.11

11111111111111.1111 ”71‘ 11' ”11111 11-111"

IIIIII III ' ' 1'11".-
I11III "JII1I * ~- ‘ ' 111111;

—.~ .4
:5.

———_—

II III ‘I11 I I I
"111 : 1-I1' 11 1'

_.r--
13'" .
‘ “l t.“ w
I v " . v
u

8.7.“ -
{:11

' W25. ._-;

t. °
7.
m ' .
u.’ . ‘
u.

1.. "
u "‘m“ rum-c9“ ‘

7’.- g
‘

J

-.-o

v -
uncoqq—

‘ L" «. _ _
4~.¢..‘ "

 

   
    
        

 
   

   
     
 

1 112. '31 1 1. {3.1 ”$13“; ‘ 1.12:1":
1 ‘1I1‘Jzy‘ 1 1.“ "IIJI1II‘I [”1” q ‘ “1;“,-
‘1. I H'EII.I11I‘11II1IJ‘ii‘1‘:,II‘III111';I§ I1|~1 ‘|" . = .:
III III "11I1'1111;"111'1'11131'?‘11'11'II' 11‘ -I._.'1.;l1
’ ‘11I 1 1 , |1I 1‘)..1‘ 1“ I.
' - 11'“ II1‘1 111: 1' “ 1' " .1111 11 11111111 1111' 1
1 .‘1'11111I111; ‘1 1 '~ 9
1 W l; I11'Ii1I11111M111H’1} 1 _ - 'II1II11
'1‘] 1111:, II1‘W.II1LII: .1 1'11‘rl' I11

I )1 (aIl'..I.1I .
1II1I .1II1I'I1III'I1I1I1I'111111‘ 1 11I III' IIII' I1111I ;I_: III’II I1.’.-4I.

.1 1:11
‘m " I" ‘11II111I11 1I1 1111111111‘II1II1 I I11111I11111IIIII111I1111| 111111111

11
1111' 1111:1III'1: .
111.1 11111.1. 111-- 1111111111 '1I111 I111II’ I1111111l1111111111111 W

THESIS

"will" l

3 1293 10692 0499 ___444

             

lllllllllll

                                       

L3“; i~m

' 7
‘fl' ‘_ i; VA!"
1 -34urauch

44“!
':.pm!_
li'nb‘bl‘u

 

:rad. £37?!" I
{13.33196 10-931: 3",. j

ls

;—-v..i.gu.c.¢;

 

7"“ 4

This is to certify that the

thesis entitled

AN EXAMINATION OF THE SOCIAL NETWORKS OF
NORMALS AND SCHIZOPHRENICS

presented by

Kenneth Lee Carrico, Jr.

has been accepted towards fulfillment
of the requirements for

 

 

 

 

Ph.D. dcgreein Counseling and
Educational Psychology
/ Major professor

 

0-7639

 

 

\Mzme 0:
.49

{Emma

 

OVERDUE FINES:

25¢ per day per item

RETURNING LIBRARY MATERIALS :

Place in book return to remove
charge tron circulation records

 

 

 

AN EXAMINATION OF THE SOCIAL NETWORKS OF
NORMALS AND SCHIZOPHRENICS

By

Kenneth Lee Carrico, Jr.

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Counseling and Educational Psychology

198l

ABSTRACT

AN EXAMINATION OF THE SOCIAL NETWORKS OF
NORMALS AND SCHIZOPHRENICS

By

Kenneth Lee Carrico, Jr.

The major purpose of this study was the comparison of the social
networks of normals and schizophrenics, the goal being identification
of psychosocial factors associated with the schizophrenic disorder.
The comparison was made along four classes of social network variables:
(a) structure, the basic morphological characteristics; (b) content,
the nature of shared activities; (c) function, the transaction of
support; and (d) emotion, the experiencing of affect. A secondary,
yet essential, purpose was the evaluation of the research assumption,
conceptually basic to much of the previous research, that self-report
data are a sufficient and reliable indicator of the actual status of
the social network.

The concepts of schizophrenic withdrawal and isolation represent
the theoretical basis of this study. The isolation hypothesis posits
that the person becomes schizophrenic as a result of being isolated
socially. The withdrawal hypothesis posits that as the schizophrenic
disorder progresses, the affected person becomes withdrawn. Support
was found in the review to substantiate social withdrawal and isola-

tion as interacting and reciprocal.

Kenneth Lee Carrico, Jr.

The normal sample was composed of 18 subjects and the schizo—
phrenic sample, of 17 subjects. All subjects were aged between l8 and
40 years and were living with family. Selection of the subjects was
designed to promote comparability along those dimensions that affect
outcome (age, sex, race, socioeconomic status, and cultural back-
ground), yet insure the identification of appropriate subjects from
both populations. Notably, the schi20phrenic sample was characterized
as willing to participate.

The methodology of the study consisted of two phases, the corrobo-
ration of self-report data within both samples and the comparison
between the two samples along l6 social network variables. In both
phases, the analyses were implemented through the use of multivariate
and post hoc univariate statistical procedures.

In the first phase, as the self-report of both samples was not
adequately corroborated, the conclusion was drawn that self-report
data are an insufficient, possibly inaccurate, and unreliable source
of information in social network analysis. This conclusion seriously
challenges the assumption made by previous researchers who viewed
self-report as sufficient and reliable. Furthermore, the conclusion
was drawn that the schizophrenic's perception of social relationships
was as reliable as the perception of the normal-~a startling conclu-
sion. Having far-reaching implications, evidence was also found to
support the contention that the lack of reliability of self-report
measures limited the power of the statistical tests to make compari—
sons between the two groups. In line with these findings, the conclu-

sion was drawn that the self-report data represent solelyifimeperceptions

Kenneth Lee Carrico, Jr.

of the subjects regarding their social networks, not the actual status
of the social networks.

In the second phase, the comparison between the social networks
of the normals and schizophrenics, significant differences were found
in relation to only two network variables, perceived network size and
reciprocity. Both results are in line with previous research. The
mean perceived network sizes of the social networks of normals and
schizophrenics were found to be 36.23 and 12.65, respectively, imply-
ing that the schizophrenics perceive themselves surrounded by fewer
"important" persons. In terms of reciprocity, the conclusion was
drawn that normals perceived their relationships as reciprocal in
terms of support, whereas schizophrenics perceived their relation-
ships as lacking in reciprocity, placing the schizophrenic in a depend-
ent position in his/her self-perception. The presence of significant
findings along only 2 of l6 variables provided weak support for the
hypotheses of withdrawal and isolation. Furthermore, the relative
lack of significant differences challenged two consistent findings
in previous studies, namely, fewer multiplex relations and negative

emotional perception in the social network of the schizophrenic.

DEDICATION

To my social network, in the context of which I came into exist-
ence and evolved as a human being. But especially my parents, the
true cornerstones of my personality, to whom I owe my creation,
preservation, and foundation--debts never paid, yet never outstanding.
And my wife, Joanne, the sunshine of my life, for her emotional
support, unswerving confidence, and all the love one man could imagine.
And my children, Dianna, Katherine, and Kenneth III, the radiant stars
of my life, ever propelling from me, yet forever close to my heart.
And my granny, whose divine love, grasp of spiritual truths, and faith
in God quickened in me a quest for knowledge and truth and instilled
in me, like three granite columns, faith in others, my Self, and God.
And last, Dr. Peter L. Giovacchini, my analyst, who freed me from
inner shackles and burdens, enabling the exquisite realization of two

major life goals, my marriage and the completion of my Ph.D.

ii

 

 

 

 

 

ACKNOWLEDGMENTS

My deepest gratitude is extended to the members of my dissertation
committee for their sponsorship of this research. To Dr. Norman R.
Stewart, my advisor and cxmmfittee chairperson, I am chiefly indebted
as he was the central figure in my studies as well as the dissertation
research. His continuing availability, openness, candor, guidance, and
constructive criticism were invaluable contributions to my project. To
Dr. William H. Schmidt, I owe much for his consultation on statistics
and methodology. On a personal level, his humor, interest, and encour-
agement were deeply appreciated. To Dr. Thomas Stachnik, I am indebted
for his insightful understanding of the goal of my project and singular
contribution to the study associated with data corroboration. To
Dr. Herbert M. Burks, I am greatly indebted for his editing and atten-
tion to detail, as well as professional grasp of nuances of style and
methodology. His commitment to my education is hopefully reflected in
this document.

The support of several institutions providing research sites for
the study was essential for the completion of the research. I am fore-
most indebted to Dr. Karl Willrich, Director of Adult Programs, and
Dr. Truman G. Esau, Medical Director, of Old Orchard Hospital who
wholeheartedly endorsed the research and continually supported my
efforts throughout the months of data collection. Notably, Dr. Karl
Willrich introduced this researcher to the clinical relevance of the

iii

family and social network, providing the initial impetus for the
direction of this research project and a strong formative influence

on my professional development. I would also like to recognize the
following persons in that they were the primary representatives of

the other participating institutions: Dr. Charles E. Turk, Medical
Director, and Doris L. Lefley, Program Coordinator, of the Day Treat-
ment Center at the Elk Grove/Schaumburg Townships Mental Health Center;
Dr. Anthony M. D'Agostino, Chairman of the Department of Psychiatry and
Medical Director of the Mental Health Unit and Alcohol Treatment
Center, and Sandra L. Spencer, Head Nurse, of the Niehoff Pavillion at
the Alexian Brothers Medical Center; and Dr. Edward J. Baranowski,
Executive Director, and Mr. Sheldon Gross, Director of the Day Care
Program, at the Main Center for Mental Health.

I am, of course, deeply indebted to the direct participants in
my research, namely the subjects, who shared their time and lives in
the continuing pursuit of scientific knowledge.

My employer, the Elk Grove/Schaumburg Townships Mental Health
Center, supported without question or equivocation the completion of
this project. The administrative and clinical staff were encouraging
and responsive to my needs for flexible scheduling, and unpredictable
and extended leave. I am particularly indebted to the members of the
Crisis/Emergency Team: Program Coordinator, Richard S. Bloom, A.C.S.N.;
and staff members, Florence Anderson, Robert J. Dziewior, Denise Elliot,
and Stephen Jzaska. These persons shared extra burdens and responsi-
bilities during my absences and endured by psychological aches and

pains.

iv

A word of appreciation seems very appropriate for my patients as
well. While attempting to cope with their emotional burdens and loss,
they were additionally confronted with my limited availability, unpre-
dictable absences, and at times, preoccupation.

I owe special gratitude to a very warm and committed friend,

John E. Laycock. As this document was completed out-of—state, neces-
sitating several commuting trips from Chicago, John repeatedly opened
his home without hesitation. He endured the unpredictability of last-
minute planning and cancellations, late arrivals, scheduling conflicts,
and precipitous departures. His hospitality and friendship were, and
are, of the highest quality. God's speed is wished to him in his new
endeavor, training for the priesthood.

Finally, I salute my wife and children, who dutifully endured
the many emotional and financial privations in our family life to
support my work and my need for solitude. Their patience and deter-
mination were sources of inspiration for me. Indeed, their task was

as demanding as mine.

TABLE OF CONTENTS

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

Chapter
I. INTRODUCTION AND REVIEW OF THE LITERATURE .......

Statement of the Problem ..............
Purpose of the Study ................
Definitions .....................
General Definitions ................
Definitions Referring to Structure ........
Definitions Referring to Content .........
Definitions Referring to Function .........
Definitions Referring to Emotion .........
Major Hypotheses ..................
Overview ......................
General Importance of the Study ...........
Social Network Analysis ...............
History and Development ..............
A Point of Controversy ..............
Conceptual Bases .................
Schizophrenic Isolation and Withdrawal: A
Theoretical Perspective ..............
The Isolation Hypothesis .............
The Withdrawal Hypothesis .............
Conclusion ....................
Principal Network Variables .............
Network Size ...................
Interconnectedness ................
Distance and Frequency of Contact .........
Uniplex and Multiplex Relationships and
Relationship Density ..............
Functional Indegree and Outdegree .........
Functional Symmetry ...............
Affective Indegree, Outdegree, and Symmetry . . . .
Investment Indegree, Outdegree, and Symmetry
Social Networks and Schizophrenia ..........
General Importance of the Social Network to
Schizophrenia ..................
Social Networks and Psychopathology ........

vi

Page
ix

xi

Page

Review of Major Studies .............. 43
Need for Replication ............... 51
Summary ....................... 56
II. METHODS ........................ 60
Subject Selection and Characteristics ........ 60
General Selection Rationale ............ 6O
Selection Criteria ................ 54
Subjects ..................... 66
Instrumentation ................... 70
Psychosocial Network Inventory .......... 70
Psychosocial Network Inventory, Modified ..... 71
Operationalization of Variables .......... 72
Consent Procedures ................. 78
Testing Procedures ................. 79
Normal and Schizophrenic Samples ......... 79
Corroborating Network Members ........... 86
Risk/Benefit Analysis ................ 88
University Authorization .............. 90
Research Questions ................. 90
Null Hypotheses ................... 92
Phase One ..................... 92
Phase Two ..................... 98
Research Design and Analysis ............ 101
Phase One ..................... 101
Phase Two ..................... 104
Summary ....................... 108
III. RESULTS ........................ 115
Results of Phase One: Corroboration of Self-Report . 115
Stage One ..................... 115
Stage Two ..................... 121
Results of Phase Two: Comparison Between the Social
Networks of Normals and Schizophrenics ...... 123
Stage One ..................... 123
Stage Two ..................... 125
Summary ....................... 127
IV. SUMMARY, CONCLUSIONS, AND IMPLICATIONS ........ 134
Summary ....................... 134
Conclusions ..................... 139
Stage One of Phase One .............. 139
Stage Two of Phase One .............. 140
Stage One of Phase Two .............. 140
Stage Two of Phase Two .............. 141

vii

Page

Discussion ..................... 142
Phase One ..................... 142
Phase Two ..................... 145
Implications for Future Research .......... 148
Significant Findings ............... 149
Suggestive Findings ................ 150
Absence of Findings ................ 152
Generalizability of Findings ........... 153
Clustering: Implications for Social Network

Analysis and Research ............... 155

Implications for Diagnosis and Treatment of the
Schizophrenic ................... 156
Synopsis of Major Contributions ........ . . . 158
In Retrospect .................... 160
APPENDICES . ......................... 163
A. INTERVIEW MATERIALS .................. 164
B. OPERATIONALIZATION OF VARIABLES ............ 187

C. PROPOSAL TO THE UNIVERSITY COMMITTEE ON RESEARCH
INVOLVING HUMAN SUBJECTS .............. 205
REFERENCES .......................... 228

viii

 

Table

10.

11.

LIST OF TABLES

Sample Characteristics .................

Major Tests of Significance for the Null Hypotheses in

Stage One of Phase One ................

Univariate Tests of Significance for the Multivariate
Null Hypotheses in Stage One of Phase One for the

Normal Sample .....................

Univariate Tests of Significance for the Multivariate
Null Hypotheses in Stage One of Phase One for the

Schizophrenic Sample

Major Tests of Significance for the Null Hypotheses

in Stage Two of Phase One ...............

Univariate Tests of Significance for the Multivariate

Null Hypothesis, Hoe, in Stage Two of Phase One . . . .

Univariate Tests of Significance for the Null
Hypotheses in Stage One of Phase Two: Main Comparison
Between the Social Networks of Normals and

Schizophrenics ....................

Major Tests of Significance for the Multivariate Null
Hypotheses Pertaining to Symmetry in Stage Two of

Phase Two .......................

Univariate Tests of Significance for the Multivariate
Null Hypotheses Pertaining to Symmetry in Stage Two

of Phase Two: Schizophrenic Sample ..........

Statistical Test Results for Stage One of Phase One:
the Analysis of the Corroboration of Self-Report

Data for the Normal and Schizophrenic Samples .....

Statistical Test Results for Stage Two of Phase One:
the Analysis of Sample Differences in Self-Report

Data .........................

ix

Page

116

118

120

122

123

124

126

127

128

130

Page

12. Statistical Test Results for Stage One of Phase Two;
the Main Comparison of Sample Differences Along
All Social Network Variables, Excepting Symmetry . . . 131

13. Statistical Test Results for Stage Two of Phase Two:
The Analysis of Symmetry for Each Sample, Normal
and Schizophrenic ............... 133

 

 

 

LIST OF FIGURES

Figure Page

1. Research Design for Phase One, Stage One, for the
Analysis of Corroboration of Self-Report Data for
Each Group, Normal and Schi20phrenic ...... . . . 103

2. Research Design for Phase One, Stage Two, for the
Analysis of Group Differences in Self-Report Data . . 105

3. Research Design for Phase Two, Stage One, for the
Main Comparison of Group Differences Along A11
Social Network Variables, Excepting Symmetry ..... 107

4. Research Design for Phase Two, Stage Two, for the

Analysis of Symmetry for Each Group, Normal and
Schizophrenic .................... 109

xi

CHAPTER I

INTRODUCTION AND REVIEW OF THE LITERATURE

The phenomenon of schizophrenia has long perplexed researchers.
Attempts have been made to relate its occurrence with various fac-
tors such as genetics, physiology, body chemistry, individual dynamics,
family process, and sociocultural influences. Each viewpoint has
had a degree of explanatory power and heuristic value; all, perhaps,
have shed light on the genesis/process of the schizophrenic disorder.
Accordingly, multitudes of theories have evolved out of the research,
making integration of the work a massive and seemingly endless pro-
cess.

In the present study, which focuses on the psychosocial aspects
of schizophrenia, a new concept and methodology developed in the field
of social anthropology will be used which shows promise in providing a
unifying framework for the psychological and social theories: social
network analysis. Social network analysis allows for the quantifica-
tion and assessment of differing levels of conceptualization, e.g.,
the personal, the familial, and the sociocultural; thus, the effects
of all investigated levels are recognized, permitting further inte-

gration.

Statement of the Problem

 

Previous investigations of the social networks of schizophrenics
have demonstrated relationships between social network variables and
the schizophrenic, suggesting associations between psychosocial fac-
tors and the schizophrenic disorder. Inchoative trends in the research
suggest that the social network of the schizophrenic tends to be
smaller than that CH the normal, nonpsychotic control. Additionally,
the relationships of the schizophrenic appear to be characterized more
by a lack of reciprocity (in particular, the schizophrenic being in a
dependent position) and a limited range of activities compared to the
control. Interpretations made from these results point out that the
schizophrenic appears to have a less unified and realistically smaller
system of social resources than the control. Furthermore, being that
this person is predominantly the dependent party in the social net-
work, this small network may be quite taxed to support this person.
Last, assuming that quality of relationships is related to the range
of activities shared, it would appear that the relationships of the
schiZOphrenic are less satisfying and enriching than those of the
control.

Although these investigations appear quite promising, the findings
are far from being conclusive because of frequent and notable incon-
sistencies in the results and the presence of a serious conceptual
error in the studies. The inconsistencies in the results prohibit a
strict relating of the specific, discrete social factors to the
schizophrenic disorder. An examination of the literature suggests

that the overall lack of concordance in the studies is related to the

 

following considerations: (a) inadequate sample descriptions and dif-
fering populations, (b) differences in conceptualization and opera-
tionalization of variables, (c) differing conceptualizations of the
social network under study, and (d) inconsistencies in the reported
data. Furthermore, a serious conceptual error calls the findings

into question. The assumption was made in three of the four pivotal
studies that the subjects' subjective analysis of the status of their
social network, i.e., the self-report data, was equivalent to the
actual status of the network. The investigators may have only been
interested in their subjects' perceptions (and such would be quite
important clinically), but inferences were drawn unequivocally to

the social network, as if self-report was indeed factual. Attempts
were not made to corroborate the reported data with more objective
measures or with reports of others in the network. Thus, the validity

of the results lacks credibility.

Purpose of the Study

 

The major purpose of this research study is to compare the social
networks of normals and schizophrenics, the goal being identification
of psychosocial factors that appear to be associated with the schizo-
phrenic disorder. In this study, the two groups, normal and schizo-
phrenic, will be compared along four major classes of variables
pertaining to social networks and relationships: structure, content,
function, and emotion. The structure-related variables convey the
basic morphological characteristics of the social network. The

content-related variables convey aspects of the content of the social

 

relationship, i.e., the types of shared activities. The function-
related variables reflect the transaction of support. And last, the
emotion-related variables assess the experiencing of affect in the
social network.

Also, a secondary purpose of the study is to evaluate the research
assumption, conceptually basic to much of the previous research, that
self-report is a sufficient and accurate indicator of the actual status
of the social network. To accomplish this task, the self-report of
five normal and five schizophrenic subjects will be compared with the
report of at least four of their network members on the status of
the shared relationship to corroborate the data and obtain some
estimation/measures associated with validity. This assessment will
be carried out prior to the main analysis as the meaningfulness of
the results depends upon it. If it is determined that the self-report
data are not truly representative of the actual status of the social
network, then the statistical results of the main analysis will not be
examined in relation to their more interactional and sociological
meaning, but rather to their strictly clinical significance, since the

observations do not correspond to a reality based on consensus.

Definitions

 

General Definitions

 

Family of origin. A social unit composed of maternal and/or

 

paternal figure, and their offspring and/or stepchildren only, in

which the focal person is an offspring or stepchild.

Focal person. The individual who is considered to be the focus
or point of origin for the analysis of a social network within this
study: The focal person is the subject, unless otherwise noted.

In the fields of sociology and social anthropology, the term ”ego"
is traditionally used to designate this focal person. General usage
of the term "ego" differs significantly from this formal definition
and is associated with psychoanalytic theory. Thus, the term ”ego” was
not chosen for usage in this research, as this study is expected to have
wide applicability and value outside the professional boundaries of
these two fields. Accordingly, to avoid initial confusion with the
more usual psychoanalytic usage of the term "ego" and possibly later
specious connotations, the term ”focal person” is being used.

Kin or kinship system. Individuals related by blood or marriage.

 

Nuclear family. A social unit composed of a maternal and/or

 

paternal figure, and their offspring and/or stepchildren only, in
which the focal person is a parent or stepparent.

Primary star. Those relationships (linkages in the social net-

 

work) that exist solely between the focal person and his/her social
network; interrelationships among the network members are thus

excluded.

Schizophrenic disorders. The group of disorders that share the

 

following essential characteristics as outlined by the American
Psychiatric Association, Diagnostic and Statistical Manual of Mental

Disorders, III (1980):

 

1. presence of certain psychotic symptoms during the active
phase of the illness, e.g., delusions, hallucinations,
poverty of and/or loosening of associations;

2. deterioration from a previous level of functioning in
social relations, self-care, and at work;

3. characteristic symptoms involving multiple psychological
processes including disruptions in content and form of
thought, perception, affect, identity, volition, relation-
ship to the external world, and psychomotor behavior;

4. onset before age 45; and

5. duration of at least 6 months.

(You are referred to the manual for an extensive discussion of the
disorders.) No differentiation relative to subtype was made.

Social network. Those individuals with whom the focal person

 

has an important personal relationship as defined by the focal person,
whether positive or negative in nature, with nuclear family, family

of origin, other blood relatives, relatives by marriage, friends,
neighbors, associates at church, work, school, etc. The usual defini-
tion of social network in the field of social anthropology applies to
a larger matrix of interrelated individuals. The given definition
strictly applies to a personal network or what is termed "immediate

network."

Definitions Referring to Structure

 

Adjacent density. This index refers to the proportion of link-

 

ages (relationships) in the network to the possible number of linkages
in the given network.

Interconnectedness. Interconnectedness refers to the relation-

 

ships that exist within the social network, including but beyond
those of the primary star. The index, adjacent density, assesses
this structure-related quality.

Structure-related variables. These variables convey the basic

 

morphological characteristics of the social network. The four
structure—related variables are size of the network, frequency of
contact, distance between the focal person and network members, and

adjacent density.

Definitions Referring to Content

 

Content-related variables. These variables convey the content of

 

the relationship between the focal person and the social network
(actually the primary star). In this study, content is strictly
defined as type of activity. Ten content-related categories (types of
activities) were selected with the intent of assuring that the areas
were inclusive of most activities: family, employment, romantic, con-
versational, social, recreational, fraternal, religious, political,
and volunteer activities. The content of the relationship (the types
of activities that characterize a relationship) is presumed to be

associated with quality within the relationship.

Multiplex relationship. A content-related variable character-

 

izing a relationship of more than one type of content, i.e., where the
focal person and the network member share more than one type of
activity.

Relationship density. A content-related variable expressing the

 

intensity and quality of relationships in the primary star of the
social network, i.e., solely between the focal person and network
members. (Note: The content of the other relationships within the
social network is not assessed, since the information would not be
considered reliable unless the others were examined.)

Uniplex relationship. A content-related variable characterizing

 

a relationship of only one type of content, i.e., where the focal

person and the network member participate in only one type of activity.

Definitions Referring to Function

 

Functional directionality. A quality of the functioning of a

 

social network that refers to the direction in which support flows
between the focal person and another in the social network. If sup-
port goes in equal measure between the focal person and the network
member, then the relationship is reciprocal and thus is character-

ized by functional symmetry. If the support goes in unequal measure

 

between the two, then the relationship is functionally asymmetrical.

Functional indegree. A function-related variable indicating the

 

degree to which functions are being served for the focal person by a

network member, i.e., the support the focal person is receiving.
The usage of this term is to be distinguished from the term as used
by Tolsdorf (1976), which refers to the numbg:_of functions served
for the focal person.

Functional outdegree. A function-related variable indicating

 

the gggy§e_to which the focal person is serving functions for a network
member, i.e., the support the focal person is providing for the network
member. The usage of this term is also to be distinguished from the
term as used by Tolsdorf (1976), which refers to the numbe:_of func-

tions that the focal person serves for the network member.

Definitions Referring to Emotion

 

Affective directionality. A quality of the emotion-related

 

variables that refers to the direction of the emotion experienced
between the focal person and the network member. If the type of emo-
tion between the focal person and the network member is identical,
then the relationship is reciprocal and thus is characterized by

affective symmetry. If the type of emotion between the two varies,

 

then the relationship is characterized by affective asymmetry.

Affective indegree. An emotion-related variable indicating the

 

types of feelings a network member experiences for the focal person
as perceived by the focal person. The range extends from all negative
to all positive feelings.

Affective outdegree. An emotion-related variable indicating the

 

types of feelings experienced by the focal person for a network member.

The range extends from all negative to all positive feelings.

10

Investment directionality. A quality of the emotion-related
variables that refers to the direction of the emotional investment
(literally, the strength of the emotion) between the focal person and
the network member. If the strength of the emotions experienced by
the focal person and the network member is equal, then the relation-
ship is reciprocal for strength of emotion, and thus is characterized

by investment symmetry. If the emotional investment varies between

 

the two, then the relationship is characterized by investment asym-

 

metry.

Investment indegree. An emotion-related variable reflecting the

 

strength of feelings (emotional investment) experienced by the network
member for the focal person as perceived by the focal person. The
range extends from weak to very strong feelings.

Investment outdegree. An emotion-related variable reflecting the

 

strength of feelings (emotional investment) experienced by the focal
person for a network member. The range extends from weak to very

strong feelings.

Major Hypotheses

 

The research study is intended to be an exploration of the inter-
personal and social processes operating in the schizophrenic disorder.
The specific goal is to isolate those properties that distinguish the
social networks of the schizophrenic person from the social networks
of the normal person. The four major hypotheses (grouped by primary

area) that identify the avenues of focus follow:

11

In terms of the structure-related or morphological variables, the
soc1a1 network of the schizophrenic person, as opposed to the normal
person, tends to be characterized by the following properties:

a.

Fewer network members; specifically, it is hypothesized that
the social network of the schizophrenic person is smaller
than the social network of the normal person.

 

A lesser proportion of interrelationships among network
members (less interconnectedness); specifically, it is
hypothesized that the social network of the schizophrenic
person is less interconnected than the social network of the
normal person.

 

A lower average distance between the focal person and the
network members; specifically, it is hypothesized that within
the social network, the members live closer to the schizo-
phrenic person than the members of the social network of the
normal person.

 

 

A relatively higher frequency of contact with the focal
person; specifically, the schizophrenic person tends to have
more contact per person with the network members than the
normal person.

 

In terms of the content—related variables (those dealing with the
nature of shared activities), the social network of the schizophrenic
person, as opposed to the normal person, tends to be characterized by
the following properties:

e.

f.

g.

A greater proportion of relationships in which only one
activity is shared (proportionately more uniplex relation-
ships); specifically, it is hypothesized that the schizo-
phrenic person tends to have relatively more relationships
than the normal person in which only one type of activity
is shared within the dyad.

 

A smaller proportion of relationships in which more than one
activity is shared (proportionally fewer multiplex relation-
shipS); specifically, it is hypothesized that the schizo-
phrenic person tends to have relatively fewer relationships
than the normal person in which two or more activities are
shared within the dyad.

A relatively smaller_prpportion of shared activities per
personTless relationship density); specifically, it is
hypothesized that the schizophrenic person tends to share
fewer activities per member of the social network than the
normal person.

 

12

In terms of the function-related variables (those dealing with the
transaction of support within the social network), the social network
of the schizophrenic person, as opposed to the normal person, tends
to be characterized by the following properties:

h.

A_greater degree of support provided by network members to
the focal person (more functional indegree); specifically,
it is hypothesized that the schizophrenic person tends to

receive more support per person from network members than

the normal person.

 

A smaller degree of supportyprovided by the focal person for
network members (less functional outdegree); specifically,

 

 

it is hypothesized that the schizophrenic person tends to
give less support per person to network members than the
normal person.

A lack of reciprocity (functional asymmetry); it is hypothe—

 

sized that whereas the normal person and the network members
tend to give and receive in equal measure, the schizophrenic
person tends to receive more from the network members than
he/she gives them; i.e., the schizophrenic person tends to
be dependent in relationships.

In terms of the emotion-related variables, broken down into two major
categories, affective and investment areas, the social network of the
schizophrenic person, as opposed to the normal person, tends to be
characterized by the following properties:

k.

A lesser degree of positive feelings from network members for

 

the focal person (less affective indegree); specifically, it

 

is hypothesized that the network members of the schizophrenic
person feel more negatively toward the schizophrenic person
than the network members of the normal person toward the nor-
mal person.

A lesser degree of positive feelings from the focal person
for the network members (less affective outdegree); spe-

 

cifically, it is hypothesized that the schizophrenic person
feels more negatively toward the network members than the
normal person does.

Affective asymmetry; specifically, it is hypothesized that

 

whereas the normal person and the network members tend to
feel similarly toward one another, the schizophrenic person
tends to feel more negatively toward the network members
than they toward him.

A greater degree of emotional investment of network members in
the focal person (more investment indegree); specifically, it

 

is hypothesized that the network members of the schizophrenic
person are more invested in the schizophrenic person than are
the network members of the normal person in the normal person.

13

o. A lesser degree of emotional investment of the focal person
in network members (less investment outdegree); specifically,
it is hypothesized that the schizophrenic person is less
invested emotionally in the network members than is the nor-
mal person.

p. Investment asymmetry; specifically, whereas the normal person
and the network members tend to be similarly invested in one
another, the schizophrenic person tends to be less invested
in the network members than they in him.

 

Before this analysis can be undertaken, the data from the subjects'
self-report must be examined to determine if they are in actuality an
accurate and sufficiently representative indicator of the functioning
of the social network. The following research hypothesis is posited:

Self-report on the status of the social network is adequately
corroborated on all designated measures by reports of the network
members in both samples: normal and schizophrenic subjects.
If the above hypothesis is not confirmed, then the following research
hypothesis is posited:
The difference between the self-report of the schizophrenic sub-
ject and the network members is larger than the difference
between the self-report of the normal subject and the network
members.
Overview

In the review of the literature, the following topics are covered
in a selective fashion to examine only those aspects relevant to the
purpose of this study, the social network analysis of normals and
schizophrenics: (a) importance of the study, (b) social network
analysis, (c) theoretical perspective--the concepts of schizophrenic
withdrawal and isolation, (d) key variables, (e) importance of the
social network to the schizophrenic disorder, and (f) relationships

between social network variables and psychopathology, in general, and

the schizophrenic disorder, in particular. The extremely disrupting

14

effect of the schizophrenic disorder upon the individual and

society is reviewed to demonstrate the general importance of the
study. Social network analysis is reviewed in terms of its history
and development, a point of controversy, and differing and inter-
related conceptualizations. Specifically, the review of history and
development indicates how the concept of the social network initially
emerged as a metaphor in the research to later develop into an analyti-
cal tool of wide applicability and value; the review of one controver-
sial point indicates that social networks have been viewed by some as
”informal residuals" unworthy of study, but by others as the key to

a new and important type of methodology in social science; and the
review of differing conceptualizations provides descriptions of the
different perspectives present in social network analysis. The review
and discussion of schizophrenic withdrawal and isolation is used to
develop a theoretical perspective of how the schizophrenic disorder

is associated with the social network. Based on this theoretical
framework, the review of the key variables provides an understanding
of their psychological and social implications which is crucial to

the analysis and interpretation of the results of the study. In the
next section, general importance of the social network to the schizo-
phrenic disorder, the components of the social network, i.e., mother,
father, family, extended family, and significant others, are reviewed
to demonstrate possible relationships to the schizophrenic disorder
and to examine the nature of those influences, thereby contributing

to the understanding of the role of the social network. The research

investigating relationships between social network variables and

15

psychopathology, in general, and the schizophrenic disorder, in par-
ticular, is reviewed to demonstrate the presence of associations
between social factors, and psychopathology and the schizophrenic
disorder, to identify the sources of inconsistency and potential
lack of credibility in the related studies, and last, to suggest

alternative, corrective methodologies.

General Importance of the Study

 

The schizophrenic disorder is a worthy area of study because of
its extremely disrupting impact upon the affected individual and
society. Generally, the schizophrenic disorder is recognized as the
most severe and debilitating psychopathological disorder. Day and
Semrad (1978) advanced the following:

The schizophrenic reactions are a group of diseases that cause

massive disruptions of thinking, mood, sensorimotor function-

ing, and behavior; they lie at the most severe end of the spec—
trum of psychopathology. Schizophrenics show a greater degree
of disturbance in intrapsychic function, character structure,
and interpersonal relationships than patients suffering from

any other disorder. (P- 199)

Arieti (1974) considered the disorder to be a psychosis which, in his
opinion, is generally accepted as a "severe or major psychiatric dis-
order" (p. 4). Reid (1975) stated regarding the schizophrenic dis-
order, "It is often catastrophic in effect, taking its toll during
the most productive years of an individual's life" (p. 299).

Not only is the schizophrenic disorder a serious psychiatric
problem for the affected individual, but it is also a critical concern

of major proportions societally. In 1969, Ullman and Krasner found

that the schizophrenic reactions compose the largest group of the

16

psychoses. Furthermore, they stated, “At the present time roughly

20 percent of all first admissions to psychiatric hospitals are cate-
gorized as schizophrenic. Even more important, roughly half of all
patients remaining in psychiatric hospitals are diagnosed as schizo-
phrenic" (pp. 356-357). Reid (1975) reported that approximately 1%
of the general population will be diagnosed as schizophrenic at least
once during their lifetime. In terms of incidence, Day and Semrad
(1978) in their review found that approximately 92,450 to 148,350 new
cases appear each year, and that between 494,500 and 1,010,500 are in
treatment for a schizophrenic disorder annually in the United States.
Last, they stated, "The total direct and indirect cost of the schizo—
phrenic disorder to the United States is estimated to be $14 billion

annually" (p. 207).

Social Network Analysis

 

History and Development

 

Social network analysis was developed in the field of social
anthropology as a methodology for the study of social relationships
(Hammer, Makiesky-Barrow, & Gutwirth, 1978; Whitten 8 Wolfe, 1973).
During the initial stages when the concept of social network was used
primarily as a metaphor, Hammer et a1. (1978) noted that the social
network generally referred to "interpersonal relationships which
crosscut the well-defined groups and sectors of pre-industrial
societies, giving a measure of integration or cohesion to otherwise
discrete segments" (p. 523). At this point, the fields of sociology

and anthropology were oriented toward the investigation of more formal

17

and well-defined groups than those determined by interpersonal
processes. Later, as interest grew, social network analysis evolved
into a more analytical tool of reseamtil which has been applied to a
wide range of social phenomena and problems, demonstrating its value
and applicability, e.g., conjugal roles (Bott, 1955/1977; 1957),
political movements (Gerlach & Hine, 1970), medical practices
(Coleman, Katz, & Menzel, 1957), etc. (You are referred to the
articles by Hammer et al., 1978, and Whitten 8 Wolfe, 1973, for a

more extensive review.)

A Point of Controversy

 

Throughout its early development, a controversy raged over the
relevance and importance of the social network concept (as the earlier
paragraph--above--might suggest). In examining the literature,
Whitten and Wolfe (1973) identified a dominant trend toward perceiv-
ing social networks as ”residuaZ--the relationships that remain
after the major structural relationships are dealt with" (p. 722).
Essentially, social relationships were not seen as important social
phenomena worthy of study, but rather as links or connections to fill
the voids between the formal units of social structure. The work of
Barnes (1954), Redfield (1956), Wolf (1966), and Boissevain (1968)
were quoted to identify this viewpoint. Other researchers, particu-
larly Mayer (1962, 1966), Mitchell (1969), and Gutkind (1965), tended
to regard the social network as a singularly important concept, allow-
ing for the integration of personal, social, and cultural variables,

documenting their independent action and interplay.

18

Conceptual Bases

Although social networks tend to be recognized now as both focal
and important, there is a great deal of variance among researchers
as how to proceed in the analysis of networks. Researchers' con-
ceptualizations tend to be based on the focus and needs of their
research. Five notable types of conceptualizations, relevant to the
present study, are reviewed below based on the following methods of
categorization: (a) set theory applied to categories of relation-
ships, (b) levels of linkage to the focal person, (c) zones of inti-
macy in relationships, (d) the objective and subjective nature/
perception of networks, and (e) clinical relevance of relationships.
Each contributes to a different perspective on the composition of
social networks.

One type of conceptualization is based on the categorization of
relationships in the social network by set theory. Wolfe (1970), for
example, has developed a "taxonomy of network concepts," based on the
structure and process operating within the social network. In his
schema, networks are first divided as to whether they are seen as
limited or unlimited. Limited networks are considered a subsection
to the total (unlimited) network and are based on some criterion
applicable throughout the total network. Five types of limited net-
works were described: (a) the personal set, limited to the links of
one person; (b) the categorical set, limited to links involving per-
sons of certain type of category; (c) the action set, limited to
links purposefully used for a specific end; (d) the role-system set,

limited to links involving a certain role system; and finally (e) the

19

field set, limited to links with certain content. The unlimited
network was conceived of without limiting criteria, and could be
expanded to include villages, societies, continents, etc. Thus
Wolfe's system is based on categorical arrangements within limited
or unlimited sets.

Hammer et a1. (1978) distinguished among three linkage levels
or orders of an individual's social network:

The immediate or personal network consists of the connections

linking a given individual with others and the connections

linking those individuals with each other;

The initia1 individuaI's second order network consists of the

connections linking the members of the immediate network with

their immediate networks; and

The extended network includes the further connections linking

these sets of individuals into larger populations. (p. 524)
Thus Hammer and her associates have developed a system which is keyed
into the structure of the social network.

Another type of conceptualization is based on zones of intimacy
and subjective importance in relationships. Boissevain (1974) has
articulated six levels or zones which radiate conceptually around the
focal person in concentric circles, constituting the person's first
order zone: The lower the number of the zone, the more intimate and
subjectively important are the persons to the focal person. The
first level, the personal cell, hypothetically contains the closest
relatives, such as one's nuclear family or family of origin, and a
few of the most intimate friends: Within the personal cell, the
contact and interaction is usually regular, frequent, and intense.

The second level, intimate zone A, is composed of very close rela-

tives and friends in which active contact and interaction are

 

 

20

maintained. The third level, intimate zone 8, consists of friends,
relatives, neighbors, and perhaps, co—workers who are emotionally
significant to the focal person, but the relations are more passive
than in intimate zone A. The fourth level, the effective zone, is
composed of those who are important "in a more pragmatic sense for
economic and political purposes and the logistics of daily life”

(p. 47). These relationships are maintained warm and congenial
because the focal person desires access to his/her friends for strate-
gic reasons; thus, their importance is based on instrumental, rather
than emotional, importance. The fifth level, the nominal zone, con-
tains persons who mean little to the focal person emotionally and
instrumentally, simply acquaintances. The names and faces of these
persons may be uncertain to the focal person. The sixth level, the
extended zone, is composed of those forgotten where recall would need
to be prompted. In conclusion, Boissevain noted, "Placement in these
zones is continually shifting: just as the transactional and emotional
balance in them is constantly shifting. From this perspective a per-
son's network is a fluid, shifting concept" (p. 48).

The objective and subjective nature of the social network, which
relates directly to analysis, is another way of conceptualizing.
Pattison (1977) described this method of dichotomizing the content
of a network. According to this approach, the objective network is
defined by the specific goal/purpose of the researcher, and this net-
work could be observed. For example, a work network is composed only
of those observed at work or a political network is composed of only

those observed or recorded as participating in a political campaign.

21

On the other hand, the subjective network can be only defined by the
focal person, as guided (not defined) by the criteria of the researcher.
As indicated by Pattison, the subjective network may include parts of
objective networks, e.g., family network, work network, friend net-
work, etc.

The fifth type of conceptualization relates to the composition of
the social network in terms of the clinical relevancy of relationships.
To this end, Pattison, de Francisco, Wood, Frazier, and Crowder (1975)
have defined a social network called the psychosocial kinship system.
Through their review, they supported the conclusion that the American
family is surrounded by a significant extended kinship system composed
of blood/marriage-related family and/or functional kin (those who
respond interpersonally as family) of friends, neighbors, work asso-
ciates, and others. Furthermore, they suggested that this identified
social group, not the family, is the basic and important social sys-
tem and accordingly, is associated with the individual's functioning.
(Unfortunately, Pattison et a1. did not distinguish by title between
the total group of the extended kinship referred to above and the
potential subgroup defined by meaningfulness and psychodynamic impor-
tance described below, as both are called the psychosocial kinship
system.) In relation to the significant extended kinship system,
they concluded that

(l) the psychosocial system does exist, (2) it exerts both

positive and negative sanctions and supports on the nuclear

family and the individual, and (3) it is a fundamental social

matrix that may prove to be either pathological or helpful
and therapeutic. (p. 1248)

22

Yet Pattison and his associates (1975) were not only interested
in investigating this social group, but also those within this group,
and perhaps, a group beyond it, who are meaningful to the individuals.
They stated,

Social psychologists have found that affinity by mere blood

or marriage does not define meaningful kin relationships and

that a causal definition of friend, neighbor, or work asso-

ciate does not define a significant psychodynamic relation-
ship. (p. 1248)

The clinical relevance of the "psychosocial kinship system, the
object of their study, was not only grounded in the social significance
of the network of family, relatives, friends, and various associates,
but in the psychodynamic significance these persons might have. In
the article, Pattison et a1. stated the following:

Our aim is to determine the psychodynamic social system that

comprises the primary social matrix of the individual. The

people in this matrix are related to the individual on the

basis of interaction and valued importance. Thus the rela—

tionships in the matrix are determined by social and psycho-

logical variables. Further, this social matrix represents

the functional kin group of the individual. Thus we term it

the psychosocial kinship system. (pp. 1248—1249)
In his personal communication, Pattison (1979, 1980) took an additional
step of inference. He argued that if the focal person has a cathexis
to another person (analytically speaking), then a linkage exists in
the psychosocial kinship system; thus, the interactional or social
component may be negligible or nonexistent. For this reason, a net-
work member may not necessarily be alive or have ever interacted
socially with the focal person, so long as the network member be
psychically alive. That is why he prefixed the term with "psycho-,"

indicating the psychological component and indexing its importance.

 

23

The social network conceptualization to be used in the present
study, chosen for its clinical importance, is the psychosocial kin-
ship system as articulated by Pattison and his associates (1975).
Contrasting this perspective with the other conceptualizations of
the social network, the psychosocial kinship system appears to be a
subjective network (Pattison, 1977), a product of the first four
zones of intimacy (Boissevain, 1974), a subgroup of the immediate or
personal network (Hammer et al., 1978) since the network consists of
connections linking the focal person with others and connections link-
ing the network members to each other, and in terms of structure, a
limited personal set as only one focal person per network exists
(Wolfe, 1970).

Schizophrenic Isolation and Withdrawal:
A Theoretical Perspective

 

 

The two hypotheses of schizophrenic isolation and withdrawal
constitute the basic theoretical perspective for the present study.
All of the social network variables can be interepreted in light of
these two hypotheses. In the following section of this chapter,
Principal Network Variables, the social network variables are examined
in relation to these hypotheses generally. Other more discrete and
measurable influences that appear to be associated with these hypothe-
sized phenomena of isolation and withdrawal, and that have been found
to be related to the variables under study, are also examined.

The hypotheses of schizophrenic isolation and withdrawal have
two basically differing but interrelated interpretations. The isola-

tion hypothesis posits that as a result of becoming isolated socially,

24

either the person becomes schizophrenic (a strict etiological inter-
pretation) or the person at high risk of schizophrenia tends to

develop this serious disturbance. This latter interpretation tends

to be supported by a related and complementary hypothesis termed the
social support hypothesis, which suggests that the presence and sup-
port of the social network is instrumental in preventing the occur-
rence of a psychotic break and maintaining the schizophrenic without
further and recurring episodes (Beck, 1978; Cobb, 1976). The with-
drawal hypothesis posits that as the disorder progresses, the affected
person tends to become withdrawn; i.e., an essential characteristic

of the schizophrenic disorder is the withdrawal from society. Each of
these propositions appears to be the converse of the other: The former states
that the schizophrenic condition arises or is precipitated by the
social phenomenon of isolation, while the latter states that with the
onset of the condition, the schizophrenic withdraws and the social
phenomenon of isolation evolves. To more clearly understand the nature
of these two influences of isolation and withdrawal, and the impact
they may theoretically have on the schizophrenic and the social net-

work of the schizophrenic, they are reviewed below.

The Isolation Hypothesis

The isolation hypothesis initially emerged as an etiological
theory, but very little empirical support was found; however, isola—
tion as a general social influence precipitating the schizophrenic
disorder has received support. Faris (1934) appears to have originated

the concept, stating that

25

Data from various sources appear to support the hypothesis

that.the"shut-in" or "seclusive" personality, which is gen-

erally considered to be the basis of schizophrenia, may be the

result of an extended period of "cultural isolation," that is,

separation from intimate and sympathetic social contacts.

(p. 155)
His evidence, in general, was based on two sets of impressionistic
findings: (a) the characteristic seclusiveness in the schizophrenic
disorder, and (b) anthropological findings relating the isolation
phenomenon to schizoid symptomatology. Kohn and Clausen (1955)
attempted to test the isolation hypothesis using the strict etiologi-
cal interpretation. Their data, focusing on the adolescent years,
were based on the retrospective impression of former, recovered
patients. As only one-third of the patients reported a period of
isolation in their adolescence, they wrote, "Our general conclusion
must be, then, that for the group here studied the data do not support
the hypothesis that social isolation in adolescence is a predisposing
factor in either schizophrenia or manic-depressive psychosis” (P. 272).
The etiological hypothesis was thus rejected. The investigators
interpreted that the social isolation was a result of inadequacies in
social relationships eventuating in alienation from peers: They
stated, "Thus, in terms of process, social isolation is to be viewed as
a sign that the individual's interpersonal difficulties have become
so great that he is no longer capable of functioning in interpersonal
relationships" (p. 273). This latter conclusion is more in line with
the withdrawal hypothesis. However, a year earlier in an epidemiologi-

cal study, Jaco (1954) interpreted the hypothesis in a less strict

fashion as a general influence operating socially. He predicted that

26

the incidence of the schizophrenic disorder would be higher in areas
characterized by more interpersonal isolation. His findings, based

on 13 variables which relate to network size and degree of social
contact (as well as employment stability and degree of mobility),
tended to support his interpretation of the isolation hypothesis.
Further, Jaco inferred that isolation acted as a precipitating factor.
Thus the research does not support a strict etiological interpretation
of the isolation hypothesis; however, an interpretation supporting a
relative degree of social isolation as instrumental in the process of
the disorder appears tenable.

The related hypothesis focusing on the relationship between
social support (loosely related to network size; i.e., with fewer
network members available, less support is possible) and the schizo-
phrenic disorder seems to support this conclusion that isolation may
act as a precipitant, and for that matter, affect the course and out-
come of the disorder. Day and Semrad (1978) have observed the type
of onset apparently associated with a chronic developing pattern.

They described the type of individual, at high risk for schizophrenia,
who nonetheless has been able to cope with his/her needs and with
society. Their gradual deterioration has been masked by a signifi-
cant dependent relationship where the inadequacies and needs were met.
With the loss of this person or persons, the deterioration becomes
evident. In general, the work of Birley and Brown (1970), Jacobs

and Myers (1976), and Jaco (1970) which focuses on stress and social
support tends to substantiate these observations. The support role

of the social network relative to course and outcome of the

27

schizophrenic disorder is cHscussed later in this chapter. (See
section, General Importance of the Social Network to Schizophrenia.)
Therefore, it appears that we can conclude that the degree of isola-
tion is associated with the onset, course, and outcome of the dis-

order.

The Withdrawal Hypothesis

 

Support for the withdrawal hypothesis consists of extensive
clinical observations and theoretical formulations. The withdrawal
of the schizophrenic person has been observed through the centuries,
but Freud (1914/1956) was the first to attempt a coherent explanation
dynamically. Basically, he viewed the schizophrenic withdrawal pro-
cess as mentally endogenous and representative of narcissistic regres-
sion. He expressed the view that the essential characteristic of
schizophrenia was the change in relationship with people. In Freudian
terminology, the schizophrenic person has withdrawn the libidinal
cathexes from others and returns to a state of primary narcissism,
totally withdrawn from others and cut off from social contact.
According to Fenichel (1945/1972), current analytic thought does not
accept such a pessimistic position, although the concept of narcis-
sistic regression is still accepted. In his opinion, the schizo-
phrenic can connect due to the presence of "residues of reality
relationships" and "the patient's spontaneous attempts at recovery“
(p. 447); therefore, the "retreat from society" is never really com—
plete. Other interpretations are reviewed by Arieti (1974): Fairbairn

(1952) interpreted the withdrawal as an escape from inner bad objects

28

as they are projected on society; Szasz (1957), as an expression of
an inner deficiency of preexisting objects which prohibits or limits
future interpersonal functioning; Becker- (1962), as a deficiency of
external objects, in that the person has been unable to find responses
which reduce problematic situations into habits, and resultingly,
develop meanings that will generalize into other interpersonal situa-
tions; Sullivan (1953, 1962), as a consequence of unhealthy and
anxiety-laden interpersonal relationships since infancy; and Cameron
(1947), as a result of a failure to acquire intelligible communica-
tion due to socially inadequate development. Arieti (1974) himself
viewed the process of desocialization as a reflection of a "concomi-
tant process occurring in the patient's inner reality“ (p. 345), in
which the schizophrenic person is regressing to earlier forms of
symbolization. Summarizing the process, he stated the following:
By giving up common or socially shared symbols he desocializes
himself. Although he may still use common symbols predominantly,
the symbols that are involved in his delusions and more intensely
experienced by him are his own paleosymbols. . . . As long as the
patient uses private symbols, at least for the life situations
that are most important to him, he cannot integrate socially.
However, even the most regressed schizophrenic will retain
expressions, words, and ways that belong to the interpersonal
world. A total abandonment of what is obtained from others is
not possible. (pp. 345-346)
Thus several theoreticians have observed the withdrawal phenomenon

and attempted to explain the process from a number of differing psy-

chological and/or social perspectives, supporting its presence.

Conclusion

 

The two processes of social isolation and withdrawalirfthe schizo-

phrenic can be seen as interacting and reciprocal. Arieti (1974) stated:

29

Inner life and external life are constantly interrelated:
abnormal external relations early in life trigger intrapsy-
chic mechanisms that disturb the inner 1ife. In its turn a
disturbed inner life causes alterations in relating to others.
A vicious circle thus originates. (P- 345)

Additionally, support for the interaction of social and psychological

events is generally provided by the theories of social psychology.

Principal Network Variables

 

The network variables discussed in this section relate directly
to the stated hypotheses of the study. Each is categorized and
evaluated as to the psychological and social implications associated
with each variable. Understanding of these variables is crucial to
the analysis and interpretation of the results: In contasting the
two samples, these variables can be perceived as describing the psy-
chological and social impact in type and degree due to the isolation

and withdrawal of the schizophrenic.

Network Size

 

Network size, a structure-related or morphological variable,
appears to be associated with the two hypotheses of schizophrenic
isolation and withdrawal. This variable, in fact, would be a direct
measure of the possible joint effect of the two hypotheses. It is
predicted that due to the hypothesized impact of either or both influ-
ences (an interaction) acting upon the schizophrenic, the social net-
work of the schizophrenic would be smaller than the social network of

the normal.

3O

Interconnectedness

Interconnectedness, a structure-related or morphological variable
which refers to the sparsity or density of relationships (links or
connections) within a social network, appears to be related with
length of and closeness in relationships, and stability and support in
social networks likewise affected by withdrawal and isolation.
(Although researchers vary widely as to how this variable is opera-
tionalized, it is considered to be a significant characteristic of
social networks [Whitten & Wolfe, 1973].) The interconnectedness of
close ties in a social network has been found to be positively corre-
lated with the mean duration of contact in these relationships, indi-
cating that forming highly interconnected networks takes considerably
more time than loosely interconnected networks (Hammer & Schafer,
1975). In a review of this article, Hammer et a1. (1978) concluded,
"Thus, for any individuals--including, of course, schizophrenic indi-
viduals--a relative lack of long-term ties would tend to preclude
anything more than peripheral participation in any interconnected
network of close relationships" (p. 525), suggesting that a social
network of relatively short duration would tend to be loosely inter-
connected and contain members that are barely involved with each other.
Further, a positive relationship has been found between interconnected-
ness and maintenance of contact with close relationships in social
networks during crisis with a psychiatric population (Hammer, 1961,
1963-64); specifically, patients whose closest relationships were with
persons who were closely interconnected with each other tended to

maintain their relationship during the crisis and ensuing

31

hospitalization. The implication is that the more highly intercon-
nected the social network, the more stable, durable, and, accordingly,
supportive (as the network does not deteriorate) the network tends to
be. Therefore, as a variable, interconnectedness appears to vary with
the length of the network's life, the degree of involvement and close—
ness in the network, and the stability and support present in time of
stress.

The present study is designed to test if the degree of intercon-
nectedness varies between the social networks of the schizophrenic
person and the normal person, possibly due to differences in the life,
degree of involvement, closeness, stability, and support characteris-
tic of their respective social networks. The social network of the
schizophrenic is predicted to be less interconnected as compared to

the normal.

Distance and Frequency of Contact

 

The distance and frequency of contact between the focal person
and the network members are likewise structure-related variables
that appear to be related to the influences of schizophrenic with-
drawal and isolation. Therefore, one would suspect that for the
schizophrenic person, the distance and frequency of contact between
the focal person and the network members would be lower than for the
normal person; however, as the size of the network of the schizophrenic
person is expected to be lower, the frequency of contact per person in
the social network of the schizophrenic person is expected and pre-

dicted to be higher.

32

It should be noted that an increased degree of contact within
the social network may not always be better for the schizophrenic.
For example, Vaughn and Leff (1976) have found that increased con-
tact between the schizophrenic person and a family characterized by
highly critical and emotional conflicts results in deterioration.
Thus frequency of contact is not an independently related variable
to the schizophrenic disorder.

Uniplex and Multiplex Relationships
and Relationship Densipy

 

 

Uniplex and multiplex relationships and relationship density are
content-related variables which appear to be associated with the
quality of the personal relationship between the focal person and the
network members; specifically, the more varied the activities within
a relationship, the greater the degree of closeness and satisfaction
(Bott, 1955/1977, 1957; Brim, 1974; Mayer, 1966). For example, one
would expect that a marital relationship in which several activities
(family, sexual, social, conversational, recreational, religious) are
shared would be closer, more enriching, and satisfying than one in
which only a few are shared. Accordingly, the uniplex relationship
would lack the quality of the multiplex relationship, and on the
network level, the social network with low relationship density would
tend to lack quality in its relationships generally, as opposed to the
social network of high relationship density. This relationship is
also obviously dependent upon other factors such as, and especially,

the feelings experienced within the dyads.

33

The present study is designed to determine whether quality
varies between the social networks of the schizophrenic and the normal
as measured by the uniplex and multiplex relationship and relationship

density.

Functional Indegree and Outdegree

 

Functional indegree and outdegree are function—related variables
which, by definition, identify the degree of affective and instrumen-
tal support given and received by the focal person in the social
network. Functional indegree and outdegree would appear to be
affected by the influences of social isolation and withdrawal,
respectively. Therefore, in terms of support, one would suspect that
the schizophrenic person would tend to receive and give less within
the context of the social network than the normal person. However,
the additional factor of size would affect this relationship in that
with the reduced size of the schizophrenic's social network, more per
person (more functional indegree) would be required to support the
schizophrenic person. Due to the withdrawal, though, the schizophrenic
person is still expected to give less to each network member, on the
average, than the normal person.

The present study is designed to determine if the degree of sup-
port given and received in the social network of the schizophrenic
differs from that of the normal person, which may be associated with
the influences of social isolation and withdrawal as well as network

size.

34

Functional Symmetry
Functional symmetry is a function-related variable in social
network analysis which relates to and has its meaning in the basic
network concept of reciprocity. According to Hollander (1967),
"interaction operates in terms of reciprocity; that is, the expecta-
tion that a benefit given will be returned" (p. 207). So in many
ways, the exchange or transaction can be seen as defining a relation-
ship. Supporting this conceptualization, Whitten and Wolfe (1973)
wrote:
For network analysis, the important aspect of exchange theory,
with its concept of reciprocity, is its demonstration that any
exchange can forge an interpersonal link, and interpersonal
links can connect individuals in series of communicative,
economic, manipulative, and other types of strands. (p. 731)
The degree of reciprocity in a relationship can be used to char-
acterize the nature of transactions and satisfaction in a relationship.
The work of Thibaut and Kelley (1959) indicated that the social
exchange should be rewarding for both parties for the interaction to
be satisfying and continue: this hypothesis can be used to examine
the following system of classifying transactions. Sokolovsky, Cohen,
Berger, and Geiger (1978) and Cohen and Sokolovsky (1978) focused on
the elements of directionality and measure of support within dyadic
relationships. In their system, which is defined relative to the
focal person, when aid flows from the focal person to a network mem-
ber, it is termed as "instrumental“; when aid flows in equal measure
between the focal person and a network member, it is termed as "recip-

rocal"; and when aid flows from a network member to the focal person,

it is termed as "dependent." Summarizing the above, then, relationships

35

that are characterized solely by instrumental or dependent transac-
tions, i.e., lacking in reciprocity, would be seen as unsatisfactory
and possibly as fragile.

The present study is designed to determine how the social net-
works of the schizophrenic person and the normal person differ rela-
tive to the concept of reciprocity and relate the differences to the
degree of satisfaction and fragility in the social network. The
links in the social network of the schizophrenic are expected to be
characterized by the dependent transaction.

Affective Indegree, Outdegree,
and Symmetry

 

 

Affective indegree, outdegree, and symmetry are emotion-related
variables in social network analysis, associated with the influence
of social isolation and withdrawal in schizophrenia in that they are
related to the degree of satisfaction and closeness in relationships.
Due to these two influences, it is expected that the schizophrenic
person cares less for and is cared for less (less positive feelings)
by network members than the normal person (low affective indegree and
outdegree). Further, it is also predicted that the schizophrenic per-
son feels less positively for the network members than they for him/
her (affective asymmetry), whereas the normal person and the network
members tend to feel similarly toward one another (affective symmetry).

The present study is designed to determine if affective indegree
and outdegree are both less positive in the social networks of the
schizophrenic person than of the normal person, and if affective

asymmetry exists for the schizophrenic person, while affective symmetry

36

exists for the normal person. These variables, associated with the
influences of social isolation and withdrawal, are related to the
degree of satisfaction in the relationship.

Investment Indegree, Outdegree,
and Symmetry

 

 

Investment indegree, outdegree, and symmetry are emotion-related
variables that appear to provide the closest estimate of the degree
of isolation and withdrawal of the schizophrenic person. Investment
indegree within the social network of the schizophrenic may initially
be expected to be lower than that of the normal, due to the phenomenon
of isolation; however, considering the smaller network size of the
schizophrenic, as well as the tendency to be dependent, it appears
that the network members of the schizophrenic are more invested than
the network members of the normal person. Investment outdegree may
reflect the relative degree of withdrawal of the schizophrenic from
the social network. Investment asymmetry is expected for the schizo-
phrenic due to relatively large degree of withdrawal in the schizo-
phrenic, especially since the network members are predicted to be in
the position of providing much more support to the schizophrenic than
the reverse. Notably, this investment imbalance suggests that at
this point in the development and life cycle of the target sample,
the role of withdrawal in the schizophrenic, endogenous influences,
predominates over the role of isolation, exogenous social influences,
within the psychosocial sphere. This statement certainly parallels

the thought of Arieti (1974). Last, investment symmetry is predicted

37

for the social network of the normal; i.e., the emotional investment
would be mutual.

The present study is designed to determine how the social net-
works of normal and schizophrenic persons differ relative to investment
indegree, outdegree, and symmetry and further, to relate the differ-
erences to the influences of isolation and withdrawal of the schizo-

phrenic.

Social Networks and Schizophrenia

 

In this section, the studies exploring the influence of various
components of the social network, and the associations between social
network variables and psychopathology and schizophrenia, are reviewed
to clarify the role of the social network in the schizophrenic dis-
order. The findings of the most relevant studies are summarized and
critically examined to identify weaknesses and sources of inconsis-
tency. Based on the review, the need for replication is supported,
and last, an alternative, corrective methodology is outlined.

General Importance of the Social
Network to Schizophrenia

 

 

Although only a few studies have examined the direct links
between social networks and the schizophrenic disorder (to be reviewed),
a plethora of studies has explored the relationships between the schizo-
phrenic disorder and the various components of the social network,
i.e., the mother, father, family, extended family, and significant
others. In this section, the impact of these components upon the

schizophrenic disorder 'Hs reviewed briefly to demonstrate their

38

importance. This review is not intended to be exhaustive, but rather
to identify major trends in the literature. This is the first step
in supporting the contention of this research study that interper-
sonal and social factors are associated with the schizophrenic dis-
order.

The relationship between the schizophrenic and his/her parents
was an early area of focus that was concentrated on parental character-
istics that tended to be associated with schizophrenic offspring.
Theorists, concerned primarily with etiology, initially directed
their speculations on the maternal relationship (Arieti, 1974),
especiallyFYjeda Fromm-Reichmann (1948), who appeared to have had
the most influence upon this movement through the development of the
heuristic phrase, "schizophrenogenic mother." This type of mother
was described in a totally negative way, e.g., over-protective, hos-
tile, rejecting, cold, distant, inconsistent, etc. Others in the
interpersonal school of thought such as Rosen (1962) and Sullivan
(1953) tended to follow her lead with one result being that the mother
was seen as totally responsible for the schizophrenic offspring;
therefore, the mother was seen in a very negative light. The work of
Lidz and his associates (see Lidz, 1973; Lidz, Cornelison, Fleck, &
Terry, 1957; Lidz, Terry, & Fleck, 1958), though strongly affected by
Fromm-Reichmann (1948L departed from this single-minded course by
focusing on the paternal relationship with the schizophrenic offspring
as well. Therefore, the father could now share the responsibility

with the mother for the schizophrenic person.

39

Later developments in the research widened this circle of influ-
ence to include the family as a whole. Early pioneers in family
research studied the family more as a unit and developed such
etiologically based concepts as pseudomutuality (Wynne, Rychoff, Day,
8 Hirsch, 1958), double bind (Bateson, Jackson, Haley, 8 Weakland,
1956), undifferentiated family-ego mass and triangulation (Bowen,
1960, 1965), and mystification (Laing, 1964). Based on the early
work and theory, present investigators are focusing on specific areas
such as communication (Ferreira 8 Winter, 1968; Mosher, Wild, Valcov,
8 Feinstein, 1972; Wild, Shapiro, 8 Goldenberg, 1975), interaction
(Mishler 8 Waxler, 1965, 1975; Riskin 8 Faunce, 1972; Waxler 8 Mishler,
1971), role structure (Wild, Shapiro, 8 Abelin, 1977), cognition and
perception (Riess, 1971), and emotional expression (Brown, Birley, 8
Wing, 1972; Leff, 1976; Vaughn 8 Leff, 1976). The results of these
studies demonstrate associations between schizophrenia and family
variables, yet as of the present none have been able to prove an
etiological hypothesis due to the difficulties with this type of
research (Reiss, 1976). Newer studies are either focusing on non-
etiological hypotheses (simply associations) or attempting to inte-
grate genetic hypotheses (Liem, 1980).

Relationships have been found between the extended family and
psychopathology, suggestive of possible relationships with the schizo-
phrenic disorder, and in at least one case, a direct connection between
characteristics of the extended family and schizophrenia. In 1962,
Bell determined that the extended family can affect disturbed families

differently than well families by supporting one side within a family

4O

argument, creating further schism, by serving to stimulate conflict,
by serving as a projection of family conflicts, and by becoming a
competing object for support or indulgence. Cohn and Talmadge (1976)
echoed this type of dynamic in a more recent clinical study. In a
study of the extended family and the schizophrenic disorder, Walsh
(1978) found that families that contained a schizophrenic offspring
experienced the death of a grandparent more commonly than families
with normal children or children with a nonschizophrenic, psychiatric
disorder. Based on clinical findings, Bowen (1961) has developed a
theoretical approach which is based on the impact of the extended
family from a historical perspective, termed a three-generational
theory. According to his approach, the psychopathology is passed
through the generations via the projection of emotional immaturity,
the end result being the schizophrenic person. It is interesting to
note that in a more recent review and update of his theoretical propo-
sitions, Bowen (1978) stated that a premature death in the family can
hasten the development of schizophrenia, a thought consistent with
the findings of Walsh (1978). Therefore, it appears that the char-
acteristics of the extended family may have an impact upon the func-
tioning of the family and quite possibly on the familial setting of
the schizophrenic person.

The relationship of the schizophrenic person with significant,
but non-blood/marriage related, persons has been shown to have bearing
upon the onset of the disorder (psychotic break), prognosis of the
disorder, response to psychological and psychotropic treatment, and

nature of posthospital functioning. In relation to the psychotic

41

onset, Birley and Brown (1970) and Jacobs and Myers (1976) found that
some disruption in the social network of the schizophrenic person,
namely a recent and stressful life event, preceded a symptomatic
relapse. (See also related articles by Allisi, 1969, and Feldman 8
Schertz, 1967, which suggest that treatment may be sought due to
ineffective networks.) In terms of prognosis, Strauss and Carpenter
(1972, 1977) in two studies utilizing two- and five-year follow-ups
found that the amount of social contact just prior to hospitalization
was a key predictor of outcome: These studies confirmed the earlier
work of Gittelman-Klein and Klein in 1969. Further, Kayton, Beck,
and Koh (1976) found that a "favorable convalescent environment," one
in which peers as well as parents or spouses "provided nonintrusive
emotional support through their attitude toward the patient and
through sensitively timed interpersonal contact” (p. 1270), was asso-
ciated with a positive outcome. Regarding psychotherapy, clinical
studies intimate that significant others in the social network may
have positive impact by supporting the family (Reuveni, 1979; Speck,
1967) or substituting for the family functions (Minuchin, 1974;
Minuchin, Montalvo, Guerney, Rosman, 8 Schumer, 1967). Dealing with
the interface of drug and psychosocial treatment, Gunderson (1977)
concluded after an exhaustive review that the interpersonal milieu

of the hospital affects the need for psychotropic medication. Last,
the reports of healing communities (rehabilitation centers) demon-
strated how the functioning of the schizophrenic person is facili-
tated by a supportive and responsive interpersonal environment

(Gunderson, 1980; Mosher 8 Keith, 1980) For clinical descriptions of

42

community support programs, see Almond, 1974; Beard, Malamud, 8
Rossman, 1978; Budson 8 Jolley, 1978; Dincin, Selleck, 8 Streicker,
1978; Fairweather, Sanders, Maynard, 8 Cressler, 1969). This review
also identified those who doubt seriously the efficacy of these
approaches (Klein, 1980; Torrey, 1980). So, much of the literature
suggests that onset, prognosis, response to treatment, and even remis-
sion of the schizophrenic disorder are influenced by the nonkin sig-
nificant others.

In closing and in line with the conclusions of Caplan (1974), it
appears that all components of the social network, i.e., family,
extended family, significant nonkin, can have impact on the life of
the schizophrenic, either in a protective and supportive way or in a

provocative and deteriorating manner.

Social Networks and Psychopathology

 

Direct relationships have been found between the presence of
psychopathology and the characteristics of social networks in three
clinical studies. As reviewed by Pattison, Llamas, and Hurd (1979),
Ratcliffe and Azim (1975) discovered that in comparing normals and
psychiatric patients the latter were less satisfied with their per-
sonal relationships, depended more on "involuntary" relationships with
relatives, and exhibited a lack of voluntary or friendship relation-
ships. A comparison of outpatient psychiatric patients and family
practice patients by' Silberfeld (1978) indicated that the social net-
works of psychiatric patients were impoverished in terms of number of

relationships and time spent and closeness within these relationships.

43

Contradicting the study by Ratcliffe and Azim (1975), Silberfeld

found that psychiatric patients seemed to have a greater proportion
of their relationships with friends. Kleiner and Parker (1974)
reported that alienation from the social network correlated positively
with three measures of psychopathology. In sum, the results of these
studies suggest that the type, quantity, and quality of relationships

within the social network appear to be linked to psychopathology.

Review of Major Studies

 

In the following section four studies are reviewed that have
explored the social network of the schizophrenic person as compared to
the social network of normal/nonpsychotic controls. Each is examined
in light of the major objectives of this study, namely, the identi-
fication and relevance of the four major classes of social network
variables: structure, content, function, and emotion. Their findings
are summarized in terms of consistencies and inconsistencies. Short-
comings and inherent methodological and design differences in the
studies which complicate the interpretation and integration of the
findings are presented in the section that follows this one.

Cohen and Sokolovsky (1978) have reported on their assessment
of the social network of schizophrenic persons who are living in
"single room occupancy" (SRO) hotels in Manhattan, New York, away
from family and kin. Three groups of tenants were identified for
study according to degree of pathology: (a) SR group--schizophrenia
with moderate or severe residual symptoms; (b) S group--schizophrenia

with minimal or no chronic symptoms; and (c) NP group--those with no

44

known psychotic history. The population studied included both males
and females, and whites and nonwhites. The data were collected
through combined participant observation, logs of daily activity,
extensive biographical interviewing, and the use of a "Network
Profile" questionnaire. The authors distinguished between "interac-
tional characteristics" of networks, the specific variables being
uniplex relations, multiplex relations, and directionality (the direc-
tion in which support flowed between the focal person and the network
members), and "morphological features” (the specific variables being
size and interconnectedness).

Significant differences were found in the Cohen and Sokolovsky
study (1978). In terms of structure, the results indicated that those
with a nonpsychotic history had social networks twice the size of
those who had a serious psychotic history (the SR group); no signifi-
cant differences were found between the two psychotic groups (the S
and SR groups): the total number of relations for the SR, S, and NP
groups were averages of 10.3, 14.8, and 22.5, respectively. However,
it was observed that a progressive increase in network size was asso-
ciated with decrease in psychopathology. Further, they observed a
trend, though statistically nonsignificant, toward a higher degree
of interconnectedness in the less disturbed groups (the S and NP
groups), indicating that the most disturbed group (the SR group) had
networks that were generally less intertwined. In terms of content,
significant differences were found in number of multiplex relations
between the NP group and the SR group, but not between the two schizo-

phrenic groups. The NP group had an average of 12.1 multiplex

45

relations, while the figures for the S and SR groups were 6.7 and
4.3, respectively. In terms of function, the authors concluded,

It was found that schizophrenics with residual deficits were

impaired in their ability to form instrumental relationships.

Furthermore, both categories of schizophrenics engaged in

significantly more dependent interactions than the NP group.

(p. 551)

Emotion as a social network variable was not examined. Not directly
associated with the objectives of this study, they also found that
rehospitalization was dependent upon the degree of psychopathology
and hotel network size, the latter being more significant with the
schizophrenic group with minimal or nonchronic symptoms (the 5 group),
suggesting that for those without severe symptomatology of overwhelm-
ing proportions the size of the social network can play a mediating
role.

