VERBAL REINFORCEMENT OF CLIENT
DEPENDENCY IN THE INITIAL
STAGE OF PSYCHOTHERAPY

Thesis Ior Hm Degree GI DI). D.
MICHIGAN STATE UNIVERSITY

Philip F. Caracena
1963

THESIS

This is to certify that the

thesis entitled

Verbal Reinforcement of Client Dependency
in the Initial Stage of Psychotherapy

presented by

Philip F. Caracena-

has been accepted towards fulfillment
of the requirements for

Pho o i- (t
D degree in Psye..olouy

 

Major proTessor

Date 27 August 1965

0-169

LIBRARY

Michigan State
University

 

ABSTRACT

VERBAL REINFORCEMENT OF CLIENT DEPENDENCY
IN THE INITIAL STAGE OF PSYCHOTHERAPY

by Philip F. Caracena

Dependency of clients upon psychotherapists is a
frequent and important occurrence in the early hours of
psychotherapy. This study examines the phenomenon as an
effect of therapists' approach and avoidance responses to
clients' verbalizations of dependent content. The hypoth-
eses state that when therapists approach dependency, cli-
ents continue the topic but when therapists avoid, then
clients discontinue. Aside from the elicitatiOn value of
approach, these responses are assumed to reinforce discus—
sion of the topic. Learning occurs if (1) the probability
of a client's continuing the topic increases as the inter-
view progresses and (2) the probability of a client's in-
itiating dependency statements increases over time. Con-
versely, therapists' avoidance of dependency reinforces
the tendency for clients to discuss non-dependent topics.
The study also examines the longer-term effects of reinforce-
ment upon staying in therapy.

Further hypotheses state (1) that therapists learn
to approach dependency as an initial technique of therapy
and (2) that therapists learn to reinforce selectively var-

ious types of dependent statements.

Philip F. Caracena

A content analysis of 72 recordings of early psy-
chotherapy interviews with clients at a university counsel-
ing center shows that approach to dependency elicits further
discussion of dependency and avoidance elicits discontinu-
ance. For clients whose dependent statements are predom-
inantly approached, the probability that approach elicits
further discussion and that clients initiate dependency
statements does not increase significantly as the hour pro-
gresses, although there are trends in the eXpected direc-
tions. Due to the high frequency of approach responses
made to all clients, effects of avoidance are inadequately
measured. Relatively infrequent approach, however, does
not reinforce the competing tendency to discuss non-depend-
ent topics. The data suggest individual differences in
inhibition of each habit independent of the therapists'
responses measured here.

Terminating and remaining in therapy are unrelated
to the percentage of approach and avoidance therapists gave
when clients discussed dependency. This fails to support
previous findings. However, more "likable" clients are very
likely to remain in therapy.

Staff level therapists more frequently approach
than avoid dependency in contrast to practicum and interne
level therapists. Approach to subtypes of dependency is
not substantially affected by experience level. ‘

Limitations inherent in the study are discussed and
implications are noted for further research.

Approved—MW Date 117 46174417163

Committee Chairman'

VERBAL REINFORCEMENT OF CLIENT DEPENDENCY
IN THE INITIAL STAGE OF PSYCHOTHERAPY

By H i, I:
fie

v
Philip FFICaracena

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Psychology

1963

 

Tga?o44
4/24/44

Dedication

To Virginia,

Chris and Kurt

ii

ACKNOWLEDGMENTS

The writer is deeply grateful to his chairman,
Dr. Charles Hanley, for his proficient direction of the
study and his generous personal help and encouragement.
The writer wishes to express his appreciation to Dr. M. Ray
Denny especially for his theoretical ideas which are evi-
dent in the study; to Dr. Barry Grater for his valuable
suggestions and support; and to Dr. Bill L. Kell for his
continuous belief in the writer.

Special appreciation is extended to Dr. C. L. Winder
who was unable to continue direction of the study. His
ideas originated and sparked this study.

Acknowledgments are made to David Kopplin for his
hard work as a judge and his help throughout the study.
The writer wishes to thank the therapists and clients of
the Michigan State University Counseling Center for their
cooperation in the collection of data.

***

iii

CHAPTER

TABLE OF CONTENTS

I. INTRODUCTION . . . . . .

A.
B.
C.
D.
E.
F.
G.

Statement of the Problem .
Dependency in Psychotherapy.
Learning and Psychotherapy .
Verbal Conditioning Studies.

Approach—Avoidance as Contingen

Selective Therapist Responses.

Statement of Hypotheses.

II. METHOD . . . . . . . .

A.
B.
C.
D.
E.
F.

III.

A.
B.
C.
D.
E.
F.

Clients. . . . .
Therapists . . . .
Stage of Therapy .

Duration of Treatment:
Scoring of Interviews.
Scoring Reliability.

RESULTS 0 0 O O O O O 0 0

Hypothesis I .
Hypothesis II.
Hypothesis III
Hypothesis IV.
Hypothesis V .
Hypothesis VI.

0 O O 0 O O
0

IV. DISCUSSION . . . . . . .

Approach-Avoidance as Contingent
Hypotheses I, V, and VI.
Approach-Avoidance as Consistent
Hypotheses II, III, and IV .

v O SUWRY O O O O O O O C 0

REFERENCES .
APPENDICES .

iv

0

t

O O O 0

00.000 0.0)....

ooI—looooo

O O 0 O O O

Stimuli

coco-co

Elicitors

coo-co ooHaOooo

O O 0 O O

45
50
56
58
62

LIST OF TABLES
TABLE Page
1. Descriptive Summary of the Sample . . . . . . . 22

2. Interjudge Reliability for Percentage Continu-
ance of Dependency Following Approach and
Avoidance . . . . . . . . . . . . . . . . . . . 29

3. Interjudge Reliability of Scores in the Two
Segments of the Interviews. . . . . . . . . . . 33

4. Change Over Segments in Percentage Continuance
of Dependency for High and Low Approach Groups. 35

5. Change Over Segments in Percentage Continuance
of Dependency Following Approach (D Ap 0%) for
High and Low Approach Groups. . . . . . . . . . 35

6. Change Over Segments in Percentage Continuance
of Dependency Following Avoidance (D Av 0%) for
High and Low Approach Groups. . . . . . . . . 36

7. Correlations of Percentage Continuance of De-
pendency Following Approach (D Ap C%) with
Therapist Responses . . . . . . . . . . . . . . 38

8. Correlations of Percentage Approach to Depend-
ency (D Ap%) with Continuance Following Ap-
proaCh (D AP 0%). O O O O O O O O O O O O C O O 39

9. Interjudge Reliability of Percentage CID in the
Two Segments of the Interviews. . . . . . . . . 4O

10. Change Over Segments in Percentage CID for High
and Low Approach Groups . . . . . . . . . . . . 41

11. Change Over Segments in High and Low Segment
One Percentage CID for High and Low Approach
Groups 0 O O O O O O 0 O O O O O O O O O O O 0 41

I2. Differences of Percentage Approach to Depend-
ency between Experience Levels. . . . . . . . . 42

13. Interjudge Reliability for Percentage Approach
to Subcategories of Dependency. . . . . . . . . 43

14. Client Likability and Remaining in Therapy. . . 52

LIST OF APPENDICES

APPENDIX

A.
B.
C.
D.

Data Sheet for Therapists

Follow-Up Questionnaire .

Client Likability Scale .

Scoring Manual.

vi

Page

62
63
64
65

I. INTRODUCTION

A. Statement of the Problem

This study investigates effects of certain thera-
pist variables upon behavioral change in clients first en-
tering psychotherapy. The client change considered is ex-
pression of dependency as shown in two manifestations: (l)
verbalization of dependent content, and (2) remaining in
psychotherapy. The therapist variables include: (1) verbal
approach and avoidance of client statements and (2) exper-
ience level. Our method of investigation is a content an-
alysis of verbal interaction during psychotherapy.

Hypotheses are derived from learning theory; change
in client dependency is presumed to depend upon the thera-
pist's verbalized approach or avoidance of the client's
dependency expressions. Thus, therapists' approach state-
ments to client dependency are hypothesized as consistent
elicitors of dependency. As such, they increase the prob-
ability of such further behavior from clients. Therapists'
avoidant statements to dependency are considered inhibitors
of this behavior.

An essential part of therapeutic skill is to retain
an accepted client in treatment as long as necessary regard-
less of his initial state of mental health. One technique
may be to approach, and thereby reinforce, dependency when

it is expressed in the client's verbal repertoire. Should

approach be effective, more experienced therapists may have
learned to utilize this technique more than inexperienced
therapists.

If the overall therapeutic situation fits the operant
conditioning paradigm, i.e., one in which the therapist
selectively reinforces the client's behavior according to
its proximity to the "correct" form, it follows that thera-
pists selectively reinforce different types of dependency
statements during the course of psychotherapy. Which types
the therapist chooses to reinforce (approach) and which he
chooses to inhibit (avoid) may vary according to the thera—
pist's own experience with the effect of differentially

reinforcing subclasses of dependency.

B. Dependency in Psychotherapy

The dynamics of dependency were described by Freud
in 1905. He traced dependency motivation to feeding frus-
trations in the oral psychosexual stage of life. Orthodox
psychoanalysts continue to theorize that adult dependency
stems from the longing for relief from distress that the
infant directs toward the parents. Fenichel (1945) describes
a continuing dependence-independence conflict, lasting through-
out the child's attempt to master his environment. Psycho—
analytic treatment regenerates the original dependency con—
flict experienced during childhood (Fromm-Reichmann, 1950;
Alexander & Ross, 1952).

Dependency is defined rather consistently in the

literature. Traditionally, dependency is a learned motive

to be taken care of, helped and nurtured (e.g., Murray,
1954, 1956). Dollard & Auld (1959) distinguish a dependent
relationship from a symbiotic one by stressing its unilat-
eral character: one person relies upon another for support,
maintenance or help. Wolberg (1954) describes a patient's
experience of dependency as a feeling of helplessness, of
wanting someone to relieve distress, give support, guidance
and direction.

The importance of client dependency in the beginning
phases of psychotherapy has been discussed by a number of
writers. Fenichel (1945) regards dependency as a component
of the more general phenomenon of transference. In describ-
ing the role of dependency in psychoanalysis, he observes
that the patient enters the treatment situation with a host
of repressed and partially repressed needs, among which are
commonly found oral—dependent strivings. One of the first
duties of the analyst is the very practical one of creating
and sustaining a "workable" relationship in which the pa-
tient abandons defensive maneuvering in order to permit
emergence of unconscious needs. The means by which the
analyst fosters such a change is the interpretation of pos-
itive and negative transference toward himself. The patient
ascribes to his therapist somewhat omnipotent powers to
help, heal and protect. By interpreting the transference
during treatment, the analyst makes the emergent dependent
strivings explicit.

Alexander & Ross (1952) emphasize the necessity of

developing a dependent relationship to enable the patient
to relive and once more face unresolved conflicts with the
parents; to combat previously formed infantile reactions
by reproducing them in the transference relationship. A
likely consequence of failure to deal with dependency is
the generation of hostility and resistance detrimental to
further therapeutic contact (Wolberg, 1954).

Thus psychoanalysis regards dependent wishes, at—
titudes and behavior as desirable and necessary in the be-
ginning of psychotherapy, and the role of the therapist
is one of allowing dependency to emerge.

The goal of more actively eliciting client depend-
ency early in treatment appears in other schools of therapy.
Fenichel and Alexander & Ross discuss this phenomenon in
non-psychoanalytic therapies, criticizing them for continu—
ing to capitalize upon dependency far beyond the goal of
uncovering. Successes are labeled "transference cures,"
i.e., cures maintained only through a continuing dependent
relationship with the therapist. Despite the further uti-
lization or management of dependency, therapists are com-
monly concerned with fostering at least its discussion in
early therapy interviews.

Rogers (1942) noted the eventual disappearance of
transference in his client-centered treatment. The dissi-
pation or lack of further development of dependency on the
therapist is ascribed to its being dealt with in the first

place, e.g., by means of reflection, clarification of

feelings, etc. Perhaps the first two of Rogers' (1958)
seven stages of therapy comprise the portion of therapy
in which dependency first emerges. A general "loosening"
of feelings and expression characterizes these stages.

Dollard & Miller (1950) note the desirability of
inducing the patient to express more fully existing depend-
ency needs in the early phase of treatment. Shoben (1949),
in his interpretation of therapy as a learning experience,
alludes to dependency, describing a part of the first step
in counseling as a matter of making therapist responses
which ” . . . most effectively further the bringing into
communicability of repressed impulses . . . " without de-
stroying the relationship.

In most therapies, then, the emergence of depend-
ency motive (gig dependent behavior) appears to occur, whether
by active, passive or unclear means. Certainly the continu-
ation of the client—therapist relationship itself is evidence
of utilization of client dependency (Dollard & Miller, 1950).
Empirical identification of significant client, therapist
or situational variables associated with the intentional
or incidental development of client dependency has been
largely neglected in light of its apparent universality
to the therapeutic process and its importance to some the-
ories of psychotherapy. Skinner (1953) theorizes, "If the
patient is to return for further counsel, the psychother-
apist must make sure that the behavior of coming to him is

in some measure reinforced." (p. 74) Rogers (1951) tentatively

hypothesizes several general mechanisms (e.g., therapists'
evaluative remarks) by which a dependent relationship may
develop. But research that would allow evaluation of these
hypotheses is scant. More speculative attention has been
paid to techniques for the management of dependency in later
stages than to the circumstances surrounding its initial
appearance. Following Winder gt_al (1962), this study ex-
amines the first stage in which the phenomenon occurs in

therapy.

C. Learning and Psychotherapy
The behaviorists Dollard & Miller (1950) and Skinner
(1953) view the therapeutic process as one in which the
therapist teaches and the client learns and unlearns. The
therapist is assumed to have at his command a number of
techniques for selectively shaping the client's behavior.
Ferster (1958) presents a similar viewpoint:
It is possible that many of the symptoms which bring
the patient to therapy are largely a by-product of
inadequate positively reinforced repertoires;'that
the disposition to engage in the psychotic, neur-
otic, and pathological behaviors may seem strong
when compared to weak existing repertoires but would
disappear as soon as alternative effective ways of
dealing with some accessible environment are gen-
erated. (p. 117)
The therapist's role, then, is to reinforce selectively.
In all "talking" therapies, what is dealt with dur-
ing the therapeutic hour is verbal behavior. Implicit in
our use of learning theory is the assumption that verbal

response is accounted for by the same principles as are

non-verbal responses: that verbal expression, while having
symbolic referents, is a behavioral response and, as such,
is amenable to experimental manipulation. Skinner (1957)
presents this viewpoint despite what he refers to as sym-
bolic or "second level" characteristics. Whether verbal
conditioning occurs because of cognitive mediation, as
Spielberger & Levin (1962) contend, or through an automatic
reinforcement of response is not put to test here. However,
the dependency motive of clients in psychotherapy may gen-
erate an expectancy or learning set somewhat similar to
that induced in straightforward learning experiments.
Whether the client can generalize his verbal learning to
other behavior is, of course, a relevant issue to be dis-

cussed later.

D. Verbal Conditioning Studies

Comprehensive reviews by Krasner (1958), Salzinger
(1959) and Greenspoon (1962) indicate the deluge of recent
efforts to apply conditioning procedures to language, mostly
in experimental or quasi-therapeutic situations. These
reviews point up a number of unsettled issues.

Our concern focusses upon whether verbal learning
occurs in psychotherapy. Learning is taken to mean either
(1) the strengthening of an existing (measurable) S-R bond
over time or (2) the establishment of a new S-R connection.
The present study attempts to demonstrate learning of the

first type during the therapeutic hour, that is, an increase

in the tendency of a stimulus to elicit a response. Accord-
ing to elicitation theory (Denny & Adelman, 1955). the in-
crement depends upon the consistent elicitation of the re-
sponse in a given stimulus complex. Learning occurs, then,
when a stimulus consistently evokes a response over a period
of time. A consistent elicitor lg a reinforcing stimulus.
In the context of this study, verbal approach by the ther-
apist is a consistent elicitor of dependent statements and
avoidance, a consistent elicitor of non-dependent statements.

In the operant conditioning paradigm, a contingent
stimulus is one which is introduced when the subject makes
a desired response. The contingent stimuli may be but are
not necessarily consistent elicitors.

Greenspoon's (1950, 1955) original work in the ver-
bal conditioning field well illustrates the conditioning
paradigm and clearly demonstrates learning. The experi-
mental procedure is (l) to measure the operant level of
the to-be—conditioned response (the strength or probability
of some existing S-R bond), (2) to introduce the contingent
stimulus, (3) to omit the contingent stimulus after repeated
presentation, and (4) to measure again the operant level
of the response. Positive change represents conditioning
or learning: an increased probability of the response's
occurrence in the absence of the eliciting stimulus. The
stimulus complex which surrounded the elicitor has now ac-
quired increased elicitation value. The contingent stimu—

lus, through consistent elicitation of the response has been

proven to be a reinforcer.

Salzinger & Pisoni (1958) report that the contingent
stimulus, "Um-hmm," is an effective "reinforcer" of affect
responses in schizophrenics but, also, that its omission
leads to a decrement in frequency of affect responses to
the level of the control group which received no contingent
stimulus. Two interpretations are possible: (1) that re—
inforcement occurred but was extremely short-lived, or (2)
that the stimulus "Um-hmm" had no reinforcement effect but
simply evoked affect responses when it was presented. It
is not known whether subjects interpreted "Um-hmm" as en-
couragement to continue responding as they had been. Con-
troversial findings with this particular stimulus lend sup-
port to "Um-hmm" having a number of possible meanings for
different subjects in various situations (Hildrum & Brown,
1956). The factor which may determine whether or not "Um-hmm"
reinforces may be the set of the subjects. Greenspoon's
use of "Um-hmm" was more obviously in a context of a learn-
ing situation for the subjects. As such, "Um—hmm" may be
reinforcing.

A host of studies have used contingent stimuli whose
meanings are less ambiguous and less controversial than
"Um-hmm." Both therapists' approach and avoidance responses
have been found to be elicitors of client responses (Bandura,
Lipsher, & Miller, 1960; Winder, Ahmad, Bandura, & Ban,
1962). The demonstration that they reinforce as contingent

stimuli has not been made. For instance, it is quite likely

10

that a response similar to the previous response will be
given following the stimulus: "Tell me more about that."
Once this probability is established in the interview, note
can be made of its becoming more likely when the client
receives various proportions of approach and avoidance.
That such elicitors do, in fact, reinforce can be determined
in this way. Often, studies have noted only that certain
contingent stimuli tend to elicit consistently without in-
vestigating whether learning occurs which can be attributed
to the elicitor. In some studies, if the to-be-conditioned
response is measured over time, the frequency of presenta-
tion of the elicitor is either statistically uncontrolled
or it is highly correlated with the response's occurrence.
Thus, it is not known whether the frequency or the proba-
bility of the response has increased.

Among the few studies of actual psychotherapy from
a learning position, an investigation by Murray (1956) pro-
vides an example of the kind of ambiguity mentioned. An-
alyzing Rogers' case of Herbert Bryan, Murray classified
therapist statements as mildly approving or disapproving
and client statements according to their content. Those
categories of client statements which were immediately fol-
lowed by mild approval increased in frequency over the course
of therapy while disapproved categories decreased in fre—
quency. Murray reports that selective verbal reinforcement
produced the changes:

The number of approvals the therapist gave on a
given hour was highly correlated with the percentage

ll

of the rewarded categories (r = .97, p.<<.01).
This was also true of disapproved categories
(r = .97, P.<<.05). Therefore, the behavior of
the therapist was consistent with the appearance
of the categories. (p. 11)

and further:

The approvals grew more frequent and stronger
as therapy progressed.

Since the elicitor was presented with increasing
frequency, the concomitant increase of the response cate—
gory is not evidence of learning but is evidence that ap-
proval does elicit. Needed here to show learning is a

demonstration of increased probability of the S-R bond

 

over time. This requires some kind of control over the
frequency of presentation of the elicitor.

Greenspoon (1962) reviews a number of studies in
which the omission of contingent stimuli immediately de-
creases the frequency of the desired response. Other
studies show short—lived effects upon the chosen response
from increasing and decreasing the frequency of the "rein-
forcing" stimuli. Few studies report change in Operant
level over time.

Waskow (1962), in a quasi-therapeutic situation,
responded by selectively reflecting feeling (F), content
(C) and a mixture of the two (FC). She found significant
differences in the mean percentage of each type of material
according to the type receiving reflection. The only con—
clusion which can be drawn is that an increase in frequency

of reflection for each category, in effect, evoked those

12

categories more often. An analysis of mean percentages
across time (through the first to the fourth interview and
through sixths of the first interview) shows only one con-
sistent increase in percentage of responses, this occurring
for the C response category in the first interview.
Examination of the graphs plotting sixths of other
interviews suggests that the percentage of C re-
sponses for the three groups remains fairly stable
within each of the interviews following the first
one. (p. 15)
No indication of the internal consistency of thera-
pist responses is given. Reflection of content may have
increased in the first interview and thereafter remained
constant. If this were the case, learning was not measured.
If, however, the number of therapist reflections did not
increase along with the rise in C, reinforcement may have
occurred and reached an asymptote in the first interview.
Greenspoon (1962) concludes his discussion of the
limitations and importance of verbal conditioning studies
by stating:
The research on verbal conditioning in both therapy
and quasi—therapy settings generally suggest that
the verbal behavior of the patient and/or subject
can be modified. (p. 544)

and more specifically:
. . . the results of the research in verbal condi-
tioning strongly suggest that the therapist may
bring the verbalization of the patient under his
control by the judicious usage of certain contingent
stimuli. (p. 548)

E. Approach-Avoidance as Contingent Stimuli

Although most of the basic work guided by learning

l3

theory has been restricted to simulated therapeutic situa-
tions, beginning application to actual therapy has been
promising. Most relevant are two investigations out of
which the present grew. The first deals with the effect
of therapists' approach and avoidance to clients' statements
of hostility (Bandura, Lipsher & Miller, 1960). Therapist
approach statements are defined as " . . . verbal responses
that were primarily designed to elicit from the patient
further expressions of hostile feelings, attitudes, and
behavior." Avoidance reactions are " . . . those verbal
responses designed to inhibit, discourage, or divert the
patients' hostile expressions."

In the second study, Winder, Ahmad, Bandura, & Ban
(1962) analogously define approach by the therapist as
" . . . the reactions of the psychotherapist which are de-
signed to elicit from the patient further verbalization
of the topic under discussion." Therapist avoidant reac-
tions are those " . . . which are designed to inhibit, dis-
courage, or divert the patient from further verbalization
of the topic under discussion." We shall adopt Winder's
definitions, making minor changes in his subclassifications.
The complete Operational definition of approach and avoid-
ance used in the present study comprises all of the subcate-
gories listed under each term in the Scoring Manual (Appendix
D). Categories of patient responses similarly adopted from
Winder §t_§l are also given in Appendix D.

These categories of contingent stimuli used by

l4

therapists are uniquely broad in that they attempt to cover
most therapist statements likely to occur in the Clinical
setting. Furthermore, a global definition of the contingent
stimuli seems apropos of the social situation. Skinner
(1935) points out the generic nature of generalized stimuli
and responses, conceiving of a class, members of which have
certain common characteristics. In this instance, the com-
munality is the eliciting or inhibiting effects approach

and avoidance are thought to possess. These are deduced

from social verbal reinforcement theory as developed by
Dollard & Miller (1950) and Skinner (1957). The type of
content included in each category resembles that which Murray
(1956) includes under "approval" and "disapproval" and Rogers
(1951) categorizes as "evaluative" therapist remarks.

The Bandura gt_al (1960) content analysis of early
therapy interviews clearly indicates that approach responses
are followed more often by the client's discussion of hos-
tility rather than by his dropping the topic. Conversely,
avoidance statements are followed immediately by fewer ex-
pressions of hostility. "Approach statements encourage
the patient to express further hostility, whereas avoidance
reactions serve to decrease or inhibit such expreSsion. . . . "
These client responses may well be strongly connected to
approach and avoidance, the connections being learned through
past social interaction. Thus approach and avoidance can
be exploited by the therapist, but it is not known that they

reinforce client responses. The therapist may be presenting

15

eliciting stimuli which trigger off ready—made response
patterns in the client without strengthening S-R bonds.

The data themselves do not attest to the reinforcing effects
of therapist approach and avoidance since, again, no change
in behavior is demonstrated over time in response to the
elicitor, to the therapist himself or to the therapeutic
setting.

When Winder gt_al examined the first two tape re—
corded sessions of 23 patients in therapy, they noted the
same effect of approach and avoidance as found by Bandura
gt_al for hostile and dependent statements. The therapists'
approach elicits and his avoidance inhibits these same re-
sponses. The question, however, remains concerning rein-
forcement value.

The present study attempts to utilize two criteria
for measuring the reinforcement effect of approach and avoid—
ance: (1) increase in operant level of responses to thera-
pist approach and (2) increase in operant level of responses
to the therapeutic situation. The first of these measures
the changing probability that dependent statements immedi-
ately follow approach and avoidance as the interview pro—
gresses. Clients who are more consistently approached on
dependent statements and show a positive change in continu-
ance of dependency from the first to the second segment of
the therapeutic hour give evidence of reinforcement or learn-
ing in the interview. The second measure examines temporal

differences in Operant level of dependency statements

l6

initiated by the client, i.e., dependency statements not
in immediate response to therapist approach or avoidance.
If the probability of "client initiated" dependency state-
ments increases during the interview when the client is
predominantly approached on dependency, then the stimulus
complex surrounding the contingent stimuli has acquired
increased elicitation value through reinforcement by thera-
pist approach.

The Winder g£_§l study is one of the few (c.f.,
Ullman, Krasner & Collins, 1961) that show behavior changes
in situations outside of the laboratory or therapeutic hour
as a result of verbal manipulation. Patients whose thera-
pists approached rather than avoided dependency tended to
remain in treatment, whereas patients receiving a greater
avoidance more often terminated the relationship prematurely.
This attests to the reinforcing effect of approach stimuli
upon dependent behavior in a long-term situation. As Winder
‘g£_§l caution, however, there is some question of the con-
taminating effect of therapist personality upon continuation
and termination, since three of the therapists had none but
terminating patients (although three other therapists had
both terminators and remainers). In the present study, the
factor of therapist personality is controlled by securing
both terminators and remainers from as many therapists as

possible.

F. Selective Therapist Responses

Consistent with the point of view guiding this study

17

is the assumption that the therapist, as a teacher, func-
tions toward a goal, having a definite, if not explicit,
purpose in his management of the client. Much has been
written concerning the overall goals of psychotherapy and
conclusions vary. Some overriding desideratum such as
"better mental health" may be a universal aim in therapy.
Should this be a therapist's ultimate goal, one might find
him selectively responding to patient behavior judged more
characteristic of this ideal. Not all his behavior would
be consistent with this goal; to make it so would necessarily
preclude his accepting everything about the unhappy client.
Goals probably change according to the phase of treatment,
the sophistication of the therapist, etc. That therapists
have varying effects upon patients according to the thera—
pists' level of experience is attested to by a number of
investigations (e.g., Strupp, 1955a, 1955b; Fey, 1958;
Sullivan, 1958).

One simple goal early in therapy is to retain the
client in treatment; one method of retaining him is to rein—
force (approach) dependency. Theoretically, the experienced
therapist has learned the effectiveness of approach and
avoidance responses to dependency, i.e., has himself been
taught by experience with continuing patients. The experi-
enced therapist with a goal of retaining a client has learned
to approach dependency more, if approach does reinforce,
where a less experienced therapist would approach dependency

less. A second possibility is that experienced therapists,

18

having learned the differential values of reinforcing various
types of dependency, approach different kinds of dependency
expressions than do the less experienced. Sears g£_§l (1953,
1957) present theoretical foundations for expecting various
types of dependency bids to have varying potency for differ-
ent people. Inspection of the Patient Statement categories
found in the Scoring Manual (Appendix D) suggests a number

of dimensions along which the seven types of dependency bids
may vary (e.g., congruence with a concept of "good mental
health," degree of social acceptance, reality adaptation,

freedom from anxiety, etc.).

G. Statement of Hypotheses

Hypotheses I and II constitute the portion of this
study which explores the generality of the findings of Winder
et al (1962). The first hypothesis deals with the elicita-
tion value of approach and avoidance within the therapeutic
setting. The second hypothesis concerns the long-term ef-
fects of approach and avoidance upon staying in therapy.

I. If therapists approach client expressions of depend—
ency early in treatment, then clients continue such
expressions: if therapists avoid, clients tend to
discontinue immediately following dependency expres-
sions.

II. If therapists approach clients' expressions of de-
pendency, clients remain in treatment: but if ther-
apists avoid dependency, clients prematurely term-
inate.

Hypotheses III and IV bear on the reinforcement

value of approach and avoidance. Both state that learning

occurs within the hour.

III.

IV.

19

Therapist approach statements to client dependency
tend to increase the probability of the clients'
immediately following expressions of dependency

as the interview progresses whereas therapist
avoidance tends to decrease the probability of
immediately following dependency expressions.

If the therapist approaches clients' expressions
of dependency, then the probability of clients'
initiating dependency statements increases over
time, but if the therapist avoids, then client
initiated dependency decreases.

Hypotheses V and VI deal with the effect of thera-

pist experience upon approach and avoidance of client de—

pendency.

V.

VI.

Experienced therapists approach dependency state-
ments more often than less experienced therapists.

Experienced therapists tend to approach different
types of dependency statements than do less exper-
ienced therapists.

II. METHOD

A. Clients

The data were obtained from tape-recorded psycho-
therapy interviews with clients at the Michigan State Uni-
versity Counseling Center. Participating psychotherapists
were requested to turn in to the investigator recordings
of the first and second interviews of all new therapy cases
begun over a period of five school quarters.

Frequently a client's original purpose in seeing
a counselor is vague. For this reason, clients were accepted
for the study whether they initially requested therapy or
decided to enter therapy after a number of non-therapy in-
terviews, e.g., educational, vocational, informational.
Clients were excluded if they had had previous psychotherapy,
as determined by therapists' reports (see Appendix A, Data
Sheet for Therapists, and Appendix B, Follow-Up Question-
naire). A total of 72 tapes from 60 clients met the cri-
teria for inclusion in the sample.

If any characterization can be made of the sample
as a whole, it might be that of a late adolescent group
experiencing newly-found independence from home. Depend—
ency problems are likely to have been precipitated by the
reality of their situational separation at college. Cases
with immediately apparent acting-out symptoms or severe

personality disorganization are usually not seen at the

 

21

Center. The sample differs from those of Winder g£_§l and
Bandura gt_§l, who used older patients who had been strongly
urged to undergo therapy primarily because their problem
children were being seen at the same clinic.

Most of the clients were assigned to therapists by
the Counseling Center receptionist who had very little basis
for assignment other than time availability. Selection
entered into this process (1) when students requested and
were able to see particular therapists, or (2) when prac-
ticum student therapists were assigned cases by their super-

visors after intake interviewing.

B. Therapists

0f the therapists asked to participate, 30 contrib-
uted tapes. Six were staff members having from four to ten
years of therapy experience, 12 were internes having an
average of one year of supervised therapy experience, and
12 were practicum students carrying their first to fourth
supervised case.

Attempts to obtain all recordings from every ther—
apist met with practical difficulties. No fixed number of
cases was set; recordings were collected whenever possible.
Selective factors, therefore, may have entered into which
therapists turned in cases and which cases therapists turned

in. Table 1 presents the distribution of tapes and therapists.

C. Stage of Therapy

Due to the vagueness of purpose in many initial

22

Table 1. Descriptive Summary of the Sample

 

 

Therapist Number Number Sex of Therapy Number Number
Experience of Ther— of Clients Interview of of Re-
Level apists Cli- M F 1st 2nd Both Term- mainers
ents inat-
ors
Staff 6 ll 6 5 9 5 3 4 5
Interne 12 29 ll 18 25 10 6 11 15

Practicum 12 2O 7 13 14 9 3 6 9

Total 30 6O 24 56 48 24 12 21 29

 

23

interviews, the hour which comprised the first therapy in-
terview was decided upon immediately after its conclusion
by the client's therapist. Prescribed criteria were sup—
plied to the therapists in the form of a brief Data Sheet
included with the unrecorded tapes (see Appendix A). No
client in the sample saw a therapist more than three times
befOre the therapist reported that psychotherapy had begun
and most therapy started the first time that a client was
seen. Mutual agreement to begin therapy was usually explicit
early in the relationship. A check was made on this in the
Follow-Up Questionnaire (see Appendix B). If sufficient
doubt appeared that the client had actually started therapy,
the case was dropped from the sample.

While first and second interviews were used by Winder
.2£_§$ (1962), this method was impractical in the present
study. Therapists often were unable to record interviews.
Thus, the sample consists of 36 clients for whom only first
interviews were available, 12 clients for whom only second
interviews were available, and 12 clients who had both first
Egg second interviews recorded. To justify analyzing second
interviews along with first interviews, a comparison was
made of first and second interview scores from the 12 cli-
ents who contributed both interviews. Of 11 scores, only
the total percentage of therapist approach statements was
found to increase from the first to the second interview
(Sign test, two-tailed p.<f.05). Total approach percentage,

however, is not relevant to any of the hypotheses. The

24

sample, then, consists of an admixture of first and/or second
interviews. In some analyses presented in the Results Sec—
tion, certain reduced sample sizes may reflect the removal

of second interview data for clients who also contributed

first interviews to the sample.

D. Duration of Treatment

Counseling Center policy encourages short-term psy-
chotherapy, suggesting that therapists, ideally, should at—
tempt to terminate clients within 20 interviews when pos-
sible. However, treatment often exceeds this duration.
While longer-term clients are accepted and continued, cli-
ents terminated after 10 to 20 interviews frequently are
deemed successful cases, keeping Counseling Center goals
in mind. Much of the treatment might be described as re-
lationship therapy, other as brief psychotherapy.‘ Consider—
ing these characteristics, "premature terminators" are those
who entered treatment but who terminated before the sixth
interview without their therapists' approval. The "remainer"
group consists of clients still in therapy after 10 therapy
interviews. Copies of the Follow-Up Questionnaire (Appendix
B) were distributed to the therapists several months after
treatment began in order to assess the status of their cases.
In this way, clients terminating earlier than the sixth
interview for reality reasons (e.g., being dropped from
enrollment in the University) and those satisfactorily com-
pleting therapy before the sixth interview were eliminated

from the analysis dealing with terminators and remainers.

25

There were 29 remainers and 21 terminators, distributed

as shown in Table l.

E. Scoring of Interviews

All recorded interviews were scored according to
a modification of content analysis manuals employed by Winder
g£_§l (1962) and Bandura g£_§l (1960) (see Appendix D). The
basic unit of analysis is the "statement," generally defined
as the total verbalization of one speaker, bounded by the
preceding and succeeding speeches of the other speaker.

An "interaction sequence" consists of three such units:

a "patient statement," a "therapist response," and the fol-
lowing "patient response." The patient response both con-
cludes one interaction sequence and begins a new one, thereby
becoming the next patient statement.

Each scoreable patient statement is classified as
one or more of the following: (1) one or more of seven
types of dependency statements, (2) a hostility statement,
or (3) an "other" type of expression. Each subsequent ther-
apist response, the contingent stimulus, is classified as
an approach or an avoidant statement in reference to the
patient statement category preceding it. In the case of
approach or avoidance of dependency or hostility, the ther-
apist response is categorized exclusively into one or more
of eight approach 2; five avoidance subcategories. The
concluding patient responses are scored as continuing or

not continuing with the dependent or hostile content of the

26

preceding patient statement. Additionally, the patient
response is scored as a new patient statement.

In order to minimize contamination in scoring, the
judges code each unit before listening to any subsequent
unit. Scoring therapists' responses to the hostility and
"other" categories of patient statements enables alternative

hypotheses to be tested and thereby serves as a control.1

F. Scoring Reliability

Two judges coded interviews: Judge B for the most
part unfamiliar with the therapists involved in the study,
and both Judge A and Judge B ignorant of the duration of
treatment each client received. The judges established
familiarity with the scoring system by mutually coding a
series of interviews. Several independently coded inter—
views were then assessed for trial agreement. None of these
tapes were used subsequently.

Interjudge reliability on the scores used in the
testing of hypotheses was determined as follows:

Over a period of several months, Judge A scored all
72 tapes. The reliability pool consists of a randomly se—
lected sample of ten tapes taken from the first 35 scored
by Judge A and another set of ten tapes selected at random
from the last 37 scored by Judge A. Judge B coded his first

ten early in the process and his last ten several months

 

lThe hostility data are used more directly in a
M.A. thesis in preparation by David A. Kopplin.

27

later. This procedure results in a reliability measure

reduced by practice effects of Judge A and forgetting ef-
fects of Judge B. Product-moment correlations of various
scores were computed to obtain interjudge reliability co-
efficients. They are presented under appropriate sections

in the next chapter.

III. RESULTS

To facilitate the presentation of results, each
hypothesis will be dealt with separately. In the subsec-
tions dealing with each hypothesis, the first describes
the scores used in the analysis, the second presents inter-
judge scoring reliability, and the third section comprises

the analysis and outcome of the hypothesis tested.

A. Hypothesis I

Hypothesis I states that when dependency statements
are approached by the therapist, the client continues the
topic under discussion, but when the therapist avoids, the
client does not continue.

Scores. For this test, the percentage of dependent
statements which were approached rather than avoided by the
therapist (D Ap%) was determined for 72 tapes. The percent-
age of units in which clients continued with further depend—
ency when they were approached was then computed (D Ap 0%).
The parallel percentage for continuance of dependency fol-
lowing therapists' avoidance (D Av 0%) was also determined.

Reliability. Table 2 indicates that interjudge

agreement on these percentages is reliable. All coeffici-

ents of correlation computed in this study are product-moment

correlations unless otherwise noted. The variation in size
of the reliability sample and the entire sample is due to

the absence of units in one or more tapes.

29

Table 2. Interjudge Reliability for Percentage Continuance
of Dependency Following Approach and Avoidance

 

 

Score N r p.
D Ap 0% 20 .81 <.Ol
D Av 0% 18 .76 <.Ol

 

Outcome. The mean percentages of D Ap 0% and D Av 0%
are 63 and 26, respectively, for 68 tapes. In 90 per cent
of the tapes, D Ap 0% is greater than D Av C%. The sign
test shows D Ap C% to exceed D Av 0% (p.<:.OOOl, one-tailed
test). Approach to dependency clearly elicits further depend-
ency while avoidance inhibits continuance of the topic.

To investigate whether approach has high elicitation
value and avoidance low elicitation value within experience
levels of therapists, sign tests were made within staff, in-
terne and practicum groups. The same superiority of approach

in eliciting dependency holds for each level (p.<:.0001).

B. Hypothesis II

The second hypothesis states that approach is asso-
ciated with the clients' remaining in treatment and avoidance
with premature termination of the relationship.

Sgggg. The percentage of therapist approach responses
to dependency (D Ap%) was computed for each of the 29 re—
mainers and 21 terminators seen by 27 therapists.

Reliability. The interjudge reliability coefficient
of D Ap% for 20 interviews is .61 (p.<:.01).

5O

Outcome. The Mann—Whitney U test reveals no dif-
ferences in D Ap% between the two groups (p. = .41, one-
tailed test).

Winder g£_§l (1962) had used only student therapists.
Since their sample was less heterogenous than the sample
of therapists in this study, analyses within experience
levels were performed. 0f the terminators and remainers
meeting the criteria, seven were treated by staff members,
28 by internes and 15 by practicum students. D Ap% fails
to differentiate terminators from remainers within any of
the three levels.

It is important to mention that the above samples
may have been biased by therapist personality characteris-
tics in that 21 out of 27 therapists contributed tapes of
only terminators or only remainers. The distribution of
clients into the two groups, therefore, is based not only
upon continuation or termination, but also upon which thera-
pists the clients had seen. Controlling for this factor
by using pairs of terminators and remainers seen by the
same therapist yields only nine pairs of clients seen by
six therapists. The Wilcoxin test for differences in D Ap%
between the nine pairs of clients shows no trend in the ex-
pected direction.

There is still another difference between the pres-
ent sample and that studied by Winder gt_§1. Sixteen of
their clients had D Ap% higher than 50 per cent and seven

clients received less than 50 per cent approach to dependency.

31

The therapists in the present study approach dependency
much more frequently (Mdn. D Ap% = 76). Only two therapists
approached dependency less than 50 per cent of the time;
one terminator's D Ap% was 47 and one remainer's was 43.

It is possible that differences in D Ap% are ob-
scured by the definition of terminating and remaining in
treatment, since the cut-off point in this study differs
from Winder's. We set a six—interview maximum for term-
inators and a lO-interview minimum for remainers. Winder
g£_§l used a lO-interview maximum and a 20-interview min-
imum. Setting a more stringent cut—off point of not more
than six and not less than 20 interviews (yielding 21 term-
inators and 13 remainers), however, does not differentiate
the groups on D Ap% (Mann-Whitney U test).

Other percentages were examined in an attempt to
separate remainers from terminators. For all 50 clients,
the following scores do not make the discrimination: D Ap 0%,
D Av 0%, H Ap% (approach to hastility), Total Ap%, and a
difference score reflecting the change in D Ap 0% from the
first to the second segment of the interview. Reliability
coefficients for all these scores are adequate and are pre—
sented under sections dealing more directly with the scores.

Using the nine pairs of clients matched for thera—
pists, the Wilcoxin test does not separate terminators from
remainers on the basis of the following scores: D Ap 0%,

D Av 0%, H Ap%, H Ap 0%, H Av 0%, Total Ap%, and the dif-
ference between D Ap 0% in the first and second segment of

the interview.

32

Terminators and remainers differ on D Ap 0% for
the six staff members (Mann-Whitney U, p.<:.02, two-tailed
test). Terminators continue dependency more when approached
than do remainers. However, three therapists had none but
remainers and two therapists had none but terminators.
Therapist personality differences could not be statistic—
ally controlled within experience levels because of the

shrinking sample size.

0. Hypothesis III

Hypothesis III deals with the reinforcement effect
of approach; if therapists approach dependency, the prob-
ability of further discussion of dependency increases over
the course of the interview. If therapists avoid depend-
ency, continuance of dependency decreases.

Scores. Results presented under Hypothesis I con-
firm that therapist approach responses to dependency con-
sistently elicit further dependency and avoidant responses
elicit non-dependent statements. Hypothesis III investi-
gates learning due to the consistent elicitation. When
therapists approach dependency, then, an increment should
be found in D Ap 0% from the first segment of the interview
to the second.

To insure that the frequency of therapist responses
to dependency is held constant from segment to segment,
each of the 72 interviews is divided into two equal segments

on the basis of the number of dependent statements made by

55

clients. The Wilcoxin test shows no significant differences
in central tendency of D Ap%from segment 1 to segment 2
(p. = .29, two-tailed test). In addition, the frequency
of D Ap does not increase from the first to the second seg-
ment (sign test, p. = .46, two-tailed). Therefore, D Ap%
and frequency of D Ap do not change but remain relatively
constant over the interview. The Piersonian correlation
coefficient between D Ap% in the first and second segment
of the interview is .38 (p.<:.01). With these controls, a
change in D Ap 0% cannot be attributed to changes in the
number of dependent statements made, their percentage, or
the frequency of approach responses from segment to segment.
D Ap 0% and D Av 0% were computed separately for
each segment of the 72 interviews.
Reliability. Interjudge reliability coefficients
of scores used in this section are shown in Table 3.

Table 3. Interjudge Reliability of Scores in the Two Seg-
ments of the Interviews '

 

 

Segment 1 Segment 2
Score r p. N r p. N
D Ap 0% .66 .01 2O .50 .05 19
D Av 0% .76 .Ol 17 .52 .05 17
D Ap .63 .01 20 .43 .05 19

 

Interjudge reliability for the number of dependency
statements (used to divide interviews into segments) scored

in the 20 reliability tapes is .90 (p. (.01). For 20 tapes,

54

interjudge reliability on the two control scores in Table
7, approach to hostility (H Ap%) and Other Ap%, is .87 and
.70, respectively.

Outcome. D Ap 0% in segment 2 does not exceed
D Ap 0% in segment 1 when the sign test is applied to the
entire sample. Of 67 tapes, 35 decreased and 32 increased
in percentage continuance following approach. Clients do
not tend to continue dependency more in the second segment
when the analysis ignores differences in D Ap%.

It is impossible to obtain a predominantly approached
group distinct from a predominantly avoided group since,
as described under Hypothesis II, the entire sample is highly
approached on dependency. The restricted range, therefore,
obscures differences in the following tests. The hypothesis
pertains to consistent approach and avoidance rather than to
consistent and inconsistent approach. Therefore, an analysis
of subgroups is called for. The "High Approach" group, then,
is composed of clients receiving the mean D Ap% score (73%)
or greater. The "Low Approach" group are clients approached
on dependency less than 75% of the time. Table 4 indicates
that in comparison to the Low Approach group, the High Ap-
proach group does not change significantly in D Ap 0% or
D Av 0% from segment to segment. The directions of changes,

however, are consistent with the directional hypothesis.

55

Table 4. Change Over Segments in Percentage Continuance
of Dependency for High and Low Approach Groups

 

 

 

 

Chi Square
Score Group Increase Decrease p.*
High Approach 21 17
D Ap 0% <.10
Low Approach ll 18_
Total 52 55
High Approach 11 7
D Av 0% ‘<.15
Low Approach 11 14
Total 22 21

 

*One-tailed test

The data for D Ap 0% are analyzed in more detail
in Table 5, which presents the comparison for High and Low
Approach groups when clients start out in the first segment
with either low or high D Ap 0% scores, i.e., D Ap 0% scores
above and below the mean 63%.
Table 5. Change Over Segments in Percentage Continuance

of Dependency Following Approach (D Ap 0%) for
High and Low Approach Groups

 

 

Seg. 1 Fisher's Exact
Score Group Increase Decrease p.*
Low High Approach 13 2
D Ap 0% .05
Low Approach 8 5
High High Approach 8 15
D Ap 0% .25
Low Approach 5 l5

 

*One-tailed test (Tocher's modification)

56

As hypothesized, the High Approach group tends to
score higher on D Ap 0% than the Low Approach group when
the percentage of continuance is low in the first segment.
This must be interpreted as a minimal confirmation of the
hypothesis because of the split imposed upon the High Ap—
proach group. In addition, without Tocher's modification
of Fisher's Exact test, the difference is not significant.
(The modification makes the test less conservative.)

Similarly, the Low Approach group which begins the
interviews with high D Ap 0% scores tends to decrease in
continuance, but not significantly more than the High Ap-
proach group. The results of both tests, then, fail to
confirm the hypothesis, but weak learning trends appear.

The same kind of analysis, applied to the groups
when first segment D Av 0% scores are low and high, i.e.,
above and below the mean 26% gives the results shown in
Table 6.

Table 6. Change Over Segments in Percentage Continuance

of Dependency Following Avoidance (D Av 0%) for
High and Low Approach Groups

 

 

Seg. 1 Fisher's Exact
Score Group Increase Decrease p.*
Low High Approach 8 l
D Av 0% .78
Low Approach 9 1
High High Approach 3 6
D Av 0% .25
Low Approach 2 13

 

3Pane—tailed test (Tocher's modification)

57

While there are similar learning trends for the
"high beginners," the differences are not significant.

The hypothesis is not confirmed.

Both Table 5 and Table 6 show a clear tendency re-
gardless of D Ap% for "low beginners" in continuance scores
to continue more in the second segment and for "high begin-
ners" to continue less as the interview progresses. The
Chi Square test shows significant directional changes to
occur for both D Ap 0% and D Av 0% scores (p. (.001, two-
tailed test). The product-moment correlations between first
segment scores and second segment scores of the high begin-
ners and of the low beginners, for both D Ap 0% and D Av 0%,
are all below .05. There is no association between magni-
tude of continuance scores from segment to segment when
clients start low or high on continuance. Correlations
of the total D Ap 0% and total D Av 0% scores from segment
to segment are also not significant.

In interpreting the slight changes in continuance
noted in the above two tables, it is relevant to investi—
gate the relationship of the percentage continuance of de—
pendency to therapists' approach to the category and thera—
pists' approach to other categories of client behavior.
Table 7 indicates that, as the interview progresses, only
D Ap% becomes more closely associated with continuance of
dependency following approach. For the entire sample, the
relationship is low but significant in the second segment.

Other scores remain unrelated in the second segment.

58

Table 7. Correlations of Percenta e Continuance of Depend-
ency Following Approach TD Ap 0%) with Therapist

 

 

Responses
D Ap% H Ap% Other Ap%
D Ap 0% N r N r N r
Segment 1 72 .02 72 .04 72 .00
Segment 2 71 .29* 71 .00 71 .12

 

*p..<.01, two-tailed test

The significance of the difference between the cor-
relations of D Ap% with the first and the second segment
D Ap 0% scores is at the one-tailed 5% level (2 = 1.64).
In the second segment of the interview, discussion of de—
pendency is more related to the therapists' apprbach re-
sponses than it is in the first segment.

D Av 0% fails to show a similar relationship with
D Ap%. The correlation of D Ap% scores with D Av 0% in
the first segment is .09, while in the second segment it
is .07. As discussed earlier, the Low Approach group, in
fact, received more approach than avoidance of dependency.

Table 8 presents correlations of D Ap% scores with
D Ap 0% in each segment for the group which shows signif-
icant increases in continuance (clients whose continuance
of dependency following approach is initially low). Al-
though there is an increase in continuance, D Ap% is not

significantly related to the change.

59

Table 8. Correlations of Percentage Approach to Dependency
(D Ap%) with Continuance Following Approach

 

 

(D AP 0%)
D Ap 0% N* r p.
Segment 1 3O .22 N.S.
Segment 2 3O .26 N.S.

 

1FNgincludes two tapes that are not included in Table 5 be—
cause D Ap 0% neither increased nor decreased.
D. Hypothesis IV

Hypothesis IV concerns reinforcement of dependency
statements initiated by the client. When therapists approach
dependency, clients initiate more dependency statements,
and when therapists avoid, client initiated dependency de-
creases during the course of the interview.

S3253. Client initiated dependency (CID) statements
are clients' dependency expressions which are not in immed-
iate response to therapists' approach or avoidance. For
a unit to be scored as a client initiated dependency state-
ment, then, the interaction sequence which precedes it con-
tains only Hostile or Other response categories. When ther-
apists respond to these categories, and simultaneously at-
tempt to introduce discussion of dependency, the judges
score "Dependency initiated by therapist." Consequently,

a succeeding dependency statement cannot be scored as CID.

CID scores are expressed as the percentage of units
in which the client initiates dependent statements in com-

parison with the opportunity he has to initiate them.

4O

"Opportunity to initiate" is the number of non-dependent
(Hostile or Other) units which do not immediately follow
therapists' D Ap, D Av, or initiation of dependency responses.
In the scoring system, therefore, CID% is independent of
the frequency and percentage continuance of dependency.
CID% was computed in first and second segments of
interviews. Each segment of an interview contains half
of the total number of dependency units in that interview.
Results presented under Hypothesis III indicate, for the
entire sample, that D Ap% is stable and that the frequency
of D Ap does not increase from segment 1 to segment 2.
Reliability. Interjudge scoring reliability of
CID%, shown in Table 9, is adequate in each segment.

Table 9. Interjudge Reliability of Percentage CID in the
Two Segments of the Interviews

 

 

CID% N r p-
Segment 1 18 .75 ‘<.Ol
Segment 2 19 . 74 < . Ol

 

Outcome. High and Low Approach groups were formed
by dividing the sample at the mean D Ap% score (73%). Table
10 indicates no significant difference between High and
Low Approach groups in the change of CID% from the first

to the second segment.

41

Table 10. Change Over Segments in Percentage CID for High
and Low Approach Groups

 

 

Score Group Increase Decrease Chi Square p.*
High Approach 19 2O

CID% (315
Low Approach 9 2O

 

tOne-tailed test

For the entire sample, the product-moment correla-
tion between D Ap% scores and CID% in segment 1 is .01.

In segment 2, the correlation increases to .15. However,
the difference between the two correlations is not signif-
icant. CID%, then, does not appear to vary in association
with D Ap%.

Table 11 presents changes in CID% scores from seg-
ment to segment for clients who received segment 1 CID%
scores below and above the mean, 32%. The findings are
not significant although the CID% scores of clients begin—
ning the interview with low CID% scores tend to change in
the hypothesized directions.

Table 11. Change Over Segments in High and Low Segment One
Percentage CID for High and Low Approach Groups

 

 

Seg. 1 Chi Square
Score Group Increase Decrease p.*
Low High Approach 16 8
CID% <.15
Low Approach 9 ll .
High High Approach 5 12
CID% <32O
Low Approach 0 9

 

zOne-tailed test

42

Product—moment correlations between D Ap% and CID%
scores in segment 1 and segment 2 for the high and low first
segment CID% scores do not reach significance. The corre-
lation for the total group CID% scores and D Ap% is simi-

larly not significant.

E. Hypothesis V

Hypothesis V states that experienced therapists
approach dependency statements more than less experienced
therapists.

S2252. Only one interview from each therapist is
included in this analysis. This procedure prevents unequal
weighting of D Ap% scores by individual therapists within
each experience level. The tapes from therapists were chosen
on the basis of the alphabetical order of clients' names,
thus giving six D Ap% scores from staff members, 12 from
internes, and 12 from practicum students.

Outcome. The Mann-Whitney §,test indicates that
staff level therapists approach dependency statements more
than either internes or practicum students. Table 12 shows
the significance levels of the differences.

Table 12. Differences of Percentage Approach to Dependency
Between Experience Levels

 

 

Direction of Difference One—tailed p.
Staff > Interne . 02
Staff >’Practicum .03

Practicum >>Interne .29

 

45

F. Hypothesis VI

Hypothesis VI predicts that experienced and inex-
perienced therapists differ in the percentage of approach
responses to subtypes of dependency statements made by
clients.

Reliability. Interjudge scoring reliability is
satisfactory for only two of the eight subcategories.
Table 13 presents the reliability coefficients. Due to
the small number of cases, Spearman rho coefficients are
computed in this analysis.

Table 13. Interjudge Reliability for Percentage Approach
to Subcategories of Dependency

 

 

 

Dependency Subcategory N rho

Problem Description 12 .54*
Help Seeking 18 .22

Approval Seeking 18 ' .32

Company Seeking 14 ,53*
Information Seeking 16 .32

Agreement 12 .06

Disapproval Concern 13 .OO

Initiative Seeking 5 .OO

*p.< .05

Part of the difficulty in obtaining satisfactory
interjudge reliability is the very low frequency of some
of the subcategories in each tape. In the extreme instance,
initiative seeking, units appear in only five of the 20
tapes assessed for reliability. Of these five tapes, the
mean number of such units per tape was only 6 for Judge A

and 2.4 for Judge B.

44

Interjudge disparity in scoring for approach and
avoidance of subcategories is statistically magnified since
scores are expressed in percentages. There were some tapes
for which one disagreement about approach and avoidance
affected the score by 50% or 100%.

Outcome. Mann-Whitney U tests determine differences
between experience levels in per cent approach to the Prob-
lem Description and Company Seeking subcategories. There
are 16 practicum, 21 interne, and 11 staff tapes included
in the analysis of approach to Problem Description, while
17 practicum, 32 interne, and 13 staff tapes are used in
computing approach to Company Seeking. Practicum students
tend to approach Problem Descriptions more than internes,
while staff therapists tend to approach Company Seeking
statements more than internes. However, both of these dif-
ferences fail to reach significance (p.<:.lO, two-tailed
tests). No other differences were found.

When only one tape from each therapist is used,

N becomes too small for a meaningful statistical test.

Sign tests show no differences between experience
levels in the effectiveness of approach in eliciting fur-
ther dependency (D Ap 0% and D Av 0%).

IV. DISCUSSION
Approach-Avoidance as Contingent Stimuli:
Hypotheses IprLpand VI

Hypothesis I pertains to the immediate effects of
therapists' approach and avoidant responses to clients'
expressions of dependency. In the present study, thera-
pists' responses that are designed to promote or to dis-
courage discussion of dependency do have these effects upon
clients. The results support Winder 23_§l (1962). Wider
generalization of the eliciting effects of approach and
avoidance is justified, since the populations of clients
and therapists in each investigation differ in several re-
spects. Although no measures are available for clients
in either sample, the dependency conflicts of late adoles-
cent students coming to a Counseling Center would seem cen—
tral to their stage of personality development. On the
other hand, parents seen at a Child Guidance clinic are
not necessarily experiencing a transition to independence.
Their dependency perhaps is more situational or problem-
centered, being motivated by a wish to get help for their
children. Approach and avoidance effectively elicit and
inhibit expressions of dependency no matter which source.

The habit of responding to approach and avoidance
is not confined to dependency. Parallel results regarding

the elicitation of immediately following hostility discussion

45

46

are noted by Kopplin (1963), who confirms findings of Bandura
g£_§l (1960).

Some importance is attached to the repeated confirma—
tion with different client and therapist populations of the
hypothesis that therapists have a direct and effective role
in determining the course of discussion in early therapy
interviews. As Murray (1956) found, even the non-directiv-
ity of Rogers is "directive" in the sense that clients are
swayed by mild therapist approval and disapproval.

Therapists appear to exploit a strongly entrenched
habit of the client to follow direction as early in therapy
as the first hour. The habit may be generalized from or-
dinary social interaction or from prototypes of the depend-
ent relationship (parent-child, teacher-pupil). The second
possibility seems more likely.

Demonstrations of the leading role which therapists
play, together with the resulting dependent role of clients,
explicate at least one aspect of psychotherapeutic technique.
Other questions arise concerning (1) the occurrence and
the effectiveness of this direction in various stages of
psychotherapy, (2) possible client and therapist differences
determining approach and avoidance and (3) the influence
of therapist's goals and values upon his manipulation of
clients.

According to Hypothesis V, experience teaches ther-
apists to approach dependency. Significant differences

in the present study support the hypothesis in regard to

47

staff level therapists compared to internes and practicum
students. The difference in D Ap% between the latter two
groups is slight, as is the difference in their degrees

of experience. Although these findings must be regarded
with care because of the small sample size, the differences
suggest that exploitation of a well—learned habit is a
therapeutic technique learned through experience with the
effects of approaching and avoiding dependency.

In addition to quantitative differences in percent-
age of approach responses, there may be substantial qual-
itative change occurring with increasing experience. Ap-
propriateness of the approach response within the broad
category of dependency may be one of these. There is some
evidence that therapists' experience is not related to their
sensitivity to clients' expressions, as rated by experts
(Rosenberg, 1962). The present study finds that the effec-
tiveness of approach responses does not differentiate ex-
perience levels. All levels were equally able to elicit
dependency when they approached and to inhibit dependency
when they avoided dependent statements. The elicitation
value of approach and avoidance seems to lie within the
verbal response itself, rather than in the experience level
of therapists who use it.

It would be interesting to determine the relative
potencies of subcategories of approach and avoidance in
eliciting continuance. This type of analysis may disclose

qualitative differences in the elicitors and in therapists'

48

use of them. Other dimensions of responses are not meas-
ured by this system. Therapist style, defined by Strupp
(1957) and studied by Rottschafer & Renzaglia (1962) is one
such factor known to influence client dependency. Through-
out the scoring of the tapes, both judges were concerned
with unscorable aspects of therapists' responses. For ex-
ample, while one therapist's "exploration" approach was
inquiring, another's was tinged with hostility.

Although differences in the tone of the therapist's
responses may be crucial in their cumulative effects upon
clients, the scoring manual emphasizes explicit verbal con-
tent; minor shifts in content and affect are not scored.

Not all approach and avoidance responses are as mutually
exclusive as the manual allows. Adhering to verbal content
in scoring may increase reliability and support the approach-
avoidance model at the cost of sensitivity to other import-
ant variables.

Whether manipulation of verbal expressions of de-
pendent content is therapeutically relevant or effective
is not known. The distinction between "talking about" one's
own or another's dependency and "being" dependent is a prob-
lem in validity of measurement for any study utilizing con-
tent analysis. Strupp (1962) provides a thoroughgoing crit—
icism of the technique, most of his objections pertaining
to questions of validity. Janis (1943) bases his defense
of content analysis upon its usefulness in establishing a

set of meaningful and lawful relationships in the data to

49

which it is applied. As such, he regards the test of con-
tent analysis as similar to the proof of a theory; validity
lies, in part, in heuristic value.

The portion of the study most hampered by low scor-
ing reliability is Hypothesis VI. Only two subcategories
of dependency can be tested for differences in approach
percentage for therapists of different experience levels.
The trends that were found are obscured by unequal repre-
sentation of D Ap% scores within each experience level.

In an attempt to raise scoring reliability, the
investigator tried to establish mutually exclusive subcate-
gories of dependent statements. However, the criteria for
each type became impractically narrow, since client verbal-
izations typically are complex. Consequently, subcategories
were used as they had been by Winder g£_§l, some being too
iinclusive and others too rare in the present sample.

Inherent in the scoring subcategories is a factor
which may explain the sparseness of units in some classi-
fications. Twelve therapists were asked to rank the depend-
ency subcategories on a continuum of "mental health." They
show strong agreement on independent rankings of most types
of dependency hide, the infrequently scored Disapproval
Concern and Initiative Seeking subcategories being judged
least healthy. Thus, the relatively healthy student popu-
lation sampled may preclude use of these categories. This
appears to have happened in transfering a manual derived
from a population of clients somewhat different from those

studied here.

50

In line with the Operant conditioning model, anal-
yses of psychotherapy interviews throughout the duration
of treatment may reveal that approach and avoidance become
distributed according to the rankings of dependency sub-
categories on the continuum of "mental health." Because
the present study examines only early psychotherapy, such
an analysis cannot be made.

Approach—Avoidance as Consistent Elicitors:

Hypotheses II, III, and IV

Hypothesis II deals with the long-term reinforce-
ment effects of approach and avoidance upon remaining in
psychotherapy. The findings show no difference between
terminators and remainers in terms of therapists' responses
to dependent statements. One side of the hypothesis is
not adequately tested, however, since the sample lacked
clients whose dependency was avoided consistently. Rather,
the Low Approach group was inconsistently approached. The
data are more accurately applied to a test of the effects
of consistent versus inconsistent elicitation, than to a
test of two consistent elicitors of competing responses.

One might speculate about the reasons that thera-
pists in this study predominantly approach dependency in
contrast to therapists' behavior in the Winder 33_§l study.
They write:

Further investigation suggested that the form of
the expressions of dependency by terminators and

remainers may differ. It is suggested that some
forms of expressions may provoke avoidance if the

51

psychotherapist follows conventional notions about

prOper psychotherapeutic technique. (p. 134)
Clients in the present study are typically self-referred
adolescents experiencing conflicts about their transition
from dependency. Clients studied by Winder g£_§l were adults
seeking help primarily for their children and only tangenti-
ally for themselves. Their personal problems encompassed
a wide range. In conventional psychotherapy, the first type
of dependency would seem to be more "approachable" than the
second. In addition, therapists may recognize and want to
deal with the more central problems in their clients. No
comparisons are available along these dimensions in the
present study.

Additionally, the means by which therapists get
clients may help to explain the failure of D Ap% to differ-
entiate terminators from remainers. Where some therapists
in the Winder 33_§l study were assigned cases following
diagnostic and intake interviews by a different therapist,
the staff members and internes in the present study were
free to choose their own clients, to retain or terminate
them, to refer them to a waiting list, or to send them to
other therapists. Therapists who incur some formal obliga-
tion to see a particular client have greater reason to avoid
dependency than do therapists who are free to avoid the
entire relationship.

A small scale study bears on this point. During
the earliest phase of data collection, the investigator

asked therapists to complete a short rating scale after

52

seeing new clients. The scale is adapted from Stoler (1961)
and is reproduced in Appendix C. Therapists rated on a
scale from 1 to 6 their feelings of liking or disliking

the client they had just seen and tape recorded for the
present study. The scale was discontinued after the first
28 tapes were obtained. The results are shown in Table 14.

The mean score is 3.2 for terminators and 1.9 for remainers.

Table 14. Client Likability and Remaining in Therapy

 

 

Client . Fisher's Exact
Likability Terminators Remainers ‘ p.*
Liked (1-3) 1 15

.05
Disliked (4-6) 7 7

 

1Two-tailed test

Clients may remain in response to their therapists'
initial liking of them, therapists may retain clients whom
they like, or some third factor may affect both variables.
In any case, both therapists and clients in the present
study had relatively unobstructed choice regarding continu-
ation of therapy. This may not have been true with Winder
91.51-

Hypotheses III and IV are not adequately tested,
since they deal with two mutually contradictory elicitors.
As noted earlier, consistent avoidance of dependency
(D Ap%<:50) characterizes only two therapists in this

sample. Therefore, the tests made are between effects of

55

consistent versus inconsistent elicitation (High and Low
Approach groups).

The study fails to confirm the hypothesis that the
fact that the therapist exploits a pre-existent habit (con-
tinuing a subject when approached) makes his ability to
manipulate the client's verbalizations more powerful as
the interview continues. Trends in the predicted directions
are found with both High and Low Approach groups; as the
interview progresses, continuance of dependency becomes
more probable with approach, and discontinuance increases
with avoidance. However, there is no significant learning
effect when the High and Low groups are compared. Corre-
lations were computed to investigate the relationship be-
tween therapists' responses and the measures which showed
the slight learning trends, D Ap 0% and D Av 0%. The asso-
ciation between D Ap% and D Ap 0% from first to second seg-
ments of interviews increases significantly, while thera-
pists' approach to Hostility and Other categories of client
responses remain unrelated to continuance following approach
to dependency throughout the interview.

Thus, over the course of the interview, there is
a small but significant build-up of association between
therapists' approach responses and the degree to which cli-
ents continue the topic following approach. The number of
clients for whom the probability of continuance increases,
however, is not significant. The relatively low correlation

coefficient obtained suggests that there are contaminating

54

factors being measured along with one or both of the vari-
ables. Measurement of the qualitative differences in ther-
apists' approach responses and shifts in content within

the broad category of dependency may be accountable. These
unmeasured variables also may have confounded the testing
of increased continuation in the second segment.

Essentially the same trends are noted between ap—
proach to dependency and client initiated dependency but
low significance levels lead to the rejection of the fourth
hypothesis.

An analysis of continuance scores which are low
or high in the first segment of the interview reveals for
both groups a general regression toward the mean in the
second segment, regardless of the percentage approach cli-
ents receive. "Low beginners" increase continuance and
"high beginners" decrease from the first to the second seg—
ment. The increase for low beginners may be due to clients'
becoming adapted to the situation and a subsequent weakening
of initial inhibition to react in a dependent manner. The
decrease for high beginners may reflect stimulus-satiation
to the tepic or accumulating dependency anxiety resulting
in a build-up of inhibition to discuss dependency. In either
case, the changes are unrelated to therapists' approach to
dependency as measured in this study.

Although learning is not demonstrated in this study,
the consistency of trends is suggestive. Larger and more

varied samples of clients and therapists, together with

55

more refined scoring categories are needed. Other content
categories and more extensive time sampling are necessary
in order to investigate the applicability of the indicated

trends to psychotherapy in general.

V. SUMMARY

Dependency of clients upon psychotherapists is a
frequent and important occurrence in the early hours of
psychotherapy. This study examines the phenomenon as an
effect of therapists' approach and avoidance responses to
clients' verbalizations of dependent content. The hypoth-
eses state that when therapists approach dependency, cli-
ents continue the tepic but when therapists avoid, then
clients discontinue. Aside from the elicitation value of
approach, these responses are assumed to reinforce discus-
sion of the topic. Learning occurs if (1) the probability
of a client's continuing the topic increases as the inter—
view progresses and (2) the probability of a client's in-
itiating dependency statements increases over time. Con-
versely, therapists' avoidance of dependency reinforces
the tendency for clients to discuss non-dependent topics.
The study also examines the longer-term effects of reinforce-
ment upon staying in therapy.

Further hypotheses state (1) that therapists learn
to approach dependency as an initial technique of therapy
and (2) that therapists learn to reinforce selectively var—
ious types of dependent statements.

A content analysis of 72 recordings of early psy-
chotherapy interviews with clients at a university counsel-

ing center shows that approach to dependency elicits further

56

57

discussion of dependency and avoidance elicits discontinu—
ance. For clients whose dependent statements are predom-
inantly approached, the probability that approach elicits
further discussion and that clients initiate dependency
statements does not increase significantly as the hour pro-
gresses, although there are trends in the expected direc-
tions. Due to the high frequency of approach responses
made to all clients, effects of avoidance are inadequately
measured. Relatively infrequent approach, however, does
not reinforce the competing tendency to discuss.non—depend—
ent topics. The data suggest individual differences in
inhibition of each habit independent of the therapists'
reSponses measured here.

Terminating and remaining in therapy are unrelated
to the percentage of approach and avoidance therapists gave
when clients discussed dependency. This fails to support
previous findings. However, more "likable" clients are
very likely to remain in therapy.

Staff level therapists more frequently approach
than avoid dependency in contrast to practicum and interne
level therapists. Approach to subtypes of dependency is
not substantially affected by experience level.

Limitations inherent in the study are discussed

and implications are noted for further research.

REFERENCES

Alexander, F., & Ross, H. Dynamic psychiatry. Chicago:
Univ. of Chicago Press, 1952.

Bandura, A., Lipsher, D. H., & Miller, P. E. Psychother—
apists' approach-avoidance reactions to patients'
expressions of hostility. J. consult. Psychol.,
1960, 21, 1-8.

Denny, M. H., & Adelman, H. M. Elicitation theory: an
analysis of two typical learning situations. P y-
ChOlo Rev. ’ 1955, .62, 290-2960 -

Dollard, J., & Auld, F., Jr. Scoring human motives: a
manual. New Haven, Conn.: Yale University Press,
1959.

Dollard, J., & Miller, N. E. Personality and psychotherapy.
New York: McGraw—Hill, I950.

Fenichel, O. Tgepsychoanalytic theory of neurosis. New
York: Norton, 1945.

Ferster, 0. B. Reinforcement and punishment in the control
of human behavior by social agencies. Psychiat.
Res. ReptSO ’ 1958’ 2.9-, 101-118.

Fey, W. F. Doctrine and experience: their influence upon
the psychotherapist. J. consult. Psychol., 1958,
22: 405-409-

Fromm-Reichmann, F. Principles of intensive sychotherapy.
Chicago: Univ. 0 Chicago Press, 195 .

Greenspoon, J. The effect of a verbal stimulus as a rein-
forcement. Proc. Ind. Acad. Sci., 1950, 59, 287.
(Abstract).

Greenspoon, J. The reinforcing effect of two spoken sounds
on the frequency of two responses. Amer. J. Psy-
chol., 1955, 68, 409-416. ,

Greenspoon, J. Verbal conditioning and clinical psychology.
In A. J. Bachrach (ed.), Ex erimental foundations
of clinical psychology. New York: Basic Books,

9 , 510-555

58

59

Hildrum, D. 0., & Brown, R. W. Verbal reinforcement and
interview bias. J. abnorm. soc. Psychol., 1956,
22, 108-111.

Janis, I. L. Meaning and the study of symbolic behavior.
Perchiatryu 1943. _6_. 425-439-

Kopplin, D. A. Hostility of patients and psychotherapists'
approach-avoidance responses in the initial stage
of psychotherapy. Unpublished master's thesis,
Michigan State University, 1963.

Krasner, L. Studies of the conditioning of verbal behavior.

Murray, E. J. A case study in a behavioral analysis of
psychotherapy. J. abnorm. soc. Psychol., 1954,
i2, 305-310 .

Murray, E. J. The content-analysis method of studying psy-
chotherapy. Psychol. Monogr., 1955:.19 (13, Whole
No. 420 .

Rogers, 0. R. Counseling and Psychotherapy. Boston:
Houghton Mifflin, 1942.

Rogers, 0. R. Client-centered therapy. Boston: Houghton
Mifflin, 1951.

Rogers, 0. R. The necessary and sufficient conditions of
therapeutic personality change. J. consult. Psy-
ChOJ-O , 1957’ 22;, 95-1030 .

Rogers, 0. R. The process conception of psychotherapy.

Rosenberg, E. H. Correlates of a concept of therapeutic
sensitivity. Unpublished doctoral dissertation,
Michigan State University, 1962.

Rottschafer, R. H., & Renzaglia, G. A. The relationship
of dependent-like verbal behaviors to counselor
style and induced set. J. consult. Psychol. 1962
gs, 172-177. ' ’

Salzinger, K. Experimental manipulation of verbal behavior:
a review. J.,genet. Psychol., 1959:.él: 65-95.

Salzinger, K., & Pisoni, S. Reinforcement of affect re-
sponses of schizophrenics during the clinical in-
terview. J. abnorm. soc. Psychol., 1958, 21, 84-90.

.R.

60

Sears, R. R., Macoby, E., & Levin, H. Patterns of child
rearing. Evanston, 111.: Row,_Peterson, 1957.

Sears, R. R., Whiting, J. W. M., Nowlis, V., & Sears, P. S.
Some child—rearing antecedents of aggression and
dependency in young children. Genet.4psychol.
Monoaru 1953. _4_Z. 135-234.

Shoben, E. J. Psychotherapy as a problem in learning theory.
Psychol. Bull., 1949, 4Q, 366-392.

Skinner, B. F. The generic nature of the concept of stim-
ulus and response. J. genet. Psychol., 1935, lg,
40-6 3 o

Skinner, B. F. Science and human behavior. New York:
Macmillan, 1955.

Skinner, B. F. Verbal behavior. New York: Appleton-Cen-
tury-Crofts, 1957.

Spielberger, C. D., & Levin, S. M. What is learned in ver-
bal conditioning? J. verb. learn. verb. behav.,
1962, l, 125-152.

Stoler, N. Client likability as a variable in the study
of psychotherapy. Ps chiatric Institute Bulletin,
Univ. of Wisconsin, 19 1:.l: 1-9.

Strupp, H. H. An objective comparison of Rogerian and psy-
choanalytic techniques. J. consult. Psychol., 1955,

g, 1-7.

Strupp, H. H. Psychotherapeutic technique, professional
affiliation and experience level. J. consult. Psy-
Ch010 ’ 1955, 2.2, 97-1020

Strupp, H. H. A multidimensional system for analyzing psy-
chothegapeutic techniques. Psychiatry, 1957: 22:
295-50 .

Strupp, H. H. Patientpdoctor relatiOnShips: the psycho-
therapist in the therapeutic process. In A. J.

Bachrach (ed.), Ex erimental foundations of clin-
ical psychology. New York: Basic BooEs, I962,
7-150

Sullivan, P. L., Miller, 0., & Smelser, W. Factors in length
of stay and progress in psychotherapy. J. consult.
P8: ChOle , 1958, a, 1-90

61

Ullman, L. P., Krasner, L., & Collins, B. J. Modification

Waskow,

Winder,

of behavior through verbal conditioning: effects

in group therapy. J. abnorm. soc. Psychol., 1961,
pg, 128-132.

I. E. Reinforcement in a therapy-like situation
through selective responding to feelings or content.

0. L., Ahmad, F. 2., Bandura, A., & Rau, L. C.
Dependency of patients, psychotherapists' responses,

and aspects of psychotherapy. J. consult. Psychol.
1962, gs, 129-134. ’

Wolberg, L. R. The technique of psychotherapy. New York:

Grune and Stratton, 1954.

APPENDIX A

Data Sheet for Therapists

Please write in and tape record the information asked for
in items 1 through 5:

1. Tape SIDE NUMBER:

 

2. 'Client's full NAME:

 

3. Therapist's position: Staff Interns Prac—
ticum

4. DATE of this interview:

 

5. Interview number:

 

The purpose of this questionnaire is to identify
tapes of the first two psychotherapy interviews which you
have with new clients. This excludes only these interviews
which you judge to be primarily educational, vocational or
testing oriented. It may be difficult to specify just which
interview you start therapy in but we would include inter-
views in which you were assessing the client's potential
for therapy by means of "trial therapy" procedures.

If you judge your first interview as a therapy in-
terview, then go on to record the next interview which you
have with the client. If the recorded interview was not a
therapy interview, please hold the tape to record a subse-
quent interview in which you begin therapy or to record
another client.

6. By the above criteria, was this interview one in which
you were primarily engaged in therapy? YES ’ NO

6a. If "YES," was it the First or Second therapy inter-
view? (Circle one) '

6b. If "NO," what was the nature of the interview?

7. Has this client had psychotherapy prior to his seeing
you? YES NO

Please return this sheet with the tape when you have

collected two therapy interview recordings. If you need
the tape for some reason, it will be available.

62

APPENDIX B

Follow-Up Questionnaire

We are inquiring about the status of some of the clients
when you tape recorded for study. Please fill in the ap-
propriate information and return this form to the reception-

ist.

CLIENT'S NAME:

1.

2.

’Has the relationship been terminated?

 

How many hours of therapy have you had with this client?

 

How many hours of "non-therapy" contacts have there
been? What was the nature of those hours?

 

Are you currently seeing him in therapy? '
Seeing him for some other reason? Explain.

 

 

a. If "YES," did the client terminate despite your ex-
plicit or implicit feeling that he needed more ther-
'apy, i.e., did you regard the termination as prema-
ture? Explain the reasons and circumstances
of termination.

 

b. Was the termination due to unavoidable situational
factors? What were they?

 

 

Was there an agreement made by you and the client to
begin therapy? When was the agreement made?
(e. g., 2nd hour)’ If there was no
mutual agreement on therapy, please explain the basis
upon which you saw him.

 

Has the client had psychotherapy prior to seeing you?

 

65

APPENDIX 0

Client Likability Scale

Upon completion of the there interview, i.e., if you
checked "YES" to question 6 on the data sheet attached
to this form, make a rating on the continuum of liking
to disliking of the client.

You may make a mark at any place along the scale: you
are not confined to the points that are numbered.

Scale point (1) is for a positive liking reaction to the
client, while a check at (6) would be a disliking reaction
to the client. Marking any place along the scale between
these two points will represent the magnitude of your lik-
ing or disliking, depending on the closeness to the end

of the scale.

Often it is our experience that we have feelings and reac-

tions to people, but do not necessarily draw our attention

to these feelings. This rating task asks you to look at

the specific liking;pr disliking feeling that this client

brings out in you. Ratings made of the same client after
f erent interviews may, of course, differ.

64

APPENDIX D

Scoring manual

(This manual is a modification of manuals used in the fol-
lowing studies: Winder, 0. L., Ahmad, F. 2., Bandura, A.,
& Rau, L. C. Dependency of patients, psychotherapists'
responses, and aspects of psychotherapy. J. consult.,Psy-
chol., 1962, gg, 129-134; Bandura, A., Lipsher, D. H., &
Miller, P. E. Psychotherapists' approach-avoidance reac-
tions to patients' expressions of hostility. J. conSult.

Psychol., 1960, _2_«_1._, 1—8.)

A. Scoring Unit and Interaction Sequence
1. Definition. A unit is the total verbalization of
one speaker bounded by the preceding and succeeding
speeches of the other speaker with the exception
of interruptions.
There are three types of scoring units: the
"patient statement" (P St), the "there ist response"
(T R), and the "patient response" (P R . A sequence
of these three units composes an "interaction se-
quence." The patient response not only completes
the first interaction sequence but also initiates
the next sequence and thereby becomes a new patient
statement.
Example:
P. T can3t understand how you can stand me.
P St

T. You seem to be very aware of my feelings.
T R

P. T am)always sensitive to your feelings.
P R

2. Pauses. Pauses are not scored as separate units.
The verbalization before and after the pause is
considered one unit. Therapist silences are scored
as prescribed under Part D2e of this manual. There
are no patient silences in this system.

3. Interruptions. Statements of either therapist or
patient which interrupt the other speaker will be
scored only if the content and temporal continuity
of the other speaker is altered by the interruption.
Then, the interrupting verbalization becomes another
unit and is scored. A non-scored interruption is
never taken into account in the continuation of the
other speaker.

65

66

Interruption scored as one unit:
P. I asked him to help me and--

T. Why was that?
P. --he refused to even try.

Non-interruption scored as 3 units, one inter-
action sequence:
P. I asked him to help me and--

T. Why was that?
P. I don't know.

Verbalizations such as "Um hmm" or "I see"
are ignored in scoring unless they are so strongly
stated as to convey more than a listening or
receptive attitude.

Patients' requests for the therapist to
repeat his response are considered interruptions
and are not scored. However, therapists' re-
quests of this sort are scored as units (as
approach or avoidance of the patient statement).

Categories of Patient Statements and Patient Responses

There are three categories: Dependency, Hostility,
and Other. They are scored as exhaustive categories.
All discriminations are made on the basis of what is
explicitly verbalized by the speaker in the unit under
consideration. One statement may be scored for several
categories.

When dependency and/or hostility units occur, the
object of the patient's behavior is also scored as either
psychotherapist or other.

77A coding of self (SI is given if the patient refers

to his own behavior and a coding of other (0) is given
if the client refers to someone else's behavior.

1. Hostility category. The subcategories of hostil-
ity listed below are not differentiated in the
scoring but are listed here to aid in identifica-
tion of hostility.

a. Hostility. Hostility statements include
description or expression of unfavorable,
critical, sarcastic, depreciatory remarks;
oppositional attitudes; antagonism, argument,
expression of dislike, disagreement, resent-
ment, resistance, irritation, annoyance,
anger; expression of aggression and punitive
behavior, and aggressive domination.

1. Anger:

67

P. I'm just plain mad!

P. I just couldn't think--I was so
angry.

P. My uncle was furious at my aunt.

Dislike: expresses dislike or describes
actions which would usually indicate
dislike
P. I just don't get interested in
them and would rather be somewhere
else.

P. I've never ever felt I liked them
and I don't suspect I ever will.

P. He hates editorials.

Resentment: expresses or describes a

persistent negative attitude which does

or might change to anger on a specific

occasion

P. They are so smug; I go cold when-
ever I think about having to listen
to their 'our dog' and 'our son.‘
Boy!

P. -They don't ever do a thing for
me so why should I ask them over?

P. Dad resents her questions.

Antagonism: expresses or describes
antipathy or enmity
P. It's really nothing definite, but
we always seem at odds somehow.

P. There is always this feeling of
being enemies.

Opposition: expresses or describes

oppositional feelings or behavior

P. If he wants to do one thing, I
want to do another.

P. It always seems she is against
things. She is even against things
she wants.

P. No, I don't feel that way (in re-
sponse to T's assertion).

2.

68

Critical attitudes: expresses negative
evaluations or describes actions which
usually imply negative evaluations

If I don't think the actors are
doing very well, I just get up
and walk out.

There is something to be critical
about in almost everything anyone
says or does.

Aggressive actions: acts so as to hurt
another person or persons, either phys-
ically or psychologically

P.

He deserves to suffer and I'm mak-
ing it that way every way I can.

I can remember Mother saying: 'We
slap those little hands to make it
hurt.‘

b. Hostility anxiet . A statement including
expression of fear, anxiety, guilt about
hostility or reflecting difficulty express-
ing hostility ‘

P.

P.
P.

I just felt so sad about our argu-
ment.

I was afraid to hit her.
After I hit her I felt lousy.

c. Hostility acknowledgment or a reement. A

 

 

statement agreeing with or acknowledging

the therapist's approach toward hostility
is scored as further hostility. May give
example.
simply agree with therapist's response.

T.
P.

May convey some conviction or may
You were angry.

Yes!

Dependency categories.
a. De inition. Any explicit expression or de—

scription of help—seeking, approval-seeking,
company-seeking, information-seeking, agree—
ment with others, concern about disapproval,
or request that another initiate discussion
or activity.

b. Scoreable categories: The subcategories
listed below are scored exhaustively.

69

Problem Description: States problem

in coming to therapy, gives reason for

seeking help, expresses a dependent

status or a general concern about de-

pendency

P. I wanted to be more sure of myself.
That's why I came.

P. I wanted to talk over with you
my reasons for dropping out of
school next quarter.

P. Part of the reason I'm here is
that everything's all fouled up
at home.

P. I depend on her, am tied to her.

P. I want to be babied and comforted.

Hela-seekin : Asks for help, reports

 

 

 

asking for help, describes help-seeking
behavior ‘
P. I asked him to help me out in this
situation.

P. What can you do for him?

P. I try to do it when he can see
it's too hard for me.

Approval-seeking: Requests approval

or acceptance, asks if something has

the approval of another, reports having
done so with others, tries to please
another, asks for support or security.
Includes talk about prestige. EXpresses
or describes some activity geared to
meet his need ‘

P. I hope you will tell me if that is

what you want.

P. If there was any homework, I did
it so Dad would know I was study-
ing like a good girl.

P. Is it alright if I talk about my
girl's problem?

P. That's the way I see it, is that
wrong?

P. I asked him if I were doing the
right thing.

7O

Company-seeking: Describes or expresses
a wish to be with peeple, describes
making arrangements to do so, describes
efforts to be with others, talks about
being with others
P. It looks as if it'll be another
lonely weekend.

P. Instead of studying, I go talk
with the guys.

P. I only joined so I could be in a
group.

P. We try to see if other kids we know
are there, before we go in.

Information-seeking: Asks for cogni-

tive, factual or evaluative information,

expresses a desire for information from

others, arranges to be the recipient

of information '

P. I asked him why he thought a girl
might do something like that.

P. I came over here to see about tests
you have to offer. I want to know
what they say.

P. I'm planning to change my major.
I'd like to know how to do it.

Agreement with another: Responds with
ready agreement with others, readily
accepts the therapist's reflection.
Often illustrates therapist's remarks
with examples, draws a parallel example
to indicate agreement. May accept pre-
ceding statement on authority or if
preceding statement was a therapist
approach to Dependency, may simply agree
with it.

P. Oh, yes! You're absolutely right

about that. '

P. Immediately I felt he was right
and I had never thought about it
that way.

T. Then you wanted to get some help?

P. Yes.

71

7. goncern about disapproval: Expresses
fear, concern, or unusual sensitivity
about disapproval of others, describes
unusual distress about an instance of
disapproval, insecurity, or lack of
support. Little or no action is taken
to do something about the concern

P. She didn't ever say a thing but
I kept on wondering what she doesn't
like about me.

P. My parents will be so upSet about
my grades, I don't even want to
go home.

P. It seems like I always expect I
won't be liked.

P. I can't understand how you can
stand me when I smoke.

P. I'm sorry I got angry at you.

8. Initiative-seeking: Asks the therapist
or others to initiate action, take the
responsibility for starting something
(to start discussion, determine the
topic). Arranges to be a recipient
of T's initiative. May solicit sugges-
tions

P. Why don't you say what we should
talk about now?

P. If you think I should keep on a
more definite track, you should
tell me.

P. I got my advisor to pick my courses
for next term.

P. Tell him what to do in these cir-
cumstances.

3. Other cats or . Includes all content of patient's
verSalIzations not classified above
0. Categories of Therapist Responses
Therapist responses to each scored patient statement
are divided first into two mutually exclusive classes,
approach and avoidance responses. When both approach

and avoidance are present, score only the portion which
is designed to elicit a response from the patient.

72

Approach responses. The following subcategories
are exhaustive. An approach response is any ver-
balization by the therapist which seems designed
to elicit from the patient further expression
or elaboration of the Dependent or Hostile (or
Other) feelings, attitudes, or actions described
or expressed in the patient' 8 immediately preced—
ing statement, i.e., the pa of the preceding
statement which determined its placement under
Dependency, Hostility or Other. Approach is to
the major category, not specific subcategories.
a. Approval: Expresses approval of or agree-
ment with the patient's feelings, attitudes,
or behavior. Includes especially strong
"Mm-hmm!", "Yes"

P. May I just be quiet for a moment?

T. Certainly.

P. I have my girlfriend's problems on my
mind. Could we talk about them?

T. Why don't we talk about that?

b. Exploration(probing): Includes remarks
or questions that encourage the patient to
describe or express his feelings, attitudes,
or actions further, asks for further clar-
ification, elaboration, descriptive infor-
mation, calls for details or examples. Should
demandnmore than a yes or no answer; if not,
may be a "label"

P. How do I feel? I feel idiotic.
T. What do you mean, you feel idiotic?

P. I can't understand his behavior.
T. What is it about his behavior you can't
understand?

c. Reflection: Repeats or restates a portion
of the patient's verbalization of feeling,
attitude, or action. May use phrases of
synonymous meaning. Therapist may sometimes
agree with his own previous responSe; if the
client had agreed or accepted the first ther-
apist statement, the second therapist state-
ment is scored as a reflection of the client
statement.

P. I wanted to spend the entire day with
him.
T. You wanted to be together.

d.

‘73

P. His doing that stupid doodling upsets
me. .
T. It really gets under your skin.

Labeling: The therapist gives a name to the
feeling, attitude, or action contained in

the patient's verbalization. May be a tenta-
tive and broad statement not clearly aimed

at exploration. Includes "bare" interpreta-
tion, i.e., those not explained to the pa-
tient. May be a question easily answered

by yes or no

P. I just don't want to talk about that
any more. ’
T. What I said annoyed you.

P. She told me never to come back and I
really did have a reaction.

T. You had some strong feelings about that--
maybe disappointment or anger.

Inteppretation: Points out and explains
patterns or relationships in the patient's
feelings, attitudes, and behavior: explains
the antecedents of them, shows the similari-
ties in the patient's feelings and reactions
in diverse situations or at separate times

P. I had to know if Barb thought what I
said was right.

T. This is what you said earlier about
your mother . . .

Generalization: Points out that patient's

feelings are natural or common

P. I want to know how I did on those tests.
T. Most students are anxious to know as
soon as possible.

P. Won't you give me the scores?
T. Many students are upset when we can't.

g. Support: Expresses sympathy, reassurance,

or un erstanding of patient's feelings.

P. It's hard for me to just start talking.
T. I think I know what you mean.

P. I hate to ask favors from people.
T. I can understand that would be diffi—
cult for you.

74

h. Factual Information: Gives information to
direct or implied questions. Includes gen-
eral remarks about the counseling procedure.

P. Shall I take tests?
T. I feel in this instance tests are not
needed.

P. What's counseling all about?
T. It's a chance for a person to say just
what's on his mind.

Avoidance responses. The following subcategories
are exhaustive. An avoidance response is any
verbalization by the therapist which seems designed
to inhibit, discourage, or divert further expres-
sion of the Dependent, Hostile, or Other patient
categories. The therapist attempts to inhibit

the feelings, attitudes, or behavior deacribed

or expressed in the immediately preceding patient
statement, i.e., the part of the preceding state-
ment which determined its placement under Depend-
ency, Hostility, or Other. Avoidance is avoidance
of the major category, not specific subcategories.

a. Disapproval: Therapist is critical, sarcas-
tic, or antagonistic toward the patient or
his statements, feelings, or attitudes, ex-
pressing rejection in some way. May point
out contradictions or challenge statements

P. Why dAn't you make statements? Make
a statement. Don't ask another question.
T. It seems that you came here for a reason.

P. Well, I wonder what I do now?‘

T. What do you think are the possibilities?
You seem to have raised a number of log-
ical possibilities in our discussion.

P. I'm mad at him: that's how I feel.
T. You aren't thinking of how she may feel.

b. Topic Transition: Therapist changes or intro-
duces a new topic of discussion not in the
immediately preceding patient verbalization.
Usually fails to acknowledge even a minor
portion of the statement

P. Those kids were asking too much. It
would have taken too much of my time.

T. We seem to have gotten away from what
we were talking about earlier.

C.

d.

75

P. My mother never seemed interested in
me.

T. And what does your father do for a
living?

Ignoripg: Therapist responds only to a
minor part of the patient response or re-
sponds to content, ignoring affect. May
under- or over-estimate affect. May ap-
proach the general topic but blatently ig-
nore the affect verbalized

P. You've been through this with other
peeple so help me out, will you.
T. You are a little uneasy. ‘

P. You can see I don't know what to do
and I want you to give me advice.

T. Just say whatever you feel is import-
ant about that.

P. My older sister gets me so mad I could
scream.
T. Mm-hmm. How old did you say she was?

Mislabeling: Therapist names attitudes,
feelings, or actions which are not present
in the actual verbalization preceding the
response

P. I just felt crushed when she said that.
T. Really burned you up, huh?

P. I don't know how I felt--confused, lost--
T. I wonder if what you felt was resentment.

Silence: Scored when it is apparent that the
patient expects a response from the therapist
but none is forthcoming within 5 seconds after
the patient staps talking. If the therapist
approaches after 5 seconds have elapsed,
silence cannot be scored and the therapist's
response is merely "delayed"

P. If you think I should keep on a more
definite track, tell me because I'm
just rambling.

T. (5 second silence)

P. It is very confusing to know what to
do.

76

Dependency and Hostility initiated by therapist:
Scored whenever the therapist introduces the topic
of Dependency or Hostility, i.e., when the patient
statement was not scored as the category which

the therapist attempts to introduce

P.
T.

P.
T.

P.
T.

Last week I talked about Jane.
You've mentioned a number of things
you have done to please her.

(Enters office)
Now, how may I help you?

I was late for class this morning.
I wonder if you dislike the teacher
or the class?

I like to run around in blue jeans.
You hate your mother.

J “I“ ‘2»

F ’t" 9‘ "7

-‘--—'III) it"‘f
1'". 4':

'34va i886

i... “5““ ,5.

:.~' I?

{In-$[I I
13w .1