Shaping Of Verbal Behavior In An Interview Setting

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content SHAPING OF VERBAL BEHAVIOR
IN AN INTERVIEW SETTING
by
James Patrick Carnevale
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Education)
January 1971
UNIVERSITY O F SO UTHERN CALIFORNIA
TH E GRADUATE SCHOOL
U N IV E R S ITY PARK
LOS ANGELES, C A L I PORN IA 8 0 0 0 7
This dissertation, w ritten by
J.AMES...EATRLCIL.CARNEVALE.........
under the direction of h i s . . . . Dissertation Com
mittee, and approved by a ll its members, has
been presented to and accepted by The G rad u
ate School, in p a rtial fu lfillm en t of require
ments of the degree of
D O C T O R O F P H IL O S O P H Y
Dm *
D a te J.anuaxy...l9.7.1
TABLE OF CONTENTS
CHAPTER PAGE
I. BACKGROUND OF THE PROBLEM...................... 1
Introduction
Statement of the Problem
II. REVIEW OF THE LITERATURE......................... 34
Can Verbal Behavior be Influenced?
Variables Influenced by Reinforcement
Conditioning, Extinction and Lasting
Effect
Effects of Interviewer Characteristics
Effects of Interpersonal Relationships
Effects of Client Characteristics
Effect of Reinforcement Scheduling
Generalization Effect
Attitude Change
Awareness Factor
Summary
A Word About Design
III. PROCEDURE AND METHODOLOGY....................... 96
Procedure
The Task
The Process
Hypotheses
IV. ANALYSES OF RESULTS................................ 104
V. SUMMARY AND DISCUSSION.............................. 117
Summary
Conclusions and Discussion
BIBLIOGRAPHY................................................ 126
APPENDIX.....................................................135
ii
LIST OF TABLES
Table
1
2
3
4
5
The Proportions of Sentences Using "She"
as the Subject of the Sentence for
Trials 1 and 2..................................
Standard Score Values for Individuals Com
paring the Proportion of Sentences Using
"She" as the Subject of the Sentence for
Trials 1 and 2..................................
The Proportions of Sentences Using "She" as
the Subject of the Sentence and a Neutral
Verb as the Predicate for Trials 1, 2, and 3.
Standard Score Values for Individuals Com
paring the Proportions of Sentences Using
"She" as the Subject and a Neutral Verb
as the Predicate in Trials 1, 2, and 3. . . .
Proportion of "She/Neutral/Him" Responses
for Base Rate, "She" Reinforcement Condi
tion, "She/Neutral" Reinforcement Condi
tion, and "She/Neutral/Him" Reinforcement
Condition ..................................... .
Standard Score Values for Individuals Comparing
the Proportions of Sentences Using "She" as
the Subject, A Neutral Verb as the Predicate,
and "Him" as the object of the Sentences for
Trials 1, 2, 3, and 4 .........................
Page
105
107
108
110
113
114
CHAPTER I
BACKGROUND OF THE PROBLEM
Introduction
In attempting to obtain a perspective of the fields
of counseling and therapy today, one is amazed to find the
myriad of "schools" of counseling and therapy which abound,
all of which have their strong supporters. Freudians, Ro-
gerians, Adlerians, Existentialists, Behaviorists, Gestalt-
ists, all have their theoretical constructs which they be
lieve provide the necessary guidelines which are most im
portant to the successful process of counseling or therapy.
The distinctions among these various theories are many; some
of them minor, some of them of major importance. This wri
ter believes that the most important division of thought is
one which divides not along "school-lines" but which divides
the field of counseling or therapy into two camps in rela
tionship to the issues of behavior based upon freedom and
choice or behavior which is completely determined.
The writer realizes the danger of being criticized for
false dichotomizing, but it is his belief that most major
theorists of counseling and therapy today constitute what
the statisticians would characterize as a bimodal distribu-
2
tion along the continuum from a free will to a deterministic
view of man. For purposes of this discussion, the writer
has abstracted what he believes to be two contrasting sub
groups of theorists who speak to the issue of free will and
determinism. The writer further believes that their con
trasting viewpoints present the student of counseling or
therapy, as well as the practitioner, with a great dilemma
as he attempts to understand the therapeutic process and
the part which he plays in it.
What kind of creature is man? Why does he do what he
does? Why does he change what he has been doing? What,
indeed, is man's nature? These questions are basic to the
counseling process and the answers given to them are
crucial to counselors.
One group answers that man is simply another organism
responding to its environment, and nothing more. A man is
what his experiences with his environment have made him. In
the truest sense, he deserves neither praise nor blame for
what he has become. If he is "good" it is merely fortunate
that his environmental experiences have been such that he
has become what is found to be pleasing. If he is "bad,"
then the environmental influences have been unfortunate,
and they must be changed in order to produce a better pro
duct. The criteria for that product must be, and can only
be, the individual's overt behavior.
3
The second group answers that man is a unique crea
ture who has the ability to choose what he is or what he
is doing. He can choose what he will be or what he will do
in the future. Man is a creature aware of being aware, ca
pable of self-reflection. Man has the great gift of in
sight and choice; he therefore is responsible for his exis
tence- -his being and his behavior. Any changes to be made,
he, the individual, must make. He cannot leave it to his
environment to produce the change; he must create it by his
choices.
Michael and Meyerson, in the 1962 Fall issue of the
Harvard Educational Review (52), present a representative
statement of the behavioristfs viewpoint and its relevance
to counseling. Focusing upon behavior as the primary con
cern of those in the counseling field, they insist that
behavior is a function of the interaction of heredity and
environmental variables, and nothing more. The idea of
inner determiners which are neither hereditary nor environ
mental is rejected. If behavior is the only appropriate
concern of the counselor, and if behavior is nothing more
than the interaction of hereditary and environmental var
iables, and since inherited genetic determiners are not
under the control of the behavioral scientist, the only
channel open to the counselor for influencing behavior is
4
through changes in the environment which influences the be
havior.
Coupled with the above postulate is the insistence
that all behavior is lawful and related systematically to
antecedent and attendant events. All behavior is deter
mined. Free will is dismissed by Immergluck as "merely a
necessary illusion" (36). Speaking for the absoluteness of
deterministic psychology, Immergluck points to human beha
vior as the last stronghold of those who insist on a capa
bility in the universe, and a dualistic world of mind and
matter. He insists that a determined, lawful world is the
only basis for a discipline to call itself a science, and
he denounces any attempt to contaminate the science of psy
chology with the metaphysics of free will (36).
Admittedly, trying to study the world of a man in be
havioristic terms is a difficult task. B. F. Skinner
speaks of resorting to such terms as "choosing," "purpose,"
or "understanding" as a capitulation to the temptation of
simplifying a phenomenon:
The world in which man lives may be regarded as
an extraordinarily complex set of positive and
negative reinforcing contingencies. In addition
to this physical environment to which he is sen
sitively attuned and with which he carries on an
important interchange, one has to contend with
social stimuli, social reinforcers and a network
of personal and institutional control and counter
controls, all of amazing intricacy. The contin
gencies of reinforcement which man has made for
5
man are wonderful to behold. But they are by
no means inscrutable. In any social situation,
we must discover who is reinforcing whom, with
what, and to what effect. (75)
The second part of the bimodal subgroup to which the
writer is referring can be labled humanistic psychology.
Men like Gordon Allport object to psychology's narrow use
of words such as reactor, response, reinforcement, respon
dent, and the like as representing the primary characteris
tics of being human. Allport prefers such characteristics
as pro-action, progress, production, problem solving, which
are so lacking in the literature of behaviorists as display
ing a disregard for the person's phenomenology of the fu
ture, for his sense of personhood and sense of freedom--
which is not an illusion:
The trouble with our current theories of learn
ing (behavior) is not so much that they are
wrong, but that they are partial. They fit
best the learning of animals and of children.
The concept of conditioning and reinforcement
seem hollow when the counselor tries to apply
them to his work . . . Supplementary theories
in terms of organizational, biographical, and
propriate learning are needed. (2:380)
Carl Rogers, in an article on man's nature, describes
man as:
. . . a basically trustworthy member of the
human species, whose deepest characteristics
tend toward development, differentiation, and
cooperative relationships, whose life tends
fundamentally to move from dependence to in
dependence; whose impulses tend naturally to
harmonize into a complex and changing pattern
6
of self-regulation; whose total character is
such as to tend to preserve and enhance him
self and his species, and perhaps to move it
towards its further evolution . . . I am in
clined to believe that fully to be a human
being is to enter into the complex process of
being one of the most widely sensitive, re
sponsive, creative, and adaptive creatures
on this planet. (66:201)
Implicit in the above statement is the concept of man
"becoming," creative man, a creature of independence and
choice. Here and in other writings, Rogers affirms his be
lief in man being more than a reactive organism. Rather,
man is viewed in the process of developing,,becoming, being
true to his nature rather than the reinforcing stimuli of
his environment (63, 64, 65).
Carlton Beck, in his book exploring the philosophical
foundations of guidance and counseling, does not deny the
validity of dynamisms and the study of behavior patterns;
rather, he believes that man is more than the above, and
can only be understood as an existing being whose fate it is
to choose the structure of his existence (6).
Within the two opposing points of view presented above
are several additional subgroups which have, through their
unique positions further divided, but perhaps have further
contributed to, the attempts to understand the cause of
man's behavior and its relevancy to counseling. It is the
writer's intent to present in greater detail the viewpoints
)
7
of these subgroups within the two main schools of thought
in order to appreciate more fully the basis upon which the
determinists and the humanists in general have developed
their opposing points of view and their relevancy to the
counseling or therapeutic process.
Learning theorists primarily concerned with stimulus-
response conditioning compose a large part of the determin
istic camp. B. F. Skinner, among the foremost of the con
ditioning theorists, presents a very thorough discussion of
man's behavior in his book, Science and Human Behavior
(74). In this work, much time and space are devoted to re
jecting many of the ideas or concepts which are commonly
accepted by men in general, and humanistic psychologists
in particular. "Ideas," "insight," or "thoughts" are terms
which are used to explain human behavior by "inner causes"
which provide spurious explanations of observable behavior
(74:11-42).
This rejection of inner causes of behavior is basic
to the Skinnerian point of view: the inner/outer man con
struct; the inner man wills an action, the outer man exe
cutes it; the inner man loses his appetite, the outer man
stops eating; the inner man wants, the outer man gets; the
inner man has an impulse, the outer man obeys (74:29). All
of the above infers the existence of a spontaneity of the
8
person, a freedom to move outside the "lawful behavior" im
posed by the contingencies of environment and heredity.
Such a concept for the cause of a man’s behavior, to Skinner
is merely an attempt to deny that man is only what he is:
an organism responding to his environment within the limits
imposed by his inherited characteristics.
For the behaviorist, the attempt to look inside a per
son to understand why he does what he does is to introduce
complexity and difficulty that is unnecessary and unhelpful.
If it is known that hard running, hot, dry conditions and no
water have been the lot of a horse; and if taken to water,
the horse now drinks, then it is completely unnecessary to
use the explanations that (1) he drinks because he is
thirsty, or (2) he feels like drinking, or (3) he wants to
drink (10:34). All that need be known are the contingencies
of running, dry heat, and water deprivation to lead one to
expect or predict that a horse in these circumstances, being
led to water, will drink. What the horse thought, or felt,
or desired really does not add much to our understanding of
the "why" of his drinking behavior. If a man instead of a
horse is substituted in the above situation, we still have
no need for further information about the man choosing to
drink, feeling thirsty, or wanting water. The aforemen
tioned contingencies and the drinking of water tell the
9
behaviorist what he needs to know about the cause of the
man's drinking behavior. And these contingencies are not
guesswork; they are not inferred, they are not unseeable;
they are visible, observable, measurable, manipulative var
iables that are the stuff of science (74:33). And it is the
science of human behavior that Skinner and fellow determin-
ists are about.
The external variables, of which behavior is a
function, provide for what may be called a cau
sal or functional analysis. We undertake to
predict and control the behavior of the indi
vidual organism. This is our dependent variable--
the effect for which we are to find the cause.
Our independent variables --the causes of be
havior--are the external conditions of which be
havior is a function. Relations between the
two--the cause and effect relationship in be
havior- -are the laws of science. A synthesis
of these laws expressed in quantitative terms
yields a comprehensive picture of the organism
as a behaving system.
This must be done within the bonds of a natural
science. We cannot assume that behavior has any
peculiar properties which require unique methods
or special kinds of knowledge. It is often argued
that the act is not so important as the intent
which lies behind it, or that it can be described
only in terms of what it means to the behaving in
dividual or to others whom it nray affect . . .Al
though such terms as "meaning" or "intent" appear
to refer to properties of behavior, they usually
conceal references to independent variables. This
is also true of "aggressive," "friendly," dis
organized," "intelligent," and other terms which
appear to describe properties of behavior but in
reality refer to its controlling relations. (74:37)
10
The independent variables must also be described in
physical terms . . . the events affecting an organism must
be capable of description in the language of physical sci
ence. It is sometimes argued that certain social forces or
the influences of culture or tradition are exceptions. But
we cannot appeal to entities of this sort without explaining
how they can affect both the scientist and the individual
under observation. The physical events which must then be
appealed to in such an explanation will supply us with al
ternative material suitable for a physical analysis . . . In
dealing with the directly observable data we need not refer
to either the inner state or the outer force (74:77).
With the concept of emotions, Skinner expands upon
the James-Lange explanation of the subject by stating that
emotions are a reaction to a behavior which is a reaction to
a situation and are of no real concern to the scientist try
ing to control and predict behavior. All that need interest
the scientist are the observable contingencies and the be
haviors accompanying them. If emotion has any value at all
it is in the adjective, description usage. Beyond that,
they are of no real concern (74:160-170).
"Choice" and "will" are similarly treated as unscien
tific attempts to explain the variables contingent upon the
behavior in a given situation. Because of all the histori
cal contingencies as well as present ones bearing upon a
11
situation demanding behavioral response, the observer sel
dom has anything like adequate knowledge of them all. He
therefore resorts to a simple explanation of an individual
"choosing" or "willing" to act in a certain way--hardly a
satisfactory manner of dealing with the problem scientifi
cally (74:113).
"Deciding" is described as looking at all the contin
gencies until one set of contingencies wins" (74:243).
"Problem-solving may be defined as any behavior which,
through the manipulation of variables, makes the appearance
of a solution more probable (74:247). "Original thinking"
is spoken of in the following manner:
Man is now in much better control of the
world than were his ancestors, and this
suggests a progress in discovery and in
vention in which there appears to be a
strong element of originality. But we
could express this fact just as well by
saying that the environment is now in
better control of man. Reinforcing con
tingencies shape the behavior of the in
dividual and novel contingencies generate
novel forms of behavior. Here, if any
where, originality is to be found. (74:255)
Even the concept of a "self" is dismissed as another
attempt to take the easy way out of understanding behavior.
The self is most commonly used as a hypothe
tical cause of action. So long as external
variables go unnoticed or are ignored, their
function is assigned to an originating agent
within the organism. If we can't show what
is responsible for a man's behavior, we say
that he himself is responsible for it . . .
The practice resolves our anxiety with re-
12
spect to unexplained phenomena and is per
petuated because it does so. (74:283)
Skinner then, is quite consistent in his view of MAN.
He is indeed only another organism among many which acts or
behaves in response to the contingencies of his environment.
The study of man, in the view of Skinner, is the study of
man's behavior and the contingencies which cause it.
The above comments illustrate the theoretical stance
of the S-R learning theorists relative to the cause of be
havior. The relevance of this stance to counseling in oper
ational terms must now be viewed along two lines of S-R
learning: operant and respondent learning or conditioning.
E. R. Guthrie's basic postulate about respondent
learning may be the best introduction to this subject. "A
combination of stimuli which has accompanied a movement,
will on its reoccurrence, tend to be followed by that move
ment" (33:23). This principle might be paraphrased by say
ing if one does something in a particular situation, the
next time he is in that situation one will likely do that
thing again. As long as stimulus and response occur to
gether, learning occurs and behavior is influenced. The
last thing that occurred in a situation is that which will
occur again in a similar situation. If one wishes to change
the behavior in a given circumstance, he must arrange the
desired ending for the circumstance. The arranged ending
13
will be the one most likely to happen, should the situation
arise in the future.
Guthrie further developed his basic postulate into
three specific methods of application toward the breaking
of habits and the establishing of new patterns of behavior:
(1) the threshold method, (2) the fatigue method and (3)
the method of incompatible stimuli. The method of thresh
old is illustrated by one method of training the horse for
riding by applying a light weight pressure on the back of
the horse, light enough so as not to elicit bucking. One
then gradually increases the load (one blanket, two blan
kets, a saddle, a saddle with increasing weight, and finally
a saddle and rider) so that the horse never feels the abrupt
increase in weight, as when one begins with saddle and ri
der (e.g. the bronco buster).
In the counseling process this is sometimes called
successive approximations and can be illustrated by the
following strategy. With a student who suddenly panics
when presented with examinations, a schedule of experiences
is developed for the student which gradually crosses differ
ent thresholds of tolerances (sitting in the exam room,
talking to peers; sitting in the exam room talking to the
teacher; then writing a few ideas for the teacher to dis
cuss ; then writing some answers which the teacher has
14
raised during the discussion; etc.) until the student is
able to take an actual written exam.
The method of fatigue is well illustrated by the
bronco busting process. The undesired response (bucking)
is elicited again and again until the horse is so tired
that he stops and walks. This last response, walking is
then the response likely to occur when the stimulus (rider)
is presented again.
An illustration of this method in the counseling pro
cess is often used by behaviorists in dealing with tics.
The client is asked to place himself in the environment in
which the tic most commonly occurs. He is then told to de
liberately go through the tic behavior again and again until
exhaustion of the muscles used for that reaction occurs. It
is likely that the tic will have exhausted itself, both then
and in the future.
With the method of incompatible stimuli for an unde
sired response, certain stimuli are presented along with
stimuli that can be counted upon to produce a different in
compatible response. The original stimuli then become at
tached to the new response. An example of this method of
influencing behavior is the case of Johnnie, who was afraid
of furry animals. While Johnnie was eating delicious candy,
a rabbit was placed in the room, but at a great enough dis
15
tance that Johnnie wasn't frightened. Then while Johnnie
was enjoying eating the candy (a response incompatible with
fear) the rabbit was moved slightly closer--but only
slightly. Thereafter, on successive occasions, the rabbit
was moved increasingly closer while Johnnie was responding
with the joy of eating the candy. At length the rabbit was
placed in Johnnie's lap while he was eating his candy, the
original fear response having been extinguished because of
its incompatibility with the pleasant response of eating.
Lazarus and Wolpe (92) have called this a process of
reciprocal inhibition, and have built a therapeutic prac
tice primarily based on this process. Wolpe, in conversa
tion with this writer even ventured the hypothesis that
most therapies are in fact based upon this process, whether
by the therapist's intent or otherwise. In the warm, loving
relationship of therapy (generating the client's positive
response) the client gradually moves closer and closer to
indirectly experiencing through verbalizing some traumatic
event in his life. The usual response to this trauma is in
compatible to the positive responses of the therapeutic re
lationship and is gradually extinguished.
The other line of S-R learning is that of operant con
ditioning with an emphasis on the "shaping of behavior." It
is believed that behavior is learned not only because a
16
stimulus and a response occur together, but also because of
the effect which follows the response. It has been deter
mined that if a stimulus is followed by a response and then
by a positive effect, the stimulus-response bond is strength
ened. If, however, the stimulus is followed by a response
and then a negative effect, the stimulus-response bond is
weakened (82).
The shaping of behavior is achieved by the selective
reinforcing of certain responses which approximate the re
sponse desired. Through the reinforcing of successive ap
proximations, the behavior is brought closer and closer to
the desired response, until finally the desired pattern of
responses is achieved.
For example, suppose one wished a rat to pull a
string, get a marble, pick it up with its paws, carry it
across a cage to a verticle tube rising two inches above
the floor, lift the marble and drop it down the tube. He
might very well wait a lifetime for this particular sequence
of events to occur before he was able to reinforce the rat
for this behavior, thereby increasing the probability that
the rat would repeat the process. Using shaping techniques,
however, a rat could be taught to do this in a relatively
short period of time. To accomplish this the rat would
first be magazine-trained (i.e., would learn that a sounded
17
click meant that food was available in the food tray). Then
a click would be sounded whenever the rat would touch a
string, then whenever he pulled the string. A marble then
would be released upon his pulling the string. Then the
click would be sounded only when the rat pulled the string
and touched the marble; then when he lifted the marble, un
til the entire sequence of desired behavior had been com
pleted, each step a refinement of the preceeding step, with
greater precision of behavior elicited with each succeeding
step.
Many behavior therapists such as Ulman and Krasner
(86) have specialized in the application of these principles
in the therapeutic process. In conversation with this
writer, Krasner suggested that much of all therapeutic suc
cesses may well be the result of operant conditioning, with
or without the deliberate intent of the therapist. Through
the subtle application of reinforcement in the form of ther
apist's interest, approval, or concern, the client may well
be shaped up to the therapist's value system or to behavior
patterns which the therapist desires.
Another deterministic subgroup is the Freudian school
of thought. One is destined to be a continuation of his
past experiences; the unconscious provides the primary mo
tivation of one's life style; the individual's behavior is
18
the result of the constant battle of id, ego, and superego,
over which he has no control. It is believed that the in
stinctual drives of the id will be manifest in some form,
socially acceptable or otherwise, and the individual cannot
really be held responsible for his behavior. The living dy
namics of Freudian theory are certainly of a different kind
than those of the conditioning theorists, but the essence of
determinism is the same. The focus of responsibility for
behavioral change of the client rests with the therapist to
provide new experiences which will more fully develop the
ego or superego (whichever one is currently overruled by the
other) in order that new responses to the urgings of the id
will result. Even the overcoming of repression is to allow
either the ego or superego more "grist for the mill." No
real choices are involved in becoming more than one is, at
least as the idea of choice is usually used by humanistic
psychologists (34:6).
Still another group which this writer places within
the deterministic camp is that of phenomenology. The major
spokesmen representing phenomenology in the United States
are generally acknowledged to be Snygg and Combs, who place
themselves almost in contradiction to "traditional condition
ing theories" which they view as extremely limiting. When
discussing the behavioristic schools of psychology they are
described as:
19
. . . having given rise to a mechanistic
conception of human beings as physical
objects whose behavior is the result of
forces acting upon them. It has largely
dehumanized psychology, making of human
beings little more than objects to be man
ipulated at will . . . man is thought of
as a passive automation, buffeted about by
the circumstances surrounding him. (76:310)
By the above quotation it is apparent that Snygg and
Combs, like the Freudians, do not see themselves as a part
of the deterministic, behavioristic school of psychology.
They view their stance as a compromise between the points of
view of man being totally responsible for his behavior by
his decisions and choices and the viewpoint which speaks of
man as completely dependent upon his environmental stimuli.
Man, in their view, is "part controlled and in part control
ling of his destiny" (78:310).
However, when looking directly for the phenomenol-
ogists concept of choice, one finds them more in the deter
ministic school than might be expected in light of the
above. Carlton Beck, in his examination of different defi
nitions of choice, declares that the phenomenologist’s con
cept of choice is a pseudoconcept (6:66). To support his
declaration, Beck quotes the following passage from Snygg
and Comb's Individual Behavior:
. . . we might say that the individual appears
to be engaged in a continuous process of making
choices. As a matter of fact, no choice what
ever exists. He attempts that which appears to
him to be self-enhancing and attempts to avoid
20
that which appears to him as threatening.
What he does is dependent upon the differ
entiation he can make in his phenomenal
field. Ordinarily this process occurs with
a minimum of disturbance to the organism.
It is seldom clearly differentiated by the
individual. Occasionally two or more differ
entiations may appear in the field and the
individual may, himself, describe the situa
tion as making a choice. If we could see the
situation as he saw it at the moment of his
behavior we would probably discover that he
made no choice at all but behaved as he ha?
to behave to maintain his phenomenal self. (6:66)
It would appear that the phenomenologist criticizes
the behaviorist for depicting man as determined only by his
environmental stimuli; but they in turn depict man as de
termined by his phenomenological field and the organismic
drive of self-enhancement. In either case, behavior seems
determined. Choice is really a semantic illusion.
For the counselor or therapist working with the phe
nomenological model, the task is to help the client see his
world differently. As figure and ground are modified, self
enhancing behavior is elicited in relationship to the new
environment and, hopefully, greater satisfaction results
for the client.
The differing deterministic schools of thought have
each accounted for at least some of the benefits of coun
seling or therapy. Each has its successes. Each has con
tributed to the helping profession with its own theoretical
constructs with at least one common base of agreement:
choice is not the acceptable explanation for man's behavior,
21
nor for the success of counseling or therapy.
But for the "choice" theorists there is also great
diversity of thought with at least one common base of agree
ment: choice is the basis of man's behavior and an inte
gral part of the process of counseling and therapy.
Albert Ellis and Robert Harper constantly reiterate
the need for rational choices. A Guide to Rational Living
is, in fact, an instruction manual for making the kinds of
choices which the author believes will lead to a more sat
isfying and fulfilling life (24). For Albert Ellis and his
followers, the cognitive processes are seen as the means of
an individual finding a sense of the "good life" through
therapy. The emotional, affective responses are to be con
trolled by reason. The emotions are not seen as having
their own autonomy. Rather, they are viewed as follow-up
responses of the individual's thinking process. If one is
feeling angry or sorrowful, it is because he is thinking
angry or sorrowful thoughts. To feel better, it is neces
sary to choose better kinds of thoughts.
Emotions directly stem from ideas, thoughts,
attitudes, or beliefs, and can usually be
radically changed by modifying the thinking
processes that create them . . . man can live
his most self-fulfilling, creative, and emo
tionally satisfying life by intelligently or
ganizing and disciplining his thinking. (24:13)
For most therapists, the good life is that life which
is experienced as emotionally satisfying. For Albert Ellis,
22
this will be possible only through having satisfying
thoughts, and this in turn is made possible by internalized
sentences which one verbalizes to himself. The individual
can choose the kinds of things he verbalizes to himself. He
therefore can choose the kinds of things he is thinking.
He therefore can control, to a large extent, how he is feel
ing. He therefore can choose the good life (24:28).
Albert Ellis does seem to hold a modified phenomenol
ogical view of reality--"Men are not disturbed by things,
but by the view which they take of them" (24:33). The mod
ification of phenomenology is in the stated belief that one
chooses what reality he creates, rather than being a passive
part of the development.
In the therapeutic process, one of Ellis' patients
discovered that her bad feelings did not derive from her
unsuccessful marriage, or her divorce, but from her evalua
tion of these and other events in her life. Through ther
apy she was able to change the internalized sentences she
usually associated with certain events. She was able there
by to change the kinds of thoughts she was having about her
self and her life functioning. Her emotions changed from
depression, discouragement, and unhappiness to elation, en
couragement and joy (24:36).
Ellis acknowledges the effects of environmental in
fluence (24:60, 102) but goes on to say that as one matures
23
and develops his rational processes, the dependence upon
environmental stimuli should be lessened. One should be
able to choose his behavior and life style (24:63, 109).
One who is still a conditioned being is he who has submitted
to conditioning as an easy escape from responsibility
(24:137). One can control his own destiny if only he will
assume his own saddle (24:131). One can uncondition him
self, if he but will it (24:155-156). The process of coun
seling or therapy is aiding the patient to assume respon
sibility for his life and his reasons for living it as he
chooses. For those therapists who follow the teachings of
Ellis, this is accomplished by changing the kinds of sub
vocalized sentences the client is using to influence the
thinking-emotive process.
Another prominent figure in the field of psychother
apy who rejects the concept of the "determined-man" is
Viktor Frankl. In his book, Psychotherapy and Existential
ism, he again and again speaks for the freedom of the in
dividual.
The freedom of man is freedom within limits,
Man is not free from conditions, be they bi
ological, psychological, or sociological in
nature. But he is and will remain free to
take a stand toward these conditions; he al
ways retains the freedom to choose his atti
tude towards them. (27:3)
Frankl does not turn to the environment to explain
man's being. The gas ovens of Auschwitz proved to Frankl
24
that man, can indeed, rise above the limits of his environ
ment .
. . . human freedom is not a freedom from
conditions, but rather the freedom to take
a stand toward conditions . . . Through this
freedom man is capable of taking a stand
toward himself. (26:25-28)
Like Ellis, Viktor Frankl believes that the process
of successful therapy will result in the individual taking
greater responsibilities for deciding what his life stance
will be, and ultimately what he, himself, will be (26:35).
Unlike Ellis, however, is Frankl's goal of logotherapy.
Ellis would hope that his patient might stop punishing him
self with "oughts" or "shoulds" imposed upon him by others
and to rationally choose a life style that gives him, the
patient, the greatest satisfaction.
Frankl, however, is not satisfied with either a self-
centered or self-actualizing goal for counseling. Frankl
believes that the ultimate motivation for a man is his
search for meaning in life. It is to this end that therapy
must direct itself.
. . . man should not, indeed cannot, struggle
for identity in a direct way; he finds iden
tity to the extent to which he commits himself
to something beyond himself, to a cause greater
than himself. (27:9)
If Ellis tries to free his clients from "oughts,"
Frankl searches for them. "If one takes man as he is, it
makes him worse; if we take him as he ought to be, we help
25
him become it" (27:12). And while Ellis attempts to facili
tate change in his patients by changing subvocalized senten
ces, Frankl attempts the facilitation through meaningful
dialogue. Obviously there are significant differences be
tween these two theorists and in their contributions to the
field of therapy; but all of their contributions are simi
larly based upon a firm conviction of man's free will.
Still another major theorist who is opposed to a de
terministic view of man is Carl Rogers. There is much of
Roger's theory that strongly resembles the phenomenology of
Snygg and Combs, but with a significantly new dimension--
that of choice--which has been noted by Carlton Beck (6).
Outlining some of Rogers working assumptions we find
the following:
1. Each individual exists in a continually
changing world of experiences of which
he is the center.
2. The individual reacts to this field as
it is experienced and perceived. This
perceptional field is, for this indi
vidual, reality.
3. The individual has one basic tendency
to actualize, maintain, and enhance the
individual in a socializing way based
upon his view of reality.
4. When this individual creates and accepts
into one consistent and integrated system,
all his sensory and visceral experiences,
he is most likely to find himself relating
in a less defensive, more actualizing way
with himself and the world around him.
(65:485-622)
26
Assumptions one and two fit very easily into a phen
omenological construct, and so, to a degree, does assump
tion three. Assumption four, however, does not, if we use
Snygg and Combs as our referent (78). The fourth assump
tion implicitly carries with it the element of choice,
which has been discarded by Snygg and Combs, as quoted
earlier (6:66). Rogers, in his later writing, is quite ex
plicit in his insistance upon choice as a fundamental pro
cess for the person. Whilehe speaks of a positive poten
tial awaiting "only the proper conditions to be released
and expressed" (64:35), he also speaks of "choosing to be
what one is, rather than choosing to be what others would
have you be" (64:103) as one of the outcomes of successful
therapy. When Rogers describes trying to establish opti
mum therapeutic conditions, he speaks of how the client
uses his new "freedom" (64:109). He quotes Kierkegaard
as saying:
The deepest despair is to choose to be other
than oneself. To will to be that self which
one truly is, is indeed the opposite of de
spair. (64:110)
When speaking of the good life, he describes it as a
direction "selected by the total organism with the psychol
ogical freedom to move in any direction" (64:187); and he
describes the existential life as "unpredictable" (64:188).
There can be no question about Rogers' rejection of a
27
determined man, nor about his belief in a man building his
own life by his choices.
Rogers' main concern with the counseling process has
been in developing conditions for personal growth via the
counseling relationship. Instead of asking, "How can I
treat or cure or change this person?" Rogers asks, "How can
I provide a relationship which this person may use for his
own personal growth" (64:32).
No approach which relies upon knowledge,
upon training, upon the acceptance of some
thing that is taught is of any use . . .
such methods are, in my experience, futile
and inconsequential . . . If I can provide
a certain type of relationship, the other
person will discover within himself the
capacity to use that relationship for
growth, and change and personal development
will occur. (64:32-33)
This emphasis upon the counseling relationship, then,
has been Carl Rogers' unique contribution to the humanistic,
nondeterministic approach to counseling and psychotherapy.
Finally, the existential therapists, Rollo May and
Adrian van Kaam also speak out loudly on the issue of de
terminism vs. free will. Both author-therapists speak em
phatically about existential theory as being more than a
collection of techniques. Rather, it is specifically con
cerned with the nature of man and the attitude of the ther
apist or counselor toward his client as a man (51:5)(88:18).
van Kaam states:
28
My counselee is freedom; he is the possi
bility of initiative; he is the source of
meaning and existence . . . with the limits
of one's factual orientation, the possibili
ties of self-actualization are inexhaustible
. . . my counselee is basically a free person.
(88:51)
Rollo May writes that the fundamental drive of man,
the essential nature of man is to live out one's potential
by his choices (51:31). May further states:
Obviously a knowledge of the drives and mech
anisms which are in operation in another per
son's behavior is useful; a familiarity with
his patterns of interpersonal relationships
is highly relevant; information about his so
cial conditioning is, of course, to the point.
But all of this is at different levels than
the immediate living person himself . . . and
he creates his world by his perceptions and
his choices . . . (51:38)
To the existentialist, man is the being who can be-
conscious of, and therefore responsible for, his existence
. . . the only creature who can experience anxiety because
he has the freedom of choice (51:41-52). Rollo May acknow
ledges the importance of conditioning, but views it as the
establishing of limits within which man chooses:
The world includes past events which condition
my existence and all the variety of determinis
tic influences which operate upon me. But it is
these as I relate to them, am aware of them,
carry them with me, molding, inevitably forming
and building in every minute of relating. For
to be aware of one's world means to be designing
it. (51:59-60)
Conscious of the problem of explaining the will, van
Kaam, almost in response to B. F. Skinner, speaks of it as
follows:
29
When we think about the will, we portray to
ourselves a kind of isolated force of energy
that is somewhere present in our personality
like an explosive in a gun chamber. Or we
imagine that the will is a kind of person in
us who, like an engineer, controls buttons
which make us think and move. In this concep
tion, the will becomes a kind of thing, an
absolute and autonomous power which is isolated
from the whole of my personality. I become a
person who, among other things, has also a
curious possession called will . . . This view
is misleading. I do not have a will, but I am
a will. . . Willing is thus an expression, a
mode of being myself as a whole interacting
with the totality of my life situation . . .
(88:71-72)
It is the nature of man to be aware of his
life situation and to find a personal posi
tion toward the daily reality which he en
counters . . . willing is being open to the
world . . . Therefore, instead of thinking
of the will as some kind of isolated power
house . . . it is necessary to see it as di
mension of the person as a whole. One should
realize that it is the willing openness for
truth, not a kind of absolute force, that in
splendid isolation pushes the rest of our per
sonality around like a boxer in the ring knocks
around a weak opponent. (88:72-75)
van Kaam also speaks to the reality of determinism;
but he too, views it only as providing broad boundary lines
within which the individual creates his life:
Existential psychology retains awareness of
the limits of freedom revealed by the psycholo
gies of factual determinants, yet it transcends
determinism by its recognition of man's radical
freedom which is equally observable in his be
havior as are his physiological and environmen
tal determinants. (88:165-166)
The goal of existential therapy, then, becomes that of
helping the individual realize his great potential of willing
30
his life to be, of creating his life through his choices.
The existential therapist, even more than the other thera
pists in the nondeterministic camp, is less concerned with
the techniques or the "hows” of therapy and is more con
cerned with the attitude toward man and goals of the thera
peutic process--man, creating his life by his free choices:
Man will become what he makes of himself, and
nothing more. Man constructs himself through
his choices, because he has the freedom to
make vital choices; above all the freedom to
choose between inauthentic and authentic mo
dality of existence. (51:118)
As can be seen by all of the above, there are indeed
two polarized positions regarding the reasons for man's be
havior, and each position in turn has subgroups which not
only support the particular point of view, but have devel
oped quite specific attitudes and implementations for
changing man's behavior through the process of counseling
and therapy. As in the science of physics with its two dis
tinct theories of light, both of these two primary theories
of behavior seem to be valid. Their validation is indicated
by the variety of theories of counseling/therapy discussed
above, all of which apparently can be used successfully
with certain clients or patients.
It is this writer's contention that both "truths" are
evidenced in the process of counseling or therapy. Client-
change can be effected through the utilization of both di
mensions of (1) insight and choice, and (2) behavioristic
31
learning theory. It is further believed that the counselor
or therapist, to be most effective for the greatest number
of clients, must know how best to effect change in his
client's behavior through both of these dimensions.
The writer believes that some of the previously men
tioned dilemmas of the students and practitioners of coun
seling and therapy can be alleviated if they can view the
counseling or therapeutic process as one in which the
clients freely choose to be affected by the counselor/ther
apist through both insight and deterministic models of beha
vior change. Clients, in fact, freely choose to be influ
enced by therapist or counselor, and the counselor/therapist
then has the responsibility to effect that influence in the
best ethical and professional manner.
There is yet much to be learned about both of these
major theories of man's behavior and the application of
these theories to the process of counseling and therapy.
The present study is a further investigation of the applica
tion of the learning theory principle of operant condition
ing. In this case the investigation will explore the
"shaping" of verbal behavior in an interview setting using
an operant paradigm. To this point in time, numerous
studies have indicated that the subject's verbal behavior
can be influenced in a gross fashion by the interviewer's re
inforcement of a broad category of verbal responses. No one,
32
however, has actually attempted to "shape" the subject's
verbal behavior through a series of stages, each stage in
creasing the complexity of influential control of the re
inforcing behavior of the interviewer. Should this be ac
complished, greater confidence might be placed in applica
tion of this learning theory principle of shaping behavior
to the influencing of verbal behavior in an interview set
ting. Having more firmly established that this principle
can indeed be so applied, there would be greater importance
to the investigation of whether this same principle may be
found applying in the counseling or therapeutic interview.
Such, however, is not the scope of this investigation.
Statement of the Problem
It is not the intent of this paper to "prove" that
the counseling or therapy process is simply another operant
learning process. Rather, it is the intent to investigate
further the extent to which verbal behavior can be influ
enced by operant conditioning and then to explore the im
plications or applicability of these findings to the field
of counseling.
As will be seen in the following chapter, attempts
have been made to influence the verbal behavior of the sub
jects in a relatively gross way. The problem of this study
33
will be the effectiveness of operant conditioning on influ
encing the verbal behavior of subjects in a complex and spe
cific pattern of speech within a time limit of two to three
hours of interviewing or training.
CHAPTER II
REVIEW OF THE LITERATURE
As one views the literature pertaining to verbal op
erant conditioning, he sees that a wide variety of variables
within this context has been investigated, and as one might
expect, a variety of methods has been used in the investi
gations.
Can Verbal Behavior Be Influenced?
Greenspoon (32), whose experiment is usually credited
with opening the area of verbal conditioning for study, de
monstrated that verbal behavior can indeed be influenced in
a gross way by two innocuous sounds uttered by the experi
menter, one of them, "mmm-hmm," a slightly positive utter
ance, and the other, "huh-uh," a slightly negative one.
Using college students as volunteers, he assigned, at ran
dom, 75 of them into five groups of 15 subjects. Group I
was asked to "Say all the words you can think of. Say them
individually. Do not use sentences or phrases. Do not
count. Please continue until I say stop. Begin." For 50
minutes these students said words. For the first 25 min
utes, each time they used a plural noun, the experimenter
softly said "mmm-hmm." For any other words used, nothing
34
35
was said by the experimenter. Records were kept of the num
ber of "plural nouns" and "other words" used throughout the
interview. During the final 25 minutes, nothing was said by
the experimenter to either class of student response.
Group II was given the same directions for the same
task for the same period of time. With this group, however,
the slightly negative sound, "huh-uh," was uttered by the
experimenter following each use of a plural noun, during the
first 25 minutes of the interview. Nothing was said during
the final 25 minutes.
With Group III, the experimenter uttered "mmm-hmm"
after each nonplural noun used by the student under the con
ditions listed above. With group IV the sound "huh-uh" was
also used with the nonplural nouns under the same condi
tions. Group V, the control group, spent the entire 50 min
utes without any comment from the experimenter, with nothing
being deliberately reinforced as with the other four groups.
Having eliminated ten subjects who were able to iden
tify and explain the experimenter's behavior, Greenspoon
then analyzed the data to see whether there was any signi
ficant difference among the groups as to the ordinal posi
tion of the first used plural nouns. Comparison of the
mean ordinal position and F scores indicated an equal ten
dency to use plural nouns among all five groups, so other
comparisons were now valid.
36
The data were then treated in two different ways for
further analysis. The interview of 50 minutes was divided
into ten five-minute sections which were then analyzed for
variance between five groups for the ten time-blocks. A
t test was also applied to determine the locus of the dif
ference in means of plural nouns used by the five groups.
The differences in the Groups III, IV, and Control
Group V were not found to be significant, indicating that
trying to influence as broad a category as nonplural nouns
seems to be out of the question, whether it be in a posi
tive or negative direction. Not so, however, with a more
specific category such as plural nouns. Group I had signi
ficantly more plural nouns for periods 2 - 6 than did Con
trol Group V, indicating that the experimenter did in
crease the plural noun usage while reinforcing with the
sound "mmm-hmm." This influence then extinguished when re
inforcement stopped. Group II had fewer plural nouns for
all trials of 1 - 5 than did the Control Group, although
only Trial 4 reached significance. Again there was evidence
of negative influence, but not so strongly indicated as with
positive influence.
Greenspoon had, however, made his point. Verbal be
havior of a subject was able to be deliberately influenced
by an interviewer without the subject's awareness in a spe
cific and predictable direction, by means of a relatively
37
common and simple sound used as a reinforcing contingency.
Some questions might be raised about how generalizable were
the findings. Because he used volunteers, he may well have
been using a group of students who were eager to please
anyone in authority and were therefore particularly easy to
influence. But with the population he did use, it would
seem the data do support his conclusions. The students in
all groups were tested to find whether they were of a
sample equally disposed to use plural nouns; and then were
randomly assigned to the different groups of the experiment
for different treatment . . . good research procedure. The
categories of student responses were quite discrete, and
the different treatment contingencies were concise, simple
and easy to record for analysis. The use of both t test
and analysis of variance to analyze the data were quite ap
propriate statistics. Only one of the four treatments was
found to support its hypothesis to a significant level, but
two of the four failures to support are easily understood
and do not detract from the conclusion. The indication that
positive influence seems easier to obtain than negative,
raised a definite question for further study, but a break
through had been made: verbal behavior had been modified in
a predictable manner according to established principles of
operant conditioning, using only sounds from the experimen
ter as the reinforcing variable.
38
Obviously, it is a long continuum of the interview
setting from the conditions of this study and those of the
therapeutic interview. Having a subject say words in a
disorganized fashion is hardly comparable to the emotion-
packed discourse of a patient or client during a counseling
or therapy hour. And the reinforcement of plural nouns is
hardly analagous to the attending behavior of a therapist
to the words of his client. But, during the counseling in
terview the client does use words to communicate to the
therapist and the therapist does respond in subtle and dif
ferent ways to the different things the client is saying.
The question could now be asked from a different basis than
before: Is there a relationship between the kinds of re
sponses a therapist makes to different types of statements
of his client and the content of the interview of which the
therapist may be unaware? And secondly, if such a rela
tionship does exist, is it significant to the therapeutic
process?
Using a similar experimental design, Cushing (20)
found support for the hypothesis that "positive" statements
of a subject could be increased by the experimenter through
the use of the sound "mmm-hmm" as a reinforcing contingency
without the subject's awareness. However, he was unable to
increase the "negative" statements of a second experimental
group. This, Cushing attributed to the differential affec
39
tive components associated with the subject's classes of
responses or statements. Again, there was evidence of suc
cess of an interviewer's attempt to increase a class of re
sponses of an interviewee in a subtle way, without the sub
ject's awareness. It was only a limited success, and it
was evident that modification of verbal behavior was not a
thoroughly predictable phenomenon. Much remained yet to be
investigated before it even approached any real predicta
bility.
Variables Influenced by Reinforcement
Matarazzo (50) added some information to the field
when he could point to success in having significantly in
creased the number of words related to people spoken by a
group of subjects. However, he also found that he had not
been able to increase the number of plural nouns which were
non-people related. It appeared that these early studies
were pointing to what verbal behavior could and could not
be seemingly influenced by reinforcement of certain classes
of verbal responses.
Buss and Durkee (13) reported that the use of both
hostile and neutral verbs used in sentences could be influ
enced by the interviewer without the client's awareness.
This was found to be true even though there was a seemingly
initial inhibition to the use of hostile sentences, which
finally appeared to come under the influence of the inter-
40
viewer's reinforcement behavior. Women were found to have
a different level of hostile usage both before and after
conditioning, but they did increase significantly.
The design of the study differed substantially from
that used by Greenspoon. In this experiment the subjects
were asked to form sentences from the words on cards pre
sented to them. Nouns and three verbs were presented, with
the verbs' usage being the criterion for reinforcement. The
verbs were classified by the categories of hostile or neu
tral by means of interjudge agreement during the selection
of words to be used. Only those verbs which were agreed
upon as falling into one or the other of these classifica
tions were reinforced according to plan.
An additional difference lay in the use of the groups
as their own control, a base rate of using having been es
tablished for comparison with the verb usage at the end of
the experiment. Comparison was made of the mean usage of
these two verb categories by both groups during the "base-
period" during which no reinforcement by the interviewer
was deliberately given. No significant differences were
found, indicating the two groups were of the same population
regarding the usage of the verb categories. Because the
subjects had been randomly assigned to the two groups, it
would seem equality of the groups relating to the criterion
in question was obtained. Comparison was also made for
41
verb usage between groups at the conclusion of the inter
viewing and found to be significant at more than .01 level,
indicating that a large change in the usage of hostile and
neutral verbs had occurred during the course of the experi
ment. However, to analyze the data in the more specific
manner, to find which group was affected in what way, each
group was compared using the analysis of variance to its
own base rate in the testing of the hypothesis. This, of
course, is more in keeping with the research designs of
Skinner, the father of operant conditioning, where the in
dividual's own learning curve is used as the basis for in
terpretation of results. In this case, group means were
used as the learning curve for analysis.
Modification of this design and of that used by
Greenspoon provide the models for most of the research done
in behavior modification.
J. M. Rogers (67), using groups as their own control,
found that clients' self-reference could be increased in a
quasi-therapeutic setting by means of the interviewer's
careful use of verbal reinforcement. In this writer's
opinion, Rogers is certainly stretching the term "quasi-
therapeutic setting" in this study. Having subjects talk
ing in a highly structured manner is hardly a therapeutic
setting, and the qualifier "quasi" really doesn't help. But
significantly influencing an increase in the use of self-
42
referents may still be of some importance, even if the ve
hicle for eliciting these self-statements is having students
create sentences in response to verbal cues presented by
cards.
Of equal importance, or perhaps greater importance,
was the indication that external criteria could also be in
fluenced by the reinforcement of a certain class of sub
jects' responses during an experimental interview. In this
case the subjects were found to have significantly improved
their self-concept as measured by pre and posttesting with
the Haig-Butler Q-Sort, and to have lessened their level of
anxiety as measured by the Taylor Scale of Manifest Anxiety.
In the writer's opinion, it is unfortunate that a control
group was not used in this study. Though using the subjects
as their own control seems appropriate when testing hypo
theses regarding an increase in a class of subjects' re
sponses compared to their own base rate, it does not seem
appropriate if one is trying to measure change of criteria
outside the data actually being manipulated, i.e. Q-sorts or
anxiety scales. For this kind of investigation, a control
group would seem necessary if one were to try to assume that
the manipulation of reinforcement was correlated to the
change in the scores of these measures as hypothesized can
not be completely dismissed, either, and as we shall see
later, other attempts at relating the generalization effects
43
of verbal conditioning have been carried out in a more so
phisticated and statistically acceptable manner.
John Turner (85) also utilized the Taylor Manifest
Anxiety Scale, but to identify three groups of college fe
male students for high, medium, and low anxiety groupings.
His objective was to determine whether or not self-refer
ences could be increased by means of reinforcing this class
of response with the sound "mmm-hmm" following each self
reference made during a free story-telling session. Fur
ther, he hoped to ascertain whether subjects with different
levels of measured anxiety would condition at a different
rate of learning; providing, of course, they conditioned
at all.
He again used the group as its own control, but added
the dimension of extinction to ascertain whether, indeed,
conditioning had taken place. The variable of extinction
has long been used in operant conditioning to prove that
control or influence has been gained on the operant beha
vior in question. If the behavior is seen to increase and
then diminish as the experimenter wishes, when he wishes,
the chance factor is quite well ruled out of the picture.
Turner did find both conditioning of self-referent
statements in the story telling interview when "mmm-hmm"
was used as the reinforcing contingency. Extinction of
this pattern was also found when the reinforcement ceased
44
to be applied, significant at the .05 level. No differences
were found between anxiety groups, using the analysis of
various measures, for either conditioning or extinction of
self-referents.
Though Turner did not find differences in measured
anxiety to be a consideration in conditionability of sub
jects, he was able to support the hypothesis that in a rel
atively unstructured situation the subjects' verbal behavior
could be influenced in a predictable manner according to
principles of operant conditioning. While it is still true
that this was not a therapy interview, it is a situation
more unstructured than most of the studies mentioned above,
and gives us more reason for considering generalizing to
other more normal verbal relationships. The fact that ex
tinction took place so quickly should also be noted, when
one is considering the long-term consequences of this field
of conditioning, be it in a therapeutic setting or other
wise.
Conditioning, Extinction and Lasting Effect
Kinzie (43) however, did find somewhat more lasting
effect in his study. He, too, was able to demonstrate the
ability to increase self-referent statements in accordance
with operant principles; and this in a setting quite
closely resembling that of an interview. In his study he
asked the subjects to talk about things that happened to
45
them recently; and he repeated the process three times.
During the first interview, no deliberate reinforcement was
administered, and a base rate for self-referents was estab
lished. During the second interview, all self-referent
statements were followed by the sound "mmm-hmm" softly
spoken by the interviewers. During the third interview, no
reinforcement was deliberately administered.
Again, the group was its own control, and again, a
significant increase was found in the number of self-state
ments during the interview in which reinforcement took
place when compared to the base rate interview, significant
at the .01 level. Even more, the number of self-statements
during the third interview, while fewer than during the
second interview, were still greater than during the first
interview, significant to the .05 level. This would indi
cate, said Kinzie, that the conditioning which did take
place had some lasting effect.
This study, too, was done in very unstructured set
ting, in a manner definitely related to a counseling inter
view. But, the question raised in this writer's mind re
lates to the definition of self-referent statements. In
neither this study nor in Turner's was it clearly stated
what could be interpreted as a self-referent statement. It
was evidently assumed that the title was self-descriptive.
This writer disagrees! Far better would it have been to
46
delimit what could be recorded as self-referent, and then
have used several judges for the recording, reporting only
those statements with a high interjudge reliability. With
this addition to the design, both of these last two studies
would have lent greater strength to the verbal conditioning
studies.
Barbara Sarason (71) also found indications of learn
ing through the applications of operant principles to ver
bal behavior, and she, too, found a lasting effect. In
this study, 120 general medical patients in a Veterans Ad
ministration Hospital were used. Three groups were com
posed of 40 subjects, each randomly assigned to either of
the two experimental groups or to the control group. The
subjects were to spend one session with a member of the hos
pital staff who was ostensibly trying to find out how they
felt about the hospital, for purposes of administrative
evaluation. During this first interview, all groups were
given a base rate period in which the interviewer responded
with interest to everything which was said, by sounding
"mmm-hmm" approximately every ten seconds. This lasted for
ten minutes. Then Group I was reinforced by the sound
"mmm-hmm" every time something positive was said about the
hospital. Group II was reinforced each time something
negative was said about the hospital, and Group III was still
given the ten-second reinforcement for the remainder of the
47
interview. The interview was a total of 40 minutes in
length.
The analysis of this data seemed to be quite thorough.
Comparisons were made of each group with its own establish
ed base rate to see if there were significant increases in
the ratio of positive, neutral, or negative responses for
the remainder of the interview. Then, analysis of variance
was computed to measure for differences in these three cat
egories among groups. When compared with their own base
rate ratio of appropriate responses, Group I was found to
have significantly increased the ratio of positive state
ments, with no significant increase in either of the other
two categories; Group II was found to have significantly
increased the ratio of negative statements, with no signi
ficant increase in either of the other two categories; and
Group III was found to have no significant increase in the
ratio of any of the three categories.
When analyzed for variance, a significant F was found,
with Groups I and II the furthest points of difference,
giving further support to the hypothesis that the verbal be
havior would be influenced by reinforcement in the predic
ted direction.
These findings, however, were only a part of the in
vestigations. Four days after the conclusion of the above
process, the subjects were again interviewed. This time
48
they believed that the interviewer was a graduate student
in sociology who was interested in the hospital as an in
stitution of society, and was interested in their percep
tions of it. In this case, the interviewer did not delib
erately reinforce any category of response for any of the
subjects. Rather, he merely asked them to help him as best
they could by telling him about the hospital as they knew
it.
Their responses to this interview were then divided
into the original three categories and then the ratio of
each of the categories was compared to the base rate estab
lished in the earlier interview.
Still there was a significant increase in the percen
tage of positive comments from Group I, in the percentage
of negative comments in Group II, and no change in the re
sponses of Group III, indicating a lasting effect of four
days for the behavior modification established by the first
interviewer. Also, because the second interview was appar
ently viewed as a different situation than the first, be
cause the interviewers and the reasons for the interviews
were different, a generalizability of the learned behavior
was indicated. An attitude seems to have been established.
Again, the writer wishes more space would have been
devoted to explaining how standardized or reliable was the
method of assigning the subjects' responses to one of the
49
three categories, and whether the author or others did the
categorization. Further question is raised about the dif
ferences among the interviewers themselves and their abil
ity to reinforce with equal skill according to strategy.
Approximately 160 hours of interviewing are indicated in
the study, 80 hours for each set of sessions, so we must
assume that several interviewers were employed for the in
vestigation. Still, this study may have some importance.
It does (1) indicate that conditioning did take place and
was under the influence of the interviewer in a predictable
fashion; (2) that it was able to be done in a conversational
setting; (3) that its effect was still significant over a
four day period of time; and (4) that the results were gen
eralized to a different setting, indicating an attitude had
been developed.
Effects of Interviewer Characteristics
In addition to ascertaining that verbal conditioning
can indeed be accomplished, a great many studies have been
done investigating a variety of variables within the frame
work of operant learning. Binder, McDonald, and Sjoholm
(8) investigated the effects of interviewer characteristics
upon the conditioning interview. Using the own-control par
adigm, two groups were reinforced for creating sentences
using the pronoun "I" from words presented to them on 5 x 8
cards flashed by the interviewer. Since the real focus of
50
this experiment was to test for the effects of interviewer
characteristics upon the conditioning process, two very dif
ferent interviewers were used with whom the differences
were glaring. One was a diminutive, gentle, soft-spoken
and attractive female. The other was a large, loud, burly
marine sergeant-type man. The female interviewer was found
to have conditioned more clients and to a greater degree of
influence than did her male counterpart, indicating that
more than the manipulated variable must be considered when
thinking about verbal modification.
Effects of Interpersonal Relationships
Sapolsky (70) investigated the effects of interper
sonal relationships which the interviewer establishes with
the subjects upon the rate of operant conditioning. As one
group of subjects was waiting to be interviewed, a plant in
the group described the interviewer they were to see in
terms of the kind of relationship which would hopefully be
established (thereby preparing them for the interview with
a tentative mental set). The contrasting descriptions were
of the easy going, concerned individual or the authoritar
ian, unbending person. Then the interviewer (the same in
dividual for all the interviews) attempted to establish a
relationship with each individual of the groups according
to the strategy which had been developed. As was hypothe
sized, the positive relationships produced the higher rate
51
of conditioning of plural nouns as subjects of created sen
tences than did the negative relationship. The "positive"
interviews conditioned to the .05 level of significance;
the "negative" interviews did not condition to the level of
significance, although there was an increase in usage of
plural nouns.
Marder (49) replicated much of the above studies, but
added the dimension of positive and negative reinforcement
(positive: "mmm-hmm;" negative: "huh-uh"). He was there
fore not only attempting to establish the fact of condi
tioned learning but correlations between conditioning rate,
type of relationship, and type of reinforcement. Analysis
of variance indicated that a positive relationship plus
positive reinforcement resulted in the highest rate of con
ditioning. This was followed, in descending order of con
ditioning, by positive relationship, negative reinforce
ment; then negative relationship and positive reinforcement;
and finally negative relationship and negative reinforce
ment. Many questions were left unanswered in this particu
lar report regarding the reliability of the different types
of relationship and the equality of the different inter
viewers used in administering both the type of relationship
and the reinforcing contingency for the particular groups.
The writer included it for this chapter as an example of
relatively poor reporting of an interesting investigation,
52
but one which may well include important implications for
a student of this topic.
All of the above studies pointed to the importance of
the interviewer's characteristics if he is to be a signifi
cant reinforcer for the modification of verbal behavior,
and indicated the importance of other variables than the
reinforcing contingency itself in this type of endeavor.
Effects of Client Characteristics
Differences in the characteristics of subjects were
also investigated as important variables to be understood
for those concerned with the ability to control and predict
the verbal behavior of others.
I. G. Sarason (72) investigated what, if any, corre
lation existed between the personality characteristics of
Test Anxiety (T.A.), General Anxiety (G.A.), Lack of Pro
tection (L.P.), Defensiveness (D) and Hostility (H) as mea
sured by the Autobiographical Survey and the tendency to
respond to verbal conditioning. The subjects were 60 neu
rotic and psychotic patients in the intensive care ward of
the West Haven, Connecticut Veterans Administration Hospi
tal. All but two were men; all were of ages between
twenty-two and fifty-two.
The subjects were first administered the Autobiograph
ical Survey, an instrument developed by Sarason, and
scored on scales of test anxiety, general anxiety, lack of
53
protection, defensiveness, and hostility. The subjects
were then individually presented with 160 3 x 5 cards upon
which were the pronoun "he" and three verbs of equal fre
quency of occurrence, according to the Thorndike-Lorge List
of Verbal Usage. One of the verbs was classified as a
verbal-activity type (e.g., talked, whispered) and was the
reinforced class verb. Subjects were told to make a two-
word sentence when shown each card using the pronoun "he"
and one of the three words printed on the bottom of the
card.
For this study, Sarason used the subjects as their
own control and established the "before-conditioning" level
of verb usage by having the subjects respond to Trials 1 -
20 with no deliberate reinforcement contingencies emitted
by the experimenter. For Trials 21 - 160, the experimenter
uttered "mmm-hmm" whenever the subject formed a sentence
using the Verbal Activity Class verb. Comparison was then
made between rate of conditioning, if any, and the scales
mentioned previously.
In addition to the above, the therapists of the hos
pital, working individually with the patients were asked to
rate the patients in terms of compliance, hostility, and
dependence. Five hypotheses were then formulated:
1. Subjects with high scores on Test Anxiety,
General Anxiety, and Lack of Protection
54
Scales would condition faster than Sub
jects with low scores on these scales.
2. Subjects with high scores on the Defensive
and Hostility Scales would condition more
slowly than low scoring Subjects.
3. The therapist's rating of the degree of
compliance with which the Subjects accep
ted interpretations of the therapist would
be positively correlated with the Anxiety
Scale and with the speed of conditioning.
4. The therapist's rating of degree of hostility
would correlate positively with the Hostility
Scale of the Survey and would condition slower
than Subjects with low scores.
5. The therapist's ratings of dependence would
positively correlate with the Dependence
Scale of the Survey and would condition
faster than Subjects with low scores.
Following the conditioning series and before statis
tical analysts of the data were begun, a brief interview
was conducted to determine the level of the subjects' abil
ity to verbalize the basis on which the experimenter had
tried to reinforce the verbal behavior of the subjects.
None of the subjects was able to describe what process had
been in effect; though seven of them had noticed that the
interviewer had a habit of muttering "mmm-hmm." No apparent
relationship was seen between this habit and the interview,
however.
Because the number of verbal activity-type verbs was
to be correlated with the various scales of the Autobio
graphical Survey, the reader might well question the relia
bility of the survey instrument for purposes of correlation.
55
It should be noted that the Survey had been administered
twice to each subject, with a two week interval between the
administration. All scales correlated for all of this
sample with "r's" above .70 indicating a degree of relia
bility high enough for this kind of exploratory study.
For purposes of analysis, all scale scores were di
vided into lower, middle, and upper thirds. Analysis of
covariance was then performed for each scale, using trials
1-20 for the control variable and trials 21 - 60 as the
criteria variable. The following F scores were obtained:
L.P F = 5.13 (j)<.01)
T.A F = 3.49 (j><.05)
G.A F = 1.68 (j><. 20)
D F = 4.65 (£<.025)
H ............................... no significance
When comparing the therapist's ratings with the Sur
vey Scales using Biserial correlations, only the Lack of
Protection scale was significant (r * .38) although the
G.A. and T.A. scales correlations were positive and ap
proached significance.
When comparisons were made of the therapist's ratings
and rate of conditioning, low compliance subjects condi
tioned at a significantly lower level than did middle and
higher levels. The ratings of dependence and hostility
were not found to be correlated to either the Survey scales
56
or to the Subjects' conditioning performance.
Though not all of the hypotheses of this investiga
tion were supported, the study cannot be easily dismissed
as indicating nothing. The writer himself questions some as
pects of the study. Are psychotic patients fit for testing
with a survey, paper and pencil type instrument as used by
I. G. Sarason? Nothing was said about the order of verb
placement on the 3 x 5 cards. Were they placed in random
order, rotational order, or was the reinforced verb always
in the same place in the word sequence? Were the different
scales of the Survey really different? If they themselves
correlated highly, what kinds of comparisons were really
being made? This writer wonders whether there was any dif
ference in the data for those seven subjects who did notice
that the experimenter had a habit of muttering. And per
haps one might tend to question how relevant this study is
when the subjects were asked to form two-word sentences.
In spite of these questions, however, the study is not with
out merit. There seems to be firm evidence that different
characteristics of subjects as well as those of the inter
viewers are variables that must be taken into consideration
as one investigates the field of verbal modification
through operant conditioning.
Others have also tried to investigate those subjects'
variabilities which might affect the verbal conditioning
57
process. D. Anderson (3), following a format similar to
Sarason, used the TAT as the criteria measure of aggres
sive impulse and fear of punishment. It was found that the
subjects in the class designated fear of punishment condi
tioned more readily in the use of aggressive verbs than did
those designated as aggressive, though once again, the cri
teria measure would seem to leave much to be desired in the
realm of reliability, the TAT is used clinically by repu
table men, and does give a measure of personality charac
teristics that many clincians recognize and find accept
able. This study, then, does seem to augment the body of
evidence about subject characteristics and conditionability.
In a second study by Anderson, it was hypothesized
that subjects high in the need for approval would condition
more readily than subjects low in approval need, as mea
sured by the Edwards Personal Preference Schedule. Using
subjects as their own control, it was found that no differ
ence existed in the rate of conditioning for these two
groups characterized by the above dimensions. Both groups
were found to have been equally influenced to increase the
number of sentences using I or We as the subject of said
sentences; significant to the .05 level.
This is somewhat in contrast to a study by Babla-
delis (4) in which it was found than an inverse correlation
existed in the data between measured autonomy and the rate
58
of conditioning, significant to the .01 level. The more
autonomous the client, the less he is likely to be shaped
by a reinforcing operant. This study was a more sophisti
cated one than Anderson's, and perhaps is worth a more
careful reporting.
The Edwards Personal Preference Schedule was adminis
tered to 102 students in undergraduate psychology at the
University of Colorado, and the autonomy scale was espe
cially considered. The autonomy scale was arbitrarily di
vided into four segments by cutting the obtained distribu
tions at the mean and at one standard deviation on either
side. The tests of all 102 subjects were sorted on the
basis of scores into one of these segments. From a table
of random numbers, 15 subjects from each range were selec
ted, giving a total of 60 subjects; thus the entire range
of the distribution was represented.
The fifteen subjects in each section were assigned to
one of the following conditions by random procedure: re
inforcement of positive self statements, reinforcement of
negative self statements, and control condition with ran
dom reinforcement.
All interviews were conducted by the same inter
viewer and were recorded on tape. Subjects were instructed
to talk about themselves, covering both positive and nega
tive characteristics. During the first ten minutes, the
59
interviewer said nothing and made no gestures, thus estab
lishing a base rate of emission of self statements. During
the second ten minutes, the interviewer reinforced the pre
selected response unit; the reinforcement cue was the sound
"mmm-hmm," accompanied by a head nod. The control group
received reinforcement every ten seconds regardless of what
the subject was saying at the time. The interviews were
terminated at the end of twenty minutes.
Independent judges rated the subjects' statements as
per categories mentioned above and numbered them. The in
terjudge reliability was high and acceptable (97.11). When
analyses were made of the data, it was found that positive
self statements had conditioned significantly (£<.05); neg
ative self statements had not. When further analysis was
made, it was determined that with those subjects who did
show significant conditioning, the above mentioned inverse
correlation of .01 was found, giving further support to the
importance of individual characteristics to the attempts at
verbal conditioning. This writer feels that finding such
significance using only 20-minute sessions was quite remark
able, and the use of random reinforcement for the control
group rather than no reinforcement seemed to strengthen the
findings. The idea that negative self statements were not
conditioned seems to support the earlier findings of
Greenspoon (32).
60
Using as the external criteria for classification of
subjects, a measure of internally and externally controlled
people which he himself designed, Geltner (29) in many ways
replicated the study of Babladelis and came up with the
same findings. Internally controlled people did not condi
tion to a level of significance; externally controlled
people conditioned significantly better.
In a related study, Taffel (80) investigated the
anxiety level of clients and its correlation to condition-
ability. His subjects were 90 hospitalized psychotic and
neurotic patients able to read, write and complete the Tay
lor Manifest Anxiety Scale. All subjects were administered
the Taylor scales in the first of two sessions used in this
experiment. The scores for the completed anxiety scales
were randomly assigned to three groups. Group I would be
given the word "good" as its reinforcing contingency; Group
II would be reinforced by a flash of light following the
appropriate response (with no explanation given); and Group
III would be the control group, with no reinforcing contin
gency given. For purposes of analysis, those subjects
achieving scores of 0 - 24 were termed low anxiety; those
with scores of 25 - 34 were termed medium; and those with
scores of 35 - 50 were termed high anxiety.
During a second session, the subjects were presented
80 3 x 5 cards on which were the pronouns "I," "We," "You,"
61
"He," "She," "They," and a simple verb in the past tense.
The order of pronouns was randomly assigned to the cards
in order to avoid a position preference or set in usage.
The subjects were asked to create sentences using one of
the pronouns as subject and the verb as predicate. Trials
1-20 were base rate trials, and no reinforcement was
given. For trials 21 - 80, reinforcement was administered
as mentioned above, for sentences using "I" or "We" as
sentence subjects.
For statistical purposes, the data were broken into
blocks of twenty trials, and analysis of variance was com
puted to test for conditioning taking place under differ
ent reinforcement contingencies and a correlation between
anxiety ratings and rate of conditioning under either form
of reinforcement.
Group I was found to have significantly increased the
use of self statements when compared to Group III. Group
II did not. And, significant positive correlation was
found between anxiety scores and rate of conditioning
(£<.05). Following the final trial, each subject was ques
tioned in order to ascertain whether or not any of them had
become aware of the processes of reinforcement. None of
the subjects was able to verbalize any such insight.
It would appear from this study that the verbal rein
forcement was a more powerful variable than was the non-
62
verbal one, because the word "good" was demonstrated to
have been a successful reinforcing contingency and the
light was not. The writer has some difficulty accepting
the use of the word "good" to train without awareness, how
ever. The additional finding of high anxiety with a posi
tive correlation to rate of learning by means of operant
reinforcement continues to support the importance of the
individual's characteristics playing a role in verbal mod
ification.
Two more important studies in this regard were those
by Ferguson and Buss (25) and by Leventhal (46). Both of
these studies found that different methods of reinforce
ment were successful with different subjects classified as
normal, neurotic and schizophrenic. Leventhal attempted to
vary systematically, characteristics of both subjects and
reinforcing agents on the basis of diagnoses for individuals
and reinforcers involving approval and disapproval.
The subjects were 180 male patients at a state hospi
tal, diagnostically subdivided into three groups: 60 schizo
phrenics, 60 neurotics, and 60 normal patients hospitalized
for nonpsychiatric reasons. The subjects were then randomly
assigned to four groups.
The subjects were presented with 80 3 x 5 cards on
which were printed a common verb in the past tense and four
63
randomly positioned pronouns: "I," "We," "He," "They."
The subjects were asked to form sentences using one of the
top four words and the bottom word. Trials 1-20 were
used for establishing base rate emission. Trials 20 - 80
were treated in the following manner:
Group R, a partial reward group. The experi
menter said "good" after any sentence starting
with "I" or "We." Nothing was said following
sentences beginning with "He" or "They."
Group P, a partial punishment group, for which
the experimenter said "not so good" for each
sentence beginning with "I" or "We." Nothing
was said for those sentences beginning with
"He" or "They."
Group RP, a reward and punishment group, for
which the interviewer said "good" for all sen
tences beginning with "I" or "We" and "not so
good" for sentences beginning with "He" or "They."
Group C, the control group, for which the in
terviewer made no response for any sentence.
Fourteen subjects who were able to describe the mech
anics of the experiment were eliminated from the analysis
of the data.
Analysis of variance was applied to determine whether
c onditioning had taken place in the different treatment
64
groups. It was found to have occurred in Groups R, RP, and
P, but not in Group C, the control group. Analysis was
then applied to determine if there were differences within
the treatment groups according to diagnostic categories of
the clients. It was found that the normal subjects had re
sponded statistically alike in Groups R, P, and RP. The
neurotic subjects performed significantly better in groups
RP and R, but not in Groups P or C, and the psychotic sub
jects indicated significant conditioning in Groups P and
RP, but not in Groups R and C.
One might wonder about only 14 subjects being aware
of such obvious reinforcement, and how much wishing-to-
please hospital authorities played a part in the result. It
is questionable how much reinforcement took place and how
much conformity. However, the differences in the kinds of
reinforcement that were successful with the different
classifications of subjects would seem to validate the
study. Further, it would seem to focus on the importance
of diagnosis, which for a time was subordinate to behavior
modification. It appeared then, that operant learning
theorists might well ignore diagnostic categories and
merely attack the problem of behavior change with positive
reinforcement techniques. Ferguson and Buss' and Leventhal's
findings cast doubt upon that thinking.
Other personal and social characteristics were also
65
found to be significant variables to be accounted for.
Cieutat (16) found a further difference in the condition-
ability of the sexes, with women having lower rates of con
ditioning when the desired operant verbal behavior was of
a controversial nature.
Franks and Mantell (28) studied several variables and
their correlations with rates of conditioning. Using a de
sign very similar to Babladelis, they utilized both the
MMPI and Osgood's Semantic Differential as external cri
teria for analysis of the correlations of conditionality
and alcoholism/nonalcoholism, high IQ/low IQ, verbal flu
ency, vocabulary scores and extroversion/introversion. All
of these groups were found to have conditioned as predicted
through the reinforcement by the sound "mmm-hmm," following
the desired statements. No differences were found among
the classifications and their rate of learning
Effect of Reinforcement Scheduling
An early study by Kanfer (38) was done to determine
whether, indeed, the influence of verbal behavior by what
seemed to be operant conditioning was also operating under
other principles of operant conditioning. As had been
demonstrated by Skinner (75), behavior learned by 100% re
inforcement schedules was learned more quickly but also
extinguished more quickly than behavior learned with partial
reinforcement schedules. The task in this study was quite
66
different from others which have been reported in this
paper. One hundred twenty-four subjects were instructed to
estimate movement of an autokinetic stimulus. A range of
number responses at the upper end of each subject's dis
tribution of 30 initial estimates were selected for rein
forcement. The subjects were divided into 100%, 671, 50%
and 0% reinforcement groups. During acquisition, each sub
ject's critical responses were followed by verbal reinforce
ment according to the schedule for his group. After a cri
terion of six critical responses in ten consecutive re
sponses was reached, extinction was carried on for 75 non
reinforced trials for all experimental groups. Data anal
ysis was then made to determine the correlation of rein
forcement schedules and order of acquisition of the six out
of ten consecutive responses and extinction rate. Acquisi
tion was found to be correlated with the highest to lowest
percentage of reinforcement scheduling procedures in the
predicted manner. Extinction with the continuous schedule
showed greater initial resistance, but faster decline than
extinction with partial reinforcement procedures. During
extinction the total number of critical responses was found
to be greatest in the 67% group, although the 50% group
showed the slowest deceleration in responses. All of these
findings were quite in keeping with the principles estab
lished for operant conditioning.
67
A similar study by Phelan, Tang and Heckmat (59) used
60 college students divided into three groups of twenty for
an attempt to shape verbal behavior by both continuous re
inforcement and partial reinforcement scheduling and then
test for rate of extinction. They tested the following hy
potheses: (a) that subjects conditioned under continuous
verbal reinforcement will emit a significantly higher num
ber of self-referential responses than the control group
during the acquisition phase of the experiment; (b) that
subjects conditioned under partial reinforcement will emit
a lower frequency of self-referential statements than the
group conditioned under a continuous schedule; (c) that the
group conditioned under a continuous schedule of reinforce
ment will emit a higher frequency of self-referent state
ments than the control group during the extinction period;
and (d) that the group conditioned under the partial rein
forcement schedule would emit a higher frequency of self
referent statements than groups conditioned under the con
tinuous schedule of reinforcement during the extinction
period.
Each subject was randomly assigned to one of three
groups; Group I received continuous reinforcement through
out the training period; Group II received a fixed ratio
schedule of reinforcement of 2:1; and Group III, the con
trol group, received no deliberate reinforcement.
68
The subjects were shown a series of pictures and
asked to make a statement about the picture, above and be
yond mere description of the picture. Record was then made
of self statements and reinforcement made of the statements.
The self statement was operationally defined as any in
which the subject described himself, told of himself, or re
ferred to any affect which he experienced. Reinforcement
consisted of the interviewer saying Mmmm-hmmM approvingly
and nodding his head.
The experiment was divided into three periods: pre
training (pictures 1-10), acquisition (pictures 11-40), and
extinction (pictures 41-70). Following the experiment,
there was an interview to try and determine whether the
subjects were aware of the procedures which were being at
tempted. There was no indication of the subjects'
awareness.
Analysis of variance of the first period indicated no
significant differences existed among Groups I, II and III.
During the acquisition period, both Groups I and II emitted
significantly more self-referent statements than did the
control group, significant to the .01 level. The differ
ence in acquisition rate between Groups I and II was sig
nificant at the .05 level. During the extinction period,
the self-referent statements were found to have signifi
cantly lessened when compared to the period of acquisition,
69
but were still greater than the control group. But no sig
nificant differences were found in the rate of extinction
between Groups I and II, though the Group II tended to ex
tinguish at a lower rate, as predicted.
These last two studies, though different in task,
would both seem to support the idea of verbal conditioning
operating under the same principles as other forms of op
erant conditioning. More important, because intermittent
reinforcement of verbal behavior has been found to be ef
fective and resistant to extinction, the implications for
psychotherapy (in which intermittent reinforcement would be
most likely administered, consciously or otherwise) become
more sharply focused.
Generalization Effect
Another question attacked by researchers has been
that of generalization effect of verbal conditioning. "So
you have increased self-referent statements; so what?"
Rudolf Moos (53) attempted to discover whether established
complex verbal habits could be conditioned in free inter
view situations, whether this conditioning effect would be
retained, and whether it would generalize to a different,
although similar situation. Unlike the many studies using
3 x 5 cards with supplied words for sentence construction,
Moos used an open interview setting in which the subjects
were asked to talk about themselves during eight 20-minute
70
sessions covering eight days. Thirty female subjects were
assigned randomly to three treatment groups: Group I, the
independence group; Group A, the affect group; and Group C,
the control group. During the first sessions, a co-experi
menter said nothing. During the second session the inter
viewer said nothing. During four conditioning sessions
with the same interviewer, the experimenter reinforced ac
cording to strategy. The seventh session, the experimenter
said nothing. For the eighth sessions, the co-interviewer
was used again, the task remained the same, and generalized
effect to the new situation was measured.
The categories for this study were operationally de
fined in the following manner. Independence category of
response included expression of needs to be independent,
adult, mature, assertive, ambitious and similar expressions.
The affect category included expressions for needs of love,
appreciation, affection, acceptance, friendship, or the
giving of these factors.
During the training sessions, the interviewer rein
forced each of the appropriate responses with a head nod
and the sound "mmm-hmm." The control group received a
similar response from the experimenter approximately every
30 seconds.
The following hypotheses were tested:
1. The categories of independence or affection
would increase during the training sessions
71
for Groups I and A, but not for Group C.
2. The reinforced categories would be used
more frequently in the postconditioning
session with the experimenter than in the
preconditioning session.
3. The reinforced categories will be used
more frequently with a co-experimenter
than they were in the preconditioning
session with the co-experimenter.
Following the eighth session, the subjects were asked
if they could verbalize what had been taking place. None
indicated they were aware of the conditioning process.
Scoring and reliability were handled nicely. Two
judges were asked to score the appropriate categories of
response for three sessions, in order to see if these cate
gories could be judged reliably. The reliability was found
to be .92, .95, and .95. One of the independent judges
then scored four different sessions with the experimenter.
Their reliability scores for these four sessions averaged
.90. It was decided that the experimenter could reliably
condition the categories involved. The independent judges
were again used to score the categories in both pre and
posttraining sessions.
In order to assess the effect of conditioning, the
percentage scores for each subject, for each category, in
the four conditioning sessions were averaged and were com
pared to the percentage of the category in the precondition
ing session with the experimenter. Both the Wilcoxon
72
matched-pairs signed ranks test and the Mann-Whitney U Test
indicated to the .01 level that conditioning had taken
place for the appropriate categories in Groups I and A. No
change was found for Group C. Thus, Hypothesis 1 was
supported. Using similar analysis, comparison was made for
the first posttraining session and the pre-experiment ses
sion, significance to the .05 level was found, supporting
the second hypothesis. The third hypothesis was not sup
ported, and the authors’ attempt to find a generalization
phenomenon was not successful.
It is probable that the authors of this study were
quite disappointed in their failure to find an indication
of generalization effect in even this minimally different
situation. It would seem that the study gives strong sup
port to the accumulated evidence of the ability to modify
the verbal behavior of others without their awareness, and
this time in a relatively free interview setting. One
might question the effect of talking about one's self for
eight sessions, and wonder if an emotional reaction to the
situation might not be a factor in the failure to support
the final hypothesis.
In many ways, this seemed to be a well designed study.
Although it did not support all of the authors' hypotheses
it contributes much to our understanding of the field.
Ulman, Krasner and Collins (86) also attack the prob-
73
lem of showing a change in generalizing effect of verbal
conditioning, or "A change in the subjects1 behavior in a
situation other than the experimental one.” The "other
situation" used as the dependent criterion measure was the
Group Therapy Scale (GTS), an 88-item scale checked by
therapists for the patient's behavior during their most re
cent four therapy sessions. For this experiment, thera
pists rated their patients before and after the training
sessions, without knowledge of which would be receiving
what kind of treatment. The same therapists made both re
cordings of the GTS for their patients in every case.
The subjects were 30 continued-treatment patients in
a Veterans Administration neuropsychiatric hospital, ran
domly assigned to three treatment groups. Subjects were
seen for two sessions per week for two weeks, and were
asked to make up stories from TAT pictures presented to
them. Four pictures were shown during each session, and
the subjects were asked to make up stories lasting five
minutes each. The pictures used were line drawings of sit
uations common to the patients' experiences and not emo
tionally arousing in themselves.
Group I was reinforced by a head nod and the approv
ing sound "mmm-hmm." Group II was reinforced by pushing a
button attached to an electric counter which emitted a loud
click. The dial was facing the patient and he saw an
74
increased number with each click. Group III, the control
group, received no deliberate reinforcement. In a previous
study, the examiner used throughout the experiment had
listened to sixteen similar four minute tapes with two
other independent judges to test for reliability in recog
nizing the verbal category being used, i.e., emotional
words. The three judges obtained a coefficient of concor
dance beyond the .001 level of significance. It was accep
ted that the experimenter could reliably reinforce var
iables in question; the same experimenter was involved in
all the interviews; the interviews were held, and the rein
forcements were applied as mentioned above.
The mean scores and standard deviations of the three
groups on the GTS before the experimental sessions were not
significantly different. Following the training sessions,
the average gain for emotional word usage in group therapy
for Group I was 5.70 GTS points; for Group II it was 1.00
GTS points; and for Group III it was 0.05 points--a gain
for Group I significant at the .05 level.
This study underlined the advantage of personal, posi
tive reinforcement over impersonal reinforcement such as
buzzers and the like. It added to the stock of evidence
that specific behavior on the part of an interviewer can
specifically influence the verbal behavior of a subject in
a predicted manner without the subject's awareness. Most
75
important of all, this study did present evidence that the
behavior change brought about in the interview situation
can effect change in behavior outside that setting.
It is true that this type of interview situation is
still a long way from a psychotherapeutic interview; but in
only 80 minutes per client, a significant behavior change
was measured in a therapeutic setting as a result of deli
berate verbal conditioning in another setting.
Attitude Change
Wilham Scott (73), in a very different study, found
evidence that reward of certain verbal behavior resulted in
change of attitude. Students who won a rigged debate con
test with a point of view they did not believe signifi
cantly changed their opinions toward the positions they had
presented. Though not an operant conditioning experiment
as such, this study did demonstrate attitude change through
reward of verbal behavior.
Another study indicating an attitude change correla
ted with verbal conditioning was accomplished by D. R.
Slavin (77). This study, somewhat similar to Sarason’s
work (71) measured the attitude toward a hospital. Using
seventy-two hospitalized psychiatric patients, the subjects
were randomly assigned to two groups, experimental and con
trol. The subjects were told they were part of a plan to
evaluate and perhaps change the hospital system of which
76
they were a part. They were asked to talk about the hospi
tal on three different days, by three different interview
ers. The experimental groups were reinforced by a head nod
and the sound "mmm-hmm" for each positive statement that
was made. The control group was reinforced on a fixed
ratio scale. Two days following the third interview men
tioned above, a young man, ostensibly seeking employment in
the hospital, interviewed each of the subjects. Analysis
of their responses to him indicated a significantly greater
percentage of positive statements about the hospital from
the experimental group than from the control group.
This interesting study, while poorly reported,
brought many questions to mind. Did the groups begin with
a similar attitude base? Were they from the same popula
tion? How reliable was the recording of positive state
ments in the last session? How reliable were the reinforce
ment judgments in the three training interviews. What
kinds of statistics were used in the analysis? Unfortun
ately, these questions would seem to be of enough importance
as to completely devalue the stated results. This report
was included in this paper as a good example of poor re
porting on what might have been a good and important exper
iment .
A much better report was done by Coons and McEachern
(19) in which they attempted to find whether conditioned
77
positive statements on one questionnaire would generalize
to increased positive statements on a second questionnaire,
and still more, if increasing self-accepting statements
would result in an increase of statements accepting others.
Generalization was the focused question for these authors,
because "if generalization cannot be demonstrated, verbal
conditioning must remain an interesting but theoretically
and socially useless exercise."
The authors decided upon the dependent variables of
increased self-acceptant and other-acceptant statements not
only to test the generalization effect of verbal condition
ing but to further explore the theoretical concept of Carl
Rogers that greater acceptance of self is in turn accompan
ied by a greater acceptance of others.
Three hypotheses were tested:
1. Subjects who originally score low on a
scale of self acceptance can be induced
to increase their endorsement of self
accepting questionnaire terms by means
of verbal reinforcement.
2. Subjects conditioned on the items of a
self acceptance questionnaire will in
crease the number of self accepting en
dorsements on a similar test composed
of novel items.
3. Subjects conditioned only on items of a
self accepting questionnaire will show a
significant increase in the number of en
dorsements of acceptance-of-others ques
tionnaire items following the conditioning
session.
78
Berger's scale of expressed acceptance of self and
expressed acceptance of others was used to measure atti
tudes toward self and others. Because Scales A and B were
used, it is important to note that the reliability for
these two scales was found to range from .75 to .89.
Form A of Berger's scale was administered to 400 men
of the R.C.A.F. who were probably military "volunteers."
The sixty lowest scores were selected and randomly assigned
to an experimental and control group. These sixty men were
told they were now being tested on how people reacted to
the same questions asked in different ways, that they may
recognize some questions but to go ahead with the test
without worrying about consistency, since consistency and
nonconsistency were the areas of research as well as the
scale content.
They were then administered the Form A again. The
experimenter presented a 3 x 5 card with the question on
it. The subject responded by pushing one of five buttons,
indicating a continuum from nonacceptance to acceptance.
Each time a subject from the experimental group responded
with the accepting button, the interviewer would say "good"
in a flat, emotionless tone, record the response and pre
sent another card. To the control subjects the experimenter
said nothing. A performance criterion was established on a
priori basis; i.e., five conditioning trials of 18 SA items
79
or the mean performance level of the 60 highest scorers
from the original 400 "volunteers." The control subject
received as many trials as the experimental subject imme
diately preceding him. Following the conditioning proce
dure, the subjects were asked to complete a written ques
tionnaire containing 18 self-accepting items from Form A
and 18 self-accepting items from Form B and 14 accepting-
of-others items from Form B, randomly distributed.
The data were compared using the t statistic. All
three hypotheses were accepted. Experimental subjects in
creased the number of self-accepting items of the Form A
with a £<.005. Experimental subjects increased the number
of self-accepting items on Form B with a £<.05. Experi
mental subjects increased the number of accepting-of-others
items on Form B with a £<.025.
In this writer's opinion, this study is very well
done. There is always the question of the operational val
idity of attitude scales, of course, but these kinds of
scales have provided a means of measurement for important
studies in the past--Rogers' and Dymond's study providing a
prime example. Of great significance is the demonstration
of generalization of the conditioned response from one ques
tionnaire to another on items of self acceptance. Of per
haps greater significance is the effect predicted in an
other context by a nonbehaviorist, Carl Rogers, that greater
80
self acceptance would result in greater acceptance of
others. These two findings provide further evidence that
verbal conditioning can be generalized, and that verbal
conditioning can be effective in modifying attitudes which
are theoretically relevant to social interaction.
Awareness Factor
One of the big questions concerning verbal condition
ing which is often asked is that of awareness. Are sub
jects merely trying to please the experimenter or are they
truly being influenced without their own awareness?
The effects of awareness of the contingency between
response and reinforcement has been mentioned by many in
vestigators of verbal conditioning. The findings of the
investigation of awareness are equivocal. Matarazzo (50)
found that only the aware subjects showed evidence of con
ditioning. Kanfer and McBrearty (40) and Kilberg (42)
found that although both aware and unaware subjects evi
denced conditioning, the aware subjects indicated greater
conditioning. Drenner (22) and Nuthmann (55) found no ef
fect from awareness. And finally, Krasner (45) and Drennen
and Greenspoon (23) found that presence of awareness may
have resulted in decreased rate of conditioning. A more
careful look at two studies of this question might be
helpful.
Beech and Adler (7) using a design similar to the
81
study by Taffel (80), conditioned subjects to use the pro
noun "We" in the construction of sentences using verbs and
pronouns supplied via 3 x 5 cards. Their study had several
objectives. The first was to investigate differences be
tween normals, depressives, schizophrenics, and neurotics
regarding verbal conditioning. Their second objective was
to investigate the area of awareness of process more thor
oughly than has usually been the case by means of asking a
standard set of questions. The third objective was to dis
cover what, if any, relationship exists between extrover
sion and anxiety on conditioning scores.
Subjects used were classified by the staff of the
psychiatric hospital in which all but the "normals" were
being treated. The "normals" were a population of volun
teers from a college class. The measure for anxiety and
extroversion was the Maudsley Personality Inventory.
The conditioning sessions consisted of 160 cards be
ing presented by an interviewer to each subject. The
cards contained the randomly sequenced pronouns "I," "We,"
"He," "She," They;" and a simple verb in the past tense.
The subjects were asked to pause after the pronoun before
continuing with the rest of the sentence, allowing the ex
perimenter to reinforce the pronoun "We" immediately with
the word "good." The experimenter said "yes" after all
other pronouns, and "thank you" after each sentence.
82
Following the training, the subjects were administered the
awareness questionnaire.
Twenty-seven subjects were judged to be aware; eighty-
five were judged to be unaware. The t test was then used
to compare the number of "We" pronouns in the first 80
cards with the number in the last 80 cards. This was done
by a number of categories.
For the unaware group, only the schizophrenics were
found to have conditioned significantly. For the aware
groups, only normals and schizophrenics had been condition
ed. No relationship was found between anxiety or extrover
sion and either awareness or conditioning rate.
On the basis of the above, the authors concluded that
successful conditioning done with normal populations was
probably more the result of awareness and cooperation than
the result of actual conditioning. However, the manner in
which the data was analyzed raises a question about the rel
evancy of the conclusions. In so many of the preceding
studies there was evidence of early reinforcement effect.
There is every reason to believe that conditioning had be
gun before the first 80 trials were concluded. If this
were true, using the first 80 trials as a base rate would
negate the chance of finding significant increase in the
behavior in the last 80 trials, using only a t test. A
much better design would have been to use smaller blocks of
83
trials before measuring for or to use analysis of var
iance to measure change.
The method used for reinforcement was also obvious
and it is no surprise that 25% of the subjects, 60% of whom
were normals, were aware of the process. The method almost
guaranteed that only those normals who were aware would
show significant increase.
S. Levin's study (47) is probably a good one to in
dicate the ambiguity of the state of awareness. Levin sus
pected that most investigators of verbal conditioning had
not thoroughly explored the state of awareness of the sub
jects regarding response and reinforcement contingencies.
Rather, he believed that more subjects than indicated were,
in fact, aware to some degree of the reinforcing process,
which he believed would account for the success of the con
ditioning process. He attempted to investigate the aware
ness factor more thoroughly in the following manner.
His subjects were 79 male patients at a Veterans Ad
ministration Hospital, 50 of whom were hospitalized for
neuropsychiatric reasons, and 29 of whom were there for
medical disorders. His stimulus materials were 100 3 x 5
white cards with "I," "We," "You," "He," "She," "They,"
printed in random order and a simple verb in the past tense,
taken from Thorndike and Lorge's list of 1,000 commonly
used words in the English language.
84
The subjects were administered, as a group, the Taylor
Manifest Anxiety Scale and the Edwards Personal Preference
Schedule. They were then interviewed individually. The
subjects were presented with individual cards and asked to
say the first sentence that came to mind using one of the
pronouns and the verb shown. For the trials 1 - 20, no
difference was followed in the procedures for either the
experimental or control group. From trials 21 - 100, the
experimental subjects were told "good" in a flat voice fol
lowing any sentence beginning with either "I" or "We." For
the control group "there was no change from the procedures
for trials 1 - 20."
Immediately following the sessions, each subject was
carefully interviewed in order to elicit information regard
ing his awareness of the reinforcement contingencies he had
just experienced. Sixteen questions were asked of the sub
jects which were very carefully designed to discover the
least bit of awareness on the part of the subject without
suggesting the facts to him. In the writer's opinion,
Levin was quite successful in the construction of this
questionnaire. Using only the first four questions of the
questionnaire, which were similar to the kinds of questions
most probably asked in a casual attempt to find awareness,
only three of the subjects would have been detected as being
aware of the reinforcing contingencies. Using all of the
85
questionnaire, 16 additional subjects were classified as
aware. This represents a substantial increase.
For analysis of the data, the trials were divided into
five blocks of twenty trials each. The mean frequencies for
each of these blocks were then treated by analysis of var
iance. Analyzing the data of all the experimental subjects
except the first three identified by casual questioning, the
experimental group was found to have increased the target
pronoun usage significant to the .05 level; a result similar
to many previous studies of this type. The subjects' data
were then analyzed by four groups:
1. Aware Group. The 16 subjects classified
as aware by the entire questionnaire.
2. Unaware Group. The 42 subjects not class
ified as aware.
3. Totally Unaware Group. Those subjects
from the above 42 subjects who weren't
even aware of the word "good" being said.
4. The Control Group.
Analysis of variance using Lindquist's technique was
employed with the following results:
1. Aware group/control group . . . . £<.005
2. Aware group/ unaware group. . . . £<.05
3. Unaware group/control group . . . £<.20 (N.S.)
4. Totally unaware group/control
group ......... . . . . . . . . £<.005
From the above data it is clear that most of the var
iance originally found in the experimental group with the
86
first analysis can be accounted for by those subjects whose
awareness would not have been detected in the usual manner
of investigation. Using only those subjects who were class
ified as unaware of how the word "good" was being used, no
significant difference was found from the control subjects.
The strong implication is that those heretofore studies
which indicated successful training without awareness may
well have been using data that included many subjects who
were in fact aware of the process, but who were not detec
ted as such by the experimenter.
There is some room for debate however. Not all the
variance is accounted for in comparing the three groups
differently; i.e., when comparing the aware group to the
unaware group and to the control group, £ is different sig
nificantly, indicating, at least a strong direction of
change as predicted. A further analysis of subjects who
were completely unaware of the word "good" being spoken, as
well as being unaware of the correct contingency relation
ship, as differentiated from those who were aware of the re
inforcer but not of what it meant) found significant in
crease to the .005 level. The importance of detecting
those who are aware of the reinforcing contingencies has
not been sufficiently emphasized. Perhaps the expression
"mmm-hmm" would be less obvious than the word "good," but
the entire area of verbal conditioning is an ambiguous area.
87
Summary
Much effort has gone into the investigation of many
factors which are involved in and affect verbal condition
ing. While the modification of verbal behavior has proven
to be a more complex learning than was first thought to be
true, more variables are becoming known and accounted for
within this experimental model. However, while much of
verbal conditioning has been researched and found to follow
many principles of learning theory, no attempt has been
made, so far, to shape verbal behavior through several
stages of behavior, in a complete operant paradigm. No one
has, as yet, put verbal conditioning to the real test of
the operant: gradual development of specific verbal beha
vior through a sequence of planned and predicted stages.
Until this is done it is questionable whether one can as
sume that the principles of operant conditioning can be
assumed to be applicable to the procedures of verbal con
ditioning. Should it be found that a subject's verbal be
havior can be "shaped" without his awareness, then the
power of reinforcement of verbal behavior takes on a new
dimension to be considered in the process of the interview
setting. The implications for counseling become more im
portant. The intent of this study is to attempt to put
verbal conditioning to this important test: the literal
shaping of verbal behavior.
A Word about Design
In reviewing the above literature, two basic designs
were used extensively, modified to fit the particular hy
pothesis being tested.
The design by Greenspoon, testing for positive or
negative reinforcement, classification of reinforcement
contingencies, acquisition and extinction, is one of the
basic models. Five groups were used, one for each of the
tested contingencies, and a control group for comparing
the effect of the manipulated variables. Using Kerlinger's
nomenclature this design can be outlined in the following
manner:
Pre Post Extinction
Xx Yi Zi Experimental plural
nouns; mmm-hmms.
x2 y2 z2 Experimental plural
nouns; uh-huhs
Xj y3 z3
Experimental nonplural
nouns; mmm-hmms.
X- Y* Experimental nonplural
nouns; uh-huhs.
X5 Ys Z5 Control
This, according to Kerlinger (41:303), may be the
ideal design of most educational and social scientific re
search. It has the best built-in theoretical control sys
tem of any other design with one or two possible exceptions
in special cases; it is flexible, being theoretically
89
capable of extension to any number of groups with any number
of variables; it can test several hypotheses at one time;
and it is statistically and probabilistically elegant (41:
303). The control group gives the comparability required
by science. The classical experimental group control-group
design using equated groups provides such comparisons in an
efficient manner; and the randomization of subjects assured
the experimenter that his groups are statistically equal on
any variables possibly related to the dependent variable.
Buss and Durkees (13) modified this by having two ex
perimental groups, then using each group as a control group
of the other. "Whenever there is more than one experimental
group and any two groups are given different treatments,
control is present" (13:306).
B. Sarason (71) used a design similar to both that of
Greenspoon, and Buss and Durkee, in effect gaining two con
trol groups for each experimental group. In her analysis
of the data, Sarason did not take advantage of the aspect
of her design, and used their own base rate and control
group data for comparison. Other modification of Green-
spoon's design was also found in the studies by Babladelis,
fTl Xi Yx
experimental
hostile verbs
control for
each other
experimental
X2 Y2 neutral verbs
90
Taffel and Leventhal.
Base Rate Post
m
positive statements
Y. Xi Y„ about hospital \
b » control for
each other
negative statements^
Y^ X2 Yf l about hospital
control group--reinforcement
Y^ X3 Y& every ten seconds
The second basic design reported was that of J. M.
Rogers, in which the experimental group was used as its own
control.
Base Rate Post
Actually, according to Kerlinger, when a group is compared
with itself, this may be the very best design of all (.41:
26), if negative considerations almost inherent to this de
sign have been accounted for.
In his discussion of research design, Kerlinger points
out that the main technical function of a research design is
the control variance (41:280). He further expands this
statement "Our hope for a good research design is to maxi
mize systematic variance, control extraneous variance, and
minimize error variance" (41:280). He further states, "If
the same subjects are used with different experimental
treatments, we have the most powerful control of variance
91
there is" (41:26). Still later he says again "When a group
is compared with itself, there is no better control since
all possible independent variables associated with the sub
jects’ characteristics have been controlled" (41:295). So
Rogers' use of this design model is not to be dismissed im
mediately as unscientific. It is a design not generally ap
proved because of negative consideration usually not ac
counted for in most research. Campbell and Stanley (15)
have listed these factors and commented on them extensively:
1. History, the specific events occurring be
tween the first and second measurement in
addition to the experimental variable.
2. Maturation processes within the respond
ents operating as a function of the pas
sage of time, including growing older,
growing hungrier, growing more tired,
and the like.
3. Testing, the effects of taking the test
upon the scores of a second testing.
4. Instrumentation, in which changes in the
calibration of a measuring instrument or
changes in the observers or scorers used
may produce changes in the obtained mea
surements .
5. Statistical regression operating where
groups have been selected on the basis
of their extreme scores.
6. Biases resulting in differential selection
of respondents for the comparison groups.
7. Experimental mortality of differential loss
of respondents from the comparison groups.
8. Selection-maturation-interaction, etc. which
in certain multiple group quasi-experimental
92
designs might be mistaken for the experi
mental variable.
9. Reactive effect of testing in which a pre
test might increase or decrease the respond
ent's sensitivity or responsiveness to the
experimental variable.
10. Interaction effects of selection biases and
the experimental variable.
11. Reactive effects of experimental arrangements
which preclude generalization about effects
of the experimental variable upon persons being
exposed to it in nonexperimental settings.
12. Multiple treatment interference, likely to
occur whenever multiple treatments are ap
plied to the same respondents because the
effects of prior treatments are not usually
erasable.
Looking at Rogers' study we see that the first poten
tial contaminator, history, was scarcely a consideration be-
cause of the nature of the task (which in 24-48 hours was
not likely to be affected by social events) and because of
the brevity of time lapse for the entire experiment. The
second potential contaminator, maturation, was felt to be
controlled for the reasons just listed. Testing was not apt
to be a factor because the subjects were unaware of the real
testing mechanisms. Instrumentation, too, can be ruled out
as a rival explanation of results, if indeed the observer
did become fatigued or less alert, because it would work
against his scoring results contributing to his attempts to
influence verbal behavior. The length of time for each in
terview (approximately 40-60 minutes) and the specificity of
93
the task were also felt to rule out the possibility of this
variable being a factor.
The fifth confounding variable is irrelevant to this
study.
The biases resulting in differential selection of re
spondents for the comparison groups was precluded by using
the group as its own comparison. The same is true of the
variable of experimental mortality. The eighth variable was
covered by all which has preceded.
The reactive effect of testing was in a sense what is
being manipulated in this experiment without the subject's
awareness; this, too, can be discounted as an argument
against this design--it was the object of the study!
The interaction effects of selection biases and the
experimental variable appeared to be an irrelevant concern
in terms of the questions at hand. Possibly one could say
that the persons who volunteer for an experiment are apt to
be those who unconsciously try to please others and are
easily influenced, thereby casting doubt on the outcome of
this experiment. In the writer's opinion, that was not rele
vant. The attempt was to find whether or not one can
"shape" the verbal behavior of subjects in a quasi-interview
setting. The ease of influencing their verbal behavior, or
the lack of it, was not in question.
The eleventh rival hypothesis was not a relevant one.
94
Affecting verbal behavior in the experimental setting was
the only thing in question.
Rogers' use of a group as its own control would seem
to be valid. Avoiding the contaminations outlined by Stan
ley and Campbell which usually plague the one group design,
he has instead been able to capitalize upon the inherent
strength of such a design as declared by Kerlinger.
Turner (85) used three groups in his study, but ac
tually used each group as its own control and then deter
mined if one kind of group conditioned more readily than
others. In reality, this is a modification of Rogers' de
sign. The same is true of Kinzie's study (43).
Binder, McDonald and Sjoholm (8) used two groups, but
reinforcement influence was measured by comparison to each
group's own base rate, as in Rogers' design. The reason for
the two groups was for comparison of different interviewer
characteristics, but the determination of reinforcement in
fluence was based upon own group control. Much the same can
be said of Sapolsky's work and that of I. G. Sarason.
The use of a one group design, then, seems widespread
in the study of verbal conditioning, and can be justified
as long as the contaminants outlined by Stanley and Campbell
are accounted for. This area of study is, indeed, one of
the exceptions mentioned by Kerlinger which can capitalize
95
on the inherent strengths of the one group design.
In keeping with the above, the writer has therefore
decided to use this design for the present study. Though
extending the conditioning of verbal behavior through sev
eral stages of behavior shaping will contribute to the
uniqueness of this study, its similarity to J. M. Rogers
approach of investigation is strong, and the justification
for Rogers' study would seem applicable to the present
effort.
CHAPTER III
PROCEDURE AND METHODOLOGY
Procedure
Subjects
The subjects who cooperated in this study were volun
teers from several graduate education courses at San Diego
State College. All volunteers were graduate students in
their fifth year of college preparation, all of whom had
received bachelor degrees with a grade point average of 2.4
or better. It was therefore assumed that all participants
were relatively skilled in verbal communication. Nothing
in the process of this study led the writer to doubt the
validity of this assumption.
The subjects were asked during a scheduled class per
iod to cooperate in a study, the stated goal of which was
to verify, statistically, some of the theories of communica
tion which have been promulgated by the famous linguist,
Hayakawa. This simple description of the experimental task
was seemingly quite a sufficient reason for those volunteer
ing for the research, and no doubt was voiced by the parti
cipants about the reason for the study. The volunteers
came to their individual interviews at the appointed times;
96
97
the reason for their task was again reiterated, again accep
ted, and the interview task accomplished. At the conclusion
of the individual interviews, the writer continued in casual
conversation with each of the participants, allowing them
the opportunity to question him further about the study. At
the same time, he further investigated whether the subjects
were indeed aware of the real intent and process that had
been carried out during the interview. In no instance was
there evidence of doubt of the stated purpose of the study,
nor was there any apparent awareness of the actual manipu
lations carried on during the interview.
Materials
The instrument used for this experiment was three
hundred twenty white, unlined, five-by-eight-inch cards up
on which were typed, in large primer type, three lines of
words. The first line of print included four nominative
pronouns: "I," "He,” "She," "You." The second line in
cluded two verbs in the past tense. The third line of print
was composed of four objective pronouns: "Me," "You," "Her"
and "Him."
The order or sequence of the pronouns in lines one
and three was rotated throughout the series of cards so
that no one pronoun was given a fixed position on the card
other than the assignment to a particular line.
The verbs of line two consisted of verbs of two cate
98
gories. One category shall be called "hostile" or "aggres
sive" verbs. The second category shall be called "nonhos-
tile" or "nonaggressive" verbs. (See Appendix A.)
Verbs used in this study were classified as being
"hostile/aggressive" or "nonhostile/nonaggressive" in the
following manner. A list of one thousand verbs was collec
ted from Webster's Seventh Collegiate Dictionary. Each
verb was then transferred to individual three-by-five inch
white, unlined cards, and this collection of one thousand
cards was presented to three high school English teachers.
The teachers were asked, independently, to divide these
verb cards into three piles to be labeled "hostile/aggres
sive," "nonhostile/nonaggressive," and "don't know." When
a verb was so labeled by at least two teachers, it was so
categorized for purposes of this study. The verbs were
then placed on the interview cards in a rotated order so
that placement on the card would not influence usage in a
systematic way.
The Task
Upon arrival at the interview room, the clients were
seated at a desk. Before them was placed the stack of
cards to be used for the interview. Because two interviews
were to be utilized for this study, one hundred sixty cards
were placed before the student for each session. The stu
99
dents were then given the following directions:
Each card before you is to be used as a spring
board for creating a series of unrelated sen
tences. You are to look at a card, choose a
word from each line on that card, and use them
with as many other words as you wish, to create
a sentence. You are to say that sentence aloud
to me, in order that I may hear it and analyze
the sentence structure you have used. You are
not limited to using only the words on the card,
but you may do so if you wish. The only re
striction is that you use a word from line one
as the subject of the sentence, a word from line
two as the verb of the sentence, and a word from
line three as an object in the sentence.
Usually at this point a question would be asked and
examples given for the task, by the writer. The directions
were then continued:
After you have said a complete sentence, remove
the card from the original pile, place it on
another stack, and go on to the next card. You
are then to follow the same process for making
another sentence; another card, another sentence,
until you have completed the stack before you.
Go at your own pace. Do not wait for a signal
from me to move to another card. While you are
saying these sentences, I will be here at my
desk keeping record of the different sentence
components you are using. I am sure I will be
able to keep up with you, so don't worry about
me. Go at whatever pace is comfortable for
you. Are there any questions?
The students usually asked the writer something about
sentence length or speed or some other irrelevant point.
The questions were answered as simply as possible, and the
student would begin.
The Process
The student actually was working with two stacks of
100
eighty cards for each session, one hundred sixty cards for
the interview. Each stack of eighty cards was dealt with
differently in the following manner: The initial eighty
sentences were used for establishing the base rate of the
student to see what kind of distribution of word combina
tions from lines one, two, and three were "natural" for the
student without any conscious influence by the writer being
exerted. As the student finished each sentence, and card,a
record was kept of what word combinations were used with
the first eighty trials. No deliberate sounds or signals
were communicated to the student at this time.
When the student began composing sentences from the
next eighty cards (Trial 2), the interviewer made the soft
sound, "umm-hmm," each time the student completed a sentence
which had used the pronoun "she" as the subject of the sen
tence. For sentences using "he," "I," or "you" as the sub
ject, no deliberate signal was given to the student. The
interviewer merely recorded the combinations which had been
used. The introduction of the sound, "umm-hmm," was done
very unobtrusively. Every effort was made to make it appear
that the interviewer was unaware of having made the sound.
It was to appear to be a kind of habitual sigh that the in
terviewer uttered from time to time, apparently uncon
sciously.
After the second pile of eighty cards had been com-
101
pleted, the interview was terminated and another one sched
uled the following day.
During this second trial, the apparently habitual
"umm-hmmM was used only with sentences using the word ’’she"
for the sentence subiect. Thus was introduced the experi
mental reinforcement, the operant which would hopefully af
fect the verbal behavior of the student, without his aware
ness, along the direction desired by the interviewer.
During the following interview, the student was again
presented with two piles of eighty cards with which to com
pose sentences aloud. As the student used the first set of
eighty cards (Trial 3), once again the sound "umm-hmm" was
used as a reinforcer, but this time only for those senten
ces in which the pronoun "she" was the subject and in which
the nonaggressive verb was the predicate. Sentences using
"she" as the subjects coupled with the aggressive verbs
(formerly also reinforced) were treated by the interviewer
as those sentences with "I," "he," or "you" as the subjects,
i.e., no reinforcement was used; merely the recording of the
combination of words used (extinction).
With the final eighty cards (Trial 4) further refine
ment of the desired behavior was sought by reinforcing with
the sound "umm-hmm," only those sentences in which "she"
was the subject, the nonaggressive verb was the predicate,
and the objective pronoun "him" was used as the object in
102
the sentence. Any other combinations used were not con
sciously reinforced; they were merely recorded (further ex
tinction) .
When the last card had been used by the student, the
interview was officially terminated. The writer then en
gaged the student in casual conversation, thanked him for
his cooperation, and attempted to ascertain whether the stu
dent was at all suspicious about what had taken place or
whether he was satisfied that he understood what was being
attempted,as had been explained. In every case, the stu
dents appeared to be satisfied with the explanation which
had been given. In no case was there any evidence of aware
ness by the student, of the writer's having used the sound
"umm-hmm" for any purpose whatever. The student was then
dismissed and the research attempt with him was completed.
To reiterate, after determining the individual base
rate, the writer, using the sound "umm-hmm" as the operant,
reinforced first, all sentences using the pronoun "she" as
the subject of the sentence. Secondly he reinforced only
sentences with "she" as the subject and nonaggressive verbs
as the predicate. Thirdly, only those sentences using
"she" as the subject, nonaggressive verbs as the predicate,
and the pronoun "him" as the subject were reinforced. There
fore each trial was an attempt to influence the verbal be
havior of the client in a specific way; each successive
103
trial was an attempt to influence the client's verbal beha
vior in an increasingly specific or precise fashion, as
demonstrated by the combinations of words used in the com
posing of oral sentences.
Hypotheses
The following hypotheses were formulated:
1. At the end of Trial 2, there will be no
significant increase in the number of the
client's sentences with the pronoun
"she," as the subject of the sentence
when compared with the sentences of
Trial 1.
2. At the end of Trial 3 there will be no
significant increase in the number of the
client's sentences with the pronoun "she,"
as the subject and the nonaggressive verb
as the predicate of the sentence, when
compared with the sentences of Trials
1 and 2.
3. At the end of Trial 4, there will be no
significant increase in the number of the
client's sentences with the pronoun "she,"
as the subject, nonaggressive verbs as the
predicate, and the pronoun "him," as the
object of the sentence when compared with
the sentences of Trials 1, 2 and 3.
CHAPTER IV
ANALYSES OF RESULTS
In analyzing the raw data of this study, it was de
cided to test for the significance of change of the propor
tions of appropriate sentence elements used by the clients
in each trial of the interviews. Because an attempt was
made to influence the verbal behavior of each client in an
increasingly more precise direction with each trial, it was
decided to use an analysis which would indicate direction
as well as significance of proportional changes for each
trial for both individual and group data. Though the pri
mary concern of this study is with group data, the writer
felt that analysis of individual data as well as group data
would be both interesting and valuable.
The Test of Signs, though not a strong statistical
tool, was utilized in order to evaluate directional change
of the proportions of appropriate sentence elements used
for each trial. The standard score or z score was then
used to test for significance of any change in the propor
tions of appropriate sentence elements for each trial, for
both individual and group data.
Hypothesis 1: At the end of Trial 2, there
will be no significant increase in the num
ber of the client's sentences with the pro-
104
105
noun "she," as the subject of the sentence
when compared with the sentences of Trial 1.
Table 1 shows the proportions of sentences using
"she" as the subject of the sentence for Trials 1 and 2 and
the direction of change in the proportions of these two
trials. Trial 1 was the base rate trial in which each
client composed all his sentences without any known rein
forcement .
TABLE 1
THE PROPORTIONS OF SENTENCES USING "SHE" AS THE
SUBJECT OF THE SENTENCE FOR TRIALS 1 and 2
Subject Trial 1 Sign Trial 2
1 .228
+ .270
2 .175 .162
3 .235
+ .240
4 .247
+ .312
5 .228
+ .300
6 .175
+ .267
7 .235
+
.250
8 .160
+ .430
9 .408
+ .450
10 .300
+ .450
11 .272
+
.312
12 .272
+ .482
13 .284
+
.333
14 .296
+ .600
15 .222
+ .312
16 .185
+
.325
17 .049
+
.250
18 .225
+
.333
19 .235]
+
.325
20 .284
+
.412
106
Trial 2 was that trial during which the interviewer
made the sound, "umm-hmm," after each client's sentence
using "she?'as the subject of the sentence. For nineteen of
the twenty subjects, there was a greater proportion of sen
tences using "she" as the subject of the sentence in Trial
2 than in Trial 1. The Test of Signs was used to test that
this was not a chance increase. The probability of this
occurring by chance was found to be less than .001. It was
concluded that the reinforcement of "she" sentences tended
to result in a larger proportion of "she" sentences than
were found in the base condition.
Table 2 illustrates the z scores of individual sub
jects for evaluating the significance of change in the pro
portion of sentences using "she" as the subject of the sen
tence of Trial 2 compared to Trial 1. As indicated by as
terisks, only five subjects were found to have z scores
significant to the .05 level, indicating a significant in
crease in the proportion of "she" sentences under the re
inforcement condition from the base rate for those indivi
duals only. So, though nineteen out of twenty subjects
showed an increase in the proportion of "she" sentences
under the reinforcement condition, only four of the nineteen
proportional increases reached a level of statistical sig
nificance in Trial 2.
However, considering the entire population as a single
107
unit, a z score was calculated comparing the proportion of
sentences using "she" as the subject of the sentence under
the reinforcement condition with that of base rate.
z = 6.01 (Pi .001)
On the basis of this finding the null hypothesis was
rejected.
TABLE 2
STANDARD SCORE VALUES FOR INDIVIDUALS COMPARING THE PRO
PORTION OF SENTENCES USING "SHE" AS THE SUBJECT OF THE
SENTENCE FOR TRIALS 1 and 2
Subject z Score
1 1.00
2 0.00
3 0.03
4 0.77
5 0.85
6 1.00
7 0.00
8 3.57**
9 0.62
10 1.79*
11 0.40
12 1.99*
13 0.51
14 0.89
15 1.11
16 1.85*
17 3.85**
18 1.33
19 1.10
20 1.56
*p < .05
**p <.01
108
Hypothesis 2. At the end of Trial 3, there
will be no significant increase in the num
ber of client's sentences with the pronoun
"she," as the subject and nonaggressive verb
as the predicate of the sentence when com
pared with the sentences of Trials 1 and 2.
Table 3 illustrates the data for the
for Trial 3 compared with Trials 1 and 2.
TABLE 3
THE PROPORTIONS OF SENTENCES USING "SHE'
JECT OF THE SENTENCE AND A NEUTRAL VERB
ICATE FOR TRIALS 1, 2, and 3.
Test of
' AS THE
AS THE
Signs
SUB-
PRED-
Subject Base Rate Sign She Sign
She/
Neutral
1 .222
+ .263 .231
2 .062
+
.100
+
.235
3
4
.148
.124
+
.114
.175
+
+
.250
.250
5 .072
+ .162
+
.300
6 .062
+
.162
+
.175
7 .111
+
.150
+
.262
8 .074
+
.215
+
.275
9 .210
+
.304
+
.400
10 .148
+ .212
+
.312
11 .124
+
.175
+
.262
12 .161
+
.296
+
.455
13 .148
+
.210
+
.275
14 .124
+
.188
+
.212
15 .136
+
.162
+
.337
16 .074
+
.175
+
.304
17 .012
+
.112
+
.250
18 .150
+
.235
+
.337
19 .136
+
.225
+
.288
20 .148
+
.262
+
.284
109
During this trial the author reinforced with the
sound "umm-hmm," only those sentences composed by the sub
jects, which used "she" as the subject of the sentence
and a neutral verb as the predicate. The proportion of
this type of sentence from the total number of sentences
composed by the client during Trial 3 were compared to
the proportion of the same type of sentence from the total
number of sentences used by the client in each of the two
previous Trials. As will be noted, the proportion of sen
tences using "she" as the subject and a neutral verb as
the predicate of the sentence, to the total number of sen
tences used in the reinforcement condition (Trial 3) was
greater than the proportion of the similar types of senten
ces in either Trial 1 or Trial 2 in twenty out of twenty
cases. The probability of this occurring by chance alone
was less than .001. Also to be noted is the consistent
trend toward this end evidenced by Trial 2 during which all
"she" sentences were reinforced, both with neutral and ag
gressive verbs used as predicates. Though this resulted in
an increase in the proportion of sentences using "she" as
the subject and a neutral verb as the predicate of the
sentence, it was to a much lesser degree than in Trial 3,
in which only this kind of sentence was reinforced. This
may well indicate increasingly specific influence being
110
manifest under the reinforcement condition.
Table 4 shows the individual z scores comparing the
proportion of clients' sentences using "she" as subject and
neutral verb as predicate of the sentence to the total number
of sentences composed by the clients during Trial 3 to the
similar proportions of Trials 1 and 2.
TABLE 4
STANDARD SCORE VALUES FOR INDIVIDUALS COMPARING THE PROPOR
TIONS OF SENTENCES USING "SHE" AS THE SUBJECT AND A NEU
TRAL VERB AS THE PREDICATE IN TRIALS 1, 2, and 3
i 2
Subject z z
1 0.06 0.40
2 2.87** 2.11*
3 1.42 2.27*
4 3.46** 0.96
5 3.41** 1.88*
6 2.27* 0.00
7 2.26* 2.44**
8 3.25** 0.66
9 2.45** 1.10
10 2.29* 1.21
11 1.79* 1.15
12 4.05** 1.92*
13 1.78* 0.78
14 1.30 0.14
15 2.83** 2.37**
16 3.52** 1.72*
17 4.24** 2.05*
18 2.58** 1.27
19 2.17* 0.72
20 1.91* 0.40
*p < .05 z1 Compares Trials 3 and 1.
**p < .01 z2 Compares Trials 3 and 2.
Ill
The column designated "z1," is a comparison of the
proportion of sentences using "she" as the subject and a
neutral verb as the predicate, with total number of sentences
used by the client in Trials 1 (base rate) and 3 (reinforce
ment condition).
The column designated "z2," was included in this
table to give a more complete picture of the developing in
fluence of the reinforcement condition. This column shows
a standard score value of the proportion of sentences using
"she" as the subject and a neutral verb as the predicate
with the total number of sentences used by the client in
Trial 3 (the reinforcement condition in this stage of the
study) and Trial 2 (in which all "she" sentences were rein
forced) . Again, it can be seen that each succeeding trial
indicates an increase in the specificity of effect of the
operant condition.
In testing the hypothesis, however, it is the group
data that is of statistical importance. Considering the
entire population as a single unit, a z score was calculated
comparing the proportion of sentences using "she" as the
subject and a neutral verb as the predicate of the sentence
under the reinforcement condition with that of the base rate.
z - 11.44 (pi .001)
On the basis of this finding, the null hypothesis was
rejected.
112
For additional information, a similar comparison was
made to evaluate statistically, the difference in propor
tion of the type of sentence in question between Trial 3
and Trial 2. The difference was found to be significant,
again at the .001 level (z = 6.02), indicating an increase
in the sentences reinforced even when compared to a trial
in which a conditioning apparently occurred for a slightly
grosser category of sentence structure.
Hypothesis 3. At the end of Trial 4, there
will be no significant increase in the num
ber of client's sentences with the pronoun
"she" as the subject, nonaggressive verbs
as the predicate and the pronoun "him" as
the object of the sentence when compared
with the sentences of Trials 1, 2, and 3.
Table 5 illustrates the data for the Test of Signs
for Trial 4 compared with Trials 1, 2, and 3. During this
trial, the author reinforced with the sound "umm-hmm," only
those sentences spoken by the clients in which "she" was
the subject, a neutral verb was the predicate, and "him"
was the object of the sentence. The proportion of this
type of sentence from the total number of sentences spoken
by the client during Trial 4 were compared to the propor
tion of this type of sentence from the total number of sen
tences spoken by the client in each of the three previous
trials. As shown by Table 5, the proportion of sentences
using "she" as the subject, a neutral verb as the predicate,
and "him" as the object of the sentence,when compared to
113
the total number of sentences in Trial 4, was greater than
twenty out of twenty subjects in any of the previous trials.
TABLE 5
PROPORTION OF "SHE/NEUTRAL/HIM" RESPONSES FOR BASE RATE,
"SHE" REINFORCEMENT CONDITION, "SHE/NEUTRAL" REINFORCEMENT
CONDITION, AND "SHE/NEUTRAL/HIM" REINFORCEMENT CONDITION
Base She/ She/
Subject Rate Sign She Sign Neutral Sign Neutral^im
1 .025
+
.053 + .115
+
.162
2 .025 0 .025 + .062
+
.138
3 .049
-
.038 + .088
+
.188
4 .049
+
.075 + .100
+
.212
5 .038
+
.062 + .100
+
.225
6 .012
+
.025 0 .025
+
.038
7 .049
-
.038 + .138
+
.225
8 .037
+
.076 .062
+
.150
9 .124
+
.215 + .250
+
.342
10 .074
.+
.075 + .100
+
.175
11 .037
+
.050 + .088
+
.175
12 .111
+
.222 + .278
+
.338
13 .062
+
.086 + .125
+
.188
14 .074
-
.050 + .088
+
.188
15 .111
-
.088 + .188
+
.250
16 .025
+
.062 + .100
+
.173
17 .000
+
.025 + .050
+ .112
18 .038
+
.062 + .125
+
.150
19 .074
+
.100 + .125
+
.212
20 .086 .038 + .099
+
.200
The probability of this occurring by chance alone was less
than .001. Again, this table of the Test of Signs seems to
indicate the increasing specificity of influence being ex
erted by the operant condition.
114
Table 6 shows the individual z scores comparing the
proportion of clients' sentences using "she" as the sub
ject, a neutral verb as the predicate, during Trial 4 to
the similar proportions of Trials 1, 2, and 3.
TABLE 6
STANDARD SCORE VALUES FOR INDIVIDUALS COMPARING THE PROPOR
TIONS OF SENTENCES USING "SHE" AS THE SUBJECT, A NEUTRAL
VERB AS THE PREDICATE, AND "HIM" AS THE OBJECT OF THE
SENTENCES FOR TRIALS 1, 2, 3, and 4
Subject Zl z2 z3
1 2.74** 1.70* 0.62
2 2.32* 2.32* 1.32
3 2.47** 2.73** 1.61
4 2.84** 2.25* 1.74*
5 3.25** 2.70** 1.93*
6 0.52 0.00 0.00
7 3.01** 3.28** 1.23
8 2.20 1.22 1.54
9 3.09** 1.60 1.10
10 1.90* 1.67* 1.15
11 2.59** 2.25* 1.40
12 3.26** 1.47 0.63
13 2.39** 1.84* 0.87
14 1.90* 2.44* 1.61
15 2.09* 2.53** 0.77
16 2.75** 1.93* 1.09
17 2.74** 1.87* 1.16
18 2.17* 1.57 0.22
19 2.28* 1.74* 1.26
20 1.83** 2.93** 1.58
*p<.05 **p<.01
Zi Comparison of proportion of sentences using "she/neutral
verb/him" responses for Trials 1 and 4.
z2 Comparison of proportion of sentences using "she/neutral
verb/him" responses for Trials 2 and 4.
Zs Comparison of proportion of sentences using "she/neutral
verb/him" responses for Trials 3 and 4.
115
The column designated "zx" is a comparison of Trials 4 and
1; "zz" is a comparison of Trials 4 and 2; and "Z3" is a
comparison between Trials 4 and 3. The columns "z*" and
"z3" were included in this table to give a more complete
picture of the increasing specificity of influence appar
ently taking place during the reinforcement condition. Most
important for this table is the column, MZ|" in which nine
teen subjects show a z score indicating an increase in the
appropriate proportion of sentences being reinforced which
was significant to at least the .05 level.
Finally, the hypothesis was tested by considering
the population as a single unit. A z score was calculated
comparing the proportion of sentences using "she" as the
subject, a neutral verb as the predicate, and "him" as the
object of the sentence under the reinforcement condition
with that of the base rate.
z - 11.16 (£<.001)
On the basis of this finding, the null hypothesis was
rejected.
For additional information, a similar comparison was
made to statistically evaluate the difference in proportion
of the type of sentence in question between Trials 4 and 2
(Z2) and Trials 4 and 3 (zj).
z2 ■ 9.86 (jk.001)
Zi - 6.04 (£<.001)
116
Again, this would seem to indicate a constant trend of
conditioning taking place with each succeeding trial as
well as from beginning to final trial.
CHAPTER V
SUMMARY, CONCLUSIONS AND DISCUSSION
Summary
Background and the Problem
This chapter was devoted to a discussion of the dif
ferences in the free will and deterministic views of man
which are found today in the fields of counseling and psy
chotherapy. First, the broad delineation between the two
points of view was drawn, followed by a discussion of the
various theorists adhering to each. The purpose of this
discussion was to focus on the relevancy of the question
and upon the arguments of both sides of the issue as pre
sented by apparently successful, as well as intelligent men.
The need for further exploration and reconciliation of
these two points of view was noted. The writer’s own be
liefs were stated, and it was resolved to investigate more
fully the application of "shaping behavior" to the inter
view setting.
Statement of the Problem
To investigate the extent to which verbal behavior can
be influenced by operant conditioning by attempting to
shape the subjects' verbal behavior in a complex and speci
fic pattern of speech within a time limit of two to three
117
118
hours of interviewing or training.
Review of the Literature
A review of the literature was presented, describing
some of the various areas of verbal conditioning which had
been explored previously. Included in this chapter was a
brief discussion of the different research designs which
have been used in these investigations and a rationale for
the utilization of the design of the paper.
Procedure and Methodology
Subjects: The subjects used in this study were twenty
volunteer students from several graduate education courses
at San Diego State College.
Treatment: The subjects were presented with 320 white,
unlined 5 x 8 cards upon which were printed three lines of
words. Line one consisted of the nominative pronouns’fehe,"
"I," "he," "you." Line two consisted of two verbs in the
past tense, one of which was an aggressive verb, the other
a nonaggressive verb. Line three consisted of the objec
tive pronouns "him," "her,” "you," "me." All of the words
were placed upon the cards in a rotated order so that no
combination of words was likely to be "set" ipon. Clients
were then asked to look at each card and to form sentences
including a word from line one as the subject, line two as
the predicate, and line three as object of the sentence.
Sentences were not limited to these three words, however.
119
Four groups of eighty cards were presented. Trial I
was used to determine the base rate of usage, and no deli
berate reinforcement was administered as each subject
formed sentences during this trial. For Trial II, the sub
ject was reinforced with the sound "umm-hmm" each time he
used the pronoun "she" as the subject of the sentence. For
Trial III, reinforcement was given each time a sentence was
composed using "she" and the nonaggressive verb as the sub
ject and the predicate. Trial IV had reinforcement given
only for those sentences using "she" as subject, nonaggres
sive verb as predicate, and "him" as object. In this way,
the interviewer was attempting to "shape" the client's
verbal behavior by means of a Skinnerian paradigm.
Hypotheses
1. At the end of Trial II there will be no sig
nificant increase in the number of client's
sentences with the pronoun "she" as the sub
ject of the sentence when compared with the
sentences of Trial I.
2. At the end of Trial III there will be no sig
nificant increase in the number of client's
sentences with the pronoun "she" as the sub
ject and nonaggressive verb as the predicate
of the sentence when compared with the sen
tences of Trials I and II.
3. At the end of Trial IV there will be no sig
nificant increase in the number of client's
sentences with pronoun "she" as the subject,
nonaggressive verb as the predicate, and the
pronoun "him" as the object of the sentence
when compared with the sentences of Trials
I, II, AND III.
120
Analysis of the Data
The Test of Signs was used to evaluate directional
change of the proportions of sentence elements used for .
each trial. The Standard Score was used to test for sig
nificance of any change in the proportions of appropriate
sentence elements for each trial for both individual and
group data, though the hypotheses were to be evluated by
group data only.
For all trials, the Test of Signs indicated signifi
cance to the .001 level in directional change of the pro
portions of sentence elements used for each trial.
All three null hypotheses were rejected with the sta
tistical significance of .001 level being reached for each
and all of the hypotheses, indicating an influence has been
exerted in a systematic and predictable fashion.
Conclusions and Discussion
C. H. Patterson, in his discussion of verbal condition
ing and its relationship to therapy, raises two questions
that must at least be approached if there is to be a rele
vancy between the phenomenon of verbal conditioning and
therapy.
One is that of defining the response classes of
emitted verbalizations to be reinforced. Sub
jects differ among themselves and from the ex
perimenter or counselor in the limits of defi
nitions of these classes which are not natural
or neutral or universal. There is also the
121
problem of the lack or rarity of spontaneous
emission of the verbal behavior desired, or
even its approximation, for the counselor can
not reinforce expressions of a positive self
concept if they do not occur. (57:141).
In the writer's opinion these are really part of the
same question. If the experimenter can capitalize on the
process of successive approximation, he need not merely
wait for a verbal response that he wishes to reinforce; he
can actively develop it. And if this be true, the defining
of emitted response classes to be reinforced becomes a
pragmatic problem and not an intellectual one.
Prior to this study, there had been no research evi
dence to support the experimenter's assumption that he could
rely upon successive approximation to preclude the two
problems posed by Patterson. Though a variety of verbal be
haviors were apparently influenced by simple reinforcement
procedures, and a variety of variables were found to be
pertinent to reinforcing a particular, and rather gross
class of verbal behavior, no studies were reported that
would indicate that the problems of which Patterson spoke
could be accounted for. The writer believes that this
study indicates that verbal conditioning through successive
approximations is possible. It would seem that not only
are the simple, one unit class of responses modifiable;
complex patterns of verbal responses can be influenced and
developed very specifically by an experimenter using rein-
122
forcement procedures. It is therefore believed that the
two technical problems which Patterson raises about utiliz
ing verbal conditioning in the counseling interview have
been partially accounted for, at least in this setting. The
therapist-interviewer may not have to hope and wait for a
particular class of verbal responses that he wishes to re
inforce; the limits of definition of the response classes
to be reinforced need not be established and then become
static. They are themselves modifiable to the demands or
needs of the situation.
There is still the question of whether changing
speech patterns leads to any change in attitudes, feelings,
values, beliefs, or changes in overt behavior. There also
remains the primary question of whether the client in a
counseling relationship is helped because of choices he has
made due to the clarification of values, attitudes, feel
ings, beliefs, or perceptions, or whether he is helped
because of the skillful manipulation of external contin
gencies resulting in responses (be they affective, cogni
tive, or physical) that seem to be more satisfying and
functional to the individual living in his environment.
The truth may well lie, not somewhere between these
two positions, but partially in both. The writer must now
123
wonder if conditioning is not a larger part of the reality
of therapy than he had thought previously. The major dif
ficulty in trying to resolve this dilemma of choice or
conditioning seems to be the difficulty of eliminating the
alternative hypothesis while remaining in the counseling
process. Behaviorists have examined the transcripts of
Carl Rogers' therapeutic interviews and have been able to
show how what transpired can be explained within the Skin
nerian paradigm, and give that reason as the basis of suc
cess, whether it was Rogers' intent or not. Carl Rogers
has replied that in all the therapeutic studies which the
behaviorists have claimed as their own, the kind of rela
tionship which Rogers claims is the real cause for thera
peutic growth was present, and the relationship is the
real reason for the therapeutic success. In neither case
has the alternative hypothesis been accounted for. Yet, if
we take any portion out of the counseling process to ex
amine in isolation, for example, verbal behavior, then we
no longer have a counseling process, and relevancy can only
be implied or denied.
This study may well be challenged for not accounting
for two hypotheses that may contaminate the present find
ings. The first of these is concerned with the variable of
awareness and cooperation on the part of the subjects used
for this study. The writer fully believes that the subjects
124
were not aware of the reinforcing contingencies being pro
vided to the subject's verbal responses. He now believes
that greater effort should have been given to a more care
ful investigation of this possibility, along the lines sug
gested by Levin's study (47).
The second hypothesis in this category deals not only
with the possibility of the subjects being influenced by
the deliberate reinforcement which is being attempted, but
also by contingencies unknown to the experimenter, e.g. non
verbal cues, which might account for the outcome behavior.
It is therefore suggested that should this study be repli
cated, a control group from which the experimenter is ab
sent, should be used, in order to control this possibility.
In trying to unravel the Gordian knot of freedom and
determinism, the writer suggests that the counselor-scien-
tist be willing at this time to assume the validity of both
positions, and be able to utilize each of them in the ap
propriate circumstances, much as the physicists now do with
their two theories of light. At the same time, however, we
must continue to search for a greater understanding of
which aspects of the counseling process fall within each
position. In pursuing the study of verbal conditioning, it
would seem appropriate that the next step would be an at
tempt to shape the verbal behavior of a client in a more
counseling-like interview through several successive stages
125
of approximation in a predictable fashion, done in such a
way that alternative explanations of the results can be pre
cluded. This may lead us to a sophistication that will en
able us to attempt a more precise manipulation in a true
counseling interview, and still be able to preclude any al
ternative explanation for the process. Then perhaps we
would be able to tackle the question of whether shaping the
verbal behavior of a client in a counseling interview is in
any way correlated with changes in attitudes, values, per
ceptions or overt behavior. Even should we reach this
point of understanding, the question of choice of determin
ism in the counseling process may not be resolved. Perhaps
enough new data will be generated that the question can be
viewed from new perspectives. Presently we do have evi
dence that conditioning of verbal behavior is subject to
the process of successive approximation, and a start of the
procedures mentioned above can be accomplished.
B I B L I O G R A P H Y
126
BIBLIOGRAPHY
Adams, J., and Hoffman, B. "The Frequency of Self
Reference as a Function of Generalized Rein
forcement." Journal of Abnormal and Social
Psychology. 6ti, 1960, 384-389.
Allport, G. "Psychological Models for Guidance."
Harvard Educational Review, 32, 4, Fall
1962, 375-382.--------------
Anderson, D. "Personality Variables and Verbal Con
ditioning." Dissertation Abstracts, 19. 1811.
1959.
Babladelis, G. "Personality and Verbal Conditioning
Effects." Journal of Abnormal and Social
Psychology, 62, 1961, 41-43.
Bandura, A. "Psychotherapy as a Learning Process."
Psychological Bulletin, 58, 1961, 143-157.
Beck, C. Philosophical Foundations of Guidance.
New Jersey: Prentice-Hall, 1963.
Beech, H., and Adler, F. "Aspects of Verbal Condition
ing in Psychiatric Patients." Behavior Research
and Therapy, Dec. 1963, 273-282.
Binder, A., McDonald, D., Sjoholm, N. "Verbal Condi
tioning as a Function of Experimenter Character
istics." Journal of Abnormal and Social Psy-
chology. 60,T9?7, 309-314. --------------
Binder, A., and Salop, P. "Reinforcement and Person
ality Factors in Verbal Conditioning." American
Psychologist, 14, 1959, 387.
Bijou, S. "Operant Conditioning After Fixed Interval
Schedules with Young Children." Journal of Ex
perimental Analysis of Behavior, 1, 1958, 25-29.
Bourdon, Roger. "Effects of Reinforcement vs. Non
reinforcement on the Acquisition of Modeled Ver
bal Operant." Dissertation Abstracts, 29, 1418.
1968.
127
128
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
2 2.
23.
Buss, A., Breden, W., and Orgel, A. "Acquisition and
Extinction with Different Verbal Reinforcement
Combinations." Journal of Experimental Psychol-
ogy. 52, 1956, 28TB-T5T.
Buss, A., and Durkee, M. "Conditioning of Hostile
Verbalizations in a Situation Resembling a Clin
ical Interview." Journal of Consulting Psychol
ogy, 22, 1958, 415-418".
Campbell, J. M., "Verbal Conditioning as a Function of
Personality Characteristics of Experimenters and
Subjects." Dissertation Abstracts, 22, 3265,
I960. -------------------------
Campbell, D., and Campbell, Stanley J. Experimental
and Quasi-experimental Designs for Research.
Chicago: kand, McNally 8 Co., 1963.
Cieutat, V. "Sex Differences and Reinforcement in the
Conditioning and Extinction of Conversational
Behavior." Psychological Reports, 10, 1962,
467-474.
Cohen, E., Cohen, B. "Verbal Reinforcement and Schizo
phrenia." Journal of Abnormal and Social Psy
chology, 60, 1961, 443-446.
Cohen, B., Kalish, H., and Cohen, E. "Experimental
Manipulation of Verbal Behavior." Journal of
Experimental Psychology. 47, 1954, 106-li0.
Coons, W., and McEachern, B. "Verbal Conditioning Ac
ceptance of Self and Acceptance of Others."
Psychology Reports, 20, 1967, 715-722.
Cushing, M. C. "Affective Components of the Response
Class as a Factor in Verbal Conditioning."
Dissertation Abstracts, 17, 2313, 1957.
Dollard, J., and Miller, N. Personality and Psycho
therapy. New York: McGraw-Hill, lysu.
Drennen, W. T. "Transfer Effects of Verbal Condition
ing." Dissertation Abstracts, 23, 1793, 1962.
Drennen, W. T,, and Greenspoon, J. "Instruction and
Verbal Conditioning." Journal of Abnormal and
Social Psychology, 60, 1960, 443-447.
129
24. Ellis, A., and Harper, R. A Guide To Rational Living.
Englewood Cliffs, N. J." Prentice-Hall, 1961.
25. Ferguson, E. L., and Buss, A. H. "Acquisition, Ex
tinction and Counter Conditioning of Different
Verbal Reinforcement Combinations." Journal of
Experimental Psychology, 58, 1967, 94-95.
26. Frankl, V. Man's Search for Meaning. Boston: Beacon
Press, T9ZT. ------------
27. Frankl, V. Psychotherapy and Existentialism. New
York: Washington Square Press, 1967.
28. Franks, C., and Mantell, D. "Introversion and Extro
version and the Verbal Conditioning and General
ization of Meaning Responses to Nonsense Syl
lables in Normal and Alcoholic Subjects." Bri-
tish Journal of Social and Clinical Psychology,
5, 1966, 299-305.------- ---------------------
29. Geltner, H. "A Personality Determinant of Verbal
Conditioning." Journal of Personality, 34, 1966,
397-405.
30. Goodkim, R. "Changes in Word Production, Sentence
Production and Relevance in an Aphasic Through
Verbal Conditioning." Behavior Research and
Therapy, 7, 1969, 93-99
31. Grant, D., Hake, H., and Hornseth, T. "Acquisition
and Extinction of a Verbal Conditioned Response
with Differing Percentages of Reinforcement."
Journal of Experimental Psychology. 42, 1951,1-5.
32. Greenspoon, J. "The Reinforcing Effects of Two Spoken
Sounds on the Frequency of Two Responses."
American Journal of Psychology. 6 8, 1955,
409-416.
33. Guthrie, E. R. The Psychology of Learning. Rev. Ed.
New York: Harper, ld5z.
34. Hall, C. S. Primer of Freudian Psychology. New York:
Mentor Press, 1954.
35. Hildum, D. "Verbal Reinforcement and Interviewer
Bias." Journal of Abnormal and Social Psychol
ogy, 59, 1959, 162-166.--- ----------- ------
130
36. Immergluck, L. "Determinism-Freedom in Contemporary
Psychology." American Psychologist, 19, 4,
April 1964, 27trm :— ----------- —
37. Kahn, E. "Verbal Conditioning of Positive and Nega
tive Self Referent Statements Under Three Types
of Experimenter Interventions." Proceedings of
the 75th Annual Convention of the American Psy
chological Association, 1967.
38. Kanfer, L. "The Effect of Partial Reinforcement on
Acquisition and Extinction of a Class of Verbal
Responses." Journal of Experimental Psychology,
48, 1954, 424^37:--------- ---------------
39. Kanfer, F., and Karas, S. "Prior Experimenter-Subject
Interaction and Verbal Conditioning." Psycholo
gical Reports, 1959, 345-357.
40. Kanfer, F., and McBrearty, J. "Verbal Conditioning:
Discrimination and Awareness." Journal of
Psychology, 52, 1961, 115-124.
41. Kerlinger, F. Foundations of Behavioral Research.
New York: Holt, Rinehart and Winston, 1965.
42. Kilberg, J. "Differential Effects of Nonverbal and
Verbal Rewards in Modification of Verbal Beha
vior of Schizophrenics and Normal Subjects."
Dissertation Abstracts. 23, 1074, 1962.
43. Kinzie, W. "Generlization of Verbally Conditioned
Responses." Dissertation Abstracts, 27, 2137B,
1966.
44. Klein, S. "Conditioning and Extinction of Operant
Verbal Behavior in Neuropsychiatric Hospital
Patients." Dissertation Abstracts, 14, 2127-
2128, 1954.
45. Krasner, L. "Studies in the Conditioning of Verbal
Behavior." Psychological Bulletin, 55, 1958,
148-170.
46. Leventhal, M. "The Effects of Diagnostic Category and
Reinforcement on Learning Without Awareness."
Journal of Abnormal Psychology, 59, 1959,162-166.
131
47. Levin, S. M. "The Effects of Awareness on Verbal Con
ditioning." Journal of Experimental Psychology,
61, 1961, 67-TT.
48. Mathewson, R. Guidance, Policy, and Practice. Rev.Ed.
New York: Harper § Row, 1953.
49. Marder, M. "The Effects of Experimenter Attractive
ness and Negative Reinforcement on Verbal Con
ditioning." Dissertation Abstracts, 22, 1243,
1961.
50. Matarazzo, J. D., Saslow, G., and Pareis, E. N.
"Verbal Conditioning of Two Response Classes."
Journal of Abnormal and Social Psychology, 61,
1960, 190-206.
51. May, R. Existence. New York: Simon 8 Schuster, 1958.
52. Michael, J., Mererson, B. "A Behavioral Approach to
Counseling and Guidance." Harvard Educational
Review. 42, 4, Fall 1962, 282-402.
53. Moos, R. "The Retention and Generalization of Operant
Conditioning Effects in an Interview Setting."
Journal of Abnormal and Social Psychology, 6 6 ,
T9F3” , " 52-58.--------211---------- L-----
ft
54. McNair, D. "Reinforcement of Verbal Behavior.'
Journal of Experimental Psychology. 53, 1957,
40-46.
55. Nuthmann, A. "Conditioning of a Response Class on a
Personality Test." Journal of Abnormal and
Social Psychology, 54, 1957, 19-23.
56. Oaker, W., and Droge, J. "Operant Conditioning of
Responses to the Social Introversion Scale Items
on the MMPI." Psychology Reports, 6 , 1960,
223-225.
57. Patterson, C. H. Theories of Counseling amd Psycho
therapy. New York: Harper $ How, 1966•
58. Pepinsky, H. B., and Pepinsky, P. N. Counseling:
Theory and Practice. New York: Ronald Press.
I9T4. --------
132
59. Phelan, J. G., Tang, T., and Heckmat, H. "Some Ef
fects of Various Schedules of Verbal Reinforce
ment of Self Referent Responses." Journal of
Psychology, 67, 1967, 17-24.
60. Raimey, V. C. "Self Reference in the Counseling In
terview." Journal of Consulting Psychology, 12,
1948, 153-1XT.-------------- -------------
61. Rickard, H. C. "Verbal Manipulation in the Psychother
apeutic Relationship." Journal of Clinical
Psychology, 16, 1960, 364-367.
62. Rogers, C., and Dyamond, R. Psychotherapy and Person
ality Change. Chicago: university of Chicago
Press, 1954.
63. Rogers, C. "Two Divergent Trends." Existential
Psychology. New York: 1961, 85-9$.
64. Rogers, C. On Becoming a Person. Boston: Houghton
Mifflin-CoT7 T95I----------
65. Rogers, C. Client Centered Therapy. Boston: Houghton
Mifflin, do., 1951.
6 6 . Rogers, C. "A Note on the Nature of Man." Journal
of Counseling Psychology, 4, 3, 1957, 199-203.
67. Rogers, J. "Operant Conditioning in a Quasi-Thera-
peutic Setting." Journal of Abnormal and Social
Psychology, 60, 1960, 247-252.
6 8 . Salzinger, K., and Pisone, S. "Reinforcement of Af
fect Responses of Schizophrenics During the
Clinical Interview." Journal of Abnormal and
Social Psychology, 57, 1958, 84-90.
69. Salzinger, K., Pisone, S. "Reinforcement of Verbal
Affect Responses of Normal Subjects During an
Interview." Journal of Abnormal and Social
Psychology, 60, 1960, 127-130.
70. Sapolsky, S. "The Effect of Interpersonal Relation
ships upon Verbal Conditioning." Journal of
Abnormal and Social Psychology, 60, 1960,241-246.
71. Sarason, B. "The Effects of Verbally Conditioned Re
sponse Classes on Post-Conditioning Tasks."
Dissertation Abstracts, 17, 6799, 1957.
133
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
Sarason, I. G. "Interrelationships Among Individual
Difference Variables Behavior on a Psychotherapy
and Verbal Conditioning." Journal of Abnormal
and Social Psychology. 56, 1958, 339-344.
Scott, W. "Attitude Change Through Reward of Verbal
Behavior." Journal of Abnormal and Social
Psychology, 55, 1957, 72-75.
Skinner, B. F. Science and Human Behavior. New York:
MacMi11an, 1953.
Skinner, B. F. "Reinforcement Today." American
Psychologist, March, 1958, 94-99.
Sidowski, J. B. "Influence of Awareness of Reinforce
ment on Verbal Conditioning." Journal of Experi
mental Psychology, 48, 1954, 355-360.
Slavin, D. "Response Transfer of Conditioned Affec
tive Responses as a Function of an Experimental
Analogue of Rotational Psychotherapy." Disser
tation Abstracts, 28, 2632B, 1967.
Snygg, D., and Combs, A. Individual Behavior. New
York: Harper 8 Row, 1959.
Strong, S. R. "Verbal Conditioning and Counseling Re
search." Personnel and Guidance Journal, 42,
1964, 660-55T
Taffel, C. "Anxiety and Conditioning of Verbal Beha
vior." Journal of Abnormal and Social Psy
chology, 51, 1955, 496-501.
Thayer, F. "Relationship of Generalization to Aware
ness in Verbal Conditioning." Dissertation Ab
stracts , 27(6b), 2173, 1966.
Thorndike, E. L., quoted in Hill, W. T. Learning.
San Francisco: Chandler Publishing Co., 1§63.
Tillich. P. The Courage to Be. New Haven: Yale
University Press7 1952.
Truax, C. B. "Some Implications of Behavior Therapy
for Psychotherapy." Journal of Counseling
Psychology. 13, Summer 1966, 160-170.
134
85. Turner, J. "Effects of Manifest Anxiety on Verbal
Conditioning in a Quasi-Therapeutic Setting."
Dissertation Abstracts, 28, 2635B, 1967.
86. Ulman, L., Krasner, L., and Collins, B. "Modification
of Behavior Through Verbal Conditioning."
Journal of Abnormal and Social Psychology, 62,
1561, 128-132.------ ----------------
87. Verplank, W. S. "The Control of the Content of Con
versation: Reinforcement of Statements of
Opinions." Journal of Abnormal and Social
Psychology, 51, 1955, 668-676.
88. van Kaam, A. The Art of Existential Counseling.
Phi1adelphia: Dimension Books, 1966.
89. Wagner, M. "Reinforcement of Verbal Productivity in
Group Therapy." Psychological Reports, 19,
1966, 1217-1218.
90. Williams, R. "Verbal Conditioning and Psychotherapy."
American Psychologist, 14, 1959, 388.
91. Wolpe, J., and Lazarus, A. Behavior Therapy Tech
niques . New York: Permagon Press, 1966.
92. Wolpe, J. Psychotherapy by Reciprocal Inhibition.
Stanford: Stanford university Press, 19S&.
93. Wood, G. L. "The Effects of Verbal Reinforcement on
Depth of Intrapersonal Exploration." Disserta
tion Abstracts, 29, 3433A, 1969.
94. Zax, L., and Klein, J. "A Measurement of Personality
and Behavioral Changes Following Psychotherapy."
Psychology Bulletin, 5-7, 1960, 435-448.
APPENDIX
135
136
LIST OF VERBS USED IN THE STUDY
Trial 1 Trial 1 (continued)
encroached elevated tried thrashed
elected electrocuted taunted touched
dueled drank told whacked
educated elbowed vexed wondered
examined exasperated won wrenched
feuded entered tore tamed
escaped encouraaged walked violated
endorsed executed tussled worked
enforced emerged perceived rejected
doubled dominated resented possessed
dragged doubted purchased pushed
duplicated distorted repulsed recalled
distressed drafted practiced reprimanded
drew disrupted persuaded rebelled
distrusted drained raged photographed
eased dragged prepared pressed
divorced dressed prostrated produced
disturbed dwelled predicted repelled
drove dreaded framed gloated
tolerated throttled gnawed gained
wrote terrorized gathered gored
troubled transferred gosspied gazed
transcribed protested gestured grabbed
tyrannized placated forced favored
planned punched felt foreclosed
scorned reduced explained fined
negotiated obstructed flailed expressed
documented drubbed explored forbid
disputed doled permitted plundered
talked whipped perverted observed
threatened varied fouled found
tasted wreaked took wrecked
vanquished thought finished frightened
tempted tortured frustrated fixated
terrified taught ate ej ected
thanked warred flattered gagged
trounced whimpered followed garroted
withdrwe villified glared forecasted
tripped understood obtained pursued
provoked praised pillaged performed
twisted worshipped
137
Trial 2
seized slept
resigned shot
pledged raped
launched lusted
nagged measured
leaked molested
mugged laughed
knelt meddled
maligned labled
repeated slugged
ridiculed rose
served spat
squashed remembered
killed lighted
located knocked
issued needled
migrated lied
mangled modeled
left lured
lashed kept
jumped libeled
mashed j oined
mortified manifested
hitched hunted
infringed improved
humbled hooted
howled honored
imputed included
indicated impugned
invented kicked
mocked loved
neglected maintained
investigated mobbed
heaved idealized
humored interrogated
hurt installed
interfered inspired
implored insulted
hit inquired
heralded implicated
prohibited qualified
revenged smelled
spinned revolted
ripped revised
satisfied shoved
shook reconciled
heeded intruded
imitated hazed
impaired hauled
harvested irritated
impeached interpreted
identified haunted
j erked handled
introduced hated
hung guided
grew haggled
insisted inherited
impressed imprisoned
helped heckled
fired existed
exhibited feared
fought excited
entertained flaunted
falsified established
enj oyed exploded
exhorted enlisted
enabled enforced
expelled emerged
erected enslaved
sanctioned scourged
screamed recovered
referred slandered
smashed reported
respected spurned
menaced invoked
ravished postulated
obligated needed
peddled over-powered
offended occurred
hacked groped
138
Trial 3
contaminated
condemned
combined
clawed
circled
charged
disobeyed
disgusted
distinguished
considered
confined
conflicted
cooperated
corrupted
chanted
chased
changed
cheered
conferred
descended
cursed
dazzled
crucified
deluded
deported
derived
damned
desired
destroyed
did
disposed
demanded
demonstrated
detested
designed
deserved
denounced
growled
greeted
arrested
appealed
annihilated
apprehended
antagonized
conformed
chose
commanded
closed
choked
consoled
distributed
distilled
disfigured
convicted
composed
comforted
controverted
conducted
challenged
certified
clashed
confronted
conspired
debated
dashed
cut
danced
created
depended
desecrated
dealt
deposed
determined
disavowed
dialed
departed
demeaned
depleted
denied
deprecated
detached
gratified
guarded
applied
assailed
appointed
analyzed
amplified
appeared
appraised
argued
appeased
alarmed
amputated
agitated
allocated
aggravated
afflicted
alleviated
affirmed
impeached
accused
accompanied
persecuted
acclaimed
destroyed
abducted
continued
absorbed
abandoned
accused
bent
acquired
adhered
administered
ordered
adopted
amended
abashed
adorned
accelerated
pushed
chased
armed
aroused
alternated
ambushed
amazed
amused
appreciated
assaulted
answered
alerted
affronted
admonished
affected
advised
oppressed
accomplished
abused
accepted
remained
renounced
abased
abstained
acknowledged
acquitted
beat
interfered
mashed
admired
resisted
agonized
abided
nailed
shook
advanced
adored
139
Trial 4
pounced paid
pardoned plotted
operated poisoned
preferred raided
purged read
presented ravaged
pounded preceded
retorted reclined
recognized slashed
objected partitioned
organized persisted
obtruded offered
passed oppressed
opposed owned
numbered ordered
obliterated occupied
peeled pestered
perjured pacified
oscillated poked
blessed bilked
blended blamed
bailed blasted
beat bordered
hit accumulated
detested developed
boiled bruised
bit benefited
dispensed disdained
blinded bleached
bloomed banished
batted baked
bored banged
barred balanced
beckoned beguiled
battered believed
began battled
betrayed bent
braced bayoneted
blocked bleached
brawled bought
bombed built
calculated censored
chained compared
browsed bothered
bridged bossed
bullied calmed
blocked banded
balked behaved
digested disapproved
defended decimated
defeated declared
aspired asserted
assassinated approved
assured attacked
ceased butted
captured caused
burned canvassed
caught broke
called bumped
deferred decapitated
depicted deluded
defied denoted
delivered degraded
demolished delighted
deliberated deformed
defrauded delegated
defiled deified
deflected defamed
defaced deflated
disliked dissolved
dined disconcerted
dishonored directed
discouraged diagnosed
differed disciplined
discussed discomfited
disgraced discovered
disagreed deviated
dared decayed
deceived deduced
rammed professed

Asset Metadata

Creator Carnevale, James Patrick (author)

Core Title Shaping Of Verbal Behavior In An Interview Setting

Contributor Digitized by ProQuest (provenance)

Degree Doctor of Philosophy

Degree Program Education

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag education, guidance and counseling,OAI-PMH Harvest

Language English

Advisor Carnes, Earl F. (committee chair), Ofman, William V. (committee member), Spear, Howard S. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-468970

Unique identifier UC11362476

Identifier 7116402.pdf (filename),usctheses-c18-468970 (legacy record id)

Legacy Identifier 7116402

Dmrecord 468970

Document Type Dissertation

Rights Carnevale, James Patrick

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA