QQRSCFACE‘! {SW} 2N5}; e- _
REE -A?£OKSH31 ti: ‘2’???“ o_.

 

AND FARANGEA

Raf??? by. ”:3; :!E;é. 5?.

€963

LIBRARY

Michigan Stan:
University

 

ABSTRACT
RORSCHACH EVIDENCE CONCERING THE RELATIONSHIP
BETWEEN HOMOSEXUALITY AND PARANOIA

by Harry K. Dillard, Jr.

The present study represented an attempt to replicate
Aronson's findings that paranoids give a greater frequency
of "homosexual" responses on the Rorschach than matched
control groups. Sixty subjects, thirty paranoid psychotics
and thirty nonparanoid psychotics, were selected on the
basis of the extent of paranoid delusions in their clinical
picture. The subjects were closely matched for age, educa-
tion, intelligence, and length of hospitalization. The
Rorschach records of the sixty subjects were drawn from
their respective files and scored by two Judges for the
twenty—one signs of homosexuality suggested by Wheeler (1949)
and Reitzell (1949).

While the writer's paranoid group reSponded to a sig-
nificantly greater number of homosexual signs on the Rorschach
than did Aronson's nonparanoid psychotic group, no significant
differences in frequency of response to homosexual signs were
found between paranoid and psychotic subjects on three similar
comparisons. One reason which might, in part, account for the
lack of correspondence between the paranoid subjects of the
writer and Aronson on frequency of responding to homosexual
signs was the finding that Aronson's paranoid sample gave an

unusually large number of responses to the Rorschach. As

Harry K. Dillard, Jr.

Aronson's paranoid subjects would have had a greater Oppor-
tunity to respond to the homosexual signs, especially signs

15 through 21, which may be scored more than once, the dif-
ference between the paranoid groups in frequency of responding
to homosexual signs may have been due to Aronson having an
atypical responsive paranoid sample.

The inability of the present study to find consistent
significant results has two major implications. (a) The
phenomenon of the association between homosexuality, as meas-
ured by homosexual signs, and paranoia may be too gross to
yield consistently reliable results. (b) The usefulness of
homosexual signs as the rationale for the interpretation of
homosexual conflict on the Rorschach test is questionable as
none of the signs, taken individually, were able to signifi—
cantly differentiate between the paranoid and nonparanoid
psychotic groups.

As the results of the current study offered very
limited support to the Freudian theory of paranoia, it
appeared from Rorschach evidence that the association between

homosexuality and paranoia is not as universally applicable

L7 Z
Approved ‘22:” (;,%éé:'L/’
Major Professof

[4/4 _ 1,. ,, 1
Date ,Apt? I/Z 426~>

as implied by Freud.

RORSCHACH EVIDENCE CONCERNING THE RELATIONSHIP
BETWEEN HOMOSEXUALITY AND PARANOIA

By

Harry K. Dillard, Jr.

A THESIS

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

MASTER OF ARTS

Department of Psychology

1963

ACKNOWLEDGMENT

Acknowledgment is gratefully extended to my committee
chairman, Dr. John R. Hurley. His objective criticisms and
evaluations, continued encouragement, and generous donation
of his time as well as his expressive support were essential
in the develOpment of this thesis.

I also wish to thank Dr. B. Karon and Dr. R. McMichael
for their critical evaluations which greatly enriched the
meaningfulness of the final draft.

A special note of thanks to Mr. K. O. Schmidt and
Mr. R. 0. Olive for their generous donation of time as judges
in the present study.

My wife, Susan, is especially thanked for her continu—
ing support, encouragement, and critical proofreading in the
unfolding of this thesis.

The records of the patients of the Ionia State Hospital
were placed at my diSposal through the kindness of Dr. A.

Birzgalia, Medical Superintendent, Ionia State Hospital.

ii

TABLE OF CONTENTS

Page

ACKNOWLEDGMENT . . . . . . . . . . . . . . ii

LIST OF TABLES . . . . . . . . . . . . . . iv
Chapter

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

II. METHOD. . . . . . . . . . . . . . 7

Rating Scale. . . . . . . . . . . 7

Selection of Subjects. . . . . . . . lO

Rorschach Test . . . . . . . . . . 14

Statistical Analysis . . . . . . . . 15

III. RESULTS . . . . . . . . . . . . . 17

IV. DISCUSSION . . . . . . . . . . . . 23

V. SUMMARY . . . . . . . . . . . . . 28

REFERENCES . . . . . . . . . . . . . . . 3O

APPENDICES . . . . . . . . . . . . . . . 32

iii

TABLE OF CONTENTS

ACKNOWLEDGMENT

LIST OF TABLES

Chapter
I. INTRODUCTION.
II. METHOD.
Rating Scale. . . .
Selection of Subjects.
Rorschach Test . . .
Statistical Analysis
III. RESULTS
IV. DISCUSSION
V. SUMMARY
REFERENCES
APPENDICES

iii

Page
ii

iv

10
14
15
17
23
28
3o

32

CHAPTER I
INTRODUCTION

The role of homosexuality in the etiology of paranoia
was suggested by Freud and Breuer as early as 1895. Sixteen
years later, Freud (1911) conceptualized his theory on the
relationship of paranoia to homosexuality in a detailed
analysis of the Schreber case. In this formulation Freud
maintained that paranoia develops as a defense against un-
conscious homosexual conflicts in which all the major types
of paranoid delusions have precisely the same function, to
contradict the basic unconscious feeling that "I (a man)
love him (a man)."

Other psychoanalytic writers have supported Freud's
observation that a close association exists between paranoia
and the homosexual wish fantasy. Shockley (1914, p. 438)
went so far as to conclude that Freud's theory of paranoia
could be ”regarded as proven since it has been observed by
so many writers.” However, quantitative research in this
area have yielded conflicting results.

Gardner (1931) investigated the incidence of repressed
homosexuality in forty unselected cases of paranoid schizo-
phrenia and eighty cases of paranoid condition. As his
rationale for measuring repressed homosexuality, Gardner

1

used the relatively gross criteria of (a) homosexual acts,
(b) statements by the patients that they had been attacked
homosexually, and (c) symbolic expression of homosexuality.
He found that 55% of his paranoid schiZOphrenic patients
revealed evidence of repressed homosexuality; while among
the paranoid condition patients, he found evidence of re—
pressed homosexuality in only 40% of these cases. Thus, of
the 120 paranoid cases Gardner examined approximately 50%
revealed evidence of repressed homosexuality.

Miller (1941) studied 400 paranoid patients, of whom
152 had been diagnosed as paranoid schiZOphrenics. He found
that there were only twelve cases which were so adequately
explained by the psychoanalytic concept of paranoia that no
reasonable person could deny the applicability of Freud's
theory to those cases. In addition, he found that there
were perhaps three times as many as that number to which the
theory could be fitted. However, in the remainder of the
cases, he found that the paranoid psychosis defied explana-
tion by the psychoanalytic hypothesis. On the basis of his
research, Miller postulated that the paranoid mechanism
could be caused by other etiological factors which are:
(a) incomplete psychosexual development; (b) physical defects;
(c) impotence; (d) failing faculties; (e) organic brain
disease; and (f) life situations giving rise to feelings of
anxiety, frustration, and inadequacy.

Klein and Horwitz (1949) also used a case history

approach to study various psychosexual characteristics and

delusional preoccupation and activity in a group of forty
male and forty female paranoid patients. They found that
only about 20% of their patients showed preoccupation with
homosexuality, even at the height of their illness when the
patient's defenses were presumably weakened. The authors
reported that the patient's fears of becoming homosexual

were related to ego insufficiency, failure to reach life
goals, and a generalized distrust of people's acceptance.

For the most part, they concluded, that their subject's fears
were not related to unconscious homosexuality in the Freudian
sense. Thus, while they found support for the Freudian
theory of paranoia in some cases, they concluded that the
relationship between homosexuality and paranoia was not as
universally applicable as implied by Freud.

In an original study, Aronson (1952) attempted an
independent evaluation of the Freudian theory of paranoia by
means of the Rorschach test. Three groups of thirty subject
each, including paranoid psychotics, nonparanoid psychotics,
and normals, were given the Rorschach test. The psychotic
subjects were selected on the basis of a rating scale which
was designed to estimate the "extent to which the delusions
pervaded the patient's symptomatology,” (Aronson, 1952, p.
403). In addition, the subjects in each group were white
male war veterans who were matched with respect to age, intel-
ligence, education, and length of hospitalization (except for

the normal group).

As the rationale for the interpretation of homosexual
conflict on the Rorschach test,Aronsqnused the twenty homo-
sexual signs suggested by Wheeler (1949) plus one additional
sign suggested by Reitzell (1949). Wheeler had previously
demonstrated that his twenty homosexual signs were internally
consistent with each other and were externally consistent
with therapist's ratings of homosexuality. Moreover, while
Reitzell's data was somewhat inconclusive statistically,
she found that a group of homosexual subjects tended to
select the twenty Wheeler homosexual signs plus her addi-
tional sign more frequently than comparable groups of hysteric
and alcoholic subjects. Using these twenty-one homosexual
signs suggested by Wheeler and Reitzell, Aronson found that
the paranoids selected a significantly (P‘<Z.001) greater
number of homosexual signs on the Rorschach than did either
of the control groups.

Grauer (1954), in an attempted replication of Aronson's
study, compared the Rorschach responses of thirty-one para-
noid schiZOphrenics with Aronson's nonparanoid psychotics.

He attempted to match his paranoid subjects on three of

Aronson's control variables (age, education, intelligence).
Moreover, as twenty-eight of Aronson's thirty paranoid sub-
jects had been given a diagnosis of paranoid sxhiZOphrenia
by the psychiatric staff, Grauer used psychiatric diagnosis
and ”unequivocal” evidence of delusions and hallucinations

as his criteria for paranoid group selection rather than

U7

Aronson's rating scale. While Grauer found no statistically
significant differences in frequency of response to Wheeler's
and Reitzell's homosexual signs between his paranoid schizo—
phrenic patients and Aronson's nonparanoid psychotic patients,
the differences were in the direction of Aronson's findings.

There are two reasons, however, to think that the
Aronson-Grauer discrepancies may be due to inadequate
replication.

1. As Grauer used psychiatric diagnosis and ”unequivocal"
evidence of delusions and hallucinations for paranoid group
selection rather than Aronson's rating scale, his criteria
for subject selection were inconsistent with that of the
original study.

2. As Grauer reported insufficient data on the control
variables of age, education, intelligence, and length of
hospitalization to allow statistical comparisons between his
paranoid schiz0phrenic sample and Aronson's psychotic sample,
the adequacy of his matching could not be determined. If
the groups were not adequately matched, Grauer's results
and conclusions may reflect nothing more than procedural
errors. Moreover, Grauer did not specify if his subjects
were Caucasian. If they were not the implication is serious
as minority groups in the United States are often subject to
varying degrees of discrimination and/or persecution which
may influence Rorschach test results.

