RORSCHACH gffiNfig THENKBNG
DISORGANEZATEO‘N; AND? WETHDEAWAL
EN PRQCESS AND
REACTWE SCHIEZQPHREMCfi

Thesis few {the Dam of MI. D.
MICHEGAN STATE UNIVERSI'EY
Rosa E. Carter
3969

     
 

L I B R A R Y
Michigan Sum:
University

RT“? if"!

 
   

This is to certify that the

thesis entitled

RORSCHACH SIGNS, THINKING DISORGANIZATION,
AND WITHDRAWAL IN PROCESS AND
REACTIVE SCHIZOPHRENICS

presented by

Ross E. Carter

has been accepted towards fulfillment
of the requirements for

_2h;D;__&mnmin Dept. of Psychology

/)

/g%21

all Major professor
Date (%//’Z, //?’4
Jr

0-169

ABSTRACT
RORSCHACH SIGNS, THINKING DISORGANIZATION,

AND WITHDRAWAL IN PROCESS AND
REACTIVE SCHIZOPHRENICS

BY

Ross E. Carter

The purpose of this study was to determine whether
reactive schizophrenics tested during the acute stage pro-
duce more Rorschach protocols characteristic of schizo-
phrenics than reactive: in the non-acute stage. On the basis
of prior research it was assumed that process subjects pro-
duce schizophrenic protocols regardless of length of
hospitalization. It was furthermore assumed that differences
in levels of adjustment, or recovery, would be related to
differences in Rorschach diagnosis.

Sixty male hospitalized schizOphrenics were classi—
fied as either process or reactive on the Phillips Premorbid
History Scale. The Subjects were then divided equally ac-
cording to length of current hospitalization to form four
groups: early reactive, late reactive, early process, and
late process. All subjects were administered the Rorschach
and rated for level of adjustment on the Thinking Disorgani-

zation and Withdrawal Scales of the Psychotic Reaction

Ross E. Carter

Profile developed by Lorr, O'Conner and Stafford. A control
group was formed by taking the protocols of 15 non
schiZOphrenic psychiatric patients from hospital files for
comparison purposes. The protocols were scored by two ex-
perienced psychologists according to Beck's system. The re-
liability of the judges was .92. An additional measure of
adjustment was obtained by having the judges rate the proto-
cols on a 6 point scale for emotional health.

A survey of the literature yielded 39 different
Rorschach signs of schiZOphrenia which were employed in the
present study. The non schizOphrenics were compared with the
entire group of schizophrenics as well as with each process
and reactive group on the number of Rorschach signs of
schiZOphrenia. None of these comparisons yielded signifi-
cant differences. A tendency for early process subjects to
produce more signs than the non schiZOphrenics was noted.

All possible comparisons were made between the pro-
cess and reactive groups. None of these comparisons resulted
in significant differences in the number of Rorschach signs
of schiZOphrenia. A trend towards more signs was noted in
early process subjects when compared with late process
subjects.

Further comparisons between the groups were made for

each of the Rorschach signs. No differences were found

Ross E. Carter

between the non schizophrenics and the combined group of
schizOphrenics with respect to the presence of the signs.

The judges‘ global ratings of emotional health,
based on the Rorschach, showed the non schizophrenics to be
less disturbed than either the early reactive or early pro-
cess subjects. No differences were found, however, between
the process and reactive subgroups themselves. The same
trend was noted when the process and reactive groups were
compared on the Thinking Disorganization and Withdrawal
Scales. These findings suggest that the process and reactive
subjects may have been nearly equal in level of adjustment.
The one exception was that early process subjects showed
more thinking disorganization and withdrawal behavior than
late reactives.

Finally, the schizOphrenics and non schizOphrenics
were also compared with the Rorschach norms reported by Beck.
Schizophrenics were found to differ significantly from the
normals on 15 variables while non schizOphrenics differed on
only 6 variables. These findings tend to suggest that the
Rorschach signs of schizophrenia may not be true indicators
of schizOphrenia and that the schizOphrenics were in fact
highly deviant in their Rorschach protocols. Inferences
based on these findings are limited by the questionable para-

metric statistics used in the normative study.

Ross E. Carter

The results of this study indicate that the Rorschach
signs do not differentiate between non schiZOphrenic patients
and schizophrenics as a group or as the latter are subdivided
along the process-reactive dimension. These findings may be
limited, however, only to the specific Rorschach signs of
schizOphrenia used in the study. The fact that the subjects
in the study were taking drugs may also account, at least in
part, for the findings. The effect of drugs on the Rorschach
performance of schizOphrenics is unknown, but it is entirely
possible that they function to reduce the amount ofpathology
produced by schizophrenics in Rorschach protocols. The re-
sults also raise some serious questions regarding the utility
of the process-reactive dimension.

It is suggested that future research should exercise
caution in regards to the drug variable. All subjects should
be taking the same drugs in the same amounts, or drug—free
subjects should be used. The nature of the control group and
the extent of its pathology is another important aSpect to
be considered in future research. Attention should be given
to selecting subjects who are not disturbed and who can truly

be regarded as non schizOphrenic.

APPROVED:

DATE:

RORSCHACH SIGNS, THINKING DISORGANIZATION,
AND WITHDRAWAL IN PROCESS AND
REACTIVE SCHIZOPHRENICS
By»~
-C ‘3

Ross E. Carter

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Psychology

Fall, 1969

Ca {Va/I 7 2 .3

L/ , '2 3’ (3

To Charl,

Thank you for helping build a bookcase one gray winter day
so long ago and for being near me now that the sun has

risen. Without you it would mean little.

ii

ACKNOWLEDGMENTS

I would like to thank each of my committee members
for helping with the dissertation. Particular thanks go to
Dr. Rabin for his continued interest and guidance in carrying
out the study. He gave freely of his time, even during his
summer vacation, and was readily available when I needed as-
sistance. Without his concern and help a very important
deadline would never have been met. Dr. Karon's advice re-
garding the statistical analysis of the data was invaluable
and served to reduce a great deal of my anxiety about the
data. Dr. Thornton's suggestions regarding certain features
of the design were relevant to the findings. Dr. Grummon
aided in strengthening the final writeup.

The study could never have been completed without
the help and cooperation of the people who, because of the
nature of the research, were described as subjects. They
were, however, more than just subjects.

I would also like to thank Dr. Stewart G. Armitage
for his assistance in my collecting data at the VA hospital
in Battle Creek. Dr. Marvin Hyman, of Herman Kiefer Hospital
in Detroit, provided an additional and very important source

of data-

iii

Particular and very special thanks go to my wife
Charlene for having worked, really worked, as one of the
judges and as the typist for the difficult first draft. Her
support, encouragement and solace at times of stress were
exceptionally important to me.

Apart from the dissertation itself, I would like to
thank the committee members as well as some non-committee
members for the impact that they have had on my graduate edu-
cation and on me as a person. I was fortunate in being able
to study under Dr. Rabin whom I regard as an excellent
teacher and as an expert. Dr. Karon has always shared what
he knows and in a way which I found to be exciting and
stimulating. Dr. Grummon provided me with a great deal of
gentle guidance during the beginning stages of my experience
as a therapist. Dr. Thornton was helpful with one'of my
first publications. Dr. Lee Erlandson gave much of himself
in some very important and meaningful supervision sessions.
I regard Dr. Bill Kell as having been essential and as a
truly significant other. I will remember all of you and

keep forever what you have all so generously given.

iv

TABLE OF CONTENTS

Page
INTRODUCT I ON 0 o o o o o o o o o o o o o o o o 0' o o o 1
Problems of Validation . . . . . . . . . . . . . . 2
EXploratory Studies . . . . . . . . . . . . . 6
Specific Variable Research . . . . . . . . . . 16
Thinking . . . . . . . . . . . . . . . . . l7
Affect . . . . . . . . . . . . . . 22
Conclusions Regarding the Research . . . . . . 25
The Process-Reactive Dimension . . . . . . . . . . 26
Process-Reactive SchiZOphrenics and the

Rorschach . . . . . . . . . . . . . . . . . 27
PURPOSE . . . . . . . . . . . . . . . . . . . . . . . 35
JHYPOTHESES . . . .. . . . . . . . . . . . . . . . . . . 37
WTHOD o o o o o o o o o o o o o o o o o o o o o o o o 39
Description of Subjects . . . . . . . . . . . . . 39
SchizOphrenic Sample . . . . . . . . . . . . . 39
Non SchizOphrenic Controls . . . . . . . . . . 41
. Techniques and Instruments . . . . . . . . . . . . 42
The Rorschach . . . . . . . . . . . . . 42
The Premorbid History Scale . . . . . . . . . 44
The Ullmann-Giovannoni Scale . . . . . . . . . 47

The Thinking Disorganization and Withdrawal
Scales . . . . . . . . . . . . . . . . 51
Scoring of the Instruments . . . . . . . . . . . . 54
The Rorschach: Reliability Training . . . . . 54
The Rorschach: Scoring the Protocols . . . . 55
Treatment of the Data . . . . . . . . . . . . . . 59
The Rorschach . . . . . . . . . . . . . . . 59

The Thinking Disorganization and Withdrawal
scale 8 o o o o o e o o o o o o o o e o o o o 6 0
RESUDTS . . . . . . . . . . . . . . . . . . . . . . . 62
Comparisons on Individual RSS . . . . . . . . . . 75
Comparisons with Normals . . . . . . . . . . . . . 80

Page

ADDITIONAL FINDINGS . . . . . . . . . . . . . . . . . 84
The Rorschach . . . . . . . . . . . . . . 84
Adjustment Scales and Ratings . . . . . . . . . . 85

The Thinking Disorganization and With-
drawal Scales . . . . . . . . . . . . . . . 85
Judges' Ratings . . . . . . . . . . . . . . . 87

DISCUSSION . . . . . . . . . . . . . . . . . . . . . . 89
The Process Reactive Dimension . . . . . . . . . . 95
Suggestions for Further Research . . . . . . . . . 97

SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . lOO

APPENDICES . . . . . . . . . . . . . . . . . . . . . . 104

A. MEANS AND STANDARD DEVIATIONS FOR EACH GROUP
ON THE VARIABLES OF AGE, EDUCATION, AND

HOSPITALIZATION . . . . . . . . . . . . . . 104
B. HOSPITAL DIAGNOSIS OF NON SCHIZOPHRENIC

SUBJECTS . . . . . . . . . . . . . . . . . . 105
C. THE PREMORBID HISTORY SCALE. . . . . . . . . . 106
D. THE ULLMANN‘GIOVANNONI SCALE . . . . . . . . . 111

E. THE THINKING DISORGANIZATION AND WITH-
DRAWAL SCALES . . . . . . . . . . . . 113

F. PRODUCT MOMENT CORRELATIONS FOR THE SCORES
OF THE RELIABILITY RATERS ON THE THINKING
DISORGANIZATION SCALE AND THE WITHDRAWAL
SCALE . . . . . . . . . . . . . . . . 116

G. PRODUCT MOMENT CORRELATIONS BETWEEN JUDGES
FOR EACH RORSCHACH VARIABLE SCORED. EXP
CEPT WHERE NOTED ALL CORRELATIONS WERE
SIGNIFICANT AT A LEVEL OF .005 (28 df) . . . 117

H. RORSCHACH SIGNS OF SCHIZOPHRENIA . . . . . . . 119
I. THE NUMBER OF SUBJECTS SHOWING THE PRESENCE

OF EACH RSS IN EACH GROUP (N= 15 each
group) . . . . . . . . . . . . . . . . . . 121

vi

Page

J. RANGE OF THE RORSCHACH VARIABLES ON WHICH
THE RSS WERE BASED. SHOWN ARE THOSE
VARIABLES WHICH MAY VARY BEYOND PRESENCE
OR ABSENCE . . . . . . . . . . . . . . . . . 124

K. PRODUCT MOMENT CORRELATIONS FOR SUBJECTS'
SCORES ON THE THINKING DISORGANIZATION
AND WITHDRAWAL SCALES AND THE NUMBER OF
RSS PRODUCED BY THE SUBJECTS (N = 15 each
group) . . . . . . . . . . . . . . . . . . . 127

L. TYPE OF DRUGS AND DRUG DOSAGE IN MILLIGRAMS
BEING TAKEN BY SCHIZOPHRENIC AND NON
SCHIZOPHRENIC SUBJECTS . . . . . . . . . . . 128

REFERENCES . . . . . . . . . . . . . . . . . . . . . . 131

vii

LIST OF TABLES

Table Page

1. Median number of RSS for each group com-
parison with Chi Squares for the number
of subjects scoring above and below the
median RSS. The greater number of sub-
jects showing the RSS is underlined . - . . 63

2. The means and standard deviations of early
and late tested process and reactive sub-
jects on the Thinking Disorganization and
‘Withdrawal Scales . . . . . . . . . . . . . 69

3. Summary table of the 2 X 2 ANOVA for early
and late tested process and reactive sub—
jects on the Thinking Disorganization

Scale . . . . . . . . . . . . . . . . . . . 70
4. Summary table of the 2 x 2 ANOVA for early

and late tested process and reactive sub-

jects on the Withdrawal Scale . . . . . . . 71
5. The results of the comparisons for each RSS . 77
6. Median number of RSS for each group to-

gether with the range for the number of
RSS produced by the subjects in each

group 80
7. The means and standard deviations of the

schizOphrenics, non-schiZOphrenics, and

Beck's normals for 23 Rorschach variables . 81.
8. Judges' ratings of emotional health . . . . . 87
9. Product moment correlations for the number

of RSS produced by the subjects and the

corresponding rating of emotional health . . 88

viii

LIST OF FIGURES

Figure Page

1. The hypothesized extent of thinking dis—
organization and withdrawal behavior for
early and late tested process and re-
active subjects . . . . . . . . . . . . . . 73

2. The observed scores of the early and late
tested process and reactive subjects on
the Thinking Disorganization and With-
drawal Scales . . . . . . . . . . . . . . 74

ix

INTRODUCTION

The purpose of this study was to investigate the
Rorschach performance of process and reactive schiZOphrenics.
Specifically, the study focused on the variable of recovery
as it was related to differences in the Rorschach diagnosis
of process and reactive schizophrenics. The study was based
on the assumption that reactive schiZOphrenics do recover.
Accordingly, it was expected that reactive schizophrenics
who were tested at a point considerably past the initial
period of hospitalization would differ in terms of Rorschach
performance when compared with reactive schizophrenics
tested during the initial period of hospitalization. It was
furthermore assumed that the late tested reactive schizo-
phrenics would differ from process schiZOphrenics tested
either early or later in hospitalization.

A review of the literature on the use of the
Rorschach with schizoPhrenics reveals a great deal of varia-
bility in findings. Such variability could suggest that the
Rorschach may be invalid as a diagnostic tool for use with
schizophrenics. On the other hand, observed differences be-

tween process and reactive schizophrenics tested early or

late would also suggest a possible means for explaining such

variability.

ggoblems of Validation

The validity of the Rorschach as a measure of
personality structure and functioning has been questioned
since the instrument was first introduced. Herman Rorschach
(1942) himself was acutely aware of the unproven and experi-
mental status of his test. Pioneer workers such as Beck
(1935) and Hertz (1941) called attention to the necessity
for fundamental research on the Rorschach and to the need
for validation using objective criteria and rigorous
methodology.

Some workers have resisted efforts to deal with the
validity of the Rorschach on an objective basis and have
been content to accept personal success based on a clinical—
intuitive approach as sufficient evidence in support of
claims for validity. Resistance has been based on the
notion that the Rorschach requires such complex, subjective
skills of interpretation and integration that it is im-
possible to deal with the instrument in a rigorous manner
without violating its very nature. Beck has characterized
this position by saying that: "there has develoPed around
this test the feeling that it is a sort of mysterious instru-

ment, supra-mundane and sui generis, one resting on intuitive

insight and not required to follow the usual cannons of
scientific method" (as quoted by Benton, 1950, p. 45).

There is no doubt that the Rorschach does require
skill for its most effective use and that some workers are
more proficient with the instrument than are others. Indi-
vidual personal success, however, is insufficient as a basis
on which to claim validity for the Rorschach. The reason
for this, as Meehl (1965) has pointed out, is that it is im-
possible to determine the extent to which the test becomes
confounded with the interpreter. Validation as well as in-
validation becomes a matter of who uses the instrument and
has little to do with the actual value of the Rorschach
itself.

Another disadvantage to basing the validity of the
Rorschach on personal use is that such an approach leads to
assumptions about the generalized validity of the instru-
ment. Even though the need for validation has been recog-
nized for some time, it has only recently become apparent
that both questions and assertions posed in blanket fashion
about the validity of the Rorschach are misleading and gross-
ly oversimplified. The matter is not whether the Rorschach

is valid or invalid, but rather, for what the Rorschach is

 

valid or invalid. Assumptions about a general type of
validity lead only to misapplication of the Rorschach as,
for example, when the instrument is used to predict success

in aviation training (Guilford, 1947).

The distinctions made by the APA (French, Michael
__E__l., 1966) regarding types of validity have served to re-
inforce the importance of focusing on the kind of validity
which is of concern. Dana (1962) has stressed the need for
construct validity and concludes that questions regarding
content, predictive, or concurrent validity are inapprOpriate
for projective tests at this time. One may, however, take
issue with Dana's (1962) conclusions. The Rorschach is
widely used on the clinical level for the purposes of person—
ality assessment. Consequently, the most practical, if not
the most pressing need, is to deal with the concurrent
validity of the instrument.

Establishing concurrent validity for the Rorschach
is not a simple matter, particularly as the instrument is
considered to be a measure of psychOpathology. The most
serious difficulty is that of defining criterion groups
against which the Rorschach is to be compared. Most often
research attempts to validate the Rorschach have been made
using groups formed on the basis of traditional diagnostic
systems. The unreliability of these systems has frequently
been noted (Ash, 1949; Mehlman, 1952) and poses a perplexing
problem for validation studies, namely that of how to de-
termine whether inconclusive research results are due to the
inadequacy of the Rorschach or to the inadequacy of the

criteria-

One approach to overcoming the limitations imposed
by the use of traditional classification systems might be to
focus on making gross rather than fine distinctions between
types of pathology, and to test the validity of the Rorschach
against criteria groups merely identified as psychotic,
neurotic or normal. This method has been adopted to a large
extent in those research studies which have investigated the
use of the Rorschach with schiZOphrenics. Generalizations
across all studies are difficult to make but the general
practice has been to ignore traditional classification
systems and to deal with samples of undifferentiated schizo-
phrenics. Some studies have made noteworthy and laudable
attempts to be more specific by further identifying samples
with such clinical designations as chronic, severe, mild, or
ambulatory. For the most part, however, diagnostic terms
are used much less often than is needed for clarity in
designating sample composition.

In general, the thrust of research on the use of the
Rorschach with schiZOphrenics has taken one of two di-
rections. One direction has sought to determine what
Rorschach variables might be found to be characteristic of
schizOphrenics. The other direction has been to investigate
particular Rorschach scores or patterns of scores as they
are hypothesized to be related to some specific aSpect of
schiZOphrenia. One direction has thus mainly been exploratory

while the other direction has been toward specific variables.

Both types of research have the common purpose of
establishing the concurrent validity of the Rorschach as
used with schizophrenia. Consequently, as different types
of studies agree on Rorschach measures of schiZOphrenia,
inferences may be drawn regarding the concurrent validity of

the instrument.

Exploratory Studies ’

Included with the research which has been explora-
tory in focus are those studies which have sought to de-
termine what, if any, scores or patterns of scores might be
typical of schizophrenics.

One of the earliest exploratory studies was carried
out by Rickers-Ovsiankina (1938) who compared 37 hOSpitalized
chronic schizOphrenics with 20 normal control subjects. All
subjects were equated for age, sex and education. The re-
sults of her study showed schiZOphrenics to score higher
than normals on the variables of W, D, Dd, C, CF, A% and
shading determined responses. Schizophrenics scored lower
than normals on F +‘%, and M. The protocols of the schizo-
phrenics contained more fluctuations between good and poor
form perception and showed poorer quality W. As these
differences are interpreted according to the theory under-
lying the Rorschach, schizOphrenics appear to be disturbed
in the areas of thinking, emotional control, and perception

of reality.

