A comparison of two methods of overcoming avoidance behavior.


Material Information

A comparison of two methods of overcoming avoidance behavior.
Physical Description:
iv, 37 leaves : ill. ; 28 cm.
Goldstein, Alan Jay, 1933-
Publication Date:


Subjects / Keywords:
Avoidance (Psychology)   ( lcsh )
Learning, Psychology of   ( lcsh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis--University of Florida.
Bibliography: leaves 34-36.
Statement of Responsibility:
By Alan Jay Goldstein .
General Note:
Manuscript copy.
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000565671
notis - ACZ2090
oclc - 13571766
System ID:

Full Text





August, 1967


The author wishes to express his gratitude for the dedicated

guidance and instruction by Drs. H. C. Davis, Jr., and H. S. Pennypacker.

Appreciation is extended to Drs. H. Grater, W. D. Wolking, Kacquelin

Goldman and I. J. Gordon for their helpful suggestions in the planning

and in the reviewing of this dissertation.

The research reported herein was supported by NIH Grants

5-F1-MH-28, 099-02 and MH-08887.



A LC P LED j- l'-'ETS . . ii


INTRODUCTION .. .. .. *. 1

METIOD ......... .. .. 10



SUri-'ARY ............... .. .. 31

REFERENCES .. ... .. . 3*4





1 Number of Days of Treatment Required to
Completely Overcome Avoidance . .. .

2 Percentage Change Scores (Inches) for
Each S and Group Means . .... .




la Runway ...... ... .. ..

lb Treatment Surround ..... ........ .* *

2 Mean Avoidance Measurement (Inches) for Each Group
Throughout Phase I. Raw Scores Transformed by
Means of Three-Term Moving Average .

S. 17

3 Avoidance Measurements (Raw Data) for Each S Across
Phase I Treatment Program . .

4 Last Three Treatment Measurements, Spontaneous.
Recovery Measurement and Four Reconditioning
Measurements for Each Treatment S .. .


There are those who have put forth theories of psychological

disorders which appeal to learning concepts in explanation of etiology

(Eysenck, 1960; Ullman and Krasner, 1965). Some of these people have

derived psychotherapeutic regimes with a wide range of disorders by

extension of learning theory concepts and laboratory conditioning

paradigms (Bandura, 1961). On the other hand there are many who take

issue with these theoretical positions and are able to point to serious

difficulties when generalizing too broadly from laboratory data (Breger

and McGaugh, 1965). Nonetheless, there seems to be some consensus that

this approach has yielded some fairly effective techniques applicable

to at least a limited number of problems (Bandura, 1961; Grossberg, 1964).

In reviewing the literature, it appears that two of these techni-

ques and situations in which they are fairly consistently effective are:

1. Skinnerian operant techniques wherein reinforcement of

discreet desired behavior is the primary technique. Such techniques

have been primarily directed toward building behavior in those who show

deficits in particular areas, or extinction of undesirable discreet

responses (Ullman and Krasner, 1965). In particular, these techniques

have been extensively reported as helpful in building of more positive

sorts of behavior in inpatient settings, and although there are few

reports of its resulting in cures per se, they apparently are helpful

in ward management. They are also being used in building new behavior

in children with behavior deficits, primarily with retarded and


autistic children and in extinction of undesirable overt behaviors.

2. The second is the concept of reciprocal inhibition developed

from Pavlovian classical conditioning work. This technique is used

primarily to overcome fear and avoidance behavior. It has been found

particularly effective with anxiety neurosis in which the anxiety is

well focused such as in phobias (Wolpe, 1958). The hypothesized

therapeutic principle is that: "If a response antagonistic to anxiety

can be made to occur in the presence of anxiety-evoking stimuli so that

it is accompanied by a complete or partial suppression of the anxiety

response, the bond between these stimuli and the anxiety responses will

be weakened" (Wolpe, 1958). It is generally accepted among behavior

therapists that it is also necessary to approach the anxiety provoking

stimulus gradually so as to allow the reciprocally inhibitory response

to occur (Wolpe, 1958). In fact this same principle was applied by

Watson and Rayner (1920) and Jones (1924) wherein: (1) A rabbit was

paired with a loud noise until the child (who showed no fear or avoid-

ance prior to the pairing) showed fear of the rabbit. (2) The child

was then allowed to eat in the presence of the rabbit--with the rabbit

initially being a distance from the child. The rabbit was slowly

moved closer to the child as the child ate. This procedure resulted in

removal of fear and avoidance responses in the presence of the rabbit.

The reciprocal inhibition technique now being used is just this

sort of arrangement. The hypothesized therapeutic factors are based on

two assumptions:

1. "Counterconditioning" rather than "Extinction" is taking


1. Extinction is herein defined in accordance with Pavlov's


2. Presentation of the feared stimulus by "Progressive

approach" is necessary.

Issues now being raised about these assumptions (Lomont, 1965)


1. Since in application of this technique, all conditions for

pure extinction are present, then it may be that the countercondition-

ing agent is not in fact responsible for the effect, but that only

extinction is occurring.

2. There is little comparative evidence supporting the assump-

tion that presentation of a feared stimulus by a progressive approach

leads to quicker removal of fear or avoidance than repeated full presen-


3. Nothing is known about which procedure leaves the S least

susceptible to either spontaneous recovery or reacquisition of the fear

and avoidance. Is there a susceptibility difference in Ss following

different treatment procedures?

It is proposed that some light may be shed upon these issues

with the present experimental investigation. In the review of the

relevant literature which follows, it may be seen that there have been

(1927) description, i.e. the specific procedure of presenting a
stimulus unaccompanied by reinforcement.

Counterconditioning is herein defined as the set of operations
which include along with the extinction procedure the addition of a
new unconditioned stimulus which elicits a response antagonistic to
the existing response.

2. Progressive approach refers to the presentation of a feared
stimuli in ranked order on the dimension of quantity of fear aroused,
the stimulus evoking least fear being presented first. Also, decreasing
distance from the feared stimulus and increasing time of exposure may
be utilized in progressive approach (Wolpe, 1958).


only a very few studies directly concerned with the questions raised

by Lomont (1965). Other studies are included in the review which

indirectly deal with some of these questions.

Extinction vs. counterconditioning

Wolpe (1958) attempted to arrive at a clinically therapeutic

treatment regime through the process of applying experimental data to

clinical situations. Specifically, he conducted a series of studies in

which cats were first conditioned to fear stimuli previously associated

with feeding. In order to eliminate the fear, he then began to feed

the Ss at decreasing distances from the feared stimuli until the Ss

were able to eat freely again under the full influence of the conditioned

stimuli. The assumption here is that eating and the fear reactions are

reciprocally inhibitory so that when feeding occurs in the presence of

the conditioned fear stimuli, the fear reaction is inhibited and gra-

dually eliminated.

This paradigm was extended in the clinical situation to include

such responses as relaxation, sexual arousal and assertiveness, all of

which were to be paired with stimuli which arouse fear in the patient.

At this point the question may be raised as to whether the feeding

(or any other counterconditioning agent) itself was necessary in obtain-

ing the results. An alternative explanation may be that the feeding

merely held the Ss in the fear arousing situation long enough for

extinction to occur. These experiments do not control for separate

analysis of extinction and counterconditioning effects as there is no

pure extinction group included.

Lomont (1965) points out that in order to conclude that the

reciprocal inhibition procedure really eliminates anxiety through

counterconditioning it is necessary to demonstrate that it is effec-

tive under conditions where extinction cannot fully account for its

efficacy. Lomont (1965) reviews four animal studies in which some

attempts at controlled comparisons were made. In three of these

studies, the counterconditioning procedure consisted of confining and

feeding rats in the fear conditioning apparatus between conditioning

and regular extinction. An extinction group was confined for the same

length of time but not fed.

In two of these studies (Sermat and Shephard, 1959; Lane, 1954)

avoidance responses were measures of fear acquisition and extinction.

Sermat and Shephard (1959) found no difference between countercondition-

ing and extinction groups while Lane (1954) found the countercondition-

ing procedure more effective than the extinction procedure, however,

there was opportunity for a great deal of movement through the apparatus

and the effect of feeding could have been to expose the Ss to more of

the stimuli present in the apparatus.

Moltz (1954) compared counterconditioning and extinction groups

in such a way as to eliminate the confounding effects of differential

movement. All Ss were trained to choose one of two arms of a T-maze

and were then shocked at the choice point. The Ss were then divided

into a counterconditioning group (fed at the choice point) and an

extinction group (merely confined at the choice point). The measure

of effectiveness was the persistence of the choice behavior after the

switching of the food reward. The results indicated no difference in

efficacy of the two procedures.

The most unequivocal study covered by Lomont (1965) was that of

Sollod and Sturmfels (1965) in which defecation in rats was used as


the measure of conditioned fear. The CS was a tone and no avoidance

training was used. Another element of this study was that it more

accurately represented Wolpe's reciprocal inhibition technique in

that the presentation of the fear stimulus progressed from presumably

weak to strong through change in frequency. While both treatment

groups showed reduction of fear, the counterconditioning group under-

went a significantly faster reduction.

Rachman (1965) investigated the separate effects of desensiti-

zation with relaxation (the hypothesized counterconditioning agent)

and desensitization without relaxation (extinction) in human Ss

demonstrating fear of spiders. Marked reduction in fear was reported

in the counterconditioning group only. The feared stimulus was

approached gradually in both groups.

Gale, Sturmfels and Gale (1966) compared counterconditioning and

extinction of a conditioned emotional response (defecation). Again,

both groups received the CS by progressive approach. The results indi-

cated that the counterconditioning procedure was more effective than

extinction in terms of time required to eliminate the fear response.

Progressive approach vs. full strength presentation of the feared

Studies using rats as Ss by Moltz (1954), Elder, Noblin and

Maher (1961) and Berkum (1957) are all similar in that Ss were subjected

to fear conditioning in one alley and then counterconditioned (taught

to approach for food) in alleys differing from the original alley

either in color or in distance. The Ss were then retested for avoidance

in the original alleyway. The results across studies are consistent

in that no differences were shown where color was the dimension of

progressive approach and a difference in favor of progressive approach


was demonstrated where distance was the manipulated variable.

In another study using rats, Kimble and Kendall (1953) trained

the Ss to avoid shock by turning a wheel. One group was subjected

to extinction by progressive approach (change in cycles per second of

tone CS) while another group received ordinary extinction. Non-escape

responding occurred significantly quicker in the former group.

It should be noted at this point that the studies supporting the

hypothesis that counterconditioning is more effective than extinction

all include a progressive approach procedure as well. They are

designed so that the separate effects of these two factors cannot be

appraised. Likewise in the studies dealing primarily with progressive

approach vs. full presentation of the CS-, all include feeding (counter-

conditioning) in both groups. The exception is the study by Kimble and

Kendall (1953) discussed above. An hypothesis consistent with the above

body of literature is that a counterconditioning agent adds effectiveness

only when progressive approach is also employed, but progressive approach

facilitates both counterconditioning and extinction of conditioned fear.

Spontaneous recovery and reacquisition of fear as a function of different
methods of fear removal

The only study available which approaches this problem is the

study by Gale, Strumbels and Gale (1966) which is discussed above. The

Ss were subjected to relearning trials following testing for fear removal.

The Ss receiving counterconditioning of fear responses showed slower

relearning of fear than did the extinction Ss. This result is equivocal,

however, because of a difference between groups in continued treatment

past the point of minimum fear.

The literature bearing on the questions being asked herein is in

many cases only indirectly applicable and none of the studies separates


the effects of extinction, counterconditioning and progressive approach

so that these factors can be evaluated separately and in all possible

combinations. Such a study would require four treatment groups:

(1) extinction, (2) counterconditioning, (3) extinction plus progres-

sive approach and (4) counterconditioning plus progressive approach.

A weakness inherent in animal studies concerned with overcoming

fear is that the fear is almost always conditioned in the same labora-

tory apparatus that is used in treating the fear. This is unlike

clinical situations wherein patients present themselves for treatment

with the fears already operating and the formative conditions usually

can only be speculated. In order to better approximate clinical condi-

tions, Ss (Cebus monkeys) exhibiting a non-adaptive (in the experimental

situation) fear were chosen. The fear stimulus was a stuffed "teddy


Avoidance behavior has been shown to be an adequate measurement of

fear (Hall, 1955) and, in fact, the behavioral act of avoidance is

often the cause of the difficulties bringing phobic patients to treat-

ment. Avoidance is, therefore, the favored criterion for measurement

of change. Studies in the literature using avoidance behavior as the

measurement criterion have run into difficulty keeping the conditions

of treatment constant across groups when comparing extinction and

counterconditioning in that the counterconditioning treatment usually

introduces the probability of more movement in the apparatus than does

extinction. In addition these studies do not control for the amount

of food intake (the usual counterconditioning element) across groups.

This study is designed to hold constant all known conditions across

treatment groups except the one directly under study, while using the


favored criterion of change in avoidance behavior as the measurement

of effectiveness of treatment.

Following the above guidelines, the present study will attempt to

shed some light on the following questions:

1. What is the effect upon reduction in fear (decreased avoidance)

of counterconditioning vs. extinction?

2. What is the effect upon reduction in fear of progressive

approach vs. full strength presentation of the feared stimulus?

3. What is the differential effect of the possible combinations

of these variables in treatment of fear?

4. In terms of spontaneous recovery and/or reacquisition of

fear, is there any difference due to different treatments in overcoming

the fear?



Ten young adult male Cebus monkeys (Cebus Albifrons) with no

previous experimental experience were used as Ss. These animals are

trapped in their native habitat and previous life experience is not

known. Without any training, all animals demonstrated a fear and

avoidance of strange furry objects.


1. A straight runway 345 in. long, 24 in. high and 24 in. wide

with guillotine doors at both ends was constructed as shown in Figure la.

The runway was constructed of 1/4 in. wire mesh folded lengthwise and

secured to 2 in. by 2 in. wooden beams which served as anchors. Both

ends of the runway were left open so that housing cages could be placed

at each end.

2. A separate enclosure, designated "treatment surround," was

constructed with hardboard sides and a wire mesh top as shown in

Figure lb. The overall length was 8 ft. and a guillotine door separated

the enclosure into two compartments, one being 2 ft. and the other 6 ft.


A toy stuffed "teddy bear" which has been determined to be a

feared object by Cebus monkeys (Plotnik, 1966) was used as a fear stimu-



All Ss were placed on 23 hr. food deprivation two days before the


I 3451 f

Figure la. Runway

&- Door

Figure lb. Treatment Surround


beginning of the experiment and this schedule was maintained throughout

the experiment.

Approach training

The first procedure was that of training an approach response in

the runway. An S and his housing cage were placed at one end of the

runway while an empty housing cage was placed at the other. The S was

then allowed six trials per day in which he was allowed to move from one

end to the other of the runway to obtain one-fourth of a cube of sugar.

Before each trial the sugar was placed in the cage opposite the one

holding the S. Each trial was begun by raising the guillotine doors at

both ends,which allowed the S access to the runway and both housing

cages. Each trial was terminated by lowering the doors when the S

entered the cage to obtain the sugar. Training was stopped when the S

transversed the runway in 15 sees. or less on the first three consecu-

tive trials of that day.

Pre-treatment avoidance test

Following approach training, a pre-treatment avoidance measure

was obtained for each S by placing the stuffed bear, along with the

sugar, at the far end of the runway and releasing the S as usual at

feeding time.

The minimum distance that separated the S from the sugar and bear

within 90 secs. after being released was defined as the pre-treatment

avoidance measure.

Each S was then assigned to one of five groups, yielding an N of

two per group. The groups were matched on the avoidance-test measure


Treatment: Phase I


Phase I was designed to determine the effects of different treat-

ments in reducing fear.

Extinction (Ext) Group.--The Ss in this group were treated as

follows: Each was placed in the treatment surround separately, so that

the S was immediately in front of the movable door to the compartment

containing the stuffed bear. The S was given one-half a cube of sugar.

(The sugar is presented before the fear stimulus and therefore does not

act as a counterconditioning element, but does control for sugar intake

and stimulus properties of the treatment surround across groups.) After

an interval of about 3 min. the stimulus presentation occurred. A

stimulus presentation (trial) consisted of lifting the door, inserting

the bear into the cage with the S and closing the door for 2 min. The

door was then reopened and the bear removed.

Just prior to each treatment session (except the first) an avoid-

ance measurement as described above was obtained for each S to ascertain

the effect of the previous day's treatment.

Trials continued at the rate of one per day until the S entered

the cage containing the bear and food during an avoidance measurement

session. A maximum of 21 treatments were given to those Ss not reach-

ing this criterion.

Countercondition (CC) Group.--The Ss in this group were treated

exactly as those Ss in Ext group except that the sugar was delivered

10 secs. after each trial began rather than before the trial.

Extinction with progressive approach (Ext-PA) Group.--The Ss in

this group were treated exactly as Ss in Ext group except (1) the first

day's trial consisted of exposing the bear at a distance of 6 ft. from

the S and (2) on each subsequent day's trials, the bear was placed


2 ft. closer to the S than on the previous day's trials. On the fourth

day's trials the bear was placed in the S's cage as described for all

trials in Ext group. Trials were then continued under these conditions

to criterion.

Counterconditioning with progressive approach (CC-PA) Group.--The

Ss in this group were treated exactly as Ss in Ext-PA group except that

the sugar was delivered 10 sees. after each trial began rather than

before the trial.

Control Group.--All Ss in this group were subject to the same pro-

cedures as described for each of the groups except that during the

treatment trials the door was not raised so that the S was never exposed

to the bear.

Treatment: Phase II

Phase II was designed to determine the effects of different treat-

ment on spontaneous recovery. This was accomplished by waiting four days

after the criterion was met and at this time reintroducing experimental

Ss to the runway for an additional test of avoidance.

Treatment: Phase III

Phase III was designed to shed some light on the question of

whether or not some treatment more than others leave the S more suscep-

tible to reconditioning.

All experimental Ss were treated as follows: Each S was placed

in the treatment apparatus and the bear was exposed to each S for 30

sees. as in the previous Ext group treatment trials. However, during

this phase the bear was shaken vigorously. All Ss were given trials on

four successive days and then given an avoidance test in the runway 23

hrs. following each session in order to compare groups on reacquisition

of fear.


Phase I: Effects of treatments on avoidance behavior

Figure 2 shows the mean avoidance measure for each treatment

group and for the control group. The raw measurement scores have been

transformed by means of a three-term moving average (Hoel, 1960). As

these measurements were made between treatment sessions, they reflect

the change in the approach-avoidance balance throughout the treatment

course. The exception is the first measurement shown for each group

which was the measurement used for matching groups prior to the

introduction of treatment.

From this figure it may be seen that all groups were essentially

the same prior to treatment and that all treatment groups diverged

noticeably from the control group across the treatment course. In

addition it appears that the CC-PA treatment group showed the most rapid

overcoming of avoidance. However an analysis of variance showed a non-

significant difference across trials between groups (F(4,5)= 3.59;

p <.10].

A t test comparing the control group to all treatment groups was

highly significant (t = 3.73, p < .01). Indications are, therefore,

that all methods of treatment were more effective than no treatment

but that there was little difference in effectiveness of any one treat-

ment over the others when considering all group performances across all

treatment trials.

In terms of efficiency in time required to overcome avoidance,



Mean Avoidance Measurement (Inches) for Each Group
Throughout Phase I. Raw Scores Transformed by Means of
Three-Term Moving Average

0! -


20 f..

30 / j .'. "
4 o'

40 _
C\ I I
If \ I /I I ; i

50 I .

/ ,

7 I0:
: /:

70 "



Ext Days
....... Ext-PA
_. CC-PA


the results appear somewhat different. Table 1 shows the number of

days of treatment required to allow for complete overcoming of avoidance

for each S and the group means.

From Table 1 it may be seen that all treatments were more effi-

cient than non-treatment, that the counterconditioning component alone

added somewhat to efficiency only when progressive approach was present,

that the progressive approach component alone added somewhat to effi-

ciency only when counterconditioning was present, and that countercon-

diti.:.ig plus gradual approach (as represented by CC-PA group) was the

most efficient treatment of all in terms of number of occurrences

necessary to reach a non-avoidance criterion. A one-tail Mann-44hitney

test comparing the CC-PA group to all other groups was significant

(P < .044).

Figure 3 shows the avoidance measurements (raw data) for each S

across the entire treatment program (Phase I). It may be seen that in

no case did a S reach the non-avoidance criterion after the eighth

treatment. It seems that either avoidance was completely overcome by

this time or that the tendency was to show more avoidance as treatment

continued. The S showing the least avoidance consistently after treat-

ment eight seems to have performed asymptotically (Subject 25).

In order to relate each S's best single performance (least avoid-

ance) to the type of treatment regime, a quantitative comparison was

attempted. The avoidance measurement for each S that reflected the

least amount of avoidance for that S of any of his avoidance measurement

trials was used along with the pre-treatment measurement for each S to

generate a change score. The method of transformation followed that of

Lang and Lazovik (1963) and obeyed the following formula:


Number of Days of Treatment Required
to Completely Overcome Avoidance






20* 22 22 22 6

22 7 7 2 1

21 14.5 14.5 12 3.5


* This S did not overcome avoidance but became ill after 20
treatments and could not be treated any further.


Avoidance Measurements (Raw Data) for Each S
Across Phase I Treatment Program


S 25


s 78


S 84


s 83


s 9o

S 46

S 02

s 66



s 104



s 86




- 175



V- -
















percent change score = pre-treatment score--best performance score
pre-treatment score

Change scores for each S are presented in Table 2.

The data in Table 2 indicate some added effectiveness was gained

by the use of progressive approach whether counterconditioning was

present or not (Ext vs. Ext-PA, CC vs. CC-PA, and CC + Ext vs. Ext-PA +

CC-PA). The added variable of counterconditioning appears to contribute

to effectiveness only when progressive approach was also used (Ext-PA vs.

CC-PA). A comparison of the CC group with the Ext group indicates that

extinction was slightly more effective than counterconditioning when pro-

gressive approach was not included.

It may be seen in Table 2 that both Ss in the CC-PA group meet the

non-avoidance criterion while only one S did so in each of the other

treatment groups. In the control group both Ss showed more avoidance.

All treatments taken together and compared to the control group

showed a significant difference (P <.022) by use of a one-tail Mann-

W.hitney test.

Phase II: Spontaneous recovery

Figure 4 shows the last three treatment trials for each treatment

S, the spontaneous recovery measurement taken four days after the last

measurement of treatment effect, and the avoidance measurements follow-

ing each reconditioning trial.

For those Ss that completely overcame avoidance during treatment,

there was no spontaneous recovery of fear to a degree sufficient to

prevent complete approach on the spontaneous recovery measurements. For

those three Ss that did not completely overcome avoidance during the

treatment phase, two showed less avoidance on the spontaneous recovery


Percentage Change Scores (Inches)
for Each S and Group Means

Groups Control Ext Ext-PA CC CC-PA

Ss -16.67 30.77 66.67 22.22 100.00

-15.46 100.00 100.00 100.00 100.00

Mean -16.56 65.39 83.34 61.11 100.00


S 25


s 78


s 84


s 83

s 46

^ ~* ^

I I 1 !


,I I I I ,l l


i l l I I I I

,,/ -'

a s l I a .,


S 002

s 66



s 104




Last Three Treatment Measurements, Spontaneous
Recovery Measurement and Four Reconditioning
Measurements for Each Treatment S




DI y








165 .



Figure 4.





measurement and one showed a small increase.

Because of the variability of intrasubject measurement during

Phase I, the spontaneous recovery measurements were not subject to

evaluation for significance between groups. Indeed, it seems that the

treatment differences are of little importance, while success or failure

during treatment is a more crucial difference in appearance of sponta-

neous recovery.

Phase III: Reconditioning of fear

From Figure 4 it may seem that those five Ss that completely over-

cane avoidance during treatment showed no avoidance across reconditioning

trials, just as they showed no spontaneous recovery. Also, as in the

spontaneous recovery data, the three Ss that did not completely overcome

avoidance in the treatment phase showed an apparent reconditioning effect,

but again, this could not be said to be outside chance expectations when

considering intrasubject variability.


It is apparent that the statistical computations and trends noted

in the data must be regarded with reserve because of the small size of

the groups. Some hypotheses are suggested, however, concerning some of

the questions raised in the introduction of this paper.

1. What is the effect upon reduction in fear of countercondition-

ing vs. extinction?

The data suggest that counterconditioning was more effective

than extinction only when progressive approach was also used. This

was the case in terms of efficiency in time and also in terms of the

percentage of change effected. When the element of gradual approach

was removed (as in comparing the Ext group and CC group) then there was

no apparent difference between groups on any of the dimensions analyzed


2. What is the difference in effect upon reduction of fear between

progressive approach and full strength presentation of the feared stimu-


In terms of time efficiency, progressive approach added effective-

ness when compared with immediate full presentation only when counter-

conditioning was also present. In terms of maximum percentage of change

effected, progressive approach was more effective both when countercondi-

tioning was present and when it was not.

3. What is the differential effect of the possible combinations

of these variables in treatment of fear?



It is clear that in all comparisons made, the combination of

counterconditioning and gradual approach added to effectiveness of

treatment over any other treatment alone or any other combination of

treatment procedures.

All of the above hypotheses are consistent with the literature

covered in the introduction of this paper except our finding that

suggests that progressive approach is superior to immediate full

presentation in terms of time efficiency only when counterconditioning

is present. This finding is contradictory to the result of the study

by Kimble and Kendall (1953).

The trends in the data support Wolpe's (1958) theoretical assump-

tions that counterconditioning is an effective therapeutic element when

the anxiety provoking stimuli are presented by progressive approach so

as to allow the reciprocally inhibitory response to occur.

No light was shed on the last question raised concerning sponta-

neous recovery and reconditioning of fear as a function of treatment


As reported earlier, no S reached non-avoidance criterion between

the ninth and twenty-second treatment and there was an apparent trend

toward more avoidance after treatment nine. This finding is unexpected

in that the literature suggests gradual reduction of avoidance behavior

with continued exposure to feared stimuli when no escape is possible

(Kimble, 1961). A possible explanation is that the treatment trials

were so short and so spaced that there was insufficient exposure for

some Ss to benefit from them. Put another way, it is possible that

longer or more frequent treatment trials would have prevented the

occurrence of the observed increased avoidance. However, the fact that


control Ss showed an immediate increase in avoidance across trials

while all treatment Ss showed some decrease indicated that the treat-

ments were effective for a while at least. In fact, the increasing

avoidance by the control Ss indicates another factor is active.

While no escape was possible during treatments in the treatment

surround, it is possible that all Ss were furnished opportunity to

escape in the avoidance measurement runway. These measurement trials

often ended after the S made an approach toward the bear and sugar but

then scurried back to the start cage. Trial termination was effected

by closing the guillotine door exposing the bear, then the guillotine

door between the S and the runway. If removing visual exposure of

the bear was reinforcing as it probably was, then these were escape


If this is the explanation for the observed phenomenon in the.

control Ss, it still leaves something to be desired in explaining the

differences among treatment Ss. The above conditions of treatment and

avoidance measurement set up conflicting drive states with the balance

seemingly being tipped in favor of non-avoidance in the short run and

avoidance in the long run. These conditions are, however, very like

those existing in the clinical situation in that the patient is

treated in the office while "measurements" of treatment effectiveness

are made in other situations where avoidance and escape are possible

alternatives to non-avoidance on the patient's part. Because of the

similarity of this design to the clinical situation, means of overcom-

ing the tendency toward increased avoidance ought to be investigated

rather than attempting to design out this factor in future research.

Several approaches to overcoming this unexpected tendency toward


increased avoidance in later trials may be experimentally tested.

Experiments designed to investigate the effect of: (1) varying

lengths of exposure of the fear stimulus during treatment and (2)

increasing lengths of exposure across treatments seem to be promising


Procedurally, there were several problems encountered during the

present experiment.

Many potential Ss were eliminated because of the wide divergence

in performance at the point of taking a pre-treatment avoidance measure-

ment. While some Ss showed no avoidance, many were eliminated because

they showed too much. A more economical method of matching groups

would have included the possibility of reinstituting approach training

until an acceptable avoidance measure was obtained.

Attempts to measure spontaneous recovery in the present design

may have been frustrated by an unforeseen problem. There was a quali-

tative difference introduced between Ss that completely overcame

avoidance and those that did not. Those Ss doing so received additional

reinforcement (sugar) for the approach gradient when reaching the no-

avoidance criterion. This additional strengthening of the approach

response was not experienced by those Ss not reaching criterion.

The same problem presented itself in the reconditioning phase in

that those Ss reaching the no-avoidance criterion had additional rein-

forcement of approach responses that the remaining Ss did not have. In

addition, opportunity for reinforcement occurred on each measurement

trial between reconditioning trials. It is probable that this accounts

for the lack of effectiveness of the reconditioning stimulus with those

Ss having previously reached no-avoidance criterion. Both spontaneous


recovery and reconditioning would best be evaluated in independent

experiments in which all Ss could be treated to a specific non-

avoidance criterion that would be equal across groups.


The concept of "reciprocal inhibition" (Wolpe, 1958) has been

found to be an effective clinical technique, particularly in the treat-

ment of phobias. The hypothesized therapeutic principles are: (1) "If

a response antagonistic to anxiety can be made to occur in the presence

of anxlety-evoking stimuli so that it is accompanied by a complete or

partial suppression of the anxiety response, the bond between these

stimuli and the anxiety responses will be weakened" (Wolpe, 1958), and

(2) it is necessary to approach the anxiety provoking stimulus gradually

so as to allow the reciprocally inhibitory response to occur (Wolpe,


These assumptions, that counterconditioning rather than extinction

is taking place and that presentation of the feared stimulus by progres-

sive approach is necessary, are now being questioned. Lomont (1965)

pointed out that in the application of this technique all conditions

for pure extinction are present and that it may be that the countercon-

ditioning agent is not in fact responsible for the effect. He also

noted that there is little experimental evidence supporting the

assumption that presentation of a feared stimulus by a progressive

approach leads to quicker removal of fear than does repeated full


A review of the literature revealed that there were no published

studies which separated the effects of extinction, counterconditioning

and progressive approach so that these factors could be evaluated



separately and in all possible combinations. In order to make such an

evaluation possible, the present study employed four treatment groups

and a control group. The treatment groups were extinction, countercon-

ditioning, extinction plus progressive approach and counterconditioning

plus progressive approach.

Ten Cebus monkeys were used as Ss. Without any training, all Ss

demonstrated a fear of a toy "teddy bear" which was used as the fear

stimulus. The Ss were trained to transverse a runway to obtain sugar,

after which each S was allowed to approach the sugar while the bear was

also present at the end of the runway. The minimum distance that sepa-

rated the S from the sugar and bear was defined as the pre-treatment

avoidance measure which constituted the criterion for matching groups.

The same runway was used to determine change in avoidance behavior

throughout the course of treatment. The treatments (presentation of

the feared stimulus under the conditions appropriate to each of the

treatment groups) were conducted in a separate apparatus.

Although trends noted in the data must be regarded with reserve

because of the small size of the groups, some hypotheses were suggested.

In terms of both percentage of change effected and efficiency in time,

the data suggested that counterconditioning was more effective than

extinction when progressive approach was also used. When progressive

approach was not used, there was little difference between extinction

and counterconditioning. In terms of time efficiency, progressive

approach added effectiveness when compared with immediate full presen-

tation only when counterconditioning was also present. In terms of

maximum percentage of change effected, progressive approach was more

effective both when counterconditioning was present and when it was not.


It was noted that no S reached non-avoidance criterion between the

ninth and twenty-second treatment and that there was an apparent trend

toward more avoidance after treatment nine. Possible explanations for

this unexpected trend were discussed.

There was an attempt to shed light on the possible differences,

due to different treatments, in spontaneous recovery and reacquisition

of fear; however, the design proved inadequate to allow for such evalua-

tions. The design problems were also discussed.


Bandura, A. Psychotherapy as a learning process. Psychol. Bull.,

1961, 58, 143-159.

Berkum, 1M.M. Factors in the recovery from approach-avoidance conflict.

J. EL. Psychol., 1957, 54, 65-73.

Breger, L., & McGaugh, J.L. Critique and reformulation of "learning-

theory" approaches to psychotherapy and neurosis. Psychol. Bull.,

1965, 63, 338-358.

Elder, T., Noblin, C.D., & Maher, B. The extinction of fear as a function

of distance vs. dissimilarity from the original conflict situations.

J. abnozn. soc. Psychol., 1961, 63, 530-533.

Eysenck, H.J. (Ed.) Behavior therapy and the neuroses. New York:

Pergamon Press, 1960.

Gale, Diane S., Sturmfels, Gloria, & Gale, E.N. A comparison of reci-

procal inhibition and experimental extinction in the psychothera-

peutic process. Behav. Res. Ther., 1966, 4, 149-155.

Grossberg, J.M. Behavior therapy: a review. Psychol. Bull., 1964, 62,


Hall, Julia C. Some conditions of anxiety extinction. J. abnorm. soc.

Psychol., 1955, 51, 126-132.

Hoel, P.G. Elementary statistics. New York: John Wiley & Sons, Inc.,


Jones, Mary Cover. A laboratory study of fear. The case of Peter. J.

genet. Psychol., 1924, 31, 308.



Kimble, G. Hilgard and Marquis' conditioning and learning. New York:

Appleton-Century-Crofts, Inc., 1961.

Kimble, G.A., & Kendall, J.W. A comparison of two methods of producing

experimental extinction. J. Ep. Psychol., 1953, 45, 87-90.

Lane, Beatrice R. The reduction of anxiety under three experimental

conditions. Unpublished doctoral dissertation, 1954, Columbia


Lang, P. & Lazovik, A. The experimental desensitization of a phobia.

J. abnorm. soc. Psychol., 1963, 66, 519-525.

Lomont, J.F. Reciprocal inhibition or extinction? Behav. Res. Ther.,

1965, 3, 209-219.

Moltz, H. Resistance to extinction as a function of variations in

stimuli associated with shock. J. Exp. Psychol., 1954, 47, 418-424.

Pavlov, I.P. Conditioned reflexes. Oxford: Oxford University Press,


Plotnik, R. Changes in social behavior of squirrel monkeys after tempo-

ral lobectomy and an interspecies comparison of social dominance.

Unpublished doctoral dissertation, 1966, University of Florida.

Rachman, S. Studies in desensitization I: the separate effects of

relaxation and desensitization. Behav. Res. Ther., 1965, 3, 245-


Sermat, V., & Shephard, A.H. The effect of a feeding procedure on

persistent avoidance responses in rats. J. comp. physiol. Psychol.,

1959, 52, 206-211.

Sollod, Diane, & Sturmfels, Gloria. Reciprocal inhibition and the

conditioned emotional response. Unpublished manuscript, 1965,

Washington University.


Ullman, L.P., & Krasner, L. Case studies in behavior modification.

New York: Holt, Rinehart & Winston, Inc., 1965.

Watson, J.B., & Rayner, R. Conditioned emotional reactions. J. Exo.

Psychol., 1920, 3, 1-14.

Wolpe, J. Psychotherapy b reciprocal inhibition. Stanford: Stanford

Press, 1958.


Alan Jay Goldstein was born September 25, 1933, at Atlanta, Georgia.

In 1955 he received the degree of Bachelor of Business Administration

from the University of Georgia. He entered the University of Florida

September, 1963, and to the present he has pursued his work toward the

degree of Doctor of Philosophy.

This dissertation was prepared under the direction of the chairmen
of the candidate's supervisory committee and has been approved by all
members of that committee. It was submitted to the Dean of the College
of Arts and Sciences and to the Graduate Council, and was approved as
partial fulfillment of the requirements for the degree of Doctor of

August, 1967

Dean, College of Arts and Sciences

Dean, Graduate School
Supervisory Committee:

Co-Ch an

U Cm I

IJ -


11IIIII I1111 llll lll 1IIIII1III
3 1262 08553 6281