To the researcher, replication evidence is tenable only

to the extent that the original study had been demonstrably

reproduced. Thus, it appears that Grauer's procedural dis-
crepancies severly limit the reliability of his replication.
The inability of the investigator to reproduce the
design of an original study represents one of the major
limitations to the replication of clinical studies. While
absolute replication is a formidable, if not impossible
task, designs which more closely approximate the original
study are often experimentally feasible. The present study
constitutes an attempt to conduct a controlled replication
of Aronson's study. However, the present study's attempt at
replication is limited by the perplexing unavailability of
Aronson's criteria defining each level of his seven point
rating scale.1 It was felt, however, that if the subjects
of the present study were more apprOpriately selected and
matched in accordance with the procedures of the original
study, the findings might provide evidence to reconcile the

discrepancies of the two prior studies.

 

lAronson's original dissertation contained no further
information on his rating scale. In a personal communication
from Aronson, he did not reveal his rationale defining each
level of his rating scale.

CHAPTER II

METHOD

Rating Scale

 

Aronson used as the primary criterion for subject
selection the extent to which paranoid delusions were
paramount in the patient's symptomatology, offering three
reasons for this choice. (a) Psychiatrists differ widely
in the criteria which they utilize for diagnosing patients
as either paranoid or nonparanoid. (b) Some patients who
are not diagnosed paranoid may, nevertheless, exhibit many
paranoid delusions. (0) Diagnostic classifications do not
ordinarily indicate the extent to which paranoid delusions
are present or absent (Aronson, 1952, p. 402).

Since the rationale defining the levels of the orig—
inal scale was unavailable, the writer developed a new
rationale for each level of Aronson's rating scale (see
Appendix A). If results comparable with those of the orig—
inal scale could be obtained, it would suggest that the
rationale differences between the original and present scales
were negligible from the standpoint of delusional quantifi-
cation.

In accordance with Aronson's procedures, all clinical

material which was available at the Ionia State Hospital was
7

used to rate the paranoid and nonparanoid psychotic subjects
of the present study. Most of the available clinical data

on the patients in this hospital could be found in their

case records. Typically, all case records contain the follow-
ing information: (a) social history, (b) routine admission
note, (c) psychiatric interview notes, (d) staff evaluation
notes, (e) correspondence with other institutions, (f)
attendant's ward notes, (g) medical history, (h) commitment
papers and other legal documents, (i) censored correSpondence,
(j) psychological test protocols and interpretations
(Rorschach, D-A-P, WAIS).

The writer read through approximately fifty case
records which contained the above information, omitting test
protocols and interpretations to avoid contamination, and
selected twenty case records which sampled the range of
paranoid delusions represented on the rating scale. The
remainder of the cases were discarded. The twenty cases
were rated by the writer and then presented to two psycho-
logists at the Ionia State Hospital for independent ratings.

The judges were instructed to read through each record,
omitting test protocols and interpretations, and to give
high ratings to those patients whose entire symptomatology
was dominated by one or more of the four principal kinds of
paranoid delusions, i.e., persecution, erotomania, jealousy,
and megalomania (Freud, 1911), and to give low ratings to
those patients whose clinical picture was characterized by

an absence of paranoid delusions. Inter-rater reliability

was then computed for the highest rating on any of the four
kinds Of paranoid delusions, for each type of paranoid delu-
sion, and for total ratings by means of the Pearson product-
moment correlation.

Table 1 shows that a high agreement between the three
judges was obtained for each comparison. For the highest
rating on any of the four kinds of paranoid delusions, the
inter-rater reliabilities of .94, .91, and .90 on the present
scale compared favorably with Aronson's inter—rater reliabili-
ties of .95, .92, and .88. No other inter-rater reliabilities

were reported by Aronson. As in Aronson's study, the agreement

TABLE 1

PEARSON PRODUCT-MOMENT CORRELATION OF INTER—
RATER RELIABILITY ON RATINGS OF DELUSIONS

 

 

 

(N=20)
Judge Pa E M J H T
I' I“ I' I’ I" I"
A vs. B .93 .91 .98 .98 .91 .97
A vs. C .83 .76 .92 .96 .94 .85
B vs. c .81 .66 .91 .90 .90 .81

 

 

8P = persecution; E = erotomania; M = megalomania;
J = jealousy; H = highest rating on any of four types of
delusions; T = all ratings combined.
between the current raters on the twenty cases was considered high

enough to warrant the writer's selection of the entire sixty

cases for the paranoid and nonparanoid psychotic groups by

lO

himself. Moreover, it appeared that the use of the present
scale to select subjects would not seriously limit the
replication of the study from the standpoint of delusional

quantification.

Selection of Subjects

 

The writer obtained a list of approximately 200 psy—
chotic male war veterans who were patients at the Ionia
State Hospital. In accordance with Aronson's procedure, the
writer read through the case records of each of the fore-
mentioned patients and selected thirty subjects for each
of the psychotic groups, paranoid (PD) and nonparanoid psy-
chotic (N-PD),according to the following criteria:

1. Paranoid group——The patient must have received a
rating of four or more on delusions of persecution by the
writer.

2. Nonparanoid psychotic group—-The patient must not
have received a rating of more than two on delusions of per-
secution by the writer.

3. All subjects must be white, male war veterans, of
average intelligence, who have not completed more than
twelve grades of education, who are not over forty years of
age, and who have not spent more than 167 days in the hospital
prior to Rorschach testing.

Table 2 presents an overall comparison of the paranoid
and psychotic groups of both studies on each of the major

control variables. As can be seen from this table, both of

11

the investigator's groups had spent significantly fewer days
in the hospital prior to Rorschach testing than had Aronson's
paranoid (PA) and nonparanoid psychotic (N—PA) groups. As
reSponsiveness to homosexual signs may be a function of the
length of hospitalization, the per cent of homosexual signs
to the total number of Rorschach responses (R) was correlated
with days of hospitalization prior to psychological testing
by means of the Pearson product-moment correlation. A cor—
relation of .082 was found which suggest that reSponsiveness
to homosexual signs is not significantly related to length
of hospitalization prior to psychological testing in the

current study.

TABLE 2

COMPARISONS BETWEEN ARONSON'S AND DILLARD'S GROUPS ON THE
CONTROL VARIABLE'S MEANS AND SAMPLE VARIANCES

 

 

(N=30)
Vocabulary Education Days in
Age Scaled Score Level Hospital
Group —- 2 —- 2 —- 2 —- 2
X S X S X S X S

 

5 10.9 9.6 167.1 771.0
.4 9.5 4.9 74.6 727.2
8

PA 30.8 20.9 11.3 5
5
N-PA 28.9 20.2 10.9 3.
4

PD 30.8 28.9 10.1
11.1 5.1 149.1 393.0

N-PD 29.5 15.2 9.8 9.0 4.2 68.4 1124.7

UT

 

 

Table 2 also reveals that no significant age differences

between any of the four groups were found. As in Aronson's

study, the intellectual level was estimated from the mean

 

l2

weighted score obtained on the Vocabulary Subtest of the
Wechsler-Bellevue Scale (Wechsler, 1944) or the Wechsler
Adult Intelligence Scale (Wechsler, 1958). Aronson's non-
paranoid psychotic group was found to have a significantly
higher mean vocabulary score than the writer's nonparanoid
psychotic group (t = 2.25; P-<.05). However, there is
little reason to believe that the difference was of much
significance as both group's weighted vocabulary means were
within the average range. The N-PA group was also signifi-
cantly more educated than the N-PD group (t = 3.82; P (.001).
The N-PD group did, however, satisfy the selection criterion
in that no subject had completed more than twelve grades of
school. However, as the N-PD group's responsiveness to homo—
sexual signs may be a function of intelligence and education,
the per cent of homosexual signs to Rorschach R was corre-
lated with scaled vocabulary scores and with amount of educa-
tion by means of the Pearson product-moment correlation. The
respective correlations were .04 and .14 which suggest that
the N-PD group's responsiveness to homosexual signs is not
significantly related to intelligence and education in the
present study.

While the subjects in the PD and N—PD groups were not
originally selected with reSpect to religious affiliation,
occupation, and diagnosis, inspection of Table 3 suggest that
there are no statistically significant differences between
the PA and PD groups and the N-PA and N-PD groups on these

variables.

13

TABLE 3

COMPARISONS OF PARANOID AND PSYCHOTIC GROUPS OF DIAGNOSTIC,
OCCUPATIONAL, AND RELIGIOUS DISTRIBUTION

 

Comparison A
Distribution of Diagnoses of the Patient Groups

 

Code for Diagnosisa

 

 

Group P PS M U SR H C S
PA 28 0 l 1 0 0 0 0
PD 26 1 l 2 0 0 0 0

N—PA 0 0 5 18 1 1 2 3

N-PD 0 1 5 l7 2 1 2 2

 

Comparison B
Occupational Distribution of the Patient Groups

 

Group Unskilled Skilled Clerical Semi-Prof. Prof.

 

PA 17 6 4 2 1
PD 19 6 4 l O
N-PA 21 6 2 l O
N—PD 2O 4 3 2 l

 

Comparison C
Religious Distribution of the Patient Groups

 

 

 

Group Protestant Catholic Jew No-Pref.
PA 17 ll 2 0
PD 18 10 1 l
N-PA 22 8 0 0
N-PD 24 5 l O
2P paranoid schiZOphrenia; PS = paranoid state; M =

mixed; U undifferentiated schizophrenia; SR = schizophrenic
reaction; H = hebephrenic; C = catatonic; S = simple.

14

In summary, the PD and N—PD groups spent fewer days in
the hospital prior to Rorschach testing than did Aronson's
comparable groups; the N-PD group also had lower vocabulary
scores and less education than did the N-PA group. However,
there is little reason to suspect that the discrepancies
seriously limited the present study's attempted replication
of Aronson's study as the writer's paranoid and/or psychotic
group's responsiveness to Wheeler's and Reitzell's homosexual
signs were not significantly related to the forementioned
variables. Thus, as the investigator's subjects appeared
to be more closely selected and matched in accordance with
the procedures outlined by Arowmmthan.were those of Grauer,
the present study should provide a more adequate replication

of Aronson's study than that of Grauer's.

Rorschach Test

 

The Rorschach records of the subjects, each of whom
had previously been tested by a psychologist other than the
investigator, in the PD and N-PD groups were drawn from their
respective files, coded, and all identifying information
was removed. The protocols were then mixed according to a
table of random numbers.

Two judges, one of whom was the writer, independently
scored the protocols for Wheeler's twenty signs of homosexual—
ity plus one additional sign suggested by Reitzell. In
scoring the twenty-one signs of homosexuality for the sixty

protocols, 96% agreement was found between the two judges by

l5

dividing the sum of per cent agreement between the two judges
by the sum of per cent agreement plus the sum of per cent
disagreement between the two judges. For the total sample
each judge made approximately 1,260 judgments as to the

presence or absence of the twenty-one signs of homosexuality.

Statistical Analysis

 

As Cronbach (1949) has pointed out that the "sign
approach” on the Rorschach may result in "signs" being con-
founded with number of responses, the mean productivity of
the subjects in the PD, PA, N-PD, and N—PA groups was deter—
mined and the groups were compared on mean number of Rorschach
responses. If the groups do not differ significantly in mean
productivity on the Rorschach, they could legitimately be
compared on the mean absolute number of homosexual signs.
However, as Grauer did not report mean productivity for his
paranoid group (PG) no meaningful comparison could be made
between his PG group and the forementioned groups on mean
absolute number of homosexual signs. Both Aronson and Grauer,
however, reported the mean per cent of homosexual signs to
Rorschach R for their respective groups. Therefore, to
provide a more meaningful statistical analysis of the five
groups (PD, PA, PG, N-PD, N-PA), the per cent of homosexual
signs to Rorschach R was computed for each individual in the
PD and N-PD groups and the five groups were compared on mean

per cent of homosexual signs to Rorschach R.

16

The .05 level of statistical confidence was chosen for
the above analyses as it was felt that this significance
level would provide sufficient statistical confidence to
reflect meaningful differences between the paranoid and psy-
chotic groups if they existed. As differences in either
direction would be meaningful, a two-tailed test of signifi—

cance was used.

CHAPTER III
RESULTS

Table 4 shows that there were no significant differ-
ences between the PD, N-PD, and N-PA groups when compared
on mean productivity on the Rorschach test. However, when
the mean productivity of the PA group was compared with the
highest mean productivity of the forementioned groups, it
was found that Aronson's paranoid subjects gave significantly
more reSponses on the Rorschach than did the subjects in the

PD, N—PD, and N-PA groups.

TABLE 4

COMPARISONS 0F PARANOID AND PSYCHOTIC
GROUPS 0N RORSCHACH R (N=30)

 

 

Rorschach R

 

Group X 82 Comparison t df P
PA 33.9 396.8 PA VS. PD 2.31 58 .05
PD 22.4 356.8 PD vs. N-PD .96 36 N.S.

N—PA 22.2 84.6 PD vs. N-PD .05 36 N.S.

N-PD 18.8 83.9

 

Note: For t test with heterogeneity of variance,
Welch's correction for degrees of freedom (df) is applied.

17

18

Therefore, while the PD, N—PD, and N-PA groups could
legitimately be compared on the mean absolute number of
homosexual signs, any differences between these groups and
Aronson's paranoid group on mean absolute number of homo-
sexual signs could be attributed to the fact that the PA
group gave Significantly more responses to the Rorschach
than did the subjects in the other three groups. Moreover,
as Grauer did not report mean productivity, no comparison
on mean absolute number of homosexual signs was feasible.

Table 5 shows that there were no significant differ-
ences between the PD, N—PD, and N-PA groups on mean absolute

number of homosexual responses.

TABLE 5

COMPARISONS OF PARANOID AND PSYCHOTIC GROUPS ON
ABSOLUTE NUMBER OF RORSCHACH HOMOSEXUAL SIGNS

 

 

 

(N=3O)
Homosexual
Group Signs
'7 82 Comparison t df P
PD 3.03 8.99 PD vs. N—PD 1.64 58 N.S.
N-PD 1.93 5.17 PD vs. N—PA 1.68 50 N.S.
N—PA 1.93 3.84

 

Note: For t test with heterogeneity of variance,
Welch's correction for degrees of freedom (df) is applied.

As neither Aronson's nor Grauer's paranoid group

could legitimately be compared in the analysis of Table 5,

19

the five groups were compared on mean per cent of homosexual
signs to Rorschach R. Table 6 discloses that when the
writer's paranoid subjects and Aronson's nonparanoid subjects
were compared on mean per cent of homosexual responses to the
Rorschach, the difference barely attained the required .05
level of significance. However, the PD group was not found
to have a significantly higher mean per cent of homosexual
responses to the Rorschach than the N-PD group. Moreover,

no significant differences were found between Grauer's para-
noid subjects and the writer's and Aronson's nonparanoid

psychotic subjects on mean per cent of homosexual responses.

TABLE 6

COMPARISONS OF PARANOID AND PSYCHOTIC GROUPS ON
PER CENT OF HOMOSEXUAL SIGNS TO RORSCHACH R

 

HOmosexual
Signs

Group N Y% 82 Comparison t df P

 

 

PA 30 22.9 151.29 PD vs. N-PD 1.98 408 N.S.

PD 30 14.4 198.1 PD vs. N-PA 2.027 38 .05

PG 31 12.7 123.2 PG vs. N-PD 1.55 51 N.S.
N—PD 30 8.9 58.65 PG vs. N—PA 1.74 51 N.S.
N-PA 30 8.5 55.3

 

Note: For t test with heterogeneity of variance,
Welch's correction for degrees of freedom (df) is applied.

 

8Correction for matching Smn\/ (82ml + 82m2)(1-r2mx)

20

Table 7 Shows clearly that when the paranoid groups
of Aronson, Grauer, and the writer were compared on mean
per cent of homosexual signs to Rorschach R, there were no
significant differences between the PD and PG groups. How-
ever, the differences were highly significant when Aronson's
paranoid subjects were compared to the writer's and Grauer's

paranoid subjects.

TABLE 7

COMPARISONS OF ARONSON'S, GRAUER'S, AND DILLARD'S
PARANOID GROUPS ON PER CENT OF HOMOSEXUAL
SIGNS TO RORSCHACH R

 

 

 

Homosexual
Signs
Group N X% S2 Comparison t df P
PA 30 22.9 151.3 PA VS. PD 2.49 58 .02
PD 30 14.4 198 1 PA vs. PG 3.37 58 .01
PG 31 12.7 123.1 PD VS. PG .52 58 N.S.

 

Table 8 shows the frequency with which the twenty-one
Signs of homosexuality were selected by the subjects in the
total sample. It was found that Signs 2, 8, and 18 occurred
with equal frequency in both groups. Table 8 also reveals
that the psychotic patients responded more frequently to
Signs 16, 20, and 21 than did the paranoid patients; however,
the differences were not statistically significant as Shown

in Table 9. As in the original study, none of the signs

21

TABLE 8

FREQUENCY OF OCCURRENCE OF HOMOSEXUAL SIGNS (F)
ON THE RORSCHACH AND NUMBER OF INDIVIDUAL (n)
RESPONDING ONE OR MORE TIMES TO EACH SIGN

 

 

 

 

Homosexual Paranoid Sample Psychotic Sample
Sign F n F n
l 7 7 5 5
2 2 2 2
3 5 5 1 1
4 1 1 0 0
5 4 4 2 2
6 3 3 1 1
7 5 5 3 3
8 3 3 3 3
9 2 2 1 1
10 4 4 1 1
ll 4 4 2 2
l2 2 2 1 1
13 2 2 O O
14 4 4 l 1
l5 9 8 4 4
16 4 4 5 4
17 l 1 O O
18 2 2 2 2
19 20 8 10 5
20 2 2 5 4
21 > 5 4 8 8

 

Note: Signs 15 through 21 may be responded to more
than once.

22

occurred with great frequency except for Sign 19. Although
the paranoidsubjectsmresponded more frequently to Sign 19
than to any other sign, Table 9 shows that there were no
significant differences between the paranoid and psychotic
groups on responsiveness to Sign 19. Moreover, as the
greatest difference in frequency of response to homosexual
signs occurred between the paranoid and psychotic subjects
on Sign 19 and was nonsignificant, the remaining seventeen

signs were not subjected to statistical analysis.

TABLE 9

COMPARISONS 0F PARANOID AND PSYCHOTIC GROUPS ON
PER CENT MEAN RESPONSIVENESS T0 FOUR HOMOSEXUAL
SIGNS ON THE RORSCHACH (N=30)

 

Sign 16 Sign 19 Sign 20 Sign 21

 

 

 

Group x% S2 t x% 82 t x% 82 t x% S2 t

 

PD .13 .12 .4 .66 2.8 .9 .07 .07 1.05 .17 .21 .8
N-PD .17 .21 .33 .78 .17 .21 .27 .20

 

 

 

 

 

CHAPTER IV
DISCUSSION

The results reveal that while the matching between the
writer's paranoid and psychotic groups and Aronson's com-
parable groups was imperfect on three control variables
(intelligence, education, and length of hospitalization),
the subjects in each group of the present study were more
closely selected and matched in accordance with the procedure
of Aronson than were the paranoid subjects of Grauer. Con-
sequently, it appeared that the present study provided a more
satisfactory replication of Aronson's study than did that of
Grauer's.

When the mean number of Rorschach reSponses given by
the paranoid groups of the writer and Aronson were compared
it was found that Aronson's paranoid group gave significantly
more responses to the Rorschach than did the writer's com-
parable group. From the writer's experience in Rorschach
testing of paranoid subjects, it appeared that Aronson's
paranoid subjects were unusually productive on the Rorschach.
Rapaport (1946) and Hertz and Paolino (1960) have reported
that the mean number of Rorschach responses given by paranoid

subjects in their studies were respectively 20.85 and 26.7.

23

24

Thus, it appears that Aronson's paranoid group gave an
unusually large number of Rorschach responses. As Aronson's
paranoid subjects would have had a greater Opportunity to
respond to the homosexual signs, especially Signs 15 through
21, which may be scored more than once, this finding suggests
that the difference in responsiveness to homosexual signs be-
tween the writer's and Aronson's paranoid group may have been
due to Aronson having an atypical responsive paranoid sample.
Interestingly, Grauer also suggested that Aronson's paranoid
sample constituted an atypical sample as Aronson's twenty-
eight cases of paranoid schiZOphrenia in his sample of thirty
paranoid subjects represented only a small minority of all
the cases of paranoid schiZOphrenia he examined for his
paranoid group.

The results of the current study offered very limited
support for Aronson's results "that paranoid subjects report
an overwhelmingly greater number of homosexual signs on the
Rorschach test than do [psychotic controls]" (Aronson, 1952,

p. 409). In only one of four comparisons did a paranoid

group reSpond with a significantly greater frequency of homo-
sexual responses to the Rorschach than a matched group of
psychotics. It is important to note that the difference between
these groups only barely surpassed the critical minimum
value of t required for the .05 level of significance. However,
when the paranoid and nonparanoid psychotic groups of the writer
were compared, the PD group was not found to have a signifi-

cantly higher mean per cent of homosexual responses to the

25

Rorschach than the N-PD group. Moreover, the differences in
responsiveness to homosexual signs on the Rorschach between
the paranoid and psychotic groups in two other cross study
comparisons did not achieve the .05 level of significance.
The paranoid subjects in the three forementioned comparisons,
however, tended to respond more frequently to the homosexual
signs than did the psychotics. While these findings are
somewhat more consistent with those of Grauer than with those
of Aronson, the inability of this study to find consistent
results has two major implications.

1. While the difference in responsiveness to homo-
sexual signs between the writer's paranoid subjects and
Aronson's nonparanoid psychotic subjects attained the .05
level of significance, the finding that no significant dif-
ferences were obtained between paranoid and psychotic
subjects in three similar comparisons suggest that the dif-
ferences in responsiveness to homosexual signs are too
unreliable to serve as a satisfactory indication of the
association between homosexuality and paranoia. Therefore,
it appears that the phenomenon of the association between
homosexuality, as measured by homosexual signs, and paranoia
may be too gross to yield consistently reliable results.

2. The usefulness of homosexual signs in the diagno—
sis of paranoia is questionable as six signs (2, 8, 16, 18,
20, 21) occurred in the psychotic group with equal or greater

frequency than in the paranoid group; however, the differences

26

in frequency of response to those signs did not reach statis-
tical significance. Moreover, it was found that none of the
signs, taken individually, were able to significantly differ-
entiate between the paranoid and psychotic groups. While
these results are contrary to the findings of Wheeler and
Reitzell, they support the findings of Hooker (1958) who
found little justification for the continued use of Signs

1, 2, 3, 4, 5, 9, and 18.

While the use of Wheeler's and Reitzell's homosexual
signs as the rationale for the interpretation of homosexual
conflict on the Rorschach appears to have some supporting
evidence of validity from the studies of Wheeler (1949),
Reitzell (1949), and Ferracuti and Rizzo (1956), these writers
report that several of the signs are not entirely acceptable
as indicator of homosexual conflict. This finding received
support from the present study as the result revealed that the
use of the Wheeler and Reitzell signs of homosexuality as in-
dicators of homosexual conflict on the Rorschach appeared to
be too weak to be used effectively.

One interesting finding of the present study was the fact
that the paranoid groups of Aronson, Grauer, and the writer
consistently revealed marked individual differences with respect
to frequency of responding to homosexual signs on the Rorschach.
Approximately one-half (43%) of the writer's paranoid sub-

jects were below the nonparanoid psychotic (N-PD) group's

27

mean per cent of homosexual signs on the Rorschach. Moreover,
in the writer's paranoid group, there were four subjects who
did not respond to any of the twenty—one signs of homosexuality.
While the results are inconclusive, this finding suggest that
paranoid subjects may be quite heterogeneous with respect to
the etiological basis of their psychosis. These results are
consistent with those of Miller (1941) and Ovesey (1955) who
postulated that the paranoid mechanism may develop from
factors such as frustrations in life situations, dependency
needs and/or power needs rather than from sexual needs or
conflicts.

As the results of the current study offered very
limited support to the Freudian theory of paranoia, it
appeared from Rorschach evidence that the association between
homosexuality and paranoia is not as universally applicable
as implied by Freud. Finally, the inconclusiveness of the
results obtained on the Rorschach suggest the desirability
of using other methodological and theoretical approaches,
i.e. (Daston, 1956; Ovesey, 1955), to examine the Freudian

theory of paranoia.

CHAPTER V
SUMMARY

The present study represented an attempt to replicate
Aronson's findings that paranoids give a greater frequency
of "homosexual" responses on the Rorschach than matched
control groups. Sixty subjects, thirty paranoid psychotics
and thirty nonparanoid psychotics, were selected on the
basis of the extent of paranoid delusions in their clinical
picture. The subjects were closely matched for age, educa-
tion, intelligence, and length of hospitalization. The
Rorschach records of the sixty subjects were drawn from
their respective files and scored by two judges for the
twenty—one signs of homosexuality.

While the writer's paranoid group responded to a sig-
nificantly greater number of homosexual signs on the
Rorschach than did Aronson's nonparanoid psychotic group,
no significant differences in frequency of response to homo-
sexual signs were found between paranoid and psychotic sub-
jects on three similar comparisons. One reason which might,
in part, account for the lack of correspondence between the
paranoid subjects of the writer and Aronson on frequency of
responding to homosexual signs was the finding that Aronson's

paranoid sample gave an unusually large number of responses

28

29

to the Rorschach. AS Aronson's paranoid subjects would have
had a greater Opportunity to respond to the homosexual signs,
especially signs 15 through 21, which may be scored more than
once, the difference between the paranoid groups in frequency
of responding to homosexual signs may have been due to
Aronson having atypical responsive paranoid sample.

The inability of the present study to find consistent
significant results has two major implications. (a) The
phenomenon of the association between homosexuality, as
measured by homosexual signs, and paranoia may be too gross
to yield consistently reliable results. (b) The usefulness
of homosexual signs as the rationale for the interpretation
of homosexual conflict on the Rorschach test is questionable
as none of the signs, taken individually, were able to sig-
nificantly differentiate between the paranoid and nonparanoid
psychotic groups.

As the results offered very limited support to the
Freudian theory of paranoia, it appeared from Rorschach
evidence that the relationship between homosexuality and
paranoia is not as universally applicable as implied by

Freud.

REFERENCES

Aronson, M. L. A study of the Freudian theory of paranoia
by means of the Rorschach test. J. proj. Tech., 1952,
19, 397-411.

 

Cronbach, L. J. Statistical methods applied to Rorschach
scores: a review. Psychol. Bull., 1949, 16, 393-429.

 

Daston, P. Perception of homosexual words in paranoid
schiZOphrenia. Percept. mot. Skills, 1956, 6, 45-55.

 

Ferracuti, F. and Rizzo, G. B. Analisi del valore discrim—
inativo di alcuni segni di omosessualita rilevabili
attraverso tecniche proiettive, Bollettino d.
Psicologia e Sociologia Applicate, 1956, 13-16, 128-
134.

 

 

Freud, S. Psychoanalytic notes upon an autobiographical
account of a case of paranoia. In Collected Papers.
Vol. III, London: Hogarth, 1924.

 

Freud, S. and Breuer, J. Studienueber Hysterie. Leipzig
und Wien: F. Deuticke, I895.

 

Gardner, G. Evidences of homosexuality in one hundred and
twenty unanalyzed cases with paranoid content.
Psychoanal. Rev., 1931, 18, 57-61.

 

Grauer, D. Homosexuality in paranoid schiZOphrenia as
revealed by the Rorschach test. J. consult. Psychol.,
1954, 18, 459—462.

 

Hertz, Marguerite and Paolino, A. F. Rorschach indices of
perceptual and conceptual disorganization. J. proj.
Tech., 1960, 24, 370-388.

Hooker, Evelyn. Male homosexuality in the Rorschach.
J. proj. Tech., 1958, 22, 33-54.

 

Klein, H. and Horwitz, W. Psychosexual factors in the
paranoid phenomenon. Amer. J. Psychiat., 1949, 105,
687-701.

 

Miller, C. The paranoid syndrome. Arch. Neurol. Psychiat.,
1941,.45, 953-963.

30

31

Ovesey, L. Pseudohomosexuality, the paranoid mechanism,
and paranoia. Psychiatry, 1955, 18, 163-173.

 

Rapaport, D. Diagnostic psychological testing, Vol. II.
Chicago: Yearbook, 1946.

Reitzell, J. M. A comparative study of hysterics, homo-
sexuals, and alcoholics using content analysis of
Rorschach responses. J. proj. Tech., 1949, 13,
127-141.

Shockley, G. The role of homosexuality in the genesis of
aranoid conditions. Psychoanal. Rev., 1914, 1,
31—438.

Wechsler, D. Measurement of adult intelligence. Baltimore:
Williams and Wilkins, 1944.

 

Wechsler, D. The measurement and appraisal of adult 1
intelligence, (4th ed.)' Baltimore: Williams and
Wilkins, 1958.

 

Welch, B. L. Further note on Mrs. Aspin's tables and on
certain approximations to the tabled function.
Biometrika, 1949, 36, 293-296.

 

Wheeler, W. An analysis of Rorschach indices of male home-
sexuality. J. proj. Tech., 1949, 13, 97-126.

 

APPENDIX A

32

33

RATING SCALE

0--minimally delusional. Case record data reflects no
evidence of a delusional system in the patient's symptom-
atology.

l--very slightly delusional. Case record data reflects
negligible delusional content in the patient's symptom-
atology.

2--slightly delusional. Case record data reflects prelimin-
ary manifestations of delusional content in the patient's
symptomatology. Delusional content is expressed in the
record; however, it is sparse with respect to other
content.

3—-fairly delusional. Case record data reflects simple
delusional content in the patient's symptomatology.
Delusional content is defined, but limited; it does not
pervade throughout the record.

4—-markedly delusional. Case record data reflects distinct
delusional content in the patient's symptomatology.
Delusions are defined and the record reflects the
delusion's influence on the patient's behavior.

5--extremely delusional. Case record data reflects extensive
delusions in the patient's symptomatology. Delusional
content is unmistakable and pervades throughout the record.
Behavior is highly influenced by the patient's delusional

system.

34

6--maximally delusional. Case record data reflects compli-
cated and highly systematized delusions in the patient's
symptomatology. Delusional content predominates the

record.

APPENDIX B

36

NUMBER OF HOMOSEXUAL SIGNS (S), RORSCHACH RESPONSES

(R), AND PER CENT 0F SIGNS T0 RORSCHACH RESPONSES
FOR EACH INDIVIDUAL (%)

 

 

Paranoid Sample

 

Psychotic Sample

 

 

Subject 8 R % S R %
1 1 21 4.7 1 10 10.0
2 2 12 16.6 2 18 11.1
3 3 16 18.7 0 10 0.0
4 11 28 39.2 1 10 10.0
5 5 30 16.6 1 16 6.2
6 3 18 16.6 0 10 0.0
7 1 13 7.6 0 12 0.0
8 5 18 27.7 0 10 0.0
9 12 17 70.5 1 19 5.2
10 1 12 8.3 5 41 12.2
11 2 15 13.3 1 22 4.5
12 6 32 18.7 2 18 11.1
13 7 31 22.5 11 35 31.4
14 2 10 20.0 0 15 0.0
15 0 27 0.0 0 10 0.0
16 2 15 13.3 2 21 9.5
17 1 22 4.5 4 25 16.0
18 2 30 6.6 1 5 6.7
19 1 12 8.3 1 20 5.0
20 5 18 27.7 1 15 6.7
21 0 10 0.0 1 10 10.0
22 5 112 4.4 3 11 27.3
23 0 10 0.0 5 23 21.7
24 5 45 11.1 1 7 14.2
25 1 10 10.1 3 33 9.1
26 3 17 17.6 2 21 9.5
27 2 26 7.6 1 19 5.2
28. 2 13 15.3 2 32 6.2
29 1 20 5.0 1 17 5.8
30 0 14 0.0 5 38 13.1
Mean 3.03 22.4 14.4 1.93 18.8 8.9
Variance 8.99 356.8 198 1 5.17 83.9 58.65

 

APPENDIX C

37

38

LIST OF HOMOSEXUAL SIGNS SUGGESTED

BY WHEELER AND REITZELL

 

 

Sign Card Location Content

1 I W or W Mask or human or animal face

2 I Lower Male or muscular female torso
Center D

3 II Lower Crab or crab—like animal
Center D

4 III W or W Humans; with sex confused

5 111 W or W Humans; with sex uncertain

6 III W or W Animal or animal-like

7 IV W or W Human or animal; contorted

monstrous or threatening

8 V W or W or Human or humanized animal
Center D

9 VI Center or Objective; with implication
Top D of cleavage.

10 VII W or W or Human; female with derogatory
TOp D specification

11 VIII Lateral D Animal; several incogruous ones

or with incongruous parts

12 IX Upper Human; dehumanized
Lateral D

13 X Top Animal; attacking or fighting
Center D over central object

14 X All or Human; with blue as oral
Upper Half specification

15a Human or animal oral detail

16 Human or animal anal detail

or specification

 

39

LIST OF HOMOSEXUAL SIGNS SUGGESTED BY
WHEELER AND REITZELL--Continued

 

 

Sign Card Location Content

 

17 Human or animal described as
"back to back"

18 Human object or architecture;
with religious specification

19 Male or female genitalia
20 Feminine clothing
21b Household furnishings

 

Note: This list is based on a table by Wheeler (1949,
pp. 104-106).

8Signs 15 through 21 may occur on Card I to X; no
specific location is noted.

bReitzell's homosexual sign.

 

I“)

 

vi

 

 

 

”(1141(1))“ (11(7))(ilflLI141117S