Citation
Verbal behavior in group psychotherapy

Material Information

Title:
Verbal behavior in group psychotherapy
Creator:
Daniels, Aubrey Clise, 1935- ( Dissertant )
Schumacher, Audrey ( Thesis advisor )
Barger, Benjamin ( Reviewer )
Barry, John R. ( Reviewer )
Wolking, William D. ( Reviewer )
Hanna, Thomas ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1965
Language:
English
Physical Description:
vi, 87 leaves. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Adjectives ( jstor )
Anxiety ( jstor )
Clinical psychology ( jstor )
Correlations ( jstor )
Data collection ( jstor )
Gene therapy ( jstor )
Group psychotherapy ( jstor )
Observational research ( jstor )
Psychotherapy ( jstor )
Rate bases ( jstor )
Behaviorism (Psychology) ( lcsh )
Dissertations, Academic -- Psychology -- UF ( lcsh )
Group psychotherapy ( lcsh )
Psychology thesis Ph. D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis -- University of FLorida.
Bibliography:
Bibliography: leaves 84-86.
Original Version:
Manuscript copy.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
022567568 ( AlephBibNum )
13868303 ( OCLC )
ADA1835 ( NOTIS )

Downloads

This item has the following downloads:


Full Text






VERBAL BEHAVIOR IN GROUP

PSYCHOTHERAPY




















By
AUBREY CLISE DANIELS









A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA
December, 1965
















































3 1262 052 3875

3 1262 08552 3875












ACKNOWLEDGMENTS


The author wishes to express his gratitude to the members of the

supervisory committee for their assistance in all phases of this study:

Dr. Audrey Schumacher, Chairman, Dr. Benjamin Barger, Dr. John R. Barry,

and Dr. William D. Wolking of the Department of Psychology, and

Dr. Thomas Hanna of the Department of Philosophy. He also wishes to

thank Drs. Hugh Davis and Nathan Perry, the group psychotherapy super-

visors, without whose cooperation and assistance this study would have

been impossible. Appreciation is also extended to Dr. Thomas Weaver,

Mr. Ted Keiser, and Mr. Wayne Richard, who were the group therapists.

Special appreciation is extended to Dr. Schumacher who made a

difficult job much easier than it would have been otherwise through her

knowledge and understanding of the problem, and through her emotional

support and encouragement.

Dr. Barger's working knowledge of this kind of research and his

many practical suggestions were very helpful throughout this study.

A special word of appreciation is accorded Dr. Nathan Perry who

inspired the present study, gave invaluable assistance in the design

phase of the study, and rendered a considerable amount of technical

assistance.

Finally, thanks is due to my wife, Becky, who typed the manuscript

and whose help in many ways made the completion of this study possible.












TABLE OF CONTENTS


ACKNOWLEDGMENTS . . . . . . . . . . . .

LIST OF TABLES . . . . . .

LIST OF FIGURES . . . . . . . . . . . . .

CHAPTER

I INTRODUCTION . . . . . . . . . . .

The Present Study . . . . . . . . .

2 METHOD . . . . . . . . . . . . .

Subjects . . . . . . . . ... ..
Groups . . . . . . . . . . .
Apparatus .. ...... ......................
Measures of Patient Mood and Perceptions .. . ..
Group Perception Inventory (GPI) .. . . ....
Procedure .. . .......................
Operational Definition of Terms . . . . . .
Statistical Treatment of the Data .. . .....

3 RESULTS . . . . . . . . . . . .

4 DISCUSSION . . . . . . . . . . . .

Group Equilibrium ....................
Effect of Therapists' Verbal Behavior on
Group Equilibrium . . . . . . . .
Changes in Mood .. .. . . . .. ......
Patients' Verbal Participation and Their Perceptions
of Self, the Group, and the Therapists .. ..
Summary .. .. ..........................
Suggestions and Implications for Furture Research .

5 SUMMARY . . . . . . . . . . . .

6 APPENDICES . . . . . . . .. . .






Page

APPENDIX A. MAACL. . . .. ........... . 63

APPENDIX B. GPI. ................ . . 65

APPENDIX C. Average Frequency of Verbal Responses Per
Minute (FVR) in Five-Minute Intervals. ... . 67

APPENDIX D. Average Duration (in seconds) of a Verbal
Response in Five-Minute Intervals . ... 69

APPENDIX E. Frequency of Silences Occurring During
Five Periods of Each Session . . .. 71

APPENDIX F. Average Duration of Silences for Five
Periods of Each Session. . . . . . ... 72

APPENDIX G. Frequency of Interruptions Occurring
During Five Periods of Each Session . . ... 73

APPENDIX H. Percentage of Session Time Spent
Talking . . . . . . .. .. . 74

APPENDIX I. Frequency of Verbal Responses of Patients and
Therapists for Eight Consecutive Sessions. . 76

APPENDIX J. Total Duration (in seconds) of Verbal
Responses of Patients and Therapists . . .. 77

APPENDIX K. MAACL Difference Scores for Anxiety,
Depression, and Hostility. . . . . ... 78

APPENDIX L. List of Rank Order Correlations Relating to
Hypothesis V . . . . . . . .. . 80

REFERENCES ............... . . . . . ... 84

BIOGRAPHICAL SKETCH. ...... ..... ..... . ... . 87












LIST OF TABLES


able Page

1 Composition of Groups: Age, Sex, Education, and
Diagnosis . . . . . .. . . . . . . 17

2 List of Rank Order Correlations Relating to Hypothesis V . 80

3 Rank Order Correlations Between Patients' Perceptions of
Amount of Help They Received (GPI, Item 1) and Frequency
of Their Verbal Responses. . . . . . . . ... 80

4 Rank Order Correlations Between Patients' Perceptions of
Amount They Contributed to the Group (GPI, Item 3) and
the Total Duration of Their Verbal Responses . . . . 81

5 Rank Order Correlations Between Patients' Perceptions of
Amount They Contributed to the Group (GPI, Item 3) and
the Frequency of Their Verbal Responses. . . . . . 81

6 Rank Order Correlations Between Patients' Perceptions of
Effectiveness in Handling Their Problems (GPI, Item 9)
and Total Duration of Their Verbal Responses . . ... 82

7 Rank Order Correlations Between Patients' Perceptions of
Effectiveness in Handling Their Problems (GPI, Item 9)
and Frequency of Their Verbal Responses. . . . . ... 82

8 Rank Order Correlations Between Group Rankings of
"Benefit" (Gpi, Item 8) and Rankings of Patients on
Both Frequency and Duration of Verbal Responses. . . . 83











LIST OF FIGURES


Figure Page


1 Data Collection Sequence. .. . . . . . . . 24

2 Experimental Session #1 (El). . . . . . . .. 24

3 Experimental Session #2 (E2). . . . . . . ... 25

4 Average Frequency of Verbal Responses Per Minute (FVR)
Across Six Sessions for Group 1 and Group 2 . . ... 32

5 Average Duration of Verbal Responses Per Minute (DVR)
Across Six Sessions for Group 1 and Group 2 . . ... 33

6 Average Frequency of Silences Per Session for Group 1
and Group 2 . . . . . . . . ... ... .. 35

7 Average Duration of Silences Per Session for Group 1
and Group 2 ........ ... . ... .... 35

8 Mean MAACL Change from Pre-Group to Post-Group Testing on
Anxiety, Depression, and Hostility in Group 1 ...... 39

9 Mean MAACL Change from Pre-Group to Post-Group Testing on
Anxiety, Depression, and Hostility in Group 2 ...... 40












CHAPTER 1

INTRODUCTION


The efficacy of group psychotherapy in bringing about personality

change is relatively unquestioned by practicing therapists. There are

reports of its success in the treatment of almost every conceivable

patient group. Studies quoted in a review article by Wirt and Wirt

(1963) indicate that group psychotherapy has been conducted with many

widely divergent groups such as the blind, aphasics, and children and

adults with psychosomatic disorders. Studies using all of these groups

report positive results. In spite of this, the meaning of verbal

behavior to the group psychotherapeutic process is still quite vague

although it is probably one of the most critical variables. The

purpose of this study was to investigate two aspects of verbal behavior

in group psychotherapy. First, the function of the therapist's verbal

behavior in the group was investigated. Secondly, the relationship

of verbal behavior and patient's perceptions of the group was studied.

Group therapy in the United States is generally regarded as

beginning with Pratt's classes (1907) for tubercular patients. Little

research was done with group therapy, however, until 1940. A review by

Dreikurs and Corsini (1954) indicated that just over a hundred papers

were published from 1910 to 1930, whereas over five hundred were

published in the period from 1940 to 1949. Since 1949, increased interest

in the use of group therapy has more than doubled the number of published

papers.







The tremendous growth in the application of group therapy in the

past few years is no doubt due to a variety of reasons, not the least

of which are (1) the treatment is relatively inexpensive; (2) pro-

fessional manpower requirements are less for the number of patients

treated; and (3) group therapy represents a development in the continuing

search for new and more effective methods of treatment. Wirt and Wirt

(1963) pointed out that a primary reason for the current enthusiasm for

group psychotherapy is that its effectiveness has been demonstrated to

the point that it is being advocated more and more as the treatment of

choice, apart from any economic or manpower considerations.

Research on group psychotherapy, however, has not kept pace with

its rapid growth as a treatment method. Although more is being

written about group therapy than ever before, there is little experi-

mental literature on the subject. Wirt and Wirt (1963), in their

review article, saw most of the group psychotherapy literature as

"descriptions of therapists' experiences or recommendations." One of

the obvious reasons for the paucity of experimental data on group

psychotherapy is the extreme complexity of the process. The many

variables operating in individual psychotherapy are multiple several

times over in group psychotherapy. After reviewing the literature,

McFarland, Daniels, and Lieberman (1964) gave a much more basic reason

as to why research in group psychotherapy has not gone beyond "descriptions

of the therapists' experiences or recommendations." They state:

Lack of sufficient descriptive material, or a
language in which to insert these descriptions,
prevents the development of a fund of knowledge
based on the collective efforts of many investi-
gators. All too frequently the inventiveness of
one investigator is ineffective in making a
contribution outside the limited confines of his








own institution because others cannot, with
any certainty, duplicate his efforts.

At this time it appears that the most fruitful approach to

research in this area is to study those variables that are easily

defined, fairly stable, and reliably measured. A cluster of variables

that meets all of these criteria is found in verbal behavior.

Matarazzo (1962) has conducted several studies of changes in

interviewee speech as a function of changes in interviewer speech.

He has concerned himself with a number of non-content verbal interaction

variables and has reported high reliabilities in recording and scoring.

Kew, Rickard, and Timmons (1962) reported that group verbal

behavior was quite stable. They found that a hierarchy of verbal re-

sponding was formed in the group which persisted over time. They

suggest that because of this stability verbal measures could be used

as a dependent variable.

Saslow and Matarazzo (1959), in a study on the stability of non-

content verbal interaction variables, e.g., frequency and duration of

verbal responses, etc., found that patients who showed the most change

on these measures, over an eight-month period, had significantly more

hours in psychotherapy than those who showed little change.

Although verbal behavior seems to be a critical and much-discussed

variable in the group psychotherapeutic process, it is probably one of

the least understood. For example, the relationship between the level

of verbal activity of group members and the benefit derived from it is

not well defined. In a study relating change in group psychotherapy to

degree of verbal participation in the group, Smith, Bassin, and Froelich

(1960) concluded:








The most articulate member of the group is not
necessarily the person who gains the most, nor
is the quiet one in the group the individual
who does not achieve important attitudinal
alterations.

However, in a study by Sechrest and Barger (1961), patients per-

ceived as more beneficial those sessions in which they participated,

generally, at what was for them a relatively high level. This study

suggests a linear relationship between verbal participation and benefit

received, whereas Smith, et al. (1960) suggest a relationship between

these variables which is curvilinear. A curvilinear relationship is

also suggested by Cook (1964) who found, in a study of silence in

individual psychotherapy, that a relative lack of silence in the session

characterized the unsuccessful cases, whereas a higher percentage of

silence tended to characterize the successful ones. Successful cases

had silences occurring from 4 to 20 percent of the time, whereas un-

successful ones either had less than 4 percent silence or more than

20 percent. This seems to indicate that there is an optimal amount of

verbal participation in psychotherapy. He goes on to say that silence

seemed to be an index of the therapist's behavior and reflects the

"climate" of the interview. Although this inference is probably a bit

premature in terms of the limits of his data, it does suggest the

possibility that non-content variables can lead to a better understanding

of more psychodynamic concepts.

Some indirect support for Cook's hypothesis is found in a study

by Anderson (1960). He found that an interviewer talked more with job

applicants he accepted than those he rejected. He also found more

silences in the interviews with rejected applicants than in the inter-

views of applicants who were accepted.







It is a commonly held conviction that a patient will derive more

from group psychotherapy if he participates in it verbally. It is

implicit in many studies, and in the theories of many group psycho-

therapists, that a high degree of patient-patient interaction is

desirable. Yet, it also seems that it would be desirable for the

therapist to be able to control the group activity level when appro-

priate. For example, there may be times in a group when a silence may

be very threatening and the therapist may want to end it, or at other

times he may want to reduce the amount of interaction. Practically

every textbook on group psychotherapy discusses the desirability of

maintaining control of group action. To quote Powdermaker and Frank (1953):

It would seem that the doctor's aim in the initial
stages of therapy should be to sense when and how
to give the amount of direction and support that
each situation demands. Too little guidance often
results in intense competition for superiority,
which only upsets the patients without leading to
anything beneficial. Too much guidance, on the
other hand, may inhibit the appearance of the
neurotic patterns, the analysis of which is part
of the objective of psychotherapy.

However, as stated earlier, little is known about the meaning of the

effect of activity level in general, and less about the effect of the

therapists' verbal activity level in psychotherapy.

Although several authors have investigated the meaning of the

therapist's and interviewer's verbal activity level in a two-person

interview situation or individual psychotherapy relationship, very

little has been done along these lines with psychotherapy groups.

However, the findings in individual psychotherapy may have application

to group psychotherapy.

Matarazzo, Weins, and Saslow (1964) have found highly consistent

changes in interviewee verbal behavior by varying the duration of the








interviewer's utterances. Their results showed that the length of

interviewee utterances varied positively with the length of therapist's

utterances. They have also found that silence behavior by the interviewer

causes a decrease in frequency and duration of verbalizations by the

interviewee.

In an investigation on the effects of individual differences between

interviewers on the interaction of patients in individual psychotherapy,

Goldman-Eisler (1952) found that each of three psychiatrists had his

own individual reaction pattern regardless of the type patient he was

interviewing. Further, she found that they influenced patients'

interaction differently, i.e., the depressed patients "responded best"

to active stimulation and the active patients to the passive interview.

Lennard and Bernstein (1960) related therapists' verbal activity

level to signs of strain in individual therapy. They found that there

were fewer signs of strain in the patients of verbally active therapists

as compared to patients of verbally passive therapists. Another

interesting finding was that therapists compensate for lower patient

verbal output by increasing their own verbal output, and for higher

patient output by reducing their output. They suggest that the therapist

attempts to restore the therapy session to within, what is for him,

normal limits of verbal activity by varying his level of output. However,

"theory oriented" therapists seemed to maintain equilibrium by establishing

a constant proportion of output from session to session.

Matarazzo, Weins, and Saslow (1964) found, in a study of therapist-

patient verbal interaction, that with two different patients, one

therapist, by varying his own verbal participation, maintained a level of

verbal activity of 93 percent with each patient. The other therapist







studied showed a constant rate of verbal participation of approximately

19 percent across 12 psychotherapy sessions regardless of the patient

output. These results are clearly in line with those cited above of

Lennard and Bernstein (1960).

In a study of content-free verbal behavior in contrived psycho-

therapy groups, Timmons, Rickard, and Taylor (1960) found that

experimenters can become nearly equivalent stimuli to the patients by

following simple content-free rules. The frequency and duration of a

patient's verbal activity was not affected by the presence of different

experimenters.

Salzberg (1962), in a study of the effect of therapist activity

level and direction of action on group psychotherapy patients, found

that therapist's silence led to significantly more interaction (patient-

patient conversation) than did talking by the therapist. However,

methodological limitations of Salzberg's study restrict the generali-

zation of these results to other psychotherapy groups.

The main limitation in Salzberg's study is that the four experi-

mental treatments were not distributed throughout all sessions. The

four experimental treatments were (1) silence-redirecting in which the

therapist was silent but when asked a question would mention another

patient in answering; (2) talking-direct in which the therapist spoke

whenever conversation lagged, speaking directly to a patient, and not

referring to any other members; (3) silence-direct in which the therapist

was silent but when asked a question would speak directly to a patient;

and (4) talking-redirect in which the therapist spoke whenever the

conversation lagged, but when asked a question would mention another

patient in answering. Only combinations of silence-redirecting and







talking-direct were used in the first ten sessions. In the last ten

sessions only silence-direct and talking-redirect combinations were

used. There was a significant increase in the amount of verbal

activity by the group from the first ten to the last ten sessions,

but one does not know whether this is due to the experimental conditions

or to the possibility that patients talk more in later sessions than

earlier ones. Other problems in design are also left unsolved. If

the study is examined in terms of total verbal activity it can be

shown that inclusion of the therapists' responses in the group total

led to significantly more verbal activity for sessions in which the

therapist was talking than for sessions in which he was silent. If one

considers the fact that during the talking-direct condition the

responses of the group averaged only six seconds each, it would appear

verbal activity may have reached a saturation point. In addition, if

one takes into account the fact that the therapist averaged 99.1

responses per half hour, it is obvious that group interaction will be

reduced. If the therapist spends approximately one-third of the time

talking directly to group members and they spend approximately one-

third of the time responding to him, only a limited amount of time is

left for group interaction.

Several studies have attempted to get at the meaning of verbal

activity to the patient.

Matarazzo, Saslow, Weins, Weitman, and Allen (1964) found that

headnodding by the therapist resulted in increases in interviewee

speech duration. This, coupled with the previous findings that

interviewers' speech duration caused increased interviewee speech

duration, led the authors to speculate that the increases are due to







more subject satisfaction when the interviewer is more active.

The study cited earlier by Sechrest and Barger (1961) indicates

that the amount of help the patient feels he received, the contribution

the patient feels he made to the session, and the perceived relevance

of a session are related to the patients' verbal participation in

the session.

Although not working with therapy groups, Knutson (1963) found

that quiet and vocal groups had differences in degree of satisfaction

with their work, in initial progress, and in the quality of their

work. The vocal groups tended to express greater satisfaction with

their productionparticipation,and leadership than the quiet groups.

On the other hand, the quality of the work (preparing a public health

pamphlet) was superior in the quiet group. At least initially, verbal

fluency seemed to be the sole identifying mark of leadership. They also

pointed out that successful leadership in these two types of groups

(quiet and vocal) seemed to require different characteristics.

In Lennard and Bernstein's study it was found that patients

rated those sessions in which there was a high verbal output by the

therapists as proceeding "more easily." Patients indicated dis-

satisfaction with the sessions in which there was a low verbal output

by therapists.

All of these studies demonstrate the utility of the non-content

aspects of verbal behavior in gaining more insight and knowledge into

the therapeutic process. Although a body of knowledge is being built

up and theories proposed about these variables in individual psycho-

therapy and non-therapy two-person interactions and groups, research

at this level in group psychotherapy has only recently begun. Therefore,




10

it seems that an extension of some of the theories used in individual

psychotherapy and non-therapy two-person interaction and non-therapy

groups to group psychotherapy would be particularly profitable.

The equilibrium theory of Chapple and Coon (1942) appears to be

singularly useful in the research of the effects of verbal participation

on the group process and, more particularly, the effects of changes in

the therapists' verbalizations upon the verbal behavior of the group.

Concerned almost exclusively with the verbal interaction of individuals

and groups, the theory states that a group is in equilibrium if its

interaction rates are constant within clearly defined limits, and if,

after a disturbance takes place, the rates return to their previous

values. The longer a state of equilibrium is maintained the more

stable this state becomes.

The process of adjustment which takes place within a group after

a disturbance is marked by cyclical fluctuations on at least five

measures: (1) frequency of interaction--number of times interactions take

place; (2) amount of interaction--length of time spent interacting; (3)

origin-response ratio--proportion of actions originated to the number re-

sponded to; (4) rhythm of interaction rate--characteristic relationship of

periods of action and silence; (5) degree of synchronization or adjust-

ment in interaction--degree to which interruptions and failures to

respond occur. When the equilibrium is upset, members may interrupt each

other often, may become angry and unable to work together, or may let

many silences occur. As the group gradually regains its equilibrium the

range of fluctuations on these variables decreases and the rhythm of

routine interactions becomes more constant.

The Present Study

The purpose of this study, then, is to investigate two aspects of





11

verbal behavior in the group psychotherapeutic process. Two major aspects

of its meaning will be studied. First, the major purpose of this study

will be to explore the effects of variations in the therapists' verbal

activity on the verbal activity in an ongoing psychotherapy group.

Secondly, this study will seek to determine the extent to which verbal

participation affects the patient's perception of himself, the therapist,

and other members of the group.

As discussed earlier, Lennaro and Bernstein (1960) and Matarazzo,

Weins, and Saslow (1964) suggested that verbal interaction in the

psychotherapy session operates on an equilibrium model and that the

therapist's verbal activity is an important factor in maintaining the

equilibrium of the system. Chappel and Coon's theory provides the

conceptual tools for testing whether the therapists' verbal partici-

pation serves the same function in group psychotherapy that it does in

individual psychotherapy.

One concern of this study is that the therapist in group psycho-

therapy provides a major stabilizing role primarily through his verbal

behavior. Patients are typically people who lack the capacity for

effective interaction or who are unable to adapt to new situations

readily without help. Therefore, the patients in group psychotherapy may

often find themselves in some difficulty interacting with other patients

and lack the ability to get themselves out of it. It generally falls

upon the therapist to remedy such situations.

If the therapist is limited in this hypothesized role, the group

may show marked changes in the normal pattern of verbal interaction. As

the group equilibrium becomes disrupted the equilibrium of each of the

members will tend to be upset, causing affective changes in the members

which will lead to further disruption.








When the therapist resumes normal interaction the group should

return to its normal state. The recovery of the normal interactional

pattern would seem to be partly a function of the duration of the dis-

ruptive conditions. If such conditions were to prevail long enough the

group would either stabilize at a new level or break up completely.

It is further hypothesized that group members' satisfaction with

the group is a function, not only of their own level of verbal partici-

pation, but also of the other group members' verbal level and the level

of therapist activity. That is, they tend to judge groups in which the

verbal activity level is high as more satisfactory than groups with a

lower level of verbal output. Further, group members may feel that the

relevance of group discussion and the amount the therapist contributes

is a function of the amount of verbal participation of the therapist.

All of this is based on the assumption that group members expect to

talk about their own problems, hear others' problems, and expect the

therapist to contribute toward an understanding and solution of their

problems. Therefore, they should see themselves, the group, and the

therapist as functioning more in line with their purpose when verbal

output is higher.

More specifically, the following hypotheses are advanced:

Hypothesis I

When the therapists' verbal participation is reduced
by planned periods of silence, the equilibrium of
the group will be disrupted to the extent that sig-
nificant changes will occur on the following non-
content verbal interaction variables:

(a) Frequency of verbal responses
(b) Duration of verbal responses
(c) Frequency of silences
(d) Duration of silences
(e) Frequency of interruptions








Hypothesis Ii

When the therapists' verbal participation is reduced
by planned periods of silences, the longer the periods
of silence the more disruptive they will be to the
equilibrium of the group.

Hypothesis III

When the therapists' verbal participation returns to
normal following sessions in which their verbal
participation has been restricted, the equilibrium of
the group will be re-established at a level more
similar to the base rate level than to the level of
the sessions in which the therapists' verbal partici-
pation has been restricted.

Hypothesis IV

When the therapists' verbal participation is reduced by
planned periods of silence, significant increases will
occur on measures of patient anxiety, depression, and
hostility in the group.

Hypothesis V

Patients' perceptions of the amount of help received
from the group, the amount they contributed to the
group, and their own overall effectiveness in handling
their problems will vary positively with changes in
the amount of their own verbal responses.

Hypothesis VI

Patients' perceptions of the amount of benefit each
patient received from the group will be a function
of the amount of each patient's verbal participation.

Hypothesis VII

Patients' perceptions of the amount of help the group
received and the amount the group contributed will
vary positively with the amount of the group's
verbal activity.

Hypothesis VIII

Patients' perceptions of the therapists' contribution
to the group and the degree to which the therapists kept
the group discussion on relevant problems will vary
positively with the amount of therapists' verbal
participation in the group.












CHAPTER 2

METHOD

,e general design of this study involved ar. analysis of tne effect

of planned charges in the therapists' verbal activity (,pecified periods

of silence) on verbal participation and mood in two ongoing psycho-

therapy groups. The relationships between the patient's verbal

participation in the group and his perceptions of himself, the therapists,

and other group members, were also investigated. Since the hypotheses

were made without reference to type of group, e.g., types of patients in

the group, number of patients in the group, etc., the two groups were used

to explore, to a limited degree, the extent of the generalizabiiity of

the hypotheses across groups. No predictions were made about the

differences between the groups and consequently there were no planned

comparisons of the two groups.

In order to evaluate the effect of the therapists' activity on

verbal participation and mood, six consecutive sessions in each group

were used. The first two were used to establish base rates of respond-

ing. In the third session the therapists were silent for two periods

of 20 minutes each, and in the fourth session were silent for a

continuous period of 55 minutes. Each silent period was preceded and

followed by a 15-minute period of normal interaction. The fifth and

sixth sessions were employed to evaluate tne effects of the experimental

sessions on succeeding sessions. Adjective checklists were filled

out prior to and at the conclusion of each session and the difference

scores were used to evaluate the effect of the experimental sessions

14







on mood.


In order to investigate the relationship between verbal partici-

pation and the patient's perception of himself, other group members,

and the therapist, the six sessions described above,plus the two

proceeding sessions, were used for the analysis. At the end of each

session each group member filled out a 9-item inventory, designed

to measure his perceptions of several aspects of the session, e.g.,

benefit received, amount contributed, etc. The responses to these

items were correlated with individual, group, and therapists' verbal

activity measures.


Subjects

The subjects for this study were the members of two ongoing

outpatient psychotherapy groups in the Psychiatric Outpatient Clinic,

Teaching Hospital, J. Hillis Miller Health Center, University of

Florida. The patients of each group were heterogeneous with respect

to age, sex, and diagnosis (see Table I).


Groups

The groups met weekly in 90-minute sessions. At the time of

this study Group 1 had been meeting for three months and Group 2 for

seven months. Membership had been quite stable. Only two patients

had been added since the groups were formed. No members were added to

or dropped out of either group during the course of the data collection.

Both groups had co-therapists who were clinical psychology interns

in the Department of Clinical Psychology, College of Health Related

Professions. One intern served as co-therapist in both groups.

Average data from the two base rate sessions indicated that the







groups were quite different on most of the non-content measures of

verbal interaction. The mean frequency of verbal responses per minute

(FVR) in Group 1 was 8.36. In Group 2 the mean FVR was 5.82. Group 1

had a mean frequency of silences of 13.0 per session, and Group 2 had a

mean of 34.5 per session. Interruption occurred at the mean rate of

23 per session in Group 1, and Group 2 had a mean of only one interrup-

tion per session.

In the base rate sessions therapists in Group 1 had an average

of 161.5 verbal responses per session and talked for a total duration

of speaking of 13.6 minutes. In Group 2, for the same sessions, the

therapists averaged 176.0 verbal responses per session and a total

duration of speaking of 19.8 minutes per session.

Since there were only two base rate sessions, no adequate statistical

tests could be made on the differences between the groups. However,

across eight sessions for which data were available, Group I was sig-

nificantly lower than Group 2 on FVR and interruptions. There was also

a tendency toward significance between the two groups on the frequency

of silences in these sessions.


Apparatus

The group meetings were held in an observation room equipped with

a one-way vision glass and ceiling microphones. A loudspeaker in the

adjoining room allowed observers to hear as well as see the group

sessions. All patients were aware that they were being observed from

the time the groups were formed. The observers were staff group

therapy supervisors and interns of the Department of Clinical Psychology.

The group therapy room held eight patients and two therapists

















0
0





c .

0 10 0


0 U 0I












9 09 0 0 0 0



- C 0 AC 0


0


o






0- 0U 0- .
C > C-










0 0- 0 >9
-o o .. a ,






U 0n u 0 0-


O 0 09 -009 0

0 1 Sa
Enuum


Lu










L420 CCCC20 0 )







- C ONC-7 ,flC
CCCC'.C'CC '











- ( I Lr' S








comfortably. The patients and therapists sat in a circle. No special

seating order was maintained although some patients seemed to prefer

certain locations. The therapists were generally the last ones to

enter the room and occupied any available seat.

An Easterline-Argus Series "S", 20 channel, LabGraph Event Recorder

was used to record the verbal participation measures of frequency and

duration. This is a portable ink-pen type recorder which has a

variable speed drive, in recording the durations of verbal responses

the speed was set at six inches per minute. This allowed the duration

to be measured to the nearest one-half second. The event recorder was

operated by a keyboard arrangement which permitted each patient to be

assigned a separate channel, in order to record the interaction, the

key for the patient speaking was pressed when he began talking and

released when he stopped talking.

Since there were never more than ten people in a group, including

the therapists, the 20 channels of the event recorder made possible

simultaneous recording of the same data by two operators. This permitted

a reliability estimate for the recording of duration of verbal responses

to be obtained. Samples were taken from the beginning, middle, and end

of the first session on which simultaneous recordings were available,

covering a period of 40 minutes and 140 responses. A Pearson Product-

Moment correlation for duration of responses was r=.994. This reliability

of measurement was obtained with minimal instruction and training.

Since the measurements were found to be highly reliable, the double

recording was discontinued.








Measures of Patient Mood and Perceptions

Multiple Affect Adjective Check List (MAACL).--This checklist was

used to determine affect changes as a result of events and behavior

taking place during the therapy session. The MAACL (Appendix A),

constructed by Zuckerman, Lubin, Vogel, and Valerius (1964), is a

132-item adjective checklist which measures immediate, as well as cay

to day, changes in three moods or affects: anxiety, depression, and

aggression. Pre- and post-measures were obtained for each session

during the course of the experiment.

The MAACL was chosen as the measure of affect change because of

its demonstrated sensitivity to changes in anxiety, depression, and

aggression over short periods of time (Zuckerman, et al., 1964)

and because of the short time required to administer it (two to

three minutes).

There are 132 adjectives on this checklist, 89 of which are scored.

The remaining 43 items are buffer items. No word is used in more than

one scale. No adjective is considered to be above an eighth grade

reading level. There are plus and minus items to each scale. Plus

items are scored if the subject checks them; minus items are scored

if the subject does not check them.

The anxiety scale of the MAACL was derived from adjectives which

were answered differentially by anxious psychiatric patients and non-

anxious normals. A cross-validation of the items was done on subjects

with hypnotically induced anxiety. A validity study of the scale was

done on college students taking the scale on exam and non-exam days.

Significant rises in anxiety level were observed on exam days

(Zuckerman, 1960). A replication of this study was done by Zuckerman and








Baise (1962) which supported the original findings. Zuckerman, Levine,

and Baise (1964) found a significant rise in anxiety scores after

subjects were perceptually isolated for a period of six to eight hours.

Zuckerman and Lubin (1965) reported an unpublished study in which actors

and actresses showed significant reductions in anxiety scores when

tested prior to and after a performance. Hankoff, Rudorfer, and Paley

(1962) studied the anxiety reducing effects of several drugs and found

that chlordiazepoxide and chlorpromazine reduced anxiety scores

significantly but meprobamate and placebo did not.

The depression scale was derived from items which severely depressed

male patients checked significantly more or less frequently than normal

males. The items were cross-validated on severely depressed and normal

females. A validation of this scale by Zuckerman, et al. (1964) showed

significant increases in depression and anxiety scores of females after

viewing a film of detailed procedures in a slaughter house. The fact

that the increase for the male subjects was not significant was

explained in terms of the differential stimulus value of the film for

males and females.

The hostility scale was developed from those items on which there

was a significant change in items checked by subjects going from a normal

state to a hypnotically induced hostility state. No cross-validation

of these items was made. A validation study of the MAACL, focusing on

hostility, was done by Zuckerman, Lubin, and Robins (in Zuckerman and

Lubin, 1965). Thirty-three VA patients were rated on the Paranoid-

Belligerence factor from Lorr's Psychotic Reaction Profile, and the

ratings were correlated with the patient's MAACL scores. There was a

significant relationship between observed hostility and the MAACL








hostility scale. However, the MAACL anxiety scale was also found to be

related to hostility ratings, and it was found that the hostility scale

only discriminated at the upper end of the rating scale.

A validation study of the entire MAACL (Zuckerman, et al., 1964)

was conducted in which a "surprise exam" threat resulted in significant

increases in all three scales. Fake grades also produced significant

changes on all scales; but a real exam only caused changes in anxiety.

Zuckerman and Lubin (1965) reported that the MAACL has generally

failed to show any differences attributable to age, education, intelli-

gence, or sex in the normal population, but sex differences have been

reported among psychiatric patients.

Although the present study involved a psychiatric population, the

sex differences are not thought to be of significance because only pre-

and post-differences were used in the analysis on which each subject

served as his or her own control.


Group Perception Inventory (GPI)

This is a 9-item inventory (Appendix B) designed by the writer to

measure the patient's perception of certain aspects of the group

psychotherapy sessions as they related to Hypotheses V VIII. This

approach, which involves asking patients rather directly about their

"feelings" about the session, is very similar to that of Sechrest and

Barger (1961), and several of their items have been modified to fit this

inventory. On Items 1 through 7 the patient was instructed to indicate on

a 0-12 point scale the number which most accurately reflected his feelings

about that session. These questions inquired about (1) amount of help

the patient received from the session; (2) amount of help the group

received from the session; (3) the contribution the patient made to the








session; (4) the contribution made by the group to the patients'

problems; (5) the amount contributed to the session by the therapist;

(6) the patient's feeling relative to how he felt before the session;

and (7) the relevance of the discussion. Although there was a tendency

for the subjects to circle one of the five anchor points on each item

(Appendix B), most items received acceptable variance across subjects

and across sessions. On Item 8 the subject was asked to rank the group

from "those who you think benefited most from today's session to those

you think benefited least." On Item 9 the subject was asked to rank

the group, including himself, on overall effectiveness in handling

everyday tensions and anxieties.


Procedure

The data for this study covered a period of eight consecutive weeks

from April 20, 1965, to June 8, 1965.

The MAACL and GPI were introduced to the group in the first

session by the Clinical Psychology Department staff member who was the

supervisor for the group therapists in that group. The introduction

was as follows:

I feel I know most of you since I have been observing
this group from its beginning. The purpose of my coming
in today is to ask your cooperation in a research project
we are undertaking.

From where we sit on the other side of the window, and I'm
sure from where you sit in here, it is very difficult
sometimes to tell how the other person feels about some-
thing despite what they say they feel or by the way they
look like they feel. What we are primarily interested in
is simply getting an idea before the group starts and at
the end of the group of how you feel at that moment.

Now if any of you have objections or strong feelings about
not participating please feel perfectly free to refuse to
do it. The only people who will have access to this
information will be the therapists and the rest of the
faculty who observe the group. It will take approximately








three or four minutes before and after the group
to check off how you feel at that moment. Look the
items over, and if you have any objections you will,
of course, not have to participate (Pause). Since
there are no objections we will probably make this a
routine part of the group meeting from now on. Thank
you very much for your cooperation.

Each patient was given a clipboard at the beginning of the session

containing two copies of the MAACL, for pre- and post-measures, and a

copy of the GPI. One copy of the MAACL was filled out when the therapist

entered the room, and when the therapist indicated that the session was

over, the other copy of the MAACL and the GPI were filled out. The

patients were quite cooperative and seemed to understand what was re-

quired of them.

Since it was thought that the introduction of the checklist and

inventory would temporarily disrupt the normal routine of the group, it

was decided that the session in which it was introduced and the session

following it would probably not give an adequate estimate of the base

rates for frequency, duration, silence and interruption data. There-

fore, it was planned to use the two sessions following those as base

rate sessions.

The six sessions following these two were used to collect the

data for the evaluation of the effect of planned changes in the thera-

pists' verbal activity on the non-content measures of verbal activity.

These sessions were as follows: Two base rate sessions (B1 and 82),

two experimental sessions (El and E2), and two follow-up sessions

(Fl and F2).

The diagram below (Figure 1) outlines the data collection sequence.







In 81 and B2 the checklist and inventory data were collected and

the verbal participation measures were recorded, but no experimental

manipulations were introduced.


Date
4-20 4-27 5-4 5-11 5-18 5-25 6-1 6-8
MAACL Adjusted
and to Bl B2 El E2 Fl F2
GPI Begun Routine


Figure 1. Data Collection Sequence.



In El, for two periods of 20 minutes each, between the three 15-

minute periods of normal therapist interaction, the therapists were

silent (see Figure 2).


Period
2 2 3 4 5
Normal Therapists Normal Therapists Normal
Therapists' Silent Therapists' Silent Therapists'
Interaction Interaction Interaction
0-15 min. 16-35 min. 36-50 min. 51-70 min. 71-85 min.


Figure 2. Experimental Session #1 (El).



Silence periods of 20 minutes were chosen because it was observed

during preliminary investigation that both co-therapists occasionally

remain silent for periods of over 10 minutes but very rarely as long as

20 minutes.

In E2 the therapists were silent for a continuous period of 55

minutes between periods of normal therapists' interaction of 15 minutes

each (see Figure 3).









Period
12 3 4 5
Normal Therapists Therapists Therapists Normal
Therapists' Silent Silent Silent Therapists'
interaction Interaction
0-15 min. 16--------------------------------70 min. 71-5 min.


Figure 3. Experimental Session #2 (E2).



During all sessions the therapists were instructed to make all

interactions verbal, e.g., not to use gestures alone but use them only

in conjunction with some verbal interaction. During the silent periods

the therapists were to speak only when asked a direct question. This

occurred only two times, and both times the therapist responded in half-

second responses. Due to difficulties in timing the shift between

silence and talking, there was some talking by the therapists during

silence periods in El; however, in Group I it amounted to only eight

times for a total duration of 10.5 seconds, and in Group 2 the therapists

spoke only six times for a total of 6.5 seconds. Since most of these

responses occurred in the last minute of the silence periods, they are

felt to be of no real consequence. At no time during or after the

experimental sessions did any patient seem aware of the planned silences.

In E2 in Group 2, one of the therapists was absent because of

illness. One therapist is absent for various reasons, from time to

time, and the groups are accustomed to it. When this occurred previously,

no disruption was evident as the groups appeared to function in a normal

manner with only one therapist.

Only 85 minutes of each session were scheduled to be recorded. The

remaining five minutes were used for the group members to fill out the

MAACL and GPI. The sessions were considered as beginning when the








members all completed the forms and ending when they began filling out

the final forms. The mean length of the sessions for both groups was

83'13". Since the sessions did not always last 85 minutes,the final

period of the session (Period 5) was reduced accordingly.

Dependent variables in the study were (1) frequency of verbal

responses; (2) duration of verbal responses; (3) frequency of silences;

(4) duration of silences; (5) frequency of interruptions; and (6) patient

moods and perceptions measured by response to the MAACL and the GPI.


Operational Definition of Terms

Verbal Response.--A verbal response was defined as extending from

the beginning of a statement by a group member or therapist until some-

one else began to speak. However, if more than five seconds elapsed

after someone finished responding, and the same person began speaking

again, this was counted as another response. This was necessary because

counting the time from one person's response till the next person began

speaking would artificially inflate the duration variable. Five seconds

was chosen as a cut-off point because Matarazzo (1962) reported that

in his interview situations, if a patient responds to his own utterance,

after a pause, he does so within five seconds 65 percent of the time.

When frequency of verbal responding is reported for group data it is

reported as frequency per minute (FVR). FVR always included the

therapists' responses.

Duration.--Duration of a verbal response was the elapsed time

during which a patient or therapist was actually speaking, including

pauses of less than five seconds. When referring to duration of group

data this is the average duration of a verbal response (DVR) for the

session. DVR also included therapists' verbal responses. Total duration









refers to the sum of the duration of all verbal responses, and is

reported in relation to individuals only.

Interruption.--An interruption occurred when two or more people

talked at once. This variable was difficult to record because of an

apparent tendency of most people to anticipate an interruption and

subsequently pause as the other person interrupts. This would not show

up on the event record as an interruption. Occasionally an agreeing

response such as "yes" or "uh huh" was probably recorded as an

interruption because it was said while someone else was talking. These

facts tend to make the interruptions the least reliable of the variables

recorded. However, the operational definition used seemed the least

arbitrary and consequently the most satisfactory.


Statistical Treatment of the Data

Early in the data collection process the problem of what to do

about missing data came up as several absences occurred. Since

Hypothesis I was based on an equilibrium theory, it was not theoretically

proper to use any statistical way of estimating the missing data.

This is clearly illustrated by the fact that there is no difference,

on the variables measured, between BI and 82 in either group even though

members were absent in both groups on those days. This suggests that

the other members and the therapists "make up" for the missing member

and maintain verbal interaction at the same level. Therefore, there

seemed to be no proper way to estimate what a missing member would have

done if he had been in the group. This fact alone ruled out the use of

most parametric tests as the evaluation of correlated means requires

two scores of some kind for every subject.








Another factor that went against the use of parametric tests was

the high variance from subject to subject and within the same subject

across sessions. Because of this it was felt that the parametric tests

would not be able to make the best use of these data. Another reason,

of course, was the small number of patients in each group.

When more than two group means were compared the analysis was made

by the non-parametric Friedman 2-Way Analysis of Variance Test ( .r)

(Friedman, 1937). This test allowed the group means to be compared

without reference to individual members, which was desired. In order

to compute y. ,r each of the six sessions used.in the evaluation of

Hypothesis I was broken into five-minute intervals. This gave 17

estimates of the session mean and allowed a session to be ranked 17 times

in relation to any others involved in a particular analysis. Also, this

method controlled for any systematic changes that might occur in the

verbal activity rate during a session, as each session was compared

with all others on the same time period.

Siegel (1956) states that the y2r test compares very favorably

with the most powerful parametric test, the F test. Reviewing a study

by Friedman comparing the two tests, Siegel says "...it would be diffi-

cult or even impossible to say which is the more powerful test...."

Hypothesis IV was also evaluated by the X2r test. However, in evalu-

ating this hypothesis, missing data were supplied. The best estimate of

a subject's MAACL scores was considered to be the average of his

other scores.

Because of the small N involved, the correlations required for

the evaluation of Hypotheses V VIII were computed by the Spearman Rank

Correlation Coefficient (rs). Siegel (1956) reports the efficiency of







29


rs "...when compared with the most powerful parametric correlations,

the Pearson r, is about 91 percent."

Unless otherwise specified, the p values reported in this study

are for a one-tailed test.













CHAPTER 3

RESULTS


The results of the various statistical analyses, as they relate

to the specific hypotheses, will be presented in this chapter. The

raw data required for the computation of the various statistical tests

are presented in the appendices.

Hypothesis I

When the therapists' verbal participation is reduced
by planned periods of silence, the equilibrium of
the group will be disrupted to the extent that sig-
nificant changes will occur on the following non-
content verbal interaction variables:

(a) Frequency of verbal responses
(b) Duration of verbal responses
(c) Frequency of silences
(d) Duration of silences
(e) Frequency of interruptions

This hypothesis received some support in both groups. However,

the groups were not only affected to a different degree by the thera-

pists' silences, but showed the effects in different ways.

(a) Frequency of Verbal Responses (FVR).--ln order to evaluate

the effect of planned silences by the therapists on frequency of verbal

responses per minute, each Experimental session (El and E2) was compared

with the Base Rate sessions (Bl and B2). The periods of the Experi-

mental sessions in which the therapists were silent (El: Periods 2 and

4; E2: Periods 2, 3, and 4) were compared with the corresponding periods

in the Base Rate sessions. Also, in separate analyses, the Base Rate








sessions and the Experimental sessions were compared with the Follow-

Up sessions (F] and F2).

There was no significant difference between 81 and 82, when compared

by the Wilcoxon Matched-Pairs Signed-Ranks Test, in either Group I or

Group 2. The difference between Fl and F2 was not significant in

either group when compared by the same test. However, in both groups

E2 was significantly lower than El (Group I, p C .05; Group 2, p c .01;

two-tailed tests).

In Group I a Friedman 2-Way Analysis of Variance ( 2) involving

BI, B2, and El, was not significant (p < .50). Even a comparison of the

periods in El in which the therapists were silent (Periods 2 and 4),

with the same periods in BI and B2, was not significant (p < .50).

When Bl, B2, and E2 were compared, the Friedman test approached sig-

nificance ( 2r = 5.38; df = 2) (p C .10), with E2 having the lowest

mean. None of the other comparisons approached significance. There

was no difference between Base Rate and Follow-Up sessions or between

Experimental and Follow-Up sessions.

In Group 2 the Friedman test between Bl, 82, and El was not

significant (p < .95). The direct comparison of Periods 2 and 4 of El

with Periods 2 and 4 of Bl and 82 was also not significant (p < .70).

However, an analysis of Bl, 82, and E2 was highly significant ( 2r

18.73; df = 3) (p < .001), with E2 again having the lowest mean. The

Experimental sessions differed significantly from the Follow-Up

sessions (y 2 = 25.03; df = 3) (p < .001). The Base Rate sessions

also differed significantly from Follow-Up sessions, with the Follow-Up

sessions having the highest mean FVR, but the level of significance

was lower (x. 2r = 11.42; df = 3) (p< .01).







Figure 4 shows session means of FVR for the Base Rate, Experimental,

and Follow-Up sessions for both groups. Appendix C shows FVR in five-

minute intervals for the six sessions.

(b) Average Duration of Verbal Responses (DVR).--Since no

correlation between frequency and DVR has been reported in the literature,

Spearman Rank Order correlations between these two variables were

computed for both groups. The correlations ranged from -.842 to -.971,

with an average correlation of -.922.

SESSION

81 B2 El E2 Fl F2

10
9


W7

a-6 ...---- /
in \ /
4 '/ -- Group 1

3 -----Group 2


Figure 4. Average Frequency of Verbal Responses Per Minute (FVR)
Across Six Sessions for Group 1 and Group 2.


Although this correlation was known to exist between average

duration and frequency, the duration measure was included in the analysis

because of the possibility that the relationship might change during

the experimental sessions.

Because of the high correlation between these variables, the

analysis of DVR data yielded essentially the same results as the FVR

data. However, there were some exceptions, as noted below.

In Group 1 the comparison of Bl, 82, and El on DVR was reduced

from the p < .10 level of FVR to p 4 .30. The difference between El and








E2, which was significant on FVR (p < .05), was not significant on DVR.

In those instances in Group 1 where there was a significant

relationship or tendency toward significance (p C .10) on the FVR data,

the analyses of DVR yielded a lower level of significance. This

relationship did not hold in Group 2 where the significance levels that



SESSION

B1 B2 El E2 Fl F2

18
16
14 / \
o /
-12 / \

10 S
8





Group 2
..--.- Group 2


Figure 5. Average Duration of Verbal Responses Per Minute (DVR)
Across Six Sessions for Group 1 and Group 2.


were obtained for the FVR data were maintained in the DVR data. Figure

5 shows session means of DVR for Base Rate, Experimental, and Follow-Up

sessions for both groups. Appendix D shows DVR in five-minute intervals

for the six sessions.

(c) Frequency of Silences.--Because of the relative infrequency

of silences and interruptions in a session, breaking these data into

frequency per five-minute interval, as was done with FVR and average

duration data, resulted in too many empty cells for a meaningful

analysis. Therefore, silence and interruption data were analyzed over








the five periods of the session rather than 17 five-minute intervals.

In Group I the Friedman test comparing Bl, B2, and E2 for frequency

of silences was significant (p 4 .02), with E2 having the highest

frequency. There was no difference between 81 and B2. There were also

significantly more silences in the Follow-Up sessions than in Base Rate

sessions (p <.02) in Group 1.

For the same analyses in Group 2 there were no significant

differences. However, there was a tendency for more silences to occur

during El than in the Base Rate sessions ( X2r = 4.90; df = 2) (p 4 .10).

There was also a tendency for more silences to occur in F2 than in the

Base Rate sessions ()C2r = 5.70; df = 2) (p .10).

(d) Duration of Silences.--The length of silences occurring in

Group 1, session E2, were significantly longer than those occurring in

the Base Rate sessions (X 2r = 10.00; df = 2) (p < .001). El did not

differ significantly from the Base Rate sessions, and the Base Rate

sessions did not differ significantly from the Follow-Up sessions.

In Group 2 there were no significant differences across sessions in

the average duration of silences (see Figure 7).

(e) Frequency of Interruptions.--There were no significant

differences in any of the comparisons of the frequency of interruptions

per session. However, when Period 1 was not included in the analysis

for Group 1, a Friedman analysis of Bl, B2, and El was highly significant

( X r = 9.62; N = 4; k = 3) (p < .005),with El having the highest mean.

Leaving Period I out of the analysis seems justified here as this period

should not differ from the Base Rate sessions since this was prior to

any experimental conditions.

Interruptions occurred so infrequently in Group 2 that no statistical

test was made on those data.

















60
50

a 40

30
20

10


35



SESSION

BI 82 El E2 FI F2


/
I
I
I

/
/ *> S
/ s /


_- Group I

...- Group 2


Figure 6. Average Frequency of Silences Per Session for Group 1
and Group 2.











SESSION

B1 82 El E2 Fl F2


- Group 1

- Group 2


Figure 7. Average Duration of Silences Per Session for Group 1
and Group 2.








Hypothesis II

When the therapists' verbal participation is reduced
by planned periods of silences, the longer the periods
of silences the more disruptive they will be to the
equilibrium of the group.

This hypothesis is generally supported by the data in both groups.

In Group 1, El was significantly different foom the Base Rate

sessions on only one variable--interruptions, whereas E2 was significantly

different from the Base Rate sessions on both frequency (p < .02) and

duration (p <.001) of silences and approached significance (p c .10)

on FVR.

The differences were much more striking in Group 2. El did not

differ from the Base Rate sessions on any variable. There was a

tendency for more silences to occur in El than in the Base Rate sessions

(p < .10), but this was the only variable on which El approached a

significant difference from the Base Rate sessions; however, E2 was

significantly lower than the Base Rate sessions on FVR and DVR (p / .001).

Differences on FVR and DVR are considered to be the best indications of

disruption as they tend to be the most stable from session to session.

Hypothesis III

When the therapists' verbal participation returns to
normal following sessions in which their verbal
participation has been restricted, the equilibrium of
the group will be re-established at a level more
similar to the base rate level than to the level of
the sessions in which the therapists' verbal partici-
pation has been restricted.

This hypothesis was generally supported by the data in Group 2

with the exception of silences. This hypothesis was largely untestable

in Group 1 as few significant changes occurred in either El or E2.

In Group 1, however, Figure 1 shows that after E2 (which was the lowest)








mean) there was a gradual return to the Base Rate level. On frequency

of silences and duration of silences the differences between the Base

Rate and Follow-Up sessions were significant, whereas the differences

between Experimental and Follow-Up were not significant. This is

contrary to the hypothesis.

In Group 2, although the Follow-Up sessions were significantly

different from the Base Rate sessions on FVR ( y 2 = 11.42; df = 3)

(p < .01), the difference was of greater significance between Follow-

Up and Experimental sessions (, 2r = 25.03; df = 3) (p 4 .001). In

addition, a comparison of B1, B2, and F2 by the Friedman test was not

significant. In Group 2 there were no significant differences in

duration of silences across sessions. However, in frequency of silences

the comparison of Base Rate and Follow-Up sessions approached signifi-

cance ( X2r = 7.62; df = 3) (p < .10), whereas the comparison of

Experimental and Follow-Up was significant ( 2r = 8.10; df = 3)

(p C .05). Although no adequate statistical test can be made between

these session means because of a small N, El appears to differ from E2

and Fl appears to differ from F2. Therefore, the Experimental-Follow-

Up comparison is somewhat misleading. Consequently, it was not possible

to determine adequately the relative differences between Base Rate-

Follow-Up and the Experimental-Follow-Up sessions, although the difference

appeared to be greater between the former.

Hypothesis IV

When the therapists' verbal participation is reduced by
planned periods of silence, significant increases will
occur on measures of patient anxiety, depression, and
hostility in the group.

This hypothesis was not supported by the data as none of the

measures was significant in either group.








One person in each group was left out of this analysis because

each missed three or more sessions. For those members who were absent

two times or less, their mean pre- and post-MAACL scores, determined

from the sessions they attended, were substituted for the sessions

missed. Each subject was then ranked across the six sessions by change

scores on the MAACL for each of the three affect variables--anxiety,

depression, and hostility. A Friedman 2-Way Analysis of Variance test

was calculated for each variable and none was found to be significant

(see Figures 8 and 9).

In addition, the change scores for each individual on the three

variables were summed for each session both algebraically and without

regard to sign. Separate analyses by the Friedman test on these data

were also not significant.

Hypothesis V

Patients' perceptions of the amount of help received
from the group, the amount they contributed to the
group, and their own overall effectiveness in handling
their problems will vary positively with changes in
the amount of their own verbal responses.

This hypothesis was generally supported by the data.

In the evaluation of this hypothesis, as in Hypothesis IV, one

patient from each group was excluded from the analyses because of poor

attendance. Since this hypothesis related primarily to level of verbal

participation in general and not directly to the experimental sessions,

all eight sessions on which data were available were used.

The analysis of this hypothesis involved relating each individual's

verbal participation measures to his responses to GPI items and not to

measures relating specifically to a particular group; consequently the

subjects in both groups were pooled and one overall analysis was made.












SESSION

B1 B2 El E2 F] F2
4.00
A
,\\
3.00

2.00 / / \

1.00
/ /
/ \
1.0 / /\ 7- \

/ \


-1.00 / \
./ \ .
-2.00 /

-3.00 Anxiety
---- Depression
-4.00 -----Hostility


Figure 8. Mean MAACL Change from Pre-Group to Post-Group Testing on
Anxiety, Depression, and Hostility in Group 1.


















4.00


3.00


2.00


1.00


0


-1.00


-2.00


-3.00


SESSION

BI B2 El E2


Fl F2


"\








\
\ I
\ I
---Anxiety % I
----Depression
CN


-4.00 -.-.-Hostiity

Figure 9. Mean MAACL Change from Pre-Group to Post-Group Testing on
Anxiety, Depression, and Hostility in Group 2.








Although there is generally a high correlation in a given session

between the rankings of the patients on frequency of verbal responses

and total duration of time spoken, there is little correlation from

session to session between a patient's frequency and his total duration.

In other words, the person who talks the highest number of times in a

session is generally the one who talks longest, but the session in which

a given patient talks the highest number of times may not be the session

in which he talks longest. Since these measures may be getting at

different aspects of the subject's participation in the group, both were

used in the evaluation of this hypothesis.

When Spearman Rank Order correlations were computed between the

amount of help patients felt they received (GPI, Item 1) and the total

duration of their verbal responses over eight sessions, it was found

that they were highly correlated. Of the eleven correlations computed

between these two variables, five were significant beyond p < .05. The

combined probability of this many significant correlations occurring by

chance is p < .0001. Therefore, the null hypothesis of no relationship

is rejected.

However, there does not appear to be a significant relationship

between frequency of verbal responses in a session and the amount of

help patients feel they received. Out of the II correlations computed

between these variables, only one was significant beyond p < .05.

Both frequency and total duration of verbal responses are related

to the amount the patients felt they contributed to the group (GPI,

Item 3). Four of 11 of the correlations between frequency and "amount

contributed" were significant (beyond p 4 .05). The combined probability

of this occurring by chance is p 4 .002. Three of 11 correlations








between a patient's frequency of verbal responses and "amount contributed"

were significant (beyond p < .05). This is significant beyond p < .02.

In evaluating the relationship between total duration of speaking

and effectiveness in handling their problems (GPI, Item 9), the subject's

rankings of himself in relation to the rest of the group for the sessions

he attended were ranked, and then compared to the ranks of the total

duration of speaking, for that subject for those sessions. Although

only one patient's correlation between total duration and "effective-

ness" was significant (p < .01), there appeared to be an interesting

relationship between these variables. Of the 11 patients for which

this correlation was computed, nine were negative. The two patients

who did not have negative correlations had an r=.000, as those patients

did not change their ratings of themselves on any session. One of

these patients ranked himself last every time and the other ranked

herself in the middle every time.

Hypothesis VI

Patients' perceptions of the amount of benefit each
patient received from the group will be a function
of the amount of each patient's verbal participation.

The evaluation of this hypothesis involved averaging the rankings

by the patients of who benefited most to who benefited least in each

session (GPI, Item 8). The averages were ranked and correlated with

actual ranks of each patient's actual verbal participation measures

for that session.

This hypothesis was supported by the data on both groups as 7 out

of 15 correlations were significant beyond p .05. The combined

probability of 8 out of 15 correlations occurring by chance is p & .0001.

Of the correlations between total duration of verbal responses







over the sessions and rankings of "benefit", 4 out of 15 were

significant (p < .01).

Hypothesis VI

Patients' perceptions of the amount of benefit each
patient received from the group will be a function
of the amount of each patient's verbal participation.

This hypothesis was not supported by the data.

In order to evaluate this hypothesis the group averages on

response to GPI, item 2, were ranked across eight sessions and then

correlated with the FVR for each session.

In Group 1 the correlation between FVR and ratings of the amount

of help the group received was not significant (rs = --357). However,

in Group 2 the correlations, although not significant (p 4 .10, two-

tailed test) approached significance in the direction opposite that

predicted (rs = -.659)-

In Group 1 there was not a significant correlation between average

ratings of the "amount the group contributed" and FVR. In Group 2 there

was a significant relationship between the amount the group contributed

and FVR (p <.05, two-tailed test). Again the correlation was negative

(rs = -.771).

Hypothesis VIll

Patients' perceptions of the therapists' contribution
to the group and the degree to which the therapists kept
the group discussion on relevant problems will vary
positively with the amount of therapists' verbal
participation in the group.

This hypothesis was only partially supported by the data.

This hypothesis was evaluated in the same manner as Hypothesis VII

above. In Group 1 the correlation between the amount the patients felt

the therapists contributed (GPI, Item 5) and the total duration of




44


,erapists' verbal responses was rs = .589, which was not significant.

However, perceived therapists' contribution and frequency of therapists'

verbal responses (rs = .545) was significant (p < .0).

In Group 2 che correlation between therapists' contribution and

oLal duraJion of therapists' verbal responses (rs = .726) was

significant (p < .05). Frequency of therapists' verbal responses and

perceived tnerapisis' contribution was not significant (rs = .333).

In neither Grocp 1 nor Group 2 was there a significant relation-

ship between tie degree to which patients felt the therapists kept the

discussion on relevant problems (GPI, Item 7) and either the total

duration of therapists' verbal responses or the frequency of therapists'

verbal responses.













CHAPTER 4

DISCUSSION


Group Equilibrium

The results of this study lend further support to the concept of

group equilibrium as described by Chapple and Coon (1942) by

demonstrating its applicability to the group psychotherapeutic process.

These authors state that a group is in equilibrium if its interactions

are constant within clearly defined limits, and if after a disturbance

takes place the rates return to their previous values. Constant rates

were clearly evident in this study as the base rate sessions in the

two groups did not differ from each other on any of the variables on

which measures were obtained, i.e., frequency of verbal responses,

average duration of verbal responses, frequency and duration of

silences, and frequency of interruptions. This is in spite of the fact

that in both groups different patients were absent on the two days.

The latter part of Chapple and Coon's requirement for a group in

equilibrium can be demonstrated by what happened after the Experimental

sessions. After session E2 there was a return, or tendency to return,

to the base rate level on practically every variable. In every case,

except on silences, the Follow-Up sessions were more similar to the

Base Rate sessions in both groups than the Experimental sessions were.

The fact that silences did not show a similar return toward base rate

levels may have been the result of the therapists being more comfortable

45







with silences in E2 as a result of their experience in El, and may not

reflect what the data would look like if this possible contaminant could

have been controlled. This will be discussed further in the following

section. Therefore, the groups used in this study satisfied the general

requirements for equilibrium as defined above. The fact that the groups

maintained constant rates on the previously mentioned variables even

when different group members were absent simplified the task of

assessing the hypothesized role of the therapists in group equilibrium.


Effect of Therapists' Verbal Behavior on Group Equilibrium

In the first experimental session when the therapists' verbal

participation was reduced by two planned periods of silence of 20 minutes

each, the only significant effect in either group was that Group 1 had

significantly more interruptions in that session than in the Base Rate

sessions. Therefore, patients were able to maintain the established

pattern of verbal interaction for these relatively short periods of

time. Since the therapists did not participate verbally during these

periods, this suggests that departures from the base rate level of

interaction for short periods by the therapists may increase patient

interaction. This is in line with the results of a study by Salzberg

(1962) in which he found that therapist's silence in group psychotherapy

led to significantly more interaction (patient-patient conversation)

than did talking by the therapist. His silence periods were of only

ten minutes' duration.

The importance of the length of the time the therapist does not

participate verbally in the session is indicated by the results of the

second Experimental session in which the therapists were silent for a








continuous period of 55 minutes. In Group 2 there was a significant

reduction in the frequency of verbal responses per minute and a

corresponding significant increase in the duration of the responses.

There were, however, no significant effects on the frequency and

duration of silences. This indicates that although the level of verbal

interaction was reduced, it was not brought about by long or frequent

silences. This means that the percentage of the time someone was

talking in E2 did not vary from the Base Rate sessions. This clearly

indicates that the therapists are responsible for most of the inter-

action between patients in this group.

In Group 1 the effect seemed to be quite different. Significantly

more silences occurred in the second Experimental session than in the

base rate sessions. Although not significant there was a tendency

(p 4 .10) toward fewer verbal responses per minute during this session

than in the base rate sessions. However, the average duration of

responses did not change. Therefore, any reduction in the verbal

interaction was the result of more and longer silences, rather than

reducing interaction by increasing duration of responses as in Group 2.

Presumably, a number of these silences would normally have been termi-

nated by the therapist, but when the therapist did not respond the

group was able to keep the interaction going at a level which was only

slightly reduced. Group 2 seemed to lack this ability.

The results in E2, in both groups, indicate that when the therapists

do not verbally participate in the sessions for relatively long periods

that there is a reduction in the verbal interaction rather than an

increase. Therefore, it appears that Salzberg's results (1962) may not

be generalized beyond short periods (10 to 20 minutes) of therapist

silence.








There was evidence that suggested that some of the effects found

in the longer silence period may have been due to a carry-over effect

from the first Experimental session. Appendix A (Group 1) shows that

for E2, Periods I and 2 have lower mean FVR than in El. There were also

more silences in these two periods of E2 than in El (Appendix E).

One would not expect any differences in these periods as the conditions,

to this point, were the same for both sessions.

Although a carry-over effect cannot be ruled out, several factors

obviate its importance. The first is that the difference between

these two periods of El and E2 on FVR are not significant. The second

is that there were apparently few differences between El and the Base

Rate sessions and consequently there should be only minimal carry-over

effects. The last, and probably most significant,factor relates to

comments by the therapists about their own performance in these sessions.

All of the therapists were surprised at what happened in El, and the

general feeling was that they did not need to participate verbally as

much as they had in previous sessions. Consequently, in E2, from

remarks made by the therapists, they seemed to be more prone to let a

silence continue than they would have previously. This would not only

account for the high number of silences occurring in the first two

periods of the session in Group 1, but also explain why the number of

silences continued at a high rate through the Follow-Up session in

both groups.

From the previous discussion it is clear that both groups did not

respond in the same way to similar experimental conditions. The two

groups used in this study provided a good contrast in rates of verbal

interaction. Consequently, the results pointed up the need to control







this variable in future research, as base rates seem to be important

in understanding the effect of the therapists' verbal activity in

group psychotherapy. The data on the two groups in this study suggest

that in a highly verbal group the verbal interaction patterns may be

relatively unaffected by prolonged therapists' silences, or if they are

affected, they seem to be able to recover base rate patterns of verbal

interaction quickly. On the other hand, a low verbal group may have a

greater dependence on the therapists for maintaining verbal interaction

patterns, and when the therapists are silent for extended periods the

group interaction rate is reduced significantly.

In Group 2 there apparently was a carry-over effect after E2

because the Follow-Up sessions had a significantly higher FVR than the

Base Rate sessions. However, when Bl, B2, and F2 were compared, they

were not significantly different, indicating the recovery of Base Rate

levels on this variable within two sessions.


Changes in Mood

Hypothesis IV relating to expected affect changes, measured by

the MAACL, during the sessions in which the therapists' verbal

participation was reduced, was not supported.

This suggests that even when the pattern of verbal interactions

of the group (Group 2) were changed significantly, there were no

consistent affective changes in the members of the group. Other

evidence tended to support this result. On the GPI, Item 6, which

asked how the patient felt at the end of the session relative to how he

felt when he came, three of five patients in E2, Group 2, said they

felt better, and the two others said they felt the same way as when

they came.








A Rank Order correlation was computed for each patient between his

total change score on the three variables of the MAACL and his responses

to Item 6 of the GPI, and in spite of the low variance on both variables,

3 of the 11 correlations were significant. The probability of this

occurring in the absence of a relationship is p < .02. This seems to

add some weight to the validity of both measures.

Several factors seemed to work against a significant negative

change, i.e., feeling worse, which was predicted. One factor is that

it is thought that the difference scores from pre- and post-testing on

the MAACL were attenuated by the fact that during Period 5 the therapists

were allowed to interact freely, thereby partially restoring equilibrium,

especially in E2, and consequently probably resolved some of the

feelings that may have been aroused earlier in the session. Since

the GPI was also administered at the end of the session it was probably

affected in the same way.

Another factor that may have worked against significance is

suggested by Zuckerman's report (1960) of high and significant

correlations over an eight-day period with the MAACL in a psychiatric

population, whereas with normals the correlation for the same period

was low and insignificant. This suggests that because psychiatric

patients have a higher base rate on this test than normals, that they

may be relatively insensitive to anything but major changes in affect,

at least as measured by this instrument. Although some patients did

change markedly in a given session, the average changes that occurred

were quite small. Pre- and post-changes were also attenuated by the fact

that given a high initial response the freedom to vary upward is reduced.

However, the fact is that not only did patients not change much,







but they did not all change in the same direction. Some felt better

and others felt worse, e.g., more anxious, depressed, or hostile.


Patients' Verbal Participation and Their Perceptions of Self, the Group,
and the Therapists

The tests of Hypothesis V indicate positive relationships between

both the amount of help a patient felt he received from the session and

the amount he felt he contributed to the session with how much he

actually talked during the session.

This tends to confirm the results of Sechrest and Barger (1961),

who obtained the same results.

However, the present data, which included both frequency and total

duration of verbal responses, suggests a refinement of this relationship.

Total duration of verbal activity was highly related to perceived

benefit from the session, whereas frequency of verbal responses was not.

If we can assume that when a person has a high total duration with a low

frequency of responses he is probably talking about himself, whereas

when frequency is high he is spending a lot of time interacting with

others and may or may not be talking about himself, an interesting

hypothesis is suggested. That is, patients may feel they receive more

from a group session when they talk about themselves more. However,

they feel that they are contributing to the session when both talking

about themselves and interacting with others.

The negative correlations between "overall effectiveness in handling

one's problems" and total duration were unexpected. The initial

interpretation of these results was that the more a patient talked, the

less effective he felt in relation to the rest of the group. However,

this didn't seem to fit with the findings above. When examined in the








group context a more tenable explanation was discovered. That is,

whan a patient is silent others are usually telling about themselves in

one way or another and, consequently, he may feel that in relation to

other patient's problems, his are not very bad. Therefore, it may be

that it is not only what the patient is doing that determines his

perceptions of his effectiveness in handling his problems, but also

of what the other patients are doing.

According to the tests of Hypothesis VI, patients tend to rate

other patients as benefiting from the session according to the amount

those participate in it verbally. However, here again there seems to be

a distinction between frequency and duration as the correlations of

frequency and rankings of benefit were much more significant (p < .0001)

than total duration and the same rankings (p < .01). This suggests a

stronger relationship between the frequency and benefit than duration

and benefit. It appears that patients do not see the one who talks

most, e.g., long monologues, etc., as the one who benefits most, but

the one who interacts most, e.g., talks the highest number of times.

As discussed earlier, patients tend to feel they benefit more when

they talk more (high total duration). However, this suggests that they

do not tend to perceive other patients as benefiting more when they

talk more, but rather when they interact more.

The results relating to Hypothesis VII are quite confusing as

there was no consistent pattern across groups or within groups. There

was a tendency in Group 2 for patients' perceptions of the amount of

help the group received and FVR to be negatively correlated.

There was a significant (p < .05) negative correlation (rs = -.771)

between the group responses to "amount the group contributed" and FVR.







In Group 1 the correlation was positive but not significant.

The findings on this hypothesis generally do not coincide with the

other results obtained. The lack of significance suggests that the

patients were responding to different aspects of the session when

answering the question or were interpreting the GPI item differently.

The significant negative correlation in Group 2 cannot be adequately

explained within the limits of these data.

The relationship between therapists' verbal participation and

patients' perceptions of the therapists' contribution to the group

appears to be a positive one. As one would suspect, the relationship

between frequency and duration is not as clear in these comparisons as

it was in some of the others. In Group 1 correlation between frequency

and perceived therapists' contribution and total duration and perceived

therapists' contribution both approached significance (p < .05).

However, in Group 2 the correlation between total duration and perceived

therapists' contribution was significant (p < .05), whereas the

correlation between frequency and perceived therapists' contribution was

insignificant. These correlations suggest that whenever the therapist

is talking he is perceived by the patients as contributing to the group,

whether the duration of his responses is long or short. Therefore,

frequency and duration may measure much of the same thing as far as

therapists' verbal responses are concerned.

The hypothesized relationship between the degree to which patients

feel the therapists kept the discussion on relevant problems and both

frequency and duration of the therapists' verbal responses was not

supported. One reason for this may be the fact that most patients

tended to rate the discussion high on relevancy throughout the sessions.








It may be that patients tend to perceive any behavior by the therapist

as relevant to the problems of the group.


Summary

The results of this study show that Chapple and Coon's concept of

group equilibrium is applicable to psychotherapy groups. Base Rate

values for the non-content verbal interaction variables which were

studied varied relatively little. After the Experimental sessions had

made significant changes on several variables there was a return, or

tendency to return, to the Base Rate level in following sessions on all

of the variables except silences.

The evaluation of the effect of the therapists' verbal behavior

on group equilibrium indicated that group base rates play an important

part in determining the therapists' affects on the group. In a high

verbal group the therapist does not seem to play as important a role

in maintaining normal verbal interaction patterns as he does in a low

verbal group. At least in this study when the therapists' verbal

participation was reduced by planned periods of silence, there was a

significant reduction in group interaction in the low verbal group

(Group 2); whereas in the high verbal group (Group 1) there was a slight,

but insignificant, reduction.

The present results indicate that in evaluating the effects of

planned therapists' silences on the group an important variable is the

length of the silences. In this study two 20-minute periods of

therapists' silence between three normal periods of 15 minutes each

had little effect on either group, as the patients were able to

maintain previous base rate levels without the therapists. The only







exception to this was in Group 1 which had significantly more interrup-

tions on this day. However, in the following session in which the

therapists were silent for a period of 55 minutes between two normal

periods of 15 minutes, several significant changes occurred. Group 1

had significantly more and longer silences in this session than in the

Base Rate sessions and slight reduction in group interaction. Group 2

showed a significant reduction in verbal interaction. Therefore,

therapist silence in group psychotherapy seems to have different

effects on the group depending on its duration.

The hypothesized changes in mood, or affect, during the

Experimental sessions was not supported. Although several methodological

problems may have contributed to the lack of significance, there were

no uniform changes among the group members. Some made positive and

some made negative changes. It appears that patients' affective

responses to changes in the verbal interaction patterns of a group

differ from individual to individual. Some patients may be upset and

others may feel better.

A positive relationship was found between the amount of help a

patient felt he received from the group and his total duration of

speaking in a session. There was no relationship between the amount of

help a patient received from the session and the frequency of his verbal

responses. Since a high frequency of verbal responses tends to

indicate a high interaction with others in the group, it appears that

patients may feel they benefit most not from interacting at a high

rate, but talking at a high rate. There was also a significant relation-

ship between both an individual's frequency and duration of verbal

responses and the amount he felt he contributed to the group.







Patients' perceptions of their own overall effectiveness in

handling their problems tended to be negatively correlated with the

total duration of their verbal responses. The positive relationship

suggested from this is that the more other group members talk, the more

effective the patient feels. This tends to confirm a reason often given

by patients in group psychotherapy as to the benefit they receive from

the group. It is that by being in group they find out that they are

not the only ones with problems, and that some people have problems

worse than their own.

There was a significant positive relationship between the rankings

of the group on who benefited most to who benefited least and their

actual rank on both frequency and duration of verbal responses.

However, the relationship between frequency and "benefited" was much

stronger than duration and "benefited." This suggests that patients

feel that other patients receive the most benefit from the group when

they interact the most.

Patients' perceptions of the amount of help the group received

and the amount the group contributed did not appear to be meaningfully

related to the group's frequency of verbal responses per minute.

One explanation for this probably lies in the ambiguousness of these

items on the instrument used.

There was a positive relationship between the frequency and

duration of the therapists' verbal responses and the amount the patients

felt they contributed to the group. However, the degree to which patients

felt that the therapists kept the discussion on relevant problems, and

the amount of therapists' verbal participation in the session, did not

appear to be related.







Suggestions and Implications for Future Research

The most obvious implications for future research resulting from

this study seems to be a need for further knowledge of and control of

base rate variables of verbal interaction in group psychotherapy

research. Because of the stability of these variables, they may pro-

vide a much more reliable and effective way of detecting change in the

group over time, or as a result of experimental manipulations, than

more traditional subjective approaches. Just as Cook (1964) suggests

that percentage of silence in an individual psychotherapy session is an

index to the "climate" of the session, so may several of the base rate

variables used in this study indicate the same thing about group psycho-

therapy. In such case, a knowledge of base rates and the ways in which

given variables effect them would add immeasurably to the understanding

of group psychotherapy.

The present study indicates that the comparison of groups with

different base rates would be profitable. Knutson's study (1963) of

quiet and vocal non-therapy groups provides the general model, and

in fact asks many questions that would be applicable to group psycho-

therapy, e.g., patients' satisfaction with the group, patients' progress,

etc.

Because of the differences found in this study with regard to

the length of the therapists' silences, a more detailed study of this

variable is needed. There was an indication that short silences have

little effect on the group level of verbal interaction and may even

facilitate patient-patient interaction; whereas long silences reduce

interaction. This, of course, suggests that the relationship between

the amount of group interaction and length of planned therapists'




58


silences is curvilinear. The optimal amount of therapist's activity

for the best group interaction is not known. Although some inroads

are being made into an understanding of the influence of the therapist's

verbal activity in groups, the relationship is obviously complex and

needs considerably more research.

Several interesting relationships were suggested by the data

regarding perceived benefit and verbal participation. One of the most

interesting was the suggested relationship between frequency and

durations of an individual's responses. This study indicates that they

are probably getting at somewhat different aspects of the patient's

verbal performance in the group and, therefore, both should be used.

Several studies have used only frequency and may have seriously limited

their findings by so doing.












CHAPTER 5

SUMMARY


Two aspects of verbal behavior in group psychotherapy were

investigated in the present study. First, based on Chapple and Coon's

theory of group equilibrium, and the results of Lennard and Bernstein,

and Matarazzo, Weins, and Saslow with regard to the role of the therapist

in maintaining equilibrium in individual psychotherapy, several hypotheses

were advanced about equilibrium in group psychotherapy. In general, it

was hypothesized that the group therapist plays a major stabilizing role

in the group through his verbal behavior, and when he is restricted in

verbal activity the equilibrium of the group will be upset and increases

will occur in patient anxiety, depression, and hostility.

Secondly, several hypotheses were made about the relationship

between patients' verbal participation in group psychotherapy and their

perceptions of themselves, the therapist, and other members of the group.

Six consecutive 85-minute sessions were used to evaluate the effects

of the therapists' verbal activity on group verbal participation and

mood in two ongoing heterogeneous psychotherapy groups. Both groups had

co-therapists. The first two sessions were used to establish base rates

on the following non-content verbal interaction variables: Frequency and

duration of verbal responses, frequency and duration of silences, and

frequency of interruptions. In session three, both therapists were

silent for two periods of 20 minutes each. In session four the therapists







were silent for a period of 55 minutes. The fifth and sixth sessions

were used to evaluate the effect of the experimental sessions on

succeeding sessions. Difference scores, obtained from pre-group and

post-group testing with an adjective checklist (MAACL) were used to

evaluate the effect of the experimental sessions on mood.

The relationship between verbal participation and "patients'

perceptions" was investigated by correlating patient responses to a

group perception inventory, administered in each session with individual,

group, and therapists' measures of verbal participation.

The results of this study indicate that Chapple and Coon's concept

of group equilibrium is applicable to psychotherapy groups. Constant

base rates on non-content measures of verbal interaction were clearly

evident in both groups. Following the departure from these rates during

the experimental sessions there was a return to near previous values

on all variables except frequency of group silences.

The effect of therapist silences on group verbal activity must be

interpreted in the light of at least two variables: Base rate of

verbal responses in the group and length of the therapists' silences.

Longer silences tend to reduce group interaction, whereas shorter

silences have very little effect on the established patterns of verbal

interaction. Generally speaking, the "high verbal" group showed few

significant effects from the long silence, whereas the "low verbal"

group showed marked changes in both frequency and duration of responses

which are the most stable of the base rate variables.

There were no significant changes in anxiety, depression, and

hostility as a result of the experimental manipulations.





61


A positive relationship was found between the amount of help a

patient felt he received from a session and his total duration of

speaking in that session. There was also a positive relationship be-

tween the amount a patient felt he contributed and both the duration

and frequency of his verbal responses. There was a highly positive

relationship between the rankings of benefit from the session and the

rankings of the patients on frequency of verbal responses. The

possibility that frequency and duration may measure different aspects

of a patient's participation was discussed. There was also a positive

relationship between the amount the therapists talked and the amount the

patients thought the therapists contributed to the group.





























APPENDICES










APPENDIX A

MAACL

Name

Below you will find words which describe different kinds of moods
and feelings. Check the words which describe how you feel now. Some
of the words may sound alike, but we want you to check all the words
that describe your feelings. Work rapidly.


active

adventurous

affectionate

afraid

agitated

agreeable

aggressive

alive

alone

amiable

amused

angry

annoyed

awful

bashful

bitter

blue

bored

calm

cautious


cheerful

clean

complaining

contented

contrary

cool

cooperative

critical

cross

cruel

daring

desperate

destroyed

devoted

disagreeable

discontented

discouraged

disgusted

displeased

energetic


enraged

enthusiastic

fearful

fine

fit

forlorn

frank

free

friendly

frightened

furious

_gay

gentle

glad

gloomy

good

good-natured

grim

happy

healthy


hopeless

hostile

impatient

incensed

indignant

inspired

interested

irritated

jealous

joyful

kindly

lonely

lost

loving

low

lucky

mad

mean

meek

merry




64
APPENDIX A (Continued)


mild quiet stormy tormented

miserable reckless strong _understanding

nervous _rejected _suffering unhappy

obliging rough sullen unsociable

offended sad sunk upset

outraged safe sympathetic vexed

panicky satisfied tame warm

patient secure tender whole

peaceful shaky tense wild

pleased shy terrible willful

pleasant soothed terrified wilted

polite steady thoughtful _worrying

powerful stubborn timid young









APPENDIX B

GPI


On Questions 1-7 please circle the number on the scale that indicates
best your feelings about today's session. You may circle any number
and not just those with words or phrases under them. Use the words
and phrases to help you determine the number that best describes
your feelings.


1. From today's session I received:

12 11 10 9 8 7 6 5 4 3 2 1 0
/ / / / / / / / / / / / /
a a lot some very no
great of help little help
deal help help


2. From today's session the group received:

12 11 10 9 8 7 6 5 4 3 2 1 0
/ / / / / / / / / / / / /
a a lot some very no
great of help little help
deal help help


3. In today's session I contributed:

0 1 2 3 4 5 6 7 8 9 10 II 12
/ / / / / / / / / / / / /
nothing very some a lot a
little great
deal


4. In today's session the group contributed:

0 1 2 3 4 5 6 7 8 9 10 11 12
/ / / / / / / / / / / / /
nothing very some a lot a
little great
deal




66
APPENDIX B (Continued)


5. In today's session the therapists contributed:


12
/


11 10 9 8
/ / / /


7 6
/ /


5 4 3
I I /


2 1 0
I I


gr
d


all
very
relevant


mostly
relevant


about as
relevant
as irrelevant


mostly
irrelevant


thing


a a lot some very nol
great little
leal


6. After today's session I feel:

0 1 2 3 4 5 6 7 8 9 10 11
/ / / / I I / I I / / /
uch a little about same a little mu
-se worse as before better bet
I came


7. In today's session the therapists kept the discussion on problems
that were:

12 11 10 9 8 7 6 5 4 3 2 1
/ / / / / / / / / / / /


all
very
irrelevant


8. Rank the group (including yourself) from those who you think benefited
most from today's session to those who you think benefited least.
(Let 1 be "benefited most" and 6 be "benefited least.")

(Patients' names were listed alphabetically here.)


9. Rank the group (including yourself) from those who you think are
most able to deal effectively with everyday tensions and anxieties
to those you think are least able to do so. (Let 1 be "most
effective" and 6 be "least effective.")

(Patients' names were listed alphabetically here.)


12
I
ch
ter


I








APPENDIX C

Average Frequency of Verbal Responses Per Minute (FVR)
in Five-Minute Intervals


Group I

Sessions
Intervals BI B2 El E2 Fl F2

1 11.20 5.00 7.60 8.60 9.20 5.60

2 7.20 7.40 6.00 6.00 9.80 8.20

3 10.00 12.20 8.80 3.80 8.60 4.00

4 7.60 7.40 9.40* 5.80* 10.00 4.60

5 7.60 7.60 9.20* 0.80* 7.60 8.80

6 7.80 10.20 12.80* 5.80* 9.60 6.60

7 10.20 9.00 8.60* 6.80* 11.60 11.20

8 11.60 5.00 11.20 9.20* 4.00 9.60

9 10.20 10.60 14.60 7.80* 9.00 10.00

10 8.40 10.00 6.60 11.00* 2.60 7.80

11 7.60 8.20 8.00* 4.60* 3.60 8.40

12 9.40 6.80 9.00* 5.40* 1.60 4.40

13 5.40 13.20 6.00* 7.40* 8.40 10.40

14 5.80 7.20 8.80* 8.60* 8.40 9.80

15 4.40 6.00 6.60 10.20 8.00 13.20

16 5.00 13.00 7.40 6.00 13.40 11.00

17 8.00 8.18 9.41 6.74 7.84 10.90

Mean 8.08 8.65 8.82 6.74 7.84 8.50

*Indicates periods of therapists' silence.













Intervals 81

1 10.40

2 8.20

3 3.40

4 5.80

5 5.40

6 3.60

7 7.20

8 3.20

9 5.60

10 4.20

11 8.20

12 6.00

13 4.60

14 6.20

15 6.20

16 3.20

17 5.71


Mean


5.71 5.92 5.74 3.62 7.98 7.23


*Indicates periods of therapists' silence.


APPENDIX C (Continued)

Group 2
Sessions
B2 El E2 Fl

10.00 7.00 8.00 10.00

6.40 4.80 5.60 7.20

7.00 5.80 3.00 10.00

8.20 2.60* 2.00* 8.60

4.60 7.20* 2.40* 6.40

4.60 3.00* 3.00* 9.00

5.40 7.20* 4.80* 11.40

1.80 7.40 3.20* 3.40

3.60 6.00 1.80* 4.20

5.40 5.40 1.80* 9.80

5.20 7.80* 3.20* 9.00

7.80 4.60* 2.60* 10.60

8.20 8.00* 4.20* 6.40

6.00 3.60* 2.40* 7.60

5.20 3.80 3.80 8.20

6.60 8.40 6.20 6.40

4.69 5.74 3.62 7.98


F2

7.80

9.40

10.20

7.20

5.80

10.80

5.60

4.00

8.00

9.00

5.00

5.60

6.20

4.00

8.80

10.00

5.53







APPENDIX D

Average Duration (in seconds) of a Verbal Response in
Five-Minute Intervals

Group I

Sessions


Intervals 81

1 4.41

2 7.36

3 5.16

4 6.86

5 6.80

6 7.03

7 5.10

8 4.67

9 5.25

10 6.73

11 7.16

12 5.71

13 9.98

14 9.12

15 12.82

16 11.50

17 6.52


Mean


B2

11.54

6.32

4.02

6.80

6.53

5.05

6.33

11.80

5.38

5.65

6.89

7.76

4.20

8.10

9.57

4.26

6.48


El E2 F1

6.53 3.56 5.70

8.17 6.95 4.79

4.34 11.71 5.99

5.90* 8.79* 4.76

5.25* 56.00* 6.97

4.36* 7.59* 4.81

6.78* 5.06* 4.39

4.80 5.96* 14.08

4.09 6.54* 5.32

8.68 5.16* 19.42

6.86* 8.24* 14.25

6.30* 8.13* 15.12

9.33* 7.64* 5.54

5.22* 6.59* 6.49

7.74 5.24 5.84

6.85 8.93 4.09

5.84 7.68 7.97


7.19 6.86 6.30 9.99 7.97


*Indicates periods of therapists' silence.


F2

9.96

5.82

1 1.70

11 .83

6.49

7.14

4.72

5.74

5.51

7.29

6.30

12.98

5.56

5.07

3.70

4.25

4.87


7.00








APPENDIX D (Continued)

Group 2

Sessions

Intervals Bi 82 El E2 FI F2

1 4.76 5.33 6.16 6.49 4.59 6.32

2 5.35 6.98 14.00 8.98 5.86 3.99

3 17.21 6.10 9.38 17.83 5.09 3.95

4 10.50 4.95 14.85* 23.70* 5.76 7.36

5 8.23 12.07 5.54* 15.17" 7.84 8.19

6 13.78 11.83 19.77* 19.70* 5.60 5.09

7 7.79 8.65 7.58* 11.77* 5.18 8.89

8 17.62 30.28 7.50 18.28* 17.35 12.83

9 8.77 12.72 7.87 31.94* 12.55 6.19

10 13.40 11.00 11.45 33.17* 4.53 5.00

11 5.98 10.44 5.38* 14.59* 5.70 8.62

12 7.92 5.88 11.26* 20.27* 4.86 6.95

13 11.76 5.83 6.05* 10.43* 7.77 7.05

14 8.56 9.13 15.11* 23.96* 6.66 13.61

15 6.91 10.00 15.58 15.37 5.93 5.45

16 15.88 8.00 6.68 7.42 9.05 4.94

17 5.25 12.13 10.26 17.37 7.14 6.72

Mean 9.98 10.08 10.26 17.37 7.14 7.13


*Indicates periods of therapists' silence.







APPENDIX E

Frequency of Silences Occurring During Five
Periods of Each Session

Group 1

Session

Period Bl 82 El E2 Fl F2

1 6 4 12 19 6 7

2 2 6 5 9 8 7

3 0 1 0 2 11 1

4 3 3 4 13 8 5

5 0 1 6 7 3 8


Total 11 15 27 50 36 28





Group 2

Session

Period BI 82 El E2 Fl F2

1 7 10 11 10 12 17

2 8 15 24 9 6 12

3 5 5 6 5 8 8

4 6 3 11 11 11 22

5 5 5 5 3 4 15


Total 31 38 57 38 41 74









APPENDIX F

Average Duration of Silences for Five
Periods of Each Session


Group 1

Session


Period BI

1 5.33

2 5-75

3 0

4 5.83

5 0


Mean


B2 El E2 Fl

6.38 7.92 11.24 6.58

8.17 7.50 22.50 7.25

6.50 0 8.00 6.50

6.17 10.12 12.27 25.82

5.00 6.40 5.92 8.25


5.64 6.44 7.98 11.99 10.88 7.78





Group 2

Session

BI B2 El E2 Fl F2

7.57 9.55 8.36 7.15 8.58 6.76

8.43 8.63 6.88 11.67 6.08 7.25

6.30 8.50 6.42 7.40 7.94 9.69

8.83 8.83 7.04 10.05 7.23 9.52

13.75 6.33 6.88 6.00 12.83 7.43


8.98 8.37 7.12 8.45 8.53 8.13


Mean


-- - -- - --


--








APPENDIX G

Frequency of Interruptions Occurring During
Five Periods of Each Session


Group 1

Session

Period Bl 82 El E2 F] F2

1 2 2 0 0 2 0

2 6 6 19 5 7 4

3 4 12 19 8 2 8

4 2 8 11 9 2 3

5 0 4 5 5 3 8


Total 14 32 54 27 16 23





Group 2

Session

Period B1 82 El E2 Fl F2

1 0 0 2 0 2 0

2 0 0 0 0 2 2

3 0 0 0 0 0 0

4 2 0 0 1 5 0

5 0 0 3 1 0 a


Total 2 0 5 2 9 2








APPENDIX H

Percentage of Session Time Spent Talking


Group 1

Sessions

Intervals Bl B2 El E2 Fl F2

1 .823 .962 .827 .510 .873 .930

2 .883 .780 .817 .695 .782 .795

3 .860 .818 .637 .742 .858 .780

4 .868 .838 .925* .850* -793 .907

5 .862 .827 .805* .933* .883 .952

6 .913 .858 .930* .733* .770 .785

7 .867 .950 .972* .573* .848 .882

8 .903 .983 .897 .913* .938 .918

9 .893 .950 .995 .850* .798 .918

10 .942 .942 .955 .947* .842 .948

11 .907 .942 .915* .632* .855 .882

12 .895 .880 .945* .813* .403 .952

13 .898 .925 -933* .942* .775 .963

14 .882 .972 .765* .945* .908 .828

15 .940 .957 .852 .892 .778 .813

16 .958 .923 .845 .893 .913 .780

17 .870 .875 .916 .831 .814 .887


Mean .892 .905 .878 .806 .814 .878

*Indicates periods of therapists' silence.







APPENDIX H (Continued)


Group 2

Sessions

Intervals BI B2 El E2 Fl F2

1 .825 .888 .718 .865 .765 .822

2 .732 .745 .887 .838 .703 .625

3 .975 .712 .750 .892 .848 .685

4 .840 .677 .643* .790* .825 .883

5 .713 .925 .665" .607* .837 .792

6 .827 .907 .988* .985* .840 .917

7 -935 .778 .910* .942* .983 .830

8 .940 .908 .925 .975* .983 .855

9 .818 .763 .787 .958* .878 .825

10 .938 .990 .840 .995* .740 .750

11 .817 .905 .700* .778* .855 .718

12 .792 .765 .863* .878* .858 .648

13 .902 .797 .807* .730* .828 .728

14 .913 .913 .907* .958* .843 .862

15 .737 .867 .987 .973 .810 .800

16 .847 .880 .935 -767 .965 .823

17 .847 -950 .832 .871 .848 .620


Mean .847 .845 .832 .871 .848 .775


*Indicates periods of therapists' silence.









APPENDIX I

Frequency of Verbal Responses of Patients and Therapists
for Eight Consecutive Sessions


Group 1

Patients

Date 11 12 13 14 15 16 17 Therapists

4-20 93 213 34 44 4 100 41 161

4-27 -- 183 3 32 58 18 117 147

5-4 122 228 42 25 -- 41 -- 209

5-11 160 248 67 50 -- 30 52 114

5-18 180 240 48 -- 8 47 120 89

5-25 -- 202 49 -- 37 20 167 81

6-1 68 171 7 53 34 65 102 130

6-8 108 221 45 -- 42 77 147 89


Group 2

Date 21 22 23 24 25 26 Therapists

4-20 134 101 57 30 69 2 159

4-27 110 70 -- 20 80 1 150

5-4 56 33 151 -- 57 35 145

5-11 131 73 -- -- 68 35 207

5-18 110 116 -- 47 42 55 84

5-25 -- 68 85 24 32 60 26

6-1 136 267 -- 52 1 12 176

6-8 138 108 -- 67 13 37 249







APPENDIX J

Total Duration (in seconds) of Verbal Responses of
Patients and Therapists

Group 1

Patients

Date 11 12 13 14 15 16 17 Therapists

4-20 667.0 1261.0 103.0 126.0 6.0 640.0 193.5 917-5

4-27 --- 1612.0 4.0 128.0 361.0 110.5 1550.5 814.0

5-4 969.5 1318.5 621.5 286.0 --- 186.5 --- 1033.0

5-11 1213.5 1487.5 529.0 214.5 --- 112.0 369-5 601.0

5-18 1285.5 1431.0 376.0 --- 14.0 175.5 770.0 327.5

5-25 --- 1573.0 212.5 --- 153.0 78.5 1548.0 402.0

6-1 552.0 1007.0 14.5 206.0 190.0 218.5 989.0 750.5

6-8 586.0 1331.5 241.0 --- 403.0 322.0 973.0 660.0



Group 2

Patients

Date 21 22 23 24 25 26 Therapists

4-20 593.0 824.0 284.5 185.0 901.5 2.5 1004.5

4-27 535.5 742.5 2252.0 422.0 727-5 1.5 1685.0

5-4 144.5 113.0 --- --- 655.5 352.5 678.5

5-11 575.5 1085.0 --- --- 667.5 409.0 1697.0

5-18 674.0 1614.0 --- 562.0 232.0 484.0 548.0

5-25 --- 925.5 1646.5 319.5 533.0 748.5 90.5

6-1 615.5 1944.5 --- 576.0 4.0 52-5 978.5

6-8 529.0 646.0 --- 850.5 71.5 342.0 1504.0








APPENDIX K

MAACL Difference Scores for Anxiety, Depression, and Hostility

Group 1

ANXIETY*


Patient BI B2 El E2 Fl F2

11 -5 -4 3 -2.6 -2 -5

12 0 3 0 -4 -2 -4

13 -6 1 -1 5 -3 3

15 -3.7 -3.7 -1 -8 5 -7

16 3 1 0 0 2 1

17 -3-7 -7 1 9 -4 -8

DEPRESSION*

11 -7 4 -2 0.8 11 -2

12 4 4 -1 -3 0 -4

13 -11 1 2 11 -3 2

15 0.8 0.8 -2 0 1 -2

16 2 3 -3 2 4 -4

17 -1.2 -5 4 11 -5 0

HOSTILITY*

11 0 1 1 -0.4 -5 1

12 0 8 0 1 2 -5

13 -1 2 0 1 -4 1

15 1-7 1.7 -1 -5 -1 0

16 3 4 2 2 0 -1

17 -0.3 -1 7 9 -8 -10

*Scores reported to the nearest tenth are estimated data.








APPENDIX K (Continued)


Group 2


ANXIETY*


Patient

21

22

24

25

26

DEPRESSION*

21

22

24

25

26

HOSTILITY*

21

22

24

25

26


BI B2 El

-1 -3 -3

0 -1 1

+0.3 +.03 -5

0 -1

1 -1 0



-2 3 -3

-2 -2 -1

-1.2 -1.2 -6

2 -1 3

0 -1 0



0 -5 -6

2 0 -1

-1.5 -1.5 -5

1 -1 0

0 0 0


*Scores reported to the nearest tenth are estimated data.


F2


E2 FI

-2.1 -2

+0.1 1

0 1

2 -1

0 0



-3.2 -12

0.4 -2

-1 -4

0 1

0 -1



-1.5 -1

-0.1 -I

0 0

-3 0

+4 0







APPENDIX L

List of Rank Order Correlations Relating to
Hypothesis V

TABLE 2

Rank Order Correlations Between Patients' Perceptions of
Amount of Help They Received (GPI, Item 1)
and Total Duration of Their
Verbal Responses


Pat i ent

11

12

13

15

16

17


p .05



p .05


Patient

21

22

24

25

26


TABLE 3

Rank Order Correlations Between Patients' Perceptions of
Amount of Help They Received (GPI, Item 1)
and Frequency of Their
Verbal Responses


Patient r5

11 .058

12 -.220

13 .096

15 .617

16 .542

17 .150


Patient

21

22

24


N

7

8 p .01

6

8

8


p .05








APPENDIX L (Continued)

TABLE 4

RanK Order Correlations Between Patients' Perceptions of
Amount They Contributed to the Group (GPi, Item 3)
and the Total Duration of Their
Verbal Responses


Patiern ~s

11 .655

12 -.276

13 .741

IS .759

16 .732

17 .316


p < .05



p < .05


Patient r

21 -.164

22 .770

24 -.213

25 .548

26 .862


TABLE 5

Rank Order Correlations Between Patients' Perceptions of
Amount They Contributed to the Group (GPI, Item 3)
and the Frequency of Their
Verbal Responses


Patient rs

11 .655

12 -.030

13 .613

15 .850

16 .620

17 .158


N

6

8

8

6 p < .05

8


Patient rs

21 .331

22 .770

24 .152

25 .548

26 .716


p 4 .05





p < .01


p < .05





p <.05








APPENDIX L (Continued)

TABLE 6

Rank Order Correlations Between Patients' Perceptions of
Effectiveness in Handling Their Problems (GPI,
Item 9) and Total Duration of Their
Verbal Responses


Patient rs

11 -.866

12 -.339

13 -.974

15 -.580

16 -.401

17 -.473


N

5

7

7 p .01

5

7


Patient rs

21 -.345

22 -.231

24 .000

25 .000

26 -.591


TABLE 7

Rank Order Correlations Between Patients' Perceptions of
Effectiveness in Handling Their Problems (GPI,
Item 9) and Frequency of Their
Verbal Responses


N

5

7

7

5 p .05

7

6


Patient rs

21 -.579

22 -.151

24 .000

25 .000

26 -.449


Patient

11

12

13

15

16

17


rs
-.866

-.657

--533

-.948

-.401

-.338








APPENDIX L (Continued)

TABLE 8

Rank Order Correlations Between Group Rankings of
"Benefit" (GPI, Item 8) and Rankings of
Patients on Both Frequency and
Duration of Verbal Responses

Group 1 Grou

Dur./Benefit Freq./Benefit Dur./Benefit


.771

.897

.829*

.886*

.593

.429

-.320


S_2

Freq./Benefit


.486 .829*

.700 .900*

1.000* .900*

.350 .957

.206 .361

.412 .412

.900* .900*

.700 .900*


.771

.787

.829*

.886*

.593

.429

.000


c'p .05












REFERENCES


Anderson, C. The relation between speaking times and decision in
the employment interview. J. app. Psychol., 1960, 44,
267-268.

Chapple, E. D., and Coon, C. S. Principles of anthropology. New
York: Holt, 1942.

Cook, J. J. Silence in psychotherapy. J. counsel. Psychol., 1964,
11, 42-46.

Dreikurs, R., and Corsini, R. Twenty years of group psychotherapy.
Am. J. Psychiat., 1954, 110, 567-575.

Friedman, M. The use of ranks to avoid the assumption of normality
implicit in the analysis of variance. J. Amer. stat. Asso.,
1937, 32, 675-701.

Goldman-Eisler, Freida. Individual differences between interviewers
and their effect on the interviewer's conversational behavior.
J. ment. Sci., 1952, 98, 660-671.

Hankoff, L., Rudorfer, L., and Paley, H. A reference study of ataraxics:
A two week double blind outpatient evaluation. J. new Drugs,
1962, 2, 173-178.

Kendall, M. G. Rank correlation methods. London: Griffin, 1948.

Kew, J. K., Rickard, H. C., and Timmons, E. 0. The stability of group
verbal behavior. Psychol. Rec., 1962, 12, 323-325.

Knutson, A. L. Quiet and vocal groups. In M. M. Rosenbaum and M. Berger
(Eds.) Group psychotherapy and group function. New York:
Basic Books, Inc., 1963.

Lennard, H. L., and Bernstein, A. The anatomy of psychotherapy: Systems
of communication and expectation. New York: Columbia Univ.
Press, 1960.

Lubin, B., and Lubin, S. Bibliography of group psychotherapy 1956-1963.
ep. Psychother., 1964, 17, 177-230.









McFarland, R., Daniels, R., and Lieberman, M. Group psychotherapy.
In E. A. Spiegel (Ed.) Progress in neurology and psychiatry.
New York: Greene and Stratton, 1963.

Matarazzo, J. D. Prescribed behavior therapy: Suggestions from inter-
view research. In A. J. Bachrach (Ed.) Experimental
foundations of clinical psychology. New York: Basic Books,
Inc., 1962.

Matarazzo, J. D., Saslow, G., Weins, A. N., Weitman, M., and Allen, B. V.
Interviewer headnodding and interviewee speech durations.
Psychother., 1964, 1, 54-63.

Matarazzo, J. D., Weins, A. N., and Saslow, G. Studies in interview
speech behavior. In L. Krasner and L. P. Ullman (Eds.)
Research in behavior modification: New developments and
their clinical implications. New York: Holt, Rinehat, and
Winston, 1964.

Powdermaker, Florence, and Frank, J. D. Group psychotherapy.
Cambridge, Mass; Harvard Univ. Press, 1953.

Pratt, J. The class method oftreating consumption in the homes of
the poor. J.A.M.A., 1907, 49, 755-758.

Salzberg, H. D. Effects of silence and redirection on verbal responses
in group psychotherapy. Psychol. Rep., 1962, 11, 455-461.

Saslow, G., and Matarazzo, J. D. A technique for studying changes in
interview behavior. In A.P.A. Research in psychotherapy.
Washington, D. C., 1959.

Sechrest, L. B., and Barger, B. Verbal participation and perceived
benefit from group psychotherapy. Int. J. Gp. Psychother.,
1961, _1, 49-59.

Seigel, S. Non-parametric statistics: For the behavioral sciences.
New York: McGraw-Hill, 1956.

Smith, A. B., Bassin, A., and Froehlich, A. Changes in adult offenders
during group therapy. J. consult. Psychol., 1960, 24, 247-249.

Timmons, E. 0., Rickard, H. C., and Taylor, R. E. Reliability of content-
free group verbal behavior. Psychol. Rec., 1960, 10, 297-305.

Wilkinson, B. A statistical consideration in psychological research.
Psychol. Bull., 1951, 48, 156-158.




86






Wirt, R., and Wirt, Anne Psychotherapeutic processes. In Farnsworth, P.
(Ed.) Annual review of psychology. Palo Alto, Calif: Annual
Reviews, Inc., 1963.

Zuckerman, M. The development of an Affect Adjective Check List for the
measurement of anxiety. J. consult. Psychol., 1960, 457-462.

Zuckerman, M., and Baise, D. Replication and further data on the Affect
Adjective Check List measure of anxiety. J. consult. Psychol.,
1962, 26, 250-260.

Zuckerman, M., Levine, S., and Baise, D. Stress response in total and
partial perceptual isolation. Psychosom. Med., 1964, 26,
250-260.

Zuckerman, M., and Lubin, B. A preliminary manual and research progress
report for the Multiple Affect Adjective Check List.
Unpublished manuscript, 1965.

Zuckerman, M., Lubin, B., Vogel, L., and Valerius, E. Measurement of
experimentally induced affects. J. consult. Psychol., 1964, 28
418-425.













BIOGRAPHICAL SKETCH


Aubrey Clise Daniels was born May 17, 1935, in Lake City,

South Carolina. He attended public schools in Lake City and entered

Furman University in 1953. He graduated from Furman University in

1957 with the degree of Bachelor of Arts. At that time he also

received a commission in the United States Army. During his military

service he was stationed in Korea and at Fort Rucker, Alabama.

In September, 1960, he entered the University of Florida to work

toward the degree of Doctor of Philosophy. Since then he worked for

two years as a graduate assistant to the Dean, College of Health Related

Professions and for two years was a United States Public Health Fellow.

He completed an internship in Clinical Psychology at the J. Hillis Miller

Health Center, University of Florida, Gainesville, Florida, in June, 1965.

He is married to the former Rebecca Lanette Tapp of Greer, South

Carolina. They have a daughter, Laura Lee.








This dissertation was prepared under the direction of the

chairman of the candidate's supervisory committee and has been approved

by all members of that committee. It was submitted to the Dean of the

College of Arts and Sciences and to the Graduate Council, and was

approved as partial fulfillment of the requirements for the degree of

Doctor of Philosophy.


December, 1965



Dean, College of Arts and Sciences




Dean, Graduate School



Supervisory Committee:




Chairman


II{7K I~




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ECNR8W4WQ_0VMF9K INGEST_TIME 2017-07-19T20:40:43Z PACKAGE UF00097897_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

VERBAL BEHAVIOR IN GROUP PSYCHOTHERAPY By AUBREY CLISE DANIELS A DISSERTATION PRESENTED TO THE GRADUATE COUNQL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA December, 1965

PAGE 2

UNIVERSITY OF FLORIDA 3 1262 08552 3875

PAGE 3

ACKNOWLEDGMENTS The author wishes to express his gratitude to the members of the supervisory committee for their assistance in all phases of this study: Dr. Audrey Schumacher, Chairman, Dr. Benjamin Barger, Dr. John R. Barry, and Dr. William D. Wolking of the Department of Psychology, and Dr. Thomas Hanna of the Department of Philosophy. He also wishes to thank Drs. Hugh Davis and Nathan Perry, the group psychotherapy supervisors, without whose cooperation and assistance this study would have been impossible. Appreciation is also extended to Dr. Thomas Weaver, Mr. Ted Keiser, and Mr. Wayne Richard, who were the group therapists. Special appreciation is extended to Dr. Schumacher who made a difficult job much easier than it would have been otherwise through her knowledge and understanding of the problem, and through her emotional support and encouragement. Dr. Barger 's working knowledge of this kind of research and his many practical suggestions were very helpful throughout this study. A special word of appreciation is accorded Dr. Nathan Perry who inspired the present study, gave invaluable assistance in the design phase of the study, and rendered a considerable amount of technical ass i stance. Finally, thanks is due to my wife, Becky, who typed the manuscript and whose help in many ways made the completion of this study possible.

PAGE 4

TABLE OF CONTENTS Paqe ACKNOWLEDGMENTS ii LIST OF TABLES v LIST OF FIGURES vi CHAPTER 1 INTRODUCTION 1 The Present Study 10 2 METHOD 1^ Subjects 15 Groups 15 Apparatus 16 Measures of Patient Mood and Perceptions 19 Group Perception Inventory (GPl) 21 Procedure 22 Operational Definition of Terms 26 Statistical Treatment of the Data 27 3 RESULTS 30 4 DISCUSSION ^5 Group Equilibrium ^5 Effect of Therapists' Verbal Behavior on Group Equilibrium ^6 Changes in Mood ^9 Patients' Verbal Participation and Their Perceptions of Self, the Group, and the Therapists 51 Summary 5*+ Suggestions and Implications for Furture Research ... 57 5 SUMMARY 59 6 APPENDICES 62

PAGE 5

Page APPENDIX A. MAACL 63 APPENDIX B. GPI 65 APPENDIX C. Average Frequency of Verbal Responses Per Minute (FVR) in Five-Minute Intervals 67 APPENDIX D. Average Duration (in seconds) of a Verbal Response in Five-Minute Intervals 69 APPENDIX E. Frequency of Silences Occurring During Five Periods of Each Session 71 APPENDIX F. Average Duration of Silences for Five Periods of Each Session 72 APPENDIX G. Frequency of Interruptions Occurring During Five Periods of Each Session 73 APPENDIX H. Percentage of Session Time Spent Talking Jk APPENDIX I. Frequency of Verbal Responses of Patients and Therapists for Eight Consecutive Sessions. ... 76 APPENDIX J. Total Duration (in seconds) of Verbal Responses of Patients and Therapists 77 APPENDIX K. MAACL Difference Scores for Anxiety, Depression, and Hostility 78 APPENDIX L. List of Rank Order Correlations Relating to Hypothesis V 80 REFERENCES 8k BIOGRAPHICAL SKETCH 87

PAGE 6

LIST OF TABLES Table Page 1 Composition of Groups: Age, Sex, Education, and Diagnosis I7 2 List of Rank Order Correlations Relating to Hypothesis V . . 8O 3 Rank Order Correlations Between Patients' Perceptions of Amount of Help They Received (GPI, Item 1) and Frequency of Their Verbal Responses 80 k Rank Order Correlations Between Patients' Perceptions of Amount They Contributed to the Group (GPI, Item 3) and the Total Duration of Their Verbal Responses 81 5 Rank Order Correlations Between Patients' Perceptions of Amount They Contributed to the Group (GPI, Item 3) and the Frequency of Their Verbal Responses 8I 6 Rank Order Correlations Between Patients' Perceptions of Effectiveness in Handling Their Problems (GPI, I tern 9) and Total Duration of Their Verbal Responses 82 7 Rank Order Correlations Between Patients' Perceptions of Effectiveness in Handling Their Problems (GPI, Item 9) and Frequency of Their Verbal Responses 82 8 Rank Order Correlations Between Group Rankings of "Benefit" (Gpi , Item 8) and Rankings of Patients on Both Frequency and Duration of Verbal Responses 83

PAGE 7

LIST OF FIGURES Fiqure Page 1 Data Collection Sequence 2k 2 Experimental Session //I (El) 2k 3 Experimental Session ^2 (E2) 25 k Average Frequency of Verbal Responses Per Minute (FVR) Across Six Sessions for Group 1 and Group 2 32 5 Average Duration of Verbal Responses Per Minute (DVR) Across Six Sessions for Group 1 and Group 2 33 6 Average Frequency of Silences Per Session for Group 1 and Group 2 35 7 Average Duration of Silences Per Session for Group 1 and Group 2 35 8 Mean MAACL Change from Pre-Group to Post-Group Testing on Anxiety, Depression, and Hostility in Group 1 39 9 Mean MAACL Change from Pre-Group to Post-Group Testing on Anxiety, Depression, and Hostility in Group 2 kO

PAGE 8

CHAPTER 1 INTRODUCTION The efficacy of group psychotherapy in bringing about personality change is relatively unquestioned by practicing therapists. There are reports of its success in the treatment of almost every conceivable patient group. Studies quoted in a review article by Wirt and Wirt (1963) indicate that group psychotherapy has been conducted with many widely divergent groups such as the blind, aphasics, and children and adults with psychosomatic disorders. Studies using all of these groups report positive results. In spite of this, the meaning of verbal behavior to the group psychotherapeutic process is still quite vague although it is probably one of the most critical variables. The purpose of this study was to investigate two aspects of verbal behavior in group psychotherapy. First, the function of the therapist's verbal behavior in the group was investigated. Secondly, the relationship of verbal behavior and patient's perceptions of the group was studied. Group therapy in the United States is generally regarded as beginning with Pratt's classes (1907) for tubercular patients. Little research was done with group therapy, however, until 19^0. A review by Dreikurs and Corsini (195^) indicated that just over a hundred papers were published from 1910 to 1930, whereas over five hundred were published in the period from 19^0 to 19^9Since 19^9, increased interest in the use of group therapy has more than doubled the number of published papers.

PAGE 9

The tremendous growth in the application of group therapy in the past few years is no doubt due to a variety of reasons, not the least of which are (1) the treatment is relatively inexpensive; (2) professional manpower requirements are less for the number of patients treated; and (3) group therapy represents a development in the continuing search for new and more effective methods of treatment. Wirt and Wirt (1963) pointed out that a primary reason for the current enthusiasm for group psychotherapy is that its effectiveness has been demonstrated to the point that it is being advocated more and more as the treatment of choice, apart from any econimic or manpower considerations. Research on group psychotherapy, however, has not kept pace with its rapid growth as a treatment method. Although more is being written about group therapy than ever before, there is little experimental literature on the subject. Wirt and Wirt (I963), in their review article, saw most of the group psychotherapy literature as "descriptions of therapists' experiences or recommendations." One of the obvious reasons for the paucity of experimental data on group psychotherapy is the extreme complexity of the process. The many variables operating in individual psychotherapy are multipled several times over in group psychotherapy. After reviewing the literature, McFarland, Daniels, and Lieberman (1964) gave a much more basic reason as to why research in group psychotherapy has not gone beyond "descriptions of the therapists' experiences or recommendations." They state: Lack of sufficient descriptive material, or a language in which to insert these descriptions, prevents the development of a fund of knowledge based on the collective efforts of many investigators. All too frequently the inventiveness of one investigator is ineffective in making a contribution outside the limited confines of his

PAGE 10

own institution because others cannot, with any certainty, duplicate his efforts. At this time it appears that the most fruitful approach to research in this area is to study those variables that are easily defined, fairly stable, and reliably measured. A cluster of variables that meets all of these criteria is found in verbal behavior. Matarazzo (I962) has conducted several studies of changes in interviewee speech as a function of changes in interviewer speech. He has concerned himself with a number of non-content verbal interaction variables and has reported high reliabilities in recording and scoring. Kew, Rici
PAGE 11

The most articulate member of the group is not necessarily the person who gains the most, nor is the quiet one in the group the individual who does not achieve important attitudinal alterat ions . However, in a study by Sechrest and Barger (1961), patients perceived as more beneficial those sessions in which they participated, generally, at what was for them a relatively high level. This study suggests a linear relationship between verbal participation and benefit received, whereas Smith, et a_l_. (I960) suggest a relationship between these variables which is curvilinear. A curvilinear relationship is also suggested by Cook (196^) who found, in a study of silence in individual psychotherapy, that a relative lack of silence in the session characterized the unsuccessful cases, whereas a higher percentage of silence tended to characterize the successful ones. Successful cases had silences occurring from 4 to 20 percent of the time, whereas unsuccessful ones either had less than k percent silence or more than 20 percent. This seems to indicate that there is an optimal amount of verbal participation in psychotherapy. He goes on to say that silence seemed to be an index of the therapist's behavior and reflects the "climate" of the interview. Although this inference is probably a bit premature in terms of the limits of his data, it does suggest the possibility that non-content variables can lead to a better understanding of more psychodynami c concepts. Some indirect support for Cook's hypothesis is found in a study by Anderson (I960). He found that an interviewer talked more with job applicants he accepted than those he rejected. He also found more silences in the interviews with rejected applicants than in the interviews of applicants who were accepted.

PAGE 12

It is a commonly held conviction that a patient will derive more from group psychotherapy if he participates in it verbally. it is implicit in many studies, and in the theories of many group psychotherapists, that a high degree of patient-patient interaction is desirable. Yet, it also seems that it would be desirable for the therapist to be able to control the group activity level when appropriate. For example, there may be times in a group when a silence may be very threatening and the therapist may want to end it, or at other times he may want to reduce the amount of interaction. Practically every textbook on group psychotherapy discusses the desirability of maintaining control of group action. To quote Powdermaker and Frank (1953) It v-jould seem that the doctor's aim in the initial stages of therapy should be to sense when and how to give the amount of direction and support that each situation demands. Too little guidance often results in intense competition for superiority, which only upsets the patients without leading to anything beneficial. Too much guidance, on the other hand, may inhibit the appearance of the neurotic patterns, the analysis of which is part of the objective of psychotherapy. However, as stated earlier, little is known about the meaning of the effect of activity level in general, and less about the effect of the therapists' verbal activity level in psychotherapy. Although several authors have investigated the meaning of the therapist's and interviewer's verbal activity level in a two-person interview situation or individual psychotherapy relationship, very little has been done along these lines with psychotherapy groups. However, the findings in individual psychotherapy may have application to group psychotherapy. Matarazzo, Weins, and Saslow (1964) have found highly consistent changes in interviewee verbal behavior by varying the duration of the

PAGE 13

interviewer's utterances. Their results sinowed that the length of interviewee utterances varied positively with the length of therapist's utterances. They have also found that silence behavior by the interviewer causes a decrease in frequency and duration of verbalizations by the i ntervi ewee. In an investigation on the effects of individual differences between interviewers on the interaction of patients in individual psychotherapy, Goldman-EIsler (1952) found that each of three psychiatrists had his own individual reaction pattern regardless of the type patient he was interviewing. Further, she found that they influenced patients' interaction differently, i.e., the depressed patients "responded best" to active stimulation and the active patients to the passive interview. Lennard and Bernstein (I960) related therapists' verbal activity level to signs of strain in individual therapy. They found that there were fewer signs of strain in the patients of verbally active therapists as compared to patients of verbally passive therapists. Another interesting finding was that therapists compensate for lower patient verbal output by increasing their own verbal output, and for higher patient output by reducing their output. They suggest that the therapist attempts to restore the therapy session to within, what is for him, normal limits of verbal activity by varying his level of output. However, "theory oriented" therapists seemed to maintain equilibrium by establishing a constant proportion of output from session to session. Matarazzo, Welns, and Saslow (1964) found, in a study of therapistpatient verbal Interaction, that with two different patients, one therapist, by varying his own verbal participation, maintained a level of verbal activity of 93 percent with each patient. The other therapist

PAGE 14

studied showed a constant rate of verbal participation of approximately 19 percent across 12 psychotherapy sessions regardless of the patient output. These results are clearly in line with those cited above of Lennard and Bernstein (I96O). In a study of content-free verbal behavior in contrived psychotherapy groups, Timmons, Rickard, and Taylor (I96O) found that experimenters can become nearly equivalent stimuli to the patients by following simple content-free rules. The frequency and duration of a patient's verbal activity was not affected by the presence of different experimenters. Salzberg (I962), in a study of the effect of therapist activity level and direction of action on group psychotherapy patients, found that therapist's silence led to significantly more interaction (patientpatient conversation) than did talking by the therapist. However, methodological limitations of Salzberg's study restrict the generalization of these results to other psychotherapy groups. The main limitation in Salzberg's study is that the four experimental treatments were not distributed throughout all sessions. The four experimental treatments were (1) silence-redirecting in which the therapist was silent but when asked a question would mention another patient in answering; (2) talking-direct in which the therapist spoke whenever conversation lagged, speaking directly to a patient, and not referring to any other members; (3) silence-direct in which the therapist was silent but when asked a question would speak directly to a patient; and (k) talking-redirect in which the therapist spoke whenever the conversation lagged, but when asked a question would mention another patient in answering. Only combinations of silence-redirecting and

PAGE 15

8 talking-direct were used in the first ten sessions. in the last ten sessions only silence-direct and ta 1 k i ng-red i rect combinations were used. There was a significant increase in the amount of verbal activity by the group from the first ten to the last ten sessions, but one does not know whether this is due to the experimental conditions or to the possibility that patients talk more in later sessions than earlier ones. Other problems in design are also left unsolved. if the study is examined in terms of total verbal activity it can be shown that inclusion of the therapists' responses in the group total led to significantly more verbal activity for sessions in which the therapist was talking than for sessions in which he was silent. If one considers the fact that during the talking-direct condition the responses of the group averaged only six seconds each, it would appear verbal activity may have reached a saturation point. In addition, if one takes into account the fact that the therapist averaged 991 responses per half hour , it is obvious that group interaction will be reduced. If the therapist spends approximately one-third of the time talking directly to group members and they spend approximately onethird of the time responding to him, only a limited amount of time is left for group interaction. Several studies have attempted to get at the meaning of verbal activity to the patient. Matarazzo, Saslow, Weins, Weitman, and Allen (196^) found that headnodding by the therapist resulted in increases in interviewee speech duration. This, coupled with the previous findings that interviewers' speech duration caused increased interviewee speech duration, led the authors to speculate that the increases are due to

PAGE 16

more subject satisfaction wlnen the interviewer is more active. The study cited earlier by Sechrest and Barger (I96I) indicates that the amount of help the patient feels he received, the contribution the patient feels he made to the session, and the perceived relevance of a session are related to the patients' verbal participation in the session. Although not working with therapy groups, Knutson (19o3) found that quiet and vocal groups had differences in degree of satisfaction with their work, in initial progress, and in the quality of their work. The vocal groups tended to express greater satisfaction with their product i on, part i ci pat ion, and leadership than the quiet groups. On the other hand, the quality of the work (preparing a public health pamphlet) was superior in the quiet group. At least initially, verbal fluency seemed to be the sole identifying mark of leadership. They also pointed out that successful leadership in these two types of groups (quiet and vocal) seemed to require different characteristics. In Lennard and Bernstein's study it was found that patients rated those sessions in which there was a high verbal output by the therapists as proceeding "more easily." Patients indicated dissatisfaction with the sessions in which there was a low verbal output by therapists. All of these studies demonstrate the utility of the non-content aspects of verbal behavior in gaining more insight and knowledge into the therapeutic process. Although a body of knowledge is being built up and theories proposed about these variables in individual psychotherapy and non-therapy two-person interactions and groups, research at this level in group psychotherapy has only recently begun. Therefore,

PAGE 17

10 it seems that an extension of some of the theories used in individual psychotherapy and non-therapy two-person interaction and non-therapy groups to group psychotherapy would be particularly profitable. The equilibrium theory of Chappie and Coon (19^2) appears to be singularly useful in the research of the effects of verbal participation on the group process and, more particularly, the effects of changes in the therapists' verbalizations upon the verbal behavior of the group. Concerned almost exclusively with the verbal interaction of individuals and groups, the theory states that a group is in equilibrium if its interaction rates are constant within clearly defined limits, and if, after a disturbance takes place, the rates return to their previous values. The longer a state of equilibrium is maintained the more stable this state becomes. The process of adjustment which takes place within a group after a disturbance is marked by cyclical fluctuations on at least five measures: (1) frequency of i nteract i on--number of times interactions take place; (2) amount of i nteract ion-length of time spent interacting; (3) origin-response rat i o--proport ion of actions originated to the number responded to; (k) rhythm of interaction rate--character i st i c relationship of periods of action and silence; (5) degree of synchronization or adjustment In i nteract i on--degree to which interruptions and failures to respond occur. When the equilibrium is upset, members may Interrupt each other often, may become angry and unable to work together, or may let many silences occur. As the group gradually regains its equilibrium the range of fluctuations on these variables decreases and the rhythm of routine interactions becomes more constant. The Present Study The purpose of this study, then, is to investigate two aspects of

PAGE 18

11 verbal behavior in the group psychotherapeutic process. Two major aspects of its meaning will be studied. First, the major purpose of this study will be to explore the effects of variations in the therapists' verbal activity on the verbal activity in an ongoing psychotherapy group. Secondly, this study will seek to determine the extent to which verbal participation affects the patient's perception of himself, the therapist, and other members of the group. As discussed earlier, Lennard and Bernstein (I960) and Matarazzo, Weins, and Saslow (1964) suggested that verbal interaction in the psychotherapy session operates on an equilibrium model and that the therapist's verbal activity is an important factor in maintaining the equilibrium of the system. Chappel and Coon's theory provides the conceptual tools for testing whether the therapists' verbal participation serves the same function in group psychotherapy that it does in individual psychotherapy. One concern of this study is that the therapist in group psychotherapy provides a major stabilizing role primarily through his verbal behavior. Patients are typically people who lack the capacity for effective interaction or who are unable to adapt to new situations readily without help. Therefore, the patients in group psychotherapy may often find themselves in some difficulty interacting with other patients and lack the ability to get themselves out of it. It generally falls upon the therapist to remedy such situations. If the therapist is limited in this hypothesized role, the group may show marked changes in the normal pattern of verbal interaction. As the group equilibrium becomes disrupted the equilibrium of each of the members will tend to be upset, causing affective changes in the members which will lead to further disruption.

PAGE 19

When the therapist resumes normal interaction the group should return to its normal state. The recovery of the normal interactional pattern would seem to be partly a function of the duration of the disruptive conditions. If such conditions were to prevail long enough the group would either stabilize at a new level or break up completely. It is further hypothesized that group members' satisfaction with the group is a function, not only of their own level of verbal participation, but also of the other group members' verbal level and the level of therapist activity. That is, they tend to judge groups in which the verbal activity level is high as more satisfactory than groups with a lower level of verbal output. Further, group members may feel that the relevance of group discussion and the amount the therapist contributes is a function of the amount of verbal participation of the therapist. All of this is based on the assumption that group members expect to talk about their own problems, hear others' problems, and expect the therapist to contribute toward an understanding and solution of their problems. Therefore, they should see themselves, the group, and the therapist as functioning more in line with their purpose when verbal output is higher. More specifically, the following hypotheses are advanced: Hypothesis I Vyhen the therapists' verbal participation is reduced by planned periods of silence, the equilibrium of the group will be disrupted to the extent that significant changes will occur on the following noncontent verbal interaction variables: (a) Frequency of verbal responses (b) Duration of verbal responses (c) Frequency of silences (d) Duration of silences (e) Frequency of interruptions

PAGE 20

Hypothesi s I I When the therapists' verbal participation is reduced by planned periods of silences, the longer the periods of silence the more disruptive they will be to the equilibrium of the group. Hypothesi s I 11 When the therapists' verbal participation returns to normal following sessions in which their verbal participation has been restricted, the equilibrium of the group will be re-established at a level more similar to the base rate level than to the level of the sessions in which the therapists' verbal participation has been restricted. Hypothesis IV When the therapists' verbal participation is reduced by planned periods of silence, significant increases will occur on measures of patient anxiety, depression, and hostility in the group. Hypothesis V Patients' perceptions of the amount of help received from the group, the amount they contributed to the group, and their own overall effectiveness in handling their problems will vary positively with changes in the amount of their own verbal responses. Hypothesis Vl Patients' perceptions of the amount of benefit each patient received from the group will be a function of the amount of each patient's verbal participation. Hypothesis Vl I Patients' perceptions of the amount of help the group received and the amount the group contributed will vary positively with the amount of the group's verbal act i vi ty . Hypothesis VI I I Patients' perceptions of the therapists' contribution to the group and the degree to which the therapists kept the group discussion on relevant problems will vary positively with the amount of therapists' verbal participation in the group.

PAGE 21

CHAPTER 2 METHOD The general design of this study involved an analysis of the effect of planned changes in the therapists' verbal activity (specified periods of silence) on verbal participation and mood i r. two ongoing psychotherapy groups. The relationships between the patient's verbal participation in the group and his perceptions of himself, the therapists, and other group members, were also investigated. Since the hypotheses were made without reference to type of group, e.g., types of patients in the group, number of patients in the group, etc., the two groups were used to explore, to a limited degree, the extent of the genera 1 i zab i 1 i ty of the hypotheses across groups. No predictions were made about the differences between the groups and consequently there were no planned comparisons of the two groups. In order to evaluate the effect of the therapists' activity on verbal participation and mood, six consecutive sessions in each group were used. The first two were used to establisn base rates of responding. In the third session the therapists were silent for two periods of 20 minutes each, and in the fourth session were silent for a continuous period of 55 minutes. Each silent period was preceded and followed by a 15-minute period of normal interaction. The fifth and sixth sessions were employed to evaluate the effects of the experimental sessions on succeeding sessions. Adjective checi
PAGE 22

on mood . In order to investigate the relationship between verbal participation and the patient's perception of himself, other group members, and the therapist, the six sessions described above, plus the two preceeding sessions, were used for the analysis. At the end of each session each group member filled out a 9-item inventory, designed to measure his perceptions of several aspects of the session, e.g., benefit received, amount contributed, etc. The responses to these items were correlated with individual, group, and therapists' verbal act i vi ty measures . Subjects The subjects for this study were the members of two ongoing outpatient psychotherapy groups In the Psychiatric Outpatient Clinic, Teaching Hospital, J. Hillis Miller Health Center, University of Florida. The patients of each group were heterogeneous with respect to age, sex, and diagnosis (see Table 1). Groups The groups met weekly in 90-minute sessions. At the time of this study Group 1 had been meeting for three months and Group 2 for seven months. Membership had been quite stable. Only two patients had been added since the groups were formed. No members were added to or dropped out of either group during the course of the data collection. Both groups had co-therapists who were clinical psychology interns in the Department of Clinical Psychology, College of Health Related Professions. One intern served as co-theraplst in both groups. Average data from the two base rate sessions indicated that the

PAGE 23

16 groups were quite different on most of the non-content measures of verbal interaction. The mean frequency of verbal responses per minute (FVR) in Group 1 was 8.36. In Group 2 the mean FVR was 5-82. Group 1 had a mean frequency of silences of 13-0 per session, and Group 2 had a mean of 3^-5 per session. Interruption occurred at the mean rate of 23 per session in Group 1, and Group 2 had a mean of only one interruption per session. In the base rate sessions therapists in Group 1 had an average of 161. 5 verbal responses per session and talked for a total duration of speaking of I3.6 minutes. In Group 2, for the same sessions, the therapists averaged I76.O verbal responses per session and a total duration of speaking of I9.8 minutes per session. Since there were only two base rate sessions, no adequate statistical tests could be made on the differences between the groups. However, across eight sessions for which data were available. Group 1 was significantly lower than Group 2 on FVR and interruptions. There was also a tendency toward significance between the two groups on the frequency of silences in these sessions. Apparatus The group meetings were held in an observation room equipped with a one-way vision glass and ceiling microphones. A loudspeaker In the adjoining room allowed observers to hear as well as see the group sessions. All patients were aware that they were being observed from the time the groups were formed. The observers were staff group therapy supervisors and interns of the Department of Clinical Psychology. The group therapy room held eight patients and two therapists

PAGE 24

C J-i 4-1 QJ 0) -M 03 fU CU n3 (1) GJ C C > JD CC o

PAGE 25

18 comfortably. The patients and therapists sat in a circle. No special seating order was maintained although some patients seemed to prefer certain locations. The therapists were generally the last ones to enter the room and occupied any available seat. An Easterl i ne-Argus Series "S", 20 channel, LabGraph Event Recorder was used to record the verbal participation measures of frequency and duration. This is a portable ink-pen type recorder which has a variable speed drive. In recording the durations of verbal responses the speed was set at six inches per minute. This allov/ed the duration to be measured to the nearest one-half second. The event recorder was operated by a keyboard arrangement which permitted each patient to be assigned a separate channel. In order to record the interaction, the key for the patient speaking was pressed when he began talking and released when he stopped talking. Since there were never more than ten people in a group, including the therapists, the 20 channels of the event recorder made possible simultaneous recording of the same data by two operators. This permitted a reliability estimate for the recording of duration of verbal responses to be obtained. Samples were taken from the beginning, middle, and end of the first session on which simultaneous recordings were available, covering a period of kO minutes and 140 responses. A Pearson ProductMoment correlation for duration of responses was r=.994. This reliability of measurement was obtained with minimal instruction and training. Since the measurements were found to be highly reliable, the double recording was discontinued.

PAGE 26

Measures of Patient Mood and Perceptions Multiple Affect Adjective Check List (MAACL) .--Th i s checklist was used to determine affect changes as a result of events and behavior taking place during the therapy session. The MAACL (Appendix A), constructed by Zuckerman, Lubin, Vogel, and Valerius (1964), is a 132-item adjective checklist which measures immediate, as well as day to day, changes in three moods or affects: anxiety, depression, and aggression. Preand post-measures were obtained for each session during the course of the experiment. The MAACL was chosen as the measure of affect change because of its demonstrated sensitivity to changes in anxiety, depression, and aggression over short periods of time (Zuckerman, _et_ a_l_. , 1S64) and because of the short time required to administer it (two to three mi nutes) . There are 132 adjectives on this checklist, 89 of which are scored. The remaining 43 items are buffer items. No word is used in more than one scale. No adjective is considered to be above an eighth grade reading level. There are plus and minus items to each scale. Plus items are scored if the subject checks them; minus items are scored if the subject does not check them. The anxiety scale of the MAACL was derived from adjectives which were answered differentially by anxious psychiatric patients and nonanxious normals. A cross-validation of the items was done on subjects with hypnotically induced anxiety. A validity study of the scale was done on college students taking the scale on exam and non-exam days. Significant rises in anxiety level were observed on exam days (Zuckerman, I960). A replication of this study was done by Zuckerman and

PAGE 27

20 Baise (1962) which supported the original findings. Zuckerman, Levine, and Baise (196^) found a significant rise in anxiety scores after subjects were perceptually isolated for a period of six to eight hours. Zuckerman and Lubin (I965) reported an unpublished study in which actors and actresses showed significant reductions in anxiety scores when tested prior to and after a performance. Hankoff, Rudorfer, and Paley (1962) studied the anxiety reducing effects of several drugs and found that chlordiazepoxide and chlorpromazi ne reduced anxiety scores significantly but meprobamate and placebo did not. The depression scale was derived from items which severely depressed male patients checked significantly more or less frequently than normal males. The items were cross-validated on severely depressed and normal females. A validation of this scale by Zuckerman, et_ aj_. (1964) showed significant increases in depression and anxiety scores of females after viewing a film of detailed procedures In a slaughter house. The fact that the increase for the male subjects was not significant was explained in terms of the differential stimulus value of the film for males and females. The hostility scale was developed from those Items on which there was a significant change in items checked by subjects going from a normal state to a hypnotically induced hostility state. No cross-validation of these items was made. A validation study of the MAACL, focusing on hostility, was done by Zuckerman, Lubin, and Robins (in Zuckerman and Lubin, 1965). Thirty-three VA patients were rated on the ParanoidBelligerence factor from Lorr's Psychotic Reaction Profile, and the ratings were correlated with the patient's MAACL scores. There was a significant relationship between observed hostility and the MAACL

PAGE 28

21 hostility scale. However, the MAACL anxiety scale was also found to be related to hostility ratings, and it was found that the hostility scale only discriminated at the upper end of the rating scale. A validation study of the entire MAACL (Zuckerman, _et a_l_. , 1964) was conducted in which a "surprise exam" threat resulted in significant increases in all three scales. Fake grades also produced significant changes on all scales; but a real exam only caused changes in anxiety. Zuckerman and Lubin (I965) reported that the MAACL has generally failed to show any differences attributable to age, education, intelligence, or sex in the normal population, but sex differences have been reported among psychiatric patients. Although the present study involved a psychiatric population, the sex differences are not thought to be of significance because only preand post-differences were used in the analysis on which each subject served as his or her own control. Group Perception I nventory (GPi) This is a 9-item inventory (Appendix B) designed by the writer to measure the patient's perception of certain aspects of the group psychotherapy sessions as they related to Hypotheses V VIM. This approach, which involves asking patients rather directly about their "feelings" about the session, is very similar to that of Sechrest and Barger (I96I), and several of their items have been modified to fit this inventory. On Items 1 through 7 the patient was instructed to indicate on a 0-12 point scale the number which most accurately reflected his feelings about that session. These questions inquired about (1) amount of help the patient received from the session; (2) amount of help the group received from the session; (3) the contribution the patient made to the

PAGE 29

22 session; (k) the contribution made by the group to the patients' problems; (5) the amount contributed to the session by the therapist; (6) the patient's feeling relative to how he felt before the session; and (7) the relevance of the discussion. Although there was a tendency for the subjects to circle one of the five anchor points on each item (Appendix B) , most items received acceptable variance across subjects and across sessions. On Item 8 the subject was asked to rank the group from "those who you think benefited most from today's session to those you think benefited least." On Item 9 the subject was asked to rank the group, including himself, on overall effectiveness in handling everyday tensions and anxieties. Procedure The data for this study covered a period of eight consecutive weeks from April 20, 1965, to June 8, 1965The MAACL and GPI were introduced to the group in the first session by the Clinical Psychology Department staff member who was the supervisor for the group therapists in that group. The introduction was as fol lows : I feel I know most of you since I have been observing this group from its beginning. The purpose of my coming in today is to ask your cooperation in a research project we are undertaking. From where we sit on the other side of the window, and I'm sure from where you sit in here, it is very difficult sometimes to tell how the other person feels about something despite what they say they feel or by the way they look like they feel. What we are primarily interested in is simply getting an idea before the group starts and at the end of the group of how you feel at that moment. Now if any of you have objections or strong feelings about not participating please feel perfectly free to refuse to do it. The only people who will have access to this information will be the therapists and the rest of the faculty who observe the group. It will take approximately

PAGE 30

23 three or four minutes before and after the group to check off how you feel at that moment. Look the items over, and if you have any objections you will, of course, not have to participate (Pause). Since there are no objections we will probably make this a routine part of the group meeting from now on. Thank you very much for your cooperation. Each patient was given a clipboard at the beginning of the session containing two copies of the MAACL, for preand post-measures, and a copy of the GPI . One copy of the MAACL was filled out when the therapist entered the room, and when the therapist indicated that the session was over, the other copy of the MAACL and the GPI were filled out. The patients were quite cooperative and seemed to understand what was requi red of them. Since it was thought that the Introduction of the checklist and Inventory would temporarily disrupt the normal routine of the group, it was decided that the session in which it was introduced and the session following it would probably not give an adequate estimate of the base rates for frequency, duration, silence and interruption data. Therefore, It was planned to use the two sessions following those as base rate sessions. The six sessions following these two were used to collect the data for the evaluation of the effect of planned changes In the therapists' verbal activity on the non-content measures of verbal activity. These sessions were as follows: Two base rate sessions (Bl and B2), two experimental sessions (El and E2), and two follow-up sessions (Fl and F2). The diagram below (Figure 1) outlines the data collection sequence.

PAGE 31

24 In Bl and 82 the checklist and inventory data were collected and the verbal participation measures were recorded, but no experimental manipulations were introduced.

PAGE 32

25 Norma] Therapi sts' I nteraction Therapi sts Si lent Period 3 Therapi sts Si lent Therapi sts Si lent Normal Therapi sts ' I nteract ion 0-15 min. 1670 min. 71-85 min. Figure 3Experimental Session #2 (E2) . During all sessions the therapists were instructed to make all interactions verbal, e.g., not to use gestures alone but use them only in conjunction with some verbal interaction. During the silent periods the therapists were to speai< only when asked a direct question. This occurred only two times, and both times the therapist responded in halfsecond responses. Due to difficulties in timing the shift between silence and talking, there was some talking by the therapists during silence periods in El; however, in Group 1 it amounted to only eight times for a total duration of 10. 5 seconds, and in Group 2 the therapists spoke only six times for a total of 6.5 seconds. Since most of these responses occurred in the last minute of the silence periods, they are felt to be of no real consequence. At no time during or after the experimental sessions did any patient seem aware of the planned silences. In E2 in Group 2, one of the therapists was absent because of illness. One therapist is absent for various reasons, from time to time, and the groups are accustomed to it. When this occurred previously, no disruption was evident as the groups appeared to function in a normal manner with only one therapist. Only 85 minutes of each session were scheduled to be recorded. The remaining five minutes were used for the group members to fill out the MAACL and GPI . The sessions were considered as beginning when the

PAGE 33

26 members al] completed the forms and ending when they began filling out the final forms. The mean length of the sessions for both groups was 83']3". Since the sessions did not always last 85 minutes, the final period of the session (Period 5) was reduced accordingly. Dependent variables in the study were (1) frequency of verbal responses; (2) duration of verbal responses; (3) frequency of silences; (4) duration of silences; (5) frequency of interruptions; and (6) patient moods and perceptions measured by response to the MAACL and the GPl . Operational Definition of Terms Verbal Response .--A verbal response was defined as extending from the beginning of a statement by a group member or therapist until someone else began to speak. However, if more than five seconds elapsed after someone finished responding, and the same person began speaking again, this was counted as another response. This was necessary because counting the time from one person's response till the next person began speaking would artificially inflate the duration variable. Five seconds was chosen as a cut-off point because Matarazzo (I962) reported that in his interview situations, if a patient responds to his own utterance, after a pause, he does so within five seconds 65 percent of the time. When frequency of verbal responding is reported for group data it is reported as frequency per minute (FVR) . PVR always included the therapists' responses. Durat ion . --Duration of a verbal response was the elapsed time during which a patient or therapist was actually speaking, including pauses of less than five seconds. When referring to duration of group data this is the average duration of a verbal response (DVR) for the session. DVR also included therapists' verbal responses. Total duration

PAGE 34

27 refers to the sum of the duration of all verbal responses, and is reported in relation to individuals only. lnterruption .--An interruption occurred when two or more people talked at once. This variable was difficult to record because of an apparent tendency of most people to anticipate an interruption and subsequently pause as the other person interrupts. This would not show up on the event record as an interruption. Occasionally an agreeing response such as "yes" or "uh huh" was probably recorded as an interruption because it was said while someone else was talking. These facts tend to make the interruptions the least reliable of the variables recorded. However, the operational definition used seemed the least arbitrary and consequently the most satisfactory. Statistical Treatment of the Data Early in the data collection process the problem of what to do about missing data came up as several absences occurred. Since Hypothesis I was based on an equilibrium theory, it was not theoretically proper to use any statistical way of estimating the missing data. This is clearly illustrated by the fact that there is no difference, on the variables measured, between Bl and B2 in either group even though members were absent in both groups on those days. This suggests that the other members and the therapists "make up" for the missing member and maintain verbal interaction at the same level. Therefore, there seemed to be no proper way to estimate what a missing member would have done if he had been in the group. This fact alone ruled out the use of most parametric tests as the evaluation of correlated means requires two scores of some kind for every subject.

PAGE 35

28 Another factor that went against the use of parametric tests was the high variance from subject to subject and within the same subject across sessions. Because of this it was felt that the parametric tests would not be able to make the best use of these data. Another reason, of course, was the small number of patients in each group. When more than two group means were compared the analysis was made 2 by the non-parametric Friedman 2-Way Analysis of Variance Test ( yc ) (Friedman, 1937)This test allowed the group means to be compared without reference to individual members, which was desired. In order 2 to compute v. „, each of the six sessions used. in the evaluation of Hypothesis I was broken into five-minute intervals. This gave 17 estimates of the session mean and allowed a session to be ranked 17 times In relation to any others involved In a particular analysis. Also, this method controlled for any systematic changes that might occur in the verbal activity rate during a session, as each session was compared with all others on the same time period. Slegel (1956) states that the v^ test compares very favorably with the most powerful parametric test, the F test. Reviewing a study by Friedman comparing the two tests, Siegel says "...It would be difficult or even Impossible to say which Is the more powerful test...." 2 Hypothesis IV was also evaluated by the ^ test. However, in evaluating this hypothesis, missing data were supplied. The best estimate of a subject's MAACL scores was considered to be the average of his other scores. Because of the small N Involved, the correlations required for the evaluation of Hypotheses V VIM were computed by the Spearman Rank Correlation Coefficient {r^). Slegel (1956) reports the efficiency of

PAGE 36

29 r^ "...when compared with the most powerful parametric correlations, the Pearson r, is about 91 percent." Unless otherwise specified, the p values reported in this study are for a one-tailed test.

PAGE 37

CHAPTER 3 RESULTS The results of the various statistical analyses, as they relate to the specific hypotheses, will be presented in this chapter. The raw data required for the computation of the various statistical tests are presented in the appendices. Hypothesis I When the therapists' verbal participation is reduced by planned periods of silence, the equilibrium of the group will be disrupted to the extent that significant changes will occur on the following noncontent verbal interaction variables: (a) Frequency of verbal responses (b) Duration of verbal responses (c) Frequency of silences (d) Duration of silences (e) Frequency of Interruptions This hypothesis received some support in both groups. However, the groups were not only affected to a different degree by the therapists' silences, but showed the effects in different ways. (a) Frequency of Verbal Responses (FVR) .--in order to evaluate the effect of planned silences by the therapists on frequency of verbal responses per minute, each Experimental session (El and E2) was compared with the Base Rate sessions (Bl and B2) . The periods of the Experimental sessions in which the therapists were silent (El: Periods 2 and 4; E2: Periods 2, 3, and 4) were compared with the corresponding periods in the Base Rate sessions. Also, in separate analyses, the Base Rate 30

PAGE 38

31 sessions and the Experimental sessions were compared with the Fol lowUp sessions (Fl and F2) . There was no significant difference between Bl and B2, when compared by the Wilcoxon Matched-Pairs Signed-Ranks Test, in either Group 1 or Group 2. The difference between Fl and F2 was not significant in either group when compared by the same test. However, in both groups E2 was significantly lower than El (Group 1 , p <. .05; Group 2, p <: .01; two-tai led tests) . 2 in Group 1 a Friedman 2-Way Analysis of Variance ( -^ ) involving Bl , B2, and El, was not significant (p < .50). Even a comparison of the periods in El in which the therapists were silent (Periods 2 and h) , with the same periods in Bl and B2, was not significant (p < .50). When Bl, B2, and E2 were compared, the Friedman test approached sig2 nificance ( "X = 5-38; df = 2) {p < .10), with E2 having the lowest mean. None of the other comparisons approached significance. There was no difference between Base Rate and Fol low-Up sessions or between Experimental and Fol low-Up sessions. In Group 2 the Friedman test between Bl, B2, and El was not significant (p < .95)The direct comparison of Periods 2 and k of El with Periods 2 and k of Bl and B2 was also not significant (p < -70). However, an analysis of Bl, B2, and E2 was highly significant (% ^ = 18.73; df =3) (p < .001), with E2 again having the lowest mean. The Experimental sessions differed significantly from the Fol low-Up sessions (X ^ = 25.03; df = 3) {p < -001). The Base Rate sessions also differed significantly from Fol lowUp sessions, with the Follow-Up sessions having the highest mean FVR, but the level of significance was lower ( X ^r = '1.42; df =3) (p < -01).

PAGE 39

32 Figure h shows session means of FVR for the Base Rate, Experimental, and Follow-Up sessions for both groups. Appendix C shows FVR in fiveminute intervals for the six sessions. (b) Average Duration of Verbal Responses (PVR) .-Si nee no correlation between frequency and DVR has been reported in the literature; Spearman Rank Order correlations between these two variables were computed for both groups. The correlations ranged from -.842 to -.971. with an average correlation of -.922. SESSION Bl 82 El E2 Fl F2 —Group 1 • -Group 2 Figure k. Average Frequency of Verbal Responses Per Minute (FVR) Across Six Sessions for Group 1 and Group 2. Although this correlation was known to exist between average duration and frequency, the duration measure was included in the analysis because of the possibility that the relationship might change during the experimental sessions. Because of the high correlation between these variables, the analysis of DVR data yielded essentially the same results as the FVR data. However, there were some exceptions, as noted below. In Group 1 the comparison of Bl, 82, and El on DVR was reduced from the p < .10 level of FVR to p .
PAGE 40

33 E2, which was significant on FVR (p < .05), was not significant on DVR. In those instances in Group 1 where there was a significant relationship or tendency toward significance (p < .10) on the FVR data, the analyses of DVR yielded a lower level of significance. This relationship did not hold in Group 2 where the significance levels that SESSION El E2 Fl F2 .Group —Group 2 Figure 5Average Duration of Verbal Responses Per Minute (DVR) Across Six Sessions for Group 1 and Group 2. were obtained for the FVR data were maintained in the DVR data. Figure 5 shows session means of DVR for Base Rate, Experimental, and Fol low-Up sessions for both groups. Appendix D shows DVR in five-minute intervals for the six sessions. (c) Frequency of Si 1 ences .--Because of the relative infrequency of silences and interruptions in a session, breaking these data into frequency per five-minute interval, as was done with FVR and average duration data, resulted in too many empty cells for a meaningful analysis. Therefore, silence and interruption data were analyzed over

PAGE 41

the five periods of the session rather than 17 five-minute intervals. In Group 1 the Friedman test comparing BI, B2 , and E2 for frequency of silences was significant (p < .02) , with E2 having the highest frequency. There was no difference between 81 and B2. There were also significantly more silences in the Fol low-Up sessions than in Base Rate sessions (p < -02) in Group 1 . For the same analyses in Group 2 there were no significant differences. However, there was a tendency for more silences to occur during El than in the Base Rate sessions ( ?6 ^^ = ^.90; df =2) (p
PAGE 42

35 .Group Group 2 Figure 6. Average Frequency of Silences Per Session for Group 1 and Group 2. SESSION B2 El E2 Fl F2 Group 1 _« Group 2 Figure 7. Average Duration of Silences Per Session for Group 1 and Group 2.

PAGE 43

36 Hypothesi s I I When the therapists' verbal participation is reduced by planned periods of silences, the longer the periods of silences the more disruptive they will be to the equilibrium of the group. This hypothesis is generally supported by the data in both groups. In Group 1, El was significantly different from the Base Rate sessions on only one var iabl e-i nterrupt ions , whereas E2 was significantly different from the Base Rate sessions on both frequency (p < .02) and duration (p < .001) of silences and approached significance (p < .10) on FVR. The differences were much more striking in Group 2. El did not differ from the Base Rate sessions on any variable. There was a tendency for more silences to occur in El than in the Base Rate sessions (p < .10), but this was the only variable on which El approached a significant difference from the Base Rate sessions; however, E2 was significantly lower than the Base Rate sessions on FVR and DVR (p < .001). Differences on FVR and DVR are considered to be the best indications of disruption as they tend to be the most stable from session to session. Hypothesi sill When the therapists' verbal participation returns to normal following sessions in which their verbal participation has been restricted, the equilibrium of the group will be re-established at a level more similar to the base rate level than to the level of the sessions in which the therapists' verbal participation has been restricted. This hypothesis was generally supported by the data in Group 2 with the exception of silences. This hypothesis was largely untestable in Group 1 as few significant changes occurred in either El or E2. In Group 1, however. Figure 1 shows that after E2 (which was the lowest)

PAGE 44

37 mean) there was a gradual return to the Base Rate level. On frequency of silences and duration of silences the differences between the Base Rate and Fol low-Up sessions were significant, whereas the differences between Experimental and Fol low-Up were not significant. This is contrary to the hypothesis. In Group 2, although the Follow-Up sessions were significantly different from the Base Rate sessions on FVR (X ^ = 11.42; df = 3) (p < .01), the difference was of greater significance between Follow2 Up and Experimental sessions ( % j, = 25-03; df = 3) (p
PAGE 45

38 One person in each group was left out of this analysis because each missed three or more sessions. For those members who were absent two times or less, their mean preand post-MAACL scores, determined from the sessions they attended, were substituted for the sessions missed. Each subject was then ranked across the six sessions by change scores on the MAACL for each of the three affect var i ables--anxi ety , depression, and hostility. A Friedman 2-Way Analysis of Variance test was calculated for each variable and none was found to be significant (see Figures 8 and 9) • In addition, the change scores for each individual on the three variables were summed for each session both algebraically and without regard to sign. Separate analyses by the Friedman test on these data were also not significant. Hypothesis V Patients' perceptions of the amount of help received from the group, the amount they contributed to the group, and their own overall effectiveness in handling their problems will vary positively with changes in the amount of their own verbal responses. This hypothesis was generally supported by the data. In the evaluation of this hypothesis, as in Hypothesis IV, one patient from each group was excluded from the analyses because of poor attendance. Since this hypothesis related primarily to level of verbal participation in general and not directly to the experimental sessions, all eight sessions on which data were available were used. The analysis of this hypothesis involved relating each individual's verbal participation measures to his responses to GPI items and not to measures relating specifically to a particular group; consequently the subjects in both groups were pooled and one overall analysis was made.

PAGE 46

39 Figure 8. Mean Anxii MAACL Change from Pre-Group to Post-Group Testing on ;ty, Depression, and Hostility in Group 1.

PAGE 47

4.00 3.00 2.00 B2 40 SESSION El E2 Fl F2 .00 ,00 2.00 -3-00 4.00 Depress! on Host! ] i ty Figure 3Mean MAACL Change from Pre-Group to Post-Group Testing on Anxiety, Depression, and Hostility in Group 2.

PAGE 48

41 Although there Is generally a high correlation in a given session between the rankings of the patients on frequency of verbal responses and total duration of time spoken, there is little correlation from session to session between a patient's frequency and his total duration. In other words, the person who talks the highest number of times In a session Is generally the one who talks longest, but the session in which a given patient talks the highest number of times may not be the session In which he talks longest. Since these measures may be getting at different aspects of the subject's participation in the group, both were used in the evaluation of this hypothesis. When Spearman Rank Order correlations were computed between the amount of help patients felt they received (GP I , Item 1) and the total duration of their verbal responses over eight sessions, it was found that they were highly correlated. Of the eleven correlations computed between these two variables, five were significant beyond p < .05The combined probability of this many significant correlations occurring by chance Is p < .0001. Therefore, the null hypothesis of no relationship Is rejected. However, there does not appear to be a significant relationship between frequency of verbal responses In a session and the amount of help patients feel they received. Out of the 11 correlations computed between these variables, only one was significant beyond p < .05Both frequency and total duration of verbal responses are related to the amount the patients felt they contributed to the group (GPI, Item 3). Four of 11 of the correlations between frequency and "amount contributed" were significant (beyond p < .05). The combined probability of this occurring by chance is p < .002. Three of 11 correlations

PAGE 49

kz between a patient's frequency of verbal responses and "amount contributed" were significant (beyond p < .05)This is significant beyond p < .02. In evaluating the relationship between total duration of speaking and effectiveness in handling their problems (GPI , I tern 3), the subject's rankings of himself in relation to the rest of the group for the sessions he attended were ranked, and then compared to the ranks of the total duration of speaking, for that subject for those sessions. Although only one patient's correlation between total duration and "effectiveness" was significant (p < .01), there appeared to be an interesting relationship between these variables. Of the 11 patients for which this correlation was computed, nine were negative. The two patients who did not have negative correlations had an r=.000, as those patients did not change their ratings of themselves on any session. One of these patients ranked himself last every time and the other ranked herself in the middle every time. Hypothesis Vl Patients' perceptions of the amount of benefit each patient received from the group will be a function of the amount of each patient's verbal participation. The evaluation of this hypothesis involved averaging the rankings by the patients of who benefited most to who benefited least in each session (GPI, Item 8). The averages were ranked and correlated with actual ranks of each patient's actual verbal participation measures for that session. This hypothesis was supported by the data on both groups as 7 out of 15 correlations were significant beyond p < .05The combined probability of 8 out of 15 correlations occurring by chance i s p
PAGE 50

43 over the sessions and rankings of "benefit", 4 out of 15 were si gni f i cant (p < .01 ) . Hypothesis V! I Patients' perceptions of the amount of benefit each patient received from the group will be a function of the amount of each patient's verbal participation. This hypothesis was not supported by the data. in order to evaluate this hypothesis the group averages on response to GPi, item 2, were ranked across eight sessions and then correlated with the FVR for each session. In Group 1 the correlation between FVR and ratings of the amount of help the group received was not significant (r = -.357)However, in Group 2 the correlations, although not significant (p 4 .10, twotailed test) approached significance in the direction opposite that predicted (r^ = -.659) • in Group 1 there was not a significant correlation between average ratings of the "amount the group contributed" and FVR. In Group 2 there was a significant relationship between the amount the group contributed and FVR (p <.05, two-cailed test). Again the correlation was negative (r^ = -.771). Hypothesi s VM 1 Patients' perceptions of the therapists' contribution to the group and the degree to which the therapists kept the group discussion on relevant problems will vary positively with the amount of therapists' verbal participation in the group. This hypothesis was only partially supported by the data. This hypothesis was evaluated in the same manner as Hypothesis VII above. in Group 1 the correlation between the amount the patients felt the therapists contributed (GPI, Item 5) and the total duration of

PAGE 51

44 therapists' verbal responses was r^ = .589, which was not significant. However, perceived therapists' contribution and frequency of therapists' verbal responses (r = .645) was significant (p < .05)in Group 2 the correlation between therapists' contribution and total duration of therapists' verbal responses (r = .726) was significant (p < .05)Frequency of therapists' verbal responses and perceived therapists' contribution was not significant (r = .333)in neither Group 1 nor Group 2 was there a significant relationship between the degree to which patients felt the therapists kept the discussion on relevant problems (GPi , item 7)
PAGE 52

CHAPTER 4 DISCUSSION Group Equi 1 ibr ium The results of this study lend further support to the concept of group equilibrium as described by Chappie and Coon (1942) by demonstrating its applicability to the group psychotherapeutic process. These authors state that a group is in equilibrium if its interactions are constant within clearly defined limits, and if after a disturbance takes place the rates return to their previous values. Constant rates were clearly evident in this study as the base rate sessions in the two groups did not differ from each other on any of the variables on which measures were obtained, i.e., frequency of verbal responses, average duration of verbal responses, frequency and duration of silences, and frequency of interruptions. This is in spite of the fact that in both groups different patients were absent on the two days. The latter part of Chappie and Coon's requirement for a group in equilibrium can be demonstrated by what happened after the Experimental sessions. After session E2 there was a return, or tendency to return, to the base rate level on practically every variable. In every case, except on silences, the Fol low-Up sessions were more similar to the Base Rate sessions in both groups than the Experimental sessions were. The fact that silences did not show a similar return toward base rate levels may have been the result of the therapists being more comfortable 45

PAGE 53

46 with silences in E2 as a result of their experience in El, and may not reflect what the data would look like if this possible contaminant could have been controlled. This will be discussed further in the following section. Therefore, the groups used in this study satisfied the general requirements for equilibrium as defined above. The fact that the groups maintained constant rates on the previously mentioned variables even when different group members were absent simplified the task of assessing the hypothesized role of the therapists in group equilibrium. Effect of Therapists' Verbal Behavior on Group Equilibrium In the first experimental session when the therapists' verbal participation was reduced by two planned periods of silence of 20 minutes each, the only significant effect in either group was that Group 1 had significantly more interruptions in that session than in the Base Rate sessions. Therefore, patients were able to maintain the established pattern of verbal interaction for these relatively short periods of time. Since the therapists did not participate verbally during these periods, this suggests that departures from the base rate level of interaction for short periods by the therapists may increase patient interaction. This is in line with the results of a study by Salzberg (1962) in which he found that therapist's silence in group psychotherapy led to significantly more interaction (patient-patient conversation) than did talking by the therapist. His silence periods were of only ten minutes' duration. The importance of the length of the time the therapist does not participate verbally in the session is indicated by the results of the second Experimental session in which the therapists were silent for a

PAGE 54

hi continuous period of 55 minutes. in Group 2 there was a significant reduction in the frequency of verbal responses per minute and a corresponding significant increase in the duration of the responses. There were, however, no significant effects on the frequency and duration of silences. This indicates that although the level of verbal interaction was reduced, it was not brought about by long or frequent silences. This means that the percentage of the time someone was talking in E2 did not vary from the Base Rate sessions. This clearly indicates that the therapists are responsible for most of the interaction between patients in this group. In Group 1 the effect seemed to be quite different. Significantly more silences occurred in the second Experimental session than in the base rate sessions. Although not significant there was a tendency (p <, .10) toward fewer verbal responses per minute during this session than in the base rate sessions. However, the average duration of responses did not change. Therefore, any reduction in the verbal interaction was the result of more and longer silences, rather than reducing interaction by increasing duration of responses as in Group 2. Presumably, a number of these silences would normally have been terminated by the therapist, but when the therapist did not respond the group was able to keep the interaction going at a level which was only slightly reduced. Group 2 seemed to lack this ability. The results in E2, in both groups, indicate that when the therapists do not verbally participate in the sessions for relatively long periods that there is a reduction in the verbal interaction rather than an increase. Therefore, it appears that Salzberg's results (1962) may not be generalized beyond short periods (10 to 20 minutes) of therapist

PAGE 55

k8 There was evidence that suggested that some of the effects found in the longer silence period may have been due to a carry-over effect from the first Experimental session. Appendix A (Group 1) shows that for E2, Periods 1 and 2 have lower mean FVR than in El. There were also more silences in these two periods of E2 than in El (Appendix E) . One would not expect any differences in these periods as the conditions, to this point, were the same for both sessions. Although a carry-over effect cannot be ruled out, several factors obviate its importance. The first is that the difference between these two periods of El and E2 on FVR are not significant. The second is that there were apparently few differences between El and the Base Rate sessions and consequently there should be only minimal carry-over effects. The last, and probably most s i gni f leant, factor relates to comments by the therapists about their own performance in these sessions. All of the therapists were surprised at what happened in El, and the general feeling was that they did not need to participate verbally as much as they had in previous sessions. Consequently, in E2, from remarks made by the therapists, they seemed to be more prone to let a silence continue than they would have previously. This would not only account for the high number of silences occurring in the first two periods of the session in Group 1, but also explain why the number of silences continued at a high rate through the Fol low-Up session in both groups. From the previous discussion it is clear that both groups did not respond in the same way to similar experimental conditions. The two groups used in this study provided a good contrast in rates of verbal interaction. Consequently, the results pointed up the need to control

PAGE 56

49 this variable in future research, as base rates seem to be important in understanding the effect of the therapists' verbal activity in group psychotherapy. The data on the two groups in this study suggest that in a highly verbal group the verbal interaction patterns may be relatively unaffected by prolonged therapists' silences, or if they are affected, they seem to be able to recover base rate patterns of verbal interaction quickly. On the other hand, a low verbal group may have a greater dependence on the therapists for maintaining verbal interaction patterns, and when the therapists are silent for extended periods the group interaction rate is reduced significantly. In Group 2 there apparently was a carry-over effect after E2 because the Fol low-Up sessions had a significantly higher FVR than the Base Rate sessions. However, when Bl, B2, and F2 were compared, they were not significantly different, indicating the recovery of Base Rate levels on this variable within two sessions. Changes in Mood Hypothesis IV relating to expected affect changes, measured by the MAACL, during the sessions in which the therapists' verbal participation was reduced, was not supported. This suggests that even when the pattern of verbal Interactions of the group (Group 2) were changed significantly, there were no consistent affective changes In the members of the group. Other evidence tended to support this result. On the GPI, I tern 6, which asked how the patient felt at the end of the session relative to how he felt when he came, three of five patients In E2, Group 2, said they felt better, and the two others said they felt the same way as when they came.

PAGE 57

50 A Rank Order correlation was computed for each patient between his total change score on the three variables of the MAACL and his responses to item 6 of the GPi , and in spite of the low variance on both variables, 3 of the 11 correlations were significant. The probability of this occurring in the absence of a relationship is p < .02. This seems to add some weight to the validity of both measures. Several factors seemed to work against a significant negative change, i.e., feeling worse, which was predicted. One factor is that it is thought that the difference scores from preand post-testing on the MAACL were attenuated by the fact that during Period 5 the therapists were allowed to interact freely, thereby partially restoring equilibrium, especially in E2 , and consequently probably resolved some of the feelings that may have been aroused earlier in the session. Since the GPI was also administered at the end of the session it was probably affected in the same way. Another factor that may have worked against significance is suggested by Zuckerman's report (I960) of high and significant correlations over an eight-day period with the MAACL in a psychiatric population, whereas with normals the correlation for the same period was low and insignificant. This suggests that because psychiatric patients have a higher base rate on this test than normals, that they may be relatively insensitive to anything but major changes in affect, at least as measured by this instrument. Although some patients did change markedly in a given session, the average changes that occurred were quite small. Preand post-changes were also attenuated by the fact that given a high initial response the freedom to vary upward is reduced. However, the fact is that not only did patients not change much,

PAGE 58

but they did not all change in the same direction. Some felt better and others felt worse, e.g., more anxious, depressed, or hostile. Patients' Verbal Participation and Their Perceptions of Self, t he Group, and the Therapists "^ ~ ~ 'The tests of Hypothesis V indicate positive relationships between both the amount of help a patient felt he feceived from the session and the amount he felt he contributed to the session with how much he actually talked during the session. This tends to confirm the results of Sechrest and Barger (I96I), who obtained the same results. However, the present data, which included both frequency and total duration of verbal responses, suggests a refinement of this relationship. Total duration of verbal activity was highly related to perceived benefit from the session, whereas frequency of verbal responses was not. If we can assume that when a person has a high total duration with a low frequency of responses he is probably talking about himself, whereas when frequency is high he is spending a lot of time interacting with others and may or may not be talking about himself, an interesting hypothesis is suggested. That is, patients may feel they receive more from a group session when they talk about themselves more. However, they feel that they are contributing to the session when both talking about themselves and interacting with others. The negative correlations between "overall effectiveness in handling one's problems" and total duration were unexpected. The initial interpretation of these results was that the more a patient talked, the less effective he felt in relation to the rest of the group. However, this didnot seem to fit with the findings above. When examined in the

PAGE 59

52 group context a more tenable explanation was discovered. That is, whan a patient is silent others are usually telling about themselves in one way or another and, consequently, he may feel that in relation to other patient's problems, his are not very bad. Therefore, it may be that it is not only what the patient is doing that determines his perceptions of his effectiveness in handling his problems, but also of what the other patients are doing. According to the tests of Hypothesis Vl, patients tend to rate other patients as benefiting from the session according to the amount those participate in it verbally. However, here again there seems to be a distinction between frequency and duration as the correlations of frequency and rankings of benefit were much more significant (p < .0001) than total duration and the same rankings (p < .01). This suggests a stronger relationship between the frequency and benefit than duration and benefit. It appears that patients do not see the one who talks most, e.g., long monologues, etc, as the one who benefits most, but the one who interacts most, e.g., talks the highest number of times. As discussed earlier, patients tend to feel they benefit more when they talk more (high total duration). However, this suggests that they do not tend to perceive other patients as benefiting more when they talk more, but rather when they interact more. The results relating to Hypothesis Vli are quite confusing as there was no consistent pattern across groups or within groups. There was a tendency in Group 2 for patients' perceptions of the amount of help the group received and FVR to be negatively correlated. There was a significant (p 4. .05) negative correlation (r^ = -.771) between the group responses to "amount the group contributed" and FVR.

PAGE 60

53 In Group 1 the correlation v;as positive but not significant. The findings on this hypothesis generally do not coincide with the other results obtained. The lack of significance suggests that the patients were responding to different aspects of the session when answering the question or were interpreting the GPI item differently. The significant negative correlation In Group 2 cannot be adequately explained within the limits of these data. The relationship between therapists' verbal participation and patients' perceptions of the therapists' contribution to the group appears to be a positive one. As one would suspect, the relationship between frequency and duration is not as clear in these comparisons as it was in some of the others. In Group 1 correlation between frequency and perceived therapists' contribution and total duration and perceived therapists' contribution both approached significance (p < .05). However, in Group 2 the correlation between total duration and perceived therapists' contribution was significant (p < .05), whereas the correlation between frequency and perceived therapists' contribution was insignificant. These correlations suggest that whenever the therapist is talking he is perceived by the patients as contributing to the group, whether the duration of his responses is long or short. Therefore, frequency and duration may measure much of the same thing as far as therapists' verbal responses are concerned. The hypothesized relationship between the degree to which patients feel the therapists kept the discussion on relevant problems and both frequency and duration of the therapists' verbal responses was not supported. One reason for this may be the fact that most patients tended to rate the discussion high on relevancy throughout the sessions.

PAGE 61

54 It may be that patients tend to perceive any behavior by the therapist as relevant to the problems of the group. The results of this study show that Chappie and Coon's concept of group equilibrium is applicable to psychotherapy groups. Base Rate values for the non-content verbal interaction variables which were studied varied relatively little. After the Experimental sessions had made significant changes on several variables there was a return, or tendency to return, to the Base Rate level in following sessions on all of the variables except silences. The evaluation of the effect of the therapists' verbal behavior on group equilibrium indicated that group base rates play an important part in determining the therapists' affects on the group. In a high verbal group the therapist does not seem to play as important a role in maintaining normal verbal interaction patterns as he does in a low verbal group. At least in this study when the therapists' verbal participation was reduced by planned periods of silence, there was a significant reduction in group interaction in the low verbal group (Group 2); whereas in the high verbal group (Group 1) there was a slight^ but insignificant, reduction. The present results indicate that in evaluating the effects of planned therapists' silences on the group an important variable is the length of the silences. In this study two 20-minute periods of therapists' silence between three normal periods of 15 minutes each had little effect on either group, as the patients were able to maintain previous base rate levels without the therapists. The only

PAGE 62

55 exception to this was in Group 1 which had significantly more interruptions on this day. However, in the following session in which the therapists were silent for a period of 55 minutes between two normal periods of 15 minutes, several significant changes occurred. Group 1 had significantly more and longer silences in this session than in the Base Rate sessions and slight reduction in group interaction. Group 2 showed a significant reduction in verbal interaction. Therefore, therapist silence in group psychotherapy seems to have different effects on the group depending on its duration. The hypothesized changes in mood, or affect, during the Experimental sessions was not supported. Although several methodological problems may have contributed to the lack of significance, there were no uniform changes among the group members. Some made positive and some made negative changes. it appears that patients' affective responses to changes in the verbal interaction patterns of a group differ from individual to individual. Some patients may be upset and others may feel better. A positive relationship was found between the amount of help a patient felt he received from the group and his total duration of speaking in a session. There was no relationship between the amount of help a patient received from the session and the frequency of his verbal responses. Since a high frequency of verbal responses tends to indicate a high Interaction with others in the group, it appears that patients may feel they benefit most not from interacting at a high rate, but talking at a high rate. There was also a significant relationship between both an individual's frequency and duration of verbal responses and the amount he felt he contributed to the group.

PAGE 63

56 Patients' perceptions of their own overall effectiveness in handling their problems tended to be negatively correlated with the total duration of their verbal responses. The positive relationship suggested from this is that the more other group members talk, the more effective the patient feels. This tends to confirm a reason often given by patients in group psychotherapy as to the benefit they receive from the group. It is that by being in group they find out that they are not the only ones with problems, and that some people have problems worse than their own. There was a significant positive relationship between the rankings of the group on who benefited most to who benefited least and their actual rank on both frequency and duration of verbal responses. However, the relationship between frequency and "benefited" was much stronger than duration and "benefited." This suggests that patients feel that other patients receive the most benefit from the group when they interact the most. Patients' perceptions of the amount of help the group received and the amount the group contributed did not appear to be meaningfully related to the group's frequency of verbal responses per minute. One explanation for this probably lies in the amb I guousness of these items on the instrument used. There was a positive relationship between the frequency and duration of the therapists' verbal responses and the amount the patients felt they contributed to the group. However, the degree to which patients felt that the therapists kept the discussion on relevant problems, and the amount of therapists' verbal participation in the session, did not appear to be related.

PAGE 64

57 Suggestions and Implications for Future Research The most obvious implications for future research resulting from this study seems to be a need for further knowledge of and control of base rate variables of verbal interaction in group psychotherapy research. Because of the stability of these variables, they may provide a much more reliable and effective way of detecting change in the group over time, or as a result of experimental manipulations, than more traditional subjective approaches. Just as Cook (196^1-) suggests that percentage of silence in an individual psychotherapy session is an index to the "climate" of the session, so may several of the base rate variables used in this study indicate the same thing about group psychotherapy. In such case, a knowledge of base rates and the ways in which given variables effect them would add immeasurably to the understanding of group psychotherapy. The present study indicates that the comparison of groups with different base rates would be profitable. Knutson's study (I963) of quiet and vocal non-therapy groups provides the general model, and in fact asks many questions that would be applicable to group psychotherapy, e.g., patients' satisfaction with the group, patients' progress, etc. Because of the differences found in this study with regard to the length of the therapists' silences, a more detailed study of this variable is needed. There was an indication that short silences have little effect on the group level of verbal interaction and may even facilitate patient-patient interaction; whereas long silences reduce interaction. This, of course, suggests that the relationship between the amount of group interaction and length of planned therapists'

PAGE 65

58 silences is curvilinear. The optimal amount of therapist's activity for the best group interaction is not known. Although some inroads are being made into an understanding of the influence of the therapist's verbal activity in groups, the relationship is obviously complex and needs considerably more research. Several interesting relationships were suggested by the data regarding perceived benefit and verbal participation. One of the most interesting was the suggested relationship between frequency and durations of an individual's responses. This study indicates that they are probably getting at somewhat different aspects of the patient's verbal performance in the group and, therefore, both should be used. Several studies have used only frequency and may have seriously limited their findings by so doing.

PAGE 66

CHAPTER 5 SUMMARY Two aspects of verbal behavior in group psychotherapy were investigated in the present study. First, based on Chappie and Coon's theory of group equilibrium, and the results of Lennard and Bernstein, and Matarazzo, Weins, and Saslow with regard to the role of the therapist in maintaining equilibrium in individual psychotherapy, several hypotheses were advanced about equilibrium in group psychotherapy. In general, it was hypothesized that the group therapist plays a major stabilizing role In the group through his verbal behavior, and when he is restricted in verbal activity the equilibrium of the group will be upset and increases will occur in patient anxiety, depression, and hostility. Secondly, several hypotheses were made about the relationship between patients' verbal participation in group psychotherapy and their perceptions of themselves, the therapist, and other members of the group. Six consecutive 85-minute sessions were used to evaluate the effects of the therapists' verbal activity on group verbal participation and mood in two ongoing heterogeneous psychotherapy groups. Both groups had co-therapists. The first two sessions were used to establish base rates on the following non-content verbal interaction variables: Frequency and duration of verbal responses, frequency and duration of silences, and frequency of interruptions. In session three, both therapists were silent for two periods of 20 minutes each. In session four the therapists 59

PAGE 67

60 were silent for a period of 55 minutes. The fifth and sixth sessions were used to evaluate the effect of the experimental sessions on succeeding sessions. Difference scores, obtained from pre-group and post-group testing with an adjective checklist (MAACL) were used to evaluate the effect of the experimental sessions on mood. The relationship between verbal participation and "patients' perceptions" was investigated by correlating patient responses to a group perception inventory, administered in each session with individual, group, and therapists' measures of verbal participation. The results of this study indicate that Chappie and Coon's concept of group equilibrium is applicable to psychotherapy groups. Constant base rates on non-content measures of verbal interaction were clearly evident in both groups. Following the departure from these rates during the experimental sessions there was a return to near previous values on all variables except frequency of group silences. The effect of therapist silences on group verbal activity must be interpreted in the light of at least two variables: Base rate of verbal responses in the group and length of the therapists' silences. Longer silences tend to reduce group interaction, whereas shorter silences have very little effect on the established patterns of verbal interaction. Generally speaking, the "high verbal" group showed few significant effects from the long silence, whereas the "low verbal" group showed marked changes in both frequency and duration of responses which are the most stable of the base rate variables. There were no significant changes in anxiety, depression, and hostility as a result of the experimental manipulations.

PAGE 68

61 A positive relationship was found between the amount of help a patient felt he received from a session and his total duration of speaking in that session. There was also a positive relationship between the amount a patient felt he contributed and both the duration and frequency of his verbal responses. There was a highly positive relationship between the rani
PAGE 69

APPENDICES

PAGE 70

APPENDIX A MAACL Name Below you will find words which describe different k and feelings. Check the words which describe how you fee of the words may sound alike, but we want you to check al that descr ibe your feelings. Work rapidly. i nds of moods 1 now . Some 1 the words act I ve adventurous affect ionate af rai d _agi tated agreeable aggressi ve al i ve alone amiable amused angry annoyed awful bashful _bitter b lue bored calm cauti ous cheerful clean complai ni ng contented contrary cool cooperative cr i ti cal cross cruel dar ing desperate destroyed devoted di sagreeable di scontented di scouraged di sgusted displeased energetic enraged enthus last i c fearful f i ne _fit forlorn frank free f ri endly frightened furious .gay gent le glad g 1 oomy good good-natured grim _happy heal thy hopeless host i le mpat i ent ncensed ndignant nspi red nterested rr i tated ealous oyful kindly lonely _lost lovi ng low lucky mad mean meek merry 63

PAGE 71

64 APPENDIX A (Continued) mi Id

PAGE 72

APPENDIX B GPI On Questions 1-7 please circle tiie number on the scale that indicates best your feelings about today's session. You may circle any number and not just those with words or phrases under them. Use the words and phrases to help you determine the number that best descri bes your feel i ngs . From today's session I received: 12 11 10 9 / / / / 7

PAGE 73

66 APPENDIX B (Continued) 5. In today's session the therapists contributed: 12 11 10 98765^32 1 //////////// / a a lot some very nothing great ] i tt le deal After today's session I feel /

PAGE 74

APPENDIX C Average Frequency of Verbal Responses Per Minute (FVR) in Five-i^inute Intervals

PAGE 75

68 APPENDIX C (Continued)

PAGE 76

APPENDIX D Average Duration (in seconds) of a Verbal Response Five-Minute Intervals

PAGE 77

70

PAGE 78

APPENDIX E Frequency of Silences Occurring During Five Periods of Each Session Group Session Period BJ B2 El E2 Fl F2 1 6 4 12 19 6 7 2 2 6 5 9 8 7 3 1 2 11 1 4 3 3 4 13 8 5 5 016738 Total 11 15 27 50 36 28

PAGE 79

APPENDIX F Average Duration of Silences for Five Periods of Each Session Group 1

PAGE 80

APPENDIX G Frequency of Interruptions Occurring During Five Periods of Each Session

PAGE 81

APPENDIX H Percentage of Session Time Spent Talking Group 1 Sess ions Intervals 1 2 3 k 5 6 7 8 9 10 11 12 13 14 15 16 17 Bl .823 .883 ,860 868 ,862 913 .867 903 .893 .9^2 907 .895 .898 ,882 .9^0 958 .870 .892 B2 .962 780 818 .838 .827 .858 .950 .983 .950 • 9^2 9^2 ,880 .925 972 .957 .923 .875 .905 El .827 .817 .637 • 925.805.930-' .972-'' .897 .995 .955 .915^' .9'+5^' .933^' .765' .852 .845 .916 .878 E2 .510 .695 .742 .850--' .933.733.573.913.850-' .947.632-.' .813.942-' .945-' .892 .893 .831 .806 Fl .873 782 .858 793 .883 770 .848 .938 798 .842 .855 .403 775 .908 778 .913 .814 ,814 _F2 930 • 795 780 .907 .952 785 ,882 918 918 .948 ,882 952 • 963 .828 .813 780 .887 .878 "Indicates periods of therapists' silence. 74

PAGE 82

75 APPENDIX H (Continued)

PAGE 83

APPENDIX I Frequency of Verbal Responses of Patients and Therapists for Eight Consecutive Sessions

PAGE 84

APPENDIX J Total Duration (in seconds) of Verbal Responses of Patients and Therapists

PAGE 85

APPENDIX K MAACL Difference Scores for Anxiety, Depression, and Hostility Group 1 ANXIETY-'--

PAGE 86

ANXIETY-' 79 APPENDIX K (Continued) Group 2

PAGE 87

APPENDIX L List of Rank Order Correlations Relating to Hypotiiesis V TABLE 2 Ranl< Order Correlations Between Patients' Perceptions of Amount of Help They Received (GPi, Item 1) and Total Duration of Their Verbal Responses Pat i ent

PAGE 88

APPENDIX L (Continued) TABLE 4 Rank Order Correlations Between Patients' Perceptions of Amount They Contributed to the Group (GPi, Item 3) and the Total Duration of Their Verbal Responses Patient

PAGE 89

82 APPENDIX L (Continued) TABLE 6 Rank Order Correlations Between Patients' Perceptions of Effectiveness in Handling Their Problems (GPI , item 9) and Total Duration of Their Verbal Responses Patient

PAGE 90

83 Date 4-20 4-27 5-4 5-11 5-18 5-25 6-1 6-8 APPENDIX L (Continued) TABLE 8 Rank Order Correlations Between Group Rankings of "Benefit" (GPI, Item 8) and Rankings of Patients on Both Frequency and Duration of Verbal Responses Group 1 Dur ./Benef i t • 771 .787 . 829-'.886^v • 593 .429 .000 Freq ./Benef i t 771 897 829886-.' 593 429 320 Group 2

PAGE 91

REFERENCES Anderson, C The relation between speaking times and decision in the employment interview. J. aop. Psychol., I960, 44, 267-268. ~ — — Chappie, E. D., and Coon, C. S. Principles of anthropology . New York: Holt, 1942. Cook, J. J. Silence in psychotherapy. J. counsel. Psychol., 1964, ]±, 42-46. ~ — Dreikurs, R., and Corsini, R. Twenty years of group psychotherapy. Am. J. Psychiat ., 1954, JJO^, 567-575Friedman, M. The use of ranks to avoid the assumption of normality implicit in the analysis of variance. J. Amer . stat. Asso., 1937, 32, 675-701. Goldman-Eisler , Freida. individual differences between interviewers and their effect on the interviewer's conversational behavior. J. ment . Sci ., 1952, 98, 66O-67I • Hankoff, L. , Rudorfer, L. , and Paley, H. A reference study of ataraxics: A two week double bl i nd outpatient evaluati on. J. new Drugs , 1962, 2, 173-178. ~ Kendall, M. G. Rank correlation methods . London: Griffin, 1948. Kew, J. K. , Rickard, H. C, and Timmons, E. 0. The stability of group verbal behavior. Psychol . Rec , I962, _1_2, 323-325. Knutson, A. L. Quiet and vocal groups. In M. M. Rosenbaum and M. Berger (Eds.) Group psychotherapy and group function . New York: Basic Books, inc. , I963. Lennard, H. L. , and Bernstein, A. The anatomy of psychotherapy : Systems of communication and expectation . New York: Columbia Univ. Press, I960. Lubin, B., and Lubin, S. Bibliography of group psychotherapy I956-I963. G£. Psychother ., 1964, jj, 177-230. 84

PAGE 92

85 McFarland, R., Daniels, R., and Lieberman, M. Group psychotherapy. In E. A. Spiegel (Ed.) Progress in neurology and psychiatry . New York: Greene and Stratton, I963. Matarazzo, J. D. Prescribed behavior therapy: Suggestions from interview research. In A. J. Bachrach (Ed.) Experimental foundations of clinical psychology . New York: Basic Books, Inc., 1962. Matarazzo, J. D., Saslow, G., Weins, A. N., Weitman, M., and Allen, B. V. Interviewer headnodding and interviewee speech durations. Psychother . , 1964, j_, 54-63. Matarazzo, J. D., Weins, A. N., and Saslow, G. Studies in interview speech behavior. In L. Krasner and L. P. Ullman (Eds.) Research i n behavior modi f i cat ion : New developments and thei r cl inical impl icat ions . New York: Ho 1 1 , Ri neharTT and Winston, 1964. Powdermaker, Florence, and Frank, J. D. Group psychotherapy . Cambridge, Mass: Harvard Univ. Press, 1953. Pratt, J. The class method oftreating consumption in the homes of the poor. J.A.M.A,, 1907, 49, 755-758. Salzberg, H. D. Effects of silence and redirection on verbal responses in group psychotherapy. Psychol . Rep. , 1962, V\_, 455-461. Saslow, G., and Matarazzo, J. D. A technique for studying changes in interview behavior. In A. P. A. Research in psychotherapy . Washington, D. C, 1959. Sechrest, L. B., and Barger, B. Verbal participation and perceived benefit from group psychotherapy. Int . J. Gp. Psychother . , 1961, jM, 49-59. Seigel, S. Non-parametric statistics : For the behavioral sciences. New York: McGraw-Hill, 1956. Smith, A. B., Bassin, A., and Froehlich, A. Changes in adult offenders during group therapy. J^. consult . Psychol . , I96O, 24, 247-249. Timmons, E. 0., Rickard, H. C., and Taylor, R. E, Reliability of contentfree group verbal behavior. Psychol . Rec . , I96O, _1£, 297-305. Wilkinson, B. A statistical consideration in psychological research. Psychol. Bull., 1951, 48, I56-I58.

PAGE 93

Wirt, R., and Wirt, Anne Psychotherapeutic processes. In Farnsworth, P. (Ed.) Annual review of psychology . Palo Alto, Calif: Annual Reviews, inc., I963. Zucl
PAGE 94

BIOGRAPHICAL SKETCH Aubrey Clise Daniels was born May 17, 1935, in Lake City, South Carolina. He attended public schools in Lake City and entered Furman University in 1953He graduated from Furman University in 1957 with the degree of Bachelor of Arts. At that time he also received a commission in the United States Army. During his military service he was stationed in Korea and at Fort Rucker, Alabama. In September, I960, he entered the University of Florida to work toward the degree of Doctor of Philosophy. Since then he worked for two years as a graduate assistant to the Dean, College of Health Related Professions and for two years was a United States Public Health Fellow. He completed an internship in Clinical Psychology at the J. Hillis Miller Health Center, University of Florida, Gainesville, Florida, in June, 1965He is married to the former Rebecca Lanette Tapp of Greer, South Carolina. They have a daughter, Laura Lee. 87

PAGE 95

This dissertation was prepared under the direction of the chairman of the candidate's supervisory committee and has been approved by all members of that committee. It was submitted to the Dean of the College of Arts and Sciences and to the Graduate Council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Phi losophy. December, 1965 Dean, Coll Dean, Graduate School Supervisory Committee; Lc^ Chai rman

PAGE 96

44 17 f9