Tolsdorf (1976) compared 10 male, hospitalized first-admission
psychiatric patients and 10 male, hospitalized medical patients at a
Veterans Administration Hospital. Those in the psychiatric sample were
all diagnosed as schizophrenic. Data were obtained through a 66-item
interview conducted by the author. The methodology entailed both
quantitative and qualitative analyses. The quantitative data included
structural variables, e.g., size and interconnectedness, and content
and function variables, both paralleling Cohen and Sokolovsky's (1978)
interactional characteristics. The perception of the social network
from an emotional perspective was evaluated through the qualitative data.

Significant results were also found in the Tolsdorf (1976) study,

but only in relation to the content and function network variables.

In terms of structure, no significant differences were found in the

46

total size of the networks, although a trend was noticed with the
medical/nonpsychotic group being larger: the average network sizes
of the medical and schi20phrenic groups were 37.8 and 29.8, respec-
tively. No differences were found in adjacent density, the measure
of interconnectedness. In relation to content variables, the medical
group had a significantly higher number and proportion of multiplex
relationships: the total average multiplex relationships for the
medical and psychiatric groups were 16.6 and 5.4, respectively, and
proportioned average multiplex relationships for the medical and
psychiatric groups were .44 and .21. Further, significant differ-
ences were found on relationship density: average of 1.53 for the
medical group and average of 1.24 for the psychiatric group. In

terms of function in the networks, the psychiatric group had fewer
functional members within the social network, but these members

served several functions; thus, the support was concentrated in a
limited few. On the other hand, the network of the medical group

was characterized by more functional people overall who provided fewer
functions apiece, so the supportive functions were spread out more
evenly across the network. Additionally, the psychiatric group tended
to have asymmetric, dependent relationships with network members, and
the medical group, symmetric or reciprocal relationships. Consider-
ing a possible interaction, the author hypothesized that the psychi-
atric group tended to limit their multiplex relationships to func-
tional people in their network, while the medical group had multiplex
relationships with nonfunctional people too; therefore, the activities

of the psychiatric group were limited primarily to functional people,

47

i.e., those vnu) support them. From the quantitative data, Tolsdorf
(1976) made the following interpretative statement in conclusion
(Inferences regarding families examined in the study and contained in
the statement below were not explored in this review as they do not
pertain to the results of the present study.):

In summary, the psychiatric subjects reported fewer intimate

relationships with their network members in a network that was

more heavily dominated by family members, where functional

people were in a more controlling and dominant position, and

where overall there were relatively fewer but more powerful
functional people in the network. The medical subjects, on the
other hand, reported more intimate relationships with more people
in a network that was less dominated by family members and where
functional people were on an equal standing with subject in the

exchange of support, advice, and feedback. (PP. 412-413)

The interpretation of the qualitative data generally suggested that
members of the psychiatric group had a negative emotional perception
of the helpfulness and character of their social networks, and this
perception influenced their ability to seek support and utilize the
resources of their networks. Members of the medical group uniformly
had a positive perception of their networks, which appeared to facili-
tate the seeking of support and use of resources.

Pattison, deFrancisco, Wood, Frazier, and Crowder (1975) used
the Psychosocial Network Inventory (described in Chapter II) to
analyze the social networks of normals, neurotics (not pertinent to
this studY). and psychotics. No statements were made regarding the
sexual composition of the samples. The data were analyzed along
five variables: (a) degree of interaction, (b) type of emotion,

(c) strength of emotion, (d) instrumental base, and (9) symmetry.

Except for size and a social connectedness ratio (associated with

interconnectedness), the variables were not operationalized.

48

The results of the study by Pattison et a1. (1975) are suggestive
of associations between social network characteristics and schizo-
phrenia, but as it is a descriptive study based on impressionistic
findings, no statistical validation of any kind, "statistically sig-
nificant results" cannot be reported. In terms of the structure of
the social networks, it was found that the normal sample was composed
of 20 to 30 people whereas the psychotic sample was composed of 4 to 5
people, mostly family, so impressionistically the normal sample
appears much larger than the psychotic sample. Further, using the
social connectedness/unconnectedness ratio, it was determined that
approximately one-half to two—thirds of the network members knew
one another, while the rest did not. For the psychotic network,
approximately nine-tenths of the network knew one another. So one
could impressionistically conclude that the social network of the
psychotic was more interconnected than those of the normal. Again,
hard data were not provided on the other variables; however, it was
noted from a review of the data that the relationships appeared to
be consistently ambivalent and nonreciprocal. Pattison et a1. con-
cluded with the following provocative comment: "In other words, the
psychotic is caught in an exclusive small social matrix that binds
him and fails to provide a healthy interpersonal matrix" (p. 1249).

In a more recent study directed by Pattison (Pattison, Llamas,

8 Hurd, 1979), he has modified the results of his earlier research.
Again in this study the sample was not described in terms of composi-
tion and crucial variables were not operationalized. Furthermore,

what hard data were available were not subjected to statistical

49

validation. In regard to the results, Pattison and his associates
indicated that the psychotic network is composed of 10 to 12 network
members, whereas the normal network is composed of approximately

25 members. Not of direct relevance to this study, he found that in
line with his previous study the network of the psychotic group was
populated predominantly by kin/family members. The relationships
were characterized from a review of the data as totally intercon-
nected, negativistic, conflictual, ambivalent, and highly asymmetric.
Pattison et a1. concluded of the psychotic, "The subject is caught in
and tyrannized by a collusive closed system, with few links to the
larger communities of relationships" (p. 66/481).

This review of the four major studies demonstrates the presence
of consistencies and inconsistencies in the reported data. In terms
of network size, the results of the four studies differ on the exact
figure with the range of reported figures extending from 4 to approxi-
mately 30; however, regardless of the average figure for network size,
consistently it was found in three studies (Cohen 8 Sokolovsky, 1978;
Pattison, deFrancisco, Wood, Frazier, 8 Crowder, 1975; Pattison, Llamas,
8 Hurd, 1979) that the social network of the schizophrenic person tended
to be smaller than that of the normal person. In terms of intercon-
nectedness, the results of the studies again vary a great deal.
Tolsdorf (1976) found no differences, while Pattison et a1. (1975,
1979) reported a high level of interconnectedness and Cohen and
Sokolovsky (1978) distinguished a trend indicating that the degree of
interconnectedness fell significantly for the schizophrenic groups,

the opposite of the previous findings of Pattison et a1. (1975, 1979).

50

In terms of multiplex relationships, the investigators who examined
this phenomenon (Cohen 8 Sokolovsky, 1978; Tolsdorf, 1976) tended to
agree on their basic findings and figures: fewer multiplex relations.
In terms of relationship density, Tolsdorf (1976), the only investi-
gator to examine this variable, found significant differences. In
terms of the transaction of support within the network, the studies
reach consensus in that the schizophrenic tended to form more dependent
and asymmetric relationships than is true for the normal person.
Tolsdorf (1976) further explored this finding, concluding that those
who support the schizophrenic are also those with whom the schizo-
phrenic person shares most of his activities; these supportive persons
also appeared to be a limited subgroup within the larger network. In
terms of the emotion variable, it is unfortunate that the investigators
(Pattison, deFrancisco, Wood, Frazier, 8 Crowder, 1975; Pattison,
Llamas, 8 Hurd, 1979; Tolsdorf, 1976) ink) examined this phenomenon
did not use empirical measures, but their qualitative results tend to
agree in essence, namely that schizophrenic persons would have a more
negative view of their social network. Using only those results in
which some degree of consensus was found, it appears that one can ten-
tatively conclude that as compared to the normal person, the schizo-
phrenic person tends to relate in a more dependent and asymmetrical
fasion in a social network that is characteristically smaller, marked
by fewer multiplex relations, and perceived in a negative way emo-

tionally.

51

Need for Replication

In reviewing the research on the social networks of schizo-
phrenics, two major factors led this researcher to consider the need
for replication of the previous work: (a) the importance of the
research, and (b) confusion around the interpretationauulintegration
of the previous findings. (Statements will be made in the final para-
graph of this section as to how this study will be conducted.) First,
the social network research is quite promising as it has revealed some
findings which are consistent across studies, possibly shedding light
upon the social factors operating in schizophrenia (and, for that
matter, in other psychopathological populations). Thus, due to the
potential importance of these findings, replication is essential.

Second, the findings of the previous work are difficult to
interpret and integrate for five basic reasons: (a) inadequate sample
descriptions and differing populations, especially those factors
relating to residence with family; (b) differences in conceptualiza-
tion and Operationalization of variables; (c) differing conceptuali-
zations of the social network; (d) inconsistencies in the reported
data; and (e) importantly, a crucial methodological problem in making
inappropriate inference regarding the characteristics of social net-
works from unvalidated self-report.

Relative to sample characteristics, Pattison et a1. (1975, 1979)
did not give adequate sample characteristics. The studies of Cohen
and Sokolovsky (1978) and Tolsdorf (1976) have more adequate descrip-
tions of their samples, but they differ in sexual composition and

especially, residential status. In relation to the latter, residential

52

status, the three studies except the one supervised by Cohen and
Sokolovsky appear to have focused on schizophrenics living with
family; Cohen and Sokolovsky only utilized subjects living away

from families. Only Cohen and Sokolovsky acknowledge the implica-
tions of this sample criterion. Last, it is interesting to specu-
late that the differences in interconnectedness among the studies of
Cohen and Sokolovsky and Pattison et a1. may relate to this differ-
ence.

Integration of the work is also difficult due to the differences
in the Operationalization of interconnectedness of the social network,
and vague conceptualization of the content variable. To operational-
ize interconnectedness, Pattison et a1. (1975, 1979) used a social
connectedness/unconnectedness ratio. Tolsdorf (1976) used adjacent
density, and Cohen and Sokolovsky (1978) used density and degree,
two related variables (the latter has been used to correct for the
tendency for larger networks to produce low density systems). In
their conceptualization of content, Tolsdorf and Cohen and Sokolovsky
appear to have included distinct and differing concepts under the
content category, confusing its meaning and contaminating its signifi-
cance: Tolsdorf's definition of the content variable included areas
that relate to roles and ways of relating within the network, in
addition to areas that relate directly to activity, the apparent focus
of the content category; and Cohen and Sokolovsky assessed functional
properties and supportive roles as well as types of activity within

their content variable.

53

Notable inconsistencies in the data are found in the size and
interconnectedness of the social network, as well as other variables.
(You are referred to the earlier discussion of this subject.)

The procedures used to collect data suggest differing concep-
tualizations of the social network which compounds the task of inte-
gration of findings. Pattison et a1. (1975, 1979) used a specific
network termed the psychosocial kinship system, which is composed of
all those important to the subject, e.g., family members, relatives,
friends, neighbors, co-workers, etc., the basic criterion being per-
sonal importance to the subject. According to his personal communi-
cation (Pattison, 1979, 1980), it would contain all those of
psychodynannc significance,so anyone real or fantasied, or dead or
alive, could be included. Referring to the earlier reviewed concep-
tualizations, the psychosocial kinship system appears to be personal
network (see Wolfe, 1970), subjective in nature (see Pattison, 1977),
containing only the immediate or personal network level of linkages
(see Hammer et al., 1978), and being a product of the first two and
possibly third levels of intimacy (see Boissevain, 1974). The social
network conceptualization chosen by Tolsdorf (1976, 1980), on the
other hand, is much different because he focused on a much larger
group of alternative persons and used a more objective model. His
primary goal was to identify the largest possible group of persons
available in a realistic sense to the schizophrenic person so as to
advance some statements regarding the resource capabilities of the
social network; therefore, personal importance would appear to play

a secondary role. Of course, deceased persons would be excluded as

54

members of the social network as he conceived it. Again in relation
to the reviewed conceptualizations, the social network as chosen by
Tolsdorf (1980) was more objective in focus, although the subjective
component was high since he relied on self-report (see Pattison, 1977),
consisted of the immediate and second-order networks (see Hammer et al.,
1978), may have focused on personal as well as categorical sets--an
emphasis being on resource capabilities of the network-~(see Wolfe,
1970), and may have been directed on including all zones of intimacy,
or at least the first four (see Boissevain, 1974). Cohen and
Sokolovsky (1978) have most clearly defined their social network,
basing their conceptualization on subjective and objective data; the
networks were developed from participant observation, logs of daily
activity, extensive biographical interviewing, and the use of a
"network profile" questionnaire. Networks were broken down into

those existing outside of and inside of the SRO hotel. Their con—
ceptualization of the social network was based on subjective and
objective perspectives (see Pattison, 1977), conceived as a personal
set (see Wolfe, 1970), limited to the level of the immediate or per-
sonal network (see Hammer et al., 1978), and may include those from
the first four levels of intimacy (see Boissevain, 1974). Thus it

is seen that each researcher has from differing perspectives developed
differing conceptualizations which inhibit, but when recognized may
facilitate integration of their results. For example, the differing
conceptualizations used in the studies would certainly affect the
figures obtained for network size and, accordingly, account for the

large variability in results.

55

Last and importantly, this study is undertaken to analyze a
methodological problem stemming from reliance on unsubstantiated self-
report data. Pattison et a1. (1975, 1979) through an inventory, and
Tolsdorf (1976) through an intensive interview, assessed the subjects'
perception of their social network. They assumed that their data
accurately reflected the characteristics of the social networks. As
has been stated earlier, this is a dubious assumption, for their data
reflect only the subjects' opinion of their networks unless otherwise
proven. This was not true for the study of Cohen and Sokolovsky
(1978), who used subjective and objective measures; their analysis
may therefore present data closer to the "actual" status of the social
network.

The following steps have been taken in the present study to
respond to the difficulties noted above. First, the sample is
thoroughly described and delimited demographically, and specifically
limited to the 18- to 40-year-old age group. Second, only subjects
residing with families are included in the samples. Third, the
methodology utilizes the previous sound research techniques for
assessment of interconnectedness; and in relation to content, this
variable is strictly and categorically defined as the set of shared
activity. Fourth, to reduce confusion it is stated that the con-
ceptualization of this study is that of the psychosocial kinship
system (Pattison et al., 1975, 1979). And fifth, the present study
is designed to corroborate the subjects' perception of their social
network with the report of network members as a measure of validity,

using the same assessment instrument, the Psychosocial Network

56

Inventory, Modified. Although it seems inherently inconsistent to
choose an instrument focusing on subjective perceptions to investi-
gate objectivity of response, it is consistent with the notion in
perceptual theory that all of human existence is subjective, and as
consensus is reached, the phenomenon becomes more "objective." The
test then is to determine how able the normal and schizophrenic per-

sons are to reach consensus with their relative social networks.

Summary

The major purpose of this research study is to compare the social
networks of normals and schizophrenics, the goal being identifica-
tion of psychosocial factors that appear to be associated with the
schizophrenic disorder. The psychosocial kinship system, the particu-
lar social network conceptualization used in this study, was chosen
for its clinical significance. The two groups are compared along
the four major classes of social network variables: (a) structure,
the basic morphological characteristics; (b) content, the nature of
shared activities; (c) function, the transaction of support; and
(d) emotion, the experiencing of affect.

A secondary purpose of this study, yet of paramount importance
in the study, is to evaluate the research assumption conceptually
basic to much of the previous research, that self-report data are a
sufficient and accurate indicator of the actual status of the social
network.

Social network analysis was reviewed. Although it was found

that the concept of the social network evolved as a metaphor, later

57

the concept began to be regarded as a singularly important concept
with wide applicability and methodological value which has allowed
for the integration of personal, social, and cultural dimensions,
documenting their importance and interplay.

The theoretical concepts of schizophrenic withdrawal and isola-
tion were reviewed and were found to have joint and interacting
impact on the social network of the schizophrenic. They represent
the theoretical basis of this study. In light of the influences of
schizophrenic withdrawal and isolation, the network variables were
reviewed and evaluated as to their psychological and social implica-
tions; based on this information, predictions were made. In terms of
the structure-related variables, it appears that the schizophrenic,
as contrasted with the normal, has a social network that is smaller
and less interconnected, and network members that live closer and
participate in more frequent contact. In terms of the content-related
variables, it appears that the schizophrenic, as contrasted with the
normal, has a greater proportion of relationships in which only one
activity is shared, a smaller proportion of relationships in which
more than one activity is shared, and overall, a relatively smaller
proportion of shared activities per network member. In terms of the
function-related variables, it appears that the schizophrenic, as
contrasted with the normal, within interpersonal relationships with
the network members receives more and provides less support. Further-
more, the schizophrenic appears to generally receive more support
from network members than is given in return, placing the schizophrenic

in a functionally asymmetrical, nonreciprocal, and dependent position.

58

In terms of the emotion-related variables, it would appear that the
schizophrenic, as contrasted with the normal, is the object of less
positive (more negative), yet stronger (more emotionally invested)
feelings, and has less positive (more negative), and weaker (less
emotionally invested) feelings. Overall, the schizophrenic appears
to have more negative feelings for and to be less invested in the
network members than they in her or him, demonstrating an emotionally
asymmetrical and nonreciprocal relationship with the social network,
and a withdrawn and/or isolated position.

The four major studies investigating the relationships between
the schizophrenic disorder and social networks demonstrated consis-
tencies in their findings tending to support some of the above predic-
tions, as well as inconsistencies. Consistent findings in relation
to the normal population compatible with the predictions included the
following: (a) smaller network size for the schizophrenic; (b) a
smaller proportion of relationships in which more than one activity
is shared for the schizophrenic; (c) relatively more support from
network members for the schizophrenic; (d) asymmetrical, nonrecipro-
cal, and dependent relationships for the schizophrenic; and (e) nega-
tive emotional perception of the social network for the schizophrenic.
(One investigator found that the relationships were characterized by
fewer activities per person within the social network of the schizo-
phrenic.) Inconsistent findings were associated with the actual network
sizes and interconnectedness, the former possibly related to differences
in conceptualization of the social networks and the latter to differ-

ences in the familial characteristics of the populations sampled.

59

The findings of the four studies were, however, difficult to
interpret and integrate for five reasons: (a) inadequate sample
descriptions and differing populations, (b) differences in concep-
tualization of and Operationalization of variables, (c) differing
conceptualizations of the social networks under study, (d) inconsis-
tencies in the reported data, and (e) importantly, lack of validation
of the self-report data. The present study is designed to respond to
the difficulties noted in the previous research allowing for a clari-

fication of the findings in this area of study.

 

CHAPTER II

METHODS

Subject Selection and Characteristics

 

General Selection Rationale

 

Research investigating significant differences between popula-
tions is usually designed to promote comparability along certain dimen-
sions that could affect outcome measures. The goal is to reduce
initial differences between the samples on these relevant dimensions
that are unrelated to the basic distinctions between the populations;
accordingly, additional precision is obtained in the statistical
analysis of the data. In social science research, the influences of
the following variables may be controlled: (a) age; (b) sex; (c) race;
(d) marital status; (e) educational level; (f) vocational level;

(9) personal economic status (PES), for the purpose of this study,

an indicator of the individual's ability to be self-supportive;

(h) socioeconomic status (SES), for the purpose of this study, an
indicator of the family's economic level; and (i) cultural background.

The following discussion presents the rationale used in deter-
mining the relevance of the variables noted above and the procedures
used to control for the potential untoward effects, if appropriate.

Due to the impairment in the social functioning of the schizo-
phrenic (see Review of Literature), the variables, educational and

vocational levels, and personal economic status (PES), did not appear

60

61

appropriate for control. These variables are directly affected by
the schizophrenic process, by definition. Attempting to insure
comparability, for example, by matching along these three variables
is to exclude a priori differences inherent in the two populations.
Such "equalizing" procedures would call into question the diagnosis
of the schizophrenic sample if the subjects were functional in these
areas. The mental health of the normal sample would likewise be
questionable if the subjects were dysfunctional in these areas. To
some extent, this line of reasoning applies alsotxlthevariable,nwrita1
status, since the schizophrenic tends not to have the complex skills
needed to initiate and maintain the marital union (this variable
will also be considered below).

Although studies have shown that the lowest socioeconomic class
has a significantly higher incidence of schizophrenia (Dohrenwend 8
Dohrenwend, 1969), further exploration of the methodology and findings
shows no correlation between SES, as defined in this study (the
family's income level), and incidence of schizophrenia; however,
what is being termed PES in this study appears to be related to inci-
dence of schizophrenia (Dunham, 1965; Goldberg 8 Morrison, 1963;
Hare, 1956a, 1956b). An example would be the schizophrenic who has
difficulty maintaining employment (low PES), but who comes from a
wealthy family (high SES); herein lies the basic distinction.
Dohrenwend and Dohrenwend (1974), notably, still consider this issue
unresolved.

Although SES was not considered related to schizophrenia, it

is a significant variable in social research of schizophrenia

62

(Myers 8 Roberts, 1959), and as such, appears an appropriate variable
for control. The remaining variables of age, sex, race, and cultural
background appear appropriate for control for the same reason. Two
procedures were implemented which have impact on assuring compara-
bility. First, the samples were limited to residents of a specified
geographic locale, constraining the range of SES and differences in
cultural background. Both samples were drawn solely from the north
and northwest suburbs of Chicago. Second, subjects were chosen who
would tend to provide a representative cross-section of the population
of the designated locale, thereby allowing for the development of
samples that would be reflective of the population along the above
dimensions. The normal sample was obtained from volunteers found

in a variety of settings, i.e., employment, residential, and religious,
insuring a cross-section of the identified locale. The schizophrenic
sample was also developed from a variety of settings, as no one
psychiatric institution (public or private) services all strata of

the locale clinically.

An inherent difficulty with this type of research is that sub-
jects are not strictly selected, randomly or otherwise; they volunteer.
Therefore, sample characteristics cannot be rigidly controlled.
Research sites can be selected that contain the appropriate popula-
tion or subgroup of same for the study; however, after the qualified,
potential participants are identified within the setting, they are
invited to participate on a voluntary basis. Their involvement is
not fixed; it is a variable. Those who eventually become subjects

and follow through with the testing are a subgroup of the identified

63

original group (this process was essentially identical for both
samples). Each sample, therefore, becomes a discrete entity with
special characteristics that hopefully are representative of a cross-
section of the population and are comparable along the salient dimen-
sions of age, sex, race, SES, and cultural background, given a certain
sample size.

Another method to deal with the difficulty of samples that are
potentially skewed along salient dimensions is to increase sample
size. While it would be advantageous to have very large samples,
the following practical issues made the proposed sample size goals
difficult to meet. First, the voluntary nature of the procedure
proved to be a significant hurdle: Although exact figures were not
kept, it was calculated that approximately one-half of the schizo-
phrenics asked to participate declined. Several from the schizophrenic
population who refused to participate were characterized clinically as
distinctly more hostile or paranoid than participating subjects. Inso-
far as this subgroup was not studied, its composition was not clearly
determined. Other possible influences determining the negative response
may have included the potential time commitment of the interview and
the instrument's exploration of the personal sphere which may have
appeared threatening. In specific relation to the schizophrenic sample,
it was found that securing institution approval and support for social
science research was a generally difficult, complicated, and lengthy
process. Furthermore, obtaining individual physician authorization
for the project created lengthy delays. Legal and technical considera-

tions further complicated and delayed the data gathering. Last, the

64

the corroborating network members were difficult to reach, today's
mobile society possibly a factor.

The potential exists, however, that as subjects volunteer for
the study, the samples may become skewed along the important vari-
ables mentioned above. In the event of such an occurrence, an
alternate procedure exists that could theoretically be utilized to
promote comparability between the samples, termed ”matching." The
procedure was not chosen, based on the following rationale: Matching
along the five relevant variables would probably lead to a drastic loss
of subjects in both samples, reducing sample size below reasonable
limits for statistical evaluation. On the other hand, one could
attempt to continue searching for appropriate subjects, but consider-
ing the number of variables and difficulty obtaining matchable sub-
jects, the process was seen as close to impossible. Basically, though,
matching as a procedure was not used because the gains in precision
did not appear to justify the procedural difficulties, as the sample
differences on the dependent variable were expected to be large.
Furthermore, the procedures already taken to promote comparability,
in concert with other measures to reduce initial differences, i.e.,
the selection of subjects aged 18 to 40 years and residing with mem-
bers of their nuclear family or family of origin, only, reduced the

hypothetical variance between the two samples to tolerable limits.

Selection Criteria

 

While the previous discussion focused on promoting comparability
between the two samples on variables unrelated to the definition of

the populations, the present discussion focuses on criteria to insure

65

that the samples drawn are, in fact, distinct and representative of
the normal and schizophrenic populations under study. To assure the
mental health of the normal sample, the subjects chosen met the fol-
lowing criteria: (a) no known personal or family history of a mental
illness as established by the American Psychiatric Association, Qi_gf

nostic Statistical Manual of Mental Disorders, III (1980); and

 

(b) continuing employment and/or attending school, as evidence of
a functional or adaptive capacity. In the first criterion for
the normal sample, family history pertains to history of those
family members present in the immediate family, nuclear family, and/or
family of origin. Immediate nuclear family could possibly include
subject's spouse and children only; immediate family of origin could
possibly include subject's parents and/or stepparents, and siblings
only. Therefore, the psychiatric history of the subject's nephews
or nieces, in-laws, aunts, uncles, or grandparents would not be con-
sidered or pertinent. To assure that the other sample was schizo—
phrenic, the subjects chosen met the following criteria: (a) history
of hospitalization for a schizophrenic disorder, and (b) diagnosis
of schizophrenia made in the past three years.

Other selection criteria alluded to in the previous discussion
relative to the issue of comparability included (a) subjects aged
18 to 40 years and (b) residence with members of their nuclear family
or family of origin.

Last, due to the difficulties that can be encountered communicat-

ing with the schizophrenic, the following criterion was added: Those

66

participating in the study must not be presently psychotic, meaning

out of touch with reality and incapable of communication.

Subjects

Two different groupings of subjects were involved in the two
phases of the research. Considering phase two first, all subjects, a
total of 35, participated within the two comparison samples: The
normal and schizophrenic samples contained 18 and 17 subjects, respec-
tively. Phase one contained five subjects from each sample who per-
mitted the corroboration and at least four members from their social
networks, a total of 46 network members. (The following statements
apply to all the subjects of the study, excepting the subject/
corroborators of phase one.)

According to the selection rationale, the sample characteristics
of age, sex, race, and cultural background should be comparable to
promote statistical precision; the data found in Table 1 indicate
that such was the case in this study. The two sample distributions
of age were quite similar (see Table 1). The median and standard
deviation for both samples were identical, and the mean varied by
only four years. The sex ratios, as presented by Table 1, were roughly
comparable. Relative to race, the data on Table 1 demonstrate that the
clear majority of both samples were members of the same race (Cauca-
sian): Only one subject per sample was of another race. Last, it was
seen that the religious background (associated with cultural heritage)
of all subjects reflected the Judeo-Christian tradition; in particular,
the Protestant and Catholic religions predominate clearly in the normal

sample and were exclusive within the schi20phrenic sample.

67

Table 1: Sample Characteristics

 

 

 

 

 

 

 

 

Groups
Sample Characteristic G] Norma1 62 Schizophrenic
p_ = 18 p2 = 17
Age _
Mean X = 30 7': 26
Median M4= 30 M = 30
Standard deviation §D_= 6.0 4S9 = 6.0
2%
Male 8 11
Female 10 6
Race
Caucasian 17 16
Black -- 1
Oriental 1 _-
Marital status
Single 4 13
Married 13 3
Divorced l 1
Family residence
Nuclear family 14 3
Family of origin 4 14
Education
Highest degree:
Less HS diploma -- 3
HS diploma 4 12
Baccalaureate 8 1
Advanced degree 6 1
Summary:
Mean years of schooling 16 13
Employment status
Full-time 16 5
Part-time 2 --
Unemployed -- 12
Studenta l 2
Religious background
Protestant 7 11
Catholic 8 6
Jewish 3 --

 

aThe student category overlaps with the other three.

 

68

Also according to the selection rationale, the sample character-
istics associated with educational level, employment, personal economic
status, and marital status should vary as they are related to the
a priori differences in both samples, i.e., their population char-
acteristics; the data in Table 1 indicate that such was the case,
although the absence of random selection does not strictly permit the
drawing of inferences. Relative to educational level, it appears that
subjects within the schizophrenic sample predominantly tended to ter-
minate their education after completion of high school. Subjects of
the normal sample, however, tended to pursue and complete the bac-
calaureate and more advanced degrees. Notably, when the mean years of
schooling (see Table 1) were compared, the differences do not seem as
large; yet it is crucial to observe that although the mean years were
not so dissimilar, the number of those who actually completed the
requirements for a degree were. The category of employment status
reflects a priori designations in terms of sample criteria; nonethe-
less, the finding is striking that only 5 of 17 members of the schizo-
phrenic sample were employed at all, while all 18 normal subjects
were employed either full- or part-time (see Table 1). Personal
economic statusvdficfliis associated with vocational level and income
was difficult to determine due to inability of most schizophrenic
subjects to report their personal and family income; however, it was
found that the majority of the schizophrenic subjects had held
unskilled or semi-skilled jobs in their employment history and were
at the time of testing largely dependent upon public aid and/or their

family for financial support, whereas subjects within the normal

69

sample were in large part pursuing a career full-time in a trade,
semi-professional, or professional occupation supporting themselves
and their family. One could tentatively conclude that the personal
economic status did vary considerably between the two groups, the
normal sample having a much higher status. Last, the sample differ-
ences in terms of marital status and, accordingly, type of family
residence (see Table 1) were considerable. Curiously, while married
subjects living with nuclear family predominated in the normal sample,
single subjects residing with family of origin predominated in the
schizophrenic sample. Therefore, as predicted, the subjects differed
on variables associated with population characteristics.

These findings support the two positions taken in the selection
rationale. First, as the two samples were comparable along the vari-
ables of age, sex, race, and cultural background, the relative impact
on the social variables in this study is reduced, limiting the possible
variance. Second, because the two samples varied on educational
level, employment, personal economic status, marital status, and type
of family residence, the two samples appeared basically different.

In conclusion, the two samples appeared both relatively uniform
internally, yet inherently different, representative of the distinct
population characteristics; thus, a comparison with intelligible
results appears tenable. However, in that approximately half of the
schizophrenic persons refused to participate in the study, data were
not obtained from a particular section of the schizophrenic population,

a subgroup of unclear composition but possibly more hostile and

7O

paranoid. Therefore, the results will not reflect the influence of

this undefined subgroup.

Instrumentation

 

Psychosocial Network Inventory

 

The Psychosocial Network Inventory was developed as a research
instrument to explore the functioning of social networks by E. Mansell
Pattison (Pattison et al., 1975). As originally devised, the subject
is faced with three primary tasks. First, the subject develops a
list of all people personally important to him or her. The criterion
of importance has its basis in the clinical origins of the researcher
and, as such, places the focus on those persons who are active in the
subject's emotional life (see Chapter I). Second, the subject rates
the nature and quality of the relationship with each person specified
along five variables. The first variable is contact; questions define
the frequency of interaction between the subject and the network mem-
ber, and how close to each other they live. The second variable is

emotional intensity; the strength of the feeling between the subject

 

and the network member is assessed. The third variable is type of
emotion; questions relate to how positive, negative, or mixed the

feelings are. The fourth variable is instrumental base; questions

 

relate to how one can be counted upon for concrete assistance. The
fifth and final variable is degree of reciprocity; questions assessing
the feelings and the instrumental base are interrelated to develop an
idea of the affective and instrumental quid pro quo, i.e., give and
take. In other words, the relationship is determined to be symmetri-

cal or not, depending on how the feelings and supportive actions are

 

 

7)

expressed between the subject and the network member. Third, the
subject indicates which network members have relationships with one
another (outside their relationship with the subject), providing a
measure of the interconnectedness within the network.

Psychosocial Network Inventony,
Modified

 

The Psychosocial Network Inventory (Pattison et al., 1975) has
been modified for use in this study; the revised form, termed the
Psychosocial Network Inventory, Modified (PNIM), is distinct in seven
major ways. First, the PNIM (see Appendix A) was designed for use in
a structured interview. The degree of structure varies depending
upon the population: In this study, the degree of structure was high
for the schizophrenic sample (see Appendix A: Directions: Structured
Interview [2]) and low for the normal sample (see Appendix A:
Directions: Structured Interview [l]). Second, the subject's selec-
tion of social network members has been structured to enable a more
orderly and thorough identification process (see Appendix A: Listing
of Persons Important to you)- Third, the answer sheet has been
designed to ease recording and facilitate computer keypunching (see
Appendix A: Answer Sheet). Fourth, a response key utilizing scales
has been added to insure accuracy of response within the interview
(see Appendix A: Response Key [2]) or without interviewer support for
the normal subject (see Appendix A: Response Key [1]). Fifth, a ques-
tion has been added to assess the content-related aspects of rela-
tionships, i.e., types of shared activities within the relationship

(see Appendix A: item 3 on the PNIM). Sixth, as the concept of

 

 

 

 

 

 

 

 

 

72

reciprocity is basic to the interpretation of this instrument, the
instrumental and affective quid pro quo between the focal person and
the network member has been retained, yet redefined and expanded. In
agreement with Pattison, the instrumental and affective aspects of
relating are viewed as being supportive, thus functional to the
relationship. As designed, affective functioning can be expressed as
both behavioral support, i.e., physical affection (see Appendix A:
items 5 and 7 on the PNIM), and verbal support (see Appendix A: items
l2 and 15 on the PNIM). Furthermore, instrumental functioning can be
expressed as both behavioral support, i.e., helping by doing (see
Appendix A: items 6 and 13 on the PNIM), and verbal support, i.e.,
guidance (see Appendix A: items 4 and 11 on the PNIH). (Note: One
item assesses the type of support from the focal person to the network
member, while the other assesses the support from the other direction.)
Seventh and quite importantly, the variables within this study are

all defined operationally in terms of the items of the PNIM; there-
fore, the items of the PNIM are all operationally linked conceptually
to variables. The Operationalization of the variables used in this

study is reviewed in the following section.

Operationalization of Variables

 

The purpose of this section is to provide an overall perspective
and detailed description of the specific variables used in phase two
of the study. The major areas are defined, divided, and broken down
into their component parts. The variables are described thoroughly

and operationalized relative to the PNIM: Important computational

 

 

73

formulas are included. (Please refer to Appendix B, Phase I: Corrob-
oration, for the Operationalization of phase one; and Phase 11: Main
Analysis, for a more intensive treatment in outline form of the opera-
tionalization of the PNIM.)

Structure-related variables. The structure-related variables

 

convey the basic morphological characteristics of the social network.
The four main structural variables are size, interconnectedness or
adjacent density, distance, and frequency of contact. Size refers

to the total number within the social network: the actual number of
persons identified as important in the social network. Adjacent
density, an index of the interconnectedness, refers to the proportion
of linkages (relationships) in the social network to the total pos-
sible number of linkages in the network. The formula for computing
adjacent density is 2a/n(n + l) where a = the actual number of
linkages in the network and n = the network size. Distance refers

to the distance between the focal person and a network member in
terms of location or time of travel, and is assessed through a desig-
nated question (see Appendix A: item 2 on the PNIM). Frequency of
contact refers to the relative frequency of contact between the focal
person and a network member, and is assessed through a designated
question (see Appendix A: item 1 on the PNIM).

Content-related variables. The content-related variables convey

 

aspects of the content of the relationship between the social network
of normals and schizophrenics. In this study, content is strictly
defined as activity. Ten content areas--literally, types of activity--

have been selected for inclusion in this study and assessed by a

74

designated question (see Appendix A: item 3 on the PNIM). These
types of activities are intended to be categorical and inclusive.
They are listed below with definitive examples:

1. FAMILY activities such as meals, holidays, vacations, or
reunions.

2. EMPLOYMENT-related activities such as working with co-workers
or supervisors.

3. ROMANTIC activities such as dating, dancing, or going out to
dinner.

4. CONVERSATIONAL activities such as intimate, personal sharing
or philosophical discussions.

5. SOCIAL activities such as parties, banquets, or visiting.

6. RECREATIONAL activities such as playing cards, participation
in sports, or attendance at sports events.

7. FRATERNAL activities such as participation in clubs or other
organizations.

8. RELIGIOUS activities such as attendance at church, synagogue,
or temple.

9. POLITICAL activities such as rallies, or discussions of
politics.

10. VOLUNTEER work such as service to the community, giving
blood, or hospital work.

The content of the relationship is assumed to be a partial index
of the quality: The more varied the content, i.e., the greater num-
ber of activities shared, the higher the quality of the relationship.

To reflect this dimension, two types of relationships (with varying

75

implications about quality) have been chosen for study: the uniplex
and the multiplex relationship. In the former, only one activity is
shared in the relationship; in the latter, two or more activities
are shared.

Another content—related variable, relationship density, is a
summary index of the quality of the relationships within the social
network, actually the primary star; the formula for relationship
density is Z rx/n, where r = the number of content areas (in this
study r may equal 10) and n = the number of possible linkages in the
primary star. Thus, relationship density is an average of the content
areas of the relationships between the focal person and the social
network.

Function-related variables. The function-related variables

 

reflect the transaction of support between the focal person and the
social network: Function is designated as either affective verbal,
affective behavioral, instrumental verbal, or instrumental behavioral
support (refer to the previous section for a list of the matching
items). Two items on the PNIM correspond with each one of the four
designations, making a total of eight related items; one item of each
pair elicits the degree of support the focal person receives; and the
other, the degree of support the focal person provides.

The three function-related variables are functional indegree,
functional outdegree, and functional symmetry. Functional indegree
indicates the degree of instrumental and affective support that the
focal person is receiving (items 5, ll, 13, and 15). Functional out-

degree indicates the degree of instrumental and affective support

 

 

 

 

 

 

76

that the focal person is providing to a network member (items 4, 6,
7, and 12). For one relationship, both are computed by summing the
degree of support, whether given or received by the focal person,
across the items assessing the four dimensions. The functional
indegree and outdegree for a network are computed by summing across
the network and dividing by the total network size.

Functional symmetry is a function-related variable that assesses
the balance of support in relationships in the social network. Three
cases are noted. When the functional indegree (FI) and functional
outdegree (F0) are equal, i.e., when the support is given and received
in equal measure, the relationship is characterized by functional
symmetry, and is termed for this study, "reciprocal." Uhen the F1 is
larger than FO, i.e., when more support is given by the network member
to the focal person, the relationship is functionally asymmetrical,
and is termed relative to the focal person, "dependent," When the F0
is larger than the FI, i.e., when the focal person is giving more than
the network member, the relationship is again asymmetrical, and termed
relative to the focal person, "supportive.” Assessments of functional
symmetry are made within the social networks of normals and schizo-
phrenics in terms of the total social networks.

Emotion-related variables. Assessment of emotions is broken

 

down into two categories: types of feelings and strength of feelings.
The variables used to examine the type of emotion are affective
indegree, affective outdegree, and affective symmetry. Affective
indegree indicates the type of feelings expressed by a member for the

focal person as perceived by the focal person. The range extends

 

 

77

from all positive to all negative feelings. The index for one rela-
tionship is equivalent to the answer of the designated item (see
Appendix A: item 8 on the PNIM). The affective indegree for the
social network is a summing across the network for that item divided
by the size of the network. Affective outdegree indicates the type of
feelings experienced by the focal person for a member of the social
network. The range extends from all negative to all positive feel-
ings. The index for one relationship is also equivalent to the answer
of one designated item (see Appendix A: item 14 on the PNIM). The
total affective indegree is a summing for that item divided by n.
Affective symmetry exists when the type of emotion between the focal
person and the network member is identical, i.e., when affective
indegree (AI) = affective outdegree (AO). Affective asymmetry exists
when the type of emotion between the focal person and the network
member varies, i.e., when AI f A0.

The variables used to examine the strength of emotion are invest-
ment indegree, investment outdegree, and investment symmetry. Invest-
ment indegree reflects the strength of feelings (emotional investment)
experienced by the network member for the focal person as perceived
by the focal person. The range extends from weak to very strong
feelings. The index for one relationship is equivalent to the answer
to one designated item (see Appendix A: item 9 on the PNIM). The
investment indegree for the social network is a summing across the
network for that item divided by n. Investment outdegree reflects
the strength of feelings (emotional investment) experienced by the

focal person for a network member. The range extends from weak to

 

78

very strong feelings. The index is also equivalent to the answer to
one designated item (see Appendix A: item 10 on the PNIM). The total
investment outdegree is also a summing across the network divided by n.
Investment symmetry exists when the emotional investment experienced
by the focal person and the network member is equal. Investment asym-
metry exists when their emotional investment is different.

An exception. Certain types of network members will not be

 

included in the data analysis due to the importance and relevance of
reciprocity in the analysis of the PNIM and the necessity of cor-
roborated self-report. The analysis and interpretation of the results
of the PNIM are predicated on the assumption that the assessed rela-
tionships can be reciprocal in all respects: The presence or absence
of reciprocity in relationships has certain meanings that are central
to the purpose of this study (see chapter I). In order to be able to
make proper interpretations of the results, only those network members
capable of reciprocal arrangements will be included in the data analy-
sis. Also, network members must be capable of reporting on their
relationship with the focal person. Furthermore, they must actually
be present physically. For these reasons, the deceased and/or fan-

tasied network members will not be included in the data analysis.

Consent Procedures

 

Two sets of consent procedures have been developed to accommodate
to the intrinsically different characteristics of the two samples,
the normal and the schizophrenic, and the environments in which they
were identified. The term "schizophrenia" is a psychiatric diagnosis

that denotes a certain type of mental illness; as such, it is subject

 

79

to regulation and strict definition. Those facilities responsible
for the housing and/or treatment of the schizophrenic are ethically
and legally charged to protect the identity of the schi20phrenic and
maintain the confidentiality of the schizophrenic's records. Excep-
tions are defined legally, must be justificable, and are subject to
stringent safeguards. Accordingly, the identification of the schizo-
phrenic sample and the acquisition of informed consent for the study
are difficult and complex procedurally and involve ethical and legal
considerations, the focus explicitly being the protection of the wel-
fare of the schizophrenic. The term "normal,” used to describe the
other sample, is neither a psychiatric diagnosis nor is subject to
regulation; therefore, the consent procedures are relatively simple
and straightforward, although the focus remains protection of the

subjects' welfare.

Testing Procedures

 

Normal and Schizophrenic Samples

 

A set of parallel, but differing, testing procedures was designed
for the normal and schizophrenic samples. Each set constituted a
structured interview. Both sets of procedures contained seven inter-
view phases, one of which might not be used depending on whether data
corroboration was permitted by the subject. Four of these phases were
required for the completion of the Psychosocial Network Inventory,
Modified. Each phase corresponded to the completion of a primary task
within the interview. The primary tasks of the seven phases were

identical between the two samples. However, the manner in which the

80

primary tasks were to be accomplished differed in varying degrees,
based on the characteristics of each sample. Most prominently,

group testing and unassisted completion of the inventory (once directions
were understood) were permitted for the normal subjects, due to their
superior functioning capacity; on the other hand, all schizophrenic
subjects were tested individually and in private to respond to their
particular personal needs, lower functioning capacity, and issues of
confidentiality and privacy: In this way, the interviewer could pro-
vide immediate and continuing assistance and direction, and demon-
strate overtly a respect for the privacy of the schizophrenic subject.
Last, to facilitate the completion of the two types of structured
interviews, directions were drafted for the use of the interviewer
(see Appendix A: Directions: Structured Interview [1], designed for
the normal sample; and Directions: Structured Interview [2], designed
for the schizophrenic sample).

Phase 1. The primary task of this phase of the structured inter-
view was to provide an adequate explanation of the study to the sub-
ject. Two forms were prepared, one for each sample (see Appendix A:
Explanation of the Research Study [1], designed for the normal sample;
and Explanation of the Research Study [2], designed for the schizo-
phrenic sample). The appropriate forms were read to subjects of both
samples; subjects were encouraged to ask their questions immediately.
Members of both samples were asked if they understood the explanation
of the study. A lack of understanding or misunderstanding was explored
when either appeared. Also, each subject was asked specifically about

the corroboration aspect/option of the study to assure comprehension.

Bl

If necessary, the letter to be sent to corroborating network members
(see Appendix A) was shown to clarify any questions, especially those
pertaining to confidentiality and privacy.

Phase 2. The primary task of this phase was to obtain the
informed consent of the subject to participate in the study and, when
appropriate, to permit the contact of network members, randomly
selected, the purpose being corroboration of self-report data. Two
consent forms were prepared for use in this phase, one with additional
statements relating to the corroboration option for use with subjects
giving their approval for corroboration and from whom it was desired
by the researcher (see Appendix A: Statement of Informed Consent [1]),
and another for use when corroboration was neither approved by the
subject nor desired by the researcher (see Appendix A: Statement of
Informed Consent [2]). The researcher made the request for the cor-
roboration option of every subject in both samples who appeared
interested and open until the desired percentage of respondents was
reached. After the appr0priate form was placed in front of the
subject/s, the t0p of form (1) or form (2) in entirety was read by
the interviewer. When the interviewer was confident that the explana-
tion of the study and of the rights of the subject was understood by
the subject/s, the interviewer asked the subject/s to sign and date
the form, all interested in further information about the study after
its completion to make a mark by their name. For those subjects
interested in the corroboration option, the relevant statements found

on the lower portion of form (l) were read and explained by the

82

interviewer. At this point, the request was made for the subject/s
to sign and date this portion of the consent form.

Once the consent forms had been signed and dated, the four basic
steps in the study were outlined and the data collection began. The
four basic steps were as follows: (a) completing the General Infor-
mation Sheet, (b) listing persons in the network, (c) assessing rela-
tionships, and (d) indicating the relationships among the persons
named.

Phase 3. The primary task of this phase was the completion of
the General Information Sheet (see Appendix A). The interviewer
encouraged the normal subjects to complete this form without assistance;
however, if this wasn't a successful approach with any normal subjects,
then the interviewer would ask the questions and complete the form,
using the form as a prompter. For the schizophrenic sample, the
interviewer asked the questions, using the form as a prompter, and
recorded the responses for the subject.

Phase 4. The primary task of this phase was the development of
a list of all persons who were presently important in the life of the
subjects. Two forms were designed for use in this phase: the answer
sheet (see Appendix A) and a list of persons who might be potentially
important to the subject to be used as a prompter, insuring a complete
list (see Appendix A: Listing of Persons Important to You). All sub-
jects were encouraged to use their own definition of importance.
Additionally, subjects from the normal sample were invited to fill
out the names, as well as other identifying information (sex, length

of relationship, and type of relationship) on their answer sheet. The

83

normal subjects were also requested to circle the names of persons
named who are deceased. The normal subjects could also exercise an
option to complete the list following the interview, as long as a
partial list of a minimum of six names was completed in the inter-
viewer's presence to assure a thorough understanding of the task and
accuracy. The interviewer filled out the names and identifiers of
the complete social network for the schizophrenic subjects.

Phase 5. The primary task of this phase was the assessment of
the relationships, identified in the social network, through the use
of the Psychosocial Network Inventory, Modified (see Appendix A).

Two other forms contained the three scales for answering items 4
through 15 on the Psychosocial Network Inventory, Modified (PNIM),

one for each sample (see Appendix A: Response Key [1], designed for

the normal sample; and Response Key [2], designed for the schizophrenic
sample). After the total list of important people was developed for
both samples or, at least, a partial list of six for the normal sample,
the assessment of relationships began within the structured interview.
The interviewer initiated the assessment by taking the first person
identified in chronological order and asking the first three questions
on the PNIM. For both samples the interviewer read these questions;
however, the normal sample was encouraged to fill in the blanks with
the interviewer's assistance and the schizophrenic sample only needed
to respond verbally to the questions while the interviewer marked the
answer sheet. After these were completed for both samples for the
first person, the rest of the questions on page two of the PNIM and

the appropriate response key were presented to both samples. The

84

scales were described as a continuum, all along which answers could
be selected. Response Key (1) for the normal sample was numbered one
through nine to correspond to the answer code, to permit answers
between the modal points, and to facilitate unassisted completion
once the directions were understood. Response Key (2), identical
to the former response key except for the numbering, was presented to
the schizophrenic sample with the explanation that they could identify
their answer on any point along the scale, including the five points
offered for guidance. Once this had been fully explained, items
4 through 15 were asked of both samples. It was then added that all
questions must be completed for all identified persons. The schizo-
phrenic sample with the interviewer's assistance then continued
assessing the relationships until all were completed. In some cases,
the schizophrenic subjects opted to answer items one through three on
all persons before continuing to the second page: Such was permitted
(likewise, some normals opted for this alternative). Also, the
schizophrenic subjects were offered breaks in the testing to renew
their interest and concentration. The normal sample either completed
all assessments of the relationships during the interview or finished
at least two complete assessments of the relationships if they opted
to complete the form later, given their more difficult timetables
and scheduling.

Phase 6. The primary task of this phase was the assessment of
the degree of connectedness among the network members: The procedure
was identical for both samples. The interviewer assisted all sub-

jects to identify ongoing relationships between the network members.

 

85

To qualify for an ongoing relationship, it was required that both
persons must know each other and_have some kind of relationship with
each other outside their relationship with the subject. The assigned
chronological numbers of those identified as having a relationship
with the selected person were written in the right hand column of

the answer sheet designated "Connections." After all of the identi-
fied persons had been assessed for network connections, the structured
interview terminated unless the subject had been asked and permitted
the corroboration phase.

Phase 7. The primary task of this phase was to identify through
random selection the network members who were to be contacted to
corroborate the self-report of the subject. One form was used for
this phase to assist the interviewer in random selection of the corrobo-
rating members (see Appendix A: Selected Network Members). On this
form, the size of the identified network was associated with five
numbers of random selection which designate those identified. All
subjects were informed which five network members were chosen through
the randomizing process. Once fully informed, the subjects were free
to rescind their approval of this phase. Only two schizophrenic sub-
jects so decided at this point. The names, addresses, and phone
numbers were collected on all five members identified for corrobora-
tion; however, when an identified member met one of the criteria for
exclusion ([a] deceased and [b] a fantasied relationship), alternate
choices were made randomly on the spot. This phase completed the
structured interview. With the interview ended, the subjects were

thanked for their time, effort, and interest in the research.

86

Corroborating Network Members

 

Contact and examination of the corroborating network members,
those identified by random selection in phase seven of the testing
procedures for the subjects, involved a three-phase effort.

Phase 1. During this phase, the corroborating members were con-
tacted by phone. In the conversation they were informed of the nature

and purpose of the contact and the extent of their involvement,

 

asked if they were interested in participating in this research, and
if so, to supply or confirm their mailing address. A few declined
immediately at this point and others failed to respond after the
mailing. A few others were not contacted first by phone as the
numbers were not known.

Phase 2. During this phase, the corroborating members who were
interested in or open to considering participation or unreachable by
phone were mailed the following forms, as well as a self-addressed
stamped envelope for response: Corroborating Member Letter (see
Appendix A), Statement of Informed Consent (2) (see Appendix A), and
Response Key (1) (see Appendix A). In the letter the nature and pur-
pose of the contact and focus of the study were presented, safeguards
to confidentiality and privacy mentioned, and the request was made
that if they would be willing to participate in the study to sign,
date, and forward the enclosed consent form in the self-addressed
envelope. If there were any questions, a 24-hour phone number was
included. Last, a response key was enclosed to facilitate answering

of the questions from the PNIM.

87

Phase 3. This phase began once the consent form was received
in the mail from the cooperating network member; however, when the
corroborating members seemed unduly late in returning the form,
follow-up calls were made. With the consent forms returned, the cor-
roborating members were again contacted by phone and asked questions
about their relationship with the subject and their knowledge of other
important relationships. In relation to the former, they were asked
all of the questions on the PNIM as it applied to their relationship
with the subject. All answers were written on a standard answer
sheet which was designated for corroborators of a certain subject:
Each corroborator was assigned the number given to him/her by the
subject. In relation to the latter, they were asked how many people
they knew whom they would consider as personally important to the
subject, i.e., the network cluster. The answer to this question, the
size of the network cluster, was written under "SUM" on the right
portion of the answer sheet, the same place designated for the sum
portion of the subject's response. With this final answer recorded,
this third and final phase is completed: At least four corroborating
members were tested in this fashion, although five were initially
identified. All corroborators were informed that, upon their request,
they would be forwarded more complete information about the study and
an overview of the results. Last, all were thanked for their interest

and participation.

88

Risk/Benefit Analysis

 

An examination of the nature and design of the study suggested
the presence of potential social and psychological risks to the
subjects; physical, legal, and economic risks were not posed.

Potential social risks were associated with the request that some
subjects relinquish their anonymity to provide access to their social
network for further study of the social relationships. The risk
appeared to relate to possible misinterpretations by the contacted
network members of (a) the subject‘s motive in identifying them and
(b) the reason for the subject's participation in the study.

The social risk of misinterpretation appeared to be largely
dependent upon the degree of vagueness inherent in the explanation
given to the contacted network member. To minimize this possibility,
every effort was made to develop a clear, direct, precise, and unambig-
uous explanation. However, this safeguard is one-sided in the sense
that it can only act to increase the clarity of the statement; it
cannot be assumed to control for idiosyncratic interpretations or
misuse of the disclosure of the subject's participation by the
contacted network member.

Potential psychological risks related to the use and effect of
the Psychosocial Network Inventory, Modified (PNIM). The use of all
psychological tests is accompanied by potential risks, and the same
is true for the PNIM. The primary risk is the generation of psycho-
logical conflicts, and resulting discomfort/dysfunction, previously
controlled by the subject's defense mechanisms. This reaction could

occur in either group, but it is more likely with the schi20phrenics;

 

89

their coping abilities and defenses are much less controlled and
integrated, by definition. However, the type of test used in this
study is a highly structured, rather straightforward inventory tapping
primarily conscious levels, possibly the least threatening of all
psychological test forms; projective tests, on the other hand, tap
much deeper levels of the subject's psyche and would, therefore, be
much more likely to uncover hidden conflict and provoke an untoward
reaction.

The psychological risks were lessened through the choice of a
multiple-choice, structured test (see above); however, if such risks
materialized, plans were made to encourage the schizophrenic-~who, by
the way, was tested in a psychiatric setting--to consult with his/her
primary therapist and/or sign a release so that the reaction could be
reported to appropriate staff. If the reaction was considerable,
plans were made to report it immediately so that facility staff could
act to intervene. If difficulties arose with a member of the normal
group, plans were made to refer him or her to the local mental health
center.

Confidentiality was assured through (a) a coding procedure,

(b) locked records, and (c) destruction of all identifying information
after no longer necessary.

Last, the study was designed to be of benefit to the subject, the
profession, and society in general. Through the administration of the
PNIM, the subject might develop insights into the processes and
dynamics operating in his/her social network and perhaps, discover

the extent of available social support. Gains within the professional

90

realm include the following: (a) determination of the validity and
usefulness of self-report scales in social network analysis, clarify-
ing the value of previous contributions to the field who based their
work on self-report; (b) greater insight in the psychosocial function-
ing of the schizophrenic relative to the normal subject; and (c) recom-
mendations regarding the psychosocial management and treatment of the
schi20phrenic disorder. Society, in general, could be profited by

the knowledge as used by the profession.

University Authorization

 

Following the dissertation committee approval of the research
proposal on January 25, 1980, the University Committee on Research
Involving Human Subjects (UCRIHS) requested and reviewed the research
proposal and supporting materials focusing on the rights and welfare
of human subjects (see Appendix C), as mandated by the National
Research Act, Public Law 93-348, Section 474(a). Following their
review, Henry E. Bredeck, Chairman, UCRIHS, notified this researcher

that the committee had approved the project on April 7, 1980.

Research Questions

 

This section has been included in the chapter as well as the
section, Null Hypotheses, to provide the reader with an organizing
schema, a mind set, with which to conceptualize the intent or goals
of the study, and to grasp the extensive nature of the null hypotheses;
the Null Hypotheses section is lengthy and detailed, and yields slowly

to understanding without such a guide.

91

The purpose of the study was to compare the social networks of
normals and schizophrenics along four major classes of variables:
Structure, the dimension reflecting the morphological characteristics
of the network; Content, reflecting the sharing of activities within
the network; Function, reflecting the exchange of support within the
network; and Emotion, reflecting the experience of emotion within the
network. However, before the analysis of the social networks of nor—
mals and schizophrenics can be more fully understood, the self—report
of the subjects on their social networks must be corroborated with
the report of at least four network members to determine the degree
of agreement. This constitutes phase one of the analysis. In rela-
tionship to this determination, the following research question was
posited (stage one of the analysis within phase one): How does the
self-report of normals and schizophrenics corroborate with the self-
report of identified network members/corroborators on the joint rela-
tionship along these variables:

1. Structure, specifically size of the network cluster, distance,
and frequency of contact;

Content, specifically the number of content areas;
Function, specifically functional indegree and outdegree; and

Emotion, specifically affective indegree and outdegree, and
investment indegree and outdegree?

One further question is posited (stage two of the analysis within
phase one): Are the differences between the subjects and corroborators
significant between the normal and schizophrenic samples?

In relation to the stated purpose of the study, the following

groups of research questions are posited in phase two of the analysis:

92

1. In terms of structure, how do the social networks of normals
and schizophrenics differ relative to size, interconnected-
ness, distance, and frequency of contact? Given the differ-
ences, what do the interrelationships among the variables
suggest about the structure of the social networks?

2. In terms of content, how do the social networks of normals
and schizophrenics differ in proportion of uniplex and multi-
plex relationships, and in relationship density? Given the
differences, what do the interrelationships among the vari-
ables suggest about the content of the social networks?

3. In terms of function, how do the social networks of normals
and schizophrenics differ in functional indegree, functional
outdegree, and functional symmetry? Given the differences,
what do the interrelationships among the variables suggest
about the function of the social networks?

4. In terms of emotion, how do the social networks of normals
and schizophrenics differ in affective indegree, affective
outdegree, affective symmetry, investment indegree, investment
outdegree, and investment symmetry? Given the differences,
what do the interrelationships among the variables suggest
about the emotional dimensions of the social networks?
Last, given the differences in the above four classes of variables,
what can be concluded about the overall functioning of the social net-

works of normals and schizophrenics?

Null Hypotheses

 

Phase One

Stage one. The following null hypotheses relate to the first
stage of the analysis in phase one, the corroboration of the self-
report of the normal and schizophrenic samples with selected members
of their social network. A statement of the major null hypothesis
of stage one and phase one outlines the major intent of this stage of
the analysis and precedes the testable null hypotheses associated with
the classes of social network variables for the normal and schizo-

phrenic samples: structure, content, function, and emotion. These

93

following hypotheses are written in multivariate form, excepting the
hypotheses associated with the class of variable, content. Both con-
tain a univariate statement. The numbered variables within the null
hypotheses correspond to subordinate, univariate null hypotheses:
The coded designation of the univariate hypothesis retains the short-
hand symbol of the multivariate hypothesis, adding only the variable
number separated by a dash from the subscript. The univariate
hypotheses are written in full only within the first set of multi—
variate hypotheses for demonstration purposes. The hypotheses of
stage one of phase one are:

Major null hypothesis of stage one of phase one.

Hom: No difference exists in the self-report between the normal
subjects and network member/corroborators, and between the
schizophrenic subjects and network member/corroborators
along the major classes of social network variables:
structure, content, function, and emotion.

Major null hypotheses: structure.

H01": No difference exists in the self-report of the normal sub-
jects and corroborators in terms of structure-related vari-
ables: (a) Variable 1, size of the network cluster;

(b) Variable 2, distance; and (c) Variable 3, frequency of
contact.

Subordinate, univariate null hypotheses: structure, normalsample.

Ho]n_1: No difference exists in the self-report of the normal

subjects and corroborators in terms of the size of

the network cluster.

94

Ho]n_2: No difference exists in the self-report of the normal

Ho

Hols:

subjects and corroborators in terms of the distance.

ln-3: No difference exists in the self-report of the normal

subjects and corroborators in terms of the frequency
of contact.
No difference exists in the self-report of the schizophrenic
subjects and corroborators in terms of structure-related
variables: (a) Variable 1, size of the network cluster;
(b) Variable 2, distance; and (c) Variable 3, frequency of

contact.

Subordinate, univariate null hypotheses: structure, schizophrenic
sample.

Ho

Ho

Ho

ls-l: No difference exists in the self-report of the schizo-

phrenic subjects and corroborators in terms of the

size of the network cluster.

ls-2: No difference exists in the self-report of the schizo-

phrenic subjects and corroborators in terms of the

distance.

ls-3: No difference exists in the self-report of the schizo-

phrenic subjects and corroborators in terms of the

frequency of contact.

Major null hypotheses: content.

H02“:

No difference exists in the self-report of the schizophrenic
subjects and corroborators in terms of the content-related
variable: Variable 4, the number of content areas. To

maintain consistency the univariate hypothesis is designated

H°2n-4'

H025:

95

No difference exists in the self-report of the schizophrenic
subjects and corroborators in terms of the content-related
variable: Variable 4, the number of content areas. To main-

tain consistency, the univariate hypothesis is designated

H02s-4°

Major null hypotheses: function.

H03”:

H035:

No difference exists in the self-report of the normal
subjects and corroborators in terms of function-related
variables: (a) Variable 5, functional indegree; and

(b) Variable 6, functional outdegree. Univariate hypothe-
ses corresponding to the variables: Ho3n_5; and H03n-6'
No difference exists in the self-report of the schizo-
phrenic subjects and corroborators in terms of function-
related variables: (a) Variable 5, functional indegree;
and (b) Variable 6, functional outdegree. Univariate

hypotheses corresponding to the variables: H03s-5; and

H03s-6'

Major null hypotheses: emotion.

Ho4n:

No difference exists in the self-report of the normal
subjects and corroborators in terms of emotion-related
variables: (a) Variable 7, affective indegree; (b) Vari-
able 8, affective outdegree; (c) Variable 9, investment
indegree; and (d) Variable lO, investment outdegree.

Univariate hypotheses corresponding to the variables:

H°4n-73 ”04n-83 H04n-9; and H04n-1O'

96

H045: No difference exists in the self-report of the schizo-
phrenic subjects and corroborators in terms of emotion-
related variables: (a) Variable 7, affective indegree;

(b) Variable 8, affective outdegree; (c) Variable 9, invest-
ment indegree; and (d) Variable lO, investment outdegree.

Univariate hypotheses corresponding to the variables:

”04s-73 H04s-8; “045-9; and H04s-lO'

Stage two. The following null hypotheses relate to the second
stage of the analysis in phase one, the comparison of the normal and
schizophrenic samples in terms of the relative differences between the
subjects and corroborators. A statement of the major null hypothesis
of stage two and phase one outlines the major intent of this stage of
the analysis and precedes the testable null hypotheses associated with
the classes of social network variables for the normal and schizo-
phrenic samples: structure, content, function, and emotion. These
following hypotheses are written in multivariate form, excepting the
hypotheses associated with the class of variable, content, which con-
tains a univariate statement. The numbered variables within the
null hypotheses correspond to subordinate, univariate null hypotheses:
As in stage one, the coded designation of the univariate hypothesis
retains the shorthand symbol of the multivariate hypothesis, adding
only the variable number separated by a dash from the subscript. The

hypotheses of stage two (If phase one follow below:

97

Major null hypothesis of stage two of phase one.

Hom:

No difference exists in the normal and schi20phrenic samples
in terms of discrepancy of self-report between subjects and
corroborators along the major classes of social network

variables: structure, content, function, and emotion.

Major null hypothesis: structure.

Ho :

S

No difference exists in the normal and schizophrenic samples
in terms of discrepancy of self-report between subjects and
corroborators along the structure-related variables:

(a) Variable 1, size of network cluster; (b) Variable 2,
distance; and (c) Variable 3, frequency of contact. Univari—
ate hypotheses corresponding to the variables: Hos_1;

Ho Ho

s-2; s-3'

Major null hypothesis: content.

Ho :

C

No difference exists in the normal and schizophrenic samples
in terms of discrepancy of self-report between subjects and
corroborators along the content-related variable: Variable 4,
the number of content areas. To maintain consistency the

univariate hypothesis is designated Hoc_4.

Major null hypothesis: function.

Hof:

No difference exists in the normal and schizophrenic samples
in terms of discrepancy of self-report between subjects and
corroborators along the function-related variables:

(a) Variable 5, functional indegree; and (b) Variable 6,
functional outdegree. Univariate hypotheses corresponding

to the variables: Hof_5; and Hof_6.

98

Major null hypothesis: emotion.

Hoe: No difference exists in the normal and schizophrenic samples
in terms of discrepancy of self-report between subjects and
corroborators along emotion-related variables: (a) Variable
7, affective indegree; (b) Variable 8, affective outdegree;
(c) Variable 9, investment indegree; and (d) Variable lO,

investment outdegree. Univariate hypotheses corresponding

to the variables: Hoe_7; Hoe-8; Hoe_9; and Hoe_10.

Phase Two

Stage one. The following null hypotheses relate to the first
stage of the analysis in phase two, the main comparison between the
social networks of the normal and schi20phrenic samples. A statement
of the major null hypothesis of stage one and phase two outlines the
major intent of this stage of the analysis and precedes the testable
null hypotheses associated with the classes of social network variables:
structure, content, function, and emotion. These major hypotheses--
following below--are written in multivariate form. The numbered
variables within the null hypotheses correspond to subordinate, uni-
variate null hypotheses: The coded designation of the univariate
hypotheses retains the shorthand symbol of the multivariate hypothe-
sis, adding only the variable number separated by a dash from the
subscript. The hypotheses of stage one of phase two follow below:

Major null hypothesis of stage one of phase two.
Hom: No difference exists in the social networks of the normal and

schizophrenic samples along the major classes of social net-

work variables: structure, content, function, and emotion.

99

Major null hypothesis: structure.

Ho :

S

No difference exists in the social networks of the normal
and schizophrenic samples in terms of the structure-related
variables: (a) Variable 1, size of the network; (b) Vari-
able 2, interconnectedness; (c) Variable 3, distance; and
(d) Variable 4, frequency of contact. Univariate hypotheses
corresponding to the variables: Hos_]; Hos_2; Hos_3; and

Hos_4.

Major null hypothesis: content.

Ho :

C

No difference exists in the social network of the normal

and schizophrenic samples in terms of the content-related
variables: (a) Variable 5, proportion of uniplex relation-
ships; (b) Variable 6, proportion of multiplex relationships;
and (c) Variable 7, relationship density. Univariate hypothe-
ses corresponding to the variables: HOc-S' Hoc_6; and

“Cc-7°

Major null hypothesis: function.

Hof:

No difference exists in the social networks of the normal
and schizophrenic samples in terms of the function-related
variables: (a) Variable 8, functional indegree; and Vari-
able 9, functional outdegree. Univariate hypotheses corres-

ponding to the variables: Hof_8; and Hof_9.

Major null hypothesis: emotion.

Hoe:

No difference exists in the social networks of the normal
and schizophrenic samples in terms of the emotion-related

variables: (a) Variable ll, affective indegree;

100

(b) Variable 12, affective outdegree; (c) Variable l4,
investment indegree; and (d) Variable 15, investment out-
degree. Univariate hypotheses corresponding to the vari-

ables: Hoe_]]; Hoe_]2; Hoe_]4; and Hoe_]5.

Stage two. The following null hypotheses relate to the second
stage of the analysis in phase two, the assessment of symmetry within
the two samples. A statement of the major null hypothesis of stage
two and phase two outlines the major intent of this stage of the analy-
sis and precedes the testable hypotheses associated with two classes
of social network variables: function and emotion. These following
hypotheses are written in multivariate form. The numbered variables
within the null hypotheses correspond to subordinate, univariate null
hypotheses: The coded designation of the univariate hypotheses
retains the shorthand symbol of the multivariate hypothesis, adding
only the variable number separated by a dash from the subscript. The
hypotheses of stage two of phase two follow below:

Major null hypothesis of stage two of phase two.

Hom: No difference exists between the measures of outdegree and
indegree within the normal and schizophrenic samples along
the two major classes of social network variables: function
and emotion.

Major null hypothesis: function and emotion, normal sample.
Hofe/n: No difference exists between the measures of outdegree
and indegree within the normal sample along the two major

classes of social network variables: function and

lOl

emotion; i.e., within the normal sample, functional sym-
metry (Variable 10), affective symmetry (Variable l3),
and investment symmetry (Variable l6) exist. Univari-

ate hypotheses corresponding to the variables: Hofe/n-lO;

Hofe/n-l3; and Hofe/n-l6'
Major null hypotheses: function and emotion, schizophrenic sample.
Hofe/s: No difference exists between the measures of outdegree
and indegree within the schizophrenic sample along the
two major classes of social network variables: function
and emotion; i.e., within the schizophrenic sample, func-
tional symmetry (Variable lO), affective symmetry (Vari-
able 13), and investment symmetry (Variable 16) exist.

Univariate hypotheses corresponding to the variables:

Hofe/s-lO; Hofe/s-l3' and Hofe/s-l6'

Research Design and Analysis

 

Phase One

The focus of the research design and analysis of phase one
in the study was the corroboration of self-report. Two stages of the
research were planned to investigate this focus. Both are described
below.

Stage one. In the first stage, a comparative group design was
used for each sample to determine if the self-report data were ade-
quately corroborated by network members within the two samples. To
set up this type of design, the self-report of five subjects in each

sample was contrasted with the self-report of at least four, if not

102

five, members of their social networks on the joint relationship,

and average difference scores were computed for each subject relative
to the network members. These scores were compared to the zero
vector. This design was characterized by one fixed independent
variable, termed group differences. It has two levels, termed group
of average difference scores and group of zero differences. Further-
more, the design has 10 dependent variables, grouped by major class
(see Figure 1). Cell size equals five. In a sense, the group of
average difference scores associated with the two samples were each
compared to the average difference scores of the "ideal" group.

The data analysis was conducted by using three multivariate tests
and one univariate test for each sample. The major hypotheses cor-
responding to the three major classes of variables (structure, func-
tion, and emotion) were analyzed through the use of multivariate
analysis of variance tests. The goal of these multivariate procedures
was to consider each class of variable separately, yet its component
variables jointly, while avoiding inflated alpha levels. Each multi-
variate test was examined at the p_= .10 level, chosen due to the low
degrees of freedom (gf_= 4, l). Univariate tests were designed to
follow if the multivariate test was significant, using an approximate,
partitioned alpha level, divided by the number of dependent variables.
The partitioned alpha levels of the structure-related, function-
related, and emotion-related variables are .03, .05, and .025, respec-
tively. The major hypothesis, corresponding to the class of variable,
content, was analyzed through the use of a univariate analysis of

variance, examined at the p_= .10 level.

.uwcmcnaongum new _msco: .Qsocm some Low name ugoawcuwpmm
we cowumgonogcou mo mwmx_mcm mg» Low .mco mmmpm .mco mmmgn Low :mwmmu gucmmmmm .P «gnaw;

 

m mmcmmupzo
ucmsumm>cH

 

mmcmmucF
pamEumm>cH

 

mmcmmuuzo
w>wpum¥m<

 

wmcmmucw

a>vpuacc< ceases“

 

mmmeupzo
Pecowpuczd

 

103

mmgmmucw

chowpocau cowpuczd

 

women ucmpcou

mo consaz pcmpcoo

 

pompcou mo
xucmzcmcm

 

wucmamwo

 

 

m u.m memz—u xLome:

mo mem mgspuaegm

 

 

 

 

 

 

mmucmcmwwwo AuwcmL580ngum co FmELozv mmmpu an

mmcoum mesmemwmwo mm mpgm cm cmam
ocmN mo asogu mmmgm>< we asocw _a . > p u o

 

 

 

mmucwgmwwwo Azocw ”anmwgm> pcwucmammmm nwxwd

104

Stage two. In the second stage, a comparative group design was
used to determine if the two samples were significantly different in
relation to the corroboration of self-report. This design was char-
acterized by one fixed independent variable, type of average differ-
ence scores, with two levels, normal and schizophrenic. Furthermore,
the design has 10 dependent variables, grouped by the four major
classes of variables (see Figure 2). Cell size equals five.

The data analysis was conducted by using three multivariate tests
and one univariate test. The major hypotheses corresponding to the
three major classes of variables (structure, function, and emotion)
were analyzed through the use of multivariate analysis of variance
tests. As in stage one, the goal of these multivariate procedures
was to consider each cflass of variable separately, yet its variables
jointly, while avoiding inflated alpha levels. Each multivariate
was examined at the p_= .10 level, chosen due to the low degrees of
freedom (g:_= 4, l). Univariate tests were designed to follow if the
multivariate test was significant, also using approximate, partitioned

alpha levels: The alpha levels are identical to those of stage one.

Phase Two

The focus of the research design and analysis of phase two in
the study was the main comparison of social network characteristics
between the two samples. Two stages of the research were planned to
investigate this focus. Both are described below.

Stage one. In the first stage, a comparative group design was

used to determine if the two samples were significantly different

105

 

 

 

 

 

 

 

 

 

 

 

mud mud

 

 

 

ewesem ewcegseewwcem ewesem weswez

 

meeeem meeeeewwwe emece>< we emxw
"eweewwe> peeeceeeecw eexww

.euee “Loeweuwwmm cw
meeeecewwwe eeeem we mwmxweee esp Lew .ezp emepm .ece emece Lew :mwmee neweemem

meemeeuee
weeseme>CH

eecmeecw
peeEume>cw

ememeepee
e>wpeeww<
emgmmecw
m>wpeeww<
eewmeeuee
Pecewpeceu

mecmeecw
Fecewpecau

meeee “cmpeee

we geese:

pueucee we
zeceeemgu

museumwe

gmpmewe geezem:

we mem

.N mcemwu

:eweesm

:ewpecew

pampeeo

eezuezeum

 

mmewu xe
mmweewwe> uceecmeme

106

along the social network variables that can be compared directly.

This excludes those dealing with symmetry. This design was char-
acterized by one fixed independent variable, type of sample, with two
levels, normal and schizophrenic. Furthermore, the design has 13
dependent variables (see Figure 3). Although groupings by major class
of variable were developed in the previous phase of analysis, this
stage was not so designed as the emphasis at this point was placed on
the optimum efficiency of the statistical test as opposed to concep-
tual grouping. Also the possible degrees of freedom permitted this
procedure.

The data analysis was conducted by using one multivariate test.
All 13 dependent variables, associated with the four major classes of
variables, were analyzed through the use of a multivariate analysis
of variance. This statistical test was examined at the p_= .05 level,
chosen as an appropriate level of significance due to the high degree
of freedom (gf_= l, 33). Univariate tests were designed to follow if
the multivariate test was significant, using an approximate, parti-
tioned alpha level to avoid an inflated alpha. The partitioned alpha
level for the univariate tests, if needed, was set at p_= .005. This
value is roughly equivalent to the alpha level of the multivariate
test divided by the number of dependent variables.

Stage two. In the second stage, a comparative group design was
used to determine if symmetry within the two major classes of vari-
ables, function and emotion, exists within the two samples. To
develop the design, difference scores were computed between the measures

of functional, affective, and investment outdegree and indegree for

107

.xeeessxm mewueeexe .mmweewwe> xeezee: weweem wwe mcewe meecm
-wewwwe eeeem we cemwweeeee ewes we» Lew .mce mmepm .e2p emcee Lew cmwmee :egemmmm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

w_ u.m e. ".m

 

 

 

eweEem ewceczeerzem ewesem weELez

 

eweEem we maxw "eweewge> “ceeeeeeecw eexwe

megmmeuze pceseme>cH
meemeecw peespme>cH
megmmeuee m>weumww<
eeemeeew e>wpemww<
meemeeuae weeewueczw
meemeeew weeewpeceu
xpwmeee ewgmcewpewmm

mewzmcewpewe; xewewpwee
we :ewpeeeeee

mewcmcewpewee xewewez
we :ewpeeeege

peepcee we xeceecmgw
meeeemwo
mmmeeeueecceeeepew

xwezpe: esp we mem

.m eeemwm

coweose

:ewueezw

peepceu

egspuzwum

 

mmewu we mmweewwe> pceecemmo

108

both samples and compared to the zero vector. This design was char-
acterized by one fixed independent variable, group differences, with
two levels, group of difference scores and group of zero differences:
Furthermore, the design had three dependent variables (see Figure 4).
The data analysis was conducted by using one multivariate test
for each sample. Each statistical test was examined at the p_= .05
level, due to the high degrees of freedom ([3, 15] and [3, 14] for the
normal and schi20phrenic samples, respectively). Univariate tests
were designed to follow if the multivariate test was significant,
using an approximate, partitioned alpha level. The partitioned alpha
level for the univariate tests, if needed, was set at p_= .015. This
value is roughly equivalent to the alpha level of the multivariate

test divided by the number of dependent variables.

m

Subject selection was designed to promote comparability along
those dimensions that affect outcome, as well as to insure the iden-
tification of appropriate subjects from both normal and schizophrenic
populations. In relation Unthe former goal, the promotion of com-
parability, it was determined that steps be taken to control for age,
sex, race, socioeconomic status, and cultural background, but not for
educational and vocational levels, marital status, and personal
economic status. In regard to the latter goal, criteria were estab-
lished, all of which had to be met, for a person to qualify for member—
ship in one of the two samples. A review of the characteristics of

both samples indicated that both goals were essentially met. However,

109

.eweeeceersem ece weELe: .eseem
seem Lew zeueeazm we mwmxwece we» Lew .eze emeum .ezp emege cew cmwmee seweemem .e eeemwm

 

aeueeexm
peeeumm>ew

 

hepeesxm
e>wpeeww< eewuesm

 

 

aepeeexm
wecewpeeee eewpeeew

 

 

 

 

woweecgeercem Le weseezv
meeeem eeceeewwwo
we eeeew

mmeceeewwwo eLeN

mmewo am meweewwe> pceeceema
we ezewm

 

mmeewcwwwwo mecca ”eweewee> weeeeemeeew eexwd

110

in that approximately half of the schizophrenic persons contacted
refused to participate in the study, the results will not reflect the
influence of this undefined, though possibly more hostile and paranoid,
subgroup.

The primary assessment instrument, the Psychosocial Network
Inventory, Modified (PNIM), as its name suggests, was adapted from an
inventory developed by E. Mansell Pattison (Pattison et al., 1975).

As originally devised, the instrument presented the examinee with the
three tasks of developing a list of personally important persons,
rating the nature and quality of the relationships in this social net-
work, and finally, noting the groups of persons who have relationships
in common. The new instrument developed for this study, the PNIM,
retains much of the content and form of the earlier inventory; however,
seven improvements were made. First, it was revised as a structured
inventory, responsive to the characteristics of the target populations.
Second, the subject's selection of social network members was struc-
tured to insure a more orderly and thorough identification process.
Third, the answer sheet was designed to ease recording and facilitate
computer keypunching. Fourth, a response key was added to insure
accuracy of response within the interview. Fifth, a question was
added to assess the content-related aspects of relationships. Sixth,
the questions associated with the supportive and emotional give-and-
take within relationships, i.e., reciprocity, were designed to reflect
the verbal and behavioral dimensions, as well. Seventh, the items of

the PNIM were all conceptually and operationally linked to the

111

dependent variables in this study; the Operationalization of the
variables was reviewed.

Consent procedures were developed to accommodate to the intrin-
sically different characteristics of the two samples and the settings
in which they were identified. Although the procedures relating to
the schizophrenic, as opposed to the normal, were more complex due
to ethical and legal considerations, the focus remained on the pro-
tection of the subjects' welfare.

Separate testing procedures were also developed for both samples
and for corroborating network members; however, the normal and schizo-
phrenic subjects were all faced with seven phases of testing as
follows: (a) explanation of the study, (b) presentation and signing
of the consent forms, (c) completion of the demographic data,

(d) development of the list of personally important persons, (e) assess—
ment of relationships with the PNIM, (f) assessment of the intercon-
nectedness, and (9) selection of the corroborating network members.
Three phases were involved in the contact and examination of corrobo-
rating network members as follows: (a) initial contact including
exploration of interest and confirmation of demographic data, (b) mail-
ing of appropriate forms, and (c) second contact for the assessment of
the relationship with the focal person.

A risk/benefit analysis was undertaken, determining the presence
of potential social and psychological risks to the subjects, but on
the other hand, potential benefits to the subject, the professional
community, and to society. The potential social risks related to the

possible misinterpretation by the corroborating network members of

112

the subject's motive in identifying them and participation in the
study. Clarification in the explanation to the corroborating network
member was considered the primary and only safeguard. The psychologi-
cal risk related to the use of the PNIM in possibly provoking an
untoward reaction; however, the risk was deemed low due to the highly
structured and straightforward nature of the PNIM. Finally, the
benefits were determined to outweigh the possible risks, especially
since safeguards were planned in case problems appeared.

Considering the methods so articulated of the research study, the
University Committee on Research Involving Human Subjects approved
this project on April 7, 1980. The methodology of the research study
was directed toward two interrelated goals, the corroboration of self-
report data of social network variables within both samples and the
main comparison between the two samples along the social network
variables. Two research phases were designed to correspond to the
two goals: Two stages were required foreach phase. Research ques-
tions and null hypotheses were developed as well for both stages of
both phases. In stage one of phase one, the self-report data of both
groups, broken down into 10 variables, were contrasted with the self-
report of the corroborating network members on the joint relationship.
Average difference scores were computed and compared relative to the
zero vector, using a comparative group design. For both samples,
three multivariate analysis of variance statistical tests, examined
at the p_= .10 level, with optional post hoc univariate tests, using
partitioned alpha levels, were used to analyze those dependent vari-

ables associated with the major classes of variables, structure,

113

function, and emotion. One univariate analysis of variance, examined
at the p_= .10 level, was used to analyze the content-related dependent
variable. In stage two of phase one, the average difference scores of
both samples were compared to determine if the two groups were sig-
nificantly different in terms of degree of corroboration of self-report,
using a comparative group design. Three multivariate analysis of
variance statistical tests, examined at the p_= .10 level, with
optional post hoc univariate tests, using partitioned alpha levels,
were used to analyze those dependent variables associated with the
major classes of variables, structure, function, and emotion. One
univariate analysis of variance, examined at the p_= .10 level, was
used to analyze the content-related dependent variable. In stage one
of phase two, the main comparison relative to the social network
variables between the two samples was implemented by comparing along
13 dependent variables, using a comparative group design. One multi-
variate analysis of variance statistical test, examined at the p = .05
level, was used to analyze these 13 variables simultaneously: Group-
ing by major class was not as important conceptually as was efficiency
of the statistical test. The partitioned alpha level for the post hoc
univariate tests was set at p_= .005. In stage two of phase two, the
focus was on symmetry within the two major classes of variables,
function and emotion. Difference scores between the measures of func-
tional, affective, and investment outdegree and indegree were com-
puted in both samples and were compared to the zero vector, using a

comparative group design. For each sample, one multivariate test,

114

examined at the p_= .05 level, was used to analyze the three dependent
variables. The partitioned alpha level for the post hoc univariate

analysis of variance tests was set at p_= .015.

CHAPTER III

RESULTS

The purpose of this chapter is the presentation of the statis-
tical results. For clarity, the material is organized by the various
stages and phases of the analysis. Prior to the description of the
results, a synopsis of the focus and analysis of each stage is pre-
sented. Only significant findings are presented in this chapter;
however, in the following chapter, suggestive, though nonsignificant,
findings will be explored to determine their relevance for future

research.

Results of Phase One: Corroboration of Self-Report

 

Stage One

The focus of stage one of phase one was the corroboration of the
self-report of the normal and schizophrenic samples with selected
members of their social networks. Four major hypotheses were
developed for each sample, corresponding to the four major classes
of social network variables. The 10 dependent variables were grouped
by class within these hypotheses. Using a comparative group design,
the major hypotheses were analyzed through the use of three multi-
variate analysis of variance (MANOVA) statistical tests and one uni-
variate analysis of variance (ANOVA) statistical test for each

sample. Post hoc univariate statistical tests were planned to follow

115

116

when appropriate. The results of the analysis, summarized in

Tables 2, 3, and 4, follow below.

Table 2: Major Tests of Significance for the Null Hypotheses in
Stage One of Phase One

 

Null Class of Type of df

Hypothesis Variable Test ——- ffvalue

 

 

 

Ho1n Structure MANOVAa (3,2) lO.83*
H02n Content ANOVAb (1,4) 23.06*

Normal . ,

Sample Ho3n Function MANOVA (2,3) 42.9l*
Ho4n Emotion MANOVA (4,1) lO3.ll*
Ho1S Structure MANOVA (3,2) 61.00*
Ho2S Content ANOVA (1,4) 125.00*

Schizophrenic

Sample Ho3S Function MANOVA (2,3) 50.96*
Ho4S Emotion MANOVA (4,1) 34.24

 

aMANOVA = Multivariate analysis of variance.
bANOVA = Analysis of variance.
*p_< .10.

Corroboration within the normal sample. Significant differences

 

were found to exist between the self-report of the subjects of the
normal sample and their corroborators along all four classes of social
network variables. In terms of the major null hypothesis associated
with structure (Holn)’ the MANOVA indicated that significant differ-

ences in self-report exist between normal subjects and corroborators,

117

i.e., a lack of corroboration, £_(3, 2) = 10.83, p_< .10 (see Table 2);
however, no significant differences were found in the univariate post
hoc analyses of the three structure-related variables: (a) size of
network cluster, (b) distance, and (c) frequency of contact (see

Table 3). In terms of the major null hypothesis associated with con-
tent (HOZn’ equivalent to H02n-4)’ the ANOVA indicated that signifi-
cant differences in self-report exist between normal subjects and
corroborators, f_(l, 4) = 23.06, p_< .10 (see Table 2). The specific
content-related variable was the number of content areas. In terms

of the major null hypothesis associated with function (Ho3n), the
MANOVA indicated that significant differences in self-report exist
between normal subjects and corroborators, f_(2, 3) = 42.91, p.< .10
(see Table 2). The analysis of the univariate null hypotheses yielded
significant findings for both function-related variables, functional
indegree and functional outdegree, the ANOVA results being f_(l, 4) =

67.94, p_< .05 for Ho and f_(l, 4) = 34.20, p.< .05 for Ho

3n-5 3n-6’
respectively (see Table 3). In terms of the major null hypothesis
associated with emotion (Ho4n), the MANOVA indicated that significant
differences in self-report exist between normal subjects and cor-
roborators, f (4, l) = 103.11, p_< .10 (see Table 2). The analysis

of the univariate null hypotheses yielded significant findings for
three of the four emotion-related variables, affective indegree,
investment indegree, and investment outdegree, the ANOVA results being
f (1, 4) = 36.00, p_ < .025 for Ho4n_7, E_(l, 4) = 33.68, p_< .025 for
Ho4n_9, and f_(l, 4) = 22.50, p_< .025 for Ho4n_]0, respectively (see

Table 3). So, in summary, it appears that the self-report of normal

118

subjects was lacking in corroboration generally along the four major
classes of social network variables, i.e., structure, content, func-
tion, and emotion, and specifically, along the dependent variables of
number of content areas, functional indegree, functional outdegree,

affective indegree, investment indegree, and investment outdegree.

Table 3: Univariate Tests of Significance for the Multivariate Null
Hypotheses in Stage One of Phase One for the Normal Sample

 

 

 

 

Major Null Univariate . _
Hypothesis Hypothesis Variable 9f_ F_value
Ho]n_1 Size, net- (1,4) 8.67
work cluster
Ho1n Ho]n_2 Distance (1,4) 4.38
Ho]n_3 Frequency (1,4) 9.85
of contact
Ho Functional (1,4) 67.94**
3n-5 .
indegree
Ho3n
H03n-6 Functional (1,4) 34.20**
outdegree
H04n-7 Affective (1,4) 36.00***
indegree
H04n-8 Affective (1,4) 10.27
H outdegree
O4n
Ho4n_9 Investment (1,4) 33.68***
indegree
***
H04n-10 Investment (1,4) 22.50
outdegree

 

Note. All significance values correspond to partitioned alpha
levels.
*p.< .03.

HR < .05.
***p_ < .025.

119

Corroboration within the schizophrenic sample. Significant dif-
ferences were found to exist between the self-report of the subjects
of the schi20phrenic sample and their corroborators along three of
the four classes of social network variables, specifically structure,
content, and function; the finding relative to emotion was not sig-
nificant. In terms of the major null hypothesis associated with
structure (H013), the MANOVA indicated that significant differences in
self-report exist between schizophrenic subjects and corroborators,
i.e., a lack of corroboration, £_(3, 2) = 61.00, p_< .10 (see Table 2);
significant differences relative to self-report were also found in
all three structure-related variables: (a) size of network cluster,

f (l, 4) = 31.34, p_< .03 for Ho (b) distance, F_(l, 4) = 12.25,

ls-l’
p_< .03 for Ho]S_2; and (c) frequency of contact, F_(l, 4) = 36.00,

p_< .03 for Ho]s_3 (see Table 4.) In terms of the major null
hypothesis associated with content (H025, equivalent to H025-4)’ the
ANOVA indicated that significant differences in self-report exist
between schizophrenic subjects and corroborators, E_(l, 4) = 125.00,
p_< .10 (see Table 2). The specific content-related variable was the
number of content areas. In terms of the major null hypothesis asso-
ciated with function (H035), the MANOVA indicated that significant
differences exist between schizophrenic subjects and corroborators,
f_(2, 3) = 50.96, p_< .10 (see Table 2). The analysis of the uni-
variate null hypotheses yielded significant findings for both function-
related variables, functional indegree and functional outdegree, the
ANOVA results being F_(l, 4) = 21.05, p_< .05 for Ho3s_5 and f_(l, 4) =
120.92, p_< .05 for ”035-6 (see Table 4). The MANOVA investigating

120

the major null hypothesis associated with emotion (H045) failed to
reach significance (see Table 2). In summary, it appears that the
self-report of schizophrenic subjects was lacking in corroboration
generally along three of the four major classes of social network
variables, i.e., structure, content, and function, and specifically,
along the dependent variables of size of network cluster, distance,
frequency of contact, number of content areas, functional indegree,

and functional outdegree.

Table 4: Univariate Tests of Significance for the Multivariate Null
Hypotheses in Stage One of Phase One for the Schizophrenic

 

 

 

Sample
Major Null Univariate . _
Hypothesis Hypothesis Variable gf- £_value
Ho]S_1 Size, net- (1,4) 31.34*
work cluster
HOls Ho]S_2 Distance (1,4) 12.25*
”015.3 Frequency (1,4) 36.00*
of contact
Ho3s 5 Functional (1,4) 21.05**
' indegree
Ho3S
H035-6 Functional (1,4) 120.92**
outdegree

 

Note. All significance values correspond to partitioned alpha
levels.

*p_< .03.
**p_< .05.

121

Stage Two

The focus of stage two of phase one was the comparison between
the normal and schizophrenic samples in terms of the corroboration of
self-report. Four major hypotheses were developed, corresponding to
the four major classes of social network variables. The 10 dependent
variables were grouped by class within these hypotheses. Using a
comparative group design, the major hypotheses were analyzed through
the use of three MANOVA statistical tests and one univariate ANOVA
statistical test. Post hoc univariate statistical tests were planned
to follow when appropriate. The results of the analysis, summarized
in Tables 5 and 6, follow below.

Significant differences were found to exist in the degree of
corroboration of self-report between the two samples along two major
classes of social network variables, including content and emotion;
the findings for structure and function were not significant. In
terms of the major null hypothesis associated with content (Hoc), the
ANOVA indicated that significant differences in the degree of cor-
roboration exist between the two samples, E (l, 8) = 8.96, p_< .10
(see Table 5). The specific content-related variable was the number
of content areas. A review of the approximate confidence intervals
suggests that the differences between subjects and corroborators were
higher for the schizophrenic sample. In terms of the major null
hypothesis associated with emotion (Hoe), the MANOVA indicated that
significant differences in the degree of corroboration exist between
the two samples, f_(4, 5) = 3.64, p_< .10 (see Table 5). The analy-

sis of the univariate null hypotheses yielded significant findings

122

for only one of the four emotion-related variables, investment indegree,
the ANOVA results being f_(l, 8) = 10.00, p_< .025 for Hoe_9 (see

Table 6). A review of the approximate confidence intervals suggests
that the differences between subjects and corroborators were higher

for the schizophrenic sample. In summary, it appears that significant
differences exist in the degree of corroboration between the two
samples along the two major classes of social network variables, i.e.,
content and emotion, and specifically, along the dependent variables,
number of content areas and investment indegree; in both cases, the
degree of difference between subjects and corroborators was higher

for the schi20phrenic sample.

Table 5: Major Tests of Significance for the Null Hypotheses in
Stage Two of Phase One

 

 

Null Class of Type _
Hypothesis Variable ' of Test 9: E value
HoS Structure MANOVAa (3,6) 2.05
HoC Content ANOVAb (1,8) 8.96*
Hof Function iANOVA (2,7) 1.06
Hoe Emotion MANOVA (4,5) 3.64*

 

aMANOVA = Multivariate analysis of variance.

bANOVA = Analysis of variance.

*p_< .lO.

 

123

Table 6: Univariate Tests of Significance for the Multivariate Null
Hypothesis, Hoe, in Stage Two of Phase One

 

Univariate

 

Hypothesis Variable gj_ Ffvalue

Hoe_7 Affective (1,8) 2.65
indegree

Hoe-8 Affective (1,8) .85
outdegree

Hoe_9 Investment (1,8) 10.00*
indegree

Hoe-10 Investment (1,8) .01

outdegree

 

Note. The significance value corresponds to a partitioned alpha
level.

*9 < .025.

Results of Phase Two: Comparison Between the
Social Networks of Normals and Schizophrenics

 

 

Stage One

The focus of stage one of phase two was the main comparison
between the social networks of the normal and schizophrenic samples.
Using a comparative group design, 13 of the 16 dependent variables
associated with the four major classes of variables were analyzed
through a MANOVA statistical test. Post hoc univariate statistical
tests were planned to follow, if appropriate. The results of the
analysis, summarized in Table 7, follow below.

Significant differences were found to exist between the normal
and schizophrenic samples when the social network variables were

analyzed altogether; however, post hoc procedures revealed that the

124

 

 

 

 

 

Table 7: Univariate Tests of Significance for the Null Hypotheses in
Stage One of Phase Two: Main Comparison Between the Social
Networks of Normals and Schizophrenics.
Null Class of .

Hypothesis Variable Variable §f_ ffvalue
HoS_1 Structure Size of network (1,33) 45.86*
HoS_2 Interconnectedness (1,33) 2.31
Hos_3 Distance (1,33) 6.74
Hos_4 Frequency of contact (1,33) 3.44
HOc-S Content Proportion of (1,33) .32

uniplex relationships
HOc-O Proportion of (1,33) 2.24
multiplex relationships
HoC_7 Relationship density (1,33) 1.11
HOf-8 Function Functional indegree (1,33) .73
Hof_9 Functional outdegree (1,33) 1.17
Hoe-ll Emotion Affective indegree (1,33) 4.03
Hoe__12 Affective outdegree (1,33) 1.58
Hoe__14 Investment indegree (1,33) 1.99
Hoe_15 Investment outdegree (1,33) 3.17

 

*p_ < .005.

125

differences were significant only for one dependent variable, size of
network. The MANOVA analyzing the comparison between the two samples
along the 13 dependent variables yielded the following: F_(l3, 21) =
6.55, p <.05. (The actual significance of the ffvalue was a striking
p_= .00008.) Post hoc findings between the two samples yielded only
one significant finding for the structure-related dependent variable,

size of network, f_(l, 33) = 45.86, p_< .005 for Ho (see Table 7).

_.._.._, ._ 1

s-l
In terms of size of network, the normal group had a mean of 36.23

with a standard deviation of 13.31, and the schizophrenic group had a

mean of 12.65 and a standard deviation of 5.61.

Stage Two

The focus of stage two was the assessment of function-related and
emotion-related symmetry within the two samples. Accordingly, func-
tional, affective, and investment symmetry were grouped into a major
null hypothesis for each sample. The difference scores were computed
for the measures of outdegree and indegree for the three dependent
variables, and the data were analyzed through a MANOVA statistical
test. Post hoc univariate statistical tests were planned if approp-
riate. The results of the analysis, summarized in Tables 8 and 9,
follow below.

Symmetry within the normal sample. In terms of the normal

 

sample, no significant differences were found to exist between the
measures of functional, affective, and investment outdegree and
indegree when the variables were analyzed simultaneously; therefore,

support was not found to reject symmetry as significant differences

126

between the measures of indegree and outdegree would mean that

symmetry for the normal sample does not exist (see Table 8.)

Table 8: Major Tests of Significance for the Multivariate Null
Hypotheses Pertaining to Symmetry in Stage Two of Phase Two

 

 

 

Null
Sample Hypothesis gj_ ffvalue
Normal Hofe/n (3,15) 3.03
Schizophrenic Hofe/s (3,14) lO.91*
*p_< .05.

Symmetry within the schizophrenic sample. In terms of the

 

schizophrenic sample, significant differences were found to exist
between the measures of functional, affective, and investment out-
degree and indegree when the variables were analyzed simultaneously;
therefore, the existence of symmetry was disproven for the multi-
variate hypothesis. Significant univariate findings were found rela-
tive to functional symmetry. In terms of the major null hypothesis
associated with the schizophrenic sample (Hofe/s)’ the MANOVA indicated
that significant differences exist in the three dependent measures

of outdegree and indegree within the schizophrenic sample, f_(3, 14) =
10.91, p_< .05 (see Table 8). The analysis of the univariate null
hypotheses yielded significant findings for only one dependent vari-
able, functional symmetry, the ANOVA results being F_(1, 16) =

15.58, p_< .015 for Hofe/s-lO (see Table 9); therefore, functional

symmetry was disproven,iandaffective and investment symmetry were not

127

rejected. A review of the approximate confidence intervals for the
results related to functional symmetry indicated that functional
indegree was significantly larger than functional outdegree for the

schizophrenic sample.

Table 9: Univariate Tests of Significance for the Multivariate Null
Hypotheses Pertaining to Symmetry in Stage Two of Phase
Two: Schizophrenic Sample

 

Null

 

Hypothesis Variable gj_ Efvalue
Hofe/s-lO Functional symmetry (1,16) 15.58*
Hofe/s-l3 Affective symmetry (1,16) 4.35
h0fe/s-l6 Investment symmetry (1,16) .04

 

Note. The significance value corresponds to a partitioned alpha
level.

 

fp < .015.

Summary

The focus of stage one of phase one was the corroboration of the
self-report of the normal and schizophrenic samples with selected
members of their social networks. As presented in Table 10, signifi-
cant differences were found to exist between the self—report of the
subjects and their corroborators in the normal sample along all four
classes of social network variables and in the schizophrenic sample
along three of the four classes of variables: structure, content,

and function. In terms of the structure-related dependent variables,

128

.mpwemww pceuwwwcmwm n x
.mu—emee uceuwwwcmwmeez n o

 

 

 

 

><O><><

 

 

 

 

 

 

 

 

 

><><><><><

OOO><><

 

x

x

 

uwcewgeerzum

 

weEeoz ewcewzeercem

 

 

weELez

mmwmmeuee pcmsume>cw

emwmeecw ucmsumm>cw

ememeeuee e>wuemww<

ewcmmecw m>wueeww<
:ewpesm

meemmeuee wecewpeezw

eewmeecw Fecewpeczw
:ewueczw

meewe ucmucee we weeszz
pzeuceu

peepcee we Aeeeecmww

meeepmwe

weumzwe xee3ue: we mem

weepeeeum

 

 

<>oz<

<>ez<z
emew weewemweeem

 

mmewo xe mmweewee> pceeememo

 

meweEem eweeeceergem eee weswez we» wew eueo pweeemuwwmm we
:ewueweeewweu esp we mwmxwec< we» ”wco emege we use emeum Lew mpwzmmm awe» weewumwueum ”ow eweew

129

no significant differences were found in the normal sample, while
significant differences were found for all three variables in the
schizophrenic sample: size of network cluster, distance, and fre-
quency of contact. In terms of the content-related dependent vari-
able, number of content areas, significant differences were found in
both samples. In terms of the function-related dependent variables,
functional indegree and outdegree, significant differences were again
found in both samples. In terms of the emotion-related dependent
variables, significant differences were found in the normal sample
for affective indegree, and investment indegree and outdegree.

The focus of stage two of phase one was the comparison between
the normal and schizophrenic samples in terms of the corroboration
of self-report. As presented in Table 11, significant differences
were found to exist in the degree of corroboration of self-report
between the two samples along two major classes of social network
variables, including content and emotion; the findings for structure
and function were not significant. The two samples varied signifi-
cantly along (a) number of content areas and (b) investment indegree,
variables related to content and emotion, respectively. Finally, the
degree of difference between subjects and corroborators was higher
for the schizophrenic sample.

The focus of stage one of phase two was the main comparison
between the social networks of the normal and schizophrenic samples.
Significant differences were found to exist between the normal and
schizophrenic samples when the dependent variables were analyzed

simultaneously; however, as presented in Table 12, post hoc procedures

130

revealed that the differences were significant only for one dependent
variable, size of social network. In terms of size of network, the
normal group had a mean of 36.23 with a standard deviation of 13.31,
and the schizophrenic group had a mean of 12.65 and a standard devia-

tion of 5.61.

Table 11: Statistical Test Results for Stage Two of Phase One: the
Analysis of Sample Differences in Self-Report Data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Relevant Dependent Statistical Test
Varlab‘es MANOVA ANOVA
Structure 0
Content x
Number of content areas x
Function 0
Emotion X
Affective indegree 0
Affective outdegree 0
Investment indegree x
Investment outdegree 0
0 = Nonsignificant results.

><
II

Significant results.

131

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 12: Statistical Test Results for Stage One of Phase Two: the
Main Comparison of Sample Differences Along All Social
Network Variables, Excepting Symmetry
Statistical Test
ANOVA

Structure Size of network X
Interconnectedness 0
Distance 0
Frequency of contact 0

Content Proportion Of. . O
uniplex relationships
Proportion of 0
multiplex relationships
Relationship density 0

Function Functional indegree 0
Functional outdegree O

Emotion Affective indegree O
Affective outdegree 0
Investment indegree 0
Investment outdegree O

O = Nonsignificant results.
X = Significant results.

 

132

The focus of stage two of phase two was the assessment of
function-related and emotion-related symmetry within the two samples.
As presented in Table 13, significant differences were found to exist
between the measures of functional, affective, and investment out-
degree and indegree (when the variables were analyzed simultaneously)
in the schizophrenic sample, but not in the normal sample. Essen-
tially, though, functional, affective, and investment symmetry were
not disproven for the normal sample. In terms of the schizophrenic
sample, significant differences were associated with functional
symmetry, supporting functional asymmetry. In this case, functional
indegree was found to be significantly larger than functional out-
degree. Furthermore, affective and investment symmetry were not

rejected for the schizophrenic sample.

133

.mpwemew peeewwwcmwm

ll
><

.mpwzmew uceewwwcmwmcez

11
O

 

o zeweesxm gemEpme>cw

 

o acueeexm e>weeeww<

 

x xeuesexm Fecewuecew

 

x o xwuesezm eeeeewu

 

eweewceerzem weseez ewceeceercem weawez

 

 

 

 

 

<>ez< <>ez<z
memew weewemweeem

 

 

ewceeeeergem eee weEwez .mPQEem
seem Lew xcpeaexm we mwmxwee< mew ”egw emeee we ezw emeum Lew muwemem pmew weewpmwueum ”mp eweew

CHAPTER IV

SUMMARY, CONCLUSIONS, AND IMPLICATIONS

Summary

This study was undertaken with two purposes in mind. The major
purpose of this study was the comparison of the social networks of
normals and schizophrenics, the goal being the identification of
psychosocial factors that appear to be associated with the schizo-
phrenic disorder. The comparison was made along four major classes
of social network variables: (a) structure, the basic morphological
characteristics; (b) content, the nature of shared activities;

(c) function, the transaction of support; and (d) emotion, the experi-
encing of affect. A secondary, yet essential, purpose was the evalua-
tion of the research assumption conceptually basic to much of the
previous research that self-report data are a sufficient indicator

of the actual status of the social network. The analysis was designed
so that all questions and items of the assessment instrument could be
corroborated.

In the review of literature, six relevant topics were covered and
conclusions were drawn. First, in terms of the importance of the
study, it was concluded that the schizophrenic disorder was a worthy
area of study due to its extremely disrupting impact upon the affected
individual and society. Second, in terms of social network analysis,
it was concluded that this new methodology has wide applicability and

134

135

value, and as a concept, allows for the integration of multiple

levels of analysis. Furthermore, the type of social network con-
ceptualization was adjudged to be quite pertinent to research focus
and general perspective. Third, in terms of the theoretical perspec-
tive, it was concluded that the concepts of schizophrenic withdrawal
and isolation were the theoretical foundation of this study. Fourth,
in terms of the key social network variables, predictions which follow
below were made regarding the results of the analysis, based on the
concepts of withdrawal and isolation:

1. In relation to structure, the social network of the schizo-
phrenic, as opposed to the normal, will be characterized by fewer
network members, less interconnectedness, closer proximity of members,
and higher frequency of contact.

2. In relation to content, the social network of the schizo-
phrenic, as opposed to the normal, will be characterized by a greater
proportion of relationships in which only one activity is shared
(proportionately more uniplex relationships), a smaller proportion of
relationships in which more than one activity is shared (proportion-
ately fewer multiplex relationships), and a relatively smaller pro-
portion of shared activities per person (less relationship density).

3. In relation to function, the social network of the schizo-
phrenic, as opposed to the normal, will be characterized by a greater
degree of support provided by the network members to the focal person
(more functional indegree), a smaller degree of support provided by
the focal person for network members (less functional outdegree), and

a lack of reciprocity (functional asymmetry).

136

4. In terms of emotion, the social network of the schizophrenic,
as opposed to the normal, will be characterized by a lesser degree
of positive feelings from network members for the focal person (less
affective indegree), a lesser degree of positive feelings from the
focal person for the network members (less affective outdegree), a
lack of reciprocity in the feelings (affective asymmetry), a greater
degree of emotional investment of network members in the focal person i
(more investment indegree), a lesser degree of emotional investment [
of the focal person in network members (less investment outdegree),
and finally, a lack of reciprocity in the emotional investment
(investment asymmetry).

For the fifth of the six topics, the general importance of the
social network to the schizophrenic disorder, it was concluded that
all components of the social network, i.e., family, extended family,
significant nonkin, can have impact on the life of the schizophrenic,
either in a protective and supportive way or in a provocative and
deteriorating manner. Sixth and finally, in terms of the specific
relationships between the social network and psychopathology, in
general, and the schizophrenic disorder, in particular, the following
was concluded: (a) the type, quantity, and quality of relationships
within the social network appear to be linked to psychopathology;
and (b) smaller network size, proportionately fewer relationships of
more than one activity, relatively more support from network members,
nonreciprocal and dependent relationships, and negative perception of
the network consistently characterized the social network of the

schizophrenic across all studies. The findings associated with actual

137

network sizes and interconnectedness were inconsistent, possibly
varying due to the differences in conceptualizations of the social
networks and in the familial characteristics of the samples studied.
The decision was made to attempt a replication in the area of
social networks of schizophrenics because of the importance of the
research and difficulty with the interpretation and integration of
the previous findings due to the following five reasons: (a) inade-
quate sample descriptions and differing populations, (b) differences
in conceptualization of and Operationalization of variables, (c) dif-
fering conceptualizations of the social networks under study,
(d) inconsistencies in the reported data, and (e) most important,
lack of validation of the self-report data.
The normal sample was composed of 18 subjects, and the schizo-
phrenic sample, of 17 subjects. Selection of the subjects was
designed in such a way as to promote comparability along those dimen-
sions that affect outcome (age, sex, race, socioeconomic status, and
cultural background), yet insure the identification of appropriate
subjects from both populations. A review of the characteristics of
both samples indicated that both goals were essentially met. However,
insofar as approximately half of the schizophrenic population contacted
refused to participate in the study, the data would not reflect the
influence of this group. Although the composition of the subgroup
remains unclear and undefined as they were not studied, it is noteworthy
that several members of this subgroup were characterized as hostile and

paranoid.

_..._.-.__

138

All data were gathered through the use of the Psychosocial
Network Inventory, Modified (PNIM). In its original form, the instru-
ment presented the examinee with the three tasks of developing a list
of personally important persons, rating the nature and quality of the
relationships, and noting the groups of persons who have relationships
in common. As modified into a structured interview for this study,
the PNIM was designed overall to be more thorough in identification
of network members, facile and accurate in the recording of subject
responses, reflective of the activities shared in relationships and
the behavioral and verbal aspects of reciprocity, and last, readily
operationalized.

Separate consent and testing procedures were developed for both
samples to accommodate to the different characteristics of the two
samples, namely the differing legal and ethical considerations, and
need for structure and support in the testing situation.

The methodology of the study was directed toward interrelated
goals corresponding to two phases in the analysis: (a) phase one,
the corroboration of self-report data of social network variables
within both samples; and (b) phase two, the main comparison between
the two samples along the social network variables. In the phase one
analysis, the self-report data of five subjects from both samples and
at least four, if not five, members of their social networks were
contrasted in order to determine the degree of discrepancy. In stage
one of this phase, the analysis focused on the comparison of 10
dependent variables grouped by major class of variable within each

sample. In stage two, the analysis focused on the comparison of the

139

degree of corroboration between the two samples. In the phase two
analysis, both samples were, in general, compared along 16 dependent
variables, grouped by the four major classes of social network vari-
ables within this study. In stage one of this phase, the analysis
focused on the comparison of 13 variables directly between the two
samples, and in stage two, the comparison of the three variables
associated with symmetry within the two samples. The analyses were
carried out through the use of multivariate statistical procedures

and post hoc univariate analyses of variance.

Conclusions

 

In the following section, the conclusions of the study are pre-
sented. It is important to remind the reader that these findings do
not reflect the influence of a subgroup of the schizophrenic popula-
tion who were unwilling to participate. This subgroup of basically
undefined composition was, in large part, characterized as more hos-
tile and paranoid than those who participated. The following conclu-
sions apply by definition to those schizophrenic subjects sharing the
common attitude regarding willingness to participate in a scientific

study.

Stage One of Phase One

 

Self-report was not adequately corroborated in either normal or
schizophrenic samples. Within the normal sample, it was concluded
that self-report was not adequately corroborated, as significant dif-
ferences were found to exist in self-report along all four major

classes of social network variables and 6 of the 10 related dependent

140

variables. Within the schizophrenic sample, it was concluded that
self-report was not adequately corroborated as significant differ-
ences were found to exist along three major classes of social network

variables and six related dependent variables.

Stage Two of Phase One

 

The normal and schizophrenic samples do not appear to vary sub-
stantially in terms of corroboration of self-report, as differences
were found in only two dependent variables and, as expected, less
agreement was found to exist between subject and corroborator within
the schizophrenic sample regarding these two variables. As signifi-
cant differences were found in two major classes of social network
variables, content and emotion, but only 2 of the 10 related dependent
variables, number of content areas and investment indegree, it was
concluded that overall in relation to the degree of corroboration
the two samples do not vary substantially; however, they do vary

significantly in terms of the reported number of content areas and

investment indegree. More specifically, the differences were in an
expected direction as more disagreement, i.e., less corroboration,

in self-report was found to exist between the schizophrenic sample
and their network member/corroborators than between the normal sample

and their network member/corroborators.

Stage One of Phase Two

 

In direct comparison, the social networks of the two samples of
normals and schizophrenics appear to differ in terms of perception

of network size only; specifically, the schizophrenic appears to have

141

a smaller social network than the normal. As significant differences
were found to exist in relation to the 13 dependent variables-~not
associated with symmetry--assessed simultaneously, but in the post
hoc analysis only one dependent variable, network size, it was con-
cluded that the two samples differ in network size: a mean size of
36.23 for the normal sample and a mean size of 12.65 for the schizo-

phrenic sample.

Stage Two of Phase Two

 

In indirect comparison the following was learned: (a) for the
normal sample, relationships appear to be perceived as reciprocal in
terms of the transaction of support and mutual in terms of the experi-
encing of affect; and (b) for the schizophrenic sample, relationships
appear to be perceived as mutual relative to the experiencing of
affect, but nonreciprocal relative to the transaction of support,
placing the schizophrenic in a dependent position. For the normal
sample, this conclusion was supported by the lack of significance
found overall in relation to symmetry, the result being that func-
tional, affective, and investment symmetry exist for the normal
sample. For the schi20phrenic sample, this conclusion was supported
by the finding of significant differences in terms of symmetry gen-
erally and functional symmetry specifically, the results being that
functional asymmetry exists. In relation to the differences, func-
tional indegree was significantly larger than functional outdegree,
ostensibly placing the schizophrenic in a perceived dependent position

relative to his or her social network.

 

142

Discussion

 

Phase One

The conclusions in both stages of phase one are rather startling
initially. One might expect the degree of corroboration between
subjects and network members to be low for the schizophrenic sample
but not for the normal sample. Indeed, the results appeared less
significant for the normal sample in stage one of phase one for the
structure-related variables, although not overall. However, the con-
clusions of stage two of phase one tend to confirm that the two
samples do not vary substantially in terms of corroboration. Thus
the findings that self-report was not adequately corroborated for
both samples and that the two samples do not vary substantially in
relation to corroboration of self-report, though unexpected, appear
conclusive.

Lack of corroboration in both samples casts doubt upon the
research assumption that self-report is an accurate and sufficient
indicator of the actual status of the social network present in
earlier studies (see Pattison, deFrancisco, Wood, Frazier, & Crowder,
1975; Pattison, Llamas, & Hurd, 1979; Tolsdorf, 1976). To understand
this statement, one must remember that the data were generated from
an instrument that taps conscious-level thought, primarily on phenomena
that vary along a continuum of subjectivity and objectivity, all
assessed subjectively. The dependent variables associated with the
structure, content, and function of the social network tend to be

objective, while those associated with emotion tend to be subjective.

143

The objective variables could be assessed by observation and con-
firmed, while the subjective (emotion) variables are always products
of the subjects' private and conscious mental life, not subject to
confirmation. Accordingly, lack of consensus regarding the objective
variables tends to suggest that the subjects and corroborators have
differing assessments of what both have observed jointly, questioning
the accuracy of such data. Essentially, one does not know which might
be accurate or if either is. In terms of the subjective variables,

it tends to suggest that the subjects and corroborators have differ-
ing perceptions of what both experience internally, of course based

on external evidence. The implication is that the perception of
emotional experience between persons is not sufficiently shared.

While some degree of consensus is necessary for functioning in

reality (and not rejected here), it appears that there are signifi-
cant discrepancies existing in the subjective evaluation of both
objective and subjective aspects of the social network within both
samples; therefore, self-report data from subjects are an insufficient

and possibly inaccurate, i.e., unreliable, source of information in

 

terms of the objective variables and is not reflective of the emotional
experience of the social network members. Based on this rationale,
self-report data in this study can be meaningful, however, in that they
represent the subjects' perception of their social network, and as such
have important clinical significance.

The lack of reliability in the self-report data appears to have
impact upon the power of the statistical tests of phase two of the

analysis. By way of explanation, it is known that the reliability of

144

measures is positively associated with the power of the statistical
test in the isolation of existing significant findings. The unrelia-
bility of measures has an inflating effect upon the standard error.
Therefore, it was expected that the demonstrated lack of reliability
might have obscured statistically significant findings in phase two.

To determine if this may be the case, sample means from the group of
corroborator data were computed corresponding to 14 of the 16
variables--means could not be calculated for the variables, size and
interconnectedness, as they strictly relate to the total social net-
work. In a comparison of the corroborator data with the actual (subject)
data, major differences were identified for 8 of the means, while 20
were comparable. In a cursory, nonstatistical analysis of the differ-
ing means of the corroborators, considerable differences were found
between means associated with the normals and schizophrenics along four
variables. It is also noteworthy that along all other variables these
findings from the corroborator data were supportive of the actual
results of the study. In conclusion, then, as an assessment of the
corroborator means suggests significant findings along four additional
variables, it supports the contention that lack of reliability in the
self-report data is associated with decreases in power of the statis-
tical tests of phase two, possibly obfuscating the presence of existing
significant findings. Furthermore, the analysis not only strengthens
the argument associated with reliability, but implies that the results

may vary considerably, depending upon the perspective taken.

 

145

Phase Two

As the results represent perceptions of the social networks, as
opposed to their actual status, the conclusion of stage one of
phase two, the main comparison, would need to be reworded as follows:
The social networkscyfthe two samples differ in terms of perceptions
of network size. The conclusions of stage two of phase two have
already been stated in relation to the subjects' perceptions, since
an indirect analysis was implemented.

The finding that the social networks vary in terms of size is
shared in the previous studies. Cohen and Sokolovsky (1978) found
significant differences between two schizophrenic samples and a non-
psychotic control sample, even though the populations studied were
not living with family. These results were based on observation as
well as self-report. The two mean sizes for the schizophrenic groups
in their study, 10.30 for the more disturbed group with residual
symptoms and 14.80 for the less disturbed group with minimal residual
symptoms, are strikingly comparable to the mean size of 12.65 for
the schizophrenic group in this study. Their mean size of 22.50 for
the nonpsychotic control group, however, is not similar to the mean
size of 36.23 found for normals in this study. Tolsdorf (1976) came
close to finding significant differences between his two samples,
presumed to be living with family, recording a mean size for the normal
(medical) group of 37.80, close to the results of this study, but not
for the schizophrenic group (mean size of 29.80). One must remember

that he pressed for the largest possible number with the schizophrenic

 

146

group to determine the total amount of support available to the schizo-
phrenic. Using the Psychosocial Network Inventory, Pattison et a1.
(1975) in the early study found considerable difference between the
network size in the normal and schizophrenic samples. In relation

to this study, his figures for network size varied for the schizo-
phrenic sample (mean range of four to five) and for the normal sample
(mean range of 20 to 30). In his most recent study, Pattison et a1.
(1979) found differences again between normal and schizophrenic groups,
but the results for the schizophrenic group were closer to those found
in this study (mean range of 10 to 12). Accordingly, the results of all
previous studies are consistent regarding differences in perception of
network size. Furthermore, even though different conceptualizations of
the social network were adopted in each study, consensus is approached
in terms of overall size of the social network of the schizophrenic,
except for one researcher who attempted to secure an expanded social
network to determine the amount of available social support: Such

was not the case for the normal sample.

The conclusions regarding symmetry are as expected for the normal
sample. The findings that normals perceive relationships character-
ized by symmetry, i.e., mutually supportive and caring, certainly are
in line with the theoretical work of Hollander (1967) and Thibaut
and Kelley (1959). In essence the findings support the contention
that reciprocity in normal relationships is responsible for their
maintenance. Furthermore, it could be said that due to the presence
of reciprocity in the relationships of normals, their relationships

are rewarding and satisfying.

 

147

As expected, symmetry within the schizophrenic sample was
rejected. In the follow-up analysis, a perceived lack of symmetry
was identified in the supportive aspects of the schizophrenics' rela-
tionships; namely, the schizophrenic perceives himself or herself as
the dependent party in social relationships. This finding is in
agreement with the results of the previous studies (see Cohen &
Sokolovsky, 1978; Pattison et al., 1975, 1979; Tolsdorf, 1976). In
one study with the focus on objectivity, Cohen and Sokolovsky (1978)
inferred that the dependent position of the schizophrenic is an
actual, objective phenomenon. Accordingly, one could speculate that
in this area the observation of the schizophrenic may, in fact, cor-
respond to reality. However, in contrast to predictions, the emo—
tional aspects of relationships were perceived as mutual both in type
and strength of feeling, meaning that the schizophrenic does not per-
ceive himself or herself as having more negative and less strong
feelings for network members than are returned. These findings sharply
contrast with the impressionistic findings of Tolsdorf (1976) and of
the two studies of Pattison et a1. (1975, 1979) which suggest that
the schizophrenic has a very negative perception of her or his social
network. Cohen and Sokolovsky (1978) did not explore the emotional
aspects of the social network.

The relative absence of significance in phase two is, in itself,
a major finding which tends to raise doubts regarding some earlier
findings and the influences of withdrawal and isolation. Significant

differences were found in only 1 of 13 dependent variables in direct

148

comparison in stage one, and one of three dependent variables in
indirect comparison in stage two. These results tend to refute earlier
findings (actually based on perceptions) with regard to interconnected-
ness, relationship density, proportion of uniplex and multiplex rela-
tionships, and the type and strength of emotion. Furthermore, the
significance of only one dependent variable in stages one and two,
network size and functional asymmetry (namely, the perceived dependent
position of the schizophrenic), lends only weak support to the theo-
retical foundation of this study, withdrawal and isolation. In
explanation, it is posited that the schizophrenic's conscious per-
ception of social relationships, as reflected in the PNIM, does not
appear to be substantially different from normals. The differences

in the social networks of normals and schizophrenics may, in fact,
exist but were not perceived. Also, the PNIM may not have tapped the
appropriate level of personality to detect those differences as well.
Perhaps they could be assessed through measures of unconscious process.
This is described more fully in the following section. Last, as
suggested in the previous section, the unreliability of the measures

may be responsible as well for the lack of findings.

Implications for Future Research
The implications for future research are associated with the
four following areas: (a) significant findings; (b) suggestive,
nonsignificant findings; (c) absence of findings; and (d) generaliza-
bility of findings. In this section, the meaning and impact of each
area are explored in relation to previous research, with recommenda-

tions presented for future research.

 

149

Significant Findings

The significant findings of phases one and two in this study
should have impact on social network analysis. The conclusions of
phase one, demonstrating a lack of consensus between subject and
corroborator in both samples and a lack of substantial difference in
discrepancy of self-report between both samples, strongly suggests
that future research on social network analysis, possibly for all
types of populations, be based on objective measures, e.g., observa-
tion, various types of counting, logs of activity, etc., if the
researchers intend to infer to the actual status of the social net-
work, as opposed to subjective perception. As results based on
subjective measures cannot be considered sufficient, reliable, or
accurate, researchers can no longer use the assumption that self-report
on a social network is equivalent to the actual status of the social
network. On the other hand, subjective assessments of the social net-
work are seen as valuable when the focus is clearly placed on subjective
perception or clinical meaning; however, as stage one of phase two
results suggest that both samples of normals and schizophrenics tend
to view their social networks similarly, the PNIM does not appear to
evoke much useful clinical material; again, this may be a function
of its focus on conscious-level thought. In closing, it is noted
that social network analysis first evolved as a metaphorical phe-
nomenon, conceptualized to identify a subjective social structure,
and by definition remains in part a subjective experience. In a
sense, it cannot get away from its theoretical roots. However,

attempts can be made to develop objective and reliable measures in the

 

 

150

three more objective areas of social network analysis termed structure,
content, and function in this study, and use conceptualizations of the

social network that are assessed more easily with objective measures.

Suggestive Findings

 

Falling into both phases of the research, the suggestive, non-
significant findings, closely approaching significance levels, may
possibly support and enhance the significant findings of phase one
and phase two. Further research in each area is necessary to confirm
or disconfirm their actual significance. In stage one of phase one
for the normal sample, the analysis of the structure-related vari-
ables, size of network cluster and frequency of contact, and the
emotion-related variable, affective outdegree, yielded suggestive
findings. If these had been significant, then 9 of the 10 total
dependent variables would have supported the conclusion of lack of
corroboration in the normal sample. In stage one of phase one for
the schizophrenic sample, the analysis of emotion generally and the
emotion-related variable, affective indegree, yielded suggestive
findings. The other three emotion-related variables were significant.
Had this class of variable and related dependent variable been sig-
nificant, then all 10 variables would have supported the conclusion
of lack of corroboration in the schizophrenic sample. In stage one
of phase two, the analysis of the structure-related variable, dis-
tance from focal person, i.e., proximity of network members, yielded
suggestive findings. If it had been significant, then it would have

appeared that the schizophrenic perceives his or her social network

 

 

151

living closer to him or her than does the normal. This might possibly
relate to the schizophrenic's inability to sustain important relation-
ships over distance, as compared to normals. Finally, in stage two
of phase two, the analysis of symmetry generally for the normal sample
yielded a suggestive finding, investment symmetry being significant in
the post hoc analysis. If the finding had been significant, then some
basic ideas would have to be revised regarding perception of symmetry
in relationships with normals. In summary, pursuing further research
into the areas described above might strengthen (or weaken) the support
relating to corroboration of self-report and expand our knowledge about
the perception of the social network of normals, in terms of symmetry,
and of schizophrenics, in terms of proximity.

Given the above rationale that close, though nonsignificant,
findings may prove suggestive in future research, it seems important
to examine an opposing, though complementary, line of reasoning, namely,
that mildly significant results may be specious and misleading. From
an investigation of the levels of probability, it appears that all but
the following two measures were highly significant: (a) distance, a
variable in stage one of phase one; and (b) emotion, a class of vari-
able in stage two of phase one. In relation to the variable distance,
had it not been significant, then there would still not be any change
in the overall conclusion (lack of corroboration for the schizophrenic
sample). In relation to the class of variable emotion, had it not
been significant, then the conclusion of no substantial differences
in corroboration of self-report between normals and schizophrenics

would have been strengthened, not challenged. Therefore, these two

 

 

152

measures of mildly significant probabilities do not appear to suggest
substantial reversals in the conclusions, suggestive for further
research; in fact, in the latter case, the conclusion would be

strengthened.

Absence of Findings

 

The absence of significant subjective and objective findings in
phase two, primarily the main comparisons, may be attributed to the
design of the instrument and the diverse nature of the social network,
respectively, as well as to the unreliable data--discussed earlier.
First, the design of the instrument enables one to assess the
examinee's conscious perception of his or her social network.

This process excludes unconscious information, e.g., inner object
life, intrapsychic dynamics, fantasies, drive states, etc., that
possibly could be used to differentiate between normals and schizo-
phrenics as they relate to members of the social network. Uncon-
scious assessment of emotion may have determined differences between
the two samples, not accessible to instruments tapping conscious-
level thought. Thus, it is recommended that future researchers con-
sider the use of unconscious assessments of subjective variables in
social network analysis. Also, the diverse nature of the social net-
work may be responsible for the lack of objective results; namely, the
social network contains groups of differing nature, relation, and
importance so that the significant trends may disappear in the varying
influences. Specifically speaking, differences in structure, con-

tent, function, and emotion in the social networks of normals and

 

153

schizophrenics may appear if family, relatives, friends, co-workers,
etc., were assessed separately. Therefore, it is recommended that
future researchers consider investigating subgroups within the social
network to assess social network variables more closely. In summary,
by using instruments tapping unconscious levels and by focusing on
various subgroups of the social network in the analysis (using more
objective measures), future researchers may detect more significant
findings between social networks of normals and schizophrenics, the
two methods assessing subjective and objective realities, respec-

tively. Hopefully, at this point more definitive statements could be
made about the relative and interactive effects of isolation and

withdrawal.

Generalizability of Findings

 

The generalizability of the findings is duly limited and obscured
by the loss of potential schizophrenic subjects who refused to partici-
pate in this study. Although several of these potential subjects were
characterized as more hostile and paranoid than the participating sub-
jects, this group, nonetheless, was not studied and, accordingly,
remains undefined and vague. Therefore, except for the basic com-
monality, willingness to participate in a scientific study, the sub-
jects of the study cannot be differentiated from those who refused
participation. Notably, though, the willingness to participate, a
behavioral characteristic, is associated with the degree of investment
in other people or, stated conversely, the degree of withdrawal in

the schizophrenic. Perhaps in the methodology of the study, the

 

154

researcher created conditions whose demand characteristics separated
the two groups along the continuum of capacity for other-directed
investment. (Of course, those schizophrenics most withdrawn were not
included in the study because they were nonverbal, autistic, etc., but
this discussion does not apply to this particular population.) In
other words, the schi20phrenic who was more withdrawn and less capable
of investment in others would avoid participation. Certainly this
would be true of those characteristically hostile and paranoid. This
situation would have significant impact on the results, if those more
withdrawn were excluded a priori. Indeed, the measures were designed
to assess this dimension of withdrawal. It is suspected that, in
fact, this rationale may account for a sizable number of refusals, yet
all this remains speculative. Other variables as well could also
account for the refusals, e.g., the personal manner of the researcher,
the time commitment, other demands upon the subjects, etc. Only
further research could approach this problem, but then one is faced
with the ethical and methodological dilemmas of attempting to study
those who resist. While the group of subjects seems well-defined in
many respects, reflecting the criteria, the loss of potential subjects
obscures the sample characteristics and representativeness of the p0pu-
1ation and, resultingly, limits the generalizability. The two groups
are only distinguished behaviorally by participation in a scientific
study, possibly associated with capacity for investment in others or

degree of withdrawal as well as degree of hostility and paranoia.

155

Clustering: Implications for Social Network
Analysis and Research

 

 

At this point in the development of social network analysis,
there does not appear to be a means of reflecting the phenomenon of
clustering, namely, the naturally occurring grouping of network mem—
bers who both know each other and share a unifying characteristic.

One example of a cluster is the group of co-workers as they share
relationships and the common task of work. Other examples include
family members, relatives, neighbors, fellow church members, etc. This
lack of methodology is unfortunate because an examination of the data
identified what appeared to be highly developed clusters in the social
networks of normals and nonexistent or poorly developed clusters in
the social networks of schizophrenics. Therefore, the two samples may
actually differ in terms of clustering within the social networks.

The only variable that approaches an estimation of this phenomenon

is interconnectedness. However, this variable was based on a sim-
plistic and misleading--though heuristic--model of the social network,
the primary star. In this conceptualization, relationships are
conceived of as radiating around the focal person, with interrela-
tionships perceived of as connections among the network members
without any attention to a unifying characteristic. Accordingly, the
variable provides a raw estimate of the density of relationships
within the social network but fails to reflect clustering. Since
clustering may be a distinguishing factor between the two samples, it

is recommended that future researchers develop an appropriate

 

EA" emJ .‘o - V1

 

 

156

methodology to identify clustering and explore such differences

between the social networks of normals and schizophrenics.

Implications for Diagnosis and Treatment
of the Schizophrenic

 

 

The findings of this study pertain to the diagnostic assessment
and treatment planning of the schizophrenic. In the paragraphs that
follow, the implications of these findings are explored and, when
appropriate, reconmendations are presented.

The lack of corroboration of the self-report of the schizophrenic
has implications both for diagnosis and treatment. In terms of diag-
nosis, if one is interested in the functioning of the schizophrenic's
social network, it would appear important to interview the family and
key members of the social network, rather than depend solely upon the
schizophrenic's perspective of the important social relationships.
Furthermore, as one is corroborating the self-report of the schizo-
phrenic and gathering information about the social network, one should
fully identify all those available for support. These persons are
potentially important in the maintenance of the schizophrenic. In
planning treatment, specifically a family or social networking
approach, the lack of corroboration suggests two particular tech-
niques: the identification and the correction of misperception within
important social relationships. As both techniques are already gen-
erally recognized and accepted in the field of family treatment and
social networking, this finding tends to substantiate the need for

their use.

157

The absence of significant differences in perception of social
networks, except in relation to size, leads to questioning of the
utility of diagnostic testing limited to conscious-level evaluation
(as discussed earlier). As conscious-level testing may not lead to
the detection of clinically significant diagnostic material, the use
of projective tests is recommended for diagnostic assessment.

The findings of smaller network size and, possibly, of closer
proximity of network members have diagnostic and treatment signifi-
cance relative to the withdrawal and isolation of the schizophrenic.
These social network characteristics may imply that the schizophrenic
has difficulty investing personal importance in those around him or
her. Diagnostically, this perception of the schizophrenic matches
with the professionally held opinion. In terms of treatment, this
difficulty should be kept in mind when considering the development of
the counseling relationship and treatment goals. In short, the
schizophrenic may have problems becoming involved in the counseling
relationship. Once the relationship has begun to develop, the coun-
selor in the outpatient setting might facilitate interaction with
other treatment staff by introducing the schizophrenic to those staff
who vfill be relating to him or her. In the day treatment setting,
the staff would be ready to offer assistance in facilitating contact
with others and also be supportive of the need for distance and
privacy.

The perceived dependent position of the schizophrenic is a diag—
nostic statement with treatment implications. It is expected that

this self-perception will surface in all types of treatment with

 

158

differing manifestations, depending upon the treatment modality.

For example, the dependency may become totally focused upon the
rehabilitation counselor, psychotherapist, or psychiatrist in a
one-to-one relationship. On the other hand, within group, family,

or social networking models, the dependency would unfold within the
context of these social relationships. Possible interventions would
include identification of the perceived dependency and the behavioral
ramifications in relationships, as well as attempts to facilitate
changes in relationships to ameliorate the dependent position. Change
assumes some openness. However, some schizophrenics may prefer the
dependent position, even though it is consciously denied, and thus be

resistant to change.

Syngpsis of Major Contributions

 

The major findings, conclusions, and recommendations of phase one
are, first, that the self-report of both samples was not adequately
corroborated. The lack of consensus regarding the objective vari-
ables (structure, content, and function) seriously questioned the
accuracy of self-report data and regarding the subjective variables,
the possibility of sufficiently shared emotional perceptions. Thus it
was concluded that self-report data are an insufficient and possibly
inaccurate, i.e., unreliable, source of information in social network
analysis. Accordingly, the assumption made by previous researchers
regarding its sufficiency, accuracy, and reliability appears lacking
in support. Second, the degree of corroboration or discrepancy in

self-report in both samples did not vary significantly. Thus the

 

 

159

nonpsychotic schizophrenic's perception of his/her social relation-
ships is substantially as reliable as the normal's perception of his
or her social relationships. Evidence was also found to support the
contention that the lack of reliability limited the power of statisti-
cal tests in the main comparison. The above sets of conclusions urge
a reexamination of current methodologies in social network analysis,
recommending the use of more objective conceptualizations and reliable
measures in the assessment of structure, content, and function.

The major findings, conclusions, and recommendations of phase
two are, first, that the social networks of normals and schizo-
phrenics vary in terms of only one variable, “perceived" network
size. The mean sizes of the social networks of normals and schizo-
phrenics were 36.23 and 12.65, respectively. The absence of signifi-
cance along the other 12 variables raised the two points: First,
when nonpsychotic, the schizophrenic may not consciously perceive
his or her relationships much differently from normals. Second, the
findings challenged results shared in previous studies, namely,
fewer multiplex relations, more support from network members as com-
pared to normals, and negative emotional perception of the network.
Thus it was recommended that future researchers consider the use of
projectives to determine (subjective) unconscious differences in nor-
mals and schizophrenics. Also as the diverse nature of the social
network may be responsible for lack of results on a conscious level,
the examination of subgroups such as family, relatives, and co-workers
to detect significant differences was recommended. Next, it was found

that normals perceived their relationships as reciprocal and

160

emotionally mutual, the latter suggesting that feelings and investment
in each other were shared. For the schizophrenic, emotions were also
found to be mutual, contradicting previous studies and predictions.
However, lack of reciprocity was perceived in relationships as pre-
dicted, placing the schizophrenic in a dependent position in his or
her self-perception. This finding was in line with the results of
previous studies.

In closing, and in line with the results, the nonpsychotic
schizophrenic with a history of psychiatric hospitalization and a
willingness to be involved in a scientific study, aged 18 to 40
years and living with family, appears as capable as the normal in
assessing social relationships. However, the perception of both is
insufficient, inaccurate, and unreliable, at least for research pur-
poses. In contrast to the normal, the schizophrenic consciously per-
ceives himself or herself socially surrounded by fewer important
persons upon which he or she is dependent. Nonetheless, in the per-
ception of the schizophrenic, the emotions are mutual in these
relationships. Overall, though, the schizophrenic does not appear

to perceive his or her relationships much differently from normals.

In Retrospect

 

Following the completion of a research project, it is rather
traditional to take a backward glance to determine if the study was
actually worth the investment. In retrospect, the input of energy,
thought, and time is balanced against the output of results to arrive
at the overall value. Certainly this study was much more demanding

than imagined in its conception and planning. Three reasons can be

161

identified. First, the development of the methodology was a difficult
task due to the conceptual infancy of social network analysis.
Markedly different conceptualizations of social networks presently
exist, so much time was spent determining which was most appropriate
for use in this study. Furthermore, the variables for social network
assessment are still in a formative stage. After a social network
conceptualization was chosen, the variables, derived from the Psycho-
social Network Inventory, Modified, were operationalized, a difficult
and complex task. The second and perhaps most difficult hurdle was
the identification and selection of normal and schizophrenic subjects.
With the schizophrenic subjects, several medical and ethical barriers
had to be handled, taking a great deal of time with meetings, consul-
tation, and waiting to obtain approval (discussed earlier). Once
these tasks were completed, one was still faced with identification
of potential subjects who met the criteria and then the unpredictable
response to the request for participation. Several rejections were
received. For the normal sample, the main problem turned out to be
the major time commitment requested from this group. Some subjects
spent five or more hours compiling a list of persons and assessing
their relationships. To provide an overall estimate/indicator of the
difficulties encountered, only 17 schizophrenic persons and 18 normal
persons, identified as meeting the criteria, agreed to participate in
the study through a search lasting a period of one year in the north
and northwest suburbs of Chicago. (Future researchers should cer—
tainly keep this difficulty in mind.) Third, the data collection for

both groups, but especially the schizophrenic, proved to be a tiring,

162

boring, and repetitive task. For example, in most examinations of

the social networks of the schizophrenic sample, each question was
read for each identified network member. Also several explanations
and clarifications were given. Interviews frequently lasted between
one and three hours. As normals did not require the examiner's vigi-
lance, the interview was not as taxing. On the more positive side,
however, the results of this study, in the opinion of this researcher,
have made significant contributions to the understanding of the schizo—
phrenic (and the normal) and to the evolving development of social
network analysis (see previous section). This researcher therefore
concludes that the contributions of this study were worth the efforts
needed to conceptualize, plan, and implement this research project,

making this an extremely valuable and productive research experience.

 

APPENDICES

163

APPENDIX A

INTERVIEW MATERIALS

164

165

STRUCTURED INTERVIEW FORMAT

Explanation of the Research Study (1 or 2)
Statement of Informed Consent (l or 2)
General Information Sheet

Listing of Persons Important to You

Family
Mother/Stepmother

Father/Stepfather
Sisters/Stepsisters/Half-sisters
Brothers/Stepbrothers/Half—brothers
Husband or Wife

Ex-spouses

Daughters/Stepdaughters
Sons/Stepsons

Relatives
Aunts & Uncles
Grandparents
Cousins
Nieces & Nephews
In-laws

Friends

Neighbors

Co-workers
Colleagues
Supervisors

 

Psychosocial Network Inventory, Modified
Assessment of Connections

Contact of Network Members

Names

Addresses
Phone Numbers

166

Directions: Structured Interview (1)
Phase 1: Explanation of the research study

Now as we begin, I would like to read with you the explanation of the
study. (Explanation of the Research Study [1] is placed in front of
the subject/s: Group testing is permitted for normal subjects only.)
Please ask your questions as they develop. (The form is read.)

00 you feel that you understand the explanation of the study?
(Response of the subject/s is explored if necessary.) Would you like
to see the letter that will be sent to identified network members?
(If the subject/s would like to see the letter to clarify the nature
of the contact, the letter will be shown and explained.)

Phase 2: Statement of informed consent

 

Before the study can begin, I need to obtain your informed consent to
participate in this study. This means that you can only agree to be in
the study after you fully know how the study affects you and your
rights as a subject. (Statement of Informed Consent [l] or [2] is
placed in front of the subject/s, the selection of either being deter-
mined by researcher intent and tentative subject consent: Form [1]

is for use with subjects who will consider giving their approval for
corroboration and from whom it is desired by the researcher, and form
[2] is for use when corroboration is neither approved nor desired.)

I have placed in front of you the Statement of Informed Consent. I
would like to read this with you. Again, please ask your questions
immediately. (Top portion of form [1] or form [2] in entirety is
read.

If you feel that you understand the explanation of the study and your
rights as a subject, and you wish to participate in this study, please
sign and date the form. Make a check mark by your name if you wish to
receive further information about the study after it is completed.

Now for the option I mentioned: I would like to request that you per-
mit me to contact members of your network to further explore their
relationship with you. Let's read together the rest of the consent
form. (Lower portion of form [1] is read.) If you wish to allow me
to contact network members as I have indicated, please sign and date
the lower portion of the consent form.

Now we are ready to begin with the study. Four basic steps are
involved in collecting the data in this study: (a) completing the
General Information Sheet, (b) listing persons, (c) assessing rela—
tionships, and (d) indicating the relationships among the persons
named.

167

Phase 3: General information sheet

 

First, we will complete the General Information Sheet. (The inter—
viewer will attempt to encourage the subject/s to complete the form
independently; however, if this procedure is not fruitful, the inter-
viewer will ask questions and fill in the responses on the form,
using the form as a prompter.)

Phase 4: Listing of persons important to you

 

I would like to develop a list of all persons who are important in
your life at this moment, whether you like them or not. Use your own
definition of who is important. I have placed in front of you a list
of persons who might be potentially important to you. (Listing of
Persons Important to You is placed in front of the subject/s.)

Please tell me the first name and initial of the last name of each
person. Also I need to know their sex, how long you have known them,
and their relationship with you, that is, mother, son, husband or

wife, co-worker, etc. Last, please tell me if anyone named is deceased
so that their name may be circled. (This information will be filled
out on the answer sheet: The names will be written in the order given
and numbered consecutively with the answers to follow placed in the
appropriate column. Due to varying time commitments or time tables,
some of the subjects may exercise the option to complete the list later
with the assistance of the form; however, a partial list of a minimum
of six names must be completed in the interviewer's presence to assure
a thorough understanding of the task and accuracy.)

Phase 5: Completion of the Psychosocial Network Inventory, Modified

Now that we have developed a list--or, at least, a partial list of
six--of those important to you, we can assess the relationships you
have with them by using the Psychosocial Network Inventory, Modified.
(Page 1 is placed in front of the subject/s.) Let's begin with the
first three questions. I will read these with you and show you where
to mark your answers.

Now here are the rest of the questions. (Page two of the inventory is
placed in front of the subject/s, along with the Response Key [1]:

The top of the second page is read--notation about the response key--
and it is explained that scales on the response key represent a
response continuum. The scales are numbered one through nine to per-
mit answers between the five modal points, to facilitate unassisted
completion once the directions are understood, and to correspond to
the answer code. Once this has been explained, items 4 through 15 are
asked of the subject/s.) As you see, we will take each person, one at
a time, and answer these questions. (After the form has been completed
for a minimum of two network members, the subject/s will be permitted
to fill out the rest of the inventory without the assistance of the
inverviewer.)

168

Phase 6: Assessment of connections

(This phase begins once all network members have been assessed.) Now
that the assessment is completed, we can continue to the brief final
step. I would like you to tell me which people that you have named
are connected by an ongoing relationship, that is, which persons both
know each other and have some kind of relationship with each other,
outside their relationship with you. First, what about ?
Who else has a relationship with this person? (Network members will
be taken in order to determine who knows whom. The numbers of those
members that know the selected person will be placed in the extreme
right hand column of the answer sheet.)

 

(If the corroboration has not been approved by or desired for this
subject, then this completes the examination. Skip to the concluding
remark.)

Phase 7: Contact of the network members

 

(If contact of corroborating network members has been permitted and
after the five corroborating network members have been identified by
the interviewer through the use of the form, Selected Network Members,
the subject/s will be told which were identified randomly. At this
point, the subject/s will be fully informed and, of course, free to
rescind approval for this phase of the research. Name, address, and
phone number of each identified member will be requested. If an iden-
tified member meets the criteria for exclusion, alternate choices will
be made randomly on the spot.)

This completes the interview. Thank you very much for your time and
effort.

169

Directions: Structured Interview (2)

Phase 1: Eyplanation of the research study

Now as we begin, I would like to read with you the explanation of the
study. (Explanation of the Research Study [2] is placed in front of
the subject. All schizophrenic subjects must be tested on an indi-
vidual basis and in private.) Please ask your questions as they
develop. (The form is read.)

00 you feel that you understand the explanation of the study?
(Response of the subject is explored if necessary.) Would you like to
see the letter that will be sent to identified network members? (If
the subject would like to see the letter to clarify the nature of the
contact, the letter will be shown and explained.)

Phase 2: Statement of informed consent

 

Before the study can begin, I need to obtain your informed consent to
participate in this study. This means that you can only agree to be
in the study after you fully know how the study affects you and your
rights as a subject. (Statement of Informed Consent [l] or [2] is
placed in front of the subject, the selection of either being deter-
mined by researcher intent and tentative subject consent: Form [1] is
for use with subjects who will consider giving their approval for cor-
roboration and from whom it is desired by the researcher, and form [2]
is for use when corroboration is neither approved nor desired.) 1 have
placed in front of you the Statement of Informed Consent. I would
like to read this with you. Again, please ask your questions immedi-
ately. (Top portion of form [1] or form [2] in entirety is read.)

If you feel that you understand the explanation of the study and your
rights as a subject, and you wish to participate in this study, please
sign and date the form. Make a check mark by your name, if you wish to
receive further information about the study after it is completed.

Now for the option I mentioned: I would like to request that you per-
mit me to contact members of your network to further explore their
relationship with you. Let's read together the rest of the consent
form. (Lower portion of form [1] is read.) If you wish to allow me to
contact network members as I have indicated, please sign and date the
lower portion of the consent form.

Now we are ready to begin with the study. Four basic steps are involved
in collecting the data in this study: (a) completing the General Infor-
mation Sheet, (b) listing persons, (c) assessing relationships, and

(d) indicating the relationships among the persons named. You need only
respond to the questions I will be asking as I will record your
responses.

170

Phase 3: General information sheet

 

First, we will complete the General Information Sheet. (The inter-
viewer will ask the questions, using the form as a prompter.)

Phase 4: Listing of persons important to you

 

I would like to develop a list of all persons who are important in your
life at this moment, whether you like them or not. Use your own defi-
nition of who is important. I have placed in front of you a list of
persons who might be potentially important to you. (Listing of Persons
Important to You is placed in front of the subject.)

Please tell me the first name and initial of the last name of each
person. Also I need to know their sex, how long you have known them,
and their relationship with you, that is, mother, son, husband or wife,
co-worker, etc. Last, please tell me if anyone is deceased so that
their name may be circled. (This information will be filled out on the
answer sheet: The names will be written in the order given and numbered
consecutively with the answers to follow placed in the appropriate col-
umn.

Phase 5: Completion of the Psychosocial Network Inventory, Modified

 

Now that we have developed a list of those important to you, we can
assess the relationships you have with them by using the Psychosocial
Network Inventory, Modified. (Page 1 of the form is placed in front of
the subject.) Let's begin with the first three questions. I will read
these with you. (Answers are obtained for the first three questions.)

Now here are the rest of the questions. (Page two of the inventory is
placed in front of the subject, along with the Response Key [2]: The
top of the second page is read--notation about the response key--and

it is explained that scales on the response key represent a response
continuum.) As these scales represent a continuum of possible answers,
you may choose an answer anywhere along the scale, as well as the five
major points marked on the response key pointed out by the interviewer.
(Once this has been explained, items 4 through 15 are asked of the sub-
ject.) As you see, we will take each person, one at a time, and answer
these questions. (Questions 1 through 15 are asked on each network
member in order; however, it is permissible to ask questions 1, 2, and
3 on all network members, and then the other questions in situations
where it seems appropriate.)

Phase 6: Assessment of connections

 

(This phase begins once all network members have been assessed.) Now
that the assessment is completed, we can continue to the brief final
step. I would like you to tell me which people that you have named are
connected by an ongoing relationship; that is, which persons both know
each other and have some kind of relationship with each other, outside

171

their relationship with you. First, what about ? Who
else has a relationship with this person? (Network members will be
taken in order to determine who knows whom. The numbers of those
members that know the selected person will be placed in the extreme
right hand column of the answer sheet.)

(If the corroboration has not been approved by or desired for this
subject3 then this completes the examination. Skip to the concluding
remark.

Phase 7: Contact of the network members

(If contact of corroborating network members has been permitted and
after the five corroborating network members have been identified by

the interviewer through the use of the form, Selected Network Members,
the subject will be told which were identified randomly. At this point,
the subject will be fully informed and, of course, free to rescind
approval for this phase of the research. Name, address, and phone
number of each identified network member will be requested. If an
identified member meets the criteria for exclusion, alternate choices
will be made randomly on the spot.)

This completes the interview. Thank you very much for your time and
effort.

.4-— “— .‘ -

172

 

Code #

MICHIGAN STATE UNIVERSITY

COLLEGE OF EDUCATION ILAS'I l\\sl\\. . sin Hl\.\\ - «an

DEPARTMEVT OF COUNSELING AND amt/(nova PSYCHOLOGY

Release of Information
1. I authorize to release
(Facility)
2 O
(Specify nature of information to be disclosed)

3. Information about to

4.

(Name) (Where information to be sent) Address)

5. For the specific purpose of

6. I understand that I have the right to inspect and copy the informa-
tion to be disclosed.

7. I understand that I may revoke this authorization at any time
except to the extent that action has been taken on this authoriza-
tion. I further understand that this authorization shall expire
without my express revocation on:

, l9__.

8. I further understand that the agency which receives this informa—
tion will not disclose this information without further written
consent.

9. 10.

(Signature) (Relationship)
11. 12.
(Date) (Witness)

 

 

 

 

 

 

 

 

 

 

 

M511: an Affirmative Arno- ‘Equal Opponent; lumumn

173

Explanation of the Research Study (1)

The purpose of the research study in which you are being asked to
participate is to examine one's social network, that is, the social
relationships one has with family, relatives, friends, co-workers,
and others.

You will be asked to identify persons important to you and to respond
to a series of questions about your relationship with these persons.
The time commitment varies between an hour and an hour and a half. A
possible additional request may apply to you: With your permission,
some persons in your network will be contacted in order to further
examine their relationship with you. They will only be notified of
your participation in this research study: All of your answers, as
well as their answers, will remain in strict confidence.

174

Explanation of the Research Study (2)

The research director/connfittee of this facility selected you as a
potential participant in an approved research study. The reasons

for identifying you specifically include your age, diagnosis, and

residential status--your residing with your family.

The purpose of the research study in which you are being asked to
participate is to examine one's social network, that is, the social
relationships one has with family, relatives, friends, co-workers,
and others.

You will be asked to identify persons important to you and to respond
to a series of questions about your relationship with these persons.
The time commitment varies a great deal. A possible additional request
may apply to you: With your permission, some persons in your network
will be contacted in order to further examine their relationship with
you. They will only be notified of your participation in this research
study: All of your answers, as well as their answers, will remain in
strict confidence.

 

“.1

 

175

Code #

MICHIGAN STATE UNIVERSITY

 

COLLEGE OF EDUCAUON EAST L.\.\Sl\(; ' MICHMAN ' 4M.“
DEPARTMENT OF COUNSELING AND EDUCATIONAL PSYCHOLOGY

2120 Hassel Road, #309
Hoffman Estates, Illinois 60195
(Local address)

“"1

 

 

Dear ,
You were contacted, because identified
you as a/an while participating in a research

 

study and permitted us to contact you. The purpose of this study
is to examine one's "social network", that is, the social relation-
ships one has with family, relatives, friends, co-workers, and
others.

To clarify our data, I would like to ask you a series of questions
briefly about your relationship with
Your answers to these questions and your participation in the study
will remain confidential.

 

Please carefully read the enclosed Statement of Informed Consent.

If you agree to participate in this research endeavor, please sign
and date the form, and return it in the enclosed envelope. If you
have any questions, please leave a message at this number, 432-3102
in care of Kenneth Carrico, and I will return your call. If you
wish the results of the study and/or an additional explanation after
the study is completed, please make a note on the consent form.

The Response Key has been enclosed for your convenience in responding
to the questions I will be asking.

Thank you very much for your time and consideration.

Sincerely yours,

Kenneth L. Carrico
Doctoral candidate
Michigan State University

MSL'u all Aflumuuv Arturo £9541 Oppnrmuun Imntutmn

176

Code #

MICHIGAN STATE UNIVERSITY

 

COLLEGE OF EDUCATION EAST LA\.\|\\. - \IlLHIk-\\ ' unit
DEPARTMENT OF COUNSELING AND mL’CATIONAL PSYCHOLOGY

Statement of Informed Consent (l)

I have freely consented to take part in a scientific study being
conducted by Kenneth L. Carrico, under the supervision of Norman R.
Stewart, Ph.D., Professor, Michigan State University.

The study has been explained to me and I understand the explanation
that has been given and what my participation will involve.

I understand that I am free to discontinue my participation in the
study at any time without penalty.

I understand that the results of the study will be treated in strict
confidence and that I will remain anonymous. Within these restrict-
ions, results of the study will be made available to me at my request.

I understand that my participation in the study does not guarantee any
beneficial results to me.

I understand that, at my request, I can receive an additional explana-
tion of the study after my participation is completed.

Name .

 

Date .

 

I have freely consented to the additional request that persons in my
social network be contacted by the researcher, the purpose being
further examination of the identified relationships. Only these
persons will be informed of my participation in the research study;
in all other ways, my anonymity will be preserved.

I understand that all of my answers, as well as their answers, will
remain in strict confidence.

Name

 

Date .

 

MIL u an Aflumam-r Anton haul ()ftpu'fullh [nun-mm.

 

 

 

177

Code #

MICHIGAN STATE UNIVERSITY

 

COLLEGE OF EDUCATION EAST LR\SI\(. - “It H|(.§\ - rut-3»)
DEPARTMENT OF COUNSELING AND EDUCATIONAL PSYCHOLOGY

Statement of Informed Consent (2)
I have freely consented to take part in a scientific study being
conducted by Kenneth L. Carrico, under the supervision of Norman R.

Stewart, Ph.D., Professor, Michigan State University.

The study has been explained to me and I understand the explanation
that has been given and what my participation will involve.

I understand that I am free to discontinue my participation in the
study at any time without penalty.

I understand that the results of the study will be treated in strict
confidence and that I will remain anonymous. Within these restrict-
ions, results of the study will be made available to me at my request.

I understand that my participation in the study does not guarantee any
beneficial results to me.

I understand that, at my request, I can receive an additional explana-
tion of the study after my participation is completed.

Name .

Date -

 

MS! '1: an Alumni-u Arum- Itqui ()ppnflumm lmnutum

178

Code #

MICHIGAN STATE UNIVERSITY

 

COLLEGE OF ELL-CANON EAST Hum. - mt Humx - tam
DEPARTMENT OF COLNSELLNG AND EDUCAUONAL PSYCHOLOGY

General Information Sheet (PNIM)

Age . Sex: Male . Female .

 

Education: Last year of school completed or final degree .
Employment: Your occupation .

Work status: Employed full-time . Employed part—time .
Unemployed . If unemployed, how long ?
Never employed Student . Number of jobs
held, past 5 years .

 

Gross yearly income: Personal income . Family income

Ethnic group: White . Black . Mexican/American
Oriental . American Indian .
East Indian . Other .

 

Marital status: Single . Married . Separated .
Divorced . Widowed . Living together

 

 

Religious preference: Protestant . Catholic . Jewish .
Islam . Other . None .

 

Residence: Living alone . Living with family and/or
relatives . Living with unrelated others
In mental health facility or residential setting .

 

Number of moves in past 5 years .
Number of years at present address .

 

Major life events: Check any of the following if they have
occurred during the last year to you and/or
your family:

 

 

 

 

 

 

 

Birth . School or job change .

Death . Serious accident .

Divorce or separation . Serious illness or disability .
Family or marital conflict . Trouble with Law .

Family move . Other

Financial problems .

Marriage or remarriage . None .

Group membership: List names of all formal or informal community
associations to which you may belong, such as
church, fraternal club, service club, self-help
group, social or recreational groups, etc.,
regardless of your level of participation.

 

 

MSL n on Aflnmuu Adm ‘qu Opportumn llllllullfll

179

Gaza: Human g

Hovoncmu

 

18D

Listing of Persons Important to You

Family

Mother/Stepmother

Father/Stepfather
Sisters/Stepsisters/Half—sisters
Brothers/Stepbrothers/Half-brothers
Husband or Wife

Ex-spouses

Daughters/Stepdaughters
Sons/Stepsons

Relatives
Aunts & Uncles
Grandparents
Cousins

Nieces & Nephews
In-laws

Friends

Neighbors

Co-workers

 

Colleagues
Supervisors

181

PSYCHOSOCIAL NETWORK INVENTORY, MODIFIED

 

How often do you usually have CONTACT with this person, whether
face-to-face, or by phone or letter?

(DQOO'QJ

Daily.

At least once a week.

At least once a month.

At least once every six months.
At least once a year.

How CLOSE does this person live to you?

(DQOU'QJ

Lives in the same household.

Within the same general neighborhood or locale.
Within about a 30-minute drive from you.

Within about a 2-hour drive from you.

Beyond a 2-hour drive from you.

Which types of ACTIVITIES do you and this person share as
described in the following list?

a.

b.

 

FAMILY activities such as meal-time, holidays, vacations, or

reunions.
EMPLOYMENT activities such as working with co-workers or

 

supervisors.

ROMANTIC activities such as dating, dancing, or going out

for dinner.
CONVERSATIONAL activities such as intimate, personal sharing

 

or philosophical discussions.

SOCIAL activities such as parties, banquets, or visiting

friends or neighbors.
RECREATIONAL activities such as playing cards or participation

 

in or attendance at sports events.

FRATERNAL activities such as participation in clubs or other

organizations.

RELIGIOUS activities such as attendance at church, synagogue,
or temple.

POLITICAL activities such as rallies or discussions about
politics.

VOLUNTEER activities such as service to the community, giving
blood, or hospital work.

 

 

182

(Answers for the following items are found on the scales of the
Response Key.)

4.

10.

11.

12.

13.

14.

15.

How much do you help this person by providing GUIDANCE when needed,

 

such as giving advice on a task or decision, or giving feedback on
an action taken? (Response Key: A)

How much does this person help you by expressing AFFECTION when
you negd it, such as a hug, kiss, or pat on the back? 7(Response
Key: A

 

How much do you help this person bnyOING THINGS for them when
needed, such as helping with household tasks, providing personal
or family care, assisting on the job, or even lending money?
(Response Key: A)

 

How much do_you help this person by expressing AFFECTION when

 

needed, such as a hug, kiss, or pat on the back? (Response Key: A)

What TYPE of feelings does this person have toward you, regardless
of their strength? (Response Key: C)

 

What is the STRENGTH of the feelings and thoughts this person has

 

toward_you, regardless of their type? (Response Key: 8)

 

What is the STRENGTH of the feelings and thoughts you have toward
this person, regardless of their type? (Response Key: 8)

 

How much does this person help you by providing GUIDANCE when you
need it, such as giving advice on a task or decision, or giving
feedback on an action taken? (Response Key: A)

How much do you help this person by giving EMOTIONAL SUPPORT when
needed, such as giving praise, being a good listener, or providing
them with encouragement? (Response Key: A)

How much does this person help you by DOING THINGS for you when
you need it, such as helping with household tasks, providing
personal or family care, assisting on the job, or even lending
money? (Response Key: A)

What TYPE of feelings and thoughts doyyou have toward this person,
regardless of their strength? (Response Key: C)

How much does this person help you by giving EMOTIONAL SUPPORT
when you need it, such as praising you, being a good listener,
or providing you with encouragement? (Response Key: A)

 

 

183

 

 

 

m w w e m e m N w
e>wuwmee ww< e>wpwmee zwpmez eexwe xwwezee peee< e>wuemec xwpmez e>wuemec __<
+ + + + +
.m
m m w e m e m N w
meewum >Le> meetum eueceeez ewwz gee:
+ + + + +
.m
m w w o m a m N w
xwuceeeeww >Lm> :eewo meewmeeee eEem co awewem __e we pez
+ + + + +
.m

 

Awe wee emzeamme

184

 

 

 

e>wpwmee ww< e>wpwmee xwpmez eexwe awweeee useeq e>wpemec xwumez e>wpemme ww<
+ + + + +

.m

meeeem xwe> meeeum meeteeez ewwz xeez

+ + + + +

.m

xwuceeeeew xte> :eewo mcewmeeee eeem co xmeem wwe we pez
+ + + + +

.m

 

Ame wee mmzeamme

185

SELECTED NETWORK MEMBERS

 

 

Size of Identified Network Numbers of Network Members for Contact

1 ----------------------------------- 1

2 ----------------------------------- l, 2

3 ----------------------------------- l, 2, 3

4 ----------------------------------- l, 2, 3, 4

5 ----------------------------------- l, 2, 3, 4, 5

6 ----------------------------------- 2, 3, 4, 5, 6

7 ----------------------------------- l, 2, 3, 4, 5

8 ----------------------------------- 2, 3, 4, 5, 7

9 ----------------------------------- 2, 3, 7, 8, 9

10 ----------------------------------- 2. 4, 5, 8, 9

1] ----------------------------------- 4, 5, 8, 9, 10

12 ----------------------------------- l, 2, 6, 7, 10

13 ----------------------------------- 2, 7, 8, 9, ll

14 ----------------------------------- l, 5, ll, 12, 13
15 ----------------------------------- 3, 4, 9, 12, 15

16 ................................... 5, 6, 9, ll, 13

17 ----------------------------------- l, 8, 11, l4, 17
18 ----------------------------------- l, 8, 13, 14, 16
19 ................................... 2, 7, 9, ll, 13
20 ................................... 5, 6, 9, 13, 19
21 ----------------------------------- 2, 3, 4, 7, 17
22 ................................... 6, 10, ll, 15, 16
23 ................................... 4, 6, 14, 15, 21
24 ----------------------------------- 6, 10, ll, 15, 16
25 ................................... 11, 12, 14, 15, 17
25 ................................... 5, l6, 17, 20, 21
27 ................................... l, 2, 8, 19, 20
28 ................................... l, 17, 18, 20, 22
29 ----------------------------------- 7, 8, 9, ll, 22
3o ................................... 5, 6, 19, 23, 26
31 ................................... ll, 14, 15, 21, 24
32 ................................... 4, 8, 9, 15, 24
33 ................................... 1, 7, 16, 31, 32
34 ................................... 5, 16, 20, 21, 32
35 ----------------------------------- 5, 6, 23, 26, 34
36 ----------------------------------- 12, 14, 15, 24, 33
37 ................................... 10, 15, 16, 30, 37
38 ................................... 14, 15, 24, 36, 38
3g ................................... 2, 8, 22, 36, 37
4O ----------------------------------- 4, 18, 25, 32, 34
41 ----------------------------------- 1, l6, 18, 25, 30
42 ................................... 9, 12, 15, 20, 23
43 ................................... 13, 24, 35, 38, 4O

44 ................................... 4, 15, 17, 22, 38

 

 

 

186

 

 

Size of Identified Network Numbers of Network Members for Contact
45 ---------------------------------- 3, 10, 15, 20, 24

46 ---------------------------------- 2, 3, 9, 15, 32

47 ---------------------------------- 2, 7, 13, 15, 31

48 ---------------------------------- 4, 6, 21, 42, 44

49 ---------------------------------- 3, 9, 15, 19, 22

50 .................................. 4, 9, 24, 46, 48

 

APPENDIX B

OPERATIONALIZATION OF VARIABLES

187

 

188

 

PHASE I: CORROBORATION

STRUCTURE-RELATED VARIABLES

 

l. VARIABLE: Size of network cluster

DEFINITION: The size of the network cluster is that number of

 

persons that are known jointly by the focal person (the subject) and
by the network member/corroborator who are considered important by

the focal person.

 

INDEX: XCSy and XCCy

.KEY: szy = The size of the network cluster as determined by the
subject for a certain network member/corroborator. In the notation
used in this phase, the prefixes and suffixes are used as coordinates,
not as mathematical or statistical symbols. The "x" in the prefix
designates the figure is one estimated by the focal person (the sub-
ject). The "y” designates the number given to the corroborator as
written on the answer sheet. For instance, in an example we could
write 10SCs3’ meaning that for the schizophrenic subject, code #lOS,
we have the size of the network cluster estimated by the subject rela-
tive to the network member/corroborator, #3. If the subject was a nor-
mal with the same code number, we would have lONCs3'

xCCy = The size of the network cluster as determined by the
corroborator relative to the focal person (the subject). The "x" in
the prefix designates the code number given to the subject (the focal
person, not the corroborator). The "c" in the suffix designates that

the figure is one estimated by the corroborator. The "y" designates

the number given to the corroborator as written on the answer sheet.

189

In our example, we would write lOSCc3’ meaning that for the schizo-
phrenic subject, code #1OS, we have the size of the network cluster
estimated by the network member/corroborator whose number is #3.

RELATIONSHIP TO INVENTORY: NA

 

RELATIONSHIP TO ANSWER SHEET: Five corroborators were identified

 

on the subject's answer sheet; the responses of each corroborator were
written on the ”corroborator" answer sheet, designated by the number
given to them by the focal person. Under the ”number" column of the
subject's answer sheet (extreme left hand side), match the number of
the identified corroborator/s with the number on the corroborator's
answer sheet, same column. Follow the row to the extreme right side--
the column "SUM." The figure on the subject's answer sheet would be

C ; the figure on the corroborator's answer sheet would be xC

x sy CY.

2. VARIABLE: Distance
DEFINITION: Distance refers to the distance between the focal

 

person and a network member/corroborator in terms of location or time
of travel.

INDEX: D and xDC

—— xsy y

KEY: D and D = The estimated distance between the focal
—-—- x sy x cy
person and the corroborator; the former, subject's estimate and the
latter, corroborator's estimate. The prefix and suffix designate the
subject's code number ("x"), the source of the estimate ("5" or "c"),

and the corroborator's number ("y").

RELATIONSHIP TO INVENTORY: The estimated distance is the answer

 

to item #2.

 

190

RELATIONSHIP TO ANSWER SHEET: To locate the response on the sub-
ject's answer sheet relative to the corroborator, find the corrobora-
tor's number on the extreme left hand side on the subject's answer
sheet and follow the row across to item #2. On the corroborator's
answer sheet, the matching response is also under item #2 on the same

row as the corroborator's identifying number.

3. VARIABLE: Frequency of contact

DEFINITION: Frequency of contact refers to the relative frequency

 

of contact between the focal person and the network member/corroborator.

INDEX: FSy and XFC

y

KEY: xFSy and chy = The estimated frequency of contact between
the focal person (the subject) and the corroborator; the former, the
subject's estimate and the latter, the corroborator's estimate.

RELATIONSHIP TO INVENTORY: The estimated frequency of contact is

 

the answer to item #1.

RELATIONSHIP TO ANSWER SHEET: The responses to item #1 for both

 

subject and corroborators are found as indicated above.

CONTENT—RELATED VARIABLES

 

4. VARIABLE: Content

DEFINITION: Content refers to the number of activities shared

 

between the focal person (the subject) and the network member/
corroborator.

INDEX: XRS and XRC

Y

 

191

KEX: sty and chy = The estimated number of activities shared
between the subject and the corroborator; the former, the subject's
estimate and the latter, the corroborator's estimate.

RELATIONSHIP TO INVENTORY: The number of activities is the sum of

 

all of the activities checked under item #3, section K.

RELATIONSHIP TO ANSWER SHEET: The answers are found under the

 

"SUMMARY" section of item #3 on both answer sheets: 3-K.

FUNCTION-RELATED VARIABLES

 

5. VARIABLE: Functional indegree

DEFINITION: Functional indegree indicates the degree of support

 

that the focal person (the subject) is receiving from the network
member/corroborator.

INDEX: Q

and Q
x sy x cy

KEX; XQSy and xQCy = The estimated degree of support that the
subject is receiving from the corroborator; the former, the subject's
estimate and the latter, the corroborator's estimate.

RELATIONSHIP TO INVENTORY: Items 5, ll, 13, and 15 assess the

 

degree of support that the focal person is receiving. Items 4, 6, 7,

and 12 assess the functional indegree from the corroborator' s perspective.

 

RELATIONSHIP TO ANSWER SHEET: To obtain szy’ add the responses
from items 5, 11, 13, and 15 in the row with the same number of the
corroborator. To obtain chy’ add the responses from items 4, 6, 7,
and 12 from the corroborator's answer sheet on the same row as the

matching identifying number.

192

6. VARIABLE: Functional outdegree

DEFINITION: Functional outdegree indicates the degree of support
that the focal person (the subject) is providing the network member/
corroborator.

INDEX: XPSy and XPC

Y
5E1: xPSy and xpcy = The estimated degree of support that the

focal person is providing to the corroborator; the former, the sub-
ject's estimate and the latter, the corroborator's estimate.

RELATIONSHIP TO INVENTORY: Items 4, 6, 7, and 12 assess functional

 

outdegree from the subject's perspective. Items 5, ll, 13, and 15
assess functional outdegree from the corroborator's perspective.

RELATIONSHIP TO ANSWER SHEET: To obtain xPsy’ add the responses

 

from items 4, 6, 7, and 12 on the subject's answer sheet in the row
with the same number of the corroborator. To obtain xPcy’ add the
responses from items 5, ll, 13, and 15 from the answer sheet of the

corroborator, matching the identifying number.

EMOTION-RELATED VARIABLES

 

7. VARIABLE: Affective indegree
DEFINITION: Affective indegree indicates the type of feelings
held by a network member/corroborator for the focal person.

INDEX: xVSy and chy

KEX; xVSy and xVCy = The estimated type of feelings held by the
corroborator for the subject; the former, the subject's estimate and

the latter, the corroborator's estimate.

 

 

193

RELATIONSHIP TO INVENTORY: Item #8 assesses affective indegree

 

from the perspective of the subject. Item #14 assesses affective
indegree from the perspective of the corroborator.

RELATIONSHIP TO ANSWER SHEET: To obtain xvsy’ locate the score on

item #8 for the corroborator on the subject's answer sheet. To obtain

 

 

chy, locate the score of item #14 on the corroborator's answer sheet.

 

8. VARIABLE: Affective outdegree
DEFINITION: Affective outdegree indicates the type of feelings

 

experienced by the focal person (the subject) for the network member/
corroborator.

INDEX: Y and XYC

—— xsy y

KEY; XYSy and XYCy = The estimated type of feelings held by the
subject for the corroborator; the former, the subject's estimate and

the latter, the corroborator's estimate.

RELATIONSHIP TO INVENTORY: Item #14 assesses affective outdegree

 

from the perspective of the subject. Item #8 assesses affective out-
degree from the perspective of the corroborator.

RELATIONSHIP TO ANSWER SHEET: To obtain sty’ locate the score

 

on item #14 for the identified corroborator on the subject's answer

sheet. To obtain chy’ locate the score on item #8 on the corrobo-

rator's answer sheet.

 

194

9. VARIABLE: Investment indegree

DEFINITION: Investment indegree reflects the strength of feelings

 

(emotional investment) experienced by the network member/corroborator
for the focal person (the subject).

INDEX: szy and XGCy

KEX: xGSy and chy = The estimated strength of feelings, i.e.,
emotional investment, held in the subject by the corroborator; the
former, the subject's estimate and the latter, the corroborator's

estimate.

RELATIONSHIP TO INVENTORY: Item #9 assesses investment indegree

 

from the perspective of the subject. Item #10 assesses investment
indegree from the perspective of the corroborator.

RELATIONSHIP TO ANSWER SHEET: To obtain szy’ locate the score on

 

item #9 for the corroborator on the subject's answer sheet. To obtain

chy’ locate the score on item #10 on the corroborator's answer sheet.

10. VARIABLE: Investment outdegree

DEFINITION: Investment outdegree reflects the strength of feelings

 

(emotional investment) experienced by the focal person (the subject)
for the network member/corroborator.

INDEX: H and xHc

-—--— XS)’ y

yey; H and xHc = The estimated degree of emotional investment

X SY y
in the corroborator by the subject; the former, the subject's estimate

and the latter, the corroborator's estimate.

 

 

195

RELATIONSHIP TO INVENTORY: Item #10 assesses investment outdegree

 

from the perspective of the subject. Item #9 assesses investment
outdegree from the perspective of the corroborator.

RELATIONSHIP TO ANSWER SHEET: To obtain sty’ locate the score

 

on item #10 for the corroborator on the subject's answer sheet. To
obtain chy’ locate the score on item #9 on the corroborator's answer

sheet.

 

 

196

PHASE II: MAIN ANALYSIS

 

STRUCTURE-RELATED VARIABLES

 

l. VARIABLE: Size

DEFINITION: Size refers to the actual number of persons identi-

 

fied as important in the social network.

INDEX: n

KEY: n = an absolute value; an index is not necessary.

RELATIONSHIP TO INVENTORY: NA

 

RELATIONSHIP TO ANSWER SHEET: "n" is the total number of persons

 

identified on the left side of the answer sheet-~this number will not

include the focal person.

2. VARIABLE: Interconnectedness or adjacent density

DEFINITION: Adjacent density, the index of interconnectedness,

 

refers to the proportion of linkages (relationships) in the social
network to the total possible number of linkages in the network. The

proportion varies between .00 and 1.00.

, a 2a

KEY: a the actual number of linkages in the network.

N the network size (n) plus the focal person--n + 1.

RELATIONSHIP TO INVENTORY: NA
RELATIONSHIP TO ANSWER SHEET: The actual number of linkages in
the network, "a," is computed by first summing the sizes of the net-

work clusters (found under the SUM column of the answer sheet) across

197

all network members, designated xCS , dividing this total figure by

y
two, and last, adding network size. The formula is as follows:
n
:l
a 2 iil XCSy + n.

3. VARIABLE: Distance

DEFINITION: Distance refers to the average distance between the

 

focal person and a network member in terms of location or time of

travel.

INDEX:

d/n
1

”M:

.i
5E1; d = the degree of distance between the focal person and
network member, as assessed on the inventory.

RELATIONSHIP TO INVENTORY: "d” is assessed through item #2.

 

RELATIONSHIP TO ANSWER SHEET: Locate the #2 column on the answer

 

sheet; a = l, b = 2, c = 3, d = 4, and e = 5: the lower the value,
the closer to the focal person the member lives. The numerator of

the index would be the sum of the #2 column.

4. VARIABLE: Frequency of contact

DEFINITION: Frequency of contact refers to the average frequency

 

of contact between the focal person and a network member.

INDEX:

f/n
1

"M:

i
yey; f = the frequency of contact for an identified person in

the network.

198

RELATIONSHIP TO INVENTORY: "f" is assessed through item #1 in

 

the inventory.

RELATIONSHIP TO THE ANSWER SHEET: To find ”f," locate the co1umn

 

under #1 on the answer sheet; a = l, b = 2, c = 3, d = 4, and e = 5:
the lower the value, the more frequent the contact. The numerator of

the index would be the sum of the #1 column.

CONTENT—RELATED VARIABLES

 

5. VARIABLE: Proportion of uniplex relationships

DEFINITION: In the uniplex relationship, only one type of activity

 

is shared in the relationship. The index is a proportion.
INDEX: u/n

KEY: u = the number of uniplex relationships in the social net-
work of one subject.

RELATIONSHIP TO INVENTORY: ”u" is assessed by item #3.

 

RELATIONSHIP TO ANSWER SHEET: On item #3 the type and number of

 

activities shared in a relationship are noted. If only one activity
is checked for a relationship, then the identified relationship is

uniplex.

6. VARIABLE: Proportion of multiplex relationships

DEFINITION: In the multiplex relationship, more than one activity

 

is shared in the relationship. The index is a proportion.

INQEX; m/n

5E1; m = the number of multiplex relationships in the social
network of one subject.

RELATIONSHIP TO INVENTORY: "m" is assessed by item #3.

 

199

RELATIONSHIP TO ANSWER SHEET: If more than one type of activity

 

is checked for a relationship under #3 on the answer sheet, then the

relationship is multiplex.

7. VARIABLE: Relationship density

DEFINITION: Relationship density is an average of the content

 

areas, i.e., types of activity, between the focal person and the

social network.

INDEX:

r/n
1

11M:

1'
KEY: r = the number of content areas checked for one person in
the network.

RELATIONSHIP TO INVENTORY: (Also a part of item #3)

 

RELATIONSHIP TO ANSWER SHEET: To find "r," locate the k—column

 

under item #3, "SUMMARY"; each line contains the sum of the checked
content areas. The grand total of this column would be numerator of

the index.

FUNCTION-RELATED VARIABLES

 

8. VARIABLE: Functional indegree

DEFINITION: Functional indegree indicates the degree of support

 

that the focal person is receiving.

INDEX:
i

q/n
1

"MD

KEY; q = sum of support that the focal person is receiving from
one network member as assessed by four items (5, ll, 13, 15), each on

a scale from 1 to 9.

 

 

ZOO

RELATIONSHIP TO INVENTORY: See the table below:

 

 

Item #5 = Affective functioning, behavioral mode (physical
affection)

Item #11 = Instrumental functioning, verbal mode (guidance)

Item #13 = Instrumental functioning, behavioral mode (helping by
doing)

Item #15 = Affective functioning, verbal mode (verbal support)

Note: The higher the number, the greater the degree of functional

 

support.

RELATIONSHIP TO ANSWER SHEET: To obtain "q" for one network member,

 

add the responses from items 5, ll, 13, and 15. According to the
index, this would be done for every individual network member, the
grand total being divided by "n"; however, an easier method would be
to add the columns under 5, ll, 13, and 15 and divide the result grand

total by "n.”

9. VARIABLE: Functional outdegree

DEFINITION: Functional outdegree indicates the degree of support

 

that the focal person is providing others.

INDEX: p/n

1

"M:

i

KEY: p = sum of support that the focal person is providing one
other network member as assessed by four items (4, 6, 7, 12), each
on a scale from 1 to 9.

RELATIONSHIP TO INVENTORY: See the table below:

 

Item #4 Instrumental functioning, verbal mode (guidance)

Instrumental functioning, behavioral mode (helping
by doing)

Item #6

 

201

Item #7 = Affective functioning, behavioral mode (physical
affection)
Item #12 = Affective functioning, verbal mode (verbal support)

Note: The higher the number, the greater the degree of support of
the various kinds.

RELATIONSHIP TO ANSWER SHEET: To obtain "p" for one network

 

member, add the responses from items 4, 6, 7, and 12. According to

the index, this would be done for every network member, the grand

total being divided by "n"; however, an easier method would be to add
the columns under 4, 6, 7, and 12, and divide the result--grand total--

by n."

10. VARIABLE: Functional symmetry

DEFINITION: Functional symmetry is a function-related variable

 

that assesses the balance of support in relationships in the social
network. The balance can be equal, i.e., reciprocal, or unequal,

i.e., either dependent or supportive relative to the focal person.

EMOTION-RELATED VARIABLES

 

ll. VARIABLE: Affective indegree

DEFINITION: Affective indegree indicates the type of feelings

 

held by a network member as perceived by the focal person. The range

extends from all negative to all positive feelings.

INDEX: v/n

1

IIM:

i
KEY; v = the type of feelings of one network member for the

focal person (as perceived by the focal person) as assessed by item #8

202

on a scale from 1 to 9. The lower numbers (1-3) suggest a negative
feeling; moderate numbers (4-6), mixed feelings; higher numbers (7-9),
more positive feelings.

RELATIONSHIP TO INVENTORY: Item #8 assesses this variable.

 

RELATIONSHIP TO ANSWER SHEET: The singular answer to item #8 is

 

"v” for one network member. To obtain the numerator of the index for
one subject, total the column under item #8. The index is computed

when the amount is divided by "n."

12. VARIABLE: Affective outdegree

DEFINITION: Affective outdegree indicates the type of feelings

 

experienced by the focal person for a member of the social network.

The range extends from all negative feelings to all positive feelings.

INDEX:

y/n
1

III"):

i

.KEY: y = the type of feelings the focal person experiences for
one network member-—assessed by item #14 on a scale from 1 to 9. As
with affective indegree, the lower numbers (1-3) suggest a negative
feeling; the moderate numbers (4-6), mixed feelings; and the higher
numbers (7-9), a more positive feeling.

RELATIONSHIP TO INVENTORY: Item #14 assesses this variable.

 

RELATIONSHIP TO ANSWER SHEET: The singular answer to item #14 is

 

"y" for one network. To obtain the numerator of the index, total the
column under item #14. The index is computed when this amount is

divided by "n"--that is for one social network.

203

13. VARIABLE: Affective symmetry

DEFINITION: Affective symmetry exists when the type of emotion

 

between the focal person and the network member is identical or com-

parable.

l4. VARIABLE: Investment indegree

DEFINITION: Investment indegree reflects the strength of feelings

 

(emotional investment) experienced by the network member for the focal
person as perceived by the focal person. The range extends from a

weak to very strong investment of feeling (1 to 9).

1'1
INDEX: 2 g/n

_——__' i=1

KEY; g = the degree of investment in the focal person by a net-
work member as perceived by the focal person--assessed by item #9 on
a scale from 1 to 9. Lower scores suggest less involvement; higher
scores, greater investment.

RELATIONSHIP TO INVENTORY: Item #9 assesses this variable.

 

RELATIONSHIP TO ANSWER SHEET: The answer to item #9 is "g" for

 

one network member. To obtain the numerator of the index for one
subject, total the column under item #9. The index is computed when

this amount is divided by "n."

15. VARIABLE: Investment outdegree

DEFINITION: Investment outdegree reflects the strength of feeling

 

experienced by the focal person for a network member. The range is

identical to that of investment indegree.

204

MM:

INDEX:

h/n
l

i
KEY: h = the degree of investment in the network member by the

focal person-—assessed by item #10 on a scale from 1 to 9. Lower

scores suggest less involvement; higher scores, greater investment.
RELATIONSHIP TO INVENTORY: Item #10 assesses this variable.
RELATIONSHIP TO ANSWER SHEET: The answer to item #10 is "h" for

one network member. To obtain the numerator of the index for one

 

subject, total the column under #10. The index is computed when this

amount is divided by ”n."

16. VARIABLE: Investment symmetry
DEFINITION: Investment symmetry exists when the emotional invest-
ment experienced by the focal person and the network member is compar-

able; investment asymmetry, unequal.

 

 

APPENDIX C

PROPOSAL TO THE UNIVERSITY COMMITTEE ON RESEARCH
INVOLVING HUMAN SUBJECTS

205

206

Title of Proposed Research: An Examination of the Social Networks of
Normals and Schizophrenics

Type of Research: Doctoral dissertation

 

Presenting Doctoral Candidate: Kenneth L. Carrico
2120 Hassel Road, #309
Hoffman Estates, Illinois 60195
Phone: (312) 882-2156

 

 

University Department: Department of Counseling and Educational
Psychology, College of Education

 

Supervisor: Norman R. Stewart, Ph.D.
Professor
Phone: 353-6456

 

Date of Proposal Approval: January 25, 1980

 

 

207

TABLE OF CONTENTS

Abstract

Subject Populations

Risk/Benefit Analysis

Consent Procedures

Explanation of the Research Study (1)

Statement of Informed Consent (l)

Corroborating Member Letter Form

Statement of Informed Consent (2)

Explanation of the Research Study (2)

Release of Information

Testing Procedure

Psychosocial Network Inventory, Modified
General Information

(Test Form)

208

Abstract

The phenomenon of schizophrenia has long perplexed researchers.
Recent attempts to study its psychosocial aspects have utilized a new
method of study from the field of sociology, "social network analysis."
The investigator in the present study will attempt to replicate the
work of these recent studies by comparing the social networks of nor-
mals and schizophrenics; however, in the initial stage of the study, the
investigator will test an assumption methodologically basic to these
studies-~that the subjects' self-report is a sufficient and accurate
indicator of the functioning of the social network. If this assumption
is proven as valid, then the comparison--the focus of the second stage
of the study--will be carried out using the self-report data. If not,
then the comparison will be carried out as possible.

The social networks of normals and schi20phrenics will be com-
pared along four major variables pertaining to social relationships:
structure, content, function, and emotion. The structural variables
convey the basic morphological characteristics of the social network.
The content variables convey aspects of the content of the social
relationships, i.e., the types of activities. The function variables
reflect the transaction of support. And last, the emotion variables

assess the experiencing of affect.

Subject Populations

Sample of Normals

 

A group of individuals will be identified to represent a normal

population who meet the following criteria: (a) residence with

209

nuclear family or family of origin; (b) age ranging between 21 and
40 years; and (c) no personal or family history of a psychiatric
disorder for two generations. From this group, a sample of 20 will

be randomly selected for inclusion in this study.

Sample of Schizophrenics

 

A group of individuals will be identified who have a history of
hospitalization for a schizophrenic disorder, and who meet the fol-
lowing additional criteria: (a) residence with nuclear family or
family of origin; (b) age ranging between 21 and 40 years; (c) not
presently psychotic, meaning not out of touch with reality and capable
of communication; and (d) diagnosis of schizophrenia made within the

past 3 years.

Selection Rationale

 

Both groups will be selected for reasons of comparison and
contrast, for without this, one is unable to establish a meaningful
perspective.

Obtaining informed consent is not foreseen as a problem with the
sample of normals, but it may arise with the sample of schizophrenics.
As can be seen in the consent procedures, everything is being done to
fully inform the schizophrenic subject of his/her rights and alterna-
tives. When this information cannot be understood, then the subject
will not be included in the study as they have not met the third cri-
terion mentioned above. This decision will be a judgment made by the

researcher.

210

Risk/Benefit Analysis

 

An examination of the nature and design of the study suggests
the presence of potential social and psychological risks; physical,
legal, and economic risks are not posed.

Potential social risks are associated with the request that some
subjects relinquish their anonymity to provide access to their social
network for further study of the social relationships. The risk
appears to relate to possible misinterpretations by the contacted
network members of (a) the subject's motive in identifying them and
(b) the reason for the subject's participation in the study.

The social risk of misinterpretation appears to be largely
dependent upon the degree of vagueness inherent in the explanation
given to the contacted network member. To minimize this possibility,
every effort has been made to develop a clear, direct, precise, and
unambiguous explanation. However, this safeguard is one-sided in the
sense that it can only act to increase the clarity of the statement;
it cannot be assumed to control for idiosyncratic interpretations or
misuse of the disclosure of the subject's participation by the con-
tacted network member.

Potential psychological risks relate to the use and effect of
the Psychosocial Network Inventory, Modified (PNIM). The use of all
psychological tests is accompanied by potential risks, and the same is
true for the PNIM. The primary risk is the generation of psychologi-
cal conflicts, and resulting discomfort/dysfunction, previously con-
trolled by the subject's defense mechanisms. This reaction could

occur in either group, but it is more likely with the schizophrenics;

211

their coping abilities and defenses are much less controlled and
integrated, by definition. However, the type of test used in this
study is a highly structured, rather straightforward inventory tapping
primarily conscious levels, possibly the least threatening of all
psychological test forms; projective tests, on the other hand, tap
much deeper levels of the subject's psyche and would, therefore, be
much more likely to uncover hidden conflict and provoke an untoward
reaction.

The psychological risks have been lessened through the choice of
an objective and structured test (see above); however, if such risks
materialize, the schizophrenic-—who will, by the way, be tested in a
psychiatric setting--will be encouraged to consult with his/her pri-
mary therapist and/or sign a release so that the reaction can be
reported to appropriate staff. If the reaction is considerable,
every effort will be made to report it immediately so that facility
staff can act to intervene. If difficulties arise with a member of
the normal group, he/she will be referred to the local mental health
center.

Confidentiality is assured through (a) a coding procedure,

(b) locked records, and (c) destruction of all identifying information
after such is no longer necessary.

It is hoped that the study will be of benefit to the subject,
the profession, and society in general. Through the administration
of the PNIM, the subject may develop insights into the processes and
dynamics operating in his/her social network and perhaps discover the

extent of available social support. Gains within the professional

212

realm include the following: (a) determination of the validity and
usefulness of self-report scales in social network analysis, clarify-
ing the value of previous contributions to the field who based their
work on self-report; (b) greater insight in the psycho-social func-
tioning of the schizophrenic relative to the normal subject; and

(c) recommendations regarding the psycho-social management and
treatment of schizophrenia. Society, in general, may be profited by

the knowledge as used by the profession.

Consent Procedures

 

Two sets of consent procedures have been developed to accommodate
to the intrinsically different characteristics of the two samples, the
normal and the schizophrenic, and the environments in which they will
be identified. The term schizophrenia is a psychiatric diagnosis that
denotes a certain type of mental illness; as such, it is subject to
regulation and strict definition. Those facilities responsible for
the housing and/or treatment of the schizophrenic are ethically and
legally charged to protect the identity of the schizophrenic and to
maintain the confidentiality of the schizophrenics' records. Excep-
tions are defined legally, must be justifiable, and are subject to
stringent safeguards. Accordingly, the identification of the schizo-
phrenic sample and the acquisition of informed consent for the study
are difficult and complex procedurally and involve ethical and legal
considerations, the focus explicitly being the protection of the wel-
fare of the schizophrenic. The term normal, used to describe the

other sample, is neither a psychiatric diagnosis nor is subject to

213

regulation; therefore, the consent procedures are relatively simple
and straightforward, although the focus remains protection of the
subjects' welfare. The consent procedures for the normal sample will
be described first.

The sample of normal subjects will be chosen from groups assumed
to contain "normal" members; those selected for participation in the
study must meet the requirements for inclusion. The purpose and
nature of the study will be described to the potential subject (see
Explanation of the Research Study [1]), and if the potential subject
meets the requirements and agrees to participate in the study, then
an informed consent to participate in the study will be obtained in
writing (see Statement of Informed Consent [1]). If the subject is
involved in stage one of the study where corroboration of the data is
required, then a release to notify others in the subject's network
will also be obtained (see bottom of Statement of Informed Consent
[1]). Those identified by the subject will be contacted by phone,
informed of the reason for the phone call and nature of the study, and
asked for their consent to participate in the study. A letter describ-
ing the study (see Corroborating Member Letter Form) and an informed
consent form (see Statement of Informed Consent [2]) will be sent by
mail with a self-addressed stamped envelope enclosed. When the con-
sent form is returned, the person-~identified by the subject--will be
questioned as to the relationship with the subject either in person or
over the phone.

The examination of the schizophrenic sample may take place in a

day treatment/hospital program or an inpatient unit of a ”mental health

214

facility" or "licensed private hospital" (terms in quotation marks

are defined by Illinois statute). To secure permission for the run-
ning of the study, a research request will be submitted for approval
to the research director and/or committee of the facility. If the
request is approved, the procedure to identify the sample and gain the
consent forms would proceed as follows: First, a designee of the
research director and/or committee will identify those schizophrenics
within the facility who meet the requirements for inclusion in the
study. Second, a group will be selected randomly from the original
identified number. After they have been informed of the nature and
purpose of the study (to be determined by the facility), a release of
information (see Release of Information) which conforms to Illinois
statute will be obtained from those schizophrenics potentially willing
to participate in the study, allowing their identity, age, diagnosis,
and residential status to be released to the researcher. In a meeting
with the researcher, the potential subjects will be informed as to the
reasons for their initial selection, and the nature and purpose of

the study (see Explanation of the Research Study [2]). If the poten-
tial subject agrees to participate in the study, then a consent form
to participate in the study will be obtained in writing (see Statement
of Informed Consent [1]). If the subject is involved in stage one of
the study where corroboration of the data is required, then a release
to notify others in the subject's network will also be obtained (see
bottom of Statement. of Infbrmed Consent [1]). Those identified by
the subject will be contacted by phone, informed of the reason for

the phone call and nature of the study, and asked for their consent

 

 

215

to participate in the study. A letter describing the study (see
Corroborating Member Letter Form) and an informed consent form (see
Statement of Informed Consent [2]) will be sent by mail with a self-
addressed stamped envelope enclosed. When the consent form is
returned, the person-~identified by the subject--will be questioned
as to his/her relationship with the subject either in person or over

the phone.

 

 

216

Explanation of the Research Study (1)

 

The purpose of the research study in which you are being asked
to participate is to examine one's social network, that is to say,
the social relationships one has with family, relatives, friends,
co-workers, and others.

You will be asked to identify important persons to you and to
respond to a series of questions about your relationship with these
persons. A possible additional request may apply to you: With your
permission, some persons of your network will be contacted in order
to further examine their relationship with you. They will pply_be
notified of your participation in this research study; all of your

answers, as well as their answers, will remain in strict confidence.

217

Statement of Informed Consent (l)

I have freely consented to take part in a scientific study being con-
ducted by Kenneth L. Carrico, under the supervision of Norman R.
Stewart, Ph.D., Professor, Michigan State University.

The study has been explained to me, and I understand the explanation
that has been given and what my participation will involve.

I understand that I am free to discontinue my participation in the
study at any time without penalty.

I understand that the results of the study will be treated in strict
confidence and that I will remain anonymous. Within these restric-
tions, results of the study will be made available to me at my request.

I understand that my participation in the study does not guarantee any
beneficial results to me.

I understand that, at my request, I can receive additional explanation
of the study after my participation is completed.

Signed

 

Date

 

Code

 

I have freely consented to the additional request that persons in my
social network be contacted by the researcher, the purpose being
further examination of the identified relationships. Only these per-
sons will be informed of my participation in the research study; in
all other ways, my anonymity will be preserved.

I understand that all of my answers, as well as their answers, will
remain in strict confidence.

Signed

 

Date

 

 

218

Corroborating Member Letter Form

 

(Name and address of corroborating network member)

Dear (Contact's name) ,

 

You were contacted because (subject's name) identified you as

 

a/an (relationship) while participating in a research study and

 

permitted us to contact you. The purpose of this study is to examine
one's "social network," that is, the social relationships one has with
family, relatives, friends, co-workers, and others.

To clarify our data, I would like to ask you a series of questions

briefly about your relationship with (subject's name) . Your answers

 

to these questions and your participation in the study will remain
confidential.

Please carefully read the enclosed Statement of Informed Consent.
If you agree to participate in this research endeavor, please sign
and date the form, and return it in the enclosed envelope. If you
have any questions, please leave a message at this number, 446-1110,
and I will return your call. If you wish the results of the study
and/or additional explanation after the study is completed, please
make a note on the consent form.

Thank you very much for your time and consideration.

Sincerely yours,

Kenneth L. Carrico

219

Statement of Informed Consent (2)

I have freely consented to take part in a scientific study being con-
ducted by Kenneth L. Carrico, under the supervision of Norman R.
Stewart, Ph.D., Professor, Michigan State University.

The study has been explained to me, and I understand the explanation
that has been given and what my participation will involve.

I understand that I am free to discontinue my participation in the
study at any time without penalty.

I understand that the results of the study will be treated in strict
confidence and that I will remain anonymous. Within these restric-
tions, results of the study will be made available to me at my request.

I understand that my participation in the study does not guarantee any
beneficial results to me.

I understand that, at my request, I can receive additional explanation
of the study after my participation is completed.

Signed

 

Date

 

Code

 

220

Explanation of the Research Study (2)

The research director/committee of this facility selected you as
a potential participant in an approved research study. The reasons for
identifying you specifically include your age, diagnosis, and residen-
tial status--your residing with your family.

The purpose of the research study in which you are being asked to
participate is to examine one's social network, that is to say, the
social relationships one has with family, relatives, friends, co-
workers, and others.

You will be asked to identify important persons to you and to
respond to a series of questions about your relationship with these
persons. A possible additional request may apply to you: With your
permission, some persons of your network will be contacted in order
to further examine their relationship with you. They will ppr be
notified of your participation in this research study; all of your

answers as well as their answers will remain in strict confidence.

11.

221

RELEASE OF INFORMATION

Date prepared

 

I authorize to release
(Facility)

 

 

(Specify nature of information to be disclosed)

Information about to
(Patient)

 

 

(Name) (Where information to be sent) (Address)

For the specific purpose of

 

 

I understand that I have the right to inspect and copy the informa-
tion to be disclosed.

I understand that I may revoke this authorization at any time
except to the extent that action has been taken on this authori-
zation. I further understand that this authorization shall expire
without my express revocation on:

, 19___.

 

I further understand that the agency which receives this informa-
tion will not disclose this information without further written
consent.

10.

 

 

(Signature) (Relationship)
12.

 

 

(Date) (Witness)

222

Testing Procedure

 

The Pattison Psychosocial Network Inventory, Modified (PNIM) for
interviewing, will be pilot tested and refined as necessary before
being used in this study. For the study, the PNIM will be adminis-
tered to all subjects within a period of two months. Administration
will take place in a setting familiar to the subject.

At the administration, the examiner will make an introductory
statement explaining the purpose of the interview and obtain an
informed consent in writing. The subject will be presented with the
General Information Sheet; the examiner will answer questions and
guide as needed to insure accuracy and completeness.

A three-step process now begins. Step one entails the listing of
all “important" persons in the subject's social network. The examiner
will record the responses. An interview guide for the examiner,
designating types of possible relationships, will be used to insure
the development of a complete network. Step one ends when all names
of those in the subject's network have been obtained. Step two
involves the assessing of persons named along the items of the PNIM.
The PNIM item guide will, at this point, be placed in front of the
subject. The examiner will interview using the PNIM as a visual device
to key the subject and otherwise structure the interview. For step
three, the subject will be asked to determine which persons in the
social network know each other, i.e., have some type of relationship
with each other, outside their relationship with the subject.

After the administration of the PNIM is completed, the subject

and the examiner will identify a minimum of four key network members

223

(when corroborating information is required): two in the family;
one, a friend; and another, a co-worker. Permission will be obtained
to contact these four network members by phone, inform them of the
nature of the study, and enlist their participation. The interview
may be conducted by phone or in person, after a consent form has been
signed and received.

Three steps ensue during the corroborating interview. First,
demographic data will be gathered. Second, the corroborator will be
asked to list persons by first name and initial, known to be important
to the subject, and briefly explain why. Last, the relationship
between the corroborator and the subject will be assessed along the
items of the PNIM. Corroborating data will be gathered until an

adequate sample is obtained to test for concordance of results.

10.

224

PSYCHOSOCIAL NETWORK INVENTORY, MODIFIED

GENERAL INFORMATION

Age . 2. Sex: Male . Female .

Education: Last year of school completed or final degree

Employment: Your occupation .
Work Status: Employed . Unemployed . Retired . Student .

 

Ethnic group: White . Black . Mexican/American . Oriental
American Indian . East Indian . Other .

Marital status: Never Married . Married . Separated .
Divorced . Widowed .
Religious preference: Protestant . Catholic . Jewish

Islam . Other . None .

 

Residence:
Living arrangement: Alone__. In home with family and/or relatives___
In home with unrelated others
In a facility or residential setting__.
Number of moves in past 5 years___.

Number of years in present address

Major life events: Check any of the following if they have occurred during the
last year in your life and/or your family:

 

 

 

 

 

 

 

Birth : School or job change .

Death . Serious accident .

Divorce or separation . Serious illness or physical disability__.
Family or marital conflict . Trouble with law or legal problems—
Family move . Other (specify)

Financial problems .

Marriage or remarriage . None .

Group membership: List the names of all formal or informal community asso-
ciations that you may belong to, such as church, fraternal
club, service club, self-help group, social or recreational
groups, etc., regardless of your level of participation.

 

 

 

 

225

1. How often do you usually have CONTACT with this person, whether face-to-face,
or by phone or letter?

 

 

a. Daily.

b. At least once a week.

c. At least once a month.

d. At least once every six months.

e. At least once a year.

2. How CLOSE does this person live to you?

a. Lives in the same household.

b. Within the same general neighborhood or locale.

c. Within about 30 minutes drive from you.

d. Within about 2 hour drive from you.

e. Beyond a 2 hour drive from you.

3. Which types of ACTIVITIES do you and this person engage in according to the

following list.

3. FAMILY activities such as meals, holidays, vacations or reunions.

b. EMPLOYMENT-related activities such as working with co-workers or
supervisors.

c. ROMANTIC activities such as dating, dancing, or going out for dinner.

d. CONVERSATIONAL activities such as intimate, personal sharing. or philo-
sophical discussions.

e. SOCIAL activities such as parties, banquets, or visiting neighbors or
friends.

I. RECREATIONAL activities such as playing cards, participation in or

 

attendance at sports events.

 

g. FRATERNAL activities such as participation in clubs or other organi-
zations.

h. RELIGIOUS activities such as attendance to church, synagogue. or temple.

i. POLITICAL activities such as rallies or discussions of politics.

j. VOLUNTEER work such as service to the community, giving blood, or
hospital work.

4. How much do you help this person by providing GUIDANCE when they need it,
such as giving advice on a task or decision, or giving feedback on an action
taken?

Not aTaU Rarely On some occasions Often Very Trequently

5. How much does this person help you by expressing AFFECTION when you need

it, such as a hug, kiss, or pat on the back?

 

Very frequently Often On some occasions Rarely Not at all

226

6. How much do you help this person by DOING THINGS for them when they need
it, such as helping with household tasks, providing personal or family care,
assisting on the job, or even lending money?

 

Very mquently bfien On some occasions Rarely Not at all

7. How much do you help this person by expressing AFFECTION when they need
it, such as a hug, kiss, or pat on the back?

 

 

Not afall Rafizly On some occasions Ofien Very fiquently
8. What TYPE of feelings does this person have toward you, regardless of their
strength?
All positive Mostly positive About equally mixed MostlyT negative AU negative
9. What is the STRENGTH of the feelings and thoughts this person has toward you,

regardless of their type?

 

Weak Mild Moderate Strong Very strong

10. What is the STRENGTH of the feelings and thoughts you have toward this person,
regardless of their type?

 

Weak Mild Moderate Strong Very strong

11. How much does this person help you by providing GUIDANCE when you need
it, such as giving advice on a task or decision, or giving feedback on an action
taken?

 

Very fizzquently Often On some occasions Rarely Not at all

12. How much do you help this person by giving EMOTIONAL SUPPORT when they
need it, such as praising them, being a good listener, or providing them with
encouragement?

 

Very frequently Often On some occasions Rarely Not at all

227

13. How much does this person help you by DOING THINGS for you when you need
it, such as helping with household tasks, providing personal or family care,
assisting on the job, or even lending money?

 

Not Elk all Rarely On sortie occaBons Ofien Very frequently

14. What TYPE of feelings and thoughts do you have toward this person, regardless
of their strength?

 

All negative Mostly negative About eq'ually mixed Mostly7 positive All positive

15. How much does this person help you by giving EMOTIONAL SUPPORT when you
need it, such as praising you, being a good listener, or providing them with
encouragement?

 

Very frequently Often On some occasions Rarely Not at all

REFERENCES

228

REFERENCES

Allisi, A. S. Social work with families in group-service agencies:
An overview. Family Coordinator, I969, 18, 39I-40I.

 

Almond, R. The healing community. New York: Jason Aronson, I974.

 

American Psychiatric Association (R. L. Spitzer, M.D.: Chairperson,
Task Force on Nomenclature and Statistics). Diagnostic and
statistical manual of mental disorders (3rd ed.). Washington,
D.C.: Author, I980.

 

 

Arieti, S. Interpretation of schizophrenia (2nd ed.). New York:
Basic Books, I974.

 

Barnes, J. A. Class and committees in a Norwegian island parish.
Human Relations, I954, Z, 39-58.

 

Bateson, 6., Jackson, D. D., Haley, J., & Weakland, J. Towards a
theory of schizophrenia. Behavioral Science, I956, I, 25l-264.

 

Beard, J. H., Malamud, T. J., & Rossman, E. Psychiatric rehabilita-
tion and long term rehospitalization rates: The findings of two
research studies. Schizophrenia Bulletin, I978, 3, 622-635.

 

Beck, J. C. Social influences on the prognosis of schizophrenia.
Schizophrenia Bulletin, I978, 3, 86-IOI.

 

Becker, E. Toward a theory of schizophrenia: External objects and
the creation of meaning. Archives of General Psychiatry, I962,
7, I70-l8l.

 

Bell, N. w. Extended family relations of disturbed and well families.
Fami y Process, I962, l, I75-I93.

 

Birley, J. L., & Brown, G. Crises and life changes preceding the
onset or relapse of acute schizophrenia. British Journal of
Psychiatry, I970, 116, 327-333.

 

 

Boissevain, J. The place of non-groups in the social sciences.
Man: Journal of the Royal Anthropological Institute of Great
Britain and Ireland, I968, a, 542-556.

 

Boissevain, J. Friends of friends: Networks, manipulators and coali-
tions. New York: St. Martin‘s Press, I974.

229

230

Bott, E. Urban families: Conjugal roles and social networks. In
S. Leinhardt (Ed.), Social networks: A developing paradigm.
New York: Academic Press, I977. (Originally published, I955.)

 

Bott,]gé7 Family and social network. London: Tavistock Publications,

Bowen, M. A family concept of schizophrenia. In D. Jackson (Ed.),
Etiology of schizophrenia. New York: Basic Books, I960.

 

Bowen, M. Family psychotherapy. American Journal of Orthopsychiatry,
I96I, g1, 40-60.

 

Bowen, M. Family psychotherapy with schizophrenia in the hospital
and in private practice. In I. Bozoremyi-Nagy & J. L. Framo
(Eds.), Intensive family therapy, New York: Harper & Row, I965.

 

Bowen, M. Schizophrenia as a multi-generational phenomenon. In
M. M. Berger (Ed.), Beyond the double bind. Communication and
family systems, theories, and techniques. New York: Brunner/
MazeI, I978.

 

 

Brim, J. A. Social network correlates of avowed happiness. Journal
of Nervous and Mental Disease, I974, 158, 432-439.

 

Brown, G. H., Birley, J. L. T., & Wing, J. K. Influence of family
life on the course of schizophrenic disorders: A replication.
British Journal of Psychiatry, I972, 121, 24I-258.

 

Budson, R. 0., & Jolley, R. E. A crucial factor in community program
success: The extended psychosocial kinship system. Schizophrenia

 

Bulletin, I978, 4, 609-62I.

Cameron, N. The psychology of behavior disorders: A biosocial inter-
pretation. Cambridge: Mifflin, I947.

Caplan, G. Support systems and community mental health. New York:
Behavioral Publications, I974.

Cobb, S. Social support as a moderator of life stress. Psychosomatic

 

 

Medicine, I976, 87, 300-3I4.

Cohen, C. I., & Sokolovsky, J. Schizophrenia and social networks:
Ex-patients in the inner city. Schizophrenia Bulletin, I978, 4,
546-560.

Cohn, C. K., & Talmadge, J. M. Extended family presents. Family
Therapy, I976, a, 235-244.

Coleman, J., Katz, E., & Menzel, H. The diffusion of an innovation
among physicians. Sociometry, I957, 29, 253-270.

 

23I

Day, M., & Semrad, E. V. Schizophrenic reactions. In A. M. Nicholi
(Ed.), The Harvard guide to modern psychiatry. Cambridge:
Harvard University Press, I978.

Dincin, J., Selleck, V., & Streicker, S. Restructuring parental
attitudes--Working with parents of the adult mentally ill.
Schizophrenia Bulletin, I978, 4, 597-608.

 

Dohrenwend, B. P., & Dohrenwend, B. S. Social status and psychologi-
cal disorder. New York: Wiley, I969.

 

 

Dohrenwend, B. P., & Dohrenwend, B. S. Social and cultural influences
in psychopathology. Annual Review of Psychology, I974, 25,
4I7-452.

 

Dunham, H. W. Community and schizophrenia: An epidemiological analy-
sis, Detroit: Wayne State University Press, I965.

 

Fairbairn, R. Object-relations theory of the personality. New York:
Basic Books, I952.

 

Fairweather, G., Sanders, D., Maynard H., & Cressler, 0. Community
life for the mentally ill: An alternative to institutional care.
Chicago: Aldine Publishing Company, I969.

 

Faris, R. E. L. Cultural isolation and the schi20phrenic personality.
American Journal of Sociology, I934, 49, I55-I64.

 

Feldman, F., & Schertz, F. Family social welfare. New York:
Atherton, I967.

 

Fenichel, 0. The psychoanalytic theory of neurosis. New York:
W. W. Norton, I945. (Copyright renewed, I972.)

 

Ferreira, A. J., & Winter, W. Information exchange and silence in
normal and abnormal families. Family Process, I968, Z, 251-276.

 

Freud, S. 0n narcissism: An introduction. In E. Jones (Ed.),
Collected papers (Vol. IV). London: Hogarth Press, I956.
(Originally published, I9I4.)

 

Fromm-Reichmann, F. Notes on the development of treatment of schizo-
phrenia by psychoanalytic psychotherapy. Psychiatry, I948, 11,
263-273.

 

Gerlach, L., & Hine, V. The social organization of a movement of
revolutionary change: Case study, black power. In N. E. Whitten,
Jr. & J. Szwed (Eds.), Afro-American anthropology. New York:
Free Press, I970.

 

232

Gittelman-Klein, R., & Klein, 0. F. Premorbid social adjustment and

prognosis in schizophrenia. Journal of Psychiatric Research,
I969, 2, 35-53.

Goldberg, E. M., & Morrison, S. L. Schizophrenia and social class.
British Journal of Psychiatry, I963, 199, 785-802.

Gunderson, J. G. Drugs and psychosocial treatment of schizophrenia
revisited. Journal of Continuing Education in Psychiatry, I977,
38, 25-40.

Gunderson, J. G. A reevaluation of milieu therapy for nonchronic
schizophrenic patients. Schizophrenia Bulletin, I980, 6, 64-69.

 

Gutkind, P. C. W. African urbanism: Mobility and social network.
International Journal of Comparative Sociology, I965, 6, 48-60.

 

Hammer, M. An analysis of social networks as factors influencing the
hospitalization of mental patients (Doctoral dissertation, Colum—
bia University). Dissertation Abstracts International, I96I, 22,
968A-969A. (University Microfilms N0. 6I-3883T)2

 

Hammer, M. Influence of small social networks as factors in mental
hospital admission. Human Organization, I963-64, 22, 243-25I.

 

Hammer, M., Makiesky-Barrow, S., & Gutwirth, L. Social networks and
schizophrenia. Schizophrenia Bulletin, I978, 4, 522-545.

 

Hammer, M., & Schaffer, A. Interconnectedness and the duration of
connections in several small networks. American Ethnologist,
I975, 2, 297-308.

 

Hare, E. H. Mental illness and social conditions in Bristol. Journal
of Mental Science, I956, 102, 349-357. (a)

Hare, E. H. Family setting and the urban distribution of schizophrenia.
Journal of Mental Science, I956, 102, 753-760. (b)

 

Hollander, E. P. Principles and methods of social psychology. New
York: Oxford University Press, I967.

Jaco, E. G. The social isolation hypothesis and schizophrenia.
American Sociological Review, I954, 12, 567-577.

 

Jaco, E. G. Mental illness in response to stress. In S. Levine &
N. A. Scotch (Eds.), Social Stress. Chicago: Aldine, I970.

 

233

Jacobs, 5., 3 Myers, J. Recent life events and acute schizophrenic
psych051s: A controlled study. Journal of Nervous and Mental
Disease, I976, 162, 75-87.

 

Kayton, L., Beck, J., & Koh, S. D. Postpsychotic state, convalescent
environment, and therapeutic relationship in schizophrenic out-
come. American Journal of Psychiatry, I976, 122, I269-I274.

 

Klein, 0. F. Psychosocial treatment of schizophrenia or psychosocial
help for people with schizophrenia? Schizophrenia Bulletin,
I980, 6, 122-I30.

 

Kleiner, R. J., 8 Parker, S. Network participation and psychosocial
impairment in an urban environment. NIMH Grant MH I9897, Final
Report, I974.

Kohn, M. L., & Clausen, J. Social isolation and schi20phrenia.
American Sociological Review, I955, 29, 265-273.

 

Laing, R. Sanity, madness and the family. Baltimore: Penguin, I964.

 

Leff, J. P. Schizophrenia and sensitivity to the family environment.
Schizophrenia Bulletin, I976, 2, 566-574.

 

Lidz, T. The origin and treatment of schizophrenic disorders. New
York: Basic Books, I973.

Lidz, T., Cornelison, A. R., Fleck, S., & Terry, D. The intrafamilial
environment of schizophrenic patients: II. Marital schism and
marital skew. American Journal of Psychiatry, I957, 114, 24l-248.

Lidz, T., Terry, D., & Fleck, S. The intrafamilial environment of the
schizophrenic patient: VI. The transmission of irrationality.
Archives of Neurological Psychiatry, I958, 22, 305-3I6.

 

Liem, J. H. Family studies of schizophrenia: An update and commen-
tary. Schizophrenia Bulletin, I980, 6, 429-455.

 

Mayer, A. C. System and network: An approach to the study of politi-
cal processes in Dewas. In T. Madan & G. Sarana (Eds.), Indian
anthropology: Essays in memory of D. N. Majumdar. Bombay: Asia
Publishing Company, I962.

 

Mayer, A. C. The significance of quasi-groups in the study of complex
societies. In M. Banton (Ed.), The social anthropology of complex
societies. New York: Praeger, I966.

Minuchin, S. Families and family therapy. Cambridge: Harvard Uni-
versity Press, I974.

234

Minuchin, S., Montalvo, 8., Guerney, B., Rosman, B., & Schumer, F.
Families of the slums. New York: Basic Books, I967.

Mishler, E. G., & Waxler, N. E. Family interaction processes and
schizophrenia: A review of current theories. Merrill-Palmer
Quarterly of Behavior and Development, I965, 12, 269-3I5.

 

Mishler, E. G., & Waxler, N. E. The sequential patterning of interac-
tion in normal and schi20phrenic families. Family Process, I975,
I4, I7-50.

 

Mitchell, J. C. (Ed.). Social networks in urban situations.
Manchester: Manchester University Press, I969.

 

Mosher, L. R., & Keith, S. J. Psychosocial treatment: Individual,
group, family, and community support approaches. Schizophrenia
Bulletin, I980, 6, I0-4I.

 

Mosher, L. R., Wild, C., Valcov, A., & Feinstein, A. E. Cognitive
style, schizophrenia, and the family: Methodological implications
of contextual effects. Family Process, I972, ll, I25-I46.

 

Myers, J. K., & Roberts, 8. H. Family and class dynamics in mental
illness. New York: Wiley, I959.

 

Pattison, E. M. A theoretical-empirical base for social system
therapy. In E. R. Foulkes, R. M. Wintrob, J. Westermeyer, &
A. R. Favazzo (Eds.), Current perspectives in cultural psychiatry.
New York: Spectrum, I977.

 

Pattison, E. M. Personal communication, March I6, I979.
Pattison, E. M. Personal communication, February 5, I980.
Pattison, E. M., de Francisco, 0., Wood, P., Frazier, H., & Crowder, J.

A psychosocial kinship model for family therapy. American Journal
of Psychiatry, I975, 122, I246-I25I.

 

 

Pattison, E. M., Llamas, R., & Hurd, G. Social network mediation of
anxiety. Psychiatric Annals, I979, 9, 56-67.

 

Ratcliffe, W. D., & Azim, H. Social networks of psychiatric patients
and normals. Paper presented at the meeting of the Banff Canadian
Psychiatric Conference, I975.

 

 

Redfield, R. Peasant society and culture: An anthropological approach
to civilization. Chicago: University of Chicago Press, I956.

 

Reid, L. S. Schizophrenia and manic-depressive illness. In C. W.
Johnson, J. R. Snibbe, & L. A. Evans (Eds.), Basic psychopathology;
Ayprogrammed text. New York: Spectrum, I975.

 

235

Reiss, D. Varieties of consensual experience: I. A theory for
relating family interaction to individual thinking. Family
Process, I97I, 19, I-28.

Reiss, D. The family and schizophrenia. American Journal of
ngchiatry, I976, 122, I8I-I85.

 

 

Riskin, J., & Faunce, E. E. An evaluative review of family interac-
tion research. Family Process, I972, 11, 365-455.

 

Rosen, J. N. Directypsychoanalytic psychiatry. New York: Grune 8
Stratton, I962.

 

Rueveni, U. Networking families in crisis. New York: Human Sciences
Press, I979.

 

Silberfeld, M. Psychological symptoms and social supports. Social
Psychiatry, I978, 12, II-I7.

 

Sokolovsky, J., Cohen, C. I., Berger, 0., & Geiger, J. Personal
networks of ex-mental patients in a Manhattan s.r.o. hotel.
Human Organization, I978, 31, 5-I5.

 

Speck, R. V. Psychotherapy of the social network of a schizophrenic
family. Family Process, I967, 8, 208-2I4.

 

Strauss, J. S., & Carpenter, W. T. The prediction of outcome in
schizophrenia: 1. Archives of General Psychiatry, I972, 22,
739-746.

 

Strauss, J. S., & Carpenter, W. T. Prediction of outcome in schizo-
phrenia: III. Five-year outcome and its predictors. Archives
of General Psychiatry, I977, 24, I59-l63.

 

Sullivan, H. S. Conceptions of modern psychiatry, New York: Norton,
I953.

 

Sullivan, H. S. Schizophrenia as a human process. New York: Norton,
I962.

Szasz, T. A contribution to the psychology of schizophrenia. A.M.A.
Archives of Neurology and Psychiatry, I957, 22, 420-436.

Thibaut, J. W., & Kelley, H. H. The social p§ychology of groups.
New York: Wiley, I959.

Tolsdorf, C. C. Social networks, support and c0ping: An exploratory
study. Family Process, I976, 15, 407-4I8.

 

Tolsdorf, C. C. Personal communication, February 5, I980.

236

Torrey, E. F. Schizophrenia and civilization. New York: Jason
Aronson, I980.

 

Ullman, L. P., & Krasner, L. A psychological approach to abnormal
behavior. Englewood Cliffs, N.J.: Prentice-Hall, I969.

 

Vaughn, C. E., & Leff, J. P. The influence of family and social
factors on the course of psychiatric illness: A comparison of
schizophrenic and depressed neurotic patients. British Journal
of Psychiatry, I976, 129, I25-I37.

 

 

Walsh, F. W. Concurrent grandparent death and birth of a schizophrenic
offspring: An intriguing finding. Family Process, I978, ll,
457-463.

 

Waxler, N. E., & Mishler, E. G. Parental interaction with schizo-
phrenic children and well siblings. Archives of General
ngchiatry, I97I, 25, 223-23I.

 

 

Whitten, N. E., & Wolfe, A. W. Network analysis. In J. J. Honigmann
(Ed.), Handbook of social and cultural anthropology. Chicago:
Rand McNaIIy, I973.

Wild, C. M., Shapiro, L. N., & Abelin, T. Communication patterns and
role structure in families of male schizophrenics. Archives of
General Psychiatry, I977, 24, 58-70.

 

 

Wild, C. M., Shapiro, L. N., & Goldenberg, L. Transactional commu-
nication disturbances in families of male schizophrenics. Family
Process, I975, 14, I3I-I60.

Wolf, E. R. Kinship, friendship, and patron-client relations in
complex societies. In M. Banton (Ed.), The social anthropology
of complex societies. New York: Praeger, I966.

 

Wolfe, A. W. 0n structural comparisons of networks. Canadian Review
of Sociology and Anthropology, I970, 4, 226-244.

 

Wynne, L., Rychoff, 1., Day, J., & Hirsch, S. Pseudomutuality in the
family relationships of schizophrenics. Psychiatry, I958, 21,
205-220.