n
. ,.. “9.55.9!

 

fHESl‘S

 

r" ~ willmm:n:1W:an

 

 

 

This is to certify that the

thesis entitled
The Assessment of Dynamic Relational Patterns

Within the Family System

presented by

Jerry Stuart Adams

has been accepted towards fulfillment
of the requirements for

Master of Arts degreein Psychology

Metqw

M‘Jor professorI

Mesa“ . A 3%2‘

0-7639

 

MSU

LIBRARIES
”In.

 

RETURNING MATERIALS:
Place in book drop to
remove this checkout from
your record. FINES will

 

 

be charged if book is
returned after the date
stamped below.

 

 

 

I
t
l
iMA
107 .A103
37' 23:?

 

(Rwy I

 

THE ASSESSMENT OF DYNAMIC RELATIONAL
PATTERNS WITHIN THE FAMILY SYSTEM

By
Jerry Stuart Adams

4 // 7(JIJ’}?

A THESIS

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

MASTER OF ARTS

Department of Psychology

1982

ABSTRACT

THE ASSESSMENT OF DYNAMIC RELATIONAL
PATTERNS WITHIN THE FAMILY SYSTEM

By
Jerry Stuart Adams

The family forms a social system in which each individual's
inner psychological world is interdependently tied to the psycholo-
gical world of the other family members through the mechanism of
projective identification. In order to examine this interdependence,
twenty family triads (father, mother and adolescent son) were asked
to create stories to six TAT cards, both individually and together
as a family unit. The school records of ten of the adolescents con-
tained three or more instances of school behavior problems; the ten
controls had zero or one. It was assumed that in those families
where the adolescent had a history of behavioral problems at school,
the son was expressing the denied or split-off aggressive impulses
of the parents. It was hypothesized that the fantasy stories of
these families would differ in aggressive content from those of the
‘control families. There were few significant findings, but the pat-
terns evident in the data do generally support the theoretical model.
Intra-familial differences in aggressive content suggest that the
fathers of behavior problem adolescents are conflicted about their
own aggressive tendencies and both the son and the mother appear to

be intrapsychically involved in the father's conflict.

DEDICATION

To Iva Zoe

1'1

ACKNOWLEDGEMENTS

I would like to thank Dr. Lucy Ferguson, chairperson, Dr.
Bert Karon and Dr. Martha Karson for serving as members of my thesis
committee.

A very special thanks to Mr. Jim Brady for his endless
assistance and support, without which this study would have never
been possible. I would also like to express my gratitude to Dr.
Bert Karon for his overall support and "being there", especially
over the past summer. I would like to acknowledge the cooperation
of the families who participated in this study. An important thank
you to Jeff Blank and Yvonne Haddad, for rating the data quickly
and efficiently.

I would like to express my special appreciation to Dr. Lucy
Ferguson for her assistance and conscientious reading of the many
drafts and her aid and insights which helped me to better understand
and comprehend my results.

Finally, it is difficult to find the words which can fully
express my deepest of thanks and gratitude to Ms. Iva Katz Adams,
for her support, encouragement and eagerness to help in every possible
way, A special thanks for her willingness to share in the emotional
highs and lows which seem to be an inherent part of research. I am

forever grateful.

iii

TABLE OF CONTENTS

LIST OF TABLES.

CHAPTER I - INTRODUCTION

Relevant Clinical and Theoretical Background.
Superego Lacunae . . . . .
Scapegoating .

Object Relations Theory.

Delineations .
Irrational Role Assignment.
Projective Identification and Collusion

Summary.

Empirical Background. . .

In Support of the Theoretical Review
Interaction Testing and the Unrevealed
Differences Technique

The TAT and Family Research

Summary and Hypotheses . .

Hypotheses .
Hypothesis
Hypothesis
Hypothesis
Hypothesis
Hypothesis
Hypothesis

O‘U‘lth-‘O

CHAPTER 2 - METHOD

Participants .
Experimental Group
Control Group. .
Participant Selection .
Matching and Demographic Variables .
Assessment Procedure. . .
Individual Administration .
Conjoint Administration . .
TAT Card Selection and Rationale.

iv

Page

vi

Scoring System. .
Weighted Hostility
Overt- Covert Scale

Raters
Training.

Determination of Inter-rater Reliability .

CHAPTER 3 - RESULTS .

General Considerations .
GPA . . .
Number of Words in the TAT Stories .
Organization . .
Comparisons Between Adolescent Sons.
Hypothesis 1
Hypothesis 2
2-A . .

2-3
2- C . .
Hypothesis 3
Comparisons Between the Families.
Hypothesis 4
Hypothesis 5 .
Comparisons Between Parents
Hypothesis 6

CHAPTER 4 - DISCUSSION .

APPENDICES - APPENDIX A
APPENDIX B
APPENDIX C
APPENDIX 0

LIST OF REFERENCES

96

100
102

109

Table

LIST OF TABLES

Types of Offenses Reported in the School Records
of the Adolescents With Behavioral Problems.

Participant Variables .

Comparisons Between Adolescent Sons' and Average
Parents' Scores .

Test of the Difference Between Two Variances for
Matched Groups. Summary of the Deviation-score X
Group ANOVA . . . . . .
Comparisons Based on Family Stories

Comparisons Between Parents .

The Number of Family Pairs Needed to Have Signifi-
cant Results Given the Obtained Effect Size.

Percentage of Overt Themes for the 60 Participants

vi

Page

52
55

68

69

74
77

80
90

CHAPTER I
INTRODUCTION

This study will focus on the development of an assessment
procedure appropriate for the evaluation of family dynamics. There
is a growing body of primarily theoretical literature, clinical in
origin, that emphasizes the interdependence between an individual's
intrapsychic life and his/her family of origin. Individuals are
perceived as functioning in a family system, and relational con-
structs are employed to understand both conscious and unconscious
processes or dynamics. What was once the province of a strictly
individual, intrapsychic approach, i.e., unconscious wishes, needs,
affects and fantasies, is now under scrutiny from a family-system
perspective. Thus, there is an attempt to combine the intrapsychic
and interpersonal dimensions of human behavior. The assessment
procedure developed in the present study, which will be referred to
as the Thematic Apperception Family Interaction Test, (TAFIT), is
an attempt to offer empirical support towards this integration of

individual-intrapsychic and interpersonal-family dynamics.

Relevant Clinical and Theoretical Background
It is essential to provide a theoretical framework which
emphasizes and describes the manner in which interactions within

the family system are major determinants of individual behavior and

l

psychological functioning, focusing on what Framo (1972) refers to
as the "family's personality-forming influence". This review will
concentrate on those clinicians and theoreticians who stress the
interdependence between the intrapsychic and interpersonal dimensions
of behavior.

There are a number of basically similar theoretical concepts,
e.g., "superego lacunae" (Johnson and Szurek, 1952), "scapegoating"
(Vogel and Bell, l968), "delineations" (Zinner and Shapiro, l972),
"irrational role assignment" (Framo, 1972, 1976) and ”bound-up dele-
gate" (Stierlin, 1976), which when grouped together form a substan-

tial body of literature that presents the position that individual

behavior and intrapsychic functioning can be understood from a rela
tional-interpersonal perspective, and furthermore, that relational-
family dynamics do have a significant effect on the individual's
conscious and unconscious behavior. The underlying similarity of
these various concepts is that they imply that a given person's
implicit and explicit perceptions and interactions with another may
become part of that other's intrapsychic dynamics, e.g., self-image
and/or motivational system. In many respects Framo (l972, p. 271)
exemplifies these various concepts when he states that individual
family members "reciprocally carry part of each others' psychology
and form a feedback system which in turn regulates and patterns
their individual behavior." It is also important to point out that
these various concepts derive, both directly and indirectly, from
object relations theory which has proven extremely useful in clinical

work with families (e.g., Stewart et al., 1975).

Johnson and Szurek's (1952) work on the etiology of "anti-
social acting out", i.e., delinquent behavior, is of great historical
significance and in fact has been referred to as a "classical contri-
bution" (Stierlin et al., 1971). Even though Johnson and Szurek's
observations, dating back to 1939, originate from a psychoanalytic
intrapsychic orientation, their work is frequently referred to in
the literature, some 35 years later, by those clinicians and theore-
ticians working from a family perspective. Their work on the dyna-
mics involved in delinquency serves as a prime example of how the

psychoanalytic-intrapsychic position has been, and can be, modified

or extended into a family interaction viewpoint.

Superego Lacunae. Delinquent behavior was viewed tradition-

 

ally as the individual being excessively guilty about conflicts or
being driven by constitutionally unmanageable instinctual drives
(Giffin, Johnson and Littin, 1968). But Johnson and Szurek perceived
the child's delinquency, viewed as a circumscribed deficit in the
superego, ("superego lacunae"), as an outgrowth of family relational
dynamics. It is interesting to note that this change in perspective
was the result of a change in therapeutic procedures. Operating
from the traditional conceptual viewpoint, Johnson and Szurek

(p. 339) reported having "failed repeatedly to understand relatively
simple cases of acting out." They then engaged in "collaborative
therapy" in which both child and parent were seen in individual
dynamic psychotherapy and “the interchange between therapists (was)

regular and frequent" (Giffin et al., 1968, p. 676). Thus, while

"collaborative therapy" is not conjoint family therapy, it does
appear to have been an important step towards the gathering of
clinical observations on the interdependence in intrapsychic func-
tioning among family members. Many years later Framo (1972, p. 271)
also points out the clinical usefulness of a relational conceptual
orientation when he reports that his clinical experience in family
therapy has shown that symptoms or disordered behaviors "which
remain etiologically and dynamically obscure from the standpoint of
individual (intrapsychic) psychology, can often be decoded when
viewed within the matrix of their intimate social systems."

Johnson and Szurek's (1952) work with collaborative therapy
at the Chicago Institute for Juvenile Research led them to the
general observation "that the parental neurosis often provides the
unconscious impetus to the child's neurosis" (p. 326). It was a
logical extension when they arrived at their basic thesis regarding
delinquency, namely, "that antisocial acting out in a child is
unconsciously initiated, fostered, and sanctioned by the parents,
who vicariously achieve gratification for their own poorly integrated
forbidden impulses through a child's acting out" (Giffin et al.,
1968, p. 671). Hence it appears that the child's behavior and
intrapsychic life is intricately tied to the psychodynamics of other
family members, in this case particularly the parents, and this
interchange appears to cross even generational boundaries; the
child's superego lacunae are seen to "correspond to similar uncon-

scious defects of the parent's superego which in turn was derived

from the conscious and unconscious permissiveness of their own
parents" (Johnson and Szurek, 1952, p. 324).

Stierlin et al. (1971, p. 411) point out that this type of
conceptualizing implies a shift away from the traditional lines of
psychoanalytic theory "which held that the primary locus of influ-
ence was the child who was conceived as internalizing his/her parents
in part or whole--through imitation, introjection or identification."
The work of Johnson and Szurek, as well as others to be reviewed,
have "subsequently corrected this imbalance" by emphasizing the
parent's active contributions to such internalizations; while in
turn introdUCing, what at times appears to be the opposite imbalance
in that effects are still seen as unidirectional, i.e., parent to

child.

Scapegoating. Even though employing different terminology

 

and beginning with a focus on the family and not the individual,
Vogel and Bell (1968) have made observations similar to those stated
above. Their position is that the disturbed child is the family
scapegoat that "serves to condense a variety of social and psycholo-
gical problems impinging on the family“ (p. 416). Vogel and Bell
observed that the child was seen by the parents to possess very
undesirable traits. Even though the parents possess these same
undesirable traits, the scapegoated child serves to keep the focus

of attention away from the parents and onto the child. The parent's
individual problems are projected onto the child, so that the parents

deal with them as the child's problems rather than their own.

Congruent with the observations of Johnson and Szurek, Vogel and

Bell comment on how the parents encourage their children to act out
the parent's own repressed wishes. Additionally, it is noted that
the child is chosen as an appropriate object to symbolize conflict
and to draw off tensions which are typically the result of unresolved
conflicts within the marital relationship. Vogel and Bell state

that once the child has responded to the implicit wishes of the
family, particularly the parents, it is difficult to differentiate
the child playing the role from the child who has actually inter-

nalized the role and corresponding behaviors.

Object Relations Theory

 

The concepts of "delineations" (Zinner and Shapiro, l972)
and "irrational role assignments" (Framo, 1965, 1972, 1976), both
have their roots in the intrapsychic theory of object relations. In
order to advance the present review it will be helpful to sketch
briefly the contributions of object relations theory, since this
will help highlight the psychic mechanisms which underlie the recip-
rocal interdependence in psychological functioning found between

family members.

Delineations. The concept of "splitting" the object world

 

into good vs. bad objects is attributed to Melanie Klein. This is
a defensive mode evolving from an early infantile developmental

stage which is characterized by splitting of the ego and objects as
a defense against anxiety (Zinner and Shapiro, 1972). In order to

bring this concept into an interpersonal domain, Zinner and Shapiro

focus on Klein's concept of "projective identification". This is
a psychic mechanism which is defined as a combination of splitting-
off parts of the self and projecting them onto another person, with
the feeling of identification with the other because one has attri-
buted qualities or attributes of one's own to them. The contents
and dynamics of projective identification have been clearly examined
by Zinner and Shapiro while observing the ways in which parents
influence the adolescent's identity formation. These authors employ
the concept of "delineations" to refer to those modes of parental
perception and behavior which communicate to the adolescent the
parent's image of him/her. Zinner and Shapiro point out, however,
that many parental delineations reflect parental defensive needs
rather than a realistic appraisal of the adolescent's attributes,
i.e., this is an example of parental projective identification.
Since these "defensive delineations" are an expression of parental
defensive needs, there is a strong motivation to sustain these
perceptions rather than taking into account the adolescent's actual
behavior which might otherwise alter them. Zinner and Shapiro
(p. 524) point out that "these defensive delineations are the expres-
sion, at an individual level, of family group behavior which is
determined more by shared unconscious fantasies than by reality
considerations."

These delineating communications are very important to the
developing child in that they provide the raw material for adolescent
internalization and subsequent identity formation. Zinner and

Shapiro (p. 524) point out that the adolescent does identify with

 

(T)

these defensively distorted parental images, and thus his/her own
"subjective self—experience is likely to be affected by parental
efforts to diminish their own anxiety." It is interesting to note
the striking similarities between these observations made from an
object relations approach and those made by Johnson and Szurek from
a more traditional psychoanalytic perspective. Giffin et al. (1968,
p. 673) postulates that the dynamics involved in the transmission
of the superego lacunae involves the child's identification with the
parent, which "includes the subtleties of the parent's conscious and
unconscious image of the child." The child internalizes not only
the positive, socially acceptable attitude of the parents, "but also
the frequently unexpressed antisocial feelings."

The work of Klein has been elaborated by Fairbairn, who
postulates that the basic motivation of human behavior is the need
for a satisfying object relationship (Framo, 1972). Framo (1972,
1976) provides a schematicized review of Fairbairn's basic formula-
tions: The young child is placed in a precarious situation when
the parent's behavior is perceived as threatening or anxiety pro-
ducing, e.g., perceived as rejection, desertion and/or persecution.
The child is not in a position to give up the parents as external
objects, nor is the child frequently capable of making significant
changes in the family structure or environment. Fairbairn proposes
that the child handles such anxiety, frustration and disappointment
by internalizing the "loved-hated" parent. Thus while not capable
of modifying the external reality, the child is able to master and

control the object in the inner psychic world. Framo (1972, p. 274)

points out that "it is the emotional relationship between the self
and some external object which is internalized, not feelings as
such." The internalized object is subsequently repressed and
retained as an introject, a psychological representative. It is
essential to point out that these internal objects, undergo various
splits and become sub-identities and part of the structure of the
personality. As Framo (Ibid.) dramatically describes, the inter-
nalized objects continue to have a life of their own, significantly
affecting the individual's psychological functioning:

"The bad internal objects remain as 'internal saboteurs'

or warring forces in the inner world-~furious, guilty,

hungry, anxious, conflictual--consuming psychic energy

which the central ego should have available for evalua-

tion of reality and investment in external relationships."

Framo (1972, 1976) points out that this inner psychological
splitting of which Fairbairn speaks can be seen to have real external
counterparts, in that people will seek representatives of their bad
objects in their external relationships. People, especially inti-
mates, are often perceived largely in terms of the individual's own
needs, or as carrying one's own guilt-ladden denied, split-off
traits. Of special importance to understanding relational patterns
within the family is Framo's (1976, p. 194) statement that:

"Life situations in outer reality are not only uncon-

sciously interpreted in the light of the inner object

world, resulting in distorted expectations of other

people, but active unconscious attempts are made to

force and change close relationships into fitting the
internal role model."

Irrational Role Assignment. This provides the theoretical

 

background for Framo (1972, p. 274), who views the family as a

10

unique setting in which human needs "operate most forcefully", with
struggles over love, rejection, hurt, gratification and jealousy
"being a continuous dynamic process from one generation to the next."
Therefore, Framo's concept of "projective transference distortion"

or "irrational role assignment“--role being defined as a pattern or
type of behavior which a person builds up in terms of what others
expect and demand of him/her--"reflects unconscious attempts of the
parents to master, re-enact, or externalize their intrapsychic con-
flicts about these powerful needs, derived from relationship exper-
iences in the family of origin" (Ibid.). Similarly, Zinner and
Shapiro (1972) point out that parents' relations with their adoles-
cents contain highly conflicted elements of an object relationship
with the parent's own family of origin. While projective transfer-
ences, externalizations and vicarious participation all aid in the
avoidance of inner anxiety and the maintenance of psychological
equilibrium, Framo (1972, p. 279) suggests that "object possession"
may be the "chief motive underlying irrational role assignments"
since these mechanisms all serve the function of "recapturing the
symbolically retained family members." According to Framo, this
serves to delay the pain of loss and mourning. Zinner and Shapiro
observed that the parents' projections, as inferred from parental
delineations, contain elements of their own internalized relationships.
Thus Zinner and Shapiro (p. 526) agree with Framo when they state
that projective identification "serves not only a defensive function,
but also a restorative one to bring back to life in the form of the

offspring the parent's own lost objects, both good and bad."

11

Framo, and Zinner and Shapiro, respectively, point out that
the processes they describe take place in normal as well as "patho-
logical" families. Zinner and Shapiro state that the variables
relevant to the development of psychopathology involve the content
of the projected material, the capacity of the parents to differen-
tiate themselves from the child and the intensity of the parental
defensive requirements. Thus, depending on the nature of the inter-
action of these factors, "projective identification can endow a rela-
tionship with salutary empathic qualities or, to the contrary, gene-
rate binding attributions in which the child remains a creature of
the parental defensive economy" (Zinner and Shapiro, 1972, p. 526).
In a similar vein, Framo (1965, p. 192) states that the assigned
roles within the family are infinite. He has observed that in
multiple sibling families the various children "come to represent
valued or feared expectations of the parents, based on parental
introjects." Finally, Stewart et a1. (1975, p. 176) also point out
that while the child, seen as the identified client, becomes a
carrier or container of the split-off, unacceptable impulses of the
parents, "the child may be idealized just as he may be denigrated."

In order to complete this review it is essential to point
out that an essential part of the picture is still missing. For
example, Giffin et a1. (1968, p. 674) observed that the delinquent
child "wishes to do the things which he senses gives the parents
pleasure, even though he may be punished. There is a positive
undeniable drive towards acting in the manner which the parents

unconsciously wish, even though it be antisocial in direction."

12

Quite simply put, it is essential to ask Why? How can this obser-
vation of irrational role assignment or scapegoating be accounted
for in terms of the child, the recipient of these parental projec-

tions?

Projective Identification and Collusion. It will be helpful

 

to return to the mechanism of projective identification in order

to shed light on this important issue. Stewart et a1. (1975, p. 166)
points out that ”an integral part of projective identification is

the concept of collusion by which the recipient of the split-off part
of the partner does not disown the projection but acts upon the
conscious or unconscious message." Zinner and Shapiro (1972) also
point out that the subject's behavior in projective identification

is guided by two principles: 1) the subject interacts with that
projected part of him/herself in the object as she/he would inter-
act with the self were it internalized and 2) the subject must make
an effort to involve the object (i.e., the other person) as a collu-
sive partner in conforming with the way in which she/he is perceived.
The disinheriting of the projected part is not so complete that the
subject loses her/his capacity to experience vicariously a wide
range of the object's feelings, including gratification as well as
punishment and deprivation. Thus, for example, in the case of
delinquency, the child's behavior becomes a means through which the
"antisocial impulses of the parents are expressed through the child"
(Giffin, 1968, p. 676), as well as keeping those impulses in check

by vicariously experiencing the child's punishment. By externalizing

13

the conflict via projective identification, the parent is likely to
experience a dimunition in the intensity of anxiety compared to the
parent's internalization of the conflict.

This collusion can easily be seen in regards to the marital
relationship. The two partners engage in an implicit reciprocal
agreement, or trade-off, with each partner colluding to carry a
split-off part for the other (Dicks, 1963; Stewart et al., 1975).

In terms of the child within the family, however, the motivation to
collude is not as straightforward.

Zinner and Shapiro do provide a list of coercions and moti-
vational forces which are involved in the adolescent's colluding
with the parent's projective identification. Among these are l) the
opportunity for impulse gratification (e.g., the delinquent's behav-
ior), 2) actualization of omnipotent fantasies supported by the
child's power to determine his/her parent's self-experience, 3) tacit
parental compliance with the adolescent's defensive needs, and 4)
selective parental reinforcement of attributes of the adolescent
which conform to the parental projection. In regards to the second
alternative Stierlin (1976, p. 287) speaks of the schizophrenic
child as a "bound-up delegate", raising the possibility that the
parent's very own psychological and "perhaps even physical“ survival
now depends on the child. Thus the child, as the recipient of
massive parental projections, is given meaning, a sense of great
importance, a task and "perhaps most important" the child is given
power. The child can deliver him/her-self as "living proof of (the

parent's) own sickness and badness" (Stierlin et al., 1971, p. 425).

14

Additionally, in the extreme case, by letting him/her-self be
exploited and ruined the child gains the upper hand in the relation-
ship with the parents by operating the "guilt lever”. On the other
hand, there is also a sense of guilt which the child experiences
over hurting the parents. Therefore, the child is motivated to
collude with the unconscious parental needs and wishes, thereby
exhibiting a genuine desire to serve and help his family. The
child's collusion as an act of loyalty to the family captures, in
large part, the essence of Boszormenyi-Nagy and Spark's (1973) con-
cept of "invisible loyalties". Finally, it is interesting to note
that support for the fourth alternative listed above is provided by
Vogel and Bell (1968, p. 422). They point out that once the child
had responded to the implicit wishes of the parents and acted dis-
turbed, he/she was treated as disturbed and "the particular role
assigned to the child was appropriately rewarded."

While these motivational forces listed above are seen to
account, in part, for the child's collusion with the parent's pro-
jections, Zinner and Shapiro (1972) clearly point out that the
adolescent's collusion with the parent's defensive system is most
strongly motivated by the fear of object loss were the adolescent
not to act according to the parent's unconscious wishes and needs.
Framo points out that in some families veiled threats of emotional
or even legal disowning are used on those who have strayed from
their assigned roles. Stierlin et a1. (1971) also focus on this
same underlying dimension when they emphasize the importance of the

child's dependent emotional tie. Thus as Framo (1972, p. 302) points

15

out, for the sake of parental approval and because abandonment has
such disastrous consequences, the child will sacrifice whatever
ego integrity is called for in order to survive.

"If the price for acceptance is to absorb unrealities,

accept an irrational identity or role assignment, be

persecuted, be scapegoated, be parentified, or what

have you, this price will have to be paid; to be alone

or pushed out of the family either physically or psycho-

logically is too unthinkable."

Framo (1972, p. 280) appropriately cautions against labeling
the family processes and phenomena described in this review as
"conscious" or "unconscious". He suggests that these are "family
systems phenomena" and are not the result of intent on the part of
family members. These processes, outside the plan or control of
anyone involved, have a "rhythmic and gyroscopic force of their own"
which takes over as a regulatory mechanism. The external results
of splitting, projection and collusion become the "family way" of

seeing and doing things; they become automatic and unquestioned,

"like the air one breathes."

Summary

This review has presented clinical and theoretical support
for the interdependence between the intrapsychic and the interper-
sonal dimensions of human behavior. Neither the intrapsychic nor
the interpersonal levels of behavior can be replaced by the other,
or reduced to the others. As Framo (1972, p. 302) suggests, both
perspectives of understanding human behavior "are necessary for the
whole picture even though the enormous complexity of the relation-

ship between the two levels has only begun to be explored." Object

16

relations theory appears to be an extremely valuable tool in this
approach towards a truly "interactional psychology." As Zinner and
Shapiro astutely point out, object relations theory, and specifically
the broadly defined concept of projective identification, provide

an important and essential conceptual bridge "between an individual
and interpersonal psychology, since our awareness of the mechanism(s)
permit us to understand specific interactions among persons in terms
of specific conflicts occurring within individuals“ (p. 523).

Martin (1979) points out that there is currently a trend in
marriage and family therapy towards losing sight of the uniqueness
of each human being's personality. Conceptualizing the family
solely from a "systems” perspective omits a considerable amount of
important and crucial material. As Martin (p. 30) points out "the
family is not merely a collection of relationships." Rather it is
essential to stress the reciprocal interdependence between intra-
psychic dynamics and relational-interpersonal, i.e., family-systems,
dynamics. As noted above, the two cannot be separated. The litera-
ture reviewed in this section strongly points toward perceiving the
family as a multiperson system in which each family member carries
part of the motivation and behavior of the other; it is a system of
intimate relationships where the other can become a structural part
of the self. It appears that this perspective utilizes a systems
approach without ignoring the uniqueness of each individual's inner
world of psychological functioning. Therefore, this represents an

attempt to apply psychodynamics to a transactional field, the family,

(”I

4

a;

D!

E)!

17

where neither the individual nor the system in which he/she is
embedded is ignored.

One way to further expand our understanding of this intricate
relationship between the intrapsychic and the interpersonal is to
explore the interdependence in psychological functioning found
between family members. The clinical literature strongly suggests
that there is a reciprocal interdependence in psychic functioning
among family members. Family members collusively carry psychic
functions for each other, with mutual projection and transference
of internal objects. It is these processes which, when taken
together, appear to lie at the heart of this interdependency. In
terms of psychopathology, this perspective would lead one to examine
the ways in which a family member's abnormal behavior, e.g., delin-
quency, actually "fit" into the family system. It would be an
important step towards increasing our understanding of these pheno-
mena, if they could be brought under scrutiny and examined in a

controlled experimental setting.

Empirical Background

 

Recent reviews of the literature (Riskin and Faunce, 1972;
Jacob, 1975; Doane, 1978a) attest to the fact that there is now a
substantial body of empirical literature focused on family inter-
action research. Rather than survey this entire literature, the
present review will restrict its focus and thereby attempt to high-
light two primary objectives. 1) To determine if there is empirical/

experimental support for the theoretical positions reviewed in the

18

previous section. 2) It will be helpful to review the existent
methodology in the literature which has contributed directly to
the formation of the TAFIT. In order to reach this second objec-
tive, the present review will focus primarily on Interaction Test-
ing (Roman and Bauman, 1960) and the "unrevealed differences tech-
nique" used by Ferreira and Winter (1968), as well as a variety of
story telling techniques, especially the TAT, that have been

employed in family assessment and interaction research.

In Support of the Theoretical Review. Along with the atten-

 

tion paid to highlighting the methodological shortcomings in family
interaction research, a major goal of recent reviews of the litera-
ture (Riskin and Fauce, 1972; Jacob, 1975; Doane, 1978a), seems to
be an attempt to isolate those variables which consistently appear
to be significant factors in family interactions across a variety
of studies. Jacob (1975, p. 56) does conclude, however, that while
family interaction studies are based on a potentially "sound methodo-
logical strategy", in total the studies "have not yet isolated
family patterns that reliably differentiate disturbed from normal
groups." On the other hand, Doane (1978a, p. 372), who reviews the
same literature as Jacob, concludes that "despite the difficulties
inherent in family research, trends are evident." She cites as
important variables "coalitions", "flexibility" and "communication".
Similarly, Riskin and Faunce (1972) also notice the many contra-
dictory findings in family interaction research, but they do point

out that "humor", "agreement-disagreement", "support",

19

"acknowledgement-commitment-affirmation" and "clarity of communica-
tion" are all important family variables which represent "major
areas of agreement."

While it is not possible to resolve this current controversy,
especially the one between Jacob (1975; Jacob and Ground, 1978) and
Doane (1978a, 1978b), it is important to point out that the variables
under discussion do not easily lend themselves to the psychodynami-
cally oriented theories reviewed previously. Most of the variables
explored in family interaction research are not "depth" variables,
i.e., they do not represent family functioning on a dynamic/uncon-
scious level. For example, while communication clarity does appear
to be a significant factor involved in family life, the current
research does not attempt to assess possible underlying "depth" or
unconscious motives and dynamics which may in fact be at the root
of communication deviance. It is not the purpose of this review to
argue that one level of observation/exploration, e.g., the overt,
is more important than another level, e.g., the covert, but rather
to point out that while the two levels coexist in actual family
functioning, it is the more conscious overt level which receives
primary attention and focus in the empirical literature.

Therefore, it still appears reasonable to agree with Framo's
(1965, p. 409) conclusion, made 17 years ago, that "there is no body
of formalized literature on systematic research on family dynamics
with clear-cut stands taken on issues and specified limits from
which departures can be made." In other words, the existent empiri-

cal literature appears to be lacking in systematic research

20

approaches to the "inner dynamics of family life." It is possible
to speculate on why there is this apparent lack of attention to
family dynamics. It is reasonable to suggest that the experimen-
tally controlled exploration of, for example, "... deep, concealed,
family relationship needs, which are unconscious, infantile and
corrective of past hurts or perpetuations of past gratification..."
(Framo, 1965, p. 447) is, as an example of understatement, a formid-
able undertaking. The complexity of such a dynamic perspective,
with the heavy reliance on covert processes, appears at times to be
an overwhelming challenge for an empirical approach. Additionally,
there is the long standing problem of operationalizing, in a mean-
ingful way, the various dynamic constructs. While the empirical
approach is probably far away from capturing and manipulating these
significant family life variables in their full complexity, Framo
(Ibid.) does point out certain aspects that can be brought into
laboratory focus. He suggests that certain family-interpersonal
phenomena, e.g., double-bind situations, unconsciously shared
fantasy systems, unconscious collusion or "other admittedly concep-
tually loose principles" do need systematic exploration.

Clearly then, there is not a body of empirical literature
or support for the clinical and theoretical position reviewed in
the previous section of this study. There are limited data, however,
which can be taken as support for this perspective. Unfortunately,
these data were not collected in experimentally rigid, replicable
designs, but are rather more anecdotal and "impressionistic" in

nature.

21

A prime example of this type of support for the theory is
provided by Fisher and Mendell (1956, 1958; Mendell and Fisher,
1956, 1959; Mendel et al., 1968). These investigators approached
individual psychopathology by viewing the patient as acting out
conflicts which have concerned his or her family group over numerous
generations. This position was explored and subsequently supported
in a variety of studies all using a similar methodology. For
example, in the initial investigation (Fisher and Mendell, 1956)
there were 6 families with 3 generations of kin and 14 families
with 2 generations of kin. Every family member was given a battery
of projective tests and interviews and usually at least one from
each family was seen for individual or group psychotherapy. In this
initial study the projective data were analyzed and evaluated on a
"blind basis" and a personality profile was written out for each
person. The specific criteria or operational definitions for the
formulations of the "over-all personality pattern of each family
member" are not presented. Fisher and Mendell (1956, p. 42) state
that the first important finding to emerge from this analysis was
"that each family tends to be characterized by a special 'flavor'
or 'atmosphere'. There seems to be a fairly specific core neurotic
pattern which pervades the projective expression of members of each
given family." This line of investigation suggests that any given
family tends to be characterized by a distinguishing theme or
problem area, e.g., themes of death, fear of destructive loss of

self-control or disturbed concern with body image.

22

The fact that any given family tends to be characterized
by a distinguishing theme or problem area is also supported by a
different line of investigation. Ziegler and Muslinger (1977) are
utilizing data from Sander's (1964) extensive and rigorous study
of first-born infants and their mothers. Ziegler and Muslinger
describe their pilot study in which they have recently revisited,
some 15 years later, these same families, now focusing on the
adolescent child and his/her family relationships. During the
assessment of each adolescent and their family, (which was done
"blind" in terms of the previous findings), a central issue was
repeatedly observed. These same concerns or issues were reflected
some 15 years prior in the mother-infant pair and within the family
relationships. Ziegler and Muslinger (1977, p. 303) note that
"while many other issues were attended to, resolved and disappeared
from the data, (the early infant-mother data compared to the adoles-
cent data), one particular theme in each family surfaced repeatedly."
Furthermore, it is important to note that the index adolescent
appeared to incorporate these themes or issues into his/her own
personality functioning. For example, one adolescent girl whose
family was characterized by issues concerning limit setting and
control, revealed this similar theme or problem area during her
individual testing and interviews. She spoke of her own needs to
control different aspects of her life and the degree of apprehension
she experienced when situations seemed unclear or were not well-

defined. Even though this is clearly pilot data in that only 3 out

23

of the original 30 families had been recently revisited some 15 years
after the original study, it was possible to discern the persistence
of specific family themes or problem areas over this 15 year time
period.

It is essential to note that from a strictly empirical
perspective these studies have many significant shortcomings: e.g.,
no use of control groups, not providing of an index of reliability
for the projective data and the lack of clear operationalized defi-
nitions for the assessment of various themes. In fact, Fisher and
Mendell (1956) do point out that it would be more impressive if
their data could be Quantified instead of relying on an "impression-
istic overview". They clearly point out that it would "be more
striking if one could demonstrate in an objective quantitative
manner that the similarity in projective responses of those in each
family is greater than one would expect to find in a change aggrega-
tion of people" (p. 43), possibly utilizing Bodin's (1968) strategy
of "synthetic families".

Nevertheless, even with this "impressionistic overview",
these studies do provide some support for the stated theoretical
position. Specifically the work by Fisher and Mendell clearly
illustrates that it is possible to "identify fantasy patterns common
to a given family." This observation becomes an essential ingredient
in the formation of the present study. The fact that a key motif or
family atmosphere can be discerned from the fantasy or projective
responses of each family member strongly point towards the inter-

dependence in psychological functioning across family members.

tit;

SJ:

I“
1»-

”1917

Mr

24

Additionally, Mendell, Cleveland and Fisher (1968, p. 127) state
that these family themes "represent aspects of family interaction
that (are) forbidden or concealed.”

A closer examination of Mendell, Cleveland and Fisher's
(1968) most recent study in this area suggests even more specific
support for the theory. In this study, opportunity to "trace the
transmission of family fantasy patterns" across five generations
was provided in the examination of 27 members of one family, plus
two close relatives by marriage. The psychological tests were
"neither administered nor interpreted blind in the sense that the
subject and his test productions were unknown to the interpreter"
(p. 129).

Once again a common family theme was exhibited. This is a
dramatic finding when one considers that in addition to the immediate
family, family members sharing no physical or social proximity, with
no opportunity for collaboration, all revealed a similar theme or
problem area during their individual testing. Furthermore, there
are trends in the data which can be taken as evidence of the pheno-
mena of unconscious collusion. Mendell et al., (1968, p. 129) state
that in this family there was an unusual emphasis on "one or the
other of two related themes: either the test fantasies were con-
cerned with control, maintaining a fixed and rigid restraint over
one's impulses, or there was a loose and unstructured quality
expressed." The test reports made repeated references to family
members' concerns, e.g., "concerned with acting in a proper and

morally strict manner vs. being loose and uncontrolled." Framo

25

(1965, p. 452) points out that ”the unconscious exchange frequently
appears in the form of themes or recurrent problems between people
who are deeply related; both parties in the interchange take a posi-
tion vis-a-vis the problem." Thus, it seems very possible that the
Mendell et al., (1968) data reveal this collusion with some family
members expressing the impulse, e.g., "being loose or uncontrolled”,
while others reject the impulse, e.g., appearing "morally strict."
It is reasonable to speculate that all family members are concerned
with the exact same issue, but have worked out an unconscious rela-

tionship within which the defenses against the problem are opposite.

Interaction Testing and the Unrevealed Differences Technique.

 

These two similar procedures provide the framework for the TAFIT,
the method used in the present study. Basically, these two techni-
ques refer to the administration of a psychological test or research
questionnaire as a two-stage procedure: first the instrument is
administered individually to each member of a family or group and
then re-administered to the group as a whole, requiring a group or
family response. As Bodin (1968), in his review of family assess-
ment techniques, points out, this type of paradigm yields three
classes of data: 1) individual scores or comparisons, 2) conjoint/
family or interaction scores, and 3) individual-conjoint comparisons.
Interaction Testing (Roman and Bauman, 1960; Bauman and
Roman, 1966, 1968) utilizes this paradigm in the administration and
analysis of psychological tests. This method includes projectives,

even though the focus of this line of investigation has been on the

(J.

‘1:
It)

Ex

GS

to

Dr

26

WAIS in terms of “marital intelligence" (e.g., Bauman et al., 1967).
In the analysis of these data Bauman and Roman stress the end pro-
duct, focusing on both the individual and conjoint response. The
important assumption underlying this "interaction product analysis"
is the hypothesis of an individual-group psychological isomorphism.
This assumption, which is essential to the formation of the TAFIT,
states that the inner psychological organization of groups is funda-
mentally similar to that of individuals, e.g., both individuals and
families show memory for past learning and experience; a group pre-
serves characteristic behavior habits and structure, similar to the
concept of personality; the group possesses drives which become more
or less integrated in executive functions of nutrition, acquisition,
aggression, defense, etc. As Bauman and Roman (1968, p. 331) point
out "groups vary in dynamic integration analogously to the variations
of individuals in character." Roman and Bauman (1960, p. 95) appro-
priately caution that this assumption of individual-group isomorphism
in no way denies the existence of significant differences in struc-
ture between individuals and groups, but is offered rather as "a
conceptual approach to the study of group behavior which seems
promising."

Bauman and Roman do provide some preliminary well controlled
experimental support for the reliability, utility and the underlying
assumptions of interaction testing. For example, Bauman et al.,
(1967) administered parts of the WAIS to 50 marital couples in order
to determine the "marital IQ". Using a split-half reliability

procedure these investigators found that the interaction product,

27

i.e., the marital IQ derived from the conjoint responses, was as
statistically reliable, e.g., .85, as individual IQ scores. Addi-
tionally, they have developed four process categories in order to
analyze the individual-conjoint comparisons: l) Dominance - when
the interaction response contains one member's individual response
and the absence of the other's; 2) Combination - elements of both
members' responses are found in the interaction product/response;
3) Emergence - the presence of a new response in the interaction
product, and 4) Reinforcement - when the same response is given by
both individuals and appears as the interaction response. Further—
more, the 10 scores outcomes of the decision-making process can be
evaluated as a) +, indicating an increase in the accuracy of the
response; b) -, indicating a decrease in the accuracy of the response
or c) 0, indicating no change in terms of the accuracy of the
response. Thus, the interaction product can be evaluated in terms
of positive dominance, negative emergence, positive combination,
etc.

On the basis of these evaluative procedures Bauman and Roman
have conducted a few studies exploring the marital relationship.
For example, Bauman and Roman (1966) studied dominance in 50 marital
pairs. They found that the husband dominates more than the wife,
the more competent spouse (i.e., higher individual IQ) dominates the
less competent and that non-patients dominate more than their
patient spouses do. In a later study (Bauman et al., 1967) it was
found that post-hospital interaction products contained significantly

more reinforcement and significantly less negative emergence than

28

during the interactions when one spouse was hospitalized. The
importance of these studies is that they illustrate the usefulness
of both interaction testing and this scoring method. Thus Bauman
and Roman (1968, p. 331) take these preliminary findings as suppor-
tive of interaction testing and the underlying assumptions on which
it is based:

"The scores and profiles from consensus protocols on

traditional clinical tests are characteristic of the

group--that is, reliable--and predictive of the group's
functioning--that is, valid--in ways that are analogous

to the scores derived from the same tests when admin-

istered to individuals."

While these data on marital intelligence were collected
under experimentally controlled conditions, Roman and Bauman's
work with projectives, primarily the Rorschach but including the
TAT, is more a "one-of—a-kind case-study". Nevertheless, they do
report that their work with projectives suggests the "fruitfulness"
of this product-oriented approach to interaction testing (e.g.,
Roman and Bauman, 1960). Bauman and Roman (1968) do report that
for projectives, in this case the Rorschach, the interaction
response is found to act at least in some ways like the individual
response. They point out that one of the "dividends” from this
"interaction product analysis" is that conjoint responses can be
used to derive inferred process scores, e.g., dominance, combination
etc. Additionally, Bauman and Roman (1968) suggest it is useful
to "turn the conceptual tables" and consider the group as a model

for understanding individual behavior. In other words, interaction

testing can reveal data about the effects of the group on the

29

personality functioning of those individuals in the group. A
relevant example for the present study is provided by a case exami-
nation in which the Rorschach was used in interaction testing. In
this case (Roman and Bauman, 1960, p. 129) there is evidence to
suspect "profound unconscious collusion in which the son expresses
his mother's deeper disturbance, thus helping her to avoid expressing
it."

A similar methodological paradigm, which has been referred
to as the "unrevealed differences technique“, has been employed in
the investigation of decision-making in "normal and abnormal fami-
lies" (Ferreira, 1963; Ferreira and Winter, 1965, 1968). For this
line of investigation, these researchers developed a questionnaire
requiring preferences around "neutral content". The questionnaire
was first administered individually. Then the family was brought
together and, while not informed of each others' preferences, asked
to complete the same questionnaire, this time representing the
family's preferences. Ferreira and Winter employed the following
categorization of the data: 1) Spontaneous Agreement (SA)--repre-
senting the amount of agreement or matched preferences by comparing
the individually filled out questionnaire; 2) Choice Fulfillment
(CF)--this is a measure of the number of instances where the indi-
vidual preference was also the family preference, and 3) Decision-
Time (DT)--the time the family took to complete the joint question-
naire. In the original study Ferreira (1963) employed somewhat
different variables: majority decisions, dictatorial decisions and

chaotic decisions; the latter corresponding to "emergence".

30

This line of investigation has not only provided support
for the stability and reliability of these measures, but has also
consistently found differences in the decision-making processes
between normal and abnormal families. It has been found that normal
families have a significantly higher rate of SA, less OT, and
greater CF, even when the rate of SA is statistically controlled
for, when compared to abnormal families. Ferreira and Winter (1965,
1968) interpret their data by suggesting that normal families are
characterized by more open communication, sharing and warmth.
Additionally, they interpret the OT and CF scores as exhibiting
a more efficient decision-making process in which the family members
are more personally satisfied. On the other hand, Ferreira and
Winter characterize the abnormal families as having less communica-
tion, more conflict, less sharing of opinions and feelings and
generally less information exchange. The abnormal families' lower
CF score is seen as a reflection of their inefficiency in decision—
making and is taken as an index of less personal satisfaction and
fulfillment of needs and wishes. In order further to support this
position Winter et al., (1973) examined these same processes in
married couples as well as unrelated couples. As predicted, the
married couples have a significantly greater rate of SA, but con-
trary to expectations the unrelated group did not show a difference
in terms of CF.

Even though the overall interpretation of the decision-
making data is reasonable and well formulated, it does appear

possible to speculate that the rates for SA and CF in abnormal

31

families may also reflect underlying dynamic processes. For example,
it seems reasonable to propose that in abnormal families there are
more "delineations". Each person is more rigidly assigned a certain
role in the family, affecting the member's psychological make-up as
exhibited in their differing perceptions and attitudes or their

likes and dislikes as revealed on the questionnaire.

The TAT and Family Research. A number of studies have used

 

TAT productions in family interaction research, developing different
techniques and focusing on differing dimensions. There are a number
of studies which administer the TAT to families, either to each
family member individually or to the family as a whole, where the
primary focus of analysis is not on the stories' content/theme per
se. Rather, in these studies, there is an emphasis on either the
cognitive style exhibited in the stories or on different dimensions
of family interaction and communication as evidenced during the
formation of the family's response (Singer and Wynne, 1965, 1966;
Ferreira et al., 1966; Friedman and Friedman, 1970; Lieber, 1977).
Thus, these different investigations have used the TAT as a vehicle
by which to capture and examine characteristic patterns of relating,
communicating and/or thinking within the family. For example,
Singer and Wynne (1966, p. 262) point out that they use the TAT

not as a "specific test" but rather as "a convenient means of
sampling verbal transactions under conditions in which a particular
range and form of behavior can be expected." Therefore, while these

studies employ the TAT in an examination of a number of interesting

32

variables, they are not particularly relevant to the focus of the
present study.- Rather, it is more important to review selectively
those studies which exhibit a specifically content or thematic
orientation in the analysis of the family produced TAT data.

Sohler et a1. (1957) were able to successfully predict
family interaction from a battery of projective tests, e.g., TAT,
Rorschach, Draw-a-Person and Sentence Completion Test, administered
individually to each family member. One of Sohler et a1. (1957,

p. 207) conclusions was that the TAT "seemed to be the best source
of information about attitudes towards other members of the family."
A review of more recent literature does in fact reveal that the

TAT has been used for this purpose. For example, Werner, Stabenau
and Pollin (1970; Stabenau et al., 1965) focused on those TAT
stories produced by parents in which the manifest content revealed
parent-child interactions. The parent-child interactions were
analyzed using three general categories: a) child-centered, flex-
ible interactions; b) impersonally involved, superficial inter-
actions, and c) over involved, parent-centered interactions. 0n
the assumption that you can infer patterns in living from TAT
productions, the results were interpreted as demonstrating that
"quite distinct patterns of parent-child interactions characterize
parents whose children do well from those whose children become
delinquent; and that both differ from parents whose children become
schizophrenic" (Werner et al., 1970, p. 143). Similarly, Fisher,
Boyd,Walker and Sheer (1959) were able to differentiate the TAT

stories produced by parents of normals from the parents of neurotic

33

and schizophrenic patients by "level of aspiration" and "parental
images, i.e., attitudes towards the parental figures.

A study of Goldstein et a1. (1970) also focused on the
family's "interpersonal themes” revealed in the TAT responses.

They developed quantitative indices of interpersonal themes and
compared the TAT stories of both adolescents and their parents, as
they related to the child's specific behavioral disturbance. The
stories were rated for the structure of interpersonal relationships
(on a 5 point scale, 5 indicated a definite nuclear family unit and
l reflected an absence of any identifiable interpersonal relation-
ship) as well as for the nature of the relationship, ranging from
positive involvement to negative involvement. The results indi-
cated "certain consistencies" between the adolescent's psychopathol-
ogy and the parent's TAT stories, and to a lesser extent consis-
tencies among the various sub-groupings of the adolescent's TAT
responses, but a direct comparison between adolescent and parent
stories was not undertaken.

The few studies reviewed thus far do support the position
that the TAT is a valuable tool in family research and in detecting
differences across families. These differences appear to corres-
pond to differing nosological categories or criterion groups.
Underlying these studies is the assumption that there is a corres-
pondence between the way family relationships are seen or projected
in TAT stories and the actual relational patterns within the family.
Support for this view comes from studies (e.g., Sohler et a1, 1957;

Fisher et al., 1959; Stabenau et al., 1965) which make comparisons

34

between the TAT data and other assessment techniques, or with direct
behavioral observations, as well as those which use the TAT as a
projective instrument (e.g., Murstein, 1963). There can be a
correlation between the projected verbal response and overt behavior,
provided that the scoring or analysis is sophisticated in order to
pick up subtleties in the story's content, e.g., aggressive content
associated with much guilt or inhibition would not be a good pre-
dictor of overt behavior. On the other hand, if a study relied
exclusively on parental data, this would raise issues as to whether
the content represented present family functioning, past family
functioning (i.e., family of origin) or wishes for family function-
ing.

There is a different methodological paradigm which overcomes
the limitations of relying on the parent's responses alone. This
involves analyzing the interpersonal themes revealed by each indi-
vidual member of a given family. This allows for an examination of
the common and consistent themes, as well as providing data which
could prove helpful in putting together the various individual
perceptions which in actuality operate within the family at any
given time. For example Kadushin et a1. (1969, 1971) have employed
this strategy with the Family Story Technique (FST), which uses TAT
cards. These investigators analyze the TAT responses using Fine's
(1955) method of scoring interpersonal themes. Each individual's
stories are scored for a) affect related conditions, b) family
interactions, and c) outcomes. The scoring of the interactions

is based on Horney's categories of interpersonal relations, i.e.,

35

moving towards, moving against and moving away from. Kadushin

et a1. (1969) studied one family with this technique illustrating
the way the FST "systematically points to a dynamic description of
the family."

Minuchin et al. (1967; Elbert et al., 1964) have adminis—
tered the Family Interaction Apperception Test (FIAT) which utilizes
TAT-like pictorial stimuli. Thus the FIAT is a projective test,
administered individually to each family member, with the data
analysis focusing on the needs, wishes and affects, as well as the
intra-familial perceptions of family members. In the original
publication, a case study, these authors state that "in pooling the
tests of all family members, constellations of congruent and dis-
sonant perceptions of each other can be examined, as well as mutual
satisfaction and frustration of needs and interlocking defensive
systems" (Elbert et al., 1964, p. 886). For example, the case study
involves a family with three delinquent sons. The mother's reac-
tion to her son's aggression is revealed in one story in which she
describes the aggressive action of her son "with relish": "So this
16 year old girl starts teasing my son and he hit her and he's only
8 years old and he split her lip. I had to laugh." This same
mother, "known to be sexually delinquent", exhibited conflict and
disturbance involving her own aggressive impulses. Elbert et a1.
(1964) state that the FIAT revealed that the mother is getting
vicarious satisfaction from the children's acting out behavior.

Thus, the pooling of the family data provided these inves-

tigators with an understanding of the family dynamics. In fact,

 

36

this study could have very easily been reviewed in an earlier sec-
tion as support for the proposed theory, since it clearly illustrates
the interdependence in psychological functioning across family mem-
bers. In that previous section the work by Fisher and Mendell
(1956), which also administered the TAT to each family member, was
reviewed. The common assessment methods of these various studies
(i.e., Elbert, 1964; Fisher and Mendell, 1956; Kadushin, 1969; as
well as the preliminary work (1960) of Roman and Bauman) consistently
suggest that a fruitful methodological approach for the empirical
investigation of "depth" variables within the family is to administer
projective tests, in this case the TAT, to each family member. This
method can provide a useful avenue for the empirical investigation
of family dynamics by examining both the overt and covert relational
patterns found within families.

Furthermore, it is important to point out that the data
obtained in this way can be quantitatively scored and subjected
to statistical analysis; thus the TAT can be a valuable tool in
systematic family research. For example, the FIAT was later refined
and employed in a research project on families with delinquent
children (Minuchin et al., 1967). In this study the variables,
e.g., control, guidance, acceptance of responsibility, nurturance,
affection, cooperation, aggression and family harmony, were reported
to be reliably scored by a quantitative rating system, i.e., 86%
over-all agreement out of 110 categorized responses. Similarly,
Kadushin et a1. (1971) employed the FST in a pre- post-family
therapy study. They state (p. 71) that "scoring by the system used

37

(in their study) or by many others available in the TAT literature
provides convenient quantification for statistically oriented
studies." Also, these researchers point out that the TAT technique
provides a "fascinating fund of information“ and "reorienting the
TAT to productions of 'family stories' does not detract from, and
may possibly augment, its yield" (Ibid).

Another important, and for the present study relevant, use
of the TAT in family research is presented by Winter, Ferreira and
Olsen (1965, 1966). These investigators were able to differentiate
normal from abnormal groups of three person families on the basis
of the family's conjointly produced TAT stories. For example,
Winter et al., (1965) presented three sets of TAT cards, each set
containing three cards, and asked the family to make up a story
which would link up the three cards together in the order in which
they were presented. These family-produced stories were scored by
the method of Story Sequence Analysis developed by Arnold (1962).
Focusing on the essential meaning or import of the sequential themes
on the TAT protocols according to a complex criteria, the emphasis
of Story Sequence Analysis is on the motivational system of the
storyteller, in this case the family as a unit. The Winter et a1.
(1965) results indicated that compared to normal families, families
with abnormal children (e.g., "emotionally maladjusted", schizo-
phrenic or delinquent) are characterized by negative motivational
patterns. This method of analysis did not, however, differentiate

the three abnormal family groups from each other.

38

In the second publication on the same sample (Winter et al.,
1966) these researchers attempted to differentiate the three abnormal
groups in terms of the hostility themes revealed in the family-
produced TAT stories. The protocols were scored along two dimen-
sions: 1) Weighted Hostility - for this scale specified themes
are given weights according to the degree of hostility they repre-
sent, e.g., a weight of 4 is assigned to those stories involving
direct fighting or assault, and 2) Percentage of Overt Hostility -
each hostile theme that is assigned a weight is judged according
to criteria to be either Overt or Covert in nature. Thus this
score is the ratio, in percentage, of the number of Overt Themes/
number of Overt + Covert Themes. By focusing the analysis on the
hostility themes, Winter et a1. (1966) were again able to differen-
tiate the normal from the abnormal families. The results from this
analysis, moreover, detected differing patterns among the four sub-
groupings of abnormal families. One particularly relevant pattern
was exhibited by the male delinquent-family group. While this
group had the highest scores for weighted hostility compared to the
other male family groups, their percentage of overt hostility was
no different from the normal family group. Winter et a1. (1966)
state that the results from this delinquent group were the "least
clearly defined", and thus it appears that this pattern of hostility
scores is open to speculation. On the basis of the theoretical
positions reviewed previously, especially Johnson and Szurek (1952),
it is reasonable to suggest that the family-produced TAT stories

for the delinquent group revealed alot of conflict and disturbance

39

around issues of aggression and hostility, i.e., this would account
for the group's higher weighted hostility score. At the same time,
it is the son who unconsciously colludes and is the only family
member to act out these impulses for one or both of his parents;
therefore, as a family group, they would have a lower ratio of overt
hostility/total hostility than other types of families. Clearly,
this speculative interpretation of the Winter et a1. results needs
further experimental support and exploration.

In conclusion, it is apparent that there is not an extensive
body of literature on the TAT used with families, but the available
evidence does consistently suggest that the TAT is a valuable tool
in family research. One of the shortcomings in this existent TAT
family literature is the lack of systematic approaches which attempt
to combine the various methods, e.g., administering the TAT to
parents, to each family member or to the family as a whole. It is
only the preliminary work of Roman and Bauman (1960), reviewed in
a previous section, which combines these various data gathering
procedures into one paradigm, Interaction Product Analysis. It is
in this paradigm, similar to the "unrevealed differences technique",
that both individual and family responses can be analyzed, thus
allowing for individual-family comparisons. Therefore, the
researcher has an opportunity to explore the interdependence between
both individual and family produced fantasy patterns, thus providing
an important method for bridging the gap between intrapsychic and

family-interpersonal dimensions of behavior.

40

Summary and Hypotheses

 

There are a variety of clinical and theoretical perspectives
which perceive the family as a social system with both overt and
covert, i.e., unconscious, relational patterns. The processes of
object splitting, projective identification and psychological collu-
sion are said to account for the fact that family members "recipro-
cally carry part of each others' psychology and form a feedback
system which in turn regulates and patterns (the) individual
behavior" (Framo, 1972, p. 271). Thus, it is reasonable to suggest
that within the family there are behavioral roles, e.g., organizer,
leader or provider, as well as "psychological roles", e.g., scape—
goat or bound-up delegate, and that these psychological roles can
serve a homeostatic function for the family-system, e.g., in the
case of the delinquent child expressing and releasing the aggressive
impulses for the family.

The Systematic patterning of these psychological roles
within the family is the focus of the present study. It does appear
that the family can be characterized by a partitioning of psycholo-
gical functions and intrapsychic content among family members, and
this partitioning takes place in all family systems. In a "patholo-
gical" family system, the partitioning takes place within rigid,
nonoverlapping boundaries, while in normal families there is a more
fluid cooperative type of partitioning. Additionally, as Zinner
and Shapiro (1972) point out, it is the nature of the intrapsychic

content which is partitioned or assigned and subsequently carried by

41

one family member that can account for the development of psycho-
pathology.

These presumed covert family-system dynamics form the basic
thesis of this study: The intrapsychic functioning of each family
member is bound within a reciprocal interdependent family-system,
which is governed by both overt and covert relational patterns. It
is the goal of this investigation to subject this dynamic relational
family-system perspective to systematic empirical exploration.

Our understanding of intrapsychic functioning has been
expanded by the use of projective instruments. Covert or unconscious
family dynamics may also be explored via projective assessment pro-
cedures. In fact, the literature does point towards the usefulness
of these procedures in understanding family dynamics. The present
study is an attempt to offer support for the appropriateness and
further development of such an assessment procedure. The TAFIT,
which follows closely on the footsteps of Interaction Product Analy-
sis (Roman and Bauman, 1960), is designed to explore the covert
relational patterns/dynamics found within families.

Given the basic methodology and the study's thesis, as
stated, an essential question still remains unanswered. How can
one find empirical support for this thesis or theoretical position?
If in fact the psychological functioning of each family member is
part of a reciprocal interdependent family-system, then the intra—
psychic content, or projected content, of each individual family
member should fit into a predictable, specified pattern. If covert

relational dynamics actually govern the intrapsychic functioning of

I.

 

42

each family member, then the interdependency in intrapsychic func-
tioning will operate according to systematic principles. Therefore,
fantasy content, for example, should fit into an intelligible and
systematic pattern. If this predicted systematic pattern is in
fact obtained, this will support the stated theoretical position,
as well as offer support for the appropriateness of the TAFIT as a
family assessment procedure.

This study explores the patterns in the TAT fantasy material
produced by three person family triads, tested both individually
and conjointly, in which the child has a history of disruptive or
rebellious behavior at school. This study is not strictly concerned
with the dynamics involved in this type of family per se, but rather
this type of family provides an opportunity to begin an empirical
examination of the interdependency in psychological functioning
found between family members. This approach will further provide
a reasonable test of the efficiency and usefulness of the TAFIT.
This type of family allows for a set of predictions, following
closely the logic and rationale of Johnson and Szurek, about the
patterning of the fantasy content. If the results support the
predicted patterning of fantasy content, this won't prove that in
all families the intrapsychic world of the members is interdepen-
dently tied, but it will be a first step, as well as suggesting
that the TAFIT is a valuable tool for further empirical investiga-
tion.

The theoretical position proposes that the behavioral

problems at school reflect the adolescent's inner psychological

43

world and this inner world can be understood from a family-systems
perspective. It is proposed that the adolescent's behavior is sus-
tained by the family since he/she is unconsciously colluding to act-
out and express the denied or split-off disruptive/aggressive
tendencies of either one or both parents. This theoretical position,
which clearly exemplifies the interdependence in psychological
functioning, would make specific predictions about patterns of TAT
hostility content based on these relational dynamics. If the adoles-
cent with behavior problems at school is viewed as unconsciously
colluding to carry the uninhibited or overt aggression for the
family, it is expected that this dynamic will be revealed in the
fantasy patterns, when comparisons are made across individual mem-

bers.

Hypotheses

 

The following definitions will be useful for a clear under-

standing of the hypotheses:

Weighted Hostility Score: This score reflects the degree

 

of hostile content appearing in the TAT stories. A judgment is made
for each theme, using a scale ranging from a low of 0 (no hostile

content) to a high of 4.

Number of Overt Hostility Themes: Overt hostility is that

 

hostility which is manifest and direct. The number of overt hosti-
lity themes is the total number of themes across a set of stories

judged to be overtly hostile.

 

 

 

 

 

 

44

Number of Covert Hostility Themes: Covert hostility is that
hostility which is insidious, indirect, disguised or latent. The
number of covert hostility themes is the total number of themes

across a set of stories judged to be covertly hostile.

Percentage of Overt Hostility: This is the ratio of themes
judged to be overtly hostile to the total number of themes that
were rated, expressed in terms of a percentage, i.e., (number of
overtly hostile themes/number of overtly hostile themes + number of

covertly hostile themes) X 100.

Predicted Family Score (PFS): The PFS is the arithmetic
mean of the mother's, father's and son's individual scores. For
example, if the weighted hostility scores in a given family were
mother = 20, father = 22 and son = 24, then the PFS would be
(20 + 22 + 24)/3 or 22.

Influence Score: The "influence" score is a difference
score. It is the difference between an individual's score and
his/her family's score. For example, if a son had a weighted
hostility score of 30, and the stories produced by the family group
had a weighted hostility score of 23, then the son's influence score
would be 7. Thus the lgwgr this score, in terms of the absolute

value, the greater the individual's "influence" on the family score.

Hypothesis 1. This hypothesis predicts differences in the
content of the TAT stories produced by the adolescent's with behav-

ioral problems and the normal adolescents.

1-A.

1-B.

1-C.

45

The behavioral problem adolescent's stories will
exhibit an amount of weighted hostility significantly

different from the normal adolescent's stories.

The behavioral problem adolescent's stories will
exhibit a percentage of overt hostility significantly

different from the normal adolescent's stories.

The behavioral problem adolescent's stories will exhibit
a number of overt hostility themes significantly

different from the normal adolescent's stories.

Hypothesis 2. This hypothesis predicts differences between

 

the content of the stories produced by the adolescents with behav-

ioral problems and the content of their parents' stories and that

these differences in story content will not be found between the

normal adolescent's and their parents.

2-A.

2-B.

The stories produced by adolescents with behavioral
problems will have a number of overt hostility themes
significantly different from the arithmetic mean of
their parent's number of overt hostility themes and
this difference will not be found in the normal fami-

lies.

The stories produced by adolescents with behavioral
problems will have an amount of weighted hostility

significantly different from the arithmetic mean of

2-C.

46

their parent's weighted hostility and this difference

will not be found in the normal families.

The stories produced by adolescents with behavioral
problems will have a percentage of overt hostility
significantly different from the arithmetic mean of
their parent's percentage of overt themes and this

difference will not be found in the normal families.

Hypothesis 3. This hypothesis predicts differences in the

 

variability of the influence scores for the adolescents with behav-

ioral problems compared to the normal adolescents.

3-A.

3-B.

3-C.

The influence scores, derived from the number of
overt hostility themes, of adolescents with behavioral
problems will exhibit a variance different from the
variance of influence scores exhibited by the normal

adolescents.

The influence scores, derived from the weighted
hostility scores, of adolescents with behavioral
problems will exhibit a different variance from the
variance of influence scores exhibited by the normal

adolescents.

The influence scores, derived from percentages of
overt hostility, of adolescents with behavioral

problems will exhibit a variance different from the

47

variance of influence scores exhibited by the normal

adolescents.

,flypothesis 4. This hypothesis predicts that the family

 

stories produced by those families with an adolescent with behav-
ioral problems will have different content from those stories

produced by normal families.

4-A. The family stories produced by those families with
an adolescent with behavioral problems will have a
significantly different amount of weighted hostility

from those stories produced by normal families.

4-8. The family stories produced by those families with
an adolescent with behavioral problems will have a
significantly different percentage of overt hostility

from those stories produced by normal families.

4-C. The family stories produced by those families with
an adolescent with behavioral problems will have a
significantly different number of overtly hostile

themes from those produced by normal families.

Hypothesis 5. This hypothesis predicts that the difference

 

between the predicted family scores and the actual family scores
will be a variable that discriminates the families of adolescents

with behavioral problems from normal families.

48

5-A. The absolute difference between the predicted family
score for weighted hostility and the actual family
score (i.e., PFS for weighted hostility minus actual
weighted hostility for the family stories) will be a
variable that significantly discriminates the families
of adolescents with behavioral problems from normal

families.

5-B. The absolute difference between the predicted family
score for overt hostility (number of themes scored)
and the actual family score will be a variable that
significantly discriminates the families of adoles-

cents with behavioral problems from normal families.

5-C. The absolute difference between the predicted family
score for the percentage of overt hostile themes and
the actual family score will be a variable that
significantly discriminates the families of adoles-
cents with behavioral problems from the normal fami-

lies.

Hypothesis 6. This hypothesis predicts that the stories

 

produced by the parents of adolescents with behavioral problems will
have different content from the stories produced by "normal" parents,

i.e., the parents of the normal adolescents.

6-A. The combined weighted hostility scores of the parents

of adolescents with behavioral problems, (i.e.,

6-B.

6-D.

6-F.

49

father's score + mother's score), will be signifi-
cantly different from the combined scores of the

"normal" parents.

The parents of adolescents with behavioral problems
will exhibit a significantly different combined score
for the number of overt hostility themes from the

combined scores of the "normal" parents.

The parents of adolescents with behavioral problems
will exhibit a significantly different combined
percentage of overt hostility from that of the "normal"

parents.

The stories produced by fathers of adolescents with
behavioral problems will have a significantly differ-
ent amount of weighted hostility from those stories

produced by "normal" fathers.

The stories produced by fathers of adolescents with
behavioral problems will have a significantly differ-
ent number of overt hostility themes from those

stories produced by "normal" fathers.

The stories produced by fathers of adolescents with
behavioral problems will have a significantly differ-
ent percentage of overt hostility from those stories

produced by "normal" fathers.

6-G.

6-H.

6-1.

50

The stories produced by mothers of adolescents with
behavioral problems will have a significantly differ-
ent amount of weighted hostility from those stories

produced by "normal" mothers.

The stories produced by mothers of adolescents with
behavioral problems will have a significantly differ-
ent number of overt hostility themes from those

stories produced by ”normal" mothers.

The stories produced by mothers of adolescents with
behavioral problems will have a significantly differ-
ent percentage of overt hostility from those stories

produced by "normal" mothers.

CHAPTER 2
METHOD

Participants

 

The participants in this study were 20 family units consist-
ing of a father, mother and an adolescent son. There were three
families, (two Experimental and one Control), in which there had
been a previous separation or divorce between the natural parents.
In these three families the step-parent had been living within the
reconstructed family unit for four years, 12 years and 12 years,
respectively, thereby exhibiting involvement in the adolescent's
development. The participating families were divided into two

groups, with 10 families in each of the following groups:

Experimental Gropp. The families in this group had an

 

adolescent son attending high school whose educational background/
record contained a history of behavioral problems which can be
characterized as disobedience, disruptiveness, aggressiveness or
rebelliousness. As shown in Table l, the actual school records of
these boys exhibited a wide range of disruptive behaviors. Place-
ment in the Experimental Group required that the adolescent son have
a minimum of three such offenses recorded on his record for the
current academic year. The actual number of offenses ranged from

3 to 11, with the mean number of offenses being 5.5 (SD = 2.66).

51

f\_)

1

.3:-

(.11

20.
21.
22.

52

‘TABLE l.--Types of Offenses Reported in the School Records of

the Adolescents With Behavioral Problems

 

10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.

LII-DOOM

Insubordination

Talking back

Fighting

Shoving

Inappropriate classroom behavior
Throwing of objects in the classroom
Disruption

Foul language and/or gestures
Fire setting

Refusing to work

Door slamming and/or banging
Disobedience

Not following instructions
Horseplay

Rude and/or talking rude

Yelling

Misconduct

Disrespectful

Shoving furniture and/or kicking doors
Lying, cheating and/or stealing
Temper problems

"Mouthing"

 

53

Control Group: These ten families were selected from the

 

same high school population. The students in this group did not
have an educational history indicating behavioral problems. Four

of the students in this group did have one offense recorded on their
current school record, while the other six students had no reported
offenses. These ten families were to be matched as closely as
possible to the Experimental Group on the following two variables:
student's grade level (ranging from 9th to 12th grade) and student's
percentile rank in English and Mathematics on the Stanford Achieve-

ment Test, administered in the 8th and 10th grade.

Participant Selection

 

The participants in this study were selected from the same
high school population. The high school, with an enrollment of 611
students, is in a mostly rural farming area, (population approxi-
mately 3,000), located about 20 miles from Lansing, Michigan, the
state capital.

The principal of the high school was involved in the selec-
tion procedure. He selected a pool of families who he believed met
the criteria for the Experimental Group and then, matching as
closely as possible on the variables of student's grade level and
Stanford Achievement Test scores, selected a pool of Control Group
families. The names, addresses and telephone numbers of these
families were then provided to the researcher.

Both a letter describing the study and a cover letter written

by the high school principal were sent in the mail to those selected

54

families, (both letters appear in Appendix A). Within a few days

of mailing this introductory and informational letter, the researcher
contacted each family by telephone. The families were not told

about the specific basis for their selection. During the telephone
conversation it was determined if the family wanted to participate

in the study and an appointment time was scheduled.

Matching and Demographic Variables

 

The means and standard deviations of both the two matching
variables and the other demographic variables for both groups are
shown in Table 2. The descriptive variables were student's grade
point average (past two years of all classes taken in high school--
for 9th graders this included their course grades from the 8th
grade), family income, parent's education and the number of off-
Spring in each family. Data on these variables were obtained during
the experimental procedure, i.e., after the selection process was
completed.

A two-tailed t~test for comparing dependent samples was
performed on both the matching variables, (student's grade level
and Stanford Achievement Test scores) and on each descriptive vari-
able. With only one exception, the t-tests were not significant
(p > .05), thereby indicating sufficient similarity between the two
groups in these characteristics. The one exception was grade point
average (GPA). The Control Group had a significantly higher GPA
than the Experimental Group, E (9) = 3.00, p < .01. Finally, as

would be expected by the selection criteria, the Experimental Group

55

TABLE 2.—-Participant Variables.

 

 

 

 

 

 

 

Experimental Gropp Control Group
Mean SD Mean SD

Student's

Grade 10.1 .943 10.2 .978
SAT- a

English 45.2 24.8 49.6 27.4
SAT- a

Math 38.7 24.1 51.5 23.9
G.P.A. 1.77 .493 2.73 .300
Family b

Income 35.5 13.0 35.3 12.5
Parent's c

Education 27.4 4.36 28.4 4.54
Siblings 4.8 3.16 4.1 2.17

 

aPercentile rank on the Stanford Achievement Test.
bThe unit is dollars, multiplied by 1,000.

cThe combined total years of education f0r mothers and fathers.

56

had a significantly higher number of behavioral offenses at school

than the Control Group,_p (9) = 5.59, p_< .001.

Assessment Procedure

 

Each family met with the researcher at the high school for
one session lasting an hour to an hour-and-a-half. The researcher
knew to which group the families belonged, i.e., the researcher was
not blind as to group membership. At the beginning of the session
all participants were asked to sign a consent form and to fill out
the Family Information Sheet (both forms are in Appendix B).

Item 5 of the consent form released the student's educational
records for review by the researcher. The Family Inf0rmati0n Sheet,
also in Appendix B), provided data on family income, number of
siblings and the parents' educational background.

After these forms had been completed the TAFIT was then
administered. The administration of the TAFIT involved a two-stage
procedure in which 6 Thematic Apperception Test cards were first
individually presented to each family member, and then the same six
cards were presented to the family as a whole during the conjoint

administration.

Individual Administration. For the individually administered

 

stage of the TAFIT the following instructions were read to the
participants:

"You now have six cards with pictures of people engaged

in different activities. I would like you to look at

each picture, one at a time, and write down a story for
each picture. Write down a story about what is going on
in the picture. Make sure that each story has a beginning

57

(what led up to the picture), a middle (what is happen-
ing in the picture), and an ending (what happens after
the picture). There are no right or wrong stories, so
any story you want to make-up and write down is com-
pletely up to you. You can really let your imagination
have a good time as you create a story trying to describe
who the people are, what they are like, what are their
relationships like, how do the people behave and act and
what do the people think and feel. Please take about
five minutes for each story."

The researcher monitored the time and told the participants every
five minutes how much time had passed. At the 10 and 20 minute
points, the researcher announced which story the participants should
be working on. At 40 minutes the participants were asked to finish
their story writing. All participants completed writing the six
stories within 45 minutes, with approximately 80% of the partici-

pants completing their story writing within 35 minutes.

Conjoint Administration. The following instructions for the

 

conjointly produced stories were read to the family:

"I am now going to show you the same pictures again.
This time, though, I would like you, as a family, to
make up a story together. Try to make it a story that
the family as a unit can agree upon; a story you are
all involved in creating; one that originates and
comes from all of you. In other words, it should be
the (family's name) story. Again, the story should
have a beginning, a middle and an end. So, together
as a family let your imaginations have a good time as
you create a story trying to describe who the people
are, what they are like, what their relationships are
like, what do they think and feel and how do they
behave and act.

 

Every family goes about creating a family story in
their own unique way. It is important to remember,
though, that the story you make up should, as much as
is possible or reasonable, come from all of you.

58

This time the tape-recorder will be turned on (at this
point the tape-recorder was actually turned on) and so
when you are ready to tell me the story you have created
together, let me know."

TAT Card Selection and Rationale

 

The six TAT cards selected for this study are, in order of
presentation, 3BM, 68M, 78M, 88M, 98M and 13MF. The rationale for
choosing these cards stems from the combination of three sources:
Karon (1981), Hafner and Kaplan (1960) and Matranga (1976).

Karon (1981) presents a basic set of TAT cards which tap
significant dimensions/aspects of one's life. For clinical use he
suggests: Card 1 - childhood the way it was; Card 3BM - the most
pressing current problem; Cards 6 and 7 - relationship to parents;
Cards 4 and 13 - heterosexual adjustment. Thus it is reasoned
that these cards would tap the projected hostility content in
significant domains of the participant's life. Following this line
of reasoning, Card 98M was added to assess the projected hostility
that may be involved in peer relationships.

The scoring system for this study, to be discussed below,
is the Hafner-Kaplan Hostility Content Analysis (Hafner and Kaplan,
1960). In the development of this scoring system they used 22 TAT
cards, including 68M, 78M, 88M and 13MF. Card selection for the
present study was not limited solely to those cards used by Hafner
and Kaplan for the following two reasons: 1) an analysis of the
relationship between cards and scoring system was never undertaken
(personal communication), and 2) an examination of the scoring

system strongly suggests that it is a general scoring system and

59

not stimulus specific or stimulus bound. in other words, the
scoring system does not provide a rigid basis for card selection
since it can apparently be used to score hostile content originating
from a variety of experimental procedures. In fact, the develoPment
of this scoring system was undertaken simultaneously for use with
both the Rorschach and the TAT.

Finally, Matranga (1976) lists cards 38M, 88M, 12M and
13MF as four cards which usually elicit aggressive themes. This
is an important consideration since one dimension of the Hafner-
Kaplan scoring system is the distinction between overt and covert
hostility. It was reasoned that in order to include this as a
meaningful dimension, i.e., have the ability to differentiate the
two groups with respect to the type of hostility expressed, the use
of some cards which typically elicit aggressive themes is advisable.

A summary of the rationale for TAT card selection is as

follows:

3BM . assesses most pressing current problem (Karon);

elicits aggressive themes (Matranga)
_6BM assesses mother-son relationships (Karon); used

by Hafner and Kaplan

78M assesses father-sonrelationships (Karon used
by Hafner and Kaplan

88M elicits aggressive themes (Matranga); used by
Hafner and Kaplan

98M assesses peer relationships; used by Hafner and

Kaplan

13MF assesses heterosexual adjustment (Karon); elicits
aggressive themes (Matranga); used by Hafner and
Kaplan

6O

ScoringySystem

 

Both the individually and conjointly produced TAT stories
were scored for hostility content according to a system developed
by Hafner and Kaplan (1960). (Copies of the scoring manual pro-
vided by the authors, as well as the actual manual used in the
present study, which has slight modifications and additions, appear
in Appendix C and D). This system scores two aspects of hosti-

lity: a "weighted hostility” and an overt-covert hostility dimension.

Weighted Hostility. This dimension assesses the degree of

 

hostile content appearing in the stories. It is a scaled score
ranging from 0 (no hostile content) to a high of 4. The following
is a brief description of the criteria for assigning weights.

4 points: themes involving direct physical hostile acts
between people or towards the self

3 points: themes involving hate; thoughts, feelings,
dreams or threats of direct physical acts
between people; themes involving punishment,
permanent debilitating injury and death;
themes of direct physical hostile acts involv-
ing animals

2 points: themes involving verbal hostility; derogatory
descriptions of people; anti-social acts;
people forced by others to do things; hostile
or negative emotionality; rejections; illness
and accidents involving injury; destruction of
inanimate objects; predatory animals; destruc-
tive forces of nature; weapons

1 point: themes involving emotional deprivation; guilt
feelings; escape; misfortune; death symbols;
broken objects; the military

0 points: themes without hostile content

61

Each hostile theme is assigned a weighted score and therefore,

within any given story, multiple weights can be assigned.

Overt-Covert Scale. Each theme that was assigned a weight

 

was also judged as either overt hostility or covert hostility.
Overt hostility is hostility which is manifest and direct. Covert
hostility is hostility which is insidious, indirect, disguised or
latent. (Examples are provided in the scoring manual which appears
in Appendix C and D).

Hafner and Kaplan (1960) report the overall inter-scorer
reliabilities as .87 for the weighted scale, .76 for the overt
scale and .78 for the covert scale. All coefficients are signifi-
cant at the .01 level. Further, they report the intercorrelations
among these three scales as follows: Overt vs. Covert scale is
-.44; Overt vs. Weighted scale is .66 and Covert vs. Weighted

Hostility scale is .72.

Raters
The data was scored and rated by two undergraduates enrolled
in Psychology 490. The raters were blind as to the nature of the

study, the hypotheses and the group membership of the participants.

Training. The raters met together with the researcher for
approximately 15 hours of training. The training was conducted on
pilot data, i.e., TAT stories collected from non-participating

families. Training for each of the two scales took place

62

simultaneously, since agreement in terms of the overt/covert dimen-
sion was easily obtained once the actual weighted score was assigned.
The training emphasized familiarization and learning of the
scoring manual. The format for each training session followed a
similar course. Each rater would independently rate the same 10
to 15 stories. This was followed by a group discussion, emphasizing
the points of disagreement. Disagreement between the raters appeared
to be the result of carelessness, hostile themes in the pilot data
not specifically mentioned in the Hafner and Kaplan manual and lack of
consensus regarding the difference between a repetition of the same
theme, (which was not scored twice) compared to an elaboration of a
theme, which was scored separately. As arbitrary decisions needed
to be made, i.e., decision points not clearly spelled out in the
Hafner and Kaplan manual, the researcher made these judgements and
then appended onto the scoring manual the general decision rule
with the example(s) from the pilot data.
A test for the criterion level of inter-rater agreement,
i.e., 80%, was conducted at periodic points in the training. In
order to determine the percentage of agreement, each rater was
given the same 15 stories to score independently. When the percent-
age of agreement reached 84% for the Weighted Hostility scale, (97%

for the Overt/Covert Scale) the training was terminated.

Determination of Inter-rater Reliability

 

Since data analysis was performed on individual total scores

and family total scores, it is useful to think of the data in terms

63

of "story-packets", with each packet comprised of 6 stories. Each
packet was assigned a total score for a) weighted hostility, b)
number of overtly hostile themes, and c) number of covertly hostile
themes. There was a total of 80 story packets, i.e., 60 individually
produced packets and 20 family produced packets. A total of 65
story-packets were separately and independently scored by the raters,
with one rater scoring 32 packets and the other 33 packets. The
story packets were randomly assigned to each rater.

The remaining 15 story-packets, which were scored by both
raters independently, provided the data for calculating the overall
inter-rater reliability. The 15 packets were scored three packets
at a time by both raters at predetermined points, i.e., after each
rater individually completed 6 story-packets. Additionally, these
five occasions also allowed for a periodic check of the percentage
of agreement, which for the Weighted Hostility scale ranged from
74% to 96%.

On the basis of these 15 story-packets which were rated
independently by both raters, the inter-rater reliability coefficient
for the Weighted Hostility scale was 3 = +.83. The percentage of

agreement on the overt/covert dimension was 97%.

CHAPTER 3
RESULTS

General Considerations

 

The statistical test for each hypothesis, with the exception
of Hypothesis 3, is a two-tailed tytest for dependent/matched samples.
This is the method accepted after considering the possibility that
a co-variate type of analysis might be required, in order to control
for the influence of some independent variables systematically
related to the dependent variable. Two such variables were examined:
1) grade point average (GPA), and 2) the number of words in the TAT

stories.

GPA. As mentioned in the Methods section, the Control Group
had a significantly higher GPA than the Experimental Group, 3 (9) =
3.00, p < .01. However, calculating the correlation between the
GPA and dependent measures reveals that they are not related: a)
the relationship between GPA and the Weighted Hostility Score is.: =
-.20; b) the relationship between GPA and the Number of Overt Themes
is.[ = -.28, and c) the relationship between GPA and the Percentage
of Overt Themes is.[ = -.22. None of the three correlations, each
with N = 20, is statistically significant (p > .10). Therefore,
since GPA is not related to the dependent measures, a co-variate

type of procedure is not required.

64

65

Further, it seems on a rational basis that it would be an
error to control statistically for the difference in GPA between the
two groups. At the high school from which the sample was drawn the
students are assigned a single course grade, i.e., there is no
separate evaluation of work habits, cooperation or conduct. Conduct
or behavioral considerations are taken into account in the assign-
ment of course grades, from which the GPA was computed. Logically,
the number of behavioral offenses and GPA are related and one would
expect the Experimental Group, with many more behavioral offenses,
to have a lower GPA. In fact, the correlation between GPA and the
number of behavioral offenses, (N = 20), was statistically signifi-
cant, p = -.79, p < .01. Therefore, to control for the effect of
GPA on the dependent measures would in fact remove the effect of the
independent variable "Number of Offenses", i.e., the variable which
defines the two groups.

Additionally, given the method of assigning course grades
at the high school from which the sample was drawn, it appears that
the Stanford Achievement Test (SAT) scores probably provide a purer
or more accurate index of academic achievement or ability. As
detailed in the Methods section, the two groups do not differ signi-
ficantly in terms of their SAT scores in both English and Mathematics.
Finally, the SAT scores in English and Mathematics, respectively,
are not significantly correlated to any of the dependent measures.
Thus, academic achievement or ability, whether defined from the
point of view of GPA or SAT scores, is not significantly related to

any of the dependent measures.

66

Number of Words in the TAT Stories. An issue raised in the

 

literature, e.g., Winter, Ferreira and Olson (1966), is that the
longer the TAT story the greater the chance or opportunity for
hostile (or other) themes to be introduced into the participant's
stories. Thus this "verboseness" is a potentially important variable
that needs to be examined in terms of a co-variate type of analysis.
In the present study, however, the number of words produced in the
TAT stories is not related to the Weighted Hostility Score; for the
entire sample, (mothers, fathers, sons - N = 60) the correlation
between Weighted Hostility and the number of words in the TAT stories
was not significant, p = .14, p > .10. The same results occur when
examining the correlations just for the sons' stories (N = 20,_: =
.14, p_> .10) or the parents' stories, (N = 40,_p = .26,_p > .10).
The relationship between the number of words and the Weighted Hosti-
lity Score for the family produced stories is also not significant,
.2 = .01, p > .10. Additionally, the number of words actually pro-
duced in the TAT stories did not differentiate the two groups.
Therefore, since the variable "number of words in the TAT stories”
did not differentiate the two groups and is not related to the depen-

dent measure Weighted Hostility, a co-variate analysis is not

required.

Organization. What follows is a report of the statistical

 

results for the six hypotheses. Hypotheses 1, 2 and 3 involve
statistical comparisons between the two groups of adolescent sons,

while Hypotheses 4 and 5 involve comparisons between the two groups

67

of families and Hypothesis 6 is concerned with comparisons between
the parents of adolescents with behavioral problems and the parents
of normal adolescents. The following presentation will adhere to

this grouping.

Compgrisons Between Adolescent Sons

 

Tables 3 and 4 present the results for Hypotheses l, 2 and
3. These three sets of hypotheses involve predictions about the
content of the TAT stories produced by adolescents with behavioral
problems compared to the stories produced by adolescents without

behavioral problems.

Hypothesis 1

 

The prediction (Hypothesis l-A) that the adolescents with

 

behavioral problems would have a significantly different amount of
weighted hostility (WH) from the normal adolescents was not supported,

.3 (9) = .25 p > .05. Similarly, the prediction (Hypothesis l-B)

 

that there would be a significant difference between the two groups
in terms of the Percentage of Overtly Hostile Themes (% Overt) was

not supported, p_(9) = 1.053, p > .05. Finally, Hypothesis l-C was

 

not supported since the two groups did not differ significantly in
the Number of Overtly Hostile Themes (# Overt) produced, 3 (9) =
.913, p > .05.

Hypothesis 2
This set of hypotheses involve a comparison between each

son's score, on a given dimension, and the arithmetic mean of his

68

TABLE 3.--Comparisons Between Adolescent Sons' and Average Parents'

Scores

 

Experimental Group

 

Control Group

 

 

 

 

0"

p_ <.02
<.05
<.lO

“112%“

2 (.001

*(Mother's score + Father's score) / 2

Mean SD Mean SD t-Scorea
.HH 25.1 14.22 24.85 9.62 .05
% Overt 41.37 15.32 34.26 18.85 1.053
# Overt 4.75 3.76 3.8 2.44 .913
# Overt
Parents .725 .596 b
Sons 4.75 3.76 ............................ 3.163
Parents .975 .884 c
Sons 3.8 2.44 '''' 2’86]
HH
Parents 15.26 4.05 d
SOnS 25.1 14.22 oooooooooooooooooooooooooooooo 10883
Parents 15.42 3.75 c
Sons 24.85 9.62 °°°°°° 2'562
% Overt
Parents 8.84 7.62 5 78e
Sons 41.37 15.32 .............................. .
Parents 13.95 14.58 2 12d
Sons 34.26 18.85 """" '
a2-tailed t-test for dependent samples (n = 10; df = 9)

69

TABLE 4.--Test of the Difference Between Two Variances for Matched
Groups. Summary of the Deviation-score X Group ANOVA.

 

 

 

 

Source pf. MS F*
Influence Score
# Overt
Deviation Scores 1 5.11 1.355
(Between)
Matched Pairs 9 6.33
Group X Pairs 9 3.77
(Error)
Influence Score
WH
Deviation Scores 1 - 82.83 1.076
(Between)
Matched Pairs 9 78.55
Group X Pairs 9 76.91
(Error)
Influence Scores
% Overt
Deviation Scores 1 418.43 3.64a
(Between)
Matched Pairs 9 80.44
Group X Pairs 9 114.77
(Error)

 

*None of the 5 ratios are statistically significant at the .05 level.

ép < .10

70

parents' scores, e.g., (mother's score + father's score) / 2. The
hypotheses predict differences between the behavioral problem sons'
scores, on a given dimension, and the arithmetic mean of their
parents' scores, and that these differences will not exist within
the normal families. (The results relevant to these hypotheses

appear on Table 3.)

.3:5' The prediction that the adolescents with behavioral
problems would produce a significantly different # Overt Themes from
the arithmetic means of their parents' # Overt Themes was supported,
.3 (9) = 3.163, p < .02, but the prediction that this difference
would not exist for the Control Group was not supported, i.e., the
normal adolescents also produced a significantly different # Overt
Themes from their parents' average production, 3 (9) = 2.866,

.p < .05.

Thus, both groups of adolescent sons did produce a signifi-
cantly greater # Overt Themes than did their parents. A post hoc
analysis was undertaken to compare the two groups of adolescents
directly. In other words, this analysis was undertaken in order
to determine if the "difference scores", i.e., (son's # Overt Themes)
- (arithmetic mean of the parent's # Overt Themes), for the adoles-
cents with behavioral problems was significantly different from the
normal adolescents' "difference scores." The analysis indicates
that the "difference scores" are not significantly different for

the two groups, p (9) = 1.0978, p > .05.

71

2:8, This hypothesis, which cannot be accepted, predicted
that the adolescents with behavioral problems would have a signi-
ficantly different WH score from the arithmetic mean of their par-
ents' WH score, and that this difference would not exist for the
Control Group. In fact, the results were reversed. The WH scores
of the adolescents' with behavioral problems did not differ signi-
ficantly from their parents' average WH score, 3 (9) = 1.883, .10
> p'> .05, but the Control Group sons' scores did differ signifi-
cantly from their parents' average WH score, t_(9) = 2.562, p > .05.

Thus, both groups of sons had a higher WH score than the
arithmetic mean of their parents' WH scores, but this difference
was statistically significant only for the Control Group. Again, a
post hoc analysis was performed in order to compare the two groups
directly. The analysis indicates that the "difference scores”, i.e.,
(son's WH score) - (arithmetic mean of their parents' WH scores),
are not significantly different for the two groups, t_(9) = .069,

'p > .05. An examination of Table 3 reveals the similarities in
Means for both a) the two groups of adolescent's WH scores, and

b) the two groups of parents' WH score. Thus, the Mean of the "dif-
ference scores“ for the two groups are all but identical; the Mean
"difference score" for the Experimental Group is 9.83 and for the
Control Group is 9.43. Yet the Control Group differed significantly
from their parents, while the Experimental Group does not. What
accounts for the statistical difference in the Control Group is a
lower variance in the "difference scores" (§Q_= 11.04) compared to

the variance in "difference scores" for the Experimental Group

 

72

(_D = 15.67). However, a comparison of variances between the two
groups of "difference scores” reveals that the variances are not

statistically different.

Iglg. The results indicate that this hypothesis can be
accepted. The results support the prediction that the adolescents
with behavioral problems would exhibit a significantly different
% of Overt Hostility from the arithmetic mean of their parents' %
of Overt Hostility, t (9) = 5.78, p_< .001. As predicted, the
Control Group did not exhibit this difference, p (9) = 2.12, .10 >
p_> .05.

Thus, while both groups of adolescents exhibited a higher
% of Overt Hostility than the arithmetic mean of their parents' %
of Overt Hostility, this difference was statistically significant
only for the Experimental Group. Once again, a post hoc analysis
was performed in order to compare the two groups directly. This
analysis indicates that the "difference scores", i.e., (son's %
Overt Hostility) - (arithmetic mean of their parents' % of Overt

Hostility), are not significantly different for the two groups,

.3 (9) = 1.19, p_> .05.

Hypothesis 3

 

This set of hypotheses predicts differences, between the two
groups of adolescents in the variability of their influence scores.
A common statistical procedure to test for differences in variance

2

between two groups is $2 / S =.[. This test of variance, however,

is very sensitive to the assumptions of independence, which the

73

present study does not meet. Thus, the procedure employed to analyze
this hypothesis, one that is reasonably robust, is as follows: for
each group, the mean of the influence score was calculated. Then,
the deviations from the mean, in absolute values, were used in a
one-way ANOVA, i.e., I = influence score; Mi = groups mean of the
influence scores; 0 = / I - Mi / and it was the 0 scores that were
used in the ANOVA (for matched pairs) calculations.

As shown in Table 4, there were no significant differences
between groups in the dispersions of the influence scores. Hypoth-
esis 3-A cannot be accepted since the ANOVA did not indicate a signi-
ficant difference in the dispersion of the influence scores for the

# Overt Themes, E (1, 9) = 1.355, p > .05. Similarly, Hypothesis

 

§;§ cannot be accepted as the ANOVA did not reveal a significant
difference in the dispersion of the influence score for WH,_[ (l, 9)

= 1.076, p_> ,05. Finally, the prediction (Hypothesis 3-C) that the

 

two groups would have a significantly different dispersion in the
influence scores for % Overt Hostility was not supported, 5 (l, 9)
= 3.64, .10 > p_> .05. The Experimental Group had a greater vari-
ability of influence scores for % Overt Hostility, SD = 21.30 than
the Control Group, SD = 10.99.

Comparisons Between the Families

 

Table 5 shows the results for Hypothesis 4 and 5. These

hypotheses deal with comparisons between the two groups of families.

74

TABLE 5.--Comparisons Based on Family Stories.

 

 

 

 

Experimental Group Control Group
Mean SD Mean SD t-Score
Family's WH 19.15 3.46 23.05 4.91 - 2.03a
Family's a
% Overt 32.83 16.7 20.87 20.03 1.867
Family's
# Overt 2.9 1.51 2.15 2.0 1.23
b
PFS-Actual
WH 5.5 3.67 5.36 3.94 - .08
# Overt 1.56 .892 1.26 1.52 .155
% Overt 16.04 15.81 14.03 12.96 .346

 

a.10 > p_> .05

bThis is the difference score, in terms of absolute value, between

the Predicted Family Score and the actual family score.

75

Hypothesis 4

 

This hypothesis predicts that the family stories produced
by those families with an adolescent with behavioral problems will
have different content from those stories produced by normal fami-

lies. The prediction (Hypothesis 4-A) that the family stories pro-

 

duced by those families with an adolescent with behavioral problems
would have a significantly different amount of WH from the family
stories produced by normal families was not supported, a (9) = -2.03,
.10 >.p > .05. (The Control Group had a higher WH score than the

Experimental Group.) Contrary to what was predicted (Hypothesis

 

3:8), the two groups of family-produced stories did not differ in
the % Overt, t_(9) = 1.867, .10 >ip > .05. (The Experimental Group
had a higher % Overt than the Control Group.) Finally, the two
groups of family-produced stories did not differ significantly in

the # Overt Themes, t_(9) = 1.23, p > .05.

Hypothesis 5

 

This set of hypotheses predicted that the difference score,
in terms of absolute value, between the Predicted Family Score (PFS)
and the actual family score, would be a variable on which the two

groups would significantly differ. Hypothesis 5-A cannot be accepted

 

as there was not a significant difference between the two groups of
families in this difference score for WH, (i.e., WH PFS - Actual
Family WH), t_(9) = -.08, p > .05. Similarly, this difference score

for # Overt Themes (Hypothesis 5-8) was not significantly different

 

for the two groups, t_(9) = .155, p > .05. Finally (Hypothesis 5-C),

 

76

the two groups of families did not differ significantly in terms of

the difference score for % Overt, p (9) = .346, p_> .05.

Comparisons Between Parents

 

Hypothesis 6

 

Table 6 shows the results of the §:test comparisons made for

the two groups of parents. Hypothesis 6-A cannot be accepted since

 

the parents' score, i.e., (father's score + mother's score), for NH

was not significantly different, t_(9) - .190, p_> .05, for parents
of behavioral problem adolescents as compared to controls. Neither

can Hypothesis 6-B be accepted since the two groups of parents did

 

not differ in terms of the # Overt, t_(9) = -.0670, p > .05. Finally,
the parents' combined score for % Overt did not significantly dif-

ferentiate the two groups, (Hypothesis 6-C), p (9) = -.828, p > .05.

 

Comparisons between the fathers did reveal a significantly
different # Overt, with the fathers of normal adolescents producing

a greater number of overtly hostile themes, (Hypothesis 6-D), p (9)

 

= -2.273, p < .05. It is important to point out that the actual
number of words used in the TAT stories produced by the fathers of
behavioral problems adolescents was not significantly different from
the number of words used by the Control Group fathers, t_(9) = .676,
.p > .05. Contrary to the prediction of Hypothesis 6-E, the two

 

groups of fathers did not differ in terms of their WH score, a (9)

= -1.41, nor their % Overt (Hypothesis 6-F) p (9) = -l.53, p (for

 

both prscores) > .05. Thus, while the differences are not statisti-

cally significant, the Control Group fathers produced both a higher

77

TABLE 6.--Comparisons Between Parents

 

 

Experimental Group Control Group

 

 

 

 

Mean SD Mean SD t-Score

Parentsa

WH 30.55 8.09 29185 7.55 .190
# Overt 1.45 1.19 1.95 1.77 - .670
% Overt 17.68 15.25 27.89 29.1 - .828
Fathers

WH 12.5 6.0 16.65 6.7 -l.4l

# Overt .4 .49 1.45 1.31 -2.273b
% Overt 6.78 10.30 22.03 28.07 -l.53
Mothers

WH 18.05 8.05 13.2 4.07 1.92b
# Overt 1.05 1.15 .5 1.02 1.632
% Overt 10.91 10.81 5.86 11.72 .977
aThis is the combined mother and father score.

Pp < .05

78

WH score and a higher % Overt Themes than did the Experimental
fathers.

The predictions that the mothers of the adolescents with
behavioral problems would produce different TAT stories from that
produced by the normal mothers were not supported. The two groups
of mothers did not produce significantly different WH scores,

Hypothesis 6-G, t_(9) = 1.92, .10 > p_> .05. Even though the two

 

groups of mothers did not differ significantly, the mothers of
behavioral problems adolescents did produce higher WH scores, on
the average, than the Control Group mothers. The two groups of

mothers did not differ in either the # Overt Themes (Hypothesis 6-H)

 

.p (9) = 1.632, with the Experimental mothers producing a greater #
Overt than the Control Group mothers; nor did they differ in terms

of % Overt, (Hypothesis 6-1), p (9) = .977; for both_p—scores_p >

 

.05.

CHAPTER 4
DISCUSSION

In general, the a priori predictions about the differing
patterns of fantasy content between the two groups of families did
not gain statistical support. In attempting to account for this
state of affairs, the present section will also attempt to explore
two related issues: 1) Is the lack of statistically significant
results a strong statement against the general theoretical frame-
work from which the hypotheses were generated?; 2) While the a priori
predictions cannot generally be accepted, do the data nevertheless
I'fit" the general theoretical model? Given the exploratory nature
of this study, i.e., viewed as a ”pilot study" in size and scope,
it is legitimate to examine any trends and patterns actually exhi-
bited in the data that are of sufficient magnitude to suggest the
usefulness of further research in this area.

One important consideration relevant to the present study
is the issue of statistical power, i.e., the ability to reject the
hypothesis of no difference. The present study appears to have had
a lack of statistical power due, to a large extent, to the small
sample size. Given the actual differences in group means found
in the present study, sometimes referred to as the obtained "effect
size", Table 7 indicates what sample size, with N = matched family
pairs, would have been needed to have reached statistical

79

80

TABLE 7.--The Number of Family Pairs Needed to Have Significant
Results Given the Obtained Effect Size*

 

Comparisons Between Sons: .H**
WH 15366
% Overt 36
# Overt 47

Comparisons Between Sons 8
Average Parents' Score:

Experimental-WH 13

Control - % Overt ll
Comparisons Between Families:

WH 12

% Overt l3

# Overt 25
Comparisons Between PFS-AFSa:

WH 6002

# Overt 466

% Overt 163
Comparisons Between Combined
Parents:

WH 1064

# Overt 87

% Overt 58
Comparisons Between Fathers:

NH 21

% Overt 18
Comparisons Between Mothers:

WH 13

# Overt 17

% Overt 47

 

*The difference in group means obtained in this study.

**N is the number of matched pairs. Thus an N = 15 refers to 15
families in each group for a total of 30 participating families.

aThis is the Predicted Family Score - Actual Family Score.

81

significance for each hypothesis not actually supported in this
study. Clearly, in order to have certain hypotheses supported the
number of participants needed in the study would have been prohi-
bitive to obtain. 0n the other hand, 8 hypotheses would have

been supported if the number of matched family pairs had been 25.
This strongly suggests that certain trends do in fact exist in

the data and that these trends might have reached statistical
significance if the sample size was larger, yet still within reason-
able and acceptable limits.

For purposes of participant selection, how adequate was the
operationalized definition of "behavioral problem adolescents“? In
the present study the independent variable Number of Behavioral
Offenses (# Offenses) found in the school records was used opera-
tionally to define and discriminate the two groups of adolescent
boys. Correlations between # Offenses and each of the three depen-
dent measures yielded the following relationships (N = 20): with
WH, : = .316 (p > .10); with % Overt, p = .438 (p§< .06) and with
# Overt, p = .491 (p_ <.05). Not all the correlations are signi-
ficant, but this is due, in part, to the limited range of # Offenses
for the Control Group, which was 0-1. In fact, when the same cor-
relations are performed just for the Experimental Group, the rela-
tionship between # Offenses and the dependent measures are as
follows, (N = 10): with WH, p = .581 (pr: .09): with % Overt, p
= .761 (p< .02) and with # Overt, I: = .733 (p< .02). These corre-
lations support the relationship between behavioral indices, e.g.,

# Offenses, and TAT content, especially the relationship with %

82

Overt, and generally speak favorably about the appropriateness and
adequacy of the operational definition.

The above correlations also suggest that if the # Offenses
selection criterion for the Experimental Group had been higher, this
would have systematically enhanced the differences between the two
groups of adolescents on the dependent measures. If, for instance,
the criterion level for group membership had been 6 instead of 3
behavioral offenses, this would have systematically raised the
Experimental Groups' scores on the dependent measures, and hence
would have made the two groups of adolescents more distinct. In
fact, a post hoc analysis was performed on the % Overt score using
the three adolescent boys having 6 or more behavioral offenses and
their matched Control participants. This was a post hoc re-evalua-
tion of Hypothesis l-B, which originally indicated that the Experi-
mental Group exhibited a higher % Overt hostility, but the difference
was not significant. The post hoc analysis, however, using only
three Experimental participants, did reveal a significant difference,
_t_ (2) = 8.439, p < .02. While this can obviously be a spurious
finding, it nevertheless supports the position that the lack of
statistically significant differences between the two groups of
adolescents may have been the result of not using a more stringent
criterion for selecting the Experimental Group with behavioral
problems.

The way in which the operational definition of behavioral
problems, or "acting-out" behavior, is dealt with lies at the heart

of the answer as to whether the lack of significant results is a

83

statement against the theoretical model. The theoretical model
states that the adolescent's behavioral problems can be understood
from a family-dynamics systems perspective. It is assumed that to
the extent that one can examine a population in which the behavioral
symptom is exaggerated or extreme, then this should correspond to
an exaggeration of the hypothesized underlying family dynamics.
The analogy of magnification is useful; the greater the magnifica-
tion, the greater the likelihood of detecting a real phenomenon. A
review of those families contacted (both parents in the home and the
son having 3 or more behavioral offenses) but refusing to partici-
pate is enlightening. The range of behavioral offenses for those
16 non-participating adolescents was 3-17, with the mean number of
offenses = 8.5 (SQ = 4.1). A one tailed p-test reveals that these
adolescents had a higher number of behavioral offenses than the
actual Experimental adolescents, p (24) = 1.96, p (one-tailed) <
.05. This suggests that the Experimental Group in the present study
may in fact exhibit, on a continuum, only "moderate" behavioral
problems, while the group of "non-participants" would have ideally
been more appropriate for testing the hypotheses. On the other
hand, to the extent that the patterns and trends in the present data
with a sample of ”moderate behavioral problems" do conform to the
theoretical model, then this can be viewed as strongly suggestive
evidence in favor of a dynamic family-systems perspective.

Quite clearly, to the question as to whether the lack of
statistical results is a strong statement against the general theo-

retical framework from which the hypotheses were generated, the

84

answer is ''no", not from this single study. The present study
suffered from small sample size, and very likely a lack of
sufficient discrimination between the two populations studied.
Futhermore, the trends that are evident in the data do closely
follow patterns logically predicted and expected from the theore-
tical model. Thus, the remainder of this chapter will examine and
review these trends, in order to detenmine if they are persuasive
to the point of suggesting further research.

A useful starting point to begin the examination of the data
is with a review of the hostility content exhibited in the family-
produced stories. As can be inferred from Table 7, (as well as
Table 5 in the Results section), the differences in both the WH and
% Overt scores between the two groups of family-produced stories
approach significance and the trends and patterns of these scores
are worth noting. The data indicate that the actual amount of
hostile fantasy content produced by normal families is higher than
that produced by families with an adolescent with behavioral prob-
lems. However, the % Overt Hostility was greater for the Experi-
mental families than for the normal family units. It appears that
the hostility expressed within the Experimental families is more
direct, overt and perhaps "primitive". Thus, the expression of
aggression for these families is more dangerous since the fantasy
is loaded with direct forms and meanings, exhibiting less control
over the hostile tendency/impulse. Therefore, it is logical that
such aggression needs to be more strongly guarded against, defended

or suppressed, and the lower WH score can be taken as evidence of

85

such forces operating within the Experimental families. Overall,
the content of the family-produced stories can be seen as supportive
of the basic theoretical position: The families with behavioral
problem adolescents are in fact conflicted about the expression of
hostile/aggressive tendencies, and this conflict is generally dealt
with by a suppression of hostile thought by the family unit. The
data suggests that these families may suppress aggression since the
families' aggressive tendencies are more loaded with direct forms
and meanings.

It is important to note that the results of the family-
produced stories are contrary to what has been reported previously.
Winter, Ferreira and Olson (1966) report that the family-produced
stories of their male "Delinquent“ group had a higher WH than "any
other group", but their % Overt was "close to that of the normals";
the present results are a reversal of this finding. There are a
variety of procedural differences between the two studies which make
direct comparisons or contrasts between the discrepant results diffi-
cult. One procedural difference worth noting, however, is the
different stimulus presentation. While the present study attempted
to elicit aggressive themes, as a way of assessing how the family
deals with hostility, this was not the intent of the Winter et a1.
study. In fact, they presented only one TAT card which typically
elicits hostile themes, as well as using a variety of "GF" cards.

It is reasonable to suggest that the use of the opposite gender
TAT cards allows for less identification, or a distancing from the

stimulus. If we assume that the experimental families, in both

86

studies, are in fact conflicted around the expression of aggression,
then the Winter et a1. stimulus presentation does not activate or
elicit the conflict with the same intensity or purpose as in the
present study. Thus, it seems plausible to speculate that this
important difference in the stimulus content may account, at least
in part, for the differing patterns of scores exhibited in the
family-produced stories.

Two other hypotheses, 3 and 5, do provide an indirect avenue
for examining the family-produced stories. Unfortunately, the
results of both these hypotheses do not lend themselves to clear
interpretation. For example, Hypothesis 5 predicted that the dif-
ference between the Predicted Family Score (PFS) (i.e., (mother's
score + father's score + son's score) / 3), and the Actual Family
Score (AFS) would be less for the Control Group families. The goal
of this hypothesis was to examine whether the sum of hostile ten-
dencies within the Normal families was equally partitioned or shared
by the individual members, and thus the average of their individual
productions, e.g., the PFS, would be a good predictor of the AFS.

It was reasoned that if the behavioral problem adolescent was in
fact carrying the hostile tendencies for his family, then the PFS
would be a misleading indicator or predictor of the AFS. Underlying
this speculation was the assumption that the family-produced story
represented for the Normal families an averaging of the three indi-
vidual inputs, but this averaging process would not consistently
take place within the Experimental families since the son's input

might be rejected or barred from entering the family-system. The

87

results clearly do not support this speculation and yet this lack of
support is not a clear statement about the theoretical model's vali-
dity. What appears to be in error is the assumption of a differing
averaging process, or even perhaps the overall assumption that the
family-produced story represents additive processes. It remains a
possibility that the adolescent with behavioral problems is the
carrier of split-off aggressive tendencies, but the averaging pro-
cess underlying the family-produced stories remains the same. In
other words, the Experimental son's input, while more hostile due to
his "unique" or "special" role within the family, is not in fact
rejected, but rather accepted into the family-produced story. Hence,
the ability of the PFS to predict the AFS would be the same for the
two groups of families.

Additionally, Hypothesis 3 does not lend itself to a clear
test. This hypothesis is concerned with the son's "influence score",
i.e., son's score - family score. The actual results are difficult
to interpret since the data were actually difference scores and
I'influence" can only be indirectly inferred. It is interesting to
note, however, that the dispersion/variability of the "influence
scores” for % Overt was greater for the adolescents with behavioral
problems than for the Normal adolescents (p < .10). Furthermore,
the normal adolescents produced, on the whole, lower "influence
scores". To interpret these data as supporting the position that
normal adolescents generally had a greater influence on the family-
produced stories might very well be erroneous. What is clear is

that the Control Groups' % Overt scores were on the average closer

88

to their family's % Overt score. The data do indicate that the
behavioral problem adolescents, when compared to the normals, exhibit
a different relationship to their family's stories in terms of the

% Overt. This does closely fit the theoretical model which suggests
that the experimental sons have a special role within the family;
they are the expressors of the family's aggression, i.e., % Overt.

In a very general sense, the data on the "influence scores" for %
Overt can be seen to support this notion, but unfortunately the
results do not aid in providing a clearer understanding of the

family dynamics.

In the present study a variety of comparisons were made
between the adolescents with behavioral problems and the "normal"
adolescents. In general, the adolescents' TAT content was similar,
with the behavioral problem adolescents producing a noticeably higher
percentage of Overt Themes. Additionally, there were comparisons
made between the adolescent's score and his parents' average score,
e.g., son's score - (mother's score + father's score/2). Both
groups of adolescents produced higher scores, on all three dependent
measures, than their parents' average score. Some of these differ-
ences were significant, others were not. But the differences between
the son's and parents' average that were not significant, were all
approaching significance. Further, direct comparisons between the
difference scores reveals that they were very similar for the two
groups of adolescents. Thus, the suggested conclusion is that all
adolescents produce a greater amount of hostile TAT content than

their average parents' production, and the dynamics involved in the

89

family of behavioral problem adolescents do not significantly alter
this state of affairs.

On the other hand, using the "parents' average production'I
can be misleading. The "parents' average production" is a gross
measure which in fact obscures the real underlying dynamics, especi-
ally within the Experimental families. What is lost or obscured
with the "average production" score, or even the "combined parents'
score", is the individual parent's contribution to that score. It
is only by simultaneously examining the comparisons between fathers
and mothers, fathers and sons and mothers and sons, that the adoles-
cents' problem behavior becomes understandable from a family-dynamics
perspective. In other words, support for the psychological inter-
dependence between family members begins to emerge clearly when the
focus is on the father-mother-son triadic relationship.

Table 6 (p. 77) presents the means for the two groups of
combined parents' scores (e.g., mother's score + father's score).
All the means look very similar except for % Overt, which is never-
theless not significantly higher for the Control parents. (In fact,
if the "outlyer" score is removed, see Table 8, then the combined
% Overt mean for the Control group is 19.22, which again is very
similar to the Experimental parents.) An interesting pattern begins
to emerge from the data when examining the mothers' and fathers'
respective contributions to this combined score. On the three mea-
sures where the Experimental fathers have a lower score than the
Control group, the Experimental mothers have a higher score than the

Control mothers. This pattern suggests that the Experimental

90

TABLE 8.--Percentage of Overt Themes for the 60 Participants

—-

Experimental Families

 

 

 

Son Mother Father
1. 62.5 11.1 33.3
2. 16.66 33.33 11.11
3. 42.86 23.53 0.
4. 44.44 0. 9.09
5. 36.37 7.69 O.
6. 44.44 0. O
7. 38.5 0. O.
8. 35.7 20.0 0.
9. 22.2 10.0 0.
10. M .239. M
Mean = 41.37 10.91 6.78
Sp.= 15.32 10.81 10.30
Control Families
11. 46.15 0. O.
12. 10.0 0. 25.0
13. 18.18 30.0 0.
14. 25.0 0. lO0.0**
15. 70.57 0. O.
16. 16.66 28.57 32.14
17. 46.2 0. 14.3
18. 15.38 0. 10.0
19. 44.44 0. 16.66
20. gHLjL. ._Q;__ 22.22
Mean = 34.26 5.86 22.03**
§Q,= 18.85 11.72 28.07

 

**When the 100% Overt score is removed, the new group mean for
Control Fathers (N = 9) is 13.36 and the SD = 11.18.

91

mothers' and fathers' scores are systematically related; this is
the first clear indication of a reciprocal interdependence in
psychological functioning.

This position is further advanced by examining Table 8.
The following observations are worth noting: a) the mean % Overt
for the Experimental fathers is close to that of the Control mothers;
b) the pattern of actual scores reveals that 6 out of 10 Experimen-
tal fathers do not produce a single overtly aggressive theme; (the
only group with a greater absence of overt hostility is the Control
mothers); c) the Normal fathers % Overt score is 3 times greater
than the Experimental fathers score and even with the "outlyer"
datum removed, their % Overt score is about twice that of the Experi-
mental fathers, and d) the Experimental mothers' mean % Overt score
is about twice that of the Control mothers. Additionally, five
Experimental mothers produce more overt hostility than their hus-
bands, and this occurs only once in the Control group. A post hoc
analysis comparing the mother - father difference scores fer the
two groups reveals that the difference between husband and wife is
significantly greater for the Experimental group, t_(9) = 2.039,
p_< .05 (one-tailed). Finally, it is important to note that the
above observations, as well as the remainder of this chapter,
emphasize the patterns and results in tenns of % Overt scores since
the reciprocity in content is most clearly exhibited on this dimen-
sion. However, the patterns are similar, though less striking,

for the other two dimensions. For example, the Experimental fathers

92

as a group produced the lowest WH score and their # Overt themes
is practically the same as the Control mothers.
In our society the male is stereotypically more aggressive
than the female. This appears to be part of the socialization pro-
cess by which we are all inevitably influenced. It is the male who
is more overtly aggressive and this is generally accepted within our
society. This stereotypical pattern is in fact clearly exhibited "

within the husband-wife relationship in the Normal families. But

in the families with a behavioral problem adolescent this pattern
is not exhibited. As a group, the Experimental fathers appear to
be inadequate or deficient in the expression of overtly hostile
themes. Similarly, their lower WH scores strongly suggest the denial
or repression of any hostile content from their fantasies, or intra-
psychic world. Together, this strongly suggests that in the Experi-
mental families it is the fathers who are conflicted around issues
of aggression or hostility and have difficulty in the expression
or release of these tendencies. It seems reasonable to postulate
from the above observations that the mother/wife compensates for
her husband's intrapsychic conflict and this can be viewed as con-
sistent with Dicks' (1963) position that within the marital rela-
tionship the two partners engage in a trade-off, with each partner
colluding to carry a split-off part for the other. The pattern of
the TAT content clearly indicates that the intrapsychic worlds of
the husband and wife are interdependent.

The essential question of this study still remains unanswered:

Is the Experimental sons' behavior at school related to, or the

93

result of, the family dynamics? Is there evidence to suggest that
the school behavior is related to the fathers' proposed intrapsychic
conflict? The data, while only suggestive, strongly point towards
an affirmative answer. A post hoc comparison of sons' and fathers'
% Overt scores reveals a significant difference for the Experimental
participants, a (9) = 8.03,_p < .002; while the Normal sons also
have a higher % Overt score than their fathers, this difference is
not significant, 3 (9) = .096, p_> .05. A direct compairson between
these difference scores, (sons' - fathers' scores) shows that the
adolescents with behavioral problems differ significantly more from
their fathers' % Overt scores than do the Normal adolescents, p (9)
= 1.728, p < .05 (one-tailed). Furthermore, all of the Experimental
sons had a higher % Overt score than their fathers, while only 7

of the 10 Normal sons had a higher score than their fathers. Thus,
not only do the Experimental adolescents differ significantly more
from their fathers, but their difference scores also exhibit signi-
ficantly less variability, E (l, 9) = 5.775, p < .05. To highlight
the meaning and importance of these patterns, the same comparisons
between sons and mothers were conducted. Both groups of adolescents
produced a significantly higher % Overt themes than their mothers
and the son - mother difference scores were roughly similar. Hence,
it appears that in the % Overt score, the Experimental sons' rela-
tionship to their mothers is no different than the Normal adolescents.
This is not the case, however, in the Experimental sons' relationship
to their fathers. The data clearly indicate that in terms of the

proportion of overtly hostile themes produced, the adolescent with

94

behavior problems has a unique and different relationship to his
father that is not exhibited by the normal adolescent. The differ-
ence is in the direction strongly to support the theoretical model.

In summation, the theoretical model states that there is a
psychological interdependence between family members. The son's
problem behavior at school was conceptualized via a dynamic family-
systems framework: It is proposed that the adolescent son is the
"carrier" of the denied or split-off aggressive tendencies of one
or both parents, and his behavior at school is an "acting-out" of
these tendencies. In the present study it was reasoned that if this
theoretical model is in fact accurate, then the individual family
members' production of hostile fantasy content, viewed as a repre-
sentative of intrapsychic functioning, should systematically follow
this theoretically proposed pattern.

Even though the interpretation is speculative, the data
exhibit patterns that would be expected from the theoretical model.
It seems reasonable to propose that it is the fathers of behavioral
problem adolescents who are psychologically conflicted about their
own aggressive tendencies. These fathers appear to deny and/or
repress their own aggressive thoughts, the content of which exhibits
a striking absence of hostility that is overt in nature. It remains
a logical possibility that fathers exhibiting this low production
of hostile fantasy content might live with and be related to wives
and sons who also have an unusually low level of aggressive fantasy
content. This pattern would be contrary to the theoretical model

for behavioral problem adolescents and in fact is not supported by

95

the data. Both the wives and sons of these fathers produce a higher
amount of overtly hostile fantasy content than their Normal counter-
parts. The adolescents with behavioral problems exhibit a unique
relationship to their fathers. Unlike the relationship between the
normal adolescents and fathers, these sons produce a significantly
greater amount of overtly hostile fantasy content than their fathers.
It is reasonable to interpret this as evidence that these sons are
compensating for their fathers and the content of both fathers' and
sons' intrapsychic world, as assessed through fantasy content,
clearly exhibits a reciprocal interdependence. What the father must
deny, disregard or disown possibly as a way to decrease his own
anxiety, is carried and expressed by his son. When viewing the data
from this perspective it becomes reasonable to suggest that the
son's behavior at school may also be a compensation for, or in fact
an "acting out" of the father's split-off aggressive tendencies.
Finally, the data strongly suggest that the mother also plays an
essential role in these dynamics. She too appears to be compensating
for and/or intrapsychically involved in the father's conflict.
Compared to the normal mothers, these mothers produced a greater
amount of overall hostile fantasy, with a larger proportion of that
being overt in nature. It is reasonable to suppose that these
mothers communicate permission to their sons for the expression of
hostility. All together, the evidence points favorably towards a
reciprocal interdependence in psychological functioning that effects
the intrapsychic domain f0r each individual in the triadic relation-

ship.

APPENDICES

APPENDIX A

96

Dear Parents:

I am conducting a study that examines the family and various pro-
cesses that take place within the family unit. As you have probably
noticed in recent newspaper and magazine articles, movies and tele-
vision shows, there is a growing interest in our society regarding
the increasing variety of family units and family life-styles. The
present study is being undertaken in the hope of providing further
understanding of family relationships. I am interested in exploring
the themes of creative stories produced by high school students, as
well as the themes of those stories produced by the student's par-
ents.

I, Jerry Adams, am a graduate student in Psychology at Michigan State
University. I will be working on this study under the close super-
vision of Dr. Lucy Ferguson, Professor of Psychology at Michigan
State.

I am writing this letter to personally ask you and your son to parti-
cipate in this study. The design of this study is quite simple and
straightforward: Each participating family, (husband, wife and high
school student), will meet with me for only one session lasting for
1-14 hours. This session will be scheduled at your convenience.
During the session I will ask each family member to create a few
stories based on some pictures and then ask the family to make-up
some stories together. As can be expected, all obtained information
will be held strictly confidential. The stories will be coded to
protect the individual identity of each participant.

The principal of the high school, has approved and endorsed this
study, the results of which may provide educators, parents, students
and psychologists with important information regarding the ways in
which a family is organized and functions. As soon as the study is
completed, the results and implications will be shared with all
participating families. Past experience suggests that participating
families will enjoy their story telling session. The opportunity

to participate in this study, together as a family, should prove to
be an interesting and rewarding experience.

I will be contacting you by telephone during the upcoming week to
determine if you are interested in participating and to schedule a
convenient time for our session together. I will also be able to
answer any further questions you might have regarding this study.

Thank you,

Jerry S. Adams

 

 

97

Dear Parents:

Mr. Jerry Adams, has asked for the cooperation of this high school
in his graduate research at M.S.U. He has picked a sample of fami-
lies, including yours, to interview as part of his study. We would
like to cooperate with him as much as possible, and while you cer-
tainly are not obligated, we would appreciate your help.

Please feel free to call me at school if you have any questions or
concerns.

Thank you,

Principal

APPENDIX B

 

98

CONSENT FORM

We have freely consented to participate in a scientific study
being conducted by Jerry S. Adams, under the supervision of
Dr. Lucy Ferguson, Professor of Psychology at Michigan State
University.

The study has been explained to us and we understand the explana-
tion that has been given and what our participation will involve.

We understand that we are free to discontinue our participation
in the sutdy at any time without penalty.

We understand that the results of the study will be treated in
strict confidence and that we will remain anonymous. Within
these restrictions, results of the study will be made available
to us at our request.

We agree to let Jerry Adams review the school record of our son
in order to obtain background information. We understand that
this is solely for research purposes and that all obtained
information will remain strictly confidential.

We understand that, at our request, we can receive additional
explanation of the study after our participation is completed.

Signed by:

 

(Father)

 

(Mother)

 

(Son)

Date:

 

99

Family Information Sheet

 

Please answer all questions as completely as possible.

Father's Age:
Mother's Age:
Son's Age:

How many children are there in your family?

 

Please list the children's ages:

Boys: , , . . a 9 9
GTT‘IS: 9 3 9 9 9 a 3

 

 

Mother's relationship to son? (Please check one)
______Natural Parent
______Step-Parent
______Other (Please specify)

 

How many years have mother and son been living together?

Father's relationship to son? (Please check one)
______Natural Parent
______Step-Parent
._____Other (Please specify)

 

How many years have father and son been living together?

What is the highest level of education completed?

(Please be specific, e.g., 11th grade of high school)
Father's Education:

 

Mother's Education:

 

State your type of employment. (Please be specific, e.g.,
English teacher at Junior High School)

Father's Employment:

 

Mother's Employment:

 

The approximate annual family income is: (Please check one)

Below $4,000 $30,000-$34,000
$5,000-$9,000 $35,000-$39,000
$10,000-$14,000 $40,000-$44,000
$15,000-$19,000 $45,000-$54,000
$20,000-$24,000 $55,000-$64,000
$25,000-$29,000 $65,000-$74,000

Above $74,000

APPENDIX C

100

Hafner and Kaplan

 

TAT HOSTILITY SCALE
WEIGHTED SCORE SCALE

4 points
Themes involving direct physical hostile acts between people or

towards the self. (criminal assault; fighting; knifing of people;
murder; physical torture; rape, shooting of people; suicide, war)

3 points .
Themes involving hate; thoughts, feelings, dreams or threats of

direct physical hostile acts between people; themes involving
punishment, permanent debilitating injury, and death; themes of
direct physical hostile acts involving animals, (accidental
death; animals attacking humans; animals fighting; capital
punishment; dreams or thoughts of death, fighting, murder, self-
injury, sexual attack, and suicide; drowning, feelings of hate;
hunting; people dying; persons harming animals; punishment
involving deprivation; readiness to kill or physically attack;
revenge; threats of harm or punishment)

2 points
Themes involving verbal hostility; derogatory descriptions of

people; anti-social acts; people forced by others to do things;
hostile or negative emotionality; rejections; illness and acci-
dents involving injury; destruction of inanimate objects; pre-
datory animals; destructive forces of nature; weapons. (anger;
arguing; cheating; coercion; criminals; criticality; criticism
of others; domination; dreams of illness; drunkenness; embezzle-
ment; forgery; inconsiderate people; jealously; kidnapping;
lying; mentally retarded person; negativism; people ignoring or
snubbing one another; people smashing things; people with physi-
cal deformities; plagiarism; policemen; readiness for verbal
attack; robbery; sarcasm; self-depreciation; someone forced to
do something against their wishes; surgery; swearing; tornadoes;
truancy; verbal disagreement; verbal rebuke; verbal threat
other than physical harm)

1 point
Themes involving emotional deprivation, guilt feelings; escape,
misfortune; death symbols; broken objects; the military. (ceme-
teries and graves; people hiding; people running away from some-
thing; remorse; sadness; shame; soldiers)

0 oints
Themes without hostile content. (economic hardship; insects)

 

 

101

Hafner and Kaplan

 

TAT HOSTILITY SCALE
Overt - Covert Scale

OVERT HOSTILITY: Hostility which is manifest and direct.

 

Themes involving animals fighting; animals attacking humans;
arguing; capital punishment; cheating; coercion; criminal assault;
criticality; criticism of others; destruction of objects; dero-
gatory description of people; feelings or dreams of anger; feel-
ings or dreams of hate; fighting; harm to things of nature, such
as trees or plants; kidnapping, killing, knifing people; murder
or thoughts or dreams of murder; people snubbing one another;
people yelling at each other; persons doing harm to animals;
physical restraint; physical torture; punishment involving depri-
vations; rape or thoughts and dreams of sexual attack; readiness
to kill or for physical or verbal attack; revenge; robbery;
shooting people; someone forced to do something against their
wishes; strangling; swearing/ threats of punishments; truancy;
verbal disagreement; verbal rebuke; verbal threat including
dreams of verbal threat.

COVERT HOSTILITY: Hostility which is insidious, indirect, disguised,

 

or latent.

Themes involving accidental death; accidental injury or self-
injury including dreams and thoughts of broken objects; ceme-
teries or graves, criminals; destructive forces of nature (torna-
does, hurricanes, etc.); domination; dreams or thoughts of death;
drowning; drunkeness; embezzlement; emotional deprivation;
forgery; guilt feelings; hunting; illness--including dreams of
injury involving permanent disability; jealously; lying; mis-
fortune; negativism; people described as neurotic, psychotic or
mentally defective; people hiding; people ignoring one another;
people in the process of dying; people running away from some-
thing; people with physical deformities; people who are incon-
siderate of one another; plagiarism; policemen; predatory animals;
rejecting; remorse; sadness; sarcasm; self-depreciating; shame;
soldiers; suicide--inc1uding thoughts of surgery; thoughts of
arguing; war; weapons.

APPENDIX D

102

GENERAL GUIDELINES*

 

1. Before the actual scoring or assigning of weights, read the
entire story and then go back and score the hostility themes.

2. For every theme which is assigned a score of weighted hostility,
a score on the overt-covert dimension should also be made.

3. If a scorable object (e.g., gun) is mentioned or involved in
a scorable act (e.g., murder) only score the act.
4 Points = He picked up the gun and shot her.
If however, only the object is mentioned then score it.
2 Points anger
2 Points gun

4. If a scorable emotion or motivation is articulated, then it

should be scored, even if the emotion is part of a scorable
act or behavior.

1 Point = sad

3 Points = thoughts of suicide

He was so sad he thought about suicide.

Also remember, the emotion or motivation must be clearly stated
in order to be scored; don't score emotionality if the judge-
ment is based on inference.

5. "Or Rule": If a list of scorable acts are stated, none of
which are subsequently chosen, then score the lowest point
value act. In other words, if there is a listing of possible
scorable acts, and it remains unclear which act is actually
happening, then score the lowest act.

1 Point - sadness
Maybe this is about murder or perhaps suicide or maybe
sadness. It's not clear.
If however, one of the acts is eventually developed then score
that act.

6. If a scorable act is later articulated score both the act and
the articulation or elaboration.

1 Point = anger

3 Points thoughts about murder

3 Points feelings of hate
He was really angry. He was so angry he even thought
about murdering him, as he felt so much hatred towards
the men.

 

*These guidelines were created for the present study and were applied
to the Hafner and Kaplan scoring system.

10.
11.

12.

103

If the same scorable act takes place twice in the same story,
but is performed by two different people, score both acts.
3 Points = Man's threat of physical harm

2 Points
2 Points

Verbal argument
Boy's verbal threat

The man and the boy were involved in a verbal agrument.
The man threatened to beat the boy. The boy then threat-
ened that he would run away and never come back.

If the idea remains unclear, arbitrary or is contradicted, score
with the lowest point value that is applicable.

Be sparing with inferences; unless clear and obvious do not

infer.

If there is a caricature, treat as if it is a real person.

Drugs and Drunkenness--if within a social context do not score,
e.g., "two friends had a drink". Only score if:
1. it is a theme of excess (loss of memory, hangover,
passing-out)
2. anti-social drug use (addiction, trafficking, etc.).

Punishment -
3 Points
2 Points

2

categories apply:
all physical and threats of legal punishment
grounded, sent to room, can't use the car, etc.

 

O C
II II

4 Points

1.

3 Points

1.

104

Scoring System

 

Overtly Hostile Theme

Covertly Hostile Theme

Themes involving direct physical hostile acts between
people or towards the self:

KL!-

d-D’tfl 41¢ CLO Um

criminal assault - 0

physical fighting - 0

knifing of people - O

murder - 0

physical torture - O

rape (sexual attack, includes molestation) - O
shooting of people - O

suicide - attempt but not contemplation - C
war (only descriptions - don't include mere mention,
e.g., "looks like WWII") - C

strangling - 0

actual self injury - C

Themes involving hate: (all are Overt)

fDD-OU'Q’

Themes involving thoughts, feelings, dreams and/or threats

feeling of hate

despise

hatred as a motivation (must be clearly stated)
dreams of hate

revenge or vengeful acts (must be clearly articulated
e.g., "getting back")

 

of direct physical hostile acts between people or towards
self (nonperformed acts)

Q0 U'Dl

fighting - 0
murder - O
self-injury - C
sexual attack - O

 

2 Points

1.

e.
f.

g.

105

suicide - C
readiness to kill or physically attack - O
threats of physical harm - 0

Themes involving punishment

a.
b.
c.

d.
e.

capital punishment - O

punishment involving deprivation (usually legal in
nature, e.g., jail, prison, arrest, reform school) - O
threats of physical punishment - 0

actual physical punishment - 0

physical restraint - 0

Themes involving permanent debilitating injury:
(not the result of accident or illness)

a.
b.
c.

blindness - C
deafness - C
paralysis - C

Themes involving death:

CLOUD)

accidental death (car accident; drowning; plane crash) - C
dreams or thoughts about death - C

people dying - C

terminal or incurable disease - C

Themes of direct physical hostile acts involving animals:

CLOUQ!
oooo

hunting - C

animals attacking humans - 0

animals fighting - 0

person harming or torturing animals (with intent) - 0

Themes involving verbal hostility

LD'fiDQOC’O’

arguing - O

readiness for verbal attack - O

swearing - O

verbal disagreement - O

verbal threat other than physical harm - O
verbal provocation (teasing, insulting) - O
yelling at each other - O

 

IH1..A-_‘_ .2 1 .A
n
A m.- .

 

106

Themes involving derogatory descriptions of people

(DQOU'QJ

inconsiderate people - C

mentally retarded persons - C
self-depreciation - C

criminal - C

neurotic, psychotic, mentally defective - 0

Themes involving anti-social acts

Who—“3'40 "hm CLO Um

drunkenness or intoxication - C
embezzlement — C

forgery - C
kidnapping - O
lying - C
plagiarism - C
truancy - O
cheating - O
robbery - O
bribery - C

drug trafficking - 0

Themes involving people forced by others to do things
(includes attempts to do things, success is not a criteria)

CLOUD)

coercion - O

domination - C

grounding - sent to room without dinner - O

punishment not included under 3.3 (firing-as punishment;
expulsion from school)

Themes involving hostile or negative emotionality:

“'30-'50 an 0'01

anger - O

criticality - O

criticism of others (blaming) - O
jealousy - C

sarcasm - C

negativism - C

frustration - C

bitterness - C

resentment - C

Themes involving rejections: (a break up of a relationship
is not scored unless rejection is specifically mentioned)

a.
b.
c.
d.

people ignoring one another - C
the silent treatment - C

people snubbing one another - 0
person feeling rejected - C

 

10.

11.

1 Point

1.

107

Themes involving illness and accidents involving injury:

dreams of illness — C

people with physical deformities - C
surgery - (only in a medical sense) - C
autopsy - C

hunting or shooting accident - C

(DQOU'Q!

Themes involving destruction of inanimate objects:

a. people smashing things or objects - O

b. breaking windows - O r:-

c. smashing of inanimate objects (includes crashing e.g.,
auto accidents without injury) - 0

d. harm to things of nature (trees, plants) - 0

ill.“

Themes involving predatory animals:
a. stalking prey - C
Themes involving destructive forces of nature:

earthquakes - C
tornadoes - C
hurricanes - C
tidal waves - C
volcanoes - C

{DQOU'Q’

Themes involving weapons:

a. knives, guns, bombs, etc. - C
b. policemen - C

Themes involving emotional deprivation: (must be very
clear - more than just mention of someone 1eaving--has to
include misery or loss of person who has been left)

a. deprivation of love, support, comfort - C
Themes involving guilt feelings - all Covert
a. remorse or (bitter) regret

b. shame
c. feeling terrible about something done

 

 

108

3. Themes involving death symbols: all Covert

a. funerals

b. cemeteries - caskets and graves
c. cremations

d. widows/widowers

4. Themes involving escape - all Covert (primarily involves
escape from people or animals or physical objects (place
or situation, e.g., jail); only refers to emotional or
internal escape when it is clearly stated as such, e.g.,
"he was feeling so bad he got drunk to escape his worries."

 

a. people running away from something
0. people hiding

5. Themes involving the military - all Covert

 

a. soldiers

b. refers to a war without descriptions ("looks like WWII")
c. army base

d. naming of a branch of the military (army, marines, etc.)

6. Themes involving broken objects: all Covert
a. a clock that doesn't work
b. a door that is too rusty to open
c. flat tire

7. Themes involving misfortune: all Covert

 

a. sadness
b. crying
c. upset, depression
d. suffering
e. sorrow
f. grief
8. Themes involving illness and accidents involving injury to
animals:
a. animals with illness - C
b. animals with physical deformities - C
c. veterinary surgery - C
d. accidental injury to animals - C

0 Points

Themes without hostile content (economic hardship, insects)

 

 

L I ST 0F REFERENCES

 

 

LIST OF REFERENCES

Arnold, M. Story seqyence analysis. New York: Columbia University,
1962.

 

Bauman, G., and Roman, M. Interaction testing in the study of
marital dominance. Family Process. 1966, 5, 230-242.

 

Bauman, G., Roman, M., Borello, J., and Meltzer, 8. Interaction
testing in the measurement of marital intelligence. Journal
of Abnormal Psychology, 1967, 12, 489-495.

 

Bauman, G., and Roman, M. Interaction product analysis in group and
family diagnosis. Journal of Projective Technigues, l968,_§2,
331-337.

 

Bodin, A. M. Conjoint family assessment: an evolving field. In
P. McReynolds (Ed.), Advances in psychological assessment.
Palo Alto: Science & BéhEVior Books,71968.

 

Boszormenyi-Nagy, I., and Sparks, G. Invisible loyalties: recipro-
city in intergenerational family therapy. New York: Harper
& Row, 1973.

 

 

Dicks, H. V. Object relations theory and marital studies. British
Journal of Medical Psychology, 1963, aa. 125-129.

 

Doane, J. A. Family interaction and communication deviance in dis-
turbed and normal families: a review of research. Family
Process, l978a,_lz, 357-376.

Doane, J. A. Questions of strategy: rejoinder to Jacob and Grounds.
Family Process, 1978b,.17, 389-394.

 

Elbert, S., Rosman, B., Minuchin, S., and Guerney, B. A method for
the clinical study of family interaction. American Journal
of Orthopsychiatry, 1964,_34, 885-894.

 

 

Ferreira, A. J. Decision-making in normal and pathologic families.
Arch. Gen. Psychiat., 1963, 8, 68-73.

 

Ferreira, A. J., and Winter, W. W. D. Family interaction and deci-
sion-making. Arch. Gen. Psychiat., 1965, 13, 214-223.

 

109

110

Ferreira, A. J., Winter, W. D., and Poindexter, E. J. Some inter-
actional variables in normal and abnormal families. Family
Process, 1966, 5, 60-75.

Ferreira, A. J., and Winter, W. D. Decision-making in normal and
abnormal two-child families. Family Process, 1968, 2, 17-36.

 

Fine, R. A scoring scheme for the TAT and other verbal projective
techniques. Journal of Projective Techniques, 1955, 12,
306-316.

 

Fisher, S. and Mendell, D. The communication of neurotic patterns
over two and three generations. Psychiatry, 1956, l9, 41-46.

 

Fisher, S. and Mendell, D. The spread of psychotherapeutic effects
from the patient to his family group. Psychiatry, 1958, a1,
133-140.

 

Fisher, 5., Boyd, 1., Walker, 0., and Sheer, 0. Parents of schizo-
phrenics, neurotics and normals. Archives of General
Psychiatry, 1959, l, 149-166.

 

Framo, J. L. Systematic research on family dynamics. In I.
Boszormenyi-Nagy and J. L. Framo (Eds.), Intensive Family
Therapy. New York: Hoeber, 1965.

 

Framo, J. L. Symptoms from a family transactional viewpoint. In
C. Sager and H. Kaplan (Eds.), Progress in grogp and family
therapy. New York: Brunner/Mazel, 1972.

 

Framo, J. L. Family of origin as a therapeutic resource for adults
in marital and family therapy: you can and should go home
again. Family Process, l976,_l§, 193-210.

 

Friedman, J., and Friedman, A. Characteristics of schizogenic fami-
lies during a joint story telling task. Family Process,
1970, 9, 333-353.

 

Giffin, M., Johnson, A., and Littin, E. The transmission of super-
ego defects in the family. In N. Bell and E. Vogel (Eds.),
A modern introduction to the family. New York: MacMillan,
1968.

 

Goldstein, M., Gould, E. Alkire, A., Rodnick, E., and Judd, L.
Interpersonal themes in the thematic apperception test stories
of families of disturbed adolescents. Journal of Nervous
and Mental Disease, 1970, 150, 354-365.

 

 

Hafner, J., and Kaplan, A. Hostility content analysis of the
Rorschach and TAT. Journal of Projective Techniqpes, 1960,
24, 137-144.

 

 

111

Jacob, T. Family interaction in disturbed and normal families: a
methodological and substantive review. Psychological Bulle-
iip, 1975, 82, 33-65.

 

Jacob, T., and Grounds, L. Confusion and conclusions: a response
to Doane. Family Process, 1978, ll, 389-394.

Johnson, A. M., and Szurek, S. A. The genesis of antisocial acting
out in children and adults. Psychoanalytic Quarterly, 1952,
21, 323-343.

 

 

story technique and intrafamily analysis. Journal of Pro-
jective Techniques and Personality Assessment,71969, aa,
438-450.

 

Kadushin, P., Cutler, C., Waxenberg, S., and Sager, C. The family T.

Kadushin, P., Waxenberg, S., and Sager, C. Family story technique
changes in interactions and affects during family therapy.
Journal of Personality Assessment, 1971, l5, 62-71.

 

Karon, B. The thematic apperception test (TAT). In A. Rabin (Ed.),
Assessment with projective techniques. New York: Springer,
1981.

Lieber, D. J. Parental focus of attention in a videotape feedback
task as a function of hypothesized risk for offspring
schizophrenia. Family Process, 1977, l6, 467-475.

 

Martin, P. A. Dynamics of family interaction. J. C. E. Psychiatry,
1979, 40, 23-36.

 

 

Matranga, J. T. The relationship between behavioral indices of
aggression and hostile content on the TAT. Journal of
Personality Assessment, 1976, 49, 130-134.

 

 

Mendell, 0., and Fisher, S. An approach to neurotic behavior in
terms of a three generation family model. Journal of Nervous
and Mental Disease, 1956, iga, 171-180.

 

 

Mendell, 0., and Fisher, S. A multi-generation approach to the
treatment of psychopathology. Journal of Nervous and Mental
Disease, 1959, lag, 523-29.

 

Mendell, 0., Cleveland, S., and Fisher, S. A five-generation family
theme. Family Process, 1968, Z, 126-132.

 

Minuchin, S., Montalvo, B., Guerney, B., Rosman, B., and Schumer, F.
Families of the slums: an exploration of their structure
and treatment. New York: Basic Books, 1967.

 

112

Mustein, 8. Theory and research in projective techniques lemphasiz-
ing the TAT), New York: John Wiley & Sons, 1963.

 

 

Riskin, J., and Faunce, E. An evaluative review of family inter-
action research. Family Proceaa, l972,_ll, 365-455.

 

Roman, M., and Bauman, G. Interaction testing: a technique for the
psychological evaluation of small groups. In M. Harrower
(Ed.), Creative variations in projective techniques, Spring-
field: C. Thomas, 1960.

Sander, L. W. Adaptive relationships in early mother-child inter-
action. Journal of the American Academy of Child Psychiatry,
1964, a, 231-265.

 

Singer, M. T., and Wynne, L. C. Thought disorder and family rela-
tions of schizophrenics: III. methodology using projective
techniques. Archives of General Psychiatry, 1965, 1a, 187-
200.

 

Singer, M., and Wynne, L. Principles for scoring communication
defects and deviances in parents of schizophrenics: Rorschach
and TAT scoring manuals. Psychiatry, 1966, 22, 260-288.

 

Sohler, D. T., Holzberg, J. D., Fleck, S., Cornelius, A., Kay, E.,
and Lidz, T. The prediction of family interaction from a
battery of projective techniques. Journal of Projective
Techniques, 1957, 21, 199-208.

 

 

Stabenau, J. R., Tupin, J., Werner, M., and Pollin, W. A compara-
tive study of families of schizophrenics, delinquents, and
normals. Psychiatry, 1965, 28, 45-59.

 

Stewart, R., Peters, T., Marsh, S., and Peters, M. An object-rela-
tions approach to psychotherapy with marital couples, fami-
lies, and children. Familnyrocess, 1975, ll, 161-178.

 

Stierlin, H., Levi, L. 0., and Savard, R. J. Parental perceptions
of separating children. Family Process, 1971, l9, 411-427.

 

Stierlin, H. The dynamics of owning and disowning: psychoanalytic
and family perspectives. Family Process, 1976, lg, 277-288.

 

Vogel, E. F., and Bell, N. W. The emotionally disturbed child as the
family scapegoat. In N. Bell and E. Vogel (Eds.), A modern
introduction to the family. New York: MacMillian, 1968.

 

Werner, M., Stabenau, J. R., and Pollin, W. Thematic apperception
test method for the differentiation of families of schizo-
phrenics, delinquents and normals. Journal of Abnormal
Psychology, 1970, 15, 139-145.

 

113

Winter, W. D., Ferreira, A. J., and Olson, J. Hostility themes in
the family TAT. Journal of Projective Techniques, 1966, 39,
270-274.

 

Winter, W., Ferreira, A., and Olson, J. Story sequence analysis of
family TATS. Journal of Projective Techniques and Personality
Assessment, 1965, 29, 392-397.

 

 

Winter, W. D., Ferreira, A. J., and Bowers, N. Decision-making and
unrelated couples. Family Process, 1973, l2, 83-94.

 

Ziegler, R. G., and Muslinger, P. J. Persistent Themes: A naturali-
stic study of personality development in the family. Family
Process, 1977, lg, 293-305.

Zinner, J., and Shapiro, R. Projective identification as a mode of
perception and behavior in families of adolescents. Inter-
national Journal of Paycho-Anal., 1972, 53, 523-530.

 

ICHIGRN STATE UNIV. LIBRARIES
lIHI“1111111111111111111111111111111”11111111111
31293104515816