CHANGES III THE BEHAVIOR OF CLINIC - REFERRED
CHILDREN DDRIIID PLAY INTERACTIONS WITH
TRAIIIED RHD UR'IRAIIIED _
IIIIDERRRADIIAIES

Thesis for the Degree of M. A.
MICHIGAN STATE UNIVERSITY
LORETTA R. LAURENITIS
1974

a o-...-,.-

"-9”

 

‘Mmmc
‘ "ng 1. ,

L I83? 31 Q 4 ‘1‘";
Unifictsigx

 
     
   

 

ABSTRACT.

CHANGES IN THE BEHAVIOR OF CLINIC-REFERRED
CHILDREN DURING PLAY INTERACTIONS WITH
TRAINED AND UNTRAINED UNDERGRADUATES
BY

Loretta R. Laurenitis

The present study was designed to investigate the process
of play encounters between children and trained and untrained
high and low potential undergraduates in reference to six di-
mensions of children's behavior - Leadership, Dependency,
Affection, Aggression, Contact and Role-playing. On the basis
of theoretical assumptions, clinical observations, and re-
search findings, the development of particular patterns
for each of the six variables was predicted. Leadership
was expected to increase over sessions,diminishing in the
last phase as the child engaged in more egalitarian activi—
ties with the undergraduate. Dependency and Role-playing
were expected to decline in frequency over time, whereas
Contact was expected to increase over sessions. It was
anticipated that Aggression would also increase in frequency,
reaching a peak and then declining. For Affection, the trend
expected was a decline in frequency followed by a rise in
occurrence in the later phases as negative feelings were
expressed and resolved. The following specific hypotheses
were then formulated in reference to the six behavioral vari-
ables: 1) Significant differences over sessions would be

found for the dependent variables in the direction of the

Loretta R. Laurenitis

anticipated trends. 2) The behavior of children interacting
with trained undergraduates would more clearly fit the ex-
pected patterns than the behavior of children encountering
untrained undergraduates. 3) The behavior of those children
playing with high potential undergraduates would more closely
approximate the predicted patterns than the behavior of
children engaged with low potential undergraduates. 4) Train-
ing would tend to reduce the differences in behavior between
children interacting with high potential undergraduates and
those playing with low potential undergraduates. However,
children playing with trained high potential undergraduates
would show greater levels of the dependent variables in the
predicted directions.

Subjects were twenty—seven clinic-referred children di-
vided into the following four groups depending upon the type
of undergraduate with whom they interacted: Trained High
Potential (THP), Trained Low Potential (TLP), Untrained High
Potential (UHP), or Untrained Low Potential (ULP), Due to
the small N's and unequal cell frequencies, only within-
session statistical comparisons could be made between groups.
Thus, the first hypothesis could not be tested. Determina-
tion of which session scores to analyze statistically for
each dependent variable was made on the basis of deviations
in the mean scores between groups. In those instances where
the possibility of significant differences was indicated by
inspection of the means, a 2(trained-untrained) X 2(high

potential-low potential) analysis of variance was performed.

Loretta R. Laurenitis

An analysis of simple effects was carried out in the one case
where a significant interaction effect was found.

The results showed few significant differences between
groups, no hypothesis receiving definitive support. Inter-
pretation was difficult because the results were few and
scattered. No statistical analysis over sessions could be
performed because of the small N's and unequal cell fre-
quencies.

It was concluded that additional data is needed to
clarify the scattered findings and suggestive trends which
resulted from this study. Modifications of the scoring
system were proposed. Limitations of the present research,
such as the small number of children seen, the short time
period, the restriction of the play setting, and the need
for consideration of variables like age, sex, and presenting
problem of the child were reviewed. The need to study the
interactional relationship between the undergraduates' and
children's qualities and behaviors was discussed. Directions

for future research were explored.

CHANGES IN THE BEHAVIOR OF CLINIC-REFERRED
CHILDREN DURING PLAY INTERACTIONS WITH

TRAINED AND UNTRAINED UNDERGRADUATES

BY

Loretta R. Laurenitis

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

MASTER OF ARTS

Department of Psychology

1974

To Bob, whose caring and love
have enriched my life

To my grandparents, Zidur and Susan,
for their many sacrifices
and great love

ACKNOWLEDGMENTS

I would like to express my appreciation to Dr. Gary
Stollak for his infinite patience, understanding, and
guidance over the years of my involvement in this study.
His ideas, interests, and concern for children have made
him a constant source of inspiration and challenge.

Special gratitude is extended to Dr. Lawrence Messé
who so generously gave of his time. His advice and sta-
tistical knowledge were of invaluable help to me.

Thanks also to Dr. Martha Karson whose encouragement
and support have been consistent throughout my graduate
years.

Finally I wish to acknowledge my great indebtedness to
Bob, who has shared these last two and a half years of
struggle, learning, and growth. His supportive assistance,
kind understanding, and enduring belief in me have made it

possible to continue during the hard times.

iii

TABLE OF CONTENTS

Page

LIST OF TABLES O O I O I O O C O O O O O O O C O O 0 Vi
INTRODUCTION AND REVIEW OF THE LITERATURE . . . . .

1
Current Status of the Mental Health Field . . . l
Nonprofessionals as Psychotherapeutic Agents . . 2
College Students as Mental Health Agents . . . . 4
Play Therapy - A Theoretical Overview . . . . . 7
0

Research in Play Therapy . . . . . . . . . . . . 1
Research on the Process of Play Therapy and

Play Encounters . . . . . . . . . . . . . . . 13
Research Questions and Process Variables . . . . 14

METHOD . . . . . . . . . . . . . . . . . . . . . . . 29

Selection of Undergraduates . . . . . . . . . . 29
Control Group Activities . . . . . . . . . . . . 30
Trainee Group Activities . . . . . . . . . . . . 3l
Selection of Children . . . . . . . . . . . . . 32
Analysis of Children's Play Session Behavior . . 34

Reliability . . . . . . . . . . . . . . . . . . 35
RESULTS 0 O O O O O O O O O O O O O O O O O O O O O O 38

Analysis of the Data . . . . . . . . . . . . . . 38
Hypothesis 1

Hypothesis 2 . . . . . . . . . . . . . . . . . . 40
Hypothesis 3
Hypothesis 4

. . . . . . . . . . . . . . . . . . 46

DISCUSSION . . . . . . . . . . . . . . . . . . . . . 48

Overview . . . . . . . . . . . . . . . . . . . . 48
Hypotheses . . . . . . . . . . . . . . . . . . . 48
Patterns of the Dependent Variables . . . . . . 48
Trained vs. Untrained Groups . . . . . . . . . . 54
High Potential vs. Low Potential Groups . . . . 55
Training X Potential Interaction . . . . . . . . 57
Limitations of the Present Study . . . . . . . . 58
Directions for Future Research . . . . . . . . 72

SUMMARY . . . . . . . . . . . . . . . . . . . . . . 76

iv

TABLE OF CONTENTS (Cont'd)
Page
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . 79
APPENDICES . . . . . . . . . . . . . . . . . . . . . . 85
Appendix

A - Definitions of the Rating Scales for
Scoring Children's Behavior . . . . . . . . . 85

B - Sample Scoring Sheet for Coding Children's
Behavior . . . . . . . . . . . . . . . . . . 93

C - Mean Percentage of Agreement for Raters
With Expert 0 O O O O O O O O O O O O O O O O 94

D - Inter-Rater Reliability Coefficients . . . . 95

E - Mean Frequency and Mean Intensity Scores
for Affection and Aggression . . . . . . . . 96

F - Mean Intensity per Session of Affection
and AggreSSion O O O O O O O I O O O O O O O 9 7

G - Mean Scores of the Dependent Variables for
the Trained High Potential (THP), Trained
Low Potential (TLP), Untrained High Potential
(UHP), and Untrained Low Potential (ULP)
Groups . . . . . . . . . . . . . . . . . . . 98

H - Analysis of Variance Summaries for the
Dependent Variables . . . . . . . . . . . . .103

I - Simple Effects Analysis . . . . . . . . . . .106

Table

10

11

12

LIST OF TABLES

Subjects (N=28) . . . . . . . . . . . . . . .

Cell means for the Dependent Variables for
the Trained and Untrained Groups . . . . . .

Mean Scores of the Dependent Variables for
the High Potential (HP) and Low Potential (LP)
Groups 0 O O O O O O O O O O O O O O O O O 0

Mean Percentage of Agreement for Raters with
Expert 0 O O O O O O O O 0 O O O O O O O O O

Inter-Rater Reliability Coefficients . . . .

Mean Frequency and Mean Intensity Scores for
Affection and Aggression for Children
Encountering Trained High Potential (THP),
Trained Low Potential (TLP), Untrained High
Potential (UHP), and Untrained Low Potential
(ULP) Undergraduates . . . . . . . . . . . .

Mean Intensity per Session of Affection and
Aggression for Children Encountering Trained
High Potential (THP), Trained Low Potential
(TLP), Untrained High Potential (UHP), and
Untrained Low Potential (ULP) Undergraduates

Mean Scores of the Dependent Variables for
the Trained High Potential (THP), Trained

Low Potential (TLP), Untrained High Potential
(UHP), and Untrained Low Potential (ULP)
Groups . . . . . . . . . . . . . . . . . . . .

Analysis of Variance Summaries for Leadership
in Sessions 1, 6, and 15 . . . . . . . . . .

Analysis of Variance Summaries for Dependency
in Sessions 11 and 15 - . . . . . . . . . . .

Analysis of Variance Summaries for Affection
in Sessions 6, 11, and 15 . . . . . . . . . .

Analysis of Variance Summaries for Aggression
in Sessions 1, 6, 11, and 15 . . . . . . . .

vi

Page

35

41

45

94

95

96

97

98

100

101

102

103

&

42

99

Table

13

14

15

LIST OF TABLES

Page

Analysis of Variance Summaries for Contact
in Sessions 6, 11, and 15 . . . . . . . . . . 104

Analysis of Variance Summaries for Role-
Playing in Sessions 1, 6, 11, and 15 . . . . 105

Simple Effects Analysis Exploring the

Significant Training X Potential (AB)
Interaction for Aggression in Session 1 . . . 106

vii

INTRODUCTION AND REVIEW OF THE LITERATURE

The purpose of this research was to investigate differ-
ences in the behavior of clinic-referred children in play
encounters with trained or untrained, "high potential" or
"low potential" undergraduates. Differences between groups
were studied for the following behaviors: Leadership,

Dependency, Affection, Aggression, Contact, and Role-Playing.

Current Status of the Mental Health Field

 

Within the last ten years, mental health professionals
have become alarmingly aware of the ever-increasing demands
for mental health services and the limited supply of trained
professionals available to meet those needs (Cowen, Gardner,
and Zax, 1967; Gordon, 1965; Guerney, 1969; Hobbs, 1964;
Joint Commission on Mental Health of Children, 1970; Smith
and Hobbs, 1966). Cowen and Zax (1967) note that "Merely in
terms of demand for mental health services (rather than need
which is inferred to be many times greater), present re-
sources, measured by almost any criterion, are grossly in-
sufficient" (p. 15).

Hobbs (1964) suggested that twenty-five percent of
mental health resources be used to mount a holding action
against the problems of adults and 75 percent be devoted to

the mental health problems of children. He proposed that

1

such action would be the only way to make substantial changes
in the mental health of our adult population a generation
later.

The Joint Commission on Mental Health of Children

titled its 1970 report Crisis in Child Mental Health and

 

called our lack of commitment to the problems and needs of
children "a national tragedy." They cited National Institute
of Mental Health estimates that in 1966, nearly 1,000,000 of
the 1,400,000 children under 18 having serious difficulties
and needing psychiatric care received no treatment (p. 5).
Clearly, the mental health needs of our nation cannot
be met by the traditional approaches and the training of
more professionals. Smith and Hobbs (1966) noted that "The
present and future shortage of trained mental health pro-
fessionals requires experimentation with new approaches to
mental health services and with new divisions of labor in

providing these services" (p. 40).

Nonprofessionals as Psychotherapeutic Agents

 

One way to help alleviate the current shortage of mental
health professionals and the critical demand for services is
through the use of trained nonprofessionals to assume some
of the responsibilities and tasks of professional change
agents. Research has been conducted in this direction using
parents (Berkowitz and Graziano, 1972; Guerney, 1964;
Guerney, Guerney, and Andronico, 1966; Guerney and Stover,

1971; Hawkins, Peterson, Schweid, and Bijou, 1966; Shah, 1967;

Stover and Guerney, 1967), teachers(Becker, Madsen, Arnold,
and Thomas, 1967; Harris, Wolf, and Baer, 1964; Zimmerman
and Zimmerman, 1962), housewives and other underemployed
individuals (Harvey, 1964; Rioch, Elkes, Flint, Usdansky,
Newman, and Silber, 1963), college students (Cowen, Zax, and
Laird, 1966; Davison, 1965; Goodman, 1967; Kreitzer, 1969;
Linden and Stollak, 1969; Reinherz, 1964; Stollak, 1968,
1972, 1973a), and high school students and peers (Fellows
and Wolpin, 1969; McWilliams and Finkel, 1973; Perlmutter
and Durham, 1965).

Results of such research have been promising, Rioch,
Elkes, Flint, Usdansky, Newman and Silber (1963), in one
of the earliest studies, achieved positive results in using
married women to do therapy with adults. Carkhuff and
Truax (1965) found that they could train graduate students
and lay hospital personnel in less than 100 hours to function
at levels of therapy nearly commensurate to those of experi-
enced therapists. Poser (1966) used trained undergraduate
students as therapists for groups of hospitalized patients
and found that these lay therapists, in comparison to an
untreated control group, achieved somewhat better results
than psychiatrists and psychiatric social workers doing
therapy with similar groups of patients. Stover and Guerney
(1967) and Guerney and Stover (1971) found that mothers
could be trained to assume the role behavior of a client-
centered therapist and that their behavior effected concomi-

tant changes in the behavior of their children.

Carkhuff (1968), Cowen, Gardner, and Zax (1967), Guerney
(1969), the Joint Commission on Mental Health (1970),
Matarazzo (1971), Riessman (1965), and Rioch (1966), all
have encouraged the use of nonprofessionals and cite the
benefits which may be achieved. Carkhuff (1968), in review-
ing research findings on the use of nonprofessionals, notes
the evidence indicating that lay persons can effect signifi-
cantly constructive changes in the clients they see and
that "lay trainees function at levels essentially as high or
higher (never significantly lower), and engage clients in
counseling process movement at levels as high or higher,
than professional trainees" (p. 118). Carkhuff (1968),
Riesman (1965), and Rioch (1966) also note evidence indicat-
ing that those people giving help undergo constructive

change themselves as a result of being in the helping role.

College Students as Mental_Health Agents

 

In attempts to expand mental health services through
the use of non-professionals, college students constitute a
relatively untapped resource pool. Although their services
have been used in programs such as "Big Brothers" and various
other volunteer activities, only within the last few years
has there been recognition of their potential usefulness as
psychotherapeutic agents and studies conducted to research
this issue.

Gruver (1971), in a review of the research using college

students as mental health agents, found that very few studies

have been done and that "further, most of those investiga-
tions which have been conducted are so methodologically in-
adequate that it is impossible to draw firm conclusions about
the relative effectiveness of college students as therapeutic
agents" (p. 123). Gruver also noted, however, that there is
sufficient evidence to conclude that the therapeutic relation-
ship has a definite, positive effect upon the college student
therapist himself.

Stollak (1968) trained college students to act as play
therapists with clinic-referred children under 10 years of age.
The student's role was modeled after that of a client-centered
therapist. The basic task was to be empathic, non-directive,
and understanding in the play sessions and to convey acceptance
and understanding to the child. Training was accomplished in
ten sessions. Students discussed the nature and purpose of
their role, observed play therapy sessions, and practiced
playing with normal children, At the end of the training
each student was assigned a child who had been referred to
the Psychology Clinic at Michigan State University or to the
Lansing Child Guidance Clinic. Results obtained after ten
play sessions indicated that the undergraduates' behavior did
change significantly. "Reflection of content" and "Clari-
fication of feeling" were behaviors which increased over the
sessions. (Results obtained with regard to the children's
behavior will be discussed in a later section.)

Linden and Stollak (1969) further investigated under-

graduates and children in play encounters. They found that

those students who were in a more didactic and structured
training group (as opposed to an experiential group in which
students sought to discover their own method of responding

to a child) reflected significantly more feeling and content of
behavior, gave significantly less direction and unsolicited
help, asked fewer questions and restricted the children less.
The authors concluded that the ability to communicate em-

pathy is not an innate ability but must be specifically

taught.

Reif and Stollak (1971), in a study using "normal"
children, compared a group of nine undergraduates trained in
specific techniques of play interaction based on a client-
centered play model with an untrained control group. Results
indicated that the trainees demonstrated greater frequencies
of the following behaviors: "Reflection of verbal content,"
"Reflection of motor behavior," "Interpretation," "Compliance
clarified," and "Reciprocal participation of fantasy be-
havior," These results were in the expected direction and
were considered effective responses in interaction with a
child. The trainees were also rated significantly higher
than the control students on the measure of warmth for the
last two sessions. (Again, results with regard to child be-
haviors will be reviewed in a later section).

Stollak (1972, 1973a) conducted an expanded study in-
vestigating the behavior of trained and untrained, "high
potential" and "low potential" undergraduates in play en-

counters with clinic-referred children. (Since the present

report is based on this project, a comprehensive description
may be found in the Method section). Schreiber (1972) found
that trained undergraduates showed higher levels of "empathic"
behavior with children than did the controls, particularly
with regard to "Communication of Acceptance" of the child.
There were also some significant training effects for the
remaining two categories, "Allowing Self-Direction" and
"Involvement."

In summary, current research indicates that college
students can be trained to alter their behavior in a thera-
peutic direction. This study, however, is concerned with
the changes occurring in children's behavior during play
encounters with undergraduates. Clearly, children's be—
havioral patterns need to be researched in order to assess

the impact of the undergraduates' behaviors.

Play Therapy - A Theoretical Overview

 

A major psychotherapeutic treatment approach with chil-
dren is play therapy as developed by Allen (1942), Axline
(1947, 1955), Dorfman (1951), and Moustakas (1953, 1959).
Although often referred to as "client-centered" or non-
directive play therapy, Moustakas (1959) distinguishes his
method as "relationship therapy.“ He views "client-centered"
therapy as focusing on the therapist and child as separate
individuals, the therapist conveying empathic understanding
and unconditional regard, making reflections and clarifica-

tions. In "relationship therapy" Moustakas views the

relationship as the significant growth experience, as "both
means and end." Guerney, Guerney, and Andronico (1966)
also emphasize the importance of the relationship between
the therapist and child as the critical factor in psycho-
therapy.

Although Moustakas makes this distinction between
client-centered play therapy and relationship therapy, the
approaches as exemplified by Moustakas and Axline are very
similar and based on the same principles. For Axline (1955)
the overall objective of any psychotherapy is "to provide a
relationship with a client that will enable him to utilize
the capacities that are within him for a more constructive
and happier life as an individual and member of society"

(p. 619). Moustakas (1953) conceptualizes play therapy in
terms of "a set of attitudes in and through which children
may feel free enough to express themselves fully, in their
own way, so that eventually they may achieve feelings of
security, adequacy, and worthiness through emotional insight.
The three basic attitudes in child-centered play therapy are
faith, acceptance and respect" (p. 2).

Axline (1947) suggests eight basic principles to guide
the non-directive play therapist:

1. the establishment of a warmeriendly relationship
with the child in which good rapport is established;

2. the therapist's acceptance of the child "exactly as

. II
he is;

3. the establishment of a feeling of permissiveness in
the relationship to allow the child to feel free to express
his feelings;

4. the therapist's ability to recognize and reflect the
child's feelings;

5. a deep respect on the therapist's part for the child's
ability to solve his own problems;

6. no attempt by the therapist to direct the child's
actions or conversation, but rather allowing the child to take
the lead;

7. the therapist's view of therapy as a gradual process,
one not to be hurried; and

8. the establishment of only necessary limitations
(p. 73—74).

Within the warm, accepting, permissive environment
characterized by Axline's eight principles, it is believed
that a child can develop his potentials and innate capacities,
work through conflicts and disruptive feelings, and become a
more independent and self-confident individual.

Stollak (1973b) stresses the importance of sensitive
and effective communication with children which he feels in-
cludes the following elements:

1. a clear, unambiguous verbal and nonverbal communica-
tion indicating to the child that the adult is aware of, and
understands his or her feelings, needs, and wishes and how

the child's actions derive from such thoughts and feelings;

10

2. a clear and unambiguous communication indicating
that the adult accepts the child's feelings, needs, and
wishes as natural and valid human experiences - but not
necessarily the child's actions, which might be unacceptable
to the adult;

3. a clear and unambiguous communication indicating
how the adult thinks and feels about the child's thoughts,
feelings, and actions;

4. if appropriate (i.e., if the child's actions are un-
acceptable), a clear and unambiguous communication indicating
alternative ways for the immediate expression of the child's
thoughts, feelings, and needs; and

5. if appropriate (i.e., if the child's actions are un-
acceptable), a clear and unambiguous communication indicating
alternative ways for the child to express his inner experiences
in the future (p. vii-viii).

In summary, the conditions some consider important
aspects of adult-child encounters in which a child can ex-
press his feelings, gain self-confidence, and grow as an in-
dividual are acceptance, empathy, sensitive and effective
communication of the child's feelings, permissiveness, and

respect for the child.

Research in Play Therapy

 

In spite of the fact that play therapy is a widely used
approach in attempts to help clinic-referred children, re-

search in this area has been insufficient. Advocates of play

11

therapy espouse its effectiveness and beneficial outcomes
through theoretical dissertations and clinical reports of

the process of "successful" cases. However, as Levitt (1971)
noted, evaluation research in child psychotherapy has failed
to demonstrate unequivocally the effectiveness of child
psychotherapy. Research relating specific therapeutic inter-
ventions to specific changes in particular children and under
specific conditions still needs to be done in the field.
Ginott (1964) stressed that many of the most important
questions about play therapy, such as what the process is

and what variables affect that process, have been left un-
answered. Levitt (1971) also noted that surveys of play
therapy practices, such as Ginott's and Lebo's (1961) in-
vestigation of play therapy limits and theoretical orienta-
tion, indicate that variation in the behavior of child thera-
pists in the therapy situation is considerable and may not

be related to theoretical orientation. Therefore, it is
possible that there are broad individual differences among
therapists in effectiveness.

Levitt (1971) recognized that one reason for the dearth
of objective studies in child psychotherapy may be the
methodological difficulties unique to child therapy research.
He cited the following two:

1. The child is a developing organism. Therefore, in
children who are basically normal, symptomatic manifestations

often seem to disappear as a function of the child's develop-

ment. That is, "spontaneous remission" or the child's

12

"growing out" of problematic behaviors is quite apt to occur.
Another common occurence is develOpmental symptom substitu-
tion, "Symptoms that are pathognomic of an underlying emotion-
al illness may...disappear as a function of development, but
will then be replaced by other symptoms" (p. 477).

2. Often persons other than the child may be directly
involved in the treatment (e.g. the mother or parents). When
this occurs, a primary research problem lies in separating
effects. Both these considerations make the task of design-
ing sound research studies in individual child psychotherapy
more complex.

Lebo (1953) summarized the status of research on play therapy
as follows:

The principles and methods of nondirective play therapy
are frequently presented as though they were firmly establish-
ed. The assured manner of writing of many of the authors and
the large-scale possibilities held before the reader, tend to
make one believe that, at long last, 'the way' has been
found...Actually, this is not so. Indeed, it may not be the
specific procedures of play therapy per se that effect the
rather remarkable personality changes. The children may be
benefiting from having someone constantly and consistently
interested in their welfare...

Nondirective play therapy, while promising when evaluated
subjectively, has been seen to have rather serious methodo-
logical lacks. One cannot concur...that play work with
children...'has come to stay' until play therapy has been es-
tablished by objective means. In the long run, nondirective
play therapy should stand or fall on the results of experi-
mental studies investigating its effectiveness in relation
to other procedures.

....To be admitted to the ranks of approved therapeutic
methods nondirective play therapy needs more than enthusiasm,
belief, and the Shibboleth, 'It works, if you only try it.'
(p. 430)

13

Research on the Process of Play Therapy and Play Encounters

As already noted, the research in the area of child
psychotherapy is minimal. This section will review those
few studies which focus on the process of play therapy and
play encounters, with specific emphasis on children's be-
havior during therapy and play interactions.

Landisberg and Snyder (1946) were the first to use an
objective approach in an attempt to analyze what occurred
in client-centered play therapy. They studied the protocols
of four children (three "successes" and one incomplete case),
ages five to six, seen by three nondirective therapists.
Children's statements were categorized as to the content,
attitude or emotion expressed, and the activity.

With regard to the children's behavior, Landisberg and
Snyder (1946) reported a rise in the amount of physical
activity in the last three-fifths of treatment and increased
expression of feeling. During the first two-fifths of treat-
ment, approximately fifty percent of the children's actions
and feelings were those of emotional release. For the last
three-fifths of the process, this rose to seventy percent.
Feeling was discovered to be significantly related to the
action rather than to the verbal responses. It was found
that negative feelings were expressed with increasing fre-
quency while the expression of positive feelings remained
fixed at about thirty percent of the responses. Although
negative feelings initially comprised less than twenty

percent of the responses, they increased to almost forty

14

percent near the end of treatment and then dropped off in

the final fifth to thirty-three percent, or exactly the same
as the percentage of positive feelings. There was a marked
increase in the expression of feelings toward other persons
or situations, the major part of the children's feelings not
being directed toward themselves or the therapist. It was
noted that the children did not change markedly in the amount
of feelings expressed toward themselves or their counselors
but were able to bring out both negative and positive feel-
ings toward other individuals during the play therapy process.
This ability increased over the period of treatment. In the
final stages, these expressions fell off somewhat, the
authors speculating that the child's need for such expression
may have decreased.

The small N, lack of a control group, and lack of a be-
havioral definition of "success" for the three cases. are
serious shortcomings of this study. In particular, without
the control group, one cannot assess whether "success" was
due to the therapy and/or to the presence of an interested
adult or instead was a developmental phenomenon.

Finke (1947) analyzed the complete protocols of six
children ages five to eleven seen by six different therapists
for behavior problems. Expressions of feeling were empha-
sized in the belief that this category would reflect the
child's changing emotions resulting from the play therapy.

The results showed that similar trends emerged in the

therapy of the different children which divided therapy into

15

three stages:
(1) Child is either reticent or extremely talkative.
He explores the playroom. If he is to show aggression at
any time during therapy, a great deal of it will be exhibited
in this stage.

(2) If aggression has been shown, it is now lessened.
The child tests the limitations of the playroom. Imaginative
play is frequently indulged in here.

(3) Most of the child's efforts are now expanded into
attempted relationship with the counselor. The child tries
to draw the therapist into his games and play. (Lebo, In
Haworth, 1964, p. 423).

Lebo (In Haworth, 1964) notes that Finke (1947), like
Landisberg and Snyder (1946), found no pattern for positive
statements. A major difference from their study, however,
was that Finke also found no pattern or trends for negative
statements.

Finke's (1947) study suffered from the same limitations
noted above for the Landisberg and Snyder (1947) study.

Both, however, showed changes in the children's behavior over
the period of play.

Lebo (1952) researched the relationship between chrono-
logical age and the kinds of statements children make during
play. He felt Finke (1947) had failed to recognize that age
and maturation might effect the types and levels of verbaliza-
tions made by children. Lebo (1952) studied twenty children
in three play sessions given by the same therapist. Five age
stages (four, six, eight, ten, and twelve years) were select-
ed with two boys and two girls in each stage. Fifteen pages
of verbatim notes were randomly chosen and categorized by

three experienced play therapists according to Finke's (1947)

categories.

l6

Lebo (1952) found that age, indicative of maturation,
did have an effect and was related to the types of ex-
pressions made. As children became older, they tried less
to draw the therapist into their play, expressed more likes
and dislikes while telling the therapist fewer of their de-
cisions, and did not take as much time exploring the limita-
tions of the situation. Hence, it appears that the process
of play therapy differs in particular ways depending on the
child's age.

Moustakas and Schlalock (1955) studied therapist-
child interactions in play therapy sessions. Two groups of
four year olds (N = 5 for each group) were selected. Group
A consisted of children without emotional problems while
Group B was composed of children with severe emotional diffi-
culties as unanimously determined by three sets of ratings.
Each child was seen in two forty minute play sessions about
three days apart. Behavior over these sessions was assessed
every five seconds with respect to seventy-two categories.
Although the results of this study do not cover changes in
children's behavior over time, if one considers that the pur-
pose of play therapy is to help a child with emotional prob-
lems become less "disturbed," or achieve a healthier adjust-
ment,then one might expect that if treatment were successful
for children in Group B (those with emotional problems), they
would at the end of their play therapy sessions exhibit be-
haviors similar to those of children in Group A (children

without emotional problems).

17

The results of this study in reference to the nature of
the interaction behaviors of the children were the following:

(1) Children with serious emotional problems were more
like than different from the children without such problems
(correlation of .694 for the two groups). Their behavior
patterns, in general, were similar. Ninety-five percent of
the interaction behavior of both groups was represented by
these seven categories: Nonattention, Attentive Observation,
Statement of Condition or Action, Seeking Information, Giv-
ing Information Verbally, Recognition of Stimulation, and
Nonrecognition of Stimulation. Minor differences were re-
vealed in the approach of the two groups to the therapist.
Group A (without emotional problems) used the following be—
havior categories more frequently: Joint Participation in
Activity, Orienting the Role of the Therapist in Play, Seek-
ing Help, Directing by Suggestion, Seeking Permission, Re-
jection by Changing the Subject, and Rejection by Denying
the Validity of the Therapist's Statement. Group B children
used more often all categories of Threat to Attack, Forbid-
ding, and Physical Attack.

Moustakas and Schlalock concluded that the differences
in the behavior between the two groups seemed related to the
problem of the Group B children.

These children spent much of their time in noninter-
active behavior - in fantasy, play, and other activity that
excluded the therapist - or they responded to him in a way
that did not encourage interaction. In contrast, the

children of Group A showed considerable verbal interaction
with the therapist, talking about friends, school, home, and

18

other conditions in their lives, explaining their behavior
and giving the therapist clues to an understanding of their
behavior. ....Group A is much more verbal in a social sense,
while Group B is more often nonattentive, that is, does not
interact with the therapist, or interacts in a way that does
not elicit interactive responses from him. (pp. 148-149)

(2) The second major difference between the two groups
was the large number of hostile feelings exhibited by Group
B. However, this finding was not unexpected since the
emotional difficulties described for all children in Group
B related to hostility affecting their nursery school be-
havior. Although a significant difference was found between
the hostility rankings of Groups A and B, hostile and attack-
ing behavior was still a rather infrequent occurrence in
both groups.

Minor differences related to Group B's pattern of
greater hostile behavior were also found. Group B children:
tended to be more forbidding, more prone to attack

and threaten to attack, and more likely to use physical
barriers to block or restrain the therapist. There was
some indication that dependency, often associated with hos-
tility, was more frequently expressed by Group B, in the
sense of frequently asking questions...Along with the ten-
dency of Group A children toward more verbal interaction

on a friendly, social level, there was more Joint Par-
ticipation in Activity with the therapist and more Seek-
ing Permission. These children were also more assertive

in the sense of orienting the therapist to his role and
function, directing by suggestion, and denying the validity
of the therapist's statements or actions when his behavior
did not satisfy the child. (p. 149)

Stover and Guerney (1967) conducted a study of filial
therapy to determine

(1) whether parents of disturbed children could be

trained to assume a reflective, empathic role in play sessions

with their children and

l9

(2) whether the children's behavior would reflect con-
comitant changes as a result of the mother's new role behavior.

Two control and two experimental groups of six to eight
mothers were randomly formed from those applying to the
Rutgers University Psychological Clinic. The children had
previously been diagnosed as having emotional difficulties.
The experimental groups received training in client-centered
play therapy principles, with an emphasis on the reflective
role and the building of an empathic relationship with the
child. The control group received no training but had the
same number of play sessions with their children as the ex-
perimental group mothers. For the purposes of the research,
the fourth play sessions (third training session) were com-
pared with the initial (base-line) session.

Multivariate analysis of variance was used to study the
data acquired with regard to the children's behavior. They
found that for children in the experimental groups, nonverbal
aggression increased; whereas for control group children,
this behavior decreased. The differences between the experi-
mental and control groups were found to be statistically
significant. The experimental group children also increased in
expression of verbal negative feelings rather than decreasing
as with the control groups. The results of one experimental
group reached statistical significance. No significant dif-
ference was found for the variables of verbal leadership
and verbal dependency. Thus, some trends did emerge although

there were only four play sessions involved in this study.

20

Stollak (1968) used Guerney's filial therapy procedures
and techniques to train college undergraduates to act as
play therapists with clinic-referred children under ten.
Students had been selected as "high potential" or "low po-
tential" therapists based on several personality questions.
Children's behavior was analyzed in the first, fifth, and
tenth sessions to determine what changes occurred on meas—
ures of aggression, leadership, dependency, and negative
feelings. Due to the lack of significant differences between
the "high potential" and "low potential" therapists, data
for the children were combined. Stollak found that expres-
sion of negative feelings increased significantly from the
first to the tenth session, but not from the first to the
fifth, or fifth to the tenth. Leadership behavior also in-
creased from the first to the fifth, but not from the fifth
to the tenth sessions. There were no significant changes
over the three sessions in expressions of either aggression
or dependency.

Another study of children's behavior and the training
of undergraduates in play techniques was conducted by Reif
and Stollak (1971). Two groups of nine undergraduates each
were randomly selected to play with "normal" children (ages
4 to 7) from the community. One group, the experimental or
trainee group, was given training in play techniques based
on a client-centered model of play therapy. The control
group received no such training but played with their child-

ren once a week as the trainees did. Twenty 20 minute play

21

sessions were conducted. Four of these sessions, the first,
seventh, thirteenth, and twentieth were videotaped and used
for the analyses of students' and children's behavior. Re—
sults indicated that the experimental group children showed
overall significantly greater increases in "statements of
personal thought or behavior in the context of fantasy,"
"statements of interpersonal awareness in the context of
fantasy," "fantasy behavior," "fantasy aggression," and
"nonrecognition." Experimental children's fantasy behaviors
showed significant increases with respect to both time and
training in the final play session.

The most comprehensive study to date of the efficacy
of filial therapy and the process of play sessions conducted
by trained nonprofessionals was carried out by Guerney and
Stover (1971). The children were between four and ten years
old, all with serious emotional difficulties, all screened
very carefully from cases referred to Rutgers Psychological
Clinic and the Hunterdon Psychiatric Clinic. Of seventy-one
who began, fifty—one children completed the project. Mothers
were divided into groups of six to eight members. These
groups met two hours each week with one or two mothers con-
ducting a twenty minute play session with her clinic-referred
child. The mothers were trained to use play techniques
modeled after Rogerian client-centered play therapy, the em-
phasis being on the development of an empathic relationship

with the child. These groups were conducted for a period of

22

twelve to eighteen months. Following the demonstration
training sessions of six to eight weeks, the mothers con-
ducted weekly play sessions at home with their children.
The group meetings were used to discuss those sessions and
the mothers' feelings and concerns. Each mother also was
observed monthly in a play session with the child at the
clinic.

Children's behavior in play sessions with their mothers
was coded at the time of the diagnosis and each month
throughout the course of treatments for the entire twelve to
eighteen months of involvement in the program. The coded
process variables studied were Aggression toward the Mother,
General Aggression, Dependency, Leadership, Affection, Affec-
tion toward the Mother, Role-playing, and Contact with the
Mother. Data derived from some 49,000 fifteen-second ob-
servations occurring in some six-hundred play sessions were
assessed by thirds of the treatment period, determined in—
dividually for each child depending upon his number of re-
corded sessions. The average number of responses per child
per session for the full sample and selected sub-samples was
used.

Following are the results and variable patterns dis-

cerned:

23

Aggression toward the mother - Peaking of aggression
occurs toward the midpoint of the sessions, followed by a
decrease. At the last phase of therapy, aggression toward
the mother is slightly lower than at the start of treatment.
When data for aggressive and withdrawn children are sepa-
rated out, there is a noteworthy difference in the pattern.
For the aggressive children, aggression toward the mother
rises sharply and then declines, remaining well above the
starting point. For withdrawn children, however, aggression
toward the mother declines from the first to second period,
and somewhat further in the third.

 

General aggression - Hardly any peaking of this variable
occurs. There is, however, a very slight decline at the last
phase relative to the beginning one.

 

Affection - Expression of affection declines throughout
the treatment phase, the rate of decline being the same for
the second to third phase as from the first to the second.

 

Affection_toward the mother - This variable shows the
same pattern of straight-line decline as for affection.

 

Contact with the mother increased steadily from the
beginning to the end of the sessions.

Role~Playing declines throughout the sessions, the
largest rate of decrease being from the first to second
phase of treatment.

 

Dependency shows a sharp and steady decline over the
three phases.

 

Leadership - A small rise initially is followed by a
decline to a point slightly above the starting level.

 

Although Guerney's and Stover's (1971) research sheds
important light on the process of play interactions, because
statistical analysis was not performed, we do not know if
the changes over time are significant. Since no placebo or
no-treatment control groups of children applying to the clinic
were employed, the findings must be viewed cautiously. We do
not know how or whether the patterns occurring in the chil-

dren's behavior differ from those which would occur in sessions

led by other adults or untrained mothers over the same period

24

of time.

In summary, research related to the process of play
therapy or play encounters does indicate that some changes
occur in children's behavior. However, much.of the research
suffers methodologically. At times statistical analyses
to determine significance are lacking or control groups
have not been used, making it almost impossible to draw
definitive conclusions regarding the findings. Variables
such as the age of the child, sex of child and therapist,
and degree and type of disturbance of the child have not
usually been taken into consideration. Time sequences and
behavioral categories are not often directly comparable and
hence preclude comparisons between studies. Clearly, re-
search in child psychotherapy still leaves much to be de-
sired. Many important questions remain unanswered and in-

conclusively studied.

Research_Questions and Process Variables

 

Based primarily on theoretical writings and clinical
impressions, and in some cases supported by research, there
is a generally expected pattern of positive and negative
responses in individual client-centered or nondirective play
therapy. That process is usually conceptualized as follows:
In the warm, accepting, permissive play therapy environment,
the child begins to feel accepted and liked, to recognize
his freedom, and to express his feelings and conflicts

through his play, actions, and speech. Negative and ag-

gressive feelings are often displayed, initially diffuse and

25

pervasive in their expression. As the relationship develops,
these feelings become sharpened and more specific. That is,
they are more often focused on and directed toward particular
persons in the child's life. As these feelings are expressed,
released, and worked through in the accepting presence of the
therapist, the feelings become less intense and diminish
(Moustakas, 1953, 1955a, 1955b, 1959). With regard to posi-
tive feelings, once the initial period of socially desirable
responses passes by, the child is believed to express fewer
positive feelings since some of these are thought to have
been pseudo-positive ones. Later, as the child releases and
Works through his negative and aggressive feelings, he is
able to express more positive and affectionate emotions. As
the sessions progress and the child becomes more self-
confident, it is expected that dependent behavior, thought

to occur frequently at first, will decrease steadily, while
leadership behaviors will increase steadily. Also, as a
child's conflicts and feelings are expressed, there is less
need for the child to engage in fantasy and role-playing be-
haviors. Therefore, the occurrence of these behaviors is
expected to fall off as attempts to engage the therapist in
activities increase.

The objectives of this research were to study six child-
ren's behaviors - Aggression, Affection, Leadership, Dependency,
Contact, and Role-playing - occurring in play interactions
with undergraduates who were divided into the following four

groups: Trained High Potentials (THP), Trained Low

26

Potentials (TLP), Untrained High Potentials (UHP), and
Untrained Low Potentials (ULP). Changes over sessions as
well as differences between groups were to be assessed. We
were interested in determining whether different patterns
emerged for the different experimental groups and whether
significant changes occurred in the frequency of occurrence
of the six variables over the first, sixth, eleventh, and
fifteenth sessions of play interactions. As already noted,
methodological inadequacies and differing time intervals

make it difficult to draw definitive and precise conclusions
from previous research. However, based on Guerney's and
Stover's (1971) recent filial therapy project as well as

the generally accepted pattern of the process of play therapy
derived from the literature and clinical accounts, the follow-
ing trends were expected to occur for the six variables con-
sidered:

(1) Leadership was expected to increase over sessions,

 

diminishing in the last phase as the child engaged in more
egalitarian activities with the undergraduate.

(2) Dependenquwas expected to decline in frequency over
the sessions.

(3) Aggression was expected to increase in frequency and

 

intensity, reaching a peak, and then declining.

(4) For Affection, the trend anticipated was a decline in

 

frequency which would be followed by a rise in occurrence in
the later phases as negative feelings were expressed and re-

solved.

27

(5) It was thought that Role-Playing would diminish

 

over sessions.

(6) Contact was expected to increase in frequency over
time.

The children's play participants in this study were all
college undergraduates selected as having "high" or "low"
potential for being able to assume an empathic, sensitive,
and responsive role with children. The hypotheses studied
were as follows:

(1) Significant differences over sessions will occur
for the variables in the directions of the trends noted above.

(2) Children encountering trained undergraduates will
display the anticipated trends to a greater degree than
children playing with untrained undergraduates. That is, the
frequencies of occurrence of the six variables in sessions
will be stronger in the anticipated direction for children
interacting with the trainees.

(3) The behavior of children encountering high potential
undergraduates will fit the predicted patterns more closely
than the behavior of children whose play participants are low
potential undergraduates. That is, the frequencies of the
variables in sessions will be more strongly in the expected
direction for children of high potential undergraduates.

(4) The behavior of children encountering trained high.
potential undergraduates will most clearly conform to the
anticipated patterns in comparison to the behavior of children

encountering undergraduates from the remaining three groups.

28

That is, the frequencies of the variables will be more
strongly in the anticipated directions for children of trained
high potential undergraduates. Also, training will tend to
reduce the differences between high potential and low po-
tential undergraduateSIresulting in more similar behavioral

trends for the two trained groups.

METHOD

This section is derived from Stollak (1973).

Selection_gf Undergraduates

 

Recruitment of undergraduate students to serve as play
therapists was accomplished through an advertisement placed
in the University newspaper. Volunteers who were interested
in working with children and learning how to respond to and
communicate more sensitively with children were solicited.
Students attending the initial meeting were asked to com-
plete three inventories: the Parent Attitude Research Instru-
ment (Schaefer and Bell, 1958), a Sensitivity to Children
projective questionnaire developed by Stollak and consisting
of sixteen parent-child problem situations, and a Personality
Questionnaire, also devised by Stollak and used to determine
general "mental health" of the respondents. Undergraduates
were selected on the basis of their scores for these three
inventories. Ten malesand ten females who scored "highest,"
that is, who had more " 'child-oriented,’ 'liberal' values
and attitudes, were able to communicate understanding and
acceptance of children's needs and feelings, and presented
themselves as being within the 'average' range on various
psychological dimensions" were selected as "High Potential"

(HP) subjects. Another ten males and ten females who scored

29

30

"lowest" on the inventories were designated "Low Potential"
(LP) subjects. Ten HP's and ten LP's were then randomly se—
lected to be the experimental (trainee) group. The remain-
ing ten HP's and ten LP's became the control (untrained)
group. Equal numbers of females and males were included in

both of these groups.

Contrql Group Activities

 

Each of the twenty control group subjects met individ-
ually with Stollak and was informed of the random selection
procedure. The importance and necessity of a control con-
dition to evaluate the effect of training and supervision
on the undergraduate's and child's behavior were discussed.
Control subjects were told they would be called when there
was a clinic-referred child for them to play with, and that
although they would receive neither training nor supervision
during those sessions, an observer would be present behind
a one-way mirror. The purpose of the observer, usually a
graduate student, was to insure that neither the under-
graduate nor child were destructive toward one another. Con-
trol subjects were also informed that they could participate
in training similar to that of the experimental group once
the research was concluded. In the meantime, the control
subjects were given a list of books on play therapy which
they could read if they wanted but which they would not be

allowed to discuss with the graduate student observers.

31

TraineedGroupdActivities

The twenty trainees were randomly assigned to one of
three groups, each group consisting of six or seven trainees
and approximately equal numbers of females and males, HP and
LP subjects. Stollak met with one group while graduate re—
search assistants supervised the other two groups.* Each
group met two hours weekly.

Until the trainees began play sessions with their clinic-
referred children, their first nineteen weeks of training
consisted of various activities. The Sensitivity to Children
questionnaire was discussed. The group's task was to attempt
to reach consensus on how best to handle each of the problem
situations. Once collective answers were acquired, discus-
sion centered on the effects of each "solution" upon the
child's needs and feelings. Principles of sensitive and
effective communication with children were studied through
the readings of Axline, Moustakas, Ginott, and others.
"Behavioristic" concepts such as reinforcement were also dis-
cussed to help students become aware of how they might be re-
inforcing a child's behavior. The majority of time during
this first training phase was devoted to Discussion Topics.
One-half the group served as the "panel of experts" for
handling problems for each Discussion Topic; the other half

served as questioners. As the experts and questioners

 

*The graduate research assistants were Sharon Berliner

and Allan Scholom.

32

discussed child behavior problems related to the specific
topic, they attempted to work out solutions based on their
readings and on the principles of communication they had
learned.

Trainees also began to play with a "normal" child as
soon as they could find one. Selected readings and material
from five one-hour edited videotapes consisting of sensitive
and insensitive handling of various child behaviors served
as main discussion sources. Role-playing and use of ex-
amples and problems one might encounter with children were
helpful learning aids as well. In addition, each student
was videotaped with his or her child and these sessions were
then observed and discussed by the group. Emphasis was on
the importance of empathy, sensitive communication with the
child, and the effect of the student's behavior on the child's
actions and feelings. Trainees continued to play with their
"normal" children until assigned a clinic-referred child some

fifteen weeks into the new school year.

Selection of Children

 

Children for this project were selected from referrals
of four to nine year olds made to the Psychological Clinic.
Cases were assigned to Stollak if the Clinic staff worker
felt that there was an emotional or social problem and that
more than an assessment seemed needed. Each case was com—
prehensively evaluated by a graduate student. If the evalua-

tion indicated that the child did have emotional or behavioral

33

problems, was of average intelligence without neurological
or physical impairment, was not psychotic, and did not have
parents who were psychotic, suicidal, or homicidal risks,
then the recommendation was made for individual play assess-
ment sessions.

Upon completion of the evaluation, the graduate student
met with the parents and, if appropriate, suggested fifteen
sessions of "play interaction" at no cost with a "therapist
in training" whom the graduate student would be observing.

It was decided that the fifteen sessions would serve as a
continuing assessment of the child's feelings, needs, and
fantasy expression, as well as possibly being beneficial to
the child due to the special individual kind of attention
profiided. If the parents accepted the recommendation, the
graduate student, by means of a randomization process, called
a trainee or control subject. A time convenient to the under-
graduate, the graduate student, the parent and the child was
arranged and the play sessions were begun. The first, sixth,
eleventh, and fifteenth sessions were videotaped for the pur-
poses of the research project. Trainees received immediate
feedback and supervision from the graduate student who was
observing through a one—way mirror.

At the end of the fifteen play sessions, the graduate
student again conducted an evaluation and assessment of the
case. If at that time the parents did not wish to continue

with the procedure, or if as a result of the evaluation it

was felt that the treatment of choice should be marital

34

counseling or individual therapy for another family member,
then these issues were discussed with the parents and the
child was dropped from the project. The option to continue
individual play sessions also existed. If parents were satis-
fied with the program and it seemed that at least some pro-
gress was being made by the undergraduate participant with

the child, then another fifteen play sessions were begun.

Analysis of Children's Play Session Behavior

 

In this study, six variables were chosen for the analysis
of children's behaviors over the fifteen play sessions. The
variables selected were those developed and used by Guerney
and Stover (1971). (See Appendix A for descriptions of the
variables). They were coded as follows:

Leadership as having occurred (1) or not occurred (0).

 

Dependency as having occurred (1) or not occurred (0).

 

Affection as having not occurred (0) or having occurred

 

along a dimension of intensity from mild (l), to moderate
(2), or intense (3).

Aggression as having not occurred (0) or having occurred

 

along a dimension of intensity scored mild (1), moderate (2),
or intense (3). (Adopted from Siegel, 1956).
Contact as having occurred (1) or not occurred (0).

Roleeplaying as having occurred (1) or not occurred (0).

 

A rating was made for each variable during each fifteen
second interval of the videotape. The first, sixth, eleventh,

and fifteenth play sessions were used. Each session was thirty

35

minutes long.
Table 1 presents a breakdown by session for the subjects

for whom results were obtained.

Table 1. Subjects (N = 28)

 

' Group; Session 1 6 ll__ 15
HP - Trainees (N = 6) 5 3 5 4
HP - Control (N = 7) 7 4 4 2
LP - Trainees (N =10) 9 8 3 7
LP - Control (N = 5) 4 2 3 2

The low number of tapes is readily apparent from the
table. Mechanical failures of the video-equipment resulting
in loss of the audio and/or video portions was largely re-
sponsible for the reduction in data. Also, only twenty-eight
of forty undergraduates received clinic cases because of a
lower referral rate than anticipated. Stollak replicated the

project during 1972-73.

Reliability

Raters were given the description sheets of the variables
(See Appendix A) and asked to study the specific examples
closely, while submitting the categories and general descrip-
tions of the variables to memory. After a training period

of approximately twelve hours involving practice coding and

36

discussion, three raters* independently coded three half-hour
videotapes of undergraduate play sessions not used in this
project. Because these three coders left school before the
study was completed, two new assistants** were trained to
rate the remainder of the tapes. In order to obtain relia-
bility, one new coder rated three half-hour tapes and the
other rated six. In all five cases, scores from the coder
were compared with scores of the experimenter, designated

as "expert," and a mean percentage of agreement with the
"expert" was obtained for each rater across categories.

These scores across coders and categories ranged from 66.0

to 100 with a mean agreement of 82.1. (See Table 4 in Appen-
dix C for complete results)

Once adequate reliability was established, coders in-
dependently rated the half-hour play sessions which were
available. Coders were unaware of the purpose of the study,
the session number, or the characteristics of the undergradu-
ates. Except for two tapes which could only be coded once
each because of difficulties with the tapes, each session was
coded twice by two different raters. Inter-coder reliabili-
ties were calculated using Pearson product-moment correlations.

Note that ratings of one of the original three raters were

 

*These raters were Alex De Yonker, Barry Jay, and
John Healey.

**The new coders were Mary McCaslin and Michael Harding.

37

deleted from this study because they were too few in number

to use in calculating inter-coder reliability. These corre—
lations for each pair of coders over the six behavior vari-

ables are presented in Appendix D. They ranged from .36 to

1.00. Mean correlations for the six variables were as fol-

lows: Leadership, .83; Dependency, .84; Affection, .81;

Aggression, .96; Contact, .93; and Role-playing, .92.

RESULTS

This study assessed the behavior of children interact-
ing in play sessions with undergraduates. The dependent
measures selected for analysis were the following six behavior-
al categories: Leadership, Dependency, Affection, Aggression,
Contact, and Role-Playing. The undergraduates participating
in these play encounters formed four groups: Trained High
Potentials (THP), Trained Low Potentials (TLP), Untrained High
Potentials (UHP), and Untrained Low Potentials (ULP). Train-
ed coders independently scored the children's behavior in
videotapes of the first, sixth, eleventh, and fifteenth play
sessions for the six variables. Inter-coder reliability was
obtained by coding each session twice (except for two which

had mechanical difficulties) by two separate raters.

Analysis of the Data

Scores for each of the six dependent variables were ex-
amined by means of separate 2(trained-untrained) X 2(high
potential-low potential) analyses of variance for selected
sessions. Although Affection and Aggression had been
scored for intensity, only the mean frequency scores were
subjected to analysis because the intensity scores were so
similar to the frequency scores. Table 6 in Appendix E pro-

vides the mean frequency and mean intensity scores for

38

39

Affection and Aggression for the THP, TLP, UHP, and ULP groups.
Table 7 in Appendix F presents the mean intensity per session
for Affection and Aggression, obtained by dividing the total
of the intensity scores per session for each variable by that
variable's frequency of occurrence within the specific session
being examined.

Determination of which sessions' scores to analyze sta-
tistically for each dependent variable was made on the basis
of deviations in the mean scores between groups. Table 8 in
Appendix G presents the mean scores for the THP, TLP, UHP,
and ULP groups. In those instances where the possibility of
significant differences was indicated by inspection of the
means, a 2(trained-untrained) X 2(high potential-low
potential) analysis of variance was performed. Summary tables
of these analyses of variance for the following instances may
be found in Appendix H: Leadership (Sessions 1, 6, 15),
Dependency (Sessions 11, 15), Affection (Sessions 6, 11, 15),
Aggression (Sessions 1, 6, ll, 15), Contact (Sessions 6, 11,
15), and Role-Playing (Sessions 1, 6, 11, 15).

Due to the small N's and unequal cell frequencies, no

analysis over sessions could be performed.

Hypothesis 1

It was originally hypothesized that significant differ-
ences over sessions would be found for the dependent variables
in the direction of the anticipated trends. As noted, how-

ever, the small N's and unequal cell frequencies precluded

a statistical analysis over sessions. Any investigation of

patterns is therefore contingent upon inspection of the means.

40

Since there is no way to evaluate the significance of the
differences between means, such an inquiry would be highly
speculative.

Table 8 in Appendix G presents the means of the depend-
ent variables. Fluctuations between sessions and generally
inconsistent deviations seem apparent.

Because of the small and unequal numbers of subjects
per session, this hypothesis regarding patterns of the de-
pendent variables could not be subjected to statistical

analysis.

Hypothesis 2

 

It was hypothesized that for children encountering train-
ed undergraduates, the frequencies of occurrence of the six
behavioral variables in each session would be stronger in the
anticipated direction than for children playing with untrain-
ed undergraduates. Support for this hypothesis would be ob-
tained from significant training effects and from higher or
lower mean scores for the respective variables. Table 2
presents the mean frequencies of the six variables for the
trained and untrained groups.

No significant differences were found for the two groups

on Leadership or Contact behavior. A significant difference

 

was found for Dependency_in Session 11 (5.: 6.70, d; = 11,

 

p<<.05), revealing that children encountering untrained under-
graduates exhibit more dependency in this session than children

encountering trained undergraduates. Since dependency was

41

Table 2. Cell Means for the Dependent Variables
for the Trained and Untrained Groups.

 

VARIABLE: LEADERSHIP

Session Number
'Condition 1 6 ll 15

 

Trained 29.06 (N=14) 38.44 (N=11) 36.83 (N=8) 35.56 (N=11)
Untrained 21.61 (N=1l) 20,75 (N=6) 39.55 (N=7) 59.25 (N=4)

VARIABLE: DEPENDENCY

 

Session Number
Condition 1 6 11b 15

 

Trained 12.13 (N=14) 18.88 (N=11) 15.10 (N=8) 12.57 (N=11)

Untrained 15.99 (N=11) 12.38 (N=6) 30.08 (N=7) 16.75 (N=4)

 

VARIABLE: AFFECTION

Session Number
99ndition l 4‘ 6a 11 15

 

 

Trained 14.28 (N=14) 8.64 (N=11) 20.70 (N=8) 13.88 (N=ll)

Untrained 14.06 (N=11) 21.63 (N=6) 24.50 (N=7) 16.00 (N=4)

 

VARIABLE: AGGRESSION

 

Session Number
Condition lb 5 11 15

 

 

Trained 30.84 (N=14) 65.88 (N=11) 96.50 (N=8) 75.02 (N=11)

Untrained 83.79 (N=11) 69.13 (N=6) 58.13 (N=7) 109.00 (N=4)

 

aDifference is marginally significant (p<:.10)

bDifference is significant (p‘<.05)

42

Table 2 (continued).

VARIABLE: CONTACT

Session Number

 

 

 

 

 

 

Condition 1 6 4_7 11 15
Trained 22.74 (N=14) 33.88 (N=1l) 50.10 (N=8) 55.74 (N=ll)
Untrained 19.17 (N=1l) 55.88 (N=6) 50.96 (N=7) 85.50 (N=4)

VARIABLE: ROLE-PLAYING

Session Number

Condition 1 6 11 15a
Trained 11.84 (N=14) 18.17 (N=11) 35.47 (N=8) 13.31 (N=ll)
Untrained 2.86 (N=ll) 11.00 (N=6) 7.80 (N=7) 59.50 (N=4)

aDifference is marginally significant (p<:.10)

bDifference is significant (p<<.05)

43

expected to decrease over time and it was hypothesized that
the Trained group would fit this pattern more closely, the
significant finding for Dependency appears to be in the an-
ticipated direction. However, the significant finding of
Session 11 disappeared in Session 15.

A marginally significant difference was found for Affec—
tion in Session 6 (F = 4.09, df - 1/13, p<:.10), the Untrain-
ed group children displaying significantly more affectionate
behavior than the Trained group children. The expectation
was for Affection to decrease and later increase. Children
of trained undergraduates were expected to show the pattern
more clearly. Although somewhat difficult to interpret, the
significant finding for Affection seems to lend slight sup-
port to the hypothesis in Session 6, the children of trained
undergraduates showing less affection.

Aggression in Session 1 was significantly higher for the

 

Untrained group (§_= 4.39, d£_= 1/21, p4:.05). This finding
is difficult to interpret in regard to the hypothesis since
it pertains only to the initial session.

Rolefiplaying in Session 15 was significantly higher for

 

the Untrained group (F = 4.19, d; = 1/11, p<;.10). It was
expected that the Trained group would more closely approxi-
mate the pattern of a decrease in Role-playing over time.

The finding for Session 15 lends slight support to the hypo-
thesis in terms of the Trained group exhibiting less Role-
playing in the last phase of the play sessions. However, one

must be wary of drawing conclusions because of the small N in

44

the untrained condition.

Because the results obtained are so few and so scatter-
ed in regard to the variables, interpretation is difficult.
However, the hypothesis with regard to Training did receive
slight but inconclusive support. The results obtained must
be considered with caution since they may only be chance

findings.

Hypothesis 3

It was anticipated that the behavior of children en-
countering High Potential undergraduates would be more in
the expected direction for each variable in each session than
the behavior of children with Low Potential play participants.
Support for this hypothesis would come from significant
Potential effects and from higher or lower mean scores for
the respective variables. Table 3 presents the frequency
means of the dependent measures on the basis of Potential.

No significant differences were found for Dependengy,

 

Aggression, Contact, and Role-playing. A significant differ-

 

 

ence was found for the two groups' Leadership behavior in

 

Session 15 (§_= 11.21, df = 1/11, Efl4'01) revealing that
children encountering Low Potential undergraduates exhibit
more Leadership behaviors in the last session. This finding
appears to support the expectation that children of High Po-
tential undergraduates would fit the pattern for Leadership

more clearly than children of Low Potential undergraduates.

45

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 3. Mean Scores of the Dependent Variables for the

High Potential (HP) and Low Potential (LP) Groups

Session Number
1 (HP: N=12) 6 (HP: N= 7) 11 (HP: N=9) 15 (HP: N=6)
(LP: N=l3) (LP: N=10) (LP: N=6) __(LP: N=9)
LEADERSHIP
HP 25.86 29.25 41.38 23.88
LP 24.81 29.94 35.00 70.93
DEPENDENCY
HP 12.16 16.88 20.35 11.50
LP 15.96 14.38 24.83 17.82
AFFECTION
HP 12.93 a9.21 d12.70 C6.63
LP 15.41 21.06 32.50 23.25
AGGRESSION
HP 38.69 55.13 83.63 111.88
LP 75.94 79.88 71.00 72.14
CONTACT
HP 20.59 32.13 27.23 66.38
LP 21.32 60.63 73.83 74.86
ROLE-PLAYING
HP 6.76 22.67 20.93 50.38
LP 7.94 6.50 22.34 22.43
aDifference is marginally significant (p‘<.10)

Difference
CDifference
dDifference

is significant (p
is significant (134.025)
is significant (p<.005)

4.01)

46

Leadership was expected to diminish in the last phase as the
child engaged in more cooperative activities with the under-
graduate and did less "testing."

Significant differences between the two groups were

found for Affection in Session 6 (F = 3.40 d£_= 1/13, p<:.10),

 

Session 11 (E 14.80, df = 1/11, p<<.005), and Session 15

(§.= 8.39, g; l/ll,.p41.025). In all three instances, the
children in the Low Potential Group exhibited more affection-
ate behavior than those in the High Potential. The antici-
pated increase in Affection for the HP group in the later
phases was not observed. For Sessions 6 and 11, the find-
ings for Affection in the HP group appear to be in the ex-
pected direction.

In summary, several significant differences were found
between groups for Affection and Leadership. Findings for
Leadership appeared to be in the anticipated direction. Al-
though the HP group showed less Affection than the LP group
in Sessions 6, 11, and 15, the means do not indicate a steady
decline. No statistical differences between groups were ob-
tained for Dependency, Aggression, Contact, and Role-playing.

Again, the results are few and must be viewed with caution.

Hypothesis 4

It was expected that the behavior of children encounter-
ing Trained High Potential undergraduates would more likely
be in the expected direction in each session than the be-

havior of children encountering undergraduates from the re-

maining three groups. Also, training was expected to reduce

47

the differences between the HP and LP groups. Support for
this hypothesis would come from significant training X poten-
tial effects, and from higher or lower mean scores for the
respective variables. Table 8 in Appendix G presents the
mean frequency scores of the six variables for the four
groups.

A marginally significant training X potential interaction

was found only for Aggression in Session 1 (3.: 4.23, d£.= 1/21,

 

p<:.10). An analysis of simple effects was performed to in-
vestigate this interaction. (See Appendix I, Table 15). This
test revealed that children encountering Untrained Low Po-
tential undergraduates displayed significantly more Aggression
than those encountering Trained Low Potential undergraduates
(§.= 5.81, d§_= 1/21'.B‘4'05)' and that those children inter-
acting with the Untrained Low Potential undergraduates also
exhibited more aggression than those interacting with Un-
trained High Potential undergraduates (F = 3.88, d: = 1/21,
£41.10). Thus training appeared to reduce the effects of po-
tential. Children seen by UHP and ULP undergraduates ex-
hibited a marginally significant difference in frequency of
Aggression displayed while there was no difference between
children of THP and TLP students in Aggression. In addition,
training appears to have diminished the effect of the low
potential undergraduates with regard to children's Aggression.
That is, less Aggression was shown by children interacting
with TLP than ULP students. Again, since the results are

limited, they must be accepted with reservation.

DISCUSSION

Overview

The present study was undertaken to investigate the pro-
cess of play encounters between children and trained and un-
trained, high potential and low potential undergraduates in
reference to six dimensions of children's behavior - Leader-
ship, Dependency, Affection, Aggression, Contact, and Role-
playing. The children were divided into four groups depend-
ing upon whether they interacted with a trained high poten-
tial, trained low potential, untrained high potential, or un-
trained low potential undergraduate. Comparisons were made
between these groups, as well as between the trained and un-
trained, and high potential and low potential groups.

Due to small N's and unequal cell frequencies, only
intra-session statistical comparisons could be made.

Hypotheses

 

Patternsyof the Dependent Variables

Statistical analyses of the dependent variables over
sessions could not be performed because of the small number
of subjects and the unequal cell frequencies. These factors
also impeded use of the means to determine whether any pat-
terns of the variables had occurred. Although statistical
data are lacking, it seems important to consider the premises

upon which the expectations for the patterns were based, as

48

49

well as any relevant observational data. Inspection of the
means will be referred to as suggestive data.

On the basis of theoretical assumptions and implications
drawn from Guerney's and Stover's (1971) research, it was an-
ticipated that the following patterns would emerge for the
dependent variables: As the child experienced the permissive
and accepting environment in the play situation, he or she
would become freer in the expression of his or her needs and
feelings, approaching the situation more confidently. Hence

Leadership was expected to increase over sessions, declining

 

somewhat in the last phase as the child engaged in more
egalitarian and cooperative activities with the participant-

undergraduate. At the same time, Dependency would decline

 

over sessions as the child asserted himself or herself more,
seeking less approval and acknowledgment from the undergraduate.
Mean scores for Leadership in all conditions showed fluctua-
tions. No definitive conclusion could be drawn for that vari-
able's pattern. However, children of HP subjects exhibited
significantly less Leadership in Session 15 than children of

LP subjects - a finding which was in the hypothesized direc-

tion. In regard to Dependency, inspection of the means does not

 

indicate the anticipated steady decline, but rather fluctua-
tions which at times are suggestive of an increase followed
by a decrease. In observing some of these play sessions over
time, it did often seem that the child showed a rise in De-
pendency responses as he or she tested the limits of the new

situation, sought information, engaged the undergraduate in

50

attempts to establish a relationship, and tried to test how
responsive the undergraduate was to what the child wanted.

Thus it would seem that Dependency might well reveal an in—

 

crease and then decline, which if it did occur in the filial
sessions of Guerney's and Stover's study (1971) was obscured
by the fact that their measures were taken for thirds of
treatment time over a twelve to eighteen month period. It
may well be that fifteen sessions is too short a time to ex-
pect the patterns of the variables to approximate the patterns
established in the filial sessions. Stollak (1968) found no
significant change for Dependency over ten sessions of play
interaction between children and undergraduates while Leader-
ship increased from the first to the fifth but not the fifth
to the tenth sessions. Stover and Guerney (1967) found no
significant differences in verbal leadership or dependency
for an experimental filial therapy group and a control group
from the first to the fourth play sessions. Thus, not only
are our findings impaired by the small N's and unequal cell
frequencies, but our fifteen sessions may be too short a time
period for significant changes to occur in the behaviors
studied. It may also be that this fifteen session period
represents, at best, only the initial third of the pattern re-
ported by Guerney and Stover (1971) and hence, shows fluctua-
tions which are obscured in Guerney's and Stover's data.

Concomitant with the changes anticipated for Leadership

 

and Dependengy was the expectation that Aggression would in-

 

crease, peak, and then diminish in the last phase as the child

51

worked through some of his conflicts and feelings. It was

anticipated that Affection would decrease as the socially

 

desirable responses dropped out, followed by an increase in
the last sessions as negative feelings were resolved and more
genuine expressions of positive affect occurred. Mean data
is somewhat suggestive of an increasing trend for Aggression.
Again, our fifteen sessions are probably too short a time
period to approximate the anticipated pattern. However, the

slight increase suggested for Aggression does seem to corre-

 

spond to the expected increase for the first phase of play
sessions.

In their study, Guerney and Stover (1971) distinguished
between "Aggression Toward the Mother" and "General Aggression."
The expected pattern for aggression was found for the former
but not the latter category. There were also differences in
the patterns of Aggression Toward the Mother which emerged for
those children assessed as aggressive versus those who were
categorized as withdrawn, as well as apparent differences for
General Aggression between the general sample and the most
improved sample. Since our study did not distinguish between
types of children and presenting problems, degree of improve-
ment, or categories of Aggression, some data may well have
been obscured.

In general, inspection of the means for Affection showed

 

fluctuations and no definite patterns. What does seem evident
is that the decline of Affection followed by an increase in

the last session did not emerge as expected. Again, our data

52

may only be indicative of the early phase of play interaction
and may reveal the fluctuations hidden by more extended time
analyses. It is also interesting to note, however, that Reif
and Stollak (1972) excluded the variable Affection from their
study because its occurrence was so infrequent. Rutledge
(1974) found that over 90% of children's behavior in the first
two taped sessions (1 and 6) fell into the neutral (nonrecogni-
tion) category. In studies of protocols of play sessions,
Landisberg and Snyder (1946) found that expression of positive
feelings remained at about the same level, while Finke (1947)
found no pattern for positive statements. Guerney's and
Stover's (1971) data revealed a steady decline for Affection
over sessions. In this study, poor videotapes often made
viewing of the children's expressions difficult or hard to
determine, thus affecting our data collection and results.
clearly, further investigation ofthe variables of warmth,
affection, and expression of positive feelings is needed to
clarify just what process does occur in play sessions.

It was anticipated that as the child resolved some of

his concerns, Role-playing would decrease and Contact increase,

 

the child engaging in less distancing, child-oriented fantasy
and instead becoming more conversational and attempting to en-
gage the undergraduate jointly in more activities. Although
some fluctuation between session means did seem to occur, there
seemed to be quite consistent suggestive evidence over all
groups for the increase in Contact from the first to the fif—

teenth sessions. It does appear that over time the child

53

begins to establish some type of relationship with the under-
graduate, and participates in or shares activities with the

student. Rolefiplaying'on the other hand, seemed most commonly

 

to indicate a pattern contradictory to the one expected.
Whereas declines over sessions had been anticipated, it seems
that the pattern may instead be one of increases, followed by
a decrease in the final phase. This pattern would appear to
be consistent with the type of play encounters in which trained
undergraduates in this study were encouraged to engage. Fan-
tasy behavior on the part of the child was viewed as a posi-
tive event and trained undergraduates often participated in,
shared, and enjoyed the children's fantasies. Hence Contact
and Role-playing were not necessarily mutually exclusive cate-
gories. Reif and Stollak (1972) emphasize the association of
fantasy behavior with the process of achieving ego mastery and
coping with internal conflicts. They suggest that fantasies
involving problems of emotional integration and concerns with
identification and role-behavior (hence role-playing) may rep-
resent higher levels of fantasy. The pattern then may be that
as the child experiences the permissive environment and begins
to trust his play participant, he moves from the lower level
of fantasy (the more simple, stereotyped responses) to the
higher level which would incorporate Role-playing. It could
be speculated that as the child attains some emotional inte-
gration of his feelings and learns to express his feelings
more directly the Role-playing behavior will drop off. The

steady decline in Role-playing found by Guerney and Stover (1971)

54

may be, again, a result of the difference in number of sessions

and the time period between our data and theirs.

Trained versus Untrained Groups

 

Very few significant findings resulted from analyses of
differences between the means of the dependent variables for
children playing with trained and untrained undergraduates.
Because only intra-session analyses could be performed, these
findings are limited in terms of their usefulness for an under-
standing of the differences between the two groups.

Aggression in Session 1, Affection in Session 6, Depend-
ency in Session 11, and Role-playing in Session 15 were sig-
nificantly higher for children encountering untrained under-
graduates. Thus in these four instances training does seem
to have an effect, although for Aggression one might speculate
that the sample of children could have biased the results.

For example, did the untrained undergraduates happen to get
more acting-out children? Had we been able to include a
larger number of subjects in our study, additional results

may have emerged. Because analyses of the data over time
could not be conducted, the findings which we did obtain are
difficult to interpret with reference to specific patterns for
the variables. However, having anticipated the children of
trained undergraduates to exhibit the expected patterns to a
greater extent, the results for Dependency, Affection, and
Role-playing seem to be in the general direction. Only with
additional data from the completed replication of this project,

however, can the effect of training be clearly assessed and

55

patterns of the variables more definitely ascertained. The
lack of more significant differences between the groups may
have been largely a function of the small N's. Also, the
effect of training may not have been as critical a factor as
the amount of caring, interest, and commitment on the under-

graduates' part.

High Potential versus Low Potential Groups

 

It was anticipated that children playing with high poten-
tial undergraduates would display the expected patterns of
the variables to a greater degree than those encountering low
potential undergraduates. Again, restriction to intra-session
analyses made this difficult to determine. Significant re-
sults obtained revealed that children of low potential under-
graduates displayed more Affection in Sessions 6, 11, and 15,
and more Leadership in Session 15. These findings are some-
what consistent with our expectation for the directions of
Affection and Leadership. Children of HP undergraduates were
expected to show less Leadership in Session 15 when it was
anticipated that they would be more engaged in egalitarian
activities with the undergraduate. Affectionate behavior was
expected to decrease (children of HP subjects showed less than
children of LP subjects) but the anticipated increase in
Session 15 was not evident. However, our data may be more con-
gruent with Guerney's and Stover's (1971). They found a steady
decline in Affection over sessions. The higher level of Affec-
tion for the children of LP subjects may indicate instead more

of a desire to please and acquire the undergraduates' acceptance,

56

whereas there might be less concern with those issues for the
HP children. Again, interpretation is difficult. The small
N's and unequal cell frequencies clearly hindered determination
of the differences between the behavior of children encounter—
ing high potential and those interacting with low potential
undergraduates. In addition, the effect of potential may

have been obscured by that of training. Initial HP-LP differ-
ences may have been diminished by training, which would then
have affected the overall differences between the two cate-
gories for potential. However, it is also not unlikely that
differences between the two groups were not extensive, or had

a limited effect upon the children's behavior. Schreiber
(1972) found that the high potential and low potential under-
graduates differed significantly only on the category "Com-
munication of Acceptance" in Session 15. No significant dif-
ferences were established for the other two categories -
"Allowing the Child Self-Direction" and "Involvement." It may
be that the project was not long enough for significant dif-
ferences between the HP and LP undergraduates to emerge. This
pertains as well to the play patterns of the children. Perhaps
the high potential and low potential undergraduates were not
such disparate groups as expected, and were not experienced as
such by the children. Just by the nature of the project, one
may assume that undergraduates who volunteered were interested
in working with children. A relationship with a caring, atten-
tive individual may have been a more critical factor to the

child than the degree of sensitivity displayed by the student

57

or the specific principles of interaction which the under-

graduate employed.

Training X Potential Interaction

 

It was expected that children of HP undergraduates and
LP undergraduates would differ significantly in the untrained
condition but not necessarily in the trained. It was antici-
pated that the children of trained HP students would fit the
anticipated patterns most clearly.

Training did seem to reduce the effect of potential in
Session 1 for Aggression. Children seen by untrained low po-
tential undergraduates displayed significantly more Aggression
in this session than children encountering either trained low-
potential undergraduates or untrained high potential under-
graduates. No significant findings were obtained for any
other behavioral variables. Because children were not divided
on the basis of their presenting behaviors, we do not know
how much these significant findings were influenced by this
factor. It may be that the greater degree of aggression ex-
pressed by children of untrained low potential undergraduates
was a function of the type of presenting problem of the
children (e.g., acting-out versus withdrawn behavior) rather
than the lack of training of the undergraduates. On the basis
of inspection of the means, it is not feasible to attempt to
draw conclusions. Fluctuations and patterns inconsistent with
those expected seem apparent. It is impractical to attempt to

interpret the extent of differences between the two untrained

groups, or the extent to which the trained high potential group

58

approximates the expected pattern. Again, it is hoped that
additional data from the replicated project will enable this

hypothesis to undergo statistical analysis.

Limitations of theyPresent Study

The present study yielded few significant findings.

Those that were obtained were often difficult to interpret.
This section will consider some of the study's limitations
and possible reasons for the lack of significant results.

One very serious shortcoming of the present study was
the small and unequal number of subjects in each cell of the
experimental design. As already noted, this difficulty pre-
cluded the performance of any statistical analyses over
sessions. Thus our study of the patterns of the dependent
variables was seriously impaired. Instead of being able to
make definitive conclusions based on our data, it was only
possible to note intra-session differences between groups
which were difficult to interpret.

The small number of children and undergraduates was partly
the result of a low referral rate. Although undergraduates
were willing and anxious to begin their play sessions, some
waited over a year for a case to be assigned to them. In
several instances parents withdrew their children from partici-
pation before the fifteen sessions were concluded. Videotaping
errors and mechanical difficulties with the equipment also re-
duced the total number of tapes which were available for analysis.

Very often even when tapes had been completed for a session the

59

sound was poor or it was hard to distinguish a child's facial
expressions. Scoring of such tapes was difficult and subject
to error. Hopefully, new and more reliable videotape equip-
ment and the addition of data from the replication of this
study will help to clarify the results presently obtained.

The videotaping itself was also an interfering factor
in the sessions. The camera was clearly visible to the child.
Some children became so involved with their activities that
they completely ignored the camera's presence. Others, how-
ever, were obviously affected by the equipment. Responses
ranged widely. Some children were shy and embarrassed, and
slow to begin playing. One child hid under a table through-
out a major part of one session. Another made faces and
clowned in the playroom. Still another became angry, shot at
the camera and attempted to block its view with a hat. To
what extent the taping influenced the occurrence of the
children's behavior (e.g., the aggression noted in the previous
example) is uncertain. However, it was a factor. Since only
those sessions needed for the project were taped, they were
special times to the child. Use of videotape equipment behind
a one-way mirror would alleviate this problem.

In reviewing the research on child psychotherapy, Levitt
(1971) cited some studies which show no relationship between
outcome and number of treatment hours. However, children in
this study participated in only fifteen half-hour play en-

counters. Can we really expect that significant changes will

occur within that short a time interval? Often, too, these

60

sessions were interrupted by quarter breaks and sometimes even
the whole summer. It is impossible to assess just what effect
these interruptions had. However, they clearly interfered with
the constancy of the relationship and, quite likely, the pat—
terns of the children's behavior which emerged over time. Per-
haps we would have acquired more meaningful results and clear
patterns for the variables if we had extended a study over a
longer period, or increased the number of weekly sessions.

An important variable which was not considered in this
study was the classification of the children's presenting
problems. Children in Guerney's and Stover's (1971) study were
classified as either Withdrawn, Mixed-Intermediate, or Aggres-
sive. Withdrawn and Aggressive children showed very different
patterns of behavior in Aggression Toward the Mother. It may
be that some of our data were obscured because of the lack of
classification. It is likely that children who tend to act-
out will show the development of quite different patterns of
behavior over sessions than withdrawn children. For some vari-
ables perhaps the patterns would be similar, but the period of
time over which they evolved might differ for the two cate-
gories of children.

Sex of the undergraduate and child was another variable
which was not taken into account and which may have important
implications. In a study of the effect of children's behavior
upon parents, Osofsky and O'Connell (1972) studied 5-year old

girls in dependence - and independence-producing situations.

They found that the children's behavior did have an influence

61

upon the parents, who showed some differences in their re-
sponses. Fathers were more action oriented; they were more
likely to help the child physically or, at times, to withdraw
or become detached. In contrast, mothers were more supportive
and encouraging of the children's efforts, being more likely

to question, explain, and make comments, and to resist giving
direct help immediately. Although the children's behaviors
seemed to be generally consistent with mothers and fathers,
there were also some differences with each parent. For ex-
ample, girls in the independence-producing situation showed
more task-oriented behavior with their fathers than with their
mothers. Thus it appears that children do exhibit some contrast—
ing behaviors in their interactions with their parents, and that
mothers and fathers show some differences in their responses to
their children's behavior. These findings have direct bearing
upon the current study. Had we been able to have a much larger
sample we could have divided male and female children into dif-
ferent groups depending upon the sex of the undergraduate with
whom they were interacting. Perhaps differences in the patterns
of behavior would have emerged, dependent upon the sex of both
the child and the student. Research is needed to determine

just what impact the sex variable does have upon the children's
and the undergraduates' behavior. Do girls tend to exhibit

more affectionate responses with female undergraduates? Are
male undergraduates more likely to promote dependent behaviors

in the child? Do cultural factors predispose females, for

example, to be more empathic and sensitive to children's

62

feelings, and to touch children more often, whereas males are
more task-oriented? Is there more of a need for males to
"prove" themselves through their accomplishments, and does

this result in a competitive element with male children? I

am thinking in particular of an untrained low potential male
undergraduate who often "out-did" the child with whom he was
playing, either in their role-playing, at games, or in build-
ing with tinker toys. The child's attempts to master his feel-
ings of inadequacy were impeded. When the undergraduate was
obviously able to complete a more elaborate building project,
the child gave up on his own, immediately exhibited more de-
pendent responses (e.g., "Will you make one for me?") and
withdrew from that activity in later sessions. A high level

of Aggression occurred throughout many of these play inter-
actions, and much of the Aggression was directed at the under-
graduate. Some of this behavior must clearly have been in
response to the interactions which the child had with the
student. Even in role-playing, if the child "shot" the under-
graduate and locked him in jail, the undergraduate always es-
caped and then "captured" the child. Of course, it may be that
the lack of training and the student's selection as a "low po-
tential" participant affected these interactions, but it would
still be important to determine whether the sex of the child
and the undergraduate were factors as well. In a study of non-
verbal behaviors in the first session, Ducat (1973) found that

untrained male subjects spent less time than either untrained

female or trained male subjects in looking at a child and

63

watching the child's play activity. Thus there were male and
female differences in this instance and training altered the
male behavior. Perhaps we need to establish what other differ-
ences exist between males and females to evaluate training pro-
cedures and determine whether those procedures should be varied
somewhat for the two sexes.

An additional variable which needs to be considered in
its relation to children's behavior is the child's age. Lebo
(1952) noted that differences were apparent in children's play
when their behavior was assessed on the basis of age. Thus
children's sex, classification as to the problematic behavior,
and age are all relevant factors to be taken into consideration.

As already noted, one of the possible reasons for our
failure to acquire more significant and meaningful results is
related to our small number of subjects and the unequal cell
frequencies as well as the limited time period employed. The
selection and training of the undergraduates may be additional
factors. Perhaps the training procedures were not as pro-
ductive as expected, and may need to be altered in some re-
spects. Because the groups for the trained undergraduates
were led by a senior faculty member and two clinical psycho-
logy graduate students, it is possible that there were differ-
ences across groups. If so, these differences might have af-
fected the undergraduates' performance in the play sessions
and consequently, the children's behavior as well. It may have
been necessary to have one trainer work with all the trained

undergraduates. It may also be that different methods should

64

be employed. Perhaps some time should be spent initially in
developing listening and feedback skills with one another.
Such a procedure might stimulate the growth of trust in the
group while allowing the students to develop their sensi-
tivities with one another and become more attumed to their

own feelings. This type of method would also be helpful in
terms of teaching students how to give constructive criticism
to one another. In one particular group with which I worked
as part of the 1972-73 replication project, one individual

who obviously knew, or thought he knew, the "right way" to
respond to children (something he learned through classes and
readings) was extremely insensitive and condemning in his com-
ments about another member's answers on the Sensitivity to
Children Questionnaire. The other member felt attacked and
humiliated. It was clear that the group would feel very wary
about exposing themselves and their uncertainties, and hesi-
tant to take risks in trying out new behaviors if even one
member was going to act as a critical judge who already knew
all the answers. If undergraduates are going to learn how to
be sensitive, empathic, and non-directive with children, it
seems apparent that they should also develop these skills with
one another, and that the leader should serve as a model. In
their study of filial therapy, Stover and Guerney (1967) found
that of the two experimental (trained) groups, mothers in the
one led by the more experienced clinician showed higher levels
of reflective behavior in sessions with their children. It

seemed that the more experienced group therapist was able to

65

provide higher levels of empathic understanding in his re-
lationship to mothers in their weekly meetings. Thus the
mothers were able to explore their feelings in a therapeutic
climate similar to the one they were being taught to provide
for their children. It seemed that mothers were able to be
more helpful with one another in this group, and had a stronger
model in their leader. Thus greater learning of empathic be-
havior seemed to occur and was evident in the play sessions.
These findings support the idea that undergraduates need to
learn how to be sensitive and responsive with one another in
order to be more effectively empathic with children.

Another consideration in training undergraduates should
be some experience in observing children's behavior. Many
times undergraduates just do not seem to know what to look for
in their sessions with a child. When asked to describe a ses-
sion, an undergraduate will often respond that, "It was O.K."
or "Well, we played with the dolls for a while. Then Susie
painted. That's all." Specific references to the child's or
undergraduate's feelings are often lacking, as well as more
detailed content or an understanding of what seemed to be hap-
pening with the child in that particular session and over time.
Learning how to observe the play interaction may well be com—
bined with more experience in playing with children. Perhaps
in addition to videotaping the undergraduates' play sessions
with their "normal children," "live supervision" could be
carried out simultaneously. Feedback is usually more helpful

if it is given immediately and while the session is still "fresh"

66

in the undergraduates' memory. Perhaps, too, more role-
playing and discussion of problematic situations and childhood
disturbances would ease the transition from playing with a nor-
mal child to the experience of encountering a "clinic-referred"
child, an adjustment which may have been quite anxiety provok-
ing for some undergraduates.

Few significant differences were found with regard to
Potential. Possibly the measures used to discriminate High
Potential and Low Potential undergraduates were not as selec-
tive or sensitive as had been anticipated, or perhaps the
scoring systems were not precise enough. The measures them-
selves may not have tapped those characteristics which were
directly related to the undergraduates' performance in the
play situation. Further research is needed to determine ex-
actly what therapist or adult behaviors are related to success-
ful outcomes in play therapy and play encounters. When these
variables can be more clearly delineated, new tests and scor-
ing methods may need to be devised. In this project, it was
quite apparent that students who volunteered and were willing
to make a long-term commitment were interested in working with

and learning about children. In this sense, the undergraduates

involved were a select group. Having observed at least one

undergraduate in each of the four groups - trained high poten-
tial, trained low potential, untrained high potential, and un-
trained low potential - I was struck by their perseverance,in-

terest, and commitment. Even those untrained subjects who re-
ceived no feedback and only minimal encouragement and support

over the 15 sessions remained enthusiastic about their encounters

67

with the children. Even though some students had been selected
as "low potentials" they usually seemed just as interested in
children as the "high potentials." For example, one untrained
undergraduate, whom I later learned was a "low potential" student,
was working in a volunteer tutoring project with elementary
school children at the time of her play sessions. When

these sessions ended, she maintained contact with the child
involved by visiting and essentially becoming a "big sister."
Students I encountered also followed-up "their“ children after
the post-asseSsment to learn what recommendations had been
made. In fact, some undergraduates actively participated in
the post-assessment process. Clearly interest was high among
these students - regardless of potential. Perhaps, at least
in these fifteen sessions, for children to experience the
warmth, liking, and mere presence of an interested undergradu-
ate was more critical then properly reflective and empathic
statements. Over a longer time period, other factors might

be more influential. It seems though that the child's per:
ception of the undergraduate as a caring adult would be a
vital factor influencing his behavior and actions - and a
factor not necessarily accurately assessed by observers'
measurements. Schreiber (1972) found that trained undergradu-
ates in this project exhibited significantly more "empathic"
behavior, especially "Communication of Acceptance," than un-
trained undergraduates. However, this study revealed few
significant and conclusive differences in children's behavior

between the two groups. Perhaps other variables - such as the

68

child's perception of the undergraduate, the student's degree
of interest in and liking for the child, and the student's
warmth and commitment - were more determining factors. Only
future research can clarify this issue. Some way of assess-
ing the effect of commitment and interest seems needed. Re-
sults may have been affected also by the children involved in
the project. Post-analysis revealed that children seen by un-
trained undergraduates were perceived by the parents as hav-
ing a somewhat higher initial level of dysfunctioning (See
Stollak, Green, Scholom and Schreiber, 1973). Both the pat-
terns of the children's behavior and analyses of the findings
on the basis of training and potential could have been affect-
ed by this factor.

An additional point must be made regarding the selection
of high and low potential undergraduates. Since only paper-
and-pencil inventories and questionnaires were used - albeit
the Sensitivity to Children questionnaire presumably related
to one's behavior in a problem situation - it may be that sub-
jects answered with what they considered a proper response
rather than the actual way they would behave. Hence their
answers may not necessarily be reflective of how they would
act. Perhaps some type of interactional measures would need
to be made of the undergraduates' behavior in actual play
sessions or role-playing situations.

Rutledge (1974» in an interactional study of undergradu-
ates' and childrens' behavior in the first and sixth play ses-

sions using the same data as this study, found that 92% of the

69

children's time was spent in behavior of the neutral category -
that is,behavior exhibiting no discernible warmth or interest
in the adult present. This finding indicates that our lack
of results may have been due not only to the small N, unequal
cell frequencies, and other factors already mentioned, but to
the play situation itself. Perhaps because of the nature of
the playroom with its many different and attractive toys and
varied activities, the child focused his interest on the room
instead of interacting with the undergraduate. After all, the
child was in a strange place with a person he had only recently
met. Therefore, he was probably anxious and uncertain.
Rutledge (1974) speaks of the child's "incapacity for clear
and positive direction," and the child's ambivalence about ap-
proaching. Time is probably a critical factor. Perhaps as
the sessions increase, the child's behavior will become clear-
er and more focused towards the adult. How long it takes a
child to develop trust and enter into a meaningful relation-
ship with an adult may be an individual matter. Six or even
fifteen sessions is probably not long enough for some children.
One area which needs to be considered in evaluating play
therapy or the efficacy of play sessions is the carry-over
effect of the sessions. That is, even though a child may en-
gage in primarily neutral behavior in the play sessions, does
the child's participation in play encounters seem to affect
his behavior with family and peers? Clearly some post-session

behavioral indices or interactional measures would be needed.

70

One major issue not yet considered is the efficacy of
the dependent variables used in this study and the short-
comings of their definitions. In many instances valuable
information may have been lost because the categories used
were so global. For example, it would have been revealing
to divide Leadership into 1) Commands Directed to the Adult,
Directions or Giving Ideas About Activity to the Adult and
2) Proclamations of Intended Self Activity. In that way, re-
sponses related to independence and self-assertion on the
child's part could be distinguished from directive responses
made to the undergraduate. With regard to Dependency,
information-seeking is not necessarily a measure of dependent
behavior and should be used as a separate category. Often
specific requests for information (e.g., "What time is it?"
"Can they see us?" "Can I shoot the windows?") are indications
of curiosity, exploration, testing the limits, or simply a
desire to know something. Consequently it would be more profit-
able to use the following three categories instead of "Depend-
ency" - 1) Seeks Help, 2) Seeks Information, 3) Seeks Approval.

The variables Affection and Aggression need to be sub-
divided to provide more data. "Behavioral (nonverbal) Expres-
sion of Affection/Aggression" and "Expression of Positive/
Negative Feelings" would seem to be more productive categories.
A definite shortcoming with the global definition of Aggres-
sion used in this study was the contamination which resulted
because a separate category for mastery - for example, Reif's

and Stollak's (1972) "Behavioral Expression of Object

71

Mastery," - was not included. Thus although shooting darts
against the wall was arbitrarily established as a #1 rating
for Aggression, such activity may indicate skill mastery in-
stead, especially when the child makes a statement such as
"I'm going to try and shoot the ceiling...see if it sticks."
An additional consideration may be inclusion of some type of
interactional measure to determine whether a child's behavior
is related to some provocation or response on the undergradu-
ate's part,or to a modeling effect.

Reif's and Stollak's (1972) category "Joint Participation
in an Activity" is similar to the category "Contact" but dis-
tinguishes what role the child assumes - i.e., Dominant Role,
Submissive Role, or Discernible Role. Hence more information
is derived using their category. A basic difficulty in using
"Contact" as a variable was in distinguishing between parallel
play and joint participation or genuine sharing. For example,
a child and an undergraduate fingerpainting side-by-side are
sharing in an activity, yet there may not be any real "contact"
between the two. Often an activity such as card-playing, throw-
ing a football, or playing checkers may be a "distancing" ac-
tivity which impedes the development of closeness between the
child and undergraduate. Perhaps some measure of involvement
is needed if the category of "Contact" is to be retained.

Role-playing seems essentially to be a straightforward
category. However, Reif and Stollak (1972) discuss the differ-

entiation of various levels of fantasy behavior. More precise

categories will need to be developed to explore fantasy play.

72

Directigns»for_Future‘Research

The present study has attempted to assess the changes in
children's behavior which occur in play sessions with under-
graduates. As already noted, additional data is needed to ar-
rive at any definitive conclusions regarding this process.
For example, outcome measures need to be carefully analyzed to
determine what changes in child behavior have occurred over
time. Data analyzed from the small sample of this project
showed a trend for parents of children seen by trained under-
graduates to perceive more improvement in their children's be-
havior (Stollak, et a1,, 1973). It would be useful to study
just what improvements the parents noted, and then attempt to
coordinate these findings with results regarding the behavior
patterns in the sessions. We may find important behavioral dif-
ferences between children who are most improved, and those with
less successful outcomes. Such information would have implica-
tions for training procedures and, quite likely, for an under-
standing of the types and ages of "problem children" who might
benefit from play sessions. Perhaps, too, new measures will
need to be devised to assess changes in the child's feelings
and self-concept instead of relying only on outcome ratings
related to the perceptions of others. Although projective
tests were used in the extended project, interactional tasks,
objective scales, and some type of structured doll play
might provide more scorable information. Before and after
ratings of the children in group play sessions could be used

to assess the child's behavior and interaction with peers.

73

As Levitt (1971) has observed, one factor which presents
difficulties in studies of child psychotherapy relates to the
rapid developmental changes and growth which occur in children.
In this study all children participated in play sessions. A
no-treatment group could have been included to assess to what
extent such problem behaviors as the project children exhibited
would improve over time. Additional variations such as treat-
ment of mothers only, filial therapy, parent groups, or family
sessions might be compared, and improvements in children's be-
havior studied. Increasingly emphasis seems to be on thera-
peutic methods other than one-to-one play sessions. Clearly,
research is needed to assess the effects of various treatment
methods. Only with such information can we provide more ef-
ficient services. Also, it would be useful to place more em-
phasis in research on changes in the parents' behavior, and
to ascertain what effect such changes have upon their children's
behavior. Levitt (1971) cites studies indicating that treat-
ment is most effective when the focus is either the mother and
father, or the mother, father, and child. Additional work
such as Guerney's and Stover's (1971) is needed in this regard.

Even though future research may provide continuing sup-
port for the argument that the child should not be seen alone
in treatment, there will obviously still be cases in which the
parents will not, or perhaps cannot be involved. In such in-
stances, play sessions for the child may be helpful. If fu-

ture studies can provide clearer understanding of the play

process, and of the effect of such variables as age, presenting

74

problem, and sex, we will certainly be able to offer more
beneficial treatment. In this regard, interactional studies
of adult-child behavior and changes over time are needed. We
require more information about which adult behaviors and
therapist characteristics promote the play process and produce
the desired child behaviors. Then more appropriate and ef-
fective training procedures can be devised.

In recent years, focus has shifted increasingly away from
treatment of illness or sickness to improvement of mental
health and prevention of psychological difficulties. The
human potential movement, encounter groups, and workshops to
increase personal awareness and communication skills are promi—
nent features of our current times, Emphasis is now much more
on human growth and increasing positive functioning. Play
sessions, too, may be viewed in these terms. If we can under-
stand more precisely the changes which occur in play inter-
actions and which induce positive behavioral changes in the
child, then we will be better able to promote a child's growth
and development. A related issue is the training of under-
graduates and other nonprofessionals. Research indicates that
college students can be trained to alter their behavior in a
therapeutic direction (Schreiber, 1972; Stollak, 1968, 1973).
Stollak (1969) regards undergraduates'age, malleability, open-
ness to suggestion, and eagerness to learn as positive charac-
teristics. Gruver (1971) notes the relevance of focusing upon

college students, not only because of those inherent charac-

teristics which make them suitable for working with troubled

75

clients, but also because of the positive developmental in-
fluence such work has upon their personalities. Rioch (1966)
stressed the importance of non-professionals becoming "con-
structive, better integrated citizens themselves" through

their training and work. Guerney (1969) notes the importance
of "the education and training of parents as a way of prevent-
ing psychological difficulties, and not merely alleviating
them once they arise" (p. 391). The training of undergraduates
to be attentive to children's emotional needs and to be effect-
ive play participants will surely increase their abilities as
potential parents, teachers, and educators. If such programs
can be implemented and more widely used, it is quite probable
that we can reduce the incidence of child psychopathology
while at the same time easing the demand on mental health

professionals.

SUMMARY

The present study was designed to investigate the
process of play encounters between children and trained and
untrained, high and low potential undergraduates in refer-
ence to six dimensions of children's behavior - Leadership,
Dependency, Affection, Aggression, Contact and Role-playing.
On the basis of theoretical assumptions, clinical observa-
tions, and research findings, the development of particular
patterns for each of the six variables was predicted.
Leadership was expected to increase over sessions,diminishing
in the last phase as the child engaged in more egalitarian
activities with the undergraduate. Dependency and Role-
playing were expected to decline in frequency over time,
whereas Contact was expected to increase over sessions.
It was anticipated that Aggression would also increase in
frequency, reaching a peak and then declining. For Affection,
the trend expected was a decline in frequency followed by a
rise in occurrence in the later phases as negative feelings
were expressed and resolved. The following specific
hypotheses were then formulated in reference to the six
behavioral variables: 1) Significant differences over sessions
would be found for the dependent variables in the direction
of the anticipated trends. 2) The behavior of children inter-

acting with trained undergraduates would more clearly fit the

76

77

expected patterns than the behavior of children encountering
untrained undergraduates. 3) The behavior of those children
playing with high potential undergraduates would more closely
approximate the predicted patterns than the behavior of child-
ren engaged with low potential undergraduates. 4) Training
would tend to reduce the differences in behavior between
children interacting with high potential undergraduates and
those playing with low potential.undergraduates. However,
children playing with trained high potential undergraduates
would show greater levels of the dependent variables in the
predicted directions.

Subjects were twenty-seven clinic-referred children di-
vided into the following four groups depending upon the type
of undergraduate with whom they interacted: Trained High
Potential (THP), Trained Low Potential (TLP), Untrained High
Potential (UHP), or Untrained Low Potential (ULP). Due to
the small N's and unequal cell frequencies, only within-
session statistical comparisons could be made between groups.
Thus the first hypothesis could not be tested. Determination
of which session scores to analyze statistically for each
dependent variable was made on the basis of deviations in the
mean scores between groups. In those instances where the
possibility of significant differences was indicated by in-
spection of the means, a 2(trained-untrained))(2(high potential-
low potential) analysis of variance was performed. An analysis

of simple effects was carried out in the one case where a

significant interaction effect was found.

 

78

The results showed few significant differences between
groups, no hypothesis receiving definitive support. Inter-
pretation was difficult because the results were few and
scattered. No statistical analysis over sessions could be
performed because of the small N's and unequal cell fre-
quencies.

It was concluded that additional data is needed to
clarify the scattered findings and suggestive trends which
resulted from this study. Modifications of the scoring
system were proposed. Limitations of the present research,
such as the small number of children seen, the short time
period, the restriction of the play setting, and the need
for consideration of variables like age, sex, and presenting
problem of the child were reviewed. The need to study the
interactional relationship between the undergraduates' and
children's qualities and behaviors was discussed. Directions

for future research were explored.

BIBLIOGRAPHY

BIBLIOGRAPHY

Allen, F. Psychotherapy with Children. New York: W. W.
Norton and Company, Inc., 1942.

Axline, V. M. Play Therapy. New York: Ballantine Books, Inc.,
1947.

Axline, V. M. Play therapy procedures and results.
American Journal of Orthopsychiatry, 1955, 25, 618-626.

Becker, W. C., Madsen, C. H., Jr., Arnold, C. R., and
Thomas, D. R. The contingent use of teacher attention
and praise in reducing classroom behavior problems.
Journal of Special Education, 1967, 1, 287-307.

Berkowitz, B. P. and Graziano, A. M. Training parents as
behavior therapists: A review. Behavior Research and
Therapy, 1972, 10, 297-318.

Carkhuff, R. R. Differential functioning of lay and pro—
fessional helpers. Journal of Counseling Psychology,
1968, 15, 117-126.

Carkhuff, R. R. and Truax, C. B. Training in counseling
and psychotherapy: An evaluation of an integrated
didactic and experimental approach. Journal of Con-
sulting Psyghglggy, 1965, 29, 333-336.

 

Cowen, E. L., Gardner, E. A., and Zax, M. Emergent
Approaches to Mental Health Problems. New York:
Appleton-Century-Crofts, 1967.

Cowen, E. L., Zax, M., and Laird, J. D. A college student
volunteer program in the elementary school setting.
99mmunity Mental Health_gournal, 1966, 2, 319-328.

Davison, J. C. The training of undergraduates as social
reinforcers for autistic children. In L. P. Ulmann
and L. Krasner (Eds.) Case Studies in Behavior
Modification. New York: Holt, Rinehart and Sinston,
1965, 146-148.

 

Dorfman, E. Play therapy. In C. R. Rogers (Ed.) Client—

Centered Therapy: Boston: Houghton, Mifflin, 1951,
235-277.

79

80

Ducat, C. An examination of nonverbal behaviors of trained
and untrained college undergraduates acting as play
therapists. Unpublished M. A. thesis, Michigan State
University, 1973.

Fellows, L. and Wolpin, M. High School psychology trainees
in a mental hospital. In B. J. Guerney, Jr. (Ed.)
Psychotherapeutis_Agents: New Roles for Nonprofessionals,
Parents and Teachers. New York: Holt, Rinehart and
Winston, 1969, 274-277.

Finke, H. Changes in the expression of.emotionalized atti-
tudes in six cases of play therapy. Unpublished Master's
thesis, University of Chicago, 1947. Cited in Lebo, D.
The present status of research in nondirective play
therapy. In M. R. Haworth, (Ed.) Child Psychotherspy
New York: Basic Books Inc., 1964, 421-430.

Ginott, H. G. Research in play therapy. In M. R. Haworth
(Ed.) Child Psychotherapy. New York: Basic Books, Inc.,
1964, 431-435.

 

Ginott, H. G., and Lebo, D. Play therapy limits and theo-
retical orientation. Journal of Consulting Psychology
1961, 26, 337-340.

Goodman, G. An experiment with companionship therapy:
College students and troubled boys - assumptions,
selection, and design. American Journal oquublic
Health, 1967, 57, 1772-1777.

Gordon, J. E. Project cause, the federal anti-poverty pro-
gram and some implications for subprofessional training.
American Psyshslggist, 1965, 20, 334-343.

 

Gruver, G. G. College students as therapeutic agents.
Psychological leletin, 1971, 76, 111-127.

Guerney, B. J., Jr. Filial therapy: Description and
rationale. Journal of Consulting Psychology, 1964, 27,
304-310.

Guerney, B. J., Jr. (Ed.) Psychotherapeutic Agents: New Roles
for Nonprsfessipnals, Parents, and Teachers. New York:
Holt, Rinehart and Winston, 1969, 382-391.

Guerney, B. J., Jr. Guerney, L. F., and Andronico, M. P.
Filial therapy. Yale Scientific Magazine, 1966, 40, 6ff.

Guerney, B. J., Jr. and Stover, L. Filial Therspy. Final
report to NIMH on MH 18264, 1971.

 

81

Harris, F. R., Wolf, M. R., and Baer, D. M. Effects of

adult social reinforcement on child behavior. Young
Children, 1964, 20, 8-17.

Harvey, L. V. The use of non-professional auxiliary
counselors in staffing a counseling service, Journal
of Counseling Psychology, 1964, 11, 348-351.

 

Hawkins, R. P., Peterson, R. F., Schweid, E., and Bijou, S. W.
Behavior Therapy in the home: amelioration of problem
parent-child relations with the parent in a therapeutic
role. Journal of_Experimental Child PsycholQSY: 1966,

4, 99-107.

 

Hobbs, N. Mental health's third revolution. American
Journal of Orthspsychiatry, 1964, 34, 822-833.

 

Joint Commission on Mental Health of Children. Crisis in
Child Mental Health: Challenge for the 1970's. New
York: Harper and Row, 1970.

 

 

Kreitzer, S. F. College students in a behavior therapy pro-
gram with hospitalized emotionally disturbed children.
In B. G. Guerney, Jr. (Ed.) Psychotherapeutic Agents:
New Roles for Nonprofessionalsl Parents, and Teachers.
New York: Holt, Rinehart and Winston, 1969, 226-230.

 

Landisberg, S. and Snyder, W. Nondirective play therapy.
Journal of Clinical Psycholong 1946, 2, 203-213.

 

Lebo, D. The present status of research in nondirective
play therapy. In M. R. Haworth (Ed.) Child Psychotherapy.
New York: Basic Books Inc., 1964, 421-430.

Lebo, D. The relationship of response categories in play
therapy to chronological age. Child Psychiatry,
1952, 2, 330-336.

Levitt, E. E. Research on psychotherapy with children. In
A. E. Bergin and S. L. Garfield (Eds.) Handbook of
Psychotherapy ana_Behavior Change. New York: John
Wiley, 1971, 474-494.

 

Linden, J. I. and Stollak, G. E. The training of under-
graduates in play techniques. Journal of Clinical
Psychology, 1969, 25, 213-218.

 

 

Matarazzo, R. G. Research on the teaching and learning of
psychotherapeutic skills. In A. E. Bergin and S. L.
Garfield (Eds.) Handbook of Psychotherapy and Behavior
Change. New York: John Wiley, 1971, 895-924.

82

McWilliams, S. A. and Finkel, N. J. High School Students
as mental health aides in the elementary school setting.
Journal of Consulting and Clinical Psychology, 1973,
40,— 39-427" ‘ ‘

 

Moustakas, C. E. Children iasPlay Therapy. New York:
McGraw-Hill, 1953.

 

Moustakas, C. E. Emotional adjustment and the play therapy
process. Journalsof Genetic Psychology, 1955, 86,
79-99.

Moustakas, C. E. The frequency and intensity of negative
attitudes expressed in play therapy: a comparison of
well-adjusted and disturbed young children. Journal of
gsnetic Psychology, 1955b, 86, 309-325.

 

Moustakas, C. E. Psychotherapy with Children, New York:
Ballantine Books, 1959.

 

Moustakas, C. E. and Schlalock, H. D. An analysis of
therapist-child interaction in play therapy. Child
Development, 1955, 26, 143-157.

 

Perlmutter, F. and Durham, D. Using teen-agers to supplement
casework service. Social Work, 1965, 10, 41-46.

 

Poser, E. The effect of therapists' training on group thera-
peutic outcome. Journal of Consulting Psychology,
1966, 30, 283-289.

Reif, T. and Stollak, G. E. Sensitivity to Children: Train-
ingyand itssEffects. East Lansing: MiEhigan State
University Press, 1972.

 

Reinherz, H. The therapeutic use of student volunteers.
Children, 1964 2, 137-142.

Riessman, F. The 'helper' therapy principle. Social Work,
1965, 10, 27-32.

 

Rioch, M. J. Changing concepts in the training of therapists,
JournalsgfgConsulting Psyshology, 1966, 30, 290-292.

Rioch, M. J., Elkes, C., Flint, A. A., Usdansky, B. S.
Newman, R. G., and Silber, E. National Institute of
Mental Health pilot study in training mental health
counselors. American Journal of Orthopsychiatry, 1963,
33, 678-689.

83

Rutledge, T. M. Sequential analyses of adult—child play
encounters: the effects of training and personal
characteristics. Unpublished Ph.D. dissertation,
Michigan State University, 1974.

Schreiber, J. R. The effects of training on undergraduate
behavior in play interaction with clinic-referred
children. Unpublished M. A. thesis, Michigan State
University, 1972.

Shah, S. A. Training and utilizing a mother as the therapist
for her child. In B. G. Guerney, Jr. (Ed.)
Psychotherapeutic Agents: New Roles for Nonprofessionals,
Parents, and Teachers. New York: Holt, Rinehart and
Winston, 1969, 401-407.

Smith, M. B. and Hobbs, N. The community and the community
mental health center. Amssican Psycholpgist, 1966,
21, 499-509.

Stollak, G. E. The experimental effects of training college
students as play therapists. Psychotherapy: Theosy,
Research‘and Psactice, 1968 5, 77-80.

 

 

Stollak, G. E. An integrated graduate-undergraduate program
in the assessment, treatment, and prevention of child
psychopathology. Profsssional Psychology, 1973a
4, 158-159.

Stollak, G. E. Undergraduates as play therapists: The
effects of training and personal characteristics. Paper
presented at the third annual meeting of the Society for
Psychotherapy Research, June 16, 1972, Nashville,
Tennessee.

Stollak, G. E. What Happened Today. Iowa: Kendall/Hunt
Publishing Co., 1973b.

 

Stollak, G. E., Green, L., Scholom, A., Schreiber, J., and
Messe', L. A. The process and outcome of play encounters
between undergraduates and clinic-referred children:
Preliminary findings. Paper presented at the 1973 meet-
ing of the Society for Psychotherapy Research, Philadelphia,
Pennsylvania.

Stover, L. and Guerney, B. G. The efficacy of training pro-
cedures for mother in filial therapy. Psychotherapy:
Theory, Research aad Pragtice, 1967, 4, 110-115.

 

 

84

Zimmerman, E. H. and Zimmerman, J. The alteration of behavior
in a special classroom situation. Journal_of Experimental

 

Analysis of Behavior, 1962, 5, 59-60.

 

APPENDIX A
Definitions of the Rating Scales

for Scoring Children's Behavior

APPENDIX A

Child Behavior Ratiag Scale

 

LEADERSHIP

 

Verbal statements directed specifically to the adult
describing intention of self action or expecting action from
the adult (all within the play session). Nonverbal commands
which clearly indicate expectation of action from the adult
in the session are scored Leadership. EXCEPTIONS: (1) State-
ments made in the context of fantasy or role-play are not
scored Leadership. (2) Positive statements which are re-
quests for help or permission are scored D.

 

COMMANDSgDIREQTED T9 ADULT (EXPECTING SOME BEHAVIOR
CHANGE OR sCTION_FROM ADULT)

 

 

Score

Wait a minute. L
Give me a piece of paper. L
Look what I did. L
Give me the small part. L
Don't take it out. (plus Ag if critical) L plus Ag
Don't look. L
Get away from me. L plus Ag
You stop that! (to toy) not scored
Let me try. (This is asking permission, even if

stated positively) D

(Placing phone in front of adult with expectation

that she will pick up receiver) L
Pointing to piece of toy, expecting adult to hand

it over. L
I think you'd better help me. (Help-seeking even

if stated positively) D

PRQCLAMATIONS OF INTENDED SELF ACTIVITY

 

Now I'll shoot him.

I'm going to make a house.

I'll make a pie.

I have to do it myself.

We'd better get going.

I'm going to paint this ugly thing.

I'd better wash up.

I'll do whatever I want to do

I want to build a house. (if followed by action)
I'll get it up next time. (not followed by action) not scored

plus Ag

plus Ag

plus Ag

bt'bbbfibt'b

85

86

Appendix A (Cont'd)

 

DIRECTIONS OR GIVING IDEAS ABOUT ACTIVITY TO ADULT

 

Let's put him to bed

Not like that!

Now, you go to bed. (to toy)

Don't do it that way.

You should really be doing this.

Make believe this is a stove.

This is the father, this is the mother, he's the
boy.

This is tea.

Drink your tea, mother. (fantasy)

You're supposed to say "ouch."

You can have this one.

No, put him here.

You be the mother. (assigning doll or puppet role)

We have to make it higher.

He's taking a bath. (fantasy)

VOLUNTEERING INFORMATION IS NOT SCORED L.

 

I missed it.

See, that's how it works.
This is the front of it.

That's not a nice house.

This one is the fastest.

This is a good puppet.

ASSERTIVE BEHAVIOR NOT SCORED LEADERSHIP UNLESS

 

THERE IS AN EXPECTANCY OF AgTION FR0M_ssLF ORsADULT

 

L
L plus Ag
not scored
L plus Ag
L plus Ag
L

L

L

not scored
L

L plus Af

L plus Ag

L

L

not scored

not scored
not scored
not scored
A9
not scored
Af

I'm going to see if I can make it stick. (no action)not scored

You're not the best!

I'll give you one of these.
I don't care what you draw.
Maybe I'll shoot the Bobo. (if action follows)
You can have it.

I can't make it.

A big boy does it like this,
I have it.

I just broke the Bobo.

We need one.

No, this is thinner.

I know what we'll do.

I have it.

We need another couch.
Aren't they cute?

I think so.

I'll bet it won't stick.

not scored
not scored
not scored
not scored
not scored
not scored
not scored
not scored

not scored
not scored

87
Appendix A (Cont'd)
I want a drink of water. (stopping other action,

expecting adult to get it). L
I know what I'll do. not scored

QUESTIONAsLE COMBINATIONS

 

I'm not making the feet. I don't know how. L plus Ag
Look at this. Do you like it? (command plus
question) L plus D

Let's read some stories. All right? (if child
goes on with activity, score L. If child stops
to wait for OK score D).
How did you make yours? Make it for me. D plus L
(D for the question. L for command).
I'm doing it backwards. Do you want me to do it
over? (first sentence is evaluative statement,
not a proclamation of intended activity. Not
scored for first sentence. D for the second)

I don't want to do this anymore. (refusal not scored
if in response to adult's suggestion. If spontaneous,
score L plus Ag).

 

 

 

 

I don't have to watch. (if spontaneous) L plus Ag
You have the wrong number. (to adult on the phone) not scored
Blow it up, will you? (expectation of action,

don't score D). L

88

DEPENDENCY

 

Statements made to the adult which are in the form of a
request. This includes looking for response from the adult
before doing something (to see if it will be allowed) even
if adult doesn't respond. This kind of "looking for response"
doesn't have to be verbal though it usually is. Child may
be busily playing but covertly or overtly watching adult to
see her reaction, trying to get approval of something (this
would have to be shown by more than occasional glances).
Child must clearly be expecting response from adult.
EXCEPTIONS: (1) Questions or dependency remarks made in the
context of fantasy or role-play are not scored D. (2) Rheto—
rical questions in which the child goes right on with his in-
tended action are not scored. (3) Asking opinion in a con-
versational way, as well as any questions referring to out of
session activity, are not scored. (4) Clearly considerate or
polite statements phrased as questions are not scored.

(5) Seeking affection not scored D. (6) Approval of finished
activity, art production, etc., not scored D.

A§RING FOR HELP FOR GQIDANCE FOR ACTION (with expectation of
specific reply from adult, i.e., permission, reassurance,
approval for some intended action in the play session).

 

I wish you would do this. D
I think you'd better help me. D
You have to help! D
The air won't come out, will it? D
Can I shoot the window? D
You do it. (a command) L
Want to play with me? (seeking companionship or
affection, not help) 0 or Af
I want a drink of water. (expecting adult to get
it) L
Look at mine. (a command) L
Tell me, do you like this? (asking approval of
finished work is not D) L for the
command
Know how I do mine? not scored
I can't make mine so good (self critical but not
asking for comment or action from adult) Ag
I don't know if I can make this. Ag
I can't do it. (handing it to adult to do) D Ag
I'm not making the right kind head. Ag
Do you like me? (seeking affection) Af
‘Where do you want to sit? (politeness) not scored
You sit here. Okay? (if action stopped, waiting
for ok) D

(fully expecting adult to sit) L

89

DssENDENCY 2

 

Is this Mrs. Goober? (on the phone, role-play)

I'll take a cup of coffee, can I? (expecting
adult to pour it in role-play)

This is a good time for the Munsters to be on,
isn't it? (refers to out of session)

You know what I'm making?...it's a house. (rheto-
rical, not waiting for an answer)

Do you like these toys? (asking opinion in con-
versational way)

ASKING FOR EXPLICIT IDEAS, INSTRUCTION OR
DIRECTION IN THE SESSION

Can you do this? (this may also be asking for
help)

Where shall I put this?

Now what do we do?

What's that? (pointing to microphone or pipe in
wall)

What time is it?

Do you want me to make a rocket?

Who won?

What's behind that?

Why are we here? (questions must be related to
activity in the play session)

Can we come back again?

You know the big raft we take to the beach?

Can they see us?

Who was that lady?

What color shall I make the hat? (response ex-
pected before going ahead with the activity)

Are you going to be the mother? (explicit
instruction)

Do you like doing that? (opinion, not related to
help of or direction of child's activity)

not

not

not

not

not

UUU

UUUUU

not
not

not

not

scored
scored
scored
scored

scored

scored
scored
scored

scored

scored

90

AFFECTION

 

Affection is seen in gestures and language, both with
respect to the adult and toys.

0

1

no affection

mild affection - smiling at adult, inviting adult
to participate, sitting near adult, including adult
in activity (e.g. both should draw, sip tea, make
clay figures - on child's invitation). Evidence

of warmth - not just animated talking. Analogous
behavior with dolls (e.g. smiling at doll, etc.)

stronger, more demonstrable affection - leaning on
adult, standing next to, thSpering to adult, patting
him or her. Analogous affectionate behavior with
toys, acted out between dolls.

most demonstrable affection - climbing on adult,
hugging, kissing. Full expression and involvement -

"I love you." Analogous behavior with dolls or
Bobo.

91

AGGRESSION

 

The play of the child is observed for hostile, destruc-
tive, or aggressive behavior, either toward the adult or
toward a toy. The aggressive behavior might occur in gesture,
facial expression, physical action, or words. Included here
are hitting, spitting, fighting, shooting, hurting, pinching,
drowning, burying, throwing, with destruction in mind.

Verbal aggression is expressed when the child shows an-
noyance, anger or dislike toward an object or person. In-
cluded here are negative statements such as disparaging re-
marks or criticism, cursing, threat of injury, or accusa-
tions (attack by gesture or words).

The ratings denote the intensity of aggression. The
score for the interval represents the most extreme form of
the child's aggression, not the average.

0 - no aggression

l - mild, perhaps playful aggression - includes any play-
ful activity such as inspecting or handling a knife
or gun, or cutting clay. Mild verbal aggression -
"He's mean." Verbal criticism toward self, or
toward own activity. Shooting against the wall with
the dart gun has arbitrarily been established as
Ag 1.

3 - intense aggression in which the child is highly in-
volved, or a very hostile act, e.g. stabbing, shoot-
ing, slicing, with force. Swearing profusely at the
adult. An act might be intensely aggressive because
its aim is highly destructive or hostile, or because
it is highly forceful and shows much self-involvement.

CONTACT

Sharing an activity with the adult, e.g. card playing,
tinker toys, joint effort in clay, dolls or puppets, target
practice, even if adult initiates the joint activity. Con-
versation alone is not scored contact. Adult must be physi-
cally participating on the child's level. This is shared
activity. When the child invites the adult to play, the in-
terval at which this occurs is scored for affection (l) as
well as contact.

92

ROLE - PLAYING

 

Scored if the child assumes by voice or action another
identity: person or animal. Action must be quite overt.
Simply picking up a puppet, for example, is not role -
playing.

APPENDIX B
SAMPLE SCORING SHEET FOR CODING

CHILDREN'S BEHAVIOR

Totals

anagrc

 

 

H

 

93

Child's Name
Date
Therapist
Coder Initials

Af

 

 

 

 

 

APPENDIX C

MEAN PERCENTAGE OF AGREEMENT FOR

RATERS WITH EXPERT

Table 4. Mean Percentage of Agreement for Raters with Expert.

 

 

 

Raters:
Variables: M K __ B ‘_ A _sY"s
Dependency 94.0 97.5 82.3 81.8 88.98
Leadership 84.5 79.5 91.6 86.8 85.60
Aggression 83.0 93.1 67.7 89.0 83.20
Affection 82.4 72.0 66.0 84.2 76.15
Contact 100.0 88.4 73.6 100.0 90.50
Role-playing 87.5 78.0 88.2 95.7 87.35
X's 88.5 84.75 78.23 89.58 85.28

94

APPENDIX D

INTER-HATER RELIABILITY COEFFICIENTS

Table 5. Inter-Rater Reliability Coefficients

 

_ _—*__.‘ _ 1“

 

 

 

Coder Pairs Variables:

and _

.12 1 AS DE 9 E 1':

Sessign_Numpsrs:
M + K, #1 .79 .84 .98 .70 .86 .93 .85
M + K, #6 I .42 .86 .95 .89 .96 .99 .85
M + K, #11 .95 .91 .98 .99 .95 .96 .96
M + K, #15 .89 .85 .95 .96 .85 .97 .91
M + B, #1 .99 .96 .99 .36 .99 1.0 .88
M + B, #6 .71 .73 .92 .99 .99 .86 .87
M + A, #1 .93 .68 .95 .57 - - .78
M + A, #6 .96 .77 .99 .99 .82 .99 .92
M + A, #11 .86 .82 .92 .76 .94 .65 .83
M + A, #15 .86 .91 .97 .93 .97 .89 .92
X's .84 .83 .96 .81 .93 .92 .88

 

(-) = N too small to obtain reliabilities. Means N of less
than 5 per rater.

95

APPENDIX E

MEAN FREQUENCY AND MEAN INTENSITY SCORES

FOR AFFECTION AND AGGRESSION

 

Table 6. Mean Frequency and Mean Intensity Scores for
Affection and Aggression for Children Encounter-
ing Trained High Potential (THP), Trained Low
Potential (TLP), Untrained High Potential (UHP),

and Untrained Low Potential (ULP) Undergraduates.

__‘ _‘

Session Number

 

 

 

 

 

1 1 _l__ 6_ 11 _‘ 15
VARIABLE: AFFECTION
THP Frequency 7.00 1.83 3.70 3.38
Intensity 7.30 1.83 3.70 3.38
TLP Frequency 7.28 6.81 17.00 10.50
Intensity 7.39 7.56 17.67 10.86
UHP Frequency 5.93 7.38 9.00 3.25
Intensity 5.93 7.38 9.13 3.25
ULP Frequency 8.13 14.25 15.50 12.75
Intensity 8.13 14.25 15.50 14.75
VARIABLE: AGGRESSION
THP Frequency 14.40 28.00 48.00 43.38
Intensity 15.70 29.83 48.60 49.25
TLP Frequency 16.44 37.88 48.50 31.64
Intensity 16.83 39.50 51.33 32.50
UHP Frequency 24.29 27.13 35.63 68.50
Intensity 30.71 31.50 38.25 75.00
ULP Frequency 59.50 42.00 22.50 40.50
Intensity 61.13 42.00 24.83 42.50

96

APPENDIX F

MEAN INTENSITY PER SESSION

OF AFFECTION AND AGGRESSION

Table 7. Mean Intensity per Session of Affection and
Aggression for Children Encountering Trained
High Potential (THP), Trained Low Potential
(TLP), Untrained High Potential (UHP) and
Untrained Low Potential (ULP) Undergraduates.

Sessioa_Number

 

 

 

 

 

1 6 11 15

AFFECTION

THP 1.04 1.00 1.00 1.00
TLP 1.02 1.11 1.04 1.03
UHP 1.00 1.00 1.03 1.00
ULP 1.00 1.00 1.00 1.16
AGGRESSION

THP 1.09 1.07 1.01 1.14
TLP 1.02 1.04 1.06 1.03
UHP 1.264 1.16 1.07 1.09
ULP 1.03 1.00 1.10 1.05

 

97

 

APPENDIX G
MEAN SCORES OF THE DEPENDENT VARIABLES
FOR THE TRAINED HIGH POTENTIAL (THP),
TRAINED LOW POTENTIAL (TLP), UNTRAINED
HIGH POTENTIAL (UHP),AND UNTRAINED LOW

POTENTIAL (ULP) GROUPS

Mean Scores of the Dependent Variables for the
Trained High Potential (THP), Trained Low Potential
(TLP), Untrained High Potential (UHP), and Untrained
Low Potential (ULP) Groups.

Table 8.

 

VARIABLE: LEADERSHIP

Session Number

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Condition l1_ ‘_ ll 15-__
THP 18.00 (N=5) 19.50 (N=3) 20.00 (N=5) 11.13 (N=4)
TLP 11.06 (N=9) 18.94 (N=8) 16.83 (N=3) 24.43 (N=7)
UHP 7.86 (N=7) 9.75 (N=4) 21.38 (N=4) 12.75 (N=2)
ULP 13.75 (N=4) 11.00 (N=2) 18.17 (N=3) 46.50 (N=2)

VARIABLE: DEPENDENCY
Session Number

Condition 1 6 IL. ._ 15
THP 6.80 (N=5) 11.50 (N=3) 6.60 (N=5) 3.00 (N=4)
TLP 5.33 (N=9) 7.38 (N=8) 8.50 (N=3) 9.57 (N=7)
UHP 5.36 (N=7) 5.38 (N=4) 13.75 (N-4) 8.50 (N=2)
ULP 10.63 (N=4) 7.00 (N=2) 16.33 (N=3) 8.25 (N=2)

VARIABLE: AFFECTION
Session Number

Conditioa l 6 ll 15
THP 7.00 (N=5) 1.83 (N=3) 3.70 (N=5) 3.38 (N=4)
TLP 7.28 (N=9) 6.81 (N=8) 17.00 (N=3) 10.50 (N=7)
UHP 5.93 (N=7) 7.38 (N=4) 9.00 (N=4) 3.25 (N=2)
ULP 8.13 (N=4) 14.25 (N=2) 15.50 (N=3) 12.75 (N=2)

VARIABLE: AGGRESSION
Session Number

Condition 1 6 ‘_‘_1l 4_ 15
THP 14.40 (N=5) 28.00 (N=3) 48.00 (N=5) 43.38 (N=4)
TLP 16.44 (N=9) 37.88 (N=8) 48.50 (N=3) 31.64 (N=7)
UHP 24.29 (N=7) 27.13 (N=4) 35.63 (N=4) 68.50 (N=2)
ULP 59.50 (N=4) 42.00 (N=2) 22.50 (N=3) 40.50 (N=2)

#44 A—_ _-_

98

99

 

 

 

 

 

 

 

 

Table 8. (Cont'd)
VARIABLE: CONTACT
Session Number
Condition 1 6 ‘11 15
THPT 12.30 (N=5 19.00 (N=3) 17.10 (N=5) 25.88 (N=4)
TLPT 10.44 (N=9) 14.88 (N=8) 33.00 (N=3) 29.86 (N=7)
UHPT 8.29 (N=7) 13.13 (N=4) 10.13 (N=4) 40.50 (N=2)
ULPT 10.88 (N=4) 45.75 (N=2) 40.83 (N=3) 45.00 (N=2)
VARIABLE: ROLE-PLAYING
Session Number
Condition 1 6 11 _*_ 15
THPT 4.40 (N=5) 11.67 (N=3) 14.30 (N=5) 7.88 (N=4)
TLPT 7.44 (N=9) 6.50 (N=8) 21.17 (N=3) 5.43 (N=7)
UHPT 2.36 (N=7) 11.00 (N=4) 6.63 (N=4) 42.50 (N=2)
ULPT 0.50 (N=4) 0.00 (N=2) 1.17 (N=3) 17.00 (N=2)

APPENDIX H

ANALYSIS OF VARIANCE

SUMMARIES FOR THE

DEPENDENT VARIABLES

 

100

Table 9. Analysis of Variance Summaries for Leadership
in Sessions 1, 6, and 15.

 

 

 

 

 

Source df SS MS F
Training 1 78.84 78.84 0.61
Potential 1 1.59 1.59 0.01
SESSION
1 Training X Potential 1 233-68 233.68 1.80
Error 21 2725.33 129.78
Training 1 258.94 258.94 2.41
Potential 1 0.36 0.36 0.00
SESSION
6 Training X Potential 1 2.71 2.71 0.03
Error 13 1395.47 107.34
Training 1 402.66 402.66 2.84
Potential 1 1588.32 1588.32 11.21
(p< .01)
SESSION
15 Training X Potential 1 300.06 300.06 2.12

Error 11 1588.52 141.68

 

101

Table 10. Analysis of Variance Summaries for Dependency
in Sessions 11 and 15.

 

 

 

df SS MS F
Training 1 200.84 200.84 6.70
<p<.os)
Potential 1 17.97 17.97 0.60
SESSION
11 Training X Potential 1 0.39 0.39 0.01
Error 11 329.62 29.97
Training 1 12.51 12.51 0.64
Potential 1 28.64 28.64 1.46
SESSION
15 Training X Potential 1 33.38 33.38 1.70

Error 11 215.83 19.62

 

 

102

 

 

 

 

 

Table 11. Analysis of Variance Summaries for Affection in
Sessions 6, 11, and 15.
df SS MS F
Training 1 139.65 139.65 4.09
(24.10)
Potential 1 116.18 116.18 3.40
(34.10)
SESSION

6 Training X Potential 1 2.95 2.95 0.09
Error 13 443.95 34.15
Training 1 12.92 12.92 0.54
Potential 1 350.88 350.88 14.80

(p<.005)
SESSION "

11 Training X Potential 1 41.38 41.38 1.75
Error 11 260.80 23.71
Training 1 3.24 3.24 0.14
Potential 1 198.20 198.20 8.39

(p<.025)
SESSION ‘

15 Training X Potential 1 4.02 4.02 0.17

Error 11 259.93 23.63

 

 

 

 

 

 

Table 12. Analysis of Variance Summaries for Aggression
in Sessions 1, 6, 11, and 15.
(if SS MS F
Training 1 3981.28 3981.28 4.39
(p_< .05)
Potential 1 1970.34 1970.34 2.17
SESSION
1 Training X Potential 1 3834.28 3824.28 4.23
(B < . 10)
Error 21 19034.90 906.42
Training 1 8.74 8.74 0.01
Potential 1 153.14 153.14 0.17
SESSION
6 Training X Potential 1 20.59 20.59 0.02
Error 13 12003.06 923.31
Training 1 1317.69 1317.69 0.99
Potential 1 142.77 142.77 0.11
SESSION
11 Training X Potential 1 166.26 166.26 0.12
Error 11 14634.19 1330.38
Training 1 828.45 828.45 0.91
Potential 1 1133.13 1133.13 1.25
SESSION
15 Training X Potential 1 189.68 189.68 0.21
Error 11 9970.05 906.37

103

 

104

 

 

 

 

 

Error 11

14519.05 1319.91

Table 13: Analysis of Variance Summaries for Contact in
Sessions 6, 11, and 15.
df SS MS F
Training 1 517.55 517.55 0.84
Potential 1 672.13 672.13 1.09
SESSION
6 Training X Potential 1 1116.99 1116.99 1.81
Error 13 8025.68 617.36
Training 1 0.68 0.68 0.00
Potential 1 1943.55 1943.55 2.50
SESSION
11 Training X Potential 1 196.04 196.04 0.25
Error 11 8539.06 776.28
Training 1 635.45 635.45 0.48
Potential 1 51.57 51.57 0.04
SESSION
15 Training X Potential 1 0.20 0.20 0.04

 

 

105

 

 

 

 

 

 

Table 14. Analysis of Variance Summaries for Role-Playing in
Sessions 1, 6, 11, and 15.
df SS MS F
Training 1 114.57 114.57 0.68
Potential 1 1.99 1.99 0.01
SESSION
1 Training X Potential 1 34.02 34.02 0.20
Error 21 3554.78 169.28
Training 1 42.57 42.57 0.34
Potential 1 216.37 216.37 1.71
SESSION
6 Training X Potential 1 28.10 28.10 0.22
Error 13 1647.67 126.74
Training 1 685.25 685.25 2.37
Potential 1 1.79 1.79 0.01
SESSION
11 Training X Potential 1 136.08 136.08 0.47
Error 11 3182.83 289.35
Training 1 1530.80 1530.80 4.19
(p<110)
Potential 1 560.48 560.48 1.53
SESSION
15 Training X Potential 1 381.19 381.19 1.04
Error 11 4016.90 365.17

 

APPENDIX I

SIMPLE EFFECTS ANALYSIS

 

 

 

 

Table 15. Simple Effects Analysis Exploring the Significant
Training X Potential (AB) Interaction for Aggres-
sion in Session 1.
df SS SS F
AB A within Bl 1 277.75 277.75 0.31
A within B2 1 5265.81 5265.81 5.81
(p < .05)
BA B within A1 1 11.76 11.76 0.01
B within A2 1 3520.86 3520.86 3.88
(P < .10)
Error 21 19034.90 906.42

106

MICHIGAN STAT UNIVERSITY LI

mu III nnnm

085 681

ARIES

3 1293 0

III
3