Another study comparing schizophrenics and normals
was carried out by Friedman (1952) who differentiated be-
tween hebephrenic and catatonic subgroups within the schizo-
phrenic sample. The findings of the study, however, are
presented for the combined group of schizophrenics, implying
that no differences were found between the subgroups. The
results of the study showed that schizophrenics scored lower
than normals on the Rorschach variables of D, F + % and P.
SchizOphrenics produced significantly more W, DW, and C
responses.

Friedman (1952) also compared the responses of his
schiZOphrenic subjects to the Spiegel norms published by
Beck_gt_§1. (1950). This comparison showed schizOphrenics
to differ from normals in that schi20phrenics scored lower
on the variables of P, FC, M, H, Hd. No differences between
schizophrenic subjects and either the Spiegel norms or the
normal controls were found on the variables of A%, de, Adx,
Sex, or Anatomy. In line with the Rickers-Ovsiankina (1938)
study, the results of Friedman (1952) show schiZOphrenics to
be disturbed in the areas of thinking and affect control.

In addition, Friedman's (1952) results indicate that schizo-
phrenics withdraw from human contact and have little inner
fantasy life. These findings tend to conform to a general
clinical description of schizOphrenia and suggest that the
Rorschach has concurrent validity as a measure of schizo-

phrenic process.

Sherman (1952) used the Rorschach to compare 71
hospitalized, prognostically good schizophrenics with 66
normals on 26 formal and 17 content scores. SchizOphrenics
and normals were found to differ on 20 of the 26 formal
scores and on seven of the 17 content scores. The differ-
ences were significant at the .05 level of confidence and
included the variables of R, F +‘%, P, D, Dd, S, M, FC and
Shading as well as H and A. As part of his design, Sherman
(1952) re-analyzed his data with controls for differences in
education and differences in the number of responses. The
new analysis showed schiZOphrenics to differ significantly
from normals on only 10 formal factors and on only two con-
tent factors when level of education was controlled.
SchizOphrenics differed from normals on 15 formal factors
and four content factors when number of responses was con-
trolled. Since schizophrenics were found to differ from
normals on some of the same variables as in the Rickers-
Ovsiankina (1938) and Friedman (1952) studies, Sherman's
(1952) research provides support for the notion that schizo-
phrenics do produce different patterns of scores than nor-
mals. Sherman's (1952) study is most important, however,
for the focus it places on the need for methodological
controls.

The study which is most devastating to the notion
that it is possible to identify Rorschach scores which are

uniquely characteristic of schizOphrenics is that of Rieman

(1953). The study made two independent comparisons of
normals and ambulatory schizophrenics receiving out-patient
treatment. Significant differences between groups were
found on only five of the 86 variables studied: Average
Form Level Rating, Average Form Level Rating (non-color),

R +‘%, Confabulations, Pathogenic Verbalizations and
Perseverations. Any three of these signs diagnosed as
schiZOphrenic 27% of the normals and 65% of the schizo-
phrenics. All five signs diagnosed as schizophrenic T% of
the normals and 18% of the schizophrenics.

The findings of Rieman's (1953) study clearly call
into question the issue of whether schizophrenics can be
differentiated from normals on the basis of formal scores
since three of the five differences were found on non-formal
factors. Rieman's (1953) study also points out that a great
deal of overlap may exist between the protocols of normals
and schizOphrenics and suggests that the Rorschach is ex—
tremely weak in differentiating schiZOphrenics from normals.
The conclusion is applicable only to ambulatory schizo-
phrenics, however, and in no way applies to hospitalized
schizophrenics or even ambulatory schizophrenics not re-
ceiving treatment.

In spite of the limitations on the generalizations
which may be made from the Rieman (1953) study, it is note-
worthy that ambulatory schizophrenics were found to differ

from normals on the variables of confabulations, pathogenic

10

verbalizations and perseverations. These variables, to-
gether with position responses and contaminations, have been
understood to be highly reliable indicators of schiz0phrenia
(Beck, 1961; Kolly and Klopfer, 1939). A study by KnOpf
(1956), however, casts doubt on the absolute value of these
signs as indicators of schizophrenia since contaminations
and position responses were found to occur among groups of
neurotics, psychopaths and schizophrenics.

The results of the studies reviewed so far present a
great deal of variability in the extent to which they sup-
port the concurrent validity of the Rorschach as a diagnostic
tool for schiZOphrenia. The only conclusion which may be
drawn is that the Rorschach has been shown to produce
different results when used with different samples. To the
extent that the samples were homogeneous across studies, the
conclusion would support the inference that the Rorschach
lacks concurrent validity. In spite of their identification
as schiZOphrenic, it is doubtful that the samples in the
studies reviewed can be considered homogeneous. Descriptions
such as chronic (Rickers-Ovsiankina, 1938), hebephrenic and
catatonic (Friedman, 1952), prognostically good (Sherman,
1952), and ambulatory (Rieman, 1953) imply heterogeneity
among samples and, moreover, point up the extreme limitation
imposed on research by the lack of a reliable and standard-

ized means for identifying criterion groups.

11

An attempt to deal with in-group variance as it may
affect the Rorschach performance of schiZOphrenics was made
by Beck (1954). The procedure in this study was to develop
120 statements describing the psychodynamic behavior of
schizophrenics. Each statement was matched with a Rorschach
score or pattern of scores which was defined as related to
that behavior. Q sorts of the 120 statements were made for
each subject in a sample of 50 schiZOphrenic adults and 60
schizophrenic children.

The frequency with which each of the 120 statements
was used for each subject was ranked from highest to lowest.
A comparison of these orderings resulted in the identifi-
cation of six different types: four adult; one adult-
children; and one children type. On the basis of these
differences it was concluded that six types of schizophrenics
had been distinguished, together with matching patterns of
Rorschach scores.

In a follow-up study Molish and Beck (1958a, 1958b)
added 50 statements to the original 120 statements. Q sorts
were carried out for two samples of hospitalized schizo-
phrenics. Six types were identified in each sample by
factor analyzing the sorts. Patterns of intercorrelation
between these types and the six types found in the study by
Beck (1954) were significant, thereby supporting the as-
sumption that the types were identifiable in samples other

than the original.

12

Even though the studies of Back (1954) and Molish
and Beck (1958a, 1958b) show that it is possible to identify
various types of schizophrenia, these findings have little
practical value for use of the Rorschach itself. The
Rorschach variables associated with each of the types are
neither mutually exclusive nor internally consistent.
Identification of a type depends on the rank ordering of the
Q sort statement rather than the presence or absence of a
Rorschach score in a particular protocol. For example, Q
sort item five related to Rorschach variable "A" may be most
descriptive of one type but least descriptive of another
type. Merely observing Rorschach variable "A" does not,
however, discriminate between types.

Another approach to using the Rorschach to identify
schizOphrenics has been taken in a series of studies carried
out by Piotrowski and his co-workers (Piotrowski, 1955;
Piotrowski and Lewis, 1950; Piotrowski and Berg, 1955). The
focus of this work has been on the Alpha SchiZOphrenic de-
fined by Piotrowski (1945) and elaborated on in other
studies. The dynamics of the Alpha Schizophrenic are based
in a disparity between the amount of available energy and
the degree of control over expression of this energy. Con-
ditions of equal amounts of energy and control are found in
normal states with increases in disequilibrium related to
increases in the severity of pathology. The Alpha Schizo—

phrenic is characterized by an excess of energy over control

13

which is corrected by an increase in control. Thus, the
Alpha SchiZOphrenic appears rigid, well-defended and may be
wrongly diagnosed as non-schizophrenic.

Diagnosis of the Alpha SchiZOphrenic on the Rorschach
is based on a set of five Rorschach variables: Sum W, Sum
C, the ratio of bright shading responses to Sum C, shock re-
actions to dark gray, and F +‘% less than 70%. Scores are
assigned to the values which these variables may take on.
Scores above and below three respectively define schizo-
phrenics and neurotics with 90% accuracy (Piotrowski and
(Lewis, 1950). The magnitude of scores has been found to be
related to follow-up ratings of emotional health (Piotrowski,
1955) and scores increase with increased length of hOSpital—
ization (Abrams, 1964).

The studies which have focused on the use of the
Rorschach to distinguish among various forms of schizo-
phrenia are clear in stressing the importance of patterns
of scores rather than single signs in diagnosing schizo—
phrenia. Even though the most ideal situation would be to
discover one simple and single diagnostic indicator of
schiZOphrenia, reality and the nature of the Rorschach would
indicate that matters are somewhat more complicated. Indeed,
as the Rorschach measures a number of functions and results
in several measures of the same function, the search for
Single signs overlooks the possibility of finding signifi-

cant results .

14

In line with the importance of using sets of scores
rather than single signs, Thiesen (1952) developed five
patterns of scores to be used in identifying schiZOphrenic
protocols. The patterns are as follows. Pattern A: high
Sex and Anatomy scores; Pattern B: low F +‘% and Z scores;
Pattern C: low FC +, M, and A% scores; Pattern D: low
F +‘%, P, and A% scores: Pattern E: high DW and low PC +
scores. High and low scores were defined as either one
sigma's distance above or below the mean of Beck's (1950)
standardization sample.

Thiesen rated the protocols of 60 diagnosed schizo-
phrenics for the occurrence of the above patterns. Using
the norms reported by Beck (1950), a comparison was made be-
tween the schizophrenic subjects and normal controls. The
results showed that significantly more schizophrenic sub-y
jects produced these patterns than normals (.001). 48.4% of
the schizophrenic sample produced one or more patterns, while
only 3.2% of the normals did so. 51.6% of the schiZOphrenic
subjects produced none of the patterns, while the same was
true of 96.8% of the normals. Among the schizOphrenics, the
nest frequent patterns were patterns A and B, while the most
frequent combination of patterns was a combination of
Patterns A, B, C, and D.

Follow-up studies using Thiesen's Patterns have
failed to provide strong support for the original study.

Rubin and Lonstein (1953) analyzed the protocols of 42

15

hospitalized male schizophrenics. 16.7% of the subjects
produced one or more patterns, while 82.8% of the subjects
failed to produce any of the patterns. A comparison between
Thiesen's (1952) normal subjects and the subjects in this
study showed the only significant difference to occur on
Pattern D.

A cross—validation study of the Thiesen Patterns is
also reported by Taulbee and Sisson (1954). These investi-
gators used 62 hospitalized schizophrenics and 157 normal
controls. Significant differences were found between the
groups on the number of subjects producing Patterns A, C,
and D. The study provides practical support for the original
study by Thiesen (1952) and contradicts the work of Rubin
and Lonstein (1953).

A somewhat different emphasis on the use of the
Rorschach with schizophrenics can be found in the work of
Piotrowski and Lewis (1952). These authors were concerned
with using Rorschach scores as prognostic indicators for re-
covery. The Long Term Prognostic Index (LTPI) was developed
on the basis of 15 signs measuring intellectual maturity,
fantasy life control, creative imagination, realistic
planning and interest in others. The LTPI was applied to
the protocols of 100 diagnosed schizophrenics. subjects
were rated as improved or unimproved at a three year follow-
up and once again at a seven year follow-up. High scores on

these variables were found in only 50% of the improved cases.

16

Minor revisions of the-LTPI were carried out by Piotrowski
and Bricklin (1961) and applied to two samples of schizo—
phrenics. Successful predictions regarding improvement at
the end of a three year period were made in 89% of the cases
in one sample and 90% of the cases in the other sample.
These results were supported by still another study carried
out by Piotrowski and Efron (1966).

Stotsky (1952) has found improved schiZOphrenics to
show more fantasy life, emotional control, anxiety, interest
in human contact, and better form quality than unimproved
schizophrenics when compared on the Rorschach. A study by
Goldman (1960) confirms these findings. Remitting cases
were found to score higher on the variables of M, F dominated
responses, Y blends, H, Hd, and F +‘% than unimproved cases.

On the basis of the findings of the studies reviewed
above, it may be concluded that the Rorschach can be used to
predict future status of emotional health. The consistency
of the studies in finding the same variables related to im—
provement is important in providing a practical basis for

making decisions regarding plans for treatment.

Specific Variable Research

The research reported on in this section has focused
on investigating specific Rorschach scores or patterns of
scores as measures of particular types of schiZOphrenic

functioning. Given the number of variables which could be

17

studied, it is noteworthy that most of the research has been
restricted to those variables which are most likely to re-
flect Schizophrenic personality functioning. Some range
does exist, however, in terms of Rorschach variables which
have been studied. Rather than deal with each variable
separately, the research has been grouped in accordance with
whether the variables studied measure the personality

functions of thinking or affect.

Thinking

Disturbances of thought process, one of the prime
characteristics of schizophrenia, is indicated in the
Rorschach by a number of variables such as approach, the use
of popular responses, and form quality.

Using the popular response as a measure of thinking
disorganization, Molish (1951) attempted to differentiate
among normal, neurotic, and schizophrenic subjects. Al-
though a popular reSponse is determined not only by the con—
tent of the response but also by the blot area to which the
response is made, Molish (1951) studied the pOpular response
in terms of both content and location. Two measures were
taken. One measure, called Operation Group, was defined by
'the content of the response regardless of blot area. The
second measure, Operation Stimulus, consisted of the number
of popular responses as determined by both content and blot

area.

18

The results of the study showed normal subjects to
differ significantly from neurotics and schizophrenics on
both Operation Group and Operation Stimulus, that is normals
not only gave more popular responses, as defined by content
alone, than did neurotics or schizophrenics, but also
differed in the sense of giving these responses to the ap-
propriate blot areas. The same relationship held when
neurotics were compared with schizophrenics.

Warner (1951) investigated the popular response
using schizophrenics who were subdivided into groups of
paranoid and non-paranoid schizOphrenics, and neurotics sub-
divided into groups of anxiety and non-anxiety neurotics.

No significant difference was found among the groups on the
number of popular responses given.

The findings of the Warner (1951) study support
those of Molish (1951) so far as schiZOphrenics and neurotics
are concerned. It should be noted, however, that the mean
number of pOpular responses for the schiZOphrenic groups in
the Warner (1951) study was 4.80 responses. This number is
only slightly less than the number expected from normals.
Even though extreme scores within the schiZOphrenic group
may have functioned to elevate the group mean, the finding
that schizophrenics produce a number of popular responses
nearly equal to that of normals makes any conclusion regard-
ing the use of popular responses as a diagnostic indicator

<>f schizophrenic thinking extremely tentative.

l9

McReynolds (1966) has used the Rorschach as a means
for evaluating concept formation in schizophrenics as op-
posed to normals. The procedure used in this study was that
of presenting subjects with each Rorschach card and asking
if the blot could represent some named object. Three scores
are derived: J scores, V scores and C scores. J scores
represent a tendency to agree and V scores represent the de-
gree of deviation from a normative group. C scores repre-
sent the number of correct reSponses. Mean J, V, and C
scores for a group of 125 male schiZOphrenics were 42.33,
42.52, and 37.66 respectively. Compared to a standardized
distribution with a mean of 50, all scores were signifi-
cantly different (p < .001) indicating that schizophrenics
tend to agree and deviate from normals, while obtaining
fewer correct responses.

Although the McReynolds (1966) study did not focus
on specific Rorschach scores, it is important for the use of
the Rorschach with schizophrenics in that it supports the
notion that schizophrenics are uncritical in their thinking
and tend to give easy responses. Lack of critical thinking
and a tendency to give easy responses are two processes
measured by the Rorschach variables of A% and approach with
W'or D emphasized. No studies have been carried out di-
rectly investigating the variable of A%. The exploratory
studies reviewed above have been contradictory regarding A%

as a measure of schizophrenic functioning.

20

With regard to the variable of Approach, Stotsky

(1952) reports on a factor-analytic study of the location
scores in the protocols of 143 diagnosed schizophrenics.
Two factors were isolated, one heavily loaded on D and Dd,
and the other negatively loaded on W. These findings sup—
port the notion that schizophrenics do overuse either the
whole blot area or some detailed portion of the blot.

A study by Watkins and Deabler (1952) investigated
the uncritical aspect of schizophrenic thinking. The study
was based on the assumption that schiZOphrenics form first
impressions on the basis of minimal information and maintain
these impressions even with a later opportunity for more
information. Tachistoscopic presentations of Rorschach
slides were made to 16 chronic schiZOphrenics at speeds in-
creasingly from .001 second to unlimited exposures. More
responses were made as length of presentation was increased
but the content of these responses failed to change, there—
by indicating an inability to shift the content of responses
based on minimal information.

Another approach to the use of the Rorschach as a
measure of thinking disorganization among schiZOphrenics was
carried out by Stotsky and Lawrence (1955). These investi-
gators used two groups of schiZOphrenics pre-experimentally
defined as with and without perceptual impairment. ~Analysis
of the protocols of the two types of subjects showed those

with no impairment to score better in form determined

21

responses, to show more genetically good locations responses,
and to have more shifts in determinants from Free Associ-
ation to Inquiry. More P was noted in the impaired group
than in the unimpaired group. These findings offer strong
support for the assumption that the Rorschach does and can
diagnose disturbances in thinking.

Other Rorschach variables which may indicate think-
ing disorganization are M, R, and tempo. Scott (1956)
scored the Rorschach protocols of 37 schiZOphrenics for
movement responses as defined by both Beck (1961) and
Klopfer (1954). In spite of the fact that Klopfer's (1954)
scoring is broader and, consequently, should produce more
second responses, M scores were found to occur more fre-
quently than PM or M scores. Scott (1956) concludes that M
represents deterioration while FM and m are regressive in
nature. Adams and Cooper (1961) found the number of re-
sponses among a group of 39 schizophrenics to be inversely
related to severity of symptomatology. weiner (1962) de-
veloped four indices of deviant tempo based on the number of
responses to card pairs IV and V, V and VI, VIII and IX, and
IX and x. Significant differences in tempo were found on
the first three pairs when schizOphrenics were compared with
normals. These findings are supported by the follow-up

studies of Weiner (1964b, 1965b).,

22

Affect

Variability of emotional reactivity has long been
regarded as schiZOphrenic symptomatology. Research on the
use of the Rorschach with schizOphrenics has been carried
out to determine if the Rorschach does measure the emotional
reactivity of schizophrenics.

Bucker and Williams (1951) investigated the color re-
sponsiveness of schizophrenics using standard sets of the
five chromatic Rorschach cards, as well as black and white
reproductions. Twenty-one subjects were included in a
counterbalance design. Differences in response total were
controlled by dividing subjects into high and low groups.

No differences were found between experimental conditions
except for the average time of the first response. This
variable increased when color cards were used but decreased
when achromatic cards were used. On the basis of these re-
sults, Bucker and Williams (1951) conclude that color does
not affect the responses of schizophrenics. This con-
clusion, however, is contradictory to what the data indicate.
The median pure C response for the subjects was 2.9. The
usual interpretation is that even one pure C reSponse indi-
cates extreme emotional responsivity. Moreover, increased
reaction time to color cards is one indication of color
shock or responsivity to color.

Crumpton and Goot (1966) studied the effect of color

on the connations of the Rorschach color cards as perceived

23

by schiZOphrenic patients. Fifty male hospitalized schizo-
phrenics rated the five standard chromatic cards and five
achromatic duplicates on a 15 point Semantic Differential
Scale. Analysis of the ratings showed that color cards were
rated higher on the dimensions of evaluation potency and
activity than were their achromatic duplicates. On the
basis of these findings it may be concluded that schizo-
phrenics do respond with greater affect to colored stimuli
than to achromatic stimuli. This conclusion is supported by
the results of Taulbee, Sisson and Gaston (1956) who found
schizophrenics to differ significantly from normals on the
variable of affective ratio.

weiner (1961) investigated the diagnostic efficiency
of three Rorschach color variables: one or two CF; Sum C
between 1.5 and 3; and at least one CF or C response with no
c' responses. Two separate samples were used in the study.
Each was composed of neurotic, character disorder and schizo-
phrenic patients. Analysis of the results showed either one
or none of these signs to be characteristic of the protocols
of neurotics and diagnosed character disorders, while two or
three of the signs were characteristic of the protocols of
schizophrenics. Differences were statistically significant
(p < .001). The results of this study were replicated in a
follow-up study (Weiner, 1964a).

Considerable controversy develOped over weiner's

(1961, 1964a) findings. Klinger and Roth (1964) attempted

24

to replicate the findings regarding Sum C. Their results
showed Sum C in a group of schiZOphrenics to be less than
1.5 responses or more than 3.0 responses, not_Sum C of be-
tween 1.5 and 3.0. Orme (1964) investigated all three color
variables in a study using the same type of subjects as
Weiner (1961, 1964a). Orme found all three signs to be more
common in the protocols of neurotics. No significant
differences were found, however.

In reply to the above studies, Weiner (1965a) points
out that Klinger and Roth (1964) failed to fully report on
their data. Reanalysis, according to weiner, shows that 31%
of 69 schizOphrenic subjects produced between 1.5 and 3.0
responses. In the same article (Weiner, 1965a) critiques
Orme (1964) on the basis of his failure to include the vari—
able of Sum C in the analysis of the data. Reanalysis by
weiner showed schizophrenics to differ from normals on the
color variable at the .07 level of confidence. In the final
reply of this series of exchanges, Orme (1966) points out
that a .07 level of confidence is not significant and that,
furthermore, even with inclusion of the Sum C variable only
52% of the schizophrenics produced protocols with two of the
three signs present as compared to 32%.of the neurotics.

The difference using Chi Square was found to be significant
at the .10 level of confidence.

The importance of these studies rests not so much in

the issue of differences in levels of probability but rather

25

in the findings that schizophrenics do tend to give color

dominated responses.

Conclusions Regarding the Research

Overall, research on the use of the Rorschach has
failed to produce unequivocal evidence supporting either the
concurrent validity or invalidity of the Rorschach as a
diagnostic tool for use with schiz0phrenics. It is only a
slight exaggeration to say that for any study supporting the
use of the Rorschach there is a like study failing to concur.
Taken alone, without regard for content, the pattern of the
research is very interesting. Indeed, it is somewhat akin
to a promising maiden.

While it may be concluded that the inconsistency of
the research findings are due to the Rorschach itself, such
a conclusion is difficult to defend since it implies that
the instrument can be valid and, at the same time, invalid.
Neither can experimental design, nor statistical treatment
of the data, nor even differences in scoring systems fully
account for the variance of the research since there is no
reason to suSpect that these variables would be distributed
in a skewed fashion either favorable or unfavorable to the
Rorschach. It can be assumed, however, that the nature of
the samples Which have been used may account for the incon—

sistent research findings.

26

Indeed, it may not be feasible to attempt to es-
tablish concurrent validity for the Rorschach using such a
gross definition of schizophrenia. As already suggested,
differences in sample composition may be inferred from those
studies which have specified the type of schizophrenics
studied by such terms as chronic, paranoid, ambulatory or
remitted. Since the exact referents for these terms are un-
clear, generalization to other studies is unwarranted. More-
over, neither these descriptions nor the more general label
of schizophrenic provides any basis whatsoever for con-
trolling within group variance.

The variance between samples as well as within
samples on the defining variable of schizophrenia may be
the chief source of difficulty. What is needed is some way
to reduce this sample variance so as to produce more homo-
geneous groups. The process-reactive distinction may serve

this need very well.
_TheProcess-Reactive«Dimension

The process-reactive dimension is a bipolar classifi-
cation system which is defined on a continuum of severity of
illness. Process and reactive schizOphrenics are not de-
fined as dichotomous types but, rather, as continuous enti-
ties. The process—reactive dimension developed out of re-
search on favorable and unfavorable prognostic indicators.

(300d prognosis has been related to a rapid onset of illness

27

with some definite precipitating factor, better socio-
economic adjustment prior to illness, and to marriage. Poor
prognosis has been found to be related to the obverse of
the factors (Hunt and Appel, 1936; Malamud and Render, 1939;
Kant, 1940, 1941, 1944; Chase and Silverman, 1943). In
terms of the process-reactive dimension, those with poor
prognosis are defined as falling into the process range
while those with good prognosis fall into the reactive range.

The development of scales for use in identifying
process and reactive schizophrenics has greatly facilitated
research which has focused on the contrast between process
and reactive subjects. A wide variety of variables have
been studied including factors such as organicity (Brackbill
and Fine, 1956; Tutko and Spence, 1962), reactivity to in—
ternal stimulation (Judson and Katahn, 1963), cognitive
functioning (Elisio, 1963: Strum, 1965), and perceptual
functioning (Bleke, 1955; Zahn, 1959). For the purposes of
this review, the most important comparisons of process and
reactive schiZOphrenics are those which have focused on the
use of the Rorschach.

Process-Reactive Schizophrenics
and the Rorschach

A number of studies have been carried out using the

Rorschach to investigate the hypothesis that process schizo-

phrenics show less mature perceptual development than do

28

reactive schizophrenics. Becker (1956) hypothesized that
process subjects would show less a differentiated and more
global type of perception than reactive subjects. Rorschach
protocols of process and reactive subjects were scored using
Friedman's (1952) scoring system which measures the genetic
level of perceptual development. The hypothesis of the
study was supported and is in agreement with the results of
Fine and Zimet (1959) and Zimet and Fine (1959). A study by
Judson and Martin (1964), however, contradicts the findings
of the Becker (1956) study.

The importance of the process-reactive dimension for
use in the research on the concurrent validity of the
Rorschach is clearly indicated in a study by Kantor and
Herron (1966). The study was designated to test the hypothe-
sis that subjects classed as either process or reactive sub-
jects would vary in terms of severity of pathology. Ratings
of Rorschach protocols confirmed the hypothesis. Pathology
was more severe for process subjects than for reactive sub-
jects. Very curiously, however, the Rorschach protocols of
reactive subjects were judged as neurotic or normal, not as
schizophrenic. Since reactive subjects were found to be
less severely disturbed than process subjects, one might ex-
pect their Rorschach protocols to differ somewhat, but
hardly to the extent that they would be rated as neurotic or
normal since reactive subjects were defined as schiZOphrenic

on the basis of psychiatric evaluation.

29

The study of Kantor, Wallner, and Winder (1953) sup—
ports the finding regarding the diagnosis of reactive sub—
jects. These authors used the Rorschach protocols of pro-
cess and reactive subjects as part of a study designed to
investigate the use of a scale for distinguishing between
process and reactive schizophrenics. The Rorschach proto-
cols of process subjects were judged as psychotic while the
Rorschach protocols of reactive subjects were judged as
non-psychotic.

The findings of these two studies suggest an ex—
planation for the fact that research has failed to provide
conclusive support for the concurrent validity of the
Rorschach as a measure of schiZOphrenia. It may simply be
that reactive subjects have been included in the samples of
the various studies. If reactive subjects are not diagnosed
as schizophrenic on the Rorschach, their inclusion in re-
search samples would act as a suppressor variable and lessen
the extent to which the group as a whole did produce
Rorschach indications of schiZOphrenic functioning. More-
over, as the prOportion of reactive subjects in samples has
differed among studies, the suppression effect would also
differ. Contradictory findings would then be the rule among
studies which have used the Rorschach with schiZOphrenics.
This is, in fact, exactly the case.

The argument above would be important, however, only

in so far as the Rorschach is valid as a measure of any type

30

of schizOphrenic functioning, non-reactive or not. Finding
that reactive subjects are not diagnosed as schizOphrenic on
the Rorschach in no way whatsoever implies that the Rorschach
is valid for other forms of schizophrenia. Inconsistency in
diagnosing subjects from the same population, in fact, sug—
gest that it may be invalid for the entire population. Ac-
cordingly it is important to determine how process and re-
active subjects differ on the Rorschach and to determine
whether the instrument does or does not diagnose reactive
subjects as schizophrenic. Several aspects of the Kantor,
Wallner, and Winder (1953) and Kantor and Herron (1966)
studies suggest that it may.

With regards to the Kantor, Wallner, and Winder
(1953) study, there is some doubt about the adequacy of the
judges who were used to rate the protocols as well as the
judging procedure which was followed. Protocols were
supposedly judged by trainee psychologists "each of whom had
adequate previous training in the administration of the
test" (p. 160). The analysis of the protocols was made with
the assistance of an advanced trainee and "then reviewed by
a staff psychologist who had prior interpretation experience
with more than five hundred Rorschachs" (p. 160). Who
actually made the judgments is unknown. Moreover, the basis
for judging a protocol as non-psychotic or psychotic was not

stated. The basis may have been inadequate particularly with

31

regards to reactive subjects since the greatest amount of
disagreement came in the case of judging reactive subjects.

Kantor and Herron (1966) fail to deal with the find-
ing that reactive subjects are not diagnosed as schizo-
phrenic except to note that " . . . the Rorschach diagnosis
count as neurotic or normal the reactive schizophrenics who
compose about half the sample" (p. 77). Nething further is
mentioned by way of explanation. One is left to assume that
the Rorschach protocols of the reactive subjects were judged
as non schizophrenic without knowing the basis for judgment.
~Since the study used a scale of malignancy to rate each
protocol in relation to the hypothesis, one may assume that
the protocols of the reactive subjects were judged non-
psychotic using this scale. The scale is composed of five
items ranging from "No Schizophrenia“ to "Severe Schizo-
phrenia,“ and is shown below:

1. No Schizophrenia: 2-5 M, M/C equal, FC larger
than CF, A% ranges from 33 to 59, about 5 to 9 P.

2. Mild Schizophrenia: well integrated, adequate
reality contact, may be somewhat flat emotion-
ally, shows neurotic overlay, color shock, low C.

3. Moderate Schizophrenia: some integration of
personality, but distortion in some area such
as thought process, marked anxiety, or lessened
reality contact, low A%, M/C higher on C.

4. Strong Schizophrenia: somewhat disorganized,
poor reality contact, some disruption of thought
proceSses, affectively very flat or labile, may
shOw pure C responses, high A%, more than three
responses, abstract reactions, no M.

32

5. Severe Schizophrenia: completely disorganized,

overtly disturbed, loss of reality contact,

thought processes odd, few responses, low P,

pure C, color naming, perseveration (p. 105).

Inspection of the scale shows that as the items
indicate increasing severity, there is less consideration
given to the scoring system. All five factors in the Item
"No Schizophrenia" are taken entirely from the scoring
summary while only two of the nine factors in the Item
"Severe Schizophrenia" are so derived. No indication is
given as to the basis for judging such factors as "lessened
reality contact“ ("Moderate SchiZOphrenia"), or "completely
disorganized“ ("Severe SchiZOphrenia"). Moreover, no indi-
cation is given as to how contradictions between such
factors as "may show pure C responses" and "affectively very
flat or labile" (“Strong SchiZOphrenia") were resolved.

With reference to the factors which are included in
the Item "No Schizophrenia," it is questionable as to
whether a diagnosis of "No SchizOphrenia" required the
presence of all factors or just some, and if just some, how
many. Reference to Beck's (1945) sample protocols shows
wide variance in terms of M, M:C, FC, CF, A% and P in the
records of both schizophrenics and non schizOphrenics. In-
deed some schiZOphrenics may show some of thedfactors con-
tained in the item "No SchiZOphrenia" while some non
schizophrenics may show an absence of some of the same

factors. Accordingly, to rate a protocol as "No Schizo-

phrenia" on the basis of less than all of the factors

33

contained in this term would risk producing false negatives.
Whether this did occur cannot be ascertained from the study
as reported since the authors fail to describe the content
of the protocols judged schizophrenic or non schiZOphrenic.

One critical variable which may have been left un-
controlled in both the Kantor, Wallner, and Winder (1953)
study and in the Kantor and Herron (1966) study is the vari-
able of time between hospital admission and administration
of the Rorschach. Neither study reports on this variable.
The time lapse between hospital admission and Rorschach ad-
ministration is important since reactive subjects recover
over time. If the Rorschach records of reactive subjects
were taken at a time interval considerably past hospital ad—
mission, the findings that reactive subjects are not diag—
nosed as schizophrenic may actually be based on the protocols
of recovered subjects. Accordingly, the conclusion that
reactive schizOphrenics are not diagnosed as schiZOphrenic
on the Rorschach must be regarded as tentative. It is
questionable as to whether the findings of the Kantor gt 31.
(1953) and the Kantor and Herron (1966) studies came about
because of the invalidity of the Rorschach or because of
different levels of improvement among reactive subjects.

In view of the implications for the concurrent
validity of the Rorschach, the question of whether or not
reactive subjects are diagnosed as schiZOphrenic on the

Rorschach deserves further study. The matter should be

34

approached using a more extensive and explicit basis for
diagnosis, more experienced judges and controls for the ef-
fect of administering the Rorschach at an interval consider—
ably past the time of hospital admission. Measures of im-
provement independent of Rorschach signs themselves should
also be used in order to show a relation between recovery
and differential performance on the Rorschach. This study
was carried out under these conditions and was designed to
investigate the Rorschach diagnosis of process and reactive

schiZOphrenics as it may be affected by recovery.

PURPOSE

Finding that reactive schizophrenics are not diag-
nosed as schizophrenic on the Rorschach implies that the
concurrent validity of Rorschach indices which hypothetical-
ly measure schizOphrenia must be re-evaluated and questioned.
Indeed, the notion that reactive subjects are not diagnosed
as schizophrenic on the Rorschach clearly suggests that the
Rorschach may have little if any validity as a measure of
schiZOphrenic functioning.

The question, of course, is whether or not reactive
subjects can be diagnosed as schizOphrenic when an extensive
set of Rorschach signs indicative of schiZOphrenia is used
as the basis for diagnosis and when the time interval be-
tween hospital admission and adminiStration of the Rorschach
is controlled for the effect of recovery or improvement.

This study was conducted to answer this very
question. Rorschach protocols of process and reactive sub-
jects tested at the time of hospitalization and at a period
of from three to six months after hospitalization were com-
pared with the Rorschach protocols of non schizophrenics on
the basis of a number of Rorschach variables which indicate

schiZOphrenia. It was assumed that all groups of

35

36

schizophrenics would differ significantly from non
schizophrenics on the Rorschach signs thereby providing sup-
port for the concurrent validity of the Rorschach as a
measure of schizophrenic functioning. In order to determine
how differences in the length of hospitalization may affect
the Rorschach performance of process and reactive subjects,
comparisons on the signs were made between groups of process
and reactive subjects tested early in hospitalization and
later in hospitalization.

The Thinking Disorganization and the Withdrawal
Scales of the Psychotic Reaction Profile (Lorr, O'Conner,
and Stafford, 1960) were used as an independent measure of
improvement due to hospitalization. It was assumed that if
differences on the Rorschach signs of schizophrenia were due
to recovery or improvement during hospitalization, this ef-
fect would be observed by comparing the behavior ratings of
process and reactive subjects tested early and later in

hospitalization.

HYPOTHESES

In line with the purpose of this study, the follow-

ing hypotheses were developed.

Hypothesis 1
More hospitalized schizOphrenics show Rorschach

signs of schiZOphrenia than do non schizophrenics.

Hypothesis 2
More reactive schizophrenics tested early in
hospitalization show Rorschach signs of schizophrenia than

do reactive schizophrenics tested later.

Hypothesis 3
There is no difference between reactive and process
schizophrenics tested early with respect to the number of

Rorschach signs of schizophrenia produced.

Hypothesis 4
There is no difference between process schizOphrenics
tested early and later with respect to the number of

Rorschach signs of schizophrenia produced.

37

38

Hypothesis 5
Reactive schizophrenics tested early show more think-
ing disorganization and withdrawal behavior than reactive

schizophrenics tested later.

Hypothesis 6

There is no difference in the levels of thinking
disorganization and withdrawal behavior of reactive schizo-
phrenics tested early and process schiZOphrenics tested

early.

Hypothesis 7

There is no difference in the levels of thinking
disorganization and withdrawal behavior of process schizo-
phrenics tested‘early and process schiZOphrenics tested

later.

METHOD

Description of Subjects

Schizophrenic Sample

Hospitalized schiZOphrenics were used by the experi-
menter to form a pool of potential subjects. Individual
patients were included in the pool on the basis of infor-
mation gathered from clinical files regarding diagnosis, age,
date of hOSpitalization, level of education, and prior
hospitalizations.

Potential subjects were taken from the pool one at a
time and screened by the experimenter in order to ascertain
whether their mental status would permit them to serve as
subjects. It was necessary to reject several potential sub-
jects because they were too disturbed to participate in the
research. Those subjects judged able to serve were invited
to volunteer. The research was explained by the experi-
menter as dealing with recently admitted and long-staying
hospitalized patients. It was stressed that participation
would in no way affect either current or future hospital
status.

Those patients who agreed to serve as subjects were

immediately tested and then classed on the basis of whether

39

40

they were process or reactive schizophrenics as well as ac-
cording to whether they had just recently been hOSpitalized
or whether they had been hospitalized for some time. The
procedure described was followed until four groups of sub-
jects had been formed: reactive subjects tested early in
hospitalization; reactive subjects tested later in hospital-
ization; process subjects tested early in hospitalization;
and process subjects tested later in hOSpitalization. Each
group contained 15 subjects. Each subject carried a current
hospital diagnosis of schiZOphrenia.

Four early reactive and four early process subjects
were drawn from the population of a psychiatric receiving
hospital. All other subjects were drawn from a VA hOSpital
population. The procedure of testing subjects first and
then classifying them necessarily resulted in extra subjects
having been tested. The additional subjects were not in-
cluded with the final sample of the study.

Early reactives and early process subjects had been
hospitalized for a mean of 12.13 days (SD = 6.76) and 10.40
days (SD = 5.14), respectively. The difference between the
number of days the subjects had been hospitalized was not
significant (t = .772, p < .25, 28 df). Late reactive and
late process subjects had been hospitalized for a mean of
5.30 months (SD = 2.38) and 5.35 months (SD = 1.32), re-

spectively. The difference in the length of time these

40

41

subjects had been hospitalized was not significant (t = 0.68,
28 df).

All subjects were male and came from an urban back—
ground. The mean age for schizophrenic subjects was 35.29
years (SD = 8.92). A one-way analysis of variance comparing
the ages of the schizophrenic was not significant (F = 1.662,
p < .25, 3, 56 df). The mean level of education for schizo—
phrenic subjects was 11.05 years (SD = 1.75). A one-way
analysis of variance comparing levels of education was not
significant (F = 1.813, p < .25, 3, 56 df). Further com-
parisons were made on the variables of the number of prior
hospitalizations and length of time between last and cur-
rent hospitalization. No significant differences were found.

Detailed statistics computed on the control vari-

ables are shown in Appendix A.
Non Schizophrenic Controls

In order to test the hypotheses of this study it was
necessary to have a non schiZOphrenic comparison group. To
provide such a group the Rorschach protocols of 15 hospital-
ized, non schizophrenics were taken from the files of the VA
hospital. The diagnostic testing report as well as the
clinical file for each non schizophrenic was checked in
order to insure that they truly were not schizophrenic. ,The

majority carried a diagnosis of anxiety reaction or anxiety

42

neurosis. The diagnosis for each of the control subjects is
shown in the Appendix B.

The mean age for the non schiZOphrenic sample was
34.66 years (SD = 5.86). A one-way analysis of variance of
the ages of the combined sample of schizophrenics and non
schizOphrenics was not significant (F = 1.410, 4, 7O df).

The mean level of education for the non schizOphrenic sample
was 10.00 years (SD = 1.41). A one-way analysis of variance
of the levels of education for the combined sample of schizo—
phrenics and non schizophrenics was not significant (F =
2.317, p < .10, 4, 70 df).

Further comparisons using the combined sample of
schizophrenics and non schizOphrenics were carried out on
the variables of number of prior hOSpitalizations, length of
prior hospitalizations, and length of time between last and
most recent hospitalization. No significant differences
were found except for the variable of length of prior
hospitalizations (F = 3.695, p < .05, 4, 45 df). Schizo-
phrenics were found to have had longer prior hospitalizations
than non schizophrenics. More complete statistics for these

variables have been included in Appendix A.
TechniguesRandjgnstruments
The Rorschach

The Rorschach was administered to the subjects of

this study by the experimenter. In all cases the Rorschach

43

was given prior to rating the subjects as process or re—
active schizophrenic.

Very early in the data collection period the ex-
perimenter noted that many subjects became increasingly
anxious as they progressed through the series of cards, pre—
sumably in response to the thoughts or feelings which were
aroused by the stimulus configurations on the cards. Due to
the nature of the sample, the experimenter questioned what
impact the testing session might have on some subjects. It
seemed advisable to devise some procedure for estimating
this impact so that steps could be taken to avoid doing harm
to the subjects.

The potent capacity of the Rorschach to upset a pre-
cariously balanced adjustment was clearly indicated by one
incident with a particular subject. The subject in mind re-
sponded to the first cards of the series quickly, easily and
in a flippant manner. Card IV, however, elicited the re—
Sponse "My mother." The subject's responses became in-
creasingly pathogenic in content and were accompanied by an
increase in the subject's tension level. Areas D1, D13, and
D15 on Card x were "things coming towards me, trying to get
me." At the beginning of the Inquiry the subject threw down
Card x and began to sob. At this point the session was
aborted and the experimenter focused on calming the subject.
An offered box of kleenex was forcibly knocked from the ex—

perimenter's hand and the subject began hitting the wall

44

with his knuckles. The incident finally abated, but only
after the subject had broken an ear plug for a transistor
radio by whipping it against a brick wall and smeared him-
self with paint from a nearby working area.

Although the incident described was the only one of
its kind which occurred during the data collection period,
it did emphasize the issue of potentially dangerous impact.
Indeed, the matter was one of subjecting patients to a situ-
ation which could be stressful and which might be harmful to
them.

To forestall any damaging effect, the experimenter
adopted the procedure of inquiring regarding a subject's
willingness to complete the testing session when increased
tension was noted and continued without relief. In some
cases the procedure acted as a tension reducer and allowed
the testing session to be completed. Only one subject re-

quested that the testing be terminated.
The Premorbid History Scale

The schizophrenic subjects in the study were defined
as either process or reactive schiZOphrenic on the basis of
scores obtained by rating each subject on the Premorbid
History Scale (PMHS) of the Prognostic Rating Scale
(Phillips, 1953). The six categories of the PMHS are shown
below. Depending on the age of the subject, either category

C or Category D is used.

45

A. Recent sexual adjustment
B. Social aSpects of sexual life during adolescence

C. Social aSpects of recent sexual life: 30 years of
age and above

D. Social aspects of recent sexual life: below 30
years of age

E. Personal relations: history
F. Recent premorbid adjustment in personal relations
Each category contains from five to seven items ar-
ranged in step fashion from behaviors characteristic of re-
active schizophrenics to behaviors characteristic of process
schiZOphrenics. Subjects are rated on only one item in each
category. Each item carries a value of from zero to six. A
total score is obtained by summing the values of the items
which are descriptive of a subject. Minimum and maximum
scores are 0.and 30. In the original study (Phillips, 1953),
subjects scoring from 0 to 15 were classed as reactive schizo—
phrenics and subjects scoring from 16 to 30 were classed as
process schizophrenics. The same cut—off points were used
here. The PMHS is shown in Appendix c.
The mean ratings on the PMHS for early reactive and
late reactive subjects were 8.60 (SD = 3.720) and 10.00
(SD = 2.309), respectively. The difference between these
means was not significant (t = 1.19633, p < .15, 28 df).
The mean ratings for early process and late process subjects
were 21.86 (SD = 3.597) and 22.26 (SD = 3.419), respectively.

The difference between these means was not significant

46

(t = .301, p < .40, 28 df). All cross comparisons by time
and process-reactive status were significant (p < .001).

Ratings on the PMHS have most often been reported as
based in social history data taken from clinical files. It
was impossible, however, to follow the procedure of using
social history data to rate the subjects in the present
study. Social history data was unavailable for some sub-
jects due to recency of hOSpitalization. For other subjects,
the only social history data available was that which had
been gathered at the time of a previous hospitalization.
This data contained no information about recent sexual, so-
cial or premorbid adjustment. In still other cases, social
history data was current but incomplete in terms of the
information required for the ratings.

As an alternative to using social history data, the
experimenter chose to base the Premorbid History Scale
ratings on an interview with each subject. These interview
based ratings were checked against social history data when
it was available and useable. In order to carry out the
interviews, a series of questions were developed based on
the content of the items in each category of the PHMS. The
questions were designed in such a way that, except for the
first question, the focus of any question was determined by
the immediately preceeding answer. Accordingly, questions
could be asked with increasing focus until the amount of

information necessary for rating an item in each category

47

had been obtained. For example, with Category A (Recent
Sexual Adjustment), the answer to an initial question re-
garding marital status would determine whether the second
question focused on the information required by Item 2, Item
3, or Item 4b. The answer to the second question would de-
termine whether the third question should focus on Item 3,
Item 4a, or Item 5a, and so on.

All interviews were carried out by the experimenter.
The ratings were completed as the interview progressed. All
interviews were carried out after the Rorschach had been ad—
ministered in order to control for possible bias due to

knowing the process-reactive classification of subjects.
The Ullmann-Giovannoni Scale

A second measure used in the study to define sub-
jects as process or reactive schizophrenics was the Ullmann-
Giovannoni Scale (UGS) (Ullmann and Giovannoni, 1964). The
UGS, shown in Appendix D, was included in the study to serve
as a partial check on the procedure followed in rating sub-
jects on the PMHS. Subjects were rated on the UGS immediate—
1y after ratings on the PMHS had been obtained.

The UGS consists of 24 items answered true or false.
Item content deals with marital status, stress factors, work
history, sexual relations, social relations and prior

hospitalization. The ratings were made by the experimenter

48

who read the items to the subjects and recorded the affirma-
tive or negative responses.

The original study (Ullmann and Giovannoni, 1964)
reports the direction in which 638 reactive subjects were
found to answer each item. A subject"s score is obtained by
summing the number of items answered in the reactive di-
rection. Minimum and maximum scores are zero and 24, with
high scores defining reactive subjects. The original study
provides no information regarding the cut-off points for
process and reactive schizophrenics, or for the amount of
overlap which may be expected. For the present study, pro-
cess subjects were defined by scores of from 0-11. Re-
active subjects were defined by scores of from 14—24. An
indeterminate range was defined by scores of 11, 12, and 13.

Using the cut-off scores, 20 subjects in the present
study were defined as reactive schizophrenic and 22 subjects
were defined as process subjects. The mean scores for the
reactive and process subjects were 15.757 (SD = 1.568) and
7.875 (SD = 2.108), respectively. The mean scores for 11
reactive subjects tested early and 9 reactive subjects
tested late were 16.182 (SD = 1.466) and 15.333 (SD = 1.564).

The difference between the mean scores was not significant

(t 1.251, p < .15, 18 df). The mean scores for 10 process
subjects tested early and 12 process subjects tested late
were 7.500 (SD = 2.617) and 8.250 (SD = 1.479), respectively.

The difference between these mean scores was not significant

49

(t = .8648, p < .250, 10 df). Eighteen subjects fell into
the indeterminate group with a mean score of 11.777
(SD = 1.315).

While the scores of the process and reactive sub-
jects are in accord with what might be expected on the basis
of the original study, the main question is that of the re-
lationship between the UGS and PMHS scores. Reactive sub—
jects are defined by high scores on the UGS and low scores
on the PMHS. The converse is true for process subjects.
Negative correlations are thus expected if the instruments
are in basic agreement regarding classification of subjects.
The correlation between scores on the UGS and the PMHS for
those subjects defined as process or reactive schizophrenic
by the UGS was -.899 (p < .001, 40 df). For those subjects
defined as reactive schizophrenic, the correlation between
the scores on the two instruments was -.299 (NS, 18 df).
The correlation between the scores on the UGS and PMHS for
subjects defined as process schizophrenic by the UGS was
-.584 (p < .005, 20 df). The same relation of significant
and non-significant correlations was observed for subjects
defined as reactive schizophrenic and tested early (r =
-.252, NS, 9 df) or late (r = .342, NS, 7 df), and for sub-
jects defined as process schizophrenics and tested early
(r = —.586, p < .05, 8 df) or late (r = —.638, p < .025, 10

df).

50

The correlation of the scores on the UGS and PMHS
for those defined as either process or reactive schizo-
phrenic by the UGS would indicate a high degree of agreement
between the two instruments. Johnson and Ries (1967) com—
pared PMHS scores and UGS scores for two samples of process
and reactive subjects. The correlations between scores were
-.75 (n = 50, p < .01) for one sample and -.58 (N = 41,

p < .01) for the other. Comparing sample sizes, the corre—
lation of -.899 obtained in this study is considerably
higher than those obtained in the Johnson and Ries (1967)
study.

The relation between the PMHS and UGS would appear
to hold only for process subjects, however, since the corre-
lation between the PMHS scores and UGS scores for subjects
defined as reactive schizophrenics by the UGS, although in
the expected direction, failed to reach significance. No
comparisons between these results and those of the Johnson
and Ries (1967) study are possible since the previous re-
searchers failed to report on the correlations between the
PMHS and UGS by subject class.

The low correlation between the PMHS scores and the
UGS scores for reactive subjects as defined by the UGS may
be attributed to the fact that, in this study, more reactive
subjects than process subjects scored closer to the cut-off
point for process and reactive subjects on the PMHS.

Thirteen reactive subjects scored in the 8-15 range on the

51

PMHS, while 8 process subjects scored in the 16-22 range on
the PMHS. The mean PMHS scores of these subjects were 11.00
and 19.00 respectively, indicating that they were within the
acceptable reactive range and process range on the PMHS.
The effect of the scores of the 13 reactive subjects on the
UGS was to reduce the correlations between the PMHS and UGS.

In spite of the fact that the UGS defined fewer sub—
jects as process or reactive schizophrenic than the PMHS,
the PMHS ratings were used as the basis for identifying pro-
cess and reactive subjects in this study. The PMHS is a
widely used instrument and has been shown to correlate high-
ly with other standard measures of the process-reactive
continuum (Garfield and Sundland, 1966). Moreover, what
actually constitutes a reactive score on the UGS needs
further research, as does the issue of whether or not pro-
cess schizophrenics can be identified by a set of questions
which are answered in an opposite direction to the way in
which reactive schizophrenics answer.

The Thinking Disorganization and
Withdrawal Scales

The Thinking Disorganization and Withdrawal Scales
of the Psychotic Reaction Profile (PRP) were used in this
study as a measure of the emotional health of the subjects.
The PRP was developed by Lorr, O'Conner, and Stafford (1960)

for use by psychiatric attendants and ward nurses in rating

52

patient behavior. Minimum observation time necessary for
use of the scales is three days. Reliability for the Think-
ing Disorganization Scale and the Withdrawal Scale has been
reported as .90 and .94 respectively (Lorr, O'Conner, and
Stafford, 1960).

The Thinking Disorganization Scale consists of 18
items which are descriptive of disorientation, irrelevant
speech, hallucinations, and peculiar movements. The With-
drawal Scale consists of 38 items dealing with activity
level, interest in the environment, and interpersonal re-
lations. Each item is answered True or Not True and scores
are derived by summing the number of items which are rated
in a direction indicative of thinking disorganization and
withdrawal.

In the present study, items from the Thinking Dis-
organization Scale and Withdrawal Scale were randomly com-
bined into one composite scale. Also included in the
questionnaire was an item regarding the procedure used in
completing the questionnaire and an item asking about the
length of time the rater had known the subject.

Instructions for the use of the Thinking Disorgani-
zation and Withdrawal Scales were printed at the top of each
rating sheet used by the ward personnel. The instructions
read as follows:

Listed below are 56 items which describe types

of behavior which you have probably seen before in
working on psychiatric wards. Keeping in mind the

53

patient whose name is shown above, answer each item

according to whether it is true or not true of him.

Cross out (T) if it is true--cross out (NT) if it is
not true.

This will take about 10 minutes. Work quickly.

Put down your first impression, it's likely to be
most accurate. Don't change your answers. And
Thank you very much for your help.

The experimenter interviewed each rater before the
ratings were carried out in order to determine whether the
instructions were understood and whether the raters knew the
subjects well enough to rate them. Due to the demands on
staff time, it was impossible to give the raters training
with the scales.

As a basis for estimating the reliability of the
ratings, 23 subjects were rated by two different raters.

The product-moment correlation for the ratings on the com—
bined Thinking Disorganization and Withdrawal Scale was

.635, significant at the .005 level of confidence (44 df).
The correlation for the ratings on the Thinking Disorgani-
zation Scale was .505, significant at the .01 level of confi-
dence (21 df). The correlation for the ratings on the With-
drawal Scale was .252 and not significant (21 df).

The correlation between the length of time the
raters knew the subjects and the ratings on the Thinking
Disorganization Scale and Withdrawal Scale were not signifi-
cant. Additional correlations between ratings for subjects

by groups are shown in Appendix F. The Thinking Disorgani-

zation and Withdrawal Scale is shown in Appendix E.

54

Scoring of the Instruments

 

The Rorschach: Reliability Training

Two Ph.D. clinical psychologists were used to score
the Rorschach protocols in this study. Each judge had com—
pleted course work on the Rorschach and had also used the
Rorschach in a professional capacity. In Spite of this ex-
perience, both judges were given intensive training before
scoring the actual data of the study.

The judges were trained to score according to Beck's
(1961) system. They were also trained to score 11 variables
not formally included in the system: Card Edging, Color-
naming, Clang associations, Contaminations, Cosmic Themes,
Excessive Card Turning, Pathogenic Verbalizations, Personal
Intrusions, Perseverations, Rejections, and Vivid De-
scriptions. As part of the training each judge scored 11
Rorschach protocols of schiZOphrenics and non schiZOphrenics.
The protocols were taken from a different project and were
not a part of the data of this study. Both judges scored
the same protocols. The training protocols were grouped in
four sets of 3, 3, 2, and 3 protocols each. After scoring
each set, the judges met and discussed the scoring pro-
cedures. Correlation data computed on the scoring of the
two judges were presented during these discussions and re-
viewed as the data indicated points of agreement and dis-

agreement. The data proved invaluable as the correlations

55

quickly indicated that the largest amount of disagreement oc-
curred in scoring for Pathogenic Verbalizations, Personal
Intrusions, Contaminations, and Confabulations. Extensive
effort was made to define these variables in a rigorous and
meaningful way so that the exact definitions could be used
by the judges in judging the scoring task.

The correlation between the judgments for all vari—
ables scored on the last five training protocols ranged from
+.75 to +1.00, indicating a satisfactory degree of

reliability.

The Rorschach: Scoring the Protocols

The protocols of the non schizophrenic sample were
recopied to make them indistinguishable from the schizo-
phrenic protocols. All protocols were coded and scored ac-
cording to Beck's (1961) system and for the 11 variables
listed above. In addition, the judges were asked to make
global ratings for each protocol on a 6-point scale of
Emotional Health with 1 indicating high health and 6 indi-
cating high pathology. The judges were instructed to make
their ratings on the basis of their clinical feelings regard—
ing the protocols. No attempt was made to define emotional
health and no practice was given in making the ratings.

Thirty protocols were randomly selected from the
total 75 and designated as Reliability Protocols. In turn,

these 30 protocols were randomly assigned to three

56

Reliability sets: Set R1, Set R2, and Set R3. Each set con-
tained 10 protocols. The remaining 45 protocols were desig-
nated as Non-Reliability Protocols and randomly assigned to
six Non-Reliability sets: Set NR1, Set NR2, Set NR3, Set
NR4, Set NR5, and Set NR6. Sets NR5 and NR6 contained two
and three protocols, respectively. All other Non-Reliability
sets contained 10 protocols.

The Reliability (R) and Non-Reliability (NR) sets
were arranged in alternate fashion over a series of six
trials. Odd numbered trials were designated Reliability
Trials and even numbered trials were designated Non-
Reliability Trials. Both judges scored the same set of
protocols on Reliability Trials. Each judge scored differ-
ent sets of protocols on Non-Reliability Trials. The proto-
cols were scored as follows: Set R1 was scored on Relia-
bility Trial 1; Sets NR1 and NR2 were scored on Non-
Reliability Trial 2; Set R2 was scored on Reliability Trial
3; Sets NR3 and NR4 were scored on Non-Reliability Trial 4;
Set R3 was scored on Reliability Trial 5; Sets NR5 and NR6
were scored on Non-Reliability Trial 6.

Following each Reliability Trial and before the next
Non—Reliability Trial, product-moment correlations were com-
puted for each variable scored in the reliability protocols.
Accordingly, a continuous check was made on the judges'
scoring. This procedure also permitted the experimenter to

provide the judges with feedback information about the

57

scoring and to pinpoint those variables which required
greater scoring accuracy.

The correlation between the judges' scoring of FC.
FV, FT, Contaminations and Pathogenic Verbalizations fell
below the .05 level of significance on Reliability Trial 1.
For Reliability Trial 2 the correlations for CF, 3% and
Perseverations were below a .05 level of significance. All
other variables were significant at a level of .05 or better
on both Reliability Trials 1 and 2.

In computing the reliability of the judges' scoring
of the Rorschach variables, it was difficult to interpret
the meaning of the correlations which resulted from two
types of patterns which the scores took on. The corre-
lations resulting from the patterns referred to were termed
Indeterminate Correlations. Correlations resulting from all
other patterns were termed Determinate Correlations. The
first pattern was that of perfect agreement between the
judges regarding the absence of some variable, that is where
O—A/O-B scores were observed for some variable in all proto-
cols in a reliability set, and where O—A and O-B are defined
as zero scores assigned by Judge A and Judge B. The formula

for the product-moment correlation in this case reduced to

0 .

0

an impossible mathematical Operation since a division by

zero is not permitted. The most feasible method for

58

interpreting this case was to define the correlation coef-
ficient as Indeterminable but to report it as +1.00, thus
reflecting the important fact that both judges did agree on
the absence of the variable.

The second pattern occurred when both judges agreed
on the absence of some Rorschach variable in all but one
protocol in a reliability set, and when the variable for
this protocol was given a numerical score by one judge but
a zero score by the other judge. That is, where some vari-
able for 9 protocols was scored O-A/O-B, but X-A/O—B, or
O—A/X-B for 1 protocol, where O—A, O—B are defined as zero
scores assigned by Judge A and Judge B, and where X—A and
x-B are defined as numerical scores assigned by Judge A or
Judge B. The pattern described above resulted in a corre—
lation of 0.00 which was defined as Indeterminate and re-
ported as 0.00. It should be noted, however, that the corre-
lation failed to reflect the agreement of the judges in 9
out of 10 cases regarding the absence of the variable. As
the number of O-A/O-B scores decreased to 8 and as the
number of XeA/O—B, or O—A/XPB scores increased to 2, the
sign of the 0.00 correlation became negative. As the ratio
of O—A/O-B scores to x-A/O—B scores (or O-A/XPB scores) in-
creased beyond 8:2, the correlation took on a negative
numerical value which reflected the difference in the judges'

scoring.

59

Product—moment correlations were used to compute the
final reliability of the judges' scoring of the 30 relia-
bility protocols. Following Guilford's procedure (1965,
pp. 348-349), correlation coefficients for each variable
were transformed to z' scores which, weighted by the degrees
of freedom,were summed and divided by the total degrees of
freedom. The resulting z' score was transformed back to r,
the over-all correlation coefficient.

Combining Determinate and Indeterminate correlations
resulted in an over-all correlation coefficient of .921
(p < .001, 49 df). For Determinate correlations, an over-
all correlation coefficient of .920 was obtained (p < .001,
33 df). The over-all correlation coefficient for Indetermi-
nate correlations was .928 (p < .001, 14 df). The corre-
lations for each variable scored by the judges are included
in Appendix G.

The correlation for the judges ratings of Emotional

Health was .546 (p < .001, 28 df).
Treatment of the Data
The Rorschach

Four hypotheses of this study were concerned with
Rorschach signs of schizophrenia (RSS). A total of 39 such
signs were drawn from Beck (1945, 1954), Goldman (1960),
Kelly and Klopfer (1939), and Piotrowski (1952) for use in

this study. The RSS are shown in Appendix H.

60

In order to test the hypotheses it was first neces-
sary to determine the number of RSS in each protocol. For
the reliability protocols the mean score of Judge A and
Judge B for each variable was used. The Median Test was
used for each hypothesis. Yates Correction was applied in
the Chi Square tests which were carried out in order to test
for the significance of the difference in the number of sub-
jects in the comparison groups falling above and below the
common median. Fisher’s Exact Test was substituted for Chi
Square when the expected frequency was less than five.

Additional analyses was made for each RSS on the
basis of the number of subjects in the comparison groups
showing the presence or absence of each RSS. Chi Square
with Yates Correction was used in these analyses. Fisher's
Exact Test was substituted for Chi Square when the expected
frequency was less than five.

The Thinking Disorganization and
Withdrawal Scales

Three hypotheses of this study were concerned with
the level of adjustment of the subjects as measured by the
Thinking Disorganization and Withdrawal Scales. Level of
adjustment as measured by these scales was defined as the
number of items rated in a direction indicating thinking
disorganization or withdrawal behavior. The mean score of
the raters was used for those subjects on which the relia-

bility of the ratings was based.

61

In order to test the hypotheses, scores on the
Thinking Disorganization Scale were analyzed separately from
the scores on the Withdrawal Scale. The analysis for each
scale was carried out in two steps. The first step con-
sisted of using a 2 X 2 analysis of variance to compare the
scores of the subjects classed as process or reactive,
tested early or late. The second step consisted of com—
paring the mean scores of each of the groups using the

Newman-Keuls procedure.

RESULTS

Hypothesis 1: More hospitalized schizophrenics show
Rorschach signs of schizOphrenia than do non schiZOphrenics.

The first hypothesis was based on the notion that
while non schiZOphrenics might produce protocols which con-
tained Rorschach signs of schizophrenia (RSS), more schizo-
phrenics would produce protocols which contained RSS. In-
deed, this would be expected if the Rorschach does diagnose
schizophrenia. Even though one of the major variables of
interest in this study was the process-reactive dimension,
it was decided to compare non schizophrenics with the entire
group of schizophrenics as well as with each process and
reactive group.

Accordingly, in testing Hypothesis 1, the first com-
parison was made on the basis of the number of SChiZOphreniCS
and non schizOphrenics scoring above and below the median
RSS for the combined schizOphrenic and non schiZOphrenic
group. The median RSS and the Chi Square value for this
comparison are shown in Table 1. With 1 df, a Chi Square
> 0.9744 is clearly not significant.

A second set of comparisons was made in which each

process and each reactive group was paired with the

62

63

Table 1. Median number of RSS for each group comparison
with Chi Squares for the number of subjects
scoring above and below the median RSS. The
greater number of subjects showing the RSS is

underlined.

 

 

Group Comparison

Median

Chi Square* p

+

 

 

NOn SchiZOphrenic-Combined_§roce§§

 

 

 

 

 

 

 

 

 

 

 

 

and Reactive 9.64 0.9744 50
Non SchizOphrenic—Early Reactive 9.50 0.5333 .50
Non SchiZOphrenic-Late Reactive 8.50 0.0000 0.00
Non SchiZOphrenic-Early Process 10.00 3.3333 .10
Non Schizophrenic-Late Process 8.90 0.5430 .50
Early Reactive-Late Reactive 9.50 0.5333 .50
Early Reactive-Early Process 11.50 0.0000 0.00
Early Reactive—Late Process 9.50 0.5333 .50
Late Reactive-Early Process 10.50 0.5333 .50
Late ReactivejLate Process 9.00 0.000 0.00
_§ar1y Process-Late Process 10.00 3.333 .10

*1 df
+p values are for two tailed tests. For one tailed

tests the p values are halved.

64

non schizophrenic group. The median RSS and the Chi Square
values for each of these comparisons is shown in Table 1.
The only comparison which approached significance was the
non schizophrenic early process comparison. With 1 df, p
(Chi Square 2 3.333) = .10*. These results do not support
Hypothesis 1. Rorschach signs of schiZOphrenia were not
shown by more schizophrenics than non schiZOphrenics. The
number of early reactive, late reactive, and late process
subjects showing Rorschach signs of schizophrenia was not
significantly different from the number of non-schiZOphrenic
subjects. There was a trend, however, for a significantly
greater number of early process subjects than non

schizophrenics to show Rorschach signs of schiZOphrenia.

Hypothesis 2: More reactive schizOphrenics tested
early in hospitalization show Rorschach signs of schizo-
phrenia than do reactive schiZOphrenics tested later.

The second hypothesis, the central and the most im-
portant hypothesis of the study, is based directly on the
notion that reactive schizOphrenics do recover. Accordingly,
it was assumed that the number of reactive schiZOphrenics
producing RSS would.be a function of recovery, where re-
covery increases over time and with lengthened hospitali-

zation. It was furthermore assumed that the effect of

 

*A two tailed test was used in these comparisons.
One tailed tests were also carried out and resulted in a
significant difference for the non schiZOphrenic--ear1y
process comparison (p < .05).

65

recovery on the production of RSS would be most clearly
demonstrated by comparing reactive subjects tested early
with reactive subjects tested later in hOSpitalization.

The median RSS for the combined group of early and
late tested reactive subjects is shown in Table 1 together
with the Chi Square value for this comparison. With 1 df,
the Chi Square value is not significant. Accordingly,
Hypothesis 2 was not supported. There is no significant
difference between the number of reactive subjects tested
early and later in hospitalization who show Rorschach signs

of schizOphrenia.

Hypothe§i§_3: There is no difference between re-
active and process schizophrenics tested early with respect
to the number of Rorschach signs of schizophrenia produced.

The third hypothesis focuses on the issue of whether
process schizophrenics are in fact more severly disturbed
than reactive schizophrenics. An important factor in making
such a comparison would seem to be the period at which pro-
cess and reactive schizophrenics are compared relative to
the point of recovery for reactive schiZOphrenics. It is
reasonable to expect more similarity in degree of disturbance
when process and reactive subjects are compared early in
hospitalization than when compared later in hOSpitalization.
The very fact of recent hOSpitalization would indicate a

level of disturbance unaffected by the variable of recovery.

66

The results of testing Hypothesis 3 are shown in
Table 1. With a common median of 11.50 RSS for the combined
group of early process and early reactive subjects, the Chi
Square value was not significant. Accordingly, Hypothesis 3
was supported. The number of early tested reactive subjects
showing Rorschach signs of schizophrenia is not significantly

different from the number of early tested process subjects.

Hypothesis 4: There is no difference between process
schizophrenics tested early and later with respect to the
number of Rorschach signs of schizOphrenia produced.

The assumption underlying the fourth hypothesis was
that process schizophrenics, contrary to reactive schizo—
phrenics may never improve to any substantial degree. If
this assumption is correct, then there would be no difference
between process subjects tested at different points in the
period of hOSpitalization with regard to Rorschach
performance.

The results of testing Hypothesis 4 are shown in
Table 1. With a common median of 10.00 RSS, the comparison
between the number of early and late tested process subjects
scoring above and below the median resulted in a Chi Square
of 3.333. With 1 df, p (Chi Square 2 3.333) = .10. On the
basis of this result Hypothesis 4 was supported. There was,
however, a trend for more early process than late process

subjects to show Rorschach signs of schiZOphrenia.

67

Hypothesis 5: Reactive schizophrenics tested early
show more thinking disorganization and withdrawal behavior

than reactive schizophrenics tested later.

Hypothesis 6: There is no difference in the levels
of thinking disorganization and withdrawal behavior of re-
active schiZOphrenics tested early and process schizophrenics

tested early.

Hypothesi§_z: There is no difference in the levels
of thinking disorganization and withdrawal behavior of pro—
cess schiZOphrenics tested early and process schiZOphrenics
tested later.

Although the fifth, sixth, and seventh hypotheses of
this study were stated individually, all three hypotheses
rest on the assumption that process and reactive schizo-
phrenics differ in potential for recovery so that, over
time, recovery increases for reactive schizophrenics but is
invariant for process schizOphrenics. Implicit in this as-
sumption is the notion that there is a maximum level of re—
covery for process schiZOphrenics which is less than the
potential level of recovery for reactive schiZOphrenics. It
was not assumed, however, that the same relation would hold
true for degree of pathology. That is, process and reactive
subjects should not differ in potential for emotional
disturbance. If the assumption is correct, then different

levels of adjustment would be characteristic of process and

68

reactive subjects, depending on the point in hospitalization
at which they were tested. These observed differences

would explain differences as well as similarities in the
Rorschach protocols of process and reactive subjects tested
early and late.

Table 2 shows the mean scores and standard devi-
ations of early and late tested process and reactive sub-
jects on the Thinking Disorganization and Withdrawal Scales.
Lorr, O'Conner, and Stafford (1960) report the mean scores
of open and closed ward patients on the Thinking Disorgani-
zation Scale as 2.66 and 4.16, respectively. The mean
scores of Open and closed ward patients on the Withdrawal
Scale were reported as 14.35 and 17.13, respectively. The
Thinking Disorganization Scale scores of the late tested
subjects in the present study are most like the scores of
open ward patients, while the scores of the early tested
subjects are most like the score of closed ward patients.
The Withdrawal Scale scores of the reactive subjects in the
present study are most like those of open ward patients
while the scores of the process subjects are most like those
of closed ward patients.

Hypotheses 5, 6, and 7 were tested together and in
two steps. In the first step a 2 X 2 analysis of variance
was carried out comparing the scores of the early and late
tested process and reactive subjects on the Thinking Dis-
organization Scale. A summary of the analysis of variance

is shown in Table 3.

69

 

 

 

 

 

mom.m ¢N.m mm.h H¢.m mm.m mv.m om.s mm.> Om
mmo.ma mmm.ma omo.HH oao.ma oom.ma mmn.ma oom.oa mmv.NH new:
oamom Hmsmuccuez
Hm.m ehm.m .mh.¢ m¢.v vm.m mo.m mm.m m¢.m Om
oon.m ooo.m oom.m 00¢.v mmm.m oom.m ooa.m mm¢.v coo:
coflumuflcmmHOmHQ
assesses
pounce pounce o>euomom mmooonm (mmoooum mmoooum o>euomom o>euomom
mung manna Ham Had mama maumm oumq maumm
Hafi Ham
.moamom Hmsmucnuez can coHumNHcMmHOmwo mewxceca map :0 muomflnsm o>euom
Ion

can mmoooum coummu mama can aaumo mo mcoHume>op cumocmum can memos 029 .N wanna

70

Table 3. Summary table of the 2 X 2 ANOVA for early and late
tested process and reactive subjects on the Think-
ing Disorganization Scale.

 

 

 

Source of Variation Sum of Squares df Mean Square , F
Process - Reactive 79.35 1 79.35 6.419b
Early - Late 18.15 1 18.15 1.468
Interaction 493.45 1 493.45 39.923a
Error 692.70 56 12.36

Total 1283.65

 

ap < .01

bp < .05

The significant F for the process-reactive effect
and the interaction effect indicates that the main de-
terminants of scores on the Thinking Disorganization Scale
were the process—reactive classification in interaction with
the early-late variable. As a main effect, being tested
early or late had no effect on the scores of the process or
reactive subjects.

In order to determine which of the groups con-
tributed to the results of the analysis of variance, the
means of the early and late tested process and reactive sub-
jects were compared using the Newman—Keuls procedure (Winer,
1962, pp. 77-85). The results of these comparisons showed

the only significant difference to occur in the early

71

process - late reactive comparison (p < .05). The pre-
diction made in Hypothesis 5 was not supported. Early re-

active and late reactive subjects did not differ in terms of

thinking disorganization. Hypotheses 6 and 7 were supported.

Early reactive subjects and early process subjects, as well
as early process and late process subjects did not differ in
degree of thinking disorganization.

In the second step of testing Hypotheses 5, 6, and
7, a 2 X 2 analysis of variance was carried out comparing
the scores of the early and late tested process and reactive
subjects on the Withdrawal Scale. A summary of the analysis

of variance is shown in Table 4.

Table 4. Summary table of the 2 X 2 ANOVA for early and late

tested process and reactive subjects on the With-
drawal Scale.

 

 

Source of Variation Sum of Squares df Mean Square F
process - Reactive 813.69 1 813.69 10.444a
Interaction 17.40 1 17.40 0.223
Error 4363.00 56 77.91

Total 5197.84

 

ap < .01

72

The highly significant F for the process-reactive
effect indicates that the process-reactive classification
was the main determinant of scores on the Withdrawal Scale.
The period of hospitalization at which subjects were tested
had no effect on the scores either alone or in interaction
with the process-reactive classification.

In order to determine which of the groups contributed
to the effect of the process-reactive classification, the
means of the early and late tested process and reactive sub-
jects were again compared using the Newman-Keuls procedure.
The results of these comparisons showed that the differences
between late reactive and early process subjects and between
the late reactive and late process subjects were significant
(p < .05)-

The prediction made in Hypothesis 5 was not sup—
ported. Early reactive and late reactive subjects were not
found to differ on level of adjustment as measured by the
Withdrawal Scale. Hypothesis 6 and Hypothesis 7 were sup—
ported. Early reactive and early process subjects as well
as early process and late process subjects did not differ in
level of adjustment as measured by the Withdrawal Scale.

The relation between the hypothesized and observed
extent of thinking disorganization and withdrawal behavior
'may be clearly seen by reference to Figures 1 and 2. Al—
though Hypothesis 5 was not supported, the observed relation

between early and late tested reactive subjects was in the

73

Key‘

0 Process Subjects
+ Reactive Subjects

HIGH

Hypothesized Extent of Thinking Disorganization
and W1thdrawa1 BehaV1or

 

Low

 

Early Late

Figure 1. The hypothesized extent of thinking disorgani-
zation and withdrawal behavior for early and late
tested process and reactive subjects.

25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10

mm

Observed scores for the extent of thinking disorganization
and Withdrawal behaV1or
.s' U1 <n q

 

74

Key
0 Process Subjects
+ Reactive Subjects

Thinking Disorganization Scale
- - Withdrawal Scale

—
’—
”
—
,
’-
‘-
-
-
—
"
——
-
-
’
‘

 

Figure 2.

Early Late

The observed scores of the early and late tested
process and reactive subjects on the Thinking
Disorganization and Withdrawal Scales

75

direction hypothesized on both the Thinking Disorganization
and Withdrawal Scales. Even though Hypothesis 7 was sup-
ported, the observed scores of late process subjects do not
fall in the expected direction. It should be noted that
there were consistent differences between process and re-
active subjects as well as between the conditions of being
tested early or late. The one exception is the late process

subjects on the Withdrawal Scale.

Comparisons on Individual RSS

Even though the RSS used in this study were defined
as indicating schiZOphrenia, some of the RSS may have been
characteristic of both schiZOphrenics and non schiZOphrenics
while other RSS may have been characteristic of schizo-
phrenics only. Moreover, some RSS may have been character-
istic of only the process or reactive subgroups. Combining
heterogeneous RSS would result in cancellations and function
to reduce the overall discriminatory power of the entire set
of RSS.

Accordingly, it was necessary to make comparisons
among the groups on each RSS. Chi Square with Yates Cor—
rection was used to test for the significance of the differ—
ence in the number of subjects showing the presence or ab-
sence of each RSS. Fisher's Exact Test was substituted for

Chi Square when an expected frequency was less than five.

76

Those RSS which were found to be produced by more subjects
in one group than in another at the .05 level of signifi—
cance or better are shown in Table 5. Included also are
those differences which approached the .05 level of sig-
nificance. The notation in parentheses next to the level of
significance identifies the group containing the greater
number of subjects producing the RSS. The incidence of each
RSS in each group is shown in Appendix I.

The results of these comparisons fail to provide
strong support for the notion that any group could be
characterized by a unique pattern of RSS. Inspection of
Table 5 does show however that several RSS were consistently
found to be associated with one or the other groups over
several comparisons and at different levels of probability.

The notion that the RSS were heterogenous with some
RSS characteristic of one but not of another group is not
supported by the data reported in Table 6. Reference to
Table 6 shows that all groups were highly similar in terms
of the median RSS which were produced as well as in term of
the range of RSS which were produced. The range for each
Rorschach variable on which the RSS were based is shown in

Appendix J.

77

Ammo

 

mz mz m2 m2 m2 m2 m2 m2 m2 om. m2 0 sow A s sow
m2 m2 m2 m2 m2 m2 m2 m2 m2 mz mz_>.+m> + s + a»
+u+mussmv
>m + as + on sow
m2 m2 m2 mz mz m2 m2 m2 m2 m2 m2 mcoauoouooam
Isis
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m A o + mo sow
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 2 A mocoam 2 sum
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m A u_z sow
Arno
mz om. m2 m2 m2 m2 m2 m2 m2 m2 m2 N v 2 saw
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 oo. m.um«
Ammv comswcoo
mz mz m2 m2 m2 ca. m2 m2 m2 m2 m2 numaomounH
”mucosvom
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 “on
m2 m2 m2 m2 m2 m2 m2 m2 mz m2 m2 13
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 H N.3oo
Amqv Roms
m2 m2 m2 m2 m2 m2 mo. mo. m2 m2 m2 a N.sn
m2 m2 m2 m2 mz m2 m2 m2 m2 m2 m2 Hum A zus_v Hum
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 om A m v ma
deuce maimq amuse mules amiss menus mz.mq mz.mm maize azimm mZIm mmm

 

 

.mmm some now mcomflnmmsoo ecu mo muasmon one

.m manna

78

 

 

Ammo Ammv Ammv Ammv massage
m2 m2 m2 mo. Ho. mo. m2 m2 m2 0H. mz sumo o>emmooxm
Ammo
m2 m2 m2 m2 mo. m2 m2 m2 m2 m2 m2 messes Hoaoo
mz mz m2 m2 m2 m2 m2 m2 m2 m2 m2 mo. N.
noncommom N
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 so N.sm
Ammv
m2 m2 om. m2 m2 m2 m2 m2 m2 mz mz m m.mumaodom
Ammo Imus Ammo Imus Acme Imus
cm. on. m2 m2 mo. mo. oz mo. mo. m2 m2 soc N.sm
xmzc
m2 m2 mz m2 m2 m2 m2 m2 mmo. m2 m2 see m.sz
xmc
m2 m2 m2 m2 m2 m2 m2 mz m2 m2 om. smm m.s.+ m
Amqv Amzv noncommmm
m2 m2 m2 om. m2 m2 m2 m2 mz om. oz msoemce
u 00C mm Tum
noncommom xom
m2 m2 m2 m2 m2 m2 mz m2 m2 m2 m2 "oucmmoum
Lame Ammo om
m2 m2 m2 m2 mo. mmo. m2 m2 m2 m2 m2 u m ”acetone
Amzc
m2 m2 m2 m2 m2 mz m2 m2 m2 om. we 2 Bow A 6 gen
menus mq.sq amuse males mmumm squmm mznmq mz.mm mZumq mZImm mZIm mmm

 

 

1.o.ocooc m canoe

79

 

Lampoon no mo.v

 

m H m m a m H m m m o mooconommHo
ecconHcon
mo Honsoz Hmuoe
Amqv Amqv quv Ammv mowuomoumu
mz om. mo. m2 m2 m2 m2 m2 Ho. mo. mz ucoucoo souucz
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 rem m.mohsmH
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 mcoHudHuomoe
oH>H>
m2 m2 m2 mz m2 m2 m2 m2 m2 m2 m2 H N.mcoHuoonom
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 mochEmom
coHuHmom
Home Imus
mmo. mo. m2 m2 m2 m2 m2 m2 m2 m2 m2 mcoHumuo>omuom
Ammv mconsHucH
mo. m2 m2 m2 m2 m2 m2 m2 m2 m2 mz Hancouom
Ammv Ammv Ammv Ammv meowumNHHmQHo>
om. mz Hoo. m2 m2 Ho. mz 0H. wz m2 m2 OHcomonumm
m2 m2 mz m2 m2 m2 m2 m2 m2 m2 m2 mchom sumo
Ammv Ammo Hams
m2 m2 m2 mo. mz mo. m2 m2 m2 mo. mz mesons oHsooo
m2 m2 m2 m2 m2 m2 mz m2 m2 m2 m2 mcoHumcHsmucoo
m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 m2 mcoHum
IHOOmm< mGMHO
“Sums mHumH mmnmu .3 use muses mHumm mde mzim m2 :5 m2 new main one

 

 

A.o.ucoov m oHnt

80

Table 6. Median number of RSS for each group together with
the range for the number of RSS produced by the
subjects in each group.

 

 

 

Group Median RSS Range of RSS
Combined Process 10.00 5 — 15
Early Process 11.75 5 - 15
Late Process 9.13 6 - 11
Combined Reactive 9.50 5 — 19
Early Reactive 10.25 5 - 19
Late Reactive 8.50 6 - 14
Non Schizophrenic 7.40 5 - 15

 

Comparisons with Nbrmals

In order to provide another basis for evaluating the
Rorschach performance of the subjects in this study, the
schizOphrenics and non schizophrenics were compared with the
norms reported by Beck 23 31. (1950).

The mean scores of normals for 23 variables are shown
in Table 7 together with the corresponding mean scores for
the schizophrenics and non schiZOphrenics. The t test was
used to test for the significance of the difference between
the means. It was not possible to carry out t tests for
four of the 23 variables since Beck failed to provide infor-

mation.regarding the standard deviations. These four

81

Table 7. The means and standard deviations of the schizo-
phrenics, non schiz0phrenics, and Beck's normals
for 23 Rorschach Variables.

 

 

 

All Non
Normals SchiZOphrenics SchiZOphrenics

(N = 157) (N = 60) (N = 15)

R M 32.65 16.82a 22.07b
so 17.68 5.85 12.71

w M 5.50 4.50b 5.73C
so 3.76 2.45 4.19
ow M 0.02 0.31 0.00
SD Net Reported 0.60 0.00

o M 22.85 11.13a 14.25a
so 10.49 4.90 11.50

Dd M 3.02 0.78a 1.27b
so 3.38 1.50 2.35

c M 0.49 0.52c 0.23C
so 0.81 0.85 0.48

CF M 1.44 0.95C 1.43C
so 1.77 1.24 1.70

sc M 1.36 0.90a 1.97C
so 1.21 1.16 1.85
Sum c M 3.11 2.11a 2.40
so 2.74 1.92 2.04

M ii 3.50 1.96a 1.63b
so 3.24 1.60 1.56

Y M 1.96 0.05a 0.00.
so 2.22 0.24 0.00
V M 1.84 0.00 0.00
so 2.05 0.00 0.00

F + % M 79.25 64.30a 77.27C
so 10.20 18.70 14.48

82

Table 7 (con't.)

 

 

 

 

All Non-
Normals Schizophrenics SchiZOphrenics

(N = 157) (N = 60) (N = 15)

p M 6.79 5.08a 5.83C
SD 2.41 2.36 2.70

A% M 46.45 55.58a 48.13C
SD 13.12 5.54 15.55

H .M 4.02 2.23a 2.53C
SD 3.62 1.92 1.75

Hd M 1.78 0.91a 2.20C
SD 1.95 1.46 3.71
X M 0.13 0.00 0.00
SD Net Reported 0.00 0.00

Anatomy M 1.55 0.86b 1.80 (NS)

Responses SD 1.97 1.55 2.07
Sex M 0.03 0.06 0.00
Responses SD Not Reported 0.23 0.00

‘White M 1.90 0.65a 0.53b
Space SD 2.14 0.92 0.81
Lambda M 1.5-2.5 2.06 2.02
SD Net Reported 1.89 1.25

Afr M 0.60 0.51c 0.49b
SD 0.19 0.19 0.71

ap < .005
bp < .025

cNot significant

 

83

variables have been included for purposes of non-statistical
comparisons. Due to zero scores it was not possible to make
comparisons on two of the variables for the schiz0phrenics
and on four of the variables for the non schizophrenics.

The results of these comparisons showed schizo—
phrenics to differ from the normals on 15 of the 19 vari-
ables at the .025 level of significance or better. The non
schizophrenics were found to differ from the normals on 6 of

the 19 variables at the .025 level of significance or better.

ADDITIONAL FINDINGS

The Rorschach

Since the process-reactive dimension was the major
focus in this study, further comparisons were made among the
different groups of process and reactive subjects.

A major question which may be raised is whether more
process or more reactive subjects produce RSS when the
early-late distinction is ignored. In order to answer this
question the combined group of process subjects was compared
with the combined group of reactive subjects. This com-
parison was made using the Median Test and resulted in a Chi
Square of 0.00, with 1 df. There is no difference in the
number of process and reactive subjects producing RSS when
considered without regard for the period in hOSpitalization
when tested.

A second major question which may be asked is
whether more early tested subjects or more late tested sub-
jects produce RSS when the distinction between the process-
reactive subjects is ignored. In order to answer this
question the combined group of early tested process and re-
active subjects was compared to the combined group of late

tested process and reactive subjects. The Median Test for

84

85

this comparison resulted in a Chi Square value of 0.685 with
1 degree of freedom. There is no difference in the number
of early tested and late tested subjects producing RSS when
considered without regards for the process—reactive
distinction.

Three additional comparisons were carried out be-
tween groups of early and late tested process and reactive
subjects. The groups compared were as follows: early re-
active and late process subjects; late reactive and early
process subjects; late reactive and late process subjects.
The Median Test was used again to make each comparison. The
median RSS and the Chi Square values for the comparisons are
shown in Table 1. Reference to Table 1 will show that none

of these three comparisons was significant.
Adjustment Scales and Ratings

The Thinking Disorganization and
Withdrawal Scales
Additional comparisons for the Thinking Disorgani—
zation and Withdrawal Scales were carried out with attention
given to the process-reactive dimension and to the variable
of early and late testing. Specifically, one comparison in-
volved the combined group of early tested reactive and pro-
cess subjects and the combined group of late tested reactive

and process subjects. Another comparison was made between

86

the combined group of reactive subjects and the combined
group of process subjects.

Using t for each comparison of scores on each scale,
the only significant difference found was in the process-
reactive comparison on the Withdrawal Scale. Process sub-
jects showed more Withdrawal behavior than reactive sub-
jects (t = 4.720, pig .005, 58 df). There was a trend for
early tested subjects to show more thinking disorganization
than late tested subjects.

In a different focus an attempt was made to relate
scores on the Thinking Disorganization and Withdrawal Scales
to Rorschach performance. Produce moment correlations were
computed for the number of RSS produced by the combined
group of early and late tested process and reactive subjects
and subjects' scores on the Thinking Disorganization and
Withdrawal Scales. A correlation of .175 was found for the
Thinking Disorganization Scale and a correlation of .107 was
found for the Withdrawal Scale. With 58 df, neither corre-
lation was significant. Further correlations were carried
out for the subjects in each group. The number of RSS pro-
duced by early process subjects was significantly correlated
'with scores on the Thinking Disorganization Scale (p < .05).
For the same group the relation between the number of RSS
and scores on the Withdrawal Scale approached significance.
None of the other correlations was significant. The corre—

lations for all groups are shown in Appendix K.

87
Judges' Ratings

The mean rating and standard deviation of the
judges' global ratings of emotional health for the schizo—

phrenic and non schizophrenic groups are shown in Table 8.

Table 8. Judges' ratings of emotional health.

 

 

Early Late Early Late NOn
Reactive Reactive Process Process SchizOphrenic

Mean 4.183 ' 3.500 4.083 3.133 2.566
so 1.380 1.155 1.274 0.718 0.887

 

A one way analysis of variance carried out to com—
pare the global ratings resulted in a significant F = 5.148
(p < .01, 4, 70 df). The Newman-Keuls procedure was followed
in comparing the means of the groups. The results of this
comparison showed the ratings for the early reactive and
early process subjects to differ significantly from the
ratings for the non schizOphrenic subjects (p < .01).

A one way analysis of variance on the health ratings
for the process and reactive subjects only, however, was in-
significant (F = 2.567, 3, 56 df). A comparison of the mean
health ratings of the groups using the Newman-Keuls pro-
cedure showed none of the mean ratings to differ

significantly.

88

In order to obtain some measure of the extent to
which the judges followed the instructions to make their
ratings on the basis of their clinical judgments corre-
lations were computed for the health ratings of each subject
and the number of RSS in each subject's protocol. These
correlations are reported in Table 9. With the exception of
the late process subjects all correlations are highly

significant.

Table 9. Product moment correlations for the number of RSS
produced by the subjects and the corresponding
rating of emotional health.

 

 

 

 

 

I — —7
'__—‘ —‘

 

 

 

 

 

Group Correlation Level of Significance*
Early Reactive 0.757 .005
Late Reactive 0.686 .005
Early Process 0.520 .025
Late Process 0.137 Not Significant
Non SchizOphrenic 0.536 .025.

 

*13 degrees of freedom

The results of these comparisons would suggest that
the judges' ratings were at least in part based on some of

the formal scoring variables in the Rorschach protocols.

DISCUSSION

The results of this study clearly indicate that the
Rorschach has little, if any, concurrent validity as a diag-
nostic tool for use with schizophrenics. In none of the five
comparisons did more schizophrenics than non schizophrenics
produce Rorschach signs of schizophrenia. Regardless of
whether they were tested early or late, reactive subjects
were not found to differ significantly from non
schizophrenics. A trend was noted, however, for more early
process than non schizophrenic subjects to produce Rorschach
signs of schizophrenia.

These findings offer support, albeit weak support,
for the results of the Kantor, wallner, and Winder (1953)
and Kantor and Herron (1966) studies which showed process
subjects but not reactive subjects to be diagnosed as schizo—
phrenic on the Rorschach. An important aspect of the find—
ings of the present study is that the early tested process
subjects tended to differ from late tested process subjects
in producing RSS.

In terms of accounting for the findings of the
present study, the nature of the non schizophrenic sample

may have been a factor which Operated to bring about the

89

90

results of the non schizophrenic and schizophrenic com-
parisons. All non schizophrenics were hospitalized. For
the two subjects with a physical and a psychosomatic diag-
nosis, hospitalization may possibly have been warranted
purely on the basis of somatic reasons. According to con-
ventional practice, however, hospitalization would not have
been warranted for the remaining 13 subjects unless their
symptoms were extreme.

Indeed, the very fact of hospitalization would indi-
cate that the non schizophrenic subjects were so disturbed
as to require institutional care and custody. Accordingly,
at least 87% of the non schizophrenic subjects in the
present study could not be categorized as mild neurotic.
The implication, of course, is that the non schizophrenics
were as disturbed, or nearly as disturbed, as the schizo-
phrenics and that this factor operated in such a way as to
reduce the possibility of finding differences in the non
schizophrenic and schizophrenic comparisons.

The level of disturbance among the non schizophrenic
subjects may have, in fact, reduced the number of possible
differences. However, the findings of this study cannot
entirely be accounted for on the basis of this one factor.
It was not possible to obtain ratings for the non
schizOphrenic subjects on the Thinking Disorganization and
Withdrawal Scales since they had been discharged from the

hospital at the time of the study. The global ratings of

91

emotional health made by the judges, however, does indicate
that in spite of being hospitalized, the non schizophrenics
were the least disturbed of all of the groups. Accordingly,
the most appropriate conclusion is that the non schiZOphrenic
sample was truly different from the schizophrenic samples.

If mutually exclusive patterns of RSS had been found
to be associated with the non schizophrenic and schizophrenic
groups, the results of the non schizophrenic comparisons
might have been accounted for on the basis of a cancellation
effect occurring when the groups were compared without re—
gard for differences on individual RSS. Finding that the
non schizophrenic and the schizophrenic subjects failed to
differ to a significant extent on any of the RSS rules out
this explanation.

In spite of the fact that the RSS were by definition
indicative of schizophrenia, they were also produced by the
non schizophrenics. The data regarding the range and median
number of RSS produced by the groups clearly shows that
there was a high degree of overlap among the groups. While
this finding might be interpreted as supporting the notion
that the non schizophrenics were more severely disturbed
trian their diagnostic labels would suggest, it is more ap-
propriate to conclude that the RSS used in this study are
not exclusive signs of schizophrenia. Indeed, non

SChiZOphrenics as well as schiZOphrenics may produce

92

Rorschach variables which have been claimed to be character-
istic of schizophrenics only.

The finding that the RSS were produced by both
schizophrenics and non schizophrenics strongly suggests that
relying on signs alone in order to make a diagnosis at the
clinical level may result in a mis-diagnosis. This would be
true particularly in the case where one was concerned with
identifying some disorder as either schizophrenic or non
schizophrenic. Use of the RSS may be less severely limited,
however, if one is concerned with the degree of disturbance
rather than with the type of disturbance. Indeed, a more
valid application of the RSS might be one where the RSS were
considered not as signs of schizophrenia but rather as signs
of disturbance. In order to test this assumption it would
be necessary to determine whether the RSS, or RSD, were pro-
duced by non—hospitalized, non schiZOphrenics. The design
of the present study can be criticized because it failed to
include a second control group composed of this very type of
subject. Future research should consider this variable.

Further questioning of the notion that the RSS were
diagnostic of schizophrenia comes from the comparisons which
were made with the Beck 23 a1. (1950) norms. The differences
between the schizophrenics and normals indicate that the
schizOphrenics were highly deviant in terms of their
Rorschach performance. On the other hand, the non

schizophrenics were not as strikingly different from the

93

normals. These findings would suggest that the schizo—
phrenics and non schizophrenics were different in degree of
pathology. Accordingly, if the RSS were valid indicators of
schiZOphrenia the same relations should have been observed
when the schizophrenic and non schizOphrenic groups were com-
pared on the RSS. This was not the case.

The findings regarding the normal comparisons how—
ever are severely limited by the use of the statistic t.

The t test is appropriate only when N's do not differ marked-
ly and when variables can be assumed to be approximately
normal in distribution. The normal comparisons were made
with samples of sizes 157, 60, and 15. Furthermore,
Rorschach variables are not normally distributed as pointed
out by Cronbach (1949). The t test, however, was the only
statistic which could be used in making the comparisons

since Beck (1950) reports his data using means and standard
deviations.

Another factor which may account for the results of
the non schizophrenic and schiZOphrenic comparisons on the
Rorschach is the effect of the drugs which were being taken
by the subjects. The type of drugs and the amount of drugs
being taken by each subject in each group is shown in Ap-
pendix L. Since different subjects were taking not only
different types of drugs but also different amounts of drugs,
it is extremely difficult if not impossible to determine

what effect the drugs might have had on the Rorschach.

94

Comparisons made on the basis of dosage in milligrams is mis-
leading since several subjects were taking Haladol and
Navane, both of which are potent in effect and which are
given in lesser amounts than such drugs as Mellaril or
Thorazine.

Even though there are limitations on the inferences
which can be made about the effects of the drugs, it is
clear that regardless of type or amount, more schizOphrenics
were taking drugs than were non schizOphrenics. It is en—
tirely possible that the drugs did in fact produce a level-
ing effect such that the degree of pathology for the schizo-
phrenics was reduced to that of the non schizophrenics. How
this effect may have operated is unclear. Indeed, the ef-
fects of drugs on personality functioning is, at this time,
not completely understood.

Dramatic changes in behavior, however, have been
brought about through the use of drugs. Enelow and Wexler
(1966) report cases in which drugs have resulted in reducing
or eliminating a variety of psychotic behaviors such as so-
matic delusions, hallucinations, extreme hyperactivity, inco-
herent speech, and mental confusion.

Accordingly, it is possible that the drugs which
‘were being taken by the subjects in the present study did
affect those personality functions which were measured by
the Rorschach. Future research should give careful at-

tention to the matter of drugs and either use subjects who

95

are taking the same drugs in the same dosages, or subjects

who are not taking drugs.
The Process Reactive Dimension

One of the major thrusts of the present study was to
show that the Rorschach diagnosis of reactive schizOphrenics
would be directly related to that point in hospitalization
at which the Rorschach was administered. It was assumed
that the variable of recovery would function in such a way
that reactive subjects tested early in hOSpitalization would
be diagnosed as schizophrenic while reactive subjects tested
later would not be diagnosed as schiZOphrenic. It was
furthermore assumed that the variable of recovery would not
function to produce differences in the Rorschach performance
of early and late tested process subjects.

The results failed to support these assumptions. No
substantial differences were found in the Rorschach per-
formances of early and late tested reactive subjects. Early
and late tested process subjects did, however, tend to
differ significantly.

The lack of differences among the groups in terms of
Rorschach performance could be explained on the basis of
equal or near equal levels of adjustment. The findings of
the study regarding level of adjustment as measured by the
Thinking Disorganization and Withdrawal Scales does suggest

that the process and reactive groups were highly similar in

96

adjustment. The notion is furthermore strengthened by find-
ing no significant differences among the groups when com-
pared on the basis of the judges' global ratings of emotion-
al health.

The one exception to the notion of equal or near
equal levels of adjustment rests with the early process sub-
jects. In the early process-late reactive comparison, early
process subjects scored significantly higher on the Thinking
Disorganization and Withdrawal Scales. Moreover, the only
group showing a significant correlation between the number
of RSS produced and subjects' scores on the Thinking Dis-
organization Scale was the early process group. The corre-
lation between the number of RSS produced and scores on the
Withdrawal Scale was very nearly significant for the early
process group as well.

These findings would suggest that the early process
group was in fact the least adjusted of the process and re-
active groups, and may explain, in turn, the trend for sig-
nificantly more early process subjects than non I
schizophrenics to produce RSS.

With regard to the judges' ratings it is interesting
to note that while no differences were found when the schizo-
phrenic groups were compared, the ratings for the early
process as well as for the early reactive subjects were sig-
nificantly different from the ratings for the non

schizophrenics. These findings suggest that the judges may

97

have been making their judgments on more than a mere ap-
praisal of the Rorschach scores. .With the exception of the
late process subjects the ratings for the groups are in line
with what was expected. This may indicate, in fact, that
the best use of the Rorschach involves the use of not only
signs and scores, but also, and perhaps even more important,
that ill defined element often called clinical judgment.

When used in research on such variables as physio-
logical reactivity (Judson and Katah, 1963), perceptual
functioning (Bleke, 1955), and organicity (Brackbill and
Fine, 1956), the process-reactive dimension has proved to be
valuable as a means for distinguishing between types of
schizophrenics. The findings of the present study, however,
suggest that the process-reactive dimension may lack utility
in that type of clinically orientated research which is de-
signed to study differences in personality functioning.
Further studies investigating differences in the personality
functioning of process and reactive schizophrenics are needed
in order to be able to completely ascertain the heuristic

value of the process-reactive dimension in clinical research.
Suggestions for Further Research

Although the number of RSS found to be associated
with the process and reactive groups in the present study
was too small to rule out the possibility of chance oc-

currence, it is possible that some of the RSS may be

98

characteristic of either process or reactive schizophrenics.
Accordingly, it is suggested that further research should be
carried out in order to investigate the question of whether
process or reactive schizophrenics can be characterized by
different patterns of Rorschach variables.

Further research is needed on the Ullmann-Giovannoni
Scale. Norms for the scale are needed particularly in re-
gards to the average number of questions answered by re-
active subjects. Reliable cut—off points should also be de-
veloped for distinguishing between process and reactive
subjects.

It is also suggested that further research should
deal with the compatibility of using social history data and
the interview method for arriving at a process-reactive
classification for research subjects. Meichenbaum (1969)
has used a verbal administration method with the Ullmann-
Giovannoni Scale. No research, however, has been carried
out on the use of the same method with Premorbid History
Scale, or in comparing across methods and scales. Verbal
administration is not only a useful method but may at times
be necessary as, for example, when social history data is
inadequate, or when, in the case of the Ullmann-Giovannoni
Scale, subjects are unable to perform a pencil and paper
task.

Finally and perhaps most important, further research

should deal with the issue of the effect of drugs on

 

 

99

Rorschach performance. The best design for such research
would be Solomon's Four Group Design (Campbell and Stanley,
1967). As used in research on the Rorschach and drugs the

design can be diagrammed as follows:

Group A Pre-Test 1 Drugs Post-Test 2
Group B Pre-Test 3 No Drugs Post-Test 4
Group C Drugs Post-Test 5
Group D Test 6

The effect of drugs on the Rorschach could readily
be ascertained by the following four comparisons: Post-Test
2 and Pre-Test l; Post-Test 2 and Post—Test 4; Post-Test 5
and Test 6; Post-Test 5 and Pre-Test 3. The design could
also be carried out under the reversed experimental con-
ditions, and indeed, this procedure might be necessary if
the subjects were hospitalized patients, particularly since
the use of drugs is the rule rather than an experimental

condition.

SUMMARY AND CONCLUSIONS

The purpose of this study was to determine whether
reactive schizophrenics tested during the acute stage pro-
duce more Rorschach protocols characteristic of schizo-
phrenics than reactives in the non-acute stage. On the
basis of prior research it was assumed that process subjects
produce schizophrenic protocols regardless of length of
hospitalization. It was furthermore assumed that differ-
ences in levels of adjustment, or recovery, would be related
to differences in Rorschach diagnosis.

Sixty male hospitalized schizophrenics were classi-
fied as either process or reactive on the Phillips Premorbid
History Scale. The subjects were then divided equally ac-
cording to length of current hospitalization to form four
groups: early reactive, late reactive, early process, and
late process. All subjects were administered the Rorschach
and rated for level of adjustment on the Thinking Disorgani-
zation and Withdrawal Scales of the Psychotic Reaction Pro-
file developed by Lorr, O'Conner and Stafford. A control
group was formed by taking the protocols of 15 non
schizophrenic psychiatric patients from hospital files for

comparison purposes. The protocols were scored by two

100

101

experienced psychologists according to Beck's system. The
reliability of the judges was .92. An additional measure of
adjustment was obtained by having the judges rate the proto-
cols on a 6 point scale for emotional health.

A survey of the literature yielded 39 different
Rorschach signs of schizophrenia which were employed in the
present study. The non schizophrenics were compared with
the entire group of schizophrenics as well as with each pro-
cess and reactive group on the number of Rorschach signs of
schizophrenia. None of these comparisons yielded signifi—
cant differences. A tendency for early process subjects to
produce more signs than the non schizophrenics was noted.

All possible comparisons were made between the
process and reactive groups. None of these comparisons re-
sulted in significant differences in the number of Rorschach
signs of schizOphrenia. A trend towards more signs was noted
in early process subjects when compared with late process
subjects.

Further comparisons between the groups were made for
each of the Rorschach signs. No differences were found be-
tween the non schizophrenics and the combined group of
schizophrenics with respect to the presence of the signs.

The judges' global ratings of emotional health,
based on the Rorschach, showed the non schizOphrenics to be
less disturbed than either the early reactive or early

process subjects. No differences were found, however,

102

between the process and reactive subgroups themselves. The
same trend was noted when the process and reactive groups
were compared on the Thinking Disorganization and Withdrawal
Scales. These findings suggest that the process and re—
active subjects may have been nearly equal in level of ad-
justment. The one exception was that early process subjects
showed more thinking disorganization and withdrawal behavior
than late reactives.

Finally, the schizophrenics and non schizOphrenics
were also compared with the Rorschach norms reported by Beck.
Schizophrenics were found to differ significantly from the
normals on 15 variables while non schizophrenics differed
on only 6 variables. These findings tend to suggest that
the Rorschach signs of schizophrenia may not be true indi-
cators of schizophrenia and that the schizophrenics were in
fact highly deviant in their Rorschach protocols. Inferences
based on these findings are limited by the questionable para-
metric statistics used in the normative study.

The results of this study indicate that the Rorschach
signs do not differentiate between non schizOphrenic patients
and schizophrenics as a group or as the latter are subdivided
along the process reactive dimension. These findings may be
limited, however, only to the Specific Rorschach signs of
schizophrenia used in the study. The fact that the subjects
in the study were taking drugs may also account, at least in

part, for the findings. The effect of drugs on the Rorschach

103

performance of schizophrenics is unknown, but it is entirely
possible that they function to reduce the amount of path-
ology produced by schizophrenics in Rorschach protocols.

The results also raise some serious questions regarding the
utility of the process-reactive dimension.

It is suggested that future research should exercise
caution in regards to the drug variable. All subjects should
be taking the same drugs in the same amounts, or drug-free
subjects should be used. The nature of the control group
and the extent of its pathology is another important aspect
to be considered in future research. Attention should be
given to selecting subjects who are not disturbed and who

can truly be regarded as non schiZOphrenic.

APPENDICES

mcucOE

 

 

Q
mmmpm
coo. so.om om.mH mo.H~ mH.Na ococos
moH.H mm.m~ mo.HH He.om oo.oo cH mcoHomaHHmuHomom
HoHam mo camcoH
oHomHHm>< o.me Hm.m oo.mH ee.oH coHooaHchHomom
ooz Hs.om oH.HH oe.mH mo.mH ecouuoo one soon
coosuon mcucoz mo HonEsz
co. co». om.H owe. mo.H ecoHomNHHmeHdmom
oo.H oo.H oo.H em.H em.H eoHaa mo eonsoz
5 NH m «H m mcoHumNHHmqumom
noHum cuHs moooflsom
oHanHmea mm.m oe.o mm.H «H.m coHomNHHmuHcmom
ooz nom.m mmH.NH nmm.m m36H ocouaoo mo camcoq
He.H Hm.o so.H mm.H Hm.H
oo.oH so.oH mH.HH oe.HH mo.oH coHomooom
cm.m NH.o em.o -.m co.»

oo.em oc.om mm.mm oc.mm mm.om mom

Uecmucmonwnom O>Huomom o>HuOmmm mmoooum mmououm

:oz mung mHHmm mung Manna

ZOHB<NHA<BHmmOE 92¢ «ZOHBflUDQm amufl
mMAmGHm¢> Qua 20 mbomw 304m mom mZOHB€H>mQ QMGQZflBm 92¢ mz<m2

G Nanmmmfl

104

APPENDIX B

HOSPITAL DIAGNOSIS OF NON SCHIZOPHRENIC SUBJECTS

 

 

Diagnosis N

Anxiety Neurosis 1
Anxiety Reaction 4
Anxiety Reaction: Chronic 2
Anxiety State 1
Chronic Brain Syndrome 1
Character Disorder 1
Emotionally Unstable Personality 1
Passive Aggressive Personality 3
PsychoPhysiological Gastrointestinal Reaction 1
TOTAL.. 15

 

105

APPENDIX C

THE PREMORBID HISTORY SCALE

Shown below is the Premorbid History Scale of the

Prognostic Rating Scale (Phillips, 1953). A rating of from

zero to six is assigned for each category according to the

value of the item best describing a subject.

Subjects re-

ceiving a total rating of from zero to 15 were defined as

reactive schizophrenics.

Subjects receiving a total rating

of from 16 to 30 were defined as process schizophrenics.

Premorbid History Subscale

A. Recent Sexual Adjustment

1.
2.

Stable heterosexual relations and marriage.

Continued heterosexual relation and marriage
but unable to establish home.

Continued heterosexual relation and marriage
broken by permanent separation.

(a) Continued heterosexual relation and
marriage but with low sexual drive.

(b) Continued heterosexual relation with

deep emotional meaning but emotionally
unable to develop it into marriage.

106

Ratings

(a)

(b)

(a)

(b)

107

Casual but continued heterosexual re-

lations, i.e., affairs, but nothing more.

Homosexual contacts with lack of or
chronic failure in heterosexual
experiences.

Occasional casual heterosexual or homo-
sexual experience with no deep emotion—
al_bond.

Solitary masturbation with no active at-
tempt at homosexual or heterosexual
experience-

No sexual interest in either men or women.

Social ASpects of Sexual Life During Adolescence
and Immediately Beyond

1.

Always showed a healthy interest in girls

with a steady girl friend during adolescence.

Started taking girls out regularly in
adolescence.

Always mixed closely with boys and girls.

Consistent deep interest in male attach-
ments with restricted or no interest in
girls.

(a)

(b)
(a)

(10)

Casual male attachments with inadequate
attempts at adjustment to going out with
girls.

Casual contacts with boys and girls.

Casual contacts with boys and with lack
of interest in girls.

Occasional contacts with girls.

NO desire to be with boys and girls; never
went out with girls.

Ratings

108

Social Aspects of Recent Sexual Life; 30 years
of Age and Above

1.

Married and has children, living as a
family unit.

Married and has children but unable to es-
tablish or maintain a family home.

Has been married and had children but
permanently separated.

(a) Married but considerable marital
discord.

(b) Single, but has had engagement or deep
heterosexual relationships but emotion-
ally unable to carry it through to
marriage.

Single, with short engagements or relation-
ships with women which do not appear to have
had much emotional depth for both partners,
i.e., affairs.

(a) Single, has gone out with a few girls
but without other indications of a con—
tinuous interest in women.

(b) Single, consistent deep interest in male
attachments, no interest in women.

(a) Single, occasional male contacts, no
interest in women.

(b) Single, interested in neither men or
women.

Social Aspects of Recent Sexual Life; Below 30
Years of Age -

1.

Married living as a family unit, with or
without children.

Ratings

109

(a) Married, with or without children, but
unable to establish or maintain a
family home.

(b) Single but engaged or in a deep hetero-
sexual relationship (presumably lead-
ing towards marriage).

Single, has had engagement or deep hetero-
sexual relationship but has emotionally
been unable to carry it through to
marriage.

Single, consistent deep interest in male at-
tachments, with restricted or lack of
interest in women.

Single, casual male relationships with re-
stricted or lack of interest in women.

Single, has gone out with a few girls
casually but without other indications of
a continuous interest in women.

(a) Single, never interested in or never as-
sociated with either men or women.

(b) Antisocial.

Personal Relations: History

1.

2.

Always has had a number of close friends but
did not habitually play a leading role.

From adolescence on had a few close friends.

From adolescence on had a few casual
friends.

From adolescence on stOpped having friends.
(a) No intimate friends after childhood.

(b) Casual but never any deep intimate
mutual friendships.

Never worried about boys or girls; no de-
sire to be with boys and girls.

Ratings

110

Recent Premorbid Adjustment in Personal
Relations

1. Habitually mixed with others, but not a
leader.

2. Mixed only with a close friend or group of
friends.

3. NO close friends: very few friends: had
friends but never quite accepted by them.

4. Quiet; aloof: seclusive; prefered to be
by self.

5. Antisocial.

Ratings

APPENDIX D

THE ULLMANN‘GIOVANNONI SCALE

Shown below is the Ullmann—Giovannoni Scale (Ullmann and
Giovannoni, 1964) with the answers indicating reactive

schizophrenia. One point is given for each item answered in

 

the reactive direction. Reactive subjects were defined by
scores of 14 or more. Process subjects were defined by

scores of 11 or less.

1. When I leave the hOSpital, I will live with my

wife. (T)
2. I am married now. (T)
3. I have fathered children. (T)
4. I have been married. 9 , (T)

5. Before I was seventeen I had left the home I was
raised in and never went back except for visits. (T)

6. When I leave the hospital, I will live with one

or both of my parents. (F)
' - .
7. As a civilian I have worked steadily at one job
or for one employer for over two years. (T)

8. I finished at least one year of education after
high schoo1--trade apprenticeship, business
school, college, etc. (T)

9. Adding up all the money I earned for the last

three years, it comes to less than $700, be-

fore deductions. (F)
10. In my teens I was a member of a group of friends

who did things together. (T)

111

11.

12.

13.

14.
15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

112

I hardly ever-went out to another kid's house
after school or on weekends.

When I was in school I didn't like Physical Edu-
cation classes.

Alcohol has nothing to do with my difficulties.
I have paid regularly to buy a house.

More than once in the last year I have stayed on
after some group meeting and talked with some
other members about something that went on.

Shortly before I came into the hospital there was
some major change in my life--such as marriage,
birth of a baby, death, injury, loss of job, etc.

I have been deeply in love with someone and have
told them about it.

In the kinds of work I do, it is expected that
people will stay for at least a year.

My top wage in the last five years was less than
$1.25 an hour.

I have earned my living for longer than a year at
fulltime civilian work.

I have had to stay in a mental hospital for more
than one year at a time.

Within the last five years I have spent more than
half of the time in a mental hospital.

In my teens I was a regular member of a club or
organization that had a grownrup who came to
meetings. (Scouts, school club, 4-H, church
youth club, etc.)

In my teens there was more than one girl with
whom I had more than two dates.

(F)

(F)
(F)
(T)

(T)

(T)

(T)

(T)

(F)

(T)

(F)

(F)

(T)

(T)

 

APPENDIX E

THE THINKING DISORGANIZATION AND
WITHDRAWAL SCALES

Shown below are the 38 items of the Withdrawal Scale

and the 18 items of the Thinking Disorganization Scale taken

from the Psychotic Reaction Profile (Lorr, O'Conner, and

Stafford, 1960) as used in this study. Letters in parenthe—

ses following each item indicate the scale to which the item

belongs. Maximum scores on the Withdrawal Scale and the

Thinking Disorganization Scale are 38 and 18 respectively.

10.
11.

Never says more than three or four words at a time. (W)
Has to be helped along to stick to any activity. (W)
Says thanks when something is done for him. (W)

Talks whether anyone is listening or not. (TD)

Asks for things; doesn't wait for things to be given
to him. (W)

Is always doing something. (W)

Never asks for anything; waits for things to be given
to him. (W)

Doesn't take part in back and forth conversation. (W)
Often messey in eating habits. (TD)
Is able to talk about his own problems. (W)

Makes faces and strange movements that do not make
sense. (TD)

113

12.
13.
14.
15.
16.
17.
18.
19.
20.

21.

22.
23.
24.
25.
26.
27.
28.
29.
30.
31.

32.

33.

34.

35.

114

Seldom listens to radio or watches TV. (W)

Reads neWSpaper. (W)

Doesn't mix with other patients. (W)

Seems always busy with plans and projects. (W)
Shows no response to entertainment. (W)

Will do anything for recreation that comes up. (W)
Sometimes giggles in a silly way. (TD)

Spends a lot of time talking to himself. (TD)
Speaks softly; often difficult to hear. (W)

Smiles to himself a lot without any sensible
reason. (TD)

Is backward about talking to you. (W)

Has no friends on the ward. (W)

Seems concerned about what others think of him. (W)
Answers sensibly when talked to. (TD)

Talks to himself about imaginary or real faults. (TD)
WOuld sit all day unless directed into activity. (W)
Occasionally talks to himself. (TD)

Nearly always chatting with somebody. (W)

Usually knows what time it is. (TD)

Has little interest in the problems of others. (W)

It is difficult to understand what he is saying most
of the time. (TD)

Drifts off the subject when he talks. (TD)

Starts conversations with aides to become better
acquainted. (W)

Tries to be friendly with other patients. (W)

36.
37.
38.

39.

40.

41.

42.

43.

44.

45.

46.

47.
48.
49.
50.
51.
52.
53.

54.

55.

56.

115

Repeats words and phrases in a meaningless way. (TD)
Asks for help from other patients when he needs it. (W)
Talk is mostly not sensible. (TD)

Acts dead to the world; doesn't seem to care what is
going on . (W)

Never volunteers information about himself. (W)

Shows real friendliness towards at least one other
patient. (W)

Usually is slow moving and sluggish. (W)

Sometimes uses words that aren't understandable. (TD)
Interested in nothing. (W)

Ignores the activities around him. (W)

Shows occasional interest in news and current
events. (W)

(Likes to go for exercise. (W)

Has to be pushed to follow routine. (W)

Is slow thinking and a little confused. (W)

Sees and hears things that are not there. (TD)
Usually stays by himself.- (W)

Does not know where he is. (TD)

Laughs or smiles at funny comments or events. (W)
Is good company. (W)

Laughs if he is kidded. (W)

Does not know the names of aides. (TD)

APPENDIX F

PRODUCT MOMENT CORRELATIONS FOR THE SCORES OF
THE RELIABILITY RATERS ON THE THINKING DIS-
ORGANIZATION SCALE AND THE WITHDRAWAL SCALE

 

Thinking Disorgani-

 

Groups zation Scale Withdrawal Scale
All Groups .505c .252e
Combined Early Tested .915a .296a
Combined Late Tested .512b .2326
Combined Reactives .546d .166e
Combined Process .531e .2358
Early Reactives .943d .588e
Late Reactives .4876 .2706
Early Process ---—f -—--f
Late Process .5916 .249e

 

ap < .005
bp < .025
cp < .01

dp < .05

6Not significant

f I
r non-determinate:

2 subjects rated

116

APPENDIX G

PRODUCT MOMENT CORRELATIONS BETWEEN JUDGES FOR
EACH RORSCHACH VARIABLE SCORED. EXCEPT
WHERE NOTED ALL CORRELATIONS WERE
SIGNIFICANT AT A LEVEL OF .005

(28 df)
Determinate Indeterminate
w .961 de 1.000
0w 1.000 DdD 1.000
D .988 my -.049b
Dd .984 MV 0.00b
Approach .905 MC -.345b
Sequence .767 Y 0.00b
Afr 1.000 VF —.035b
v 0.00b
Sum M .921 TF 0.00b
M .920 T 1.00
NE .875 Sex 0.00b
Clang Associations 1.00
c .919 Cosmic Themes -.045b
CF .851 Card Edging .981
PC .871 Position Responses .049b
Vivid Descriptions 1.00
YF .612
FY .841
FV .283b
FT .123b
F+ .810
F— .802
EB .925
H .899
Rd .889
A .985
Ad .779
AN .979

117

Determinate

Secondary Content

F+%
A%

Populars

Color Naming

Excessive Card Turning
Contaminations

Pathogenic Verbalizations
Personal Intrusions
Perseverations

Rejections

ap < .025

bN'ot significant

118

.697

.691
.986
.919

.977
.909
.416a
.835
.773
.135b
.964

APPENDIX H

RORSCHACH SIGNS OF SCHIZOPHRENIA

Shown below are the Rorschach signs of schiZOphrenia

which were used in this study. Each criterion is listed

under the source from which it was taken.

A.

B.

Beck

UlubWNl-J

\lO‘

Beck

g...
ommxlmm-bwmI-a

had
AJH

l3.

F‘hfld
oxunp

17.

HF‘
«Jm

(1945)

R less than 15 or greater than 50

l or more rejections

Irregular or confused sequence

Sum M - greater than 2

Fluctuations between F+ and F- with at least 2
peaks and valleys

Card edging

Position responses

(1954)

l or more DW

1 or more DdW

Afr less than or equal to .60
FP% less than.55%

Sum M less than 2

Sum M blends greater than M

Sum M greater than Sum C

Sum CF + C greater than 3

A% equal to or less than 44%

A% equal to or greater than 60%
P equal to or less than 5
Lambda equal to or less than.57%
3% equal to or greater than 6%
x equal to or greater than .42
Excessive card turning

Clang Associations

Pathogenic verbalizations
Narrow content categories
Cosmic themes

119

120

20. Personal intrusions
21. Sex responses present
22. Anatomy responses present

C. Goldman (1960)

1. Sum FC + FY + FV greater than Sum CF + C + YF + Y
+ VF + V

D. Kelly and KlOpfer (1939)

. Emphasis on W

Emphasis on Dd

Sum C > Sum M .
W : M is less than or greater than 2 : 1
. Color naming

Contaminations

O‘U‘I-FWNH

E. Piotrowski (1952)

1. No H or Rd
2. Perseverations: 2 or more on one card or 1 or more

on two cards
3. Vivid descriptions but unclear concepts.

Total 39

 

3 6 o o m zsoonsom
o m o o o o§mAzsom
v o m m m >+m>+w+§+
o+mosomv>m+wm+omsom
m d m o o mooflumouuoam um .+m
m o m m m on+mo 83
o o o o o ZAmodonzsom
H H H o H mAusosm
m d m m o mvzsom
NH 3 .2 2 3 oo. Mama
N N c H m pomsmcoo Ho “masmoHuH ”oucoovom
H o a o m 6o
m o o o b E
o o o o o HNsoo
o e d m N HNso
3 3 3 3 ma TN A s”; v TN
6 m m o o omAmvfl
cacmuzmouwzom mmmoonm mmoooum 0>Huomom o>fiuomom mmm
coz mama aaumm mama manna

 

 

Ansonw some ma u zv
moomw mudm 2H mmm mo4m m0 muzmmmmm mma OZH30mm maombmbm ho mmmzbz Mme

H

Nanmmm<

121

122

HOO‘HI‘mI‘MNmONOONMNOQ‘

Homcmmmmamomoamaoo-fl'

NOMHQ‘NomOanF

H

H
H

H

cmoox-rm

\DLfiMd‘OMOMONV‘OLfi

H

H

IDHtDMu—l

Od‘fi‘MI‘LfiMF-lm

H

a

OOV‘IDLOQ‘I‘IDLO

mcowuoonom oumu
momcommom cowuwmom
mcoflumum>omuom
mcoflmouucH HMGOmHom
mcowumnwamnno> owcomocumm
meadow oumo

moEoca UHEmou
mcowumcwEmucoo
chwDMHUOmm¢ mamau
mcwcnoa ammo o>fimmwoxm
mCHEmz Hoaou

soNxm

m M.mumasmom

80M»?

Rooms:

xmmwxl...”
noncommmm meoumc< "mocommum

noncommom xmm «oucommum

pm I m ”woodman

 

oesonnmouflnom
coz

mmoooum
mama

mmmooum
hanmm

o>wuomom
oumq

0>Huomwm
wanna

mmm

 

 

123

 

 

H m w m m ucoucoo 3ounmz
o o o o o Nd. N momcoomom. x
o m H o m *3 v. 6383
o o o o H mGOHuQHHOmoQ ©H>H>
oHcounmouHcom mmooonm mmoooum 0>Huommm 0>Huomom mmm
coz oumH >Humm mama zHumm

 

 

 

mlo 0:02 Huo Hno Hlo m>

mlo Hlo mlo NIo «no >2

0:02 Hlo Hlo 0:02 0:02 .2

N10 muo Hlo Mlo Hlo 2%

due :10 one :10 MIC 22

mlo Hlo mlo mIo 010 02

mlo MIC MIC :10 :10 20

N10 «no 010 Hlo ml: 0

H10 NIo HIo 0:02 Hlo 00:0Hm z 25m

Nlo «(.0 «no Hno mIo I: Esm

mno mlo Mlo mlo 5:0 .2 Sam

mh.lom. mh.ImH. hm.umm. om.IOH. Hb.umo. and

0: 02 0:02 0: 02 0: 02 0: 02 Son

0:02 ~10 Hlo Huo N10 32

NumHnouo NumIHuH Numlmuo mumuouo buolouH Zflz

«mIOH hNIOH o¢I0H NNIOH mNIOH m
0H:0H:20NH:0m mm000um mm000Hm 0>Hu000m 0>Hu0002 00Hn0HHO> noncomnom

:02 0D0H hHumm 0umH >Humm

 

 

mozwwm¢ MO mnzmmmdm
QZOMmm Nm<> NR2 EUHMS mmqdem<> mmomfi HM¢ 230mm

2 NHQmemd

.Qmm¢m mmmz mmm mmB mUHmz_ZO mmfldem<> mu<muwmom Mme hO mwz<m

124

125

 

Hno 0:02 Hno Hno Hno 0:0Humu0>0mu0m
mno Hno «no Hno hno 0:0Hmsuu:H H0:omn0m
«no Hno oHne oooz eHne mcoHumuHHmnno> oHoomooumm
oooz Hno odoz Hno fine monom oumo
0:02 0:02 mno 0:02 ono m0E0:B UHEmou
Nno Hno mno Hno ono 0:0Hum:HEmu:00
0:02 0:02 Nno 0:02 0:02 0:0Hu0HOOmmm m:0HO
one one mne one fine moHouoe onmo o>Hmmooxm
Hne Hne mne odoz one mcHsmz uoHoo
0:02 0:02 0:02 0:02 0:02 000:02002 x
XHHnRo XHmnxo RBIXo XNHnmo onnRo mm
mHnm OHnm mnH oHnm mno mHMHomom
.xHthHN momanm Robnxmm .XomanH Rhonxfim XE
§oH 13¢ flowing. A“mooHnoHo o80H 18mm Xomnoomo on + .m
one Hne one one one momcodmom msoumcc
0:02 0:02 Hno hno 0:02 00m:omm02 x0m
mHno mno mno mno mno om
ono mno mno ono onH 2
0:02 0:02 Hno 0:02 0:02 >
0H:0H:20NH:0m mm000um mmooonm 0>Huom0m 0>Huom0m m0Hn0HHm> nomnomuom.
:02 090H mHHmm 0umH mHumm

 

126

 

xo.mnxho

 

xo.vLXom Rm.onxom xo.¢nmo XN.¢nxo Manson
0:02 0:02 0:02 0:02 0:02 0:0HumHuom0n ©H>H>
mno mne mne «no «no mooHuooflom
0:02 0:02 Hno 0:02 0:02 000:02m0m :oHuHmom
0H:0H:QONH:0m mm0ooum mm000Hm 0>Hu000m 0>Hu0002 m0HanHm> nomsomuom
:02 0umH >Humm 0umH >HH0m

APPENDIX K

PRODUCT MOMENT CORRELATIONS FOR SUBJECTS'
SCORES ON THE THINKING DISORGANIZATION
AND WITHDRAWAL SCALES AND THE NUMBER
OF RSS PRODUCED BY THE SUBJECTS.

(N = 15 each group)

 

 

Thinking Disorganization

 

 

Groups Scale Withdrawal Scale
Early Reactive 0.204 0.212
Late Reactive —0.399 -0.423
Early Process 0.486a 0.438
Late Process -0.l9l 0.293
All Groups 0.175 0.107
ap < .05

127

N HHm

oc oov on

m mom

H oom N mom

ooo oH hom

omH mom.

mH c oom mom

m com

oo¢ mom

NH 0 mom

co 0 Hon
m0>Huom0m 0u0H

ooH de

omH on v oom 0H:

0 ooc MHm

om oom NH:

ooH HH¢

ooo on

ooH oH omm mofl

ooH moo

mmeun 02 mod

. ooN mod

ov ooH mo¢

o d ooom cod

om mo<

omH c c omH «o:

c oom Hod

00>Hu000m MHHMm

 

0:m>02 HOOMHmm EoHHnHH 0:HN0H0:B 0:HN0H0um 0:0»H< HHHMHH0S

 

In I|Ill|
I l L

mBUHmem UHmemmONHmUm ZOZ 92¢ UHmemmONHmUm
Nm ZdeB OZHmm mZ4MOHHHHS.ZH flwflmon ODmQ QZC mUDmQ ho mmNB

H NHDmemd

128

129

000

000H

00¢

000

00H

000
00H

005

000
00MH
00m

00

noon:

mmoun

ov

om
0N

oHn
00H mon
00¢ won
00m hon
00G
mom
00m 000
00m 009
N09
Hon
00000Hm 000H

mHU

00H VHO
MHU

omH NHU
000 HHU
000 0H0
000m 000
000

00¢ 500
00D

m00

Oz V00
moo

00H N00

0 H00
mm000Hm aHumm

v 00m mHm
@Hm
N MHm

«Hm
0>Huom0m oumH

NNd‘

209000

0

fi‘fi' NN

 

0:m>02 H000H02 EsHHQHH

0:HN0H0:B

0:HN0H0um

oceans HHsmHHoz

 

 

130

 

 

 

 

oom mHm
mmsun 02 ¢Hm
mmoun 02 MHm
mm «Hm
ooH HHm
00H 0H2
00H mom
ooH mom
mocha 02 how
moo 0 won
mmsun 02 mom
mmoun 02 com
mmoun 02 now
oom mom
mmsnn 02 How
0H:0H:20NH:0m :02
:Hu:0HHn 0:HN0E0Hm Hm:000m 0:HNmu0:B HHHMHH02
mH mHn
ooo ¢HQ
ooc om MHQ
OH «HO
0 oomH HHQ
mm000Hm 0u0H
0:0>02 HoomHmm EoHHQHH 0:HN0H0:B 0:HN0H0um 0:00Hd HHHMHHoz

 

 

REFERENCES

REFERENCES

Abrams, S. A validation of Piotrowski's Alpha formula with
schi20phrenics varying in duration of illness.

American Journal of Psychiatry, 1964, 121, 45-47.

Adams, H. B., and Cooper, G. D. Rorschach response pro-
ductivity and overt psychiatric symptomology.

Journal of Clinical Psycholggy, 1961, 17, 355-357.

Ash, P. The reliability of psychiatric diagnosis. Journal
of Abnormal and Socialfgsychology, 1949, 44, 272-277.

Beck, 8. J. Problems of further research in the Rorschach
test. American J0u£na1 of Orthogsychigtry, 1935, 5,
100-115.

Beck, S. J. Rorschach's Test, vol. II. A variety of person-
ality pictures. New York: Grune and Stratton, 1945.

Beck, S. J. _The six schizophrenias. Research Monograph No.
6 American Orthopsychiatric Association, New York,
1954.

Beck, 8. J. Rorschach's Test, Vol. I. New York: Grune and
Stratton, 1961.

Beck, S. J., Rabin, A. I., Thiesen, W; G., Mblish, H. B.,
and Thetford, W. N. The normal personality as pro-
jected in the Rorschach test. Journal of Psychology,
1950, 30, 241-298.

Becker, W; A genetic approach to the interpretation and
evaluation of the.process-reactive distinction.
Journal of Abnormal and Socialjgsychology, 1956, 53,
229-235.

Benton, A. L. The experimental validation of the Rorschach

test. British Journaljof Medical Psychology, 1950,
23, 45-580

131

132

Blake, R. Reward and punishment as determiners of remi-
niscence effects in schizOphrenics and normals.
Journal of Personality, 1955, 23, 479-498.

Brackbill, G., and Fine, H. Schi20phrenia and central
nervous system pathology. Journal of Abnormal and
Social Psychology, 1956, 52, 310-313.

Bucker, S. L., and Williams, M. Color as a determinant of
responsiveness to Rorschach cards in schi20phrenia.
Journal of Consulting Psychglpgy, 1951, 15, 196-202.

Campbell, D. T., and Stanley, J. C. Experimental and guasi-
experimentgl designs_for research. Chicago: Rand
McNally, 1967.

Chase, L., and Silverman, E. Prognosis in schizOphrenia:
.An analysis of criteria in 150 schizophrenics
treated with metrazol or insulin. Journal of
Nervou§_and Mental Disease, 1943, 98, 464, 473.

Crumpton, E., and Groot, H. The "meaning" of Rorschach
color cards as a function of color. Journal of Pro-
jective Technigues, 1966, 30, 359-363.

Dana, R. H. The validation of projective tests. _Journal of
Projective Techniques, 1962, 26, 182-186.

Elisio, T. Overinclusive thinking in process and reactive
schizOphrenics. Journal of Consulting Psycholggy,

Enelow, A. J., and Wexler, M. Psychiatry in the practice of
medicine. New Yerk: Oxford University Press, 1966.

French, J. W., and Michael, W. B., £5 31. Standards for
_§ducationa1 and Psychological Tests and Manualg.
washington: American Psychological Association,
1966.

Fine, B., and Zimet, C. Process and reactive schizophrenia
and genetic levels of perception. Journal of Ab-
normal and Social Psychology, 1959, 59, 83-85.

Friedman, H. A comparison of a group of hebephrenic and
catatonic schizOphrenics with two groups of normal
adults by means of certain variables in the Rorschach.
Journal of Projective Techniqggg, 1952, 16, 352-360.

133

Friedman, H. Perceptual regression in schizOphrenia: An
hypothesis suggested by the use of the Rorschach
test. Journal ogflProjective Techniques, 1953, 17,
171-185.

Garfield, S. L., and Sundland, D. M. Prognostic scales in
schi20phrenia. Journa1_gf Consulting Psychology,
1966, 30, 18-240

Goldman, R. Changes in Rorschach performance and clinical
improvement in schizophrenia. Journal o§_Consulting
Psychglogy, 1960, 24, 403-407.

Guilford, J. P. (Ed.). Printed classification tests. AAF
Aviation Psychology Program Research Reports, No. 3.
Washington: U.S. Government Printing Office, 1947.

Guilford, J. P. Fundamental statistics in psychgloqv and
education. New Yerk: McGraw-Hill, 1965.

Hertz, M. R. The validity of the Rorschach method. American
J0urnal of Orthopsychiatry, 1941, 11, 512-520.

Hunt, R. C., and Appel, H. E. Prognosis in psychosis lying
midway between schi20phrenia and maniac depressive
psychosis. American Journal of Psychiatry, 1936,
93, 313-319.

Johnson, M. H., and Ries, H. A. Validation of the self-
report scale for process and reactive schizophrenics.
Journglygf Consulting Psycholggy, 1967, 31, 321-322.

Judson, A., and Katahn, M. The relationship of automatic
responsiveness to process-reactive schizophrenia and

abstract thinking. Psychiatric Quarterly, 1963, 37,
19-24.

Judson, A., and Martin N. Levels of personality organi—
zation and production of associative sequences in
process and reactive schizOphrenics. Journal of
Consulting Psychgloqv, 1961, 28, 208-213.

Kant, 0. Differential diagnosis of schizOphrenia in light
of concepts of personality stratification. American

Journal of Psychiatry, 1940, 97, 324-357.

Kent, 0. A comparative study of recovered and deteriorated
schizOphrenic patients. Journal of Nervous and
Mental Disease, 1941, 93, 616-624.

134

Kent, 0. The evaluation of prognostic criteria in schizo-
phrenia. J0urna1 of Nervous and Mental Disease,
1944, 100, 598-605.

Kantor, R., and Herron, W. Reactive and process schizo-
phrenia. Palo Alto: Science and Behavior Books,
1966.

Kantor, R., Wallner, J., and Winder, C. Process and reactive

schizOphrenia. Journal of Consulting Psychology,
1953, 17, 157—162.

Kelley, D. M., and Klopfer, B. Application of the Rorschach
method to research in sch120phrenia. Rorschach Re-
search Exchange, 1939, 3, 55-66.

Klinger, B., and Roth, I. Diagnosing schiZOphrenics with
Rorschach color responses. Journal of Clinical

Klopfer, B., Ainsworth, M. , K10pfer, W., and Holt, R. De-
ve10pments in the Rorschach Technigue, Volume I:
Technique and Theory. New York: Harcourt, Brace,
1954.

Kn0pf, I. Rorschach summary scores in differential diag-
nosis. Journal of Consulting Psychology, 1956, 20,
99-104.

Lorr, M., O'Connor, J., and Stafford, J. W. The Psychotic
Reaction Profile. Journal of Clinical PSYchology,
1960, 16, 241-245.

McReynolds, P. A comparison of normals and schizOphrenics
on a new scale of the Rorschach CET. Journal of
Projective Tests, 1966, 30, 262-264.

Malamud, W., and Render, N. Course and prognosis in schizo-

phrenia. American Journal of Psychiatry, 1939, 95,
1039—1057.

Meehl, P. Structured and projective tests: Some common
problems in validation. In B. I. Murstein (Ed.),
Handbook of projective techniques. New Yerk: Basic
Books, 1965.

Mehlman, B. The reliability of psychiatric diagnosis.
Journal of Abnormal and Social Psychology, 1952, 47,
577- 578.

135

Meichenbaum, D. Validation of marital status and the self-
report scale for prOCess-reactive schizophrenia.
_Journs1 of Consulting and Clinical Psychology, 1969,
33, 351-356.

Molish, H. B. The popular response in Rorschach records of
normal, neurotic and sch120phrenics. American

J0urp§1 of Orthopsychiatry, 1951, 21, 523-531.

Molish, H. B., and Beck, 8. J. Further explorations of the
"six schizophrenias": Type S-3. American Journal
of Orthopsychiatry, 1958a, 28, 483-505.

Molish, H. B., and Beck, S. J. Further explorations of the
"six schizophrenias": Type S-3 (concluded). Ameri-
can Jourppl of Orthopsychiatry, 1958, 28, 809-827.

Orme, J. E. A study of Weiner's Rorschach schizophrenic

indicators. Journal of Clinical Psychology, 1964,
20, 531-532.

Orme, J. E. A further comment on Weiner's Rorschach color
indicators. Journal of Clinical Psychology, 1966,
22, 223.

Phillips, L. Case history data and prognosis in schizo-
phrenia. Journal of Nervous and Mental Disorders,
1953, 117, 515-525.

Piotrowski, Z. A. Experimental psychological diagnosis of
mild forms of schi20phrenia. Rorschach Resspgch
Exchange, 1945, 9, 189-200.

Piotrowski, Z. A. A defense attitude associated with im-
provement in schiZOphrenia and measurable with a
modified Rorschach test. Journal of Nervous and
Mental Disease, 1955, 122, 36-41.

Piotrowski, Z. A. Perceptanalysis. New York: Macmillan,
1957.

Piotrowski, Z. A., and Berg, D. A. Verification of the
Rorschach Alpha diagnostic formula for underactive
schizophrenics. American Journalyof Psychiatry,
1955, 112, 443-450.

Piotrowski, Z. A., and Bricklin, B. A. A second validation
of a long-term Rorschach prognostic index for schizo-
phrenic patients. Journal of Consulting Psychglogy,
1961, 25, 123-128.

136

Piotrowski, Z. A., and Efron, N. Evaluation of outcome in
schiZOphrenia. In P. Hock and J. Zubin (Eds.),
Psychgppthology of Schigophrenia. New York: Grune
and Stratton, 1966.

Piotrowski, Z. A., and Lewis, N. D. C. An experimental
Rorschach diagnostic aid for some forms of schizo-
phrenia. American Journal of Psychiptry, 1950, 107,
360-366.

Piotrowski, Z. A., and Lewis, N. D. C. An experimental cri-
terion for the prognostication of the status of
schizophrenics after a three—year interval. In P.
Hoch and J. Zubin, Relation oinsychological Tests
to Psychiatry. New York: Grune and Stratton, 1952.

Rickers-Ovsiankina, M. The Rorschach test as applied to
normals and schizophrenic subjects. British Journal
of Medical Psychology, 1938, 17, 227-257.

Rieman, G. W. The effectiveness of Rorschach elements in
the discrimination between neurotic and ambulatory
schizophrenic subjects. Journai of Consulting Psy-
chology, 1953, 17, 25-31.

Rorschach, H. Psychodiagnostics. New York: Grune and
Stratton, 1942.

Rubin, H., and Lonstein, M. A cross validation of suggested
Rorschach patterns associated with schiZOphrenia.
Journpl of Consulting Psychology, 1953, 17, 371-372.

Scott, E. M. Regression or disintegration in schizophrenia?
Journal o§_C1inica1 Psychology, 1956, 12, 298-300.

Sherman, M. A comparison of formal and content factors in
the diagnostic testing of schiz0phrenia. Genetic
Psychology Monographs, 1952, 46, 183-234.

Stotsky, B. A. A comparison of remitting and non-remitting
schizophrenics in psychological tests. Jourppl of
Abnormal and Social Psychology, 1952, 47, 489-496.

Stotsky, B. A. Factor analysis of Rorschach scores of
schizophrenia. Journal of Clinical Psychology,
1957, 13, 275-278.

Stotsky, B. A., and Lawrence, J. F. Various Rorschach in-
dices as discriminators of marked and little con-
ceptual disorganization among schizophrenics.
Journal of Consulting Psychology, 1955, 19, 189-193.

137

Strum, E. Overinclusion and concreteness among pathological
groups. Journai of Consulting Psychology, 1965, 29,
9‘18 0

Taulbee, E. S., and Sisson, B. D. Rorschach pattern analysis
in schizophrenia: A cross validation study. Journal

of Clinical Psychology, 1954, 10, 80-82.

Taulbee, E. S., Sisson, B. D., and Gaston, C. Affective
Ratio and 8-9-10 percent on the Rorschach test for
normals and psychiatric groups. Journal of Con-
sulting Psychology, 1956, 20, 105-108.

Thiesen, J. W. A pattern analysis of structural character-
istics of the Rorschach test in schizOphrenia. p
Journgi of Consulting Psychology, 1952, 16, 365-370.

Tutko, T., and Spence, J. The performance of process and
reactive schizophrenics and brain injured subjects
on a conceptual task. J0urna1 of Abnormai and
Social Psychoiggy, 1962, 65, 387-394.

Ullmann, L. P., and Giovannoni, J. M. The deve10pment of a
self-report measure of the process-reactive con-
tinuum. Journal onyervous and Mental Disease,
1964, 138, 38—42.

Warner, S. J. An evaluation of the validity of Rorschach
popular responses as differential of ambulatory
schiZOphrenics. Journal of Projectiye Technigues,
1951, 15, 268-275.

Watkins, C., and Deabler, H. L. Responses of chronic schizo-
phrenic patients to tachistoscopic presentation of
Rorschach figures. Journal of Projective Techniqggs,
1957, 21, 404-409.

Weiner, I. B. Three Rorschach scores indicative of schizo-
phrenia. Journal of Consulting Psychoiogv, 1961,
25, 436-439.

Weiner, I. B. Rorschach tempo as a schiZOphrenic indicator.
Perceptual and Motor Skills, 1962, 15, 139-141.

Weiner, I. B. Pure C and color stress as Rorschach indi-
cators of schiZOphrenia. Perceptual and Motor
Skills, 1964a, 18, 484.

 

weiner, I. B. Follow-up validation of Rorschach tempo and
color use as schi20phrenic indicators. American
Psychologist, 1964b, 19, 494.

138

Weiner, I. B. Rorschach color stress as a schizOphrenic
indicator--A reply. Journal of C1inicai_Psychology,
1965a, 21, 313-314.

Weiner, I. B. Follow-up validation of Rorschach tempo and
color use as schizOphrenic indicators of schizo-
phrenia. Journal of Projectiyg Techniques, 1965b,
29, 387-391.

Winer, B. J. Statistical principles in experimental desigp.
New Y0rk: McGraw Hill, 1962.

Zahn, T. (Acquired and symbolic affective values as de-
terminants of size estimation in schizOphrenia and
normal subjects. Journal of Appormalpgpd Socigi
Psychology, 1959, 58, 39-47.

Zimet, C., and Fine, H. Perceptual differentiation and two
dimensions of schizophrenia. Journal of Nervous and
Mental Diseasg, 1959, 129, 435-441.

   

TY LIBR

MICIH'IGAIN (SIMIUEIUIIMIIVIEIWIIIH WW
3 1293 030821046