University of Southern California Dissertations and Theses

First Steps In An Attempt To Construct An Objective Test Of Character Structure

(USC Thesis Other)

First Steps In An Attempt To Construct An Objective Test Of Character Structure

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content T his d isserta tio n has been 62— 3743
m icro film ed ex a ctly as rec eiv ed
RICHARDSON, O rv ille R od erick , 1 926-
FIRST STEPS IN AN ATTEM PT TO CONSTRUCT
AN OBJECTIVE TEST OF CHARACTER STRUC
TURE.
U n iv e r sity of Southern C aliforn ia
P h .D ., 1962
P sy ch o lo g y , c lin ic a l
University Microfilms, Inc., Ann Arbor, M ichigan
FIRST STEPS IN A N ATTEM PT TO CONSTRUCT
A N OBJECTIVE TEST O F CHARACTER STRUCTURE
by
O rville Roderick Richardson
A D issertation Presented to the
FACULTY O F THE GRADUATE SCHO OL
UNIVERSITY O F SOUTHERN CALIFORNIA
In P a rtia l Fulfillm ent of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Psychology)
January 1962
UNIVERSITY OF SOUTHERN CALIFORNIA
GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES 7, CALIFORNIA
This dissertation, written by
A. ...........
under the direction of h.U.—Dissertation Com
mittee, and approved by all its members, has
been presented to and accepted by the Graduate
School, in partial fulfillment of requirements
for the degree of
D O C T O R OF P H I L O S O P H Y
Dean
Date.
January 1962
DISSERTATION COMM!
Chairman
TABLE OF CONTENTS
Page
TABLES v
Chapter
I . INTRODUCTION A N D THEORETICAL ORIENTATION 1
I I . REVIEW OF CURRENT DIAGNOSTIC M ETH O DS 16
I I I . RATIONALE A N D DESCRIPTION OF THE
EXPERIMENTAL INSTRUM ENT 36
Id e n tific a tio n
The Guilt Scale
The "Ego-Ideal” Scale
The Object Relations Scale
The Gang Scale
IV. EXPERIMENTAL HYPOTHESES A N D
EXPERIMENTATION .......................................................... 56
Experimental Hypotheses
Well socialized Individual
Age, Sex, and Ethnic Peers
Normative Range of Puhitiveness
Cultural Peer
More Punitive Parent
Social Reaction to Delinquent Behavior .
Severity of A ntisocial Behavior
Selection of the ’'Well Socialized”
Sample
Administration of the Experimental
Instrument
Selection of the "Poorly Socialized”
Sample
Administration of the Experimental
Instrument
Patterns
V. PILOT -STU D Y
63
i i
Chapter
Item Analysis
Mean Differences
Relationship Between Test Scores
and Intelligen ce
True-False Items
Analysis of Mean Differences
Relationship Between In tellig en ce
and Test Responses
Findings with Regard to Other
Miscellaneous Test C haracteristics
VI. REVISION OF EXPERIMENTAL INSTRUMENT. .
Selection of the Poorly Socialized
Sample
C haracteristics of the Poorly
Socialized Sample
General C haracteristics of the
3 Experimental G roups
Test Administration of the Poorly
Socialized Sample
Selection of the Well Socialized
Sample
Test Administration of the Well
Socialized Sample
S ta t is t i c a l Procedures
Scoring
Mean Differences
C orrelational Procedures and Tests
of the Significance of Differences
Between C orrelations
VII. EXPERIMENTAL RESULTS ....................................
Hypothesis I
Hypothesis II
Hypothesis III
Hypothesis IV
Hypothesis V
Hypothesis VI
Hypothesis VII
Unhypothesized Relationships
Overview of Experimental Results
VIII. DISCUSSION CHAFTER . ,
IX. SU M M A R Y
REFERENCES .
APPENDIX
LIST O F TABLES
Table Page
1. Mean Scores in Punitiveness Ascribed
to Self, Mother, Father, and Most
Others. Delinquent and Nondelinquent
Sam ples.......................................................................... 71
2. Average Age in Months, Experimental Groups. . 85
3. Pearson Product-Moment Correlations
Between C T M M IQ Scores and Self Scores,
Criminal and Delinquent Items ............................ 86
4. Mean Scores in Degree of Punitivfeness
Ascribed to the Self, in Well Socialized
Good Prognosis and Poor Prognosis Groups. . 95
,5. Pearson Product-Moment r fs, Showing Cor
re la tio n Between the Degree- of Punitive
ness Ascribed to the Self and the Degree
of Punitiveness Ascribed to Each Parent
in Well Socialized-, Good Prognosis and
Poor Prognosis Groups ............................................ 9$
6 . The Significance of Differences in Correla
tio n Between Punitiveness Ascribed to
Self and Punitiveness Ascribed to Each
Parent, Comparing Well Socialized, Good
Prognosis and Poor Prognosis Samples. . . . 100
7. The Degree of Correlation Between Punitive
ness Ascribed to the Self and Punitiveness
Ascribed to the More ’’S tric t" Parent and
to the More "Permissive" Parent, in Well
Socialized, Good Prognosis and Poor
Prognosis Samples .......................................... 103
8 . Correlation Between Punitiveness Assigned
to Self and Punitiveness Assigned to
Each Parent, Where Only the 20 Highest
Scores (Mother and Father considered
Separately) in Parental Punitiveness and
the 20 Lowest Scores in Parental Puni
tiveness are Included in Calculations . . . 106
v
v i
Table Page
9. Group Means Representing the Degree
of Punitiveness Assigned to Each
Parent and to "Most Others" in Well
Socialized, Good Prognosis, and Poor
Prognosis Samples ..................................... 109
10. Average Sentence Years Assigned to Each
Test Item by Well Socialized and
Poorly Socialized Groups ..................................... I l l
11. Group Means Showing the Degree of
Punitiveness Assigned to Self and to
Each Parent in Well Socialized, Good
Prognosis, and Poor Prognosis Groups . . 114
12. The Significance of the Mean Differences
in Punitiveness Between Self and Each
Parent Within Well Socialized, Good
Prognosis, and Poor Prognosis
S am p les....................................................................... 115
13. Degree of C orrelation Between Punitiveness
Ascribed to Father and Punitiveness
Ascribed to Mother in Well Socialized,
Good Prognosis, and Poor Prognosis
Samples (Degree of Perceived Parental
Agreement in • Punitiveness Towards
Antisocial B e h a v i o r ) ....................................... U S
CHAPTER I
INTRODUCTION A N D THEORETICAL ORIENTATION
Among crim inologists and in the applied fie ld of
corrections, there is general recognition of a p ersistin g
lag in the development of a conceptual system and a metho
dology which would allow fo r more adequate diagnostic ind i
viduation of offenders and more d iffe re n tia te d treatment
programs. Individually oriented treatment programs focused
on the issues which distin g u ish one offender from another
do not develop because there is as yet no unified theory
of a n tiso c ia l behavior from which adequate methodologies
of individual diagnosis and treatment could spring.
In the absence of well validated knowledge about
the a n tis o c ia l personality, confused, contradictory, and
mutually encumbering mythologies of etiology and treatment
are entertained by those d ire c tly responsible fo r the con
t r o l , care, and possible re h a b ilita tio n of the offender.
Much in the same way th at the delinquent or criminal had
e a rlie r "slipped between" in d iffe re n t, re je c tin g , or
f u t i l e l y conflicted parents, he may pass through a correc
tio n a l program in an unengaged and fo r a ll p ra ctic al pur
poses "untouched" way. I t appears th a t th is kind of
lapse or default may be explained at le a st in part as a
re su lt of the in a b ility of the in s titu tio n a l "parents" to
- 1 -
find areas of conceptual agreement, and to generate an
effectiv e, homogeneous focussing of th e ir energies.
There is then a pressing need for continuing re
search e ffo rts which system atically te s t the myriad as
sumptions underlying a confusion of approaches in the cor
rectional fie ld and which attempt to delineate, explore,
and organize new areas of knowledge about the a n tiso c ia l
personality. In th is way p o te n tia lly f r u i t f u l lin e s of
e ffo rt are plotted and concentrated upon for more rapid
development while the "deadwood” of concept and practice
is allowed to pass into limbo.
In the present study an attempt is made to t e s t an
idea for a diagnostic t e s t designed to tap dimensions of
personality which a currently viable theory of a n tis o c ia l
behavior suggests to be highly relevant in understanding
and distinguishing between varying kinds and degrees of
crim inality. The theory has suggested a method which may
have p ra c tic a l applications in finding c lin ic a lly meaning
fu l differences between offenders characterized by ex
te rn a lly sim ilar an tiso c ia l behavior p attern s. The study
attempts to te s t the effectiveness of th is method.
In reviewing the lite r a t u r e for conceptual o rien ta
tio n i t was found th a t th eories of an tiso c ia l behavior may
be broadly c la ssifie d according to the re la tiv e emphasis
upon physico-hereditary, socioeconomic, cultural-en viron
mental, and psycho-dynamic fa c to rs 0
3
Physico-hereditary hypotheses have provided perhaps
the oldest and in some respects the hardiest explanations
for criminal behavior. Common to these theories is the
postulation of irre v ersib le co n stitu tio n al or early ac
quired defects of physique, m entality, and emotionality
which determine chronic, e sse n tia lly unmodifiable a n ti
social behavior. Lombroso’s ( 1 9 4 $) view of the criminal
as an a ta v is tic , genetic "throwback” to prim itive men and
his attempts to prove th a t criminals were distinguished by
physical stigmata appropriate to th e ir prim itive inheri-
tahce represents the most rad ical expression of the
physico-hereditary hypothesis. L atter day conceptual r e
la tiv e s of Lombroso may be discerned in Hooton (1939) and
in Sheldon (1949) who explain a n tiso c ia l behavior in terms
of physical and ra c ia l fa cto rs. Objective, experimental
support for such concepts is meager and equivocal. At
present the physico-hereditary hypothesis receives l i t t l e
emphasis in the thinking of crim inologists, at le a s t in
th is country, although explanations of psychopathy in
terms of co n stitu tio n al types continues to have i t s ad
herents in the fie ld of psychiatry (Michaels, 1955).
The hypothesis th a t an tiso cial behavior is a con
sequence of poverty and generally depressed liv in g condi
tions is a more charitable, and popular explanation for
criminal behavior. This view gains experimental support
4
from a number of studies the most well known being th at of
Shaw (1929), who demonstrated a sig n ific a n tly higher in c i
dence of a n tiso cial behavior in the more poverty stricken
deteriorated areas of Chicago than in the surrounding
middle class suburbs. The author, him self, was, however,
loath to postulate a d irect relatio nship between depressed
liv in g conditions and a n tiso c ia l behavior and subsequent
experimental work has tended to v itia te the hypothesis
th at depressed socioeconomic circumstances have a primary
role in the genesis of crime. Sutherland (l949), for
example, has shown an unexpectedly high incidence of cer
ta in forms of an tiso c ia l behavior in middle class groups,
while the Gluecks (1950) demonstrated th at the quality of
emotional in terac tio n in the primary family group could
d iffe re n tia te between delinquent and non-delinquent boys
liv in g under the same socioeconomic circumstances. Common
sense suggests that i f depressed socioeconomic circum
stances were a major determinant of criminal behavior
poverty and criminal tendencies would be p ra c tic a lly
synonymous. Neither experimental data nor common observa
tio n supports th is notion.
Cultural-environmental theories of criminal behavior
are best exemplified in the ”d iffe re n tia l associatio n”
theory of Sutherland (1947). Sutherland’s view simply ex
pressed is th a t a n tiso c ia l behavior occurs in individuals
who have been exposed more frequently to experiences with
crim inally oriented in d iv id u als than have law abiding
persons. This theory s tre s s e s the influence of the
so c ia l group on individu al behavior. Cohen (1955), de
scrib in g the e ffe c t of membership in a delinquent gang
shares th is o rie n ta tio n to a s ig n ific a n t degree although
tending to in te g ra te h is p o sitio n with psychodynamic views
of behavior.
The body of experimental evidence accumulating in
criminology, psychology, and re la te d f ie ld s strongly
suggests th a t the culture-environm ental point of view
while s tre s s in g an important area is at best a p a r tia l
and incomplete conceptualization of crim inal behavior,.
The Gluecks’ successful attempt to p red ict fu tu re a n t i
so cial behavior in young children, using only indices of
emotional in te ra c tio n in the primary family group tends to
support the contention th a t so c ia l and c u ltu ra l influences
outside the family may represent important but more in
d ire c t determinants of a n tis o c ia l behavior and th a t the
primary in v e stig a tiv e emphasis might well be d irected
towards the complex emotional in te ra c tio n in the immediate
family group.
Concepts which a t the present time appear r e la
tiv e ly more adequate in dealing with th is area are found
in the so -called psychodynamic th eo ries of behavior.
These systems have in common the view th a t behavior is to
be understood in terms of a "dynamic” and conflicted play
of biological needs in in teractio n with "outer" so c ia liz
ing forces. This process of socialization which is seen
as an in teg ral aspect of personality development is sig n i
fica n tly determined by the in teractio n between the child
and his parents. In the l a t e r theories, of which Erickson
is represen tative, the role of social and cu ltu ral factors
in th is fundamentally important in teractio n between child
and parents is woven into the process by conceiving of
the parents as the "carriers" or transm itters of prevail
ing sociocultural patterns and tensions. The sig n ifican t
etiolog ical factors in the genesis of criminal behavior
are seen by the psychodynamic th e o ris ts , however, to lie
in the v icissitu d es of the primary family relatio n sh ip s.
Later experiences are given sig n ific a n tly less weight.
A log ical view is th at the most adequate theories
of an tiso cial behavior w ill provide conceptual means for
organizing and unifying these broad, diverse explanatory
emphases which have at one time, or another been singled
out and accorded paramount e tio lo g ic al significance by
social th inkers. Evidence has been adduced for and against
each of these theories so th at unequivocal rejectio n of
one view in favor of another has been d i f f i c u l t . A fami
l i a r resolution of these d if f ic u ltie s is seen in the
statement th a t one must view criminal behavior as m ulti
determined, a re su ltan t of many diverse influences. This
approach is valid in so fa r as i t goes, but i s u n satis
factory as a th e o re tic a l concept because i t does not de
scribe the in teractio n of these influences in a sp ecific,
organized fashion. The view is too broad and loose to
provide testab le hypotheses.
The psychodynamic view which provides th eo re tica l
o rien tation in the present study appears to focus r e la
tiv e ly more d ire c tly but also more inclusively on the
individual offender. I t does not necessarily controvert
other positions and appears to have some capacity to in
corporate or unify the broader social facto rs as less
direct impinging influences shaping to a greater or le s se r
extent the quality of the interpersonal relationships
within the family. I t is possible to study the family and
the individual as the apex of a set of converging physi
co-hereditary, economic, and sociocultural determinants
and to apply psychodynamic concepts in understanding the
effects of these factors upon individual personality*
These concepts, although re la tiv e ly vague and u n arti
culated when compared to the concepts of physical science,
are seen here to be more specific and more amenable to
formal experimental exploration than the concepts of even
more loosely developed physico-hereditary and sociocul
tu ra l theories of an tiso cial behavior.
s
Of the psychodynamic theories the psychoanalytic
system, selected as the conceptual orientation of th is
study, manifests a more elaborated and d iffe re n tia te d
form. I t appears to be the most developed and organized
of current theories dealing with an tiso cial behavior and
i t s current v ia b ility may be judged from the amount of
in te re s t and effo rt now being devoted to testin g i t s
basic assumptions.
A b rie f discussion of the psychoanalytic view of
an tiso cial behavior may be organized around the id e n ti
fic a tio n and superego c o n c e p t s P s y c h o a n a l y t ic theory
sp ecifies an in teractio n between co n stitu tio n al, in te r
personal, and cu ltural facto rs which in normal person
a lity development leads to progressively increasing capa
city for the delay, mastery, and modification of strong,
bio logically given, unsocialized impulses. The individual
v icissitu d e s of th is in terac tio n are seen as important de
terminants of the personality q u a litie s and ch aracteris
t ic s which d iffe re n tia te one mature personality from
another. The biological inheritance, the quality of the
interpersonal experience, and broader cultural factors
^The w riter draws heavily on a single resource:
Otto Fenichel, The Psychoanalytic Theory of Neuroses
(New York: W . W . Norton and Company, 1945), Chaps. 4, 5,
6 . Pp. 33-102.
9
exert e ffects through the child-parent relation sh ip which
is given status as the most sig n ifican t single determinant
of individual personality.
The term id e n tific a ti on is used to cover a ll of the
many instances of interpersonal behavior in which one per
son adopts an aspect of another's personality and absorbs
i t as his own. The id e n tific a tio n concept carries the
burden of accounting for a l l of those sig n ifican t person
a l i t y modifications which occur as a re su lt of one per
son's influence upon another.
An important aspect of the id e n tific a tio n mechanism
is th a t through feeling "as one" with another person,
i n i t i a l l y the parents, the child gets rid of disturbing
feelings having to do with a developing awareness of se
paration from others with a concomitant awareness of iso
la tio n , dependency, and v u ln erab ility . Later experiences
of fru s tra tio n accompanied by feelings of helplessness to
cope with one’s own needs may activate th is "taking in" of
the q u a litie s and a ttrib u te s of other p erso n a lities.
After infancy the taking of aspects of other personalities
as a part of the s e lf may occur as an attempt to reduce
psychological stress based on very complicated in terp er
sonal s itu a tio n s. I t is the psychoanalytic view that
what one is as an individual is determined largely by the
nature of one’s id e n tific a tio n s, i . e . , what one has taken
10
in from other personalities and what dynamic organization
arrangement or structuring of these segments has occurred.
Id e n tifica tio n with warm, nurturing, and adequately so
cialized parental figures is seen as a basic developmental
step which must occur i f the child is to master his unruly
instinctu al striv in g s.
The id en tificatio n s most decisive for the la te r
social adjustment of the child are said to occur during
the c r itic a l oedipal phase of personality development,
when the superego takes on a more chrystallized form. A
detailed exposition of psychoanalytic theory with regard
to the Oedipus-complex w ill not be undertaken here.
Familiar to the reader is the proposition th at in the la s t
phase of psychosexual development, placed approximately in
the years four to six, the focus of sexual energy is d i
rected toward the parent of the opposite sex. There is
said to be a desire to "possess’* th is parent, to have ex
clusive rights to the parent’s affections and attentio ns.
At the same time there arises an intense feeling of rivalry
with and h o s tility toward the parent of the same sex. This
parental riv al assumes the proportions of a fru stra to r of
considerable magnitude. The child is placed in a very
d iffic u lt position. Successful competition with the
parental riv al is not possible. The child is not power
ful enough to best his riv a l nor to command the degree
11
of g r a tif ic a tio n required from the desired parent. The
t o t a l discomfort of the situ a tio n is increased by a pro
je c tio n of fe a rfu lly powerful, punishing q u a litie s on the
r iv a l parent.
Normally the Oedipus-complex is resolved because
the c h ild 1s fe ars with regard to the parent of the same sex
become more s a lie n t than h is desire fo r g r a tif ic a tio n from
the parent of the opposite sex. As in other s itu a tio n s of
in to le ra b le th re a t or fr u s tr a tio n the id e n tific a tio n me
chanism comes into play. By id en tify in g with his own con
ception of the riv a l p a re n tTs sanctions and p rohib itions
the child is enabled to shed his anxiety provoking sta tu s
as a weak antagonist in danger of some t o t a l th r e a t. In a
sense h is former riv a l becomes a part of him, the two are
no longer opposed but joined.
This c r i t i c a l id e n tif ic a tio n with the parent of
the opposite sex in the course of the reso lu tio n of the
Oedipus-complex r e s u lts in highly s ig n ific a n t and so c ia lly
useful changes in person., l i t y . Impulse control, formerly
dependent upon fe ar of parental censor in the immediate,
concrete s itu a tio n is now b u ttressed by an "in n er” system
of pro hibition s which regulates behavior whether the par
ent is present or not. The superego, i . e . , the i n t e r
nalized parental "image” must be s a tis f ie d in order for
the child to be comfortable with him self. Behaving in a
12
manner contrary to the d ictates of the superego creates
intense g u ilt which is as painful and disturbing as the
parental disapproval was formerly. ”Good” behavior on
the other hand re su lts in a benign feeling of se lf worth
which the p aren t’s affection and approval E licited
formerly. Being on good terms with the superego becomes
as strong a motive as any of the basic in stin c tu a l s t r i v
ings. There is now a b u ilt- in regulator of behavior,
modeled in i t s own individual ch a ra c te ristic s by the
c h ild ’s image of his parental figu res, which operates
constantly and independently of the external s itu a tio n .
The resolution of the Oedipus-complex with the
establishment of a ’’healthy” superego is seen in psycho
analytic theory as the end product of benign maturational
sequences. The normal superego appears as the flower of
an optimal growth process, and is the sine qua non of well
regulated, socially adaptive behavior in the individual.
Chronic an tiso c ia l behavior which is not d irec tly
traceable to co n stitu tio n al or physiological factors
is explained by psychoanalysts in terms of disturbance in
those personality processes leading to normal superego
development. A p a r tia l l i s t of factors seen by psycho
analysts as typ ical causes of disruption in normal super
ego development and re su ltan t a n tiso cial behavior is
given below.
13
1. Severe disruption of the primary mother-child
re la tio n sh ip through death, prolonged absence, or rad ical
neglect (Spitz, 1959) (Rabinovitch, 1959).
2. Pathologically severe re je c tio n by the mother
(Kaufman, 1957).
3. Reading in an in s titu tio n a l environment without
individual nurturing in an enduring one to one re la tio n
ship (Spitz, 1959).
4. Transiency. Frequent change of nurturing
figure (Aichorn, 1957).
5. Inherited lack of capacity fo r object r e la
tio nships (Glover, I960).
6 . Id e n tific a tio n with h o s tile , s a d istic parental
figures (Glover, I960).
7. Id e n tific a tio n with weak, unpredictable, or
criminal parental figures (Freud, A., 1949).
S. Id e n tific a tio n with parental figures holding
a n tiso c ia l values (Aichorn, 1937).
9. Overly strong id e n tific a tio n with the parent
of th e opposite sex with reactive pseudo-masculinity
(Fenichel, 1945).
This representative l i s t i l l u s t r a t e s the psycho
analytic view th a t well socialized behavior depends upon
adequate object re la tio n sh ip s and strong id e n tific a tio n
with individuals who are themselves comfortable with
social demands. The non-criminal personality has developed
through re lia b le , emotionally supportive but at the same
time wisely rewarding and punishing, impulse regulating
relatio n sh ip s. The decisive id en tificatio n s have been
with parents whose own superegos enabled them to adapt
comfortably to prevailing social sanctions and prohibi
tio n s. A benign superego development insures th a t the
’’inner" m oral-ethical system w ill be strong enough and
adaptive enough to regulate those impulses which need to
be regulated in the p a rtic u la r settin g , while the a n ti
social personality is distinguished by the absence, incom
plete development,or pathological functioning of th is
inner censoring system. To the degree that the system is
fau lty , absent,or pathological in character the individual
can be expected to demonstrate q u a litie s of in fa n tile judg
ment, im pulsivity, and emotionality, q u a litie s which lead
sooner or l a te r to d iffic u lty with external authority.
Inherent in the above discussion is the p o ss ib ility
that by manipulating the concepts of id e n tific a tio n and
superego one may distinguish in a formal way between
various degrees, types, and forms of superego disturbance.
Attempts at such d iffe re n tia tio n within the group of a n ti
social p erson alities have been made by Aichorn (1937),
Glover (I960), and Karpman (1949) among others. Among cur
rent theories of criminal behavior psychoanalytic theory is
unique in i t s provision of p o s s ib ilitie s for a conceptual
15
d iffe re n tia tio n of the a n tiso cial group based on genotypi
cal concepts. The theory suggests th a t i f one is able to
assess individuals as to the quality and character of
th e ir object relatio n sh ip s, specify important id e n tific a
tio n s, and discern the q uality and character of the func
tioning superego, c lin ic a lly , meaningful discrim ination
should be possible among individuals displaying pheno-
ty p ically sim ilar an tiso c ia l behavior.
The problem th a t presents i t s e l f is to find means
of casting these provocative concepts into a form lending
i t s e l f to a process of experimental exploration and
appraisal. As a te n ta tiv e and exploratory e ffo rt with
these general goals in mind the present research sets the
task of employing the psychoanalytic concept of the super
ego and i t s re la te d concept id e n tific a tio n as conceptual
guides in the construction of an objective t e s t of
’’character s tru c tu re .” The aim is simultaneously th a t of
developing an instrument by means of which th e o re tic a l con
cepts may be tested through orthodox experimental and
s t a t i s t i c a l procedures and making a beginning step toward
a diagnostic instrument based lo g ically upon a theory of
human behavior.
In the following section a review and critiq u e of
current diagnostic instruments in popular use in the di
agnosis of offenders w ill be undertaken as an o rien tatio n
to the description of the experimental instrument.
CH APTER I I
REVIEW O F CURRENT DIAGNOSTIC M ETHODS
The aim of several research en terp rises over the
past th ir ty years has been to develop techniques fo r the
accurate and objective assessment of ’’character1 ’ dimen
sions of personality. Studies of th is nature have been
based on the assumption th at individuals carry within
themselves re lia b le response tendencies which determine
behavior in situ a tio n s of moral choice. A v ariety of
approaches have been employed in the attempt to select
the relevant personality dimensions, to design sen sitiv e
measurement techniques and to predict m oral-ethical
behavior from them.
Hartshorne and May (1930) engaged in an early, sys
tematic ap plicatio n of experimental techniques to the pro
blem of character assessment. Following Galton’s reason
ing th a t actual behavior was the best index of ’'character,”
they employed miniature behaviorial s itu a tio n s which en
abled them to observe the in d iv id u a l’s behavior in s itu a
tio n s of moral choice. They were in te re ste d in te s tin g
whether subjects would tend to display a consistent pat
te rn of m oral-ethical behavio’r in a v ariety of situ atio n s
and whether th is p attern would agree well with character
estim ates obtained from teachers and peers of the subjects.
- 16-
17
I t was found th a t the subjects showed a great deal of
v a ria b ility in th e ir m oral-ethical behavior from situ atio n
to situ a tio n , th is finding tending to contradict the hypo
thesized existence of unitary moraj. t r a i t s . The fact th a t
teachers and peers of the subjects displayed a high degree
of agreement in th e ir estimates of the ind iv id u al’s
"character” led Hartshorne and May to conclude th at repu
ta tio n was the best and most economical single index of
character development.
Havighurst and Taba (1949) correlated a number of
social and intrapersonal variables with "character repu
ta tio n scores obtained through the use of especially de
signed sociometric techniques. Estimates of the "charac
ter" of each adolescent subject were obtained from members
of his peer group and those community leaders in close
contact with him. A cogent point was made in the au th o rs’
report that subjects having d iffe rin g personality organi
zations could nevertheless have very sim ilar "character
reputations." This indication that in tra-in d iv id u al per
sonality dimensions are not congruent with consensual ex
ternal social estimates of character provides a basis for
rejectio n of "reputation scores" as useful indices of
in tra-in d iv id u al character dimensions. Such scores, how
ever, may well be useful in predicting external behavior.
In the above discussion we are touching upon a d is
tin c tio n which w ill be made repeatedly as methods employed
currently in the diagnosis and c la s s ific a tio n of offenders
are discussed. The d istin c tio n between causal and descrip
tiv e approaches to personality is made by Allport (1937),
who employs the Lewinian terms "Genotypical” fo r the causal
and h is to ric a l in te rp re ta tio n of behavior, and "Phenotypi
cal" fo r the descriptive, acausal and a h isto ric a l c la s s i
fic a tio n of external behavior. The position w ill be taken
in th is discussion th at while the phenotypical approach
may prove useful in a p ra ctic al empirical way, i t is the
genotypical approach th a t is p o ten tially the most promis
ing line to take where the goal is establishing meaningful
dimensions of "difference" between individuals who may be
long to a group which is homogeneous from some phenotypical
viewpoint, e .g ., crim inals.
Over the l a s t 25 or 30 years a number of researchers
have employed one or another of the so-called "objective,"
" tru e -fa lse ," or "paper and pencil" te s t s of personality,
in an attempt to discover dimensions of personality which
would serve to d iffe re n tia te between chronic offenders
against the law and th e ir law abiding peers. The success
of these e ffo rts is best estimated from a report by
Schuessler and Cressey (1950), who reviewed studies in
which a to ta l of 30 of the best known personality te s t s
developed before 1945 were employed in attempts to estab
lish personality differences between criminal and
non-criminal populations. Their review of these 30 studies
led them to conclude that personality t r a i t s are d i s t r i
buted in the criminal populations in about the same way as
in the general population. One must conclude then e ith e r
th a t individuals heavily engaged in criminal behavior over
long periods of th e ir liv e s are not d iffe re n t in person
a lity make-up from so cially adaptive persons or th a t these
early pencil and paper, tru e -fa ls e items were not gettin g
at the relevant dimensions. The b e tte r success of l a t e r ,
more soph isticated te s t s indicates th a t the l a t t e r a l t e r
native is more lik e ly .
The so-called projective techniques, notably the
Rorschach inkblot t e s t and th e Thematic Apperception Test
are currently extensively used in the diagnostic evalua
tio n of criminal offenders. The strength of the projec
tiv e technique l ie s in i t s capacity to e l i c i t data of a
highly individual, personal character from subjects.
These te s t s are very popular with professional clin ician s
who see them as sen sitiv e , "deep-going” instruments with
out which a tru ly intra-personal conceptualization of the
personality would not be possible except through extensive,
time consuming contact with the personality involved.
This investment in the projective technique as the
diagnostic instrument of choice continues despite p e rsis
tin g d iffic u lty in the delineation and adequate v alid atio n
of the in te rp re ta tiv e dimensions of the various techniques.
20
In contrast to the devices considered thus far,
there are several instruments now in use which do achieve
a level of discrim ination between offenders and nonoffend
e rs. The criticism of these techniques re la tiv e to the
goals of the present project w ill be th at the differences
obtained are not in terp re ta b le in terms of any organized
view of human behavior and so remain on a concrete, em
p iric a l level not applicable to any p a rtic u la r hypothesis
and so unamenable to the process of hypothesis testin g ,
v e rific a tio n , or re jectio n and revision which is essen tial
to the acquisition of precise knowledge. In actual prac
tic e the re su lts obtained serve only rough screening needs
in d iffe re n tia tin g poorly socialized individuals from
th e ir more adaptive peers and do not seem to have extended
our understanding of an tiso c ia l behavior.
The Porteus Maze (Doctor, 1954) is one such device
which in practice serves rath er effectiv ely to discrim inate
between p erso n alities inclined to break the law and those
not so inclined. The t e s t involves a series of pencil and
paper maze problems. The subject is rated for an tiso cial
tendencies not on the number or complexity of mazes ade
quately handled (the in tellig en ce index) but rath er on
q u alitativ e features. E ssentially th is scoring system
highlights features of im pulsivity, carelessness, disregard
of in stru ctio n s, overconfidence, sloppy but adequate
21
performance, and so on. The q u a lita tiv e score is only
s lig h tly re la te d to o v erall in te lle c tu a l a b ility and seems
to be a rath er good measure of a co n stellatio n of be-
havorial t r a i t s which is highly correlated with a histo ry
of a n tis o c ia l behavior, but re la tiv e ly independent of
in te lle c tu a l fa cto rs.
The c ritic ism leveled here is th a t th is single
score index of something th a t could be called "impulsive
ness” tends to leave unattended the p o s s ib ility of sig n i
fic a n t individual differences which may obtain between
delinquent individuals possessing sim ilar "impulsiveness"
indexes. One may also object to the im plicit assumption
th a t more or le s s "impulsiveness" is synonymous with more
or le ss crim in ality . To say th a t an individual has a
re la tiv e ly high or re la tiv e ly low score on a measure of
carelessness and a general "slapdash" approach to a task
re a lly provides no more and no le ss diagnostic information
than to say that he has few or many court appearances.
Both indices are s ig n ific a n tly associated with an tiso c ia l
behavior ju st as high temperatures are associated with
appendicitis and sc a rle t fever. Just as more d iffe re n
tia te d diagnostic information than the thermometer affords
is needed to d istingu ish between appendicitis and sc a rle t
fever so do we need more d iffe re n tia te d diagnostic con
cepts and methods to id e n tify conditions underlying
22
symptomatic delinquent behavior. The single score, single
concept te s t yields information at the level of the c l i n i
cal thermometer. W e can divine that the patient is sick;
we s t i l l have to ask, "What's the m atter with him?” With
the single score of lim ited connotative range we are con
fined to judgements of degree along a dimension which is
too narrow in terms of meaning tending to seduce the
c lin ic ia n into the logical e rro r of equating the score or
measured dimension with the condition i t s e l f .
An instrument which achieves a good level of d is
crimination between delinquent and non-delinquent groups
is the Minnesota Multiphasic Personality Inventory.
Studies in recent years have demonstrated th at th is t e s t
is sensitive enough to predict future delinquent behavior
in children several years before th e ir delinquent behavior
becomes a m atter of o f f ic ia l record (Hathaway, 1952). The
t e s t has also proven useful in selectin g those individuals
within the larg e r delinquent group whose delinquent be
havior may prove to be more chronic or severe.
The M M PI stands as the prime example and model
achievement of the pragmatic, em pirical, non-theoretical
approach to the construction of psychological te s ts . As
is well known the items are e sse n tia lly common psychiatric
symptoms. In the construction of the te s t a large pool of
these symptom-statements was b u ilt up and given to ra th e r
23
carefully selected c rite rio n groups, each group consisting
of patients fa llin g into one or another psychiatric diag
nostic category, e .g ., depressives might constitute one
group, hypochondriacal patients another, and so on. The
patient indicates for each item whether or not the sym
ptom- statement is true or not true of himself. The fin a l
t e s t is made up of those items from the large pool which
did serve to d iffe re n tia te between the c rite rio n groups,
i . e . , a l l items which depressive patients answered in a
way which was c h a ra c te ristic fo r th is group but uncharac
t e r i s t i c for the other c rite rio n groups were included in
the"depression” scale of the t e s t . The unique feature
here is th a t the items included in any p a rtic u la r c lin ic a l
scale, i . e . , depressive, paranoid, schizophrenic, etc.
were not necessarily the items th a t in terms of ’’manifest
content” might be expected to be c h a ra c te ristic of a
p a rtic u la r group. An item which might look lik e a s ta te
ment symptomatic of a depressive s ta te might ’’turn out”
em pirically to be accepted more often in the paranoid
c rite rio n group. Despite i t s seeming depressive connota
tions th is item would become a paranoid item and would be
included in the paranoid scale of the t e s t . The question
of why the item d iffe re n tia te d between c rite rio n groups
was not asked; i t was enough to find th a t in practice the
various scales of the te s t were effective in identifying
2k
persons suffering from the psychiatric conditions which
characterized individuals in the various c rite rio n groups
employed in constructing the t e s t . Paranoid schizophrenics
tend to respond to the items in the same way th at the o r i
ginal c rite rio n group of paranoid schizophrenic did.
One may ask; what are the possible uses and the
lim its to the usefulness of an instrument of th is empiri
cal, pragmatic, and ath eo retical nature. The c r it i c a l
view taken here is th a t because the t e s t is based on
symptoms and nosological categories, te s t re su lts can
never be useful at other than a d escriptive, c la s s ific a -
tory lev el, which can be no more refined than were the con
cepts and categories used in i t s i n i t i a l construction. W e
know from the su b je c t's responses when they correspond to
those of the majority of respondents in a p a rtic u la r
diagnostic category th at his behavior and condition may
well be sim ilar in many ways to what we know of the be
havior and condition of individuals in the orig inal c r i
terio n group. W e have placed him in a diagnostic group,
e .g ., schizophrenic, psychopathic, or what have you. The
value of th is diagnostic placement depends wholly upon how
specific and refined our knowledge is of the ’’condition"
to which the subject has now been assigned. W e know as
much about him and as l i t t l e as we know about the in d i
viduals in the original c rite rio n groups. I f schizo
phrenics or psychopaths are a l l alike in. th e ir essential
25
c h a ra c te ristic s and we know a l l about these c h a ra c te ristic s
than the diagnostic job is finished when the person has
been assigned to one group or another.
W e know n e ith e r schizophrenics nor psychopaths are
a l l a lik e , there is much th a t we don’t know about both
conditions. The best evidence in dicates th a t both schizo
phrenia and psychopathy represent ’’wastebasket” rubrics
under which a great v ariety of heterogeneous psychiatric
conditions are subsumed. The objection here is th a t the
t e s t places the person in a general group which may be
quite heterogeneous in many ways and about which we know
re la tiv e ly l i t t l e (Astin, 1959). The te s t t e l l s us l i t t l e
about how th is schizophrenic or psychopath d iffe rs from
another one nor does i t seem to aid in re fin in g our know-,
ledge of the condition i t s e l f . The fa ct th a t the t e s t is
based on symptoms and th a t i t has no basis in any p a r ti
cular theory of behavior has the effect of r e s tr ic tin g the
information i t yields and the deductions possible from th is
information to a level of inference which is no higher and
no more refined than th at which i s possible from the crude
nosological system upon which the te s t is based.
The point of view th a t emphasis on unidimensional,
phenotypical, and empirical c h a ra c te ris tic s in current
diagnostic instruments tends to in h ib it the development
of more d iffe re n tia te d , more fin e ly focused diagnostic
concepts and methods has been taken in the above discus
sion. A diagnostic technique which is more d if f ic u lt to
characterize in empirical, n o n -ratio n al, symptom-oriented
terms is the Qlueck Social Prediction Scale. In an exten
sive experimental in vestigation the Gluecks were able to
establish sig n ifican t personality, so cial, and physical
differences between a large group of delinquent boys and
a carefully matched group of non-delinquents (Glueck, 1950).
An important byproduct of th is study was the development
and statement of five social factors which taken in com
bination achieve a high level of discrim ination between
delinquent and non-delinquent groups. Several independent
studies have tended to support the contention of the
Gluecks th a t th e ir facto rs prove effectiv e in prediction
of future delinquency in children, years before there is
any o f f ic ia l involvement with a u th o ritie s (Glueck, 1959).
The five social fa cto rs prove effectiv e as a predictive
device at every socioeconomic level and without regard to
the fa cto rs of race and in tellig en ce (Thompson, 1952).
These findings suggest that the five social facto rs are
tapping a common denominator in delinquent p erso n a lities;
a level which cuts across or under facto rs taken as
primary by theories of a n tiso cial behavior which stress
socioeconomic and ethnic facto rs as primary in the causa
tion of an tiso cial behavior.
27
A review of the five factors which make up the con
tent of the Glueck scale shows th a t they focus upon the
quality of interpersonal relationship s and the quality of
d iscipline in the primary family group. The five factors
are: 1. Discipline of the boy by the fath er, 2. Super
vision of the boy by the mother, 3. Affection of the
fath er for the boy, 4. Affection of mother for the boy,
5. Cohesiveness of the family u n it.
The subject is placed in one or another descrip
tiv e category under each of the five main headings. Rat
ings on each facto r must be done by experienced workers on
the basis of interviews with the primary family group or
careful analysis of a case h isto ry which must provide the
information required to .make judgements on the five fac
to rs. The subject receives a to ta l score calculated by
adding the weights assigned by the Gluecks to each sub
category under the five main fa cto rs. The weights are
simply the percentages of the original delinquent sample
judged to f a ll into each subcategory. The weights were
added to the to ta l score and the t o t a l range of the
original group divided into score in te rv a ls. The per
centage becomes ’’the chances for future delinquent beha
v io r” of any individual whose to ta l weighted score f a l l s
into a p a rtic u la r score in te rv a l.
I t is a sig n ifican t finding th a t such a score based
on family d iscip lin e and family relatio n sh ip s, quite inde
pendent of factors external to the immediate family group
proves to have predictive power exceeding th a t of psychia
t r i c or psychological examination or of professional opin
ion based on complete knowledge of the case history
(Thompson, 1952). In terms of the present discussion a
major positive feature of the scale is seen in the fact
that the information the scale provides lends i t s e l f to
etio lo g ic al or genotypical forms of in te rp re ta tio n , and
may be re la te d to existing concepts and hypotheses with
regard to human behavior. The i n i t i a l findings, for
example, tend to be applicable to psychodynamic hypotheses
with regard to the so c ia liz a tio n process. Theoretical
positions which minimize the effect of the family i n te r
action tend to be weakened. The method has thus not
only proved effectiv e as a screening device but tends to
generate progress in the process of hypothesis v e rific a
tio n and re je ctio n which is so important to the develop
ment of re lia b le knowledge. The point is made here th a t,
th is becomes possible because the "score” obtained with the
Glueck scale has meaning (poor family relatio nships and
poor d iscipline) that is more extensible and has a b e tte r
quality of face v a lid ity than say the "impulsiveness"
score of the Porteus Maze or the "psychopathic deviate"
scale of the MMPI.
29
I t is possible to c r it ic i z e the Glueck technique on
the basis of unidimensionality. Possible important d if
ferences on a stru c tu ra l level which ex ist between in d iv i
duals obtaining sim ilar scores w ill be obscured by the
single t o t a l score. A second c r i t i c a l point involves an
element of cumbersomeness and d iffic u lty in obtaining the
data which the prediction score requ ires. Careful h isto ry
taking involving personal contact between a train ed worker
and the primary family group is the optimal condition fo r
obtaining re lia b le data. A good deal of time and s k i l l
with a large p oten tial for subjective e rro r in evaluation
of fam ilies is involved in the process of obtaining the
to ta l score under ordinary casework conditions.
A sophisticated attempt to adapt the tr u e - f a ls e ,
”objective" type of t e s t item to the d i f f i c u l t task of
iso latin g and measuring diagnostically sig n ific a n t dimen
sions of difference between offenders and nonoffenders and
more importantly for correctional purposes, establishin g
important dimensions of d iffe re n tia tio n within ghfe
population of offenders is represented by the Gough
Socialization Scale (Gough, 1954). This instrument is
unique among extant diagnostic instruments in th a t the
author combined a th e o re tic a l and empirical approach in
the development of his items and in subsequent attempts
to v alid ate the scale. I t is asserted by the author th a t
the scale measures a single, more or le ss , unitary dimen
sion which is of cardinal importance as a determinant of
an tiso cial behavior. This single t r a i t or capacity, the
a b ility to place oneself in the ’’role of the other” stems
from Mead’s th e o re tic a l speculations with regard to the
personality of the criminal offender. In th is th e o re tic a l
framework the a b ility to adapt to the social structure
depends d ire c tly on the extent to which one is able to ap
prehend the expectations of others with regard to oneself,
and to respond to those expectations fo r the se lf as
though they were a part of the s e lf . As Mead puts i t , the
individual can be characterized as well socialized when:
. . . . he acts towards himself in a manner analo
gous to th a t in which he acts towards others, taking
the role, gestures, and speech of the other in re s
ponding to i t himself as the other would do.l
The criminal can’t experience social emotions and cannot
respond in a so cially adaptive way because he cannot per
ceive, hence cannot fe e l the reaction of ’’others” to his
own behavior.
This th e o re tic a l formulation of criminal behavior,
the formal expression of the p oet’s wish to ’’see oneself
as others see u s ,” oriented Gough (194-9) in the fashioning
of a pool of tru e -fa lse items. The items were given to
-^-George H. Mead, "A B ehavioristic Account of the
S ignificant Symbol,” Journal of Philosophy. 1922, 19.160.
experimental subjects varying in degree of so c ia liz a tio n ,
and the discrim inating items saved fo r the fin a l form of
the instrument. The refined t e s t proves successful in
placing various c rite rio n groups, varying in adjustment
to social pro scrip tio n s, in correct place along a con
tinuum of "social m aturity" or "degree of s o c ia liz a tio n ."
The c rite rio n groups, ranging from nonoffenders of d i f
feren t ages, having so c ia lly sanctioned occupational h is
to r ie s and excellent social reputations to "hard core"
chronic offenders, are well selected and the attempted
gradation of "^degree of so cializatio n " has high face
c r e d i ta b il i t y . Validation stud ies, conducted by the
author (1954) and by independent researchers as well
(Reed, 1957), give impressive experimental support for the
conceptual v a lid ity of the instrument and to i t s empirical
efficien cy in d iffe re n tia tin g between noncriminals and
offenders, characterized by varying degrees of a n tiso c ia l
involvement.
In terms of the stated goals of the present study,
th is instrument and the experimental approach from which
i t stems is seen to have much value. The researchers
have arrived at a well specified dimension, based on an
e tio lo g ic a l concept with good co nno tative-theoretical pro
p e rtie s the approach tending to fu rth e r knowledge and
generate additional hypotheses and lin es of research.
There is strong pragmatic value in th at the t e s t can be
immediately useful as an aid in diagnosing the "se v e rity ”
of a p a rtic u la r an tiso c ia l behavior p a ttern .
The Gough S ocialization Scale and i t s conceptual
foundation are c ritic iz e d here on the grounds of "unidi
m ensionality." The concept of social roles and the ca
pacity to "role play" while comprehending what appears to
be an important area in the so cializatio n process neglects
or does not s u ffic ie n tly specify the varying conditions
under which the lack of ro le playing capacity may develop.
Attention is not called to p o ten tial differences between
individuals with sim ilar role playing capacity. The c r i
ticism may be made more clea r by drawing a tte n tio n to the
close s im ila rity between the "role playing" hypothesis of
Mead and the " id e n tific a tio n " hypothesis of Freud. From
Mead’s view, at le a s t as i t is operationally expressed in
the Gough sc a le , criminal p e rso n a litie s vary on only one
dimension, i . e . , th e ir a b ility to play or "in tern a liz e"
social ro le s. In Freud’s view, criminal p e rso n a litie s
d iffe r in; the strength of id e n tific a tio n s , the capacity to
id en tify , the d irec tio n of id e n tific a tio n s , and the spe
c ific nature of id e n tific a tio n s . This conceptual frame
work appears to provide more p o ten tial lin e s of difference
between individuals who may be characterized by the same
degree of extern ally evaluated severity of criminal
behavior. Another way of p u ttin g i t is to say th a t, the
Gough scale and the Mead concept allows us to place in d i
viduals on a continuum of so c ia liz a tio n , but does not a l
low us to make p o te n tia lly useful subgroupings at p a r ti
cular points on the continuuip. That there may be s ig n if i
cant s tru c tu ra l personality differences underlying sim ilar
degrees of defect in the role-playing capacity and in the
severity-chronicity of p a rtic u la r criminal behavior pat
te rn s, is not taken into account in the stru ctu re of the
Gough scale and in in te rp re ta tio n of responses to i t . The
Freudian concept of id e n tific a tio n with parental figures
as the basis of adequately socialized p erson ality, would
appear to provide a p o te n tia lly b e tte r basis for d i f f e r
e n tia tio n among individuals characterized by severe c r i
minal behavior pattern s. Some support is given fo r th is
position in the fa c t th a t in v estig ato rs with psychoanaly-'
t i c o rie n ta tio n have tended to provide the most d etailed
and individualized d escriptions of criminal personali
t i e s to be found in the lite r a tu r e (Karpman, 1949).
Certain c r it e r ia which may serve as guide lin es in
the evaluation of diagnostic instruments emerge from the
above discussion. The point of view has been advanced
th a t, the b est diagnostic te s t s w ill have a lo g ical foun
dation in some more or less formal theory of behavior*
This re la tio n sh ip must be well enough defined through
34
e ffo rts at external v alid atio n to com el a high degree of
consensual agreement as to the connotative significance
of p a rtic u la r responses to the instrument.
Another c rite rio n involves the impression th a t
effectiv e diagnostic instruments must be constructed to
tap those re la tio n sh ip s which lo g ically unify wider ranges
of behavior. This assumes th a t there are such re la tio n
ships and th at there are at present applicable conceptual
formulations to d ictate the stru ctu re and content of the
instrum ent.
A th ird c rite rio n would be to demand th a t a t e s t
must provide the kind of information which leads toward?
individuation of the subject along as many dimensions as
are needed to comprehend the relevant behavior in the most
adequate fashion. The in te rp re tiv e dimensions of the t e s t
must provide fo r individuation in terms of q uality, kind,
and in te n sity rath er than in te n s ity alone.
The general c r ite r io n which seems im plicit in the
above is th a t the information the te s t provides should be
meaningful and expressible in terms which have good con
cep tu a l-th e o retical p roperties. The Rorschach Test is
popular because i t is such a rich source of unique, id io
syn cratic, personal m aterial. The trouble is th a t, the
m aterial the t e s t provides is so in d iv id u a lis tic and
variable th at i t defies attempts to organize i t concep
tu a lly and to subject i t to processes of qu an tificatio n
and external v alid atio n . In other words, i t is hard to
get good interpersonal agreement as to ju st exactly what
meaning to attach to Rorschach responses. An instrument
which combines some of the p o sitive, "deep-going" data
getting properties of the projective techniques with the
"objective" quantifiable properties of s t a t i s t i c a l l y b e t
te r , more validateable but content-wise and conceptually
more s t e r i l e instruments is called fo r.
CHAPTER I I I
RATIONALE A N D DESCRIPTION O F THE EXPERIMENTAL INSTRUM ENT
In th is section a diagnostic t e s t to be applied in
the d iffe re n tia l diagnosis of offenders against the law
w ill be described. In constructing the experimental in
strument an attempt has been made to sa tisfy the c r it e r ia
described above for adequacy in a diagnostic t e s t . The
reader will re c a ll th a t psychoanalytic theory has been
selected as a feasib le conceptual o rientation to the task
at hand.
The psychoanalyst (Glover, I960) is lik e ly to
object to an attempt to assess "superego1 1 dimensions of
behavior by tra d itio n a l psychological techniques involving
normative s t a t i s t i c a l methods. His objection w ill be th a t
the individual character is so unique and complex in i t s
structure as to render meaningless any normative compari
son between groups of individuals, the most typical ap
proach in psychological research. At best, the analyst
w ill say that only a r t i f i c i a l l y abstracted overt charac
t e r i s t i c s which have no ascertainable, meaningful re la tio n
to the "inner structure" of character, can be iso lated by
methods other than long-term study of the individual case
in the psychoanalytic s itu a tio n . His major objection
would be th at although partly accessible to awareness, the
-36-
37
superego’s e ffe c t upon personality is largely an uncon
scious one. This being the case, a psychological t e s t
calling for a conscious report from subjects cannot tap
superego dimensions because by d efin itio n they are inac
cessible except through slow analytic work in a one to one
analy st-patient relatio n sh ip .
The psychologist p e rsists in asking, ”Is there no
way in which to subject these complex inner processes to
some form of external, objective normative evaluations?”
The point of view taken fo r purposes of the present ex
periment is th a t psychoanalytic theory i t s e l f suggests
possible avenues towards th is goal. Although the superego
i s described as in the main an ’’unconscious s tru c tu re ” of
personality, i t may exert c e rtain effects upon personality
which could conceiveably be tapped by assaying certain
conscious a ttitu d e s and th e ir in te rre la tio n sh ip s . A b rie f
review of those personality changes, which psychoanalytic
theory hypothesizes where there has been a ’’normal” in te r
nalizatio n of the parental sanctions and prohibitions,
w ill point to those more conscious, accessible aspects of
personality which psychoanalytic theory suggests cannot
exist where there has not been adequate superego develop
ment.
With the successful establishment of a ’’normal”
and adaptive superego one can hypothesize th at;
1. Anxiety over the p o s sib ility of external punish
ment or loss of parental affection changes to g u ilt fe e l
ing; moreover, that independently s e l f - c r i t i c a l or
self-punishing a ttitu d e s become possible,
2. Self-esteem is regulated not so much by evi
dence of parental approval or disapproval but rath er by
how well one meets ’’inner” standards of behavior. Models
01 behavior and ideal future goals are developed with
psychological discomfort accompanying lack of accomplish
ment re la tiv e to these goals.
3. "Tender” feelings of reciprocal attachment,
mutuality and consideration replace the former sexual and
h o stile impulses directed toward the parents.
4. Benign relatio n sh ip s with healthy superego
development re su lt in a readiness to re la te to and iden-r,
t i f y with constructive figures through childhood, ado
lescence and la t e r adult l i f e . There is increased readi
ness and capacity fo r strong "constructive” interpersonal
attachments outside the family.
5. Where the superego development has been ade
quate, the in d iv id u al’s m oral-ethical values w ill have an
adaptive degree of congruence with the m oral-ethical
expectations of his primary c u ltu ra l group.
6 . There w ill be a ’’r e a l i s t i c , ” i . e . , correct
evaluation of the probable m oral-ethical reactions of the
39
environment to forbidden or a n tiso c ia l behavior.
7e Since the decisive Oedipal id e n tific a tio n s form
the basis for the in tern alize d m oral-ethical values there
should be a sig n ific a n t residual id e n tity or congruance
between the value system of the individual and the value
system which he perceives his parents as possessing.
&. Since the "normal" superego is a product of a
strong id e n tific a tio n with the parent of the same sex,
there should be a re la tiv e ly g reater degree of "id en tity "
with the m oral-ethical values of the same sexed parent.
9. There w ill be a sense of healthy independence
and r e a l i s t i c self-confidence in one’s a b ility to master
or cope with events.
Reasoning from the above adaptation of psycho
analytic hypotheses with regard to the consequences of
benign superego development c e rtain conclusions are pos
sib le with regard to the quality and character of a n ti
social p e rso n a litie s. The c h a ra c te ristic s to be outlined
are so frequently described in the l i t e r a t u r e of criminal
psychology th a t they have emerged as a stereotype of the
a n tiso c ia l personality. Our th e o re tic a l discussion leads
us to conclude th a t the person unfortunate enough to have
had malignant object re la tio n sh ip s and negative i d e n t i f i
cation with poorly socialized adults can be characterized
by the following personality q u a litie s or t r a i t s :
4.0
1. Likely to he re la tiv e ly less s e l f - c r i t i c a l ,
le ss aware of psychological discomfort stemming from
unachieved ideal goals.
2 . More rejectin g of mentors or " id e n tific a tio n
figures" as an influence upon his own behavior.
3. More re je c tin g of or in d iffe re n t to the so fte r,
"tender" feelings of mutual affe c tio n a l attachment', and
conversely less capable of deep personal re la tio n sh ip s.
4. E ither more aware of feelin g s of in f e r io r ity
based on a sense of in a b ility to cope adequately or more
lik e ly to hold overblown, fa ls e ly omnipotent concepts of
h is potency and capacity.
5. In co rrectly apt to judge the m oral-ethical
reactions of h is interpersonal environment.
6 . Apt to be deviant from the c u ltu ra l norm in
h is own m oral-ethical evaluation.
7. Apt to be "at odds" with the m oral-ethical
value system which he perceives his parents espousing.
8. Apt to show re la tiv e ly le ss congruence between
h is own m oral-ethical value's and the values of the same
sex parent.
9 . Less apt to experience feelin g s of "inner"
psychological discomfort re la te d to fa ilu re to liv e up
to some standard of required so cial behavior, i . e . , less
able to experience and accept g u ilt.
41
The w riter has used these p a rtic u la r hypotheses as
a guide in the construction of the t e s t instrument to be
described below. While the dimensions selected do not pro
vide a direct avenue to the idiosyncratic and m ultilevel
complexities involved in individual superego development
they may serve to delineate certain continua relevant to
superego aspects of personality. The hope is th at in d i
viduals may be ranged along these continua and compared
in a normative quantitative manner: and that through
analysis of these continua and th e ir in terre la tio n sh ip s
a more focused, and d iffe re n tia tin g , diagnostic method
may emerge.
The content and form of the experimental in s tru
ment is best described in terms of operational d e fin i
tions for the psychoanalytic concepts of id e n tific a tio n ,
object re la tio n s, g u ilt, and ego ideal.
Id e n tific a tio n
As we have seen, criminal behavior may be viewed as
the re s u lt of disturbances in the development of the
"superego." This structure develops normally through a
process in which the asocial child rids himself of fears
of parental desertion or physical r e ta lia tio n by taking
the prohibitions and sanctions of the parent as his own.
P ersonalities may be characterized in terms of the in d i
vidual v icissitu d es of th is id e n tific a tio n process, thus
42
a given personality may be said to have " id e n tifie d ” with
a cruel, impulsive parent, a p ro h ib itiv e, stringent parent,
a lax, unduly permissive parent, and so on. "Criminal
p e rso n a litie s" may be d iffe re n tia te d in terms of the nature
of t h e i r id e n tific a tio n s . A man can be said to have iden
t i f i e d with a parent holding a n tiso c ia l values, to have
never made meaningful id e n tific a tio n with other in d iv i
duals, to have made weak or conflicted id e n tific a tio n s ,
and so on. Maidevelopments of the id e n tific a tio n process
may be associated with varying degrees of disturbance in
s o c ia liz a tio n , each carrying d iffe re n t im plications in so
fa r as the sev erity and m o d ifiab ility of the re su lta n t
a n tis o c ia l behavior is concerned.
In th is study the id e n tific a tio n dimension w ill be
defined in terms of the extent of agreement between the
s u b je c t’s evaluation of a n tis o c ia l behavior and h is im
aginative estimate of what evaluation each of his parents
would apply to the same behavior. The degree of " id e n ti
fica tio n " with the parent is defined as the extent of
agreement between the s u b je c t’s punishing-sanctioning
reaction to a serie s of a n tiso c ia l acts and his subjective
estim ation of what the p a ren t’s evaluation would be. This
lo g ical equating of the psychoanalytic " id e n tific a tio n "
dimension with s e lf estim ates of sim ila rity with parents
finds precedence in the experimental l i te r a tu r e of recent
43
years. A study by Cava and Eaush (1952) was designed to
te s t the hypothesis th a t boys ’’conflicted in areas of per
sonality related to the id e n tific a tio n process” w ill ’’per
ceive themselves as less sim ilar to th e ir lik e-sex parent
than w ill those who do not show co n flict in th is a re a .”
They asked subjects to f i l l out the Strong Vocational
In te re s t Blank, f i r s t fo r themselves and then as they
thought th e ir fath er would f i l l out the blank. Differences
scores were calculated between ’’s e l f ” and ’’f a th e r .”
’’Distmrbances” in id e n tific a tio n were inferred from res
ponses to the "Blacky” personality t e s t . Cava and Raush
concluded th at disturbances of id e n tific a tio n were as
sociated with lack of perceived sim ila rity between the
s e lf and the fath er.
In another study Sopchak (1952), defining id e n tif i
cation as ’’the modeling of oneself in thought, feelin g, or
action a f te r another person” asked subjects to take the
Minnesota Multiphasic Personality Inventory f i r s t for them
selves and then as they f e lt mother and fath er would res
pond to the items. The id e n tific a tio n measure was the
number of items answered in the same way for s e lf and for
fath er and mother. In accordance with his hypotheses,
larg er discrepancies between s e lf scores and parental
scores were found to be associated with re la tiv e ly greater
personality disturbance, i . e . , poor id e n tific a tio n with
44
parents was correlated with psychopathology as measured
by the MMPI.
Jourard (1957) also equating "id en tifica tio n " with
perceived sim ila rity between se lf and parent asked sub
jects to rate themselves as more or less like th e ir
parents on 40 d iffe re n t personality t r a i t s . The degree
of perceived sim ila rity on these t r a i t s was taken as an
index of the degree of id e n tific a tio n with the p a rtic u la r
parent. Jourard found th at individuals displaying a high
degree of subjective self-p aren t s im ila rity showed s ig n i
fic a n tly more positive feelin g towards th e ir parents and
also tended to have a more positive s e lf evaluation.
In a v ariatio n of the id e n tific a tio n theme, Chang
and Block (I960) attempted to measure the degree of s i
m ilarity between the a ttrib u te s of the su b je c t’s ideal
s e l f , i . e . , as he would "like to be" and the a ttrib u te s
of his parents as he perceives them. In th is study, the
higher the degree of correspondence between one’s "ego
ideal" and one’s conception of one’s parents the greater
the degree of id e n tific a tio n with the parent. Their ex
perimental instrument consisted of a l i s t of 79 adjec
tiv e s which described his "ideal s e lf ." Following t h i s
the l i s t was checked for the fath er as the subject sees
him and then fo r the mother- as the subject sees her. Of
course, the measure of id e n tific a tio n is the number of
45
items which the "ideal self" shares with the fa th e r and
with the mother. In th is study high mother id e n tific a tio n
occurred with s ig n ific a n tly greater frequency in a group
of overt homosexuals giving support, say the authors, to
the Freudian derivation of homosexuality from the dynamics
of the id e n tific a tio n process.
Several stu d ies based on the id e n tific a tio n concept
have used actual ra th e r than perceived parent child t r a i t
or a t t r i b u t e . s im ila r itie s as a measure of id e n tific a tio n .
Payne and Mussen (1956) asked an experimental sample of
fa th e rs, mothers, and sons to f i l l out three scales of
the C alifornia Psychological Inventory. They were; the
tolerance sc a le , the m asculinity-fem ininity scale, and
the social p a rtic ip a tio n scale. The measure of i d e n t i f i
cation was the number of item s' answered the same way by
child and parent. The boys in addition took other te s t s
designed to tap the nature of th e ir re la tio n sh ip s with
parental figures and were rated themselves as to various
personality t r a i t s by th e ir homeroom teachers. The authors
concluded th at fa th e rs perceived as warm and rewarding by
th e ir sons tended to in sp ire stronger fa th e r id e n tif ic a
tio n s. The re s u lts , in general, tended to suggest th at
the boys with stronger fa th e r id e n tific a tio n s were more
masculine, perceived the family as "rewarding and warm,"
and were somewhat more "calm and friendly" than boys with
weaker fa th e r id e n tific a tio n s .
Analysis of the studies just reviewed as well as a
number of other studies focused on id e n tific a tio n pro
cesses shows th at experimenters have by and large taken
id e n tity or sim ilarity of in te re s ts , t r a i t s , and a ttitu d e s
between the subject and his parent or .his projected con
ception of his parent as an index of ’’id e n tific a tio n ” with
the parent. An extended review and critiq u e of the
various d efin ition s of the id e n tific a tio n concept employed
in the experimental lite r a tu r e w ill not be entered into
here. The reader is referred to the excellent reviews of
Stokes (1950) and Sanford (1955). I t is enough for the
present purposes to s ta te clearly the d e fin itio n of
id e n tific a tio n adhered to in th is study and to specify
i t s th e o re tic a l o rigin s.
The reader w ill re c a ll from the discussion of the
superego in an e a rlie r chapter that the resolution of the
Oedipus complex occurred as a re su lt of the c h ild ’s taking
his projected conception of his p arent’s m oral-ethical.
rewarding-punishing, or prohibition-sanction system as
his own. According to psychoanalytic theory what the
child ’’in te rn a liz e s ” prim arily at th is important and de
cisive stage, is his conception of how the parents feel
about and would respond to various forms of forbidden
behavior. In the present study where the focus of interest
is on those ’ ’id e n tific a tio n s ” which enter most importantly
47
into the determination of socially acceptable behavior we
w ill define "id en tifica tio n " as the extent to which the
s u b j e c ts prohibiting-sanctioning responses correspond to
his parent’s prohibiting-sanctioning responses to the same
stimulus s itu a tio n s . The f i r s t general section of the
experimental instrument is designed to reveal the degree
of th is correspondence with each parent.
The su b ject’s task is described as follows. He is
asked to read short descriptions of 20 d iffe re n t crimes.
A rath er wide range of criminal acts is represented, in
cluding murder, rape, swindling, th e ft, racketeering, etc.
The subject is instru cted to take the role of a judge in
a court of law. For each crime he is to assume th at a
d ifferen t 25 year old, white male stands before him con
victed of the crime. As the judge he is to assign the
sentence he would give for the crime i f he were judge.
He is instru cted to respond as he personally would ra th e r
than as he imagines a professional ju r is t would sentence.
A 6 category, forced choice scale of sentence a lte rn a
tiv es is provided for each item. The sentence choices
range from probation through cap ital punishment. In
addition to assigning the sentences he himself would give
i f he were judge, the subject is asked to indicate the
sentences he imagines his mother and father would give in
the role of judge. This section is scored by adding the
scale points separately for s e lf , fath er, and mother.
1 + &
The extent of parental id e n tific a tio n in various c rite rio n
groups may be inferred from the re la tiv e size of co rrela
tio n co e fficien ts between s e lf and fa th e r, and s e lf and
mother in the groups under consideration.
In the second section of the experimental in s tru
ment the subject is asked to read a serie s of descriptions
of delinquent a c ts . The acts described are assumed to re
present ty p ical a n tiso c ia l behavior as seen in the
"teenage” offender. Again the subject is asked to evalu
ate the behavior from a prohibiting-sanctioning o rie n ta
tio n . In t h i s instance, however, sentences are not as
signed, but the subject is asked to indicate how he would
evaluate a frien d who had performed the acts described.
A 5 point scale of increasing social ostracism i s provided
for the s u b je c t’s response. This is a "face v a lid ity ”
scale which depends upon language fo r the sense of increas
ing sev erity of a ttitu d e .' No exact psychological distance
between scale points is assumed. After giving his own
responses to the item, the subject is asked to respohd as
he imagines ”most others he knows w ell” would respond to
the same items. As i t was planned to apply the experi
mental instrument to adolescent boys in the present
study, i t was f e l t th a t th is task would provide some in
formation as to the su b je c t’s m oral-ethical reactions to
a n tiso c ia l behavior which he himself or his peers would
4 9
be more lik e ly to engage in. The "most others" dimension
was introduced to e l i c i t the s u b je c t’s subjective concep
tio n of figures currently sig n ifican t to him in a "super
ego” sense (aside from parents). I t was f e l t th a t th is
task would tend to tap relationships associated with
"id e n tific a tio n figures" who had become important subse
quent to the parental influence. A large discrepancy
between self score on th is section of the t e s t and the
"most other" score thus might represent a sig nificant lack
of id e n tific a tio n with the in d iv id u al’s own subjectively
organized circ le of peers or of sig n ifican t "others"
fig u res. A number of hypotheses can be generated to deal
with fluctuations in these two scores.
In the th ird section of the t e s t the subject is ■
asked to respond to 159 tru e -fa ls e items. Based on the
psychoanalytic hypotheses with regard to the consequences
of healthy superego development (outlined above), these
items are designed to assess the dimensions of g u ilt f e e l
ing, "object re la tio n s ," and the "ego-ideal." There are
in addition a group of items which are not based on the
superego concept. These are statements aimed at measuring
the s u b je c t’s psychological propensity toward involvement
in delinquent gang a c tiv ity . ' These 159 items make up 1 +
separate scales in the plan of the experimental instrument.
The tru e -fa ls e items lik e the descriptions of
criminal and delinquent acts in the other two sections of
the t e s t were devised by the w rite r. Items from other
t e s ts were not used because of the w riter*s desire to make
an individual attempt to tra n s la te psychoanalytic theory
into t e s t items and t e s t re la tio n sh ip s. Taking items
however discrim inating from other te s ts based on empirical
or d iffe re n t th e o re tic a l grounds would tend to confound
the sta te d goal of constructing a t e s t with a clea r lo g i
cal re la tio n sh ip to psychoanalytic theory. Descriptions
of the types of items found in each of the "sc ales” of the
tru e - fa ls e section of the experimental instrument follows
below.
The Guilt Scale
The items of th is scale are designed to t e s t the
extent to which the subject experiences g u ilt in re la tio n
to his own behavior and the manner in which the per
so n ality deals with g u ilt. The items are w ritten to
stim ulate "dynamisms" or defense mechanisms commonly
evoked to handle g u ilt feelin g s, projection, denial,
ra tio n a liz a tio n , confession, and reparation are defenses
against g u ilt which the items attempt to tap.
The "Ego-Ideal" Scale
These items are designed to determine the presence
or absence or the re la tiv e importance of a ttitu d e s re
late d to the achievement of id eal goals, the s a tis fa c tio n
of model personal requirements of performance, and the
desire to "live up" to the model of ideal behavior pro
vided by admired external fig u res. Concern over one’s
progress toward achieving of so c ia lly esteemed long term
goals and gaining recognition and acceptance from others
is implied by high scores on th is scale.
The Ob.ject Relations Scale
The items of th is scale are designed to e l i c i t
responses which indicate a h isto ry of and capacity for
personal rela tio n sh ip s of a close, interdependent charac
t e r . The e sse n tia l strength of the in d iv id u a l’s emotional
t i e s to others is in question. How much does the person
care about others? How close w ill he l e t others come to
him? W e have seen in the preceding discussion of psycho
analy tic theory, how early lack of and subsequent re je c
tio n of close dependent t i e s may be a fundamental facto r
in disturbances of superego development. The tru e -fa ls e
items of the scale f a l l into several categories, a l l pre
sumed to bear on the way in which "dependency needs" are
handled by the p ersonality. The items may be grouped
under the descriptive headings; 1. Feelings of exag
gerated se lf sufficiency, 2. Acceptance of dependent
sta tu s, 3. Feelings of tender or "soft" sentiment towards
important fig u res, Acceptance of "close" t i e s . Al
though one might surmise that responses to these items
52
might easily be "faked" in a so cially desirable direction
there is support in the lite r a tu r e for the contention th a t
the type of sentiment expressed by these items is p a r ti
cularly repugnant to the person whose personality has
taken an an tiso cial bent. Such sentiments are said to
present a special kind of defensive problem fo r the "a n ti
social character." Here the delinquent person’s need to
re je c t tender or dependent sentiments is seen as a facto r
which makes i t d if f ic u lt for him to "fake" these items in
a socially desirable d irec tio n .
The Gang Scale
The items of th is scale are designed to e l i c i t en
dorsement or lack of endorsement of a ttitu d e s from which a
strong need for membership in an a n tiso c ia lly oriented
"ingroup" may be in ferred . Although i t is possible to
derive a need for such statu s from psychoanalytic theory,
the conceptual "distance" is greater for these items than
for the items of the other scales. The items of the gang
scale are included because i t is recognized th a t group
influences may be of considerable importance as independent
determinants of an tiso cial behavior. Possible empirical
relationships with other scales of the te s t were foreseen
and tempted the inclusion of these items. The items may
be subsumed under several descriptive categories. They
are:
53
1. Need for a feeling of physical safety and
emotional support provided by membership in a p a rtic u la r
"ingroup."
2. Marked feeling of "ingroup" behavioral solid a
r ity or loyalty in action, i . e . , "a ll for one and one for
a l l . "
3. Need to define the ingroup by a ttitu d e s of
intolerance fo r outsiders.
4. Self-esteem based importantly on how the "self"
is appraised by the ingroup of which the subject considers
himself a member.
5. Strong feelings of physical th rea t or th re a t.to
self-esteem based on fa ilu re to behave according to
ingroup v alu es.
6. Dependence upon and attachment to external
■symbols of ingroup membership, i . e . , stereotypes of dress
and speech.
This completes the description of the experimental
instrument. The reader, at th is point, may ju s tif ia b ly
wonder why th is p a rtic u la r conglomeration of items and
te s t tasks? In b rie fe s t summarizing terms, the w riter
hypothesizes th at the subject responding to the experi
mental instrument provides a sample of the m oral-ethical,
prohibiting-sanctioning components of his personality and
places i t in a context which enables the experimenter to
discern the degree of subjective and objective corres
pondence between the personal punishing-sanctioning a t t i
tudes and the punishing-sanctioning systems of each parent,
peers, the subcultural group i f any, and the prevailing
stan d ard -settin g c u ltu ra l group. As the t e s t is arranged
the specific content of punishing-sanctioning a ttitu d e s
is not e lic ite d but ra th e r two other important aspects of
m oral-ethical evaluation, namely; the re la tiv e in te n s ity
of the values, i . e . , r e la tiv e punitiveness and the cor
respondence in terms of punitiveness or in te n sity between
the individual value system and th a t of other important
norm-setting agents, i . e . , parents, peers, subculture, and
la rg e r cu lture. Our th e o re tic a l framework sp ec ifie s th a t
adequate social behavior depends upon individual " id e n ti
fic a tio n " with or "in te rn a liz a tio n " of the prohibitions
and sanctions of parental figures who are themselves
adaptively id e n tifie d with the prohibiting-sanctioning
values of the p a rtic u la r culture in which the individuals
must function. I f we are able to gauge the extent to
which individuals have id e n tifie d with parents and how weH
th is id e n tific a tio n has been carried over into i d e n t i f i
cation with the values of the prevailing social milieu we
should have then some idea as to how much and what kind of
superego development has occurred, and how well th is a l l
corresponds to the standards of the society in which the
individual must function.
The tru e -fa ls e items may be somewhat redundant,
however they are seen to tap a ttitu d e s and systems re la te d
to optimal id e n tific a tio n processes. Interp reted together
with the f i r s t two sections of the t e s t , these may focus,
.amplify, and d iffe re n tia te impressions gained from the
f i r s t two sections of the t e s t .
CH APTER IV
EXPERIMENTAL HYPOTHESES A N D EXPERIMENTAL PLAN
In the following section a number of te sta b le hypo
theses deriving from the psychoanalytic ratio n ale of the
experimental instrument w ill be set fo rth . In th is pre
liminary attempt at v alid atio n , considerations of economy
r e s t r i c t the experimental application to a few selected
hypotheses which are fundamental to the logic of the
experimental instrument.
Experimental Hypotheses
I t is predicted th at in responding to the experi
mental instrum ents:
1. Well socialized individuals w ill tend to judge
criminal and delinquent behavior within a
normative range of punitiveness em pirically es
tab lish ed by t h e ir well socialized c u ltu ra l,
sex and age peers. Poorly socialized in divid
uals of sim ilar age, sex, and cu ltu ral back
ground w ill tend to make judgements which f a l l
out of th is normative range of punitiveness
towards delinquent and criminal behavior.
2. Well socialized individuals w ill manifest a
s ig n ific a n tly higher degree of congruence or
’’id e n tity ” between th e ir own judgements with
regard to a n tiso c ia l behavior and th e ir pro
jected conception of each paren t’s judgement of
a n tiso c ia l behavior; then w ill poorly socialized
individuals of the same age, sex, and cu ltu ral
background.
3. Well socialized adolescent boys w ill show a higher
degree of id e n tity (as defined operationally in
th is study) with t h e i r same sex parent, re la tiv e
-56-
57
to the opposite-sex parent. Poorly socialized
boys of sim ilar age, and cu ltu ral background
w ill display a re la tiv e ly less strong "Id en ti
fic a tio n ” with the same sex parent, i . e . , th at
is they w ill be less strongly id e n tifie d with
th e ir fathers than w ill well socialized boys.
1 +. Well socialized individuals w ill show a sig n i
fic a n tly higher degree of id e n tific a tio n with
the parent perceived as most fru s tra tin g , i . e . ,
most punitive. Poorly socialized individuals
are expected to show a sig n ific a n tly le sse r
degree of id e n tity with the parent perceived
as most punitive.
5. The well socialized in d iv id u al’s projected con
ception of his parent’s punitiveness towards
an tiso cial behavior w ill tend to f a l l within a
normative range of punitiveness em pirically es
tablished by his well socialized age, sex, and
cu ltu ral peers.
6. In responding to tru e -fa ls e items poorly
socialized individuals w ill tend to display to
a sig n ific a n tly greater degree than a compa
rable group of well socialized peers, the
following;
1. Absence of or re jectio n of feelings
of g u ilt and s e lf criticism .
2. Rejection of the " s o fte r” feelings of
tenderness, dependency, and mutual
affectio n al attachment.
3. Rejection of external models and ideal
goals. Lack of future oriented, socially
acceptable striv in g s.
4. Concepts of special personal power, good
fortune, u n re a listic omnipotence.
5. Greater dependency upon an ingroup,
oriented in a delinquent d irectio n .
7. The degree of disturbance in superego development
as determined by the experimental instrument w ill
be positively related to the severity and chroni-
city of the delinquent pattern as evaluated by
professional clin ic ia n s.
53
As a f i r s t step in the exposition of the experi
mental undertaking, operational definitions of the terms
of the experimental hypotheses are provided as follows:
Well Socialized Individual
In th is experiment "well socialized" is defined by
the absence of a formal delinquent history and "good"
social adjustment as judged by personnel counselors and
the dean of boys at the high school where the subject is
in attendance. In the la te r phase of the experiment the
additional c rite rio n of adequate academic achievement
re la tiv e to psychometrically determined levels of i n t e l
lectu al a b ility is included. Subjects who have had no
sig nificant contact with law enforcement agencies, are
well regarded by o ffic ia ls in th e ir high school milieu
and are achieving in accordance with th e ir a b ilitie s are
defined as "well socialized."
Poorly Socialized Individual
In th is experiment "poorly socialized" is defined
by a history of chronic antisocial behavior. Those indi
viduals defined as poorly socialized in the present
study have uniformly had numerous juvenile court ex
periences, have failed on probation, and been committed
to an agency regarded as the community resource for per
siste n t juvenile misbehavior. A commitment of the
59
California Youth Authority is thus taken as the operational
definition of "poorly socialized."
Age. Sex, and Ethnic Peers
Subjects selected for the present study are white,
protestant males between the ages of 16 and 1$.
Normative Range of Punitiveness
Punitiveness refers to the total number of scale
points, standing for sentence-years which are assigned to
a series of criminal acts by the subjects. The normative
range of punitiveness is defined by the mean scores and
the measures of variability taken from the distribution of
punitiveness scores from a sample of individuals defined
as "well socialized" by the criteria adopted in this
study.
Cultural Peer
Subjects in the various experimental groups are
defined as cultural peers on the basis of similar social
class membership as determined by the "father’s occupa
tion" index of Hollingshead and Redlich (1953),
More Punitive Parent
Defined as the parent to whom the subject a t t r i
butes the greatest number of sentence years when project
ing a conception of how the parent would behave in
60
assigning sentences to crim inal behavior.
Social Reaction to Delinquent Behavior
Defined by t o t a l scale p oints assigned to a s e rie s
of delinquent a c ts where the scale points rep resen t in
creasing degrees of so c ia l ostracism .
S everity of A ntisocial Behavior P attern s
The r e la tiv e s e v e r ity .o f a delinquent behavior p a t
te r n is defined by the ev alu atio n of p ro fe ssio n a l c l i n i
cians engaged in the study and d iagnosis of o ffen d ers.
Assignment of an in d iv id u al to a ,fbad prognosis” group by
th e c lin ic ia n s studying h is case i s taken as the opera
tio n a l d e f in itio n of "severe delinquent p a t t e r n .”
Before a d iscu ssio n of the experim ental procedures
proper, a general experimental plan w ill be o u tlin e d . The
study i s planned in two phases: In the p ilo t study a
re la tiv e ly , larg e sample of poorly s o c ia liz e d and a com
parable sample of w ell s o c ia liz e d in d iv id u a ls are admini
ste re d th e experim ental instru m ent. This phase of the
study has two purposes; to provide data fo r an item analy
s is of the tr u e - f a ls e sectio n s of the instrum ent and to
serve as a prelim inary t e s t of the experim ental hypotheses
p e rta in in g to the concept of id e n t i f i c a ti o n (the f i r s t two
sec tio n s of th e t e s t ) , and to provide p ra c tic a l information
with regard to the a d m in istratio n c h a r a c te r is tic s of the
instrument, e .g ., i t s level of in trin sic in terest to sub
jects, time of administration, clearness of instructions,
and format, unforeseen emotional reactions to the items
and tasks, p o ss ib ilitie s of response bias, d iffic u lty in
reading and comprehending items, and the lik e . This was
seen as a preparatory phase for the purpose of learning
things about the instrument which could only be grasped
by seeing the te s t administered in a "live" situation. I t
was hoped that th is preliminary study would enable the
experimenter to sharpen or refine the instrument before
applying i t to more carefully selected c rite rio n groups.
The second phase of the study involves the above
mentioned application of the instrument to more care
fully selected and matched c rite rio n groups. Three ex
perimental samples are treated: A sample of 50 "well
socialized" subjects matched for age, sex, ethnic back
ground, and socioeconomic status with two poorly so cial
ized groups, one a sample of individuals judged to have a
"poor prognosis" for future an tiso cial behavior and the
other judged to have a "good prognosis," i . e . , as more
likely to make a socially acceptable adjustment at some
future time. Judgements as to prognosis are made by
professional clinicians with experience in evaluating
antisocial behavior patterns.
S ta tis tic a l operations to define group differences
in reponse to the experimental instrument in the f i r s t
phase of the study involve c o rre la tio n a l techniques to
t e s t the re la tio n sh ip between "se lf" scores and both
parental scores. Group differences in punitiveness are
defined by te s ts of significance between group means.
Hypotheses of d iffe re n tia l group responses to the
tru e -fa ls e items are tested by the chi square technique
as applied to each item. E ssen tially these same s t a t i s
t i c a l te s t s are applied to t e s t differences between groups
in the second phase of the study.
C H A PT E R V
PILOT STU D Y
Selection of the "Well Socialized" Sample
As an i n i t i a l step contact was made with the
administrator of a large public high school d is t r i c t .
The purposes of the study were outlined and cooperation
requested. The writer'was referred successively to
principals, deans of boys, and student counselors in three
large high schools. In each school a request was made
by the experimenter for subjects meeting the following
c rite ria ; 1. white, 2. Protestant, 3. male, 4. between
16 and 1$ years of age, 5. having fathers of semiskilled
and skilled trades background, 6. lite r a te , 7. no formal
record of delinquency (in so far as th is was known by
school o ffic ia ls ), $. a satisfactory academic record,
9. no significant history of disciplinary problems in the
school setting, 10. good social relations with peers and
adults in the school situ atio n .
Selection of subjects was made by student counsel
ors with consultation and recommendation from homeroom
teachers. One hundred th irty subjects were referred by
counselors in the three high schools. Examination of
inforrration sheets fille d out by the subjects in th is
sample at the time of te s t administration resulted in a
-63-
yield of 107 subjects meeting the c r ite r ia for the sample.
Seven of these subjects were a r b itr a r ily dropped from the
study to make a sample of 100. Of these subjects several
proved to have omitted or l e f t incomplete one section or
another of the experimental instrument. Because of th is
N’s for the various subscores of the instrument vary from
97 to 100.
Administration of the Experimental Instrument (’’Well
Socialized’’ Group)
Subjects were met at th e ir high schools in audi-
torium-size testin g rooms. They had been told th a t th e ir
cooperation was being requested in the construction of an
experimental testin g instrument. I t s purpose or applica
tio n was not disclosed. Subjects were assured th at th e ir
personal id e n titie s were not relevant in the study and
that they were to become ’’s t a t i s t i c s ” in a study being
submitted as a requirement fo r an advanced degree. Good
cooperation was secured from v irtu a lly a ll subjects. The
w rite r’s estimate was th at rapport was adequate for the
purposes of the experiment. The w riter gave an overview
of the te s t instructions stressing th at subjects were to
respond as they personally would in judging the items not
as they imagined a professional j u r is t would.
Selection of the ’’Poorly Socialized” Sample
As an i n i t i a l step in the collection of th is sample
a l e t t e r was w ritten to the Director of the California
Youth Authority. The l e t t e r outlined the purposes of the
study and provided an outline of the experimental plan.
Cooperation of the agency was requested and granted. As
a re su lt of th is l e t t e r the cooperation of personnel at
the Southern Reception Center and Clinic of the California
Youth Authority in the selection of subjects and the
adm inistration of the experimental instrument was secured.
This in s titu tio n is a reception and diagnostic center re
ceiving a ll boys committed to the Youth Authority from the
"Southern Counties" of the s ta te . Youths sent to th is
center are invariably individuals who have a well estab
lished history of delinquent behavior. They are generally
boys who have been trie d several times on probation in
th e ir home counties and have continued to engage in the
more serious forms of delinquency. Delinquent h isto rie s
extending over several years and poor school adjustment
are c h a ra c te ristic .
At th is center the experimental instrument was made
a regular part of the program of group te stin g which the
Youth Authority employs to aid in the individual study of
each boy committed to the agency. The te s t was given over
a period of six months to a ll boys entering the agency who
were l i te r a t e and over the age of 16. Some 400 te s t
protocols were collected. As the protocols were received
the experimenter selected individuals meeting the sampling
c r it e r ia . I t was necessary to adm inister the instrument
to th is large mixed group because individuals meeting the
experimental c r it e r ia could not be singled out fo r special
processing in th is busy and tig h tly organized diagnostic
center. Data collection continued u n til 100 subjects meet
ing the experimental c r it e r i a and producing properly com
pleted protocols were obtained.
Administration of the Experimental Instrument (poorly
Socialized” Sample)
Following the i n i t i a l verbal overview of the t e s t
in stru c tio n s, subjects were in stru cted to begin. The te s t
adm inistrator then toured the room encouraging questions
and cla rify in g confusion about te s t in stru ctio n s and re s
ponding to spontaneous comments and queries in a nondi
rectiv e supportive fashion. The few subjects who could
not or would not follow te s t in stru c tio n s were allowed to
s i t the te s tin g session out working on other m aterial.
Following completion of the t e s t by a l l subjects in a
group an opportunity was provided to discuss reactions and
c la rify questions. This discussion was again conducted in
a nondirective supportive fashion, providing reassurance
but revealing l i t t l e concerning the purposes or logic of
the experimental instrument. This was to avoid discussion
of the t e s t with boys in the dormitories who had not yet
taken the t e s t . The impression of the psychologist admini
stering the te s t was that good rapport and cooperation was
secured in the great majority of subjects. A satisfac to ry
level of ”ego-involvement” in the experimental task was
obtained.
S ta tis tic a l Procedures
1. Item Analysis
For each item of the tru e -fa ls e section of the t e s t
(159 items) the to ta l number of individuals answering true
and the to ta l answering false in each experimental group
was found. Chi squares were then computed on each item to
determine the significance of difference between the pro
portion of each sample answering true and answering fa ls e .
Each item was then checked against the a p rio ri psycho
analytic hypothesis predicting the direction of response
for poorly socialized and for well socialized individuals.
2. Mean Differences
Mean and standard deviations for the d istrib u tio n s
of self-crim inal, mother, fath er, self-delinquent and
’’most others” scores were next computed for both experi
mental groups. A te s t of the significance of difference
between means for each score d istrib u tio n was calculated
(between experimental groups).
6 a
3. Relationship Between Test Scores and Intelligence
As a preliminary exploration of the effect of i n t e l
ligence upon te s t scores the Pearson Product-Moment cor
re la tio n a l procedure was applied re la tin g the su b je c t’s
scores on the California Test of Mental Maturity to r,s e l f n
scores, delinquent items. This section of the te s t was
chosen in the preliminary phase of the study because i t
was f e l t th at tendencies on the part of the subjects to
use higher in tellig en ce in an attempt to fake socially
acceptable responses would be most apparent on items more
nearly c h a ra c te ristic of th e ir own behavior.
True-False Items
In b rie fe s t terms the item analysis shows th a t the
great majority of the tru e -fa lse statements comprising
th is section of the experimental instrument, do not d is
criminate between the two experimental groups. Out of a
to ta l of 159 items only 22 of the chi squares calculated
reached the 0.05 level of significance. The poorly
socialized group tends to respond to these items in about
the same way as does the well socialized group. The
fa ilu re of the groups to respond to items in the predicted
manner is highlighted by the fa ct that of the 22 discrim i
nating items, only 11 were answered in the direction pre
dicted by the psychoanalytic hypotheses guiding construc
tio n of the items. Differences even when present are
69
re la tiv ely small with great overlap between groups. The
re su lts then can be described as essen tially contrary to
the experimental hypotheses relevant to th is section of
the experimental instrument.
Interp retation of these re su lts in an unequivocal
way presents many d if f ic u ltie s . I f the items are accepted
as adequate representation or expressions of the psycho
analytic dimensions which they attempt to tap, i t becomes
possible to say th at the experimental resu lts cast some
doubt upon the psychoanalytic conceptualization of a n ti
social behavior. Unfortunately there is no objective way
to determine the success achieved by the w riter, working
without external c r ite r ia , e .g ., psychoanalysts as judges
of the items, in interpreting psychoanalytic hypotheses
and tra n sla tin g them into te s t items. This lack of ex
ternal "construct validation" for the tru e -fa lse s ta te
ments leaves the in terp retatio n of experimental resu lts
open to the subjective impression of any observer. About
a ll that i t is possible to say with any degree of exacti
tude is th at the experimenter’s in terp retatio n of psycho
analytic authors and the tra n sla tio n of th is in te rp re ta
tion into te s t items was not successful in discriminating
between experimental groups which the theory suggested
should be d ifferen tiated .
7°
Analysis of Mean Differences
Table 1 presents comparative group means fo r " se lf,"
"mother," "father," "self-delinquent," and "others" scores.
Mean differences which prove to be significant at the
0.05 level or above are marked with asterisk s.
The table shows that in judging criminal and delin
quent behavior as they would themselves and as they imag
ine each parent and "most others" would; the poorly so
cialized group is consistently less punitive than are th e ir
well socialized peers. For 3 of the 5 scores the d if
ference between groups proves to be s t a t i s ti c a l ly s ig n ifi
cant. Though the differences are re la tiv e ly small and do
not permit prediction from the individual te s t score to
the c rite rio n group some support is given to the experi
mental hypothesis that "poorly socialized" individuals
w ill tend to f a ll out of the "normative range" in th e ir
judgement of an tiso cial behavior. In terms of the psycho
analytic hypothesis, however, one might expect larger d if
ferences between the groups,, i . e . , more evidence of a
fundamentally d ifferent way of evaluating an tiso cial
behavior.
The differences obtained do not seem to be of a
magnitude to support an inference of a basic perceptual
and cognitive "judging" difference between delinquent and
non-delinquent individuals. Far too many in the delinquent
Table 1
71
Mean Scores in Punitiveness Ascribed to Self, Mother
Father, and Most Others. Non-delinquent and
Delinquent Samples
Non
delinquent Delinquent
Non
delinquent Delinquent
N M N M SD SD
♦ Self 100 77.6S 100 73.06
11.07 10.3
Mother
99 77.79
9S 75.70
11.59 10.41
Father 100 SO. 40 9S 77.97 10.71 9 .SO
♦Self 97 79.11
100
66.13 11.S9 15.54
♦Others 97 SO.39
100
73.53 13.73 15.10
Self (Criminal items) t is equal t o . . . 3.14, P value
is less than .01
Self (Delinquent items) t is equal to ...4 .6 S , P value
is less than .001
Others (Delinquent Items) t is equal t o . . . 4.01, P value
is less than .001
72
group appear to react to the experimental stim uli in about
the same way as do the non-delinquents.
At th is point i t is not possible to say whether the
sig nificant differences obtained are related to the experi
mental hypothesis or whether such a factor as the differen
t i a l experience of incarceration was the important deter
minant. In the la te r phase of the study the r e lia b i l it y
and th eo retical significance of th is finding of more per
missiveness in the delinquent group w ill be explored fu r
ther by employing refinements of the instrument and of
the experimental design.
The p ilo t study data may be studied also for the
lig h t thrown on the hypothesis th at poorly socialized
individuals w ill display a greater lack of "id en tifica -
tion" with th eir projected conceptions of the parental
a ttitu d e s towards a n tiso cial behavior, then w ill the well
socialized group. I f th is were true one would expect the
differences between mean ’’s e lf," ’’mother,” and "fath er”
scores within the poorly socialized group to be s ig n if i
cantly larg er than in the well socialized group. Although
slig h tly greater differences do appear they do not reach
a satisfacto ry level of s t a t i s t i c a l significance. A
b e tte r te s t, of course, would be to calculate correlation
coefficients between these scores (within groups) and to
te s t the significance of difference between correlations
across experimental groups. This was not done fo r the f o l
lowing reason: Inspection of the t e s t protocols showed
th a t subjects marking for s e lf , fa th e r, and mother on the
same page of the answer sheet tended most often to mark
the same scale point for a l l three scores. This could be
in terp reted as indicating a high degree of ’’id e n tific a tio n ”
between s e lf and both parental scores in both groups, how
ever, i t seems more lik e ly th a t a ’’path of le a s t r e s i s t
ance” set was operating so th a t subjects did not bother to
discrim inate between themselves and th e ir parents. I t
seemed more in te re s tin g to ’’iron ou t” th is probable a r t i
fact in the format of the experimental instrument before
coming to any conclusions about the id e n tific a tio n hypo
th e s is , and so co rrelatio n c o e fficien ts th a t seemed a l
most ce rtain to be spuriously high in both groups, were not
calculated. I t was assumed also th a t hypothesized d if
fe re n tia l strength of ’’id e n tif ic a tio n ” with th is or th at
parent, one experimental group compared with the other
could not be adequately te ste d without f i r s t dealing with
the apparent flaw in the format of the answer s h e e t. I f
a f te r attempts to correct the possible error introduced
by having s e lf, fa th e r, and mother scales on the same
page, immediately adjacent to one another were made, and
the re su lts remained b a sica lly the same, a lack of d i f
ference in ’’id e n tific a tio n " p atterns between experimental
74
groups could more re lia b ly be inferred.
Relationships Between Intelligence and Test Responses
A Pearson Product-Method coefficient of co rrelatio n
was calculated between the self-delinquent score and in d i
vidual scores on the California Test of Mental Maturity in
the poorly socialized group. The coefficient obtained
(N of 7$ cases) was 0.02, suggesting a negligible r e la
tionship between judgements of an tiso c ia l behavior and
in telligence as measured by the C TM M .
Findings with Regard to Other Miscellaneous Test
C haracteristics
A very satisfac to ry level of in te re s t even some
absorption in the te s t was found in the poorly socialized
sample, a notoriously d if f ic u lt group to handle in a
group te s tin g situ a tio n . Overt resistance to the te s t was
most infrequent and i t did not seem to be p a rtic u la rly
threatening to individuals. A number of comments from
individuals reading below the sixth-grade level of com
petence indicates a need for a downward revision of the
vocabulary employed for items and te s t in stru ctio n s.
There was no p a rtic u la r d iffic u lty in getting subjects in
eith er group to understand what was required of them
aside from th is d iffic u lty in reading. The school psy
chologist who administered th is i n i t i a l form of the te s t
noted th at f a ir ly frequently subjects (in the poorly
75
socialized sample) who were unwilling to take other group
te s ts would ask to take the "judge t e s t . ” An in te re stin g
finding was th at individuals in the poorly socialized
sample tended to be stimulated into post t e s t group d is
cussions of th e ir parents and the parental a ttitu d e s .
C H A PTER VI
REVISION OF EXPERIMENTAL INSTRUM ENT
Results of the p ilo t study indicated the poor d is
criminatory power of the tru e -fa lse section of the t e s t .
These items proved to have l i t t l e i f any value in tapping
sig n ifican t areas of personality difference between well
socialized and poorly socialized groups. Only minimal and
probably unreliable h in ts as to the possible content and
form of new items could be gained from the few tru e -fa lse
statements which did achieve a degree of discrim ination
between the two groups. Because of fa ilu re in th is sec
tio n of the te s t to display any u t i l i t y of e ith e r empiri
cal or th e o re tic a l nature, i t was decided to delete the
tru e -fa lse items e n tire ly in the revised version of the
instrum ent.
The re su lts obtained with those sections of the
t e s t having to do with the id e n tific a tio n hypotheses were
somewhat more equivocal. I f the 159 items of the
tru e -fa ls e section of the t e s t are removed, 2 sections
remain. These are the criminal items, yielding scores
for s e lf (as judge) and mother and fath er (as judges);
and the delinquent items, yielding scores for the s e lf
and r,most o th e rs .” Comparative analysis of group means
-76-
on each of these scores showed significant mean d iffe r
ences for the self-crim inal, self-delinquent, and "most
others" scores, the poorly socialized groups showing a
trend towards less punitiveness in the self evaluation of
antisocial behavior and a tendency to attribute greater
punitiveness to "most others." I t also developed that,
contrary to the identification hypothesis, the mother and
father scores were not significantly more deviant from the
self score in the poorly socialized group. Further the
data did not bear out the hypothesis that the projected
parental "images" of delinquents would differ significantly
from the parental images of non-delinquents. In both
groups the mother and father scores tended to adhere very
closely to the self scores. In looking at individual pro
tocols it was noted that the majority of subjects tended
to mark the same scale point for self, father, and mother
on item afte r item. A strong possibility that the test
format, which required the subject to answer for self,
father, and mother in closely adjacent answer spaces, was
determining a "clumping" of responses became clear. A
path of least resistance may have been taken such that in
the majority of subjects no real attempt was made to dis
criminate between the self and parental figures.
I t was because of this possibility that i t was
deemed unprofitable or wasteful to test the various identi
fication hypotheses liste d on page 56. High and probably
73
not s ig n ific a n tly d iffe re n t co rrelatio n s would have been
obtained in both experimental samples. I t was decided
not to accept th is in d icatio n of "no d iffe re n c e ” in the
id e n tific a tio n p attern s of the two groups without f i r s t
te s tin g th e e ffe c t of a change in the t e s t format.
The major need was to re v ise the t e s t in such a way
as to reduce the p o s s ib ility of automatic "clumping” of
the s e lf , fa th e r, and mother responses of the subject and
to maximize the "p ro jectiv e" nature of the task so th a t a
le s s consciously determined re la tio n s h ip between s e lf and
parental scores would be e l i c i t e d . With these aims in
mind the following changes in the experimental instrument
were made:
As a f i r s t step 40 new crim inal items were devised,
20 mother items and 20 fa th e r item s. These items were
based on the o rig in a l items and a close correspondence was
maintained in terms of content. S u p erficial changes in
working, lo c a le , and in other irre le v a n t s itu a tio n a l fa c
to rs were made, while close equivalence with the o rig in a l
items was maintained, in terms of a n tis o c ia l sig n ifican ce
and psychological impact, e.g., where the old item was
a ssau lt and b a tte ry with a rock in an a lley the new item
might be; a ssa u lt and b a tte ry with a club in a vacant l o t .
The 40 new items made i t possible to provide separate
sections of both the t e s t booklet and the answer sheet fo r
79
fo r s e lf, fath er, and mother responses of the subject.
Secondly, the order of items corresponding to the
original items was changed. For the mother items the order
re la tiv e to the self-item s was simply reversed so th at
item 1 of the self-item s corresponded to the la s t item in
the section fo r the mother. For the fath er section of the
t e s t the items for the s e lf section were reversed both ways
from the middle so th a t the most middle item becomes number
1 and the la s t item becomes the middle item. This was
simply to avoid helping the s u b je c t’s memory by presenting
him with equivalent items always presented in the same
orderc Control over possible s e ria l e ffe c ts was lo s t in
th is way, however at th is phase in the development of the
t e s t i t was thought of making a more projective instrument
out of the t e s t . Possible s e r ia l effects could be tested
at another time with a b e tte r design. The id e n tific a tio n
hypothesis would not be confounded because of th is change
since both experimental groups would be subjected to the
same possible s e ria l e ffe c ts .
Thirdly, the items of the delinquent sections of
the te s t were not changed but were simply duplicated so
th a t separate sections of the answer sheet and the t e s t
booklet may be u tiliz e d for the self-delinquent and ’’most
others" responses making i t possible to alte rn a te these
t e s t tasks with the self-crim inal and the fath er and mother
do
sections of the te s t. This tends to make re c a ll from self
items to mother and fath er items more d iffic u lt for the
su bject.
The over-all change in the new form of the in stru
ment may be illu s tra te d by th is schematic representation
of the order of appearance for te s t tasks in the t e s t book
l e t and the answer booklet.
Old Form
1. Criminal item s. . . s e l f , father, mother
2. Delinquent item s. . . s e l f , others
New Form
1. Criminal ite m s ...s e lf
2. Delinquent ite m s ...s e lf
3. Criminal item s. . .mother
4. Delinquent item s. . ."most others"
5. Criminal items. . .father
Selection of the Poorly Socialized Sample
Subjects making up the 2 poorly socialized samples
in the second phase of the experiment were selected in the
following manner: 4 sta ff social workers (M.S.W.), pro
cessing the cases of incoming boys at the Southern Recep-
*
tion Center and Clinic of the California Youth Authority,
were asked to refer to the experimenter the names of a ll
boys who met the following c r ite r ia ; 1. White, 2. Protest
ant, 3. 16 to ld l/2 years of age, 4. Having in the workers
professional opinion eith er a very poor or very good "prog
nosis." Prognosis here was defined s t r i c t l y as "a p articu
la r boy’s future chances of engaging in an tiso cial be
havior, serious enough to bring him to the atten tio n of
au th o rities and to cause his incarceration." The social
workers were asked to take the prior record, th e ir subjec
tiv e evaluation of the severity of the offenses and th e ir
professional evaluation of the boy’s personality status
into account in assigning subjects to e ith e r the "good
prognosis" or the "bad prognosis" group. For every subject
selected in e ith e r group, the social worker’s evaluation was
based on the following case m aterial; 1. A formal social
case history developed at the c lin ic , 2. Individual in te r
view with the subject, 3. Formal police records, 4 .Psycho
logical and/or psychiatric report.
As subjects were referred the experimenter engaged
in further screening, ruling out cases of il li t e r a c y or
very weak reading s k i l l, boys whose fathers were of pro
fessional, managerial, or "white collar" occupational
s ta tu s, suspected cases of brain damage, and boys judged
to be psychotic by psychiatric examination. All boys re
ferred who met the above c r it e r ia were administered the
experimental instrument. Every protocol which showed com
pletion of a l l te s t items was included in the experimental
sample. Collection of data was continued in th is way u n til
62
50 protocols in the ’’poor prognosis” group and 25 protocols
in the ’’good prognosis” group were available.
C haracteristics of the Poorly Socialized Sample
Data gathering was organized in such a way as to
provide a poorly socialized sample within which a severely
delinquent group and a re la tiv e ly less delinquent group
could be delineated. The following c h a ra c te ristic s of the
’’good prognosis” group and the ’’poor prognosis” group are
described to give more face v a lid ity to the experimenter’s
assumption th a t the two groups represent ’ ’d iffe re n t” seg
ments on a continuum ranging from less severe to more
severe. The f i r s t c rite rio n established is th a t qualified
social workers experienced in evaluating an tiso cial be
havior have selected the poorly socialized subjects for
one experimental group or the other. In addition to th is ,
certain objective c r ite r ia are available which may serve
to provide evidence for the v a lid ity of the selections
made by the judges. Four more or less formal, objective
c r ite r ia upon which to compare the ’’good prognosis” and
the ’’poor prognosis” groups are presented. These are:
Total number of offenses comprising the o f f ic ia l record of
the delinquent career, The duration of the formal d elin
quent h isto ry in terms of months, The percentage of in d i
viduals in each experimental group assigned formal psy
ch iatric diagnoses, from which a severe personality
disturbance and a high probability of future d iffic u lty
with the law may be inferred, and, the percentage of in d i
viduals in each group referred to special treatment pro
grams maintained for p a rtic u la rly d if f ic u lt an tiso cial
p erso n a lities, following th e ir six-week diagnostic study
at the Youth Authority Reception Center and C linic.
1. The mean number of known and formally recorded '
offenses in the "good prognosis” group is 3.62. In the
”poor prognosis” group the mean number of offenses is 7.23.
Thus the ”poor prognosis” on the average has almost twice
as many offenses as does the "good prognosis” group,
2. With regard to the length or duration of the
formal delinquent history in terms of months, the "good
prognosis group has an average duration of 19.95 months.
The mean number of months of delinquent h isto ry in the
"poor prognosis” group is 50.31.
3. The percentage of individuals in the "good
prognosis” assigned a "severe” diagnosis indicative of a
serious personality problem and a high probability of
continued d iffic u lty with the law is 0.041. This re
presents one subject out of a group of 25. In the "poor
prognosis” group the percentage is 0.636. Thus almost
2/3 of th is group are seen as seriously disturbed as
compared with less than one per cent in the "good progno
sis" group.
#4
4. The percentage of individuals in the "good prog
nosis" gFoup assigned to "special treatment" programs, for
problem cases, is 0.00. The percentage assigned to
"privilege" programs and to d irec t parole from the recep
tio n center is 0.66 per cent. The trend is heavily in the
opposite d irectio n in the "poor prognosis" group. Here
the percentage of individuals assigned to "special t r e a t
ment" programs is 0.56$ per cent. The percentage assigned
to placements involving special re sp o n sib ility and posi
tiv e capacity is 0.00 per cent.
The above indices demonstrate th at the poor prog
nosis group d iffe rs from the good prognosis group in th at
they have (according to the only re lia b le source a v a il
able) committed a greater number of offenses, have been
delinquent fo r a longer time, are seen as more seriously
disturbed by q ualified professionals, are fa r more often
assigned to treatment programs designed for the severest
cases tre a te d by the Youth Authority and fa r less often
assigned to placements requiring a b e tte r than average
degree of capacity for work and so cial in te ra c tio n .
General C haracteristics of the 3 Experimental Groups
One hundred twenty-five subjects were studied in
th is phase of the study. All subjects were white, male,
Protestant adolescents. Mean ages and standard deviations
for the 3 experimental groups are presented in Table 2.
$5
Table 2
Average Age in Months,
Experimental Groups
Average Age
in-Months
Standard
Deviation
Well Socialized 2 04 7.12
Poor Prognosis 207 4.06
Good Prognosis 207.5 9.65
t W.S. vs P.P. - .566, ,fnot s ig n ific a n t"
t P.P. vs G.P. - .047, "not sig n ifican t"
t W.S. vs G.P. - .291, "not sig n ifican t
86
The t - t e s t indicates th a t the experimental groups do not
d iffe r sig n ific a n tly in terms of mean age.
No effo rt was made to match experimental groups in
terms of in tellig e n ce. To t e s t possible in teractio n
effects between te s t scores and in telligen ce te s t scores,
Pearson Product-Moment correlations were calculated, t e s t
ing the strength of re latio n sh ip between IQ scores on the
C alifornia Test of Mental Maturity and to ta l self-crim inal
item scores and self-delinquent item scores in the poor
prognosis and good prognosis samples. The correlation
co efficien ts are re lia b le and indicate a negligible re
latio n sh ip between IQ scores on the C alifornia Test of
Mental Maturity and scores for the self-crim in al and
self-delinquent sections of the experimental instrument.
Table 3 presents the co rrelatio n co efficien ts for the
relationships tested .
An attempt was made to control the facto r of socio
economic status in the 3 experimental groups by r e s t r i c t
ing the selection of subjects to individuals whose
f a th e r ’s occupation f e l l in the semiskilled trad es. This
attempt was successful in a ll but a few cases which f a l l
into unskilled and ’’white c o lla r” occupations. These
cases represent a very small proportion of the to ta l
sample and are d istrib u ted proportionately in the 3
experimental groups.
87
. Table 3
Pearson Product-Moment Correlations Between C T M M
IQ Scores and Self Scores, Criminal and
Delinquent Items
N Self-Criminal Items Self-Delinquent
Items
Good Prognosis 2$
Poor Prognosis 50
.00
-.07
.00
.16
88
In general i t can be said th at the 3 experimental
groups do not d iffe r sig n ific a n tly in terms of ethnic
grouping, sex, age, socioeconomic sta tu s, or re lig io n .
There is good ' evidence th a t in tellig en ce is not a fa cto r
in the determination of scores on the experimental
instrum ent.
Test Administration. Poorly Socialized Sample
Subjects in the poorly socialized sample were tested
by the experimenter in groups of 6. Subjects were grouped
around a table in a room of medium size. C igarettes were
provided in su ffic ie n t quantity so th at subjects could
smoke at w ill during the te s tin g .
The subjects were referred to the t e s t booklet,
asked to read the in stru c tio n s for the f i r s t section of
the t e s t and to ask any questions they might have about
them. The experimenter answered whatever questions arose,
usually they were few and the group was encouraged to go
ahead with the t e s t . The experimenter monitored a l l t e s t
ing sessions from beginning to end, permitting small ta lk
between subjects but no discussion of the t e s t i t s e l f .
Rapport and subject involvement in the experimental
task seemed optimal in a l l groups te ste d . Not a single
subject refused to take the t e s t . Occassionally i t was
necessary to answer general questions about the purpose of
the t e s t . The subjects were to ld th at the experimenter
was working for an advanced degree and th is was part of
the work. Questions as to what the te s t was getting at
were answered by saying that the examiner was trying out
a way to select boys fo r d iffe re n t kinds of tra in in g pro
grams. When subjects asked why they as individuals had
been selected for the experiment, they were to ld th a t i t
was necessary in research for subjects to be of the same
age, ethnic, and socioeconomic background. They had been
selected because they had these things in common. These
kinds of questions were handled at the end of the te s tin g
session in an informal group discussion, aimed at reassu r
ing the subjects without revealing much about the logic of
the t e s t and experiment.
Selection of the Well Socialized Sample
The well socialized sample was obtained in the
second phase of the experiment through the cooperation of
a high school in the community adjacent to the Youth
Authority C linic. This is an ”upper-lowerM to "lower-mid
d le” class community quite sim ilar in many respects to the
home environment of the biggest segment of the Youth
Authority population. The dean of students at th is high
school was asked to selec t for the experimenter 60 to- 70
boys meeting the sampling c r it e r ia on page 63 with the
additional requirement that only boys having academic
achievement in ’ good proportion to in te lle c tu a l level be
included in the sample. A very careful attempt was made by
the dean and s ta f f of 3 counselors to select boys meeting
these c r it e r ia . Considerable time and e ffo rt were taken
by the s ta f f in th is screening process and i t was the
examiner’s impression th at the s ta ff was in good contact
with students selected, in most cases having rath er well
d iffe re n tia te d impression of the personalities and social
c h a ra c te ristic s of the boys selected. An informal c r i
te rio n expressed by the experimenter was that the s ta ff
did not choose th e ir model students but rath er ’’average
youngsters who seemed to be doing well and getting along
with people." Through these procedures a to ta l of 67 boys
meeting the selectio n c r ite r ia were referred to the experi
menter for te s tin g . Of th is group boys proved to have
completed the t e s t and f i l l e d out the information sheet
co rrectly . Eight of these were dropped from the sample
before scoring to make the fin a l sample of 50 subjects.
Test Administration, Well Socialized Sample
Subjects in the well socialized sample were met in
a large amphitheater type classroom during the second and
third- periods of th e ir normal school day. All subjects
had received a mimeographed notice from the dean of
students, requesting them to report for "voluntary p a r ti
cipation" in a research project.
When a l l the students had completed the te s tin g
proper, a discussion session was held in which a number of
general questions with regard to the meaning of the t e s t ,
the question of why ju st th is p a rtic u la r group had been
selected, and the lik e were handled in a reassuring
fashion. The group as a whole worked quite d ilig e n tly and
with good concentration throughout the te stin g period.
S ta tis tic a l Procedures
Scoring
The individual protocols were scored simply by add
ing the scale points checked by the subject on each item.
For each subject a t o t a l score was computed for the s e lf
as judge, the mother as judge, and the fath er as judge.
In addition, scores were obtained for the "self" evaluat
ing delinquent items and for the "most others" items.
Higher scores indicate a greater number of sentence years
assigned to the items. Scores may be described roughly in
terms of greater or le s se r punitiveness. The delinquent
items are scored in the same way. Scale points are added
to obtain a "self-delinquent" score and a "most others"
score. The scoring system is based on a nominal scale
from which only estimates of "more than" or "less than"
can be made. No assumption of equal distance or value
between scale points is assumed.
92
Mean Differences
Means were calculated for self-crim inal, mother,
fath er, self-delinquent, and most others scores in a ll
three experimental groups. The significance of mean d if
ferences between the various scores was then calculated
both within and across experimental groups. The s ig n if i
cance of mean differences within experimental groups was
calculated by the method quoted in Guilford (1950) for
estim ating the significance of differences between means
in correlated data, since positive correlations between
s e lf scores and other scores were anticipated by the
experimental hypotheses.
Correlation Procedures and Tests of the Significance
of Difference Between Correlations
The Pearson Product-Moment coefficient of co rrela
tio n was employed to explore the strength of relationships
between te s t scores in the 3 experimental groups.
A number of additional correlations were calculated
to explore the "id e n tific a tio n with the aggressor" hypo
th e s is . For th is hypothesis the parent score in the well
socialized sample and in the poor prognosis sample were
separated into " s tr ic t" and "permissive" samples, place
ment in the p a rtic u la r group determined by which parent
was perceived as more punitive by the subject. Self
scores were then correlated with scores for s t r i c t parent
and permissive parent. As a fu rth er step in te s tin g th is
hypothesis the 20 most s t r i c t fa th e rs, 20 most s t r i c t
mothers, and 20 most permissive fa th e rs, 20 most permissive
mothers were culled from the well socialized and the poor
prognosis samples. The se lf score was then correlated
with the parental scores in each of these categories.
To calculate the significance of differences between
correlations within and between experimental groups i t was
necessary to convert Product-Moment correlation c o e ffi
cients to F ish e r’s Z -C oefficients. This was done since
the standard e rro r of a Pearson Product-Moment correlation
co efficient is unreliable where co efficien ts are of con
siderable magnitude (McNemar, 1955). The t-T est was then
employed to calculate the significance of differences be
tween Z -C oefficients.
C H A PTER VII
EXPERIMENTAL RESULTS
Experimental findings may be ordered by presenting
s t a t i s t i c s as they re la te to the experimental hypotheses
presented on page 56. The experimental hypothesis w ill be
presented together with its related s t a t i s t i c a l hypothesis
and the experimental outcome.
Hypothesis I
W e hypothesized that i f poorly socialized individ
uals d iffe r from well socialized individuals in terms of
th e ir prohibiting-sanctioning a ttitu d e s , i . e . , in terms of
the structure and content of th e ir ,fsuperegos,T t one should
be able to show th at the two groups d iffe r in th e ir
punishing-sanctioning behavior.
The s t a t i s t i c a l hypothesis sta te s th a t the mean of
sentence years assigned to a series of criminal acts by a
group of poorly socialized individuals taking the role of
"judge” w ill d iffe r sig n ifican tly from the mean of sentence
years assigned to the same criminal acts by a group of well
socialized individuals taking the role of "judge." A mean
difference reaching the .0.05 level of confidence w ill be
accepted as evidence supporting the hypothesis. Thble 4
presents mean scores and standard deviations for the 2
-9 4 -
95
Table 4
Mean Scores in Degree of Punitiveness Ascribed
to the Self, in Well Socialized, Good Pro
gnosis, and Poor Prognosis Groups
N
Self
(Criminal
Items) SD
Self
(Delinquent
Item s) SD
W.S. 50 74.96 S.21 73.30 10.64
G.P.
25 74 *44
60 75.03 16.56
P.P. 50 71.63
S.79 75.64 15.14
’’self" scores, the se lf acting as judge in a court of law
and the s e lf evaluating delinquent behavior on a scale of
increasing social ostracism of the offender. The 3 experi
mental groups are designated in the table as the well
socialized group (W.S.), the good prognosis group (G.P.),
and the poor prognosis group (P .P .).
None of the mean differences between the s e lf scores
in the 3 experimental groups reaches the 0.05 level of
s t a t i s t i c a l significance. Although the poorly socialized
means are in a l l cases below the well socialized mean, the
differences are slig h t and in th is sample do not support a
hypothesis of a fundamental difference in punishing-sanc-
tioning a ttitu d e s among the 3 experimental groups. These
closely sim ilar means and standard deviations indicate
th at the groups are more alike than d iffe ren t in th e ir
expressed punitiveness towards an tiso cial behavior. The
evidence does not support the hypothesis that poorly
socialized individuals, in gen eral," a re unable to evaluate
an tiso cial behavior in terms which are congruent with the
values of the more law-abiding segments of the population.
On the most conscious ra tio n al level at le a s t, a great
many poorly socialized individuals are well able to judge
unacceptable behavior in the same terms as do th e ir well
socialized peers.
97
H yp oth esis II
As a second hypothesis, we predicted th a t poorly
socialized individuals who are presumed to have a ’’weaker,”
less integrated superego than well socialized individuals
w ill show a s ig n ific a n tly weaker ’’id e n tific a tio n ” with the
perceived punishing-sanctioning a ttitu d e s of the parental
figures than w ill th e ir well socialized peers.
The s t a t i s t i c a l hypothesis states that the correla
tio n co efficien ts between the self-crim inal items score
and the two parental scores w ill be sig n ific a n tly larger
in the well socialized group than in the two poorly
socialized groups. The greatest differences in the
strength of the correlations should be between the well
socialized sample and the poor prognosis group. The data
w ill be accepted as supportive of the hypothesis i f the
differences between correlations reach the 0.05 level of
probability, and the difference is in the predicted direc
tio n . Table 5 presents the correlations obtained.
The experimental' findings are in accord with the
hypothesis. For both the self-mother and the se lf-fa th e r
relatio n sh ip the highest correlations are found in the
well socialized group. For the self-mother relatio n sh ip
(W.S. vs HP.) the probability th a t the difference between
correlations obtained could have occurred by chance is
less than 0.01. The self-mother relationship also appears
98
Table 5
Pearson Product-Moment r !s, Showing C orrelation
Between the Degree of Punitiveness Ascribed
to the Self and the Degree of Punitive
ness Ascribed to Each Parent in Well
Socialized, Good Prognosis, and Poor
Prognosis Groups
N Self-Mother Self-Father
W.S. 50
.83 .77
G.P.
25
.80 .60
P.P. 50
to
-4 -
«
.54
99
to d iffe re n tia te the good prognosis (delinquent) group
from the poor prognosis (delinquent) group. Here the ob
tained difference is sig n ifican t at 0.05 level of confi
dence. For the s e lf-fa th e r re la tio n sh ip (W.S. vs P.P.) t
the difference between co rrelatio n co efficien ts is s ig n i
fic a n t at the 0.05 level of confidence. Correlations ob
tained for the s e lf- fa th e r re la tio n sh ip do not serve to
distinguish the good prognosis sample from e ith e r of the
other two samples, although the co efficien t obtained does
f a l l into log ical place between the well socialized sample
and the poor prognosis sample. These findings give de
f i n i t e experimental support to the hypothesis th a t well
socialized individuals are more closely id e n tifie d with
the punishing-sanctioning systems or a ttitu d e s which they
perceive in the parental figures, than are th e ir poorly
socialized peers.
Table 6 presents the t values obtained and corres
ponding probability lev els for each re la tio n sh ip te ste d
for the experimental hypothesis.
Hypothesis I I I
Another prediction was th a t since the most optimal
■superego development presumably re su lts from id e n tific a tio n
with the perceived punishing-sanctioning a ttitu d e s or sys
tems of the same-sex parent, then well socialized individ
uals with re la tiv e ly more complete and b e tte r integrated
100
Table 6
The Significance of Differences in Correlation Between
Punitiveness Ascribed to Self and Punitiveness
Ascribed to Each Parent, Comparing Well
Socialized, Good Prognosis, and
Poor Prognosis Samples
Self--Mother t P Self-Father t P
W.S. vs P.P. 3.21 <.01 W.S. vs P.P. 2.02 <.05
P.P. vs G.P. 2.21 <.05 P.P. vs W.S. NS
W.S. vs G.P. *NS G.P. vs W.S. NS
* NS = "Not Significant"
101
"superego" w ill show a sig n ific a n tly greater id e n tific a tio n
with the father than w ill poorly socialized individuals
and w ill show a re la tiv e ly stronger id e n tific a tio n with
the father than with the mother. The stronger fath er
id e n tific a tio n w ill not be expected to appear among poorly
socialized individuals.
S ta tis tic a lly , we should expect th at in the well
socialized group the correlation of the s e lf score with
the fath er score w ill be sig n ific a n tly larg e r than the cor
re la tio n of the se lf score with the mother score, and th at
the s e lf-fa th e r relatio n sh ip w ill be s ig n ific a n tly stronger
in the well socialized group than in the poorly socialized
groups. A sig n ific a n tly larg er s e lf-fa th e r correlation
re la tiv e to the self-mother correlation is not expected in
the poorly socialized group. The reader may re fe r to
Table 6 for the correlations relevant to th is hypothesis.
I t is readily apparent that there is no evidence of
a stronger same-sex id e n tific a tio n in the well socialized
sample. The strength of the self-mother relatio n sh ip even
appears s lig h tly , though not sig n ifican tly stronger than
the s e lf-fa th e r relatio n sh ip in the well socialized group.
In none of the groups can i t be said th a t there is ev i
dence of a stronger id e n tific a tio n with one parent than
the other although a trend towards greater "mother id e n ti
fica tio n " is seen in the good prognosis group. The only
102
statement which the data support is th at id e n tific a tio n
with both parental figures is stronger in the well so c ia l
ized group than in the severely delinquent group. The
hypothesis of stronger same sex id e n tific a tio n in the well
socialized sample is therefore rejected .
Hypothesis IV
A fourth hypothesis bearing on psychoanalytic
theory assumes that the most ’’decisive” id e n tific a tio n in
the development of the superego is with the parent per
ceived as most powerful and threatening in the family
con stellation, i . e . , with the parent who is perceived as
’’the source of the decisive f r u s tr a tio n s .” The hypothesis
is tested in th is study by making an equation between
”most decisive f r u s tr a to r ” and ’’parent perceived as most
p u n itiv e .” For each subject the parent assigned the
greatest number of sentence years w ill be seen as the
parent perceived by the subject as responsible for the
most decisive fru stra tio n s in his experience.
In s t a t i s t i c a l terms, we should expect th a t correla
tions of the se lf score with the score of the parent as
signed the greater number of sentence years w ill be sig n i
fic a n tly larger for a l l experimental groups than co rrela
tion s of the s e lf score with the scores of parents assigned
the le sse r number of sentence years. Table 7 presents cor
re la tio n s of the se lf score with. 1. ’’more s t r i c t ” parent
103
Table 7
The Degree of Correlation Between Punitiveness
Ascribed to the Self and Punitiveness
Ascribed to the More tfStrict" Parent
and to the More "Permissive" Parent,
in W e.11 Socialized, Good Prognosis,
and Poor Prognosis Samples
N S e lf-S tric t Self-Permissive
W.S. 50 .S3 .SO
G.P. 25 .7S .69
P.P. 50 .59 .49
1 0 4
and 2. "more permissive" parent.
In no case do the correlations between self and
" s tr ic t" parent and self and "permissive" parent prove to
be sig n ifican tly d ifferen t, one from each other. The
relationships between self and " s tr ic t" parent a ll tend to
be somewhat larger but not sig n ifican tly so. The hypo -
thesis that greater id e n tific a tio n w ill occur between self
and the parent perceived as most s t r i c t or punitive is thus
not supported by the experimental findings. The only valid
statement possible with regard to the data is again that
well socialized individuals are characterized by a stronger
id en tificatio n with the perceived punishment-sanction a t t i
tudes of the parental figures than are poorly socialized
individuals. With the experimental te s t described above
the factors of sex of parent and re la tiv e ly perceived
punitiveness of the parent do not appear to have any d if
fe re n tia l effect on the strength of th is basic re la tio n
ship.
As a further te s t of the hypothesized relationship
i t was decided to obtain correlations in the two groups most
extreme on a continuum of so cializatio n , i . e . , the well
socialized sample and the poor prognosis sample, using sub
jects who perceived th e ir parents as eith er unusually puni
tive or unusually permissive towards antiso cial behavior.
There was the p o ssib ility that the hypothesized re la tio n
ship might hold for the extremes of the distrib u tio n s but
105
not for the center of the d istrib u tio n where differences
in perceived parental punitiveness were re la tiv e ly s lig h t.
Accordingly, the 20 most punitive fath ers and mothers and
the 20 most permissive fa th e rs and mothers were selected
and new co rrelatio n s were obtained with the respective s e lf
scores. Table 8 presents the co e fficien ts obtained.
With smaller N’s of 20 only 3 of the co rrelatio ns
in the tab le reach a sa tis fa c to ry level of r e l i a b i l i t y .
They are a l l in the well socialized sample: s e lf with
strict-m other; with permissive-mother; and with permis-
siv e -fa th e r.
The " id e n tific a tio n with the aggressor” hypothesis
prescribes th at the highest co rrelatio n s with the s e lf
score w ill be with parents perceived as s t r i c t e s t in t h e i r
a ttitu d e s towards a n tiso c ia l behavior. The co rrelatio ns
obtained again do not support th is hypothesis. Differences
where the r e l i a b i l i t y of correlations permits an estimate
of significance are not great enough to warrant the con
clusion th at g reater perceived parental punitiveness and
strength of id e n tific a tio n with the parent are positiv ely
re la te d . In the well socialized group, considering only
the re lia b le correlation s the co e fficien ts are of about
the same degree of magnitude whether the parent is per
ceived as s t r i c t or permissive. This comparison in the
poor prognosis group is not possible because the
106
Tkble 8
Correlation Between Punitiveness Assigned to Self and
Punitiveness Assigned to Each Parent, Where Only the
20 Highest Scores (Mother and Father Considered
Separately) in Parental Punitiveness and the
20 Lowest Scores in Parental Punitiveness
are Included in Calculations
S elf- Self-
S tric t S tric t
Mother Father
Self-
Permissive
Mother
Self-
Permissive
Father
W.S. „66* ,40 .56* .63*
P.P. -.11 .24 .14 .39
107
co efficien ts obtained are too small re la tiv e to the size of
the sample to warrant such comparison. The experimental
findings do not confirm the ’’id e n tific a tio n with the
aggressor” hypothesis.
Hypothesis V
This hypothesis deals with the psychoanalytic a sse r
tio n th at good id e n tific a tio n with parental figures whose
own values are congruent with the prevailing standards of
the social milieu w ill insure th a t the in d iv id u a l’s evalua
tion of prohibited behavior and his anticip atio n s as to the
probable punishing-sanctioning reactions of important
figures in his environment w ill be in harmony with the pre
vailing an ticipations and varlues characterizing the well
adjusted individuals in his social environment. Individ-
auls who have formed weak id e n tific a tio n s or have id e n ti
fied with poorly adjusted individuals w ill be character
ized by punishing-sanctioning a ttitu d e s and anticipation s
which are not congruent with such a ttitu d e s and a n ticip a
tion s among individuals well adjusted to the social
environment.
In view of these considerations, we should expect
the poorly socialized in d iv id u a l’s projected conception
of his paren t’s punitiveness towards an tiso cial behavior
and his projected conception of the social reactions to
an tiso cial behavior of people he ’’knows w ell’’ w ill tend
108
to deviate from a "normative range of punitiveness" e sta
blished by his well socialized peers.
S ta tis tic a lly , means calculated for mother, fa th e r,
and "most others" scores in the poorly socialized sample
w ill deviate sig n ific a n tly from sim ilar mean scores cal
culated in the well socialized sample. Differences sig n i
ficant at the 0.05 level of confidence w ill be accepted as
supportive of the hypothesis. Table 9 presents means and
standard deviations for the scores relevant to the
hypothesis.
None of the mean differences calculated reached a
satisfac to ry level of s t a t i s t i c a l significance. Judging .
from the present data i t must be concluded that poorly
socialized and well socialized groups do not d iffe r in the
degree of p u n itiv ity which they project upon parental f i
gures and upon a generalized group of "others" who are
designated as those the subject knows "most w ell." Anti
cipations concerning the probable punishing-sanctioning
reactions of th e ir social milieu in the poorly socialized
group do not d if f e r sig n ific a n tly from those of th e ir well
socialized peers. The hypothesis is not confirmed.
Although sig n ifican t differences in mean punitive
ness between the 3 experimental groups do not appear in
comparing to ta l s e lf, parent, and "most others" scores,
there is the p o s sib ility th at means calculated for
109
Table 9
Group Means Representing the Degree of Punitiveness
Assigned to Each Parent and to "Most Others" in
Well Socialized, Good Prognosis, and Poor
Prognosis Samples
"Most
N Mother SD Father SD Others" SD
W.S. 50 79.94 10.9 S i.12 9.15 77.26 14.94
G.P. 25 SO.60 S.2 SO.04 S .99 77.S4 17.36
P.P. 50 77.16 S .l 77.96 9.26 7S..40 17.33
110
individual items may prove more discrim inating. As a pre
liminary exploration of th is p o s s ib ility the mean puni
tiveness score fo r each item was found in the well so c ia l
ized and poor prognosis groups. These means are presented
in Table 10.
I t would be uneconomically time consuming to explore
th is unhypothesized issue by verifying the significance of
observed differences in the average sentence years assigned
by each experimental group to each item; however, the
reader w ill recognize in these figures th at a detailed
item analysis might enable the experimenter to cu ll out a
group of items which would discrim inate between the experi
mental groups.
The question of group and individual v aria tio n s in
punitiveness on p a rtic u la r items represents an in te re stin g
problem for future research.
Hypothesis VI
The fa ilu re of the tru e -fa ls e items to d iffe re n tia te
between the experimental groups is fu lly discussed in the
" r e s u lts ” section of Fhase One of the study.
Hypothesis VII
Further speculation assumed that the t e s t dimen
sions and re la tio n sh ip s hypothesized to d iffe re n tia te
between poorly socialized and well socialized groups w ill
Table 10
Average Sentence Years* Assigned to Each Test Item by Well Socialized and
Poorly Socialized Groups
N
Bad Checks 4.9
Assault - Black Jack 7.2
Bookkeeper - $3,000.00 S. 2
Gang Fight - Death 29.S
Knife Threat - Rape 27.6
Home Lots - $25,000.00 9.4
Hit and Run - Death 21.5
Kidnap - Death 47.4
Rape - Female, Age 30 29.9
Poison Attempt - Wife 21.7
Armed Robbery - Death 43.0
Beating - Policeman 6.6
Acid Blinding 22.9
Adult Homosexual 7.2
Kidnap-Sex Acts, Female 22.3
Forced Homosexuality-Boy 10 years 16.3
old
Exhibitionism - minor g ir l 12.0
Investment Broker 9.12
Beating - Extortion 12.7
Union Official. 12.1
Self Mother Father
D-PP
JL
D-PP N D-PP
2.56
S.4
5.96 9.S S.4S
6 .OS
11.3
12.56
15.1 13.4S
4.32 10.1 7.52 S. 6
7.24
25.04
32.1 32.2 3S.6 42.60
30.64 33.S
32.04 26. 0S 2S.40
5.64 12. S 9.SS
S.3 9.64
20.24
36.S 35.S 33.6 33.0S
49.20 50.S 52.2 52.2 51. SO
26.0S
34.7 24.11 30.5 24.64
2S.60
19.5
23. SO
25.3 26.40
43.60
46.9
4S.00
42.4
42.60
10.40 10.3 7.96
S.7 S.36
IS. 04
26.1 26.60 2S.2 27.00
S.2S 16.1 11.96 15.6 13.16
21.24 2S.4
30.40 27.2 25.60
24.32 21.S 2S. 44 23.3 2S.40
10. S4
17.0
21.24 22.5 17. SO
6. S4
S.4
5.4S 11.16 11.16
9.40
15.4
11.4S
IS .4 13.32
3.44
11.1 S. 40 12.3 9.96
* /
Scale points are converted into years by a r b itr a r ily assigning a value of 60 years to
c a p ital sentences and 45 years to l i f e sentences.
112
correctly place a less severely delinquent group between
the well socialized and the poorly socialized groups on a
continuum of severity and chronicity of antiso cial be
havior. To meet th is criterio n optimally the data should
show s t a t i s ti c a l ly significant differences on each dimen
sion (degree of id en tificatio n , mean differences in self
and projected parental punitiveness) distinguishing the
good prognosis group from both the well socialized group
and the poor prognosis group. Actually we have only one
significant difference between the well socialized group
and the poor prognosis group. This one difference indi
cates th at the id en tificatio n with the mother in the good
prognosis group is sig nificantly stronger than the iden
tif ic a tio n with the mother in the poor prognosis group.
However, i t should be noted th a t, although the majority of
measures do not distinguish the good prognosis group from
the other 2 experimental groups on a satisfactory level of
s ta t i s ti c a l significance, the mean scores and the correla
tion coefficients lie as would be anticipated at a point
intermediate to the figures obtained for the other two
groups. This suggests that whatever dimension is being
measured the good prognosis group lie s in position between
the well socialized group and the poor prognosis group as
would be predicted. The pattern of differences obtained
does suggest that the good prognosis group has more
113
resemblance to the well socialized sample than to the poor
ly socialized example (See Tables 3, 4, 7 .).
Unhypothesized Relationships
Although mean self-mother and father-scores were
not found to be sig n ifican tly d iffe ren t when compared
across the 3 experimental groups an in terestin g re la tio n
ship did emerge when the mean se lf score was compared with
father and mother scores within experimental groups.
Table 11 presents mean self-mother-and father-scores for
the three experimental groups. The respective standard
deviations are also lis te d . The reader’s attention is
called to the fact that in each experimental group, the
mean father and mother scores are of greater magnitude
than the s e lf score. Simply stated th is indicates that
individuals in a l l experimental groups tended to see the
parental figures as more punitive than the se lf in reacting
to an tiso cial behavior. All the differences between the
self and both parental figures are sig nificant at beyond
the 0.05 level. The t values with corresponding probabi
l i t y level for each relationship are presented in
Table 12.
A general statement can be made to the effect that
white, protestant, adolescent boys of the upper-lower to
lower-middle-class as a group, perceive th e ir parents as
more punitive in judging an tisocial behavior than'they
114
Table 11
Group Means Showing the Degree of Punitiveness Assigned
to Self and to Each Parent in Well Socialized,
Good Prognosis, and Poor Prognosis Groups
N Self SD Mother SD Father SD
W.S. 50 74.96 S .21
79.94
10.90 31.12
9.15
G.F.
25 74*44
S. 60 30.60
3.19 30.04 3.99
P.P. 50 71.63 3.79
77.16
3.07 77.96 9 . 2 6
115
Table 12
The S^gnifcance of Mean Differences in Punitiveness
Between Self and Each Parent Within Well
Socialized, Good Prognosis, and
Poor Prognosis Samples
W . S.
Self vs Father t = 7.25 P = < .001
Self vs Mother t = 5.66 P = < .001
G.P.
Self vs Father t = 3 .4$ P = < .05
Self vs Mother t = 5.65 P = < .05
P.P.
Self Vs Father
Self vs Mother
t = 5 .0 6
t — 6
P = < .01
P = < .05
116
perceive themselves to be. This phenomenon of g re ater per
ceived punitiveness in the parental figures prev ails in
both well so cialized and poorly socialized groups.
A second empirical re la tio n sh ip which appears strong
enough to mention involves the correlations calculated be
tween se lf-p a re n ta l-sc o re s where the 20 most punitively and
20 most permissively perceived fa th e r and mothers were con
sidered. In these calculations only the well socialized
sample and the poor prognosis sample are considered. The
N of 25 in the good prognosis group is too snail to provide
a workable sample of extreme cases.
A comparison to Table & (page 106) with Table 5
(page 9$) suggests th a t th is cu llin g out of parents who
could be called "deviant” in terms of e ith e r permissive
ness or punitiveness, i . e . , both extremes on both dimen
sions, re su lted in obtaining much lower co rrelatio n s be
tween s e lf and parent scores. The reduction in the
strength of the re la tio n sh ip between the self-sc o res and
the parental scores, i . e . , reduction in strength of " id e n ti
f i c a t i o n ,” was p a rtic u la rly marked in the poor prognosis
group. The suggestion is th a t parents perceived as e ith e r
extremely permissive or extremely punitive are associated
with lower self-p aren t id e n tific a tio n p a rtic u la rly in the
poorly socialized group. Differences between the two
groups are marked where the e ffe c t of a punitively per
ceived mother on the strength of self-mother
"id en tifica tio n " is considered. In the well socialized
group where the mother is perceived as highly punitive the
table shows th at the self-mother relatio nship remains
quite strong. In the poorly socialized group however the
strength of the self-mother relatio n sh ip appears to be
markedly reduced where the mother is perceived as highly
punitive. The difference between the two correlations is
sig n ifican t at the 0.02 level of confidence. I t is not
possible to te s t the significance of differences between
other correlation coefficients in Table #. With NTs of 20
in each group, the other co efficien ts are too small to
sa tisfy c r ite r ia for r e l i a b i l i t y . Inspection of the
Table, however, suggests th at the greatest differences
between the two groups are to be found in the self-mother
relatio n sh ip , rath er than the s e lf-fa th e r re la tio n sh ip .
A th ird non-hypothesized finding emerges when mother
and fath er scores are correlated with each other and com
pared across experimental groups. The correlation c o e ffi
cients are presented in Table 13. I t is seen that the
degree of "perceived agreement" between parents is greatest
in the well socialized and good prognosis groups and dimi
nishes in the poor prognosis group. The difference be
tween the correlation in the well socialized sample and
the correlation in the poor prognosis sample is significant
at the 0.01 level of confidence. The correlation in the
118
Table 13
Degree of Correlation Between Punitiveness Ascribed
to Father and Punitiveness Ascribed to Mother in Well
Socialized, Good Prognosis, and Poor Prognosis
Samples (Degree of Perceived Parental
Agreement in Punitiveness Towards
Antisocial Behavior)
Correlation
N .: Mother versus Father
W.S. 50 .80
G.P. 25 . .80
P.P. 50 .61
119
good prognosis group though fa llin g in correct logical
place is not sig n ific a n tly d iffe re n t from the correlations
in eith er of the other groups. This finding suggests th a t
perceived parental "harmony'1 in terms of punishing-sanc-
tioning systems or a ttitu d e s is related sig n ific a n tly both
to higher self-parent id e n tific a tio n and to an adequate
social adjustment. Perceiving parents as disparate in
terms of th e ir punishing-sanctioning a ttitu d e s is asso
ciated with lower self-paren t id e n tific a tio n and inadequate
or delinquent social adjustment.
Review of Experimental Findings
The following declarative statements appear to be
ju s tif ie d on the basis of experimental findings:
1. Non-delinquent boys tend to be more s ig n if i
cantly "identified " with th e ir projected conception of
th e ir p aren t’s punishing-sanctioning a ttitu d e s to a n ti
social behavior than are delinquent boys of the same age,
race, re lig io n , and socio-economic background.
2. Non-delinquent boys considered as a group, are
not more closely id e n tifie d with the projected conception
of the f a th e r ’s punishing-sanctioning a ttitu d e s . Stronger
fa th e r-id e n tific a tio n , re la tiv e to the m other-identifica-
tion does not distinguish between delinquent and non-delin
quent groups. Stronger id e n tific a tio n with both parental
figures is c h a ra c te ristic of well socialized boys.
120
3. Poorly socialized boys, considered as a group,
do not d if f e r markedly from well socialized boys, consi
dered as a group, in the strength of expressed punitiveness
towards a n tiso c ia l behavior. Projected conceptions of
parental punitiveness also do not discrim inate between
groups of well socialized and poorly socialized boys. At
le a s t, on a conscious, ra tio n a l lev el poorly socialized
boys, when considered as a group, are able to project a t
titu d e s of punitiveness to a n tiso c ia l behavior which are
within normative lim its of punitiveness established by
a well so cialized group of adolescent boys; and th e ir a n ti
cipations of the punitiveness of parental figu res and a
group of generalized "o thers” do not vary notably from
lik e projected an ticip atio n s of well socialized boys where
group measures are concerned.
4. Greater perceived parental s tric tn e s s or puni
tiveness is not generally associated with stronger id e n ti
fic a tio n with the projected parental punishing-sanctioning
system in e ith e r poorly socialized or well socialized boys.
5. The most sen sitiv e discrim inator between groups
of adolescent boys, varying in the chronicity and severity
of th e ir a n tiso c ia l behavior, proves in t h i s experiment to
be the dimension, "strength of id e n tific a tio n with the
projected punishing-sanctioning system of the mother."
6. Non-delinquent boys tend to project conceptions
121
of the parental punishing-sanctioning systems which are
more in agreement with each other, i . e . , b e tte r correlated
than the projected parental punishing-sanctioning systems
of delinquent boys (in other words, non-delinquent boys
see th e ir mothers and fathers as more lik e each other in
th e ir punishing-sanctioning a ttitu d e s than do delinquent
boys).
7* Adolescent boys, whether delinquent or non-delin
quent, as a group are characterized by tending to see th e ir
parents as more punitive than themselves in T fjudging”
an tiso cial behavior.
&. In adolescent boys the strength of the self-id si-
t i f i c a t i o n with projected conceptions of the parental
punishing-sanctioning systems tends to be reduced where
the parent is perceived as e ith e r extremely punitive or
extremely permissive towards a n tiso c ia l behavior. This
is most p a rtic u la rly true in the delinquent group where
self-parent id e n tific a tio n s become negligible, when the
parents are perceived as extreme in eith er punitiveness
or permissiveness.
9. The most marked difference between well so cial
ized and poorly socialized boys in strength of id e n tific a
tio n with the parental punishing-sanctioning system occurs
where the parental figure is a mother perceived as highly
punitive. Poorly socialized boys are characterized by
122
very poor id e n tific a tio n with the projected punishing-sanc
tioning system of mothers perceived as highly punitive.
By contrast well socialized boys who perceive th e ir mothers
as highly punitive maintain a high degree of id e n tific a tio n
with th e ir projected conception of her punishing-sanction
ing system.
CHAPTER V I I I
DISCUSSION
Experimental expectations based on the psychoanaly
t i c theory of superego development were th a t groups vary
ing in the quality of th e ir adjustment to social proscrip
tions should be discriminable in terms of: 1. the
strength of th e ir id e n tific a tio n with projected parental
punishing-sanctioning systems, 2. the strength of id e n ti
fic a tio n with the projected fath er re la tiv e to the mother,
3. the strength of id e n tific a tio n with the' salien t "d is
cip lin in g ” parent re la tiv e to the other parent, and 4. the
degree of deviation from the norm of punitiveness in a
well socialized group in assignments of punishment to
a n tiso c ia l behavior and in estimates of parental punishing
reactions to a n tiso cial behavior. What statements can be
made with respect to these expectations on the basis of
experimental findings?
Perhaps the most strik in g discrepancy between
th e o re tic a l expectations and experimental outcomes involves
the re la tiv e roles of father and mother in the so c ia liz a
tio n of the child. Oriented by a c la ssic a lly Freudian
conception of the fa th e r as the patriarch of the family,
the ’law-giver" and the agent of d iscip lin e, one might ex
pect th a t reactions to the fa th e r as "authority fig u re”
-123-
would prove to be the most discrim inating dimension in
d iffe re n tia tin g between groups of boys varying in terms
of th e ir adjustment to the rules of society. One might
also expect th a t, as the "authority figure" the fa th e r
would be perceived by adolescent boys as somewhat stern ,
more uncompromising, more demanding in terms of his stan
dards of conduct than the mother. Our th e o re tic a l posi
tio n then, leads us to expect th a t delinquent boys d iffe r
most sharply from non-delinquent boys in the strength of
id e n tific a tio n with a firm, standard settin g paternal
figure. Instead, we find th at s t a t i s t i c a l l y sig n ifican t
differences between experimental groups are found in the
degree of id e n tific a tio n with the punishing-sanctioning
system of the mother, who is perceived as equal to the
fath er, or a t le a s t, not d iffe re n tia te d from the fath er
in terms of the strength of her punitiveness towards
unacceptable behavior. About the same level of punitive
ness and the same degree of id e n tific a tio n (Tables 9 and
5) are found with the mother as with the fa th e r in a l l
experimental groups, i . e . , i t is the strength of id e n t i f i
cation with the parents not the direction of the i d e n ti f i
cation which appears to d iffe re n tia te the delinquent group
from the non-delinquent group.
These findings appear to contradict the conclusions
of other studies in which stronger fa th e r "id en tifica tio n "
125
was found to be correlated with more positive social t r a i t s
and re la tiv e absence of pathological trends. For example,
Sopchak (1952), using sim ila rity between s e lf responses to
the M M PI and projected (role playing) parental responses
as an operational d efin itio n of id e n tific a tio n , showed
that stronger s e lf- fa th e r id e n tific a tio n re la tiv e to
self-mother id e n tific a tio n was correlated in college males
with an absence of pathological trends as measured by the
MMPI. In another study Lazowick (1955), employing the
semantic d iffe re n tia l, found th a t high "semantic sim ilarity "
(his operational d e fin itio n of "id en tifica tio n ") between
s e lf and fath er in college student males was correlated
with low anxiety as measured by the Taylor Manifest
Anxiety Scales, while the converse was true with males
showing low s e lf-fa th e r semantic sim ila rity and high
self-mother semantic s im ila rity . Lazowick compared the
s u b je c t’s responses with the actual responses of each
paternal figure rather than the su b ject’s projected con
ception of what the parent’s response would be. Payne
and Mussen (1956), also using actual sim ilarity between
se lf and parents in responding to selected scales of the
California Personality Inventory, found high father-identi-
fic a tio n re la tiv e to m other-identification to be correla
ted with calmness, warmth, and frien d lin ess and g re ater
"m asculinity." Father id e n tifie d boys more often per
ceived th e ir fam ilies as "warm and rewarding."
126
The discrepancy between these studies and the find
ing of the present study, th a t high m other-identification
is at le a s t equal to high fa th e r-id e n tific a tio n as a cor
re la te of well socialized behavior, leads to the specula
tio n th a t experimental findings and conclusions with regard
to id e n tific a tio n processes may vary depending upon the
d e fin itio n of id e n tific a tio n employed, the experimental
technique, and the nature of c rite rio n groups.
In comparing the findings of the present study with
the studies ju st reviewed i t develops th at where the d e fi
n itio n of id e n tific a tio n involves perceived self-p aren t
s im ila rity in terms of punishing-sanctioning behavior and
the c rite rio n is freedom from a n tis o c ia l behavior, high
m o th er-identification is an advantage, i . e . , i t tends to
be correlated with non-delinquent behavior. However,
where the c rite rio n of general psychopathology is absent
and the measure of id e n tific a tio n is perceived self-p aren t
s im ila rity in answering M M PI items, high mother id e n tif ic a
tio n is found to be a disadvantage, i . e . , i t is correlated
with increased pathology in males. Sim ilarly high r e la
tiv e congruence between s e lf and fa th e r in responses to
t r a i t type questions, e .g ., tolerance, social p artic ip a tio n
and m asculinity-fem ininity items on the C alifornia Per
sonality Inventory, is associated with positive social
q u a litie s .
127
The impression is gained that " id e n tific a tio n ” with
a parental figure may not be an " a ll or none" phenomenon
such th at persons can be designated as "fath er-id en tified "
or "m other-identified." Perhaps one could specify what
aspect, t r a i t , or habit in the parental figure is involved
in the " id e n tific a tio n ," and study what effect id e n tific a
tion with p a rtic u la r t r a i t systems with p a rtic u la r parents
may have on personality and behavior. The above discus
sion suggests th at the parent is not "taken in" altogether,
ra th e r i t appears, th at a given child could iden tify with
one parent’s punishing-sanctioning a ttitu d e s, the other
parent’s a ttitu d e s towards food and somebody e l s e ’s a t t i
tudes about work. To break i t down in another way, the
child might identify with the mother’s punishing-sanction
ing a ttitu d e s around sex and the f a th e r’s punishing-sanc
tioning a ttitu d e s around physical aggression. The idea is
th a t, id en tific a tio n s may represent more p a r tia l, segmented,
d istrib u ted phenomena than presently viewed and treated in
theory and in experimental work.
The question remains; what " id e n tific a tio n ," i f
any, may be important in the prevention of an tiso cial
behavior: In so far as engaging in an tiso cial behavior is
concerned, i t appears not to matter which parental value
system is id en tified with just as long as the id e n ti
fic a tio n made is strong enough to insure adaptation to
social demands. Boys may be id e n tifie d with th e ir mothers,
th e ir fa th e rs, or have a kind of mixed father-mother
id e n tific a tio n . As long as adequate standards have been
internalized (e ith er mother’s or f a th e r’s) there would
seem to be less likelihood of protracted an tiso cial be
havior. The most optimal conditions fo r well socialized
behavior would appear to be strong id e n tific a tio n with
both parents who are perceived as harmonious in th e ir
a ttitu d e s to forbidden behavior.
The reader is reminded th at only the e ffe c t of
id e n tific a tio n with the parental punishing-sanctioning
system on ’’law breaking” behavior has been considered in
th is study. Other personality t r a i t s , c o n flic ts ,o r d is
turbances which do not involve v io latio n of legal pro
h ibitions may be correlated with specific patterns of
id e n tific a tio n s, e .g ., strong m other-identification but
th is hypothesis has not been te s te d . The experimental
findings support the statement th a t, strong id e n tific a
tion with perceived parental punishing-sanctioning systems
(without regard to the sex of the parent) is po sitively
related to law-abiding behavior. There are h in ts in the
present study th at the role of the mother in ’’superego”
development may be underestimated by c la ssic psychoanaly
tic theory. Changes in family structure and sex roles
which have occurred since the psychoanalytic theory of
129
so cializatio n was developed can be invoked in speculating
about th is experimental finding th at mother’s a ttitu d e
towards unacceptable behavior may have a lo t to do with a
boy’s a b ility to keep out of trouble. The fact th at the
psychoanalytic theory was developed in a strongly p a tr i
archal, European cu ltu ral milieu d ifferin g in sig n ifican t
features from patterns of family in teractio n in th is coun
try may also have relevance h e re .
Some ten tativ e remarks w ill now be made with re
gard to unexpected experimental findings which bear on the
mechanism or process of id e n tific a tio n . There is some
evidence (Table & ) th a t, contrary to the "id e n tific a tio n
with the aggressor" hypothesis, extremes of punitiveness
(and permissiveness) tend to weaken self-p arent i d e n t i f i
cation. To fu rth er v i t ia te the simple " id e n tific a tio n
with the aggressor" hypothesis of id e n tific a tio n , there is
ten ta tiv e evidence th a t re la tiv e punitiveness may have
d iffe re n tia l effects upon the strength of id e n tific a tio n
depending upon the sex of the parent. I f we assume th at
the trends suggested in Table & are re lia b le and would
prove sig n ific a n t with a larg er sample of subjects, i t
would look as though high punitiveness in the fa th e r has.
a weakening effect on s e lf- fa th e r id e n tific a tio n , while
high punitiveness in the mother does not have th is weaken
ing effect or a t le a st not to the same degree as in the
130
fa th e r. In the well socialized group, where parents per
ceived as extremely permissive or extremely punitive are
considered, the highest self-parent correlations are ob
tained with the most punitive mothers and the most per
missive fa th e rs. To fu rth e r complicate the picture,
while high punitiveness in the mothers of well socialized
boys is associated with strong id e n tific a tio n , in the
poorly, socialized group i t is associated with a v irtu a lly
complete lack of id e n tific a tio n with the values of the
mother. These rath er tenuous findings are most useful in
pointing up the complex nature of processes of value as
sim ilation and stimulating fu rth er questions and lin es of
research. Perhaps, the d iffe re n tia l effect of highly
punitive mothers in the two experimental groups may be
explained in terms of a fa c to r not considered d irec tly in
the present design. This would be, the affectio n al r e la
tionship with the mother. That a combination of awe and
respect with love and warmth would achieve the strongest
id e n tific a tio n s appeals to common sense and is supported
by researdh findings. Why perceived stric tn e s s or puni
tiveness in the fath er would tend to weaken id e n tific a
tion where i t does not have th is effect on the self-mother
relatio nship in well socialized boys is an open question.
These d iffe re n tia l effects suggest some form of
in teraction between sex of parent, re la tiv e punitiveness,
131
strength of id e n tific a tio n and quality of social adjustment.
The p o s s ib ility th at social adjustments, personality in te
gration, and re la tiv e strength of id e n tific a tio n with each
parent may vary with d ifferen t patterns of father-mother
in teractio n also seems to be raised by these findings. The
impression is th a t i t would be useful to study the parents
as an in te ra c tin g unit rath er than as separated figures
and te s t the in terac tio n between combined parental e ffe c t
on the child and re la tiv e strength of id e n tific a tio n , quali
ty of social behavior, degree and kind of pathology, and
so on. As an example, one might t e s t the strength of iden
t i f i c a t i o n with each parental figure under the following
conditions: fath er very punitive, mother very permissive;
both parents highly punitive; mother highly punitive,
fath er very permissive; both parents very permissive; and
so on, re la tin g each of these conditions to social ad ju st
ment, degree and kind of psychopathology, strength of
id e n tific a tio n , or whatever external correlate is hypo
thesized.
The unhypothesized finding th at adolescent boys, in
general, perceive th e ir parents as more punitive than they
perceive themselves to be in th e ir a ttitu d e towards a n ti
social behavior, is harmonious with psychoanalytic ac
counts of the psychodynamics of adolescence. Such psycho
analytic w riters as Anna Freud (1937), Josselyn (1952),
and Fenichel (1945) speak of the renewal of in fa n tile
c o n flic ts and the in te n s ific a tio n of basic aggressive and
sexual striv in g s in the ’’sturm und drang” period beginning
at puberty. Josselyn, in p a rtic u la r, speaks of the
adolescent’s need to project punitive images of the pa
re n ts, s u p e rfic ia lly as a ra tio n a liz a tio n fo r emancipatory
struggles against the parent’s auth ority, but more deeply
to provide themselves with a sense of strong external
environmental controls opposed to th e ir own feared in
s tin c tu a l s triv in g s. The projection of s t r i c t a u th o rita
tiv e a ttitu d e s on the parents may provide the adolescent
with the fe elin g th a t his parents are strong enough to
prevent him from ’’losing c o n tro l,” a basic fe ar in c h il
dren of th is age. In th is sense, one may hypothesize th at
the perception of the parental figures as more punitive
than the s e lf is an index of adolescent immaturity, i . e . ,
of a p e rsistin g in secu rity with regard to the capacity of
the s e lf to exercise re lia b le , independent control over
the impulses and a continuing ambivalent dependency upon
the needed lim its and control. The findings of th is study
are th a t, well socialized and poorly socialized adolescent
males tend to have more in common the need to project an
a u th o rita tiv e or punitive image of the parent. They d if
f e r only in the strength of th e ir ’’id e n tific a tio n ” with
the punitive image projected. I t would be in te re s tin g to
see whether punitive parental image would disappear in a
133
group of well socialized ad u lts, but p e rsis t in adults
c la s s ifie d as neurotic or poorly socialized .
The suggestion from psychoanalytic theory th a t well
socialized and poorly socialized individuals could be ex
pected to d iffe r in th e ir m oral-ethical or punishing-sanc
tioning a ttitu d e s towards an tiso c ia l behavior and d iffe r
also in the punishing-sanctioning a ttitu d e s projected upon
the interpersonal environment tends to receive only mini
mal and ra th e r equivocal support in t h i s study. The data
do support a statement to the e ffe c t that with large NTs
poorly socialized groups w ill show on the average s lig h tly
but s ig n ific a n tly more permissive a ttitu d e s towards a n ti
social behavior than w ill well socialized groups. Judging
from the present study, however, many individuals with a
h isto ry of extensive involvement in a n tis o c ia l behavior
are able to project punishing-sanctioning a ttitu d e s which
are quite congruent with lik e a ttitu d e s in th e ir well .
socialized peers.
There is some evidence from our data th a t there is
some v aria tio n in punitiveness in the 3 experimental
groups as a function of the type (as opposed to the seve
r ity ) of the criminal act considered. The fact th at the
most severe, poorly socialized group gives an average sen
tence of 24.32 years fo r adult homosexual m olestation of
a ten year old boy while the well so cialized group assigns
134
an average sentence of 16;3 years fo r the same crime is
suggestive in th is regard. This difference suggests th a t
loading the te s t with such homosexual items would re su lt
in re la tiv e ly higher average punitiveness in the poorly
socialized group, the reverse of the trend reported.
The impression - is th a t, extensive item analysis
employing many d ifferen t types of crimes might well turn
up psychodynamically in te re stin g and from a diagnostic
standpoiht useful differences in the sentencing behavior
of various c rite rio n groups. As an i l lu s tr a tio n of th is
p o te n tia lity for fu rth e r research with the experimental
instrument, the re a d e r’s atten tio n is called to the in
te re stin g fa c t th a t adolescent males in the well socialized
group assign an average sentence of 30 years in prison for
the crime of simple rape without severe harm to the v ic
tim, while assigning a le s s e r average of 22 years fo r the
crime of attempted murder by poisoning. This re la tiv e ly
more intense reaction to aggressive, uncontrolled sexual
ity in well socialized adolescent males tends to invoke
speculation with regard to d iffe re n tia l cu ltu ral sanctions,
psychoanalytic notions with regard to the adolescent’s
fear of sexuality, defenses against the in s tin c ts , and
so on.
Perusal of Table 10 also suggests th a t crimes of
great severity tend to e l i c i t sim ilarly severe punitive
135
responses in a l l 3 experimental groups while crimes of .
le sse r sev erity are responded to in a more variable way.
One could hypothesize that the social response to gross
criminal acts is so well defined or sa lie n t as to o b lit
erate the determining factors of subtle intraindividual
personality facto rs which might emerge in less well
structured stimulus s itu a tio n s. In constructing a re
vised Instrument then one would avoid the severe crime
and focus on more innocuous offenses.
With regard to psychoanalytic concepts of the
so cializatio n process, perhaps the most adequate generali
zation is that the psychoanalytic emphasis on T ,id e n tific a -
tio n ” with the punishing-sanctioning a ttitu d e s , which
children perceive in th e ir parents, as an important de
terminant of the quality of social adaptation has been
given experimental support. Experimental re su lts have,
however, not tended to confirm classic psychoanalytic
concepts dealing with the mechanism or process through
which adequate id e n tific a tio n and healthy superego de
velopment occurs. Suggestions for fu rth e r experimental
exploration of the processes of value assim ilation, using
the experimental instrument have been advanced.
I t seems possible to sta te th at the present experi
mental findings indicate a degree of progress in the a t
tempt to cast a meaningful set of genotypical concepts
into a form amenable to normative, q u an titative experi
mental procedures. A q u an titative variable, strength of
id e n tific a tio n with the projected parental punishing-^sanc-
tion in g system, has been iso la te d and shown to be a sig n i
fica n t dimension of difference between well socialized and
poorly socialized individuals. This dimension is not
em pirically developed but derives from and has a re la tiv e ly
well defined log ical re la tio n sh ip to psychoanalytic hypo-
theses dealing with the process of so c ia liz a tio n . The
logical-conceptual properties of the instrument allow for
a d ire c t application to th e o re tic a l problems in the area
of id e n tific a tio n , value assim ilation, child-parent in te r
action, and so on.
The instrument provides fo r investig atio n of a num
ber of independently varying but re la te d dimensions, e .g .,
id e n tific a tio n re la tiv e to mother, fath er; level of puni
tiveness re la tiv e to mother, fath e r, group norms; per
ceived harmony between parental value systems; parental
values versus normative values, and so on. Individual
id e n tific a tio n patterns incorporating several d iffe re n t
scores and amenable to expression in p ro file form provide
a basis for a b e tte r degree of individuation than is pos
sible with single score t e s t s . This b u ilt- in v aria tio n
on the single theme of id e n tific a tio n with perceived
punishing-sanctioning systems tends to widen the
137
instrum ent’s potential a p p lic a b ility to the exploration of
hypotheses with regard to d iffe re n tia l id e n tific a tio n pat
tern s in various c rite rio n groups and could serve as a
basis fo r the derivation of a lo g ically integrated system
of diagnostic categories i f experimental validation of
’’id e n tific a tio n p attern s” showed th a t the lin e s of cleav
age had predictive value. A claim to statu s as a ’’pro
jectiv e technique” can be made for the instrument since
in the absence of well defined group ’’norms” for sen
tencing behavior, the respondent must draw on his own more
subjective a ttitu d e s in assigning sentences. There is at
the same time a well defined q u antitative element in th at
t e s t responses can be expressed in terms of sentence
years.
Experience with the experimental instrument in th is
study has suggested th a t i t could be improved in i t s d is
criminatory power by fu rth e r item analysis, and possibly
by fu rth e r changes in te s t format. The finding, for
example, th at the c rite rio n groups did not respond d if
fe re n tia lly to crimes of great severity indicates th at
the differences found between the c rite rio n groups, both
in terms of re la tiv e punitiveness and in strength of
id e n tific a tio n , would be enhanced by elim inating descrip
tions of severe crimes and amplifying items more innocu
ous in content. The finding th a t differences were
enhanced by making re call from s e lf "judgments" to "judg
ments" for the parents more d if f ic u lt suggests that i t
would be worthwhile to diminish this re call factor s t i l l
further by providing completely d ifferent items for se lf,
mother, and father sections and gauging the strength of
id en tificatio n with parents by calculating discrepancies
between standard scores. To a c lin ic a lly oriented person
th is diminishing of reference points by which the subject
may guide his response is best communicated as an attempt
to increase the projective properties of the instrument.
Projective features could be further enhanced by in s tru c t
ing subjects to write in th e ir sentences (as a p articu lar
number of years) rather than selecting a point on a scale.
This would, of course, have the effect of diminishing
reference points considerably while increasing, range or
variation of response.
In concluding, an important reservation with regard
to the application of the instrument to the experimental
exploration of psychoanalytic hypotheses must be made.
This has to do with the question of the various "levels"
of consciousness postulated by psychoanalytic theory and
the manner in which conscious or preconscious responses
may be integrated with unconscious striv in g s. A very
knotty and perhaps, at this point, irresolvable problem
lie s in specifying what re la tio n the present experimental
d efin itio n of ’’id e n tific a tio n ” has to the kind of id e n ti
fic a tio n psychoanalysts speak of in describing the course
of superego development. May i t be assumed th at a great
discrepancy between one’s own punishing-sanctioning a t
titu d e s and one’s projected conception of one’s f a th e r ’s
punishing-sanctioning a ttitu d e s is indicative of a r e la
tive fa ilu re to id en tify with the fath er in childhood?
Does the discrepancy between the ’’s e l f ” punitiveness score
and the ’’fa th e r” punitiveness score re la te d ire c tly to the
situ a tio n prevailing in the personality at the most in
tr in s ic or fundamental level of personality structure?
An objection may be made to the effect th at ’’id e n tific a
tio n ” is an unconscious phenomenon. What appears in
conscious behavior may, because of defensive d isto rtio n ,
be completely misleading in so fa r as the ’’tru e ” situ atio n
prevailing in the unconscious is concerned. Thus, close
id e n tific a tio n withthe fath er as defined in the present
experiment might actu ally represent a defensive a ttitu d e
based on ’’weak” unconscious id e n tific a tio n with the fath er.
Contrariwise, one might surmise th at sharp co n flict between
one’s own a ttitu d e s towards forbidden behavior and the pro
jected conception of one’s f a th e r ’s a ttitu d e s on the punish-
ment-reward dimension might represent strong unconscious
id e n tific a tio n with the fath e r. The most dismal probability
of course, is th at responses to the te s t have nothing what
ever to do with id e n tific a tio n processes as they are
conceived by psychoanalytic theory.
This study explores the reasonable hypothesis th at
optimal id e n tific a tio n with parental figures in ch ild
hood may be inferred from evidence of the presence of a
quasiconscious or preconscious core of agreement between
a su b je c t’s own punitiveness and his projected conception
of his parent as a ’’punisher.” In th is context, ’’id e n ti
f ic a tio n ” is conceived of as a necessary l a t e r residual
product or derivative of the e a r lie r Oedipal id e n tific a
tio n .
To view parents as b a sic a lly ’’reasonable” in th e ir
punitive a ttitu d e s , to share a common area of agreement
with parental figures, where forbidden behavior is evalua
ted, is tre a te d here as the probable outcome of optimal
id e n tific a tio n with the parents. ’’I d e n tif ic a tio n ,” as
operationally defined in th is study, may be viewed as a
necessary re su ltan t or l a t t e r day re fle c tio n of e a r lie r ,
now unconscious or preconscious events in personality
development. Whatever the v ic issitu d e s in the play of
drive and defense on the unconscious le v e l, i t is assumed
th a t the fin a l product of optimal reso lu tio n of the in
fa n tile striv in g s is a functional harmony between the se lf
and the parent where a ttitu d e s towards forbidden behavior
are concerned. This is the hypothesis upon which the
logic of the experimental instrument r e s ts .
141
The conceptualization of a "reasonable" judging
o rien tatio n , congruent with projected parental values and
with em pirically defined "normative" lim its of punitive
ness in a p a rtic u la r social m ilieu, as evidence of benign
superego development does seem to be supported by psycho
analytic l i t e r a t u r e .
I t is hypothesized th a t the experimental task w ill
serve to c a ll out not completely consciously determined
responses but ra th e r quasiconscious sets and determining
tendencies which are rooted in unconscious processes and
are not necessarily "known" by the subject on a completely
conscious le v e l. Evaluation as to the v a lid ity of th is
unsupported assertio n re s ts on experimental te s ts of the
hypotheses upon which the t e s t is based.
Further work with the experimental instrument may
demonstrate th at responses to i t are best evaluated as
re la te d to the most "conscious levels" of the superego
stru ctu re ra th e r than as d ire c t expressions of uncon
scious re la tio n sh ip s. This does not obviate the p ossib i
l i t y th a t differences in "superego content" and emphasis
at the most surface "layers"of personality may have
c lin ic a l significance.
The indication th at a sp ecific psychoanalytic
hypothesis may su ffer varying and contradictory fa te s
under d iffe re n t experimental conditions, varying opera
tio n a l d e fin itio n s, and d iffe re n t measuring instruments
emphasizes the need to avoid a too direct and speedy ap
p licatio n of experimental findings in choosing whether to
accept or re je c t psychoanalytic hypotheses. The sugges
tio n is th at positive conclusions with regard to theore
t i c a l assertions should be based on an accumulation of
findings from many experiments in which varying opera
tio n al d efin ition s and experimental instruments are tested
against appropriate c rite rio n groups. Both tools and con
cepts can be refined in th is way.
CHAPTER IX
SU M M A R Y
Problem
In t h i s experiment an attempt was made to develop
and p a r tia lly v alid ate a t e s t instrument which would d is
criminate in a c lin ic a lly meaningful way among individuals
characterized by a n tiso c ia l behavior p attern s.
Method
The stru ctu re and content of the experimental in
strument were determined through an adaptation of the
psychoanalytic concept of id e n tific a tio n . For purposes
of the study, strength of id e n tific a tio n with each parent
was defined in terms of the degree of co rrelatio n between
a s u b je c t’s punitive responses to various a n tiso c ia l acts
and his imagined estimate of each p aren t’s punitive re s
ponse to equivalent a n tiso c ia l a c ts. -
S pecifically , each subject was asked to place him
s e lf in the role of a judge.in a court of law and to as
sign an appropriate number of sentence-years to imaginary
offenders, found.guilty of various criminal ac ts. The
subject, having ’’sentenced" the prisoner as he would i f
he were the judge, was then asked to estimate what sen
tences his fa th e r would give in the role of judge and
-143-
144
l a s t l y what sentences his mother would assign in the role
of judge.
This technique yields a to ta l "punitiveness" score
for " s e lf ," "mother," and " fa th e r." Strength of id e n t i f i
cation with each parent is then operationally defined in
terms of the degree of c o rrelatio n between s e lf and fath er
scores and between s e lf and mother scores.
Experimental hypotheses were based on the psycho
analy tic theory of superego development. Specific pre
dictions were th at a nondelinquent group of adolescent boys
would be d iffe re n tia te d from a delinquent group by;
1. stronger id e n tific a tio n with th e ir own projection of
each p a ren tTs punishing-sanctioning a ttitu d e s , 2. stronger
same-sex or " fa th e r-id e n tific a tio n " re la tiv e to "mother-i-
d e n tif ic a tio n ," 3. stronger id e n tific a tio n with the
parent perceived as the most s a lie n t d iscip lin in g agent
(the parent perceived as most pu nitive), and 4* by the
degree of punitiveness projected on the parents.
The experimental sample consisted of 125 Caucasian,
protestant boys, of sem iskilled and sk ille d labor back
grounds, between sixteen and eighteen years of age. A
t e s t of the re la tio n sh ip between responses to the experi
mental instrument and in te llig e n c e provided evidence th at
in tellig en ce was not a fa c to r in the determination of t e s t
scores, however a l l subjects were of at le a s t average
145
in tellig e n ce. The to ta l experimental sample was divided
into three subgroups. The f i r s t consisted of 50 highschool
boys with no known delinquent h isto ry , of good highschool
reputation, and sa tisfa c to ry academic standing. A second
group consisted of 50 delinquent boys who had been com
mitted to the C alifornia Youth Authority and judged as
’’severe" or "poor prognosis" delinquent p erso n alities by
experienced c lin ic ia n s. The th ird experimental group was
comprised of 25 delinquent boys, committed to the Youth
Authority and judged to have re la tiv e ly "less severe"
an tiso cial p erso n alities and a re la tiv e ly "good pro
gnosis."
A s t a t i s t i c a l analysis of mean differences in
punitiveness within and between experimental groups and
calculations of the significance of group differences in
correlation of"Self" with "mother" and "father" scores
was performed.
Results
As predicted nondelinquent boys show stronger
’’identification" with projected parental punishing-sanc-
tioning a ttitu d e s than delinquent boys; and delinquent
boys judged to have less serious delinquent patterns show
stronger id e n tific a tio n with parents than boys judged to
have more serious delinquent p atterns.
146
There is no experimental evidence of a stronger
same-sex or " fa th e r-id e n tific a tio n " in the nondelinquent
group. Also experimental re su lts do not support the hypo- *
th e sis th at nondelinquent boys w ill show a stronger iden
t i f i c a t i o n with the most s a lie n t d iscip lin in g parent (the
most punitively perceived parent) than with the other
parent.
With regard to the question of d if f e r e n tia l puni
tiveness in the three experimental groups, re s u lts suggest
th a t there are no marked differences between delinquent
and nondelinquent boys in the in te n s ity of s e lf-p u n itiv e -
ness towards a n tis o c ia l behavior nor in the in te n s ity of
projected conceptions of parental punitiveness towards
a n tiso c ia l behavior. I t is only in the "strength of
id e n tific a tio n " with the punishing-sanctioning a ttitu d e s
projected on parents th a t the experimental groups d iffe r,
not in the absolute level of in te n s ity of such a t t i
tudes.
There were several unhypothesized findings. The
f i r s t of these involves a marked trend in both delinquent
and nondelinquent groups towards viewing parents as more
punitive than the s e lf in th e ir a ttitu d e s towards a n ti
social behavior. A second unhypothesized finding was th at
delinquent subjects who perceived th e ir mothers as highly
punitive towards a n tiso c ia l behavior tended to display
147
negligible id e n tific a tio n with the punitive a ttitu d e s
projected on the mother. Nondelinquent boys on the other
hand maintained a very high degree of id e n tific a tio n with
mothers perceived as highly punitive in th e i r a ttitu d e s
towards criminal behavior. The th ird unpredicted finding
was th at nondelinquent boys tend to see th e ir parents as
more in agreement with one another or more harmonious in
th e ir a ttitu d e s towards criminal behavior than do d elin
quent boys.
Conclusions
In general the experimental findings were in t e r
preted as supportive of the psychoanalytic emphasis on
id e n tific a tio n processes as important determinants of the
quality and character of social adaptation. Experimental
re su lts did not, however, tend to confirm the psychoana
ly tic concept of "id e n tific a tio n with the father-agressor"
as the basic mechanism upon which adequate id e n tific a tio n s
and healthy superego development are alleged to depend in
male children. The unpredicted finding th a t strong id e n ti
fic a tio n with the punishing-sanctioning a ttitu d e s of the
mother is at le a s t as sig n ifican t a correlate of nondelin
quent behavior as id e n tific a tio n with the father, sug
gested a maternal facto r in the so cializatio n of male
children, not adequately emphasized in classic Freudian
theory. The fact that the psychoanalytic theory of
14S
superego development originated in a strongly p a tria rc h a l,
European cu ltu ral m ilieu, d ifferin g in s ig n ific a n t features
from patterns of family in te ra c tio n in th is country was
seen to have relevance in th is context.
The study supports the statement th a t strong id e n ti
fic a tio n with perceived parental punishing-sanctioning
a ttitu d e s (without regard to the sex of the parent) is
p o sitiv e ly re la te d to law-abiding behavior in adolescent
males. The reader is reminded th a t only the effect of
id e n tific a tio n with parental punishing-sanctioning a t t i
tudes on "law-breaking" behavior was considered in the
experimental design. Other personality t r a i t s , c o n flic ts,
or disturbances which do not involve v io la tio n of legal
proscriptions may be correlated with specific patterns of
id e n tific a tio n , e .g ., strong m other-identification, but
th is hypothesis was not te ste d in the present study.
The finding th a t both delinquent' and nondelinquent
adolescent boys see th e ir parents as more punitive than
themselves in "judging" a n tiso c ia l behavior was i n t e r
preted as a possible index of adolescent in secu rity with
regard to the capacity of the s e lf to exercise re lia b le ,
independent control over the impulses and a continuing
ambivalent dependency upon the parents and other au th o ri
t i e s to provide a sense of external lim its and control.
I t was fin a lly concluded that a p o te n tia lly useful
149
dimension of discrim ination among individuals characterized
by a n tiso cial behavior patterns had been established. A
quantitative variable, "strength of id e n tific a tio n with
projected parental punishing-sanctioning a ttitu d e s" was
iso lated which appeared to be worthy of fu rth e r experi
mental investigation .
A review of studies indicating th at a specific
psychoanalytic hypothesis may su ffer varying and contra
dictory fates under d iffe ren t experimental conditions,
varying operational d e fin itio n s, and d iffe re n t measuring
instruments emphasized a need to avoid a too d irec t and
speedy application of experimental findings in choosing
whether to accept or re je c t psychoanalytic hypotheses.
The suggestion is th a t positive conclusions with regard
-to th e o re tic a l formulations should be based on an accumula
tion of findings from many experiments in which varying
operational d efin itio n s and experimental instruments are
tested against appropriate c rite rio n groups.
REF ERENCES
REFERENCES
Aichorn, A. Wayward youth. London: Putnam, 1937.
A llport, G. W . Personality, a psychological in te rp re -
p re ta tio n . New York: Holt, 1937.
Astin, A. W . A fa cto r study of the M M PI psychopathic
deviate scale. J. consult. Psychol., 1959, 23,
116- 118.
Cava, Esther L. & Rausch, H. L. Id e n tific a tio n and the
adolescent boys perception of his fa th e r. J. abnorm.
soc. Psychol., 1952, 42.> 855, 856.
Chang, Judy & Block, J. A study of id e n tific a tio n in
male homosexuals. J. consult. Psychol., I960, 24.
307-310.
Clinard, M.B. Sociology of delinquent youth. New York:
Harper & Rinehard & Co., In c ., 1958.
Cohen, A. K. Delinquent boys, the culture of the gang.
Glencoe, 111.: The Free Press, 1955.
Dennis, W . Readings in general psychology. New York:
Prentiss Hall In c., 1949.
Doctor, R. F. & Winder, C. L. Delinquent versus
non-delinquent performance on the Porteus q u alitativ e
maze t e s t . J. consult. Psychol., 1954, 1$, 71, 73.
Erikson, E. H. Childhood and so ciety . New York:
W . W . Norton & Co., 1950.
Fenichel, 0. The psychoanalytic theory of neurosis.
New York: W . W . Norton & Co., 1945.
Frank, L. K. Toward a projective psychology. J. pro.j.
Tech. . I960, 24, 2 4 6 , 253.
Freud, Anna. The ego and the mechanism of defense.
Hogarth Press, 1937.
Freud, Anna. Certain types and stages of social malad
justment. In K. R. E issler (Ed.) Searchlights on
delinquency. New York: In t. Univer. Press, 1949.
-151-
Friedlander, Kate. Latent Delinquency and ego develop
ment. In K. R. E issler (Ed.), Searchlights on
delinquency. New York: In t. Univ. Press, 1949.
Glover, E. H. The roots of crime. New York: In t.
Univ. Press, i 9 6 0 .
Glueck, S. & Glueck, Eleanor. Unraveling .juvenile
delinquency. Cambridge, Mass.: Harvard Univer. Press,
1950.
Glueck, S. & Glueck, Eleanor. Predicting delinquency and
crime. Cambridge, Mass.: Harvard Univer. Press, 1959.
Gough, H. G. A sociological theory of psychopathy.
Amer. J. Sociol. . 1947, 53., 359, 366.
Gough, H. G. Systematic validation of a te s t for
delinquency. Amer. Psychologist. 1954, £, 381
Guilford, J. P. Fundamental s t a t i s t i c s in psychology
and education. New York: McGraw-Hill, 1950.
Hartshorne, H. & May, M . A. Studies in the nature of
character and d e c e it. New York: McMillan, 1930.
Hathaway, S. & Monachesi, E. D. The M M PI in the study of
juvenile delinquency. Amer. Sociological Rev.. 1952,
12, 704, 7 1 0 .
Havighurst, R. J. & Taba H. Adolescent character and
p erso n ality . New York: John Wiley & Sons, 1949.
Hollingshead, A. B. & Redlich, F. C. Social class and
mental illn e s s : a community study. New York: Wiley,
Hooten, E. A. Crime and the man. Cambridge, Mass.:
Harvard Univer. Press, 1939.
Josselyn, Irene. The adolescent and his world. New York
Fam. Serv. Ass. Amer., 1952.
Jourard, S. M . Id e n tific a tio n , parent-cathexis and self
esteem. J. consult. Fsvchol.. 1957, 21, 375, 380.
Kanner, L. Types of delinquents. Psychodynamic Monogr.
S e r ., 1959, 138-139.
153
Karpman, B. Psychopathic types, the symptomatic and the
idiopathic. J. crim. Psychopath.. 1949, 3, 141,152.
Kaufman, I. & Heims, Lora. The body image of the
juvenile delinquent. Amer. J. Othopsychiat., 1957,
125, 132.
Lazowick, L. M . On the nature of id e n tific a tio n . J.
abnorm. soc. Psychol.. 1955, 51, 175, 1B3.
Lombroso, C. Crime, i t s causes and remedies. In M . B.
Clinard (Ed.), Sociology of deviant behavior. New
York: Rinehart, 195$, 142-143•
McNemar, Q. Psychological s t a t i s t i c s . (2nd ed.) New
York: Chapman & Hall, 1955.
Mead, G. H. A behavioristic account of the sig n ifican t
symbol. J. P h il.. 1922, 19, 157, 162.
Michaels, J. J. Disorders of ch aracter. Springfield,
111.: Bannerstone House, 1955.
Mussen, P. & D istler, L. Masculinity, id e n tific a tio n
and fath er son re la tio n sh ip s. J. abnorm. soc. Psychol..
1959, 59, 350, 356.
Payne, D. E. & Mussen, P. Parent child re la tio n s and
father id e n tific a tio n among adolescent boys. J.
abnorm. soc. Psychol.. 195o, 52, 35^-362.
Rabinovitch, R. D. Psychogenetic fa cto rs. Psychodynamic
Monogr. Ser.. 1959. Pp. 11-14.
Reed, C. F. & Curdra, C. A. The role taking hypothesis
in delinquency. J. P h il., 1957, 21, 254-261.
Sanford, N. The dynamics of id e n tific a tio n . Psychol.
Rev.. 1955, 62, 106, U S.
Schachtel, E. G. Notes on Rorschach te s ts of five hundred
juvenile delinquents and a control group of five
hundred non-delinquent adolescents. J. p ro j. Tech..
1951, 15, 144, 171.
Schuessler, K. F. & Cressey, D. R. Personality charac
t e r i s t i c s of criminals. Amer. J. Psychol.. 1950, 55,
476-4^4.
154
Shaw, C. R. & McKay, H. D. Delinquent a re a s . Chicago:
Univer. of Chicago Press, 1929.
Sheldon, W . H. V arieties of delinquent youth. New York:
Harper, 1949.
Sopchak, A. L. Parental id e n tific a tio n and tendency
towards disorders as measured by the MMPI. J. abnorm.
soc. Psychol.. 1952. 47. 159-165.
Spitz, R. A. Possible in fa n tile precursors of psycho
pathy. Psychodynamic Monogr. Ser.. 1959, pp. 20-27.
Stokes, S. An inauiry into the concept of id e n tific a tio n .
J. genet. Psychol.. 1950, 76, 163-1&9.
Sutherland, E. H. White c o lla r crim in ality . Amer.
Sociological Rev.. 1940, 5., 1-12.
Sutherland, E. H. P rinciples of criminology. P h il.,
Penn.: Lippincott, 1947.
Thompson, R. E. A v alid atio n of the Glueck social
prediction scale for proneness to delinquency. J . crim.
law & Criminol.. 1952, 43, 451-470.
APPENDIX
EXPERIMENTAL INSTRUMENT, ABRIDGED FORM
THE EXPERIMENTAL INSTRUM ENT
(Abridged Form)
Imagine th at you are a judge in a court of law. A twenty
five year old white male stands before you, convicted by a
jury. Sentence the man for each crime as you would i f you
were the judge.
Read the crime, then give your sentence by using the
numbers shown below:
1. I f you c irc le number one (1), i t means a sentence
of Probation.
2. I f you circ le number two (2), i t means a sentence
of Six months to One year.
3. I f you c irc le number three (3), i t means a sentence
of Five years.
4. I f you c irc le number four (4), i t means a sentence
F ifteen years.
5. I f you c irc le number five (5), i t means a sentence
of Life in Prison.
6. I f you c irc le number six (6), i t means a sentence
of Capital Punishment or Death.
-1 5 6 -
157
White male, age 2 5 ...
1. Entered several markets over a period of three
months and cashed bad checks in the amount of
$500,00. Second conviction of the charge.
2. Stopped a man in an a lle y , strik in g him with a
blackjack and taking his w allet. Victim had mile
concussion. Second conviction th is offense.
3. Working as a bookkeeper over a period of five year^
took the sum of $3,000 to his own pocket. Second
conviction th is offense.
4. Was in gang fig h t, stabbing another man. The
victim died of the wounds.
5. Entered house threatening the housewife with a
knife, assau ltin g her sexually.
6. Collected $25,000 down payments on home lo t s ,
when there were no such lo ts .
7. Driving under the influence o f alcohol, struck
a pedestrian and fled the scene. Victim died of
his in ju rie s . Second offense fo r drunk driving.
15S
For th is part of the t e s t , read the items below and
say how you would fe e l about these s itu a tio n s .
To shorten the t e s t and make things a l i t t l e e a s ie r,
you are to give your opinions about the items by
number. For each item you are to c irc le a number of
from one to fiv e . Each number means something d if
fe re n t. The meaning for each number follows below.
I f you mark 1 (one) i t means you believe:
The behavior is not very serious. This kind of
thing happens a l l the time. You wouldn’t fe e l too
strongly about i t , and i f you were frie n d ly with the
person i t wouldn’t change your friendship in any way.
I f you mark £ (two) i t means you believe:
You disapprove of th is and might t e l l the person you
disapprove, but i f he was a frien d you would not hold
i t against him and would continue to be frie n d ly .
I f you mark 3 (three) i t means you believe:
This behavior is d e fin ite ly wrong. I f you were a
frien d of the person who did th i s , you would fe e l
quite d iffe re n tly about him. You would fe el
strongly enough about i t to stop your friendship
with the person.
I f you mark ^ (four) i t means you believe:
This is re a lly seriously wrong. You would not
fe e l frie n d ly toward the person who did th is
and would not tr u s t him. You d e fin ite ly would
not spend time in his company.
I f you mark £ (five) i t means you believe:
The behavior is so serious th a t the police would
have to be called in . You would fe el th a t the per
son should be sent to some kind of school or
in s titu tio n .
159
1. Jim and Bob age 17 have a fig h t which is stopped
by a teacher. Bob l a t e r damages jim ’s car by
b a tte rin g the engine, ra d ia to r and fuel pump with
a heavy wrench.
2. B ill age 17, driving 55 miles per hour in a 35
mile zone, is stopped by an o ffic e r. He is given
a tic k e t for speeding. As soon as the o ffic e r
leaves, B ill again accelerates the car to ”55”
in the same 35 mile zone.
3. In a shop class, the in s tru c to r asks Jack age-17
to sweep up some wood shavings. Jack re p lie s ,
"Pick on somebody e lse . I ’m not the ja n ito r ."
4. Jim age 16 is a good mechanic. He rep airs a
frie n d ’s car, but s u b stitu te s some of h is own
second hand auto parts and keeps the newer parts
from his frie n d ’s la te model car.
5. In a f i s t fig h t with B ill, 17, Jack also 17 sud
denly picks up a rock and s trik e s B i l l ’s head
with i t . B ill drops unconscious to the ground.
160
In th is part of the t e s t , imagine again th at a twenty-
five year old white male stands before a judge. This
time imagine that the judge is Y O U R , M OTHER. Give the
sentence you fe el Y O UR M OTHER would give fo r the crime
described. (If your mother did not raise you, answer
as you fe e l the woman or women you knew best when you
were a child would answer.)
As before, give the sentence by using the numbers
below:
1. Probation.
2. Six months to one year.
3. Five years.
A. Fifteen years.
5. Life in prison.
6. ’Capital Punishment1 (Death)
USE T H E A N SW E R SHEET
161
White male age 2 5 ...
1. D eliberately weakened dosage of medicine needed
to pull wife through a serious illn e s s . Discovered
by a friend, thus preventing w if e ^ death.
2. Attacked a young woman, standing on s tre e t corner
at night, forcing her to an act of sexual in te r
course.
3. Kidnapped a small child and l e f t i t in an empty
house. Child died of exposure and lack of food.
4. Got drunk at a party and, while driving home,
struck a person standing in a safety zone. Did
not stop, although he knew he had h it the person.
The victim died of his in ju rie s . Second convic
tio n for drunk driving.
5. Sold 27,500 worth of stock in a frich gold mine1.
Biyers found th a t there was no such mine.
ITT
162
For th is part of the t e s t , read the items below and tr y to
imagine how M OST PEOPLE Y O U K N O W would fe e l about these
s itu a tio n s .
Mark your answers the same way you did before, only th is
time you are marking as M OST O F THE PEOPLE Y O U K N O W
would answer.
I f you mark 1 (one(), i t means most others b elieve:
The B ’ ehavior i s not very serious. This kind of
thing happens a l l the time. They wouldn’t fe e l
too strongly about i t , and i f they were
frien d ly with the person i t wouldn’t change th e ir
friendship in any way.
I f you mark 2 (two), i t means most others believe:
They disapprove of th is and might t e l l the person
they disapprove, but i f he was a frie n d , they
would not hold i t against him and would continue
to be frie n d ly .
I f you mark 3 (th re e), i t means they believe:
This~behavior is d e fin ite ly wrong. I f they were a
friend of the person who did th is , they would fe e l
quite d iffe re n tly about hina. They would fe e l
strongly enough about i t to stop th e ir friendship
with the person.
I f you mark (four), i t means most others believe:
This i s re a lly seriously wrong. They would not
fe e l frien d ly toward the person who did th is
and would not tr u s t him. They d e fin ite ly would
not spend time in his company.
I f you mark 5 (fiv e ), i t means most others believe:
The Fehavior is so serious th a t the police would
have to be called in. The person should be sent
to some kind of school or in s titu tio n .
USE THE A N SW ER SHEET
163
9. B ill age 19 forces Ju lie age 17 to a sexual act in
sp ite of her struggles and p ro te sts.
10. Angry because they are asked to leave a party they
have crashed, Jim and B ill age 17 and IB draw pocket
knives and slash curtains in the liv in g room of
B etty’s house.
11. Jim, working at a gas s ta tio n week-ends, earns extra
money by sellin g new ra d ia to r hoses. He makes many
sales by showing the customer cracks in the hose
and pointing out how worn i t i s , even though he
knows the hose is good fo r thousands of m iles.
12. M r . Smith, the histo ry teacher, le c tu re s a group of
boys fo r t h e i r behavior in class. Angry at him they
go to h is car, and he l a t e r finds the surface of h is
car dented and chipped and his tire s, slashed.
13. B ill age IB takes Sally 16 out to a movie. Knowing
she has never had anything to drink, he talk s her
into drinking considerable liq u o r. When she is
quite drunk, he forces her to have sexual intercourse
with him
164
In th is part of the t e s t you are again to imagine th a t a
twenty-five year old white male stands before a judge con
victed by a jury. This time you are to imagine th a t the
judge is YOUR FATHER. Give the sentence you fe e l YOUR
FATHER would give for the crime described. (If your
fa th e r did not raise you, answer as you fe el the man or
men you knew best when you were a child would answer.)
As before, give the sentence by using the number as shown
below.
1. Probation.
2. Six months to one year.
3. Five years.
4. F ifteen years.
5. Life in prison.
6. ,T Capital'7Punishment” (Death)
USE TH E A N SW ER SHEET
165
6. Forced female, age 30, to an act of sexual interco urse.
7. Exhibited his private p arts to a minor g i r l . Third
offense.
&. Subject, a druggist, t r i e d to k i l l his wife by-
putting small q u an titie s of poison in medicine
she had been taking re g u la rly . This was discovered
and the wife recovered.
9. By threatening a ten year old boy with a beating,
forced him to various homosexual a c ts .
10. Collected $30,000 from customers a f te r convincing them
he had a machine which would make one liv e longer.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Self-Worth, Future Goals, Mood And Time Perception

PDF

Measuring Thought Process As An Ego Function In Schizophrenic, Mentally Retarded And Normal Adolescents By Means Of The Rorschach

PDF

Prediction Of Therapeutic And Intellectual Potential In Mentally Retardedchildren

PDF

Delinquency As A Function Of Intrafamily Relationships

PDF

Sex-Role Preferences Of Early Adolescents In Relation To Adjustment

PDF

Psychological Test Changes In Schizophrenic Patients Under Brief Stimuluselectroconvulsive Therapy

PDF

Behavioral Seriousness And Impulse-Control Balance In Delinquency

PDF

Psychological Variables In The Mother Related To Infant Feeding Patterns

PDF

Effects Of Different Treatment Procedures On Reading Ability And Anxiety Level In Children With Learning Difficulties

PDF

Ego Diffusion In Women With Behavioral Disorders And The Integrating Effects Of Psychodrama In Identity Consolidation

PDF

Perceptual Defense And Somatization: A Comparison Of The Perceptual Thresholds Of Obese And Peptic Ulcer Patients

PDF

Methodological Problems In The Identification Of Suicidal Behavior By Means Of Two Personality Inventories

PDF

An Investigation Into The Communication Style Of Suicidal Individuals

PDF

Dental Student Selection Through Handwriting Analysis

PDF

A Comparison Of Two Methods Of Adapting Self-Instructional Materials To Individual Differences Among Learners

PDF

The Use Of Multiple Objective Measures In Survey Research As A Means For Predicting Respondent Future Behavior

PDF

Improvement Rate In Tuberculosis As A Function Of Accessibility To Repressed Emotional Disturbance

PDF

An Experimental Study Of Aggression Habit Patterns And Sex Role

PDF

The Effect Of Irrelevant Environmental Stimulation On Vigilance Performance

PDF

A Study Of The Attitudes Of Mothers Of Mentally Retarded Children As Influenced By Socioeconomic Status

Asset Metadata

Creator Richardson, Orville Roderick (author)

Core Title First Steps In An Attempt To Construct An Objective Test Of Character Structure

Degree Doctor of Philosophy

Degree Program Psychology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,Psychology, clinical

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Seward, Georgene H. (committee chair), Ruch, Floyd L. (committee member), Weckler, Joseph E. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-266598

Unique identifier UC11358665

Identifier 6203743.pdf (filename),usctheses-c18-266598 (legacy record id)

Legacy Identifier 6203743.pdf

Dmrecord 266598

Document Type Dissertation

Rights Richardson, Orville Roderick

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA