An Investigation of Operant Elements in Desensitization: A
Comparison of Differential Reinforcement of Other
Behaviors and Desensitization in the Reduction
of Phobic Responses in Rats
ARTHUR M. WELLS, JR.
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 acknowledge his indebtedness and to ex-
press his gratitude to those who gave so willingly of their guidance,
support, and patience. First of all appreciation is due to the
Supervisory Committee Chairman, Dr. Harry A. Grater, Jr., and to all
members of the Supervisory Committee, Dr. Benjamin Barger, Dr. Hugh
C. Davis, Jr., Dr. Mary lou McEver, and Dr. Henry S. Pennypacker.
Appreciation is also due Dr. Frederick A. Kins for his kind assistance
with experimental equipment and to Dr. Edward F. Malagodi, Jr., for
his lean of equipment and laboratory facilities.
Thanks are due also to Mrs. Irma Smith for her careful
preparation of the manuscript.
Finally, special gratitude goes to the author's wift, Anastasia,
to whom this study is dedicated. She provides an invaluable source of
meaning to this work and to the achievement it represents.
TABLE OF CONTENTS
ACKOOW;LIGEQaE7TS ....................... iii
LIST OF TABLES ..... . . . . ......... v
LIST OF FIGURES ........... . . . . .. vi
AB3TRACT ........... . . . vii
INTRODUCTION. . ............ . .. ...... 1
METHOD .. .. .. . . . . .. .. . 13
RESULTS ...... . . . . ..... 25
DISCUSSION .. . . . . .. . . .. . 'i1
RK-EI CESCI ............ . . . ..* 51
APPENDICS ......... . .... .. ... 54
APPENDIX A........... . . . . . 55
APPENDIX 3B. .... ....... ** ** ..* 56
APPIE:DIX C ......************ 57
BIOGRAPHICL SKTC . . . . . . . . 58
LIST OF TABLES
1. ONSET AN DURATION OF TONE AtD FOOD PRESENTATION IN
TREATMEIT .... . . . . . .. 19
2. ANALYSIS OF VARIANCE OF AVOIDANCE RESPONSE OF THREE GROUPS
ON THE FIFTH DAY OF ACQUISITION TRAINING . . . 27
3. DAILY TOTAL IIUMlBER OF PHOBIC RESPONSES MADE BY THE
OPERANT ANIMALS OIER FIVE TREATMENT DAYS ........ 29
4. MEANS AND STANDARD DEVIATIONS OF NUM.BE OF PHOBIC RESPONSES
MADE DURING THE FIVE DAYS OF TEL TRIALS PER DAY
EXTINCTION PHASE ............... ... 32
5. ANALYSIS OF VARIANCE OF TUO BER OF PHOBIC IESPONSES iJ)DE
BY THE THREE GROUPS OVER FIVE DAYS OF EXTINCTION ... . 33
6. POST.-HOC COIIPARISONS OF THE THREE GROUPS IN NUMBER OF
AVOIDANCE RESPONSES DURING EXTINCTION. . . . .. 34
7. MEANS A.ND STANDARD DEVIATTIOS OF CS CR LATEN0CIIS OF
THREE GROUPS OVER FIVE EXTINCTION DAYS . . . ... 38
8. ANALYSIS OF VARIANCE OF CS CR LATENCIES OF THREE GROUPS
OVER FIVE EXTINCTION DAYS. .... . . ..... . 39
9. POST-HOC COMPARISONS OF THE THREE GROUPS IN LATENCY OF
AVOIDANCE RESPONSE DURING EXTINCTION ....... . . 40
LIST OF FIGURES
1. Mean number of avoidance responses of three groups of
nine subjects over five days of conditioning ..... 26
2. Mean number of avoidance responses of three groups of
nine subjects over five extinction days . . . . 31
3. Average time between CS onset and phobic response of
three groups of nine subjects over five
extinction days ................... 37
Abstract of Dissertation Presented to the Graduate Council in
Partial Fulfillment of the Require-ents for the
Degree of Doctor of Philosophy
at the University of Florida
AI INVESTIGATION OF OPERANT ELEI.H3NTS IN DE3SISIZTIATION:
A COMPARISON OF DIF F.a:TIAL RITF1ORCIE;;T OF OTHER
BEHAVIORS AND DE3SE:SITIZATIO10 IN THE REfDUCTION
OF PHOBIC RESPONSES IN RATS
Arthur M. Wells, Jr.
Chairman: Dr. Harry A. Grater, Jr.
Major Department: Psychology
An animal analogue study was designed to compare the effects of
desensitization and operant conditioning in reducing phobic behaviors.
The intent of this comparison was to investigate oper,-nt icients in
desensitization rather than to evaluate the relative efficacy of the
On the basis of previous research an explanation of desensiti-
zation in terms of operant rather than classical conditioning principles
was thought to be feasible, leading to the following hypotheses:
1. An operant treatment procedure similar to an operant pro-
cedure possible in desensitization and presented within the framework of
desensitization facilitates the extinction of phobic behavicrs more so
than a desensitization treatment in which operant elements are minimized,
thereby demonstrating the dependence of desensitization on operant con-
2. Both the operant and desensitization proceJures facilitate
the extinction of phobic behaviors to a greater extent than does a
control procedure, thereby demonstrating treatment
To test these hypotheses, rats were conditioned to hurdle jiap
in a shuttle box to avoid electric shock which followed a 500 Hz.
tone (CS). After five days of avoidance conditioning the rats were
assigned to two treatment and one control groups, matched according
to their strength of avoidance responding.
Shock then was discontinued and treatment began. Both treat-
ment groups received equal exposure to the CS tones (phobic stimuli)
which were presented in hierarchial arrangement. Initially the treat-
ment CS was an 8,000 Hz., 66 db tone presented for two seconds, follow-
ing two seconds of a seven second food presentation. On subsequent
treatment days the CS became lover, louder and longer. By the fifth
(final) treatment day it was equal to the original CS tone used in
avoidance conditioning. Food presentation increasingly followed CS
onset in the operant procedure and was made contingent upon the animal's
making various inhibitory responses involved in confronting the CS for
increasing periods without making an avoidance (phobic) response. HoweTer,
in desensitization the tone onset followed and was concomitant with food
Following treatment all three groups received five days of
extinction, during which the number and latencies of the phobic responses
were recorded. Statistical analyses revealed the following:
1. The hypothesis of superiority of the oporant procedure as
compared with desensitization in reducing phobic behaviors was rejected.
In fact, the desensitization procedure demonstrated a stronger treat-
ment effect-though not significantly so.
2. Both treatments resulted in less resistance to the
extinction of phobic behaviors than did. the control procedure.
The results of this work are consistent with the view that the
treatment of phobic behavior may be accomplished either by rewarding
phobic object approach behaviors at the motor level or by the desensiti-
zation of anxiety at a more internal level. These results further
support a unified view of behavior and experience which would regard
changes occurring at one level of overt or covert behavior as having
implications for changes at other such levels.
A future determination of the most effective of these two
treatments could be accomplished within the experimental framework
here introduced by rcmeving certain restrictions which were placed on
the operant procedure. These restrictions were regarded as necessary
in this present study in order to make the operant treatment more
comparable to the operant procedure involved in desensitization.
Desensitization with humans could perhaps be made more effective
by employing and maximizing conditioned positive reinforcers, making
rewards contingent upon the subject's report of successful progression
through the hierarchy of threatening stimulus situations.
Desensitization techniques as introduced by Wolpe (1958) have
been successfully employed by many therapists for the relief of pri-
marily psychoneurotic symptoms (Wolpe, 1958; Lazarus, 1961; Lang and
Lazovik, 1963; Lazarus, 1963, and Paul, 1966). This study represents
an attempt to further our understanding of the processes by which
desensitization success is achieved in reducing phobic behaviors.
Fears and phobic behaviors are common distressful problems and the
treatment of these problems is worthy of investigation.
The term phobic behavior is used here in agreement with
Wolpe's (1958) definition as behavior which is persistently mal-
adaptive in view of changed environmental. conditions of reifiorceanct.
According to classical conditioning theory (Pavlov, 1927) sach phobic
behavior is established when a previously neutral stimulus situation
is followed closely by punishment. The neutral stimulus will be
referred to as the conditioned stimulus (CS) while the punishing situ-
ation will be regarded as an unconditioned stimulus (UCS). After re-
peated presentations of the CS, followed closely by the UCS, the
subject (S) reacts to the CS in a manner similar to his reaction to
the UCS. This overt avoidance behavior, termed the conditioned response
(CB), is generally accepted as a measure of fear on the part of the S
learning this phobic behavior. The persistence of phobic behavior in
response to the CS, lorg after removal of the UCS, may be regarded as
unadaptive behavior. Treatment of phobic behavior will be defined here
as any process which facilitates, or otherwise results in, the reduction
of the overt phobic behavior as a response (CR) to the CS when the
UCS ceases to occur. In this study shock (UCS) will be used to con-
dition aulmals to jump (CR) over a hurdle from one end of a shuttle
box to the other in response to a tone (CS). This jumping behavior
following onset of the tone persists long after the shock is discon-
tinued. The jumping is regarded as a phobic response rather than an
avoidance response under these conditions. A particular interest here
is with the explanation of treatment of phobic responses by desensiti-
Counterconditioning or Desensitization
Counsterconditioning is the term sometimes used in describing
experimental laboratory work with animals which is analogous to
desensitization in clinical treatment with humans. However, both terms
describe essentially equivalent processes (Wolpe, 1952; 1953). Wolps
bases his desensitization technique on early animal studies using
shock as the UCS and using eating responses on the part of the S as the
reciprocal inhibitor to the previously conditioned fear (Wolpe, 1952;
1958). This approach was similar to that employed by Watson and Rayner
(1920) and Jones (1924); a previously conditioned phobia in a child was
treated with food given to the child at progressively diminishing
distances to the phobic object.
In a similar vein, flooding (Weinberger, 1965) is the laboratory
animal study equivalent of implosive therapy in treating humans (Stampfl
and Levis, 1967). This latter treatment involves repeated and continued
presentation of the CS until the CR diminishes, in a massed extinction
The treatment procedure of desensitization, as described by
Wolpe (1958), will be employed here with animals. This treatment in-
volves having the phobic animal experience the phobic stimulus while
eating food. The CS or phobic stimulus is presented along a dimension
of stimulus generalization ranging from weakest to strongest in terms
of CR evocative potential. While the animal is eating, the CS is
presented in increasingly stronger forms until the original CS form Is
presented. In working with humans the first step in desensitization is
construction of a hierarchial arrangement of stimulus situations which
evoke the phobic behavior, arranged from weakest to strongest in terms
of their evocative potential. After reli3.ation training, the subject
is asked to iEngine these stimulus situations while remaining relaxed.
Wolpe cautions that the new response to be learned (relaxation) must be
stronger than tle unadaptive response to be elicinatmd (anxiety). Thus,
by presenting weal: portions of the CS first, and by not progressing to
a stronger CS iten until all anxiety in response to that CS item has
been reciprocally inhibited by the reciprocal inhibitor (relaxation),
desensitization will be maintained without hindrance of sensitization
(conditioning instead of removing anxiety in response to the CS). This
caution on Wolpe's part is important, though not necessarily for the
reasons which Wolpe would offer. Gale, Sturmfels, and Gale (1966) found
that though this use of progressive approximations in presenting these
threatening stimuli (CS) facilitates extinction, this method is most
effective when used in conjunction with desensitization. Goldstein's
(1967) carefully controlled study suggests that desensitization is more
effective than extinction alone LonL when progressive approximation in
presenting the CS is employed; without hierarchially arranged presentation
of the CS, there is no significant difference between extinction and
counterconditioning. Thus, the procedure of progressive approximations
in presenting the CS is of therapeutic importance, particularly in
conjunction with the desensitization treatment method.
Classical Conditioninp Exlanation of
Wolpe (1958) offers an explanation of desensitization based
primarily on classical conditioning principles. Volpe's explanation of
desensitization treatment of phobic behavior primarily involves the
"masking" or reciprocal inhibition of anxiety by a reciprocal in-
hibitor, i.e., a response state which is antagonistic to anxiety and
which prevents the occurrence of the anxiety. In desensitization, cues
which originally evcked anxiety and phobic behavior come to be associ-
ated with and come to evoke relaxation, which being antagonistic to
anxiety comes to eliminate the anxiety and phobic behavior.
Wolpe (1961),using his principles of reciprocal inhibition,
clearly suggests that in desensitization it is the inhibiting properties
of the incompatible responses (relaxation) that reduces anxiety and thus
eliminates phobic behavior: "if a response inhibitory 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
response will be weakened" (Wolpe, 1961, p. 189). Note that ';olpe
believes that :'hat is treated is the anxiety and not its outward mani-
festations in the forr of observable phobic behavior. Also note that
Welpe believes it is the associations between the CS and internal
response states, such as anxiety and relaxation, which are changed in
As previously noted, counterconditioning of laboratory animals
is the equivalent of the desensitization process. In his research
vith animals, Wolpe (1952) used eating responses as a reciprocal in-
hibitor and argued that the neural and physiological responses associ-
ated with eating and those associated with fear cannot occur simul-
taneously because the first are antagonistic to the latter. Thus, in
desensitization, eating responses reciprocally inhibit fear and thereby
decrease and eliminate those behaviors based on such fear. This study
investigates desensitization with animals employing eating responses
as the reciprocal inhibitor.
Cautela (1966) offers an explanation of the desensitization
process in terms of Pavlovian theory. While relaxed, the subject's
cortex is thought to be in a predominantly inhibitory state when the
excitatory stimulus (CS) is presented. The excitatory stimulus thus
loses its anxiety arousing potential after desensitization. This ex-
planation is similar to Wolpe's in that cue-anxiety bonds are weakened
as a result of strengthening of cue-relaxation bonds.
Recidrocal Inhibitors as Rewards
Relaxation responses are nost commonly used in desensitization
as a reciprocal inhibitor of anxiety. However, a wide variety of other
sucd inhibitors have been employed including assertive responses
(Wolpo ad Lazarus, 1966), drugs which induce relaxation (Wolpe and
Lazarus, 1966), a comforting relationship (Bentler, 1962), emotive or
enjoyable imagery (Lazarus and Abramovitz, 1962), playing responses
with children (White, 1959), and eating responses (Wolpe, 1952; 1958).
An element common to all these reciprocal inhibitors is their potential
reinforcement value in operant conditioning.
Solyom and Killer (1967) have used withdrawal of punishment
(termination of shock) instead of relaxation as a "reciprocal inhibitor"
in treating phobias. The fact that reciprocal inhibitors can also be
positive reinforcers suggests the possibility of operant processes
in desensitization treatment of phobic behavior.
OpOrant Processes in Desensitization
Desensitization involves having the subject imagine increasingly
threatening scenes while he is in a state of relaxation. The sequence
occurs as follows: relaxation--therapist's suggestion of the imagination
of an anxiety provoking inage (the CS)-subject's report of having
successfully imagined the scene without disturbance-therapist's
suggestion that the subject relax (Wolpe and Lazarus, 1966, p. 81). The
subject reports whether or not he is disturbed over imagining the scene
by signaling with raising his finger. By slowly progressing through
the hierarchial list of threatening scenes, from least to most threaten-
ing, anxiety experiences on the part of the subject can be eliminated
or kept to minimal levels.
Although Wolpe offers an explanation of this process in terms
of classical conditioning principles as previously described, an
alternate explanation in terms of operant learning principles appears
to be tenable and more parsimonious. Relaxation instructions or any
other "reciprocal inhibitor" would serve as a conditioned reinforced
in such an operant explanation. The behavior rewarded would be the
subject's self-report of having imagined a threatening scene. As
previously noted this self-report is given via finger signal in the
desensitization procedure. As the subject reports that he has suc-
cessfully imagined threatening scenes without experiencing disturbance
he is then operantly conditioned to progressively continue in this
process until he has imagined the most anxiety provoking scene in the
hierarchy. If the subject reports being disturbed in the desensitization
procedure then he is presented with previously given discriminative
stimuli in the hierarchy and must progress back through the hierarchy
again. This procedure could be regarded as operant conditioning of
finger elevatlon (representing the person's report of successful
imagination of the threatening discriminative stimulus) in the presence
of a fading procedure wherein the discriminative stimuli progress in
presentation along a stimulus dimension from least to most threatening
in terms of phobic response evocative potential.
The alternate explanation of desensitization in operant terms
involves modifying a response by bringing behavior under stimulus
control rather than by hypothesizing the reduction or elimination of
anxiety, upon which the phobic behavior is based. The sequence of
events in desensitization plus the reward potential of reciprocal
inhibitors laLe such an alternate explarntion feasible if it can be
shown that reduction of anxiety is not a necessary element in eliminating
phobic behavior. Wolpe, as previously described, attributes treatment
success of phobic behavior as due to reduction or elimination of
anxiety. The successful modification of phobic behavior using an
operant procedure would show that reduction of anxiety is not a
necessary element in the reduction of phobic behavior, though it might
still be regarded as a sufficient element.
Studies Sunportive of an Operant
Studies which have attempted to analyze and study the components
of desensitization have shown that whereas relaxation has no treatment
value alone (Cooke, 1968; Johnson and Sechrest, 1968), presentation o^
the CS in hierarchical form without using relaxation does have treatment
value (Cooke, 1968; Rachman, 1968; and Fachan and Hodgson, 1967).
Thus, it appears that relaxation to counteract anxiety is not a
necessary treatment condition; anxiety does not need to be directly
eliminated in the successful treatment of phobic behavior.
Wolpin and Raines (1966) found that having subjects simply
imagine the original full CS (most threatening scene in the hierarchy)
for more than two minutes resulted in extinction of the phobic behavior.
Implosive therapy, which involves the technique of flooding whereby
the subject is repeatedly presented with the CS in massed and prolonged
extinction trials, has been successfully used in eliminating phobic
behavior (Hogan and Kirchner, 1968; Stampfl and Levis, 1967). These
procedures also make no attempt to minimize or reciprocally inhibit
anxiety. folpo and Lazarus (1966, p. 140) believe that when such
treatment is sometimes successful, it is only because of "transmarginal
(protective) inhibition . the diminution of response that is ob-
served when stimulus intensity exceeds a certain limit." However, this
does not explain the lasting effect of such treatment to all degrees
of stimulus intensity in CS confrontation.
Reduction of phobic behavior is facilitated or phobic behavior
is successfully treated when the subject is prevented from making the
unadaptive response with repeated presentations of the CS without
any presentations of the UCS (Carlson and Black, 1959; Weinberger, 1965).
In these studies the subject was prevented from making the unadaptive
response (phobic behavior) by utilizing a barrier in a shuttle box,
where jumping front one end of the box to the other upon presentation
of the CS was defined as the phobic behavior. An operant method of
reinforcing the subject to stay in the situation when the CS is presented
without the UCS should be equally successful in facilitating climina-
tion of such phobic behavior, demonstrating the possibility of operant
processes in desensitization.
Lang and Lazovik (1963) reported that following desensitization
of snake phobia, using relaxation as a reciprocal inhibitor, some sub-
jects were able to demonstrate treatment success by touching a snake,
in spite of the fact that after treatment they reported no decrease
in their pre-treatment anxiety concerning touching a snake. This Is
exactly what would be expected in terms of an operant learning view of
desensitization technique. Other subjects who reported a decrease in
anxiety levels following treatment may have been responding to the
demand characteristics of the experimental situation (Orne, 1962) or
the expectations of the experimenter (Rosenthal, 1966). Another
possibility for decrement in anxiety is that anxiety was experienced
and then extinguished in the absence of the UCS during desensitization
Results of studies cited in this section all demonstrate that
anxiety does not have to be reciprocally inhibited in the treatment
of phobic behavior. This fact coupled with an alternate explanation of
events in the desensitization procedure in terms of operant conditioning
leads to the hypothesis that at least part of desensitization success
is due to operant procedures.
Formulation of the Hypotheses
Treatment of phobic behavior may involve facilitation of ex-
tinction processes. These extinction processes do not always operate.
When the subject makes an unadaptive response (CR) in response to the
CS, this may preclude the subject from experiencing the absence of the
UCS. Furthermore, the frequently short latency between onset of the
CS and the occurrence of the CR may serve to prevent an anxiety reaction,
thus conserving conditioned anxiety from undergoing extinction (Deltde
and Carlson, 1964 ; Solomon and Wynne, 1954). Desensitization, operant
techniques, flooding/barrier methods, and other procedures used in
treating phobic behavior are considered here to be successful to the
extent that they allow, encourage or force the subject to experience
both the CS and the lack of the occurrence of the UCS, so that extinction
Wolpe's explanation of desensitization in terms of a primarily
classical conditioning paradigm, in which anxiety must be eliminated to
eliminate phobic behavior, is questioned here. This explanation
requires inferences as to unobservable internal processes which have
not been experimentally established (Breger and McGraugh, 1965).
This present study concerns an attempt to show that phobic
behavior can be treated effectively by operant methods, concerned only
with bringing the subject's observable behavior under stimulus control,
without attempting to reduce or eliminate internal anxiety. This bears
on the question of the necessity of reciprocally inhibiting anxiety
in treating phobic behavior. The first hypothesis is that an operant
treatment procedure, presented within the framework of the technique
of desensitization, will result in significantly less resistance to
extinction compared to the disuse or controlled rest condition.
Previously an alternate explanation was offered of the desensiti-
zation process in terns of operant principles rather than in terms of
classical conditioning theory. Such an alternate explanation would
find support if an operant procedure proved to be a more effective
treatment procedure than a desensitization procedure in which operant
processes were minimally operative, i.e., restricted to a continuous
and simultaneous presentation of reward with the presentation of the
CS (so that confrontation with the CS is possibly rewarded). The
second hypothesis is that an operant treatment procedure, presented
within the frame'-ork of desensitization, will result in significantly
less resistance to extinction comparedtoadesensitization procedure, in
which operant procedures are minimally involved.
It is assumed that these treatment procedures can be applied
to the previously conditioned phobic behavior of animals in the follow-
ing way. First, using a 500 Hz. tone as the CS, a conditioned avoidance
response to shock (UCS) can be established. This response can then be
treated by using the progressive approximations method in presenting the
CS without the UCS, starting with an 8,000 Hz. tone which does not evoke
the avoidance response, then progressing to the original CS (500 Hz.
tone). Such progressive presentations of the CS, in hierarchial arrange-
ment from weakest to strongest item, can then be combined with either
a desensitization treatment or an operant treatment. In the desensiti-
zation treatment the hierarchially arranged conditioned stimuli can
be gradually introduced in increasing approximation to the original CS
while the animal is eating. The various responses involved in this
eating behavior should serve to reciprocally inhibit the anxiety aroused
by presentation of the CS tone. In the operant treatment the food is
presented in a way such that obtaining it and making an avoidance
response are mutually exclusive. In this operant procedure, obtaining
food requires that the animal inhibit his making an avoidance response,
making instead various other responses which are competitive with the
avoidance response, for increasing periods of time.
Both of these procedures can be called effective treatment
procedures only if they result in significantly less resistance to ex-
tinction than the disuse or controlled rest conditions. Thus, the
third hypothesis is that the desensitization treatment will also
result in significantly less resistance to extinction compared to the
disuse or controlled rest condition.
Thirty experimentally naive, initially 80 days old Long-Evans
male rats were used as subjects. These 30 rats were selected from a
group of 42 on the basis of their overall performance during the
Acquisition Phase. Selection was accomplished by takinC out the four
fastest and eight slowest rats in terms of average reaction tine
between CS onset and CR or hurdle jumping. In the middle of this
study three rats died, leaving data on 27 animals (see discussion
of matching Ss in oeatnent section belod).
Each animal was separately housed in a hanging, metal _age in
a controlled temperature environment. Water was given ad lib iln all
phases of the experiment. Food was presented ad lib for five days
prior to experimentation and then the animals were reduced to 80 per
cent of their normal weight. Nor'cal weight was determined by taking
the median of the daily weights over the last three days of this five
day pre-experimental period. After normal body weight was determined
the animals went on their diet for two days prior to begirning
acquisition training. They were maintained at 80 per cent of normal
body weight during the Acquisition and Treatment Phases of the experi-
nent, but were again placed on an ad lib food schedule during the
Extinction Phase of the experiment. The rationale for placing the
animals on an ad lib focd schedule during the extinction trials is
presented in the Discussion section. All animals were fed each night
at the end of the experimental work Purina Laboratory Chow was used.
The wooden shuttle box was painted black and had interior
dimensions of 24 inches long by 6 inches wide by 18 inches high. The
floor consisted of 4 copper alloy bars spaced one-half inch apart.
The grid floor was divided in half by a four and one-half inch high
wooden hurdle which was in place at all times. Only one end of the
floor was wired for shock. This same "hot" end also contained a food
tray which could be pushed into the side of the box and which was
centered between the hurdle and the end of the box.
During the Treatment Phase, 20 mg n!oyes food pellets were
presented by sliding this tray into the box. The empty tray was kept
out of the box, the end of the tray covering the box tray opening,
during the Acquisition and cEtinction rhases. A plexiglass sheet
covered a rectangular opening on one side of the box and permitted
full observation of St behavior. The other side of the box contained
numerous one-quarter inch holes to permit sound to enter from an
The shock for the grid was delivered by a Lehigh Valley
Electronics Shock Generator and Scrambler unit which reversed the
polarity randomly among the 23 grid bars on the hot side of the box.
Level of shock was set at one and two-tenths milliamperes (D.C.) as
measured by the Simpson meter on the shock unit.
An AR4x speaker was used adjacent to one side of the shuttle
box with the woofer near the "hot" end of the box. The tones were
supplied from a Hewlett-Packard, model 200AB audio Oscillator. Tones
were measured by a General Radio Company Sound-Level Meter, type 1551-B,
inside of the shuttle box with background noise at a 35 db level. A
Hunter interval timer was used to measure event duration to the nearest
hundredth of a second. Manually operated switches controlled the
shock, tone and timer. When appropriate the switches from the tone
generator could be yoked to the timer to measure the duration of the
tone. The running timer was also used to measure food presentation
During this Phase each of the 42 rats received 20 trials of
acquisition training for each of five consecutive days. On the first
trial of each day the animal was placed in the hot end of the shuttle
box and allowed to roam for a period of two minutes. The 500 Hz., 90 db
tone (the original CS) was then turned on If the animal was in the hot
end of the box. If the animal was not in the hot end after the two
minute period then he was gently placed there. After the CS tone had
been turned on for a period of ten seconds shock was delivered. The CS
tone (and shock if turned on) was immediately terminated and time
recorded as soon as the animal jumped, withdrawing all contact from the
hot grid floor. In effect then the animal quit receiving both the
shock and the tone as soon as he jumped. He did not receive shock
prior to ten seconds from the tone onset.
The latency between the onset of the CS tone and the jumping
or avoidance response was recorded to the nearest hundredth of a
second. Recording of latency was accomplished by having the CS tone
duration switch yoked to the timer so that the duration of the CS
tone (which terminated with Ss jumping clear of the hot grid floor)
was timed. A latency of ten seconds or greater indicated that the
animal did not jump before receiving shock.
On trials following the first trial of each day during the
acquisition training, the animals were allowed 20 seconds in the
"cold" end of the shuttle box following hurdle jumping and CS teriina-
tion which ended a trial. They did not receive the CS, marking the
beginning of a new trial, rntil at least ten seconds after being placed
back in the hot end of the box. Animals jumping within ten seconds of
being placed in the hot end of the box to begin a new acquisition
trial did not receive either the CS or the UCS (shock) but were gently
placed back into the hot end to begin the trial again. Following the
20 daily trials each animal was weighed, returned to his cage and fed
an amount of Purina Laboratory Chow based on the deviation of his weight
from 80 per cent of his normal weight.
By the fifth day of training, all animals reached the acquisition
criterion by hurdle jumping at least 75 per cent of the time before the
duration of the CS reached ten seconds (thus avoiding the shock).
Following the five days of acquisition training all 42 animals
were ranked according to their overall average reaction time (between
the CS onset and jumping behavior). The four animals with the fastest
reaction times as well as the eight slowest reacting animals were dis-
carded from further experimentation. Three groups of ten animals each
were then drawn for the two treatment and control groups. The Control
group was comprised of the following ranked animals: 1, 6, 8, 12, 14,
16, 20, 22, 27 and 30. The Operant treatment group was made up of
animals with these ranked numbers: 2, 4, 9, 10, 15, 17, 21, 23, 25
and 29. The Desensiti nation treatment group was composed of the follow-
ing rank numbered animals: 3, 5, 7, 11, 13, 18, 19, 24, 26 and 28.
The average reaction time over the 100 trials of acquisition training,
computed for each group of ten rats aft-r the anisials were formed into
the three groups, was as follows: Control group, 4.93 seconds; Operant
group, 4.94 seconds; and the Desensitization group, 4.94 seconds. The
groups of animals were thus thought to be well matched in terms of the
overall average time it took them to jump the hurdle upon onset of the
CS during acquisition training.
On the third day of treatment three of the experimental animals,
one from each of the three groups, were found dead. Although no autopsy
was performed the presence of blood traces near the external urinary
region of these animals suggested a urinary tract infection. The remain-
Ing 27 Ss survived the experiment, apparently in good health.
The elimination of these three animals did not appear to upset
the previous matching. The ranked positions of these Ss, in terms of
overall reaction time in the acquisition trials, were ?1, 18 and 20.
The average reaction times of the three groups changed as follows with
the elimination of the three animals: Control group, dropped from 4.93
to 4.90 seconds; Operant group, remained the sane at 4.94 seconds; and
the Desensitilation group, dropped from 4.94 to 4.93 seconds. Since
matching appeared to be maintained the study progressed as scheduled with
the remaining 27 Ss.
Treatment began the day after completion of the acquisition
trials. Both the Operant and the Desensitization treatment groups
received equal exposure to the CS stimuli components during treatment.
Both groups also received equal exposure to the food pellets used in
treatment. The tonal hierarchy presented during treatment in accord..
ance with the method of progressive approximations consisted of the
following five tones: an 8,000 Hz., 66 db tone presented for two
seconds; a 4,000 Hz., 72 db tone presented for three seconds; a 2,000
Hz., 78 db tone presented for four seconds; a 1,000 Hz., 84 db tone
presented for five and four seconds; and a 500 Hz., 90 db tone (original
CS) presented for four, six and seven seconds. These tones were pre-
sented over the five day Treatment Phase period in a total of 84 treat-
ment trials. Food was also presented. The timing of the food presenta-
tion in relation to the tone presentation defined and distinguished the
Operant and Desensitization treatment groups. Table 1 provides a
detailed description of the tonal hierarchy and the presentation of
these CS tones in relation to food pellet presentation in the two treat-
,. 4-1 .
N CI CI C 0- N
I I I I I a
CN C\ ClN N -
N I I I I I
I I I N N
t .-"I .C
N C -4
0 0) -^
o' 4u cS
f rt-i (
o 'C C
I I I
1 0 e'
3 V3 I
~t 0 C> '-
N ,- .
N 'r\ '3
0, \0 *N 0,C
8 NhO r-I
As can be seen from Table 1, both treatment groups received the
same stimuli conditions during the 16 trials of the first treatment day
and the first four trials of the second treatment day. On the first
treatment day each animal was initially put into the hot end of the
shuttle box (no longer receiving shock, however) and immediately the
filled food tray was pushed inside the box. The animal was allowed
either to eat or to roam freely, but as soon as he had eaten from the
tray for 30 seconds the food tray was withdrawn. After a ten second
delay the tray was reintroduced and two seconds after the animal had
again started eating the 8,000 Hz., 66 db tone was turned on for two
seconds. After completion of this tone the food tray remained in the
box for an additional three seconds and then was withdrawn. There was a
ten second interval between trials and S was always in place in the hot
end of the box at the beginning of a trial. These first 20 trials con-
bined both operant and desensitization treatment elements but on subse-
quent and succeeding trials food presentation became increasingly de-
layed in the operant treatment in a shaping procedure designed to rein-
force S's confronting the CS stimulus component, making various other
and unmeasured competitive responses, without making the phobic response.
The Desensitization group never received the CS tone without prior and
concomitant food presentation.
Beginning with the twenty-first trial, which was the fifth
trial of the second treatment day, timing of the onset of food presenta-
tion began with the introduction of the food tray and not, as previously,
with initiation of eating behavior on the part of S. It is impossible
to reliably obtain a constant delay of food presentation relative to
previoustone onset when food presentation is defined in terms of S's
initiation of eating behaviors, but food presentation defined in terms
of food tray introduction can easily be delayed at a fixed period after
a tone onset in a consistent and accurate fashion. There was no
assurance that S would begin eating as soon as the food was presented.
in the initial trials because though deprived he had no previous ex-
perience with the food's availability. This made it necessary to
define the start of food presentation in terms of S's initiation of
eating behaviors in these first 20 trials. Stimulus conditions in
these first trials made such a definition quite feasible in regards to
procedural reliability. After the first 20 trials, So had all become
acquainted with the food tray's location and availability such that
introduction of the food tray was practically equivalent to S's
initiation of eating behavior.
After the first treatment day all treatment trials were con-
ducted in the following manner. First an animal was introduced into
the hot end of the shuttle box and after a period of at least ten
seconds, and while the animal was in the hot end of the box (being
gently put back there if he had roamed over the hurdle during the ton
second period), the treatment stimuli were introduced. In the case of
the operant procedure the treatment stimuli introduction consisted of
the onset of the CS tone component followed later by food presentation
which was contingent upon the animal's confronting the CS without
making an avoidance response. In the desensitization method the onset
of the CS tone component followed the food presentation.
As treatment progressed each S was exposed to tones which
increasingly approached the original CS tone in terms of frequency or
pitch, volume level and duration. Thus, after beginning treatment
with a relatively soft, high-pitched and short tone, which had very
weak CR evocative potential, treatment ended with the loud low-
pitched and long lasting tone used as the CS in the acquisition
training. No shock was administered to any animal at any tine follow-
ing the completion of the acquisition training; thus hurdle jumping on
the part of S in response to the tone fulfilled the experimental
definition of a phobic response. Food was presented for a period of
seven seconds in all trials except the last eight trials where it was
presented for eight seconds. The stimulus conditions for the two treat-
ment groups are described more precisely in Table 1.
During the Treatment Phase the control animals were transferred
from their cages to another cage prior to their daily weighing. This
was done to equate all animals for handling. The control animals were
not otherwise manipulated and were kept in the same isolated environment
receiving the saeo stimulus exposure as the treatment animals, except
of course for the treatment stimuli.
The extinction trials began on the day after the fifth (final)
treatment day, and consisted of ten trials per day for five consecutive
days. These trials were similar to the acquisition trials in most
respects. Of course neither UCS (shock) nor food was presented as was
the case respectively in the acquisition and treatment trials. Initially
on each day of extinction the animal was placed in the hot end of the
box, then allowed to roam for two minutes. During extinction the
empty food tray was always pulled in the "out" position so that the
end of the tray covered the tray opening in the side of the shuttle-
box. At the end of the two minute exploratory period the animal was
placed again in the hot end of the box (if he was not already there)
and the CS tone (500 Hz., 90 db) was presented and concomitantly was
timed. As soon as the animal made a phobic response (by withdrawing
all physical contact from the grid floor on the hot side of the box)
the CS tone was terminated and simultaneously the timer stopped. If
the animal did not jump at the end of 20 seconds, chosen as the
infinite latency period, then the CS tone was terminated and his
latency was recorded as 20 seconds (indicating that no phobic response
was made). This procedure provided measures for the two dependent
variables, e.g., latency or delay between onset of the CS and the
occurrence of a CR, and whether or not the phobic response or CR was
made. After the first trial of each extinction day S was allowed to
freely roam for a period of 30 seconds before the CS presentation for
the next trial. If an animal made a CR during this pre-trial period
then ten seconds later he was placed back into the hot side to begin
the 30 second exploratory inter-trial period again.
Data obtained from each of the three phases of this study
will be presented in sequential correspondence to the experimental
The latency between CS onset end CR occurrence was recorded
over the 100 trials (20 trials per day) of acquisition training. A
recorded latency of ten seconds or greater indicates that shock was
administered on that trial. These data then provide two measures of
CR acquisition strength in the speed with which a CR was evoked and
in the frequency of CR occurrence prior to UCS onset (or within ten
Figure 1 demonstrates these data graphically showing the mean
number of avoidance responses (CR's) made by each of the three groups
for each day of acquisition training.
Inspection of Figure 1 indicates similarity of conditioning
rates among the three groups. All of the animals achieved the criterion
of avoidance training by making a CR prior to UCS onset at least on
15 of the final 20 acquisition trials.
A one-way analysis of variance on the number of avoidance
responses for the final day of acquisition training supported the
contention that the three groups of Ss had been successfully matched
1 2 3 4 5
Acquisition Phase: Five Days, Twenty Trials Per Day
Fig. i.-Mean number of avoidance responses of three groups of
nine subjects over five days of conditioning.
on the strength of acquisition variable. This analysis failed to
indicate significant differences anong the three groups in uunber of
avoidance responses for the last 20 acquisition trials, as can be
seen in Table 2 below.
ANALYSIS OF VARIANCE OF AVOIDANCE RESPONSE OF THRE GROUPS ON
THE FIFTH DAY OF ACQUISITION TPJINING
Sources of Variation df Mean Square F
Between 2 .45 .44
Error 24 1.03
df (2, 24) .05 = 3.410 .01 = 5.61
Matching was accomplished not on the basis of the number of
avoidance responses made prior to the UCS, but rather on the latency
of these avoidance responses. The average CS CR latency by group
over all five days of acquisition training was as follows: Operant
group, 4.94; Desensitization group, 4.93; and the Control group, 4.90
seconds. These mean latencies for the fifth and final day of acquisition
training were as follows: Control group, 3.63; Operant group, 3.27; and
Desensitization group, 3.22 seconds. Note that although the Control
group demonstrates the fastest overall mean reaction time, indicating
strong acquisition of the avoidance response, the Control group's
reaction time for the fifth and final day of acquisition training (just
prior to treatment attempts to increase this reaction time) is the
slowest of the three groups. As in the case of the number of avoidance
responses made during acquisition, the speed with which these responses
were made is essentially equivalent among the three experimental groups.
The only recording made during the treatment trials was the
number of avoidance or phobic responses made by Ss during CS tone
component presentation. In the Desensitization group, which received
food prior to and continuing throughout exposure to the CS tones, only
two of the nine animals made phobic responses. These two animals only
made one such response each for a total of two, and these two responses
were both made only on the first treatment day. In the Operant group,
which received food usually seconds after tho onset of the CS tones,
eight of the nine Ss made a total of 21 (of a possible 756 total
trials) phobic responses. Table 3 describes the distribution of these
21 phobic responses over the five treatment days. Examination of the
data in Table 3 reveals that the operant fading procedure employed was
generally successful in maintaining the desired behavior, and that a
small percentage of phobic responses occurred.
Both treatments appear to have been accomplished successfully.
The nearly total prevention of CR occurrence in the desensitization
procedure attests to the effectiveness of the hierarchy used. Probably
the only way to have eliminated CR occurrence in the operant treatment
would have been to make food presentation less delayed relative to CS
onset, but this would have blurred the distinction between the two
DAILY TOTAL NUMBER OF PHOBIC RESPONSES MADE BY THE OPEBRAT ANIMALS
OVER FIVE TRElTMiET DAYS
Treatment Day Number and Nunber of Treatment Trials
Per Each Animal
Day 1 Day 2 Day 3 Day 4 Day 5
Animal 16 Trials 16 Trials 16 Trials 16 Trials 20 Trials
1 0 0 0 0 1
2 1 1 2 1 0
3 0 0 0 0 2
4 2 0 1 0 3
5 0 0 0 1 2
6 1 0 1 0 0
7 0 0 0 0 0
8 0 0 0 0 1
9 0 0 0 0 1
treatment groups; the operant treatment would have become more of a
Treatment effects were measured during the extinction trials
both by the number of phobic responses made and by the latency in
responding. These data can be found in Appendices A, B and C.
Considering the number of avoidant responses made by the three
groups in the extinction curves in Figure 2, differences among the
groups are readily noted. These differences are further reflected
in the means and standard deviations of the avoidance responses as
contained in Table 4. Both treatment groups made less phobic responses
during post-treatment extinction trials than did the nontreated con-
trols. This difference was most evident on the first day of extinction,
immediately following the last treatment day. A two-way, mixed model
(Hays, 1966) analysis of variance is summarized in Table 5. These
results require the rejection of the hypothesis that the three groups
were sampled from the same population. To further describe differences
attributable to treatment, a Newman-Keuls test is presented in Table 6.
These results indicate that both the Desensitization and Operant treat-
ment groups are significantly different from the Control group, but
that the Operant and Desensitization groups are not significantly
different from each other. Thus, the significant treatment effect is
primarily due to differences between the treated and untreated groups.
Changes over "days" of the extinction trials are to be expected; the
significant effect of this variable is of little concern to this
1 2 3 4 5
Extinction Phase: Five Days of Ten Trials Each
Fig. 2.-Mean number of avoidance responses of three groups of
nine subjects over five extinction days.
MEANS AND STAhDARD DEVIATIONS OF sITIBER OF PHOBIC RESPONSES lADE
DURING THE FIVE DAYS OF TEN TRIALS PER DAY EXTINCTION PHASE
Group 1 2 3 4 5
Desensitization T 3.2 4.7 6.8 5.4 5.6
SD 3.18 3.64 2.47 1.72 2.70
Operant X 4.2 6.0 7.3 6.7 6.7
S 3.35 2.58 2.99 2.79 2.94
Control X 8.9 8.6 8.7 9.3 9.4
SD .89 1.61 1.59 1.03 1.14
ANALYSIS OF VARIANCE OF LUMBER OF PHOBIC RESPOISSS MADE Bf THE
THREE GROUPS OVER FIVE DAYS OF EXTINCTION
Source of Variation df Mean Square F
Trcatnent 2 177.80 36.14-a
Days 4 19.85 2.93*b
Interaction 8 4.92 .73c
Error 120 (128)d 6.90 (6.78)d
adf (2, 8) .05 = 4.46 .01. 8.65
bdf (4, 128) .05 = 2.42
Cdf (8, 128) .05 = 2.01
dPooled MS error, used to test Days and Interaction effects
POST-HOC COIMPA~PSONSa OF THE THREE GROUPS IN NUMiBE OF AVOIDANCE
RESPONSlS DIUrIING EXTINCTION
zation Operant Control
Treatment Totals 231 278 404
Differences 47 126
t/ H3error 2.72b 7.15-'c
aNewman.ecum s Test
bQ.95 (2, 120) = 2.80
CQ.99 (2, 120) = 3.70
investigation. It should be noted that complete extinction did not occur,
and that the treatment effects were measured during but not throughout
the extinction process. The "days" factor in the analysis of variance of
number of phobic responses was regarded statistically as a random factor,
sampled from all possible extinction days.
Turning now to the second dependent variable, phobic response
latency, differences among the three groups can be seen in the ex-
tinction curves of Figure 3. These differences are most evident, as
in the case of the phobic response occurrence variable, immediately
after treatment on the first day of extinction. Also, as in the case
of the first dependent variable, both treatment groups appear to be
clearly superior to the Control group in decrement of phobic behavior
and the densensitization treatment appears perhaps more effective than
does the operant procedure. The means and standard deviations of these
latency data are presented in Table 7. Analysis of the variance of
these results, summarized in Table 8, again as in the case of the
first dependent variable, demonstrates a strong treatment effect. A
two-way, mixed model analysis was employed with an error term for test-
ing interaction and row effects (the random effects variable) derived
by pooling the interaction and error mean squares, again as was done in
the first dependent variable analysis of variance.
To more precisely delineate the effect of treatments on latency
of the phobic response during extinction another Newman-Keuls test was
calculated (Table 9). These results again parallel those obtained with
the CR occurrence data previously described in that either the Desensiti-
zation or Operant groups were significantly slower in responding to the
CS than was the Control group, but that the two treatment groups mis.
being significantly different.
The effect of desensitization and operant procedures in the
reduction of two behavioral indices of phobic behavior appears to be
highly facilitative as compared with no treatment procedure at all,
but the hypothesis that the two successful treatment procedures
differed in effectiveness escaped rejection and the differences found
in this study must await future investigation for clarification.
Operant o-- - 0-
16 Desensitization -..........
S 12 '" "
1 2 3 4 5
Extinction Phase: Five Days of Ten Trials Each
Fig. 3.-Average time between CS onset and phobic response of
three groups of nine subjects over five extinction days.
fANS AND STANDARD DEVIATIONS OF CS CR LATENCIES OF THREE
GROUPS OVER FIVE EXTINCTION DAYS
Group 1 2 3 4 5
Desensitization y 16.09 14.83 11.97 13.63 12.44
s 4.68 5.35 3.69 3.50 6.03
Operant i 15.24 13.73 11.10 12.31 11.16
SD 3.91 3.45 4.08 3.48 4.01
Control y 7.55 9.72 8.87 8.85 7.09
SD 2.18 4.33 2.74 3.58 4.58
ANALYSIS OF VARiUACE OF CS CR L~TENCIES OF THREE GROUPS
OVER FIVE EXTINCTION DAYS
Source of Variation df Mean Square F
Treatment 2 356.08 28.72**
Days 4 39.08 2.20b
Interaction 8 12.40 .70c
Error 120 (128)d 18.12 (17.76)d
adf (2, 8) .05 = 4.46 .01 = 8.65
bdf (4, 128) .05 = 2.42
Cdf (8, 128) .05 = 2.01
dpocled MS error, used to test Days and Interaction effects.
POST-H0C COIQtARISO"iS OF TEt THREE GROUPS IN LATENCY OF
AVOIDAIXCE RESPONSE DURING EXTINCTIONI
zation Operant Control
Treatment Totals 620.63 572.85 381.43
Differences 47.78 191.42
is error 1.67a 6.70 -b
120) = 2.80
120) = 3.70
Analysis of the acquisition data indicates that successful
matching of groups was achieved in this investigation so that the three
groups of animals had approximately equal conditioned avoidance
response tendency prior to the treatment phase. Thus the three groups
were probably nearly equal in the strength of their phobic response
tendency, and were probably also nearly equal in their ability to
"unlearn" the previously conditioned avoidance response, when treat-
Data obtained during the Treatment Phase indicate that both
treatments changed the rate of the avoidance responses of the eaimals
during treatment trials. Both treatment groups received a nearly
equal exposure to the CS throughout this experiment, so that the treat-
ment effect involved in CS exposure was equivalent among treatment
groups. Treatment effects due to extinction were considered minor since
the phobic behaviors of the Control group on the fifth (final) day of
extinction trials were nearly equal to those on the first day of ex-
tinction, and since the phobic behaviors of the Control group on the
final day of extinction were much greater than those of either the
Operant group or the Desensitization group on the first day of extinction.
Both treatment groups also received an equal opportunity to eat during
treatment. All Ss were maintained on similar food deprivation schedules
prior to and during the first two phases of this study, and all the
groups were fed ad lib during the final phase.
The reason for changing to an ad lib food schedule in the final
experimental phase was to achieve what was thought to be a more pure
measure of the effects of treatment by reducing deprivation effects on
response maintenance. Under food deprivation conditions the operant
treatment would have measured out as more effective than was the case.
The fact that the measures of phobic behavior on the final two days
of extinction, when any food expectation should have extinguished to a
large degree, are very similar to these same measures over all five
days of extinction suggests that food expectation played a minor role
in the effects of treatment (as measured with vell-fed animals).
Effects of operant proceduroc are typically measured after
removal of previous reinforcement, but not necessarily while needs
satisfied by the reward are reduced to a state of satiation. In attempt-
ing to measure more lasting effects of treatment, rather than temporary
and perhaps artificial effects, the operant treatment was Beapered in
The operant procedure employed was a discriminative differential
reinforcement of other behaviors. The desired event in the operant
procedure, not jumping, involved a variety of inhibitory responses
(largely superstitious in nature). It was these varied urneasured
inhibitory responses which were actually reinforced. These reinforced
responses, which were somewhat incompatible and antagonistic to the
phobic response, were brought under the control of discriminative
stimuli presented along stimulus dimensions (frequency, loudness, and
duration of tone) through a fading procedure. The tone came to serve
as a discriminative stimulus for food presentation in the operant
procedure. During the extinction trials, when the animals were satiated
for food, the tone was irrelevant as a discriminative stimulus, making
the operant animal nuch more likely to make the phobic response than
would have been the case had the food-deprivation schedule been continued.
However, by providing all animals with food during the Extinction Phase,
the operant treatment procedure more closely approximated the hypothe-
sized operant elements in the desensitization treatment. The desensitized
animal demonstrates his treatment effect whether food deprived or food
satiated and any operant procedure involved in the desensitization
treatment must also stand up under food-satiation conditions.
A one-way shuttle box was used Each that one side consistently
received shock during acquisition, food during treatment, and the CS
during all experimental phases. If a two-'.ay shuttle box had been
used, in which both sides alternately received shock, then both ends
of the box would have taken or an aversive quality for the anzials.
This potential source of confusion for the animals was avoided, though
thereby probably making treatment more difficult, by consistently offer-
ing the animals a sanctuary in the no-Ehock end of the box. Thus the
quick phobic response was always clearly associated with avoidance of
shock, and was always an available course of action to the animal.
The operant treatment involved the reinforcement of responses
involved in an animal's confronting the CS without making a phobic
response. These reinforced responses were shaped in that their occurrence
over increasing periods of tine was required in order to obtain food.
Operant treatments usually attempt to increase the rates of more active
and positive responses. However, in the experimental procedure employed
here, the reinforcement of a single, active response might have intro-
duced desensitization elements in that the active responses would have
involved inhibition of anxiety. Wolpe (1958) describes hoIT an active
motor response can reduce anxiety, even though this motor response
does nothing toward the avoidance of the threatening stimulus. The
choice of rewarding a passive CS confrontation response avoided these
problems Paid thus more clearly offered a test of motor level operant
conditioning versus more central (perhaps) autonomic level anxiety
inhibition. If an active, positive, clearly defined competing motor
response had been positively reinforced then the operant treatment would
probably have measured out as more effective than was the case. Eut
again this does not correspond to the hypothesized role of an operant
procedure in desensitization in that desensitization does not involve
the occurrence of an active, positive noter response. In human de-
sensitization, finger raising serves as a signal rather than as a compet-
ing response antagonistic to the phobic response. Thus, to make the
operant procedure most comparable to the operant procedure that might
operate in desensitization, no clearly defined rewarded response was
Results from the Extinction Phase suggest that contrary to
our hypothesis the operant treatment is not superior to the desensitization
procedure. However, and as hypothesized, both treatments were signifi-
cantly effective as compared to a no treatment condition. Before dis..
cussing possible implications of these results it must be pointed out
that the measures of treatment effect were taken from extinction trials
that did not continue to completion. The question arises as to whether
the treatment differences demonstrated would have continued in the
same form if the trials had gone on to total extinction. Statistically
this was controlled to an extent by regarding the "days" factor in the
analysis of variance as having a random rather than fixed effect,
thus taking the five days measured as a sample rather than as exhaustive
of the universe of "Days." It is most likely that data obtained from
all possible extinction trials would have yielded the same results as
the data actually obtained. The most probable exception to this is
that the desensitization treatment night have been shown to be sig-
nificantly superior to the operant procedure employed in terms of
reducing number of phobic responses made, since this was nearly so
with the data obtained.
A major finding in this study is that an operant procedure,
which most closely approximated such a possible procedure in desensiti-
zation, was found to be slightly less effective (though this was not
statistically significant) than a desensitization procedure in the
reduction of phobic behaviors. All other sources of treatment effect,
not specifically and uniquely a part of the two treatment procedures
employed, were equivalent in the two treatments. Such sources of
treatment effect which were equivalent in both treatments include
(1) extinction possibilities, (2) change from conditioned aversive
qualities to conditioned adient qualities of the hot end of the shuttle
box through food presentation in treatment, and (3) opportunities for
making various motor responses. The CS tone was of different stimulus
value for the two treatment groups, serving as a discriminative stimulus
for food presentation in the operant procedure while serving as a CS
for response states involved in eating (presumedly incompatible with
previously conditioned anxiety response states) in the desensitization
procedure. However, these differences in the tone's value, primarily in
its value as a conditioned reinforcer, are differences inherent in the
differences between the two treatment procedures. And whereas the CS
tone's value as a discriminative stimulus was much reduced during the
extinction trials (when the food schedule became ad lib) this again
more closely approximated any possible operant procedure involved
in desensitization. Therefore, the desensitization treatment effect
may involve some possible operant conditioning treatment effect, but
it is not dependent upon such operant elements. At any rate this study
supports the view that either the operant or the desensiti nation pro-
cedure is similarly effective in reducing phobic behavior. This find-
ing has some implications for further elucidation of the desensitization
process and for treatment of phobic behavior in general.
In regard to the nature of the desensitization process, contrary
to the view initially offered here, it appears that the desensitization
treatment effect involves more than either rewarded confrontation
with the phobic object or reinforced induction of competing motor
responses. The operant treatment contained these two elements plus the
additional advantage of a more efficient use of reward in using fading
and shaping procedures designed to more strongly condition motor responses
than could be accomplished by the continuous and simultaneous "reward"
offered as part of the desensitization treatment. The additional
treatment effect element, which is necessarily a part of the desensiti-
zation treatment to a greater extent than in the operant treatment,
could be a deconditioning of anxiety previously conditioned to the
stimulus represented by the phobic object. This deconditioning might
occur as a result of two processes: the weakening or reciprocal in-
hibition of conditioned bonds between anxiety affect and the phobic
stimulus, and a simultaneous strengthening or reconditioning of bonds
between the phobic stimulus and pleasant affect (a process which could
also serve to reciprocally inhibit anxiety). This night involve a
response state substitution to the sane stimulus via classical con-
But regardless of such conjecture, it doos appear that phobic
behavior can be treated by changing motor behavior in the direction of
approach rather than avoidance, without dealing with anxiety directly.
Phobic behavior treatment is also enhanced by reducing phobic anxiety,
perhaps by a process involving the substitution of conditioned adient
affect for previously conditioned aversive affect.
There is currently a continuing debate between behavior modifi-
cation learning theorists and analytically oriented theorists regard-
ing the necessity of dealing with internal "causes" in treatment of
neurotic symptoms (Ullmann and Krasner, 1965). Related to this issue
is the differing view of phobias held by psychoanalysts and behavior
modifiers. Psychoanalysis traditionally regards phobias as generated
from anxiety which is displaced onto the phobic object, whereby avoid-
ance of the phobic object while being fearful of that object becomes a
means by which the anxiety is kept in check. Analytic treatment of
phobias is accomplished through uncovering the source of the anxiety
in terms of the unconscious dynamics and intrapsychic conflicts in-
volved. After this underlying material is brought into full cognitive
and emotional awareness then the anxiety is dispelled and the phobia is
reduced. In contradistinction to this view, the behavior modification
approach would regard phobias in terms of behavior that originates
from prior aversive conditioning which is best treated by elimination of
the phobic behavior or modification of verbal behavior. Thus Eysenck's
(1960) dictua: "Get rid of the symptom and you have eliminated the
neurosis" (p. 9).
Wolpe, in arguing for the progressive deconditioning or in-
hibition of the phobic anxiety in treating phobic behavior stands some-
where between the analyst, who would resolve the underlying conditions
which give rise to the displaced phobic anxiety, and the behavior
modifier, who would change the phobic (notor level) behavior. The
differences in these three approaches to the treatment of phobic behavior
can then be regarded in terms of the level of the phobic experience at
which attempted change is primarily focused. The results of this
research suggest that either motor level behavior or more central
levels of response states can be the primary change target in effective
treatment of phobic behavior. However, it must be remebered that our
definition of phobic behavior is based on a learning theory model and
not on the psychoanalytic "phobia." There may well be different types
of phobias corresponding to the differing definitions of phobias found
in psychoanalysis and learning theory.
Treatment of phobic behavior may best be accomplished by in-
ducing changes at all levels of experience and behavior. But in
recent years there has been increasing support for the view that be-
havior, however conveniently studied at isolated and restricted levels,
is very much unified.
Social psychologists, particularly the cognitive dissonance
researchers, have well demonstrated how changes in overt behavior can
bring about internal changes in cognition, attitude, and affect. The
operant animal in this study who was rewarded for confronting threat-
ening stinuli probably experienced a fear reaction which likely under-
went some extinction so that changes in affect followed motor level
changes. Likewise, the subject whose fear was diminished was enabled
to make subsequent rotor level changes. Freud (1925) found it most
helpful in treating phobias to make his phobic patients expose them-
selves to previously avoided and painful situations. The results of
this study are consistent with a unified view of behavioral change
such that change at any level of experience has important implications
for subsequent changes at other levels of overt and covert behavior.
The method presented here, of equitably comparing operant with
desensitization methods, would seem worthy of further extension and
exploration. Hopefully, with more sophisticated measuring techniques
and more refined procedures, future research will be able to demonstrate
how much the treatment of phobic behavior depends on the resolution of
anxiety engendering deep-seated conflicts versus reduction in specific
conditioned anxiety versus change in motor level behavior. It might
also be possible to more precisely determine ho;w much desensitization
depends on the inhibition of general anxiety versus specific conditioned
anxiety, and to determine how much desensitization involves instrumental
conditioning of either approach behaviors or competing motor responses.
Such research is particularly neededvith human subjects.
Although the purpose of this study was not to determine the
comparative efficacy of the two treatments employed, such an evaluation
could be accomplished within the experimental framework here intro-
duced. In all fairness to the operant procedure, such a study would
require tie reinforcement of a clearly defined, active motor response
(competitive to the phobic response). In the same vein, food deprivation
conditions should continue throughout the extinction trials in any
such treatment comparison.
Since operant procedures are effective in bringing about con-
frontation with a phobic stimulus, desensitization with humans could
perhaps be modAe nore effective by maximally employing verbal reinforce-
ment which could be iade contingent upon the subject's self-report of
having imagined an item in the hierarchy of threatening stimulus situ-
ations. Therapist reinforcement of the subject's actual confrontations
with the phobic object could also be combined with desensitization in
an attempt to bring about changes at many levels of behavior and ex-
periencein eliminating human phobias.
Bentler, P. An infant's phobia treated with reciprocal inhibition
therapy. Journal of Child Psychology and Psychiatry, 1962, 3,
Breger, L. and McGraugh, J. L. Critique and reformulation of "learning
theory" approaches to psychotherapy and neurosis. Psycological
bulletin, 1965, 338-358.
Carlson, V. J. and Elack, A. H. Traumatic avoidance learning: note
on the effect of response prevention during extinction.
Psychological Renorts, 1959, 5, 409-412.
Cautela, J. The Pavlovian basis of reciprocal inhibition therapy.
Conditioned Reflex, 1966, 1, 293-330.
Cooke, G. Evaluation of the efficacy of components of reciprocal in-
hibition psychotherapy. Journal of Abnormal Psycholog, 1983,
Cooper, A., Furst, J. 2,, and Bridger, '. H. A brief conmentary on the
usefulncs of studying fears of snakes. Jourrail of Abnoral
PsicpoloL, 1969, 74, 413-414.
Delude, L. A. and Carlson, N. J. A test of the conservation of anxiety
and partial Irreversibility hypotheses. Canadian Journal of
Psychology, 1964, 18, 15-22.
rseonck, H. J. (Sd.) Behavior therapy and the neurosis. Oxford:
Pergamon Press, 1960.
Freud, S. Turnings in the way of psychoanalytic therapy. In Collected
anSerS,, Yol. 3, 392-402. London: Hogarth Press, 1925.
Gale, Diana, Sturofels, Gloria, and Gale, E. N. A comparison of
reciprocal inhibition and experimental extinction in the
psychotherapeutic process. Behavior; Research and Therany,
1966, 3, 209-219.
Goldstein, A. J. A comnrison of two methods of overcomingavoidance
behavior. Doctoral disoertation, University of 'lorida, 157.
Eays, W. I,. Sttlistics for sycholo-ists. New York: Holt, Rinehart
anid ;iston, 193
Hogan, R. and Kirchner, J. Implosive, electric, verbal and bibliotherapy
in the treatment of fears of snakes. Behavior Research and
Therapy, 1968, 6, 167-171.
Hussian, A. Behavior therapy using hypnosis. In Wolpe, J., Salter, A.
and Rayna, L. J. The conditioning therapies. New York: Holt,
Rinehart and Winston, 1964.
Johnson, S. and Sechrest, L. Comparison of desensitization and progres-
sive relaxation in treating test anxiety. Journal of Consult-
in and Clinical Psychology, 1968, 2, 280-2 '6.
Jones, Mary C. A laboratory study of fear: the case of Peter. Journal
of Genetics and Psycholoj, 1924, 31, 308.
Lang, P. and Lalovik, A. The experimental desensitization of a phobia.
Journal of Abnormal and Social Psychologr, 1963, 66, 519-525.
Lazasus, A. Group therapy of phobic disorders by systematic desensiti-
zation. Journal of Abnormal and Social Pschology, 1961, .3,
Lazarus, A. and Abrmovitz, A. The use of "emotive imagery" in the
treatment of children's phobias. Journal of Mental Science,
1962, 108, 191-195.
Lazarus, A. The results of behavior therapy in 126 cases of severe
neurosis. Behavior Research and Therapy, 1963, 0, 69-79.
Orne, M. On the social psychology of the psychological experiment:
with particular reference to demand characteristics and their
implications. American Psychologist, 1962, 17, 776-783.
Pavlov, I. P. Conditioned reflexes. London: Oxford University Press,
Paul, G. Insipht versus desensitiation in psychotherapy. Stanford:
Stanford Press, 1966.
Rachman, S. The role of muscular relaxation in desensitization therapy.
Behavior Research and Therapy, 1968, 6, 159-166.
Rachmni, S. and Hodgson, R. Studies in desensitization: IV. Optimum
degree anxiety reduction. Behavior Research and Therapy,
196?, J, 249-250.
Roseathal, R. Exorimenter effects in behavioral research. New York:
Solomon, R. L. and Wynne, 1. C. Traumatic avoidance learning: the
principles of anxiety conservation and partial irreversibility.
Psychological Revie, 1954, 61, 353-395.
Solyom, L. and Miller, S. Reciprocal inhibition by aversion relief
in the treatment of phobias. Behavior Research and Therapy,
1967, i, 313-324.
Stampfl, T. and Levis, D. Essentials of implosive therapy: A learn-
ing theory based psychodynamic behavioral therapy. Journal of
Abnormal Psycholooy, 1967, 72, 496-503.
Ullmann, L. P. and Krasner, L. (ils.) Case studies in behavior
modification. New York: Holt, Rinehart and Winston, 1965.
Watson, J. B. and Rayner, Rosalie. Conditioned emotional reactions.
Journal of Fxnerimental Psychology, 1920, 3, 1-14.
Weinberger, N. M. Effects of detainment on extinction of avoidance
responses. Journal of Comnarative and Physiological sycholo ,
1965, 6, 135-138.
White, J. The use of learning theory in the psychological treatenr.t
of children. Journal of Clinical Psychologg, 1959, 15,
Winer, B. J. Statistical Drincioles in experimental desisn. New
York: IMcraw-Hill, iT -.-
Wolpe, J. Experimental neurosis as learned behavior. British Journal
of Psycho 1952, 43, 192-206.
Wolpe, J. Psychotherapy byrecinrocal inhibition. Stanford: Stanford
Press, 1?5 .
Wolpe, J. The systematic desensitization treatment of neuroses. Journal
of Nervous and Mental Disease, 1961, 132, 139-203.
Wolpe, J. and Lazarus, A. Behavior therapy techniques. New York:
Pergamon Press, 196.
Wolpin, 1. and Rales, J. Visual imagery, expected roles, and extinction
as possible factors in reducing fear and avoidance behavior.
Behavior Research and Therany, 1966, 4, 25-37.
APPE DIX A
EXTINCTION PHASE: DAILY AVERAGE IATENCY RECORDED TO THE NEAREST
rHUNDPETH OF A SECOND AND SUICED NUMBER (GIVER PAiRET.HETICALLY)
OF AVOIDANCE RESPONSES MADE BY THE CONTROL GROUP
OF TEN TRIALS OVER FIVE DAYS
Animals 1 2 3 4 5
1 10.85 (8) 11.32 (8) 9.01 (9) 12.38 (9) 5.56 (10)
2 4.42 (10) 5.38 (10) 8.73 (9) 8.76 (9) 6.46 (9)
3 5.69 (9) 1.82 (10) 5.68 (10) 3.43 (10) 1.94 (10)
4 7.58 (10) 9.72 (10) 10.40 (10) 9.74 (9) 7.88 (10)
5 5.79 (9) 7.98 (10) 3.93 (10) 7.04 (10) 2.48 (10)
6 6.89 (9) 12.74 (6) 12.50 (6) 8.53 (10) 12.22 (9)
7 9.75 (8) 14.01 (6) 6.41 (10) 12.89 (10) 9.55 (9)
8 10.57 (8) 14.65 (8) 13.07 (5) 10.82 (8) 11.10 (8)
9 6.38 (10) 9.86 (10) 9.90 (9) 8.82 (9) 6.60 (10)
EXTINCTION PHASE: DAILY AVERAGE LATENCY RECORDED TO THE NZEIUST
HUNDMREDTH OF A SECOND A D SUIMIED IUMBER (GIVEl PARE'iTTICALLY)
OF AVOIDANCE RESPOlSIS MADE BY TID OPERANT GROUP
OF TEN TRIALS OVER FIVE DAYS
~ ____n__ ___Days __________
Animals 1 2 3 4 5
1 17.33 (2) 12.33 (6) 11.25 (8) 16.23 (4) 10.80 (7)
2 10.68 (8) 9.62 (8) 9.14 (10) 11.24 (8) 6.60 (10)
3 10.11 (9) 12.07 (7) 6.40 (10) 11.36 (9) 8.26 (8)
4 18.70 (1) 15.01 (7) 8.76 (9) 9.81 (8) 11.34 (8)
5 15.59 (5) 8.98 (8) 8.33 (9) 10.08 (8) 7.53 (10)
6 20.00 (0) 18.42 (3) 10.68 (?) 10.79 (7) 8.20 (7)
? 20.00 (0) 20.00 (0) 20.00 (0) 20.00 (0) 20.00 (0)
8 13.50 (7) 13.22 (7) 12.96 (7) 8.12 (10) 12.98 (5)
9 11.24 (7) 13.90 (7) 13.41 (6) 13.19 (6) 14.69 (6)
EXTINCTION TRIALS: DAILY AVERAGE LATENCY RECORDED 10 THE tFEAREST
HUIDREDTH OF A SECOND AND SUTLMED N~IBER GIVEE PARIOTHSTICALLY)
OF AVOIDANCE PRTPONSES IIADE BY THE DHSEiSITIZATION
G3OUP OF TE1 TRIALS OVER FIVE DAYS
Animals 1 2 3 4 5
1 13.17 (4) 17.69 (2) 9.01 (9) 12.38 (7) 18.90 (1)
2 19.57 (1) 16.88 (8) 12.30 (6) 18.55 (2) 18.33 (1)
3 19.53 (1) 8.59 (7) 13.60 (6) 11.02 (7) 10.86 (7)
4 19.53 (1) 17.05 (4) 15.71 (5) 14.06 (5) 8.74 (8)
5 20.00 (0) 20.00 (0) 12.36 (8) 19.30 (1) 13.10 (4)
6 20.00 (0) 19.40 (2) 14.20 (5) 16.34 (4) 12.00 (6)
7 7.69 (9) 5.89 (10) 7.16 (10) 11.59 (8) 8.72 (9)
8 15.c9 (3) 20.00 (0) 17.53 (2) 11.12 (7) 15.00 (4)
9 9.61 (7) 8.00 (9) 5.88 (10) 8.27 (8) 14.69 (9)
Arthur I. Wells, Jr., was born on October 9, 1939, at Birming-
han, Alabama. He moved to Florida in 1949, where he attended public
schools. He was graduated from Ft. Lauderdale High School in June,
1958. In August, 1962, he received the degree of Bachelor of Arts
from the University of Florida. During the summer of 1961, he worked
with maladjusted children under David Wineman while attending the
University of Michigan. In August, 1965, he received his Master of
Rehabilitation Counseling degree from the University of Florida.
After working for one year in Miani, Florida, as a Rehabilitation
Counselor, he returned to the University of florida to work toward
his doctorate in psychology. During his graduate studiess in psychology,
he vas a Vocational Rehabilitation Administration fellow. In August,
1969, he completed a one year's internship in clinical psychology at
the J. Hillis Hiller Health Center. Since September, 1969, he has
worked as a psychologist at the Northeast Florida State Hospital in
He was honorably discharged from service in the U. S. Air
Arthur M. Wells is married to the former Anastasia E. Lawlor.
He is a member of Psi Chi and Phi Kappa Phi.
Thi.s dissertation was prepared under the direction of the
cda.irman of the candidate's supervisory committee and hs been approved
by all members of that committee. It was submitted to the Dean of the
College of Arts and Sciencesanl to the Graduate Council, and was
approved as partial fulfillment of the requirements for the degree of
Doctor of Philosophy.
Dean, College of Arts and Scip -
Dean, Graduat6 School
-m y /'