39-1.5- iii; iaiégéafia 2.. . .

 

.Ffla ”

 

 

 

 

 

 

 

 

ABSTRACT
IWJAGFI’IFNT OF HOSTILITY AND VERBAL FUNCTIONING
IN COLLEGE MALES

by Richard A. Westmaas

The purpose of this study was to investigate relationships
between individual differences in the management of hostility and
verbal functioning in a situation designed to arouse hostility.
Paychoanalytic formulations of the processes involved in impulse
inhibition suggest that individuals characterized by a relatively
inhibitory method of managing hostility experience difficulty
in producing words associated with the expression of hostility.
That such individuals may also be at a disadvantage in producing
words not inherently associated with aggression was inferred from
the possible generalization of inhibitory processes. The hypotheses
formulated to test the proposed relationships predicted a positive
relationship between the expression of hostility and the produc-
tion of (l) hostile words and (2) neutral words in a frustrating
situation.

i In testing these hypotheses, 80 male college students were
randomly assigned to either a high arousal (HA) or low arousal
(LA) conditian. Prior to verbal production tasks, the to HA
subjects were frustrated by contriving their failure on a test
said to predict creative potential, and by the eXperimenter's
insulting remarks. The hO LA subjects were allowed to pass this
test. Oral production of words in restricted categories provided

the dependent measures. The hostile word category included

 

 

Richard A. Westmaas

words used to describe someone whom one hates or dislikes. The
neutral word category consisted of words ending in a—t-i-o-n.
Measures of hostility were obtained near the end of the experiment,
immediately following a procedure for inducing aggression in all
subjects. These measures consisted of ratings on a post-experi-
mental questionaire which provided for the expression of hostility
toward the experimenter, the experiment, or the testing situation;
and extrapunitive responses on a modified version of the Rosen-
zweig Picture Frustration Test.

That HA subjects were considerably more hostile than LA
subjects on questionaire ratings was taken as evidence of the
effectiveness of the arousal procedure and as support for the
validity of the questionaire. Failure of the PF Test to reliably
discriminate betwem the groups was interpreted as further
questioning the validity of a technique that has shown seriom
weaknesses in other studies.

The hypotheses to be tested concerned only the results in
the HA group. Results regarding the hypotheses were:
mgthesie E: Total hostility scores derived from questions-ire
ratings failed to relate positively to production of hostile words.
A separate analysis of the questionaire items revealed that ratings
of unfriendliness to the experimmter were highly associated with
the production of hostile words (p < .01), whereas ratings providing
more indirect expressions of hostility tended to relate negatively
to hostile word production. It was suggested that it is not the
indiscriminate expression of hostility, but the ability to aggress

directly against the instigator which is predictive of flumcy

 

 

.q

4 ~"1-' I‘m}! ‘1

.‘7..

 

 

 

I
\

.1» “ -

\

 

Richard A. Westmaas

with hostile words when frustrated. These results were seen as
consistent with psychoanalytic theory, and with a variation of
Miller 'a approach-avoidance paradigm which assumes individual
differences in avoidance tendmcies. Attention was called to the
possible significance of proficiency with hostile verbalization

as affecting patterns of managing hostility. The non-significant

 

relation of PF measures to hostile word production was takm as
a reflection of the relatively low validity of this instrument.

It was pointed out that the sample chosen represents a group in

 

which hostility is largely confined to verbal aggression, and that
generalizations to other groups should be avoided. Suggestions
for further research were made.
mthesis 3: None of the measures of hostility were significantly
related to production of neutral words, suggesting that individu-
als who inhibit hostility are not characterized by lack of fluency
with words unrelated to aggression.

Other findings were that, among IA subjects, non-significant
trends of the same relationships exhibited in the HA group were
obtained. Comparison of Hi and IA groups showed that frustrated

subjects produced more words of both categories than non-frustrated

 

subjects. That the motivating effect of frustration was more
transient for neutral words was seen as consistent with the
assumption that greater inhibition acccmpanies production of

hostile words in the presence of the instigator.

 

 

 

 

l-IANAGE-ENT OF HwflHTY
AND VERBAL FUNCTIONING
IN COLLEE MALES

By

 

‘ f‘
4\
Richard A} Westmaas

ATHEIS

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

DOCTOR OF PHILOSOPHY
Department of Psychology

1962

 

 

 

— -‘*‘—-.

 

 

 

 

 

to my wife, a kind but critical listener,
who graciously supported and assisted

in so many ways;

to Joel and Heidi, who were patient when
asked to play elsewhere so that their '
Daddy could work on his thesis;

and to Amy, whose smiles were encouraging,
even though she is too young to understand.

ii.

 

 

 

 

 

 

ACKN OWIEDG Elf-{EN TS

This research was done under the general direction of
Dr. Charles F. Winder, Dr. Joseph Rayner, Dr. Bill Kell, and
Dr. M. Ray Denny, to whom the author is grateful for advice and
encouragement. Thanks are also due to Dr. Terrence M..Lllen
who made significant contributions to the earlier stages of
this research. I am especially grateful for the lasting con-
tributions of Dr. Gerald F. King, whose death during the initial
stages of this study was not only a.deeply felt personal loss,
but still more a loss to the field of psychology.

The author deeply appreciates the contributions of the
students who served as subjects, and who generously forgave
him for the injustice he inflicted on than. >

This investigation was carried out during the tenure of
a Predoctoral Research Fellowship from the National Institute
of Mental Health, united States Public Health Service. The
receipt of this financial support greatly enhanced the execution
of this study by freeing the author to concentrate his efforts

solely on this investigation.

iii.

 

 

 

 

 

 

 

 

 

 

TABIE OF CONTENTS

WTER

I.

II.

III.

IV.

INTRODUCTION...................
Conceptsanddefinitions....... ......
The modification of hostile impulses . . . . . . .
The management of hostility . . . . . . . . . . .

Effects of inhibition of hostility on verbal
functions..................

MANAGEMENT OF HOSTILITY: RESEARCH FINDINGS . . .
Effects of punishment on expression“ of hostility
Anxiety and the management of hostility . . . . .
Individual differences in management of hostility
“Test hostility"I and response to frustration . . .
MANAGEMENT OF HOSTILITY AND VERBAL FUNCTIONS:
RESELRCHFINDINGS ..............

Management of hostility and performance of
mullmtmmkSeeeeeeeeeeeeee

Depression and verbal productivity . . .. ... .... ..

Managmt of hostility and verbal behavior of
psyChOtherapiStaoeoeooaoeeeeeee

Anxiety and verbal functioning . . . . . . . . . .
Newcomb's hypothesis of autistic hostility . . . .
Measurement of verbal functions . . . . . . . . .
Management of hostility and verbal productivity:
apilotstudy ................
THE PROBLEM, HYPC'IHESEB, AND METHOD . . . . . . .

Theproblm.o................o

iv.

PAGE

VIC'HH

10

13
18
22
2h .

29

29
30

31
32
36
37

39

1:5

 

 

 

CHAPTER

HyPOthesea e e e e e e e e e e e e e e e e e e e

Other relationships to be explored in the study

1‘48 thOd O O O O O O O O O C C O O O I O O O O O 0

Summary of the experimental design . . .
SMflafis .. ... ... ... ... ..
Procedures and measuring instruments . .

Outline of experimental procedures . .

Independent variables . . . . . . . .

Situation designed to arouse hostility’ . .

Measurement of the management of hostility.

The NOdified PF TGSt e e e e e e
Post-experimental questionaire .
Depmd w t vmablu O O O 0 O O O O O

HOStilB word CCtegory e e e e e e e

"Neutral" word category .

V. REULTS . O O O O O O O O O O O O O O O O O 0

Effectiveness of the procedure for instigating

hOStility e e e e e e e e e e e e e e e e
Findings regarding the hypotheses . . . . .

Findings based on the post-experimental
qu93tionairo e e e e e e e e e e e e e

Findings based on the Rosenzweig Picture
FrUStration TBSt e e e e e e e e e e e

Intercorrelations among verbal productivity and

other variables . . . . . . . . . . . . . . .

Other findings: General effects of frustration

onverbalproductivity ...........

Ye

PAGE

h?
1:8
ha
19
So
So
52
53
SS
56
58
S9
60

62

67

67

76

79

 

 

 

 

 

I
I

iaiia J‘III'K!’
. ' A- ' ‘
x g I e a

 

 

     

CHAPTER PAGE

VI.DISCUSSION...................o. 88
The arousal of hostility ............ 88
Thehypotheses................. 89

Mothesisl................. 90

HypOthOSiB 2 e e e e e e e e e e e e e e e e e 91
Otherfindings................. 100

VIIe SUMMARY e e e e e e e e e e e e e e e e e e e e e e 103
REFERENCES e e e e e e e e e e e e e e e e e e e e e . . 109

APPENDIX A: ROSI'NZWEIG PICTURE FRUSTRATION TEST
SCORE SHEET e e e e e e e e e e e e e e e e e e e 117

APPENDIX B: POST-EXPERIMENTAL QUESTIONAIRE . . . e 118

 

 

m we- - ~7?flw3‘d§r¥ '

mt»

 

 

     

 

7.7-... -1..-.— -.-.——~ 77w

 

LIST OF TABLES

Table Page

1. Rank order correlations of TIES with measures of
verbal productivity e e e e e e e e e e e e e e e e e hl

2. Comparison of Low Arousal and High Arousal groups on
age, verbal intelligence, and parents education . . . SO

3. Comparison of Low Arousal and High Arousal groups:
Mean ratings on the post-experimental questionaire. . 63

h. Mean pre-frustratim and post-frustration H‘ extra-
punitive scores of HA and LA groups . . . . . . . . . 65

S. Direction of PF-E change in HA and u groups . . . . . 65

6. Mean number of hostile and neutral words produced by
gs above and below the median in total hostility .. ,. 69

7. halysis of variance of verbal production: Levels
of hostility determined by TH scores on the
pOBtCCxperiment‘l questianairl e e e e e e e e e e e 69

8. Mean production of neutral and hostile words by gs
scoring above and below the median on each of
- the ratings on the evaluation questionaire . . . . . 72

9. Analysis of variance of verbal production:- Levels
of hostility determined by ratings of friendly-
unfriendly reactions to the experimsnter . . . . . . 7h

10. Production of hostile and neutral words for Se
varying in the direction of extrapunitive 'c'hange
f0110"1n8 frustration e e e e e e e e e e e e e e e e 78

ll. Intercomlations among verbal production and other
variables-Highirousalgroup........... 81

12. Intercorrelations ammg verbal production and other
variables-MAI'Ouaalgmupeeeeeeeeeeee 82

13. Mean number of hostile and neutral words produced
by'§P in the HA ‘nd Lt grOUPB e e e e e e e e e e e e 85

vii.

 

Figure Page

1.

3.

Representation of the design employed in the
investigation.................... ’49

Comparison of LA and HA groups on rate of producing
hostilcwords.................... 86

Comparison of LA and HA groups on rate of producing
neutralwords...............'.....

V111 .

 

 

 

 

 

CHAPTER I

INTRODUCTION

The tepics of hostility and of verbal functioning have
both become the objects of increasing interest among investiga-
tors in the last few years. Interest in the area of hostility
is indicated by the appearance of several instruments for assessing
hostility (Siegel, 1956; Buss and Durkee, 1957; lake and Halters,
1959), and by an increasing number of studies with various in-
terests and orientations. Interest in verbal functioning has
received impetus from experimental sources in the operant con-
ditioning studies, now being designed to simulate the clinical
interview situation (Buss and Durkee, 1958; Sarason, 1961;
Babladelis, 1961). From the clinical side, increasing interest
centers around both content and style of verbal communication
in the therapy hour as well as in personality correlates of
verbal behavior in the interview.

The present study is concerned with relationships between
the arousal of hostility and a specific aspect of verbal func-
' tioning, the continuous production of words in two categories;
one related to the expression of hostility, the other neutral
with regard to hostility. Major emphasis will be placed on individ-
ual differences in the management of hostility as related to verbal

functioning in a situation in which hostility has been aroused.

Concepts and Definitions

 

The terms hostility and aggression are not generally given

1

 

 

.—— - -
.i . ...-u“.

 

 

 

 

precise definition, and are frequently used synonymously.

Both terms serve at times to describe either motivational states
or overt behavior. In what is probably the most systematic

and influential treatment of the subject, Dollard, et a1. (1939)
define aggression as l". . . an act whose goal-response is in-

jury to an organism (or organism surrogate)". While this

 

definition appears to describe overt behavior, it is clear that.

in other portions of their work aggression is seen as covert

or attitudinal. For example, "Aggression is not always mani-
fested in overt movements but may exist as the content of a
fantasy or dream or even a well thought out plan of revenge"

(p. 10)". The expression of aggression may be inhibited through
fear of punishment, and changes in form or object of aggression may

occur (p. Mr ff). ' Implicit here is the asstmption that aggression

 

is a relatively enduring state of the organism which, once
established may be the antecedait of a variety of consequents.
In gmeral usage, the term hostility is frequently applied to
the relatively enduring motivational-emotional complex inferred
from overt behavior or antecedent events. The terms aggression
or hostile behavior refer to overt behavior which is injurious
to the interests or well being of others (English and Exglish,
1958).

The term hostility is preferred in the present research,
since attention will be focused on responses to situations which
are assured to arouse hostility, but in which overt responses
vary in the extent to which they represent hostile behavior.
Variations in response to frustration are assumed to reflect

 

...—.... “fizz-e3!!!- Ares. »-

 

 

 

 

3

(among other things) characteristic differences between individ-
uals in their control or expression of hostile motivation.

The term management of hostility is employed here to refer
to the use of devices which control or regulate the expression
of hostile impulses. Such devices are exemplified in the defense
mechanisms of daiial, repression, projection, and turning against
the self. No attempt is made to claim.that all differences
in the overt display of hostile behavior in the face of frus-
tration are a function of variations in the management.of hostility.
Individuals may differ in their vulnerability to frustration
as a function of constitutional differences and their previous
history. Thus Dollard et al. (1939) speak of the number of
frustrated response sequences, and Roserxzweig (1915) preposes
the concept of frustration tolerance as a significant variable.
The notion of ego strength, and the related concept of capacity
to delay gratification also imply relatively enduring character-
istics of the organism which relate to the outcome of frustration.
Rosenzweig also suggests that individuals may have specific
areas of vulnerability or low frustration tolerance corresponding
to psychoanalytic "complexes."

From the foregoing, it is apparent that a variety of var-
iables may conceivably operate to determine individual differences
in hostile behavior. It is here preposed that a central and
significant variable in determining reactions to frustration
lies in the area of what has been called the management of hostil-
ity. The concept of management of hostility is a broad one

which includes all the regulatory mechanisms and habitual strategems

 

 

 

relevant to the expression or inhibition of hostile impulses.

Also included are the various socially acquired skills of modu-
lating the expression of hostility so as to avoid the censure

of the superego and the disapproval and retaliation of others.

The Modification g Hostile Iyulses

 

The development of control mechanisms to regulate hostile
tendencies is a prerequisite for social living. Unrestrained
expression of hostility is discouraged at an early age within
the family, and punished by law in the broader society. In the
frustration-aggression theorists' treatment of the expression of
hostility, fear of punishment is made one of the primary variables
affecting the appearance of aggression. With the internalization
of societal standards, and the institution within the personality
of a censoring agent, the individual also comes to experiaice
guilt and discomfort from within when his aggression exceeds
the bounds of personal standards. Because of the opprobrium
with which hostile behavior is regarded, particularly in the
middle and upper social classes, various adjustments are made
by the individual in his expression of hostility.

One result of society's attempts to discourage 0pm hostility
is to make the expression of hostility less direct and more
diversified. 0pm rebellion may give way to a more subtle passive
resistance; fantasied revenge may replace direct retaliation, and
verbal dmunciation may substitute for physical violence. Thus,
Siegel (1956, 1957) proposes that people may differ not so much
in the amount of hostility they possess, but in the way in which

they express it.

 

     

l

.2:

‘I
Z .1
I .-
.
V
.-

 

 

5

A widely recognized contribution to the literature on the
management of hostility is that of Rosenzweig (l9hh). He calls
attention to differences in the direction of hostile expressions.
The arousal of hostility may eventuate in the expression of
aggression outwardly toward the frustrating agent (extrapunitive),
inwardly toward the self (intropunitive ), or the frustration
and hostility may be denied or minimized (impunitive response).
The work of Funkenstein and King (19510 and King and Henry (1955)
suggests that the concept of the direction of hostility has mean-
ing on the physiological as well as the psychological level.

They found that subjects who respond to a stress situation with
self-blame or anxiety tend to have an epinephrin-like response
on the physiological level, whereas subjects responding with
outwardly directed anger have a norepinephrin-like reaction.

In addition to the variety of forms in which overt'aggres-
sion appears, the presence of social. censure or superego pro-
hibitions may effect a displacement of aggression, i.e., a change
in the object of hostility. The hostility aroused by a frustrating
agent in one situation is not necessarily specific but may be
expressed toward individuals in other situations who had nothing
to do with the frustration.

The preceding paragraphs indicate that hostile impulses
may undergo a variety of modifications both in form and object.
These modifications are assumed to result from factors inhibiting

the direct expression of hostility.

The Management 93 Hostility

 

Whether the individual responds to frustration in an Openly

 

 

 

 

hostile or an inhibited fashion depends on a number of factors

both internal and external to the individual. The impulse to
aggressively attack the frustrator is countered by fears of his
retaliation or withdrawal of love. Situational factors, such as
the status and power of the frustrator are potent factors in
determining the level of hostile response (Berkowits, 1959).
Miller's approach-avoidance conflict theory (l9hh) offers
a framework for discussing the expression and inhibition of
hostility. The principles of approach-avoidance theory may be
briefly stated as follows: Miller posits that the tendency to
approach or to avoid a goal is stronger the nearer the subject
is to the goal. The strength of avoidance tendencies increases
more rapidly with nearness to the goal than that of approach.
That is, the slope of the avoidance gradient is steeper than
that of approach, so that the two may cross. Further, the
strength of tendencies to approach or avoid varies with the
strength of the drive on which the tendencies are based. In-
creased drive raises the height of the entire gradient. In the
context of hostility, approach tendencies are based on the
strength of hostile drives, and avoidance tendencies are based
on fears of retaliation. In the situation where hostile drive
and fears of retaliation are of approximately equal strength so
that the two gradients cross, a certain amount of hostility may
be expressed at some distance from the goal. However, as the
goal is approached, the expression of hostility becomes more and more
conflictual until the avoidance tendencies become stronger than
those of approach, and the subject withdraws. If the level of

hostile drive is increased while retaliation remains constant,

 

the gradient of approach no longer intersects the avoidance

gradient, and the hostile behavior is culminated.

In contrast to the psychoanalytic model, Miller's theory,
and the experimental evidence he presents to support it, refers
almost entirely to uternal variables in dealing with the
expression of hostility. Little attention is given to individual
differences in response to the same hostility provoking situation.
The present research stresses the fact that individuals vary
widely in their responses to frustrating situations. While
adrdtting the effects of the external situation on the expression
and inhibitiontof hostility, we posit that the individual's
pattern of responses to frustration are to some extent character-
istic of that individual from one situation to another.

In order to account for individual differences in the
expression of hostility within Miller's framework, one might
posit differences behvem individuals in their level of hostile
drive, or in fears cf retaliation, or both. Thus the relative
absence of a hostile response to frustration could be detemined
by either a low level of hostile drive or by a high level of
fear of retaliation. The concept of management of hostility
assumes the presence of hostility, and places the emphasis on
individual differences in fears of retaliation, or aggression
amdety, as determining individual differences in the amount of
hostility that is expressed in a frustrating situation.

From a dynamic viewpoint, methods of. managing hostility
may be divided into two classes: those which pemit impulse

expression, and those which block the expression of the impulse.

 

 

 

 

 

 

 

 

The paradigm of impulse inhibition would be the complete repression

of hostility and denial of frustration. Other methods of managing
hostility, whether the manifestation is verbal or physical,
subtle or obvious, direct or displaced, allow for some degree
of impulse expression. A consideration of certain methods of
dealing with hostility reveals some difficulties in the assign-
ment of behavior to one of the two classes if these classes are
thought of as discrete and mutually exclusive. Theorists disagree
as to the drive reducing Qualities of fantasied aggression
(thbeck, 1955; Berkowitz, 1960), and Dollard et a1. (1939)
assert that self directed aggression is less drive reducing than
outwardly directed because injury to the self constitutes additional
frustration.

The position taken here is that expression and inhibition
of hostility are relative terms which describe the extremes of
a continuum. Thus it is possible to speak of a given manifesta-
tion of aggression as containing greater or lesser amounts of
inhibition. The outcome of exposure to frustration can be
placed on the inhibition-expression continuum, with massive
repression and denial of frustration at one end, and physical
assault on the frustrating agent at the other. Normally, a
response to frustration represents a compromise between inhibi-
tion and complete expression. This viewpoint is in agreement with
dynamic theory, which sees behavior as the outcome of a synthesis
between opposing forces.

The behavior of the individual whose typical response to

frustration approaches either the inhibition or expression

 

 

  

I
I'_ u .
u \ !- " . -

f

 

 

 

 

9

extreme is considered pathological. The person who fails to
exercise some inhibition of aggression comes into conflict with
the mores and laws of society and soon attracts the attention of
the authorities. Behavior approaching the extreme of inhibition
of hostility may be less dramatic, but the consequences to the
individual may‘be just as far reaching. Such a person is handi-
capped in social relations by his inability to meet his frustra-
tions constructively, to compete effectively, or even to recognize
the sources of his frustration.

The basic mechanism involved in the extreme inhibition of
hostility is known in psychoanalytic theory as repression.
Repression occurs when pain becomes associated with the expression
' of an impulse, and the motive of avoiding pain becomes stronger
than the pleasure of gratification. The essence of repression,

- according to Freud, "lies simply in the function of rejecting
and keeping something out of consciousness“ (1959, p. 865. Two
phases of repression are distinguished. The inhibition of the
ideational presentation of the instinct itself is termed primal
repression. The second phase of repression, termed repression
21322: has to do with-derivatives of the repressed instinct-
presentation, such as trains of thought originating elsewhere,
which have come into associative connection with the instinct-
presentation. The radiation of repression is one of the patho-
genic features of this defense. ‘Whole tracts of mental func-
tioning may be withdrawn from effective interaction with the
environment, producing a sterile or “pollyannish” personality.
.Further, the inhibition of aggression is itself frustrating.

Thus there deveIOps a vicious circle of increasing hostility

 

 

10

which demands still more energy to prevent its expression and

its intrusion upon awareness.

Effects gf inhibition of Hostility on verbal Functions

 

 

The existence of a relationship between inhibiting processes
and verbal functions has been postulated by many observers,
notably those with a psychoanalytic orientation. In psycho-
analytic therapy, a blocking of the flow of free associations
is assumed to result from the activation of repressive forces
signaling the approach to highly conflictual material. The
effects of repression on the flow of speech in normal conversa-
tion have been described by Freud in his Payphopgthology of
Everyday EEES (1935). Disturbances in verbal functioning are
held to be particularly related to conflicts over hostility.

The stutterer is seen as a hostile person whose speech has an

anal sadistic significance. The stuttering reflects the conflict be-
tween the expression and inhibition of obscene and hostile

impulses (Penichel, l9h5).

Some insight into the nature of the proposed relationship
between disturbances in verbal functioning and inhibitory processes
may be gained from a consideration of repression as elaborated
by Freud. Freud (1959) speaks of repression proper as an "after
expulsion" of ideas which were once conscious, but which have
been given the connotation of danger either by association with
the forbidden instinct or by virtue of an irradiation of the
repressed instinct-presentation. Because of the intimate con-
nection of ideation and language, verbal functioning must also

be affected by repression. For if ideational elements are eliminated

 

from awareness through repression, the verbal symbols which serve as

vehicles for that ideation must also be involved in the process.

The phenomenon of blocking, the temporary disappearance of
a word or name from memory3 is cited as a prime example of the
manifestation of repressive processes in speech. The speaker is
about to recite an incident or mention a name, when he is
suddenly stymied by his inability to recall a key word or name.
Freud's (1938) analysis of these blockages suggests that language
is intricately involved in the course of repression. For in
tracing the connection of the blocked (derivative) words to
their sources, it is similarities and contrasts in sound and
meaning that fonm the connecting links. Thus it becomes apparent
that there is an interaction between repression and language.
Not only does repression affect speech, but the peculiarities
of language play a part in the fonmation of the derivatives
which are repressed.

In the case of hostile impulses which have undergone the
fate of repression, the influence of repression upon speech
may be especially pronounced. Social expectations place a strong em-
phasis on verbal means of expressing hostility in contemporary
middle class society. Thus the inhibition and rejection of
hostile impulses and ideas becomes, to a great extent, the
inhibition and rejection of verbal communication with hostile
connotations.

Attempts to investigate these relationships in an experi-
mental setting have been notably lacking. The foregoing analysis
however, suggests several hypotheses which are capable of experi-

mental investigation: the individual with excessive repression

 

 

w. .. w... ut-w Margy-6" "7* 655

 

 

12

centering around hostile impulses may be expected to have at his
command relatively few words which are related to the expression
of hostility. Should such an individual be placed in a situation
demanding the use of hostile words, he would speak less fluently
than an ordinary individual. The stimulation of hostile impulses
by frustration would presumably have a similar effect, since
with the increased cathexis of hostile impulses, greater inhibi-
tory forces must be brought to bear, producing blockages in the
flow of speech. It is conceivable that the inhibition of
hostility may also have a more general retarding effect on verbal
fluency. The necessity for constantly guarding against the
expression of hostile tendencies may lead to a continuous verbal
screwing process, with a resulting drop in the spontaneity and
rate of speech. ‘

The present experiment is designed to provide evidence by

which some of the above hypotheses may be supported or rejected.

3

 

CHAPTER II

MANAGEMENT OF HOSTILITI: RESEARCH FINDINGS

The formulations of Dollard et a1. (1939) serve as a reference
point for much of the research done on the inhibition and expres-
sion of hostility. In their view, frustration provides the
necessary and sufficient stimulus for aggression. However,
aggression may be inhibited by anticipation of punishment for
the aggressive act, theistrength of inhibition varying positively
with the amount of anticipated pmistmsnt. Anticipation of
punith as used in frustration-aggression theory closely
approximates the definition of anxiety as a warning signal in
psychoanalytic theory (Fenichel, 1916). Predictions of the
expression of hostility are complicated by the additional assump-
tion in frustration-aggression theory that interference with
aggressive responses is itself frustrating and provides further

instigation to aggression.

 

The inhibition of selected respmses through punishment is
amply illustrated in avoidance conditioning studies. In an
attempt to simulate the process of repression, Ericksen and
Keuthe (1956) demonstrated that it is possible to eliminate
pro-selected responses in a word association test by the use of
shock, and that the Subject (§_) is not necessarily aware of his

suppressing responses. in experiment done with pre-school

13

 

Children illustrates the inhibition of aggressive responses through
punishment (Chasdi and Lawrence, 1955). These investigators
employed 23 children, placing 12 in an experimentally punished
group and the remainder in a control group. Both groups had

four doll play sessions, with the experimental group receiving
punishment for the expression of aggression in session II. The
frequency and intensity of aggression in session III was signifi-

cantly lower for the experimental group than for the controls.

 

This difference persisted in session IV, but was no longer sig-
nificant. I

A variety of studies with adults consistently find that
less hostility is expressed when the frustrator is vested with
power and authority than toward lower status frustrators (Cohen,
1955; Graham et al., 1951; Thibaut and Reichen, 1955). Roberts
and Jessor (1958) confirm these results, and in addition find
that the status of the frustrator has a greater effect on the
responses of those receiving high authoritarian scores on the
F-Scale than on low authoritarians. Also relevant here is the

,finding in verbal conditioning studies that the learning rate
for hostile verbs is slower when the experimenter is punitive
than when he or she is neutral (Binder et al., 1957; Ferguson
and Buss, 1960).

The expectation of relationships between parental treatment
of hostility and individual differences in the management of
hostility has guided several investigators. The findings in this
area are illustrative of the complexity of the problem of

predicting the inhibition or expression of hostility. McCord and

 

15

his colleagues (1961), drawing on extensive materials on 1714

non-delinquent lower class boys and their families collected

in the Carnbridge-Somerville youth project, attempt to idmtify
familial correlates of aggression. Their results indicate that
boys classified as aggressive were most likely to have emerged
from an environment that simultaneously instilled a high level
of aggressive urges and failed to provide conditions for con-
trolling aggression. As compared withnonaggressive boys, the
aggressive boys were treated in a more punitive, rejecting
fashion, had fewer demands placed on them, and were disciplined
in a more inconsistent fashion. Both ”over-control" and “under-
control" was found in the behavior of the aggressive boys'
mothers, whereas the nonaggressive boys were likely to be normally
controlled or "overcontrolled."

Sears and his colleagues (1953) argue that frequent punish-
ment for hostile behavior produces aggression anxiety which
interferes with the expression of aggression. But inhibition
of the aggression is itself frustrating, and results in a still
higher level of hostile drive. Thus, severity of punishment may
be positively related to the amount of expressed aggression.

In confirmation of this line of reasoning, Sears and his colleagues
(1953) found that mothers' statements of how punitive they were
regarding the expression of aggression correlated +.SO with
teachers' ratings of aggressiveness in 21 nursery school boys,

and +.60 with the observed frequency of aggression. For the

19 girls in this study, however, there was a negative relation-

ship between strength of maternal punishment and observed aggression

In

3&3"! 0".M35" ‘,.

— 4. ”W'JW

 

 

 

 

 

 

 

16

511 the nursery school. Their data also suggest that severity
of punishment for aggression has a curvilinear relation with the
amount of aggression displayed in the nursery school. A moderate
amount of punishment is associated with a high level of aggressive
behavior, while both greater and lesser amounts produce less
overt aggression. In a more extensive study (Sears et al., 1957),
including 379 boys and girls, measures of parental punishment
and children's aggression in the home were derived from inter-
views with the mothers. In this study, the amount of reported
aggression in the home was found to be related positively to both
the amount of punishment received and the amount of permissive-
ness by the parents. This relationship held for both sexes.
I report of a recent follow-up study of 160 of these younsters
at age 12 (Sears, 1961) presents further complications. Data
collected by means of a self report rating scale yielded negative
correlations with reported severity of parental punishment for
both sexes. Sears summarises the various relationships as
follows: "For both sexes, high punishment for aggression in the
first 5 years of life was associated with a high level of reported
aggression in the home, but at age 12 it was associated with a
low level of anti-social aggression. In boys, at both ages,
high permissiveness was associated with high aggression; in girls
this effect was clearly evident only at the earlier age, with a
mere suspicion of it at the later" (1961, p. h78).

While all of these studies demonstrate relationships between
parental punishment of aggression and the management of hos-

tility, the obtained correlations are by no means consistent.

 

1?

Punishment by itself is not predictive of the inhibition or
expression of hostility. Other factors, such as the amount of
frustration present in the home, the consistency of punishment
for aggression, permissive or restrictive attitudes toward the
child, the setting in which aggression is measured, the objects
of aggression, and the method by which aggression is measured
must also be taken into account in predicting the management of
hostility. The results also indicate that relationships between
punishment and the management of hostility may be quite different
for boys than for girls. Sears and his colleagues (Sears et al.,
1953, Sears, 1961) offer explanations for these sex differences,
but it will not be necessary to detail them here. The studies of
parental punishment and children's aggression may be said to
agree with the proposition that punishment may serve the dual
function of interfering with aggressive acts and providing further
instigation to aggression. The difficulty is that the studies
provide little information as to when and why one or the other
outcome occurs .

The variable of parental attitudes toward aggression has been
applied with some success to the perennial problem of predicting
overt aggression from performance on projective tests. Lesser
(1957) found a significantly positive correlation between fantasy
(aggression on a modified TAT and overt aggression as indicated
by peer ratings for 23 elementary school boys whose mothers,
on a questionaire, indicated that they were relatively supportive
of aggression. For the 21 boys whose mothers were relatively

discouraging of aggression there was a significant negative

=14.V‘AW=’ " d‘ :

 

 

 

18

relationship between fantasy and overt aggression. Similarly,
contending that controls on overt aggression are more lax in
lower class families, Hussen and Naylor (1951;) find a signifi-
cantly positive relationship between overt aggression and number
of aggressive TAT themes for a sample of 2b lower-class juvenile
delinquents, as predicted. These authors also found some support
for the hypothesis that fear of punishment as reflected in the
TAT is associated with less overt aggression. ’The above findings
are in contrast to the studies which fail to find a predicted
positive relationship between fantasy and overt aggression
(Gluck, 1955; Kagah, 1956). Both Gluck and Kagan suggest that
the variable of aggression anxiety, which went largely uncontrolled
in their studies, may account for discrepancies between amount

of fantasy and overt aggression.

Anxiety End the management of Hostility
It is generally accepted that the frequent association of

punishment with the expression of hostility leads to the antici-
pation of punishment or "aggression anxiety” when hostility is
aroused. Most theorists also agree that aggression anxiety may
inhibit acts of aggression, although there is some disagreement
as to the role of anxiety as the principle factor in the
reduction of aggression (Berkowits, 1958; Feshback, 1955). There
are several lines of evidence to indicate that the inhibition of
aggression is associated with anxiety.

In a study employing physiological procedures. (Funkenstein and

King, 19Sh), 69 male college students were subjected to frustration

 

l9

and their reactions were studied by means of ballistocardiograph
and blood pressure measures. Tape recordings taken of a sub-
seQuent interview served as the basis for classifying the §s'
psychological reactions to frustration as; "anger out," "anger in,"
I'anxiety," or miscellaneous. The various psychological responses
were found to be related to the type of physiological response.
Most of the "anger out" §s responded in a manner similar to the
reaction to noradrenalin, whereas the reaction of the "anger in"
and “anxiety“ §s were similar and resembled the physiological
response to the injection of adrenalin (p'<.001). These findings
are replicated in a rather inadequately reported study by King
and Henry (1955) with 111 male college §s. thder Linda's (1955)
finding of increased physiological tension among migraine patients,
who inhibited aggression when subjected to frustration, also
confirms that the inhibition of aggression is accompanied by
anxiety indicators.

1n the follow-up study of 160 12 year old children mentioned
earlier, Sears (1961) also concerned himself with the antecedents
of aggression anxiety and with its relation to antisocial aggres-
sion. His self-report questionaire was constructed to yield
separate scores for aggression anxiety and antisocial aggression.
Correlations with child rearing variables at age five confirmed
the predictable effect of punishment as a determiner of aggression
anxiety in girls, who also scored higher on aggression anxiety
than the boys. Nonpermissiveness of aggression toward the parents
was associated with aggression anxiety in both sexes. In boys,

withdrawal of love by the parents was associated with aggression

 

 

 

 

 

 

20

anxiety, but punishment for aggression was not. Sears suggests
that the development of aggression anxiety proceeds along quite
different lines in the two sexes. His interpretations of the
obtained sex differences appear to be carried far beyond his
data, however, and there is no recognition of the possibility
of obtaining the significant correlations of aggression anxiety
with some of the 22 measures of child rearing antecedents merely
by chance. The intercorrelation of antisocial aggression with
aggression anxiety on the questionaire at age lé is more substan-
tial, however (-.33 for boys, -.h8 for girls), and offers confirma-
tion of the association of anxiety with the inhibition of aggression.
In a correlational study which pertains to the association
of anxiety with the inhibition of aggression, Goodstein (195h)
hypothesised a negative relationship between anxiety measures on
the group Rorschach, manifest Anxiety Scale (HAS), and Iowa
Multiple Choice Picture Interpretation Test (IPIT) and hostility
measures of the group Rorschach and IPIT for 57 college students.
The predicted relationship was found between anxiety scores on
the MAS and Rorschach hostility. However, no significant rela-
tionship existed between anxiety and hostility as measured by the
IPIT, nor were the Rorschach measures on anxiety and hostility
inversely related. The low intercorrelations among the anxiety
measures suggests that in a study of this kind, an attempt should
be made to confine measurement specifically to aggression anxiety.
A few studies directly incorporate the variable of aggression
anxiety in relating fantasy productions to overt aggressive

behavior. The results of the study by Mussen and Naylor (19Sh)

 

 

 

 

21

on the TAT performance of delinquents, cited earlier, are pertinent
here. Among their juvenile delinquents, those ranking lowest

in ratings of overt aggression showed greatest fear of punishment
on the TAT. Prediction of overt aggression from the TAT was
enhanced when both fantasy aggression and fear of punishment

were taken into account. Purcell (1956), comparing 57 soldiers
with high, middle and low ratings of behavioral aggression, found
that the antisocial §s produced a significantly larger number of
undisguised fantasy aggressions on the TAT, and also significantly
fewer themes of anticipated internal and external punishment.

He emphasised that the same attention must be given to repressive,
inhibitory forces manifested in the data as to the impulse or
drive system.

In an attempt to,establish that frustration is followed by
increased punishment expectancy, Crandall (1951) randomly assigned
30 male college §s to frustrated and control groups. The increase
in punishment expectancy as reflected in the TAT was significantly
greater for frustrated than for control §s. These findings are
in agreement with the generally accepted notion that in the
socialized adult, the internalization of societal standards
(superego formation) results in internal controls which make
for the inhibition of tabu impulses even in the absence of ex-
ternal controls. Accordingly, we should expect that, with the
instigation to aggression held constant, individuals characterized
by strong internal controls on hostility should express less overt

hostility than those with more lenient controls.

v ._V,7 2"

 

 

 

 

\‘

.0

22

 

Individual Differences in the Management of Hostility

 

 

vendor Linde's (1955) study of the management of hostility
in migraine headache patients is an example of successful pre-
diction of differential response to frustration based on a know-
ledge of characteristic internal controls. On the basis of the
literature on the psychosomatics of migraine postulating that
migraine-prone individuals characteristically suppress their
hostility, vander Linde predicted that his 26 migraine patients
would exhibit greater physiological stress and less verbal
expression of hostility in response to frustration when compared .
to a control group similar except for the absence of a history
of migraine headache. In confinnation of his hypotheses, the
experimental group exhibited significantly greater muscular tension
and a faster heart rate following frustration. The experimental
group also gave significantly fewer aggressive responses on the
Rosenzweig Picture Frustration Study and on a post-experimental
questionaire than did the controls. 1 feature lacking in most
studies, but of real interest here, is that aggression was
assessed both before and after frustration by means of equated
halves of the P-P Study. The differences in response_to aggres-
sion highlighted by this technique were striking. Out of 26
control §s, 22 gave an increased number of extrapunitive responses
following frustration, whereas only h of the 26 migraine patients
gave more extrapunitive responses after frustration than before
(;)<.Ol). Interestingly enough, the experimental group produced
slightly more axtrapunitive responses than the controls before they

were subjected to frustration. Assuming that migraine-prone

 

 

23

individuals are characterized by a high level of aggression anxiety,
the results of this study are in agreement with the concept of
aggression anxiety as a significant inhibiting variable for the
expression of hostility. This assumption is given credence here

by the finding of relatively greater physiological signs of

tension in the migraine group during frustration.

A recent study by veldman and Worchel (1961), employing 80
male college Ss, also attempted to predict individual differences
in the management of hostility in a frustrating situation. 2
Drawing on Rogers' self-concept theory, the authors describe four
types of personalities and predict different responses to frus-
tration accordingly. Measures of self and ideal self discrepancy
(SI), and of defensiveness as indicated by performance on the
K scale of the MMPI, served as a basis for categorizing the Se
in one of the four personality types. Four groups of 20 Se were thus
obtained: low defensive, low SI (adjustive); high defensive,
low SI (repressive); low defensive, high SI (anxious); and
high defensive, high SI (distorters). The _S_s were subjected to
a frustrating "intelligence test," involving failure, distraction,
and insult. Dependent measures of aggression and aggression
anxiety, assessed by self report rating scales and sentence
completion items, were obtained immediately for half of the
§s and after a 20-minute interpolated neutral task for the others.

Predictions of differential response to frustration were in
general, supported by the results. The adjustive Se expressed

the strongest feelings of anger and the repressive group the least.

 

--N-(fr-v... g ‘ . ,. _.

4'1W-Q‘0

 

 

 

 

2h

In the delay condition, the repressive group received the highest
aggression anxiety scores, and was the only group that increased
in aggression anxiety after delay, suggesting that in time these
§s became somewhat aware of their hostile impulses. The High

SI group displaced more hostility than the low SI §s, as predicted.
These results, together with those of Vander Linde offer support
for the notion that different methods of managing hostility

exist, and that they are measureable and predictable.

' "lest Hostility" a_n_d Response to Frustration

The studies reported thus far testify to the importance of
both suppressive and impulse or drive factors in predicting
hostile behavior. The level of hostility displayed presumably
varies from one situation to another as a function of the balance
betwem these forces. There is some evidence to indicate that
results obtained from tests of hostility in a neutral situation
offer little clue to the level of hostility displayed subsequently
in a frustrating situation, and that individuals displaying little
" test hostility“ may actually react with greater than average
aggression when frustrated. The results of a study by Hokanson
and Gordon (1958), utilizing Miller's approach-avoidance conflict
framework are illustrative. In this study, half of a group of
L0 male college §_s representing extreme scores on Siegel's
Manifest Hostility Scale (MP5) were placed in a high hostility
arousal situation and half in a low arousal situation. S_s were then
permitted to express hostility in fantasy (TAT) and overtly
(shocking the E for wrong predictions about S on a personality

test). The situations for expressing hostility also varied in

 

25
their stimulus value for evoking hostile responses. Predictions

were that low test hostility (on MES) §s would manifest a net
decrease in hostile expression both in fantasy and overtly as
arousal conditions for hostility increased, and that high test
hostility gs would increase their expression of hostility with
greater arousal. Results showed significant differences in the
direction opposite of that predicted for fantasy aggression, with
no significant differences for overt aggression. Thus the low
test hostility gs showed a significantly increased amount of
fantasy aggression when aroused, while high test hostility gs
showed a significant decrease. These results are comparable to
those obtained by Thibaut and Coulee (1952), who found that initially
hostile gs, after receiving an insulting note, communicated a
significantly smaller amount of aggression to the supposed
insulter than the initially friendly §s who were insulted.
The findings of vander Linde (1955) cited earlier are similar.
His migraine patients gave slightly more extrapunitive responses than
the controls prior to frustration, but following frustration most
of the controls became more extrapunitive while the migraine
patients did not.

Berkowitz (1960) succeeded in replicating the results of Thibaut
and Coulee and Hokanson and Gordon, utilizing male college §s
in pairs, ostensibly with the purpose of investigating first
impressions. The degree of instigation to aggression was varied
by means of a standard set of false notes substituted for those
written by the §s. The results showed that §s scoring high on

Siegel's MES initially displayed significantly more hostility in

 

‘

‘ ...-1w" “W"

26

‘their communications, but that after insult, the low test hos-
tility §s increased their hostility more than the high test
hostility gs.

The consistency of these results prompted Rosenbaum and
Stanners (1961) to attempt a replication with the difference that
the § was subjected to personal attack in contrast to the in-
direct arousal procedures of the other studies. The results
failed to confirm the previous findings, as §s high in manifest
hostility also expressed more aggression on the TAT when aroused,
and in the low arousal situation, no differences were found be-
tween high and low test hostility §s. The authors suggest that
the discrepancies between their results and those of the earlier
studies reflect procedural differences, particularly, the use
of direct insult which aroused stronger hostility than the
indirect methods of arousal.

Berkowitz (1958, 1960) and Hokanson (1961) have attempted
to provide a theoretical rationale for the results they obtained.
Following Sears (1953), Hokanson assumes that frequent or intense
punishment for aggression produces aggression anxiety which in-
‘hibits aggressive acts, but that this inhibition constitutes
additional frustration. Thus the person characterized by a high
level of anxiety over aggression may actually have a higher level
of hostility than those with less aggression anxiety, and may
score higher on tests of hostility taken in a non-threatening
situation. ‘When faced with a situation where hostility is aroused,
however, aggression amxiety is also aroused and little overt

aggression appears. In a test of this interpretation, Hokanson

 

-v—“« "r.

. ($9.751!

Jua- :9

 

 

 

 

(1961) subjected high and low test hostility gs to insult and
failure, and found that high test hostility'és showed signifi-
cantly greater physiological signs of anxiety (systolic blood
pressure and GSH increases) than low test hostility §s, thus
confirming his hypothesis that high test hostility gs manifest
greater aggression anxiety when aroused. Since the arousal
procedures used in this study appear no less direct and intense
than those employed by Rosenbaum and Stanners, little support
is found for the eXplanation of their discrepant results as due
to procedural differences.

While the findings are not completely consistent and no
fully satisfactory explanation has been advanced for all the
results obtained, the discrepancies between "test hostility“
and reaponse to frustration revealed in these studies have sig-
nificant implications. We may conclude that performance on
tests of hostility in a conventional setting is not necessarily
representative of performance in situations where hostility is
aroused. The traditional expectation of a positive correlation
between test hostility and aggressive response to frustration
appears to be unfounded.~ In some cases there may even be a negative
correlation.

All of this suggests that certain precautions should be
observed when attempting to identify individual differences in
the management of hostility. In order to separate those who
inhibit hostility from those who express hostility when frustra-
ted, it is not sufficient to rely on measures of hostility taken

in a neutral situation. The individual's test performance when

 

 

... .r 2-753:- ‘ y..-

muas' ~:-' "

 

 

 

a.“

 

28

:frustrated would seem to conform more closely to other indications
of his management of hostility in a frustrating situation than
test performance in a conventional testing situation. The ideal
means of detecting the individual's method of managing hostility
would be to secure measures of hostility before and after frus-
tration. The pre-frustration measures provide a means of con-
trolling for individual differences in level of hostile drive.
Increases and decreases in the level of hostility displayed
after frustration as compared with pre-frustration measures would
then reflect the use of eXpressive or suppressive methods of
managing hostility respectively. EXperimental support for the
validity of this approach is found in Vander Linde's study
(1955), in which migraine patients did not differ appreciably
from controls in the amount of extrapunitive aggression prior

to frustration. 'When pre— and post-frustration tests were comp
pared, however, significantly fewer of the migraine group in-
creased in extrapunitiveness, as demanded by the hypothesis. In
addition, the pre— to post-frustration changes also agreed‘with
other indications of differences in the management of hostility
in the two groups; the migraine group was significantly less
hostile on a post-experimental questionaire, and exhibited sig-
nificantly more physiological tension (aggression anxiety?)

during and after frustration.

 

CHAPTER III

MANAGEMENT OF HOSTILITY AND VERBAL FUNCTIONS: RESEARCH FINDINGS

The research areas included in this section of the review
are scattered and varied. There is little research bearing
directly on the topic of management of hostility as it relates
to verbal functioning. However, each of the studies reviewed

below has some contribution to make.

Management of Hestility and Performance 2f Intellectual Tasks

 

 

Uhder the assumption that aggression anxiety produces re-
spouses which interfere with the expression of aggression, we
should expect the effects of frustration to very with the strength
of interfering reaponses. That such interfering reSponses can
have inhibitory effects on behavior other than aggression is
suggested in a study by Child and Waterhouse (1953). In their
study, the effect of frustration for college students whose
response to a rating scale indicated the presence of interfering
responses following frustration, was to produce a slight de-
crement in performance on an intelligence. among I‘low inter-
ference'I §s, the frustration produced a significant increase
in performance. Comparable findings are reported by Sarason
(1961), also among college students. Sarason was concerned with
relationships between various measures of anxiety and performance
on difficult anagram tasks under conditions of threat and no
threat. He found that high anxious subjects as classified by

scores on the Test Anxiety Scale and Lack of Protection Scale,

29

 

 

many-1.! .- j . -

 

 

 

adien threatened, performed at a significantly lower level than

low or middle anxious §s. Under conditions of no threat, these
relationships were reversed. These studies suggest the presence
of relatively enduring individual differences in patterns of
response to frustration, and further, that these differences

may be linked to the operation of anxiety. Of significance for
our purposes is the finding that patterns of response to frustra-
tion have implications for broad areas of intellectual function-
ing. .

The findings of Lants (l9h5), who tested nine year old children
under conditions of frustration and no frustration, suggest that
the general effect of frustration is an inhibition of verbal
associations. In this study, the task of “naming all the words
you can think of" was more disturbed by frustration than any of
the other subtests on the Stanford-Binet Intelligence Test.
Unfortunately for our purposes, this study did not attempt to dis-
cover whether the disruptive effect varied with the individual's
pattern of reaction to frustration, as in the studies by Child

and‘Waterhouse, and by Sarason.

Depression 229 Verbal Productivity

 

The verbal behavior of individuals in states of depression
is relevant to our topic if we accept the interpretation of
depression as indicating the presence of guilt or internalized
hostility. That depressed states are frequently associated with
diminished verbal activity or even muteness is indicated in
clinical and textbook descriptions of depression. Also, reduced

reSponsiveness and increased reaction time on the Rorschach is

 

31

'taken as one of the signs of depression (Beck, 19h9; KlOpfer

et al., l9Sh). ‘Welch and his colleagues (l9h6) compared elated
and depressed patients of all classifications on the number of
associations given to nonsense syllables. The depressed patients
were significantly less productive than the elated patients,
with almost no overlap between the groups. Compared with the
association scores of normals obtained from previous data, de-
pressed patients were below the mean for normals in verbal pro-
ductivity. Thus, there is some evidence that extreme inhibition
and internalization of aggression,'as exemplified by depression,

is associated with lowered verbal output.

hgnagement pf Hostility and Verbal Behavior of Paychotherapists

 

 

The eXpanding literature on verbal behavior in psychotherapy
provides further data on the management of hostility and verbal
functioning. Bandura (1956) reports a significant negative
relationship between ratings of the therapist's-anxiety over
hostility and psychotherapeutic ccmpetance. In a second study
(Bandura et al., 1960), this relationship is made more explicit.
Twelve therapists in training were rated by their supervisors as
to their characteristic way of managing hostility. Mere than

one hundred taped interviews from 17 patients were then analyzed,
- and the therapists' responses to hostile statements made by
patients were scored as approach or avoidance responses. In
support of the hypotheses, it was found that therapists who
typically eXpressed hostility directly and had low needs for
approval‘were more likely to recognize and encourage the communi-

cation of hostility than therapists who did not express direct

 

 

 

 

 

32

liostility and who had high needs for approval. Therapists who
'were rated as characteristically inhibiting hostility and
sensitive to social disapproval tended to avoid hostility ex-
pressed by the patient by remaining silent, by ignoring or
mislabeling the hostility, or by changing the t0pic. The effect
of this avoidance by the therapist was a stifling of further
communication of hostility or a change in the object of hostility.
These findings suggest that anxiety over the expression of
hostility is associated with avoidance of verbal materials assoc-
iated with hostility. ‘

lnxiety and verbal Functioning

Somewhat further removed, but still relevant to the purposes
of this review are several recent studies which deal with.relations
between anxiety level in psychotherapy and verbal functioning.
Mahl (1956) reports the development of a measure of speech dis-
turbance which includes sentence incompletions, corrections,
repetitions, stuttering, omission of words, and repetition of
“ah.” He found this measure of speech disturbance to vary pos-
itively with anxious-conflictual phases of the therapy hour as
judged by the author for two of his patients. Boomer and Goodrich
(1961) attempted to replicate Mahl's findings with different
judges, and found a satisfactory inter-rater reliability for the
speech disturbance measure. However, a significant relationship
between anxiety and speech disturbance was obtained for only one
of the two therapist-Judges. When five eXperienced clinicians other
than the patients' therapists were used to judge the patient's

anxiety level during the interview, no significant relationships

 

m
F‘t‘!

 

 

 

 

 

 

33

\oetween anxiety and speech disturbance were obtained.

In a slightly different approach, Panek and Barclay (1959)
related speech disturbances in therapy patients to emotional
arousal as indicated by GSR deflection. They used tape record-
ings and synchronized GSR recordings on four clients with four
interviews each. Dividing the interview into intervals of 30
seconds, they found a significantly positive association between
speech disturbance and emotional arousal. Dibner (1956, 1958),
working with a similar measure of speech disruption, provides
some evidence for regarding speech disruption as indicative of
transient anxiety, but failed to find a significantly positive
relation between speech disruption and skin conductance measures.
This discrepancy from Panek and Barclay's results may be due to
differences in units of measurement. Dibner's GSR measures were
based on mean changes during the interview, whereas in the Panek
study, momentary changes were recorded. The more gross comparisons
of five minutes reported by Panek and Barclay also failed to show.
a significant association between skin conductance and speech
disturbance. While the evidence is far from conclusive, these
studies lend some support to the hypothesis that transient anxiety
is accompanied by disturbances in verbal functioning. This con-
clusion is of importance to our interests if we assume, from our
discussion in the previous chapter, that the inhibition of
aggression is associated with aggression anxiety. On this basis, the
individual who inhibits aggression may be expected to show dis-

turbances or non-fluencies in speech when frustrated.

 

3h

The studies reported up to this point agree to some extent
in showing that the presence of anxiety is associated with lowered
verbal productivity and inefficiencies in verbal functioning.
There is some evidence, however, to show that anxiety may some-
times facilitate rather than inhibit verbal productivity.

Davids and Eriksen (1955), Operating from a Hullian framework,

3

y r
test the hypothesis that high anxiety §s produce a greater number f
of associations to stimulus words than low anxiety 58. In support §
- "

of this hypothesis, they found a significantly positive correla- é ;

tion between scores on the Manifest Anxiety Scale and productivity
of associations for to male college students. The authors also
show that the relationship was not due to mutual correlations
of their measures with intelligence.

' In a recent series of studies, Kanfer (1958, 1959, 1960)
has attempted to show that momentary anxiety increases the rate
of verbal responding. In his first study, Kanfer found that the
effect of a warning signal preceding a noxious shock was to
increase the rate of continuous verbal responding, suggesting that
anxiety (of a certain type) and verbal production are compatible
and not competing responses. In a second study (1959), 20
college students were asked to talk for four minutes on each of
five topics which were assumed to vary in amount of anxiety
aroused. No group differences were found between topics, nor
were there any significant differences in verbal rate for groups
of gs rated well-adjusted or poorly adjusted. However, on two
of the tepics (family relations and sex), poor adjustment ratings

correlated positively and significantly with verbal rate, in

 

 

.- .znv 7'“'Z_;';_,._l‘_ 1 '

.‘
i
'i!

 

 

 

..qalk‘

 

35

Shipport of predictions. The third study (Kanfer, 1960) employed
36 recently hospitalized female psychiatric patients. The design
called for an investigation of relationships between verbal rate,
content (tepics), and eyeblink rate in eXperimental interviews.
In addition, MMPI scores were obtained, and each § was rated by

a psychiatrist on the degree of conflict in each of the four

: 5..
topic areas. Contrary to findings in the previous study, verbal E g
rate was found to differ significantly across topics. The highest 5
rate was found on the tepic “reasons for hOSpitaliaation," the '
topic chosen SEEEEEEi as most anxiety provoking. Eyeblink rate i

did not differ significantly from tOpic to topic, contrary to

the hypothesis, although the predicted positive correlation

between psychiatric ratings of conflict and blink rate was ob-

tained. The author considers his findings as supportive of his major
hypothesis, although the evidence is far from convincing. No
evidence is presented to support the selection of “reasons for
hospitalization" as the most conflictual topic, nor do the psychiatric
ratings and eyeblink data suggest that this is the case. Furthermore,
no attempt is made to equate the topics on the supply of words ap-
propriate to each topic, a variable of considerable importance

in determining rate of verbal reaponse (Bousfield, l9hh). The
negative correlations obtained in this study between the clinical
scales of the MMPI (including the anxiety scale) and verbal rate

are contrary to the findings of Davids and Eriksen (1955), and

offer some support for the position that general anxiety is

associated with a lowered rate of verbal reaponding in an inter-

view situation.

v" -t .." -3" " g‘
. . . ..

 

 

 

 

 

36

The studies relating anxiety to verbal functioning offer
somewhat contradictory results. It appears that anxiety is
sometimes associated with increased verbal productivity (Davids
and Eriksen, Kanfer), and sometimes with non-fluencies and
disturbances in Speech (Mahl, Dibner, Panek and Barclay). These

findings would not be discrepant if it could be shown that in-

E:
t r
creased verbal productivity is compatible with an increase in 5
speech disturbances. However, an inspection of data presented 6
‘ 5 I
by Wagner and Williams (1961), who employ both output and Speech 3 5
f“ é

disturbance measures, suggests that these measures are indeed
inversely related. One way of reconciling these findings, if

we accept their reliability for the moment, is to assume that at
lower levels, anxiety has a facilitating effect, but that as
anxiety is increased it comes to have an inhibiting effect on
verbal functions. It is not immediately apparent, however, that
the Mahl, Dibner, and Panek studies employed higher levels of
anxiety than those of Kanfer and Davids and Eriksen. The evidence
is far from complete, and more conclusive formulations of the
relationship of anxiety and verbal functioning will have to be

deferred, pending a more systematic investigation of this area.

Newcomb's hypothesis g£_Autistic Hostility,

 

 

As we have seen, few studies in the research literature deal
directly with relationships between the management of hostility
and verbal functioning. Newcomb's (19h?) concept of autistic
hostility bears on our problem, but has thus far received only
incidental experimental investigation. Newcomb posits that the

arousal of hostility frequently creates barriers to further

 

 

37

communication (autistic hostility). As a result, the hostile
attitude becomes persistent because the blocking of communication
(either totally by withdrawal, or partially by avoiding the
sensitive tOpic) prevents alterations in the relationship. This
withdrawal from communication can also function as a means of
expressing hostility, as in the case of snubbing.

Thibaut and Coules (1952) present findings partially support-
ing Newcomb's hypothesis. On the basis of their initial descrip-
'tion of their assigned co-workers, hl college students-who served
as §s were categorized as initially friendly or initially hostile.
The initially hostile §s produced a significantly smaller volume
of written communication to their partners prior to insult as
well as after receiving an insulting note. Also in line with
expectations from Newcomb's hypothesis, the novice therapists
studied in the previously cited study (Bandura et al., 1960)
tended to avoid responding to their patients' hostile remarks
when these attacks were directed against the therapist himself.
When the patients' attacks were less direct, the therapists with high
aggression anxiety continued to avoid the patients' hostility,
while those with less anxiety tended to encourage further com-
munication of hostility. These findings suggest the presence of
relatively stable individual differences in the degree to which

individuals reapond to hostility by the avoidance of communication.

Measurement of Verbal Functions

 

 

The settings and samples employed in the studies reported

in this section of the review are varied. Considerable variation

 

 

33

also exists in the measures of verbal functioning. Several studies

utilize naturalistic observation of change in verbal behavior as

it relates to indices of emotional arousal in the therapy hour

(Mahl, 1956; Boomer and Goodrich, 1961; Dibner, 1956, 1958).

Somewhat more control is offered in the studies which create an

experimental interview situation and attempt to vary the arousal 5
value of the topics (Kanfer, 1959, 1960). However, the measures
of verbal rate obtained in such studies offer no indication of
the relative efficiency of verbal functioning, as the § may be 2 %
avoiding the issue or repeating himself. Only one study (Lantz, '
19h5), done with children, investigates the effects of experimental
frustration on verbal productivity. Other studies which employ
experimental frustration along with individual difference measures
of intellectual functioning (Child and waterhouse, 1953; Sarason,
1961) offer suggestive evidence that the effects of frustration
varies significantly among individuals. However, these studies

did not use verbal functions as the dependent variables.

In investigating the effects of frustration on verbal function-
ing, there are obviously many aspects of verbal functioning which
could serve as dependent variables. The study of samples of
'verbal behavior obtained from interviews offers some advantages
111 being similar to real life situations. The use of psycho-
liJlguistic measures (Dollard and Mowrer, 19L?) also involves
a-Inore or less spontaneous verbal sample such as the interview

.PIVTVides. However, the interview, by its very nature, involves
some loss of experimental control. Even in the " experimental

interview" in which the s is asked to talk for a period of time

 

 

39

on specified tepics, there is no way of insuring that he confines
himself to the area of interest to the experimenter. Thus the

§ who produces words in great quantity may merely be adept at
"beating around the bush." Such behavior is a defensive maneuver
commonly seen in therapy clients and is a skill not unknown in
other settings.

A form of verbal behavior which offers somewhat greater
experimental control, while still allowing the retention of an
interpersonal situation, is that of continuous production of words a
within a restricted category. This is a readily quantifiable i ‘
form of verbal behavior, and one which has received the attention
of previous investigators (Bousfield, l9hh; Bousfield and
Sedgewick, l9hh). By specifying the category of words to be
produced (e.g., words used to convey hostility or dislike), the
experimenter can study verbal productions in areas of specific
interest, with some assurance that the S confines himself to the
selected area. The total number of apprepriate words produced
in a standard unit of time thus offers a measure of effective
verbal functioning within a given area. The use of individual
experimental sessions, and oral rather than written production
(1f words, provides the essential elements of an interpersonal
stituation, a necessary feature if there is any intention to

irrterpret the findings in terms of interpersonal dynamics.

Pkniarement of Hostility and Verbal Productivity: ‘3 Pilot Study

 

Data bearing directly on the relation of the inhibition of
hOSVtility and verbal fluency was gathered by the author in a

P131C>t study. In this study, 1h college males were asked to

 

produce words in neutral categories (US cities and words ending
in a-t-i-o-n) and in a category related to hostility (words
which might be used to describe someone whom one hates or dis-
likes). Following these tasks, the experimenter related in an
insulting and disparaging manner so as to provide instigation
to hostility. The S was then asked once again to produce words
in a neutral category ( a-t-i-o-n words) and in a hostile category.
At the conclusion of these tasks, further insulting remarks were
made and the S was then asked to complete the Rosenzweig Picture-
Frustration Test according to standard instructions. The total
number of extrapunitive responses (TBS) on the P-F Test served as a
measure of the degree of inhibition and expression of hostility.
The general hypothesis was that the level of inhibition of
hostility is related to verbal fluency. Specifically, it was
predicted that the TES would be positively related to verbal
'fluency measures both in a conflict-related area and in areas
unrelated to conflict. Another hypothesis was that stimulation
of hostility would differentially affect §s who inhibit and Se
who express hostility. Specifically, low TES §s were predicted
to show smaller gains from Pre- to Post-Insult than High TBS gs.
In analyzing the data, the §s were ranked according to their
performance on the verbal fluency tasks and on their TBS. The
rank order correlations of extrapunitive scores with the various
measures of verbal productivity are presented in Table 1. This

table also shows the order in which the verbal tasks were

presented.

 

p..- .... It-u “ab-W.“

 

- _a.mm“f§z;tuz'1ermw “'6" ' '- A

”...-i. i ‘

 

 

bl

Table 1

Rank Order Correlations of TBS with Measures
of Verbal Productivity (N - 11.)

 

Pre-Insult Post-Insult
U.S. Hostile a-t-i-o-n Hostile a-t-i-o-n
Cities werds Nerds Words Words
-011 .16 052* 058* 055*

 

*Significant at the .05 level.

An additional analysis was carried out concerning the gain in
verbal productivity from pre- to post-insult. The gs were divided
on the basis of the TES into high-inhibition and low-inhibition
groups, with seven §P in each group. The gain in number of words
produced in each category was combined to yield a "total gain
score,” and these scores were ranked. (Gain scores for the separate
categories could not be used because of too many tied ranks.) The
Mann-Whitney U test was applied to the rankings and the results
permit the rejection of the null hypothesis at the .036 level.
These findings provide support for the hypothesis that stimulation
of hostility has a predictable differential effect on verbal
productivity. The verbal production of hostile inhibitors is
diminished relative to hostile expressors.

The results bearing on the effects of inhibition of hostility
on verbal productivity in areas unrelated to the area of conflict
are equivocal. The TES measures are unrelated to the production of
names of US cities, as indicated by the non-significant cor-

relation of -.11. However, significant correlations were obtained

 

'w-‘M Jar-tau haul-r.

 

 

 

h2

with the production of a~t-i-o-n words both before and after
insult. The suggested interpretation is that an activation of
conflict and anxiety was produced in the inhibited Es first by
the task of reciting hostile words, and secondly by the stimula-
tion of hostility in the insult situation. The stronger anxiety
in the inhibited §s resulted in diminished productivity. Thus,
verbal fluency in neutral areas may be affected only when con-
flicts are activated. This activation may occur by directly
stimulating hostility through frustration and insult, or less
directly by the production of words with hostile connotations.
It is obviously necessary to vary the order of hostile and neutral
words systematically, and to vary conditions of frustration
independently in order to test this interpretation.

A similar interpretation would seem applicable regarding the
correlations obtained between hostile words and the TES. Prior
to the eXperimental stimulation of hostility there appears to be
little relation between inhibition of hostility and production
of hostile words, as indicated by the correlation of .16. At
this point the §s task may be viewed somewhat dispassionately
by him, since it does not involve his current emotional state.

It is after the stimulation of hostility that the effects of
inhibition on production of hostile words becomes evident. Here
the task of producing hostile words becomes emotionally congruent
for those who eXpress hostility relatively freely, but increasing-
ly conflictual for those whose defenses inhibit the expression

of hostility. 'When challenged to produce more hostile words

after the insult, the expressors seemed to respond with some

 

 

 

L3

enthusiasm, while the inhibitors appeared to cringe from the
task. The findings here are not unlike those of Hokanson and

Gordon (1958), Berkowitz (1960), and Vander Linde (1955) reported

earlier. In the initial or test taking situation (here the pre—insult

situation), the §s who later demonstrate the ability to aggress
when frustrated are not distinguished by high scores on indices
of aggression, and may even receive low scores. 'With the intro-
duction of genuine motives for aggression, those who were able to
express their hostility received high scores on the P-F Test and
also produced more hostile words than those whose scores on the
P-F Test indicated an inhibition of hostility.

The difficulty in the preceding interpretations is that it
is assumed that while conflicts in the production of hostile
words prior to frustration are not strong enough togroduce dif-
ferential effects on hostile word production, the resulting
anxiety did produce differential effects in inhibitors and ex-
pressors on the subsequent neutral task. In order to avoid a
contradiction, an additional assumption must be made, that of a
guilt reaction which occurs during and after the production of
hostile words. This guilt becomes stronger among inhibitors and
interferes with the succeeding task of producing a-t-i-o-n words,
but has not develped sufficient strength during the production
of hostile words to inhibit hostile word production. This inter-
pretation finds some support in the study by veldman and WOrschel
(1961) who found that repressive subjects increased in anxiety
during a waiting period following frustration and insult.

The data of this pilot study certainly do not permit the

assertion of the above conclusions and interpretations with any

 

I!

'xYI-D

.wvuwhuhzm ..lh ‘ior . (333.... we

 

 

 

 
 

«fin-

 

~_ .3, _. s
1 ., .'.:.-1/ I ..‘f ._
1 .' g? -
bl:

degree of confidence. More questions are raised than answered
by the data. Still, the correlations obtained between individual
differences in the management of hostility and verbal fluency
offer encouraging support for the line of thinking develOped

in this review. The introduction of frustration was found to be

a means of producing predictable differences in verbal functioning.

 

 

 

 

CHAPTER IV

THE PROBLEH, HYPOTHESES, A} HETHCD

THE PROBLEM

It is the purpose of this research to investigate relation- 3
ships between the variables of frustration, individual differ-
ences in the management of hostility, and verbal productivity

in neutral and hostile word categories. Underlying the plan of

W ‘“m 'qT-M‘Q' .’
a

this research are several general assumptions:

1. Individuals react quite differently to objectively
similar frustrating situations.

2. The patterning of reSponses to a frustrating situation
are to some extent characteristic for the individual.

3. These reSponse patterns may be ordered on the dimension
of eXpression vs. inhibition of hostility.

The general thesis of this investigation is that the effects
of frustration on verbal productivity are variable, depending
on the individual's characteristic mode of managing the hostility

engendered by frustration.

I‘IYPO THE: EB

Psychoanalytic formulations of the processes involved in
the inhibition of forbidden impulses have suggested that the
inhibition of hostility is accompanied by a relative inability

to produce words associated with the eXpression of hostility.

hS

 

 

'“

".

 

 

 

146

Tfixis same prediction can be made from a learning theory view-
point. Assuming that frequent punishment for aggression has
produced the reaponse pattern of inhibiting aggression in frustra-
ting situations, it is to be expected that verbal responses
related to the eXpression of aggression should also be inhibited
either as a result of having been punished directly, or by
reSponse generalization, or both.

The results of the pilot study indicated that differences
in verbal output of groups designated as high and low inhibitors
of hostilitywere maximized following frustration. These results
were predicted from psychoanalytic theory on the basis of Freud's
assertion that a heightened cathexis of the repressed instinct
presentation elicits anxiety and greater counter-cathectic forces
which normally reduce the likelihood of a derivative gaining
consciousness. Another way of viewing the problem is in terms
of Miller's approach-avoidance conflict theory as applied to

individual differences in the management of hostility. In this

framework, the avoidance gradient of the § who responds to the experi-

mental frustration with relatively little aggression is more
elevated than that of the § who reaponds aggressively. Upon the
arousal of hostility, he soon reaches the point where avoidance
tendencies based on fears of retaliation become stronger than
[aggressive approach tendencies. Aggression anxiety, and the
.resulting avoidance tendencies, may be assumed to interfere with
\rerbal production, particularly when the words are those which
(“annote aggression. Thus it is to be expected that upon the

arousal of hostility, the individual inhibiting hostility will

 

.11“, J

. J

n
It.
i .4.

‘3‘, ém—ib' $.73} I‘ I‘m") “as. .1! ism“

 

 

 

L7

be at a disadvantage in producing hostile words.

On the basis of theoretical considerations, and from the
results of the preliminary investigation, therefore, the first
hypothesis can be stated as follows:

gypothesis l: Upon exposure to an hostility-arousing situation,
individuals identified as inhibiting hostility produce

fewer hostile words in a standard unit of time than

individuals who tend to express hostility.

The results of the pilot study indicated that, following
frustration, the production of neutral words was inhibited as
'well as that 6f hostile words in the subjects who inhibited
hostility. These results are predicted from the assumption that-
the interfering effects of aggression anxiety generalize to neutral
as well as hostile areas of verbal production. 5 second hypothe-
sis, based on the assumed effects of aggression anxiety and on

results obtained in the pilot study may therefore be stated:

Hypgthesis 3; Upon exposure to an hostility-arousing situation,

individuals identified as inhibiting hostility produce fewer
neutral words in a standard unit of time than individuals
who tend to express hostility. '

Other Relationshipngxplored in_this Study

 

~ The design of this experiment permits the exploration of
other relationships in addition to those predicted in the major
hypotheses. These relationships concern: (1) the general effect
of the frustration manipulation on verbal productivity, exclusive
of individual differences, and (2) the management of hostility
and verbal productivity in a non-frustrating situation. Neither
existing theory nor available evidence permit definite predictions

to be made for these relationships. To some extent, opposite

 

 

 

 

 

 

 

h8

‘prwmdictions are possible on the basis of present information.
Tfirerefore, statements will be in the form of questions rather

than predictions of outcome.

1. General effects of frustration on verbal productivity:

What is the general effect of frustration (ignoring individ-
ual differences) on the production of:

a. hostile words?
b. neutral words?
2. Relationships between individual differences in the‘manage-
ment of hostility and verbal productivity in a non-frustrating

situation.

Are individual differences in the management of hostility
signi ioantly related to the production of:

a. hostile words in a non-frustrating situation?

b. neutral words in a non-frustrating situation?

METHOD

Summary of the Experimental Design

 

The design called for 80 §s to be randomly assigned to
high arousal and low arousal groups, with hO §s in each group. Each
§ was required to produce words of a hostile nature and words
not inherently related to the expression of hostility. The order
of presentation of these verbal categories was counter-balanced in
both the high and low arousal groups, with 20 gs in each group
producing hostile words first, the other 20 neutral words first.
Within both groups, the §s were categorized, on the basis of
procedures described below, as to their method of managing

hostility. Twenty gs in each group were designated as having a

 

.....r ! " _ ~——
smi‘m- —
M. " .-

-, 5""

 

U9

punedominantly expressive method of managing hostility, and the

(other 20 as having a predominantly inhibitory method of managing

hostility.

The essential features of this design can be readily vis-

ualized by referring to Figure 1.

Figure 1

Representation of the Design Employed in the Investigation

 

 

- High Arousal Group

(N - to)
Management
of Hostile Neutral
Hostility Words Words

 

Egg Arousal Group
0; - to)
Management
of Hostile Neutral
Hostility' werds Nerds

 

 

Inhibitors ** 4*
(N - 20)

EXpressors ** as
(N - 20)

 

g
i
5
I

Inhibitors ** **
(N - 20)

Expressors ** as
(N - 20)

A.“

HDependent measures - number of words produced.

Sublects

Eighty subjects were drawn from the population of students

enrolled in introductory psychology classes at Michigan State

University by means of a volunteer sign-up sheet. The sample

was confined to male Se between the ages of 18 and 25 who were

native born. §s were seen in individual sessions of approximately

hS minutes duration over a period of five weeks. Assignment of

the §_to either the High Arousal or Low Arousal Group was made

prior to his appearance for testing by means of a table of

random numbers. In this way, LO Se were placed in each group.

 

 

“Mums-1m; am M,
PM %

 

50

Table 2 shows that the Se in the High Arousal and Low
Arousal groups were quite similar with regard to age, verbal
intelligence (as assessed by the verbal scale of the College
Qualification Test), and a rough measure of social class origin
consisting of the combined number of years of education received

by both parents.

Table 2

Comparison of Low.arousal and High Arousal Groups
on Age, Verbal Intelligence, and Parents' Education

 

 

 

verbal Parents'
Group. Age Intell. Educ.
High arousal Mean 20.20 h7.h7 25.88
Low Arousal Mean 19455 h7o35' 26.05
(N ' ’40) SeDe 10143 11009 11096
t 1.87 .05 .16
p .10 .90 .90

 

Procedures and Measuring Instruments

. In order to provide an overview of the sequence of events
in the eXperimental session, an outline of the experimental pro-

cedure is shown below.

Outline 2f Experimental Procedure

 

l. S is seated across the table from the experimenter, near
the microphone of a tape recorder. Obtain Se name, birthdate,
etc.

2. Administer Rosenzweig Picture Frustration Test (first half).

3.

5.

 

51

‘Instructions: "Each of the pictures I will show you contains
two or more persons. One person is always shown saying
certain words to another. For each picture, I want you
to tell me the number of the sentence that is most like
what you would actually say or really reply if you were
in these situations. Work as quickly as you can and avoid
being humorous. When you have answered a card, go right
on to the next one."

Administer Arthur Stencil Design Test.

Instructions: "This next test has been used for some
time as an intelligence test. Recently there have been
some indications that the test also relates to creative
ability. You can make all of these designs with these
cards. The important thing about this test is to look
for the pattern. Put the cards back after finishing each
design."

a. High Arousal gs: Administer designs 1-3 and 16-20.

(During last five designs) "30 seconds left . . . 5
seconds left . . . time's upl"

(Repeatedly and with increasing exasperation, while S
is working on the designs) "Watch for the pattern."

(Before last design) "Maybe ii? you would try a little
harder for once."

(At conclusion, while pushing the test materials impatient-
ly aside) "If you would have listened to me and looked
for the pattern you wouldn't have done such a lousy job."

b. Low Arousal Se: Administer designs l-B and h-8. Allow
ample time to complete each design. Mild commendation at
conclusion.

Verbal productivity tasks. Odd numbered So say neutral words
first, then hostile words. Even numbered Se say hostile words,
then neutral words.

Task is introduced as requiring "the ability to produce words
and ideas at will; an essential ingredient in any form of
creativity involving verbal communication." See instructions
for hostile and neutral word production below.

Arousal procedure administered to all Se (both high and
low arousal). Task is introduced as one which requires
both mental agility and control.

 

mean: 3727's? * '. -.

 

 

52

“I want you to count backwards from one hundred to one
by three's as quickly as possible. One hundred, ninety-
seven, and so forth."

(At first hesitation) "Not so slow, hurry] Faster now."
(When S makes an error) "You goofed on __. It should be __."

(At about 20 to 30 seconds) "You're already over the time
limit. YOu'll have to go back to the beginning. Now faster
Ufistnmfl

Repeat same procedure on second trial, making corrections or
saying the numbers for S when he hesitates, in a sarcastic
tone of voice.

On third trial, harassment continues, but 8 is allowed to
count all the way down to one. Errors and—time are noted,
and the experimenter says, "The practice didn't help much.
Is that all the faster you can count?"

6. Administer P-F test (second half) "Now do these."

7. Pest-Experimental Questionaire. I'I'll read the instructions
to you to be sure you understand." The instructions are then'
read to S'in a sarcastic tone of voice.

8. S is told of the real purpose of the experiment and assured
Es to the adequacy of his performance. S is invited to
describe his feelings and reaction to the experiment.
Further catharsis is provided in an interview in which S
is asked to describe his responses to a variety of other
frustrating situations.

A. Independent variables.

 

The design of this experiment calls for the manipulation of
two independent variables. One of these is an experimental or
treatment variable-the arousal of hostility. The other is an
individual difference or subject variable--the management of
of hostility. The procedures involved in the manipulation of 1
these two variables are of considerable importance, for the

interpretation of the data as related to the hypotheses depend

 

 

 

 

 

   

  

y/

 

 
  

l for is"
e ' 0 . I"
53

)‘3 El large extent on the assumption that these variables have been

effectively manipulated .

Situation designed £3 arouse hostility.

 

 

A variety of techniques designed to elicit hostility have been
described in the literature. An assumption common to each of
these techniques and supported by considerable evidence is that
the frustration of significant needs leads to hostility and
tendencies toward aggression. This assumption is also made in
the present study. The essential feature of frustration is the
blocking of a goal-oriented activity. The strength of the
instigation to aggression is assumed to vary directly with the
strength of the motive that is thwarted. '

In studies employing college students, a common procedure
is to thwart the subject in a task labeled as an.intelligence
test, under the assumption that college students are highly
motivated to obtain a high rating on intelligence, and failure
to do so is an injury to self-esteem. The success of any such
procedure depends upon the degree to which the task is perceived
as relevant to the needs of self-enhancement. In order to promote
ego involvement, the credibility of the task as related to the
ability it purports to measure must be established. Also, the
individual should exnerience some success at the outset so as
to insure his c00peration and raise eXpectations (later thwarted)
of a successful and self-enhancing performance. Another re-
quirement, often neglected, is of special importance in this
study; the frustration must be of such a nature that the source

of frustration may be legitimately perceived as external, so as

 

 

 

 

< ”no“? 3

 

———‘ — .‘-,..'4.E..".~ '2

 

 

,-..
. all "

2/

     

Sh

IVDtv'to bias the response toward an intrepunitive reaction based
on apparent limitations in ability.
A procedure which meets the above Specifications has been

described by Weiss and Fine (1956) and has also been used suc-

 

cessfully by Rosenbaum and Stanners (1961). The procedure in-
volves the use of the Arthur Stencil Design, Test I, one of the
subtests of the Arthur Point Scale of Performance Tests, Revised
Form II (19L7). The test consists of a series of 20 colored
designs of increasing complexity, which the é'is to reproduce.
by assembling the stencils of various colors and patterns. The
normal administration allows for a time limit of four minutes
for each design. In the modification of this test for the
arousal of hostility, all §s were asked to complete three easy
designs (1-3), with appropriate praise for their successful
performance. Subjects in the hostility arousal condition were

then civen five difficult designs (16-23) with a time limit of

one minute for each design. (None of Rosenbaum and Stanners' gs were

able to complete these designs in this time limit.) The g

meanwhile, contributed distractions, calling out the number of

seconds remaining, and urging the §.on with useless advice.

This was followed by criticism for not heeding his advice.

In the low-arousal condition, the three easy designs were followed

by five relatively easy designs (h-B), with no time limit.

Mild commendation was given these Es at the conclusion of the test.
An important consideration in the experimental arousal of

hostility is that the § should not be aware that the frustration

is part of the experimental treatment. In order to minimize this

possibility, and to distract the g from guessing at the real

 

 

SS

purposes of the experiment, the following deception was employed.
In recruiting students to serve as §s in this study, the experi-
ment was labeled as an investigation of creativity. The Stencil
Design Test was introduced as a test known to be useful in
detecting creative potential, an illusion that is enhanced by the
colorful stencils and the requirement of analyzing and constructing
the designs. None of the §s appeared to question the alleged

purpose of this test.

Measurement of the management of hostility

 

Our review of the literature has shown that discrepancies
are frequently found between "test hostility" and hostility
expressed in response to frustration. From these results we
concluded that performance on tests of hostility is not necessarily
representative of performance in situations where hostility is
aroused. Accordingly, it was suggested that the proper identifi-
cation of individual differences in the management of hostility
requires that measures be secured at the time the individual is
frustrated.

In order to achieve the proper conditions for the measure-
ment of individual differences in the management of hostility,
therefore, it was necessary to stimulate hostility in all §P
prior to testing for hostility. Thus, in addition to the procedures
for the experimental manipulation of hostility described above,

a second frustration procedure was undertaken following the
verbal production tasks and prior to testing for hostility. This
procedure was designed to provide further instigation to aggres-

sion in the high arousal group and to arouse hostility in the

 

 

 

56

low arousal group for the first time. The procedure consisted
of asking the §’to count backwards by three's from 100 to l as
quickly as possible. While the g was counting, he was being
subjected to repeated interruptions and sarcastic remarks by the
£5 and was twice told to go back and start again at 100. This
procedure was found to be effective in producing hostility by
Hokanson (1961) .

In order to perfect the details of the procedures for arousing
hostility, several tape recorded sessions were held with practice
Es prior to the collection of data for the study. The play-
back of these recorded sessions served to point up and correct
initial inconsistencies in the rather difficult role-playing
required of the eXperimenter.

Two instruments were chosen to measure individual differences
in the management of hostility. The first is a modification of
the Rosenzweig Picture Frustration Test which permits an assess-
ment of the direction in which hostility is expressed before and
after frustration. The second is a questionaire which provides
the § with an opportunity to express aggression fairly directly
in response to questions asking the § for his reaction to the E,

his evaluation of the eXperiment, etc.

a. The modified P-F test

 

The Rosenzweig Picture-Frustration Test was Specifically
designed to assess an individual's characteristic mode of handling
frustrating experiences. Heeponses to the frustrating situations
depicted in this test are categorized as expressing aggression

outwardly toward the frustrating agent (Extrapunitive), inwardly

 

toward the self (Intropunitive), or as minimizing the frustration
and denying hostility (Impunitive). In scoring symbols, these
responses are labeled E, I, and M respectively. For the purposes
of this study, the intropunitive and impunitive responses are
both regarded as indicating a relatively suppressive method of
managing hostility. Since our concern is with the expression
and. inhibition of aggression directed toward the frustrating
agent, the extrapunitive scores of the P-F test seem to be an
appropriate measure.

vander Linda's (1955) adaption of the P-F test includes
several changes which enhance the suitability of the instrument
for the purposes of this research. In this modification, the
pictures are presented individually, accompanied by three alter-
native responses chosen from the P-F test manual and actual pro-
tocols to represent extrapunitive, intropunitive, and impunitive
responses to the frustrating situations. The individual is asked
to select from these reSponses the one most resembling the
response he would make_to the situation depicted in the picture.
The use of forced choice rather than free response allows for
more objective scoring, and also eliminates the element of
verbal production from the independent variable. The major
advantage of Wander Linde's modification is the division of the
test into equated halves, each consisting of 12 items, which
permits the assessment of hostility with the same instrument
before and after frustration. The items included in each half,
in the order of their appearance, are shown in.Appendix A. A

more complete description of the procedures used in constructing

 

 

 

 

‘fl “I: \‘g‘ ~ .

 

 

 

 

       

t!

//l. v" '
4 - u
.. ." f i I __
ll ‘ ' \‘
r . r
I .' . "

 

58
the modified P-F test may be found in Vander Linda (1955).

In the present study, the order of administration of the two
halves of the modified P-F test was the same for all gs. The
gain in number of E responses from pre- to post-frustration
served to identify individual differences in the management of
hostility. A large gain signifies a relatively direct expression
of aggression, and smaller gains or decreased E reaponses signify

the inhibition of direct aggression.

' b. The post-egperimental questionaire

 

The purpose of the post-experimental questionaire was to
provide Opportunities for the §_to express hostility in fairly
direct ways which could be readily recorded and quantified. The
questionaire is so constructed that the §fs hostility may be
expressed in the fom of unfriendly reactions to the g, in
unfavorable evaluations of the procedures used in the experiment,
and of. the value of the experiment. Opportunity is also given
for the §_to indicate his unwillingness to cooperate with.§
in a follow-up study. In order to eliminate the element of
verbal production in these measures, responses were in the form
of ratings on a continuous scale 100 millimeters long, anchored
at both ends, rather than verbal responses. Scores are obtained
by measuring the distance in millimeters from the check marks to the
favorable (non-hostile) pole. Ratings on each scale may thus
range from zero to 100, with higher scores representing greater
amounts of hostility.

Although aggressive acts can be carried out with little

awareness of feelings of anger, the awareness of feelings of irritation

 

 

59

and anger are usually considered a part of the eXpression of
hostility. In order to assess the subjective components of the
management of hostility, the post-experimental questionaire also
included an Opportunity for the §_to rate his feelings, as described
by several adjectives, on a continuous scale from low to high,
with a possible range of from zero to 100, as with evaluation
questionaire ratings. Self-ratings on the adjectives irritated,
anggy, and annoyed served to indicate the degree of awareness of
feelings Of hostility.

In view of the significance attributed to anxiety in de-
termining both the expression of hostility and verbal produc-
tivity, four adjectives relating to feelings of anxiety (uneas ,

distressed, vulnerable, and apprehensive) were also included

 

 

in the self-rating scale.

The complete post-experimental questionaire can be seen in
Appendix B. Since the post-experimental questionaire provides
a measure of the expression of hostility independent of the
P-F test, two independent measures of hostility were available,
both of which were expected to relate to verbal productivity as

predicted in the hypotheses.

B. Dependent Variables

 

The design of this eXperiment calls for two categories of
verbal production, one directly related to the eXpression of
hostility, and one relatively neutral with regard to hostility.
The task of producing words in restricted categories lends it-

self nicely to the demands of this experiment. By specifying

 

 

__'T'¢N'ge - -
3n"; -3"?

 

 

the category of words to be produced, it is possible to control
the area of verbal discourse, while permitting maximum freedom
for the emergence of individual differences in verbal productivity
within the defined area. In requiring the §_ to Speak the words
to the eXperimenter in a face to face situation, the elements of
an interpersonal situation are maintained. The recitation of
hostile words in the presence of the E who has just frustrated
him may easily be given the connotation by the § of expressing
hostility toward the E. The total number of appropriate words
produced in each category, written down by E as they are produced
provides measures of effective verbal functioning in the pre-
selected areas.

The verbal categories used in this study were the same as
those employed in the pilot study. A time limit of three minutes

was set for verbal production in each category.

. Hos tile word category

 

The instructions for producing words in the hostile cate-
gory were intended to elicit words directly related to the
eXpression of hostility. Instructions for the production of
hostile words were:

"Say all the words you can think of which might be used

to describe someone whom one hates or dislikes. For ex-

ample; stupid, foolish, ugly, son of a bitch, and so forth.

Amr questions?"

Whether or not §_ replies, 22 continues:

"Say just as many words as you can think of which might be
used to describe someone whom one hates or dislikes."

"Neutral" word ca tegoiz

The instructions for the production of words in the

 

 

”I.

l"?-

 

61

"neutral" category were:

"Say all the words you can think of which end in the

letters a~t-i-o~n. For example: station, nation, elevation,

reputation, and so forth."

The reason for the quotation marks surrounding the word
neutral above is that the particular category of words was chosen
to represent a verbal category which is only relatively neutral
with regard to hostility. Although the structure of this task
permits the production of a wide variety of words not related to,
aggression, there are several words within this category which
are associated with hostility and which may be emotionally
loaded for.the individual with conflicts centering around the
expression of hostility. Examples are: accusation, damnation,
devastation, constipation, and castration. It wasanticipated
that associations such as these which.might be elicited (particu-
larly in the high arousal situation) would be warded off by §s
with inhibitory methods of managing hostility, with a resulting
decrement in total production of words in this category. The
selection of this particular category was thus an attempt to
provide maximal Opportunity for the generalization of inhibitory

processes.

 

 

.... ...-r»

 

CHAPTER V

RESULTS

Effectiveness of the Procedure for Instigating Hostility

 

 

A comparison of the High Arousal (HA) and Low Arousal (LA)
groups indicates that the stencil design procedure for instigating
hostility was quite successful. ley eight of the LO §s in the
HA group succeeded in passing any of the last five (difficult)
designs within the one minute time limit. Of these eight, six
passed only one design, one passed two, and one passed three
designs. Among the LA group, all of the gs were given ample time
in which to pass all of the last five (moderately difficult)
designs. Thirty-two of these Es passed at least four of the
designs within one minute.

The effectiveness of the arousalgrocedure is indicated by
comparing the two groups on measures of hostility and axziety
derived from the post-experimental guestionaire. It can be seen
from Table 3 that the HA is were significantly more unfavorable
in their evaluations on the post-experimental questionaire, and
also rated themselves as being si~nificantly more angry and
anxious during the interview than their LA counterparts. he HA
is also rated themselves as feeling significantly less competent
during the eXperimental session than the LA is, further confirming
that the arousal procedure had the desired effect.

The presence of significant differences between HA and LA

groups on both hostility and anxiety raises the question of how

62

 

 

wgz't KO .

 

J'fl ‘

 

 

63

 

 

 

 

mac. moo. mo. Ho. A. Hoe. a
mH.N so.mu ©0.H Hm.~ m4.m s
os.ma oo.mm Ho.ea me.mm mo.HH .a.m Hamaota
. :oq
oa.e- mm.e4 nm.m4 om.~m m:.mm cam:
Aimed 00.0m H14: :mém «wig 53m .3992
, \ awe:
.3. we. mm Hg. Om 2.. m4 9me not:
mpwawomom cosmooasoo hpoHXd< mafiaflpmom ouwwcoapmesj macaw
pwz apogeeuoaom apogee-oaom ptoaoe-oamm coauaaam>m

 

33483353 Hmpahmsflueofiwpmom one so mmcabom new:
"330cc 1350.3 33 new House: 2mg mo comaumasoo

m magma

 

, ran—... ‘

‘

FETIV‘ * 9'. h..'

 

 

6L

quzcessfully the arousal procedure elicited hostility as compared
yiith a more general emotional arousal. In order to determine
whether the arousal procedure had the effect of elevating hos-
tility to a greater extent than anxiety, each §'s mean rating of
anxiety adjectives was subtracted from his mean ratings on the

evaluation questionaire and hostility adjectives combined» The

 

difference obtained in subtracting mean anxiety scores from mean

hostility scores provides a measure of net hostility, the extent to

 

which hostility is predominant over anxiety fer the individual é. i ,
A comparison of the two groups on net hostility (Table 3) shows EJ
that the HA group scores significantly higher than the LA group.
This finding suggests that the arousal procedure elevated hostility
scores to a significantly greater extent than anxiety scores.
Within the HA group, the overall level of hostility ratings
is approximately equal to that of anxiety. For the LA gs, the
level of hostility is significantly lower than that of anxiety
(D - 7.10; t - 2.87). Thus, in terms of the absolute values of
the ratings on the post-eXperimental questionaire, the HA group
may be described as both highly hostile and highly anxious. The
LA gs, by comparison, are moderately anxious and significantly
less hostile in response to the eXperimental procedures. However,
in the absence of a third group in which no frustration occured,
there is no baseline for anxiety and hostility ratings in a non-
threatening situation. Thus the absolute levels obtained must be
interpreted with caution.
A comparison of the HA and LA groups on FF test extrapunitive
(E) scores before and after frustration is presented in Table h.

The groups do not differ significantly either before or after

 

 

65

g-'I'\).st.ration in the number of E responses. The relatively skewed

distribution of PF-E scores makes the use of theutn ratio somewhat

questionable in these comparisons, however.

Table 1.

Mean pre-frustration and post-frustration PF extra-
punitive scores of High Arousal and Low Arousal groups

 

 

 

Pro-frustration Pos t-frustration
S ,D. 1.288 1.5 92
LA New 10625 20075
3.13. 1.1.61 1.889
t .81 .77

 

A comparison of the HI and Li groups on the direction
of. changes in the number of E scores from pro-frustration to
post-frustration is presented in Table 5. This table shows the
number of §_s in each group whose scores increased, decreased, or

remained unchanged from the first to the second administration.

Table 5

Direction of PF-E change in HA and IA groups

 

 

m

 

Increase No Change Decrease Total
m 22 S 13 1.0

LA 15 1h 11 to

 

Chi square - 5.75, p. 05< .10

j 97m film-‘9“... Imp .-

 

 

 

Tflie resulting chi square of 5.75, which approaches significance
at the .05 level, provides some evidence of a systematic differ-
ence between the HA and 1A groups in the direction of PF-E change
after frustration. However, the differences are not entirely
consistent with expectations. More of the HA §s than LA §s
show an increase in extrapunitiveness after frustration, as
expected. But a few more HA §s also show a decrease in E scores
after frustration, a difference opposite of that expected.
Comparing gs within groups, we note that gs in the HA group
tended to either increase or decrease in extrapunitiveness, with
only five §s giving the same number of E responses after
frustration as before. In the LA group, the §s divide themselves
almost equally between increased, decreased, and unchanged E
responses.

In summary, the effects of experimental frustration were to
produce a large and significant elevation of hostility scores
on the post-experimental questionaire. While anxiety ratings on
the questionaire were also elevated by the frustration procedure,
the differences in hostility scores were significantly greater
than those of anxiety. These findings not only reflect favorably
on the adequacy of the frustration procedure, but also on the
validity of the post-experimental questionaire in discriminating
between groups which have been subjected to treatments varying

in degree of instigation to aggression. The effects of the
frustration procedure on the direction of change in PF-E scores
from pre- to post-frustration are less substantial and less

consistent. The effect of the frustration procedures on this

 

 

 

I! 0"”

 

6?

measure of hostility was a non-significant tendency for gs

tn) become either more or less extrapunitive, with very few §s

in the HA group giving the same number of E reaponses on the

second administration. Thus the PF-E change measures are relatively
insensitive to the differences in the amount of instigation to
aggression in the two groups, and support for the validity of F

the PF test in this study is somewhat tenuous.

 

Findings Regarding the Hypotheses

 

The hypotheses to be tested predicted a positive relation-

W3- :hrlznaa *

ship between the expression of hostility and the production of
(l) hostile words and (2) neutral words in an hostility arousing
situation (i.e., within the HA group). Two separate measures of
the eXpression of hostility, the post-eXperimental questionaire
and the Rosenzweig P-F test, were available for testing each
hypothesis. Since the findings are somewhat different for these

two measures, the results will be reported separately.

A. Findings based on the post-experimental questionaire.

 

The first hypothesis stated that upon exoosure to an
hostility arousing situation, individuals identified as eXpressing
relatively little hostility will produce fewer hostile words in
a comparable length of time than those who eXpress greater amounts
of hostility. The second hypothesis predicted that the same
relationship would obtain with regard to the production of neutral
words.

These hypotheses can be re-stated in terms of the Operations

used to measure the expression of hostility on the post-eXperimental

 

u!

.I‘ .0

 

Questionaire as follows:
Operational Statement of the Hypotheses (A)

Se in the HA group whose hostility scores on the post-experi-
mental questionaire are below the median will produce (1) fewer
hostile words and (2) fewer neutral words than Se whose hostility
scores are above the median.

Two analyses were conducted by way of testing these Operational
hypotheses. In the first analysis, the ratings on the evaluation
questionaire (evaluation hostility) and on the three hostility
adjectives (self-report hostility) were averaged to provide a
measure of total hostility (TH). The TH scores were then used to
divide the Se in both the HA and LA groups at their respective

medians into high hostility (HH) and a low hostility (LH) group,

 

with 20 Se in each group. The mean number of words produced in
the hostile and neutral categories for each of the four groups
' thus obtained is presented in Table 6. Support for the hypotheses

was contingent on finding a larger number of words produced in

the HA, HH group than in the HA, LH group. As can be seen from

Table 6, the tendency of the results is in a direction opposite

of that predicted. However, in the LA group there is a tendency

for HH §s to produce more words of both categories than LH So.
Since the verbal production scores were approximately nor-

mally distributed, statistical procedures involving the assumption

of normality were possible. The results of an analysis of variance

based on the TH breakdown are presented in Table 7. That the

difference in verbal productivity between HH and LH Se is neg-

ligible when the level of arousal is disregarded is shown by the

 

   

69

Table 6

Mean Number of Hostile and Neutral Words Produced
by S3 Above and Below the Median in Total Hostility

 

 

 

 

 

 

 

 

 

 

 

 

Hostility High Arousal Group Hostility fl Arousal Group
Level 54 NW Level m NH
High Hostile 18.55 114.00 High Hostile 18.50 13.05
(TH scores :50-82) (TH scores :36-65)
Low Hostile 22.20 15.85 Low Hostile 16.20 12 .140 1
(TH accruals-149) (TH scores :12-36) '
Total Group 20.38 1h.92 Total Group 17.35 12.72
Table 7
malysis of Variance of Verbal Production: Levels
of Hostility Determined by TH Scores on the
Post-Dcperimental Questionaire
Source df SS TB 1" ratio
107618 or hostility (L7) 1 13076 . 13076 .22
h‘ x LV 1 188.52 188.52 3078
Between Se in Lv x Ar 76 3793.32 10.91
Groups (Error "between”) ‘
Total Between 79 11268.60
Hostile-Neutral Words 1 1015 .06 1015.06 119.01;
H." x h. 1 6.& 6.82 033
H-N x Lv l 2.83 2.83 .11;
H." I ‘3' 3: L7 1 190,46 190,46 09h
H-H 1 Se in Lv x 1:- 76 1573 .33 20.70
Groups (Error "within")
Total Within 80 2617 .50
Total 159 6886.10

 

Net..- P.05, df 1, 76 ‘ 3.98

 

“Smm—slgnificant F values of .22 for the main effects of levels

of hostility and .IL for the interaction of H-N x levels of hos-
tility. The significance of the differences in verbal produc-
tivity between HH and LH §s, taking arousal level into account,
is evaluated by the F test for the interaction of Arousal x Levels
of hostility and by the triple interaction of H-N x Arousal 1
Levels of hostility. Neither of these interactions are found to
have an F ratio significantly different from chance expectation,
although the Arousal x Levels of hostility interaction approaChee
significance at the .05 level. An inspection of the cell means
in Table 6 shows that differences in verbal production in the LA
group contributed as much to the interaction as did differences
in the HA group.

The results of this analysis thus fail to support the
hypothesis of a positive relationship between the expression of
hostility on the post-experimental questionaire and verbal
productivity within the HA group. On the contrary, a non-sig-
nificant trend was found for HA Se scoring high on total question-
aire hostility to produce fewer words in both hostile and neutral
categories than low scoring gs.

A second analysis, which takes account of the differences
existing between the different scales of the evaluation questionaire,
indicates that the combination of ratings into total hostility
scores overlooks significant differences among the scales of
the evaluation questionaire. An inspection of the evaluation
questionaire reveals that the items differ in the directness

with which hostility is expressed. Only one of the scales refers

 

u -

1‘ .0

71

 

directly to unfriendliness toward the experimenter. The other
scales are less directly related to the frustrating agent in that
they provide for hostility to be expressed in unfavorable evalu-
ations of the testing situation, of the value of the experiment,
or in unwillingness to participate in another eXperiment with

the experimenter. The remaining scale occupies an intermediate

 

position with regard to directness of hostility in that it pro-
vides for unfavorable evaluations of the effects of the experi-
menter's behavior on performance and does not include the ele-

ment of unfriendliness.

 

Under the assumption that ratings of unfriendliness toward
the experimenter represents the most direct expression of
hostility on the rating scales, the prediction was made that
ratings on this scale would relate positively to production of
(1) hostile words, and (2) neutral words in the HA group. To
test this prediction, the Se in the HA and LA groups were divided
at their reSpective medians into "friendly" and "unfriendly"
groups, with 20 Se in each group.

A comparison of the number of hostile and neutral words
produced by So scoring above and below the median in unfriendli-
mess to the g is presented in Table 8. It can be seen that, in the

HA group, significantly more hostile words were produced by Se

 

whose reactions were unfriendly than by the "friendly" Se.
However, the difference between friendly and unfriendly Se in the
production of neutral words is negligible.

The mean number of hostile and neutral words produced by Se

falling above and below the median on each of the other scales of

 

 

l .W‘ ',4. I

   

72

Table 8

Mean Production of Neutral and Hostile Words by §s Scoring
Above and Below the Median on Each of the Ratings on the
Post-Experimental Evaluation Questionaire

 

 

High Arousal Group £3! Arousal Group

 

 

 

 

Questionaire Item HN NW HM NW
Reaction to E Range 5
Friendly (Range 3-57) 16.85 1L.75 {2-25) 16.25 12.00
unfriendly (59-87) 23.90 15.10 (28-67) 18.u5 13.15 3
t 3.22 .25 1.01 .87
Effect of E's behavior EJ
Favorable (Owéé) 20.00 1h.95 (3—50) 16.25 12.90 ,1.
Unfavorable (67-90) 20.71 1h.90 (51-98) 18.h5 12.55
t .29 -.03 1.01 -.27
Effect of test. situation
Favorable (15-60) 22.10 16.00 {2-33) 17.7h 12.79
Unfavorable (61-98) 18.65 13.85 (3S-8h) 17.00 12.71
t ”loll? -1059 -03. -006
value of EXperiment
valuable (Ll-SO) 21.90 1.6.15 (11-38) 17070 13035
Worthless (51-96) 18.68 13.70 (39-9h) 17.00 12.00
t -l.36 -l.82 -.32 -l.06
Participate Again
Willing (0-25) 22.25 15.25 (2~13) 18.50 1b.05
thilling (31-100) 18.50 lu.60 ( 978) 16.20 11.b0
t "1.61 "oh? “1006 .2015

 

Note.- t.05 - 2.09; t 01 - 2.70 (two-tailed)

 

 

.... 41%!9-“1

 

..14

 

tfu? evaluation questionaire is alsc presented in Table 8. With
the exception of ratings of the effect of Eds behavior on per-
formance, all of the other scales consistently show negative
relationships with the number of words produced.

The results of separately relating the different scales of
the evaluation questionaire to verbal productivity thus reveal
important relationships which were obscured by the combination
of the items with self-report hostility in a total hostility
score. The scale providing for the expression of hostility
directly toward the experimenter is the only one which shows a
significantly positive relation to verbal production. The scales
in which hostility is expressed more indirectly tend to relate
negatively to verbal productivity. The scale intermediate in
directness of hostile evaluations presents a relation with verbal
productivity which is somewhere between those of the direct
hostility scale and the indirect hostility scales. The relia-
bility of these findings is supported not only by the size of the
obtained difference in hostile word production between "friendly"
and "unfriendly" §s in the HA group, but also by the finding of
a similar pattern of differences in the LA group.

As a means of further eXplicating the relationship between
the expression of hostility toward the eXperimenter and verbal
productivity, an analysis of variance was performed, using the median
split on ratings of unfriendliness to E'as the basis for consti-
tuting levels of hostility. The results of this analysis are

presented in Table 9. The F ratio of 6.35 for the main effects

of Levels of Hostility, which is significant at the .025 level,

 

 

 

 

 

7h

Table 9

Analysis of variance of Verbal Production:
of Hostility Determined by Ratings of Friendly-
Unfriendly Reactions to the Experimenter

 

Levels

 

 

 

 

Source df SS HS F p

Arousal (Ar) 1 273.00 273.00 5.68 .025
Levels of Hostility (Lv) 1 305.30 305.30 6.35 .025
Ar x Lv 1 35.10 35.10 .73 N.S.
Between §s in Lw'x Ar 76 3655.20 h8.lO

groups (Error "between")

Total Between 79 b268.60

Hostile-Neutral Words 1 1015.06 1015.06 56.7h .001
H-H x Ar 1 6.82 6.82 .38 N.S.
H-N x Lv 1 139.0h 139.08 7.77 .01
H-N x Lw’x Ar 1 98.23 98.23 5.b9 .025
H—N x §s in LN x Ar 76 1358.35 17.89

groups (Error "within")

Total Within 80 2617.50

Total 159 6886.10

 

 

 

 

 

75

ixudicates that significantly more words were produced by §s
unfriendly to the E than by the "friendly" §s. The significant

F ratio for the interaction of Hostile-Neutral categories with
Levels of Hostility indicates that the effect of Levels of
Hostility is significantly greater for hostile words than for
neutral words. The finding of a significant F ratio for the
triple interaction of Hostile-Neutral categories by Arousal by
Levels of Hostility represents the finding that the difference

in verbal output of the "friendly" and "unfriendly" E? is greatest
for the production of hostile words in the High Arousal group.

The results of the analysis based on ratings of unfriendli-
ness toward the eXperimenter thus yield support for the first
hypothesis but not for the second. Upon exposure to an hostility
arousing situation, individuals expressing relatively little
hostility directly toward the experimenter produced significantly
fewer hostile words, but only slightly fewer neutral words than
those who expressed greater amounts of hostility.

In summary, hostility scores on the post-experimental ques-
tionaire were used to test the hypotheses by means of two separate
analyses. The first analysis made use of Total Hostility scores
which were obtained by averaging all of the ratings of the
evaluation questionaire together with the three adjectives re-
lating to hostile feelings on the self-report questionaire. The
results of the first analysis produced no support for either
hypothesis and the trend of the results is in a direction Oppo-
site of that predicted. In the second analysis, the different

scales of the evaluation questionaire were separately analyzed. The

s; ...-an- ....- :_

m
w...-
I

 

 

76

Iirediction of a positive relationship between the expression of
hostility directly toward the experimenter and the production of
hostile words in the H1 group found support in a highly signifi-
cant difference between "friendly" and "unfriendly" gs. The
negligible difference between "friendly" and "unfriendly" §s

in production of neutral words failed to provide support for

the second hypothesis, however. An uneXpected finding in the
analysis of post-experimental questionaire ratings was that ratings
on scales which provide for the eXpression of hostility without
refering directly to the experimenter (unfavorable evaluations of
the testing situation, the value of the eXperiment, and unwilling-
ness to be associated with the E in further experiments) show

a strong tendency to relate negatively to verbal productivity.

8. Findings based on the Rosenzweig Picture Erustration Test.

 

 

Extrapunitive scores on the PF Test provided a second kind
of measure of the amount of hostility eXpressed in reSponse to
the experimental procedures. In this study, the PF Test was adminis-
tered in two equated halves, one half before, the other half after
frustration. The change in number of extrapunitive responses
from the pre-frustration to post-frustration administration was
taken as a measure of the amount of hostility expressed in
response to the frustrating situation. Changes in the direction
of an increased number of E responses are assumed to indicate a
relatively hostile reaction, whereas a decreased number of E
responses indicates a relative inhibition of extrapunitive

hostility. The hypotheses under investigation can then be

 

—wr'—~_—I‘nv.rm. U I.

s- .....— .c..-!~'.t’~.‘!'5" - ‘.
"vi-r. s.9~:-~.v ..

 

 

77

restated in terms of the Operations used to measure the expression

of hostility on the PF Test as follows:

Operational Statement of E motheses (E)
g- in the HA group who show a relative decrease in extra-
punitiveness on the Rosenzweig PF Test after frustration will
produce (1) fewer hostile words, and (2) fewer neutral words

than go showing a relative increase in extrapunitiveness.

The distribution of PF change scores did not permit an
equal division of the groups at the median. The procedure used
was to divide the §s into high hostility (HH) and low hostility
(LH) groups on the basis of an increase vs. a decrease or un-
changed number of extrapunitive responses. This procedure pro-
vides 22 HH §s and 18 LH §s in the high arousal condition, with
15 HH and 25 LB §s in the low arousal condition.

The mean number of hostile and neutral words produced by
each group is presented in Table 10. It can be seen that the
differences in verbal production obtained between HH and LB §§
in the HA group are in the predicted direction for hostile words,
but not for neutral words. Among IA §p, those who show a gain
in extrapunitiveness produce more words of both categories than
LH §p. However, an evaluation of the differences in verbal pro—
duction of HH and LH §s in the HA group by means of a "t"
test shows that differences of this magnitude could easily have
occurred by chance. In the LA group, the difference in hostile
word production approaches significance. Thus, the hypothesis of

a positive relationship between increased extrapunitiveness and

?.jr.

“may; “-mfl‘ev

 

 

Verbal productivity in an hostility provoking situation is not

reliably supported by the data.

Table 10

Production of Hostile and Neutral Words for §s varying
in the Direction of Extrapunitive Change
Following Frustration

 

 

 

 

 

 

High Arousal Group L ! Arousal Group

PF Change aw NW PF Change HH NW
Increase Mean 21.18 1h.59 Increase Mean 19.66 13.20
(I - 22) 3.1). 9.07 5.03 (N - 15) 3.13. 7.11 b.18
Decrease Mean 19.39 15.33 g Decrease Mean 15.96 12.hh

or i or
NO Change SoDe 5071 3071 : “0 Change 80D. 601‘s 11023
(N - 18) 3 (N - 25)

t .71 -.52 i t 1069 .56
p1 .25 N.S. p2 .10 N.s.

1One-tailed probability.
2Two-tailed probability.

Further information about the relationship of PF extra-
punitiveness and verbal productivity can be gained by considering
the pre-frustration (PFqA) and post-frustration (PF-B) E scores
separately. It was contended earlier that prediction of behavior
from tests of hostility is enhanced when the measures of hostility
are obtained in a situation in which hostility is aroused. From
this we would expect the predicted relationship between extra-
punitiveness and production of hostile words to be stronger on
the PF-B measures than on the PFqA measures. Confirmation of this

reasoning is found by comparing the production of hostile words

(q p i

I
. -
‘uvaj, <7

1 Emma

 

 

among gs with two or more B responses (High E) with that of _S_s
giving zero or one E response (Low E)o

In the high arousal condition, the mean n ber of hostile
words produced by the 21 PF-A High E §s is 20.68, as compared with
an almost identical 20.10 for the 19 Low E §§. For the 28 §s
giving two or more B reSponses on the PF-B, the average number
of hostile words was 21.75, as compared with 17.17 for the 12
Low E §s. This difference yields a "t" ratio of 1.77, which is
significant at the .05 level. Thus, the rationale of administer-
ing tests of hostility in a situation in which hostility is
aroused is supported by the data.

This finding also finds some support among §s in the low
arousal condition. Here, the PF-A High E §s tended to produce
fewer hostile words than Low E §s (15.7h vs. 18.81; N - 19 and
21 respectively), whereas on the PF-B, High E §s tended to pro-
duce more hostile words than Low B.§§ (18.10 vs. 16.53; N - 21
and 19 respectively). Although neither of these differences
are statistically significant,'the results are in the expected
direction, i.e., measures of hostility on the PF Test relate
positively to the production of hostile words only when the test

is given after frustration.

Intercorrelations Among verbal Productivity and Other variables
The results obtained in the above analyses suggest that
relationships between hostility and verbal production differ for
different measures of hostility and also differ somewhat from the
HA to the LA condition. An intercorrelational analysis, presented

in Tables 11 and 12, provides an overview of some of these

 

 

 

80

relationships .1 Also, the relation of different measures of
hostility to each other and to anxiety, verbal intelligence, and
parental education are presented.

In can be seen in Tables 11 and 12 that the hostility scores
derived from average ratings on the post—experimental question-
aire show negligible correlations with verbal productivity in
both the HA and LA groups. The marked differences in the rela-
tion of the separate items of the evaluation questionaire to
verbal productivity are obscured by the combination of.items.

The other measure of hostility, PF extrapunitive change, shows

a tendency, falling short of statistical significance, to correlate
in the predicted direction with the production of hostile words

in both groups and with the production of neutral words in the

LA group. This result parallels the findings of the earlier
analysis based on an increase vs. no change or decrease comparison.

An unexpected result of correlating anxiety adjective ratings
with verbal productions is the finding that anxiety tends to
correlate positively with verbal productivity in the HA condition,
and negatively in the LA condition. Another interesting inci-
dental finding is the difference between HA and lA.groups in the
correlation of verbal intelligence with verbal productivity.
Verbal intelligence is shown to correlate significantly with
production of neutral and hostile words only in the HA group,

indicating that differences in verbal ability affected verbal

1Thanks are due to James Clark of the PSU Computer laboratory
for assistance in the preparation of the data for analysis by
the MISTIC computerv

 

 

u .\ . ‘2‘ OIL-8332' alto-bu.»
,.

REF}. 1

 

1

 

 

war: I" ,_~

1'.

0

... iV.»

 

Aoedaaouocov on. I mo.u “ApOHHeulozpv Hm. I mo.u n.0poz

 

no.+ OH.I wo.+ mH.+ mH.+ HH.+ ao.+ oao+ .odom Hopcouom

mO.I NN.+ HO.+ mo.+ Ho.t ano+ wm.+ .HHopsH Hwnho>

HH.| Ho.+ mo.I mOo+ mo.+ NH.+ omndco mm

mm mm.+ mmo+ A mHo+ woo+ mmo+ .Nca whoaomuwaow

me.+ sm.+ eo.- Ho.- apeaesmom fleece
mm.+ ma.+ mo.+ .pmom ouoaom-eaom
o~.n 50.: .pmom :owumSHe>m

ms.+ autos Haussoz

 

.afieeca omcaso «new me :umm m-w 32 2:
Henup> am

A04 u 2v
macho HomoOhd swam I moHpofiuo> hospo one
COHposUOAm Hdnam> mcos< ncoapmaoauoohoch

Hfl oHnaB

 
 

 

 

 

 

82

Aeoaempumeov om. u mo.u afieoaaasuozsv an. . mo.u -.osoz

 

 

 

 

+3.1. OH... Om... om... aw... 0H... 40... so... .38 anaconda
No.- NN.- mo.o m0.. m0.. mH.+ No.+ .HamssH Hashes
ma.- ma.+ mo.+ aH.+ o~.+ oN.+ enemas as
om... fin... hm}. 0m.l NH: .3 €88“..an
00.. mm.+ No.u mo.+ apaaasmom Hmpoe
em.+ Ho.u OH.+ .smom ptoammmaaom
mo.+ 00.. .pmom coasmsampm
am.+ mete: emanate
.HamscH «memes «-mm me mumm mum 32 2:
dashes an
Ac: . zv

95.5 Honda: 33 .. moflpmfiaob $50 one
"8333.5 gong wcofid ecoflmaoauoohofifi

NH 0.3.3.

 

83

behavior under conditions of psychological stress but not in the
non-threatening situation. The relation of the educational level

of the §s' parents to verbal production is insignificant, suggesting
that the differences in social class origin reflected by this
measure contributed very little to differences in verbal produc-
tivity.

The intercorrelations among the various measures of hostili-
ty shown in Tables 11 and 12 are generally rather low. There is,
however, a tendency for §s in the LA condition to be somewhat more
consistent than HA §s in their expression of hostility on the
different measures. The correlation of only .28 between average
ratings on the evaluation questionaire and self-report ratings
of hostile feelings shows that there were wide discrepancies
between subjective feelings of hostility and the expression of
hostility among gs in the HA condition. Correlations between PF
extrapunitiVe change and the questionaire measures of hostility
are so low as to be negligible in both the HI and LA groups. This
finding, together with the finding that at P!“ study failed to
discriminate between the HA and LA groups casts further doubt
on the validity of the PF study.

In both the HA and the LA groups, there are significant
positive correlations between ratings of anxiety and hostility
adjectives on the self-report questionaire. This finding reflects
the presence of a generalized arousal level, but the correlations
are doubtlessly also influenced by the existence of rating scale
response sets which are independent of the particular adjectives

being rated. Some discrimination in the self-report ratings is

   

implied, however, in that correlations of subjective anxiety with
evaluation hostility are lower than those of subjective hostility
with evaluation hostility.

Other M: General Effects of Frustratig 2n Verbal Productivity.

 

 

The question of the general. effect of frustration on verbal
productivity is give: a relatively clear cut answer by the data.
Kore words of both hostile and neutral categories are produced by
the Eu in the high arousal condition than in the low arousal con-
dition. The significant 1" ratio for the main effects of arousal.
which was found in the analysis of variance (Tables 7 and 9)
indicates that the obtained differmoss are reliably different from
chance Uh.) neutral and hostile word productions are considered
together. The non-significant PHI 1: Arousal interaction indicates
that the sane effect is presmt in both hostile and neutral word
production. An evaluation of the differences between HA and IA
groups on neutral and hostile words separately by means of 't"
ratios is presented in Table 13.

Since no predictions were made in advance, the two-tailed
test is appropriate. The obtained values of the I't" ratio permit
the rejection of the null twpothesis at the .05 level for the
diffemcs in the production of neutral words, but not for hostile
words, despite the greater differmcs bet-teen the groups in the
production of hostile words. We may conclude that the effects of
frmtration are to increase verbal productivity slumg college
sales. This increase is statistically significant only for
the producticn of neutral words, however.

:----r_v'.r

,- a."
—— “

 

 

 

 

 

 

 

Table 13

Mean number of hostile and neutral words
produced by gs in the HA and LA groups

 

 

Hostile Neutral

Group Words Words
High Arousal Mean 20.38 lh.92
(N - 1.0) 5.13. 7.62 24.32
low Arousal Mean 17.35 12.72
(N - 140) 3.1). 6.79 h.12

t 1.88 ‘ 2.32

p (two-tailed) .10 .05

 

me differences in the number of words produced by the HA and
Lt groups over 30-second time intervals is plotted in Figures 2
and 3. Figure 2 shows that the HA and LA groups produce about
the same number of hostile words in the first 30-second interval.
The differmce between the groups appears in the succeeding inter-
vals where the HA §_s maintain a higher rate of verbal production
than the Li gs. The difference in the number of hostile words
produced during the last two minutes approaches statistical
significance (t - 1.91). Thus the arousal procedure had a
motivating effect on hostile word production which tended to
persist over the three-ndnute interval.

Figure 3 shows that the motivating effects of frustration
were relatively transient for the production of neutral words.
The HA §_s produce a greater number of words during the first
minute, but thereafter the two curves cross and remain at about

the same level. A comparison of the number of neutral words

 

 

 

'1".
n3" ,3‘
I. .. '?.' .
. I 1,.
I” I..-

 

 

86
8
L
7 ‘\
Number 6'— \; Low Arousal -----
as We
Words S __ \K
\\
\\
h +- \
\
\\\
3 ” \\ \
\
\ \\i\\\\\
\
\
2 *- x.. - .. .\:\~_...__..
1 1 L L i :1:
30 6O 90 120 150 180
Time in seconds
Figure 2
Comparison of 1A.and HA Groups on
Rate of producing Hostile Verde
7..

 

 

 

 

 

Number Lwirousal - - - - -
Negiral High Arousal
Nerds
2E“ \"_.. \ \
1 1 1 L ..T--—T
30 60 90 120 150 180

Time in seconds

Figure 3

Comparison of La and HA Groups on
Rate of Producing Neutral'Words

 

 

 

TJI'"

 

 

 

87

produced by the two groups in the last two minutes yields a non-

significant “t" ratiO.

 

r1-'
v—“v w. _~~.V

 

 

 

CHAPTER VI

DISCUSSION

Tne major purpose of this investigation was to test the
hypothesis that individual differences in patterns of managing
hostility are significantly related to verbal functioning in a
situation in which hostility has been aroused. Specifically,
the investigation involved an attempt to demonstrate that those
individuals who are relatively incapable of expressing hostility
show a relative inability to produce words with hostile connota-
tions, as well as neutral words, when compared with others who are
gm... to express apprOpriate hostility when the situation warrants
it. Toe results of the investigation provide evidence that the
management of hostility is related to verbal functioning in a
frustrating situation, but that the relationship holds only
for the production of hostile words.

_'1_'_h_e_ Arousal 3f Hosting.

The formulation of the problem required. that verbal. produc-
tion take place in a situatiOn in which hostility was aroused.
There can be little doubt that this requirement was fulfilled in
the high arousal condition. From a definitional standpoint, the
presume of frustration was assured in that the §_s were thwarted
in their efforts to successfully complete the stencil design
patterns, and were subjected to harassing and disparaging remarks.

In addition, the E's unreasonable, demanding, and accusatory

88

 

 

 

 

.3: a W.

 

 

89

attitude was calculated to invite an aggressive reaction toward
himself and to minimize any tmdency toward self-blame for failure
on the test. That these procedures had the required effect of
arousing hostility is shown by the highly significant differences
between the HA and the LA _S_s in their post-experimental evaluations
and in their ratings of subjective hostility. Furthermore, the
item showing the greatest discrimination betwem the groups is
that of unfrimdliness to the experimenter, HA §_s giving an
average rating of Sh as compared with 31 for the LA §_s. The highly
reliable differences between the doubly frustrated In §s and the
is in the LA group, who were subjected to a single arousal procedure,
suggests that frustration had a cumulative effect in the HA group.
The qumtitive findings are supported by informal observations
of the §_s in the HA condition, and by their commute at the con-
clusion of the experiment. While no pmical violmce was attempted,
signs of anger and irritation were frequently evidmt in tone of
voice and facial expressions, in tmsed muscles, and more rarely,
in muttered words. One of the §s, at the conclusion of the
experiment, fairly well described what appeared to be the most
typical reaction when he said, aSeam like my face is a little red.
I really was angry, but tried not to show it.“ Another HA g
clearly revealed his struggle with hostile impulses when he
marked that he had not done very well on the task of producing
hostile words because, "I was too busy trying to get over being

and."

__Th_e_ Mtheses
Throughout the investigation, an emphasis was placed on the

 

conflictual nature of aggressive tendencies. The remarks just

quoted illustrate the conflict in which the HA §_s found them-
selves. It was assumed that diffeer methods of managing hostility
would be exhibited in the solutions of the conflict. Formulations
derived from psychoanalytic theory and from Miller's approach-
avoidance conflict theory led to the hypothesis that individuals
characterised by a relatively inhibitory method of managing
hostility would have difficulty in producing words associated

with the expression of hostility. On the basis of a presumed
generalizing effect of the processes associated with the inhibi-
tion of hostility, it was further hypothesized that such individ-
als would also manifest a relative deficit in the production of
words not associated with aggression. In testing these hypotheses,
measures of hostility takm immediately following a second
frustrating experimce were used to assess the individual 's

method of managing hostility. The results bearing on the two
hypotheses are summarized below:

%thesis 1. Upon exposure to an hostility arousing situation,
identified as inhibiting hostility produce fewer

hostile words than those who tend to express hostility.

a. Total hostility scores derived from ratings on the
post-experimental questionaire failed to show a positive
relation with the production of hostile words. On the
contrary, this analysis showed a non-signifcant trend
for low scoring _S_s to produce more words than high
scoring _S_s.

b. A separate evaluation of the items on the evaluation
ques tionaire revealed that ratings of unfriendliness

 

r“ ya?"

I

 

 

 

 

91

to the experimenter were highly associated with the
production of hostile words (P<.Ol). Ratings providing
more indirect expressions of hostility showed a tendency
to relate negatively to hostile word production.

c. §_s who obtained an increased number of extrapunitive
responses on the Rosenzweig Picture Frustration Test after
frustration do not differ significantly from other _S_s
in the number of hostile words produced, although the
obtained differmces are in the predicted direction.

figthesie 2. Upon exposure to an hostility arousing situation,

identified as inhibiting hostility produce fewer
neutral words than those who tend to express hostility.

a. Neither total hostility scores from the post-experi-
mental questionaire nor any of the separate items of the
evaluation questionaire show a significantly positive
relation to the production of neutral words.

b. No appreciable relation was found between increased

- extrapunitiveness on the PF Test and production of

neutral words.

It is apparent that the findings fail to provide any support
for the second hypothesis. Nothing in the data indicates that
individuals characterized by a relatively inhibitory method of
managing hostility differ from other individuals in the number
of neutral words produced when frustrated. This finding indicates
that the processes involved in the inhibition of hostility ap-
parently do not generalize to all types of verbal production.
Since only one neutral word category was anployed in this study,

 

 

92

it is possible that this category was atypical and that a generali-
ution of inhibition might be found with other neutral word
categories. While this possibility deserves exploration, it is
felt that the present category of neutral words provided maximal
opportunity for such generalization to occur. Although not
inherently related to the expression of hostility, words with
a-t-i-o-n endings may easily become involved in' the process of
expression and inhibition of hostility whm produced in a situation
provocative of aggression. Thus, along with such innocuous words
as station, nation ,elevation, and recreation, words related to
hostility (e.g., dalmatian, devastation, accusation, indieation)
and to guilt or anxiety (e.g., consternation, trepidation, w
m, castration) were sometimes produced. It was assumed that
stimulation of hostile impulses would elicit repressive forces
among the inhibitors which would result in blocking such conflictu-
al associations, and that the net effectowould be lowered produc-
tivity of words ending in a—t-i-o-n. One would suspect that
verbal categories less susceptible of association with hostile
meanings (e.g., names of 0.8. cities, or names of birds) would
be even less likely to show a relationship with methods of managing
hostility. The absence of any supporting findings in the data thus
presents rather damaging evidence against the proposition that
those who tend to inhibit hostility are characterized by a lack
of fluency with words not inherently related to aggression.

The tests of the first hypothesis provided some unanticipated

results, the sum of which are interpreted as offering support for
the existence of a relationship between methods of managing

 

 

 

 

um *1-"3' 1 w, '. . {1'

 

 

hostility and fluency with hostile words, in the population from
which the sample was drawn. In view of the relatively tenuous
support for the validity of the Rosenzweig Picture-Frustration

test in this study (as in others; of. Lindsey and Goldwyn, 1951;), it
is perhaps not euprising that comparisons based on this measure
failed to relate significantly to hostile word production. That

the obtained differences are in the predicted direction is, at

best, an indication that this data is consonant with the findings
based on the more valid post-experimental questionaire.

The tendmcy for total hostility scores on the post-experi-
mental questionaire to relate negatively to hostile word produc-
tion is, at first glance, apparent evidmce for the rejection of
the first hypothesis. teeming that production of hostile words
in the arousal situation represents the expression of aggression,
these findings also point to an inconsistency (not the first to
be reported in the literature!) between the expression of aggression
in two different situations. This apparent'lack of consistency
betwun the expression of aggression in verbal production and in
questionaire ratings led to a closer inspection of the questionaire
itcns. It was reasoned that such lack of cmsistmcy could re-
sult because of a difference in the degree of directness of
aggression in the two situations. The recitation of hostile words
in the presence of the frustrating experimenter may be regarded
as potentially very direct aggression. However, most of the
item on the evaluation questionaire provided for indirect expres-
sions of aggression, and the self-reports of feelings of hostility

were not expressions of aggression but descriptions of subjective

 

 

, “enizrfi'mwu ~'~ " ”

 

 

kiostility. Thus it was hypothesised that the questionaire item
providing for the most direct expression of aggression (unfriendly
reactions to 2) would show a positive relation to the production of
hostile words. The marked differences in hostile word production
between "friendly" and "unfrimdly" §s provides highly signifi-
cant support for this hypothesis.

The finding that of the five scales, only one is found to
relate significantly in the predicted direction may be grounds
for sane skepticism. Certainly these findings require replication
in another study. However, there are several reasons for accepting
the present findings as something more than a random, one chance
out of five, success. First of all, the comparison was based on
a prediction derived from external evidence of a difference be-
tween the test items and consonant with the reasoning developed
elsewhere in the study. Secondly, the significance level of the
obtained difference is far beyond the minimal level needed for
rejection of the null hypothesis. Thirdly, the differences
betweal "friendly“ and "unfriendly" §s is in the direction opposite
that of the general trend for the other scales and represents a
relationship which could not be assumed from the relation of total
hostility scores to hostile word production. Finally, the
presence of a uniformly similar pattern of findings in the La
group represents essentially confirming results in a replication
at a low level of arousal.

The major finding in relation to the first hypothesis is
that only the expression of aggression toward the frustrating

agent is found to relate positively to fluency with hostile words in

 

 

 

 

 

 

 

 

 

a frustrating situation. It is not the total amount of aggression,
but the amount of aggression focused directly on the instigator
that matters.

These results can be interpreted in terms of Miller's (l9bb)
approach-avoidance conflict model in relation to individual dif-
ferences in avoidance tendencies. Miller posits that in a con-
flict situation, the tendency to approach or avoid a goal is
stronger the nearer the §'is to the goal. The slepe of the avoidance
gradient is steeper than that of approach, so that the two may
cross. When the gradients intersect at some distance from the goal,
avoidance tendencies become,increasingly predominate with proximity
to the goal, and goal-oriented behavior is replaced by retreat from
the goal. In the context of the present study, approach tendencies
are based on the strength of hostile drives, and avoidance tenden-
cies on fears of retaliation (either internally or externally
imposed). Since the questionaire measure of total hostility did
not relate positively to the production of hostile‘words, the
findings cannot be interpreted as showing consistent individual
differences in the level of hostile drive. An explanation con-
gruent with the findings is provided by assuming the presence of
consistent individual differences in the elepe of the avoidance
gradient. In this framework, the presence of relatively strong
avoidance tendencies among certain §s inhibited their recitation
of hostile words in the presence of the frustrating experimenter.
These sane strong avoidance tendencies prevented these _S_s from
expressing a great deal of unfriendliness to the eXperimenter,

whereas unfavorable ratings of the testing situation or of the

 

 

 

 

.W'?.!fifi ... .r ' ' '

 

 

 

96

value of the experiment, being farther removed from the goal, were
not so inhibited. The tmdency for the unfavorable ratings not
directly focused on the experimenter to relate negatively to
hostile word production is accounted for in this framework by
assuming that the avoidance gradients of "expressors" and "inhibi-
tors“ intersect. Thus, as the distance from the goal increases,
avoidance tendencies drOp off more rapidly in the "inhibitors,"
reaching the point where “inhibitors" express more hostility than
l'expressore.‘| These formulations would explain findings of both
consistency and lack of consistency between the expression of
hostility in two different situations. Strongly emphasized in
this approach are the interactions of individual differences with
situational variables in determining the expression or inhibition
of hostility.

The findings of this study may also be interpreted in the
context of psychoanalytic theory. From the psychoanalytic stand-
point, it may be assumed that the _s_ brings into the eXperimental
situation a pattern of relating to authority figures which re-
flects earlier patterns of relating to the primary authority
figures of childhood. The frustrating and punitive experimenter
brings out conflicts in relating to authority, the solution of
‘which may be eXpected to resemble earlier patterns. Rapressive
elements in these solutions are reflected in the inability to
produce words with hostile connotations (derivitives of hostile
impulses). In addition, the self-imposed limitations on the
ability to express aggression in the production of hostile words

constitutes additional frustration, which contributed to a greater

 

 

 

level of hostility in the second frustrating situation. This

increased drive level is reflected in the tendency for overall

questionaire hostility to be associated with the lowered production

of hostile words. But in the questionaire situation, repressive ’

forces are still at work to prevent expressions which would bring

hostile impulses toward the experimenter (as a respres entative of

the primary authority figures) into awareness. Such repressive L
forces would be exemplified in a relative inability to adequately ;
identify the authority figure as the source of frustration. Thus E
as with low ratings on the scale which clearly involves the
experimenter tend to be those who produced relatively few hostile
words. The greater level of hostile drive in the repressive §_s,

and the mobility of the hostile impulses in seeking expression,

is manifested in a diaplaced or more diffuse hostile reaction,

vim, in ratings not carrying the label of the experimenter.

It is apparent that, despite differences in terminology, the
integration of the results into these two theoretical orientations
show many common features. In fact, each formulation is enriched
by the other. The bridge which was needed to relate the situational-
1y oriented approach-xvoidance theory to the phenomena of con-
sistent individual differences was the assumption of subject
differences in avoidance tendencies-«nan assumption implicit in
psychoanalytic theory. On the other hand, the relative inattmtion
to systematic foundations of situational determinants in psycho-
analytic theory is supplemented by borrowing from approach-avoidance
conflict theory the situational variable of neamess to the goal.

That the present findings are congrumt with such formulations is

 

98

rust to be taken as proof of their validity, since the marked
<iifferences between the questionaire items were not anticipated
prior to undertaking the study. It remains for further research
to determine whether hypotheses drawn from these formulations will
stand the acid test of empirical prediction. In view of the
findings in the study, it is suggested that a sound foundation
for such research requires careful preliminary attention to
determining successive degrees of nearness to the goal in the
particular setting in which aggression is to be measured.

In this study, the requirement of securing measures of hos-O
tility after frustration made it necessary to measure hostility
near the end of the experiment, after verbal production had taken
place. Hence it is not possible, strictly speaking, to say that
verbal functioning was predicted from questionaire ratings. Instead,
the relationships noted between the production of hostile words
and hostility expressed on the questionaire are taken as evidence
of relatively consistent patterns of managing hostility in two
situations, one providing for verbal expression of hostility, the
other for non-verbal expression. Independent and dependent varies
bles are somewhat relative in this situation. The data would be inter-
preted in much the same way if the production of hostile words
was used as the basis of predicting the expression of hostility
on the questionaire. Viewing the ability to produce hostile
'words as an independent variable, it is possible that proficiency
with hostile words in a frustrating situation may represent an
important clue to methods of managing hostility. In view of the

emphasis on handling aggression verbally in our society, the

 

‘ tu-r -.\_-—..-.—._ .
I 5*" , ~. _-.

 

 

 

 

 

 

 

99

individual who has an inadequate supply of appropriate words to
convey his hostility is at a disadvantage in expressing aggression
toward an instigator. Thus he may be forced to rely on less
direct means of expressing hostility (at a greater distance from
the goal), such as displacement, self-punishment, or withdrawal
from communication. 0n the other hand, if the instigation per-
sists, the inability to drain off hostility verbally may lead to

a more direct physical expression of hostility. The production

of hostile words.may thus eventually prove to be of some value

as a clinical tool.

 

One of the reasons for selecting a sample from a college
population in this study was the assumption that social expecta-
tions in this group tends to confine aggression largely to verbal
aggression. In a population in which less emphasis is placed
on the verbal expression of hostility, the relation between the
production of hostile words while frustrated and indications of
the management of hostility derived from other measures may be
less substantial or even non-existent. Thus, any generalization
of the findings to groups other than the population from which
the sample was drawn is unwarranted. Further research is needed to
explore the extent to which the present relationships obtain
among other groups. Since strong sex differences appear in much
of the research done on hostility, there is also a need to extend
the study of management of hostility and verbal functioning to

females as well as males.

 

   

100

 

(3ther Findings
In addition to the prediction of a relationship between the
management of hostility and verbal functioning in a frustrating
situation, this study was also designed to answer the question of
‘whether the same relationships are exhibited when verbal pro-
duction occurs in a relatively non-frustrating situation. The
results of the analyses in the low arousal situation show con-
sistent trends of relationships similar to those in the HA group,
but the differences in no case reach acceptable levels of statis-
tical significance. In interpreting these results, it should be
pointed out that failure to obtain significant relationships in
the LA group may have resulted from a less successful differen-
tiation of methods of managing hostility. The available evidence
in this study and in others (Berkowitz, 1960; Vander Linde, 1955)
suggests that proper differentiation of methods of managing hos-
tility may be enhanced when the measures of hostility are secured
at the time hostility is aroused. This is, of course, especially
applicable to an instrument such as the post-experimental question-
aire. 'While an attempt was made to arouse hostility in all §s
just prior to testing for hostility, the findings clearly show
that greater’hostility'was aroused in the doubly frustrated HA
§s than in the La §s who were frustrated only once. Since
measures of hostility in the La group were secured in a situation
of relatively mild arousal, it would not be surprising if these
measures do not reflect differences in the management of hoe-
tility as well as those obtained from the HA §s. In view of this

possibility, the failure to obtain significant relationships

 

 

—r—va- W\w_':x‘1,~“

 

 

   

101

 

‘between measures of the management of hostility and verbal functioning
in the lA.group does not present strong grounds for rejecting the
possibility of such relationships. The presence of consistent
trends in the data suggest that further exploration here may prove
fruitful.

Another question posed in this study'was that of the general
effect of frustration on verbal productivity. The findings in
this study contrast with those of Lants (19h5) who found that
frustration lowered the production of words unrestricted as to
category in nine year old children. However, the differences in
‘age of subjects and in the nature of the tasks and experimental
situation make comparisons between the two studies difficult
to interpret. That frustration increased production of both neutral
and hostile words in the present study, would suggest that the
increased drive produced by frustration did not selectively affect ver-
bal production which was drive relevant. These findings are con-
gruent with the Hullian tenet that all habit tendencies activated
by a given stimulus are multiplied by the drive strength than
operating. However, the results must also be seen in the light
of inhibitory factors, the effects of which appear to be of equal
significance with drive variable for the production of hostile
‘words. Taking into account the higher level of inhibition associated
with the production of hostile words in the presence of the frus-
trator, the effects of frustration may be interpreted as generating
both greater drive and greater inhibition for the production of
hostile words. The net result, therefore, is that hostile word

production shows no greater increase in the HA group than neutral

 

 

     

102

word production, which was not so affected by inhibitory forces.
The differences in rate of verbal production, shown in'l’igures
2 and 3, are congruent with this interpretation. The continuing
higher rate of production of hostile words by Hi §_s may be inter-
preted as an indication that the inhibition associated with the
production of hostile words in the presence of the instigator

contributed to the maintenance of drive. The merging of the

I

“E

.1 ml
\

a

production curves for neutral words after one minute suggests that,
in the absence of inhibition, the drive effects of frustration

on neutral word production were transient.

 

 

CHAPTER VII

SUVMARI

The purpose of this studwaas to investigate the relation-
ship between individual differences in patterns of managing hos-
tility and verbal functioning in a situation in which hostility
is aroused. As compared with studies emphasizing the situational
determinants of the expression of hostility, the present inves-
tigation focused attention on individual differences in response
to an objectively constant hostility-provoking situation. In
this stuns, individual differences in the expression of hostility-
are seen primarily as manifestations of different methods of man-
aging hostility, and are assumed to represent relatively enduring
patterns of behavior. The existence of a relationship between the
managenmt of hostility and verbal functioning in a frustrating
situation was inferred from psychoanalytic formulations of the
processes involved in impulse inhibition, as well as from.Miller's
approach-avoidance conflict model. These formulations suggested
that individuals characterized by a relatively inhibitory method
of managing hostility experience difficulty in the production
of words associated with the expression of hostility. That such
individuals may also be at a disadvantage in producing words not
inherently related to aggression was inferred from.the possibility
of a generalization of the processes involved in the inhibition
of hostility. Although such relationships have been observed in

clinical settings, there has been no previous attempt to investigate

103

 

 

 

b.1133”?

 

 

 

 

these relationships in a controlled eXperiment which provides
for the objective measurement of the variables involved. This
study was therefore an attempt to provide a more rigorous test
than has previously been conducted of the prOposed relationships
between the management of hostility and verbal productivity.

No hypotheses were formulated as a means of subjecting the
proposed relationships to an empirical test:

”thesis 3:: Upon exposure to an hostility arousing sit-
nation, individuals identified as inhibiting hostility will produce
fewer hostile words in a standard unit of time than individuals
who tend to express hostility.

Hgothesis _2_: Upon exposure to an hostility arousing sit-
nation, individuals identified as inhibiting hostility will pro-
duce fewer neutral words _in a standard unit of time than individuals
who tend to express hostility.

In testing these hypotheses, a sample was drain from along.
male studmts enrolled in introductory psychology classes at
Michigan State University. Eighty §_s were seen. in individual
experimental sessions. _S_s were randomly assigned to one of two
experimental conditions, termed the high arousal (HA) and low arousal
(LL) conditions. Prior to verbal production tasks, the hO HA §_s were
frustrated by contriving their failure on a test said to predict
creative potential, and by the experimenter's harassing and in-
sulting behavior. For the ho gs in the Lt group, frustration
prior to verbal production was minimal. The verbal production
tasks consisted of oral production of words in restricted cate-

gories. The hostile word category consisted of words used to

.A ready-tn,

 

105

describe someone whom one hates or dislikes, such as: dOpe, stupid,
ugly, etc. The neutral word category consisted of words ending
in a-t-i-o—n, such as: nation, station, elevation, etc. The
number of words produced in each category within a three—minute
period constitutedthe depaldmt measures of verbal productivity.
Measures of the management of hostility were obtained near the end
of the experiment, iJmnediately following a procedure for inducing
hostility which was standard for all §s. These measures consisted
of: (1) ratings on a post—experimental questionaire which pro-
vided for the expression of hostility in unfriendly reactions to the
experimenter and in unfavorable evaluations of the testing situa-
tion or the eXperiment, and (2) extrapunitive responses on a
modified, multiple choice version of the Rosenzweig Picture
Frustration Test, which was administered in equivalent halves
before and after frustration.

A canparison of the HA and LA groups on the post-experimental
questionaire ratings indicated that the HA gs were considerably
~ more hostile than the Lt §s in response to the eXperimental pro-
cedures. This finding was taken as evidence of the effectiveness
of the arousal procedure in producing the required hostility in
the HA group, and also as support for the validity of the post-
experimental questionaire. The failure of the T? Test to reliably
discriminate between HA and LI gs was interpreted as casting
further doubt on the validity of a technique that has been shown
by other investigators to have serious weaknesses.

The hypotheses to be tested concerned only the relationships

between hostility scores and verbal production measures in the

 

 

 

 

 

 

106

HA group. The method of testing the hypotheses was to compare the
number of words produced by gs whose hostility scores fell above
the group median (expressors) with that of go who received hos-
tility scores below the group median (inhibitors). Support for
the vaotheses was contingent upon finding that inhibitors pro-
duced fewer hostile words and fewer neutral words than expressors.
The results bearing on the hypotheses may be summarized as
follows:

Mothesis l: Total hostility scores derived from ratings

on the post—experimental questionaire failed to show a positive

 

relation with the production of hostile words. On the contrary,
this analysis revealed a non-significant trend for low scoring
§_s to produce more words of both categories than high scoring §s.
A separate analysis of the itmes on the evaluation questionaire
revealed that ratings of unfrimdliness to the experimenter were
highly related to the production of hostile words (P (.01),
whereas ratings providing more indirect expressions of hostility ‘
showed a tmdmcy to relate negatively to hostile word production.
as who obtained an increased number of extrapunitive respmses on
the PF Test after frustration did not differ significantly from
other §s in number of hostile words produced, although the obtained
differences were in the predicted direction.

nthesis _2_: Neither total hostility scam from the post-
experimmtal questionaire nor any of the separate item of the
evaluation questionaire showed a significantly positive relation
to the production of neutral words. No appreciable relationship
was found betwem increased extrapunitivsness on the 1’? Test and
production of neutral words.

 

     

107

The negative findings in regard to the second hypothesis were
interpreted as casting serious doubt on the prOposition that the
individual who tends to inhibit hostility is characterized by a
relative lack of fluency with words not inherently related to the
expression of hostility.

In interpreting the results bearing on the first hypothesis,
it was considered significant that ratings expressing hostility
directly toward the instigator relate positively to the production
of hostile words, whereas more indirect expressions of hostility
tend to relate negatively. These findings were regarded as pro-
viding some support for the first hypothesis, while at the same
time suggesting an additional refinement of this hypothesis. On
the basis of these findings, it was suggested that it is not the
indiscriminate eacpression of hostility, but the ability to
express hostility directly toward the instigator that relates
positively to the ability to produce hostile words when frustrated.
These results were seam as consistent with psychoanalytic theory
and with a variation of Miller's approach-avoidmce conflict
model which assume the presence of individual differences in
avoidance tmdencies. Attention was called to the possible sig-
nificance of proficiency with hostile verbalizations as affecting
patterns of managing hostility. The failure to obtain the signifi-
cant differences predicted in comparing §_s who showed different
directions of change in the number of PF extrapunitive responses after
frustration was taken as a reflection of the relatively low
validity of this instrument. It was pointed out that the sample

chosen represents a group in which social expectations confine

.r-‘Rfi. -HLHAW .‘—. __
v - . ' - .1:
‘ a ‘F

 

.411.“

 

 

 

108

hostility largely to verbal. aggression, and that generalizations
from the findings of this study must be confined to the population
from which the sample was drawn. Present findings are regarded

as indicating the need for further research in this area.

The design of this study also permitted comparisons within
the IA group. An evaluation of the extmt of the relationship
between the management of hostility and verbal productivity among
LA is revealed non-significant trends of essentially. the same
relationships which were exhibited in the HA group. In discussing
these results, conclusions were qualified by the possibility
that a less valid identification of individual differences in the
maganent of hostility was secured in the LA condition.

A comparison of .the HA and LA groups also permitted an eval-
uation of the effects of frustration on verbal productivity, ex-
clusive of individual differences. The results showed that
frustrated §_s produced more words of both categories than non-
frustrated §_s. That the motivating effect of frustration was more
transient for neutral words was sea: as consistmt with the assump-
tion that greater inhibition is associated with the production of

hostile words in the presence of the instigator.

 

 

1.

3.

b.

S.

7.

8.

9.

10.

   

REFERENCES

Arthur, Grace. A point scale 9_f_ performance tests. Revised

 

f_0_11_g E. New York: Psych010gica1 Corp. , 19147.
Babladelis, Georgia. Personality and verbal conditioning
effects. 9;. 93192“: 932. Psychol" 1961, _6_2_, Isl-1&3.
Bandura, A. Psychotherapists anxiety level, self insight, H
and psychotherapeutic competence. g. £39355. soc. Psychol.,
1956, _53, 333-337.

Bandura, 5., Lipsher, D. , & Miller, Paula E. Psychothera-

 

‘. f.._l1‘\-uv “‘1: ..‘AQ:

‘3 37,95 vi 4 '
5
"'J‘

pists approach - avoidance reactions to patients' eXpressions ' é:

of hostility. g. consult. Psychol" 1960, .219 1-8.

 

Beck, 8. J. Rorschach's test E. a variety 9_f_ Emonality
pictures. New York: Grune and Stratton, 19h9.
Berkowitz, L. The expression and reduction of hostility.

PSIChOle Bull" 1958, 25-, 257.2830

 

Berkowits, L. Some factors affecting the reduction of overt
hostility. _J_. m. soc. Psychol., 1960, @2, Ill-21.
Berkowitz, L. Manifest hostility level and hostile behavior.
,1. soc. Psychol., 1960, .S_2_, 165-171.

Binder, 1., McConnell, & Sjoholm, Nancey A. Verbal condi-
timing as a function of experimenter characteristics. ,1.
m. 193. Psychol., 1957, §_S_, 309-3124.

Boomer, D. 5., & Goodrich, D. W. Speech disturbance and

Judged anxiety. ,1. consul}. Psychol., 1961, 32, 160-614.

 

Bousfield, W. A. An empirical study of the production of

lffOCtively toned it”. :10 E10 PBIZCbOIQ’ 19M, 29-, 205'2150

109

   

13.

1h.

15.

16.

1'7.

19.

20.

110

Bousfield, W. A., a. Sedgewick, C. H. W. An analysis of
restricted associative responses. g. 592' Psychol., 19%,
3:, 119-155.

Buss, A. H., & Durkee, Ann. in Inventory for assessing
differmt kinds of hostility. g. consult. Psychol., 1957,
31, 3143-448.

 

Buss, A. H.,& Durkee, Ann. Conditioning of Hostile ver-
balization in a situation resembling a clinical interview.

:1. consult. Pszgh01., 1958, 2.2.! h15-18.

- Chasdi, Eleanor H., 8: Iawrmce, Margaret S. Some antecedents

of aggression and effects of frustration in doll play.

In D. McClelland (Ed), Studies _i_n_ motiviation. New York:

 

Appleton-Century-Crofts , 1955 .

Child, I. L., & Waterhouse, I. K., Frustration and the

quality of performance: III an experimental study. .1. Pers.,

1953, _2_1_, 298-311.
Cohen, A. R. Social norms, arbitrariness of frustration and
status of the agent of frustration in the frustration-ag-

gression hypothesis. g. abnorm. soc. P_s_ychol., 1955, a,

 

222-26.

Crandall, V. J. Induced frustration and punishment-reward
expectancy in thematic appercep'tion stories. :1. consult.
Psychol., 1951, _];S_, hOO-hOh.

Davids, A., & Eriksen, C. W. The relation of manifest anxiety

to association productivity and intellectual attainment.
_._J_. consult. Psychol., 1955, 12, 219-22.
Dibner, A. S. Cue—counting: A measure of anxiety in inter-

views. ,1. consult. Psychol" 1956, 23, 1:75-73.

 

 

 

'l-vn, "-1. ‘0': - g:

 

21.

22.

23.

25.

26.

27.

28.

29.

30.

 

 

Dibner, A. S. Ambiguity and Anxiety. i. abnorm. s_0__c_. Ps hol.,
1958, g, ices-"It.

Bollard, J., Miller, N. 3., Doob, L. W., Mowrer, O. 11., &

Sears, R. R. Frustration 9.99. a ression, New Haven: Yale

 

Univ. Press, 1939.

Dollard, J., & Mawrer, O. H. A method of measuring tmsion
in written documents. ,1. m. 393. P8221101" 19M, 93,
3-32.

English, H. 8., 8: English, Ave C. _A. comprehmsive dictionary
of psychologicgl and psychoanalytical terms. New York:

 

 

Longmans, Green 8: 00., 1958.

Eriksen, C. 11., 8: Kuethe, J. L. Avoidance conditioning of
verbal behavior without awareness: a paradigm of repression.
g. m. s_o_c_. PsyghoL, 1956, 2, 203-209.

Fenichel. L. The psychomalytgic theory of neuroses. New

 

York: w. H. Norton, 191:5.

Ferguson, D. 0., & Buss, A. H. Operant conditioning of
hostile verbs in relation to experimenter and subject char-
acteristics. g. consult. Mabel" 1960, _23, 32h-27.
Feshback, S. The drive reducing function of fantasy behavior.

:1. Mom. SOCe PEICh01., 1955, £9" 39.11.

 

 

Feshback, S. The stimulating vs. cathartic effects of a
vicarious aggression activity. i. w. 933. Psyghol"
1961, g}, 381-85.

kaenstein, D. 11., King, S. H., & Drolette, M. The direction
of anger during a laboratory stress-inducing situation-

Psychosom. 593., 1951., $5.: hot-13.

 

 

 

 

31.

32.

33.

35.

36.

37.

38.

39.

to.

’41.

 

Freud, 8., PsychOpathology of everyday life. In The basic

 

writings of; Sigmund Freud. New York: Random House, 1938.

Freud, s. "Repression“ in Collected Papers, Vol. 1;. New

 

York: Basic Books, 1959.

Cluck, M. R. Rorshach context and hostile behavior. _J_.
consult. Psychol" 1955, 12, 1175-780.

Goodstein, L. D. Interrelationships among several measures
of anxiety and hostility. _J_. consult. Psychol., 1951., 18,
35-39. .

Graham, Frances K., Charwat, Wanda A., Honig, Alice 8., &
waltz, Paula C. Aggression as a function of the attack and
the attacker. 51. m. _s_o_g. PsEhol., 1951, 56, 512-20.
Hokanson, J. E., & Gordan, J.’E. The eXpressicn and inhibi-
tion of hostility in imaginative and overt behavior. :1.
M. _s_o_<_':_. Psychol., 1958, EL, 327-33.

Hokmson, J. E. Vascular and psychogalvanic effects of
experimentally aroused anger. g. £11393" 1961, _2_9_, 30-39.
Kagan, J. The measurement of overt aggression from fantasy.
1. abnorm. soc. PsEholn 1956, 53, 390-93.

Kanfer, G. H. Effect of a warning signal preceding a noxious
stimulus on verbal rate and heart rats. J. 352. Psycholq
1958, 52, 73-80.

Kanfer, F. H. Verbal rate, content and adjustment ratings
in experimmtly structured interviews. ,1. m. egg.
PM" 1959, _s_g, 305-11.

Kanfer, F. N. Verbal rate, eyeblink, and content in struc-
tured psychiatric interviews. _J. abnom. soc. 231921.,
1960, _63, 31414.7.

 

 

 

 

 

 

h2.

M.

M.

h5.

h6.

1:7.

148.

149.

50.

51.

.. ._ A 35:1 '13

113

'1 "
I

King, 8. H., 8: Henry, A. F. Aggression and cardiovascular
reactions related to parental control over behavior. _J.
Mag. P321101" 1955, 52, 206-10.

Klopfer, 13., Ainsworth, M. D., Klopfer, W. G., & Holt, R. R.
Developments in the Rorschach technique: _V_o__1_. _l. Technigue
931d M' Yonkers: World Book 00., 19514.

Lantz, Beatric. Some dynamic aspects of success and failure.
Psychol. Monogr" 1915, 52, No. 271.

Lesser, G. S. The relationship between overt and fantasy
aggression as a function of maternal response to aggression.
_J. ahnom. soc. Esychol" 1957, 25, 218-21.

Lindsey, G. , & Goldwyn, R. H. Validity of the Rosenzweig
Picture Frustration Study. _J. £251., 1951;, _2_2_, 519-147.
Mahl, G. F. Disturbances and silences in the patients'
speech in psychotherapy. _J. abnorm. soc. Payer-101., 1956,
15;, 1-15.

 

McCord, N., McCord, Jean, 8: Howard, A. Familial correlates

of aggression in nondelinquent male children. 3. m. 103.
M" 1961, §__2_, 79-93.
Miller, N. bcperimental studies in conflict. In J. Mo V.

Hunt (ed.) Personality and the behavior disorders. New York:
Ronald Press 00., 19131;.

Mussen, P. R., & Naylor, H. K. The relationships between
overt and fantasy aggression. J. abnom. soc. Psycholu
195h, 92, 235-40.

Newcomb, T., Autistic hostility and social reality. 5132.

 

Relate, 19,47, 2;, 69'860

 

 

 

 

..

“‘i’zifi L

n 2"" TE

 

 

S2.

53.

Sh.

SS.

56.

S7.

58.

S9.

60.

61.

11h

Panek. D. R., & Barclay, M. The relationship between GSR

and Speech disturbances in psychotherapy. 3.1 . abnorm. soc.

PsEholu 1959, 5g, hoz-hos.

 

Purcell, K. The TAT and antisocial behavior. ._J_. consult.
Psychol., 1956, 39, 119-56.

Roberts, a. H. & Jessor, R. Authoritarianism, punitivmess,

 

 

and perceived social states. _J. abnom. soc. Psycholu ‘ A I}
‘1': a1
1958, ié, 311.115e I I 1

Rosenbaum, M. E., a Stanners, R. F. Self esteem, manifest

hostility, and eXpression of hostility. g. abnom. soc.

 

 

P32011010, 1961, 92., 6116.50.
Rosenzweig, S. An outline of frustration theory. In J.

McV. Hunt, (Edd, Personality g the behavior disorders.

 

New York: Ronald Press, 19th.

Rosenzweig, S., Fleming, Edith E., & Clarice, Helen J.,

Revised scoring manual for the Rosenzweig Picture-Frustration
study. J. Psychol., 19157, _23, 165-208.

Sarason, I. G. Interrelationships among individual difference
variables, behavior in psychotherapy, and verbal conditioning.
_J_. m. 335. Psychol., 1958, 56, 339-bit.

Sarason, I. G. The effects of anxiety and threat on the
solution of a difficult task. ,1. m. _s_o_<_:_. PsEhol.,1961. §_2_, 165-68.
Sears, R. R., Whiting, J. W. R., Nowles, V., & Sears,

Pauline 8. Some child rearing antecedents of aggression &
dependency in young children. Genet. Psychol. Monogr.,

1953, [41, 135-2314.

 

Sears, R. R., Maccoby, Eleanor E., & Levin, H. Pattems of

child rearing. Evanston: Row, Peterson, 1957.

 

 

 

62.

63.

6h.

65.

67.

68.

69.

7O .

 

1.15

Sears, R. R. Relation of early socialization experiences to
aggression in middle childhood. ._I_. M. £3. Peycholu
1961, _6_3_, 11116-92.

Siegel, S. R., The relationship of hostility to authoritar-
ianism. g. m. 29.30 Psychol., 1956, 53, 368-72.

Siegel, S. 24., Spilka, B., & Miller, L. The direction of
manifest hostility: its measuranent and meaning. Paper pre-
sented at Amer. Psychol. Ass., New York, September, 1957.
Thibaut, J. W. & Coulee, J. The role of communication in the
reduction of interpersonal hostility. J. m. £03.
ream" 1952, 51, 770-77.

Thibaut, J., & Reichen, R. luthoxitarianism, stems,

and the communication of aggression. _H_u£. Re_l_., 1955, _3_,
95-120.

Vendor Linda, L. C. Management of hostility in adult males 7
with migraine headache. Unpublished doctoral disserteion,
Boston miversity, 1955.

Veldnan, D. J. , & Worchel, P. Defensiveness and Self-ac-
ceptance in the management of hostility. _J. M. §_o_c_.
PsyghoL, 1961, 63, 319-25.

Wagner, R. P., a William, J. E. An analysis of speech behavior
in groups differing in achievemmt imagery and defensiveness.
J. 1053., 1961, 22, 1-9.

Weiss, V., & Fine, B. J. The effect of induced aggressiveness
on opinion change. J. M. £93. Psychol., 1956, 53,
109-111;.

Weisskopf-Joelson, Edith, Asher, E. J., & Albrecht, K. J.

 

72.

73.

 

An eXperimental investigation of " label avoidance“ as a
manifestation of repression. J. M. 3332., 1957, .232,
88-93.

Welch, L., Diethelm, 0., 6: Long, L. Measurement of hyper-
associative activity during elation. g. Psycholn 19h6,
23:, 113-26.

Zaks, Misha S., 8; Walters, R. R. First steps in the con-
struction of a scale for the measurement of aggression.

g. Psychol" 1959, 131, 199-208.

 

"In

A

V'

P“- -. “‘11-'ffll‘-‘“- ‘-‘ “...-V
. ' ~p .

u, ......

 

 

 

APPENDIX A

Rosenaweig Picture Frustration Test

SCORE SHEET

Name

uperinmt No.

 

 

 

Date

Carton:
number

Cartoon
number

 

 

1.

ll.

1.

l.

13.

1h.

3.

9.

3.

7.
12.

h.

3.

h.

5.

6.
2h.
21.

5.

h.
22.

7.

7.

19.
15.

8.

9.

10.

20.
23.

9..

10.

1.6.

10.

11.

17.

ll.

18.

12.

S.

12.

2.

 

 

117

 

 

 

 

APPENDIX B

Post-Experimental Ques ti onaire
No .

A. EVALUATION QUESTIONAIRE

In research of this kind, it is necessary to get the subject's
viewpoint and evaluations inorder to get a complete picture of the
meaning of the results. It is recognized that a variety of
extraneous factors, including the nature of the testing situation,
the personality of the miner, and the subject's reactions to
him, add significant dimensions to the subject's performance.

In this questionaire, you will be given an opportunity to express
your own evaluations of a variety of these factors. Be as frank
and honest as possible. Make _a_ check mark (I) anywhere along
each of the scales below if the point that best represents, your
evaluations. -

1. The effect of the testing situation on av perfomance was:

 

1 1
7— 7
Very Very
Favorable Unfavorable

2. The effect of the experimenter's behavior on nur performance was:

 

1— ~71
Very Very
thavorable Favorable

3. W reactions to the experimenter are:

 

t :1
Very Very
Friendly , mfriendly

h. In an? opinion, the value of this eXperiment is:

 

1— .1
Entirely Very
Worthless Valuable

5. If I were given a chance to participate in a follow-up study
with this eXperimenter, I would be:

 

.1
l r
V f Very
ery Ikxwilling
Willing

118

.....__ . , _. .. ”up" 14-:
'T'.-=-e}_mr§”*1_,._s.,.‘~ - ".3...

   

119

B. SELF-RA’I'IE‘IG SCALE

Psychological testing of this type is accompanied by various
subjective feelings and reactions. This rating scale does not
have amr right or wrong answers; you are asked only to report
your feelings and reactions as accurately as possible.

Please make a check mark (X) anywhere along each of the scales
below at the-point that best indicates your typical feelingL
during this experiment.

 

 

 

 

 

 

 

 

 

 

 

 

 

Low High
cor-mama # - u 1%
IRRITATED r1: ~ ~ —%
messy 1— _TL
SATISFIED 7% i T‘
ANGRY 7% “ r‘
cor-(Poem) rl— ”r M “6
DIS TRESSED {—— . - w 4
ANIJOYEZD 1L~ -_ v Jr
. __ J
VUINERABLE 1r "r
1 __1
CAI.“ T— ‘— T
g v i l
mama-tens .— v" A T
LOU High

awmhe -

777* ...—- -mwwu “‘1'"er \ ~ ‘ - '. ‘4 . .'

 

 

D"
.

 

 

 

 

 

 

 

 

Egg.

 

 

 

 

 

 

 

      

RIEs

muMilli“‘