TERRUL ME} TERRILL'S MBHUQ FOR STUDYENG
FAMILY COMMUMCATION: DGES iT PROVIDE A
MEANINGFUL ABBREVEAWGN {BF COMMUNJCATiM?

Thesis fer the Degree of M. A.
MiCHiGAN SYA'EE UNIVERSE?!
ROBERT M. TEN KER;

1%?

 

 

a:

  

‘ ' "- .as’l‘dl’Tl-rt

1' FY? ’17" ‘ ':"'-'.'
.1

L) l I“). ’. is

.l, \'

.ll: . . "V

l'. ‘ ""‘ ' .' I
.. ., ‘. ,, a!
.3. "~.." ‘3‘

Univ 03.2,; Y

  

 

ABSTRACT

TERRILL AND TERRILL'S METHOD
FOR STUDYING FAMILY COMMUNICATION:
DOES IT PROVIDE A.MEANINGFUL ABBREVIATION OF COMMUNICATION?

by Robert H. Tinker

While Terrill & Terrill have published a method
for abbreviating and studying family communication, there
has been no information presented which indicates to what
degree such a method provides a "meaningful" or "useful"
abbreviation of family interaction.

The present research is an attempt to demonstrate
the extent to which Terrill & Terrill's (T&T's) method is
a meaningful or useful abbreviation, by comparing the author's
(E's) rankings of the functionality of family interaction with
another judge's rankings of the same interaction. _§'s rank-
ings were based on all available information about the fam-
ily's interaction, in contrast to the judge's rankings, which
‘were based only on T & T's ratings of 100 speeches of inter—
action placed in chart form. These ratings of the family
interaction into T & T's categories had been completed by
two independent raters. If the rankings of family inter-
action by.§ and the judge were significantly correlated,
it would indicate that rankings based only on the raters'

pooled and charted ratings of speeches might be "as good as"

Robert H. Tinker

rankings made directly from tapes and typescripts of family
interaction. Thus, such significant correlations would
indicate that T & T's method provides a meaningful or use—
ful abbreviation of family interaction.

Subjects were six four-member families. Interaction
data for each family consisted of 100 speeches taken from
a structured family interview (a speech is a relatively
continuous utterance).

It was also possible to determine the validity of E's
rankings, which essentially had served as a criterion by
which to evaluate the judge's rankings. The interaction of
the six families had been ranked from least to most patholog—
ical in a previous study (Moore, 1966), using a method which
discriminated at highly significant levels between a group
of eight normal and eight abnormal families. .g's rankings
were correlated with Moore's ranking. If they correlated
significantly this would provide some evidence for the valid—
ity of E's ranking.

A ranking of the judge's "overall" functionality of
family interaction based on charts of 100 speeches rated
into T & T's categories, was also compared with Moore's rank—
ing, based on 90 minutes of interaction. If the judge's
ranking correlated significantly with Moore's ranking, this
would furnish further evidence that T & T's method is "as

good as" Moore's method for this particular ranking task.

Robert H. Tinker

The specific hypotheses tested were:

1. The rank order correlations between E' s rankings
and the judge's rankings of the functionality of
the dyadic interaction within each family will
be less than or equal to zero.

2. The rank order correlation between E's rankings
and the judge's rankings of the functionality of
the mother—father interaction across families
will be less than or equal to zero.

3. The rank order correlation between_§'s rankings
of the functionality of the father-child inter—
action across families will be less than or equal
to zero.

4. The rank order correlation between E's rankings
and the judge's rankings of the functionality
of the mother-child interaction across families
will be less than or equal to zero.

5. The rank order correlation between E's rankings
and the judge's rankings of the "overall" function-
ity (excluding child—child interaction) of the
six families will be less than or equal to zero.

6. The rank order correlation between the judge's
"overall" ranking, and the ranking of the six
families based on Moore's (1966) "family path—
ology score" will be less than or equal to zero.

7. The rank order correlation between E's "overall'I
ranking and the ranking of the six families based
on Moore's (1966) "family pathology score" will
be less than or equal to zero.

Hypothesis one was not rejected, but hypotheses two
through seven were rejected (pEE.OS; one-tailed tests, right—
tail critical). These results indicated that E's rankings
and the judge's rankings of the functionality of dyadic in-

teraction within families were not significantly correlated.

The results do indicate, however, that when the functionality

Robert H. Tinker

of dyadic interaction was ranked across families, the rank-

ings byig and the judge were significantly correlated. The

results further suggest that E's rankings were a valid criter—

ion to use to assess T & T's method, in that E's "overall"
ranking of the functionality of family interaction was sign—
ificantly correlated with Moore's (1966) ranking of these
families based on his "family pathology score." Finally,
the results indicate that the judge's "overall" ranking
based only on charted ratings derived from only four minutes
of family interaction, were significantly correlated with
rankings based on Moore's (1966) "family pathology score,"
derived from 90 minutes of family interaction.

It was thus concluded that T & T's method of studying
family communication does provide a meaningful abbreviation

of family communication and could well be used profitably in

//“/]/// ‘
x4 {/2 wa//
/

L

//’c/ /'// ”aft—7

future research in interaction.

 

TERRILL AND TERRILL'S METHOD
FOR STUDYING FAMILY COMMUNICATION:

DOES IT PROVIDE A MEANINGFUL ABBREVIATION OF COMMUNICATION?

BY

Robert H. Tinker

A THESIS

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

MASTER OF ARTS

Department of Psychology

1967

To Marti, who happens to be a lovable wife.

AC KNOWL EDGM EN TS

I would like to express my appreciation to the peOple
whose helpfulness and assistance had a direct bearing on
the production and the completion of this thesis.

John Hurley, Ph.D., Lucy Ferguson, Ph.D. and Dozier
Thornton, Ph.D., as members of my committee were most in—
fluential. A special thanks goes to John Hurley, who, as
committee chairman provided much help, both as a participant
in the research, and as an interested person who furnished
much encouragement and guidance.

The research could not have been completed without
the many hours of work that Dick Does and Dick Brown gave
as raters. Their participation, comments and suggestions
were of great help in determining the rating procedure.

Finally, my thanks go to Marvin Moore, who made his
tape recordings of the family interviews available for the

present research.

iii

TABLE OF

STATEMENT OF THE PROBLEM. . . .
REVIEW OF THE LITERATURE. . . .
METHOD. . . . . . . . . . . . .

Subjects. . . . . . . . . . .

Procedure . . . . . . . . . .
RESULTS . . . . . . . . . . . .
DISCUSSION. . . . . . . . . . .
APPENDIX. . . . . . . . . . . .

BIBLIOGRAPHY. . . . . . . . . .

iv

CONTENTS

Page

38
38
39
54
59
66

86

LIST OF TABLES

Page
1 T & T's Categories and Their Definitions. . . . . 43

2 T & T's Ratings by Two Raters in Chart Form . . . 49

3 Percentagescf Agreement Between Two Raters
Separately Using T & T's Method. . . . . . . . . 56

STATEMENT OF THE PROBLEM

While Terrill & Terrill (1965) have published a
method for studying family communication, there has been
no information presented which indicates whether such a
method provides a "meaningful" or "useful" abbreviation
of family interaction. When family members interact to—
gether, the resulting communication occurs so quickly, is
so complex and interrelated, that it is necessary to limit
one's focus to only a relatively few aspects of the total
matrix of interaction. Thus it seems necessary to abbrev-
iate the family interaction in order to study it effectively.

A previous study, by Raush, Dittman & Taylor (1959)
provided suggestive evidence that a method similar to Terrill
& Terrill's (T & T's) furnishes useful information about
interaction patterns of young boys in a residential treat-
ment center. The method presented by Raush et al., and
the method presented by T & T are similar in that they
both are adaptations of Leary's (1957) interpersonal system,
using an interpersonal circle of eight categories arranged
around two orthogonal axes of dominance-submission, and
affection—hostility. T & T's method for studying family
communication and the method of Raush §£_al., seem to be
unique in that they focus upon the interpersonal aspects

of communication. Most otherresearch in family interaction

2
does not focus on such interpersonal aspects, or if it
does focus upon interpersonal aspects, does not do so in
a systematic fashion.

T & T's method and the method of Raush et al., also
seem to be unique in that they capitalize on the two axes
that other research indicates are major factors in inter-
personal behavior: dominance-submission and affection-
hostility (Adams, 1964; Foa, 1961, and Shaefer, 1959, 1961).

It would thus seem that research to determine the
meaningfulness, usefulness or validity of T & T's method would
be important; first because Terrill and Terrill have given
no definite evidence with respect to the validity or mean—
ingfulness of the method; secondly because the method focuses
on interpersonal aspects of family interaction which are def—
initely important, but virtually unresearched in any system—
atic way.

The purpose of the present study, therefore, is to
provide information on the degree to which T & T's method
abbreviates family interaction in a meaningful or useful
fashion. The study thus will provide information concern-
ing whether it would be profitable to use T & T's method in
future interaction research, or whether the method should

be modified or abandoned.

REVIEW OF THE LITERATURE

This review will be addressed to the question: "What
is it important to attend to in family communication from
the standpoint of picking out aspects of interaction which
are associated with interactional disorders?" Analyzing
research with this question in mind will (hOpefully) indicate
the current status of interaction research, and lead to ideas
as to what is needed in future research.

In attending to various aspects of interaction, in
order to test which aspects might differentiate between
say, families with a normal child and families with a
schiZOphrenic child, it is noticeable that some categories
(reported in research in the literature) are non-inferential
categories of interaction, such as who follows whom in
speaking, or the total amount of silence for a given period
of interaction. Other categories involve the use of very
low level inferences such as the number of disagreements in
a given interaction task. Still other categories could be
classified as making use of higher level inferences such
as a scale which rates communicational style with respect
to how communications focus upon a given topic.

Research in communication can be roughly grouped into
those studies which use categories which require predominantly

3

4
no inferences of low level inferences in their application,
and those studies which predominantly use categories needing
higher level inferences.

First, let us consider the studies which make use
of higher level inferences, Since there are only a small
number of these studies and they seem to have been espec—
ially productive, they will be considered in more detail.

An especially exciting piece of research of this
nature was done by Morris & Wynne (1965). In their research,
parental styles of communicating were studied by a judge
who had no diagnostically relevant data about the family
offspring. It was found that different forms or styles of
parental communication patterns (in tape recorded excerpts
from family therapy) could be discriminated from one another,
and that these differentiations could be used to accurately
deduce the psychiatric diagnosis and the form of communicat-
ing and expressing affect in the offspring. The deductions
were found to be significantly correct when compared with
independent clinical ratings of the offspring member.

Subjects were eight families with a young adult off-
spring member who had been hospitalized for psychological
reasons. The judge listened to tape excerpts from family
therapy. There were five or six excerpts about five minutes
each, which were chosen as being periods of minimum participa-

tion by the offspring member. His comments were erased from

5
the tape excerpts and paraphrased by another person. After
making ratings of parental communication, from these tape
excerpts, the judge made blind deductions about the off-
spring member, knowing only that the patient had been hospital—
ized for psychological reasons, his approximate age, and sex.
The judge's deductions from the parental interaction, about
the patient, were compared with evaluations made by a psy-
chiatrist by means of an individual interview with only the
patient. Using the psychiatrist's evaluations of the indiv-
idual patient as a criterion, the judge's deductionscf global
diagnosis for all eight offspring were "correct." Four
were called schizophrenic, and four were called non-schiZOph-
renic. Deductions of "form of thinking" were also completely
"correct," with agreement that one patient showed amorphous
(loosely organized) thinking; one mixed thinking (both
amorphous and fragmented characteristics); and two patients
showed fragmented thinking (thinking that is fairly clear,
but lacks integration). While Morris & Wynne call this
category "form of thinking" it would seem that a more accur-
ate description of the category would take into account
that it is communications that they are attending to, and
that they are judging these communications with respect
to how consistently clear and precise they are about any
given topic. Deductions about the severity of the patient's
psychotic tendency were made using five categories for level

of severity. In six cases, deductions were exactly accurate,

6

while in two cases, there was a one step error. Lastly,
deductions were made concerning the form of "affect dis-
order." Again, it might be emphasized that this category
is attending to verbal communications, and the category
is actually concerned with vocal modifiers from which one
infers apathy, depression, excitement, egg. A five step
scale is used ranging from apathetic (communications) to
turbulent (communications). Again, six deductions were
exactly l'correct," two deductions being one-step errors.

In each of the four categories above, the correlations
between the judge and psychiatrist were significant. There
thus seems to be a link between parental forms of communica-
tion and the communicational "disorder" that their disturbed
offspring manifests. Further, this disturbance was shown
to be deducible in this study, from the parental interaction.
Limitations of this particular study are that only one judge
has demonstrated such a deductive ability and only on a
small number of parent-offspring sets. No "normals" were
included, and an assumption seems to be that if it is pos—
sible to distinguish between two forms of emotional disturb-
ance, it will be possible to distinguish between a "normal"
group and groups of various "disturbed'l parents. This as-

sumption really ought to be empirically tested.

7

Given the interesting results of this study,what, in
the family communications, did Morris & wynne regard as
important to attend t0? What they felt was most important
in making these deductions was their "attention" scale,
by which they attempted to measure "the manner or style in
which family members share, or fail to share a focus of at—
tention on the same ideas, feelings, or tasks." It might
be emphasized this "attentional focus" was evaluated entirely
through the verbal communications of the persons concerned.
It thus might more accurately be called something like "com—
municational focus," rather than "attentional focus."

A ten-point scale was used, ranging from (1) undirected at-
tention, to (9) overfocusing of attention, and (10) manipul-
ation of meaning.

Not only does it seem that these deductions about the
disturbed offspring are possible from the actual parental
interaction, but it also seems that it is possible to make
such deductions from individually administered projective
techniques given to the parents, when the parental responses
to the projective stimuli are considered jointly as verbal
transactions which would have an effect on an offspring mem—
ber of their family (Wynne & Singer 1963a, 1963b; Singer &
Wynne 1965a, 1965b). In this study, by Singer & Wynne,
which is reported in four articles, much the same criteria

are used for evaluating communications as Morris & Wynne

8

(1965) used in their more recent study. In the Singer &
Wynne research, deductions were made about thirty-five off—
spring patients from the typescripts of tests given to their
parents. Secondly, tests from thirty-three offspring patients
and three siblings were matched with tests from their parents.
Of the thirty-five offspring patients, twenty were young
adult schiZOphrenics, nine were borderline schiZOphrenics,
and six were severely neurotic young adults. All were dis—
turbed enough to be hospitalized. A battery of tests were
used, with the Rorschach and TAT being the main tests. De—
ductions about the offspring member from his parents' tests
were as follows: The deductions of global diagnoses (schiZOph-
renic, borderline schiz0phrenic, non-schizophrenic) were
significantly correct (p<:.OOl). The deductions of I'forms
of thinking" (amorphous, mixed, fragmented or constricted)
were significantly correct (p‘<.001). Deductions of severity
of psychotic disorganization (a five-step scale) were sign-
ificantly correct (p (.001). The criterion for "correctness"
was based on agreement with evaluations of the individual
offspring by hospital staff who were well acquainted with
the patients.

In the matching part of the study, tests from two to
five families were studied as a set (socioeconomic status
of parents, age and sex of offspring being constant within

each set). Then the psychologist was given the tests from

9
the patient members of these families and attempted to match

them blindly with their parents.

No. of families No. of sets
per set No. of sets perfectly matched
2 2 2
3 6 4
4 l l
5 2 l

The matchings of the offspring (thirty-three patients
and three siblings) with parents was significantly correct
(p =.000002). This furnishes further evidence that what the
researchers associated with Lyman Wynne are attending to in
communications is diagnostically important.

Block, Patterson, Block & Jackson (1958) gave indiv—
idually administered projective tests to the parents of
twenty schiZOphrenic children and to the parents of twenty
neurotic children. Tests used were the Rorschach, TAT and
MMPI. The two groups were very carefully matched (on age
of child, age of parents, number of children in the family,
the educational and socioeconomic level of parents, and
participation of the mothers in psychotherapy). It took
the researchers four years to obtain these matched samples.
Analyses of the tests were made and the analyses of the
mothers of schiZOphrenics were compared with those of mothers

of neurotics; the same procedure being followed with the

lO

fathers of schizophrenics and neurotics. Essentially no
differences were found between the two groups of mothers
and fathers.

However, it is interesting to note that Singer &
Wynne (1963), reanalyzed this data, adding another group
of matched subjects (parents of twenty young adult schizo-
phrenics) and got some positive results. A main difference
in the two analyses of the same data was that Singer & wynne
considered the parents as a unit (who both would influence
the child through their verbal transactions), comparing
parents of schizophrenic children with parents of neurotic
children, instead of comparing fathers with fathers and
mothers with mothers as Block et al., did. The subjects
in this study thus consisted of the parents of twenty
schizophrenic children; the parents of twenty neurotic
children (ten acting out children, ten withdrawn); the
parents of twenty schizophrenics who became overtly ill in
late adolescence of young adulthood. Considering the
protocols of forty sets of parents of childhood schiz0phren-
ics and neurotics, thirty-four of the forty sets of parents
were correctly differentiated, which is significantly dif—
ferent from chance expectation. Parents correctly identif-
ied as parents of neurotic children, were then different—
iated as to whether they were parents of acting-our or with—
drawn neurotics, the differentiations again being significantly

COI'I'ECt.

11

The authors then set up TAT and Rorschach criteria
for differentiation between the different groups of parents.
Essentially these criteria seem to be preliminary groupings
which Singer & Wynne modified and refined in their later
research (Singer & Wynne 1965a, 1965b).

Most recently, Singer & Wynne (1966a) have published
Rorschach and TAT scoring manuals, setting forth in a very
complete and objective fashion, the scoring principles that
they used in their most recent research. The primary em-
phasis is upon the "attention" scale mentioned earlier. In
fairness to these researchers, it might be mentioned that
they no longer use the term "thought disorderll in their most
recent research, but write of "communication defects and
deviances," certainly a more accurate terminology. Yet to
be published, is a study (Singer & Wynne, 1967) comparing
parents ofschizophrenics, neurotics and normals, using the
Rorschach and TAT scoring manuals. This yet unpublished
study should clear up a number of questions which their
earlier research has left unanswered: Can more than one
rater make accurate differentiations of the groups of par-
ents; how do parents of normals differ in their communica-
tions from parents of neurotics and schiZOphrenics?

In general, it would seem that what Morris, Singer
and Wynne are attending to, in a person's communications,
have been well substantiated by their research as being im-

portant communicational variables; important in the sense

12
that these specific aspects seem to be significantly as-
sociated with specific communicational and nosological dis—
orders in the family offspring.

It might be noted that giving projective techniques
to parents of families, who would interact to produce responses,
has been suggested by several researchers (Fisher, Boyd,
Walker & Sheer, 1959; Haley, 1960; Winter, Ferreira &

Olson, 1965; 1966). This procedure would eliminate one
level of inference, in that the researcher would not have

to deduce characteristics of say, parental interaction from
the individually administered parental protocols, butcould
work directly from the actual interaction itself. Using a
test, such as the Rorschach or TAT, to elicit interaction
would have the advantages provided.by a standardized situa-
tion, plus a good deal is already known about these tests
which might be put to good use in the study of family inter-
action.

Winter, Ferreira & Olson (1965 and 1966) have ad-
ministered family TAT's to 126 three-member families. There
were fifty families with normal children, forty-four with
emotionally maladjusted, sixteen with schizophrenic, sixteen
with delinquent children. The families were asked to produce
conjointly three TAT stories based on nine cards, whidh
were scored by the Arnold system of Story Sequence Analysis

(1962). In this system of analysis, the judge rated

13
sequential themes on a five point scale of imputed emotional
maturity. Results revealed that the Arnold Score success-
fully differentiated Normal from Abnormal families, but
that the three Abnormal groups did not differ from each other.

A second study using the same subjects and their
stories, analyzed the data with respect to the relative
amount of weighted hostility and the percentage of overt
hostility in the themes, based on the Hafner-Kaplan system
(1960). Analysis of the results revealed that the normal
and schizophrenic groups produced stories which were low
both in weighted hostility and overt hostility, whereas the
stories of the families with emotionally maladjusted children
were high in both variables. The delinquent child families
scored high in weighted hostility but close to the normals in
overt hostility.

It seems that the results of these two studies are
difficult to interpret and are rather unspecific and gross
analyses of the family interaction. One gets the impression
that it would be very interesting to apply Wynne & Singer's
Scoring for communication defects and deviances as set forth
in their TAT manual (1966) to this data.

Another study which uses higher-level inferential
categories is one by Levin (1966). This study contrasts
the communicative behavior of a group of schizophrenic fam—

ily members with a control group. The experimental group

l4

consisted of seven schizophrenics, twelve fathers of schizo—
phrenics, fourteen mothers of schiZOphrenics. The control
group consisted of thirty subjects matched with respect to
age and sex, but the control group was of a higher socio-
economic status. This calls into question the validity of
the results, but the categories used remain of interest.
In this study, information about social interaction was sought
by isolating the subject and asking him to make a tape record—
ing which might be played subsequently to some specific
other person in the subject's family. His message would
enable that other person to carry out some simple task, us—
ually to produce a geometric figure. The tape recordings
generated are the data for later analysis. The subjects
performed the task twice. In the experimentalgroup, the
parents spoke "to" their spouse and "to" the identified
patient. Patients spoke "to" a parent and "to" a friend.
Controls spoke "to" a friend and "to" an unidentified eight
year old boy. Two independent raters coded eighty-six
completed explanations. The major dependent variable was
unclarity, which was measured in three ways. The hypothesis
that the experimental group would produce more ambiguous
and less adequate explanations than the control group was
supported. It was confirmed at all three indices of unclarity.

While the methodology of this study calls its results
seriously into question, that Levin's focus was on the clarity

of communications is of interest. Morris, Singer and wynne,

15
it can be recalled, found their "attention" scale most diagnos-
tic, and this attention scale was concerned with how clearly
communications focus around a given tOpic.

Another study which attempts to assess the clarity of
communications, is one by Beavers, Blumberg, Timken & Weiner
(1965), in which they compared the communication of nine moth-
ers of schizophrenics with the communication of nine mothers
of children with passive aggressive personality disorders.
These mothers were given an individual open-ended, semi-
structured interview, 45-60 minutes long, focusing on feel—
ings of the mother concerning the patients birth, early child-
hood and adolescence, socialization and heterosexual contact,
and the immediate interview. Transcripts of the taped inter-
views were scored by two independent judges. Scorable items
consisted of responses to questions by the interviewer con-
cerning the feeling state of the mother. These scorable items
were placed into one of three categories:

1. Definite responses--were clear and related to
the question.

2. Evasions--did not clarify the feeling state.
3. Shifts of meaning-~of two kinds:
a. The mother would indicate a feeling state, but
when the interviewer tried to confirm this,
an entirely different feeling state would be

given.

b. A direct contradiction of the feeling state
from a previous answer.

16
It was found that the mothers of non—schizophrenics had a
higher percentage of definite responses, and a lower percent—
age of shifts and evasions. These findings suggest again,
that measures of communicational clarity can discriminate
parents (mothers) of schizophrenics from parents (mothers)
of children with other psychological disturbances.

Another study that used high-level inferences derived
from family interaction, is a doctoral dissertation by Moore
(1966) Subjects were eight "normal" families, and eight
clinic families (these families had a child referred to MSU
Psychology Clinic for psychological reasons). The families
all had four to five members each, and clinic and nonclinic
groups were matched on the usual variables except that fathers
of clinic families had more education (1.2 years more) than
nonclinic fathers. Families participated twice in an hour
and a half structured family interview, given eight to ten
weeks apart. Two trained raters rated each family after each
session on a Family Rating Scale consisting of forty-five
five point Likert Scales (l=most; 5=least pathological). An
hypothesis that ratings of family interaction observed in
the standardized interview would reveal differences between
normal and clinic families was confirmed for both interviews
(for both p=.005). It was tested by summing the consistently

rated Family Rating Scale items as an overall pathology score.

17
Group means were then compared by a t-test. A second hy—
pothesis, that ratings of family interaction observed dur-
ing interview one are relatively similar to ratings made dur-
ing interview two, was only partially supported. Correlations
were calculated between interview one and two for each family
on the seventeen items rated consistently in both interviews.
Seven of fifteen families showed significant stability in
ratings. When group pathology scores were compared for the
two interviews, normal families were rated as significantly
less pathological in the second interview, while abnormal
families were not. This finding fits in very well with com-
ments in the literature that disturbed families seem to be
more rigid in their interaction than normal families, which
seem more flexible and adaptable.

Since this study used global ratings after ninety
minutes of family interaction, a problem arises in that item
content is not always associaUaiwith a specific behavior oc—
curring during the interview. There seems to be a halo effect
operating in the ratings, in that there were high intercor-
relations between the items which discriminated between the
two groups. It might be interesting to use more minute, dis—
crete, less global ratings of the tape recordings to measure
the temporal stability of the interaction. However, this dis-
sertation is the gnly study which has attempted to measure

the temporal reliability of family interaction thus far.

18

Although there was a halo effect, it might be interest-
ing to look at what variables this study found it important
to attend to, in the family interaction. There were six
items which significantly discriminated between the clinic
and nonclinic groups in bggh interviews:

1. Degree of consensus with the family spokesman

(was higher for nonclinic families in both inter—

views).

2. Overall emotional maturity of father (higher
for nonclinic fathers).

3. Cohesiveness (higher for nonclinic families).

4. Severity of negative sanctions by mother (less
severe for nonclinic families).

5. Warmth (higher for nonclinic families).

6. Happiness (higher for nonclinic families).

A largernumber of studies have concentrated their at-
tention on lower level inferences about communication processes
than the above studies, which emphasized higher level infer-
ences. Studies making use of low level inferences have pro—
duced some very interesting results as well as promising further
usefulness.

One of the earliest and most interesting studies on
family or parental interaction was done by Fisher, Boyed,
Walker & Sheer (1959). Subjects for their study were the
parents of twenty male schizophrenic V.A. patients, parents
of twenty male neurotics V.A. patients and parents of twenty

non—psychiatric patients. The parents were given a number

19

of diagnostic measures, including individual interviews,
projective questions, Rorschach and TAT, and a measure of
parental interaction (parents interacted to produce a joint
story to one TAT card). Of the twelve different methods used
to analyze these tests and individual interviews, only the
analysis of the sample of interaction discriminated signifi-
cantly between the three groups of parents (other measures
discri;inated only a normal-abnormal dichotomy). It was
found that the parents of normals were most often in definite
agreement in reaching a mutually agreeable story, and the
parents of schizophrenics were least often in definite agree—
ment. The authors concluded that since the individual meas-
ures didn't discriminate between the three groups of parents
and the measure of interaction did, that it made more sense
to study parental interaction rather than the individual
parents, such as the "schiZOphrenogic mother." This study
suggests that the amount of agreement may be an important
variable to attend to in communication.

Farina, in a study of the communication of parents
of schiZOphrenics (1960) used three samples: The parents
of twelve good premorbid schiZOphrenic sons, the parents of
twelve poor premorbid sons, the parents of twelve sons
hospitalized for TB. The parents were interviewed to deter-
mine the premorbid adjustment of their sons (Phillip's scale
of premorbid adjustment was used). Then the mother and

father separately answered twelve hypothetical problem

20
situations, after which they interacted to resolve their
disagreements. Indices of dominance and conflict were used
to analyze the interaction. It was found from the inter-
action indices, that maternal dominance was most marked for
the poor premorbid sons; paternal dominance most marked for
the good premorbid sons. Extent of conflict was least for
controls, greatest for poor premoribd sons. Indices of
dominance were as follows: number of times person speaks
first; number of times person speaks last; total of first
and last; passive acceptance of solution; total time speak-
ing, relative to other person; yielding. Indices of conflict
were as follows: frequency of simultaneous speech, duration
of simultaneous speech, interruptions by mother, interrup-
tions by father, interruptions total, disagreements and ag—
gressions by mother, disagreements and aggressions by father,
disagreements andaggressions total, failure to agree, verbal
activity. Farina & Dunham (1963) have also used these in—
dices of dominance and conflict in later research.

Caputo (1963) studied the parents of male chronic
schiZOphrenics, using for subjects, parents of twenty white
male, chronic schizophrenics, and parents of twenty "normals"
who had never been hospitalized for psychiatric reasons.
Measures used were the Parent Attitude Inventory (PAI), con—
sisting of fifty true-false opinion items; the semantic dif-
ferential, consisting of ratings of "myself," "my son,""my

marriage," etc. The PAI and the semantic differential were

21
given individually to the parents. Then the parents discussed
their first ten disagreements on the PAI, with a View to-
wards agreeing on a true or false Opinion. Caputo found that
the results from the parental interaction were in startling
contrast to the benign picture of the family members given
by the semantic differential. The parents of the schiZOphren—
ics were less able to agree on a Single response than the
parents of controls (p( .001). The Bales Interaction Process
Analysis was applied to the parental interaction and its
results support the notion of a hostile atmosphere in the
home of the schizophrenic. Caputo concludes that both parents,
in contrast to the mother alone contribute to the schiZOphreno-
genic character of the home. Caputo's study lends corroborat-
ing evidence to the finding of Fisher et al., that the ability
to reach agreement on a conjoint test, is an interactional
variable of importance in discriminating different disagnostic
groups of parents.

In a very interesting and well done doctoral dissert—
ation, Lerner (1964) specifically studied how different groups
of parents resolve intrafamilial conflict. The subjects were
thirty-six pairs of parents, divided into three groups of
twelve couples each. Twelve couples were controls, parents
of sons hospitalized for non-psychiatric reasons. Initially
the parents of twenty—four schiZOphrenic sons were dichotom—

ized on the basis of their son's social maturity scores

22
(determined by means of a Social Competence Scale by Zigler
& Phillips). Later the twenty-four sets of parents were
redichotomized on the basis of their son's "thought disorder"
(determined by Becker's Rorschach Genetic Level Score). The
parents were individually given a thirty item questionnaire
and asked to indicate agreement or disagreement with items
about childrearing, husband-wife relationships, etg., and
subsequently asked to resolve by discussion ten items on
which they had had discrepant Opinions. This "revealed dif-
ference" technique was originally used by Strodtbeck (1951),
and is frequently used in interaction research. These tape
recorded discussions were scored for five categories dealing
with the resolution of intrafamilial conflict:

1. Compromise (both parents change from their original
position).

2. Role induction--mother (mother is induced to change
to the father's questionnaire position).

3. Role induction--father (father is induced to change
to mother's position).

4. No agreement (conflict unresolved).

5. No agreement with distortion (the conflict is
unresolved but the couple acts as if it has been
resolved).

The two categories for role induction were further scored

for three categories:

1. Yielding (both parents state their original
position before one of them yields).

23

2. Passive acceptance of solution (one parent
acquiesces without stating his original position).

3. Masking (a position is stated which is at variance
with the questionnaire response).

When the parents of schizophrenics were dichotomized
on the basis of their son's social competence, it was found
that the control group was characterized by their signific-
antly more frequent use of compromise, the absence of dis-
proportionate patterns of parental dominance and submission,
and the tendency to openly acknowledge disagreements between
family members. The high socially competent group was
characterized by less compromise, disproportionate patterns
of paternal dominance and maternal submissiveness, and the
tendency to fully recognize differences between family members.
The low socially competent group was characterized by less
compromise, disprOportionate patterns of maternal dominance
and paternal submissiveness, and the tendency to distort
or not recognize disagreements between family members. The
low socially competent group also differed significantly
from the controls, the low group making greater use of masking.
The high socially competent group used intermediate amounts
of masking.

When the parents of schizoPhrenics were re—dichotom-
ized on the basis of their son's genetic level score, Lerner
found that compromise was used most frequently by the control

group and least frequently by the low genetic level group.

24
In contrast, "No Agreement with Distortion" was used most
frequently by the low genetic level group and least frequently
by the control group. Also the low genetic level group used
masking significantly more often than the controls, with the
high genetic level group intermediate.

It is very interesting to note that when the schizo-
phrenic sample was dichotomized on the basis of a social com-
petence score, disprOportionate patterns of parental domin-
ance and submission appeared, and the sex of the dominant
parent appeared to be related to the patient's premorbid level
of social maturity. However, when the schiZOphrenic sample
was dichotomized on the basis of the Rorschach Genetic Level
score, similar patterns of disprOportionate dominance and
submission failed to emerge. One interpretation offered for
this finding is that rather than the entire intrafamilial
environment exerting an indiscriminant influence, family
role structure (e.g. maternal or paternal dominance) could
be specifically related to the offspring's social competence,
while family masking and distortion could be specifically
related to the patient's "thought" disorder. (Again "commun-
ication disorder" would be a better term than "thought dis-
order").

Also Lerner mentions that when the conflict categories

were dichotomized into "reached a final agreement" and "reached

25
no final agreement" categories, that the control group and
the low genetic level group differed significantly. This
parallels the finding by Caputo (1963) who found that the
parents of male chronic schizophrenics were significantly
less able to reach agreement on a revealed difference test
than were matched controls.)

Another study which explored the family interaction
of the schizophrenic was by Cheek (1964) who used a re—
vealed difference technique to elicit family interaction and
used the Bales Interaction Process Analysis to categorize
the interaction. Subjects were sixty-seven family triads
with a schiZOphrenic offspring member and fifty-six partially
matched family traids. Cheek found higher mother-son agree-
ment, more withdrawn and underactive interaction in schizo—
phrenic families. Also the schiZOphrenic's mothers' higher
support-permissiveness of the child (per questionnaire)
turns out behaviorally to be merely greater tolerance in con-
trast to active support which normal mothers gave.

According to Haley (1964), the ideal data in family
research should be the recording of Observable events.
Further he suggested that there should be no inference in
this primary data, but that inferences should be made_§££§£
the primary data has been collected. Applying this principle
to research, he used an instrument recording of who followed
whom in speaking in family triads. Subjects were forty

normal family traids and forty disturbed family triads (one

26
or more members had sought or had been recommended for psy-
chological treatment). The family interaction was elicted
by means of a discussion of a questionnaire and by means
of joint efforts to make stories to TAT cards. Reasoning
that two-person sequences of speaking would be more flex-
ible (random) in normal families and more rigid (less random)
in disturbed families, it was found that the normal group
differed from the disturbed group at the .00003 level of
significance, the normal families being at the more flexible
end of the continuum. Although this is a very interesting
finding, research which will be subsequently discussed has
not substantiated it.

Riskin has discussed a methodology for studying family
interaction (1963) and has developed a set of scales for
categorizing family interaction (1964). Riskin has applied
these scales to the interaction of nine families ranging from
"normal" volunteer families to families with a schizophrenic
member. While statistical data has not been presented,
descriptive data has been presented, and Riskin states that
through examination of only the interaction ratings, it is
possible to make many inferences about family interaction
and functioning. Each family, among other things, was asked
to plan something to do together as a family, and was left
alone in the room to do this. The first seventy-six and
last seventy-six speeches were used as data (a "speech" is

a relatively continuous utterance by an individual). One

27

hundred fifty—two speeches comprise four to six minutes
of interaction. The six, essentially tri-partite scales are
the following: clarity, topic change, commitment, agreement,
intensity, and relationship. The judge would rate each speech
on each of the six scales. In using the relationship scale,
for example, he would label the speech as friendly, neutral,
or attacking. More rigorous, less descriptive research with
Riskin's scales remains to be reported.

Stabenau, Tupin, Werner & Pollin (1965) compared
the interaction patterns of families of schiZOphrenics with
families of delinquents and families of normals, using a
semi-structured family interview, individual interviews with
family members, and individual testing of the family members
(the Rorschach, MMPI. TAT, the Object Sorting Test and a
revealed differences test were used). All families were
four member families, one child being the patient,the other
child serving as a control. Fifteen families participated,
five in each diagnostic classification. Most significant
of the findings was an analysis of the TAT stories of the
parents, which differentiated the three diagnostic groups
(p<;001). This TAT analysis focused on the nature of the
parent-child interaction contained in the stories. While
the normal parents, in their stories allowed the child a

realistic amount of independent action, the parents of

28
delinquents used strict discipline and expected immediate
obedience, and the parents of the schiZOphrenics produced
stories which suggested that the child primarily filled the
emotional needs of the parents. No statistically signifi-
cant differences were found on interaction measures from the
revealed differences test. The interaction measures used
were the following: interactiantime, total number of times
family members spoke, overlaps per minute, interruptions per
minute, pauses per minute. These were recorded for four of
the revealed difference questions. However, a global esti-
mation of communication clarity was made, which, when the
results were rank ordered, revealed that the schiZOphrenic
families were least clear, the normal families most clear
in their verbal communications. This final result corresponds
with the previously cited research which suggested that
measures of clarity are important to attend to.

Ferreira & Winter (1965) in another study of family
interaction were particularly interested in three rather
simple questions:

1. How much agreement was there among family members

with respect to what they liked and what they
didn't like?

2. How much time did the families take to reach all
seven decisions required in the study?

3. How apprOpriate were these family decisions
in terms of fulfilling the wishes of the individual
family members?

29

The subjects were family triads, consisting of fifty
normal family triads and seventy-five abnormal triads, with
the abnormal family traids broekn down into fifteen schiZOph-
renia producing families, sixteen delinquency producing
families and forty-four maladjusted families. Each family
member separately answered a questionnaire which consisted
of a number of situations for which there were ten alterna-
tives from which they indicated the three they best liked
and the three they least liked (e.g. ten famous peOple you
might like to meet). The results indicated that normal famil-
ies differed in specific ways from the abnormal families.

The normal families had a much greater agreement in what
their members liked and disliked, prior to any exchange of
information; had to spend less time in reaching family de-
cisions; and were able to arrive at more appropriate de—
cisions in terms of a better fulfillment of the family
members' individual choices.

Lennard, Beaulieu & Embrey (1965) have reported on
some partially completed research, comparing schiZOphrenic
family interactionwith the family interaction of controls.

In their research they are mainly concerned with the amount,

rate and direction of intrafamilial communication. Who speaks,
and to whom is recorded. Also rate and success of intrusions
are recorded. An intrusion is defined as an entry by a third

person into a dyadic interaction, which is not requested or

30
elicited by the interacting dyad. Two other aspects of
interaction attended to, are the interpersonal referents
of parental communication in interaction with the child
(does the parent focus on himself, the child, or someone
else?), and whether the parents' communications concern
feelings, experiences, thoughts of their own, their child's,
or someone else's. These last two measures seem especially
interesting, with respect to the interpersonal learning im-
plications for the child.

Ferreira, Winter & Poindexter (1966) author another
study focusing on low level inferences from family communica—
tions. In this study their interest is on the questions:

How much, and in what ways, do abnormal families differ from
normal families with respect to who talks the most, who talks
the least, how much overlap there is, how much time the fam-
ily remains silent. Subjects were 126 family triads, fifty
normal family traids, seventy—six abnormal triads (sixteen
schiZOphrenic, sixteen delinquent, and forty—four maladjusted
families). Interaction was elicited by the family members
interacting to produce three stories to nine TAT cards. It
was found that abnormal families required more time than normal
families to perform the family task of arriving at a TAT
story, and that there was greater silence in the abnormal
group than in the normal group (p<3001). It thus seems that

abnormal families talk less and exchange less information

31
per unit of time than normal families. Haley's earlier
finding (1964) that abnormal families areless random (and
less flexible) in their interaction sequences than normal
families was not supported.

Mishler & Waxler (1966), while they have not yet
reported their results, have reported on their research
design, which is about the most complete design so far
reported in the literature. The design in this study in—
cludes male and female schiZOphrenics with good and poor
premorbid histories and their parents, as well as normal
families of both male and female offspring. Intrafamilial
controls were also provided with the inclusion of a well
sibling of the patient, of the same sex and approximately
the same age, in a separate experimental session with the
parents. A revealed differences technique was used to elicit
forty-five to sixty minutes of discussion. A number of dif-
ferent scales are applied to categorize the interaction, and
the design of the study allows for rather complex analyses of
data.

Goodrich & Boomer (1963) have develOped a Color Match—
ing Technique which has been used to elicit couple inter-
action (Ryder & Goodrich, 1966). In this technique, follow—
ing each of twenty color matching tasks, the experimenter
asks the couple to discuss their choices and arrive at an
agreement as to the best possible match, stressing that only

their agreements count toward their score in the experiment.

32
In half of the series, the couple is deceived; in these in-
stances the colors are so arranged that no agreement is
possible. The data consists of the couple's discussions
as they search for a mutually acceptable solution for each
of the twenty color matching tasks. Subjects were forty—
nine white, middle class, volunteer couples, each of whom
had been married three to four months. The final typescripts
of the recorded sessions were coded independently by two
raters for forty-seven variables, seventeen of which are dis—
cussed inthe study. It was found that there is a striking
tendency to alter or distort choices in order to avoid dis-
agreements (p<:.001 for both husbands and wives). A factor
analysis revealed four factors, three of which have been
replicated in subsequent research (Ryder, 1966); affectivity-
rationality; verbal fluency, and husband versus wife lead-
ing or assertive behavior. Many of the categories that Ryder
& Goodrich use are specific to the experimental set up of
the color matching technique and will not be discussed in
the present review.

Turning again to our question, "What is it important
to attend to in family interaction from the standpoint of
picking out aspects of interaction which are associated with
interactional disorders?" it becomes evident from a close

appraisal of the literature that measures of clarity and

33
measures of agreement seem to have been most clearly re—
lated to specific interactional disorders.

With respect to clarity, Morris & Wynne (1965), Singer
& Wynne (1965 & 1966), Beavers et al., (1965), Lerner (1964)
and Levin (1966) have found that different measures of clar-
ity differentiate between parents of schiz0phrenics, neurotics
and normals, with parents of schiZOphrenics being least clear
in their communication and parents of normals most clear.
With respect to agreement, Fisher et al., (1959), Caputo
(1963), and Lerner (1964) have found that parents of schizo-
phrenics have most difficulty in reaching definite agreement
on various tasks, and that parents of normals are most able
to reach definite agreement.

Although family interaction is definitely an inter—
personal process, it is evident from this review of the
literature, that the categories which have been used so far,
do not stress the interpersonal nature of this interaction.
Rather interest has centered on such structural characteris-
tics as clarity and agreement. While these variables are
of demonstrated importance, it would seem that attention to
how persons in the family relate to each other and how they
seem to feel about each other, as derived from their com-
municational behavior, could be a very productive approach.

While some researchers have directed attention in this

34
direction (most notably Stabenau et al., who found that TAT
stories, when analyzed with an eye to the nature of parent—
child interactions, discriminated between parents of schizo-
phrenics, delinquents and normals), there has been little
evidence accumulated so far about the value of an interpersonal
analysis of family interaction.

There are, however, two papers which indicate on a
descriptive level that the interpersonal analysis of family
communication might be a very productive approach. Guerney
and Guerney (1961) posited that the understanding of a given
family member depends upon understanding his interpersonal
interactions with the other family members, and suggested
that Leary's interpersonal theory (1957) was most appropriate
for this. They applied Leary's concepts to the case of a
family with a "disturbed" nine year old girl. Jackson, Riskin
& Satir (1961) presented a blind analysis of a tape recorded
five minute segment of parental interaction, and tried to
deduce the characteristics of the son of these paraits. Their
analysis focused upon the individual's perception of self,
his perception of the person he is talking to, and his per—
ception of the other in relation to the self. Descriptive
statements about the parents and the son seemed to be fairly
accurate.

There also are two researches which have utilized Leary's
system in a somewhat more controlled methodology. The first

is by Raush, Dittman & Taylor (1959), in which the authors

35

used Leary's system to obtain descriptions of the quality,
frequency, and intensity of behaviors of six disturbed boys,
age eight to ten, in a residential treatment center. Ob-
servations of behavior were made two times in six different
settings for each of the boys, with the same observational
procedure being followed eighteen months later. One observer
watched thechild for about eight minutes in one setting, then
dictated his observations. His observations were coded into
Leary's system by two raters working together. By comparing
interaction from the first group of observations, with the
interaction from the observations made eighteen months later,
it was possible to obtain a measure of interaction change.
It was found, for example, that the major changes in inter-
action occurred with adults, with a decrease of hostile-
dominant behavior, and an increase in friendly-passive be-
havior directed toward adults by these boys. The authors
concluded that Leary's schema for coding interpersonal be-
havior has some measure of utility. While this study has a
number of methodological shortcomings, its emphasis on changes
in interpersonal behavior as a result of treatment (instead
of changes on less direct measures, such as Q-sorts, psy—
cholOgical tests) is excellent.

Terrill & Terrill (1965) have modified Leary's system
in order to apply it more readily to tape recordings of family

interaction. Terrill & Terrill's method is based on rating

36
each"speech" (a speech is a relatively continuous utterance
by an individual), into one of eight interpersonal categor-
ies. These eight categories are arranged in a circle around
two orthogonal axes: a dominance-submission axis and a
hostility-affiliation axis. A number of reviews of the liter-
ature (Adams, 1964; Foa, 1961 and Shaefer, 1959, 1961) have
suggested that a large number of studies converge in their
findings toward such a simple ordered structure of inter-
personal behavior. According to Adams (1964), the "findings
suggest a circumplex structure around two orthodogal axes of
dominance-submission and affection—hostility." Terrill &
Terrill added four neutral categories for rating speeches
which couldn't be rated into one of the eight interpersonal
categories, and applied their system of interpersonal analysis
of communication to ten families. Two raters had an average
agreement of 78% in their ratings. The authors also gave a
case illustration of the method.

The purpose of the present study is to determine to
what extent Terrill & Terrill's analysis is a meaningful or
useful abbreviation of the total matrix of family interaction,
by comparing the author's (E's) rankings of the functionality
of family interaction with another judge's rankings of the
functionality of family interaction. E's rankings were based
on all available information about the family's interaction
(E Observed each family's entire structured interview, used

the tape recordings and the typescripts of the interaction),

37
in contrast to the judge's rankings which were based only
on Terrill & Terrill's (T & T's) ratings of interaction
placed in chart form. These ratings of the family interaction
into T & T's categories had been completed by two independent
raters. Assuming such rankings of family interaction are
useful or meaningful to psychologists, if‘g and the judge
are in significant agreement hitheir rankings of family
interaction, it could be concluded that T & T's scales are
"meaningful" abbrefiations of the total matrix of family in—
teraction. If.§ and the judge are not in significant agree—
ment in their rankings, this might call into question the
usefulness of T & T's approach to family interaction (or it
might call into question various other aspects of the re-
search design).

The specific hypotheses tested were:

1. The rank order correlations between E's rankings
and the judge' s rankings of the functionality of
the dyadic interaction within each family will
be less than or equal to zero.

2. The rank order correlation between E's rankings
and the judge's rankings of the functionality of
the mother-father interaction across families

will be less than or equal to zero.

3. The rank order correlation between E's rankings
and the judge's rankings of the functionality of

the father- child interaction across families will
be less than or equal to zero.

4. The rank order correlation between E's rankings
and the judge' s rankings of the functionality of
the mother—child interaction across families will
be less than or equal to zero.

38

5. The rank order correlation between E's rankings
and the judge's rankings of the "overall" function-
ality (excluding child-child interaction) of the
six families will be less than or equal to zero.

6. The rank order correlation between thejpdge's
"overall'I ranking, and the ranking of the six
families based on Moore's (1966) "family pathology
score" will be less than or equal to zero.

7. The rank order correlation between E's "overall"
ranking and the ranking of the six families based

on Moore's (1966) "family pathology score" will
be less than or equal to zero.

M ETHOD

Subjects

The subjects were six, four—member families (father,
mother, and two children). These families had lived together
for at least four years prior to their participation in the
research, had children ranging from ages eight to sixteen,
and had at least one male child between those ages. The fam—
ilies ranged from "normal" (no member ever recieved or had
been recommended to receive any type of psychological treat—
ment for an emotional or behavioral disorder), to "abnormal"
(families which were waiting for psychotherapy at the MSU
Psychological Clinic, involving both parents and sometimes
one or more children). The abnormal families initially con-
tacted the clinic because a male child between age eight and
twekmahad been referred for underachievement and/or lack of

behavior control in school. Of the six families in the

39
present study, three were classified as normal families
and three as abnormal families. These six families were
selected from a larger research project (Moore, 1966), in
that they were the first six families employed in the
doctoral research, having just two children. Table A in the
Appendix gives more complete sociological information about

each family.

Procedure

 

The family interaction of the four—member families
was tape recorded in a ninety minute structured interview
in the MSU Psychological Clinic, conducted by a doctoral
candidate in clinical psychology. All families interacted
to perform a number of tasks, two of which furnished the
interaction data for the present study. In the first task,
the whole family was instructed to "Plan an activity you
could all do together; it should be something you might
actually do" (Moore, 1966). In the second task, the family
was asked to discuss the following question with the inter-
viewer present: "Suppose and (used names of
children) are very angry with each other; how should they
let each other know how angry they are?" (Moore, 1966).

Typescripts were made by‘E from the two relevant parts
of the tape recordings in order to facilitate the rating
process. It required about two hours to transcribe five to
tau minutes of family interaction. It was relatively dif—

ficult to disentangle the complex conversational strands of

40
family interaction, as this process required painstaking
listening and re-listening.

The behavioral unit of analysis for making ratings in
the present study was that of a "speech," according to the
definition used by Terrill & Terrill (1965): "A scorable
speech consists of a relatively continuous utterance by_an
individual which is either uninterrupted, or if interrupted,
(apparently uninfluenced by the interruption. In this defini—
tion, the term "relatively continuous" means that there are
no pauses or breaks in the speech longer than four seconds.
It also means that the speaker does not abruptly shift from
speaking to one person to speaking to another. Thus, if a
person starts out talking to another person but then abruptly
turns to someone else and addresses him, this would be counted
as two separate speeches, even though the speech is uninter-
rupted. The term "apparently uninfluenced" refers to situa-
tions where a person begins a sentence or a train of thought,
is briefly interrupted, but continues with the same sentence
or thought as if he had not heard the interruption."

-§ broke up the interactions into speeches, consecutively
numbering each speech on each family typescript, starting
with the first speech in the "plan something" excerpt, and
ending with the last speech in the "anger expression" dis-
cussion. The first fifty and the last fifty speeches in this
combined sample furnished the rating data for each family.

One hundred speeches comprises about four minutes of interaction.

41
By listening to the tapes, and by using the typescripts, E
also determined who was speaking to whom. When it was unclear
to whom the speech was addressed, the speech was considered
directed to all family members.

Two independent raters, both advanced graduate students
in clinical psychology, rated each speech according to Terrill
& Terrill's (T & T's) system. Practice ratings were made on
one four—member family (excluded from the statistical analy-
sis Of the present study) in order to familiarize the raters
with the interactiai categories. The preliminary training
on the scales consisted of fifteen to twenty hours spent rating
about three hundred speeches, and reconciling differences in
the application of the categories. It was found that the inter-
rater agreement quickly reached 50-60% agreement on ratings
of one hundred speeches, but improved little with further
practice. Because prolonged training was resulting in only
very small improvements in reliability it was decided to pro-
ceed with this limited degree of interrater agreement.

In T & T's analysis of communication, each speech is
placed into one of Leary's eight interpersonal categories.

T & T have written formal definitions, based on the two
principal dimensions of the interpersonal circle, affection-
- hostility and dominance-submission, in addition to using
adjectival category descriptions, which are more open to

ambiguity. T & T considered the vertical axis of the circle

42
as a status dimension which indicated the status position the
speaker was taking in a given speech, and the horizontal
axis as indicating the emotional attitude which the speaker
was expressing toward the person to whom he was speaking.
Each speech is considered to have both a status and an emotional
component. In using T & T's complete Scoring Guide (obtained
in a personal communication from James Terrill,)one must
first decide which of fourteen descriptive categories (e.g.
ask question, answer question) the speech falls into before
rating it into one of the eight interpersonal categories.
The Scoring Guide is presented in Table B in the Appendix.

As preliminary ratings suggested that this seemed to
be unnecessarily complicated, in the present research the
rater primarily inferred the status and emotional dimensions
of each speech in order to rate the speech into one of the
eight interpersonal categories. The categories, a brief des-
cription, and the status and emotion components as per T & T's
formal definitions are presented in Table 1.

In order to rate each speech, the rater first assigned
a preliminary rating to it, using only the typescript. He
then listened to that same speech on the tape recording and
if necessary, modified his original rating on the basis of
the vocal elements of the speech, When still in doubt about
the prOper category, the rater employed the complete Scoring

Guide.

43

Table 1

T & T's Categories and Their Definitions

 

 

Category Description

Status

Emotion

 

AP

BC

DE

FG

HI

JK

NO

Leading, controlling
behavior

Defensive, compet-
itive

Critical, attacking
Passively resistant,
passively critical
Self-effacing, yield-
ing

Positive conforming-
following

Agreeable—affilia-
tive

Supportive-benevol-
ent leading

High or dominant

High or dominant
Slightly high to
equal

Low or submis-
sive

Low or submis-
sive
Low or submis-
sive

Slightly low to
equal

High or dominant

Neutral to
slightly pos-
itive

Negative

Negative

Negative

Slightly neg-
ative to neu—
tral

Neutral to
positive

Positive

Positive

 

As Terrill & Terrill found that some speeches did

not fit readily into the eight interpersonal categories,

they added four neutral scoring categories as described in

Appendix C to make their coding system more comprehensive.

Thus, a total of twelve mutually exclusive categories.

the eight interpersonal plus these four neutral, were utilized

for rating speeches.

44

The six families were randomly assigned the letters A
through F and were presented in random order EDCFAB to Rater
A, and in random order CADBFE to Rater B. The one hundred
speeches for each family were separately rated to ascertain
interrater agreement.

Then, each rater, using the tape recordings of the
one hundred speeches of interaction for each family, the type-
scripts of the same speeches, and his own (untabled) ratings
of the speeches for each family, made a number of rankings
about the functionality of the interaction in the six families.
When it was found, that the two raters had no agreement in
their rankings of family interaction, the research design was
modified. With the goal of clarifying the criteria for
making rankings of intra-and inter-family interactions,
minor changes were made in the ranking tasks. Secondly, E,
who had rather extensive familiarity with the families (he
had observed the ninety minute family interviews and made
the typescripts from the tape recordings of the family inter—
views), made rankings of the interaction of the six families.
Theserankings made by.§ were subsequently used, and the rank-
ings made by the two raters were discarded.

The new research design compared E's rankings of
family interaction with another judge's ranking of family
interaction. _§'s rankings were based on the total matrix
of family interaction; that is, the tapes and typescripts of

the one hundred speeches for each family, plus a general

45
impression of each family's interaction recalled from ob—
serving the ninety minute interview. The judge's rankings,
in contrast, were based solely on the two raterS' pooled and
charted ratings.

This research design permitted an assessment of the
extent to which T & T's method furnished a "meaningful ab-
breviation" of the more complete family interaction, by com-
paring rankings made from listening to the actual family
interaction with rankings made from T & T's charted ratings.
Thus, ifIE and the judge had significantly correlated rank-
ings, it would suggest that making rankings from T & T's
charts might be "as good as" making rankings from listening
to the actual interaction. That is, the abbreviation which is
involved in rating each speech into T & T's categories before
making rankings would not result in the loss of pertinent in—
formation, making T & T's method inferior to rankings made
from listening to the tapes of the actual interaction.

Thus, if the rankings from T & T's charts correlate
significantly with E's rankings (the criterion rankings), it
would suggest that the abbreviation of interaction to the
manageable proportions that T & T's analysis affords, is a
"meaningful" or "useful" abbreviation. However, it might
be noted that a demonstration of its meaningfulness is
limited, first of all, by the meaningfulness of the ranking

tasks, secondly by the meaningfulness of the criterion itself.

46

Some evidence will be presented indicating to what extent

the criterion rankings (E's rankings) are meaningful. The

usefulness or meaningfulness of the ranking tasks will de—

pend upon the specific research done in the future, but being

able to rank family interaction on some sortof continuum from

good to bad would seem to be of general importance for a

wide variety of researches which might be done.

The rankings which were made by‘g and the judge were

as follows (the judge was a Ph.D. clinical psychologist, who

had extensive experience and acquaintanceship with Leary's

system, having previously done research with it):

1.

2.

Rank order the six dyads in each family from most
to least functional.

Rank order the mother-father dyads (M-F; F-M)
from most to least functional across the six
families.

Rank order the father-child dyads (two for each
family) across the six families. This entails
comparing the functionality of two dyads in one
family (e.g. F-D, D—F and F-S, S-F) with the
functionality of the two dyads in each other fam—
ily to obtain one final ranking from onetx>six.

Rank order mother-child dyads (two for each fam-
ily) across the six families from most to least
functional. One final ranking from one to six
is obtained.

Sum the three rank scores assigned to each family
for the father-mother, father-child and mother-
child interaction and Obtain an "overall" function-
ality-dysfunctionality ranking for each family.

(A ranking was not obtained for child—child inter—
action because the amounts of child-child inter?
action were uniformly so small--eight speeches

or less--as to preclude an indication of the

nature of the interaction.

47
In considering the functionality-dysfunctionality
dimension in making the preceding rankings, a three-part
scale of functionality was used, based upon the quality
and the quantity of the interaction:

1. Most fiinctional dyadic interaction is considered
as interaction which is predominantly positively
toned (but Optimally containing both positively
and negatively toned interaction), and consisting'
of "appropriate" amounts of interaction for that
dyad. Suggested general standards for "appropriate—
ness," based on observations of family interaction
are as follows:

a. Parents interact with each other for at least
one-sixth of the speeches (there are six dyads
in each family). If the parents interact in
less than one-sixth of the speeches, this is
taken as indicative of interpersonal distance
between the parents.

b. Parents interact in approximately equal amounts
with their two children.

c. Small amounts of child-child interaction are
considered as appropriate in this interaction
setting.

2. Moderately functional dyadic interaction is con-
sidered as predominantly negative interaction (qual—
ity) which occurs in appropriate amounts (quantity),
.95 predominantly positively toned interaction
(quality), occurring in too small amounts (quantity).

3. Dysfunctional dyadic interaction is considered as
predominantly or completely negatively toned in—
teraction, occurring in inapprOpriate amounts, ie.
too much or too little.

.E's rankings and the judge's rankings were essentially

independent, as the judge's rankings were made on the basis
of T & T's charted ratings of the family interactions, while

.E's rankings were made from his own evaluation of the tapes

and typescripts. It remains to explain how E made his evaluation

48
of the tapes and transcripts and to describe the charts
that the judge employed in making his rankings.

As the functionality-dysfunctionality continuum was
concerned with the quality and quantity of interaction, E
summed the number of speeches given by, say, the mother to the
father, and by the father to the mother. This was done for
each dyad in the family. Then after listening to the quality
of the interaction in the dyad, E made a summary statement
concerning the dyadic interaction, i.e., both the mother and
the father made predominantly positively toned speeches to
each other; e.g.:

M-F 12 predominantly positively toned

F-M '18 predominantly positively toned

30 total speeches
These summary statements about the quality of the interaction
in the dyad and the amount of interaction in the dyad fur-
nished the data used by.§ for ranking the dyads within each
family and for ranking the dyads among the six families.

After the raters had rated each speech for each fam-
ily, their ratings were placed in chart form; one chart for
each family. The chart for Family A is presented in Table 2.
These six charts provided the data used by the judge for mak-
ing his rankings of the functionality of the family inter—
action. All six of the charts are presented in Table D of

the Appendix.

49
Table 2

T & T's Ratings by Two Raters in Chart Form

 

 

Family A
S1 age 12
82 age 8
Speeches Given by Father to:
M S1 82 All Sl&32 Total

A B A B A B A B A B A B
AP 1 2 6 l 6 l 3 14
BC 1 2 l 2 2
DE 1 l 0
FG 0 0
HI 1 1 l l l 3
JK l l l l 3 1
DH 2 2 l 2 5 2
NO 1 5 4 6 4
LI 2 2 l l 3 l 6 4
UM 0 0
NT 0 0
NX l 2 3 0
TTL 7 4 ll 7 l 30 30

Speeches Given by Mother to:

AP 2 3 2 7 l 2 3 14
BC 1 l l l 3 2 5
DE 1 l 5 l 6 2
FG 1 1 0
HI 1 1 l 1
JK O 0
1M. 1 l 0
NO 2 2 6 3 l 11 3
LI 2 2 2 2
UM 0 0
NT 0 0
NX 0 0
TTL 5 3 l7 2 27 27

. 50
Table 2 Continued

Speeches Given by Sonl to:

F M S All Total

A B A B A B A B A B

AP 1 l l l l l 2 4
BC 1 l l 1
DE 1 l l l 2
FG 1 1 1 1 2
HI 1 3 l 4 1
JK l l l 2 2 3
LM 0 0
NO 1 l 0
LI 1 2 2 2 2 5 4
UM 2 0 2
NT 0 0
NX 1 l 2 0
TTL 5 7 4 3 l9 l9

Speeches Given by Son2 to:

AP 5 8 2 4 l l 8 13
BC 1 3 2 3 3
DE 1 l 0
FG l l 0
HI 1 l 1 l 2
JK 2 2 2 2
LM 1 l 2 0
NO 0 0
LI 1 l l l 2 2
UM l O 1
NT 0 0
NX 3 l l 4 l
TTL 11 ll 1 l 24 24

 

The first of the two columns under each heading = Rater A
The second of the two columns under each heading = Rater B

51

In making comparisons between E's rankings and the
judge's rankings, five null hypotheses were made. For
each of the five rankings, the hypothesis was that the cor-
relation between E's rankings and the judge's rankings would
be less than or equal to zero. One-tailed tests, right tail
critical, were used, in that maximum power was desired to
reject the hypothesis if the rank-order correlation was greater
than zero. That is, only a significant positive correlation
would indicate that T & T's system is a useful shortened form
of the total matrix of family interaction; a large negative
correlation was assumed to be meaningless.

An external criterion for determining the extent to
which T & T's rating method is a useful abbreviation of
family interaction was also available. Moore's research
project (1966) included all six families used in the present
research. He used a Family Rating Scale (FRS) as his princ-
ipal method of rating family interaction, which was based on
global judgments made by raters immediately after observing
the ninety minute family interview. The FRS was shown to
discriminate between eight normal and eight abnormal famil-
ies (employing only consistently rated items) at the .005
level of significance, on each of two separate interviews.
FRS ratings were also employed to obtain an overall "family
pathology score." It was possible to obtain a ranking of
the six families using this "family pathology score," and

compare this ranking with an "overall" functionality ranking

52
Offamily interaction from the present study. Again the null
hypothesis was that the rank order correlation would be less
than or equal to zero. If the null hypothesis was rejected,
i.e., if the correlation was greater than zero, it would
indicate that, for this specific ranking Of family interaction,
T & T's analysis of interaction based on only £93£_minutes of
interaction, furnishes data highly correlated with Moore's
FRS, which was based on ninety minutes of interaction, and
which discriminated between groups of normal and abnormal
families at a highly significant level. Positive evidence of
this kind would support T & T's method as a useful condensa-
tion of family interaction.

It would also be possible to compare E's "overall"
ranking of functionality with Moore's "family pathology score"
in order to determine the extent to which E's rankings could
serve as a valid criterion by which to evaluate T & T's method.
If there were a significant correlation between E's ranking
and Moore's ranking, it would suggest that E's rankings pro-
vided a valid criterion, in that they would be correlating
with a method which discriminated between a group of normal
families and a group of abnormal families, on two different
occasions, at a highly significant level.

The specific hypotheses were as follows:

1. Hypothesis one: The rank order correlations be-
tween E's rankings and the judge's rankings of the functional-
ity of the dyadic interaction within each family will be less

than or equal to zero.

53

2. Hypothesis two: The rank order correlation between
“E's rankings and the judge's rankings of the functionality
of the mother-father interaction across families will be less
than or equal to zero.

3. Hypothesis three: The rank order correlation be—
tween E's rankings and the judge's rankings of the function-
ality of the father-child interaction across families will
be less than or equal to zero.

4. Hypothesis four: The rank order correlation between
.E's rankings and the judge's rankings of the functionality
of the mother-child interaction across families will be less
than or equal to zero.

5. Hypothesis five: The rank order correlation between
_§'s rankings and the judge's rankings of the "overall" func—
tionality (excluding child-child interaction)of the six fam—
ilies will be less than or equal to zero.

6. Hypothesis six: The rank order correlation be-
tween the judge's I'overall" ranking, and the ranking of the
six families based on Moore's (1966) "family pathology score"
will be less than or equal to zero.

7. Hypothesis seven: The rank order correlation be—
tween E's "overall" ranking and the ranking of the six fam—
ilies based on Moore's (1966) "family pathology score" will

be less than or equal to zero.

54
Edwards' (1961) formulas and tables were used to

determine the significance of the rank—order correlations.
RESULTS

The amount of agreement that the two raters demonstrated
on each individual speech is presented in Table 3. An exam-
ination of Table 3 indicates that the two raters had rather
limited exact agreement on each speech, ranging from 50%
exact agreement with Family C, to 32% with Family E; having
a mean of 40%. However, when one-step errors were added
to the exact agreements, the mean amount of agreement went
up to 62%, and when one and two-step errors were added to
the exact agreements, the mean amount of agreement went up
to 74%. This suggests that while the raters had low exact
agreement on each individual speech, they were often in
essential agreement, in that their ratings were frequently
in closely adjacent categories on the interpersonal circle.

In fact, complete disagreements (three or four-step errors

were considered as complete disagreements), on assigning speeches
to different interpersonal categories ranged only from 8-15%,
with a mean of 10%, this being a fairly small percentage of
complete disagreement among interpersonal categories. It

should be mentioned, however, that there were somewhat more

conflicts where one rater assigned a speech to one of the

55

four neutral categories, while the other rater assigned the
same speech to one of the eight interpersonal categories.
This sort of discrepancy ranged from a low of 8%.with Family
B to a high of 25% with Family C; and a mean of 16.3%. In
general then, the raters were in essential agreement on 73.7%
of the speeches rated, and disagreed on 26.3%, orabout one—
quarter of the rated speeches.

With respect to the seven hypotheses formulated in
the present study, the following results were obtained:

Hypothesis one: The rank order correkfiions between
.E's rankings and the judge's rankings of the functionality
of the dyadic interaction within each family will be less
than or equal to zero. This hypothesis could not be rejected.
No correlation reached significance at the .05 level. The
correlations for families A through F were as follows:
—.829, +.600, +.7l4, -.771, -.257 and +.l43. Since this
hypothesis could not be rejected, it could mean that T & T's
ratings are not "useful" abbreviations of family interaction,
for this particular task, or it could mean that other aspects
of the research design are at fault (If, for example, a
second judge had ranked the family dyads using the same charted
interactions as the first judge, and their rankings were not
significantly correlated, then it would call into question,
not T & T's system, but rather the process by which the judges

had derived rankings from it).

56

 

 

 

mmnm mHue ewuso moumm omnmm mmcmm
m.OH o.OH h.mn N.N© om.mm smmz
NH NH or am wm M
NH NH mo Ho Nm m
NN m me He em Q
mN m no No om o
w mH pm No we m
NH v em mo ms m
Hmuosum uswEmmumm
muswewmummmHQ moumlv Ucm nouum mwum uswEmwum¢
HMGOmummHmu Imv muswE IN paw mwum souum moum uswEomum¢ >HHEmm
IsH .m> Hmuuswz IwmnmmmHQ IH .uomxm IH pew Domxm Domxm
ponuwz m.B O B msHmD meumummwm
mnmumm 039 somzumm usmEmmumm mo mommDswoumm

m OHQMB

57

Hypothesis two: The rank order correlation between
lg's rankings and the judge's rankings of the functionality
of the mother—father interaction across families will be less
than or equal to zero. This hypothesis was rejected. The
rank order correlation was+389, a correlation significantly
different from zero (p:<.02).

Hypothesis three: The rank order correkation between
.E's rankings and the judge's rankings of the functionality
of the father-child interaction across families will be less
than or equal to zero. This hypothesis was rejected. The
rank order correlation was +.77, which is significantly dif-
ferent from zero (p=.05).

Hypothesis four: The rank order correlation between
.E's rankings and the judge's rankings of the functionality of
the mother—child interaction across families will be less
than or equal to zero. This hypothesis was rejected. The
rank order correlation was +.83, which is significantly dif—
ferent from zero (p < . 03) .

Hypothesis five: The rank order correlation between
Ig's rankings and the judge's rankings of the "overall" func-
tionality of the family interaction (excluding child-child
interaction) across families will be less than or equal to
zero. This hypothesis was rejected. The rank order cor—
relation was +.99, which is significantly different from

zero (p<.008).

58

Hypothesis six: The rank order correlation between
the judge's I'overall" ranking, and the ranking of the six
families based on Moore's (1966) I'family pathology score"
will be less than or equal to zero. This hypothesis was
rejected. The rank order correlation was +.79, which is
significantly different frOm zero (p<<.05).

Hypothesis seven: The rank order correlation between
_§'s I'overall'I ranking and the ranking of the six families
based on Moore's "familypatholOgy score" will be less than
or equal to zero. This hypothesis was rejected. The rank
order correlation was +.83, which is significantly different
from zero (p< .03) .

All actual rankings are presented in Table E in the
Appendix.

These results indicate that E's rankings and the
judge's rankings of the functionality of dyadic interaction
within families were not significantly correlated. The
results do indicate, however, that when the functionality of
dyadic interaction was ranked across families, the rankings
bylg and the judge were significantly correlated. The re-
sults further suggest that E's rankings were a valid criter—
ion to use to assess T & T's method, in that E's "overall"
ranking of the functionality of family interaction was

significantly correlated with Moore's (1966) ranking of these

59
families based on his "family pathOIOgy score." Finally
the results indicate that the judge's "overall" ranking based
only on charted ratings derived from only four minutes of
family interaction, were significantly correlated with rank-
ings based on Moore's (1966) "family pathOIOgy scores,'l

derived from ninety minutes of family interaction.
DISCUSSION

What conclusions can be drawn when rank-orders based
on different methods of analysis of family interaction are
significantly correlated? It would seem that the two dif—
ferent methods of analysis used to make the rankings, could
be regarded as equivalent, at least for the particular rank-
ing tasks on which the correlations were significant. Thus,
when it was found that E's rankings of interaction across
families were significantly correlated with the judge's rank-
ings, it could be concluded that the methods used by'g and
the judge for making the rankings were equivalent, at least
for those particular ranking tasks. In this case, it could
be concluded that T & T's method of analysis was equivalent
to the method used by E, on the ranking of interaction across
families. Or, putting it another way, little information
was lost when rankings were made from T & T's ratings placed
in charted form, in comparison to rankings made directly from

the actual tape recordings of family interaction.

60

Not only did the judge's "overall" ranking correlate
significantly with E's "overall" ranking, but the judge's
ranking also correlated significantly with Moore's ranking
of these six families. In other words, for a ranking task
involving ranking the overall family interactioi from most
to least functional (pathological) for the six families,

T & T's method, based on four minutes of interaction, was as
good a method as Moore's FRS, based on ninety minutes of in-
teraction. This is rather striking evidence that T & T's
method is an efficient abbreviation of a small amount of fam-
ily interaction. Thus, one does not need ninety minutesto
make an assessment of the overall functionality of family
interaction, but can make such an assessment from merely

four minutes of interaction, using T & T's method.

How valid a criterion was provided by using E's
rankings as a criterion in the present study? Using
Moore's "family pathology score" as a basis by which to
evaluate the validity of E's Rankings, it was found that
.E's "overall" ranking of the interaction of the six famil-
ies correlated significantly with Moore's ranking (p‘<.03).
Since E's "overall" ranking was a composite of the separate
rankings of the different dyads in the family (except for
the child-child dyad), this gives some evidence that E's
rankings across families were valid. However, it gives no

informatia1 as to how valid E's rankings within families were.

61

It thus seems logical to conclude that T & T's method
is an abbreviation of family interaction which is meaning-
ful or useful. It is an "abbreviation" in that it attends
only to a specific facet of family interaction, that of the
relationship expressed. It is "meaningful" in that it cor—
relates with a method which is not so abbreviated and about
which there is external evidence of validity.

One limitation of the present study is that it demon-
strates that T & T's analysis provides useful information
.iny with respect to the specific ranking tasks used in the
present study. If the sort of ranking tasks used in the
present study are not useful in future research, then the
present research provides less tenable evidence of the useful—
ness of T & T's method.

A second limitation in the present study is that of
the small sample size. Since only six families were used,
it perhaps would be better to interpret the results cautiously
and regard the present study as providing only suggestive evid—
ence concerning the meaningfulness of T & T's method.

A third limitation of the present study is in the
nature of its methodology. The study sought only to demon-
strate whether T & T's system of analysis provides a useful
abbreviation of the larger matrix of family interaction.

In this task, it falls short of being a study of concurrent

validity, unless the term is stretched somewhat. If

62
concurrent validity is involved with correlating the results
of a rating method with an external criterion, it can be
noted that the present study had no criterion which is purely
external. Such a criterion would involve a different segment
of interaction for each of the two persons making rankings.
In the present research B's and the judge's rankings were
based on the same one hundred Speeches, and these speeches
were part of the ninety minute interview. Since Moore (1966)
had two interviews with each family, spaced eight to ten weeks
apart, and the present study used segments only from the first
interview with each family, Moore's ranking of the families
from the second interview could serve as a criterion, with
which the overall ranking of the present study could be com-
pared. There is only one problem with this: It would be a
good criterion only if the interaction were perfectly stable
over time. Moore's research indicated that it was not.
Actually, a rank order correlation was computed between E's
and the judge's combined "overall" ranking of the family
interaction in the present study, and the ranking of families
based on their "family pathology score" from the second inter-
view. This correlation was +.48, which is non-significant.
It is probably non-significant because the family interaction
is not stable; but this is not definitely known. In sum-
mary, though, the present research falls short of being a
study of concurrent validity, which involves comparing rank-
ings with an external criterion, because there is no purely

external criterion in the present research.

63

In comparison with T & T's study on interrater agree—
ment, the present study's average agreement of 40% is rather
paltry compared to their 78%. Reasons for this difference
can only be speculated at, but here are a few speculations:
First of all T & T do not mention how much pre-training
their raters had, or whether they participated in the develOp-
ment of the rating system. The raters in the present study
spent about twenty hours in the pre—training, and although
difficulties still weren't completely ironed out between the
two raters, time limitations dictated going ahead with the
rest of the study. Thus it may be that the raters started
out with less pre—training On the system or at least less pre—
study agreement on the rating system than T & T's raters.

One difficulty that posed problems for the raters in
the present study is one related to the nature of inferences
that can be made about an individual speech. If for example,
the mother begins the "plan something" interaction by asking
her child what he wants to do, is she primarily encouraging
her child; or merely trying to gain time to think about the
situation while he answers, or is she really doing something
else? After listening to many interactions, one could probably
get a pretty good idea about what she's "really" doing. But
raters, because of personal bias or different personal ex-

periences, may easily interpret such a speech differently,

64
even with extensive training. Maybe the thing to do is get
raters who ”think alike."

Another speculation would involve the extent to which
the scoring guide adequately presents what T & T's raters
were really doing. Maybe their understanding or use of the
guide was clearer than that of the raters in the present
study. For example, the scoring guide was unclear about what
to do when a person gives an opinion (which would mean using
one column of the Scoring Guide) but also praises and agrees
with another person (which would mean using another of two
columns in the Scoring Guide). Which of these three columns
should be used to determine the interpersonal category?

After having used the rating system, some of its short-
comings are clearer and it is possible to suggest modifications
or improvements for it.

There seem to be a number of problems involved in using
adjectival descriptions of categories. They are often vague
and open to misunderstanding. Some interpersonal behaviors
might not be included in the list of adjectives, might not
seem to fit a cluster of adjectives, or might seem to fit in
several categories on the basis of the descriptions. It
might improve the system to do away with adjectival descriptions
entirely, and instead, rate each speech on the two axes,
(dominance-submission and affection-hostility) which have been

validated by researches previously cited. One way this might

65
be done is to divide the dominance-submission scale into
high, medium, low status categories, and the affection—
hostility scale into negative, neutral, positive,emotional
categories. This means the system would have nine cate-
gories ranging from high status, negativeemotion through med-
ium status, neutral emotion, to low status, positive emotion.
Speeches could thus be rated just using the status and emotion
axes without using adjectives to define the categories.

A problem with the rating of each speech, is whether
each speech is equally meaningful or important. It does
seem that some speeches are more meaningful or important than
others. What does one do about thiS?

Despite the limitations of the present study in design
and size of sample, and despite the limitations of the par-
ticular rating system used, it does seem that analysis of
interaction is a productive approach in psychology, and that
emphasizing the relationship aspects of such interaction is
also productive. The findings of the present study suggest
that T & T's method of analysis is useful in the present
form, and could beput to use in future research where the
aim is to analyze the interpersonal aspects of verbal inter-
action. It is further suggested that with modifications,
such an approach as T & T's might be made an even more ef-
fective research tool. Ultimately it may be possible to law-
fully specify the interpersonal consequences of various types

of antecedent interpersonal behavior.

APPENDIX

Table A

Specification of Family Characteristics

"' 3-:
4

Father , Mother Children
Family Occupation Education Occupation Education Sex Age Group

 

A Mechanical 17 Housewife 14 M 8 Abn.
Engineer M 12

B Postal 12 Secretary 12 M 12 Nor.
Clerk F 16

C Tool and Die 12 Housewife 12 M 11 Nor.
Maker M 14'

D Pet Store 15 Housewife 12 M 8 Abn.
Manager M 12

E Bricklayer 12 Housewife 12 M 15 Nor.
F 10

F Graduate 17 Housewife 13 M 10 Abn.
Student F 13

 

67

Table B

Terrill & Terrill's Complete Scoring Guide

68

 

 

 

 

 

 

 

 

—" “HLT AL DI.LCT ATTL ::rr CIVL DILLC TICN TLACH, I ZBORI.., ADVISE,
41%; . ”0 1g,uiOe,pelsnade, lead INSTRLCT
i .J(x LL11 urge, expor't -
;.151 1JL EXPERT LLADLRSIIP
1
r z“ “
§ REJECT, L}T LUBE LLEEVD POSITION NARCI SISTIC, CCI.‘FLT 121
l ' rerulse icea or hut doesn t have __ACTS, showing Off, beast
.ann inATIVh person, to be hostile ing
3 ESTABLISH IALLA DLPI SIVL AC1 ION CONFETITIVL: override
; ILNLLNCL on guard, ration- conversation, interfere
‘ alising _ w/ SPECQh
(ChASTIE’JL) NO---BY T
1 T TLRLATLAING ICTIC\ (MILD CRITICISA)
i3 » L.. m_, VLRbAL ACIIKL
___ - ' 1Lfi’hilfl AFN\A(K, belitLle,
“ “"““ ridicule, sarcasm
{PASSIVE ACRLE SI»; PASSIVE W/HOIDING PASSIVE RLLLLLTCN
ACTION, thwarts, Aflc ION, "cool", " in response:perceived
‘1" f‘” H balks, obstructs, "Lard to please, authority, noncompliant
*N" LUA IVE w/holds resources hard to get" , re- unwilling, stubborn,
refuses ST<to other naining close- negative
9"”‘1 . mouthed, silent "—
. - SABPTICAL, DISTRUST
AVOID STAND PASSIVL‘CONPIAINT
. 1 DOLBT
1_ I \
tLIGH‘i’Ll’ 'SI'DTJISSIVE FEB-x- SELF CRITICAL SIGNS or mum'sss, FEAR,
LLCATIVL P0153 T0 ALTnOR- 'LACT IONS self INABILITY ,
Lea T0 ITY, allow self effacing ' - .- . ;
{-.LL'TRAL be ta lked down, ' ' 1 ._ ~ I
surrender, yield, TIAID, SHY ' }
submit, apologize '
. - .‘ , O
1 ’TRLT LAPLLSLICAS or LLILADLAT ACTs,'.
-1 “fl;0'“ RESPECT, admir- place responsi-
V , * L ation, awe, es- bility on other ,
F \‘QTTIV . . . 1 p
‘ ”* “ teem _ . . _‘ 1
' ATTEMPTS T0 - 1 E
GREETINGS.AND LSTAJLISH - GIVE CRLDIT . - 5
*fl._‘ --. FRIEI'DLXP G7 ST'IF" S SH WM LT? IhUSIAlu EOR
L-.1_:;-.1.i.- " 5' L ‘--' ALLIA‘KCLS, talk OTHERS VIEWS
LC; COOPEPITIV“ 2X" ‘ as equals . -
FOSITIV3 ‘* “L " “ F Irr IX'LAUC"TLR 15
.g L” AL CHANGES ACTS or ATTEACTICN R “D n j
, ' ' LOVE, f1FFECTiON ' j
OFFER HELP, ELAsyiaig SU110F:TIVE ACFION ATTL‘5FS DY
- _ . AS SISI MA'CE,,4 SLPPORT, in form of reflection im LL'HNCE 30
-LGJ EQSITIVE GIVL ST“ " TREAT CENTER clarifying, indirect L1D f::\D1
. ”“ ‘ PATEFFAL, encouragement Ibii'LfierT
SHARING -HATEPFAL CH’EN'EFDE D Ng§* ACTS AEBIT PATIO:
- . * DlfiLCT‘LJH‘ DIRLCL'IVL 15A. AND COB'CILA‘ITCI
1 L COURACBMEUT AED C;UESTIOA S

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

,, 3 £3“ . ASK w l. ANSWER 'ASK *
\O V [‘3‘ QUESTION QUESTION OPINION I,
,-; \\‘ ' ‘
. :\ 61o T’ (SLL Nbb‘l‘fiim, CATLCOLY)
Juinuggft’J. In demanding, as eJCIwez-t3 88 Viv: gefignding
. - D 7, H "3Y: pushing spokesman for - yi a
hIGi’i JLLGHLLY for answer; others 1680.61"
IOSITIVE has”ST do w/ . ‘ ‘
‘ grp task -
HIGH “EGATIVR
SLIGHTLY . ' '_'.
HIGH .- - ~ -
its'EGA‘L‘IVEJ" . ,
TO EQUAL J ‘7- "
as passive #
criticism;
caSt doubt,
L0?! [EGAl 'IVL fjnd fault . _
SLIGHTLY ,-
LOW KLGATIVL
TO {
NLUTRAL‘ , . .
* '- --*-w' - ~-~ --— -..-..L_.-‘ _......‘-‘-....,._.---_..--..,.__,.___,__ -L ’
,Tinkiormfof in dependgnt .-
‘ ias ngo or., way; ask.aid;
IEUTRAI.épermiS ion, in look to other
LOW Idependent way; as eXpert-" ,,,,,,
wOSITIVL look to other , '
as-expert
BLIGHBLY , W as a bidI‘Or
I'Low , " (ML, be MK) {alliance ' L, _
POSITIVE -. ' . ' '
“1‘0 EQUAL
L“___ ,A l . ' I __
in friendly'" E. in nurturi e F '
- - supportive E §e-ré£k3 suprertivengay; ‘
21131, pQSITIVL ‘égziicomem WI (Men! in: NX) indirect encour-
ngs; en- a“ ‘ ' a ement; in ’- '
couraging res- N farm of sugges-‘
l ponse. I [.- htion - '

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EOSITIVL

awm Wiw‘b mfim; an.“ 4

s a bid for .
‘riendly allianee_

'33)

as’a bid for

friendly alliance :

I :«IW' ‘1 . ewe
F FQVJ 5 ,\5‘ } GIvE OPINIGM . GIVE SUGGESTION INFORMATION
‘6'! 1510” 5 ' -
’ - I“ a A -
6 ; 'IsnTpp Li £18 CXI‘ CI‘ in 33 expel-At , 811th“ 8 S exr'ert , Lilith"
E “TO . lyzositive,£ assert- ority, in posi- oyity, W/“elab‘
1H,:- ‘ fl: fTVG 1.LY1;tive way, as tive, assertive oration, 88361‘6 ‘
; . * £351??th gstatenwent Way tive‘, pOSltl V0 1
* I 3 . I =
4 2 ,
é gas personal yref- . #_
; ‘erence w/ hostil-
§ fiity or attention ;
....... ELJGATIVngettiHG tone; ' ' '
g 'f na’rcissistically i
; s . _
3 5% -
ell-3431...}: E z ;
”IQ-i; ; i
1-. u‘JATI‘IL 3 ‘ . :
:3 24079.; i g : , .
a E ' l i ,
! five and take .j ""_ ’ 6
g . gbaek idea by find- I
I ging fault w/ idea 1
193 .nGATIVn§(ST ace6mpanied by
‘ laughter) \
i . .
i ‘ . . i :
ESLIGH'Dngfi-VC and take back _ f -. :
EZQGATIVL’idea by diSparag- ‘ j i
_\, TO gins self E t
IEETRALE . .fi _ g
D ‘ ‘ r“ - "
_ g ' t. g: V * _r‘;4
i? ‘.;
-04 TO % .g
POSITIVE g
23
-——~ 'E
g , 1 I 7“

_‘ammm-‘pwv ‘ ‘ 'l. '

 

.
O.-.-.

 

0
f iin offering way {in offering WBY: in w/ elaboration
; :rather than as 1questioning, tone and w/ fr encly
' ”Itatement; tentative, rather than as statement tone

i

i

EOSITIVL :nith pauses, halting,

untnsstbere is t0p1.e

» 1

"How about

"L J’ha t a bOllt. o o

 

 

gshift, then AP

I”

Smwblttm,,

" ""Can we.;"

i

 

 

 

 

 

 

 

 

 

 

 

 

 

1' I “I?“ . 5 GIVE ‘ . .CONT-RDL . ,
’ “)5 .. 5%" PERMISSION . :OTRER CORRECT CRITISIZE
I .1 ‘ vi}. . .' o .
z "‘1 ‘h av ' 91'._o:.i’.:. .,.f".". d W- ....- '
L Vf‘.”ve;"v.V'I
I ,,“ fl” . G m"' . f a
1 NEUVRAL in pdsitive n‘ as leader, boss as expert, in straightfo
f\ To .. assertive way; _leso, dictate; informative ~ard,hpqsl
?’ HIIHi :LJGHTLY grants,-ccn~¢ routine control WHY) deal W/' idea ratta-
‘ 031TIVLsents, gives requests routine; idea ~than yerso:
freedom to do as Spokesman; ' ‘raise obje
~ 'delagate authority tion to so
3 k Esticn
autocratic and '
j exploitive, w/ .
: hostility, direct chastize
f“ hIGr _IN'..GIZ." l J. CL -
U
i
3
i A I
ELIGHTLY; .as attack on I.direct, 0V4
hIGh ; a 3- .person hostile or;
jf% {‘ISGAQIVt tws’person
“W10 EQYAL: } can he milc
5 'too
w 5 '
i ' - ~ ' 4:
Q 5 ‘33 passiVe W/y as-fassive by comelaiz
E . holding act resistant, Ob- ingitfincir
.4 Low -;I.'GA'BIVL " - structionistic d356, 13.5“"
J; \; i80t 7 fr C ’
j 1 I" . YES. 0 0
j ' -
I i " " ‘
smegma" .. _ ' a
E lthATIVL
‘ : my: To
. . :IEIITRAI.
F: ‘ { ' . v
f 1‘1AJULRI3L ‘, -
5 L02 TO " t
x ' POSITIVE :
_, -m1i - ....... ._-._.:-__;..-..._,.._.~__
SI”GETLY’ - .
LU? '
POSITIVE
TC- 37575 L .
/
.1 ~ 0 ' ‘ - _ , c
g: igiiggij’in- as eenevolcnt in friendly, ’ benevolent
iiCH 9CSI”1VLdu§ient‘wE§ ' I leader; deal.ujt11 benevolent {ver§ a.e‘
. o .33.. pEIson s feelincs ' i

 

 

 

“rauher~than ideas

.t

 

way

 

 

 

V} ‘

‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

W ‘2 V «J 5 I 5 -- -. I
(\) t) \ PRAISE BLAME AGREE DISAGREE
. \5 .\)fi ,
fl ' 2 . .
j (A
:LDTRAL‘as expert, as eXpert, frmn ‘88 expert; and
T0 ' leader,+tws position of authr giving infO;
.lGh PLIGhTLY ic:ea more ority] give_info, twd idea, w/ no
{OSITIVE ttmn person confirm ‘ hoStility
w/ condescension,
as personal pref
:~¢ 7: , erence, in nar-
‘*” ‘““ATIVL cissistic way;
tllthLY in overtly overtly, directly
IIGh hostile way w/ idea; cen be
KLGATIVL - mild in form, and
1% E VAL Ewe person
' ‘aS-pESSiVG M/'whiney tone,
fomplsint; W/holdi.ng, obstrJ
- . ._p _ '. ndirect' tionistc way;
. L0. . uATlVL { . ,' indirect
‘in self effacing
SLIGHTLY WEYF yield
LC}; 7‘.;-.CATIW:J ' ' ._
5 TO :.
INEUTRAL
-___ JL_ 1 .--+ 5 a
in respect- ' as requested, pos-
- ful way, w/ itive conformity;
FEUTRAL awe ~ with resp_ect __ _ -_ _
LOW TO '
POSITIVE ;
"T ‘ AW ,___ __. ....__L .__.~.- — -. ..-...«.-...- 5......- - - - -~ 0 — -
érrww in affilia— .' ” in affiliative'
when“?! . . . ’
3 L0' tive way '_ siding way; concurJ
-} " ‘T m .. assent, approve
\KJL; FLSJOSILIVE (to person)
"\.__ ll L ' - -
g in sugportive,i in encouraging,
E indulgent way,‘ supportive way
I _7 _. t, (I ..,.. direct, We ‘1'
f '* J“ b #91 iv;- .erson (‘51.. .M—M 1: ; - fl
3

 

Table C

T & T's Neutral Scoring Category Descriptions

1. Neutral Exchange Speeches (NX) ". . . consists of those
speeches which appear to be oriented primarily toward further-
ing the group task and which involve either the seeking of or
the giving of factual information. These messages are delivered
in such a way that the speaker appears to be taking a neutral
position in regard both to his own status and to the value
he attributes.to the person being addressed. The following
are examples of two such speeches:
Son to Father: How long would it take us to drive
from here to Santa Cruz? (Spoken in
a neutral tone)
Father to Son: About one hour, depending on the traf—
fic conditions. (Also spoken in a
neutral tone)."
2. Unclear Meaning Speeches (UM) ". . . consists of speeches
delivered in a relatively neutral tone which are unclear in
meaning because of the language used or because of the ap-
parent incongruity with the context of the conversation.
Because the rater does not understaid exactly What content
meaning the speaker is trying to convey, the rater is unable
to make a judgment about the interpersonal aspects of the mes-
sage. Speeches in this category may range from those which
are vague or garbled without seeming peculiar or unusual to

those which not only are unclear, but also seem quite bizarre

in terms of their lack of fit with the context of the

74

75
discussion. The following is an example of a speech which
is unclear, but not particularly bizarre:

Son to Parents: We don't—-we don't go-—where do we
go? You say we've got to have something
for the family. Where are we going to
g0? We're going to go someplace--
some relative's house. (Spoken in
relatively neutral tone)."

3. Neutral or Ambiguous Tone Speeches (NT) ". . . consists

 

of those speeches which seem relatively clear in their content
(lexical) meaning, but in which the interpersonal aspects are
ambiguous primarily because of the flat or neutral tone of

the speech.

In some instances, the ambiguous tone of the speech may
make it difficult for the rater to decide just how the speaker
intended the statement to be taken. For example, one mother
in the context of a discussion about a trip to Mexico, sug-
gested that the family go "below.Mexico City." The oldest
son responded to this suggestion with the comment, "Let's
get out in the sticks." The flat, ambiguous tone of this
comment made it difficult for the raters to know whether he
was really supporting his mother's suggestion, or whether he
was being sarcastic and critical.

Another kind of neutral tone speech consists of re—
marks which might be described as "musings," or "asides."

For example, one daughter, after the family had been talking

about the possibility of a visit from the grandparents, said,

76
"I wonder if she'll (grandmother) come out here?" This
speech was delrered in a quiet, musing tone, almost as if
she were talking to herself or thinking out loud. In some
cases, these musings or "asides" seem to represent an ack-
nowledgement of or a reflection upon the preceding speech.
The tone of the speech is such, however, that the speaker
appears to be avoiding any clear-cut position in regard to
the previous speech.

Although the neutral tone category may appear to in-
clude a rather wide variety of speeches, they do seem to
have a common prOperty in terms of the interpersonal stance
taken by the speaker. The ambiguous or neutral tone employed
seems to represent either a toning down or an avoidance of a
clear-cut interpersonal position vis a vis the person being

addressed."

4. Speeches Lacking Information (LI) ". . . consists of those
speeches which lack sufficient information for the rater to
judge the interpersonal meaning. They may be either too
brief, (one or two words long), or incomplete (the entire
speech may be just a fragment of a sentence). An example of
the former would be, "Uh, oh,'I and an example of the latter

would be, "Well, that's the way. . . " (trails off).

Table D

Speeches Rated Into T & T's Categories and Placed
in Charted Form; for Families A Through F

77

S1 - Rick, 12

Family A

Speeches Given by

Mother to:

Father to:

All Total
A B A B

S2
B

Total

All Sl&52
A B A B

82

 

14

9-10
mm

78

Nc3r40<3c1N<3c>o
®r4F40r4F4NCDC>Of\
H N
H N
.4 H mcv
flqul F4©CV h
[—1
H N m
H
N m
ocociHCVdH¢c>o<D
r40r4fl7mKDKDOCDWTO
m
H H
H
Hcv m t\
H
H qu
F4
F: H
H n4 F4¢
thcv N
FthvFaN h
mcbham C>H Eié .
CImIEEJESZFJEEZ u

To

 

All Total

Total

All

Son2 to:

SIMCDCDNCVCDOCVF40F4

mffirhfirchvc>N<3cfi¢g

I—IN

u—lmv—l

F4
F4
H .4
H '4”
Hrfl
mu—l
H H

¢r40JNr4W10CDVWVCDO

NHHH¢N0HmooN2

H

H
r-ir-l
r-lr-l
H
H

N
N m
N
N e

N
H «HF‘

H ram

OF4 B '

zqézxg

E-I

79

 

 

5N o N o mH MH m m h .uOB

o o o 0 x2
0 o o 0 92
H H H H o H H SD
n n v v H H N N N N H H H H HA
H H H H o 0 02
n v H H N H H v H m m H H H H N :3
N m H H N o o 0 Mb
H o H o o Hm
H o H H H H H mm
H N N H o m N H mm
N m H N N N H H H H 0m
v H H m H m N H v N mN
"Cu com "Ou Hounmsmn

Houoe HHN Q 2 m HmuOB HHN m z m

NN H m m m w mm m NH 5 OH.uOB
o o o N N x2
0 o o 0 Hz
0 o o H H ED
v m H N N N H H v m H N m m HA
H H H H m m m H N N m oz
m m H H N H H a v H N m N N H Sq
N o H H H h H N H m H Mb
N m H H H H H o v H N H Hm
H m H H N o m H H N H H N H om
o H H H N H H H mm
c m H H H H N m H H H H om
m H H N H N w o N m m N N H H mN
m N m N m N m N m N m N m N m N m N m N m N

Houoe Q o m HHN m D m Hobos HHN m o 2

"cu umnuoz "Cu umnumm

"Na cm>Ho meomem

oH .ea>mo u m
SH .ammsm u o
m NHHsmm

80

 

 

NH m m e mN v m o NH.uoa
N o N m H o s H m x2
0 o o H H 92
o o o N N so
H H H H m N H H N m H H H N HH
0 H H o 0 oz
N o H H e H H m H sq
m N H N H H m H H N H Mb
H N H H H N o H H Hm
o o o H H mm
o o o N N we
e N m H H H m H H H N om
e m H N v s m m H H N N m N m<
"Cu Ncom "Ou Hcom
Hmuoe HHe Hm z m Hmuoe HHN Nm 2 m
Hm H o m e mH mN s m mH .uoa
H s H H m N m N m xz
o o o o 92
H m H H H H o 0 so
N N H H H H o s N N HH
e o H H H H H N N H o H H oz
N o N o 0 2H
m m H N m H o H Mb
H H H H o H H Hm
o N H H H o H em
0 N H H H H H H mo
N o H H N o N H e um
OH m m H N H H e o mH mH H m H H HH HH ma
m a m 4 m a m m m m m a m a m a m 4 m a m «
Hmooe NmsHm HHN Nm Hm m Hmooa HH< Nm Hm 2
«CU nguoz «0U Hwfiwumm
“>9 Gm>Ho monomwmm N
HH .mme I m
eH .m>mum n Hm

U mHHEmm

S1 - Tony, 12

Family D

Speeches Given by

Mother to

Father to

Total

All S &S
A B A B A B

S2
A B

F
A B A B

Total
B

A

All
A B

B

 

81

Hc3r40r4r406VFHDCDO

F40CDCDOCDPMHGWOCDWWm

H
H H
H
H H:
H HqH
Hr4 94m
H
H H:
H rdH
H Nro

SIDfllfirfidH4F4MCDCDO

choaor4raocnew4c>¢Ln
H .4 4

Fl

H H

hrflrim 04Hr4

4 sq m r: oqo
H

©r\0dw m

BCVFJ r~4 HCV
N

m N

H F4 (V FJm

m anorqm CH4 B

Beamsizsszé

Tot.

 

Total

All

F

Total

All

«wmr4oachoaoroc>O<3

VH¢C>O

Or4CDOF\C>O'¢C)

N
H H
N N00
H
3 r4H
U
gunww) CVFJNCV N
O
(D e<w H m mr~
H
(nu1or4cwmr0rawcpoao
hrfifiwsrivrdrHfiFJOrdm
N
N N
H H
H F:
H H N
8 CV“) r4c3vrfi N
Hamrn4~ qur4p4mr4 r4N
G .
o n40EQCDHtfi 0+4 B
m Nanak—.525") 2.452%:

B

.uOB

Ch
H
N
KO
ON
N
N
._|k0
N
H
N‘

82

FIN
N
H

H H N
an
N
v mN

"ob nounmsmn "Ou com

Hence aw HHN m z m Hmaos HHm o z m

[\aio<wc>oc3cmunwo O
inch40r4r4chcw4 o
rdH
KONHDUWHcocqocorio O
cnuchorh4r4nw¢<3c><rN

H
HH
MNI—lu-l

 

82

m .uOE
NZ

H 92

H H ED

H HH

OZ

:3

Mb

Hm

Um

H mm

N 0m

N H N N H mN

<f'
KO
L!)
H

HH OH

H
N
Ln
l—l

mmowommooooo
HMKON
N

Q's—l
Hm
Ln
\0

 

m N
HN Q m m

m N m N m N m N
HHN Q m 2

mm VPNOHHMu—IONNHO
EiN dworiNrdanchCDCDOtx
Ha
r-INH

Fh4r4Nch4

m
N
Him <rH
N
m
N
m
N
‘8

H

"ob Hoauoz no» Hmnumm
“an cm>Hw mosommmm
oH .oHcmgdmum u o

mH .QHuumz I m
m NHHEmm

83

 

 

5N v m NH m mN N m m mH.uOB
o o o m H N x2
0 o o 0 Hz
0 o o 0 ED
N m H m H N m h H H H H N N H m HQ
0 m N H o N H H oz
0 m H o m H v v H H H m N SH
0 N H w H m H m H Mb
N N N N H H H H Hm
o m N H H o H mm
H H H H o m H N mm
H N H H H o N H H m H H mm
m N m H N H m N m o H H H N m H v N mN
"Ou com "ob umuamsmm
kuoe HHN Q 2 m Houoe HHN m E m
0N N m m m H vN o o m m .uos

o H H o o xz
o o o 0 Hz
0 o o 0 ED
H N H H H H H H H HH
H n H v H N N H H H H oz
5 m H N H H H v m H N N N H H :3
N H H H H m h v v H N H Mb
0 o o m H H H N Hm
m H N H o H H om
o o o 0 mm
H N N H N H H H H om
m H H H H H 0H m N N m N H mN

m N m N m N N m N m N m N m N m N m N m N m N

HouOE Qfim HHN m Q m HmuOB me HHN m Q 2
"Cu Hmnuoz "ob ”maumm
"Na Qm>HU mwsomwmm
HH .808 I m
mH .coumm I Q

m NHHEmm

Table E

Actual Rankings of the Functionality of Family Interaction

Rankings of the six dyads in each family from most to least

functional: 1 = most functional; 6 = least functional
Family Dyad .E Judge r' Family Dyad .E Judge r'
A F-M 5 l -.829 D F—M 6 3 -.771
F-S1 3 3 F—Sl 5 l
F-82 1 5 F-82 4 2
M-Sl 4 4 M-S1 3 4
M—Sg 6 2 M-Sg l 5
81—22 6 81-22 6
B F—M 5 6 +.6OO E F-M 2 2 —.257
F-D 2 2 F-S 3 4
F-S l l F-D 5 l
M-D 6 3 M—S l 6
M-S 3 5 M-D 6 5
D-S 4 4 S-D 4 3
C F—M l l +.7l4 F F-M 6 6 +.l43
F-S1 2 4 F-D 1 4
F-32 5 3 F-S 3 l
M-S1 4 5 M-D 5 3
M—S 3 2 M-S 4 2
81—32 6 6 D-S 2 5

Rankings of the various dyads across the six families from
most to least functional: "Overall"-all dyads

M-F F-C Dyad M-C Dyad (excl. child-child)

Family Dyad .E Judge ‘E Judge .E Judge .E Judge

A 4 3 4 5 6 5 5 5

B 3 4 l 2 2 2 2 2

C l l 3 l 1 l l l

D 6 5 6 6 5 4 6 6

E 2 2 5 4 4 6 4 3.5

F 5 6 2 3 3 3 3 3.5

r' +.89 +.77 +.83 +.99

84

85

Comparison of E's and the judge's rankings of the "overall"
functionality of family interaction with Moore's (1966) ranking.

Family Judge Moore _E
A 5 5 5
B 2 l 2
C l 3 l
D 6 6 6
E 3.5 4 4
F 3.5 2 3
r' +.79 +.83

BIBLIOGRAPHY

B I BLIO GRAPHY

Adams, H. B. "Mental Illness or Interpersonal Behavior?"
American Psychologist, 1964, 19, 191-197.

Arnold, M. Story Sequence Analysis. New York: Columbia
University Press, 1962.

Beavers, W. T., Blumberg, S., Timken, K. R. & Weiner, M. F.
"Communication Patterns of Mothers of Schizophrenics."
Family Proc., 1965, 4, 95-104.

Block, Jeanne, Paterson, Virginia, Block, J. & Jackson, D. D.
"A Study of the Parents of SchiZOphrenic and Neuro—
tic Children, Psychiat., 1958, 21, 387-397.

Caputo, D. V. "TheParents of the SchiZOphrenic," Fam. Proc.,
1963: 2: 339—356.

 

Cheek, Frances E. "The Schizophrenogenic Mother in Word
and Deed," Fam. Proc., 1964, 3, 155-177.

Edwards, A. L. "Statistical Methods for the Behavioral
Sciences. New York: Holt, Rinehart & Winston, 1961.

 

Farina, A. "Patterns of Role Dominance and Conflicts in
Parents of Schizophrenic Patients," J. Abnorm. Soc.

Farina, A. & Dunham, R. M. "Measurement of Family Relation-
ships and Their Effects," Arch. Gen. Psychiat., 1963,
91 64-73.

Ferreira, A. J. & Winter, W. D. "Family Interaction and
Decision Making," Archieves of General Psychiatry,
1965, 13, 214-223.

Ferreira, A. J., Winter, W. D., & Poindexter, E. J. "Some
Interactional Variables in Normal and Abnormal
Families," Fam. Proc., 1966, 5, 60-75.

Fisher, 8., Boyd, 1., Walker, D. & Sheer, D. "Parents of
Schizophrenics, Neurotics and Normals, Arch. Gen.
Psyghiat., 1959, 1, 149-166.

Foa, U. G. "Convergences in the Analysis of the Structure
of Interpersonal Behavior," Psychological Review,
1961: 5: 341-355.

Goodrich, W. & Boomer, D. "Experimental Assessment of
Marital Modes of Conflict Resolution," Fam. Proc.,
1963, 3, 15-24. 87

88

Guerney, B., Jr., & Guerney, Louise. "Analysis of Inter-
personal Relationships as an Aid to Understanding
Family Dynamics: A Case Report. J. Clin. Psychol.,
1961, 17, 225-228.

Hafner, J. & Kaplan, A. "Hostility Content Analysis of the
Rorschach and TAT," J. Proj. Tech., 1960, 24, 137-144.

Haley, J. "Research on Family Patterns: An Instrument
Measurement," Fam. Proc., 1964, 3, 41-65.

Jackson, D. D., Riskin, J. & Satir, Virginia. "A Method
of Analysis of a Family Interview," Arch. Gen.
Psychiat., 1961, 5, 321-339.

Leary, T. F. Interpersonal Diagnosis of Personality. New
York: Ronald Press, 1957.

Lennard, H. L., Beaulieu, M. R. & Embrey, N. C. "Inter-
action in Families With a Schizophrenic Child,"
Arch. Gen. Psychiat., 1965, 12, 166—183.

 

Lerner, P. M. "Resolution of Intrafamilial Confict in
Families of SchiZOphrenics," Unpublished doctoral
dissertation, University of Illinois, 1964.

Levin, G. "Communicator-communicant Approach to Family
Interaction Research, Fam. Proc., 1966, 5, 105-116.

Levy, J. L. & Epstein, N. B. "An Application of the Ror-
schach Test in Family Investigation," Fam. Proc.,
1964, 3. 344-376.

Loveland, Nathan T., Wynne, L. C. & Singer, Margaret T.
"The Family Rorschach: A New.Method for Studying
Family Interaction," Fam. Proc., 1963, 2, 187—215.

Mishler, E.G. & Waxler, N. E. "Family Interaction and
Schizophrenia: An Approach to the Study of Family
Interaction and Schizophrenia," Arch. Gen. Psychiat.,
19661 15' 64-69.

Moore, M. "Consistency of Interaction in Normal and Clinic
Families," Unpublished doctoral dissertation, Michigan
State University, 1966.

Morris, G. O. & Wynne, L. C. "SchiZOphrenic Offspring and
Parental Styles of Communication," Psychiat., 1965,
28, 19-44. .

89

Raush, H. L., Dittmann, A. T. & Taylor, T. J. "The Inter-
personal Behavior of Children in Residential Treat-
ment," J. Abnorm. Soc. Psychol., 1959, 58, 9-36.

 

Riskin, J. "Methodology for Studying Family Interaction,"
Arch. Gen. Psychiat., 1963, 8, 343-348.

Riskin, J. "Family Interaction Scales: A Preliminary
Report," Arch. Gen. Psychiat., 1964, 11, 484-494.

Ryder, R. G. "Two Replications of Color Matching Factors,"
Fam. Proc., 1966, 5, 43-48.

Ryder, R. G. & Goodrich, D. W. "Married Couples Responses
to Disagreement," Fam. Proc., 1966, 5, 43-48.

Schaefer, E. S. "A circumplex Model for Maternal Behavior,"
J. Abnorm. Soc. Psychol., 1959, 59, 226-235.

Schaefer, E. S. "Converging Conceptual Models for Maternal
Behavior and for Child Behavior." In J. C. Glidewell
(ed.) Parental Attitudes and Child Behavior, Spring—
field, Ill.: Charles C. Thomas, 1961.

Singer, Margaret T. & Wynne, L. C. "Differentiating Char—
acteristics of Parents of Childhood Schizophrenics,
Childhood Neurotics, and Young Adult SchiZOphrenics."
Amer. J Psychiat., 1963, 120, 234-243.

Singer, Margaret T. &‘Wynne, L. C. "ThoughtDisorder and
Family Relations of Schizophrenics IV, Results and
Implications," Arch. Gen. Psychiat., 1965, 12, 201-212.

Singer, Margaret T. & Wynne, L. C. "Thought Disorder and
Family Relations of Schizophrenics III, Methodology
Using Projective Techniques," Arch. Gen. Psychiat.,
1965, 12, 187-200.

Singer, Margaret T. & Wynne, L. C. "Principles for Scoring
Communication Defects and Deviances in Parents of
SchiZOphrenics: Rorschach and TAT Scoring Manuals."
Psychiatry, 1966, 29, 260—288.

Singer, Margaret T. & Wynne, L. C. "A Comparison of Parents
of SchiZOphrenics, NeurOtics and Normals; Results
with Rorschach and TAT Manuals Scoring Communication
Defects and Deviances," (in preparation) 1967.

Stabenau, J. R., Tupin, J.,‘Werner, Martha, & Pollin, W.
"A Comparative Study of Families of SchizoPhrenics,
Delinquents and Normals," Psychiat., 1965, 28-48-59.

9O

Strodtbeck, F. L. "Husband-wife Interaction Over Revealed
Differences," Amer. Sociol. Rev., 1951, 16, 468-473.

 

Terrill, J. M. & Terrill, Ruth B., "A Method for Studying
Family Communication," Fam. Proc., 1965, 4, 239-290.

 

Winter, W. D., Ferreira, A. J. & Olson, J. L. "Story Seq-
uence Analysis of Family TAT's," J. Proj. Tech.
& Pers. Assessment, 1965, 29, 392—397.

 

 

Winter, W. D., Ferreira, A. J. & Olson, J. L. "Hostility
Themes in the Family TAT," J. Proj. Tech. & Pers.
Assessment, 1966, 30, 270-274.

 

Wynne, L. C. & Singer, Margaret T. "Thought Disorder and
Family Rleations of SchiZOphrenics: I. A Research
Strategy," Arch. Gen. Psychiat., 1963, 9, 191-198.

Wynne, L. C. & Singer, Margaret T. "Thought Disorder and
Family Relations of SchiZOphrenics: II, A Classifica-
tioncf Forms of Thinking," Arch. Gen. Psychiat.,
1963, 9,199-206. '