues concerning rejected life styles may no

longer be personally relevant and, therefore, may not be effectively

salient. A lesser level of IA may indicate that the person has not

yet integrated identity elements into a style of life or ego identity.

Thus, several alternative life styles, each with its own constellation

of psychosocial issues, may still be personally relevant and effectively

salient.

The higher IA person, perhaps reflecting a more stable, completed

personality system, may not be as effectively reactive to as wide a

range of psychosocial stimuli as the lower IA person. The lower IA

person, perhaps reflecting a more open, evolving personality system,








would be expected to be associated with less overall mood variability

than lesser IA.

This conceptualization of IA, as providing a reflection of the

effectiveness of integration of identity elements into a life-plan,

provides a possible explanation for the present findings concerning IA

and overall mood variability. Further, the present findings, as well

as this theoretical rationale, provide support for Wessman and Ricks'

(1965) finding that an "open vs. closed to experience" dimension was

related to mood variability. It will be recalled that Wessman and Ricks

(1965) characterized their less variable subjects as closed, completed

personality systems, while their more variable subjects were described

as open, evolving personality systems. Although Gorman and Wessman

(1974) reported that their "open vs. closed to experience" factor was

related to mood level, and not to mood variability, as noted earlier,

several of the marker variables for this factor may be construed as

measures of adjustment. Further, mood level has been traditionally

conceptualized as providing a reflection of adjustment. It is con-

ceivable, then, that Gorman and Wessman's (1974) "open vs. closed to

experience" factor may provide a measure of adjustment. Thus, to

summarize, it appears reasonable that IA, conceptualized as an indicant

of the effectiveness of integration of identity elements into a life-

style, may underlie the "open vs. closed to experience" dimension which

has been reported to be related to mood variability.

Effectiveness of Integration as a Higher Order Construct

However, this formulation of IA raises a theoretical problem. If

IA is, indeed, a valid indicant of the effectiveness of personality

integration, why was IA not related to overall mood level as predicted?








There is considerable evidence that supports the notion that mood level

provides a reflection of effectiveness of functioning or adjustment

(Wessman & Ricks, 1965). Further, effectiveness of functioning or

adjustment can be conceptualized as a reflection of effectiveness of

personality integration.

Effectiveness of personality integration, however, is a higher

order construct which can be employed to describe a formal relationship

among a set of elements in any system. Effectivenss of integration, then,

may be viewed as a formal construction, which derives much of its sub-

stantive meaning and implications from the domain or process to which it

is applied. Thus, effectiveness of integration can be conceptualized

and operationalized from multiple perspectives. The relationships

between effectiveness of integration and criterion variables, then,

may vary according to the domain or process engaged.

In the present investigation, effectiveness of integration was

conceptualized and operationalized as reflected by level of IA. In

retrospect, the implications of this operationalization may be different

from an operationalization of personality integration in terms of

adjustment. Identity achievement and adjustment may be related, but

still relatively independent constructs. Theoretically, IA concerns a

normative stage in ontogenetic development. Adjustment emphasizes a

more temporally (and structurally) stable characteristic of individual

functioning.

Indeed, according to Erikson (1968), the identity crisis and

adjustment refer to different, though certainly interrelated, processes.

The identity crisis refers to a normative period of growth, heightened

potential, and heightened conflict. Maladjustment refers to a relatively








stable, self-perpetuating process of ineffective functioning. In

maladjustment, there is not a surplus of energy available for growth,

but rather, a paucity of energy, due to the investment of energy in

defensive processes. Thus, although the processes of identity

formation and general effectiveness of functioning (adjustment) are

certainly related, these two processes are clearly not functionally

equivalent. Thus, theoretically, the nomological network between IA

and adjustment should provide evidence for both convergent and dis-

criminant validity. One direction for future research might be to

more clearly demarcate the convergent and discriminant relationships

that IA and adjustment possess with criterion variables, and, thus,

with one another. To summarize, this discussion of the relationship

between effectiveness of integration, IA, and adjustment must be

considered speculative, pending the outcome of further research.

Differential Implications of IA for Mood Variability

Another implication of the present findings is that there may be a

differential effect of IA for greater and lesser differentiated persons.

Inspection of mean MVPCl scores for Groups 1-4 indicated that Group 1

(high differentiation/more effective integration) showed the smallest

overall mood variability, while Group 4 (high differentiation/less

effective integration) showed the greatest variability. The less

differentiated groups, Groups 2 and 3, showed intermediate levels of

mood variability. To speculate, this may indicate that the process of

identity formation may be a more turbulent time for the more highly

differentiated female than for her less differentiated counterpart.

Perhaps, to speculate further, the less highly developed interpersonal

skills associated with greater differentiation, concomitant with highly




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs