Title Page
 Table of Contents
 List of Tables
 List of Figures
 Biographical sketch

Title: Three responses to frustration and their effects upon subsequent aggressiveness
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00097747/00001
 Material Information
Title: Three responses to frustration and their effects upon subsequent aggressiveness
Physical Description: 1 online resource (ix, 65 leaves) : ill. ;
Language: English
Creator: Wells, Anastasia E., 1933-
Publisher: s.n.
Place of Publication: Gainesville FL
Publication Date: 1970
Copyright Date: 1970
Subject: Frustration   ( lcsh )
Aggressiveness   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 62-64.
Statement of Responsibility: Anastasia Wells.
General Note: Manuscript copy.
General Note: Vita.
General Note: Description based on print version record.
 Record Information
Bibliographic ID: UF00097747
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 004744233
oclc - 465403368


This item has the following downloads:

PDF ( 3 MBs ) ( PDF )

Table of Contents
    Title Page
        Page i
        Page i-a
        Page ii
        Page iii
    Table of Contents
        Page iv
    List of Tables
        Page v
    List of Figures
        Page vi
        Page vii
        Page viii
        Page ix
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
    Biographical sketch
        Page 65
        Page 66
Full Text








The vwiter ,ioulc li,:e to expreLsc her appreciation

to all those vwho contriljbte.d t.! the preparation of thii

dissc rtatioro.

Thaxnis arc d n to Dr. ,.yro CunniJnihai, iDr. lugh

Davis, D)r. hca-ry Gr.atr, Dr. Stephen i. LsirJo'Lis, an Dr.

l',-iacelaine Rory who served as n-embers of the author's su--

pervisory coo-i'tsec. r. CiOuigijhasni anrd Dr. l'avjs ;ce

cspciA:lly ec]pful in supplying support ascn direction witii

the t; ~'. -argulijs not only assisteCd in j.-.fiijL the

problem. bub ri o offered encouragement to test it out.

Dr. _raney va -relied upon for statistical --.'istance, and

also contributed with her time anti interest whenever it

nas necceanary.

'o Dr. C:.-j:'?y Gr.ter, chairman of the supervisory

comimuittee, the -'.nhor wishrs to eCpress a eep ..ppreciation.

Dr. GrCter suipplJicO hor with suppob aid guisdance not only

on this disser'"ation tut throughout her entire L.raduate


A Ujfc-L -ote of thn-ots is also a.d'ed to Dr. *oarrcIn

ic -'1, .r. ro.rt nu;. D': .,-arvin ;h a for their aCLcid --

ance- 'icth so:' of the. equipmct. App:rciation is also

'iven to .th 5 outho-"cL hunbrnc', -ti.ur, without whose co.-.

ope ration :at 1.one sti s-I- sy coci not: c.ve been completeC.

, i IJ >^ L G ,Jj ..



,.............. ........................ iii

LIST OF TABLES ...................................... v

LIST OF PTIGL ES ..... ... .................... ..... vi

ABSTRACT ............................................ vii


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

II METHOD ...................................... 24

III RESULTS ..................................... 33

IV DISCUSSION .................................. 47

V SU NJARY ..................................... 56

APPENDICES ................................... ...... 59



REFERENCES .......................................... 62

BIOGRABPHICAL SIKETCH ................................ 65


Table fage

1 Me'.ns and Siandard Deviations of Base
Level Systolic Blood ireasu'e ............ 34

2 Oieans sand S-'.andara reviatio:)r of .AgEreslion
Score Used as a Pr-e--mecsure, o0 Ag re s-
siveness ................................... 34

3 I;eans of Systolic 3lood Press.,ce Jithir
rl Groups and Condiitions ............... ..... 36

4 Sum ary Td.b'Le for the Analysis of
Varia nc:e: Systolic Blocd r.ss-rs .......... 37

5 arcns ior Aggression ScoreO for sil Groups .. 42

6 S-un:.tary Table for Analysis of Varc'lace:
Agg'recsion Sc:ores .............. .......... 43


Figures Pages

1 M.ean Systolic Blood Pressure at Three
Ecadings at Each Level of Frustration ......... 38

2 eanr Systolic Blood Pressuresat Three
Readings at Each Level of Response
to Frustration .. ............................ 39

3 Mean Aggression Scores for All Responses
to Frustration on Both Pre-measure and
Post-measure for Both Levels of Frustration ... 44

Abstract of Dissertation Presented to the Graduate Council in
Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy



Anastasia E. Wells

March, 1970

Chairman: Dr. Harry Grater
Major Department: Psychology

Hokanson and his colleagues (1961, 1962a, 1962b,

1963, 1966) were among the few investigators that looked

into the physiology of catharsis. This paper is essen-

tially a replication of Hokanson's major findings as well

as an extension of his work. Hokanson and his colleagues

found that experimentally produced frustration consist-

ently resulted in increased physiological arousal. While

several indices were tested, systolic blood pressure was

found to be the most reliable indicator of this arousal.

In addition, they found that the direct physical and verbal

expression of aggression towards the frustrator consist-

ently resulted in a return of systolic elevations to pre-

frustration levels. Aggression following frustration was

found to be tension-reducing. Those individuals who were

frustrated and not allowed to aggress were found to main-

tain their systolic elevations.

This paper investigated three responses to frustra-

tion: (1) counteraggression against a frustrator; (2) hav-

ing no opportunity to aggress following frustration; (3)

the discussion of frustration. The efficacy of these re-

sponses in reducing tension was studied. Following this,

the overall effect of these three responses to frustration

upon subsequent aggressiveness was determined. It was pre-

dicted that counteraggression and discussion of frustration

will be found to be equally effective in reducing frustra-

tion-produced tensions. However, it was hypothesized that

the counteraggression will result in an increase in subse-

quent aggressiveness, while the discussion will be followed

by no change in aggressiveness.

Seventy-two male college age subjects were randomly

assigned to two separate three factor designs. Three inde-

pendent variables were employed: (1) Frustration vs. no frus-

tration; (2) Three responses to frustration--counteraggres-

sion, no aggression, and discussion of frustration; and (3)

Repeated measures of (a) systolic blood pressure taken before

and after frustration and after the response to frustration,

and (b) aggression scores taken before and after the experi-

mental procedures. The systolic blood pressure readings and

the aggression scores served as the primary dependent var-

iables. Essentially the same procedure employed by Hokanson

was used in this study, with the addition of the discussion

condition and the pre- and post-measures of aggression.


Two three factor analyses of variance were carried

out after which several post hoc comparisons were made on

certain pairwise cell means. The results obtained supported

all the hypotheses. Hokanson's major findings wero repli-

cated. In addition, the discussion of frustration was found

to be as effective as counteraggression in reducing frustra-

tion-produced tensions. Insofar as subsequent aggressive-

ness was concerned, counteraggression was found to increase

it, while the discussion of frustration produced no change

in it. The discussion of frustration appears to be an ef-

ficient mechanism for handling frustration. Several alter-

native hypotheses and limitations of the study were offered

which supply a direction for future research in this area.



The chief aim of this study is to investigate the

effects that certain responses to frustration have upon sub-

sequent aggression. The paper focuses on locating an effec-

tive method for decreasing the potential for aggression.

When Dollard, Doob, Miller, IMowrer and Sears (1939) in their

classic book, Frustration and Agrression, re-examined the

Freudian notion of catharsis, they paved the way for an abun-

dance of research in this area. According to the catharsis

hypothesis postulated by them, Dollard and his colleagues

(1939, p. 50) insist that the expression of aggression re-

sults in at least a momentary reduction of subsequent aggres-

sion. Aggression was viewed as an appropriate response to

frustration. It was said to relieve frustration as well as

to produce a reduction in subsequent aggression. Many stud-

ies support this hypothesis (Thibaut, 1950; Thibaut & Coulos,

1952; worchel, 1957; Pepitone & Reichling, 1955; Feshbach,

1955, 1961, 1965; Rosenbaum 6 deCharms, 1960). However,

still other studies refute the hypothesis that subsequent

aggression is reduced (kenny, 1953; Siegel, 1956; LOvaas,

196]; Msussen & Rutherford, 1961; Bandura & Huston, 1961;

Bandura, Ross, & Ross, 1963). In fact, the latter studies


show consistently that subsequent aggression is increased

rather than reduced following aggressive behavior.

Until researchers began to look into the physiology

of catharsis (kahn, 1960; Hokanson, 1961; Baker, 1967, 1969;

Gambaro & Rabin, 1969), a stalemate existed with catharsis

research. During recent years, hokanson and his colleagues

have performed a series of studies investigating the nature

of the cathartic effect. A review of these studies will

follow, but for the moment a comment will be made concern--

ing their findings. Hokanson and his colleagues found that

experimentally produced frustration consistently resulted

in increased physiological arousal. wIhile several indices

were tested, systolic blood pressure was found to be the

most reliable indicator of this arousal. In addition, they

found that the direct physical and verbal expression of

aggression toward the frustrator consistently resulted in

a return of the systolic elevations to pre-frustration or

base level. C4ggression following frustration was found to

be tension-reducin- Those individuals who were frustrated

and not allowed to aggress were found to maintain their

systolic elevations.

The contribution that Hokanson's work makes to ca-

tharsis research is that it permits us to predict a physio-

logical effect that can be measured following certain frus-

tration-aggression sequences. The tension-reduction aspect

of counteraggression against a frustrator has certainly

found support in Hokanson's investigations. However, whether

or not (1) there exists non-aggressive method for dealing

with frustration that is equally tension-reducing; and (2)

whether or not subsequent aggression is actually decreased

or increased as a result of this tension-reduction remains

to be investigated. These are the areas of exploration for

this dissertation.

Considering the overall effect of tension-reduction

upon subsequent aggression, traditional catharsis researchers

would predict a decrease in subsequent aggression. They

would hypothesize that following frustration and counter-

aggression, aggressive energy is drained, drive is reduced,

and aggressive behavior decreases. However, a more parsimo-

nious prediction, based on learning theory, would be that

following the frustration-aggression sequence, subsequent

aggression increases. Considering the reinforcement value

of tension--reduction, the probability is high that aggression

will increase. This is one of the hypotheses to be tested

in this paper. If support can be obtained for an increase

in aggression following tension-reduction from frustration

and countcraggression, it can have provocative ramifications

for dealing with frustration. The hypothesis suggests that

if one is encouraged to respond to frustration by verbal or

physical aggression, one becomes more aggressive subsequently.

In order to reduce the potential for subsequent aggression,

then, a non-aggressive method for dealing with frustration

may have to be employed. From clinical observation, it is

known that the simple clarification and expression of feel-

ings often bring relief from emotional tensions. Clinical

experience, therefore, suggests that the discussion of frus-

tration may be an effective non-aggressive response to frus-


This paper investigates three responses to frustra-

tion: (1) counteraggression against a frustrator; (2) hav-

ing no opportunity to aggress following frustration; (3)

discussion of the frustration. The efficacy of these re-

sponses in reducing tension will be studied. Following

this, the overall effect of these three responses to frus-

tration upon subsequent aggressiveness will be determined.

It is predicted that counteraggression and discussion of

frustration will be found to be equally effective in reduc-

ing frustration-produced tensions. However, it is predicted

that the counteraggression will result in an increase in

subsequent aggressiveness while the discussion will be fol-

lowed by no change in aggressiveness.

Review of Literature

Hokanson's work. Hokanson (1961) studied the ef-

fects that (a) level of frustration, and (b) anxiety over

punishment had on overt aggression. His study provided a

more direct test of some of the classical frustration and

aggression hypotheses as they applied to overt, physical

aggression. Typically, "overt" aggression in catharsis

research was measured by means of questionnaires, ratings,

interviews concerning the feelings of unfriendliness toward

the frustrator, and a number of other indices representing

displaced aggression, e.g., doll play and TAT content. Such

questionnaires and ratings are vulnerable to the subject's

conscious control over his responses in a socially desirable

direction. Furthermore, unfavorable ratings of the experi-

menter, or aggressive TAT content, may express merely nega-

tive evaluations of the frustratcr, and not "aggression" in

the strict sense required by the specification (Dollard

et al., 1939) that the goal response of the instigation to

aggression involves inflicting injury to an organism.

Hokanson attempted to overcome these difficulties by using

the application of direct, physical pain to the frustrator

(electric shocks) as measures of aggression. Furthermore,

his equipment enabled him to look at the aggressive response

along three dimensions: number, duration, and pressure of


Hokanson's (1961) study employed an orthogonal design,

using 80 undergraduate males, varying (a) test hostility;

(b) level of experimental frustration; and (c) level of

'retaliatory anxiety. The test hostility was determined

by the subjects' responses to Siegel's (1956) i;anifest Hos-

tility Scale and to three TAT protocols vhich were scored

for hostile content. Basically, subjects who consistently

scored above the median on these rojeaures were assigned to

the High test hostility group, while those who scored below

were assigned to the Low test hostility group. The frustra-

tion manipulation was established by asking the subjects to

count backwards from 100 to 1 by three's as quickly as pos-

sible. Subjects assigned to the High frustration condition

were then exposed to repeated interruptions, insults, and

sarcastic remarks by the experimenter. In the Low frustration

condition, the experimenter omitted these remarks and allowed

the subjects to reach zero. The retaliatory anxiety variable

was manipulated by leading half of the subjects to believe

that the experimenter had the potential for inflicting phys-

ical pain to them via electric shocks if they did not co-

operate fully during the experiment. Following these experi-

mental manipulations, the subjects were given a number of

socially sanctioned opportunities to aggress against the ex-

perimenter by administering electric shocks to him. The sub-

jects' behavior with respect to the plunger activating the

shock was analyzed along three dimensions: number of shocks,

average duration per shock, and the mean pressure exerted.

In addition to all these data, Hokanson included in his study

several physiological measures. He reported on systolic

blood pressure readings recorded at several critical points

during the experiment: (1) after an adaptation period; (2)

after the anxiety manipulation; (3) after frustration; (4)

during aggression, and (5) after aggression. He believed

that changes in systolic blood pressure between any

of these levels might provide informative correlates of

aggressive behavior.

Among his major findings were: (a) that both the

number of shocks and the pressure were found to be equally

sensitive indicators of frustration-produced instigation to

overt aggression; (b) that the High frustration condition

servedto increase the instigation to aggression; (c) that

the High retaliatory anxiety condition served to increase

the instigation to aggression; (d) that among combinations

of High frustration and High retaliatory anxiety conditions,

a significant negative correlation was found between systolic

elevations (during frustration) and the increase in the in-

stigation to aggression; and, finally, (n) that a negative

correlation existed between vigor of aggression following

frustration and post-aggression elevation in systolic blood

pressure. The more aggressive the subjects were, the less

elevated their blood pressure following the expression of

aggression. This last finding interested Hokanson, and he

explored this area further in the following studies.

Later, Hokanson and Shetler (1961) attempted to test

directly the hypothesis that the expression of aggression,

after a frustration, produces greater reduction in physio-

logical arousal (systolic elevations) than having no oppor-

tunity to aggress. Hokanson believed that this was a more

feasible question to test than the traditional question

emanating from the catharsis hypothesis as to whether or not

the expression of aggression leads to a reduction in aggres-

sive behavior. The central problem in the study of this

catharsis question has been the inability of investigators

(Berkovitz, 1958) to "demonstrate clearly that the decrease

in [post-aggression] hostile behavior is due to drive reduc--

tion and not to response inhibition [p. 274]." Hokanson

and Shetler addressed themselves to the first step of deter-

mining whether or not tension-reduction occurred. In their

study, they also varied the status of the frustrator to

assess any effects this may have upon tension-reduction.

Fifty-six undergraduate males were exposed to either

high vs. low frustration by a high or low status experimenter

with a subsequent opportunity vs. no opportunity to aggress

physically (via electric shocks) towards the frustrator.

Systolic blood pressure was measured before and after the

frustration manipulation, and after the expression of aggres-

sion. The frustration manipulation was performed in exactly

the same way as in the preceding study. Subjects were asked

to count backwards from 100 to 1 by three's and differential

treatment was given to high and low frustrated groups. Im-

mediately following the frustration manipulation, half of

the subjects were given an opportunity to aggress physically

against the frustrator by administering shock to him. This

situation w'as structured by telling the subject that the

task involved an interpersonal guessing game. The subject

was to think of a number between one and ten and the exper-

imenter was to guess the number. If the experimenter guessed

incorrectly, the subject was to signal this error by admin-

istering shock to him. The experimenter was purported to

be studying the effect of pain upon subsequent guessing.

Subjects in the no opportunity to aggress condition went

through the same procedure except that they signalled errors

to the experimenter by flashing an electric light rather

than using shock.

Among their findings were: (a) frustration led to

significally greater systolic increases than the Low frus-

tration control condition with both High and Low status

experimenter; (b) subjects frustrated by the Low status

experimenter and given an opportunity to aggress against

him manifested a return of blood pressure to pre-frustration

levels which was not significantly different from that of

non-frustrated subjects; whereas, (c) subjects frustrated

by the Low status experimenter and given no opportunity to

aggress against him manifested significantly greater systolic

elevations at the conclusion of the experiment than either

the frustrated-opportunity to aggress subjects or the Low

frustrated subjects; (d) subjects frustrated by the High

status experimenter manifested a return of blood pressure

to pre-frustration levels which was not significantly dif-

ferent from those of non-frustrated subjects in both the

opportunity to aggress and the no opportunity to aggress

conditions. The authors discuss their results as offering

support for the hypothesis that under certain conditions

overt aggression has tension-reducing qualities.

In extending the work in this area, Hokanson and

Burgess (1962a) systematically investigated the type of ag-

gression. Systolic blood pressure and heart rate were meas-

ured before and after frustration and after each of three

aggression conditions: physical, verbal, and fantasy.

There was a no aggression control and a no frustration con-

trol. Eighty college age subjects were introduced to the

experiment as one involving physiological responses to per-

formance on intellectual tasks. To aid in establishing this

deception, subjects were administered the Picture Completion

subtest of the WAIS following an initial 8-minute adaptation

period. During the second task, the frustration conditions

were introduced. The counting technique, employed in the

previous studies, was used again. Immediately following the

frustration procedures, the aggression phase of the experi-

ment took place. The 20 subjects in the Physical Aggression

condition were instructed along the same lines as the pre-

vious study. They were informed that the task involved an

interpersonal guessing problem. Electric shock was to be

administered to the experimenter if he guessed incorrectly.

The 20 subjects in the No Aggression condition went through

exactly the same procedure but with the exception that a

light and not shock signalled the incorrect guesses to the

experimenter. The Verbal Aggression condition consisted of

having the subject fill out a brief questionnaire evaluating

the experimenter's capabilities as an experimenter, in the

experimenter's presence. Finally, the 20 subjects in the

Fantasy Aggression condition were asked to create a story

to Card 8T, of the TAT. Heart rate was measured continuously

and systolic blood pressure at about 2-minute intervals over

the entire procedure, with the critical measurements being

before and after frustration and immediately after the oppor-

tunity to aggress.

Their results indicated: (a) that frustrated sub-

jects who were given an opportunity to physically or ver-

bally aggress against the frustrator manifested returns on

both physiological measures to the levels of the Low frus-

tration control subjects; (b) that frustrated subjects in

the Fantasy and No Aggression conditions exhibited signif-

icantly elevated systolic pressures and heart rates at the

post-aggression recording of these measures. The results

were discussed by the researchers as offering support for

the hypothesis that the expression of Physical or Verbal

Aggression towards a frustrator tends to reduce general,

physical arousal. They cautioned against concluding that

their findings reflect a reduction in aggressive drive.

They only found a reduction of tension and did -' ol measure

the effect this had upon subsequent aggression.

Hokanson and Burgess (1962b) investigated the frus-

tration and status variables. Eighty--four college age sub-

jects of both sexes were placed in one of three frustration

conditions: ego threat, blocked goal, or no frustration.

They were subsequently given either an opportunity to


verbally aggress (via a questionnaire) or no opportunity to

aggress against the frustrator. A third orthogonal variable

investigated was the status of the frustrator, with half of

the subjects seen by an experimenter who introduced himself

as a new faculty member (high status), while the remaining

subjects were seen by a psychology student (low status).

Systolic blood pressure and heart rate were measured before

and after the frustration manipulation and immediately after

the aggression phase of the experiment. Subjects in the Ego

Threat frustration condition were asked to count backwards

from 100 to 1 by two's as quickly as possible. Following a

standardized procedure during this counting, the experimenter

repeatedly interrupted and harassed the subject for counting

too slowly, having the subject start over three times. After

one minute, the experimenter asked that the counting be

stopped, stating that the data could not be used owing to

the subject's uncooperative attitude. The subjects in the

Goal Blocking condition were also asked to count backwards

from 100 to 1 by two's with the admonition that maximum ef-

fort and speed were important in this part of the experiment.

The subjects were further instructed that since speed was so

critical, this project had received funds with which to

pay the subjects $4.00 if they completed the counting within

a specified time limit. At the completion of the counting,

the subject was told that while he had finished within

the time limit, it was the experimenter's impression that

he had not devoted maximum effort to the task, and there-

fore he would not receive the $4.00. The subjects

in the No frustration condition were allowed to count back-

wards to zero without interruption.

The results which were consistent on both physio-

logical measures were: (a) both Ego Threat and Blocked Goal

frustrations produced a significant increase in systolic

blood pressure and heart rate relative to the No frustra-

tion control group; (b) for both types of frustration, hav-

ing a post-frustration opportunity to verbally aggress

against the Low status frustrator results in a return of

vascular processes to levels not significantly different

from those of a No frustration control group; whereas, hav-

ing no opportunity to aggress with a Low status frustrator,

results in maintenance of the frustration-produced eleva-

tions; (c) with a High status frustrator, opportunity to

aggress, on the whole, did not substantially reduce vascular

elevations relative to comparable opportunity to aggress

with a Low status frustrator; (d) subjects who were in one

of the frustration conditions or were seen by the High sta-

tus experimenter had significantly elevated vascular proc-

esses at the close of the experiment when compared to sub-

jects in the No frustration or Low status experimenter con-

ditions. In this study, a comparison of the verbal aggres-

sion questionnaires between subjects frustrated by the High

and Low experimenter indicated no significant difference in

the amount of aggression expressed. Thus, differences in

overt expression of aggression can be ruled out as a factor

in the differential vascular reactions found in this research.

The purpose of Hokanson's (1963) next study with

Burgess and Cohen was to test the hypothesis that displaced

aggression also results in a degree of arousal reduction

roughly proportional to the similarity of the substitute

target to the original frustrator. A related hypothesis

tested was that the intensity of the aggressive response

will likewise be proportional to the similarity of the sub-

stitute target and frustrator.

Eighty college age subjects were placed in a 2 by 5

factorial design experiment involving High or Low frustration

experience followed by an opportunity to physically aggress

towards targets varying in similarity to the frustrator.

lost-aggression systolic blood pressure elevation relative

to pre-frustration base level was the primary dependent meas-

ure. Essentially the same procedure was carried out as in

the preceding study for the adaptation period, the frustra-

tion manipulation, and the opportunity to aggress. The only

exception was that the opportunity to aggress was divided

into five groups: (1) no opportunity to aggress; (2) aggress

against the frustrator; (3) aggress against the frustrator's

assistant; (4) aggress against a psychology student; (5)

aggress against an undergraduate student. Aside from the

systolic measurements, a record was kept of the pressure

exerted and the number of electric shocks given.

Their results indicated: (a) that among frustrated

subjects, the expression of aggression directly to the

frustrator resulted in a significant degree of arousal redac-

tion relative to a no aggression control; whereas, (b) the

same amount of aggression expressed to substitute targets

did not significantly reduce blood pressure. The authors

believed that this raised a question concerning the general-

ity of the traditional catharsis hypothesis; i.e., that the

expression of aggression, in and of itself, is tension-re-

ducing. They felt that this conclusion was underscored by

these additional findings: among frustrated subjects, there

was no significant difference between the amount of aggres-

sion they expressed regardless of the object; however, the

arousal reduction took place only when the object of the

aggression was the original fruLstrator. The authors con-

clude by stating that their results have implications for

a theory of interpersonal aggressive behavior compatible

with proposals by Berkowitz (19o2). Following an anger

provocation a general physiological arousal is produced,

which most likely is associated with a behavioral disposi-

tion to aggressive behavior. VWhen given an opportunity,

the disposition to aggression will be completed against an

available target; however, concomitant reduction of physio-

logical arousal will take place only when the aggressive

behavior in some way is perceived to affect the original

frustrator and is deemed appropriate to the situation.

Hokanson continued to investigate the conditions

under which this arousal and rccuction phenomenon occurred.

In a study .with Edelman (196b), he changed the manner in

which frustration was manipulated and offered subjects not

only an opportunity to aggress or not to aggress but also an

opportunity to be friendly or to ignore the frustrator. He

also looked at sex differences among subjects. The frustra-

tion manipulation was accomplished by having the subject

receive several noxious shocks from a confederate posing as

a fellow subject. The subject was allowed to (1) counter-

aggress (2) not to aggress,(3) reward the frustrator, or (4)

ignore the frustrator. The results revealed that: (a) the

noxious shocks by a fellow student produced systolic eleva-

tions; (b) that the counteraggression opportunity was fol-

lowed by a rapid return of systolic elevations to pre-frus-

tration levels; (c) friendly or counterresponses were fol-

lowed by a relatively slow return to base line comparable to

that of controls who were given no opportunity to respond;

(d) that the above results were obtained with systolic blood

pressure and a vasomotor response but not with diastolic

blood pressure or heart rate, and (e) that these results

were obtained on male subjects only. Females showed no dif-

ferential recovery rates. Hokanson concludes his discussion

by reviewing all the studies he has carried on so far and

stating that there are a variety of circumstances under

which the arousal and reduction phenomenon does not occur:

with a High status frustrator (Hokanson and Burgess, 1962a);

with displaced aggression towards a person unrelated to the

frustrator (Hokanson et al., 1963); with fantasy aggression

(Hokanson and Burgess, 1962b); and among college age females

(Hlokanson and Edelman, 1966). He attempted to place these

results under a learning theory framework. In each instance

where the phenomenon did not occur, he believes it can be

assumed that the subject had previously learned that under

these particular social conditions an aggressive counter-

response to a provocation was not an appropriate behavior.

That is, aggression will not bring the interpersonal exchange

to a rewarding conclusion and thus the elevated vascular

processes are maintained. Similarly, he maintains that the

subject has learned that, under certain other condiTiions, an

aggressive response is instrumental in terminating noxious

social stimulation, thereby also being associated with a

relatively rapid reduction of autonomic processes. Hence,

Hokanson believes that his series of studies has in effect

been identifying the complex discriminative stimuli under

which aggressive behaviors have or have not been reinforced

in our culture.

As indirect evidence for his theorizing, he discussed

an unpublished study in which fantasy aggression toward a

High status frustrator produced the arousal reduction effect

just as direct aggression toward a Low status frustrator

produces it. He presumed that for those subjects direct

aggression against a High status person had not proven to

be reinforcing in the past, whereas covert aggression had.

Similarly, fantasy aggression toward a Low status frustrator

had not been as successful as direct verbal or physical

aggression. Hokanson believes the failure to obtain the

arousal reduction effect with females in his study with Edel-

man (1966) is even more comprehensible. He claims that it

is virtually a truism that middle class females in our cul-

ture receive little reward or training with respect to phys-

ical aggression; that is, it has not become instrumental in

terminating noxious stimulation. Under these circumstances,

it is not surprising to him that female subjects in his

study, even when moderately reinforced to utilize an aggres-

sive counterresponse, manifested no associated physiological


Summary of Hokanson's work. in summarizing the work

that Hokanson and his colleagues have performed, among his

major findings are: (a) that experimentally produced frus-

tration produces increases in systolic blood pressure; (b)

that under certain conditions, opportunity for overt phys-

ical and verbal counteraggression towards the frustrator

reduces systolic elevations to pre-frustration levels. The

arousal and reduction phenomenon that they have repeatedly

found strongly suggests that the frustration-aggression

sequence has a tension-reduction effect that can be reliably

measure What they have not investigated is: (1) what

effect this tension-reduction has upon subsequent aggres-

sion, and (2) whether or not a non-aggressive response to

frustration exists that is equally effective as aggression

in reducing tension.

Discussing frustration as a counterresponse to frus-

tration. There is a trend that exists among some psychother-

apists today that the clarification of'and verbal expression

of feelings assist their clients in obtaining relief from

anxieties as well as in establishing an integrated self.

This trend emanated primarily from the work of Carl Rogers

and his associates. In his early work, Rogers (1965) estal-

lished an elaborate personality theory as well as developed

a psychotherapeutic approach often referred to as client-

centered therapy. Briefly stated, in client-centered ther-

apy, the therapist assumes a non-directive role and merely

reflects back to his client any feelings or thoughts that he

believes the client is trying to express. In this way, the

therapist attempts to assist his clients to become more

aware of their feelings and experiences which, according to

Rogers, helps the clients to integrate themselves. Rogers

(1957) writes of his new approach:

For the client, this optimal therapy has meant an
exploration of increasingly strange and unknown
and dangerous feelings in himself; . Thus, he
becomes acquainted with elements of his experience
which have in the past been denied or awareness as
too threatening, too damaging to the structure of
self. He finds himself experiencing these feelings
fully, completely, in the relationship, so that for
the moment he is his fear, or his anger, or his
tenderness, or his strength. And as he lives these
widely varied feelings, in all their degrees of
intensity, he discovered that he has experienced
himself, that he is all these feelings. lie finds
his behavior changing in constructive fashion in
accordance with this newly experienced self [p. 487].

Not only does Rogers see value in expressing one's feelings,

but he also suggests that once they are expressed, more

constructive behavior may follow.

Walker, iRblcn, and Rogers (1960) developed a scale

to measure process changes in psychotherapy. One of their

seven categories was the expression of feelings. They found

that unrecognized and unexpressed feelings were identified

with poor progress in therapy while feelings fully experienced

and expressed were identified with excellent progress in ther-


Additionally, Rogers' writings suggest that if feel-

ings are not expressed a situation arises that is a basis

for psychological tension. Rogers (1965) states: "Psycho-

logical maladjustments exist when the organism denies to

awareness significant sensory and visceral experiences,

which consequently are not symbolized and organized into

the gestalt of the self-structure. When this situation

exists, there is a basis or potential psychological tension

[p. 476]." This suggests that some Lype of tension might
exist if an individual was frustrated and not able to ex-

press this frustration] In the present study, both counter-

aggression and the discussion of frustration are viewed as

couniterrosponses to i.'rustration that permit the feeling to

be expressed and, consequently, relieve tension. On the

other hand, having no opportunity to aggress or discuss the

frustration creates a situation in which feeling-, could be

denied; hence, a condition of tension will persist. The

overall implication for this study is that if feelings of

frustration can be identified and verbally expressed, as

they are in client-centered therapy, then perhaps this can

lead to more constructive behavior in dealing with the frus-

tration. This study investigates (1) whether or not discus-

sion of frustration is an effective way of handling frustra-

tion; and (2) if it is effective, does this method lead to

more constructive behavior which, in this case, is identified

as a decrease in subsequent aggression.

Increase in subsequent aggressiveness. Some studies,

cited earlier, suggest that subsequent aggression is increased

rather than decreased after one engages in aggression. Berko-

witz (1962) has suggested several situational conditions

that could account for the increase in aggression. He be-

lieves that such conditions as social facilitation or permis-

siveness contribute in weakening one's inhibition to aggress

against another person (i.e., Siegel, 1956). In addition to

these situational conditions, Berkowitz believes that inves-

tigators have not always differentiated between counterag-

gression to frustration and an individual's habitual level

of aggressiveness. Research guided by the Freudian energy

model does not need to make such a distinction; they gener-

ally assume that every time a person acts aggressively, re-

gardless of the circumstances, he does so because of his

aggressive "energy' (Hartmann et al., 1949). Berkowitz

contends, however, that counteraggression to frustration must

be differentiated from one's habitual level of aggressiveness.

He maintains that aggressiveness is a learned habit and,

like any other habit, is readily occasioned by relevant cues.

The implication of such a formulation is obvious: "Provid-

ing an opportunity to express hostility may lessen the frus-

tration-engendered instigation to aggression, but could also

evoke and/or strengthen a person's habitual hostile tenden-

cies [Berkowitz, 1962, p. 203].1 The present study tests

this statement. The design includes pre- and post-measures

of an individual's aggressiveness. Interpolated between

these measures are several conditions, one of which involves

a frustration-countcraggression sequence. Following Berko-

witz, it is predicted that following frustration-counterag-

gression, there will be a relative increase in subsequent


Hyopothe ses
The first three hypotheses replicate Hokanson's

major findings regarding frustration and aggression:

Hypothesis I: Frustration produces elevations in
systolic blood pressure.

Hypothesis II: Following frustration, counterag-

gression against a frustrator produces a return of systolic

elevations to pre--frustration levels; whereas,

iHgothesis III: Systolic elevations are maintained

following frustration if there is no opportunity to aggress

against the frustrator.

Based on clinical evidence adduced by Carl Rogers

and his associates, the following two hypotheses seem


Hypothesis IV: Following frustration, having an

opportunity to discuss the frustration produces a return

of systolic elevation to pre-frustration levels.

Hypothesis V: Frustration-reduction achieved

through the discussion of frustration results in no change

in aggressiveness relative to base levels.

Following Berkowitz that a frustration-aggression

sequence could result in an increase in one's level of

habitual aggressiveness, the following hypothesis appears


HIypothesis VI: Frustration-reduction achieved through

the frustration-counteraggression sequence results in an in-

crease in subsequent aggressiveness relative to base levels.




Seventy-four undergraduate students at the University

of Florida selected from a pool of students taking an intro-

ductory or intermediate course in psycholoLg served as sub-

jects. The first two subjects had to be discarded since the

electric stimulator was not properly set up during their

trials. Students were fulfilliJng course require-lents to

participate in psychology experiments.


(1) Digit symbol subtest of the ~AIS

(2) Mercury sphygmomanometer and stethoscope to

measure systolic blood pressure.

(3) Hunter interval timer which measures duration

of time in hundreds of a second.

(4) Electric stimulator which had electrodes for

receiving shock, a plunger for activating shock,

and a dial for setting the intensity of shock

at intervals of 1 to 10 millia peres.

elxperimpntal De si{fn

Three independent variables were employed in the


(1) frustration vs. no frustration, (2) three couinterre-

sponses to frustration which were counteraggression, discus-

sion of frustration, and having no opportunity to aggress,

(3) three periods of systolic readings which were before

and after frustration and after the counterresponse to frus-

tration; and two periods of aggression measures which were

pre- and post-experiment.

Two sets of data were the primary dependent variables:

(1) systolic blood pressure readings taken at three critical

periods--before and after frustration and after the counter-

response to frustration; (2) measures of aggression taken

before and after the experimental procedures.

A three factor analysis of variance was used to test

for differences in systolic blood pressure that may have oc-

curred among the various experimental conditions. The ex--

perimcntal conditions generated the following design:

(Repeated TLeasures)
Responses Subjects Base Post-Frus- Post-Responses
to Frus- Level tration to Frustration



aggression 12 "
No Aggression 12 I
Discussion 12 "

aggre ( sion 12 "
No AC'r-rescion 12 I
Discussion 12 "

N = 72

Another three factor analysis of variance was used

to test for differences in aggression score that may have

occurred among the various experimental conditions. The ex-

perimental conditions generated the following design:

(Repeated Measures)
Responses to
Frustration Subjects Pre-heasure Post-L-hasure

aggression 12 "
trated No Aggression 12 "
Discussion 12 "


aggression 12 "

No Aggression 12 "

Discussion 12 "

N = 72

Following the two analyses, several one-tail t tests were

performed to test the significance of the differences between

certain cell means. In order to test hypotheses I and III,

two individual t tests were made cn two pairwise cell mean

comparisons from the data in the first analysis. To test

hypothesis VI, one t test was made on data from the second

analysis. The justification for employing multiple t tests

to test differences between cell means is based on the fact

that these mean comparisons were planned in advance of the

analyses and that they are orthoecnal. Hypotheses II, IV,

and V did not require any statistical tests since the differ-

ences of the pairwise cell means in question were equal to

less than zero.

In addition to the t tests that were employed to

test the hypotheses, two Tukey post hoc comparisons were

made on two additional pairwise cell mean comparisons.


Six undergraduate students, sirrilar in age and sta-

tus to the subjects, served as experimental assistants and


The experiment was introduced to the subject as one

involving blood pressure response to working on five routire

intellectual tasks. The subject was led to believe that he

would determine what five tasks he performed by selecting

five sealed envelopes from approximately one hundred con-

taining instructions for a variety of tasks. The selection

of tasks for each subject was actually fixed so as to-ran-

domly assign subjects to one of six conditions: (1) frus-

trated-counteraggression; (2) frustrated-no opportunity to

aggress; (3) frustrated-discussion; (4) not frustrated-

counteraggression; (5) not frustrated-no opportunity to ag-

gress; and (6) not frustrated-discussion.

Techniques empnloyd for obtaining dependent meas-

ures. The techniques for taking and recording the blood

pressure measures as well as the aggression scores were

standard for all subjects. The blood pressure readings

were made in the following manner: upon entering the exper-

imental room, the assistant placed the cuff on the left arm

of the subject 'and it remained there throughout the session.

Whenever a reading was required, the assistant inflated the

cuff cutting off the flow of blood in the arm momentarily.

The stethoscope was then placed over the vein directly below

the inflated cuff. As the cuff ,was deflated, the pressure

required to force the blood flow back through the arm reg-

istered on the mercury sphyngomanometer. The point at which

the blood flow returns is heard as a throb through the steth-

oscope and is designated as the systolic blood pressure.

Several readings were taken during the session by the assist-

ant with critical measures before and after frustration and

after the opportunity to respond to the frustration.

The e, : .., scores were recorded in the following

manner: the subject selected a shock intensity from one to

ten milliamperes by setting the dial on the electric stiru-

lator. Once the subject set the dial, he depressed the

plunger to administer the shock. The plunger activated a

Hunter interval timer which measured the duration of shock

in hundreds of a second. The assistant recorded both the

intensity and duration of the shock that the subject admin-

iutercd. The mean score of the intensity x duration of the

shocks that were administered was taken as pre- and post-

measures of the subject's aggressiveness. The rationale

for using this measure of aggression is discussed by

Bandura (1964). He discusses a study which used both (1)

intensity alone, and (2) intensity x duration as measures

of aggression. Under certain conditions, it was found that

although the subjects administered virtually identical in-

tensities of shock, the subjects differed significantly on

the intensity x duration indicator of aggression.

Adaptation period. All subjects selected for their

first task the Digit Symbol subtest of the VIAIS. Subjects

were allow-'ed to perform this task at their o.n pace. The

purpose of this task was to allow a five-minute acclimation

period to the experimental setting.

Pre-n-easure of aggression. Following the adaptation

period, all subjects selected for their second task an inter-

personal guessing game. The task was structured as follows:

subjects were told that this task involved an interpersonal

guessing situation in which the subject was to think of a

number between one and ten, following which another person

was to guess the number. The subject was told that the ex-

perimenter was studying the effect of pain on subsequent

guessing. For the pre-measure of aggression, the person

guessing was a confederate posing as a fellow subject. If

the confederate's guess was wrong, the subject was to signal

this error by administering electric shock to him. Although

the person was wired to an elaborate shock apparatus, he

did not actually receive shock, but merely behaved as if he

did. If the confederate war correct, no shock was administered.

The subject was led to believe that there would be about 15

guessing trials, but the experimental assistant actually

terminated the trials when the subject had administered ten

shocks to the confederate. The confederate was allowed to

guess two correct, therefore, 12 trials in all were run.

The subject was not informed of the nature of milliamperes;

however, the dial on the stimulator was clearly calibrated

and labeled in milliamperes. The subject was informed that

he was free to vary the intensity of the shock from one to

ten, with one being the least noxious and ten being the most

noxious. The subject was further instructed on how to ad--

minister the shook. The mean score of the intensity x

duration of the ten shocks that he administered to the cor--

federate was taken as the pre-measure of the subject's ag-


Frustration manipulation. Following the pre-mieasure

of aggression, all subjects selected for their third task a

counting situation which required them to count backwards

from 100 to 1 by three's as quickly as possible. Frustra-

tion was induced in one half of the subjects by a standard-

ized procedure of having the experimental assistant inter-

rupt, harass, and not allow them to complete the task. The

No Frustration condition was established by allowing the

remaining half of the subjects to complete the counting with-

out interruption and at their ovwn pace.

Pollowinsg the fruistrationl man i.pulation, the subjects

were assigned in one of three co,,t erresponse conditions:

1. Counteraggres, sion. Twenty-four subjects selected

for their fourth task the same interpersonal guessing sit-

uation described in the pre-meesure of aggression condition.

The identical procedure was carried oJt with the exception

that the person guessing the numbers and receiving shock was

the experimental assistant (frustrator).

2. o opor2Stunity to aggress. Twenty-four subjects

selected for their fourth task the same interpersonal guess-

ing situation described in the pre-measure of aggression

condition. The identical procedure was carried out with

two exceptions. The person guessing the numbers was the

experimental assistant (frustrator) and errors were signalled

verbally rather than by using shock.

3. Discussion. Twenty-four subjects selected for their

fourth task a discussion situation which required that they

discuusf the experiment for five minutes with the experimenter.

The discussion was tape-recorded and the experimental assist-

ant (frustrator) w as not present. The experimenter attempted

by means of an unstructured interview to get subjects to

identify and express any feelings that they had towards the

experimental assistant (frustrator). The subjects were not

allowed to evaluate or derogate the experimental assistant

in any way. The experimenter nas a trained doctoral student

in clinical psychology. The experimenter ran the subjects

blind in this condition. Even thouLh feelings were empha-

sized, some subjects did mention that they were not permitted

to finish the counting task.

Post-m.easure :f aggression. Follow-ing the response

to frustration conJitions, all subjects selected for their

fifth and final task the same interpersonal guessing gane

described above in the pre-measure of aggression condition.

The identical procedure was carried out with the confederate

posing as a fellow subject guessing the numbers and receiv-

ing shock for incorrect guesses. The experimental assistant

recorded the intensity and duration of each of the ten shocks

that were administered. The mean score of the intensity x

the duration of the ten shocks was taken as the post-measure

of the subject's aggressiveness.

gQuestiontaire and debriefing_. following the experi-

mental procedure, all subjects were asked to fill out t;wo

five-point scales to determine their level of guilt and

anxiety over administering shock to another person. On each

scale, point one indicates high guilt and anxiety, and point

five indicates no guilt or anxiety. Copies of these scales

are in the appendix.

Before leaving the session, all subjects were de-




Initial Comparability

The hypotheses under test in this study are based on

changes in both systolic blood pressure and measures of ag-

gression as a result of the experimental manipulations.

Hence, it must be determined if the groups did or did not

differ significantly on base level systolic blood pressure

and on the pre-measure of aggression. In Table 1, the means

and standard deviations for the base level systolic blood

pressure for the experimental groups are listed. By inspec-

tion, results reveal that very little difference exists

among the groups on base level systolic blood pressure.

Blood pressure readings in the sample ranged from

about a low of 105 to a high of 147. Average blood pressure

for most people is approximately 120 which is also the grand

mean of this sample.

In Table 2 are the means and standard deviations of

aggression scores used as a pre-measure of aggression. By

inspection, results reveal very little difference among the

groups on the pre-measure of aggression. The pre-measure

of aggression scores ranged from .07 to 5.82.


Means and Utaudard Deviations of Base Level
Systolic Blood Pressure

Counteraggression No Aggression Discussion

X 119 X 120 = 119
s = 6.3 s = 11.6 s = 11.0

X = 120 X 122 X = 121
Not Frustrated
s = 6.0 s = 8.2 s = 9.8


Means and Standard Deviations of Aggression Scores
Used as a Pre-measure of Aggression

Counteraggression No Aggression Discussion


Not Frustrated

s = .68 s : 1.15

x = 1.33

s = .84

S- 1.36

= 1.14

s 1.02

X= 1.58

X = 1.56

s l.lb

x- 1.33

s = .75

~`----1 --~1--~11--- --- -~------`-I' ~--~-

Analyses of Blood Pressure Measures

In Table 3, the mean levels of systolic blood pres-

sure for each group are shown.

The summary table for the three-way analysis of var-

iance with repeated measures that assessed changes in systolic

pressure is in Table 4.

In Figure 1, the interaction effect obtained for the

Frustration x Systolic Blood Pressure readings (SBP) is

presented graphically. Inspection of the graph reveals that

the non-frustrated group changed very little over the three

periods of systolic readings. On the other hand, the frus-

trated group showed a sharp increase following frustration and

a return to pre-frustration or base level following the re--

sponse to frustration.

In Figure 2, the interaction effect obtained for the

Response to Frustration x Systolic Blood Pressure readings

(SBP) is depicted graphically. Inspection of the graph

reveals that both the counteraggression and discussion groups

followed the same sequence over the three periods of blood

pressure readings. Both groups showed a significant in-

crease following frustration and a return to base level fol-

lowing the response to frustration. On the other hand, the

group having no opportunity to aggress showed an increase

following frustration, but this increase did not return to

base level following the response to frustration. This

group maintained systolic elevations.


'Means of Systolic Blood Pressure Vithin All
Groups and Conditions

Base Post-
Level Frustration



Co)un ter-
aggression 119 128 118

iAggression 120 129 127

Discussion 119 129 118

aggression 120 120 119




X : 121 = 125 X 2 121


Summary Table for the Analysis of Variance:
Systolic Blood Pressure

Source df Mis F

Frustration 1 322. 67 .. 24

Response to Frustration 2 121.48 .47

Systolic blood pressure
readings (iBP) 2 432.92 23.29**

Frustration x Response to
Frustration 2 43.04 .1b

Frustration x SBP 2 483.76 2o.58x-

Response to Frustration x SBP 4 55.71 3.00 -

Frustration x Response to
Frustration x SBP 4 38.28 2.10

Subs. w;/,roups 6b 259.24

SBP x Subs. w/groups 132 18.20

* p< 0.05
** p/ 0.01


Not Frustrated



Base Post-
Level Frustration

Post -Response
to Frustration

Figure 1. Mean Systolic Blood Pressures at Three
Readings at Each Level of Frustration

125 -



Counteraggreosion =
No Aggression =
Discussion =


125 .-

120 -

115 -

Base Pot- Post-Response
Level Frustration to Frustration

Figure 2. Iean Systolic Blood Pressures at Three
Readings at Each Level of Response to
Prustrat ion

In order to test hypotheses I and III, t tests were

employed to make two pairwise cell mean comparisons on the

data shown in Table 3.

Hypothesis I states that frustration produces eleva-

tions in systolic blood pressure. The t ratio for the com-

parison of the post-frustration mean (7 = 125) and of the

base level mean (7 121) was 5.63. This ratio is signifi-

cant beyond the 0.01 level (df = 132) indicating there was

an increase in systolic blood pressure from the base level

to the post-frustration measure of systolic blood pressure.

Hypothesis I is, therefore, confirmed.

Hypothesis II proposed that following frustration

counteraggression against a frustrator produces a return of

systolic elevations to pre-frustration levels. The differ-

ence between (1) the frustrated-counteraggression-post-

treatment cell (7 = 118) and (2) the frustrated-counter-

aggression-base level cell (7 = 119) was less than zero;

therefore, no statistical test was required. This finding is

consistent with the proposal that no chances remained in sys-

tolic blood pressure following counteraggression relative to

the base level.

Hypothesis IIT proposed that systolic elevations are

maintained followinL frustration if there is no opportunity

to ggress against the frustrator. The t ratio that was

obtained when the means of the frustrated-no aggression-

post-response cell (X = 127) and the frustratcd-no aggression--

base level cell (7 = 120) were compared was 4.04. This

ratio was significant beyond the 0.01 level (df = 132) and

indicates that among frustrated subjects haing no opportu-

nity to counteraggress, systolic elevations remained at the

post-response measure relative to the base level measures.

Hypothesis III, therefore, is confirmed.

Hypothesis IV proposed that following frustration

having an opportunity to discuss the frustration produces a

return of systolic elevations to pre-frustration levels.

The difference between (1) the frustrated-discussion-post-

treatment cell (2 = 118) and (2) the frustrated-discussion-

base level cell (X = 119) was less than zero; therefore, no

statistical test was required. This finding is consistent

with the proposal that no changes in systolic blood pressure

remained following discussion relative to the base level.

Analyses of Aggression Scores

In Table 5, the cell means for aggression scores for

all groups are shown.

The summary of the three-way analysis of variance

with repeated measures used to test for changes in aggres-

sion scores is shown in Table 6.

In Figure 3, the interaction effect that was obtained

for the Response to Frustration x Periods of Aggression meas--

ure conditions for each level of Frustration is presented

graphically. Inspection of the graph indicates that among


Means for Aggression Scores for All Groups

Pre-.Leasure Post--leasure



Frus I.rated

aggression 1.33 3.51

Aggression 1.14 2.31

Discussion 1.50 1.20

aggression 1.36 2.19

Aggression 1.59 1.51

Discussion 1.33 1.33



Sumirary Table for Analysis of Variance:
Aggression Scores

Source df IFS F

Frustration 1 5.03 1.64

Response to Frustration 2 b.61 2.16

Periods of Aggression measure 1 17.99 20.6b9*

Frustration x Response to
Frustration 2 1.09 .36

Frustration x Periods of
Aggression measure 1 7.34 8.44xx

Response to Frustration x Periods
of Aggression measure 2 8.64 9.932-1'

Frustration x Response to
Frustration x Periods of
Frustration measure 2 3.77 4.33>

Sub, w/groups 66 3.06

Periods of Aggression measure x
Subs. w/groups b6 .87

p <0.025
p< 0.01

Counteraggression-= n ...
No Aggression- :_.,.
DUiscussion .........,e



3.00 /


1.50. *1

1.00 t,

Pre- Post-
Measure Measure


Pre- Post-
MIeasure Jleasure

Figure 3, er- AIggre;sion Scorcs for All Responses to
Frastra-tioiJ on Both Pre- and post-mrcansu! '
for Both Levels of Frustrat.ioi

frustrated subjects, both the counteraggrLession and the no

aggression groups showed increases in the post-measure of

aggression relative to their pro-measure. Confirmation of

hypothesis VI, shova below, indicates that the increase for

the counteraggression group was significant; however, a Tukey

post hoc comparison was made on the No aggression group, and

the q ratio obtained (2.00) was not significant. The dis-

cussion group among frustrated subjects showed little or

no change between pre- and post-measures of aggression,

Among the non-frustrated subjects, the graph indicates that

the only group to show an increase in the post-measure of

aggression is the counteraggression group. A Tukey post hoc

comparison made on the no frustration-counteraggression

group was not significant (qj 1.00).

Hypothesis V states that frustration reduction

achieved through the discussion of frustration results in

no change in aggression relative to base level. The differ-

ence between (1) the frustrated-discussion-post-measure cell

(, -= 1.20) and (2) the frustrated-discussion-pre-measure

cell (*x 1.53) was less than zero, therefore, no statis-

tical test was required. This finding is consistent with

the proposal that no changes tool place between pre-measure

and post--measure of te aggression for -he discussion group.

'HyUothesis VI proposed that frustration reduction

achieved thro-ugh the frustration-count oragression sequence

results in an increase in subsccuent a. rxession relative to

base levels. The t ratio that was obtained ihen the means

of the frustrated-ccunteraggression-post-nmeasure cell (X "

3.51) and the frustrated--coutteraEgression-pre-measure cell

(X = 1.33) were compared was 6.23. This ratio was signifi-

cant beyond the 0.01 level (df 66) and indicates that a

significant increase occurred between the pre-measure and

post-measure of aggression for frustrated subjects who were

allowed to countcraggress against the frustrator. This sup-

ports hypothesis VI.

Questionnaire Results
Ninety--eight percent of the subjects checked either

four or five on the five-point scale on both the guilt and

anxiety measures. This means that the subjects, on the whole,

reported little or no guilt or anxiety resulting from admin-

istering shock to another person.



Replication of Hokanson's Findings

The confirmation of hypotheses I, II and III.repli-

cates the major findings of IHokanson and his colleagues

regarding the arousal and reduction phenomenon of the frus-

tration-aggression sequence. That is, experimentally pro-

duced frustration causes systolic elevations and being able

to counteraggress against the frustrator under certain cir-

cumstances almost guarantees a return of the elevations to

pre-frustration levels. Post-frustration counteraggression

was found to be tension-reducing. Hokanson's results con-

cerning frustrated subjects who are not permitted to coun-

teraggress were also replicated. Subjects in the no oppor-

tunity to aggress group maintained their systolic elevations.

Discussion as a Response to Frustration

Confirmation of hypothesis IV allows us to conclude

that the discussion of frustration is an effective non-

aggressive method for relieving frustration-produced ten-

sions.I Discussion of frustration was as effective as coun-

teraggression in reducing frustration-produced systolic

elevations. Post-frustration tension-reduction, therefore,

need not be directly related to inflicting injury


(Berkowitz, 1962) or merely having the opportunity to aggress

(ilokanson, 1961). Tension relief may simply be related to

the acknowledgment of the feeling by the indiv:iduel rather

than upon the necessity of inflicting pain to the frustrator.

If future research should reveal that the relief

from frustration can consistently be obtained by becoming

aware of our frustrations and acknowledging them, this

method would no doubt be an efficient and socially accepta-

ble way of dealing with frustration. Presumably, both

sexes and not only males would have access to this mode of

response. and, furthermore, this mode of response does not

require the physical presence of the frustrator.

Effec If on SunRSienJe Ak; :ilcsniveess

Confirmoiati.on of hypothesis VI Jnaicates that althouLg

there i.: tcnsion-red.uction followingr the frus-tration-counter-

aggression sequence, the overall, effect upon a:tgressiveness

is to increase it. This is i. acco:r wiLth whavt Berkowitz

(1952) proposes that frustrstion-aggress:ion could result in

an increase in one's level of habitual aggressiveness. This

finding could account for the findings in so many of the

classic studies in thi-,s area in which subsequent agLrcssion

was ircreast u rather than decreased. lore important, how,-

ever, confirmiation of hypothesis V indicates that thel dis-

cuscion of frust.rati n is an efficient non-agsgressive means

for dealing with frustration. It not only achieves tension-

reduction, but it r.';o produces no increase in subsequent

aggressive'e; 2

The efficacy of a discussion response to frustration

in reducing tension is compatible with the theory presented

by Berkowitz (1962) and elaborated by iokanson (1966).

Berkowitz maintains that frustration produces general, phys-

iological arousal. This arousal predisposes us to aggres-

sive behavior. Further, there is a tendency to complete an

aggression sequence and, therefore, if permitted, aggression

will occur against an available target. Hoakeson elaborates

this by stating that the aggression will be tension-reducing

only if it is perceived to affect the frustrator and only if

it has been learned in the past that it is an appropriate

response to make.

,What is refreshing about Berkowitz' view is that it

offers various stages at which aggression may be interrupted

and possibly prevented. This is in contrast to the tradi-

tional Freudian model (Freud, 1959) which argues that within

each individual there is a reservoir of aggressive energy

that must be discharged; hence, aggression cannot be prevented.

The present results suggest that placing a discussion of frus-

tration at an early stage in the frustration-aggression se-

quence reduces the likelihood of subsequent aggressive behav-

ior. In other words,aggression does not have to be ex-

pressed in order to relieve frustration. Discussion of frus-

tration may help the individual re-evaluate the frustrating

experience and permit him to adopt another nchavioral dis-

position to the situation]

One of th* main contribution's that Hokanson's (1966)

study makes to the present finding concerning the discussion

of frustration is his attempt to include his results within

a learning theory framework. As reviewed earlier, i.okanson

believes that people learn appropriate techniques with which

to obtain relief from frustration. If okS:anson is correct,

then it can be assumed that individuals can learn to use the

discussion of frustration in dealing with frustration.

Recently, Stone and iiokanson (1969) found that subjects can

learn to be self-punitive rather than to aggress against the

frustrator. The self-punitive behavior produced arousal

reduction for these subjects. Clearly, if discussions are

efficacious in reducing frustration and if one car- learn to

use this techique, then an effective method for decreasing

the tendency for aggressive behavior towards self and others

has been located,

Alternati;ce Hyivootheses

One way in which the efficacy of the discussion

might be explained is to consider this condition in terms

of verbal aggression. Certainly, Hokanson and Burgess

(1962) have sho;-n that verbal aggression .i as effective

as overt, physical aggression in reducing frustration-pro-

duced tensions. however, this alternative hypjot.eis is

unten-.ble because the subjects ini the discussion group were

never allowed to evaluate the frustrator in a negative man-

ner. 'hc discuo.sion involved feelings; that the subjects

may have had towards the frustrator, such as reseentc ntn, Co

disgust, or anger. The subjects were not permitted to eval-

uate or derogate the frustrator in any way.

Another interesting alternative hypothesis that may

be offered to explain the increase in aggression scores con-

cerns the report that most of the subjects gave concerning

their level of guilt and anxiety over administering shock.

One of the most prominent explanations of catharsis is that

guilt and not tension-reduction reduces subsequent aggression

(Berkowitz, 1962). Cuilt and/or anxiety over our aggressive

behavior inhibits our aggressing further. If the question-

naires concerning guilt and anxiety are accurate, then the

subjects in this study were biased in this respect. There

was no guilt or anxiety and, therefore, there was no inhibi-

tion to aggress. In order to determine whether or not the

present findings are related to low guilt and anxiety, future

research should take this into account. In a very recent

study, Gambaro and Rabin (1969) considered guilt to be an

important factor in aggressive behavior. Based on this

interpretation of catharsis, it is possible to explain that

the increase in aggression for the counteraggression groups

was a result of low guilt and anxiety rather than tension-

reduction or a practice effect.

A type of practice effect offers yet another alter-

native explanation for the increase in aggression for the

coiuteraggression groups. In Figure 3, it ccan be noted that

the non-frustrated and frustrated counteraggression groups

showed an increase between the pre-measure and post-measure

of aggression. What could account for the increase in ag-

gressiveness? Both counteraggression groups (frustrated and

non-frustrated) experienced at least 20 trials of adminis-

tering shock before the post-measure of aggression was taken.

All other groups received only ten trials before the post-

measure was taken. It may be that the increase in aggression

that was found with both counteraggression groups is due sim-

ply to engaging in more aggression and not to tension-reduc-

tion as originally postulated. Before a more definite state-

ment can be made concerning the overall effect of arousal

reduction upon subsequent aggressiveness, the practice ef-

fect has to be controlled.

One final but important way in looking at the effi- /

cacy of the discussion group in reducing subsequent aggres-

siveness is to consider it in terms of a social learning

theory proposed by Bandura (1964). Bandura maintains that\

if a frustrated person should become immersed in new activ-

ities that supersede the preoccupying, brooding frustrating

event, a noticeable degree of tension-reduction will take

place and no aggression will follow. Bandura attributes

this reduction of aggression to attentional shifts or stim-

ulus change rather than to energy discharge. He believes

that subjects would experience equally salutary effects from

getting involved in an absorbing book, a movie, or a stage

play containing few or no aggressive stimwli. It is diffi-

cult to say ,whether this alternative hyothesis applies to

the present study since the discussion group did not actually

shift attention to other matters nor did they really remove

themselves from the frustrating event. Rather, the discus-

sion focused on the frustration as such. One way to deter-

mine the effects of attentional shifts is to include an ad-

ditional discussion condition in which subjects are actually

required to get involved in discussion on topics irrelevant

to the frustrating event.


Certain limitations exist in the present study that

prevent concluding positively that the discussion of frus-

tration produced tension-reduction from frustration. Since

no group was included that was allo-wed to discuss a frustra-

tion irrelevant topic, it is difficult to pinpoint the ef-

fect of a discussion as such. Another study should inves-

tigate this condition in more detail taking into considera-

tion both. verbal aggression as well as attentional shift as

mechaninss of tension-reduction.

The results of this study appear to contradict the

catharsis hypothesis (Dollard et al., 1939) which predicts

a reduction of aggression follo--.'wing aggressive behavior.

Even though, the results sho; an increase in subsequent ag-

6gressiveness for several groups, -:hat about the possibility

that aggression may hav-e been mon.entarily reduced durin,-

the counteraggression? For example, during the post-frustra-

tion counteraggression, the subject was allowed ten trials

in which to shock the frustrator. A test to determine whether

or not aggression was momentarily reduced would be to look at

the difference between the first five trials and the last

five trials. Unfortunately, no data were collected during

the counteraggression trials. The only statement that would

be safe to make is that following the frustration-aggression

sequence, aggression may be momentarily reduced but the over-

all effect is an increase in subsequent aggressiveness.

Another limitation is that the post-measure of aggression

was taken approximately twenty minutes after the experimental

conditions. Future research should investigate whether or

not increases or decreases of aggressiveness are maintained

over a longer period of time.

One final but important limitation of this study is

that subsequent aggression towards the actual frustrator

was not tested. The shock that subjects administered during

the post-measure of aggression was always towards the con-

federate and not the frustrator. There is no way of knowing

if these subjects would become more aggressive had the orig-

inal frustrator been available for shock during the post-

measure. It can only be assumed that the confederate that

was used was similar to the frustrator and would have been

an appropriate target for the displaced aggression. This

assumption certainly obtains support from the frustrated-no


aggression group when they showed a substantial increase in

aggressiveness from pre-measure to post--measure. Not having

an opportunity to aggress against the frustrator, this group

displaced their aggression on the confederate on the post-

measure. This is similar to what Hokanson and his colleagues

found (1963) when they investigated displaced aggression.



Tihe arousal and reduction phenomenon of the frustra:-

tio.-:-. c-.gessioon sequence that okainson has consistently found

in is Vwork has been replicated in this present study. It

was found thaL- frustration produced significant elevations

.n systolic blood pressure and having the opportunity to coun-

tern-{ res against the frustrator reduced these elevations.

These who w'vere frustrated and not allowed to a-gress main.-

taind,,. svyto tlic elevations. In addition to this replication,

a discussion of frustration response was permitted, and .it

a;s found to be equally effective in reducing arousal from


Eiven though i.he counteraggression and discussion

re's,.onse; .,'ere fond to be similarly effective in reducing

.rustrti.jion-produced tensions, they differed significantly

in their effects upon subsequent aggreosiveness, Gounter-

aggreas ionv waso shown to increase sube uc ent a;,ression tbile

tile di.scu:sion of frustration produced ro chan-lesc in aiCtres-;

oji. C ir' Ci.Jlc:.ision of frou:;t'ration, was lcv;ecd as ,n effi-

cient non-a'15.os.'L; method for dcaliin; ',ith fri osira en

escci:Jiflly if ojne v- i'.s; lnt.rested in decreasin. { the pot,;ntiji.

for a:. .s'e s;sion._

Several alternative hypotheses were discussed. A

possibility exists that the efficiency of the discussion

condition could b(e attributed to (1) verbal agecrnssi on, or

(2) attentional shifts or stimulus change. Due to the nature

of the discussion condition in this study, these hypotheses

do not seem tenable, but they should be investigated in

future research.

Two alternative explanations wore offered to account

for the increase in aggression for both counteraggression

groups. One of these involved a practice effect since both

counteraggression groups engaged in ten more trials of ad-

ministering shock than any other group. The other explana-

tion was related to the fact that low ogilt and anxiety over

administering shock to another person were reported by the

majority of subjects. According to Berkowitz (1962), guilt

inhibits our aggressing further. Since these subjects had

little guilt, they could have had little inhibition to ag-


Certain limitations of the study were also discussed.

One involved the inability to determine what it was about

the discussion condition that made it effective. Another

involved the short duration of time in which the post-measure

of aggression was taken. The most important, however, in-

volved the fact that subsequent aggression towards the actual

frustrator in the discussion groups was not tested. In the

post--measure, shock was administered to a confederate and


not the actual frustrator. The assumption vias made that the

con.frder.ate vas an appropriate target for displaced aggres-

sion for this lro.p if they had wished to aggress.

The alternative interpretations as well as the limi-

tations thr..t vorc discussed offer a direction for future

research in taini area.




The PIc.LT TITi I shocked someone, I felt:(CIRCLE ONE IiULBEER)

1 2 3 4 5
Very Anxious No Anxiety

The SECOND TILiE I shocked someone, I felt: (CIRCLE OiE NUiIER)

1 2 3 4 5
Very Anxious No Anxiety

The THIRD TIU., I shocked someone, I felt: (CIRCLE ONE NUMBER)

1 2 3 4 5
Very Anxious No Anxiety



The FIRST TIES I shocked someone, I felt: (CIRCLE ONE NU,'BER)

Very Cuilty

No Guilt

The SECOND T.IME I shocked someone, I felt: (CIrCLE ONE NUMBER)

Very Guilty

No Guilt

The 5TIRPD TIl',E I shocked someone, I felt: (CIRCLE ONE NUTIBER)

Very Guilty

No Guilt


Baker, John. The effects of four tygoes of a gression on
physiological _adsychological arousal. Doctoral
dissertation, 'est Virginia University, 1967.

Baker, John. Effects of aggressing "alone" or "with another"
on physiological and psychological arousal. J. pers.
soc. psychol o y, 1969, 12, 80-86.

Bandura, A., arid Huston, Aletha C. Identification as a
process of incidental learning. J. abnorm. soc.
Psychol., 1961, 63, 311-318.

9Bandura, A.; Ross, Dorothea; and Ross, Sheila A. Imitation
of film-mediated aggressive models. J. abnorm. soc.
Psychol., 1963, 66, 3-11.

Bandura, A. Vicarious one trial learning. In L. Berkowitz
(Ed.), Advances in Experimental Social PsychologiL II.
New York: Academic Press, 19b4.

iBerkowitz, L. The expression and reduction of hostility.
Psychol. Bull., 1958, 55, 257-283.

Berkowitz, I. Agegession: A Social Psychological AnalLyis
New York: LcGra.-Hill, 19c2.

Dollard, J.; Doob, L.; miller, N.; iMower, 0.; and Sears, R.
Frustration and Aggression. New Haven: Yale, 1939.

Feshbach, S. The drive-reducing function of fantasy behav-
ior. J. abnorn. soc. Psychol., 1955, 50, 3-11.

Feshbach, S. The stimulating versus cathartic effects of
a vicarious aggressive activity. J. abnorm. soc.
Psychol., 1961, 63, 381-385.

Feshbach, S. The function of aggression and the regulation
of aggressive drive. Psyjchol._Rev., 1965, 71, 257-272.


Ganbaro, S., and Rabin A. Diastolic bloodpressure responses
following direct and displaced aggression after anger
arousal in high- and low--guilt subjects. J. pers.
soc. Psychol., 19b9, 12, 87--94.

Freud, S. Why war? Letter to Professor Einstein. In J.
Strachey (Ed.) Collected papersL of Sigmund Freud.
Vol. 5, New York: Basic Books, 1i59..

.-Hartmann, H.; Kris, E.; and Loewenstein, R. ,L. Notes on
therapy of aggression. Psychoanalytic Study of the
Child, 1949. Vols. 3-4. Neew York: International
Universities Press, 1949.

Hokanson, J. The effects of frustration and anxiety on overt
aggression. J. abnorm. soc. Psychol., 1961, 62, 346-

Holans on physiological tension level. J. abnorm. soc.
Psychol., 1961, 63, 446-448.

SHokanson, J., and Burgess, 71.. The effects of status, type of
frustration, and aggression on vascular process.
J. abnorm. soe. Psychol., 1962a, 65, 232-237.

Holkanson, J., and Burgess, 'i.o The effects of three types of
aggression on vascular processes. J. abnorm. soc.
1ychol., 1962b, 61, 446-449.

Hioklannon, J.; Burges, I,.; and Cohen, M. Effects of disp aced
aggression on systolic blood pressure, J. abnorm.
SaO.Ph :_Jlo_, 1963, _67, 214-218.

Hokanson, J., and Edelman, E. The effects of three social
responses on vascular processes. J. per. soc.
Psy hol., 1966, 3, 442-444.

Akahn, I.i. A polygraph study of the catharsis of aggro sien.
DocT'oril Tiss er-ctaion, Harvard University, lyb0.

,hahr, 1 The physiology of catharsis. J. per. soc. Psyc.ol,
1906, 3, 27'--20u.

kenny, D. An experimental test of the Cathari.i thecry of
ire on, oot.or.l d1:i. s:;s~ 'rti.ion, Unlive ~s ity of
\'ashington, 39593.

Lovaas, O. Effects of exposure to symbolic aggression on
aggressive behavior. Child. Dev., 1961, 32, 37-44.

IMussen, P., and Rutherford, E. Effects of aggressive cartoons
on children's aggressive play. J. abnorm. soc.
Psychol., 1961, 42, 461-465.

Pepitone, A., and Reichling, G. Group cohesiveness and the
expression of hostility. Hum. Relat., 1955, 8, 327-

SRogers, C. R. The phenomenological theory of personality.
In 'W. S. Sahakian (Ed.), Psychology of Personality:
Readings in Theory. Chicago, Iil.: Rand ,cNally,
1965, pp. 473-493.

Rogers, C. R. Rogers' self theory. In C. V\. Hall and G.
Lindzey (Eds.), Theories of Personality. New York:
John Wiley, 1957, pp. 476--501.

-Rosenbaum, II., and deCharns, R. Direct and vicarious
reduction to hostility. J. abnorm. soc. Psychol.,
1960, 40, 105-111.

Siegel, Alberta E. Film-mediated fantasy aggression and
strength of aggressive drive. Child Dev., 1956, 27,

Stone, L., and Hokanson, J. Arousal reduction via self-
punitive behavior. J. pers. soc. Psychol., 1969,
12, 72-79.

Thibaut, J. An experimental study of cohesiveness of under-
privileged groups. Hum. Relat.i 1950, 3, 251-278.

Thibaut, J., and Coules, J. The role of communication in
the reduction of interpersonal hostility. J. abnorm.
soc. Psychol., 1952, 41, 770-777.

Walker, A. L.; Rablen, R. A.; and Rogers, C. R. Develop-
ment of a scale to measure process changes in psy-
chotherapy. J. of clinical psychology, 1960, 16 (1).

Worchel, P. Catharsis and the relief of hostility. J.
abnorm. soc. Psychol., 1957, 55, 238-243.


Anastasia E. Wells was born on January 20, 1933, in

Iew York, New'; York. She was graduated from Cathedral High1

School, Pew York, New York, in June; 1950, and was subse-

queltly employed in private industry for several years. In

February, 1957, she enrolled in Hunter College of the City

University of hew York. After eight years of evening sension

work, she received her Bachelor of Arts degree cum laude in

June, 1965, with honors in both economics and psychology.

During her junior year at Hunter College, she was elected

to Phi Eta Kappa and Psi Chi.

In September, 1966, she enrolled in the Graduate

School of the University of Florida, and in June,1968, she

vwa awarde- c her laster of Arts degree. She is currently fL,--

filljing the requirements for her Ph.D. degree. During her

-graduate studies at the University of Florida, she was a

U.S.P.Hi. ?'ellovw for one year and a Veterans Administration

trainee for two and a half years. She also served a one

year's internship in clinical psychology at the J. Hills

E;iller Ledical Center. Upon graduation, she will be employed.

as a staff psychologist at the Veterans iiospita] in Gainef.-

ville Florida

Anastasia iells was married to Arrthur Li. ellsls, Jr.,

in April, 1967.

This di;scrtutiion ;was p'rpared under the direction

of the chairman of the cand idate's supervisory co,ir..ittee and

had been approved Ly all n nar.rs of that coniLiittee. It vas

outmitted to the Dean of the College of Arti and Sciences

and to the Graduate Council, and was approved as partial ful-

fillment of the requirements for the degree of Doctor of


l]arch, 1970

Dean, College of Arts and sciences

Dean, Graduate School

Supervisory Coommittee:


.._./^^^-.,L--------- --

^7^'"~ /^^'~-~

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs