A COMPARISON OF TWO METHODS OF
OVERCOMING AVOIDANCE BEHAVIOR
ALAN JAY GOLDSTEIN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
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.
TABLE OF CONTENTS
A LC P LED j- l'-'ETS . . ii
LISTS OF TABLES AND FIGURES . iv
INTRODUCTION .. .. .. *. 1
METIOD ......... .. .. 10
RESULTS . 15
DISCUSSION . 26
SUri-'ARY ............... .. .. 31
REFERENCES .. ... .. . 3*4
BIOGRAPHICAL SKETCH .. . . 37
LIST OF TABLES
1 Number of Days of Treatment Required to
Completely Overcome Avoidance . .. .
2 Percentage Change Scores (Inches) for
Each S and Group Means . .... .
LIST OF FIGURES
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 .
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
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-
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
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
Figure lb. Treatment Surround
beginning of the experiment and this schedule was maintained throughout
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
The minimum distance that separated the S from the sugar and bear
within 90 secs. after being released was defined as the pre-treatment
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
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
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;
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
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
30 / j .'. "
C\ I I
If \ I /I I ; i
50 I .
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
percent change score = pre-treatment score--best performance 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-
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
^ ~* ^
I I 1 !
,I I I I ,l l
i l l I I I I
a s l I a .,
Last Three Treatment Measurements, Spontaneous
Recovery Measurement and Four Reconditioning
Measurements for Each Treatment S
III I Ill
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-
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
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
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,
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
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
Dean, College of Arts and Sciences
Dean, Graduate School
U Cm I
UNIVERSITY OF FLORIDA
11IIIII I1111 llll lll 1IIIII1III
3 1262 08553 6281