' '1",
.l. .

f'.
.

35‘!

5...
.5:

 

fififimfiiapv

it?
i... .1
3th..

e ..:7
.i «I. J .
I II. 2‘. i
:3... [It
33:23

{LII}, I

{-91.1

In

2. .
1.. S? i.

n. 2303135..

i 6.13:2.-. .

.5... 1.55.; 13.1....

’I»

ksxrramnfsq u?
I a v3

4.5.72.1?! .
I. {itzzugnifi
5.4:)»: nary)
1r: . 7 . o
Ellyn... 5:

$1.. :1
53>.& ‘ n»

V . .2.
453.....137
, 1.....I‘talst
.x:v;1i.;.\.t. .
5‘. I. i

inhuman.

‘155? availz.
.I..V1.b.t:1
il::ll$)‘$£l¢t!lllxlai

:
.)

3.1.19’
is .uv...
iii:

.2; : ) c... I
nixxnxanstil (2:1
.- I. . 1’

 

an.
. v . a : f...
:25. I 1 $5.91...

.2.

:I ..

it 96‘: «‘19:...iaa’kft

Lat.) ‘
.SVO\§13A;I
IrivAQI;

.42).“.

i 1.2.x;
2d; . .1333»:
?11 .UI! VAL" )0

v!!!

. V 3‘»!
n. .n yhnvar-S.»

2.33%.]. an...
. I. It»?!

i« :5.
1! (ill. Iv}!!!

. 135:! I.

YR). 711’!!va “pl.
liliiaOPAvD sill!
at: 1..» 1.3";(3
n \bs...lx7$vv

VESIR ILTV LIIBRAR IES

WWilliilmlilililiWWII ill ll

"‘ 3 1293 014172

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

Depression and Anxiety:

Distinctions and Commonalities

presented by

Marilyn Bleiweiss Charles

has been accepted towards fulfillment I
of the requirements for

Doctoral degree in Clinical Psychology

 

Date 8/3/95
I r

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

Mm A_~—.__.-
.. n v.—

 

 

LIBRARY
Mlchtgan State
University

 

 

 

PLACE II RETURN BOX to remove We checked Irom your record.
To AVOID FINES return on or before dete due.

DATE DUE, DATE DUE DATE DUE

PRU; 92080 ‘ ‘
'1
1
l.

   

  

   

 

 

    
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DEPRESSION AND ANXIETY:
DISTINCTIONS AND COMMONALITIES

By

Marilyn Bleiweiss Charles

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Psychology

1995

ABSTRACT
DEPRESSION AND ANXIETY: DISTINCTIONS AND COMMONALITIES
By

Marilyn Bleiweiss Charles

Overlap of symptoms and high comorbidity rates have made it difficult to discriminate
anxiety and depression as discrete entities. Self-report instruments which broadly sampled
symptoms associated with these two disorders were administered to college students in an
attempt to determine whether there are distinct clusters of symptoms which can
discriminate between these two constructs. Factor analyses linked somatic symptoms of
motor tension and autonomic hyperarousal most specifically to anxiety, and fatigue and
hopelessness most specifically to depression, supporting the recent findings by Watson and
Clark (1995). Further analyses did not support models in which symptoms of anxiety and
depression are viewed as largely expressions of a single factor of “neuroticism” or
“negative afi‘ectivity;” the analyses pointed to clear and potentially meaningful distinctions
between symptoms, in spite of their high correlations. A content analysis of widely used
instruments for measuring anxiety and depression suggested that the anxiety scales under
consideration are more representative of the relevant syndrome than are the depression

scales. This may be due in part to the greater homogeneity in the construct of anxiety.

Dedication:

To my family, who have taught me to value the important things in life:

my parents, who taught me to value Integrity,
my sisters, who taught me of real Friendship,
Bruce, who taught me of Love with no limits,
Devon, who taught me the Joy of the interpersonal world,
Justin, who taught me Faith in the spiritual world,
and Jonathan, who taught me Courage: to believe in myself and those around me.

iii

ACKNOWLEDGEMENTS

I acknowledge with gratitude the efi‘orts of my committee members for their
assistance in completing what has been for me both an arduous and affirming process. As
I moved through my internship year, and began to make the transition from “student” to
“professional” in my clinical work, I was aware of parallels with the dissertation, which
felt stalled and unwieldy. Much as in a fairytale, I moved toward completing this process,
hitting barriers which had to be slain or hacked through in order to reach the castle; these
barriers were largely my own fears of what lay beyond. It brought to mind a fairytale I
learned from my mother about Diana, who was so swift that no one could catch her. She
was challenged to many races, until, finally, an admirer threw golden apples in her path.
As she stooped to catch them, and dragged them along in her skirts, she was slowed
down. At this point in time, I don’t recall whether she won the race; I know that she
finished. I have been thrown many “golden apples” in my life, and have found that the
heaviest and most difiicult to shoulder often contained the most enduring and precious
gifts. And so, I would like to express my deep appreciation toward the people who have
helped me to shoulder my burdens, and in that process, to believe in myself.

Susan Frank was the first professor at State who really invested her time and

energy toward facilitating my growth. Her remarkable intellectual and clinical acuity set a

iv

standard which both drew and intimidated me, and ultimately encouraged me to set my
own. Susan opened many doors for me both clinically and academically; she challenged
my limits and helped me to find and develop my own unique resources. Most importantly,
she provided a receptive and constructively critically ear, helping me to refine and
structure the ideas which motivated this dissertation. Her faith and friendship have meant
a great deal.

Jack Hunter has devoted himself intensively and without reservation to this
project. He has challenged me to clarify, refine, and articulate my thoughts and to then
stand by them. His investment and enthusiasm in this project, his intense intellectual
curiosity, and his passion for understanding have helped to revitalize this project and to
keep me engaged even in the face of discouragement and fi'ustration.

John Hurley has been a source of support and encouragement throughout my time
at State. His warmth, humour, and enthusiasm initially brought me here, and he has
always been ready with assistance: from theoretical issues, to editing, to moral support.
His unswerving belief in me has brought me through many of the hills and valley I have
encountered in this process.

Ellen Strommen deserves special acknowledgement for providing the impetus to
finally move this process to completion. She has helped to humanize my graduate

experience by speaking for, providing a forum for, and exemplifying the values of

relationship and care. Her warmth, humour, optimism, and common sense have helped to
turn many apparent mountains back into hills.

There are many other individuals who have supported my grth over the past six
years, and have helped me to believe in, embrace, and utilize my strengths and gifts.
Notable amongst these are Ed Gibeau, Ruth Rosenthal, and Gordon Williams; their
unswerving faith has helped me to find my own: a tremendous gilt.

Finally, I would like to acknowledge the patience, understanding, and support of
my family, who have struggled through this whole process with me. I appreciate their
willingness to accept all the challenges which this life has brought them: to work through

them, and to grow through them.

vi

TABLE OF CONTENTS

Page

LIST OF TABLES ............................................................................................................... ix
Chapter

I. INTRODUCTION ......................................................................................... 1

II. THEORETICAL BACKGROUND ............................................................... 3

Symptoms .......................................................................................... 3

Diathesis of Anxiety and Depressive Disorders ................................ 6

Physiological Studies ....................................................................... 13

“Learned Helplessness” and “Serotonin Driven Depression” ......... 17

Assessment ...................................................................................... 23

Common Concomitants .................................................................. 24

Anaclitic and Introjective Depression ............................................... 28

Difi‘erences Between Anxiety Symptoms .......................................... 40

Summary ......................................................................................... 41

III. PLAN OF ANALYSIS .............................................................................. 43

IV. METHOD ................................................................................................... 46

Participants ..................................................................................... 46

Instruments ..................................................................................... 46

Procedure ...................................................................................... 53

V. RESULTS .................................................................................................... 54

Hierarchical Structure of the Data Analysis .................................... 54

Items and Symptoms ....................................................................... 62

Symptoms and Syndromes .............................................................. 76

Results Pertaining to a One Factor Model ...................................... 97

vii

Chapter Page

VI. DISCUSSION ........................................................................................... 105
Negative Afi‘ectivity: A General Distress Factor? ......................... 106
Confirmation and Disconfirrnation of Claims ................................ 115
Current Anxiety and Depression Scales ........................................... 120
Symptoms of Anxiety .................................................................. 122
Heterogeneity in the Construct of Depression ............................. 123
Anxiety and Depression .............................................................. 124
Conclusions as to the Measurement of Anxiety and Depression... 127
APPENDIX A ....................................................................................................... 129
APPENDD( B ...................................................................................................... 144
REFERENCES ................................................................................................... 153

viii

Table

10.

11.

12.

LIST OF TABLES

Page
Rates of Anxiety and Depressive Disorders in Children and Relatives ..................... 9
Contingency Table for the Anaclitic and Introjective Scales ................................. 32
Proposed Proportions of Anaclitic and Introjective Depression in the
General Population per 100 People ...................................................................... 33
Proposed Proportions of Anaclitic and Introjective Depression in a
Patient Population per 100 Depressed People ...................................................... 34
Proposed Proportions of Anaclitic and Introjective Depression in an
Outpatient Population in Which There are 40% Depressives ................................. 35
The Correlation Between Anaclitic and Introjective Style Scales as
Reported in Various Studies ................................................................................. 36
Meta-Analysis of Correlations Between Anaclitic and Introjective Style
Scales .................................................................................................................... 37
Predicted Sums of Levels for Identification of Anaclitic and Introjective
Depressives ........................................................................................................... 39
Number of Items and Reliabilities for Symptoms Retained for Further
Analysis ................................................................................................................. 75
The Exploratory Factor Analysis of the Symptom Construct Correlation
Matrix ................................................................................................................... 78
Provisional Clusters Suggested by Exploratory Factor Analysis ........................... 83
The Confirrnatory Factor Loadings for Construct Correlation Matrix
Defined by 10 Symptom Clusters ......................................................................... 85

ix

Table Page

13. The Confirmatory Factor Loadings for the Symptoms Not Used to Define
Clusters in the Analysis of the Symptom Construct Correlation Matrix

Defined by 10 Symptom Clusters ............................................................................ 87
14. Correlations Between Syndromes ........................................................................ 89
15. Correlations Between Need for Others and Mixed Depression, Severe

Symptoms, and Alienation ...................................................................................... 9O
16. Symptom Construct Correlation Matrix (Corrected for Attenuation) ...................... 93
17. Symptom Scale Correlation Matrix (Not Corrected for Attenuation) ................... 95
18. Correlations Between Pure Anxiety, Hostility, and Pure Depression ..................... 97
19. Content Analysis of the Items Considered in the Watson and Colleagues

(1995b) Study ..................................................................................................... 108
20. Number of Items from Each Cluster Represented in Anxiety and Depression

Instruments ............................................................................................................ 121
21. Contingency Table Relating Pure Depression and Pure Anxiety ........................... 125
22. Correlation Matrix for all Scales Defined by Content Analysis ........................... 144

CHAPTER I
INTRODUCTION

Depression and anxiety are each complex syndromes, with diverse symptoms,
etiological factors, and treatments. This complexity has made it difficult to understand
and treat these disorders. There are two factors in particular which have made it difficult
to discriminate anxiety and depressive disorders as discrete entities. These factors have
complicated research on the etiology and treatment of these disorders. One factor has to
do with the high correlations reported between anxiety and depression. There are several
reasons for these high correlations. First, there are overlapping symptoms between the
two disorders. Second, because of overlapping symptoms, there are measurement
problems in distinguishing these two syndromes. There have also been arguments for
common etiological factors underlying these disorders (Merikangas, 1990). High
comorbidityrates have been reported for these two disorders (Kashani et. al., 1987a;
Kashani et al., 1987b; Weissman et al., 1987); researchers have found that as many as
87.5% of individuals diagnosed with major depression had also had an anxiety disorder at
some time (Weissman, Leckman, Merikangas, Gammon, & Prusofi‘, 1984). There is now
strong evidence for a theory that anxiety has a direct effect on depression (van Praag,
1994). In addition, some researchers have proposed that the high correlations are due to
a common factor, often called “neuroticism” or “negative affectivity” (Watson & Clark,

1984)

The other major factor which has made it difficult to understand and treat anxiety
and depressive disorders is their heterogeneity. There appear to be distinct subtypes
within each general category which have different implications for etiology and treatment
(van Praag et al., 1988; 1990b). The heterogeneity of depressive and anxiety disorders
suggest that it may be important to discriminate between different subtypes in order to
make sense of paradoxical and inconsistent empirical findings. Because of this
heterogeneity, it may be crucial to focus more specifically at the symptom level in order to
understand important distinctions and commonalities between the anxiety and depressive
disorders.

The current study examines whether there is a unique cluster of symptoms
describing depression that can be distinguished from a unique cluster of symptoms
depicting anxiety. One possibility is that there are in fact two distinct clusters. In that
case, one can then evaluate whether widely used self-report instruments represent a good
sampling of those symptom clusters uniquely associated with the relevant disorder. A
second possibility, suggested by many authors, is that there are more than one cluster
defined by symptoms of anxiety or depression, or by symptoms of both of these disorders.
In that event, one can evaluate whether those patterns correspond to predictions
associated with a number of theoretical views to be discussed regarding the heterogeneity

within, and overlap between, anxiety and depression.

CHAPTER II

THEORETICAL BACKGROUND

Many reasons have been suggested to account for the strong relationship between
anxiety and depressive disorders. First, a subset of the same symptoms are often used to
diagnose both of these disorders. AS a result, many of the instruments used to assess these
disorders lack discriminant validity (Gotlib & Cane, 1989). Second, there appear to be
common etiological factors, as suggested by family, genetic, and attachment studies.
Third, there is evidence of common underlying physiological substrates. Fourth, research
on “learned helplessness” has pointed the way to a theory of anxiety driven depression, in
which anxiety causes depression (van Praag, 1994). Fifth, assessment issues have
complicated attempts to understand these disorders. Finally, there are claims that both
anxiety and depression involve maladaptive cognitive processes (Clark & Beck, 1989;
Kendall & Ingram, 1989), or alternatively, a general distress factor, which is postulated as
the source of the high correlations (Watson et al, 1995a; 1995b)

Smptoms
Two unipolar depressive disorders are described in the Diagnostic and Statistical
Manual of Mental Disorders (DSM-III-R; American Psychiatric Association, 1987):
major depression and dysthymia. The essential features of major depression are depressed

mood and/or loss of interest or pleasure. Other symptoms include change in weight or

appetite, sleep disturbance, psychomotor retardation or agitation, fatigue, feelings of
worthlessness, inappropriate guilt, cognitive deficits such as diminished concentration or
indecisiveness, and suicidal ideation. Additional symptoms associated with dysthymia
include irritability, low self-esteem, and feelings of hopelessness (American Psychiatric
Association, 1987). One essential feature which distinguishes between major depression
and dysthymia is duration: in the DSM-III-R, dysthymia is defined as a chronic condition
with persistent or intermittent symptoms, of at least two years duration. However, in
some ways this last distinction may be somewhat artificial, in light of studies which show
high rates of dysthymic-major depressive comorbidity Akiskal & Weise, 1992). Reviews
of the relevant literature suggest that dysthymia is both a predisposing factor toward (i.e.,
a prodromal stage of) and a consequence of major depressive disorder (Akiskal, 1994;
Akiskal & Weise, 1992).

Several anxiety disorders are described in DSM-III-K including panic disorder,
phobias, obsessive compulsive disorder, and generalized anxiety disorder. The symptoms
associated with generalized anxiety disorder include unrealistic or excessive anxiety or
apprehension, such as fear of dying, going crazy, or of doing something uncontrolled. In
DSM-III-R, the remaining symptoms are grouped under three categories: motor tension,
autonomic hyperactivity, and vigilance and scanning. Motor tension includes such
symptoms as trembling, muscle tension, aches, restlessness, and fatigue. Symptoms of
autonomic arousal include shortness of breath, dizziness, palpitations, sweating, trouble

swallowing, abdominal distress, flushes or chills, and frequent urination. Symptoms

associated with vigilance and scanning include feeling keyed up or on edge, exaggerated
startle response, concentration difficulties, difiiculty falling or staying asleep, and
irritability. The individual may also experience feelings of depersonalization or
derealization. Panic disorder adds unpredictable attacks of panic to the preceding,
whereas phobic disorders are characterized by persistent and irrational fear associated with
avoidance of the dreaded object or situation.

As is evident from the previous descriptions, there is a great deal of overlap among
symptoms of depression and generalized anxiety disorder. Both anxiety and depression
may be accompanied by irritability, dysphoric mood, feelings of tension, apprehension,
worry, concentration difficulties, fatigue, and self-preoccupation (Gotlib & Crane, 1989;
Sarason, 1985; Spielberger, 1972). Torgerson (1985) found that irritability and anger are
more common in individuals with symptoms of both anxiety and depression than in either
group alone. Unfortunately, he did not distinguish between irritability and anger in his
study, and so these results may be confounded. Mook, van der Ploog, and Kleijn (1990)
found indications that the high correlations between anxiety, anger and depression may be
due largely to the high correlations between anxiety and anger, and anxiety and
depression, respectively.

There are also important differences between anxiety and depressive disorders.
Whereas sadness is most often reported by depressed individuals, fear appears to be the
predominant affective experience associated with anxiety (Bartlett & Izard, 1972; Izard,

Blumber, & Oyster, 1985; Ollendick & Yule, 1990). In a review of descriptive studies,

Breier, Chamey and Heninger (1985) found that, across studies, the symptoms which best
discriminated depression from anxiety were depressed mood, early morning awakening,
suicidal ideation, and psychomotor retardation. The symptoms which best discriminated
anxiety fiom depression were panic attacks, agoraphobia, and compulsive features (Breier
et al., 1985). Loss of interest or pleasure, and pessimism have also been noted to
discriminate between these disorders (Clark, 1989). These latter findings are consistent
with a recent review of studies with children and adolescents, in which Brady and Kendall
(1992) found that symptoms of depression, such as anhedonia and low self-esteem,
discriminated depression from anxiety. In contrast, symptoms of anxiety were less usefirl
in discriminating between groups (Brady & Kendall, 1992). This may be due, at least in
part, to overlap of items between scales in frequently used inventories for children.
DiaLesis of Anxietyand Depressive Disorders

Comorbidity

Clinical studies afiirm that the high correlations between anxiety and depression
are not merely due to problems with self -report. High comorbidity rates using diagnostic
criteria have been interpreted by some to imply a common diathesis for these disorders.
The diagnosis of depression is often a secondary diagnosis among patients with anxiety
disorders (Barlow, 1985, Kashani et al., 1987). In a cross-sectional study, researchers
found strong associations between anxiety and depression (Angst, Vollrath, Merikangas,
& Ernst, 1990). They found the strongest association between panic disorder and 1

depression: comorbidity rates were five times higher than would be expected by chance.

Longitudinal findings from the Angst and colleagues (1990) study are consistent
with other studies which have suggested that symptoms of anxiety are more likely to
precede manifestations of depression in individuals who develop symptoms of both
disorders (Schatzberg et al., 1990). Angst and his colleagues (1990) found that 62% of
those individuals who were eventually diagnosed with both disorders first manifested
symptoms of anxiety, compared to 18% who initially manifested symptoms of depression.
Pure depression tended to remain stable across the 7 year follow-up period, whereas
individuals initially diagnosed with only anxiety tended to manifest symptoms of
depression by follow-up. Although these differences were not Statistically significant, they
may still be meaningfirl; the small sample sizes in clinical studies tend to favor statistical
Type II errors. In the Angst and colleagues (1990) study, half of the purely anxious
subjects developed major depression or recurrent brief depression during the follow-up
period. This is consistent with other reports which have shown a greater tendency for
individuals with anxiety disorders to develop symptoms of depression than for depressed
individuals to develop symptoms of anxiety (I-Iagnell & Grasbeck, 1990).

Family Studies

Family studies suggest a common diathesis for anxiety and mood disorders. These
disorders tend to co-occur in families as well as in individuals. Panic disorder, but not
generalized anxiety disorder, appears to have a high specific family prevalence and genetic
transmission (Breier, Chamey, & Heninger, 1985; Maier, Buller, & Hallmayer, 1988).

However, there have been inconsistent reports in the literature as to the heritability and

etiology of mood and anxiety disorders. This may be due, in part, to the practice in some
family studies of employing diagnostic exclusion criteria. Studies which have employed
exclusion criteria have not supported a relationship between anxiety disorders and
depression (Crowe, Noyes, Pauls, & Slymen, 1983), whereas studies which have ignored
diagnostic exclusion criteria have supported a strong relationship between panic disorder
and major depression, in particular (Leckman, Weissman, Merikangas, Pauls, & Prusoff,
1983). When rates from previous studies which had employed exclusion criteria were
recalculated to include secondary depressive disorders, the relationship between anxiety
and depression was again supported (Clark, 1989).

In a review of family and genetic studies of parental depression and child
psychopathology, Weissman (1990) reported that children of depressed parents are at
increased risk for diagnosis of both mood and anxiety disorders. She found no statistically
significant differences in frequencies of depression and anxiety in children of parents who
had been diagnosed with a mood or anxiety disorder. No significant difi‘erences in
transmission of anxiety and depression have been found in children even when diagnosis
and comorbidity of parents was taken into account (W eissman, Leckman, Merikangas,
Gammon, & Prusoff, 1984). However, the small sample sizes in clinical studies often
make it difiicult to detect meaningful differences.

Studies which have examined the relatives of children with anxiety and depressive
disorders have found high rates of both depression and anxiety disorders. In contrast to

the previous studies, these studies do show specificity of transmission. For example, when

Puig-Antich and Rabinovich (1986) examined the rates of major depression in relatives of
proband children, they found that 39% of anxious children and 55% of depressed children
had relatives with major depression (See Table 1). In another study, looking at rates of
disorder in mothers of children diagnosed with anxiety disorders, 77.4% of the mothers
were diagnosed with anxiety disorders, whereas 42.2 % of the mothers were diagnosed

with major depression (Last, Francis, & Hersen, Kazdin, & Strauss, 1987).

Table 1

Rates of Anxiety and Depressive Disorders in Children and Relatives

 

 

 

 

CHILD ADULT DIAGNOSIS
DIAGNOSIS ANXH’ZTY DEPRESSION

ANXIETY 77% 39%
DEPRESSION 42% 5 5%

 

 

 

 

 

In conclusion, the evidence fi'om family studies shows a strong relationship
between depression and mixed anxiety and depressive disorders. There are conflicting
results regarding the specificity of the transmission of these disorders, which may best be
resolved by a thorough meta-analysis of the relevant studies. Generalized anxiety has been
linked to particularly high rates (70%) of secondary depression. These findings have been
taken to indicate that this disorder may be more highly associated with a “general distress”

factor than are more specific anxiety disorders (Dohrenwend, 1990; Noyes, Clarkson,

10

Crowe, Yates, & McChesney, 1987). However, there are also reports of links between
panic disorders and depressive symptoms, which would support a “spectrum” model of the
mood disorders, in which anxiety is included as one pole of mood dysregulation, with
depressed mood describing the opposite pole (Lopez-Ibor, 1990). “Kindling” models
suggest that over time the specificity or severity of stressors becomes less important in
producing symptoms (Gold, Goodwin, & Chrousos, 1988).
Twin studies

Twin studies also point to similarities in etiology, and provide a better opportunity
to differentiate between genetic and environmental factors implicated in the etiology of
anxiety and depressive disorders. Twin studies have suggested that differences between
individuals in measures of anxiety and depression can best be explained by differences in
genes and individual environmental experiences, rather than shared environmental
experiences (Jardine, Martin, & Henderson, 1984). Early studies suggested that there
were underlying nonspecific hereditary factors which may lead to a predisposition to both
anxiety and depressive disorders (Clifford, Hopper, Fulker, & Murray, 1984; Jardine,
Martin, & Henderson, 1984), as well as to neuroticism, more generally (Andrews,
Stewart, Allen, & Henderson, 1990). However, recent advances in behavior genetics
have provided new methods for testing some of these models. Carey and DiLalla (1994)
reanalyzed data fi'om Eaves, Eysenck, and Martin (1989) in an attempt to evaluate causal
models regarding neuroticisnr, anxiety, and depression. Their results did not support a

causal link from neuroticism to anxiety or depression, nor did it support the existence of a

11

common higher order factor linking these three constructs.

In a review of recent twin studies, Torgerson (1990) pointed out that, in addition
to the common genetic factors which increase susceptibility to symptoms of both anxiety
and depression, there are also genetic factors which are only linked to specific symptoms
of anxiety. For example, Martin, Jardine, Andrews, and Heath (1988) found that feelings
of panic appear to be shaped by genetic influences which do not affect other symptoms
associated with anxiety.

Aflchment Perspectives.

Research has shown a correlation between the quality of early interactions between
parent and child and later development (Murray & Trevarthen, 1985). These correlations
have been taken as causal; however, most of the research fails to control for genetics.
According to attachment theorists, the caretaker moderates the young child’s experience,
keeping the child from becoming overwhelmed by strong affect. Over time, the child
takes over more and more of these regulatory functions. Early interpersonal experiences
become the framework for understanding both self and other via internal representations
or ‘working models’ which guide expectations and actions (Bretherton, 1985). By the
end of the first year of life, the child has begun to develop complex working models of
human interaction (Kraemer, Ebert, Schmidt, & McKinney, 1991), which facilitate the
development of affective self-regulation and the modulation of impulses (Schwalbe, 1991).
These models of relationships are believed by many to have longstanding ramifications for

the quality of later experiences as well as the ability to moderate affect in later years, with

12

particular implications for anxiety and depression (Kobak & Sceery, 1988; Kobak, Sudler,
& Gamble, 1992).

Recent advances in neurobiological research suggest that adverse early experiences
not only impede the individual’s ability to moderate their experiences, they also have
structural implications for the developing organism; interactions which are responsive to
the child’s needs appear to facilitate the normal physiological development of the neural
structures which underlie affective modulation and well-being (Schore, 1994). Adverse
early experiences may impair the individual’s ability to self-regulate affective experiences,
leaving the individual more vulnerable to becoming overwhelmed by strong affects such as
sadness, shame, or fear. The ability to self-regulate affect efficiently allows the individual
to cope with stress with fewer costs (Schmale & Engel, 1975). The underlying structure
which appears to be critical in the development, storage, and regulation of internal
representations linked to the regulation of affective information appears to be
dopaminergic (Joseph, 1988; Schore, 1994). This is consistent with research which links
deficits in goal seeking behaviors (Swerdlow & Koob, 1987) and “positive emotionality”
(Depue, Luciana, Arbisi, Collins, & Leon, 1994) to the dopaminergic system (Swerdlow
& Koob, 1987), suggesting that this monoamine may be particularly important in
depressions in which anhedonia or psychomotor retardation is a major symptom (van
Praag, 1980b). Chrousos and Gold (1992) have delineated two distinct forms of stress
system dysregulation: one associated with hyperarousal, and the other taking the form of

hypoarousal. These two responses to acute stress are consistent with the vigilance and

13

arousal activation associated with anxiety, on the one hand, and the psychomotor
retardation, fatigue, and anhedonia linked to depression, on the other.
Physiological Studies

Evidence from epidemiological and animal studies biochemically links anxiety and
depression. Findings suggest that indices of noradrenergic and neuroendocrine function
may be disturbed in both disorders, albeit somewhat differently (Leckman, Weissman,
Merikangas, Pauls, & Prusoff, 1983; van Praag, 1994). Some variants of each disorder
respond to monoamine oxidase (MAO) inhibitors and tricyclic antidepressants (Uhde,
Roy-Byme, Vittone, Boulenger, & Post, 1985; Weissman, Leckman, Merikangas,
Gammon, & Prusoff, 1984). This is not true, however, of generalized anxiety disorder,
which, unlike obsessive compulsive disorder or panic disorder, does not respond to MAO
inhibitors or tricyclic antidepressants. There is also evidence which suggests that it may be
possible to distinguish between subtypes of depression on the basis of differential
neurophysiological underpinnings. For example, Weiss and Sirnson (1985) suggest that
anxiety which occurs in the face of uncontrollable stressors produces an accompanying
depressive state. This type of depression has been conceptualized as “anxious depression”
(Weiss & Simson, 1985) or “S-HT (serotonin)-related, anxiety-driven depression” (van
Praag, 1994).

It is more likely that specific symptoms or clusters of symptoms, rather than
complex syndromes such as anxiety or depression, will be linked to specific transmitter

systems. For example, van Praag and his colleagues (1990a; 1990b) have connected the

l4

initiation and maintenance of goal directed behaviors to the dopaminergic system, hedonic
firnctions associated with reward coupling to the norepinephrine system, and the affective
regulation of aggression and anxiety to the serotonergic system. Viewing symptoms in
this way helps to clarify why drugs which affect the serotonergic system have been found
to be useful in the treatment of both anxiety and depressive disorders, in both of which the
modulation of arousal may be an important component, and may be experienced either as
hostility, or anxiety. Van Praag (1994) later noted that serotonin related drugs work
poorly with some patients; Katz and his colleagues (1994) noted that about one third of
depressed patients do not respond to these drugs.

Clinical studies help point to ways in which symptoms of anxiety and depression
are tied together at the neurotransmitter level, but can be more clearly delineated when the
components are broken down into discrete symptom clusters. For example, Katz and his
colleagues (1994) found that biochemical changes in the serotonergic system were more
strongly linked to mood aspects of depression, such as hostility and anxiety, whereas
changes in the noradrenergic system were more strongly linked to behavioral aspects of
depression associated with psychomotor retardation and arousal, such as anxiety,
agitation, and somatic symptoms.

Katz and his colleagues (1994) found very different results when considering
unipolar versus bipolar responders to the intervention under study. In unipolar responders
lower levels of norepinephrine were associated with lower levels of hostility and smaller

decreases of norepinephrine were associated with reductions in psychomotor retardation.

15

Smaller decreases of a serotonin (S-hydroxyindoleacetic acid (5-HIAA); the major
metabolite of serotonin) and dopamine metabolite were associated with decreased
anxiety. Interestingly, smaller decreases in the norepinephrine metabolite were associated
with more positive outcomes, suggesting that it may be crucial to look at actual
neurotransmitter levels rather than increases or decreases per se. Bipolar responders
showed a somewhat different pattern: smaller decreases in norepinephrine were also
associated with decreased motor retardation. However, lower levels of 5-HIAA were
associated with decreased depressed mood, and greater levels of the dopamine metabolite
were associated with decreased hostility (Katz et. al., 1994). These results support other
studies which have affirmed the importance of considering subtype when looking at mood
disorders.

These findings also support earlier suggestions (Katz et. al., 1987) of a stronger
link between the serotonergic system and anxiety than with dysphoric mood. There is still
controversy whether decreases in 5-HIAA in the cerebrospinal fluid is a reflection of
enhanced serotonergic transmission (Ericksson & Humble, 1990; Meltzer, 1990). If this
link does exist, it would make sense of differences between unipolars and bipolars in the
therapeutic action of serotonergic drugs which have been reported (Katz et al., 1994).
For bipolars, greater reductions in a S-HIAA were associated with reductions in anxiety
and depressed mood, whereas for unipolars, smaller reductions in S-HIAA were
associated with reductions in anxiety (Katz et al., 1994). This suggests that the role of the

serotonergic system is different in unipolar depression, or else that it is not relevant to the

16

therapeutic action of the drug. In this context, it is notable that anxiety appears to play a
larger role in unipolar than in bipolar depression (Katz, Robins, Croughan, Secunda, &
Swann, 1982). Serotonergic drugs tend to be more effective in the treatment of disorders
in which anxiety plays a major role, than in those in which psychomotor disturbance is
prevalent (Deakin, Guimaraes, Wang, & Hensman, 1991;1nsel, 1991). These findings
affirm the importance of focusing on symptom clusters when trying to understand
important differences between anxiety and depression.

Drug trials support the importance of looking at specific subtypes or symptom
clusters of anxiety. Consistent with results from family and genetic studies, drug
treatment studies support panic disorder, obsessive compulsive disorder, and generalized
anxiety disorder as separate nosological entities (Heninger & Chamey, 1988).

The literature exploring the concomitants of learned helplessness extends our
understanding of some of the physiological underpinnings of anxiety and depressive
symptomatology by looking at the effects of chronic or uncontrollable stress. The
behavioral deficits associated with learned helplessness include the anhedonia,
helplessness, and despair often associated with depressive disorders (Maier & Seligrnan,
1976; Schutz, Schutz, Orsingher, & Izquierdo, 1979). Severe stress leads to monoamine
dysregulation which includes the dopaminergic and noradrenergic systems.

In an attempt to better delineate the functions of these two monoamines,
Dubovsky (1993) linked underlying neurological substrates to specific psychobiological

firnctions, rather than to discrete diagnostic syndromes, and found that the noradrenergic

17

system is most often associated with arousal, orientation to danger, alerting, learning,
memory, and sympathetic nervous system functioning. Symptoms associated with
noradrenergic dysregulation include agitation, arousal, fearfirlness, vigilance, insomnia,
and withdrawal (Dubovsky, 1993). Dubovsky links the dopaminergic system to
movement, reward, and motivation. Symptoms associated with dopanrinergic
dysregulation include the psychomotor retardation, anhedonia, helplessness, and despair
often associated with depression (Swerdlow & Koob, 1987). Serotonergic dysfirnction
has been linked to anxiety, aggression, motivation, memory, skeletal muscle function, as
well as to regulatory firnctions such as mood, sleep, appetite, body temperature, and
sexual behaviors (Dubovsky, 1993; Cloninger, 1986; van Praag et al., 1990a) . Associated
symptomatology includes impulsivity, aggression, suicidality, sadness, anxiety, as well as
sleep and appetite disturbances (Apter et al., 1990; Dubovsky, 1993; Soubrie, 1986; van
Praag, 1990b).

“Learned Helplessness” and “Serotonin Driven Degemn”

The best defined distinct subtype of unipolar depression is a type called “serotonin
driven depression” by van Praag (1994). This form of depression was originally
hypothesized from results on experiments called “learned helplessness” experiments. This
section reviews that research, which strongly links anxiety and depression. The reader is
warned that the label “learned helplessness” is now known to be a very misleading label

for “learned helplessness” experiments.

18

The Label “Learned Helplessness”

The original “learned helplessness” experiments took on the following form.
Animals were exposed to uncontrolled and inescapable stress. Following this stress, the
animals showed performance deficits in escape learning. For example, consider a situation
in which untreated animals easily learn a response to escape shock. Following stress,
many of these animals were unable to learn the escape response, or they were very slow in
learning it. That is, they acted “helpless” in the face of shock. Researchers early on noted
that this is similar to the pattern shown by severely depressed patients (Maier & Seligman,
1976; Schutz, Schutz, Orsingher, & Izquierdo, 1979). The “learned helplessness”
experiments have subsequently been used as animal models for depression, although there
have also been human studies done within the “learned helplessness” paradigm.

The label “learned helplessness” was coined by certain theorists who had formed a
similar theory for depression disorders. This theory claims that depression is caused by
feelings of helplessness. These feelings are claimed to be produced by a belief in the lack
of contingency between one’s actions and potential outcomes (Gerber, Miller, & Seaman,
1979). Others (Abramson, Metalsky, & Alloy, 1989) have further claimed that feelings of
helplessness are caused by self-blame or “internal” attributions for negative events. These
theorists interpret the findings of the “learned helplessness” experiments as showing that
people learn helplessness from self-blame attributions for stress.

The problem with the label “learned helplessness” is that it is got learned (c.f. Paul,

 

1988, p. 15 ). There are several forms of evidence which show this, including findings on A

l9

duration and findings on the induction of learned helplessness by drugs rather than by
stress (Weiss & Simson, 1985).

Consider duration: Learning without counterlearning lasts for periods of days,
months, and years. The behavior deficits for “learned helplessness” experiments wear off
completely in 72 hours and are largely gone in 24 hours. Furthermore, Weiss and his
colleagues (1981) have noted that effects past 4 hours are probably due to conditioned
anxiety which reproduces the anxiety state, which then reproduces the depressed state.

Consider newer studies on the induction of “learned helplessness” behavior deficits
using other drugs: Petty, Kramer, and Moeller (1994) reviewed studies showing that
“learned helplessness” can be induced by injections of anxiogenic drugs, such as
haloperidol. It can be induced by other anxiety producing drugs, as well, such as
benzodiazapine receptor ligands (Drugan, Maier, Skolnick, Paul, & Crawley, 1985). That
is, there need be no stress manipulation to get the effects of “learned helplessness.”
“Learned Helpless_ness” finding

The key finding of the “learned helplessness” studies can be restated as this: A
state of high anxiety produces a delayed state of high depression, in which the subject feels
dysphoria, irritability, and anhedonia. The anhedonia eliminates the emotional effects of
reinforcement (Ettenberg, 1989) and thus makes it hard to learn from successfirl
experiences, such as the relief from fear produced by a successfirl escape experience.
Anhedonia also makes it hard to elicit previously learned responses that were rewarded by

positive reinforcement.

20

There have also been many studies done on the brain chemistry of the depression
produced in “learned helplessness” experiments. These Studies suggest that the serotonin
effects produced by anxiety result in depleted norepinephrine in the locus ceruleus area of
the brain (Paul, 1988), and that it is this depletion which produces the anhedonia of the
depressed state.

Anxiety driven depression

The data in “learned helplessness” studies show that a state of high anxiety will
produce a state of high depression (Barlow, 1991). That in itself says nothing about trait
anxiety and trait depression. But consider the implications of this finding for people with
high state anxiety. A person with high state anxiety often experiences states of intense
anxiety. Each such experience produces a state of depression which lasts about 4 hours.
Thus, a person who experiences a high rate of intense state anxiety will automatically
experience a high rate of associated depression.

Van Praag (19803) made this inference at an early date in this line of research and
concluded fiom this that there should be a subtype of depression caused by serotonin
dysregulation. In later reports, he thought that data had disconfirmed this hypothesis (van
Praag et al., 1990a). However, in van Praag (1994), he pulled together all the data and
showed strong support for this hypothesis. Depressed patients who show low levels of
serotonin metabolites in their cerebral spinal fluid are much more likely to respond
positively to antidepressants than patients whose first presenting symptoms are of anxiety.

Those whose first and only symptoms are those of depression usually do not respond to

 

811‘

581'!

asp

WOU

21

tricyclic antidepressants. Furthermore, there is a temporal progression to the recovery.
They first show a reduction in anxiety and then later Show a reduction in depression.

The one key difference between the van Praag (1994) theory and the simple
anxiety-depression hypothesis is that van Praag noted the importance of hostility and
aggression to this theory. Serotonin has an even stronger influence on hostility and
aggression than on anxiety. Van Praag noted that in the studies done, hostility was equally
a part of the pattern. That is, in the patients who appear to be serotonin driven
depressives, hostility was an early presenting factor just as often as anxiety. In addition,
hostility - like anxiety - showed early reduction in the patients who responded positively to
tricyclics.

The van Praag theory is closely related to the epidemiological findings of an
asymmetry in the development of anxiety and depressive disorders. That is, those who
first develop anxiety fi'equently go on to develop either firll or partial depressive disorders.
Those who first develop depression usually do ao_t go on to develop problems with
anxiety. Those who develop depression following anxiety disorders would be the
serotonin driven depressives. Those who develop depression first and who do not develop
anxiety problems are a second type of depressive: those not likely to respond to
serotonin-relatedantidepressant drugs.

It would be interesting to know if hostility/aggression problems show the same
asymmetry in time for aggression as that found for anxiety. The van Praag (1994) theory

would predict this. However, the epidemiological studies rarely report on hostility and

22

aggression problems.
Quantitative Theories of Anxietyaand Depression

Quantitative theories of state and trait depression consider comparisons rather than
extreme cases. Consider then two people, one of whom is higher than the other in trait
anxiety. The person who is higher in trait anxiety will be more likely to experience states
of intense anxiety which will produce the accompanying state of depression. Thus, the
person who is higher in trait anxiety will also be higher in trait depression.

If all depression were caused by serotonin dysregulation, then there would be a
near perfect correlation between trait anxiety and trait depression, even though the two
are conceptually distinct entities. However, the studies reviewed by van Praag (1994) also
show that there are depressed patients who do not respond well to tricyclics and who
show a very different temporal pattern in the development of depression.

The implication for the joint relationship of anxiety and depression is this: The data
should Show that as anxiety goes up, depression goes up. However, as anxiety goes
down, depression need not disappear. Rather, those whose depression is not caused by
serotonin dysregulation could Show up as people with low anxiety, but high depression.
Thus, the contingency table for anxiety and depression would be predicted to show an
asymmetry: (a) no cases of high anxiety without high depression, but (b) many cases of

high depression without high anxiety.

23

Assessment

An underlying dilemma in defining depression and anxiety has been the multiple
meanings of these terms, ranging from affective experiences, to syndromes, to disorders.
Many theorists now view each of these disorders along a continuum extending from
common affective experiences to diagnosable disorders (Beck & Clark, 1988). Much of
the clinical, family, and psychophysiological literature supports this view (Paul, 1988; van
Praag et al., 1988). Quantitative studies find no break in the distribution of anxiety and no
break in the distribution of depression.

Self-report instruments.

Most anxiety self-report instruments describe features of generalized anxiety
disorder (Gotlib & Cane, 1989). Most depression instruments describe symptoms
associated with unipolar depression (Gotlib & Cane, 1989). Many anxiety and depression
scales correlate between .70 and .90 with other scales measuring the same construct,
which would provide good evidence for convergent validity. Correlations between
constructs of between .40 and .60 (with a range between .27 and .94; Clark, Beck, &
Stewart, 1990) have been reported. Gotlib and Cane (1989) have interpreted the high
correlation between anxiety and depression in self-report instruments as a lack of validity
in self-report. As mentioned previously, there is considerable overlap between symptoms.

If the high correlation were due to overlapping symptoms in the scales, that high
correlation might be interpreted as poor discriminative validity. However, there is also

high comorbidity between these two disorders. Thus, it is not clear whether the high

24

correlation between anxiety and depressive instruments is due to poor discriminant validity
or to the high comorbidity rates for these disorders (Regier, Burke, & Burke, 1990).

Mountjoy and Roth (1982) found that depression was more often described as
persistent in depressive patients, whereas it was more often described as mild or episodic
in patients with a primary diagnosis of anxiety.

There were no significant differences between anxious and depressed patients in
their self-reports of tension. Some people have accused depressives of exaggerating
reports of level of symptomatology and general distress (Prusoff & Klerrnan, 1974). Kelly
and Walter (1969) found that agitated depressives rated their own anxiety as more severe
than did anxious patients whose autonomic arousal levels were higher. Even non-agitated
depressed patients rated their anxiety as high as did anxious patients, in spite of
physiological evidence to the contrary. However, it should be noted that autonomic
arousal is not the same as fear and vigilance; a depressed individual’s experience of
ruminative anxiety may indeed be more severe to that individual even though there may be
a lower level of autonomic arousal.

Common Concomitants
Personality Factors

Another method for understanding the comorbidity of anxiety and depressive
disorders is by looking at commonalities and differences in personality characteristics.
Cloninger (1987; Cloninger, Martin, Guze, & Clayton, 1990) has delineated three

dimensions of personality: novelty seeking, described as impulsive versus constrained;

ha
rer
ass
ho:
19?

de;

per

due
rnor
fion
diso

Cleve

Ofier
<fifltt
beha
aSan
dePre

to int:

25

harm avoidance, described as apprehensive or cautious versus fearless or uninhibited; and
reward dependence, described as sensitive to social cues versus detached. Strong
associations have been found between harm avoidance and negative mood states, such as
hostility, anxiety, depression, fatigue, and confirsion (Svrakic, Przybeck, & Cloninger,
1992). Cloninger (1986) contends that individuals high in harm avoidance and reward
dependence, and also low on novelty seeking, are more likely to develop either anxiety or
depressive disorders (1986; 1988a). Cloninger (1986) suggests that similarities in
personality types account for the high comorbidity of these two disorders (1986; 1988a).

There is evidence that the overlap between these two disorders may be primarily
due to secondary features; depression is often a sequelae of chronic psychopathology
more generally (Cloninger, Martin, Guze, & Clayton, 1990). There are some indications
from family studies that there may be little overlap between primary anxiety and depressive
disorders, but rather, individuals with anxiety disorders may be at higher risk for
developing secondary depression (Cloninger, Martin, Guze, & Clayton, 1981).

Some theorists have suggested that there is a general subjective distress factor,
often termed "neuroticism," which has been defined as ”a broad dimension of individual
differences in the tendency to experience negative, distressing emotions and to possess
behavioral and cognitive traits" (Costa & McCrae, 1987, p. 301). Neuroticism is viewed
as an enduring emotional instability, and correlates .59 with anxiety and .51 with
depression in trait measures (Eysenck & Eysenck, 1968). These correlations are difiicult

to interpret. There is no independent way to measure neuroticism; the scales which

26

purport to do so are largely saturated with anxiety and depression items, along with other
associated symptoms, such as hostility.

Recovered anxious and depressed patients have been found to be more neurotic
than control subjects (Reich, Noyes, Hirschfield, Coryell, & O'Gorman, 1987).
“Neuroticism” is also a strong predictor of chronicity (Hirschfield, Klerman, Andreasen,
Clayton, & Keller, 1986). In a high-risk study, Hirschfield and his colleagues (1989)
found that factors associated with neuroticism, such as decreased emotional stability and
poorer accommodation to stressfirl situations predicted later depressive episodes.
However, the high recidivism rates for depression suggest that these individuals may be
moving from major depressive episodes to dysthymia; “recovery” may mean alleviation of
acute symptoms (Akiskal & Weise, 1992). Not surprisingly, neuroticism, much like
anxiety and depression, appears to be largely genetically determined (Henderson, 1982),
and genetic factors also appear to be important in predicting the covariation of
neuroticism, depression, and anxiety (Jardine, Martin, & Henderson, 1984; Martin,
Jardine, Andrews, & Heath, 1988).

More recently, this purported tendency to experience negative emotional states has
been conceptualized as a pervasive mood disposition, and has been termed "negative
affectivity" (Watson & Clark, 1984). Negative affectivity is measured by symptoms of
negative affective states including nervousness, tension, anger, guilt, and sadness.
Individuals high in negative affectivity are those who have a heightened sensitivity to life

stressors and a negative self-image (Watson & Clark, 1984). Watson and Clark (1984)

 

h.

 

27

claim that it is positive affectivity, or the disposition to experience positive affective states,
which may best discriminate between anxiety and depression. In contrast to Watson and
Clark, whose conceptualization of both depression and anxiety includes high negative
affectivity, Tellegen (1985) has conceptualized only anxiety as high negativity, positing
depression as a deficit in positive affectivity. In this way, Tellegen is viewing depression
as anhedonia, one predominant symptom of the disorder. Tellegen's view is consistent
with indications from the literature that mild depression is more strongly defined by a lack
of positive self-evaluations than with a predominance of negative self-evaluations (Beck &
Clark, 1991).
Maladaptive Cognitive Processes

Empirical findings link personality characteristics associated with negative
temperament to poor outcomes, such as lower self-esteem (Block & Robins, 1993 ),
anxiety, and depression (Watson & Clark, 1984). This may be due, at least in part, to
overlap between items in these scales. There are indications that cognitive states can have
important implications for affective experiences. For example, self-focus can intensify
affective states (Gibbons, 1991). Depression includes a tendency to perseverate on
negative events, increasing the experience of negative, but not positive, affect.
Alternatively, there are also indications that the absence of positive focus may be more
focal than the presence of a negative focus (Beck & Clark, 1991). The rumination and
negative expectations often associated with depression may replace active engagement and

constructive problem solving. Cognitive models also provide a means for understanding

28

how anxiety may lead to depression; Carver and Scheier (1991) suggest that when
feelings of anxiety disrupt attention, positive versus negative expectations determine
whether or not the activity is resumed. Within this framework, a paucity of favorable
expectancies leads to disengagement.

Both anxiety and depression are characterized by maladaptive cognitive processes
and irrational beliefs. However, there is evidence that the cognitive content of
maladaptive cognitions may differ between depressed and anxious individuals (Beck,
Brown, Eidelson, Steer, & Riskind, 1987). In anxiety, maladaptive cognitions tend to
have an orientation toward a firture in which threat or harm is anticipated, whereas in
depression the focus tends to be toward past perceived losses, failures, or degradation.
These differences in orientation also appear to have implications for how individuals
process information. The anxious individual may be hypervigilant for signs of threat or
harm, overestimate the possibility of danger, and tend to recall information which is
associated with threat or anxiety. In contrast, the depressive may tend to differentially
process and recall information associated with loss and failure. Depressed individuals also
tend to discount positive self-referential information (Kendall & Ingram, 1989). These
types of irrational beliefs and maladaptive cognitions may also be important in helping us
to understand and distinguish between distinct subtypes of depression.

Anaclitic and Introjective Depression
In his original and theoretical work, Blatt (1974) claimed that depressed people

can be split into two types: anaclitic and introjective depressives. However, when he

29

moved into research, there was a subtle shift in theories: a shift from types to “styles.”
This shift obscured the problems for the original theory. The data seem to contradict the
original theory.

This section will present the original theory. The empirical work will be briefly
reviewed with an emphasis on the two scales for anaclitic and introjective depression.
Selected findings for the scales will be reviewed. This evidence seems to disconfirrn the
original type theory. The evidence may disconfirrn the use of the term “depression” in
naming the scales. Finally, some hypotheses for this research will be framed.

B_latt’s Theory: Subtypes of Depresm

Blatt (1974) hypothesized that there are two types of depressives: anaclitic and
introjective depressives. His theory is developmental. Depression can be produced fiom
either of two kinds of experience: (a) helplessness and dependency or (b) feelings of
inferiority and resulting self-criticism. He posited two types of depressive. First, the
anaclitic depressive is characterized by feelings of helplessness, weakness, and
dependency. Second, the introjective depressive, presumed to be developmentally more
advanced, is characterized by feelings of inferiority, guilt, and fears relating to perceived
failure to live up to standards or expectations. Many authors have focused on this
dichotomy in their attempts to understand depression (Arieti & Bemporad, 1980; Blatt &

Zuroff, 1992; Grinker, Miller, Sabshin, Nunn, & Nunnally, 1961; Pilkonis, 1988).

3.

ar

111

de

all

3O

BlatL’a Reaearch: Styles

Blatt devised an instrument using items which were constructed to reflect
experiences often reported by depressed individuals, and were not specifically associated
with any particular theoretical orientation (Blatt, D’Afilitti, & Quinlan, 1976a). A factor
analysis of these items results in three factors. The anaclitic, or dependent, style was
described as externally directed with regard to interpersonal relations. The themes
included abandonment, loneliness, helplessness, rejection, dependency, and managing
negative affect to avoid the loss of the other. The introjective, or self-critical style, was
described as more internally directed. The themes included guilt, emptiness, hopelessness,
insecurity, dissatisfaction with self, self-blame, failure to meet standards or expectations,
and being threatened by change or responsibility. The third factor was identified as
“efficacy,” with no recognition of the fact that self-efficacy and self-criticism should be
negatively correlated with each other. This third factor consisted of items which indicated
a sense of confidence regarding one’s inner resources and capacities. The themes included
a focus on high standards, personal responsibility, a sense of autonomy, and pride in one’s
accomplishments.

The key output fi'om this research effort is two scales: one scale to measure
anaclitic depression and the second scale to measure introjective depression. However,
there is a key element missing from that research: no cutoff scores to identify the two
depressive types. It may be that the researchers looked at the distributions of scale scores

and saw no natural break in that distribution. Thus, they may have been hesitant to choose

31

such a point. However, there is also no natural break in the distribution of scores on
depression scales either: no break in scores for self-report scales such as the Beck
Depression Inventory and no break in scores for psychiatric interview scales such as the
Hamilton Depression Scale. Nonetheless, usefirl cutoff scores have been developed for
research and practice in regard to depression.

In discussing research, Blatt and his colleagues frequently make a subtle, but
important, shift in language: a shift from the language of types to the heavy use of the
word “styles.” This seems to be an indirect recognition of a shift from considering types
to the consideration of quantitative dimensions. However, when this shift is done
implicitly, there can be many errors in logic. Type theories are very different in content
from dimension theories.

The Correlation Between Anaclitigarnd Introjective Depressioa

Perhaps the most important finding for the Anaclitic and Introjective depression
scales is the fact that the two scales are positively correlated. This would seem to
contradict the original theory of two types of depression.

In the original theory, there are three types of people:

AD= anaclitic depressives

ID = introjective depressives
ND = non-depressives.

32

Consider the measurement theory behind the development of the Anaclitic depression
scale:
Anaclitic depression scale ranges:

High: AD persons
Low: ID and ND persons

Consider the measurement theory behind the development of the Introjective depression
scale:
Introjective depression scale ranges:

High: ID persons
Low: AD and ND persons

In a contingency table for the two scales considered together, we should find the entries

depicted in Table 2.

Table 2

Contingency Table for the Anaclitic and Introjective Scales

 

Introjective Depression i

 

 

 

Anaclitic Low High
D es '0
epr sr n High AD
Low ND ID

 

 

 

 

 

 

The number of people in each cell would depend on the severity of depression used

to define the depressed type and it would depend on the population studied. In clinical

33

terms, the level of severity may be defined by whether “minor” depression should be
considered or only major depression (“minor” depression is n_ot minor in terms of either
impairment or prognosis as noted by Akiskal and Weise, 1992). Consider the general
population. If 10% of the population were considered as suffering from serious
depression, and if the two types of depression were equally likely, then the table for 100

people would look like the contingency table presented in Table 3.

Table 3

Proposed Proportions of Anaclitic and Introjective Depression in the General Population
per 100 People

 

Introjective Depression

 

 

 

Low High
Anaclitic ,
Depression H1311 5
Low 90 5

 

 

 

 

 

 

Correlation: r = -.05.

That is, in a general population, we would expect a low negative correlation between the
scales of about 1; = -.05.
On the other hand, consider a patient population. If we considered only depressed

people to begin with, the contingency table for 100 people would be as shown in Table 4.

be

qu
5U;
5111

Th

81(

34

Table 4

Proposed Proportions of Anaclitic and Introjective Depression in a Patient Population
per 100 Depressed People

 

 

 

 

Introjective Depression
Low High
Anaclitic ,
Depression H‘gh 50
Low .. 50

 

 

 

 

 

 

Correlation: r= -1.00

In a population of depressed patients, we would expect a perfect negative correlation
between the scales of r = -l .00.

Many studies are done with less well defined “outpatient” samples. The key
question is: how many depressives are in the particular outpatient sample? For example,
suppose that in the outpatient population, 40% of the patients are depressives. That is,
suppose that in the outpatient population, 83% of the non-depressives are screened out.
The contingency table would look like Table 5.

That is, in an outpatient population with 40% depressed patients, we would expect

a low negative correlation between the scales of r = -.25.

35

Table 5

Proposed Proportions of Anaclitic and Introjective Depression in an Outpatient
Population in Which There are 40% Depressives

 

 

 

 

Introjective Depression
Low High
Anaclitic .
Depression H‘gh 20
Low 60 20

 

 

 

 

 

 

Correlation: r= -.25

To summarize; in the general population, anaclitic and introjective depression
should have a low negative correlation of about r = -.05. In a population with only
depressed patients, that negative correlation would be a perfect .12 = -1.00. In an
outpatient population, the correlation would be somewhere between -.05 and -1.00
depending on the percentage of outpatients that are depressed. For an outpatient
population with 40% depressed patients, the correlation should be about [ = -.25.

Eight studies were located with data on the correlation between the Anaclitic and
Introjective style scales. In some studies, the sample was an outpatient sample. In some
studies, the sample was a student sample that would be similar to the general population.
The Blatt, Quinlan, Chevron, McDonald, & Zuroff (1982) study had both outpatients and
students. Those two values are listed separately and will be analyzed as if they were from
separate studies. Unfortunately, one of the studies did not fiilly report the correlations.

Zuroff, Moskowitz, Wielgus, Powers, and Franko (1983) computed the correlation for

36

three samples, but report the value for only one sample (N = 39; r = .30). The other two
correlations were merely described as “less than .15,” so that study could not be used.
The studies and the results that could be used are listed in Table 6.

The values in this table vary considerably from one study to another. This is
expected in small sample studies because of the large sampling error in such studies.
Meta-analysis was used to obtain estimates of the distribution of population values.

Several analyses were done and the results are presented in Table 7.

Table 6

The Correlation Between Anaclitic and Introjective Style Scales
as Reported in Various Studies

 

 

 

 

 

 

 

 

 

 

Authors Sample Size Correlation Subjects
Brown & Silberschatz, 1989 60 .51 outpatients
Klein, Harding, Taylor, & Dickstcin, 1988 132 .27 outpatients
Blatt, Quinlan, Chevron, McDonald, & Zuroff, 1982 I97 .30 outpatients
262 .10 students
Smith, O’Keeffe, & Jenkins, 1988 188 .04 students
Fuhr & Shean, 1992 150 .47 students
Shapiro, 1988 l l 1 .26 students
McCranie & Bass, 1984 86 -. l 3 students

 

 

 

 

 

37

Table 7

Meta-Analysis of Correlations Between Anaclitic and Introjective Style Scales

 

 

 

 

 

 

 

 

 

Number of Number of
Studies Subjects Mean rho SD rho
ALL STUDIES 8 1186 .21 .16
OUTPATIENTS 3 3 89 .32 .02
STUDENTS 5 797 . 15 . 16

 

In contrast to the predicted negative correlation between scales, the empirical

studies have found a positive correlation between the two style scales. The meta-analysis

across all 8 studies found an average correlation of r = .21. This finding simply

disconfirrns the type theory.

There is more variation in the correlations than is explained by sampling error (the

standard deviation of population correlations is estimated to be SD = .16). Variation

would be predicted by the type theory since the correlation is predicted to be much more

highly negative for an outpatient population than for a student population. Separate

analyses were done for the two types of populations: outpatients and students. For the

outpatient studies, the average correlation is a positive I = .32 with virtually no variation

 

across studies (estimated SD = .02). For the student samples, the average correlation is a
positive _r: = .15. That is, the correlation is positive for both populations. The correlation

is a relatively low +. 15 for students and a much higher +.32 for outpatients. These

38

findings are completely inconsistent with the type theory.

There was one lone study which did find a negative correlation. McCranie and
Bass (1984) report a sample correlation of r = -.13 for a sample of N = 86 students.
However, this is a very small sample and the sample correlation could be negative by
chance. The 95% confidence interval for this study is -.34 to +08. Since the confidence
interval does extend into the positive region, it is possible that the negative value is a
fluke. If the negative value for this one study is due to sampling error, then all 8 studies
Show positive correlations.

To summarize; the type theory of anaclitic and introjective depression predicts a
negative correlation between the Anaclitic and Introjective depression scales, a low
negative correlation for the general population correlation, and a much larger negative
correlation for an outpatient population. The fact is that both correlations are positive.
The average correlation for students is a positive +. 15, while the average correlation for
outpatients is an even higher positive +.32. Thus, there is a massive departure between
the scale findings and the implications of the original type theory.

Correlations with Depreasfira

There is another potential problem for the type theory. Suppose the two scales are

scored 0 and 1 for “low” and “high,” respectively. The two scales considered together

should perfectly identify depressives (see Table 8).

Table 8

Predicted Sums of Levels for Identification of Anaclitic and Introjective Depressives

 

 

 

 

 

 

 

 

Person Type Level on Anaclitic Level on Introjective Sum of Levels
AD 1 0 l + 0 = 1
ID 0 l O + l = 1
ND 0 O O + O = O

 

 

This table shows that there should be a perfect multiple correlation for depression
as predicted fiom the two scales. While there is evidence of positive correlation between
each scale and depression scales, there appears to be no check to see if the correlations are
as high as required to have a multiple correlation of 1.00.

Research Objectives

Research is needed to establish the exact relationship between depression and the
Anaclitic and Introjective depression scales. Since the two scales are positively correlated,
the original type theory cannot be correct. However, there is an alternative theory.
Anaclitic style and introjective style might be symptoms of depression. If this were true, it
would be consistent with the original anecdotal observations that generated the type
theory. However, it would also be true that as symptoms of depression, the two would be
expected to be highly correlated with each other. As symptoms rather than types, perfect
prediction of depression would no longer be predicted.

Consider anaclitic style and introjective style as symptoms. Existing scales for

40

depression have given little consideration to these traits. This may be an error of
omission. On the other hand, the two scales do not appear to be equally linked to
depression. The two most frequently used depression questionnaires may be the Beck
Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961) and the
lung Depression Scale (Zing, 1972). Both are more highly correlated with the
introjective style than with the anaclitic style (Blatt, Quinlan, Chevron, McDonald, &
Zuroff, 1982). This could mean that the anaclitic style is not a symptom of depression or
it could mean that dependency is a key symptom which has been omitted from current
scales.

As it happens, dependency might be a symptom for anxiety rather than depression.
The Anaclitic scale is more highly correlated with fearfulness and anxiety than is the
Introjective scale (Persons, Burns, Perloff, & Miranda, 1993; Zuroff & Mongrain, 1987).
Consistent with this, a dependent, or anaclitic, style has been linked to both depression and
anxiety disorders in outpatients, whereas the introjective style was only associated with
depression (Bagby et al., 1992).

Differences Between Anxiety Symptoms

Research also suggests that it may be important to distinguish between different
types of anxiety. Panic disorder is defined by DSM-II-R as generalized anxiety disorder
plus panic attacks. Panic disorder is more highly associated with depression than are
generalized anxiety disorders. There are specific symptoms of panic disorder, such as

panic attacks, which may facilitate definition. It may be more difficult to distinguish the

41

psychological symptoms of anxiety disorders versus symptoms of depression.

There are indications that it may be usefirl to distinguish between cognitive
symptoms of anxiety and somatic symptoms of anxiety. Cognitive anxiety is characterized
by apprehension and worry associated with specific cues, muscle tension, Shyness,
fatiguability, and behavioral inhibition, whereas somatic anxiety is linked to somatic
complaints, high autonomic arousal, and impulsivity (Cloninger, 1988b). Factor analytic
studies have shown these to be relatively distinct clusters of symptoms, with some
different risk factors (Cloninger, 1988b). For example, Cloninger (1986) found that
criminality is associated with increased susceptibility to somatic anxiety and decreased
susceptibility to cognitive anxiety. Autonomic and somatic symptoms appear to be also
useful in distinguishing between anxiety and depressive disorders (King, Margraf, Ehlers,
& Maddock, 1986).

Summag

One way to understand the conflicting reports regarding consistencies versus
inconsistencies between anxiety and depression is to look at the themes and specific
symptoms associated with each disorder. Some subtypes of depression may be more
highly associated with specific symptoms of anxiety than others. For example, somatic
symptoms have been more highly linked to anxiety than to depression (King, Margraf,
Ehlers, & Maddock, 1986), and yet they have also been linked to a dependent depressive
style (Beck, Epstein, & Harrison, 1983). It is important to remember that the “disorders”

under consideration in this study are created, to a large extent, by our attempts to classify,

42

understand, and treat individuals who present with a diverse array of symptoms. These
classification structures have evolved along with our understanding, which is always
somewhat arbitrary and limited by our preconceptions and limitations. Persons (1986)
noted the importance of first delineating the actual psychological phenomena under study
in order to be able to build theories which account for those phenomena. In this study, an
attempt is being made to delineate the symptoms, or themes, associated with anxiety and

depression in order to better understand their interrelationships.

CHAPTER III

PLAN OF ANALYSIS

In this study I will review symptoms of anxiety and depression described in the
DSM-III-R, ICD-9, Diagnostic Interview Survey, Research Diagnostic Criteria, and
NEO-PI, along with depictions by theorists, such as Beck (1967; Beck, Epstein, Brown, &
Steer, 1988), Blatt (1974), Nurcombe and his colleague (1988), Gotlib & Cane (1989),
Spielberger (1972), Izard (Bartlett & Izard, 1972; Izard, Blumberg, & Oyster, 1985),
Watson and Clark (1984; Clark, 1989), and Tellegen (1985). I will also compile a list of
symptoms from two self-report measures of depression (the Beck Depression Inventory;
BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961, and the Symptom Checklist-90;
SCL-90; Derogatis, Lipman, & Covy, 1973) and two measures of anxiety (the Beck
Anxiety Inventory; BAI; Beck, Epstein, Brown, & Steer, 1988, and the SCL-90) which
were specifically chosen because they have been found to have good discriminant validity
(Gotlib & Cane, 1989). Symptoms will also be compiled from scales which assess guilt,
shame, dependency, self-criticism, interpersonal sensitivity, mastery, self-efficacy,
self-esteem, hostility, and negative and positive affectivity (see below for a detailed
description of these instruments). These constructs have all been depicted as important
aspects of anxiety and/or depression. The symptoms and items will be the data which will

be used in subsequent data analyses to determine whether there are unique symptoms

43

44

related only to depression versus unique symptoms related only to anxiety, and whether
current instruments adequately measure those symptoms.

A carefirl psychometric analysis of these symptoms and items will make it easier to
begin to evaluate the instruments, as well as some of the theories which have been
presented. It is possible that I will be able to differentiate groups of symptoms which are
uniquely associated with either anxiety or depression. In that event, I will be able to
evaluate whether the instruments commonly used to measure those constructs contain only
relevant items, and also contain an adequate sampling of relevant items. It is also possible
that I will be able to differentiate subgroups of anxiety (or depression) symptoms which
relate to anxiety (or depression), but relate less well to each other, supporting the
heterogeneity hypothesis. A third possibility is that I will be unable to distinguish
meaningfirl clusters which distinguish symptoms of depression fiom those of anxiety. This
would lend support to those theories which have conceptualized anxiety and depression as
a continuum of mood disturbances.

For example, Watson and Clark (1984) have suggested that we can better
understand the commonalities and distinguishing features of depression and anxiety by
looking at aspects of temperament, such as affective dispositions. They have argued for a
model in which one higher order category, “negative affectivity,” is used to explain the
high correlations between anxiety and depression, and the category “positive affectivity”
is used to explain the imperfect correlation between those two syndromes. Other theorists

(e. g. Tellegen, 1985) have argued that the general category of trait negative affectivity is

45

uniquely associated with anxiety, whereas it is low positive affectivity which uniquely

characterizes depression.

IO

SO:

Erb
Self.
then
SeVe.

Irma

CHAPTER IV

METHOD

Participants

Participants were 179 male and 187 female students between the ages of 17 and 29
(mean = 19) at a large midwestem university. Most of the students were Caucasian
(83%), from middle to upper-middle class families. They were recruited from
introductory psychology courses, and received credit toward course grades for
participation.

Instruments

Demographic Information

Demographic information was collected fi'om the students, including age, religion,

social status, and ethnic background.

Depression and Anxim
The Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, &

Erbaugh, 1961) was used to assess symptoms of depression. This scale consists of 84
self-evaluative statements which are grouped into 21 categories, Within each category,
there are four item choices which increase in severity fi'om neutral ("1 do not feel sad") to
severe ("I'm so sad or unhappy that I can't stand it"). Categories include mood, guilt,

irritability, crying spells, and sleep and appetite disturbance. Items are scored fiom 0 to 3.

46

47

Scores of 24-63 reflect a severe level of depression; 16-23 suggest more moderate levels;
10-15 indicates a mild level of depression, and 0-9 indicates no depression (Shaw, Vallis,
& McCabe, 1987).

Although good split-half reliabilities have been reported (.86, on the average),
factor analyses of the BDI with patients have resulted in three primary factors: sad
mood/negative self-image, somatic complaints, and psychomotor retardation (Beck &
Beamesderfer, 1974; Vredenburg, Krames, & Flett, 1985). Factor analysis of the current
data was roughly consistent with this pattern, although researchers have not consistently
found this pattern in nonclinical samples (Golin & Hartz, 1979; Lips & Ng, 1985). The
BDI correlates well with both other self-report measures of depression, and clinician's
ratings (Schwab, Bialow, & Holzer, 1967).

The Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988) was
used to measure anxiety. The BA] is a 21 item scale derived from an item pool of anxiety
symptoms. The items were selected to reflect a broad range of cognitive ("fear of dying"),
somatic ("faint”; "feeling hot"), and affective (”scared"; "nervous") symptoms of anxiety,
and to reduce overlap with symptoms of depression. The scale has shown good internal
consistency (alpha = .92), and has been found to correlate more highly with other
measures of anxiety (I = .51) than with measures of depression (1 = .25) in previous
studies (Beck et al., 1988).

The Symptom Checklist-90 (SCL-90; Derogatis, Lipman, & Covy, 1973) is a

90-item instrument designed to measure symptoms of psychological distress. The

48

individual is asked to rate on a 5-point scale (from 1 = M to 5 = xtremely) how
much that problem has distressed them during the past 7 days. Five of the nine possible
clusters were included in the inventory in this study: somatization ("headaches";
"faintness or dizziness"), depression ("blaming yourself for things"; "feeling no interest in
things"), anxiety ("feeling tense or keyed up"; "the feeling that something bad is going to
happen to you"), hostility ("feeling easily annoyed or irritated"; "having urges to beat,
injure, or harm someone"), and interpersonal sensitivity ("your feelings being easily hurt";
"feeling very self-conscious with others"). Although some researchers have suggested that
this instrument primarily measures one general dimension of psychopathology (Brophy,
Norvell, & Kiluk, 1988; Cyr, McKenna-Foley, & Peacock, 1985), both statistically and
substantively meaningful factors have been found (Shutty, DeGood, & Schwartz, 1986).
This instrument has been found to have good internal consistency and test-retest reliability
over short periods of time (Derogatis et al., 1973), and has been found to correlate
strongly with other self-report instruments (Gotlib & Cane, 1989).
PersonalfityDimenaipps

The Depressive Experiences Questionnaire (DEQ; Blatt, D'Afilitti, & Quinlan,
1976a) was used to assess differences in depressive dimensions. This instrument does not
assess symptoms of depression, but rather was designed to reflect differences in attitudes
toward self and others believed to be relevant to depression. This inventory consists of 66

items which are rated on a 7-point scale (fi'om l = strongly disagree to 7 = trongly

aggee). In previous studies, three primary factors have been identified and replicated:

49

Dependency, Self-Criticism, and Efficacy (Blatt, D'Afflitti, & Quinlan, 1976b). These
factors also show strong test-retest stability (Zuroff, Moskowitz, Wielgus, Powers, &
Franko, 1983). Construct validity was established in previous studies in which the
dependent and self-critical scales predicted anaclitic and introjective depressive states,
respectively Zuroff, Igreja, & Mongrain, 1990; Zuroff & Mongrain, 1987). For each of
the three factors, those items fi'om the manual with factors loadings of over .40, which did
not also load significantly on any of the other two factors, were selected for use in the
current study.

The Self-Criticism, or Introjective, scale consists of 12 items reflecting
dissatisfaction with self, ambivalence toward others, and failure to meet standards or
expectations ("there is a considerable difference between how I am now and how I would
like to be"; "very frequently, my feelings toward someone close to me vary. There are
times when I feel completely angry, and other times when I feel all-loving toward that
person").

The Dependency, or Anaclitic, scale consists of 13 items which reflect fears of
being rejected or hurting another person, in which pleasing others is very important and
strong negative emotions are experienced as threatening to relationships ("I find it very
difficult to say "No" to the requests of fiiends"; "if someone I cared about became angry
with me, I would feel threatened that he (she) might leave me").

The Efficacy scale of the DEQ consists of 6 items which reflect achievement

strivings and a sense of personal accomplishment ("I set my goals and standards as high as

50

possible"; "What I do and say has a very strong impact on those around me").

Affectivity was measured by the General Temperament Survey (GTS; Clark &
Watson, 1989). The GTS consists of 3 scales which measure distinct aspects of
personality. Items are rated as either true or false. The Negative Temperament scale
consists of 28 items reflecting negative mood as well as self-concept. Watson and Clark
(1993) report that individuals with high scores on this scale often experience sadness,
anxiety, and guilt, as well as other negative affects ("I often feel nervous and ‘stressed’";
"I am often troubled by guilt feelings"; "I have days that I'm very irritable"). In addition,
they are likely to be pessimistic and to have difficulties adjusting to failure or frustration
("I sometimes get too upset by minor setbacks").

In contrast, the Positive Temperament Scale measures the individual's tendency to
experience positive affective states. This scale consists of 12 items reflecting positive
affect ("I am usually enthusiastic about the things that I do"; "I often feel lively and
cheerful for no particular reason"), 12 items indicative of high energy ("Most days I have a
lot of "pep" or vigor"), and 3 additional, nonspecific, positive temperament items ("I am
usually alert and attentive"; "I get excited when I think about the firture"; "I get pretty
excited when I'm starting a new project"). Individuals who score highly on this scale tend
to be cheerfiil and enthusiastic, and report more satisfying social interactions (Watson,
1988; Watson & Clark, in press).

The Disinhibition versus Constraint Scale of the GTS consists of 35 items tapping

traits thought to be relevant to the construct of disinhibition. Sample items include "I

51

rarely, if ever, do anything reckless" (scored negativelY); "I really enjoy beating the
system"; and "taking care of details is not my strong point". In preliminary investigations,
a longer version of this scale correlated highly with scales measuring Impulsivity (.68);
Irresponsibility (.63); Risk Taking (.59); (Low) Persistence (.56); Playfulness (.54); Norm
Rejection (.49); Danger Seeking (.47); and Disorganization (.40; Watson & Clark, 1993).
Individuals who score highly on this scale are likely to be easygoing and somewhat
irresponsible; to enjoy spontaneity, change, and excitement; and to have little motivation
to conform to rules or to pursue traditional values.

Preliminary investigations suggested that these three scales correlate well with
instruments measuring similar constructs. In previous investigations, the Negative
Temperament scale correlated .72 with the Negative Emotionality scale of the
Multidimensional Personality Questionnaire (MPQ; Tellegen, in press), and .84 with the
Neuroticism scale of the Eysenck Personality Questionnaire (EPQ; Eysenck & Eysenck,
1975; (Watson & Clark, 1993; Watson & Clark, in press). Positive Temperament
correlated .76 with the Positive Emotionality scale of the MPQ, and .68'with the
Neuroticism scale of the EPQ. The preliminary measure of disinhibition correlated -.56
with the Constraint scale of the MPQ, .55 with the Psychoticism scale of the EPQ, and
-.61 with the Conscientiousness scale of the NEO Personality Inventory (NEO-PI; Costa

& McCrae, 1985) (Watson & Clark, 1992).

52

Affective Self-Evaluations

Measures of self-evaluations, which have been enumerated as symptoms of
depression were included in an Affect Inventory composed of items from four scales.
Self-esteem was measured with the Rosenberg Self-Esteem Scale (Rosenberg, 1965).
Items are endorsed on a 5-point scale from strongly disagree (1) to strongly agree (5), and
include "on the whole, I am satisfied with myself".

Mastery was assessed using the Mastery Scale from the Offer Self-Image
Questionnaire (Offer, Ostrov, & Howard, 1989). This scale consists of 10 items which are
endorsed on a 5-point scale from strongly disagree (1) to strongly agree (5), and include
"IfI put my mind to it I can learn almost anything". This scale has shown adequate
concurrent validity (Ostrov et al., 1989).

Guilt was measured by the Chang and Hunter Guilt Scale (Chang, 1988). This
scale consists of 9 items based on a definition of guilt in terms of perceptions of self as
causing harm to others and efforts toward making reparations when harm has been done.
Items include "Sometimes I cannot forgive myself for having caused deep pain in those I
love or care for". Items are rated on a 5-point scale from strongly disagree (1) to sgo_ngly
m (5).

Global shame was measured by items fiom the Intemalized Shame Scale (ISS:
Cook, 1985) as selected by Chang and Hunter (Chang, 1988). The revised Shame Scale
consists of the 11 items found to measure shame (e. g., "I feel like I am never quite good

enough"), eliminating those which Chang found to address other constructs. The items

53

are negatively worded, and are rated on a 5-point scale ranging from myer (1) to a1n_ro_§r
am (5).
Procedure
The self-report measures described previously were administered to the students in
a group setting. Participants were informed that the purpose of the study was to better
understand the affective experiences of college students. They were also informed as to
their rights as subjects. To ensure anonymity, data were identified only by code numbers.

Participants received extra course credit for their participation.

lv—d

CHAPTER V

RESULTS

Hierarchical Structure of the Data Analysis
The analysis of the data for this dissertation is hierarchical in nature. Since
hierarchical analysis is unfamiliar to many readers, this first section of the results section
will outline the steps in the hierarchical analysis. The specific steps in the actual analysis

will then be presented in detail.

Terminology

The items selected were those which various authors have used to define "anxiety
and "depression". These items can be considered using two different psychological
terminologies. The psychiatric orientation focuses on a short term way of thinking,
perhaps best captured by the word "symptom". Symptoms are assumed to vary
considerably over time depending on environmental stress as well as the underlying status
of the individual; the latter may be especially variable in the case of people diagnosed with
bipolar disorders. Much of the present data is discussed in these terms.

The second terminology being utilized may best be conceptualized using the
language of trait theory. For example, most psychologists would expect shame and self-
esteem to be long-term personality traits showing only slight variation over short

durations, such as a week or a month. For such items, the natural terminology would be

54

tem

item

8an
dysl
Ofie
the

SCC(

mes
«Van

depr

55

terms such as "characteristic" or "trait". The terminology used here will vary from one
item to another depending on the typical usage pattern in clinical psychology, though at
this point there is no attempt to sort the items in terms of etiology or assumed time
variation.

The Hierarchical Structure of Symptoms

The single items collected in this study are directed at specific experiences or
specific facts. The concept of " symptom" is a higher order concept. A symptom such as
"feelings of dependency" could be expressed in a variety of experiences. Thus, we would
predict that items would form clusters in which the items within a given cluster all measure
the same symptom as expressed by different Specific experiences.

The concepts of "anxiety" and "depression" are higher order concepts. For
example, depression is thought to be expressed in various symptoms such as anhedonia or
dysphoria. There is no established language for this higher level, although the word most
often used is "syndrome". In this way, items can be grouped in a hierarchical pattern. At
the first level, items can be clustered in terms of the symptoms they measure. At the
second level, symptoms can be clustered in terms of the syndrome that is expressed.

Some authors (cf. Watson & Clark, 1992; Watson, Clark, & Carey, 1988) have
argued for a still higher order factor. In the case of the symptoms considered in this
thesis, the relevant highest order concept is that of "neuroticism" or "negative affectivity".
Many current authors (cf. Clark, Watson, & Mineka, 1994) consider both anxiety and

depression to be expressions of this higher order trait, along with loneliness, shame, self-

56

consciousness, and other negative affects. Some authors include hostility in this higher
order trait.

The firll hierarchical clustering scheme could be considered as a nested
classification of the items at the following levels:

Level 1 - items are clustered by symptom
Level 2 - symptoms are clustered by syndrome
Level 3 - syndromes are clustered by major personality factor.

In this thesis I conducted data analyses at all of these levels.
Items and Symptoms or Characteristics

I began with 290 items. However, there was never any consideration of the
possibility that there are 290 different symptoms or characteristics to be measured.
Rather, it was assumed that items could be clustered into sets in which the items within a
set measure the same symptom or characteristic. For simplicity, this section will be
written in the language of symptoms, even though characteristics will be also be
considered in the analysis.

At the first level of analysis, the items were clustered into symptoms. This was
initially done using content analysis. The clusters were then evaluated using confirmatory
factor analysis. A major problem revealed by the confirmatory factor analysis was with
items using a binary response format. Items using a binary response format suffer from a
response set problem that was not anticipated when the study was designed; a key article

detailing this problem by Green, Goldman, and Salovey (1993) had not been published

until

prese
him
the er
of net

discar

39 ch
chara<
measu
the nu
items '

the lie

the lon
items C
II is lhl
SWPIO
cOnsisn

the Squz

57

until after the data were gathered. This issue will be discussed in detail following the
presentation of the content analysis. However, due to the problems encountered with the
binary items, those items were dropped from the analysis. That is, in spite of regrets over
the content lost thereby, the items from the Watson and Clark inventory measuring aspects
of negative affectivity, positive affectivity, and disinhibition versus constraint were
discarded from this analysis.

Without the Watson and Clark items, there are 200 items that were classified into
29 clusters. That is, the items under analysis cover 29 different symptoms or
characteristics identified by the content analysis. On average there are about 4 items
measuring each symptom or characteristic. However, there is a great deal of variation in
the number of items written to cover each symptom or characteristic. Once the binary
items were removed, the confirmatory factor analysis showed good fit for the clustering of
the items.

Consider the items in a cluster measuring one symptom, such as "fear" (although
the longer word "apprehension" would fit the content more closely). With the binary
items discarded, there were 11 items identified in the content analysis as measuring fear.

It is thus possible to use these items to define a fear scale. This scale would measure the
symptom of fear, although that measurement would be imperfect. The internal
consistency of the fear scale was .89, so the measurement is quite good: a correlation of
the square root of .89 (.94) between the scale score for fear and the fear construct itself.

Authors differ in the extent to which they have considered various symptoms in

defining
slight c
exampl
Hyper;
items 2
tepres:
is only

reliabi;

COITel.‘
Corral;

Scales

537nm
be let
the co
adlUSt:
Clues,

58

defining the constructs of anxiety and depression. Some symptoms have been given only
slight consideration by any authors. This was very much evident in the present data. For
example, the cluster measuring Psychomotor Tension had 10 items and Autonomic
Hyperarousal had 17 items; however, there were many symptoms represented by only two
items and some symptoms were represented by only single items. The symptoms
represented by the latter items were regarded as poorly measured.

Symptoms: ScaLesimd Constructs

The actual symptom of fear is the construct measured by the Fear scale. The scale
is only an imperfect measurement of that symptom. If more fear items were used, the
reliability would go up and the scale would be a better measure of the symptom.

Using confirmatory factor analysis, it is possible to obtain estimates of the
correlations between symptoms rather than the correlations between scales. These
correlations are corrected for the attenuation produced by error of measurement in the
scales (see Appendix B).

The problem with the estimated correlations between constructs is that many
symptoms were measured by only one or two items. The reliability of such scales tend to
be very low and thus correction for attenuation requires a large adjustment in the size of
the correlation. This greatly exaggerates the problem with sampling error. A large
adjustment greatly increases the sampling error in the estimated correlation. In some
cases, the sampling error can cause an estimated correlation to be larger than 1.00: an

outcome which is primarily troublesome for those not well trained in reliability theory. A

popula'
error 0
Thus, 1

interpr

"depre

S}ndt<

Slmpt

depres

been:

 

59

population correlation Cannot be larger than 1.00, but a sample correlation corrected for
error of measurement can be larger than 1 because of sampling in the estimation process.
Thus, for those symptoms measured by very few items, the corrected correlations must be
interpreted with great care.

Smptoms and Syndromes

The key questions in the controversies over the definition of "anxiety" and
"depression" can be stated at the next level of the hierarchical analysis; are there separate
syndromes for anxiety and depression? If so, which symptoms define each syndrome?

The key statistical analysis for this question starts with the correlations between
symptoms. Are there two clusters of symptoms that can be identified as anxiety and
depression, respectively? If so, which items belong to each cluster?

One method by which such questions have been considered in the literature has
been exploratory factor analysis: sometimes referred to simply as "factor analysis".
Advocates for exploratory factor analysis claim that this method will identify the
underlying factors and the corresponding clusters of variables that measure each factor.
Exploratory factor analysis was done for the symptom correlations: one analysis of the
symptom construct correlations (i.e., corrected for attenuation) and another analysis of the
symptom scale correlations (i.e., not corrected for attenuation).

There are two fundamental problems with exploratory factor analysis: (a) it
assumes uncorrelated basic traits and (b) it ignores causal relations between variables. If

these two conditions are not met, then exploratory factor analysis may give very

nine
theo

confi

nlcu
case.
phyn
nnkl
sufltr

Create

Inathe
Shdn
been:

nil b.

\3

cOnsid
chfllac

plOble

6O

misleading results. For highly correlated constructs that are causally related to each other,
the only form of factor analysis that gives good estimates of factor correlations is
confirmatory factor analysis.

Consider anxiety and depression: By whatever form of measurement, the literature
has consistently shown these two syndromes to be very highly correlated. Furthermore,
all current theories predict that they should be highly correlated (or "co-morbid" in the
case of extreme cases). High anxiety should cause depression because of the neural
physiology of the anxiety process. High anxiety causes a depletion of norepinephrine
which results in state depression (Weiss & Simson, 1985). On the other hand, people who
suffer depression also suffer many impairments of social and work life. This in turn would
create stress and accompanying anxiety.

In this way, current theory regarding anxiety and depression predicts that the
mathematical factors of exploratory factor analysis will be far removed from the actual
syndromes that cause the symptoms. For this reason, the exploratory factor analysis will
be considered only as a hypothesis generating device. The real test of the syndrome model
will be made using confirmatory factor analysis.

The Structure of the Results Section

The results themselves will be developed in segments. The first segment will
consider the process of clustering items to form scales that measure symptoms and
characteristics. This section will present the content analysis and briefly discuss the

problems caused by the binary item format used in the Watson and Clark inventory. The

keylnu
and otl

to beu

snnptc
that an.
lnaddl

conela

higher 1
'pure"
”mixed

defined

Slndror
75 ben
morbidj

cOmelat

c(”Bidet

all-em”

61

key outcome from this segment is that the 200 non-binary items measure 29 symptoms
and other characteristics. The correlations between these symptoms are the basic results
to be used to test the various syndrome hypotheses for anxiety and depression.

The second segment presents the results of the analysis of the correlations between
symptoms. This will include a discussion of the exploratory factor analysis, even though
that analysis was flawed by the high correlations between the basic constructs measured.
In addition to the high correlation between anxiety and depression, there were also high
correlations between those syndromes and other syndromes and characteristics.

The confirmatory factor analysis shows that the 29 symptoms do indeed define 10
higher order syndromes and traits. Two of these syndromes are tentatively defined as
"pure" measures of anxiety and depression while two other syndromes are defined as
"mixed" measures of anxiety and depression. There are 6 other syndromes or traits
defined by the 29 basic symptoms and characteristics.

The third segment of results will start with the correlations between the 10
syndromes identified by clusters of symptoms. A key fact is the very high correlation of
.75 between pure anxiety and pure depression, which matches the high level of co-
morbidity found in epidemiological studies. Various causal interpretations of these
correlations can be considered.

Some authors have argued that all of the syndromes defined in this thesis can be
considered as expressions of one underlying trait called "neuroticism" or "negative

afi‘ectivity". This hypothesis will be considered and a factor analysis of the 10 syndromes

nil

OK

pC

an

I?

In

Ct:

62

will be considered in this regard. Much important information is lost if only the higher
order factor is considered and the individual syndromes are ignored.
Itemnd Symptoms

This segment presents l) reliabilities of the original scales fi'om which the item
pool was derived and 2) an analysis of the 290 items into clusters that measure symptoms
and other characteristics.
Reliabilities

Depression and anxieg. Internal consistency for the BDI was .90 in the current
sample. In the current study, the BDI correlated highly with the SCL-90 depression scale
(.72), yet also correlated highly with anxiety scales (.65; .67). This is consistent with
other researchers who have found poor discriminant validity between the EDI and
measures of anxiety (Mullaney, 1987). The reliability coefficient for the BAI in the
current sample was .93. The BAI correlated highly with the other measure of anxiety (_r_ =
.82), but did not shOw good discrimination fi'om the measures of depression (rs = .65;
.74). Internal consistency for the scales of the SCL-90 in the current study were .88
(somatization), .87 (anxiety), .91 (depression), .79 (hostility), and .86 (interpersonal
sensitivity).

Personalig dimensions. Internal consistency findings for the DEQ were .82 for the
Introjective scale, .73 for the Anaclitic scale, and .75 for the Efficacy scale. Internal
consistency findings for the GTS were .88 for the Negative Temperament scale, .85 for

the Positive Temperament, and .79 for the Disinhibition scale.

63

Affective self-evaluations. Internal consistency for the Rosenberg Self-Esteem

 

Scale was .86 for the present sample. Internal consistency for the Mastery Scale was
.71. Internal consistency for the Chang and Hunter Guilt Scale was .77. Internal

consistency for the revised Shame Scale was .92.

Items and Smptoms

The first step was a content analysis of all items included in the study (See
Appendix A). This was followed by a confirmatory factor analysis to test the content
model (See Appendix B).

The confirmatory factor analysis showed three sets of problems. The most severe
problem was caused by the binary response format for the General Temperament Survey
(GTS) Inventory. These data showed the same response set problem reported by Green,
Goldman, and Salovey (1993), who showed how the use of forced-choice format can
distort results because of systematic response bias. As a result, several clusters were split
into binary and non-binary clusters. For example, an initial "won'y" cluster was split
between "worry" and "worry-binary". There are ways to statistically reduce the effect of
the binary response set, but the GTS inventory does not have the right mix of items to use
those methods. So ultimately all clusters measured using the binary format were dropped.

The confirmatory factor analysis also showed problems for two other original
clusters. The SCL-90 cluster for "interpersonal sensitivity" was not questioned in the
original content analysis. However, the confirmatory factor analysis showed that it is

multidimensional and needed further analysis. This ultimately produced two clusters

QB.

dl
31‘.
fat

fat

64

called "Unliked" and "Self-Conscious." The confirmatory factor analysis also showed
problems for the "Worry" cluster.
A Priori Content Analysis

The initial content analysis focused on symptoms associated with anxiety and
depression. Exemplars were taken from manuals of clinical diagnostic criteria, such as
research diagnostic criteria, and theoretical and empirical papers, with special attention
given to the DSM-II-R criteria to ensure that the symptoms under study reflected the
dimensions which have been depicted as important in diagnosing these disorders. Special
attention was also given to those theorists who have delineated explicit theories about
factors associated with depression and/or anxiety, and developed instruments using those
factors, such as Beck. These sources were chosen, as mentioned previously, because the
instruments in question have been found to have good discriminant validity (Gotlib &
Cane, 1989). An attempt was made to sample the relevant literature broadly, across
theoretical perspectives, for symptoms which would be accessible through self-report.

The symptoms, or themes, were then divided into 3 groups: those associated with
depression only, those associated with anxiety only, and those associated with both
depression and anxiety. The symptoms most specifically associated with depression
included dysphoria, low self-esteem, indecisiveness, hopelessness, psychomotor
, retardation, anhedonia, feelings of guilt or worthlessness, and suicidal ideation. Those
symptoms most specifically linked to anxiety included feelings of worry, anxiety,

disquietude, and irritability; vigilance; fears of dying or being out of control; somatic

65

symptoms, such as shortness of breath, dizziness, chest pains, aches, flushes, numbness,
and an exaggerated startle response; depersonalization or derealization; and shyness.
Symptoms which were linked to both anxiety and depression included appetite disturbance
and abdominal distress, sleep disturbance, fatiguability, concentration difficulties, and
agitation.

A content analysis of the items from all scales was performed, which yielded the
following themes: Fear, Motor Tension, Autonomic Hyperarousal, Vigilance, Irritability,
Insecure, Upset, Tense, Worry, Dysphoric Mood, Fatigue, Energy, Enthusiasm, Sleep
Disturbance, Appetite Disturbance, Cognitive Disturbance, Dependency, Insufficiency,
Need for Approval, Self-Critical, Anhedonia, Self-Esteem, Efficacy, Unstable, Guilt,
Anger, Hostility, Negative Afi‘ectivity, Positive Affectivity, Hopelessness, Suicidality,
Interpersonal Sensitivity, Impulsivity, Irresponsible, Playful, Antisocial, Sensation Seeking,
Disorganized, and Ambitious. Also included in the analysis were the clusters formed by
the Rosenberg Self-Esteem Scale (SE), the Mastery Scale (Mastery), the Cook Shame
Scale (Shame), and the Chang-Hunter Guilt Scale, respectively. Each of these scales was
homogeneous enough to be included in the analysis as complete clusters. A complete list
of items in each cluster is included as Appendix A.

The “Fear” cluster consisted of five items from the Beck Anxiety Inventory (BAI),
five items from the Anxiety subscale of the Symptom Checklist (SCL—90), and one item
from the Depression subscale of the SCL-90: “feelings of being caught or trapped.” The

items within this cluster depicted general feelings of fearfirlness and apprehension, such as

ln'

ch;

reh

EX;

gene

66

“the thought that something bad is going to happen to you,” as well as specific fears, such
as “fear of dying” and “fear of losing control.”

The “Motor Tension” cluster consisted of four items from the Somatization
subscale of the SCL-90, three items from the Anxiety subscale of the SCL—90, and three
items from the BAI. The items depicted many of the types of motor tension symptoms
often associated with anxiety, such as aches, pains, shakiness, and trembling.

The third cluster, “Autonomic Hyperarousal,” consisted of 11 items from the BAI,
five items fi'om the Somatization subscale of the SCL-90, and one item from the Anxiety
subscale of the SCL-90. It consisted of items associated with autonomic hyperarousal,
such as dizziness, abdominal discomfort, heart pounding or racing, and hot or cold spells.

Cluster 4 was defined by symptoms of “Vigilance,” often associated with anxiety
disorders. It was composed of two items from the BA]: “unable to relax” and “nervous,”
and one item fi'om the Anxiety subscale of the SCL-90: “feeling tense or keyed up.”

Cluster 5, “Irritability,” was composed of one item from the Beck Depression
Inventory (BDI), and one item fiom the Hostility subscale of the SCL-90, which were
characterized by feelings of extreme or excessive irritability.

The sixth cluster was characterized by “Insecurity,” especially as regards
relationships, and consisted of two items from the Introjective subscale of the Depressive
Experiences Questionnaire (DEQ).

Clusters 7 through 9 consisted of items describing feelings often associated with

general distress, or “negative affectivity.” Cluster 7 was termed “Upset,” and consisted of

67

four items from the NA subscale of the GTS, which described a tendency to become
overly upset at small setbacks. Cluster 8 was termed “Tense.” It consisted of three items
from the NA subscale of the GTS which described a tendency to feel tense, stressed, or
“on edge.” The ninth cluster was termed “Worry.” It consisted of one item fiom the BDI:
“I am so worried about my physical problems that I cannot think of anything else,” and
one item from the Depression subscale of the SCL—90: “worrying too much about things.”

Clusters 10 and 11 described symptoms often associated with depression. Cluster
10, termed “Dysphoric Mood” consisted of 3 items fi'om the Depression subscale of the
SCL-90 and 2 items from the BDI. These items described many of the feelings often
associated with depression, such as loneliness, sadness, and tearfulness. Cluster 11, on
the other hand, consisted of symptoms of the “Fatigue” which is often associated with
depression. This cluster consisted of two items from the Depression subscale of the SCL-
90, two items from the BDI, and one item from the NA subscale of the GTS. The items
depicted lethargy, tiredness, and the feeling that life “feels like a big struggle.”

Clusters l2 and 13 consisted of items from subscales of the Positive Temperament
(PA) subscale of the GTS. Cluster 12, “Energy,” consisted of all 12 items of the “Energy”
subscale, and described an active, energetic, fast-paced lifestyle. Cluster 13,
“Enthusiasm,” consisted of three items from the Positive Temperament subscale and two
items from the general Positive Affect subscale. The items depicted an enthusiastic and

excited attitude toward life.

Clusters 14 through 16 described some of the disturbances often associated with

68

depression. Cluster 14, “Sleep Disturbance,” consisted of one item from the EDI and one
item from the NA subscale of the GTS, describing sleep difficulties most often associated
with endogenous, or more severe, depression. Cluster 15, “Appetite Disturbance,”
consisted of two items from the BDI describing lack of appetite and extreme weight loss.
Cluster 16, “Cognitive Difficulties,” consisted of two items from the NA subscale of the
GTS describing mental confusion and troubling thoughts or ideas.

Clusters 17 through 19 each consisted of items from the DEQ which focus on
interpersonal needs. Cluster 17, “Dependency,” consisted of eight items from the
Anaclitic subscale of the DEQ, describing intense needs for interpersonal relatedness and
difficulties in being alone. Cluster 18, “Insufliciency,” consisted of two items from the
Introjective subscale of the DEQ which describe feelings of helplessness and emptiness.
Cluster 19, “Need for Approval,” consisted of five items from the Anaclitic subscale of the
DEQ which are characterized by a focus on attempting to please others as well as fears of
being criticized by or offending others.

Cluster 20, “Self-Critical,” consisted of three items from the EDI and one item
from the Introjective subscale of the DEQ which are characterized by self-blame, self-
hatred, and feeling of failure.

Cluster 21, “Anhedonia,” consisted of three items from the BDI and two items
from the Depression subscale of the SCL-90, and is characterized by feelings of
disinterest, boredom, and a lack of positive investment in living.

Clusters 22 and 23 focused on feelings of self-esteem and self-efficacy. Cluster 22,

69

“Self-Esteem,” consisted of three items from the Self-Efficacy subscale of the DEQ, one
item from the Interpersonal Sensitivity subscale of the SCL-90, and one item from the
Depression subscale of the SCL-90. These items address a sense of inner strength and
worth. Cluster 23, “Efficacy,” consisted of three items from the Self-Efficacy subscale of
the DEQ, which describe high goals, standards, and expectations.

Cluster 24, “Unstable,” consisted of two items from the Introjective subscale of the
DEQ, which describe extreme variability in feelings toward self and others.

Clusters 25 through 28 each was characterized by negative affect. Cluster 25,
“Guilt,” consisted of one item from the EDI, and one item from the NA subscale of the
GTS which described feelings of guilt. Cluster 26, “Anger,” consisted of three items from
the NA subscale of the GTS describing fi'equent, uncontained, or irrational anger. Cluster
27, “Hostility,” consisted of five items fiom the Hostility subscale of the SCL-90 and one
item from the Interpersonal Sensitivity subscale of the SCL—90 and consisted of items
describing feeling critical toward others, temper outbursts, and aggressive urges. Cluster
28, “Negative Affectivity,” consisted of two items from the NA subscale of the GTS
which describe general negative affectivity in terms to vague to be included in those
clusters which describe more specific negative feelings.

Cluster 29, “Positive Affectivity,” consisted of six items from the Positive Affect
subscale of the GTS, which characterize a positive, active investment in life and living.

Clusters 30 and 31 depicted some of the more severe symptoms associated with

depression. Cluster 30, “Hopelessness,” consisted of one item from the BDI and one item

Lo)

m0:

the (

70

from the Depression subscale of the SCL-9O describing feelings of hopelessness. Cluster
3], “Suicidality,” consisted of one item from the BDI and one item from the Depression
subscale of the SCL-9O describing the desire to end one’s life.

Cluster 32, “Interpersonal Sensitivity,” consisted of eight items from the
Interpersonal Sensitivity subscale of the SCL-90, describing feelings of self-consciousness
and uneasiness with others, as well as fears of being disliked or regarded as inferior.

Clusters 33, 34, and 35 consisted of items from the Disinhibition scale of the GTS.
Cluster 33, “Impulsivity,” consisted of eight items which are characterized by a somewhat
reckless, incautious, and unreasoned stance toward life. Cluster 34, “Irresponsibility,”
consisted of one item which is reverse scored: “1 am a serious-minded person.” Cluster
35, “Persistence,” consisted of one item which is reverse scored: “1 work just hard
enough to get by.” This item was also included in Cluster 40.

Cluster 36, “Playful,” consisted of three items fiom the Positive Affect subscale of
the GTS which are characterized by feelings of enthusiasm and playfulness.

Clusters 37 through 40 consisted largely of items from the Disinhibition subscale
of the GTS. Cluster 37, “Antisocial,” consisted of 11 items characterized by a lack of
regard for accepted rules and standards for social behavior, as well as a willingness to hurt
others to obtain a desired goal. Cluster 38, “Sensation Seeking,” consisted of eight
Disinhibition items, as well as one Positive Affect item, which are characterized by a desire
for excitement, novelty, and thrills. Cluster 39, “Disorganization,” consisted of one item:

“taking care of details is not my strong point.” Cluster 40, “Ambitious,” consisted of two

71

items which describe a willingness to work hard to achieve desired goals.

Clusters 41 through 44 consisted of the “Rosenberg Self-Esteem Scale,” the
“Mastery Scale,” the items from the “Cook Shame Scale,” and the “Chang-Hunter Guilt
Scale,” respectively. These scales have been described in detail in the Method section.
Confirmatogg Factor Analysis of Items

A confirmatory factor analysis tested empirically the item clusters identified by
the content analysis. The results of the confirmatory factor analysis are considered cluster
by cluster. For each cluster, it was possible to check to see if the items in that cluster form
a coherent or "unidimensional" set.

In order for a set of items to be unidimensional, they must all measure the same
construct. For this data, that means that all items must measure the same symptom or the
same trait. This is tested in confirmatory factor analysis by examining each cluster for
internal consistency - the pattern of correlations among items in the same cluster - and for
parallelism - the pattern of correlations between the items in that cluster and the other
symptoms and characteristics measured.

The confirmatory factor analysis also computes the alpha or Spearman-Brown
reliability of each cluster.

Binagg format response set. The main clusters manifesting severe problems were
clusters with items fi'om the General Temperament Survey (GTS) which uses a binary
response format. The binary items were not parallel to the other items. The binary items

showed much stronger correlations with all GTS clusters than did the non-binary items.

72

The binary items across different clusters such as "Irritability" or "Fear" correlated much
more highly with binary items from the other clusters than did the other, non-binary,
items.

This pattern of correlation for binary response format items has been found in
other studies (as reviewed in Green, Goldman, & Salovey, 1993). The fact that this
pattern derives from a binary format response set was proven in data gathered for that
purpose by Green and his colleagues (1993). Had that study been published when the data
for this study was collected, the design would have been changed and a non-binary format
would have been used.

There have been some heuristic methods used in previous studies to reduce the
effects of the binary response set, but because of the specific structure of the GTS, these
methods could not be used in this study.

The net results were these: First, each cluster defined by both binary and
non-binary items was split to form two clusters; one with binary and one with non-binary
items. Second, for the main purposes of the study, all binary item clusters were dropped
from fithher consideration.

Interpersonal sensitivity. The SCL-9O has a scale of items devoted explicitly to
"interpersonal sensitivity". The confirmatory factor analysis showed that this scale is n_ot
unidimensional. A detailed content analysis of the “Interpersonal Sensitivity” items was
then done. This identified two specific clusters of items which were labeled "Unliked" and

"Self-Conscious". The "Unliked" cluster consisted of three items stating that the person

73

feels unliked by others. The "Self-Conscious" cluster consisted of three items asking if the
person often feels self-conscious.

There were three other items in the "Interpersonal Sensitivity" scale. The item that
read "feeling critical of others" was added to the Hostility cluster. The other two items
were just dropped (i.e., put in the "Residual" cluster).

Wm The "Worry" cluster is defined by only two non-binary items. These items
initially seemed similar in content, but the empirical pattern of correlations was dissimilar
for the two items. When correlations between the "worry" scale and other scales are
corrected for attenuation, many of the estimated correlations are larger than 1.00. So the
initial "Worry" cluster does not work.

The two Worry items are :

149. SCL-DEP Worrying too much about things.

128. BDI I am so worried about my physical problems that I cannot think
about anything else.

The problem seems to be with the BDI item. It is specific to physical problems,
whereas many people are more worried about social or personal problems. Furthermore,
there is an element of obsessiveness in the item that is inconsistent for an item written as a
depression item rather than as an anxiety item. This cluster was dropped from further
analysis.

The Final Item Clusters
The final confirmatory factor analysis began with 49 item clusters. The items for

each cluster are listed in Appendix A. The binary format GTS items are those with

74

numbers 201 to 290. In cases where a content cluster would have included items from the
GTS and items from the non-binary scales, that cluster was split so that the binary items
were put in a separate cluster. Because there seems to be no way to solve the binary
response set problem in these data, all clusters using the binary format were reluctantly
dropped from further analysis.

Clusters 45 through 48 consisted of NA items which were not included in the
previous clusters of the same names, because of differences in response set between the
Likert and binary formats. Cluster 45, “Irritability (GTS),” consisted of two NA items
describing extreme irritability. Cluster 46, “Worry (GTS),” consisted of three NA items
which describe frequent or excessive worrying. Cluster 47, “Fear (GTS),” consisted of
two NA items characterized by fearfirl apprehension. Finally, Cluster 48, “Sleep
Disturbance (GTS),” consisted of one NA item: “I often have difficulty sleeping because
of my worries.”

Cluster 49, “Alienated,” consisted of three items from the Interpersonal Sensitivity
scale of the SCL-90, characterized by feeling easily hurt and that others are unfiiendly or
unsympathetic. The 29 symptoms retained for firrther analysis are listed and described in
Table 9.

The correlations among symptoms can be computed either using the items as
scales or using confirmatory factor analysis to produce construct correlations. The
confirmatory factor analysis construct correlations are the same as the scale correlations

corrected for attenuation due to random error of measurement: i.e., the same as the scale

75

Table 9

Number of Items and Reliabilities for Symptoms Retained for Further Analysis

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SYMPTOM NUMBER OF ITEMS RELIABILITY"
Fear 1 1 89
Motor Tension 10 85
Autonomic Hyperarousal 17 92
Vigilance 3 77
Irritability 2 51
Insecure 2 47
Worry 2 35
Dysphoric Mood 5 80
Fatigue 4 73
Sleep 1 100
Appetite 2 34
Dependency 8 65
Insufficiency 2 59
Need for Approval 5 48
Self-Critical 6 73
Anhedonia 6 76
Self-Esteem 5 64
SelfiEflicacy 3 56
Unstable 2 60
Guilt 2 50
Hostility 6 80
Hopelessness 2 67
Suicidality 2 73
Self-Conscious 3 72

 

 

76

 

 

 

 

 

 

 

Table 9 (cont’d).
SYMPTOM NUMBER OF ITEMS RELIABILITY
Rosenberg Self-Esteem 10 85
Mastery 10 71
Shame 10 91
Chang-Hunter Guilt 9 76
Alienated 3 76

 

 

 

 

 

"' All decimals omitted.

correlations corrected using the alpha reliabilities.

The symptom correlation matrix will be presented later since it is easier to consider
after the symptoms have been arranged by syndrome.

Smptoms and Synfdromes

The next step in the analysis was to see how the symptoms related to each other.
The key question is this: Are there two syndromes corresponding to anxiety and
depression? The second key question is this: Are the other characteristics claimed to be
measures of depression really the same as depression or are they separate syndromes or
traits in their own right?

An earlier generation of psychologists was dominated by psychometricians who

believed that the dimensionality of a set of variables could be determined by a method now

77

called "exploratory factor analysis". However in recent years, most psychometricians have
joined earlier critics and come to the belief that dimensionality is better determined by a
method that is now called "confirmatory factor analysis".

Both exploratory and confirmatory factor analysis were applied to the symptom
correlations to determine the number and identity of syndromes. The exploratory factor
analysis fit the name "exploratory"; it suggested rough syndrome categories but those
categories had to be revised to meet the more stringent criteria set by confirmatory factor
analysis.

The following segment will describe the steps taken to form the final syndromes.
The results of the final syndrome analysis will then be presented.

Exploratogz factor analysis

Exploratory factor analysis is often called just "factor analysis" in many articles,
especially older articles. By far the common method of exploratory factor analysis is a
two step process: (a) a principal axis factor analysis done with communalities, (b)
followed by VARIMAX rotation.

The beginning symptom correlation matrix can take one of two forms: (a) the
correlations among symptom scales or ”(b) the correlations among symptom constructs.
Both analyses were done and proved to have similar results. The exploratory factor
analysis of the symptom construct correlation matrix is presented in Table 10.

Two factors identified symptom clusters that were very similar in content. Two

factors identify clusters with only slight modifications. Two factors identify specific

 

[Mr \ .hru \ mpr\ \ k S 9|. HI. a . .\
«a. r... \ . . .11 Mn .
. Y .r ARK l a 1 VI“

The Exploratory Factor Analysis of the Symptom Construct Correlation Matrix.

78

Table 10

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SYMPTOM l 2 3 4 5 6 7 8 9 10
Autonomic Hyperarousal 85* ~8 37 15 8 ~1 18 4 ~8 6
Motor Tension 84" ~18 35 26 14 10 14 3 ~5 1
Fear 66"I ~27 45 32 20 18 l ~7 7 ~7
Vigilance 65* ~15 13 46 22 25 12 ~6 l6 ~1
Hostility 62" ~23 38 30 12 ~3 14 21 ~6 16
Self-Esteem (Rosenberg) ~l 1 90" ~26 ~5 ~22 1 ~12 1 ~7 ~5
Mastery ~16 90" ~22 ~13 ~25 ~7 ~5 ~7 l ~l
Self-Esteem ~30 72* ~32 ~30 ~12 2O ~9 6 38 ~1
Shame 19 ~70* 17 28 33 16 9 20 ~3 0
Self-Critical 21 ~54* 52 35 34 l 2 12 7 ~9 3
Guilt 31 ~27 89* 16 12 2 10 16 ~2 6
Appetite Disturbance 42 ~26 85" 15 ~12 ~7 33 ~2 ~13 10
Suicidality 40 ~28 75* 21 3 ~9 -5 2 -1 ~25
Anhedonia 35 ~26 64* 39 20 ~6 16 2 5 24
Hopelessness 29 ~41 54* 44 30 10 15 ~18 8 12
Alienated 47 ~18 25 74* 19 22 8 8 ~7 ~9
Dysphoric Mood 42 ~18 40 67* 27 10 l 1 ~14 ~2 ~1
Initability 41 ~25 39 66“ 30 13 l3 l6 ~2 19
Fatigue 43 ~22 36 55* 22 10 2O 1 0 42
Self-Conscious 47 ~29 1 5 50" 20 3O 22 l 8 ~16 ~7

 

 

 

 

 

 

 

 

 

 

 

 

 

' Identifies symptom’s highest positive or negative correlation.

All decimals omitted.

 

79

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 10 (cont’d).
SYMPTOMS I 2 3 4 5 6 7 8 9 10
Unstable 9 ~20 ~2 21 91 " 24 6 1 3 8 ~4
Insufliciency 30 ~33 1 l 25 76* 33 9 4 ~S 5
Insecure 14 ~44 23 12 74‘I 32 4 -7 ~6 10
Need for Approval 9 ~3 ~5 14 23 92* O 14 5 ~7
Dependency 11 ~18 18 18 41 78* 4 ~12 ~19 14
Efiicacy ~2 45 ~3 ~3 l 1 62* ~7 3 41 1
Sleep Disturbance 24 ~14 15 15 8 ~l 67* ~1 ~1 3
Guilt (Chang-Hunter) 13 ~27 20 5 5 33 ~6 41"I 3 0

 

 

‘ Identifies syrnptom’s highest positive or negative correlation.
All decimals omitted.

symptoms that seem to have no clear syndromal meaning. Two factors have no large
correlates. However, there are two factors where the statistical clusters are not well
defined by considering only the highest loadings. The discussion below considers the
simplest cases first.

No symptoms had their highest correlations with Factors 9 or 10. On the other
hand, the factor analysis for 10 factors is clearer than the analysis for 8 factors. In this
case, these "extra" factors seem to reduce the impact of sampling error.

Factor 8 has a modestly high correlation only with the trait Guilt construct. Trait

80

Guilt does not show a pattern in common with any other symptom.

Factor 7 has a modestly high correlation only with Sleep Disturbance. This
symptom has weak correlations with other factors.

Factor 2 has very high correlations with the five symptoms related to Shame or
Self-Esteem. These symptoms are used to define a syndrome of Shame (to focus on the
negative end).

Factor 5 has very high correlations with the three symptoms related to feelings of
Insecurity. These symptoms are used to define a syndrome of Insecurity.

Factor 6 has very high correlations with the symptoms that tap feelings of Need for
Others (i.e., Need for Approval and Dependency). Feelings of self-efficacy also correlate
highly with Factor 6, but Efficacy clearly differs from the two measures of Need for
Others in that Efficacy has a high correlation of +.45 with the Self-Esteem factor (Factor
2) while the Alienation measures correlate -.03 and ~.18. Thus, only the two symptoms
that tap Need for Others are used to define a syndrome.

Factor 1 has very high correlations with the symptoms most closely identified with
anxiety: i.e., Autonomic Hyperarousal, Motor Tension, Fear (or apprehension), and
Vigilance. Hostility also correlates highly with Factor 1, but is clearly different in content.
The four conventional symptoms are used to define a syndrome of Anxiety. The symptom
of Hostility is so important that it is used by itself to define a syndrome.

Factors 3 and 4 are complicated. Both factors have very high correlations with

symptoms associated with depression. Examination of the firll set of factor loadings

81

shows that neither factor lines up with clearly delineated clusters. This seems to represent
a cases where exploratory factor analysis rotates to a misleading position.

The two symptoms that tap feelings of Alienation (feeling Alienated and Self~
Conscious) both have high correlations with Factor 4, though they line up at opposite
ends. On the other hand, consideration of the full pattern of loadings shows that the
Alienation symptoms have a very different pattern of correlations than do the other three
symptoms that correlate highly with Factor 4; namely Dysphoric Mood, Irritability, and
Fatigue. These three symptoms are classic symptoms for depression. In particular, these
three symptoms all correlate much more highly with Factor 3 (the other Depression factor)
than do the Alienation symptoms.

All five symptoms that correlate highly with Factor 3 are classic symptoms of
depression. However, while Anhedonia and feelings of Hopelessness are specific to
depression, the other three symptoms are also fi'equently found in other disorders.
Furthermore, while the firll pattern of loadings is extremely similar for state Guilt, Appetite
Disturbance, and Suicidality; the other two symptoms are not similar. This is especially
clear in regard to the other depression factor, Factor 4. Anhedonia and feelings of
Hopelessness correlate much more highly with Factor 4 than do Guilt, Appetite
Disturbance, or Suicidality. Close examination of the symptoms highly correlated with
Factors 3 and 4 suggests three clusters that might define syndromes. First, the two
symptoms that tap feelings of Alienation form a closely knit cluster. Second, the three

symptoms of Guilt, Appetite Disturbance, and Suicidality seem to form a closely knit

82

cluster. Third, the other symptoms more closely tied to the definition of depression can be
put into a cluster: namely Dysphoric Mood, Anhedonia, Fatigue, Irritability, and
Hopelessness. That is, there is one cluster specifically suggested by Factor 4, one
specifically suggested by Factor 3, and one cluster of items that are very highly correlated
with both factors.

Initial Confirmatory Fact—or Analysfi

The provisional clusters suggested by the exploratory factor analysis were tested
by confirmatory factor analysis. The clusters are shown in Table 11.

The confirmatory factor analysis showed that most of these symptom clusters are
unidimensional and can be used to define a syndrome. However there were clear
departures fi'om simple equivalence for both the Anxiety and Depression clusters.

Anxiety subclusters. Close examination of the Anxiety cluster showed that the
four symptoms can be subdivided into two pairs. The two arousal symptoms have much
lower correlations with Depression than do the two cognitive symptoms. Thus for further
analysis, the Anxiety cluster was split into two Anxiety subclusters. Autonomic
Hyperarousal and Motor Tension were put in one cluster while, Fear and Vigilance were
put in the other. Since the arousal symptoms have relatively low correlations with
Depression, that cluster is called "Pure Anxiety" while the other cluster is called "Mixed
Anxiety."

Mon subclusters. Close examination of the Depression cluster showed that

the five symptoms can be subdivided into two subclusters. Fatigue, Irritability, and

83

Table 11

Provisional Clusters Suggested by Exploratory Factor Analysis

 

CLUSTER SYMPTOM

 

Anxiety Autonomic Hyperarousal

 

Motor Tension

 

Fear (apprehension)

 

Vigilance (psychological tension)

 

 

Depression Dysphoric Mood

 

Anhedonia

 

Fatigue

 

Irritability

 

Hopelessness

 

 

Severe Symptoms State Guilt

 

Appetite Disturbance

 

Suicidality

 

 

Alienation Alienated

 

Self-Conscious

 

 

Need for Others Need for Approval

 

Dependency

 

 

Shame Mastery (reverse scored)

 

Rosenberg Self-Esteem (reverse scored)

 

Self-Esteem (reverse scored)

 

Shame

 

 

 

Self-Criticism

 

 

84

 

 

 

 

 

 

 

 

 

Table 1 1 (cont’d).
CLUSTER SYMPTOM
Insecurity Insufficiency
Unstable
Insecurity
Hostility Hostility

 

 

Hopelessness have relatively lower correlations with Anxiety than do Dysphoric Mood
and Anhedonia. Thus for fiirther analysis, the Depression cluster was split into two
subclusters. Fatigue, Irritability, and Hopelessness were put in one subcluster, while
Dysphoric Mood and Anhedonia were put in the other. Since the energy symptoms have
relatively low correlations with Anxiety, that cluster is called "Pure Depression" while the
more directly mood related cluster is called "Mixed Depression".

The 10 Working Clusters.

The subdivision of the large Anxiety and Depression clusters produced a set of 10
provisional clusters to be considered as syndromes. This set of clusters was tested using
confirmatory factor analysis. The factor loadings for the confirmatory factor analysis
using the symptom construct correlations are reported in Table 12. This analysis was
extended to include the three symptoms that were not used in the definition of the
confirmatory factor clusters. The results for the unused symptoms are shown in Table 13.

Examination of the results shows good fit for the unidimensionality of each cluster.

85

Table 12

The Confirmatory Factor Loadings for Construct Correlation Matrix
Defined by 10 Symptom Clusters.*

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Factor/ CLUSTERS CONFIRMATORY FACTORS
Pure Mixed Mixed Pure Shame Need for Severe insecurity Hostility Alienated

Anxiety Ninety w W Others 8y“
Pure Anxrety
MOTOR 98 95 85 79 57 29 74 47 79 81
TENSION
AUTONOMIC 98 82 77 71 47 15 73 34 78 69
HYPERAROUSAL
Mixed Anxiety
FEAR 85 90 86 81 65 37 76 54 76 81
VIGH-ANCE 77 9O 79 76 48 47 49 54 63 82
Mixed Depression
DYSPHORIC 73 87 87 92 62 36 69 59 65 88
MOOD
WWW 7O 73 87 88 67 12 84 49 74 73
Pure Depression
FANGUE 74 80 94 94 63 35 65 55 68 8O
WARM“ 75 84 99 101 71 42 71 65 85 92
HOPELF—SSNESS 65 81 99 85 78 35 76 63 63 72
Shame
SELF-ESTEEM 43 so 59 62 99 28 54 63 48 55
(ROSENBERG)
MASTERY 35 47 55 58 9O 18 54 55 42 42
SELF-ESTES“ 44 58 65 67 85 41 49 69 52 65
SHAME 59 S7 73 68 89 5 68 45 62 67
SELF-CRITICAL 59 68 84 86 86 34 75 69 65 73

 

 

 

 

 

 

 

 

 

 

 

' All decimals omitted.

 

86

Table 12 (cont’d).

Factor/ CLUSTERS CONFIRMATORY FACTORS

Pure Mixed Mixed Pure Shame Nwd for Severe Insecurity Hostility Alienated
Anxiety Anxiety w burn-i- Others 9mm-

Need for Others
NEED/ APPROVAL
DEPENDENC Y
Severe Symptoms
GUILT
SUICIDALITY
APPETITE.
Insecurity
UNSTABLE
INSUFFICIENCY
INSECURE
Hostility
HOSTILITY
Alienation
ALIENATED

SELF—CONSCIOUS
‘ All decimals omitted.

 

87

Table 13

The Confirmatory Factor Loadings for the Symptoms Not Used to Define Clusters in the
Analysis of the Symptom Construct Correlation Matrix Defined by 10 Symptom

 

 

 

 

 

 

 

 

 

 

 

 

Clusters.*

CONFIRMATORY RESIDUAL (i.e., unused) ITEMS
FACTORS Sleep Efficacy Guilt
Pure Anxiety 44 ll 30
Mixed Anxiety 40 ~5 37
Mixed Depression 47 13 37
Pure Depression 46 12 39
Shame 36 42 44
Need for Others 10 ~53 47
Severe Symptoms 41 29 31
Insecurity 25 ~12 53
Hostility 42 15 33
Alienation 42 1 41

 

 

 

 

 

 

‘ All decimals omitted.

88

That is, the symptoms within each cluster follow the pattern required for statistical
equivalence.

Since the clusters delineate unidimensional sets of symptoms, the confirmatory
factor model shows good fit. Thus, essentially all of the information in the symptom
profile is captured by the 10 confirmatory constructs. These will be provisionally called
"syndromes" for short. The correlations between syndromes are shown in Table 14.

One correlation shows a problem in differential validity: the estimated correlation
of 1.04 between the two Depression subclusters. The fact that this correlation is larger
than 1.00 is due to sampling error in the estimation process. However, if we revise that
estimated correlation down to 1.00 it still poses a problem. Is there any difference
between the two Depression subclusters?

There are two ways in which the subclusters differ. First, they differ in their
correlations with Anxiety. For Pure Anxiety, the correlations are .77 for the Pure
Depression factor and .83 for the Mixed Depression factor (See Table 15). For Mixed
Anxiety, the correlations are .87 for the Pure Depression factor and .92 for the Mixed
Depression factor. Second, the two factors differ in the extent of correlation with Need
for Others, where the correlations are .40 for the Pure Depression factor and .28 for the
Mixed Depression factor.

So there is some evidence for differentiation between the two depression
subclusters. If they are different, then it would mean that the estimated correlation of 1.04

is a sampling error departure from some number smaller than 1.00. This is certainly

89

Table 14

Correlations Between Syndromes“

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SYNDROME 1 2 3 4 5 6 7 8 9 10
1. Pure Anxiety 100 91 83 77 53 23 75 42 80 77
2. Mixed Anxiety 91 100 92 87 63 47 69 60 78 91
3. Mixed Depression 83 92 100 104 75 28 89 62 80 93
4. Pure Depression 77 87 104 100 76 40 76 65 77 87
5. Shame 53 63 75 76 100 29 67 67 6O 67
6. Need for Others 23 47 28 40 29 100 7 67 17 54
7. Severe Symptoms 75 69 89 76 67 7 100 30 71 60
8. Insecurity 42 6O 62 65 67 67 30 100 42 62
9. Hostility 80 78 8O 77 60 17 71 42 100 73
1‘0. Alienation 77 91 93 87 67 54 60 62 73 100

 

" All decimals omitted.

 

possible with the sample size for this study though there is no way to estimate just how
high the population correlation might be.

If we assume a two dimensional structure for the four subclusters, it is possible to
generate an algebraic estimate of the correlation between the Depression clusters from the
other five correlations. That estimate is 1_' =94. This value is within the confidence
interval around 1.04 and is thus consistent with the data.

There are some other very high correlations in this table as might be predicted
fiom the exploratory factor findings. There is a very high correlation of .92 between

Alienation and Mixed Depression. There is a very high correlation of .89 between the

90

Severe Symptoms factor and Mixed Depression.

Is there a difference between Mixed Depression, Alienation, and the Severe

Symptoms factors? One sharp difference is in how the three factors relate to Need for

Others (See Table 15).

Table 15

Correlations Between Need for Others
and Mixed Depression, Severe Symptoms, and Alienation

 

 

 

 

SYMPTOM Need for Others
Mixed Depression .28
Severe Symptoms .07
Alienation .54

 

 

 

 

The difference between r = .54 for Alienation and r = .07 for the Severe Symptoms
factor is very large. Furthermore, while both Alienation and the Severe Symptoms factor
have very high correlations with Mixed Depression, they have a much lower level of
correlation with each other: ; = .60.

There is another difference between the Severe Symptoms factor and Mixed
Depression. The Mixed Depression factor shows a large difference in its correlations with
Pure Depression (nominal ; = 1.04; estimated [ = .94) and with Pure Anxiety (I = .83).
The Severe Symptoms factor shows little difference in its correlations with Pure

Depression (I = .76) and with Pure Anxiety (1 = .75).

91

The Smptom Correlations

Of the 28 symptoms that could be measured in this study, 25 were used to define
syndromal clusters and 3 were not used. Thus, the symptom correlation matrix is a bulky
25 x 25 in the primary symptoms are considered and a bulky 28 x 28 if all symptoms are
considered. The symptom correlations are much easier to look at if the symptoms are
organized. The good fit for the confirmatory factor model shows that the 10 working
clusters provide a good way to organize the clusters.

Symptoms can be correlated at one of two levels. The items in each cluster can be
used to generate a scale to measure that symptom. These scale correlations are attenuated
by error of measurement in each cluster. The key to reducing error is to use many items.
This was not possible in this study since so many different symptoms were considered.
For small clusters, the scale will have low reliability and the scale correlations for that
scale will be very much attenuated.

Confirmatory factor analysis can be used to estimate the size of the correlation
between symptom constructs themselves. These are the correlations that are assumed
mentally when we think about "the correlation" between two symptoms in terms of theory
or clinical practice. That is, confirmatory factor analysis estimates the size of the
correlation that would be obtained from a perfect measure of the symptom: i.e., the
correlation if we could measure each symptom with a very long scale. These construct
correlations are the same as the scale correlations corrected for attenuation using the scale

reliabilities.

92

Table 16 presents the symptom correlation matrix organized by clusters. Table 18
presents the symptom scale correlation matrix; i.e., the size of the correlations for the
imperfect symptom measures in this study.

Scatter plots of the total scores for anxiety and depression for each individual
revealed an extreme skew which can not be discerned from the correlational data. This
skew was more extreme for the anxiety scores than for the depression scores. It is
possible that we are looking at one end of a normal distribution. If so, a more balanced
scale, in which both ends of the continuum were represented, would show a clearer picture
of the actual constructs. Unfortunately, in the current study, those items which might best
describe the positive end of a continuum moving from fatigue to energy are best
represented by items from the GTS, which is wrought with measurement problems. It is
not clear at this point whether anxiety can best be described as a continuum with
contentment at the opposite pole, or whether anxiety may be better conceptualized as a
unipolar construct in which contentment, or relaxation, is merely the absence of anxiety.
Anxiety Driven Depression

The van Praag (1994) theory of serotonin driven depression predicts that high trait
anxiety will produce high trait depression. However, because of the effect of serotonin
dysregulation on hostility and aggression, his theory also predicts an equally strong
relationship between hostility and depression. This prediction can be tested using multiple

regression. The multiple regression analysis for the present data is presented in Table 18.

93

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 2 8 8 8 8 8 8 8 8 8 8 8 Ma 8 8 2. 2. z 8 8 2. 2. 2. 8 2. 8 8 326.3 .8
z 2. 8 8 8 8 8 8 a 8 8 8 a :. 8 8 2. 2. 8 8 8 8 8 8 8 8 8 8 8:8 .8
8 8 8 8 8 8 8 2. 8 8 8 8 8 8 2. z. 2. 2. 8 8 8 8 8 8 8 8 8 8 38.8.28 .8
8 8 8 8 8 8 8 .8 .8 8 8 8 8 8 2. 2. 2. 8 8 8 8 8 .8 8 8 8 8 8 tan: .8
8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 9.3863882.» .8
8.38

2 2 8 8 8 8 8 8 8 2. 2. 2. 8 8 8 8 8 8 8 2. 8 8 8 8 8 8 8 8 all}: .8
8 . 3 2. a 8 8 2. 8 8 8 z. 8 8 8 8 8 8 8 8 8. 8 2. 8 2. 2. 8 8 £335 .8
z o 8 8 2. 8 8 8 8 8 8 2. 8 8 2. 8 8 8 8 8 8 8 8 8 2. 2. 2. z. .88.. .:
zoanuuaouui

8 2 8 8 8 2. 8 8 8 a 2. 8 8 o 8 8 8 8 8 8 2. 8 2. 2. 8 2. 8 2. g .8
8 . 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 2. 2. 2. 2. 8 2. 8.: 38.1.8 .2
zoaganaaa

8 z. 8 2. 2. 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 2. 2. 8 . 2. 8 8 8 8 88.25 .8
8 a 8 8 8 2. 8 8 8 8 2. 2. 8 8 2. 8 8 8 8 8 2. 2. . 2. 2. 8 8 2. 8 a... ._
Egan:

8 2 8 8 8 2. 8 8 a 8 8 8 2 2 8 8 8 8 8 8 8 2. 8 8 8 2. 8 8 1358:3583 .n
8 a 8 8 8 8 8 8 8 2. 8 2. 8 8 8 8 8 8 8 8 8 z. 2. 2. 8 8 8 8 82.3.88: 8
Egg

8 8 z 8 8 8 8 2 8 2 8 8 2.. 2 8 8 8 8 8 8 n : 8 8 v _ n « 528255.30

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2282 228—280 88885 aneim

$282233. .8 2202.53

2 035.

 

94

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 8. : 8 8 8 8 8n .8 8 8 8 .8 2.8 8 8 8 .8 8 8 88 8 8 8 8 8 8 8... 2.85353 .88
8. . 8. 2 a 2 2. 2.. 8.. 8 8 2 2. 8. 2 2. 8 8 8n 2 8 a 2 8 2. n 2 8 8852.8 .8
.. 8. a 8 8 8 8 8 .w 8 8 8 8. 8 .8 8 8 8 8 8 .8 8 8 .8 8 8 8 88 asp-32......» .8..
53982.

8 8 8 8. 82 8 8 8 .8 8 2.8 8 8 8 8 8 8 8 8 8 2. 8 8 8 2. 8 8 2. nan-8.2.8 8
2 2. 8 8. 8. 8 8 8 8 82 8 8 8 8 8 8 8 8 2.8 8 8 2. 8 8 2. 8 8 2. 3.8.2 .8
2058882.?

8 2 8 8 8 8. 8 ...8 8 2. 88 8 8 8 8 8 8 8 8 8 8 8 82. 8 .8 2. 2. 2. 2.22.18. .8
22.588:

8 2. 8 8 8 8 8. 8 8 8 8 8 . 2. 28 8 8 8 8 8 8 8 8 8 8 8 8 8 8 58a. .8
8 2.. 8 8 8 8 8 ... 8 8 8 88 82. 8 8 8. 82. 8 8 8 .2. 8 8... 8 8 8 2.8 8 22.-18...... .2
8 ... 2 8 8 8 8 8 8. 8. 2.. 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 835...: .88
22.22582...

8 8 8 8 82 2. .2 8 8. 2: a 8. 2 n. 8 8 8 8 8 2.2. 2. 8 8 8 8 8 8 2. 8.1.323 .2
8 8 8 8 8 8.. 8 8 2.. .2 2. 8 8 8. 8 8 2.8 8 8 2. 8 8 22. 8 2.8 82. 8 8 2.33.8 .8
8 2 8 8 8 8 8 88 8 8. 8 8 2.. 2. 8. 8 8 8 8m 2. 82. 2. 8. 8 8 2. 8 .2. 2.6.18 .8
88023822826»

.8 2. 8. 8 8 8 2. 82. 8 2 8 2.. 8 8 8 8 8 8 8 8 8 88 8 8 8 8 2 8 8382.2 .2..
8 8.8 8 8 8 2 8 8 8 8. 8. 2. 8 8 8 ... 8 8 8 8 8 8 2. 8 8 8 2 8 388.8823: .2
8.25.29.23.22

88 8 8. 8 8 8 8 2 8 2 .n 8 2.. 2 8 8 8 8 .8 8 n 2 8 2 8 . n 8 332.82.23.33

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6.28. o. 28...

 

95

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3 2 3 3 3 3 3 3 3 3 3 3 3 2 3 3 3 3 3. 3 3 3 .3 3 3 3 3 3 3304...“. .8
2 3 3 3 3 3 3 3 2 3 3 3 2 o. 3 .3 3 3 3 3 3 3 3 3 3 3 3 3 and .3
3 2 3 3 3. 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 .3 3 3 3 3 3855.. .3
3 3 3 3 3 3 3 3 3 3 3 3 2 3 3 3 3 3 2. 3 3 3 3 3 R 3 3 3 ban: .3
3 z z 3 3 3 3 3 3 3 3 3 3 2 3 3 2. 2. 3 3 3 3 3 3 3 3 3 3 938863835.» .3
3.33

2 a 3 3 3 3 3 3 3 3 3 3 3 .: 3 3 3 3 3 3. 3 3 3 3 3 3 3 3 3.9.2.: .3.
3 3 3 3 3 3 3 3 3 3 3 3 3 2 3 3 3 3 .3 3 3. 3 3 3 3 3 3 3 33.33 .3
3 3 3 .3 3 3 3 3 3 3 3 3 3 2 3 3 3 3 3 3 3 3 3 3 3 3 .3 3 ...-u... .:
293.533:

2 2 3 3 3 3 3 3 3 3 3 3 2 o 3 3 3 3 3 3 3 3 3 3 3 3 . 3 3 383.2 .3
z 3 3 3 3 3 3 3 3 3 3 3 2.. 2 3 3 3. .3 3 3 3 3 .3 3 3 3 3 3 3.23338 2
20335938:

3 .. 3 3 3 3 3 3 3 3 3 3 3 3 .3 3 3 3 3 3 3 3 3 3 3 3 3 3 8.335 .3
3 3 3 3 3 3 3 3 3 3 8 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 .... ._
Egan:

: 3 3 .3 .3 s 2 3 2 3 3 3. z o 3 3 3 3 3. 3 .3 3 3 3 3 3 3 3 1:253:32 .n
3 o 3 3 3 2. 3 3 .n 3 3 3 3 z 3 3 3 on 3 3 3. 3 3 O S F 3 8 833.584 .u
tugs—2

3 3 z 3 3 3 3 2 3 2 3 3 2 2 3 3 .3 3 3 .3 2 3 2 3 2 3 . .n u it;

 

 

 

 

 

 

€232.82 Sm 362.50 875
5.53.4 noun—0300 2.8m Boa—gm

E 035.

 

96

 

on

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ n. 2 3 3 3 on R 3 2 3 3 3 3 2 2 3 on 3 3 3 3 3 3 3 3 3 2.33.383 .3
a. . : _ o : 3. ... a. 2 2 2 8 2. 2 3 2 3 3 g a. 3 2 n a. u a . b.8553 .3
2 : 2. 3 3 3 2 3 z 3 3 3 : n 3 3 3 3 3 .3 3 3 3. 3 an 3 3 3 8.333993» .3.
.5393

a _ 3 8 3 2 2 av. 3 3 3 3 3 3 3 3 «a 3 3 3 3 3 3. 3 3 3 a. 3 31.883 .3
a o 2 8 3 2 .n 3 .n 3 3 3 3 3 a 3 3 3 3 3 3 3 a c. 3 3 3 3 3:32 .3
29.5533

3 : 3 2 2 8. 2 on 3 3 3 3 2 n a 3 3 3 3 3 3 a 3 2 on 3 8 3 $39: .3
E33:

3 3. 2 on .n a .2 3 3 2 3 3. 3 3 3 «n 3 3. 3 3 a 2.. 3 3 3 3 R 3 a 23.... .3
a 2.. 3 3 3 2 3 2 a 2 3 3 3 3 3 2.. 3 3 3 3 an 3.. 3 3 .7 3 3 3 33 .2
2 a. z 3 z 2 e a 8 n. : z a 3 a 2 3 3 3 3 an 3 3 3 a 3 .2 3 .335 3
E5082.

2 2 3 2 3 3 2 2 n. 3 3 3 o u. 3 2 .3 3 3 3 8 3 3 3 3 3 3 3 again; .2
2 2 3 n 3 3 8 3 : 3 a a 3 _. z 3 3 3 3 3 3 8 .9. 8 3 8 3 3 5333 .3
3 2 3 a 3 3 3 3 z 3 a 3 2 n a S 3 3 3 3 3 3 a 3 on 3 3 3 5.6.16 .3
”20.53533”

.2 2. : a 2 : c 3 z o 3 s 3 e 3 z 3 2 3 3 3 8 2 3 a on 2 3 38.3899 .2
a 2. n 3 3 n 3 3 3 n. _. n 3 3 2 o. 3 3 2 2 2 2 o 2 3 3 3 2 13.52.2382 .2
gueauz

3 3 z 2 3 3 3 2 3 2 3 3 2 2 3 3 3 3 3. 3 n : 3 2 v . ... 3 Egg

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

€253 : 2.3

 

97

Table 18

Correlations Between Pure Anxiety, Hostility, and Pure Depression'“

 

 

 

 

 

Pure Anxiety Hostility Pure Depression
Pure Anxiety 100
Hostility 80 100
Pure Depression 77 77 100

 

 

 

 

 

‘ All decimals omitted.

Multiple regression:

Beta for anxiety: .43
Beta for hostility: .43

Multiple correlation: .81

The multiple correlation is consistent with the prediction derived fi'om the van
Praag (1994) theory. The correlation between Hostility and Pure Depression is as high as
the correlation between Pure Anxiety and Depression. The beta weight for Hostility is just
as high as that for Pure Anxiety. The multiple correlation increases from a high .77 to a
still higher .81.

Results Pertaining to a One-Factor Model

One Higher Order Factor?

Some advocates of the Big Five factor model have argued that symptoms
considered in this study are largely expressions of a single factor called “neuroticism” or

“negative affectivity.” This hypothesis can be tested statistically at two levels. Those tests

98

will be considered in this segment. The results will show that considerable information
would be lost if the symptoms were replaced by one global factor.

The strongest test of the one-factor model would consider the prediction of
variables outside the domain of neuroticism. No such variables were measured for this
study, and so the strong test of this model cannot be performed here. However, weaker
tests can be done which consider just the structure of correlations between items, between
symptoms, or between syndromes. It will be shown that the one factor model fails using
even these weaker tests.

The FlaAOne-Factor Model

Many researchers have argued that the symptoms collected for this study are
largely expressions of one underlying factor, traditionally called “neuroticism.” This
model can be tested statistically.

The flat model makes no distinction between symptoms. Thus, the unit of analysis
for this model is the item. That is, the flat model predicts that each item differs only by
random error from the single dimension of neuroticism. If this model were true, then all of
the symptom clusters would be equivalent to one another. For any two symptoms, the
population construct correlation would be 1.00. The average sample construct correlation
would thus be 1.00. Half of the sample correlations would be expected to be larger than
1.00, and half would be expected to be smaller.

The symptoms construct correlation matrix was presented in Table 14. There is

one sample correlation between Mixed Depression and Pure Depression which is 1.04.

99

All other correlations are less than 1. The correlation between Need for Others and the
Severe Symptom cluster is only .07. The average correlation between symptoms is only
.65; far less than 1.00.

The flat factor model is clearly disconfinned.

The Hierarchical One-Factor Model

Big Five theorists have a more sophisticated model which is hierarchical in
structure. This model acknowledges that items in the domain of neuroticism form clusters
such that the correlations within each cluster are larger than would be predicted by the
general factor of neuroticism. The factor defined by each cluster is called a “facet.” The
theorists acknowledge that facets have specific variation above and beyond variation due
to the general factor. However, they argue that the only significant variation in the facets
is the variation due to the general factor.

In the data for this study, there are actually two lower levels considered, not just
one. First, items can be collected in clusters that define symptoms. Second, the symptoms
can be clustered to form syndromes. There is no recognition of this double level of
structure in the current Big Five literature.

The hierarchical factor model can be tested at either level. That is, the concept of
“facet” can be considered where “facet” denotes “symptom” or where facet denotes
“syndrome.” The model was tested at both levels and the results follow.

Test at the symptom level. Suppose that the concept “facet” is considered to mean

“symptom.” In this study I identified 28 measurable symptoms. If the word “facet” is

100

defined by symptoms, then the hierarchical one factor model would claim that the
symptom correlation matrix should be explained by one factor. That is, if all significant
variation in facets is explained by neuroticism, then the size of the correlations between the
facets must be explained by the general factor. This hypothesis can be tested using the
data from the current study.

Consider first the implications for the factor analysis of the symptom correlation
matrix. According to the one factor model, the symptom correlation matrix should have
only one common factor. The exploratory factor analysis reported in Table 10 found 10
factors, where 6 factors had high correlations with more than one symptom. That
disconfirrns the one factor model.

It is mathematically possible for the multiple factor structure in the data to be due
to sampling error. In clinical psychology and psychiatry, there are some studies with
samples so small that this becomes a real concern. In this study, the sample size was I)! =
366, which makes this concern unreasonable. However, this remote possibility can be
tested statistically by using confirmatory factor analysis to do a significance test on the
departures from the predictions of the one, factor model. This was done using the
symptom scale correlation matrix. The deviations from the one factor model generated a
chi square value of 33340.53 with 377 degrees of freedom. This is equivalent to a _2_ value
of 108, and is significant at far beyond the .000001 level. Thus, the one factor model is
totally disconfirmed at the symptom level.

The deviations from the one factor model are statistically massive and totally

lOl

beyond question. Are the deviations substantively significant? In this segment, the
substantive meaning of the deviations is not considered. However, it is possible to
consider simply the size of the deviations. Are the deviations large enough to be of
practical significance? An abstract way to approach this question is through partial
correlation. The one factor model is fitted to the data and a correlation is computed
between each symptom and the general factor. The factor correlations can then be used to
compute partial correlations. If the one factor model were true, then each such partial
correlation would be 0 if computed for population data. The sample partial correlation -
would differ from 0 only by sampling error.

The partial correlations were computed for the symptom scale correlation matrix.
Ifthe population partial correlation is 0, and the sample size is E = 330, the standard error
of the sample population correlations is .055. The probability of a sample correlation as
large as .20 is only .00014, or 1 in 7143. Yet, among the 378 symptom partial
correlations, there were 76 correlations larger than .20, of which 24 were larger than .30,
10 were larger than .40, and 5 were larger than .50.

Lipsey and Wilson (1993) compiled meta-analyses across 18,000 treatment studies
in psychology, and found an average treatment correlation of r = .23. By that standard,
76 of the 378 partial correlations were large enough to be of practical significance. Thus,
considerable information would have been lost if the symptoms were replaced by a single

neuroticism factor.

102

Test at the syndrome level. Examination of the content names for the facets used
in the current Big Five inventories shows names like anxiety, depression, self-
consciousness, etc. Thus, the facets currently used are closer in level to syndromes than
to symptoms. The one factor model can also be tested at the syndrome level.

If syndromes are the “facets” of the hierarchical model, then the syndrome
correlation matrix should be explained by one factor. That is, if all significant variation in
facets is explained by neuroticism, then the size of the correlations between the facets
must be explained by the general factor. This hypothesis can be tested using the data from
this study.

Consider first the implications for the factor analysis of the syndrome correlation
matrix. According to the one factor model, the syndrome correlation matrix should have
only one common factor. To test this hypothesis, the syndrome scale matrix was
computed, and confirmatory factor analysis was used to fit a one factor model to the data.
The deviations from the one factor model generated a chi square value of 357.35 with 44
degrees of freedom. This is equivalent to a ; value of 33.40, which is significant at far
beyond the .0001 level. Thus, the one factor model is disconfirrned at the syndrome level.

The deviations from the one factor model are statistically massive and beyond
question. Are the deviations large enough to be of practical significance? An abstract
way to approach this question is to compute the partial correlations between syndromes
with the general factor held constant. If the one factor model were true, then each such

partial correlation would be 0 if computed for population data. The sample partial

103

correlation would differ from 0 only by sampling error.

The partial correlations were computed for the syndrome scale correlation matrix.
If the population partial correlation is 0, and the sample size is E = 330, the standard error
of the sample partial correlations is .055. The probability of a sample correlation as large
as .20 is only .00014, or 1 in 7143. Yet, among the 45 syndrome partial correlations,
there were 12 correlations larger than .20, two of which were larger than .30.

Using the average treatment correlation of .23 found by Lipsey and Wilson (1993),
12 of the 45 partial correlations were large enough to be of practical significance. Thus,
considerable information would be lost if the syndromes were replaced by a single
neuroticism factor.

The one factor model does not fit the data. The flat item model fails; the
hierarchical model using symptoms as facets fails; and the hierarchical model using
syndromes as facets fails. Thus, no substantive construct was found that corresponds to
the concept of “neuroticism.”

On the other hand, the correlations between syndromes are all positive: some
quite large. So it is possible to define mathematical general factors which are composites
of the neurotic syndromes. Since the one factor model does not fit the data, the
mathematical factor is not uniquely defined and the exact content of the factor will vary
somewhat from one study to another, depending on the specific set of syndromes
considered in any given study. However, the neuroticism factors defined in different

studies will be very highly correlated with one another.

104

This mathematical neuroticism factor provides a useful first approximation to the
data and may be all that is needed in some applications. A person who is high on the
mathematical neuroticism factor will tend to be high on all of the symptoms; however,
that individual will be higher on those symptoms more highly correlated with the
neuroticism factor than on those symptoms less highly correlated with the neuroticism
factor.

Since the one factor model does not fit the data, there may be no substantive
neuroticism factor as such. On the other hand, there may be a substantive neuroticism
factor which is only one cause of the high correlations between certain symptoms.
However, there are now many research findings that show that many of the high
correlations between symptoms and syndromes are due to causal rather than structural
relations. In view of that fact, researchers should exercise extreme caution in using
“neuroticism” as an explanatory variable. Use of such terminology must be regarded as

speculative at this point.

CHAPTER 6

DISCUSSION

In this study, I looked at clusters of symptoms associated with anxiety and
depression in an attempt to better understand commonalities and distinctions between
these disorders. The literature suggests that depression should be linked with dysphoric
mood, low self-esteem, indecisiveness, hopelessness, psychomotor retardation, anhedonia,
feelings of guilt or worthlessness, and suicidal ideation. Those symptoms most specifically
linked to anxiety have been feelings of worry, anxiety, disquietude, and irritability; fears of
dying or being out of control; somatic symptoms, such as shortness of breath, dizziness,
chest pains, aches, and an exaggerated startle response; depersonalization and
derealization; and shyness. Those symptoms linked to both depression and anxiety include
appetite disturbance and abdominal distress, sleep disturbance, fatiguability, concentration
difficulties, and agitation.

The current findings support some of these expected links, but not others.
Depression was linked most specifically to fatigue and hopelessness, as anticipated, but
also linked to dysphoric mood and anhedonia. Anxiety was linked most specifically to
somatic symptoms of motor tension and autonomic hyperarousal, and also to fear and
vigilance. This replicates previous findings linking somatization to anxiety, but not to

depression (King, Margraf, Ehlers, & Maddock, 1986). Contrary to expectations,

105

106

irritability and worry were linked to depression rather than anxiety. These findings will be
discussed in greater detail in ensuing sections, with one exception; because of the
problems associated with the Worry cluster, the findings regarding worry should be
considered as tentative and will not be elaborated.

It is notable that each syndrome was best distinguished by two clusters: one
psychological and one physiological. Anxiety was best distinguished first by physiological
symptoms associated with activation of the arousal system, and second by psychological
symptoms associated with fear and vigilance. Depression was best distinguished by
physiological symptoms associated with decreased energy, and by psychological
symptoms associated with dysphoric mood and irritability.

Mtive Affectivity: A General Distress Factor?

At this point, we can begin to answer empirically questions about some of the
models which have been presented. The findings from this study were in many ways
highly consistent with those found by Watson and colleagues (1995b) ‘in their factor
analysis of symptoms of anxiety and depression. However, it should be noted that neither
my results nor their results were consistent with their conclusions. In the Watson and
colleagues (1995b) study, the authors interpreted the high correlations between symptoms
of anxiety and depression as evidence of a general distress factor. However, their results
showed no evidence for a separate general distress factor. Rather, the symptoms that they
placed in the General Distress category each follow one of two patterns: the pattern for

the anxiety items or the pattern for the depression items. In terms of those items which

107

were most discriminative, Watson and colleagues (1995a) found that those items best
distinguishing anxiety fiom depression were symptomatic of “Anxious Arousal.” Items
which Watson and colleagues (1995a) claimed best distinguished depression from anxiety
were best characterized by “Loss of Interest” and “High Positive Affect.” Their
interpretation of depression as anhedonia led them to incorporate the High Positive Affect
items into their design. However, their findings show that the High Positive Affect items
are no more highly correlated with loss of interest than with the other depression items. A
perusal of the factor loadings of the items used by Watson and colleagues (1995b) showed
that their “General Distress” items were better markers of either anxiety or depression
than they were for a general distress factor.

A content analysis of Watson and colleagues’ (1995b) items revealed a great deal
of overlap across the categories that they formed (See Table 19). This is not surprising as
the categories were formed on the basis of lists of items thought to be linked to the
disorder in clinical practice, rather than on the basis of either a content analysis or their
empirical loadings.

The “Anxious Arousal” factor appears to be the cleanest, consisting of 12 items
associated with autonomic hyperarousal, four items associated with motor tension, and
one item associated with fear. The “Loss of Interest” category had four Anhedonia items,
two items measuring Fatigue, one Suicidality item, and one Shame item (measuring Self-
Criticism).

The “General Distress: Anxious Symptoms” category had four items measuring

108

Table 1.9

Content analysis of the Items Considered in the Watson and Colleagues (1995b) Study.

 

SCALE / Item from Watson et al., 1995b

Symptom Category from Current Study

 

GENERAL DISTRESS: MIXED SYMPTOMS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Worried a lot about things Worry

Trouble concentrating Cognitive Disturbance

Felt dissatisfied with things Self-Critical

Felt confused Cognitive Disturbance

Felt irritable Irritability

Trouble making decisions Cognitive Disturbance

Trouble paying attention Cognitive Disturbance

Felt restless Motor Tension

Felt something awful would happen Fear

Got fatigued easily Fatigue

Trouble remembering things Cognitive Disturbance

Trouble falling asleep Sleep Disturbance

Trouble staying asleep Sleep Disturbance

Loss of appetite Appetite Disturbance

Slept very well (-) Sleep Disturbance
GENERAL DISTRESS: DEPRESSIVE SYMPTOMS

Felt depressed Dysphoric Mood

Felt discouraged Self-C ritical

Felt sad Dysphoric Mood

Felt hopeless Hopelessness

Disappointed in myself Self-Critical

Felt like crying Dysphoric Mood

Felt like a failure Self-Critical

Felt worthless Self-Esteem (-)

 

 

109

 

Table 19 (cont’d).

 

SCALE I Item from Watson et al., 1995b

Symptom Category from Current Study

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

GENERAL DISTRESS: DEPRESSIVE SYMPTOMS (Cont’d)
Blamed myself for things Guilt
Felt inferior to others Self-Esteem
Pessimistic about the future Hopelessness
Felt tired or sluggish Fatigue
GENERAL DISTRESS: ANXIOUS SYMPTOMS
Felt tense: “high strung” Vigilance
Felt uneasy too vague to categorize
Felt nervous Vigilance
Felt afraid Fear
Felt “on edge,” keyed up Vigilance
Unable to relax Vigilance
Lump in my throat Autonomic Hyperarousal
Upset stomach Autonomic Hyperarousal
Tense or sore muscles Motor Tension
Felt nauseous Autonomic Hyperarousal
Had diarrhea Autonomic Hyperarousal
LOSS OF INTEREST
Felt unattractive Self-Critical
Felt that nothing was enjoyable Anhedonia
Felt withdrawn from others Anhedonia
Took extra efl'ort to get started Fatigue
Felt slowed down Fatigue
Nothing was interesting or fun Anhedonia
Felt bored Anhedonia
Thought about death, suicide Suicidality

 

 

 

110

 

Table 19 (cont’d).

 

SCALE / Item from Watson et al., 1995b Symptom Category from Current Study

 

ANXIOUS AROUSAL

 

Felt dizzy, lightheaded

Autonomic Hyperarousal

 

 

 

 

 

 

 

Was trembling, shaking Motor Tension
Shaky hands Motor Tension
Trouble swallowing Autonomic Hyperarousal
Short of breath Autonomic Hyperarousal
Dry mouth Autonomic Hyperarousal
Twitching or trembling muscles Motor Tension
Hot or cold spells Autonomic Hyperarousal

 

Cold or sweaty hands

Autonomic Hyperarousal

 

 

 

 

 

 

 

 

 

 

 

 

 

Felt like I was choking Autonomic Hyperarousal

Felt faint Autonomic Hyperarousal

Pain in chest Motor Tension

Racing or pounding heart Autonomic Hyperarousal

Felt numbness or tingling Autonomic Hyperarousal

Afraid I was going to die Fear

Had to urinate frequently Autonomic Hyperarousal

Easily startled Autonomic Hyperarousal
HIGH POSITIVE AFFECT

Felt really lively, “up” ’ Dysphoric Mood (-)

Felt really happy ‘ Dysphoric Mood (-)

Felt I had a lot of energy ‘ Fatigue (-)

Was having a lot of fun ‘ Anhedonia (-)

 

Felt I had much to look forward to ’

Hopelessness (-)

 

Felt good about myself ‘

Self-Esteem

 

 

I had many interesting things to do ‘

 

Anhedonia (-)

 

 

111

 

Table 19 (cont’d).

 

SCALE / Item from Watson et al., 1995b

Symptom Category from Current Study

 

HIGH POSITIVE AFFECT (Cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Felt confident Self-Esteem
Looked forward to things ‘ Hopelessness (-)
Felt I had accomplished a lot ‘ Self-Critical (-)
Was proud of myself ‘ Self-Critical (-)
Felt cheerful ' Dysphoric Mood (-)
Felt successful Self-Critical (')
Felt optimistic ' Howlessncss (-)
Felt really talkative Anhedonia (-)
Moved quickly and easily ‘ Fatigue (-)
Felt hopeful about the future ' Hopelessness (-)
Able to laugh easily Dysphoric Mood (-)
Felt like being with others Anhedonia (-)
Felt very clearheaded Cognitive Disturbance (-)
Thoughts came to me very easily Cognitive Disturbance (-)
Felt very alert Fatiguc (-)
Could do everything I needed to Anhedonia (-)
Felt I didn’t need much sleep Fatigue (-)
(-) Reverse scored.

' Selected as reverse keyed item for the Anhedonic Depression scale by Watson et al, 1995b.

 

112

Autonomic Hyperarousal, one item measuring Motor Tension, four items
measuring Vigilance, one item measuring Fear, and one final item (“felt uneasy”) that was
too vague to categorize effectively.

The “General Distress: Depressive Symptoms” category had five items measuring
Shame (three measuring Self-Criticism and two measuring low Self-Esteem), three items
measuring Dysphoric Mood, two measuring Hopelessness, and one measuring Guilt.

The “General Distress” category had 15 items: five measuring Cognitive
Disturbance, two measuring Depression (one Fatigue item and one Irritability item), one
measuring Shame, one measuring Worry, one measuring Appetite Disturbance, and one
measuring Sleep Disturbance.

They also had a category that is radically different in content from the symptoms
usually listed for anxiety and depression: the “High Positive Affect” category (which they
also termed “Anhedonic Depression”) had 14 items. These items were designed to
measure feelings of happiness, energy, and positive self-regard. The items can also be
classified as the bipolar opposites of some of the symptoms for depression. Looking at the
scale this way produced the following analysis: two items measuring Anhedonia, three
measuring Fatigue, three measuring Hopelessness, three measuring Dysphoric Mood, and
two measuring Shame (i.e., low Self-Esteem). Thus, of the 14 “Anhedonic Depression”
items, only two items specifically consider Anhedonia; the other 12 tap different
dimensions.

Watson and colleagues (1995a) ultimately reverse scored the “High Positive

113

Affect” items and combined them with the “Loss of Interest” items to form a measure
intended to be specific to depression and separate from General Distress. However,
careful examination of the factor loadings shows that there is no basis for this
combination. The remaining depression items correlate just as highly with the Positive
Affect factor as do the “Loss of Interest” items. Furthermore, there is only a modest
correlation between the “High Positive Affect” item construct and the depression item
construct (r = -.50). However, it is possible that that correlation may be lowered due to
nonlinearity. The real correlation may in actuality be much higher.

Because strong markers of both anxiety and depression were included in the
General Distress category, it is not surprising that it correlated highly with both anxiety
and depression factors. However, invoking a “General Distress” category does little to
tease out the salient components of depression and anxiety, once arousal and anhedonia
have been factored out. A content analysis of the items within the categories supports the
contention that there are two major categories of symptoms within the category of
anxiety: the first characterized by primarily physiological symptoms associated with
Autonomic Hyperarousal and Motor Tension, the second characterized by primarily
psychological symptoms associated with Fear and Vigilance. In my study, when I split the
original Anxiety cluster into two separate clusters, depending on the level of correlation, I
found that the physiological symptoms defined one factor, which I termed “Pure” anxiety
symptoms, whereas the psychological symptoms defined a second factor, which I termed

“Mixed” symptoms of anxiety.

114

The construct of Depression appears to be somewhat more complex. When I
broke down the Depressive symptoms into those which were relatively “Pure” versus
those which were more “Mixed” markers of depression, I found that my results, in this
case, were largely at odds with those of Watson and colleagues (1995b) in that their
analysis did not distinguish between anhedonia and fatigue. To the contrary, my best
markers of “Pure” Depression were the clusters of Fatigue, Irritability, and Hopelessness,
whereas the superordinate cluster of “Mixed” Depression contained Dysphoric Mood,
Anhedonia, and Worry. Notably, in the current study the high correlation between the
categories of “Pure” and “Mixed” Depression appears to be due largely to the high
correlations between the various variants of dysphoric mood: Dysphoric Mood, Wony,
and Irritability, rather than the relatively lower correlations between Fatigue and
Anhedonia. Results fi'om the current study do not support Watson and Clark’s
conceptualization of depression.

Consistent with my own data, Watson and his colleagues (1995b) found that items
associated with Shame, such as self-criticism and low self-esteem, tended to correlate
more highly with items associated with depression than with items associated with anxiety.
However, including these items within their “General Distress: Depressive Symptoms”
may be a poor solution. My findings support a correlation between Shame and
Depression (r = .76); however this correlation is low enough that the two constructs may
best be conceptualized as distinct and separate.

The results fi'om the current study suggest that there are meaningful distinctions

115

between subjective experiences of anxiety and depression in spite of their similarities.
These differences may be obscured when “General Distress” is invoked as a means for
explaining those similarities.

Confirmaticmnd Disconfinnation of Clalrrs

The “Interpersonal Sensitivity” Scale

 

The “Interpersonal Sensitivity” scale of the SCL-90 purports to measure what has
often been called “rejection sensitivity” in the literature. This construct is often depicted
as an aspect of depression (Blatt, 1974), and therefore could be presumed to correlate
more highly with depression than with anxiety. However, the “Interpersonal Sensitivity”
scale is not homogeneous. A content analysis of its items shows that three concede
feelings of being unliked or “Alienated” from others. Three of the items may best be
described as “Self-Conscious.” One item described feeling critical of others, consistent
with items describing “Hostility.” The remaining items did not clearly fit with any of the
other symptoms, and were not included in further analyses.

The “Interpersonal Sensitivity” scale was also not internally valid; the content did
not fit the title. If this had been a valid measure of the construct, we would have expected
higher correlations between “Interpersonal Sensitivity” and Depression than Anxiety. The
actual differences found were modest: .85 versus .73. Depictions in the literature of
rejection sensitivity as an important correlate of anaclitic depression (Blatt, Quinlan, &
Chevron, 1990) would lead us to predict higher correlations between “Interpersonal

Sensitivity” and Need for Others than Insecurity. This would be consistent with previous

116

findings that link the anaclitic style to feelings of being unliked or unpopular (Blatt, Hart,
Quinlan, Leadbetter, & Auerbach, 1993). These expectations were not born out;
“Interpersonal Sensitivity” was actually more strongly associated to Insecurity (the
current study cluster composed of introjective items; 1 = .59) than to Need for Others
(the cluster composed of anaclitic items; r = .47). We would also have expected to find
higher correlations between the “Interpersonal Sensitivity” scale and Irritability and
Hostility than were actually found (Becker & Lesiak, 1977).

Taken together, these findings suggest that this scale is not a good measure of the
construct. Irritability may be a better measure of rejection sensitivity than the symptom
clusters contained within the “Interpersonal Sensitivity” scale of the SCL-90.

Hostilig and Irritability

High correlations between anxiety, depression, and correlates of anger, such as
hostility and irritability, were found in the current study. These findings support previous
work in which significant correlations of these three constructs were found at both the
trait (Mook, van der Ploeg, & Kleijn (1990) and state levels (Gotlib & Meyer, 1986;
Zuckerman, Persky, Eckman, & Hopkins, 1967). The high correlations found in the
current study are more consistent with correlations reported at the state level, which tend
to range between .67 and .87 (Gotlib & Meyer, 1986; Zuckerman, Persky, Eckrnan, &
Hopkins, 1967), than with those found at the trait level, which tend to be somewhat lower
(Mook, van der Ploeg, & Kleijn, 1990). These high correlations are consistent with

expectations predicted by models of depression which focus on the long-term effects of

117

stress (Gold, Goodwin, & Chrousos, 1988), and may be most indicative of specific
subtypes of depression in which anxiety is an integral factor in the development of the
disorder (Weiss et. al, 1981).

Correlates of anger, such as hostility and irritability, have been found to be highly
linked to depression (Becker & Lesiak, 1977; Biaggio & Godwin, 1987). A closer
analysis of the data shows that, in spite of the strong association between Hostility and
Irritability (r = .85), their patterns of correlations show them to be distinct cOnstructs that
have distinctive links to other affective experiences. In the current study, Irritability was
associated more highly with Depression (1 = .97) than with Anxiety (1 = .78). In contrast,
Hostility was linked similarly to both Depression (1; = .73) and Anxiety (I. = .76).
Irritability was associated with greater distress across all factors except Mixed Anxiety;
Mixed Anxiety correlated .73 with Irritability and .80 with Hostility.

Quit

The guilt items in the current study were not homogenous; two items had a
distinctly different pattern of correlations than the other nine. An analysis of these items
showed that the two distinct items were taken from the BDI. The instructions for this
instrument ask the respondent to answer in terms of their experience over the past week.
In contrast, the other items were taken from the Chang-Hunter Guilt scale, in which the
items are depicted as general statements. In this way, it is likely that the first two guilt
items depict a more transient, or state, measure of guilt, whereas the other nine depict a

more enduring, or trait, measure of the construct.

118

In contrast to expectations that guilt would be more highly linked to depression
than to anxiety, neither state nor trait Guilt distinguished well between Anxiety and
Depression. In the current study, the factor describing symptoms of trait Guilt was
correlated .30 with Pure Anxiety and .34 with Pure Depression. By contrast, state Guilt
was correlated .69 with Pure Anxiety and .77 with Pure Depression. This large contrast is
consistent with previous findings which have suggested that state guilt measures may not
be good representatives of more general, or trait, guilt (Charles & Levine, 1995).

Shame, Depression and Anxieg

Many theorists have linked shame to depression. The current results support a
stronger link between Shame and Pure Depression (r = .71) than Pure Anxiety (I = .50).
However, it is noteworthy that what I have termed “mixed” symptoms of the constructs
do not discriminate well; these correlations are .68 and .62 for Depression and Anxiety,
respectively. Shame, therefore, may be linked to specific aspects of the depressive
experience, or it may be that somatic symptoms are less relevant to the experience of
shame than are more psychological experiences of distress.

Anaclitic and Introjective Dimensions

For the purposes of the current study, I selected a subsample of the items which
Blatt listed as loading most strongly on each factor, which did not also load significantly
on another factor. Using this method produced an “Anaclitic” scale which consists
entirely of the items in the Dependency factor, consistent with Blatt’s (1974) depiction of

the anaclitic dimension as “dependent.” However, it should be noted that the two clusters

119

which make up this factor in the current study are Dependency and Need for Approval.
This is contrary to Blatt’s (1974) original conceptualization of these two styles, in which it
is the introjective style which is linked to need for approval.

In contrast to the homogeneity of the “Anaclitic” scale, the “Introjective” scale
that I derived is more diverse. It is composed of all of the items in the Insecurity factor,
which included the clusters Unstable, Insufficiency, and Insecure; plus Mo items from the
Self-Critical Cluster; and four items which were difficult to classify and were not included
in the content analysis. Two of the omitted items were considered to be too vague for
inclusion in the content analysis, but were consistent with a self-critical theme (“There is a
considerable difference between how I am now and how I would like to be” and “I tend
not to be satisfied with what I have”). Two other omitted items were difficult to classify
at all (“No matter how close a relationship between two people is, there is always a large
amount of uncertainty and conflict” and “I tend not to be satisfied with what I have”).

The heterogeneity of this scale is consistent with Blatt’s diffuse depiction of the
introjective style as a mixture of self-dissatisfaction, ambivalence toward others, and
failure to meet expectations.

Because of the heterogeneity of the Introjective scale and the contamination of the
Anaclitic scale with a predominant theme fi'om the original conceptualization of the
introjective style, there are severe problems of construct validity in regard to these two
scales. Researchers who are trying to understand relationships between dependency and

self-criticism and other constructs would be advised to look carefully at the actual content

120

of these scales before attempting to use them as a measure of those constructs.
Current Anxiety and Depression Scales

We can also begin to evaluate the instruments from the study which are used to
measure anxiety and depression. For example, are they overweighting relevant items or,
perhaps weighting irrelevant items? To address this question, I looked more closely at the
anxiety and depression measures utilized in the study to ascertain how well they sample
the symptoms which appear to be most relevant to these constructs (See Table 20).

A Content analysis of the BAI shows that two thirds of the items are linked to
Pure Anxiety, and only one third are linked to Mixed Anxiety. That is, the BAI consists of
11 items from the Autonomic Hyperarousal cluster, 3 items associated with Motor
Tension, 5 items associated with Fear, and 2 items associated with Vigilance. Perusal of
the SCL—Anxiety scale shows that half of the items are characteristic of Pure Anxiety, and
half are linked to Mixed Anxiety; this scale consists of 1 item from the Autonomic
Hyperarousal cluster, 4 items fi'om Motor Tension, and 5 items associated with Fear.

A content analysis of the BDI shows that 19% of the items are linked to Pure
Depression; more specifically, there are 2 symptoms of Fatigue, 1 of Irritability, and l of
Hopelessness. In addition, 33% of the items are linked to Mixed Depression; 2 of
Dysphoric Mood and 4 of Anhedonia. Twenty-three percent of the items are linked to
Severe Symptoms; 2 of Guilt, 2 of Suicidality, and 2 of Appetite Disturbance. The
remaining 23% fall into other categories; 1 of Worry, 4 of Self-Criticism, and l of Sleep

Disturbance.

Number of Items from Each Cluster
Represented in Anxiety and Depression Instruments.

12]

Table 20

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BA] SCL- BDI SCL-
Anxiety Depression

PURE Autonomic

ANXIETY Hyperarousal l l 1
Motor Tension 3 4

MIXED

ANXIETY Fear 5 S 2
Vigilance 2

MIXED

DEPRESSION Dysphoric Mood 2 3
Anhedonia 4 3
Worry 1

PURE

DEPRESSION Fatigue 2 2
Irritability 1
Hopelessness l l

SEVERE

SYMPTOMS Guilt 2 l
Suicidality 2 l
Appetite
Disturbance 2

SHAME Self-Criticism 4
Selchsteem 1
Sleep

RESIDUAL Disturbance 1 l

 

 

 

 

 

 

 

122

A content analysis of the SCL-Depression scale shows that the items in this scale
are not as representative of the relevant symptoms as the items in the BDI. Only 23% of
the items were linked to Pure Depression; more specifically, 2 of the items were
symptoms of Fatigue and 1 of Hopelessness. In addition, 38% of the items were linked to
Mixed Depression; 3 of Dysphoric Mood and 3 of Anhedonia. Fifteen percent were
linked to Severe Symptoms; 1 of Guilt and 1 of Suicidality. The remaining 28% were
linked to other categories; 1 of low Self-Esteem, l of Sleep Disturbance, and 2 of F ear.

Examination of the clusters which define the anxiety and depression factors in the
current study allows a more specific evaluation of how representative the scales are of the
symptoms which appear to be most relevant. A content analysis suggests that the anxiety
scales are more representative of the relevant syndrome than are the depression scales.

Smptoms of Anxiety

The present findings regarding symptom clusters associated with anxiety were
consistent with previous factor analytic studies in which psychological and somatic
clusters were found (Cloninger, 1988b; Buss, 1960). Notably, the symptom clusters of
Motor Tension and Autonomic Hyperarousal, which I have termed “Pure Anxiety,” were
consistent with what Cloninger (1986; 1988b) described as somatic anxiety, and the
symptom clusters of Fear and Vigilance which I termed “Mixed Anxiety” were consistent
with what Cloninger (1986; 1988b) described as cognitive anxiety.

Somatic symptoms were not as highly linked to dependency as would have been

123

predicted by previous findings (Beck, Epstein, & Harrison, 1983). This link may be more
likely to be found in clinical studies which are able to focus differentially on distinct
subtypes of anxiety.

Drug trials support the view that somatic and psychological symptoms of anxiety
should be seen as distinct; Benzodiazapines, which affect a subsystem of the gamma-
aminobutyric acid (GABA) system, were effective in treating somatic symptoms and in
decreasing hypervigilance. In contrast, imipramine, which blocks reuptake of
norepinephrine and serotonin affected psychological symptoms, such as anxiety,
interpersonal sensitivity, anger-hostility, paranoid ideation, and obsessive compulsive
symptoms (Hoehn-Saric, McLeod, & Zimmerli, 1988).

Heterogeneig in the Construct of Depression

The current data suggests that the construct of depression is quite heterogeneous.
This is in striking contrast to the symptoms of anxiety, which were (1) more easily
described using relatively few factors, and (2) better represented by items in existing
instruments. Depression was represented most clearly by the symptoms of fatigue,
hopelessness, and irritability. Symptoms of depression with higher correlations with
anxiety included dysphoric mood and anhedonia, often considered to be th_e defining
symptoms of depression. In addition, the cluster which I have termed “Severe Symptoms”
was largely composed of symptoms associated with depression, such as guilt, appetite

disturbance, and suicidality. Notably, the factor I termed “Alienation,” composed of items

associated with rejection sensitivity, was also highly correlated with symptoms of Mixed

124

Depression and Severe Symptoms. These symptoms were not included as part of a larger
mixed factor because of striking differences in their links to other constructs under study.
Notably, Dependency was strongly linked to Alienation (r = .54), and not at all linked to
Severe Symptoms (r = .07). The link between Dependency and Mixed Depression fell
between these two extremes (r = .28). i

It is notable that Dysphoric Mood, a symptom linked conceptually very strongly to
depression, was not as strong a marker of this syndrome as were other symptoms. This
finding is consistent with suggestions in the literature that depression may present,
particularly in adolescence, without dysphoric mood as traditionally conceptualized
(Nurcombe et. al., 1989). There are numerous indications in the literature that there are
distinct subtypes of depression which may vary considerably in terms of predominant
affective experience. For example, van Praag (1994) suggests the existence of a distinct
subtype of depression characterized primarily by anxiety and/or aggression deregulation as
, the primary symptoms, associated with diminished serotonergic metabolism. Within this
syndrome, depressed mood is seen as a derivative, rather than a primary, symptom.

Anxieg and Depression

The correlation between Pure Anxiety and Pure Depression is a high .77.
However, epidemiological studies typically find an asymmetry in the relationship between
them. Patients who are first diagnosed with an anxiety disorder with no severe symptoms
of depression are often later given a diagnosis of depression (Angst, Vollrath, Merikangas,

& Ernst, 1990). On the other hand, patients who are first diagnosed as depressed without

125

severe symptoms of anxiety are rarely later diagnosed with anxiety disorders (Angst et al.,
1990)

Does the regression plot for Pure Anxiety and Pure Depression show any similar
asymmetry? The contingency table for Pure Depression as a function of Pure Anxiety is
shown in Table 21. Note that this table is constructed using scale scores which are not
perfect measures of the two constructs. The error of measurement produces a blurring of
position in this table. In particular, the columns show a larger variation than would be

shown for a perfect measure of Pure Depression.

Table 21

Contingency Table Relating Pure Depression and Pure Anxiety

 

 

 

 

 

 

 

 

 

PURE ANXIETY
P o 1 2 s 4 Total
U
R 5 1 1
E
4 6 1 1 1 9
D
E 3 5 5 3 4 1 18
P
R 2 34 26 10 1 71
E
g 1 76 14 2 92
1
o o 122 4 126
N
Total 237 55 16 6 3 317

 

 

 

 

 

 

 

 

 

 

126

Close examination of this table shows that as anxiety goes up, there is always an
increase in depression. That is, this table is consistent with the findings that high levels of
anxiety tend to lead to high levels of depression at a later point in time.

On the other hand, consider the data for 0 anxiety. There is a considerable range
for depression. The highest levels of depression are not seen for this level of anxiety, but
very high levels are seen in the sample. Furthermore, the lack of observations of high
depression must be considered in relation to the fact that the onset of depressive disorder
tends to be later than onset for anxiety. In this college population, there are many students
who will become severely depressed later in life. These may be the missing cases of
extreme depression for the low anxiety group.

Thus, there are cases of high depression even for people with no anxiety. This is
the asymmetry. High anxiety always accompanies high depression, but high depression
can occur without high anxiety. This asymmetry is consistent with research regarding
anxiety driven depression. Because at the state level anxiety produces depression, if there
is anxiety, then there will be depression (Barlow, 1991). In these cases, serotonergic drugs
have been found to be therapeutic. However, many people do not respond to drugs which
affect the serotonergic system, suggesting that these individuals are depressed for reasons
other than anxiety (van Praag, 1994). Those individuals who are low in anxiety and high
on depression are those who tend to report depression first. That group tends to remain
stable over time. These findings are not consistent with the learned helplessness findings,

and suggest that there is at least one type of depression other than that which is anxiety

127

driven.
Conclusions as to the Measurement of Anxietmd Depressipp

The results of this study in part support recent work (Watson et. al, 1995a; 1995b)
which suggests that vigilance/arousal best distinguishes anxiety whereas irritability and
psychomotor retardation best distinguish depression. The findings do not support the
utility of invoking a general distress factor to explain differences between the two
syndromes. Relatively high correlations do not tell the whole story; in spite of
commonalities between the negative affects, this study points to clear and potentially
meaningful distinctions between these symptoms, as well.

Future research with clinical and older adult populations will help to elucidate
whether the current findings will generalize beyond a non-clinical young adult sample.
The literature would suggest that anxiety and depression should show greater
difl’erentiation in clinical samples (Nurcombe et al., 1989) and greater comorbidity with
increasing age (Brady & Kendall, 1992).

Overall, the results of the analyses point to validation problems when measuring
complex constructs. There were often no clear distinctions between state versus trait
measure of the same construct, and many of the constructs were erroneously named,
inadequately sampled, or confounded in widely-used instruments.

In light of the variety of treatments available to address symptoms of mood
disorders, including both anxiety and depressive syndromes, it becomes particularly

important to distinguish clusters of symptoms which may have implications for treatment

128

choice. The severity and the seriousness of certain associated symptoms, such as
suicidality, make it particularly important to clarify symptom-treatment links. At this point
in time, there are numerous attempts to reconcile clinical understanding with
neurophysiological research, and to make sense of conflicting findings regarding
psychotherapeutic versus medical interventions (Schore, 1994). This type of integration
of theory and research may help us to focus our understanding toward more efficient and

effective intervention.

APPENDIX A

601-
143.
151.
172.
177.
179.
142.
184.
188.
193.
195.

196.

602-MOTOR TENSION
130.
136.
146.
158.
131.
139.
175.

187.

A

SCL-ANX

SCL-ANX

SCL-ANX

SCL-ANX

SCL-ANX

SCL-DEP

BA1

BA1

BA1

BA1

BA1

SCL-SOM

SCL-SOM

SCL-SOM

SCL-SOM

SCL-ANX

SCL-ANX

SCL-ANX

EA]

129

APPENDIX A

ITEMS BY THEME - ALL SCALES

Suddenly scared for no reason.

Feeling fearful.

Spells of terror or panic.

The thought that something bad is going to happen to you.
Thoughts and images of a fiightening nature.

Feeling of being caught or trapped.

Fear of the worst happening.

Terrified.

Fear of losing control.

Fear of dying.

Scared.

Headaches.

Pains in heart or chest.

Pains in lower back.

Soreness of your muscles.
Nervousness or shakiness inside
Trembling.

Feeling so restless you couldn‘t sit still.

Unsteady.

130

603-MOTOR TENSION (cont’d).

1 91 . BAI
1 92. BAI
603-AU'TONOMIC
132. SCL-SOM
1 56. SCL-SOM
159. SCL-SOM
160. SCL-SOM
162. SCL-SOM
1 55. SCL-ANX
180. BAI
1 81 . BAI

182. BAI

185. BAI

186. BAI

190. BAI

l 94. BAI

197. BA]

1 98. BAI
l 99. BAI
200. BAI
604-VIGILANCE
165. SCL-ANX
1 83. BAI

Hands trembling.

Shaky

ERAROUSA

Faintness or dizziness.
Nausea or upset stomach.
Trouble getting your breath.
Hot or cold spells.

A hunp in your throat.
Heart pounding or racing.
Numbness or tingling.
Feeling hot.

Wobbliness in legs.

Dizzy or lightheaded.

Heart pounding or racing.
Feelings of choking.
Difficulty breathing.
Indigestion or discomfort in the abdomen.
Faint.

Face flushed.

Sweating (not due to heat).

Feeling tense or keyed up.

Unable to relax.

131

604-VIGILANC E (cont’d)

1 89. BAI

605- ABILITY

1 35. SCL-HOST

l 18. BDI
606-m SEQ QB“: X
35. INTRO]
37. INTRO]
607-UPSET

207. NEGAF F
229. NEGAFF
254. NEGAFF

263. NEGAFF

608-TENSE
210. NEGAFF
236. NEGAFF

257. NEGAFF

609-WO Y
149. SCL-DEP

128. BDI

Nervous.

Feeling easily annoyed or irritated.

I don't get irritated at all by the things that used to irritate me.

I never really feel secure in a close relationship.

Often, I feel threatened by change.

I sometimes get too upset by minor setbacks.
I can get very upset when little things don't go my way.
I don't get very upset when things go wrong. (-)

Little things upset me too much.

I sometimes feel ”on edge” all day.
I would describe myself as a tense person.

1 often feel nervous and ”stressed."

Worrying too much about things.

I am so worried about my physical problems that I cannot think about
anything else.

132

610-DYSPHORIC MOOD

140. SCL-DEP
147. SCL-DEP
1 17. BDI
108. BDI

148. SCL-DEP

61 l-FATIGUE

l 37. SCL-DEP
1 7 l . SCL-DEP
1 22. BDI

124. BDI

612-ENERGY

202. ENERGY
205. ENERGY
226. ENERGY
238. ENERGY
245. ENERGY
248. ENERGY
253. ENERGY
256. ENERGY
260. ENERGY
274. ENERGY
279. ENERGY

288. ENERGY

Crying easily.

Feeling lonely.

I used to be able to cry, but now I can't cry even when I want to.
I am so sad or unhappy that I can't stand it.

Feeling blue.

Feeling low in energy or slowed down.
Feeling everything is an effort.
I can't do any work at all.

I am too tired to do anything.

I sometimes rush from one activity to another without pausing for rest.
I lead an active life.

Other people sometimes have trouble keeping up with the pace I set.
I put a lot of energy in everything I do.

I can work hard, and for a long time, without feeling tired.

My pace is usually quick and lively.

Most days I have a lot of "pep" or vigor.

People would describe me as a pretty energetic person.

In my life, 1 would rather try to do to much than too little.

I am sometimes ”on the go" so much that I wear myself out.

I have more energy than most people I know.

People sometimes tell me to slow down and "take it easy."

133

6 3- NTHUSIASM

215. POSTEMP I get excited when I think about the future.

262. POSTEMP I get pretty excited when I'm starting a new project.

235. POSTEMP I am usually alert and attentive.

217. POSAFF People would describe me as a pretty enthusiastic person.

290. POSAFF I am usually pretty excited about the things that I do.

614-SLEEP DISTURBAN C E

123. BD1 1 wake up earlier than I used to and can't get back to sleep.

615-APPETITE DISTURBANCE
125. BDI I have no appetite at all any more.

126. BDl I have lost more than 15 pounds.

616-COGNITIVE DIFFICULTIES
244. NEGTEMP Sometimes life seems pretty confusing to me.

247. NEGTEMP I am sometimes troubled by thoughts or ideas that I can’t get out of my

mind.
617-D PENDENCY
2. ANACL Without support from others who are close to me, I would be helpless.
9. AN ACL The lack of permanence in human relationships doesn't bother me. (-)
20. ANACL I would feel like I'd be losing an important part of myself if I lost a very
close friend.
23. ANACL I often think about the danger of losing someone who is close to me.
38. ANACL Even if the person who is closest to me were to leave, I could still "go it

alone." (-)

134

617-DEPENDENCY (cont’d)

50. ANACL
55. ANACL
65. ANACL
618-INSUFFICIENCY
1 1. INTRO]
l6. INTRO]

Ifsomeone I cared about became angry with me, I would feel threatened
that he (she) might leave me.

After an argument, I feel very lonely.

Being alone doesn't bother me at all.

Many times I feel helpless.

There are times when I feel ”empty" inside.

619-NEED FOR APPROVAL

34.

45.

12.

26.

32.

GZESELE-CRITICISM
1 10.

114.

115.

7.

121.

62.

ANACL

ANACL

ANACL

ANACL

ANACL

BDI

BDI

BDI

INTRO]

BDI

INTRO]

I frnd it very difficult to say "no” to the requests of friends.

I wony a lot about hurting or offending someone who is close to me.

I seldom worry about being criticized for things I have said or done. (-)
I am not very concerned with how other people respond to me. (-)

I constantly try, and very often go out of my way, to please or help people
I am close to.

I feel I am a complete failure as a person.

I hate myself.

I blame myself for everything bad that happens.

I often find that I don't live up to my own ideals or standards.
I believe that I look ugly.

I am very satisfied with myself and my accomplishments.

GZI-ANHEDONIA
133. SCL-DEP

1 50. SCL-DEP

1 1 l. BD1

1 19. BDI

129. BD1

1 20. BDI

622-SELF-ESTEEM

33. SELF -EFF
59. SELF-EFF
60. SELF-EFF
157. SCL-INTP
176. SCL-DEP
623-EFFICACY
1. SELFEFF
15 SELFEFF
24. SELFEFF
6 4-UNS B E
36. INTRO]
58. INTRO]

135

Loss of sexual interest or pleasure.
Feeling no interest in things.

I am dissatisfied or bored with everything.
1 have lost interest in other people.

I have lost interest in sex completely.

1 can’t make decisions at all anymore.

I have many inner resources (abilities, strengths).

What I do and say has a very strong impact on those around me.
I sometimes feel that I am “special.”

Feeling inferior to others. (-)

Feelings of worthlessness. (-)

I set my personal goals and standards as high as possible.
I feel I have many responsibilities 1 must meet.

Other people have high expectations of me.

The way I feel about myself frequently varies: There are times when I feel
extremely good about myself and other times when I see only the bad in
me and feel like a total failure.

Very frequently, my feelings toward someone close to me vary: There are
times when I feel completely angry, and other times when I feel all-loving
towards that person.

136

635-STATE GUILT

1 12. BD1 I feel guilty all the time.

1 13. BD1 I feel I am being punished.

626-AN GER

214. NEGAF F My anger frequently gets the better of me.
233. NEGAFF I often take my anger out on those around me.
283. NEGAF F I sometimes feel angry for no good reason.

627-HOSTILITY

134. SCL-INTP Feeling critical of others.

144. SCL-HOST Temper outbursts that you could not control.
168. SCL-HOST Having urges to beat, injure, or harm someone.
169. SCL-HOST Having urges to break or smash things.

174. SCL-HOST Getting into frequent arguments.

178. SCL-HOST Shouting or throwmg things.

6Z§-NEGATIVE AFFECTIVITY

204. NEGAFF I often experience strong emotions such as anxiety or anger without
knowing why.
281. NEGAF F Things seem to bother me less than most other people. (-)

629-POSITIVE AEEEgflgfljx

201. POSAF F C I have the ability to approach tasks in such a way that they become
interesting or fun.
21 l. POSAFF C I lead a very interesting life.

223. POSAFF C In my life, interesting and exciting things happen every day.

137

629-POSTTIVE AFFECT IVITY (cont’d)

23o. POSAFFC
242. POSAFFC
286. POSAFFC
W
109. BDI

163. SCL-DEP

W

1 16 BDI
138. scuorzp
632-srer-c0Nscrous
167. SCL-INTP

17o. SCL-INTP

173. SCL-INTP
gas-mummy
209. DISINHIB

216. DISINHIB

227. DlSINI—IIB

237. DISINI—IIB

249. DISINHIB

258. DISINHIB

I live a very full life.
It takes a lot to get me excited. (-)

1 often feel lively and cheerful for no particular reason.

I feel that the future is hopeless and that things cannot improve.

Feeling hopeless about the future.

I would kill myself if I had the chance.

Thoughts of ending your life.

Feeling uneasy when people are watching you or talking about you.
Feeling very self-conscious with others.

Feeling uncomfortable about eating or drinking in public.

I often stop in the middle of one activity to start another one.
Before I make a decision I usually try to consider all sides of the issue. (~)

The way I behave often gets me into trouble on the job, at home, or at
school.

I rely on careful reasoning when making up my mind. (-)
I always try to be fully prepared before I begin working on anything. (-)

I am not an "impulse buyer.” (-)

138

W (cont’d)

270. DISINHIB

273. DISINHIB

634-

261. DISINHIB

635-PERSISTENC E
222. DISINHIB
*##(Also in 640)***
636-P YFULNESS
218. POSAFF C
241 . POSAFF C
27 1 . POSAFF C
637- SOCIAL
2 l 9. DISINHIB

234. DISINHIB

250. DISINHIB
252. DISINHIB
267. DISINHIB
275. DISINHIB

278. DISINHIB

282. DISINHIB

SPONSIBIL

When I'm having a good time, 1 don't worry about the consequences.

1 am a cautious person. (-)

I am a serious-minded person. (-)

I work just hard enough to get by.

People would describe me as a pretty enthusiastic person.
1 can make a game out of some things that others consider work.

1 often feel playful around other people.

I believe in strictly playing by the rules. (-)

I greatly dislike it when someone breaks accepted rules of good behavior.
H

I would not use others' weaknesses to my own advantage. (-)

I really enjoy beating the system.

Lying comes easily to me.

I've done a lot of things for which I could have been (or was) arrested.

When I decide things, I always refer to the basic rules of right and wrong.
(-)

I often get out of things by making up believable excuses.

139

637-ANTISOCIAL (cont’d)

284. DISINHIB 1 get the most fun out of things that others consider immoral or illegal.
285. DISINHIB ' 1 would never hurt other people just to get what 1 want. (-)
289. DISINHIB At times I've done some petty thievery.

638-SENSATION SEEKING

213. DISINHIB If 1 had to choose, 1 would prefer having to sit through a long concert of
bad music to being in a bank during an armed robbery. (-)

225. DISINHIB I rarely, if ever, do anything reckless. (-)

228. DISINHIB 1 get a kick out of really scaring people.

231. DISINHIB IfI had to choose, I would prefer being in a flood to unloading a ton of
newspapers from a truck.

240. DISINHIB 1 would much rather party than work.

277. DISINHIB I spend a good deal of my time just having fun.

287. DISINHIB I don't ever like to stay in one place for long.

264. DISINI-IHB I like to show-off.

269. POSAFF C I like to stir up some excitement when things are getting dull.

Q39-DISORGANIZATION

280. DISINHIB Taking care of details is not my strong point.
640- AMBITIOUS
222. DISINHIB 1 work just hard enough to get by. (-)

255. DISINHIB I've been told that I work too hard.

140

641-ROSENBERG SELF-ESTEEM SCALE

67. SE
70. SE
73. SE
74. SE
80. SE
82. SE
84. SE
88. SE
91 . SE
93. SE
64 -MASTER
71 . MAST
76. MAST
78. MAST
79. MAST
81 . MAST
83. MAST
85. MAST
86. MAST
90. MAST
94. MAST

On the whole, I am satisfied with myself.

I feel that 1 have a number of good qualities.

I feel I do not have much to be proud of. (-)

I feel that I am a person of worth, at least on an equal plane with others.
At times I think I am no good at all. (-)

I am able to do things as well as most other people.

I take a positive attitude toward myself.

All in all, I am inclined to think I am a failure. (-)

I certainly feel useless at times. (-)

I wish 1 could have more respect for myself. (-)

If 1 put my mind to it I can learn almost anything.

I find life an endless series of problems with no end in sight. (-)

Most of the time I think that the world is an exciting place to live in.

My work in general is at least as good as the work of the guy next to me.
When I decide to do something, I do it.

I feel that I have no talent whatsoever. (-)

I repeat things continuously to be sure that 1 am right. (-)

When 1 want something, I just sit around wishing I could have it. (-)

I feel that I am able to make decisions.

I am fearful of growing up. (-)

141

643-COOK SHAME SCALE

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

SHAME
SHAME
SHAME
SHAME
SHAME
SHAME
SHAME
SHAME
SHAME

SHAME

I feel like 1 am never quite good enough.

I feel somehow left out.

I think that people look down on me.

Compared to other people I feel like I somehow never measure up.

1 scold myself and put myself down.

I see myselfas being very small and insignificant.

1 say to myself, "how could anyone really love me or care about me?"

I feel defective as a person, as if something is basically wrong with me.
I feel intensely inadequate and full of self-doubt.

I see myself striving for perfection only to continually fall short.

644-CHANG - HUNTER GUILT SCALE (Ifll’T GUILI)

68.

69.

72.

75.

77.

87.

89.

92.

95.

GUILT

GUILT

GUILT

GUILT

GUILT

GUILT

GUILT

GUILT

GUILT

1 often cannot forgive myself for having caused deep pain in those I love
or care for.

I feel horrible for having hostile feelings toward other people.
1 have felt very guilty for letting down those close to me.

It bothers me that I have not done more for my parents or family
members.

I have felt very guilty for not being there when someone close to me
needed me.

Sometimes I cannot forgive myself for how I have treated others.

Sometimes I hurt people I love or care for and feel very guilty about it
afterwards.

When I let my anger out, I often feel very guilty afterwards.

I often feel guilty for being better off than my family members.

142

645-IRRITABILITY (GTS)

220. NEGTEMP

259. NEGTEMP

646-WORRY (GTS)

212. NEGTEMP
239. NEGTEMP

272. NEGTEMP

647- FEAR (GTS)
221 . NEGTEMP

268. NEGTEMP

Small annoyances often irritate me.

I have days that I’m very irritable.

I frequently find myself worrying about things.
1 often worry about things I have said or done.

1 wony too much about things that don’t really matter.

Sometimes I will suddenly feel scared for no reason.

I worry about terrible things that might happen.

648- SLEEP DISIQQANCE (GTS)

251 . NEGTEMP

649- ALIENATED
1 52. SCL-INTP
1 53. SCL-INTP

1 54. SCL-INTP

RESIDUAL ITEMS

141 . SCL-INTP
l 57. SCL-INTP
276. NEGTEMP

251 . NEGTEMP

I often have difficulty sleeping because of my worries.

Your feelings being easily hurt.
Feeling others do not understand you or are unsympathetic.

Feeling that people are unfiiendly or dislike you.

Feeling shy or uneasy with the Opposite sex.
Feeling inferior to others.
Often life feels like a big struggle.

I often have difficulty sleeping because of my worries.

143

W (cont’d)

265. NEGTEMP I am often troubled by guilt feelings.

APPENDIX B

144

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3. a. ... .... ... 2. 9. ..n on 2. 2. 8 2. 2 an an 25288. .2
8. a o. 2. 2. .... 8 3 mm «m .... x 3. 6 .~ on 22.8.25 2.2.58 .2
.n 8. a .m- ... 8 S 8. a em .m :. 2. u. 2. 8 8:35.29 658.2 .2
o. R 8. on. .N. 2. 8 we ... 2 2. 8 em .1... 3. 2. 8:25.25 ..8_m ....
.... .m- 8. 8. 8 as 8. .2 2. 2. 3.. an. .m. R. em. 3.. .aa........... .2
2. ... .u. 8 8. 3. an. 2. n. a or «a. 2. :. .... 2. 183...”. .N.
9. 8 2. «a 3. 8. ... ... S .... x 2. 2. :. «a «a 2.2.: .:
9. 8 9. a... on. ... 8. 8. 2. a. an 2. 2. S 2. 2. 862 6.5.328 .2
x 8. we .2 mm- ... 8. 8. g 8 2. a. 8. 8 8 ea 2.53 .6
mm 6 2 2. n- S 2 E 8. 2 on ..n 2. on 2. 9 .85; ...
mm 8 2 2. s 9. me we 2 8. 2. S 2 mm R 28. ea: .6
.m .m em 3. 9.. 3 mm 2 on 8 8. ..n 2. .m 2. R 2:82.. .6
4m .5 9. an. S. 2.. no a. ..m S ..m 8. 2. we 2. as 2.55.... .m
a. 2. on .m. 2. 2. 2. .... 2. mm 2. 2. 8. we 2. ... 288: 8535 ...
e 2 2. R. ..- .2. 2.... 8 on ..N .m 8 we 8. e6 8 Essaém 6.5222 .m
.N 2 3. 8. 2- 2. ea 8 2. R ea 2. 2. ea 8. 6.. 8.2.6.2262 .m
on 8 2n 3.. 2- 2. ..e e... a. an R 2 ... E 2. 8. 632. .82 ..
2 2 v. 2 a. : o. a m a e m e m N . ”Emmy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2335‘ 58:00 .3 3:on 8.8m =a do.“ .632 5222.50

NN 03am.

 

145

 

16' II.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

VN av cv N. N. a0 on an we ow On or Z. N@ Mb ac 3052285923 .mm
m N _a 0N am. a 7 av mo 2. _N N X.” No hv 50 av E. £3225 .Nm
em 2. 3. 9.... ”N. cm 3 va on an we 3 we 8 mm. ca 3053—232 . _ m
w 7 ON. N. .a ma 0v. 3.. on- _N- NN. 00. MV. Nm. ~N. hm- mm. 492.83 2.28m .2”
mm 5N N be. mm- 3 No .a N@ no. 5 a0 we Nm av 2e. eb§80b< 0339.2 .aN
N _ uh Nv an. NN. am 3 ma em 3. me an no wk. as on .9532 .mN
Nv :4 2 ”N. 5.. am 94 On mm mm 3 mm mm mm :4 3 5.092. .hN
hN oo— om 0N. m T Na no a N 2 av 3. av no 2. as. :50 .oN
2 w. v C. N. m— a C E N_ a— NN o- N v N. 98:22an .N
av n- w. _ m. N. Nv mv en an Nm Na an 3. ON N am 03839 .N

b an. 3. ”N ov a. w. m- o_ v w m- N_ N _ . a. N. 5359:an .MN
@ 7 av. mm- av mv ow. N? M». NM- mm- an- no. mvt 3.- No. 5. Eooamm..=om .NN
mm am g 3. N. mm 3. a on 2.” 0m E. an 8 Nb or «Eaves—2 . _ N
am ow 3 on- am. Nb on va 9.. mm 2. mm cm 3 No m5 3:20.23 .oN
hv N 3 3. NT cm 3 mm av am 3— we mm VN ov «V _c>oaa< 3.“ .302 .a_

G N ”N mm. 5N. S we as 8 mm 5 on. G $4 co 8 5:22.685 .3

2 m— 3 m— N— 2 o— a w h o n v m N _ . -....- NEWER.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.9283 2 63a...

 

146

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

vv vM FM aN- ON. Mb Na 5— OM NM NM Ma ah MM MN. 0M coca5=< .av
MN cN vM Z. _ 7 NM NM MM MM 5M cv NM MM NN vN hN .02—3565 moo—M .Mv
MM 0M MN vM- M _- _v vM_ MM 2. MM MM oM NM MM Mv 2- 13m .vv
2. vN N— M7 :- hv Mv an Mn. MM _M vM MM aN ov ov 20:03 .Mv
MM MT v M- v vM MM bN vs 2: v FM hM v— M- M_ .b:58t.: .Mv
vv CN N. a. v. NM _M . av MM 5M _M Mv NM MN MM MM €3.52 d 93:9 =50 .vv
5M Mv cM NM. NM. 5M 5M Mb hv Mv MM MM OM 5M cM MM 0.008 082M .Mv
MN. NM- “M. «M cM Mv. Mv. MM- _v. MN- MM. _M- MM- ~M- aM. Mv- Cobov @232 .Nv
bN- MM- MN. vv Mv _M- av. MM- Nv- MM. aM- bM- NM- MM. sv- MM- 3.32805 Eoofim-hom ..v
MN 5 _. a MN- MM- MN 2 M_ M _ MN MN N T M- ~ _ SOME—5‘ .ov
N M_ 2 MT 2. N_ N. ON M. o a v N a M 2 335985 .aM
o— b v. M v— a. 2. MT 2 M M M M- N- v. N. Mcfioom cocoa—3M .MM
— 5M N— Mv. MM. 2 M a M- N a M 2. 2 M. v— _a_ooacc< NM
7 MN. M _- Na Mh Mv. Mv. ov- aN- a _ - MM. MN- M T _N- aN- MM- 332% .MM
M. M v _ 7 v7 M _ T 2. MT M- v N- _ M- N. 0365395 .MM
a MM v— Mv. _N- a— 3 N N_ v- OM _N M 2 a. vN 3.22:5 .vM
M _ M ~ 3 M— N— _ _ o— a M h M M v M N — mzmmh

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.5253 mm 23-

 

147

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 a. a. 8 mm 8 2 m. 8 8 a- 2 9 8 E 8. 55823 .2
a 8 a- 8 a. a. R 2 8 b 2- 8 8 S 8 3. 3:855 3880 .2
a 2. on- R a. :. 8. a- m. 8- 8. 8 8 8 R 2 8:35.39 289?. .2
8m 8 a. «a a. 2 8 v a 3. mm. s. z. z 8 z 85925 8%. .2
8- 8- a S. 8. 8. on. 2- a- 8 8 8. 8. 3. mm. 2. .23358 .2
2- 8- 3 mm. «a. w 2. a- 8. 8 8 8m. 8- 2. R. w .328 .2
a. 8 8. 8 8 an a. 2 S a. 8. 8 up 8 8 8 2.2.: .:
8 3 3.- 8 8 8 8 a 8 a. N9 2. 2 ... 8 8 802 0885 .2
8 a 8. a no 8 8 2 8 n. m. a 3 8 2. 8 $53 a
a 8 a. S x 8 8N : 8. 2 mm. 8 8 8 8 8 .35 .w
8" 2 a. S. a. 2 w. 2 8. v mm. 2 mm on 8. 3 .025 s
3. 8 9w 8 8 8 8 2 a a 8- 8 fl 8. S 2. 238.5 .8
8 8 a. 8 8 8 z. «N a w. 8. 8 8 8 2 8 .8ng .n
c. 8 an- 8 8 mm a. 2 S 2 8. 8 8 8 8 8 A35: 85.35 .v
8 8 a. 8 8 8 8 N 8 2- 8. 8 3. a S 2 .aBEém 252.23 ....
8 8 R. 8 2 z. z. e 8 .... «v Q. 8 8 8 um :22: as: .m
z. 8 mm- a 8 3. 8 2 8 u- xx 2. m. 8 8 8 6an 5m ._
8 a 8 on 8 R 8" mm x mm a a on 2 2 2 88.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9:53 8 29¢

 

148

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6v No pm. No 3 av mm 3 an N. no. 3 3 we 3 E 8053285928 .mm
8— Nb an. 3. 3 NV ow _ _ S on- no. on _ no _N on o b=anSm .Nm
Nb 2: an- Vm mw hv Mb 3 av n 7 «c. on 5 3. . no we 80280—082 Am
mm. an- 2: cv- mm. 3... mm- m _. hm- mm an 9»... rm. 5 T N... a. .b_>=oo.€< Eggnog .om
we 3 ow. 2: 3 co on S on m. .m. 3 8 no 2. we .b_>:oo.c< 3.3502 .am
3 mo mm- Va 8. oo 8 fl mm n... No. 2. no 94. mv mm £533 .3
NV 5v 3. ca 8 2: on mm on an. no. mm mm an 3 cm Low—2 KN
ow Mb mm. an 00 an 8. 2 mm a T he. an cw MN 3 2 =30 6N
Z 2 m7 C m_ «N N. 8. o- 3. mm. a mm m. 2 m— «cocoafihom .mu
5 av hm- ow an on mm 2 8. m- an. .m em 3 3 an 03883 .3
mm. m 7 mm m. n 7 mm- m _. .m- m- 2: E. 0.. a T R 2 NM xoaomcmfiom .MN
8. no. 3 3- no. no. no. mm. mm. X. 8. hw. mu. m T _n- _N. Evofiméom .NN
on on on. 3. I. an on a Tm o 7 so. o2 g 2 mm m— £5352 . _N
we hm rm. 8 3 nm em mm on 3. mm. E. 2: on on Na. .32t0..=om .8
—N 3. 2- «w on mm mm 2 3 pm m. _- m— an 8. 3 on 3553‘ 8.. v82 .2
om no me. 2. mv 3 vv 2 no N. 3. Na or 3 2: E. 55.0535 .3
Nm :m on a mu 5N 3 mm vn mm mm a. on 2 3 S NEE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.253 mm 23.

 

149

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 2. 8- 8 8 8 8 2 8 2 8. :- 8 8 8 8 88.22 .8
2 8 2. 8 2 8 8 m 8 2- 8- 8 8 2 8 8 .8555 2.88 .8
8 8 8- a 8 2. 8 8 8 8. 2w 8 8 8 8 8 .8". .2.
8 8 8- 8 8 8 8 z a m 8. 8 2. 8 8 8 .253 .2.
m- 8 2. 8 8 2. 2.. 2 R : 2. n 8 8 8 8 .5352— .8
8 8 2- 8 8 8 8 2 8 m 8. 8 8 8 8 :- 35: a. 2.2.2 86 .8
8 8 8- : 8 2. 8 8 R 8. 2.. 8 8 8 2. 8 382258 .8
21 2v 8 21 2.. 8. R- 8. 21 8 8 8- 2.- 8. 8- 8- 828 522 .2.
a. 8. 8 8- 8. G- 8- 8. 8. 8 8 8- ;- 8- 27 8- 228.802 8882.8 .8
_. 8 8. 2 o : v 2. 8 8- 8. n 8 8 2 8 328:2 .8
a 2 2- 8 2 2 a a 2 2. 2. 8 2 8. 2 2 88.985 .8
n n 2 a. 2 8 n 8 2 :- o _- 2. a 2. o. 2588 2.2.38.9. .8
2 .- 8. 2 8 8 o 8 _ 8. 8- 2 2 2.- 2 o 388:5. .8
8- 8. 2. 8. 8. 8. 8- :- 2- 8 8 8- 8. 2 8. - 52: .8.
2 2. - w. v N a 2 2 8. - o a 2- m. N 0.22.8822 .8
8 :- 8. 8 8 8 2 8 2 8. 8. 8 :- 2 8 2 828.5 .8
8 8 8 8 8 8 8 8 8 8 8 8 8 2 2 2 2228

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.5282 8 228

 

150

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8 8 8 8 8 2.. 8 8- 8. 8 2 a. 22 8. a 2 28 588225 2
8 8 8 8 8 8 8 8- 8- 8 8 2 2 2- 8. a 8 8022282225 322288 .2
8 8 8 8 2- 8 8 8- 8- 2- 2 p 8 8- 8 8 8 8:35:20 32222.? .2
8 8 8 2 v 2 8 2m- 8- a 22 2.. 2 2- 2. 2 8 82.22.25 .288 .2
8- 22- 8.. 2- 8. a. 8- 8 8 8- 2- a 2x 8 22- 8. 8- .5823225 .2.
8. 22. 2. 22- 2. 2.. 8- 8 8 8- 22- v2 8. 2. :- 8- 8- 18.22 .2
2. 8 22. 2. 8 8 .8 8. 28. 8 2 a. 2 8. m 2 8 0822.2 .22
8 8 8 8 8 2m 8 8. 2. 2 2. 2- 22 21 2 2.2 8 8822 0.8222229 .2
S2 8 8 2. 8 2- 8 8. 3 2 8 2- a 21 2- 8 8 .803 .8
8 8 2. 8 8 8 2... 22.. 2a m 2 2 8. 8- 2- 2 8 .8228 .8
8 8 8 8 82 8 2. 8- 8- 2 22 8 8 2- 2- 2.2 2. 28.2: .8
8 8 8 28 2.. 28 8 8. 8. 8 2.. 8 o 8- m- 8 8 238:2 .8
8 8 8 8 8 8 8 28- 8- 8 .- 22 22 8- .. 8 8 .2833 .8
2. 8 8 8 8 8 8 8- 8- 8 n 8. 2- 2- 8- 8 :- 2882 8522225 .2.
8 8 8 8 2 8 8 2m. 8- 2- a N- 2 8- 2 2 8 2322239922 02:52.23. .8
2. 8 8 8 2 8 8 8- 2.. m- 8 2.. 2 8. m- 2 2. .5238 as: .8
8 8 8 8 2 8 8 8. 8- 22 22 N. 2 8- 2. 8 8 636 .82 .2
8 8 2. 8 8 8 8 2. :- 8 8 8 8 8 8 8 8 8,288

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2252.82 8 02,28

 

151

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

w». on 8, an an R S 21 x. cm 2 w. a a. o. 2 8. ”8558883 .2
8 z 3 a m. on S 2.. a. _- a m 2 mm- _ cm 2 $.35 an
on on. 3. 3- 8 mm 3 3 $- 2 2 n v 2.. b :- S $058202 .2
on. 2. mm- on. 2. 2- on. S 2. an. 2. 2 mm. a: u. m". R. .2385 02:82 .8
2. on a 2 2 2‘ g N? 2.- 2 o a. 2 2”. x. mm 8 .2585 2,522 .8
8 2 2. mm 3 mm mm 2.. 2.. a 2 2 a an- v «M 3 £58: .2
2- : 2 G 2 mm 2- 2.. a. : 2 R R on. m 2. o.- ..omé .R
2 an S 2- h- 2. a a... x- v a m m 2. a 2 2 =50 .3
2 m on 2 2 2 on «N. a. 2.. a on a :- 2 8 z .85-Eon .3
2‘ cm :- 3 R 3 a 8. on. a 2 2 _ :- 2 2 mm 2.325 .3
_ 2- «a- m : m on- 2‘ mm mm- 2- :- 21 2 mm. 3.. N- 5823.8 .2
$- 2. 8. mm. 2. a. 2.. 2. 3 mm. 2- c on- on N. R. 8. 88.-meow .NN
: a 8 R m 3. 3. an- 2- m on _. 2 2- o mm 3 288.3- ..N
8 3 2 2. a 2. 2 2- 5- 3. 2 \- 2 an. m z. 2. 32.92% .8
an 2 :- 2. a on 2- R- mm- mm 9 m 2.. 2 2. 2 8 ESE-«8.. ..82 .2
8 mm 2 a 3. 2 :- 3- $- 2 2 _- 2 a. m. cm 2. 2822.235 .2
2- 2- :- 2. 2- 3 2‘ S S 8 an an R on 2 3 an El

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.282 mm 032-

 

152

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.2 -.R R 2., mm on on on. 2.. 2 o o 2 mm- m. 2 3 8.385.. .2.
R 8. 3 2. 2 R R 8. R- e o ... 2 ._ 2 2 2 on 3:35.30 28a .2.
R 3 c2 3 2. 2. 2 m2 8. 2 2 «.- o R.- N 3 8 18.-2 .2.
2 2. z. 8. 2. 2. «m 3.. 2.. a m _. 2 .R. :- : an 1an .2.
mm mm 2. 2. 8. mm R 2- R. u 7 : R. : 2- v R .2.st .2.
an o“ 2. 2. mm 2: : .... 2.. 2 : 2 .- 2 m R R 35:38:03.5 .3.
on R 2 2 am 3 8. R. 2. R 2 a 2 2.. R a 3 38805.5 .2.
on- on. 2 32 2- .... R- 8. 8 R. R- ...- 2 G ...- :1 2.. 2..-299.82 .2.
2.. R- o2 2.. R. 2.. m2 8 8. .2 R- 2 2- 2 :- :.. ..2 gxcaoé 58.2.23 .:.
2 o 2 o R 2 R R- R. :2 v on x a. on 2. on 32.35.. .2.
o ... 2 m 2. : 2 R- R. v 8. .. on R. 2 .m 2 82:28.5 .3
o ... m- _- : 2 a m. 2 on .. .2 .0 mm R 2. 2 2.28m 8:35 an
2 2 o 2 R- .- 2 2 2- 3 cm S .2 mm- 8 2. x 382.5. .R
R- 2.- .R. .R. : 2 2.. G 2 a. on. mm mm- 2: _ 2- 2m- .22: .3
m- 2 R :- 2- m R :- :- on 2 R on _ 8. 2. 2 0.22885 .2
2 2 x : v R R 3.. :.. 2. R 2. 2. 2- 2. 8. 2 222.5 ...m
2. 2. S 2. 2. 3. 2. 9 :. 2. an an R 3. mm ..m mm 8222.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.9283 mm uBu-H

 

REFERENCES

REFERENCES

Abramson, L. Y., Metalsky, G. 1., & Alloy, L. B. (1989). Hopelessness depression: A
theory-based subtype of depression. Echological Review. $0), 358-372.

Akiskal, H. S. (1994). Dysthymia: Clinical and external validity. Acta Psychiatrica
Scandinavica, §_9_(Suppl. 383), 19-23.

Akiskal, H. S., & Weise, R. S. (1992). The clinical spectrum of so-called “minor”
depressions. American Jourgal of Psychotherapy, 313(1), 9-22.

Arnerican Psychiatric Association (1987). Diagnostic and statistical manual of mental
disorders (Third ed., rev). Washington DC: American Psychiatric Association.

Andrews, G., Stewart, 6., Allen, R., & Henderson, A. S. (1990). The genetics of six
neurotic disorders: A twin study. Journ_al of Affective Disorders. _1_9_, 23-29.

Angst, J ., Vollrath, M., Merikangas, K. R, & Ernst, C. (1990). Comorbidity of anxiety
and depression in the Zurich cohort study of young adults. In J. D. Maser & C. R.

Cloninger (Eds), Comorbidig of mood and anxiety disorders (pp. 123-137).
Washington, DC: American Psychiatric Press.

Apter, A., van Praag, H. M., Plutchik, R, Sevy, S., Kom, M., & Brown, S. L. (1990).
Interrelationships among anxiety, aggression, impulsivity, and mood: A
serotonergically linked cluster? Psychiatry Research, 3_2_, 191-199.

Arieti, S., & Bemporad, J. R. (1980). The psychological organization of depression.
American Journal of Psychiatg, lfl, 1360-1365.

Bagby, R. M., Cox, B. J., Schuller, D. R., Levitt, A. J., Swinson, R. P., & Joffe, R. T.
(1992). Diagnostic specificity of the dependent and self-critical personality
dimensions in major depression. Journal of Affective Disorders ;_6_, 59-64.

 

B'axlow, D. H. (1985). The dimensions of anxiety disorders. In A. H. Tuma & J. D.

Maser (Eds), Anxieg and the anxiety disorders (pp. 497-500). Hillsdale, NJ:
Erlbaum.

Bartlett, E. S., & Izard, C. E. (1972). A dimensional and discrete emotions investigation
of the subjective experience of emotion. In C. E. Izard (Ed), Patterns of

emotions: A new analysis of anxiety and depression (pp. 129-175). New York:

Academic Press.

153

154

Beck, A. T. (1967). Depression: Clinical. experimental_and theoretical aspects. New
York: Holber.

Beck, A. T., & Beamesderfer, A. (1974). Amount of depression: The Depression
Inventory. In P. Pichot (Ed), Psychological megurements in
psychopharmacology (pp. 151-169). Basel: Karger.

Beck, A. T., Brown, 6., Eidelson, J. I., Steer, R. A., & Riskind, J. H. (1987).
Differentiating anxiety and depression: A test of the cognitive content-specificity

hypothesis. Jorml of Abnormal Psychology, 26(3), 169-183.

Beck, A. T., & Clark, D. A. (1988). Anxiety and depression: An information processing
perspective. In R. Schwarzer & R. A. Wicklund (Eds), Anxieg and self-focused

attention (pp.4l-54). Chur, FRG: Harwood.

 

 

Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for
measuring clinical anxiety: Psychometric properties. Journal of Consulting and

Clinigrl Psychology, i616), 893-897.

Beck, A. T., Epstein, N., & Harrison, R. (1983). Cognitions, attitudes and personality
dimensions in depression. British Journi of Cpgnitive Ther_py, 1, 1-16.

 

Beck, A. T., Ward, C. H., Mendelson, M., Mock, J ., & Erbaugh, J. (1961). An inventory
for measuring depression. Archives of General Psychiatg, 1;, 57-62.

Becker, E. W., & Lesiak, W. J. (1977). Feelings of hostility and personal control as
related to depression. Journal of Clinical chhology, 33(3), 654-657.

Biaggio, M. K., & Godwin, W. H. (1987). Relation of depression to anger and hostility
constructs. Psychological Reports, fl, 87-90.

Blaney, P. H. (1977). Contemporary theories of depression: Critique and comparison.
Journal of Abnormaflsycholcgy, §6_(3), 203-223.

Blatt, S. J. (1974). Levels of object representation in anaclitic and introjective

depression. Psychoanalyaic Study of the Child, 12, 107-157.

Blatt, S. J ., D'Aflliti, J. P., & Quinlan, D. M. (1976a). Experiences of depression in

normal young adults. Journal of Abnormal Psychology, QM), 383-389.

Blatt, S. J ., D'Afiliti, J. P., & Quinlan, D. M. (1976b). Depressive Experiences
Questionnaire, New Haven, CN: Yale University.

 

155

Blatt, S. J., Hart, 8., Quinlan, D. M., Leadbetter, B., & Auerbach, J. (1993).
Interpersonal and self-critical dysphoria and behavior problems in adolescents.
Journal of Youth m Adolescence. 2_2(3), 253-269.

Blatt, S. J ., Quinlan, D. M., & Chevron, E. S. (1990). Empirical investigations of a
psychoanalytic theory of depression. In J. Masling (Ed), Empirical studies of
psychoanalflic theories (Vol. 3, pp. 89-147). Hillsdale, NJ: Analytic Press.

Blatt, S. J., Quinlan, D. M., Chevron, E. S., McDonald, C., & Zuroff, D. (1982).
Dependency and self-criticism: Psychological dimensions of depression. Journal

of Consultinwd Clinical Psychology, $0), 113-124.

Blatt, S. J ., & Zuroff, D. (1992). Interpersonal relatedness and self-definition: Two
prototypes for depression. Journal of Consulting and Clinical Psychology Review.
1;, 527-562.

Block, J ., & Robins, R. W. (1993). A longitudinal study of consistency and change in
self-esteem from early adolescence to early adulthood. Child Development, Q4,
909-923.

Brady, E. U., & Kendall, P. C. (1992). Comorbidity of anxiety and depression in children
and adolescents. Psychological Bulletin, m0), 244-255.

Breier, A, Albus, M., Pickar, 1)., Zahn, 'r. P., Wolkowitz, o. M., & Paul, s. M. (1987).
Controllable and uncontrollable stress in humans: Alterations in mood,
neuroendocrine and psychophysiologic function. American Jgumal of Psychiatg,
L43, 1419-1425.

BI’eier, A., Chamey, D. S., & Heninger, G. R. (1985). The diagnostic validity of anxiety
disorders and their relationship to depressive illness. American 101M
Psychiatg, 1_42(7), 787-797.

B1‘eslau, N., Davis, G. C., & Prabucki, K. (1987). Searching for evidence on the validity
of generalized anxiety disorder: Psychopathology in children of anxious mothers.

Psychiatry Research, m, 285-297.

Bl‘etherton, I. (1985). Attachment theory: Retrospect and prospect. In I. Bretherton &
E. Waters (Eds), Growing points of attachment theory and research. Monogpaphs

of the Society for Research in Child Development, 59(1-2; Serial No. 209), 3-35.

Bl‘ewin, C. R., & Fumham, A. (1987). Dependency, self-criticism and depressive
attributional style. British Journal of Clinical Psychology, 2Q, 225-226.

156

Brophy, C. J., Norvell, N. K., & Kiluk, D. J. (1988). An examination of the factor
structure and convergent and discriminant validity of the SCL-9OR in an outpatient
clinic population. Journal of Personality Assessment. 2(2), 334-340.

Brown, J. D., & Silberschatz, G. (1989). Dependency, self-criticism, and depressive
attributional style. Journal of Abnoml Psychology, 28(2), 187-188.

Buss, A. H. (1960). Two anxiety factors in psychiatric patients. Journal of Abnormal

Sociology, 2_3, 56.

Carey, G., & DiLalla, D. L. (1994). Personality and psychopathology: Genetic
perspectives. Jourfll of Abnormal Psychology, $0), 32-43.

Carver, C. S., & Scheier, M. F. (1991). A control-process perspective on anxiety. In R
Schwarzer & R. A. Wicklund (Eds), Anxiety and self-focused altention (pp. 3-8).
Chur, FRG: Harwood Academic Publishers.

Chang, S. S. (1988). The measurement of shame and its relationship to other personality
Mtg. Unpublished doctoral dissertation, Michigan State University.

Charles, M., & Levine, R. (1995). Measuring guilt and shame: Construct vaJigigpan_d
Lemma» with associated variables. Manuscript in progress. Michigan State
University, East Lansing, MI.

 

Chrousos, G. P., & Gold, P. W. (1992). The concepts of stress and stress system
disorders: Overview of physical and behavioral homeostasis. WAC.
American Medical Association 267 1244-1252.

 

Clark, D. A. & Beck, A. T. (1989). Cognitive theory and therapy of anxiety and
depression. In P. C. Kendall & D. Watson (Eds), Anxiety and depression:
Distinctive and overlapping features (pp. 379-411). San Diego, CA: Academic
Press.

Clark, D. A., Beck, A. T., & Stewart, B. (1990). Cognitive specificity and
positive-negative affectivity: Complementary or contradictory views on anxiety
and depression? Journal of Abnormal Psycholpgy, 22(2), 148-155.

Clark, L. A. (1989). The anxiety and depressive disorders: Descriptive psychopathology
and differential diagnosis. In P. C. Kendall & D. Watson (Eds), Anxing Ed
depression: Distinctive Mverlapping features (pp. 83-129). San Diego, CA:
Academic Press.

157

Clark, L. A., & Watson, D. (1989). The General Temperament Survey. Unpublished
manuscript, Southern Methodist University, Dallas, TX.

Clark, L. A., Watson, D., & Mineka, S. (1994). Temperament, personality, and the mood
and anxiety disorders. Journal of Abnormal Psychology, EU), 103-116.

Clifl‘ord, C. A., Hopper, J. L., Fulker, D. W., & Murray, R. M. (1984). A genetic and
environmental analysis of a twin family study of alcohol use, anxiety, and
depression. Genetic Epidemiology, 1, 63-79.

C loninger, C. R., (1986). A unified biosocial theory of personality and its role in the
development of anxiety states. Psychiatric Developments, 2, 167-226.

Cloninger, C. R. (1987). A systematic method for clinical description and classification of
personality variants: A proposal. Archives of General Psychiatg, fl, 573-588.

Cloninger, C. R. (1988a). A unified biosocial theory of personality and its role in the
development of anxiety states: A reply to commentaries. Psychiatric
Davelopments, 2, 83-120.

Cloninger, C. R. (1988b). Anxiety and theories of emotion. In R. Noyes, Jr., M. Roth, &
G. D. Burrows, (Eds), findboci of anxiety. Vol. 2: Classification. etiological
factors and associated disturbjances. Amsterdam: Elsevier.

Cloninger, c. R., Martin, R. L, Guze, s. B, & Clayton, P. J. (1981). A blind follow-up
and family study of anxiety neuroses: Preliminary analysis of St. Louis 500. In D.
F. Klein, & J. Rabkin (Eds), Anxiety: New reseflh and changing concepts (pp.
137-150). New York: Raven.

Cloninger, C. R., Martin, R. L., Guze, S. B., & Clayton, P. J. (1990). The empirical
structure of psychiatric comorbidity and its theoretical significance. In J. D. Maser

& C. R. Cloninger (Eds), Comprbidig pf mood and anxiety disorders (pp.
439-462). Washington, DC: American Psychiatric Press.

Cook, D. R. (1985). The Intemalized Shame Scale. Unpublished manuscript, University of
Wisconsin, Menomonie.

C08m, P. T., & McCrae, R. R. (1985). Th; NEQ Personality Inventogy manual. Odessa,
FL: Psychological Assessment Services.

C()sta, P. T., & McCrae, R. R. (1987). Neuroticism, somatic Complaints, and disease: Is
the bark worse than the bite? Joamal of Parsonality, fl, 299-316.

158

Crowe, R. R., Noyes, R., Pauls, D. L., & Slymen, D. S. (1983). A family study of panic
disorder. Archives of GenerarlPsychim, fl, 1065-1069.

Cyr, J. J ., McKenna-Foley, J. M., & Peacock, E. (1985). Factor structure of the SCL-
90R: Is there one? Journal of Personality Assess_ment. fi(6), 571-578.

Deakin, J. F. W., Guimaraes, F. S., Wang, M., & Hensman, R. (1991). Experimental
tests of the 5-HT receptor imbalance theory of affective disturbance. In M.

Sandler, A. Coppen, & S. Hamett (Eds), S-Hydromgptamine in psychiatu: A
apactrum of ideas (pp. 143-156). Oxford: Oxford University Press.

Depue, R. A., Luciana, M., Arbisi, P., Collins, P., & Leon, A. (1994). Dopamine and the
structure of personality: Relation of agonist-induced dopamine activity to positive
emotionality. Journal of Personalig and Social Psychology, QB), 485-498.

Derogatis, L. R., Lipman, R. S., & Covy, L. (1973). The SCL-90: An outpatient
psychiatric rating scale. Psychopharmacology Bulletin, 2, 13-28.

Dohrenwend, B. P. (1990). Notes on some epidemiologic studies of comorbidity. In J.
D. Maser & C. R. Cloninger (Eds), Comorbidity of mood and anxiety disorders
(pp. 177-185). Washington, DC: American Psychiatric Press.

Drugan, R. C., Maier, s. F., Skolnick, R, Paul, s. M., & Crawley, J. N. (1985). An
axiogenic benzodiazapine receptor ligand induces learned helplessness. European

Journal of Pharmacology, m, 453-457.

I)‘Jbovsky, S. L. (1993). Approaches to developing new anxiolytics and antidepressants.
Journal of Clinical Psychiatu, 5_4(5, Suppl), 75-82.

Iiril(sson, E., & Humble, M. (1990). Serotonin in psychiatric pathophysiology: A review
of data from experimental and clinical research. In R. Pohl & S. Gershon (Eds),

The biplogical basis of psychiatric treatment. In Progress in basic clinical
pharmacology (Vol. 3, pp. 66-119). Basel: Karger.

 

Ettenberg, A. (1989). Dopamine, neuroleptics and reinforced behavior. Neuroscience
and Biobehayioral Reviews. 1_3_, 105-111.

Eysenck, H. J ., & Eysenck, S. B. G. (1968). Eysenck Personalng Qpestionnaire manual.
San Diego, CA: Educational and Industrial Testing Service.

Fllhr, S. K., & Shean, G. (1992). Subtypes of depression, efficacy, and the Depressive
Experiences Questionnaire. The qumal of Pachology, _1_2§(5), 495-506.

159

Garber, J ., Miller, W. R., & Seaman, S. F. (1979). Learned helplessness, stress, and the
depressive disorders. In R. A. Depue (Ed), The psychobiology of the depressive
disorders: Implications for the effects of stress (pp. 335-363). New York:
Academic Press.

Gibbons, F. X. (1991). Self-evaluation and self-perception: The role of attention in the
experience of anxiety. In R. Schwarzer & R. A. Wicklund (Eds), Anxiety and
self-focused attention (pp. 15-25). Chur, FRG: Harwood Academic Publishers.

Gold, P., W., Goodwin, F. K., & Chrousos, G. P. (1988). Clinical and biochemical
manifestations of depression: Relations to the neurobiology of stress: II. New
England Journal of Medicine. 319(7), 413-420.

Colin, 8., & Hartz, M. A. (1979). A factor analysis of the Beck Depression Inventory in
a mildly depressed population. Journg of Clinical Pachology, 25, 308-319.

Gotlib, I. H, & Cane, D. B. (1989). Self-report assessment of depression and anxiety. In
P. C. Kendall & D. Watson (Eds), Anxiety and depress_ion: Distinctive an_d
overlapping features (pp. 131-169). San Diego, CA: Academic Press.

Gotlib, I. H., & Meyer, J. P. (1986). Factor analysis of the Multiple Afl‘ect Adjective
Checklist: A separation of positive and negative affect. Journal of Personalig and

Social Pachology, fl, 1161-1165.

Green, D. P., Goldman, S. L., & Salovey, P. (1993). Measurement errors mask bipolarity
in afi‘ect ratings. Jouma1 of Personality and Social Psychology, fi(6), 1029-1041.

Grinker, R. R., Miller, J ., Sabshin, M., Nunn, R., & Nunnally, J. (1961). The phenomena
of depressions. New York: Hoeber.

Hague“, 0., & Grasbeck, A. (1990). Comorbidity of anxiety and depression in the
Lundby 25-year prospective study: The pattern of subsequent episodes. In J. D.

Maser & C. R. Cloninger (Eds), Comorbidig of mgod and anxieg disorders (pp.
139-152). Washington, DC: American Psychiatric Press.

Henderson, N. D. (1986). Human behavior genetics. Annual Review of Psychology, 22,
403-440.

Heninger, G. R., & Chamey, D. S. (1988). Monoamine receptor systems and anxiety
disorders. Psychiatric Clinics of North America, 11(2), 309-326.

Hirschfield, R. M. A., Klerman, G. L., Andreasen, N. C., Clayton, P. J ., & Keller, M. B.

160

(1986). Psycho-social predictors of chronicity in depressed patients. British
Journal of Psychiatg, 1g, 648-654.

Hirschfield, R. M. A., Klerman, G. L., Lavori, P., Keller, M. B., Griffith, P., & Coryell,
W. (1989). Premorbid personality assessments of first onset of major depression._

Archives of General Psychiatty, @, 345-350.

Hoehn-Saric, R., McLeod, D. R., & Zimmerli, W. D. (1988). Differential effects of
alprazolam and imipramine in generalized anxiety disorder: Somatic versus :
psychic symptoms. Journal of Clinical Psychiatgg, 42(8), 293-301.

 

Insel, T. R., (1991). Serotonin in obsessive-compulsive disorder: A causal connection or
more monomania about a major monoamine? In M. Sandler, A. Coppen, & S.

Harnett (Eds), S-Hydromtyptamine in psychiatty: A spectrum of ideas (pp. 228-
257). Oxford: Oxford University Press.

Izard, C. E., Blumberg, S. H., & Oyster, C. (1985). Age and sex differences in the
pattern of emotions in child anxiety and depression. In J. T. Spence & C. E. Izard

(Eds), Motivatioa, emotion, and personality. Amsterdam: North-Holland.

Jardine, R., Martin, N. G., & Henderson, A. S. (1984). Genetic covariation between
neuroticism and the symptoms of anxiety and depression. Genetic Epidemiology,

1, 89-107.

Joseph, R. (1988). The right cerebral hemisphere: Emotion, music, visual-spatial skills,
body-image, dreams, and awareness. Journal of Clinical Psychiatry, fl, 630-673.

Kashani, J. H., Beck, N. C., Hoeper, E. W., Fallahi, C., Corcoran, C. M., McAllister, J.
A., Rosenberg, T. K., & Reid, J. C. (1987a). Psychiatric disorders in a
community sample of adolescents. American Journal of Psychiatry, 133, 585-589.

Kashani, J. H., Carlson, G. A., Beck, N. C., Hoeper, E. W., Corcoran, C. M., McAllister,
J. A., Fallahi, C., Rosenberg, T. K., & Reid, J. C.. (1987b). Depressron,
depressive symptoms, and depressed mood among a community sample of

adolescents. American Journal of Psychiatty, 114(7), 931-934.

Katz, M. M., Koslow, S. H., Berman, N, Secunda, S, Maas, J. W., Casper, R., Kocsis,
J ., & Stokes, P. (1987). Multivantaged approach to the. measurement of
behavioral and affect states for clinical and psychobiological research.

Ps cholo 'cal Re orts, 55, 619-671.
Katz, M. M., Maas, J. W., Frazer, A., Koslow, S. H., Bowden, C. L., Berman, N., Swarm,

 

161

A. C., & Stokes, P. E. (1994). Drug-induced actions on brain neurotransmitter
systems and changes in the behaviors and emotions of depressed patients.

Neuropsychopharrnacology, 3(2), 89-100.

Katz, M. M., Robins, E., Croughan, J ., Secunda, S., & Swann, A. (1987).Behavioural
measurement and drug response characteristics of unipolar and bipolar depression.

Psychological Medicine, Q, 25-36.

Kelly, D., & Walter, C. J. S. (1969). A clinical and physiological relationship between
anxiety and depression. Britis_h Journal of Psychiatg, 1Q, 401-406.

Kendall, P. C ., & Ingram, R. E. (1989). Cognitive-behavioral perspectives: Theory and
research on depression and anxiety. In P. C. Kendall & D. Watson (Eds),
Anxiety and depressipn: Distinctive and overlapping features (pp. 27-53). San
Diego, CA: Academic Press.

King, R., Margraf, J ., Ehlers, A., & Maddock, R. (1986). Panic disorder; Overlap with
symptoms of somatization disorder. In 1. Hand & H.-U. Wittchen (Eds), Panic

and phobias: Empirical evidence of theoretical models and longterm effects of

behavioral treaanents (pp. 72-77). Berlin: Springer-Verlag.

Klein, D. N., Harding, K., Taylor, E. B., & Dickstein, S. (1988). Dependency and self-
criticism in depression: Evaluation in a clinical population. Journal of Abnormal

Psychology, fl(4), 300-404.

Kobak, R. R., & Sceery, A. (1988). Attachment in late adolescence: Working models,
affect regulation, and representations of self and others. Child Develmpment. 52,

135-146.

Kobak, R. R., Sudler, N., & Gamble, W. (1992). Attachment and depressive symptoms
during adolescence: A developmental pathways analysis. Development and
Psychopathology, 2, 461-474.

Kraemer, G. W., Ebert, M. H., Schmidt, D. E., & McKinney, W. T. (1991). Strangers in
a strange land: A psychobiological study of infant monkeys before and after
separation from real or inanimate mothers. Child Develo ment, 6_2_, 548-566.

Last, C. G., Francis, G., Hersen, M., Kazdin, A. E., Strauss, C. C. (1987). Separation
anxiety and school phobia: A comparison using DSM-III criteria. Amancan

Journal prsychiatty, 14:1, 653-657.
Leckman, J. F., Weissman, M. M., Merikangas, K. R., Pauls, D. L., & Prusoff, B. A.

 

162

(1983). Panic disorder increases risk of major depression, alcoholism, panic, and
phobic disorders in affectively ill families. Archives of General Psychiatg, fl,
1055-1060.

Lips, H. M., & Ng, M. (1985). Use of the Beck Depression Inventory with three
nonclinical populations. Canadian Journal of Ewioural Science. 18, 62-74.

 

Lipsey, M. W., & Wilson, D. B. (1993). The efficacy of psychological, educational, and
behavioral treatment. American Psychologist, £02), 1181-1209.

Leonard, B. E. (1994). Serotonin receptors- Where are the going? International Clinical
Psychopharmacology, 2(Suppl. 1), 7-17.

Lopez-Ibor, J. J ., Jr. (1990). The spectrum of mood disorders. Journal of Clinical
Psychiatry, 21 (9; Suppl), 62-64.

 

Maier, S. F ., & Seligman, M. E. P. (1976). Learned helplessness: Theory and evidence.
Journal of Experimental Psychology: General, fl, 3-46.

Maier, W., Buller, R., & Hallmayer, J. (1988). Comorbidity for panic disorder and major
depression: Results from a family study. In 1. Hand & H.-U. Wittchen (Eds),
Panic and phobias: Treatments and variables affecting course and outcome (pp.
180-185). Berlin: Springer-Verlag.

 

Martin, N. G., Jardine, R., Andrews, G., & Heath, A. C. (1988). Anxiety disorders and
neuroticism: Are there genetic factors specific to panic? Acta Psychiatrica
Scandinavica, fl, 698-706.

McCranie, E. W., & Bass, J. D. (1984). Childhood family antecedents of dependency and
self-criticism: Implications for depression. Journal of Abnormal Psychology,
$(1), 3'8-

Meltzer, H. Y. (1990). Role of serotonin in depression. In P. M. Whitaker-Azmith & S.

J. Peroutka (Eds), Annals of the New York Academy of Sciences: The
neuropharrnacology of serotonin (Vol. 600, pp. 486-499). New York: The New
York Academy of Sciences.

Merikangas, K. R. (1990). Comorbidity for anxiety and depression: Review of family
and genetic studies. In J. D. Maser & C. R. Cloninger (Eds), Comorbidig of

mood and anxith disorders (pp. 331-348). Washington, DC: American
Psychiatric Press.

163

Mook, J ., van der Ploeg, H. M., & Kleijn, W. C. (1990). Anxiety, anger and depression:
Relationships at the trait level. Anxiety Research, 2, 17-31.

Mountjoy, C. Q., & Roth, M. (1982). Studies in the relationship between depressive
disorders and anxiety states: Part 1. Rating scales. Journal of Affective
Disorders, 4, 127-147.

Mullaney, J. (1987). Measurement of anxiety and depression. In G. Racagni & E.

Smeraldi (Eds), Anxious depression: Assessment and treatment (pp. 7-19). New
York: Raven Press.

 

Murray, L., & Trevarthen, C. (1985). Emotional regulation of interactions between two-
month-olds and their mothers. In T. M. Field & N. A. Fox (Eds), Social
perception in infants (pp. 177-197). Norwood, NJ: Ablex.

 

Noyes, R., Clarkson, C., Crowe, R. R., Yates, W. R., & McChesney, C. M. (1987). A
family study of generalized anxiety disorder. Ameriaan Journal of Psychiatpy,
m, 1019-1024.

Nurcombe, B., Seifer, R., Scioli, A., Tramonta, M. G., Grapentine, W. L., & Beauchesne,
H. C. (1988). Is major depressive disorder in adolescence a distinct diagnostic

entity? Jouraal of the American Academy of Child and Adolescent Psychiatiy,
2_8(3), 333-342.

 

Ofl‘er, D., Ostrov, E., & Howard, K. (1989). The Offer Self-Ima e estionnaire for
Adolescents. Chicago: Michael Reese Hospital.

Ollendick, T. H., & Yule, W. (1990). Depression in British and American children and its

relation to anxiety and fear. Journal of Consulting and Clinical Psychology, 3(1),
126-129.

Paul, S. M. (1988). Anxiety and depression: A common neurobiological substrate?
Journal of Clinical Psychiatty, 42(10, Suppl), 13-16.

Persons, J. B. (1986). The advantages of studying psychological phenomena rather than
psychiatric diagnoses. American Psyghologist, £01), 1252-1260.

Persons, J. B., Burns, D. D., Perloff, J. M., & Miranda, J. (1993). Relationships between
symptoms of depression and anxiety and dysfunctional beliefs about achievement

and attachment. Journal of Abnormal Psychology, 192(4), 518-524.
Petty. F., Kramer, G., & Moeller, M. (1994). Does learned helplessness induction by

164

haloperidol involve serotonin mediation? Pharmacology, Biochemistty, and
Behavior, £6), 671-676.

Pilkonis, P. A. (1988). Personality prototypes among depressives: Themes of
dependency and autonomy. Journal of Personality Disorders. 2, 144-152.

Prusoff, B. A., & Klerman, G. L. (1974). Differentiating depressed from anxious neurotic
outpatients. Archives of GeneraiPsychiath, 3, 302-309.

Puig-Antich, J ., & Rabinovich, H. Relationship between affective and anxiety disorders in
childhood. In R. Gittelman (Ed), Anxing disorders of childhood. New York:
Guilford.

Regier, D. A., Burke, J. D., & Burke, K. C. (1990). Comorbidity of affective and anxiety
disorders in the NHVIH Epidemiological Catchment Area program. In J. D. Maser
& C. R. Cloninger (Eds), Comorbidityaof mood and anxietva disorders (pp. 113-
122). Washington, DC: American Psychiatric Press.

 

Reich, J ., Noyes, R., Hirschfield, R., Coryell, W., & O'Gorman, T. (1987). State and
personality in depressed and panic patients. American Journal of Psychiatg,
l4_4(2), 181-187.

 

Rosenberg, M. (1965). Society and the adolescent self-imags. Princeton: Princeton
University Press.

Sarason, I. G. (1985). Cognitive processes, anxiety and the treatment of anxiety

disorders. In A. H. Tuma & J. Maser (Eds), Anxieg and the anxing disorders
(pp. 932-946). Hillsdale, NJ: Erlbaum.

Schatzberg, A. F., Samson, J. A., Rothschild, A. J ., Luciana, M. M., Bruno, R. F ., &
Bond, T. C. (1990). Depression secondary to anxiety: Findings from the McLean
Hospital Depression Research Facility. Psychiatric Clinics of North America,
QM), 633-649.

Schmale, A. H., & Engel, G. L. (1975). The role of conservation-withdrawal in
depressive reactions. In E. J .Anthony & T. Benedek (Eds), Depression and
human existence. Boston: Little, Brown.

Schore, A. N. (1994). Affect regalationana the origin of the self: The neurobiology of
emotional development. Hillsdale, NJ: Erlbaum.

 

Schutz, R. A., Schutz, M. T. B., Orsingher, O. A., & Izquierdo, I. (1979). Brain

165

dopamine and noradrenaline levels in rats submitted to four different aversive
tasks. Psychopharmacology 63 289-292.

,—9

Schwab, J. J ., Bialow, M. R., & Holzer, C. E. (1967). A comparison of rating scales for
depression. Journal of Clinical Psycholggy, 2a, 225-226.

Schwalbe, M. L. (1991). The autogenesis of the self. Joumad of the theory of social
behavior, 21 , 269-295.

Shapiro, J. P. (1988). Relationships between dimensions of depressive experience and
evaluative beliefs about people in general. Personality and Sociafl’sychology
Bulletin, 14(2), 388-400.

Shaw, B. F., Vallis, T. M., & McCabe, S. B. (1985). The assessment of the severity and
symptom patterns in depression. In E. E. Beckharn & W. R. Leber (Eds),
Handbook of depres_sion: Treatment. assessment. and reseam (pp. 372-407).
Homewood, IL: Dorsey Press.

Shutty, M. S., DeGood, D. E., & Schwartz, D. P. (1986). Psychological dimensions of
distress in chronic pain patients: a factor analytic study of Symptom Checklist-90
responses. Journal of Con_sulting and Clinical Psychology, 5_4(6), 836-842.

Smith, T. W., O’Keeffe, J. L., & Jenkins, M. (1988). Dependency and self-criticism:
Correlates of depression or moderators of the effects of stressfiil events? Journal

of Personalig Disorders, 2(2), 160-169.

Soubrie, P. (1986). Reconciling the role of central serotonin neurons in human and
animal behavior. The Behavipral and Brain Sciences, 2, 319-364.

Spielberger, C. D. (1972). Anxiety as an emotional state. In C. D. Spielberger (Ed),
Anxieg: Current trends in theogg and research (Vol. 1; pp. 23-49). New York:

Academic Press.

Svrakic, D. M., Przybeck, T. R., & Cloninger, C. R. (1992). Mood states and personality
traits. Journal of Affective Disorders, 2_4, 217-226.

Swerdlow, N. R., & Koob, G. F. (1987). Dopamine, schizophrenia, mania, and
depression: Toward a unified hypothesis of cortico-striato-pallido-thalamic
fiinction. Behavioral and brain sciences, m, 197-245.

Tellegen, A. (1985). Structure of mood and personality and their relevance to assessing
anxiety, with an emphasis on self-report. In A. H. Tuma & J. Maser (Eds),

166
Anxiety and the anxiety disorders (pp. pp. 681-706). Hillsdale, NJ: Erlbaum.

Tellegen, A. (in press). The Multidimensional Personality Questionnaire. Minneapolis

MN: University of Minnesota Press.

Torgerson, S. (1985). Hereditary differentiation of anxiety and affective disorders.
Britis_h Jourratl of Psychiatty, m, 530-534.

Torgersen, S. (1990). A twin study perspective of the comorbidity of anxiety and
depression. In J. D. Maser & C. R. Cloninger (Eds), Comorbidity of mood and
anxiety disorders (pp. 367-3 78). Washington, DC: American Psychiatric Press.

Uhde, T. W., Roy-Byme, P. P., Vittone, B. J ., Boulenger, J.-P., Post, R. M. (1985).
Phenomenology and neurobiology of panic disorder. In A. H. Tuma & J. Maser
(Eds), Anxiety and the anxiety disorders (pp. 557-566). Hillsdale, NJ: Erlbaum.

van Praag, H. M. (1980a). Central monoamine metabolism in depressions: I. Serotonin
and related compounds. Comprehensive Psychiatgz, 2_1(1), 30-43.

van Praag, H. M. (1980b). Central monoamine metabolism in depressions: II.
Catecholamines and related compounds. Comprehensive Psychiatg, £0), 44-54.

van Praag, H. M. (1994). 5-HT-related, anxiety- and/or aggression-driven depression.
International Clinical Psychopharmacology, 2(Supp1. 1), 5-6.

van Praag, H. M., Asnis, G. M., Kahn, R. S., Brown, S. L., Korn, M., Harkavy Friedman,
J. M., & Wetzler, S. (1990). Monoamines and abnormal behavior: A
multiaminergic perspective. British Journal of Psychiatgg, 1_51, 723-734.

van Praag, H. M., Asnis, G. M., Kahn, R. S., Brown, S. L., Korn, M., Harkavy Friedman,
J. M., & Wetzler, S. (1990). Nosological tunnel vision in biological psychiatry: A
plea for a functional psychopathology. In P. M. Whitaker-Amnith & S. J.
Peroutka (Eds), Annals of the New York Academy of Sciences: The
neurophamracology of serotonin (Vol. 600, pp. 501-510). New York: The New
York Academy of Sciences.

 

van Praag, H. M., Kahn, R. S., Asnis, G. M., Wetzler, S., Brown, S. L., Bleich, A., &
Korn, M. L. (1988). Beyond nosology in biological psychiatry. S-HT
disturbances in mood, aggression, and anxiety disorders. In M. Briley & G. Fillion
(Eds), New concepts in depression (pp. 96-119). Basingstoke, England:
Macmillan.

167

Vredenburg, K., Krames, L., & Flett, G. L. (1985). Reexamining the Beck Depression
Inventory: The long and short of it. Psychological Reports, as, 767-778.

Watson, D. (1988). Intraindividual and interindividual analyses of positive and negative
affect: Their relation to health complaints, perceived stress, and daily activities.

Journal of Personality and Social Psychology, 24, 1020-1030.

Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience
aversive emotional states. chhological Bulletin, 26, 465-490.

Watson, D., & Clark, L. A. (1992). On traits and temperament: General and specific
factors of emotional experience and their relation to the five-factor model. Journal
of Personality, @(2), 441-476.

 

Watson, D., & Clark, L. A. (in press). Extraversion and its positive emotional core.
Chapter to appear in S. Briggs, R. Hogan, & W. Jones (Eds), mam;
Personalig Psychology. San Diego, CA: Academic Press.

Watson, D., & Clark, L. A. (1993). Behavioral disinhibition versus constraint: A
dispositional perspective. In D. M. Wegner & J. W. Pennebaker (Eds), Handbook
of mental control. New York: Prentice-Hall.

Watson, D., Clark, L. A., & Carey, G. (1988). Positive and negative affect and their
relation to anxiety and depressive disorders. Journal of Abnormal Psychology, _91,
346-353.

Watson, D., Clark, L. A., Weber, K., Assenheimer, J. S., Strauss, M. E., & McCormick,
R. A. (1995a). Testing a tripartite model: 1. Evaluating the convergent and
discriminant validity of anxiety and depression symptom scales. Journal of

Abnormal Psychology, 194(1), 3-14.

Watson, D., Clark, L. A., Weber, K., Assenheimer, J. S., Strauss, M. E., & McCormick,
R. A. (1995b). Testing a tripartite model: II. Exploring the symptom structure
of anxiety and depression in student, adult, and patient samples. Journal of

Abnormal Psychology, 124(1), 15-25.

Weiss, J. M, & Simson, P. G. (1985). Neurochemical mechanisms underlying stress-
induced depression. In T. M. Field, P. M. McCabe, & N. Schneiderman (Eds),
Stress and goping (pp. 93-116). Hillsdale, NJ: Erlbaum.

Weiss, J. M., Goodman, P. A., Losito, B. G., Corrigan, S., Charry, J. M., & Bailey, W. H.
(1981). Behavioral depression produced by an uncontrollable stressor:

168

Relationship to norepinephrine, dopamine, and serotonin levels in various regions
of rat brain. Brain Research Review. 2(2), 167-205.

Weissman, M. M. (1990). Evidence for comorbidity of anxiety and depression: Family
and genetic studies of children. In J. D. Maser & C. R. Cloninger (Eds),
Comorbidity of mood disorders (pp. 349-365). Washington, DC: American
Psychiatric Press.

Weissman, M. M., Leckman, J. F ., Merikangas, K. R., Gammon, G. D., & Prusoff, B. A.
(1984). Depression and anxiety disorders in parents and children: Results fi'om
the Yale family study. Archives of General Psychiatg, 41, 841-852.

Weissman, M. M., K. R., Gammon, G. D., John, K., Merikangas, Warner, V., Prusoff, B.
A., & Sholomskas, D. (1987). Children of depressed parents. Archives of

General Psychiatg, 44, 847-853.

Zung, W. W. (1972). How normal is depression? Psychosomatics, 13, 174-178.

Zuckerman, M., Persky, H., Eckman, K., & Hopkins, T. R. (1967). A multi-trait
multimethod measurement approach to the to the traits (or states) of anxiety,

depression, and hostility. Journal of Projective Techniques and Personalig
Assessment, 21 , 39-48.

Zung, W. W. (1972). How normal is depression? Psychosomatics, 1_3_, 174-178.

Zuroff, D. C., Igreja, I., & Mongrain, M. (1990). Dysfunctional attitudes, dependency,
and self-criticism as predictors of depressive mood states: A 12-month

longitudinal study. Cogm'tive Therapy and-Research, fl(3), 315-326.

Zuroff, D. C., Moskowitz, D. S., Wielgus, M. S., Powers, T. A., & Franko, D. L. (1983).
Construct validation of the dependency and self-criticism scales of the Depressive

Experiences Questionnaire. Journal pf Research in Perspnality, 12, 226-241.

Zuroff, D. C., & Mongrain, M. (1987). Dependency and self-criticism: Vulnerability
factors for depressive affective states. Journal of Abnormal Psychology, $0),
14-22.

nrcuran sran UNIV. LreRnRrEs
lllllllllllllllllllllllllllllllllllllllllllll11111111111111
31293014172542