ABSTRACT

PARAMETERS OF NORMAL FAMILY
COMMUNICATION IN THE DYAD

BY
Robert Allen Mark

The purpose of this study was to explore the com-
munication patterns of "normal" husband-wife dyads. Based
on research begun by Gregory Bateson on Complementarity
and Symmetry, an interaction coding scheme was developed
which codes (a) speaker, (b) speech, and (c) that speech
as a response to the preceding speech. A set of rules was
created to reduce the resultant eighty-nine potential cat-
egories down to nine relational codes. The resulting rela-
tional frequencies were derived from pairs of category
codes or dyadic exchanges.

An exploratory study was conducted with fifteen
upper and fifteen lower class Caucasian couples, with chil-
dren under ten years of age. They discussed three relevant
tOpics requiring a decision on a plan to action to deal

with them. The tOpics, which they had to rank-order by

salience, asked what they would do if: (1) their child was
overheard discussing sex with a friend using misinformation;
(2) night-time television programming began to offer pro-
grams for which their children wanted to stay up late; and
(3) what they would do with their family in case of a Civil
Defense emergency.

Data from these ninety dialogues was analyzed by
time, salience, class, sex, and pattern differences on both
aggregated and individual cases. It was found that couples
possess patterns or rules of communication and that these
patterns are not only discernable in the normal population
but consistent for these dyads over topics varying in sa-
lience. Strong evidence is presented indicating relational
consistency in both aggregate and individual cases. Rela-
tional differences between upper and lower class respondents
demonstrate that they do, in fact, communicate to their
spouses differently. Upper class respondents demonstrated
a greater variance in pattern useage than their lower class
counterparts. There was a significant difference, also, in
the frequency of overt disagreements among lower class
couples, while upper class couples had significantly more
covert disagreements (disconfirmations). In addition, sev-
eral major interaction patterns were found in this sample.
These included: stable complementary, stable symmetrical,
asymmetrical competition toward one-down and symmetry, es-

calating symmetry, and fluid.

PARAMETERS OF NORMAL FAMILY

COMMUNICATION IN THE DYAD

BY

Robert Allen Mark

A THESIS

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

DOCTOR OF PHILOSOPHY
Department of Communication

1970

\L COpyright by
ROBERT ALLEN MARK

1971

To Carole

ACKNOWLEDGMENTS

A dissertation is a difficult task and takes many
divergent pieces, stuck together with patience, will-power,
knowledge and faith to make it into an understandable whole.

This dissertation might have still been locked in-
side my head without the guidance of my Committee Chairman
Dr. R. Vincent Farace. Certainly, I would still be working
on gathering and coding the data without the very able as-
sistance of Minna "Minnow" Schneir, Steve Moore, and Steve
Dupres. Thanks also go to my office-mate, Bob Zimmerman,
for many heavy "rap" sessions on the subject. But, all of
this would have been a heavier burden without the atmos-
phere and environment which my wife, Carole, provided.

Thanks to all of you.

iv

TABLE OF CONTENTS

ACKNOWLEDGMENTS O O O O O I O I O I O O O O O O O 0

LIST OF TABLES O O O O O O O O O O O O O O O O O 0

LIST OF

Chapter

I.

II.

III.

FIGURES O O O I O O O O O O O O C I C O O 0

INTRODUCTION 0 I O O O O C O O C O O I O 0

Statement of the Problem . . . . . . . .
An Approach to the Study of Dyadic
Communication . . . . . . . . . . . . . .

RELEVANT LITERATURE O O C ' C O C C C O O O O

The Contributions of Small-Group
Research to Family Communication
Leadership . . . . . . . . . .
Group Cohesiveness . . . . . .
Power . . . . . -.- . . . . . .
Summary . . . . . . . . . . . . . .
A Review of Family Interaction Studies
Methodological Problems Connected with
Measuring Interaction in the Family .
Sampling Situations . . . . . . . .
Interaction Units . . . . . . -,- .
Selection of Subjects . . . . . . .

RESEARCH DESIGN . . . - . . . . . . . . . .

General Study Design . . . . . . .
Sampling Procedure . . ._. . . .
Data Collection . . . . . . . . .
Operationalization 0f Variables .
Analysis Plan . . . . . . . . . .
Summary . . . . . . . . . . . . .

Page
iv
vii

viii

13

13
14
16
17
20
21

26
26
29
30

32

32
33
34
36
49
50

Chapterw Page

IV. FINDINGS . . . . . . . . . . . . . . .‘. . 51
Introduction . . . . . . . . . . . .'. . 51
Coder Reliability Check . . .*. . . . 55
Description of Normal Population
"Speeches" . . . . . . . . . . . . . . 56
The Significance of Time . ... . . . . 61
Salience Findings . . . . . . . . . . . 62
Class Findings . . . . . . . . . . . . 65
Sex by Class Findings . . . . . . . . . 70
Successive Interactions . . . . . . . . 74
Relational Analysis . . .‘. . . . . . . 79
Individual Consistency over Topic
Salience . . .,. . . . . . . . . . 84
Individual Cases . . . . . . . . . . . 86
Summary . . . . . . . . . . . . . . . . . 92
V. SUMMARY AND CONCLUSIONS . . . . . . . . . . 94
Review . . . . . . . . . . . . . . 94
Other Statistical Analysis . . . . . . . 99
Future Research Suggestions . . . . . . . 102
Short-term Studies . . . . . . . . . . 102
Long-term Developmental Investigation . 104
Larger than Dyadic Groups . . . . . . . 105
Conclusions . . . . . . . .,. . . . . . . 107

BIBLIOGRAPHY o o o o o o o o ' o o o o o o o o o o o 109

APPENDIX 0 O O O O O O O O O I O O O O O O C O O O 115

vi

Table

1.

10.

11.

LIST OF TABLES

Topic salience by class . . . . . . . . . .
Categorizing reliability check . . . . . .
Normal population code frequencies . . . .
Length of time talked . . . . . . . . . . .
Code frequencies by salience . . . . . . .
Code frequencies by class . . . . . . . . .
Significant sex by class findings . . . . .
ANOVA table for code 26 . . . . . . . . . .

Frequencies of paired interactions
(significant only) . . . . . . . . . . .

Rational code frequencies by class and sex

Male-female comparison over topic salience,
by Class O O O O O O O O O C C C O O O 0

vii

Page
54
56
58
62
63
66
72

74

76

81

82

LIST OF FIGURES
Figure Page

1. Comparison of class variances (an
approximation) . . . . . . . . . . . . . . 80

viii

CHAPTER I

INTRODUCTION

Statement of the Problem

 

Recently, the direction in Psychiatry, Clinical
Psychology, and Communication has shifted from an emphasis
on the individual and his internal state(s), to the study
of interpersonal processes. Historically, the shift re-
ceived its initial impetus in the late thirties with the
emergence of object-relations and transaction theory.
Stimulated by the belief that Freud's View of man failed
to conceive of man in relation to a world peopled by others
who act as centers of reorientation to the objective uni-
verse, George Herbert Mead developed the concept of the
"generalized other" which served to mediate "my" view of
"myself";1 Cooley conceived of "the looking-glass self";2

and Dewey formulated a "transactional" view of man in

society.3 Sullivan's view of personality--"the relatively

 

1G. H. Mead. Mind, Self, and Society. Univ. of
Chicago Press, Chicago, Ill., 1934.

2C. H. Cooley. Human Nature and the Social Order.
Free Press of Glencoe, 1956.

 

3E. Becker. The Birth and Death of Meaning. Free
Press of Glencoe, 1962.

enduring pattern of recurrent interpersonal situations
which characterize a human life"1 was the next step toward
the emergence of the study of communication and communica-
tion systems as a unique process.

But what is it that is different about this latest
approach? Perhaps the most important change is that of
viewing a relationship as an interaction rather than in
stimulus-response or object-relation terms. Philosophic-
ally, the meaninglessness of the category "I" without its
complementary category "you" was developed by Martin Buber.2
Experimentally, it suggests that the smallest unit of
experimentation is not the individual, but the dyad, and
data must be gathered on that system as the elemental unit.3

Another central difference of this new approach to
the study of man, is the concentration on relationships not
as isolated and independent units for analysis, but as an

integral part of communication systems.4 Hall and Fagen

view a system from a General System Theory perspective as:

 

1H. S. Sullivan. Conceptions of Modern Psychiatry,
reprinted from Psychiatry: Journal of the—fifology and
Pathology of Interpersonal Relations, 3, 8, Feb. 1940 and
May, 1945.

. 2M. Buber. I and Thou. Charles Scribner's Sons,
N.Y., 1958.

 

3K. Krippendorff. "On Generating Data in Communi-
cation Research," presented at the National Society for the
Study of Communication, February, 1969.

4A. D. Hall and R. E. Fagen. "Definition of Sys-
tem." General Systems Yearbook, L:18-28, 1956.

a set of objects together with relationships between
the objects and between their attributes.

For a communication system, based on minimal Gen-‘
eral System Theory requirements, the necessary elements
would appear to be two mutually exclusive "sub-systems,"2
sharing, at least in part, a common sign-code system with
relationship rules. The system maintains itself via these
relationship rules which characterize the systems inter-
dependency and feedback mechanisms.

A study of relationship rules involves the study
of a matter-energy transformation which is both patterned
and symbolic3 involving the observation and analysis of
the total system without reduction of said analysis to
component or subsystem terms--one of the methodological
faults of most current studies (to be discussed later).

A system which is maintained by the passage of patterned
and symbolic matter-energy transformations via some
message-medium over time can be termed a communication
system. One set of these transformations can be termed an
interaction. A series of these sets would identify a

transacting communication system.

 

lIbid., p. 18.

2J. G. Miller. "Living Systems: Cross-Level
Hypotheses," Behavioral Science, Oct. 10, 1965.

3D. K. Berlo. "The Context for Communication
Study," a recent essay, Michigan State University, 1969.

 

Implicit to a system is a span of time. By its very
nature a system consists of an interaction, and this
means that a sequential process of action and reaction
has to take place before we are able to describe any
state of the system or any change of state.1

The importance of the systems point of view to
communication is twofold. First,it focuses attention on
the interacting system, whether it be the dyadic friend-
ship or relationship, the family, office, or organization,
etc., rather than solely on an individual, he and his
"object relations," or each of the individuals within that
system taken separately. Secondly, it views systems as
possessing certain similar characteristics across levels.
Thus, much of what may be found at the family level, where
much of this interaction systems analysis has been con-
ducted, should also be found in other, more complex
systems.

The study of communication behavior in ongoing
relationships provides an opportunity for a new type of
experiment, with the potential for the discovery of var-
iables significant to the understanding of human behavior.
These variables are eliminated when interaction studies
focus on individual behaviors. This thesis will attempt

to describe the patterns of communication utilized by

normal husband-wife dyads when they resolve conflict and

 

1H. L. Lennard and A. Bernstein. The Anatomyyof
Psychotherapy. Columbia University Press, N.Y., 1960,
pp. lj-14o

 

reach decisions. The only concepts which can be considered
are those which (1) offer a systems level perspective, (2)
offer a communication orientation, or (3) are useful in
extending our understanding of groups-with-a-history to

the discovery of order in their communication behavior.

An Approach to the Study of
Dyadic Communication

 

Much of the current family research has been con-
ducted with abnormal or pathogenic families. These fam-
ilies, in theory, utilize a more rigid and restricted set
of communication patterns which are consistent over topics.1
In the so-called "normal" families these same patterns are
found. However, for them, theoretically, the ability to
choose among patterns for problem-solving (coping) remains
a possibility.

As the family, or other small groups with a history,
communicate at two levels--content and relationship--2 they
pose a special problem for the researcher.

Most standard existing assessment techniques are

unsuitable for assessment of groups-with-a-history where

 

1D. Jackson and P. Watzlawick. "Development of a
Structured Family Interview," final Progress Report to the
National Assoc. for Mental Health, Jan., 1966.

2P. Watzlawick. An Anthology of Human Communica-
tion. Science and Behavior Books, Inc., Palo Alto,
CaIifornia, 1964.

 

the focus is on such things as relational communication
and its effect on the emotional stability, message inte-
gration, and system maintenance. A new technique has been
emerging from family therapy, and it will be the purpose
of this section to present the concepts and methods that
have led to it. This approach is distinguished by its
evaluation of interaction patterns of whole and partial
families as systems. It requires that researchers simul-
taneously focus on two or more family members in terms of
their transactions.1 The aim here is to find out not only
how the individuals characteristically respond to certain
kinds of stimuli, but also what types of stimuli these
family members characteristically present to one another,
and in response to what. The process of the interaction
is of prime concern. Arthur Bodin suggests that in this
kind of analysis:
family members are viewed as participants in interac—
tion sequences that cannot be understood in purely
individual terms, because such event chains cannot
occur in isolated individuals, except perhaps at an
imaginary level. Though the fantasied relationships
of family members may be interesting and important,

their actual relationships are at least as important,
and cannot be investigated merely by attempts to

 

lI. Boszormenyi-Nagy. "A Theory of Relationships:
Experience and Transaction," pp. 33-87 in Intensive Family
Therapy. I. Boszormenyi-Nagy and J. L. Framo (eds.),
Harper and Row, Pubs., N.Y., 1965.

integrate individual family members' fantasies into a
coordinated picture of the crucial family facts.

Of greatest concern to those studying communica-
tion interactions, is the structuring of a situation most
conducive to revealing the group's natural mode of inter-
action. Haley has created perhaps the most exhaustive
list of criteria for family experiments:

(l)- The experiments must deal with the responses of
family members to each other rather than their indi-
vidual responses to stimuli from the experimenter.
The experimenter must require family members to
interact with each other.

(2) At least some of the experiments must be of such.
a nature that any one family will behave in a consist-
ent way in that experiment over a period of trials.

(3) The experiments must be of such a nature that it
cannot be argued that intelligence, education, or
manual dexterity of the family members was a major
determinate of the results.

(4) The experiments must be such that it cannot be
argued that because one of the members is a Schizo-
phrenic (or otherwise mentally 111) that the results
must inevitably follow.

(5) It must be a type of experiment which a family
will participate in, willingly or not. Thus, the task
must be something that every member of the family can
do.

(6) The experiment must be of such a nature that it
does not impose patterns on the family by forcing them
to change under duress their typical patterns.

 

1A. M. Bodin. "Conjoint Family Assessment," in
Advances in Psychological Assessment. P. McReynolds
Ted.). Science and BEhaviorBCokE, Palo Alto, California,
1968, p. 223.

(7) The experimentation must involve multiple exper-
iments to measure multiple factors in families.

(8) The experiments must show extreme differences
between types of families, granted the sampling prob-
lems 1n this sort of study.

The most traditional research approach to the
study of the family's communication behavior has been to
administer "paper-and-pencil" or other similar conven-
tional scales to individual members of the family.2 Sev-
eral techniques, while relying on these individual
measures, have also aggregated them. This produces indi-
vidual scores, which can be compared with measures of
other members of the group, so as to produce a meaningful
picture of interpersonal patterns, though not actually
interactional. Such methods include the Self-Disclosure
Questionnaire3 and the Interpersonal Method.4

Another approach to the analysis of family commu-

nication patterns has been the "conjoint" or interactional

approach. Within this focus studies have utilized

 

lJ. Haley. "Family Experiments: A New Type of
Experimentation," Family Process, 1, 1962.

2J. L. Framo. "Systematic Research on Family
Dynamics," in Igtensive Family Therapy. I. Boszormenyi-
Nagy and J. L. Framo (eds.), Harper and Row, Pubs., N.Y.,
1965.

 

 

3S. M. Jourard. The Transparent Self. D. Van
Nostrand Co., Princeton, N.J., 1964.

4R. D. Laing, H. Phillipson and A. R. Lee. Inter-
personal Perceptions, Springer Publ. Co., N.Y., 1966.

 

 

primarily subjective methodological techniques. These
rely heavily on the ratings of judges. The Bales' inter-
action categories have often been used for these purposes.1
Leik utilized this method to study interaction in
terms of sex-role differentiation and effects of consensus
and satisfaction.2 O'Rourke utilized it to investigate
the premise that family interaction would vary as a func-
tion of social content.3 Although the former experiment
was exciting because of its approach to integrating the
work conducted on authentic families, artificial families
and 29.222 groups, and the latter because it attached the
effects of the variable "situational context" on interac-
tion, both rely on the Bales' observation system for data.
The Bales' I.P.A. Technique is a means of classi-
fying behavior, act by act, into twelve categories, which
can be combined into special ratios.4 This system forces
each unit of behavior into a single category, even though
the ambiguity of that behavior may be maddening to some

family members. Also, Winter and Ferreira suggest that

 

1R. F. Bales. Interaction Process Analysis.
Addison-Wesley, Cambridge, Mass., 1950.

2R. Leik. "Instrumentality and Emotionality in
Family Interaction," Sociometry, 26, 1963, pp. 131-145.

3J. F. O'Rourke. "Field and Laboratory: The
Decision-Making Behavior of Family Groups in Two Exper—
imental Conditions," Sociometry, 26, 1963, pp. 422-435.

4

 

 

 

Bales, op. cit.

10

the categories are multidimensional in meaning.1 According
to their research, the Bales' system doesn't separate
instrumental and expressive dimensions in order to make
coding less inferential. Furthermore, the Bales' observer,
if he is performing prOperly, considers only the preceding
act in classifying the present one--thereby excluding
emotional overtones generated by an interrelated sequence
of behaviors.2 For the purpose of family interaction
analysis the tracing of an interaction process via sequen-
tial events is of prime concern. Flanders has maintained
this sequential approach in his Verbal Interaction Category
Scheme used in classroom situations.3 With some modifica-
tion it may even be useful in the analysis of family
communication.

Another approach to conjoint family interaction
analysis uses data derived from more objective and rig-
orously quantifiable measures. This methodological tech-
nique attempts to cut through the shared history, subtle

"shorthand" ways of communicating and mutual dependence

 

1W. D. Winter and A. J. Ferreira. "Interaction
Process Analysis of Family Decision-Making," Family
Process, 6, 1967, pp. 155-173.

2R. Turner. "The Ethnography of Experiment,"
American Behavioral Scientist, April, 1967, Sage Publ.,
Inc., pp. 26-29.

3R. Flanders. "Interactional Analysis," in Inter-
action Analysis, Amidon and Hough (eds.), MacMillan PuEI.
Co., N.Y., 1963.

 

 

11

and quantify the multiple levels of communication so pre-

“‘--..~A ,

valent in the interaction of "traditional" groups. Much
of Haley's work in coalition formation andflflexability,1
and Speech sequences in family triads;2 Jackson's theoret-
ical conceptualization of family homeostasis,3 its exper-
imental validation by Verwey,4 and the subsequent explora-

tion into the family ggid pro guo rule by Jackson;5

 

Strodtbeck's revealed differences technique requiring
reconciliation of differences of interpretation of indi-
vidual evaluations of experiences;6 and the theoretical
orientation of Watzlawick to interactional patterns, meta-

7

patterns and relationship rules with an attempt at

 

lHaley, op. cit.

ZJ. Haley. "Research on Family Patterns: An
Instrument Measurement," Family Process, 3, 1964, pp. 41-
65.

 

3D. Jackson. "Family Interaction, Family Homeo-
stasis, and Some Implications for Conjoint Family Psycho-
therapy," in Science and Psychoanalysis. J. H. Masserman
(ed.), Vol. 5, Grune andEStratton, N.Y., 1959.

4N. E. Verwey. "Relationship of Adaptability to
Interactional Contingency and Interpersonal Prediction,"
unpublished dissertation, University of Washington, 1962.

5D. Jackson. "Family Rules: The Marital Quid Pro
Quo," Archives of General Psychiatry, 12, 1965, pp. 589-
594.

 

 

6F. L. Strodtbeck. "Husband-Wife Interaction Over

Revealed Difference," American Sociological Review, 16,

7P. Watzlawick, J. H. Beavin and D. Jackson.
Pragmatics of Human Communication, W. W. Norton and Co.,
N.Y., 1967.

 

 

12

Operationalization by both Sluzki and Beavin,l and Veron2

have all been methodological approaches of a more objec-
tive systems quantifiable nature. The constructs of com-
plementarity and symmetry which emerged from this theoretic
orientation focus on different forms of control based upon
a particular systems stereotypic communication behavior.
The importance of this latter approach to communi-
cation rests on its ability to conceptualize communication
from the systems level rather than from the individual
level. At the individual level our concern is the indi-
vidual and his conceptually isolated behavior. The alter-
native systems perspective is illustrated by Scheflen:
We notice three people standing on a corner facing
each other and talking. A number of abstractions is
possible. Each has on a brown suit. Thus, brownness
is abstractable. So is humanness, or standingness, or
twoleggedness. It is also possible to make another
kind of abstraction. We can abstract relatedness,
e.g. proximity, kinship, c00peration, and so on. Once
we abstract relatedness we no longer have organismic
wholeness or individuality. We have a concept. There
are only qualities or behavioral arrangements in a
concept; there are no peOple.
Interaction or communication analysis must proceed

with this systems perspective if meaningful progress is to

be made.

 

1G. E. Sluzki and J. H. Beavin. "Simetria y
Complementaridad: una Definicion Operacional y una Tip-
ologia de Parejas" (Symmetry and Complementarity: An
Operational Definition and a Typology of Dyads) Acta
Psiquidtrica_y Psicologica de America Latina, 11, I965,
pp. 321-330.

2G. E. Sluzki, J. H. Beavin, A.Tarn0polsky and
E. Veron. "Transactional Disqualification," Archives of
General Psychiatry, 18, 1967.

3A. E. Scheflen. "Stream and Structure of Commu-
nicational Behavior," Behavioral Studies Monograph #1,
E. Penn. Psych. Inst., 1965, p. 6.

CHAPTER II

RELEVANT LITERATURE

The Contributions of Small-Group Research
to Family Communication

 

 

There is considerable dove-tailing of the research
interests as expressed by small-group and family communi-
cation researchers. Their differences are significant to
the study of communication systems. In fact, it would be
(difficult, at best, to understand the dynamics of the
family system without some of these small—group concep-
tualizations. On the one hand, small-group researchers
have explored such concepts as leadership, power, coali—
tion formation, group cohesiveness, and their effects on
communication networks in so-called "ad hoc" groups
(groups without a history or any chance for one); however,
they ignore the varieties and subtleties of levels and
intensities of interaction found to occur in longer-term
groups, such as the family. On the other hand, while the
family communication researcher is exploring more signif-
icant communication concepts, his current methodologies
lack the precision of the small-group approach. Thus,

there is something to be gained from both kinds of

13

14

research--although my emphasis will be on family communi-
cation and the way small-group research dove-tails with it,
when it does. It is ironic, however, that so many small-
group researchers agree that the family is the most basic
of primary groups,1 yet a study utilizing the family as
its subject remains a rarity in this area. As Watzlawick
et a1. state:
Besides their practical importance as social or cul-
tural institutions, such vital-groups-with-histories
are of particular heuristic significance to the prag-
matics of communication. . . .Stranger groups or
chance encounters may provide interesting idiosyncratic
material, but unless one is interested in singular
artificial, or novel phenomena, such interaction is
not so valuable as that of a "natural" network in
which we assume the properties and pathologies of
human communication will be manifested with clearer
pragmatic impact.2
The purpose of this examination of small-group
research concepts is to afford an opportunity to under-

stand the dynamics of "vital-groups-with-histories" which

influence the nature of the interaction process.

Leadership

Leadership is a concept central to much of the
work done in small groups. However, as Cartwright and

Zander state:

 

1A. P. Hare. Handbook of Small-Group Research,
The Free Press, Glencoe, N.Y., 1962.

 

2 . .
Watzlaw1ck et al., op. c1t., p. 130.

15

It is unfortunate that most of the carefully controlled
studies of leader behavior have been conducted with
temporarily organized groups where, almost of neces-
sity, members are not concerned with the preservation
of the‘group.1
In discussing leadership, they state that "nearly every
conception of leadership contains the notion that a true
leader exerts more influence on the group and its activ-
ities than does the average member."2 Unfortunately, they
make no distinction between positive and negative styles
of influence and leadership, something of great concern to
long—term groups. This aspect of "leadership" is brought
out in some of the family communication studies of path-
ological families, to be discussed later.

Another leadership pattern relevant to the study
of the family and other communication systems is that of
passive mastery of the group. This means of leadership
has not yet been adequately explored by small-group lit-
erature though quiet leadership utilizing barely notice-
able, often subtle, signals is of prime concern. In the
family, this leadership may follow prescribed role pat-

terns as suggested by Parsons.3 This hypothesis was

recently examined in a family setting by Murrell and

 

lD. Cartwright and A. Zander. Group Dynamics:
Research and Theory, Row Peterson, Evanston, I ., 9 2,
p. 496.

2

 

Ibid., p. 493.

3T. Parsons and R. F. Bales. Family Socialization
and Interaction Process, Free Press, Glencoe, 111., 1955.

 

 

16

Stachowiak. They found the significant leadership differ-
ences between clinic and non-clinic families to be
between cooperative and effective leadership versus
an ineffective leadership which results from a lack
of support and c00peration on the parts of the par-
ents. Effective family leadership seems to require
that the parents have the greater influence, and that
one parent take the more dominant leadership role,

with the other parent cooperating and supporting this
leadership.l

Group Cohesiveness

Another concept of concern to small-group re-
searchers has been "group cohesiveness." This term
usually includes the conditions under which groups have
appeal, command loyalty, strive toward common goals, and
are united against extra-group attack. Cartwright and
Zander postulated that a group's attractiveness rests on
the extent to which it is need satisfying to individual
members.2 Others, however, have found that interaction
in 29.222 groups characterized by high self-oriented need
produces more conflict, less cohesion, and less satisfac—

tion among members.3 Perhaps Gross best resolves this

 

1S. A. Murrell and J. G. Stachowiak. "Consistency,

Rigidity and Power in the Interaction of Clinic and Non-
clinic Families," Journal of Abnormal Psychology, 72,
1967, pp. 265-272.

2Cartwright and Zander, op. cit.

3T. Caplow. Two Against One. Prentice-Hall
Sociology Series, N.Y., 1968.

 

l7

conflict, when in reference to "symbiosis" in groups (i.e.,

people cohering as a group when each has something needed

to give the other) he states:
As long as those needs persist, and so long as each
has no easy alternative of satisfying those needs,
then the two will be linked. This does not mean they
will necessarily like each other; it does mean that
they will remain united whether they like each other
or not. And herein lies the strength of the symbiotic
tie.

This description certainly fits the dynamic pattern of the

family and, perhaps, other long-term associations, as well.

Power

When small-group researchers explore the concept
of social power, they become concerned with the changes in
behavior produced by different kinds of power--reward,
referent, expert, or legitimate.2 Thibaut and Kelly bring
this concept closer to the actualities of vital interact-
ing groups by distinguishing between fate control and
behavior control:

If by varying his behavior, A can affect B's outcomes

regardless of what B does, A has fate control over B
. . . (whereas) if, by varying his behavior, A can

 

 

lE. Gross.‘ "Symbiosis and Consensus as Integra-
tive Factors in Small Groups," American Sociological
Review, 21, 1956, pp. 174-179.

2J. R. P. French, Jr. and B. H. Raven. "The Bases
of Social Power," in Studies in Social Power, D. Cart-
wright (ed.). Ann Arbor, Mich., Univ. of Michigan Press,
1959, pp. 118-149.

18
make it desirable for B to vary his behavior too, then
A has behav1or control over B.

One of the most potent forms of such behavior
control is simple conformity to the behaviors and commu-
nications sent by a high power person--one who others
perceive as possessing one of the above kinds of power.
Kelman2 has studied power and manipulated the sources of
power in several ways. In one of his studies, power was
manipulated such that in one condition power was defined
as means control, in another as high attraction and in a
third as high credibility. Kelman's findings in-
dicate that internalization from a high credible source
is a more effective, longer-term method of control than
means control or attractiveness, unless the "power" source
is present at all times, to insure compliance or identi-
fication, respectively. These findings hold great sig-
nificance for groups-with-a-history as the possibility of
means control and its potential effect is even greater in
a group of this sort.

Power involving means control can be communicated

in the small group by both overt demands and subtle,

 

1J. W. Thibaut and H. H. Kelley. The Social Psy-
cholo of Groups, John Wiley and Sons, Inc., N.Y., 1959,
pp. 182-103.

2H. Kelman. "Compliance, Identification and
Internalization: Three Processes of Attitude Change,"
Journal of Conflict Resolution, 2, 1958, pp. 51-60.

 

19

covert directions. The dimensions of this kind of power
increase in on-going groups. In the family, persuasive
techniques include everything from physical punishment to
subtle and sometimes manipulative threats of love with-
drawal. Imagine the dilemma faced by'a child pressured
into taking sides in a parental argument, with the impli-
cation of rejection by the parent he doesn't support. The
extreme of this pattern is found-in the schizophrenic
child, forced to deny and distort reality by the double
binds communicated by his parents.1

Studies of formal communication patterns within
small-groups have stressed the power-and-status dimension
as being crucial in developing communication channels and
networks.2 Lyle has studied this in small task groups
and found more communication in "democratic" than "authoritar-
iafi'group, when subjects performed both relevant and irrele-

vant tasks. Utilizing the forced structuring developed by Leavitt4

 

lJ. Haley. Strategies of Psychotherapy, Grune and
Stratton, Inc., N.Y., 1963.

2H. H. Kelley. "Communication in Experimentally
Created Hierarchies," Human Relations, 4, 1951, pp. 39-66;
and J. Ruesh, J. Block and L. Bennett, "The Assessment of
Communication," Journal of Psychology, 35, 1953, pp. 59-80.

3J. Lyle. "Communication, Group Atmosphere, Pro-
ductivity and Morale in Small Task Groups," Human Rela-
tions, 14, 1961, pp. 369-379.

4H. J. Leavitt. "Some Effects of Certain Communi-
cation Patterns on Group Performance," Journal of Abnormal
and Social Psychology, 66, 1951, pp. 38-50.

 

 

 

 

 

 

20

("circle," "chain," "wheel" and "completely-connected"
networks), Cohen sought to determine what effects a change
in networks would have. His findings indicate that past
experience with a network helped speed problem solving.1
However, communication networks in families and other
groups with histories would appear to demand a more com-
plicated conceptualization. These kinds of groups exist
as intricate, interrelated systems, possessing unique
styles and subtle rules of communication, affording little

opportunity for self-generated change.

Summary

Small-group research has focused on experimenta-
tion within §d_hgg groups where they can control the var-
iables to a high degree. By manipulating amounts of
information, nature of credentials, quantity and mode of
communication, weighting participants, etc., these re-
searchers have been able to explore small group leadeship,
cohesiveness,power,and networks. Their findings work:h1§§.
Egg game situations in which the costs and involvement are low.

They have found that "leaders" exert more influence on the group;

 

1A. M. Cohen. "Changing Small Group Communication
Networks," Administrative Science Quarterly, 6, 1962,
pp. 443-462.

21

that people cohere to satisfy needs; that behavior is
controllable by utilizing different kinds of power mechan-
isms; and that different networks vary in efficiency and
according to previous network experience.' These findings
have rarely been examined in existing small groups. None-
theless, they afford us an understanding of the dynamics
of these groups. Of particular relevance are the concepts
of leadership, and power or control mechanisms which can
be viewed from a systems level. These concepts, as will
be seen in the material that follows, are of great use in
understanding the behaviors of family and other small

groups with a history.

A Review of Family Interaction Studies

Research conducted in the family has resulted in
the drawing of only a few conclusions about "normal" and
"abnormal" family interactions. This is particularly true
when we consider those findings relevant to a General
Systems Theory view of communication. Though some have
hypothesized the impossibility of husband-wife dyads dis-
guising their communication patterns in public; while
others have advised researchers to only use situations
specifically relevant to a family in order to avoid the
camouflaging of disagreement; at least one researcher
has found that families will attempt to give the impres-

sion that they are in full agreement, even when in an

22

obvious state of disagreement, while in a laboratory
setting.1 Wynne gp_§1., referred to this as "pseudo-
mutuality."2 This finding may indicate a particular
stereotypic coping behavior couples utilize to protect
private behavior from public viewing.3 Thus, this re-
sponse pattern may indicate either a methods effect or,
more importantly, may be due to the lack of centrality or
importance of the decision. It is interesting to note
here that Bachove and Zubaly found that peer groups tended
to agree and disagree more frequently than did families.4
Ferreira, in his studies of family decision—making,
found several different macro-styles. Those that emerged
from his study are: (a) unanimous decisions--where the
family choice corresponded with the individual choices of
every member; (b) majority decisions--where the family
choice corresponded to the individual choices of two mem-
bers; (c) dictatorial decisions--where the family choice

equaled only one member's preference; and (d) chaotic

 

lP. Lerner. "Resolution of Intrafamilial Role
Conflict in Families of Schizophrenic Patients," Journal
of Nervous Mental Disorders, 3, 1965, pp. 342-351.

2L. C. Wynne, I. M. Ryckoff, J. Day and S. I.
Hirsch. "Pseudo-Mutuality in the Family Relations of
Schizophrenics," Psychiatry, 21, 1958, pp. 205-220.

3R. Ryder and D. Goodrich. "Married Couples Re-
sponses to Disagreement," Family Process, 5, 1966, pp.
30-42 0

 

 

 

4J. L. Framo, op. cit.

23

decisions--where the family choice corresponded to no
individual preferences.1 This study, as replicated in
1965, included two other variables--decision-making time
and decision appropriateness. Results demonstrated that
normal families reached more spontaneous agreements, took
less time, and made more apprOpriate decisions than abnor-
mal families.2

McLeod, Chaffee and Wackman developed a series of
studies also concerned with decision-making. Their stu-
dies focused on the politicalization of the young and the
macro-interaction patterns utilized. Their fourfold
typology resembling, in part, that of Ferreira's (above)
includes: (a) protective--where the child must steer
clear of controversy and is prohibited from expressing
dissent; (b) consensua1--where the child is exposed to
and allowed to express controversy, but is constrained to
develop concepts and values identical to his parents: (c)
pluralistic--where the child may explore, express contro-

versial ideas and reach his own conclusions; and (d)

 

1A. J. Ferreira. "Decision-Making in Normal and
Pathological Families," Archives of General Psychiatry,
8, 1963, pp. 68-73.

2A.J. Ferreira and W. D. Winter. "Family Inter-
action and Decision-Making," Archives of General Psy-
chiatry, 13, 1965, pp. 214-223.

 

 

24

laissez-faire--where the child is not prohibited, but also
is not exposed to information or controversy.1

A merging of small-group and family communication
Vresearch was attempted by Bodin. To do this, samples were
selected to include §§_hgg triads of strangers and actual
family triads--forming problem, normal and synthetic family
triads. The three family types studied differed very
little in their overall game strategy, but produced dis-
tinctive interaction patterns demonstrating more agreement
and more efficient joint decision-making in real than in
artificial families, and greater parent-child agreement,
maternal influence and maternal role perception distortion
in normal than in abnormal families.2

Another set of significant findings was produced
by Jay Haley with his investigations into family triads of
ordinary communication. While allowing them to predeter-
mine their own plan of coalition control prior to the
experiment, Haley only allowed them to form experimental
alliances by simultaneously pushing their "coalition but-

tons." Success was measured by execution of a plan by

 

1J. M. McLeod, S. H. Chaffee and D. B. Wackman.
"Family Communication: An Updated Report," 1967, a paper
presented to the Theory and Methodology Division Assoc.
for Ed. in Journalism.

2A. M. Bodin. "Family Interaction: A Social-
Clinical Study of Synthetic, Normal and Problem Family
Triads," a paper read at the Western Psychological Assoc.,
1966.

25

which an agreed upon family member would win by accumulat-
ing the greatest time spent in coalition. Haley found a
significant difference between normal and abnormal
families, with the abnormal families unable to carry out
their plans.1 Thus, coalition flexibility may indicate
the homeostatic nature of such a group with a history.

In Haley's 1964 study of speech sequence in family
triads (who speaks after whom), he found a more random
state of interaction in normal than in clinic families.

As he states:
Organization means limitation, and the more patholog-
ical the more limited. Therefore, on this frequency
count the more normal families will use more of the
possible sequences more often, and the disturbed fam-
ilies use fewer of the possibilities and use some of
them more often than others. Therefore, on a scale of
deviation from random behavior, the normals will tend
toward randomness and the disturbed will tend away
from randomness.2

Other findings from family interaction studies
include: (1) more rejection and a higher expectation of
rejection in pathological, than in normal, families:3 (2)

more time and greater amounts of silence required by

pathological families to create TAT stories;4 (3)

 

1J. Haley, op. cit., 1962.

2J. Haley, op. cit., 1964.

3Ferreira, 0p. cit., 1963.
4A. J. Ferreira, W. D. Winter and E. J. Poindexter.

"Some Interaction Variables in Normal and Abnormal Fam-
ilies," Family Process, 5, 1966, pp. 60-65.

 

26

Schizophrenia may result from the combined maladjusted
communication behavior of both parents as they interact
with their child;1 and (4) patterns of interaction within
families tend to remain consistent and stable over time
and situation.2 These findings have significant implica-
tions for the continued study of interaction in groups

with a history.

Methodological Problems Connected with
Measuring Interaction in the Family;

 

 

This section deals with several selected method-
ological issues stemming from the previous review of the

literature.

Sampling Situations

 

Given the potential for family defensiveness to
the uncovering of conflict (labeled "pseudo-mutuality" by
Wynne) and the resulting high need for families to present
themselves to the social scientist as "normal," it would
appear that researchers must attempt to circumvent these
c0ping behaviors. Much of the family interaction research
to date functions under the assumption that by utilizing

either an ambiguous stimulus and/or a difference to

 

1G. Bateson. "Minimal Requirements for a Theory
of Schizophrenia," Archives of General Psychiatry, 2,

2

 

Murrell and Stachowiak, op. cit.

27

resolve, the family can only resort to its most typical

patterns of behavior to cope with this stress. As

Watzlawick states:
A family's specific homeostatic processes do become
increasingly apparent to the observer, since in the
course of a series of interviews stresses occur which
force all family members to fall back upon their
typical interactional patterns. . . .It is possible
to create such stress situations deliberately, rather
than wait for them to take place spontaneously, and
this can be done within the framework of task
performance.

However, there are several factors in the research
setting which may lead one to question the validity of»
that assumption. The presence of either the experimenter
or his tape recorder during a family interaction is cer-
tainly a powerful stimulus not yet investigated by family
researchers. If the experimental setting is outside the
home this effect may be compounded. Given that family
members will have to live with each other after the inter-
view is over, one can only assume that the less the ex-
perimental situation is like their private interaction
territory the more their public responses will be tempered.
With the ethical problem of "bugging" homes, the research-

er, at best,-can only hope to reduce his effect by gath-

ering his data as unobtrusively as possible.

 

lP. Watzlawick. "Development of a Structured
Family Interview," January, 1966, a final report to the
National Assoc. for Mental Health, p. l.

28

Even if the researcher can obtain natural communi—
cation behavior in his sample, he must cope with another
issue. Will the experimental tasks employed tap the
issues which are particularly significant to family
living? As no one has yet specified typical tasks and
situations relevant to families, it is only accidental
that a particular manipulation will represent a conflict
or topic of concern family X is coping with at that point
in time. Perhaps for now, the best the researchers can do
is present stimuli to a target audience that with high
predictability that family will encounter in the not-too-
distant future. For example, families with young children
might be presented with the problem of discussing how they
will teach "sex-education" to their children. Another
example would be to have the family plan some event to-
gether. Framo suggests two other approaches to this
problem:

A meaningful experiment would require that each
family be presented with the controversies it is
inherently struggling with, not with abstract contro-
versies which result in polite play-acting. Prelim-
inary study of the family should reveal its Achille's
heel . . . (and)

Investigation of symptom-free families in a series
of exploratory sessions, dealing with the normal crises

every family has to deal with, is one research project
worth doing.

 

1J. L. Framo. "Systematic Research on Family
Dynamics," in Intensive Family Therapy: I. Boszormenyi-
Nagy and J. L. Framo (eds.)i Harper and Row, N.Y., 1965,
pp. 433-455.

 

29

Interaction Units

Even if the researcher can reduce his effect and
select topics that are both relevant and capable of un-
covering the family‘s normal communication behaviors,
there still remains the task of appropriate selection and
measurement of the interactional behavior. Few attempts
have been made to categorize interactions beyond a content
level. The problems inherent in the Bales system, though
perhaps the most extensively used system of interaction
analysis, have already been discussed. I have also al-
ready suggested the necessary criteria for a systems
approach to interaction analysis; what remains is the
careful selection of variables, their operationalization
and an attempt to explore the practicality and significance
of using these variables for the analysis of interactions.

Probably the most successful attempts to describe
an interaction and make predictions from it will utilize
some combination of variables aimed at analyzing both the .
state of the system and the components of that system.
For instance, a combination of individual voice quality or
tonal characteristics with a measure of length of time,
sequencing and interruptions coupled with a systems anal-
ysis of the interaction without reduction to individual
measures may tell us more about relational rules, homeo-
static functioning in stress situations and patterns of

interaction than any of these taken individually.

30

Another factor to be considered in the study of
interactions is that communication in a family or other
group-with-a-history may be considered as both a dependent
and an independent variable. It is dependent in that the
background factors, personality factors, and topics of
conflict all influence the communication process. It is
an independent variable in that the interpersonal rela-
tionships established, the relational rules, are somewhat
a result of the process of communication. Communication
as a dependent variable seems logical enough. Communica-
tion as an independent variable requires us to change our
viewpoint. Locke offers support for this conceptualization
when he states:

A decided reduction of intimate communication between
a husband and wife generally results in a decrease in
family unity. This does not mean, of course, simply
a decline in the number of words spoken, for a glance
or a caress may convey more meaning than a large num-
ber of words. It refers primarily to less face-to-
face communication, resulting from the husband and
wife having different work schedules or living away
from each other. Under such conditions, it is ex-
tremely difficult to maintain companionship relations
and relatively easy for misunderstandings to develop.

Continuous communication is apparently necessary for
the maintenance of emotional attachments.l

Selection of Supjects

 

The sampling techniques available for selection of

subject families poses yet another methodological problem.

 

1H. J. Locke. Predicting Adjustment in Marriage,
Henry Holt and Co., N.Y., 1951, p. 247.

31

We assume "normality" when we find families in the so-
called "real" world who have not undergone therapy. Like-
wise, we only consider them abnormal if they are 13
therapy. Certainly a better Operationalization of these
terms must be develOped.

Two other major assumptions underlie the selection
of experimental subjects. The first is that those people
who allow a researcher to examine their private modes of
communicating are not unlike those who refuse such intense
experimental examination, and the second is that of con-
sistency or stability of family interaction over time.
Though these assumptions are necessary for the continuance
of study in this field, they remain, nonetheless, assump-
tions awaiting validation.

Given an awareness of the methodological problems
inherent in the study of communication in the family,
research should continue to determine the nature of that
communication, as well as resolve the methodological prob-

lems of research in this setting.

CHAPTER I I I
RESEARCH DESIGN

General Study Design

Though the search of the literature indicates that:

Family interaction is not frequently studied
through direct observation. Perhaps this has had to
do with the privacy of the home, with the high cost,
with the fact that observation must deal with present
history and does not delve into the past.1

this study attempted just that. Discussion-decision-
making tasks were administered to couples and their in-
teraction recorded on audio tape. Since the purpose of
this study was to examine the apprOpriateness of several
interaction constructs for the purpose of interaction
analysis and the description of "normal" family communi-
cation, no hypotheses were made. (Several expectations

are projected later in this Chapter, however.)

 

1C. C. Beels and A. Ferber. "Family Therapy: A
View," Family Process, 8, March, 1969, pp. 281-329.

 

32

33

Sampling Procedure

The respondents in this exploratory study were 30
husband-wife dyads from the Lansing metrOpolitan area
purposefully selected to have the following characteristics:

(1) each dyad was considered "normal" if no member
of the family had participated in either "in-clinic" or
"out-patient" therapy with any practicing psychologist,
psychiatrist, or psycho-therapist.

(2) each dyad had at least one child.

(3) each dyad had no children over the age of ten.

(4) fifteen families came from lower-middle to
lower socio—economic backgrounds while fifteen families
came from upper-middle to upper socio-economic backgrounds.
(This was determined by the location of their dwelling
unit and subsequent questions dealing with income, educa-
tion, job title, and job activity. The class divisions were
based on previous work conducted by E. Bergel.1)

Interviews were conducted during evenings and
weekends of the second, third and fourth weeks in April,
1970. The interviewers were six undergraduates instructed

to gather taped interviews from area residents for a public

Opinion survey.

 

lE. Bergel. Social Stratification. McGraw—Hill
Book Co., N.Y., 1962, p. 272.

 

34

Data Collection

The dyadic interviews carried out in the reSpondents'

home typically began with some preliminary remarks by the
experimenter--i.e., "I am a doctoral candidate in the De-
partment of Communication at Michigan State University

and I am currently conducting a study in this neighborhood
and would be interested in having you and your (husband)
wife participate in it. Here is a letter of introduction."
Given permission, the experimenter entered the home (or
made an appointment for some later time), made some pre-
liminary remarks aimed at setting the couple at ease,

and introduced the couple to the task, namely, the dis-
cussion of several topics dealing with problems which
families may encounter.

The discussion tOpics were selected to meet
several criteria. (1) They had to be sufficiently inter-
esting to the subjects to insure some discussion. (2)
They had to be questions for which there were no "correct"
answers. Ideally, (3) they had to permit several defens-
ible solutions and allow for differences of Opinion. And,
(4) they could be rank ordered from 1 to 3 in order of
their saliance (importance) to the dyad.

The interviewer read (and the subjects were given)
the following instructions:

You will be presented with a series of situations

which couples encounter from time to time. These
situations can be handled in a variety of ways. We

35

are interested in your opinions on these matters and
would like for you to discuss as many of the alterna-
tives that seem important to you. From these dif-
ferent views, select an alternative which you both
agree on.

There are no right or wrong solutions . . . only
what both of you agree on as the most apprOpriate
answer for the situation. Do not worry about the time,
as you may take as much or as little time on each of
the following three situations as you need. Do you
have any questions?

After reading the instructions, the interviewer
answered any questions that the respondents asked. At that
point the interviewer, after turning on the recorder, phy—
sically withdrew from the immediate discussion area, though
not out of sight. Upon completion of the discussion tOpics
the couples were instructed to reach a mutually agreed upon
decision as to which of these 3 t0pics was felt to be most
saliant, important or relevant to them, and which of the 3
the least saliant, important or relevant. At this point
the interviewer answered any further questions they had,
and left.

The three discussion topics, presented to respondents
in one of five orders, were:

Imagine that it is now September, 1970. Your child

(children) are just beginning a new school year. The
major television networks (NBC,CBS, and ABC) have de-

cided to run more "educational" programming in the
evening hours. In addition to this, there are several

36

new series on television that are pOpular among your child's
(children's) friends and they have asked to be allowed to
stay up later than they do now to view these programs. How
would you, or you and your spouse handle this situation?
Please discuss the possible alternatives and decide on a
course of action.

Imagine that your child (children) are about 12
years old and have been receiving some information
about human reproduction in the classroom, as well as
from other children. You accidentally overhear them
discussing it with a friend and you realize that their
information is quite incorrect.and misleading. They
are not aware that you overheard them. How would you,
or you and your spouse handle this situation? Please
discuss the possible alternatives and decide on a
course of action.

 

Imagine that you, your spouse and child (children)
are at home together on a Sunday afternoon, and you
are watching television or listening to the radio when
a Civil Defense alert is broadcast. This broadcast
informs you that you have only 15 minutes to prepare
and take shelter prior to disaster. How would you, or
you and your spouse handle this situation? Please
discuss the possible alternatives and decide on a
course of action.

Operationalization of Variables

In choosing a set of variables to Operationalize
and use in this exploratory study of "normal" family com-
munication patterns, this author examined many of the
current constructs. (See previous review of the litera-
ture.) As others have indicated:

There is no single systematic and comprehensive theory
of family process and of relationship of family process
to the development and sustenance of individual be-
havior . . . (likewise) the complex, relevant phenomena

in family interaction have yet to be adequately
described or conceptualized.1

 

1A. Ferber and M. Mendelsohn. "Training for Family
Therapy," Family Process, 8, March, 1969, pp. 25-26.

 

37

Or as Beels and Ferber put it in another article:
. . . the problem of accounting intellectually for the

interplay of events and sequences in a family session
is one for which there is not yet a good language.

It is toward the development of this "language" capable of‘m
meaningfully describing interpersonal communication in
groups-with-a—history that this dissertation is directed.
Though I agree with a recent review of the psychiatric
literature by Frank in which he states that pg_factors
have been found which differentiate between psychopath-
ological and normal families, I cannot agree with his
statement that:

Apparently, the factors which.play a part in the dev-

elopment of behavior in humans are so complex that it

would appear that they almost defy being investigated

scientifically and defy one's attempt to draw meaning-

ful generalizations from the exploration which has

already been done.2

Based on the review of the literature, an inter-

action coding scheme was developed. The Criteria used in
picking the constructs were: relevance, potential payoff,
descriptive power,and probable utility in a theoretical ap-

proach to interaction analysis. Some of the axioms guiding

construct selection were originally stated by Watzlawick as:

 

1Beels and Ferber, op. cit.

2G. H. Frank. "The Role of the Family in the
DevelOpment of Psychopathology," Psychological Bulletin,

 

38

(1) You cannot not communicate; (2) a message sent is not
necessarily a message received; and (3) every message sent
contains both content and relationship elements. These
axioms were updated by Beels and Ferber to include:

(1) All behavior is communicative. It is impos-
sible not to communicate, since even the refusal to
send or receive messages is a comment on the relation-
ship between peOple who are in contact.

(2) Messages have 'report' and 'command' functions.
Thus, 'it's raining' is a report, but depending on the
context, inflection, and relationship of Speaker to
hearer, it may also be a command to remember an
umbrella.

(3) Command messages define relationships. The
command aspect of communication is the troublesome
part, because it is the medium through which relation-
ships are shaped, and in this process, ambiguity, mis-
understanding and duplicity are possible . . .

(4) In families, command messages are patterned
as rules. If two or more peOple are in a relationship
for a long time, the multiplicity of commands they
exchange assumes a pattern from which rules for the
relationship may be derived. These rules constrain
and order the behavior of family members in patterns
of mutual influence which have cybernetic properties
(5) Change and stability.- If a member of a family
wants to change the relationship, the regulating re-
sponse of others which stabilizes the system by reduc-
ing change, makes it appear that the 'governor' or
conservative element in the system resides in the
person or persons resisting change. . . .1

Based on the axiom that messages contain both content and
relationship aspects, the analysis scheme allows

us to focus on the relational aspects. If we under-
stand that every message in an interaction serves as either

the definition, reinforcement or redefinition of the nature

 

lBeels and Ferber, op. cit.

39

of a relationship, then it should be possible to determine
the modal or typical interaction pattern of any dyad.
Several of the concepts offered by a group of therapists
known as "systems purists" (those who "see in the family
a system of countervailing power—-a network of influence
governed by rules which shape and constrain it."1) includ-
ing Haley, Jackson, Bateson, Watzlawick, Beavin, Sluczki
and Veron, are relevant here.
The primary concepts offered by these researchers
have been symmetry and complementarity (see Chapter I).
To analyze an interaction as being one or the other of
these necessitates a unit of measurement. A single mes-
sage is of no help as a judgment as to symmetry or com-
plementarity cannot be made without reference to preceding
or succeeding messages. As Bateson and Jackson stated:
There is, strictly speaking, no such thing as a
complementary piece of 'behavior!‘ To drOp a brick
may be either complementary or symmetrical; and which
it is depends upon how this piece of behavior is
related to preceding and subsequent behaviors of the
V1s-a-V1s.
Thus, a transaction, or the relation between two contig-

uous messages is the smallest unit of analysis to get at

relationships--i.e., in a speech sequence between

 

1Beels and Ferber, op. cit.

2Bateson and D. D. Jackson. "Some Varieties of
Pathoqenic Organization," in Disorders of Communication.
David Mck. Rioch (ed.), Research Publ. A.R.N.M.D. 42:
pp. 270-283, 1964.

 

40

individuals A and B the transaction units would be Al/Bl'
Bl/AZ’ A2/Bz, etc. This two-message unit will be used,
although Bateson and Jackson describe every item in a
sequence as simultaneously a stimulus, response, and a
reinforcement, indicating a triad of messages as the basic
unit. The choice of a two-message unit does not deny
Bateson and Jackson's belief in the 3-fold nature of a
message, it merely places that in the context of two, two-
message units--Al/Bl, and Bl/AZ' (For the first time,
2+2=3.)

For the analysis of the interaction data gathered
in this exPloratory study we will be concerned with the

structure of the content or, if you will, the "how“ and

 

not the "what." To be more specific, content may be:
structured as question, referential statement, instruction
or order, negation, or acceptance, etc. These may be
considered the generalized form in which specific content
is transmitted. Therefore, disregarding what is said, it
is possible to determine how a dialogue progresses via a
combination of the basic grammatical forms (interrogative,
declarative, and imperative) in addition to the metacommu-
nicational categories of confirmation (agreement), rejec-
tion (negation), and disconfirmation (non-existence of
relationship given a particular definition of that rela-

tionship at a particular point in time).

41

According to this scheme, some examples of symmet-
rical and complementary transactions have been listed by
Sluczki and Beavin. These include:

giving/taking instruction=complementary (giving=one-up,-
taking=one-down)

asking/answering=complementary (asking=one-down, ans-
wering=one-up)

asserting/agreeing=complementary (asserting=one-up,
agreeing=one-down)

referential statement/referential statement=symmetrical

agreeing/agreeing=symmetrical

giving instructions/countering with instructions=sym-
metricall

Given this as a starting point, this author became
concerned with other aspects of the interaction which ap-
peared to be potentially as important as what Sluczki and
Beavin were working with. Their system only allows for
coding alternating Speeches. This ignores the whole prob-
lem of mobilization, or how an individual becomes a prin-
cipal speaker.

It appeared that there were styles of mobilization
which went beyond merely waiting for the other individual
to pause or complete a thought. Individuals tend to also
get into a conversation by "talking over" someone else
until they finally "get the floor." There are still others
who "talk over" but never succeed. This aspect of an
interaction cannot be ignored. Likewise, the person who
answers his own question or ends every statement in a

question is an important form.of relational communication

that was included.

 

1Sluczki and Beavin, op. cit.

42

With these concerns in mind, a coding scheme was
developed,aimed at coding each Speech (a "talk" by an
individual of any length, beginning with his first word,
and continuing until the other party in the dyad speaks)
in the interaction. Each Speech had to be coded as to
who Spoke it (male or female), what kind of Speech was it
in terms of a modified grammatical format incorporating
other "what was it" aspects, and what it was relative to
the statement that came before it (for the purpose of
first level relationship analysis and sequencing). This
was accomplished by using a three-digit coding form as
follows:

1st digit code: male = 1
female = 2

2nd digit code: question = 1
assertion = 2
instruction = 3
orders = 4
talking over = 5
assertion and question =
question and assertion =
other = 8
laughter = 9

\lm

3rd digit code: agreement = l
disagreement = 2
extension = 3
answer = 4
disconfirmation = 5
topic change = 6
agreement and extension = 7
disagreement and extension = 8
other = 9
laughter = 0

Code all silence over 2 seconds as 000; add another 000

for every additional 5 seconds of silence.

43

Given these three digits it becomes possible to
code an individual, what his speech was and in relation
to what. An individual could only succeed himself in the
Speech sequence by either successfully interrupting (for example,
a 153 code followed by a123 code) or by making a major
shift in his speech--i.e., following an assertion or ques-
tion with an instruction or order.‘ The-first digit refers
to Speaker's sex and the second digit followsthe grammatical
rules for forms of Speech. This is true for all but
"talking over," where the "second" individual is speaking
while the "first" individual is continuing the speech he
or She began prior to the "second" individual's speech;
"other" which refers to speeches that might qualify as
"filled hesitation pauses"1 and consist primarily of "well,"
"uh," "but," "but, well, I mean . . .," and laughter,
which when it appears in the second digit comes as a
response to something the individual himself has said.

The third digit refers to what a particular speech
comes in response to. Here, again, many of the categories
refer to common sense categories needing little or no
explanation--i.e., agreement, disagreement, extension,
answer, or combination of those. Disconfirmation, a term

used extensively in psychotherapy, refers to the complete

 

lG.‘L. Trager. "Paralanguage: A First Approxima-
tion," in Studies in Linguistics, 13, 1958, pp. 1-12.

44

ignoring of something individual A has said to individual

B by individual B. It amounts to B saying symbolically by
his actions, "you (A) don't exist, given that definition

of our relationship." A disconfirmation occurs after a
statement has been made which demands a response to it by
the other individual and he then does not respond to the de-
mand either by silence, leaving the field, or a disconfirming
response. Neither Silence nor leaving the field were
codable, as silent disconfirmation would require consid-
erable inferential abilities on the part of the coders and
leaving the field was restricted by the necessity to com-
plete the interview. Disconfirming responses, on the other
hand, were coded--ex.

Wife: . . . well, now that you've heard what I

think what do you think?"
Husband: Go on to the next topic!

This interaction would be coded as: 263/145--a
female, assertion followed by a question which was an
extension of something that came before, followed by a
male order which was a disconfirmation.

A topic change, though at first glance potentially
overlapping with a disconfirmation, was used to refer to
the introduction of a new idea after discussion of the
other. In addition, as disconfirmations were limited only
to those overt verbal exchanges where a demanded response

was requested but ignored, topic changes could refer to

45

any change in a subject following a speech which did not
demand or request a response to it.

The category "other" as a response was used typ-
ically in two ways. First, as the last digit in the three
digit number of the individual who spoke first. It was
also used in conjunction with Speech nonfluencies (an 8 in
digit two) or unsuccessful attempts to interrupt (a 5 in
digit two) where it was impossible to discern agreement,
disagreement, extension, etc.

A laughter code of 0 in the 3rd place, accompanied
by an 8 in the 2nd place refers to laughter as a response
to what the other person has said.

For example, the following interaction would be
coded:

Wife: That's a very interesting situation, but I
really don't know how I'd handle it, what about you,
honey?

Husband: Well, I think that in the case of an
emergency we would probably first tune in the radio
and listen for directions and at the same time begin
to gather up some food . . . (Wife: and clothing . . .)
to take to the basement (Wife: we'd have to take
blankets) where we'd . . .

Wife: because we'd need the warmth in the base-
ment, and besides the children would need it.

Husband: right, that sounds about right.

Wife: O.K., why don't you turn the page and go on
to the next one.

269/124/253/123/253/223/121/236

Thus, we can reduce each speech to a three—digit

code and the whole dialogue to a sequenced set of three-

digit codes. This allows us to determine frequencies of

46

interaction units as they vary from normal to abnormal
subjects or across socio-economic stratifications, as well
aS probability estimates of one code being followed by
another. In addition to this, it is possible to obtain a
relational score or pattern for each dialogue or interac-
tion.

This last alternative can be accomplished by com-
bining pairs of three-digit alternating Speaker codes--
i.e., first a 1-- and a 2--; then that 2-- and the next
l--; then that l-- and the next 2--; etc. Each pair gets
assigned to a category representing one of the nine pos-
sible combinations of the three categories Bateson devel-
Oped--one-up, one-down, and symmetrical. These nine
possibilities are one-up, one-down ++; one-up, one-up ++;
one-up, symmetrical +S; one-down, one-up ++; one-down,
one down ++; one-down, symmetrical (S; symmetrical, one-up
5+; symmetrical, one-down 8+; and symmetrical, symmetrical
SS. The frequency of a particular categories occurrence
is tabulated by intering its occurrence for the individual
who comes first in the sequence so that a particular in-

teraction might be profiled as:

Male Female
++ 6 1
++ 3 0
+8 0 0
++ 0 5
++ 0 5
+8 0 0
8+ 0 0
8+ 4 0
SS 2 3

47

This would be accomplished by following a set of
rules for the second and third digits only, listed below.

1. If the second and third digits are identical
in the pair of numbers it is SS. This is also true for
pairs combining -51, -81, or -21.

2.‘ All -1-'S are coded as + (one-down) except for
-12, -15, and -18 which are + (one-up).-

3. Codes with a two in the second digit are coded

as follows:

21 = + (one-down) (except when rule 1 applies)

22 = + (one-up)

23 =} (one-up) unless preceded or followed by a

}+ 23 or a 24 in which case it is coded as S

24 =} (symmetrical)

25 = + (one-up)

26 = + (one-up)

27 = first coding is as a 21; second coding is
as a 23.

28 = first coding is as a 22; second coding is
as a 23.

29 = + (one-down) unless it is the first speech in
the interaction in which case.it follows
the rules for a 23.
4. All -3-'S and -4-'s are coded as + (one-up).
5. All -5—'s are coded as'+ (one-down) unless the
first digit in the second three-digit number is identical

to that of the first digit of the first three-digit number

(indicating a successful interruption), in which case the

 

. - mil-d
a

48

-5- number is skipped, the score is f (one-up) and the
second three-digit number is used for the next score.

6. All -6-'s are coded as -2-'s the first time
and -l-'s the second time.

7. All -7-'s are coded as -l-'s the first time
and -2-'s the second time.

8. All -8-'S are coded as + (one-down).

9. Skip all -9-'s and 000's.

10. A pair combining a -51 and a -52 are coded as
-21 and -22.

Based on this coding scheme, the previously coded
dialogue would have a relational score of two ++'s and a
++ for the female and a ++, ++, and ++ for the male.

One might say that this study will omit much val-
uable information by analyzing only the audible aspects of
these interactions. However, as long as video-tape or "on-the-
scene" categorizing of non-verbal behaviors are as obtrusive
as they currently are,efforts to code the audible band re-
main a valuable first step. In addition to this,there is no
evidence to indicate what one might gain by also coding the
non verbal band of normal communication systems. Until unob-
trusive video-tape facilities can be created the coding of
all the bands of communication will remain problematic.

The coding form reflects both the communication
and interaction concerns as well as those of the small

group researcher interested in the leadership and power of

49

small groups. In fact, with substantial validation of
this coding form it may be adaptable to small groups of
up to ten or, by adding a fourth digit, up to one hundred.

Some of these possibilities will be discussed later.

Analysis Plan

Once data were gathered an inter-rater reliability
check was conducted (to be discussed more fully later).
Respondents were separated by socio-economic status (high
and low) and their discussions separated on the basic of
saliance (high, medium and low) or importance as perceived
by the individual couple.

Given these data, this author expected to find the
following:

1. The amount of time spent discussing these topics
will vary inversely to saliance, with more time being spent
on low saliant t0pics.

2. There will be a difference in the interaction pat-
terns utilized by socio-economic classes.

3. There will be less total time and more silence
over all topics in the lower class than among the upper
class respondents.

4. There will be more interaction pattern variance
among the upper class than among the lower class couples.

5. A couple's pattern of interaction will be con-

sistent over tOpics varying in saliance.

50

6. Sequences within the patterns of interaction
will repeat themselves often enough to be separated out
as distinctive patterns--e.g., perhaps like Sluzki and
Beavin's theoretical patterns of Stable Complementarity,
Stable Symmetry, Asymmetrical Escalation toward One-down

and Symmetry, etc.

Summary

Human communication occurs simultaneously at sev-
eral different levels, along various channels, all of
which carry what we call "information." These levels have
been previously labeled audible-linguistic, audible-para-
linguistic, non-audible paralinguistic or kinesics, and
contextual. Some researchers have found it impossible to
describe any of these levels, several have attempted to
explain one level. Given that there is some interrelation
among these levels, this author hOpes to administer a
stimuli to dyads with-a-history and attempt to describe
their communication behavior at both the audible-linguistic
and audible paralinguistic levels. The assumption is that
the interrelationship of these factors may adequately
describe the interaction, determine the nature of the
relational communication, and expose some of what might be
occurring at the other two levels where progress in anal-

ysis is more incomplete than at these levels.

CHAPTER IV

FINDINGS

Introduction

 

This exploratory study of normal family communica-
tion was an attempt to gain a greater understanding of
family communication behavior and develop a coding scheme
capable of unlocking some of these behaviors for the social
scientist. To accomplish this a set of constructs was de-
veloped as a result of a review of the literature and new
coding procedures were devised to operationalize the constructs.

Though the conceptual base developed by Batesonl
and refined by Jackson2 and Watzlawick3 was considered the
most potentially fruitful approach to the family as a
system, it still left some large problems, the most fun-
damental being operationalization. The on1y_0perational

approach to this point has been Sluzki and Beavin's4

 

1G. Bateson, op. cit.

2D. Jackson, op. cit.

3Watzlawick, Jackson and Beavin, op. cit.

4Sluzki and Beavin, op. cit.

51

52

article attempting to operationalize the variables symmetry
and complementarity. The results of their efforts were a
useful first step, but left out many of the subtler verbal
aspects of communication.

In an attempt to overcome these difficulties, a
three digit coding scheme, capable of coding (a) the
speaker, (b) the speech and (c) the speech as a response
to a previous stimulus, was developed and operationalized.
This study represents the first use and reliability checki
on this method of interaction analysis. Used to analyze
the discussions of fifteen upper and fifteen lower class
husband and wife dyads with children under ten years of
age, as they discussed and arrived at decisions on three
ambiguous topics, the coding scheme helped uncover many
significant relationships to be discussed later.

The study lasted for most of the month of April,
1970, although the fifteen upper class couples were
obtained within the first two weeks. The' sample
was obtained by door -to -door _ interviewing in neigh-
borhoods classified as upper and lower class during pre-
vious community studies. Nonetheless, subjects were split
into upper and lower class categories on the basis of
income (upper class being over $8,000 per year), education
(upper class having some college education), and job title

and description.1 Dwelling unit turned out to be an

 

1E. Bergel, op. cit.

53

excellent indicator of social class. No "class" deviants
were found.

To maintain a comparable set of couples, only
American, Caucasian, married couples were allowed in the
sample. In addition, those couples responding affirma-
tively to the question: "Have you or your family ever
obtained psychiatric or psychotherapeutic assistance--
either individually, or as a family?" were eliminated from
the sample.

To obtain the thirty couples finally used in this
study, seventy-three upper class homes and well over one
hundred and fifty lower class homes were approached. Of
the upper class homes, twenty-eight upper class couples
made appointments resulting in the taping of eighteen
couples. Three upper class couples were eliminated because
of previous contact with a psychiatrist or other mental
health agent. Of the lower class families, though some
thirty-four agreed to appointments for taping, only six-
teen actually participated. One lower class couple was
not used because of psychiatric contacts.

Of the thirty remaining couples, each was instructed
to discuss three topics, as outlined in Chapter III. They
were to discuss the relevant alternatives and arrive at a
decision regarding what they as a couple or as individuals
would do to handle these situations. The three t0pics-—

Civil Defense planning, Sex Education, and child

54

Television viewing times--were later rank ordered by im-
portance or salience.
Table 1 gives the frequencies for ranking these

topics after the discussions.

Table l.--T0pic Salience by Class.

 

 

Upper Class Lower Class
Hi Middle Low Hi Middle Low
Salient Salient Salient Salient Salient Salient

 

Civil

Defense 0 l 14 2 7 6
Sex

Education 8 7 0 11 3 l
TOVO

Viewing 7 7 l 2 5 8

 

Part of the reason for the higher incidence of
ranking Civil Defense as middle salient and high salient
among lower claSS respondents, was because of two major
tornado alerts during the last two weeks of interviewing.
These last two weeks were spent entirely in the lower
class communities. The resultant Shift in attitude was
probably primarily caused by this change in their environ-
mental information. It was originally believed that Civil
Defense would be the low salient topic for the majority of

the respondents. This belief was based on data gathered

55

in 1969 for the Office of Civil Defense.1 Having made the
earlier decision to rate all the topics of similar salience
level together, rather than forcing them into some precon-

ceived ranking, became all the more desirous.

Coder Reliability Check

After approximately one-half of the tapes were
obtained coding began, based on the three-digit coding
scheme (see Chapter III for full explanation). After
about two hours of training, the four coders were sent
off in pairs of two to independently rate three tapes
(nine dialogues).

Three reliability checks were made for these nine
dialogues--a unitizing, categOrizing and total reliability
check. The unitizing check on just the first digit 0r sex
code, aimed primarily at determining if the coders agreed
on the number of alternating Speeches. This unitizing
check approached the 100% level as only two Speeches in
two hundred and seventy five speeches were not agreed
upon. The categorizing reliability check, used to deter-
mine if coders were using identical coding categories,
resulted in Significantly positive results, as well (see

Table 2). In fact, summing over the nine dialogues for a

 

1"Home Fallout Protection Survey for the State of
Michigan--l969," A study conducted for the Office of Civil
Defense by the Department of Communication, publ. April
1970.

56

total reliability measure yielded a prOportion of .9163
indicating a high coder reliability. In addition, coders
reported this form to be considerably easier to handle,
more precise and retaining more information than the
Sluzki-rBeavin method which was initially tried, though

unsuccessfully.

Table 2.--Categorizing Reliability Check.

 

 

Tape 1 Tape 2 Tape 3
Lower Class Lower Class Upper Class

 

Coder 20 25 57

High Frequencies

Salient 19 24 52
Proportion .95 .96 .91
Coder l 20 79

Middle Frequencies

Salient l 17 71
Proportion 1.0 .85 .89
Coder 3 9 61

Low Frequencies

Salient 3 9 56
PrOportion 1.0 1.0 .91

 

Overall reliability - .9163

Description of Normal Population "Speeches"

After all tapes were sequentially coded (see Chapter

III), a frequency count of speech categories was made.

57

This is shown in Table 3 and represents the occurrence of
various types of statements in the normal pOpulation, irre-
spective of social class, sex, or topic salience. This
table is cumulative over the three tOpics.

As indicated on the table, 31% of the statements
made by reSpondents were assertion-extensions (code 23),
or in more common language--"picking-up on what someone
else was saying." Silence (code 00) accounted for another
large segment of the interaction frequencies (9% of the
total codes). Twelve percent of the speeches were what

1‘but what this author called

have been labeled "supportivef'
agreement or reinforcement. This 12% was divided equally
between assertion-agreements and talking-over agreements.
The category 29 (5%) indicates that primarily the first
person to speak asserted something rather than asking a
question, giving an instruction, etc. Of the other fre-
quencies with a high N, we find almost 5% of the Speeches
were assertions-in disagreement (code 22): 4% were questions
of a direct nature (code 19); while another 4% were agree-
ments which were neither talking-over nor an assertion,

but primarily consisted of statements following another's
speech on the order of "Right," "uh huh," "Sure," etc. At
the other end of the continuum, it is interesting to note
the low frequency of occurrence, in the normal pOpulation,

of all orders and instructions, which sum to only about 1%

of the total.

 

1M. Komarowsky, Blue-Collar Marriage, New York:

Random House, 1962

 

58

Table 3.--Norma1 POpulation Code Frequencies.

 

 

 

wLast 2 Digits of 3-Digit Code Frequency Percentage
11 Question-in-Agreement 2 .ll
12 Question-in-Disagreement 9 .50
13 Question-as-Extension 20 1.12
16 Question-as-Topic Change 8 .44
19 Question-is-Other-Typical 72 4.04
21 Assertion-in-Agreement 108 6.06
22 Assertion-in-Disagreement 83 4.66
23 Assertion-as-Extension 544 30.56
24 Assertion-in-Answer 66 3.70
25 Assertion-as-Disconfirmation 31 1.74
26 Assertion-as-Topic Change 47 2.64

27 ASsertion-in-Agreement-and-
Extension 48 2.65

28 Assertion-in-Disagreement-
and-Extension 15 .84

29 Assertion-is-lst statement

or other 88 4.94
20 Assertion-and-Laughter 2 .11
32 Instruction-in-Disagreement l .05
33 Instruction-as-Extension 4 .22
35 Instruction-as-Disconfirmation l .05
36 Instruction-as-Topic Change 3 .16
42 Order-in-Disagreement 1 .05
45 Order-as-Disconfirmation 1 .05

46 Order-as-ToPic Change 5 .28

59

Table 3.--Continued.

 

 

 

Last 2 Digits of 3-Digit Code Frequency, Percentage
49 Order-is-Other 2 .ll
51 Talking-Over-in-Agreement 107 6.01
52 Talking-Over-in-Disagreement 33 1.85
53 Talking-Over-as-Extension 62 3.48
54 Talking-Over-in-Answer l .05
55 Talking-Over-as-

Disconfirmation 17 .95
56 Talking-Over-as-Topic Change 7 .39
57 Talking-Over-in-Agreement-

and-Extension 14 .78
59 Talking-Over-but-Unintelligible 32 1.79
50 Talking-Over-and Laughing l .05
61 Assertion-in-Agreement,

followed by a Question 1 .05
62 Assertion-in-Disagreement,

followed by a Question 1 .05
63 Assertion-as-Extension,

followed by a Question 11 .61
64 Assertion-as-Answer,

followed by a Question 1 .05
65 Assertion-as-Disconfirmation,

followed by a Question 1 .05
66 Assertion-as-Topic Change,

followed by a Question 3 .11
67 Assertion-as-Agreement and

Extension, then a Question 1 .05
68 Assertion-as-Disagreement and

Extension, then a Question 1 .05

60

Table 3.--Continued.

 

 

Last 2 Digits of 3-Digit Code Frequency Percentage

 

69 ASsertion-and-Question-as-

Other 7 .33
73 Question-and-Assertion-in-

Agreement 10 .51
76 Question-and-Assertion-as

Topic Change 2 .10
79 Question-and-Assertion/Other 6 I .33
81 Statement-of-Agreement 79 4.43
82 Statement-of-Disagreement 2 .10
83 Statement-as-Extending 5 .25
87 Statement-in-Agreement and

Extension 3 .15
80 Laughter-in-response-to-other 17 .93
99 Laughter-was-Se1f-Induced 12 .66
90 Laughter-for-both 7 .39
00 Silence 148 8.39

 

If we sum all assertion statements (all codes that
are -2-'s), we find they account for approximately two-
thirds of the speeches ( 6 2% ). Likewise, all talking-
over statements (all codes that are -5-'s), account for
about one-sixth.( l7 %) of the utterances, and questions
(all codes that are -1-'s) account for about one-fourteenth
(7% ) of them. By summing the occurrence of agreement

over code categories (i.e., 11+21+51+61+81=16.66%) and

61

disagreement over code categories, we find the ratio of
agreement to disagreement in the couples is two to one.
This last statement, however, ignores the relational as-
pect of the last two digits which rules out placing them

together for anything other than a descriptive tendency.

The Significance of Time

Prior to discussing class, salience and sex dif-
ferences it should be noted that time spent discussint
these tOpics may have become a factor in differences in fre-
quency of code occurrence by class. A treatment by subjects-
replications analysis of variance1 on the discussion topic
times Split by social-class and tOpic salience, indicates
a Significant difference in length of time talked by the
two social-classes, but no significant difference over topic
salience. This difference in length of time talked indicates
that, on the whole, the upper class couples talked 1.5 times
longer than lower class couples, irrespective of tOpic sali-
ence. Later, we will see that this matches quite closely
to the factor representing the difference between the total
number of speeches coded by social class (speech factor=1.58).
Therefore, though tOpic salience indicated no time difference,
class does indicate a time differential of approximately 1.5

to 1.6 and will be taken into account for subsequent analysis.

 

lQuinn McNemar, Psychological Statistics, John Wiley
& Sons, Inc., N.Y., 1962.

 

62

Table 4.--Length of Time Talked.

 

 

Social Class
Lower Upper Factor Diff.

 

 

High 84.80 132.60 1.71
Topic
Salience Middle 92.00 107.20 1.20
(Averages)
Low 70.86 124.33 1.56
Total
Speeches 687 1090 1.58
F Col = 10.6010 significant at the .05 level
F Row = 0.3283 not Significant

Average time factor = 1.49

At this point, however, one should note that it is
possible to discuss these findings on several different
conceptual bases. One can refer to what happened in
thirty dyadic interactions over ninety dialogues and/or
one can refer to the 1,777 coded alternating speeches,
and/or the sum total of sequentially paired speeches (as
we shall do later). I will do each of these and attempt
to demonstrate the differences resulting from these

viSWpoints.

Salience Findings

Table 5 below indicates the frequencies by salience

|_

and chi-squares for code categories within a social class. V
As can be seen from the table, there is no significant

difference in code frequencies over topics varying in

63

 

 

 

NN.N N m m oo. o o o mm
oo. o H o oo. o o o am
AH. 4H SH SH mH.H a v N mm
oo. 4 m N mo.m N NH 4 Nm
NN. ON NN NH mm. 4H mH AH Hm
oo. o H o oo. o o H NV
NH. o m N oo. o o o ow
oo. o o o oo. H o 0 me
oo. H o o oo. o o O NH
oo. H H H oo. o o o mm
oo. o o o oo. o H 0 mm
oo. o H o oo. H H H mm
oo. o o o oo. o H o Nm
oo. H o H oo. o o o ON
NH. NH SH mH Hm. HH mH mH AN
co. m N m cm. s N H NN
oe.m A AH mH oo. N m m AN
vv.m AH o m mm. m A 4 SN
H4.H o m NH G.H o m H mN
GH.H «H HH N mN. m NH NH «N
mm.N HNH mOH mHH mN.N ow Am NA mN
NA. m a m NN.¢ 4N AN 4H NN
mm. NN AH NH Am.v HH 8H NN HN
oo. HH OH OH Nm. HH 8H 4H NH
NN. N N a oo. o o o SH
Am. A m H GH. o N H NH
oo. N H o oo. N N N NH
oo. H o H oo. o o o HH
Nx 30H mchHz anm Ax 30H mHnon anm mmHuommumo
mcoo

mmmHU momma mmmao Hm3oq

 

 

.mUGMHHmm an mowocmswmum m©OOII.m manna

64

.nmann no mm.m u Hm>mH mo. may no

HamoHNHcmHm.

 

 

 

 

«mv. Am wH mm mv.H ON NN mN oo
oo. N N m cc. o o o om
co. m m m oo. o H 0 mm
Nm.N m N H co. m m N om
mH.m OH m OH oo. H o H mm
oo. H H H oo. o o o Am
oo. H H N oo. o o H mm
oo. o o H 00. H o 0 Na
em.N NN mH mN co. m o m Hm
oo. H o N mo.H o m o mA
oo. N o o oo. o o o mA
mH. m H N mm. o N N mA
oo. H H H mH.H v o 0 mm
cc. o o o oo. o H 0 mm
oo. o o o oo. o o H Aw
oo. o o H oo. o H H mm
oo. o o H oo. o o o mm
oo. o o H oo. o o o em
oo. H N H AH. m m H mm
00. o o o oo. H o o No
oo. o H o oo. o o 0 Ho
oo. H o o oo. o o o om
mo.N m mH m HV. 0 N m mm
co. m w 6 oo. o o H Am
AH. m m H 00. o o o mm
mx 30H mecHz amHm Nx 30H mecHz smHm mmHnommumo

mmmHU Hmmmb mmMHO,H030H opoo

 

 

AumscHuaoucu-.m «Hams

65

salience. With a required Chi-square level of significance
being 5.99 at the .05 level, only silence (code 00) for

the upper class reaches that level and this appears to be
primarily the result of a low cell Size for middle salient
topics. The net result of this is an indication that dyads
tend to use the same content codes with a similar frequency
regardless of t0pic salience. This would indicate support
for the expectation of consistency in the way dyads with a

history relate, at least at this basic content-code level.

Class Findings

To this point, we have determined that there is a
time difference in discussions by class and no content-
code difference by salience. In discussing the frequency
of codes by class, I will refer to adjusted lower class
frequencies. These frequencies have been adjusted by a
factor of 1.6, suggesting that our base of analysis here
is the total number of speeches. Utilizing a Chi-square
test of significance for this ordinal data (with Yate's
Correction when cell size dropped below 15), yields eight-
een significant and two near Significant differences by
class (see Table 6). What are these differences?

First, questions: upper class dyads asked more
topic changing questions than lower class dyads, while
lower class dyads asked more general questions (either

beginning, or asking a question related to what was being

66

 

«oo.A o o A mm
oo. o o H SS
.HS.S S.SN SH SS HS
st.mH v.mm «N m Nm
mH.H m.MA mv Hm Hm
co. S.H H H as
oo. S.H H S SS
So. S.H H S SS
SS. S.H H o Nm
SS. o o N SN
HS.H S.NS mm as AN
NS. N.HH A S SN
«mm.HH v.vH m mm AN
HS. S.SN SH Hm SN
soA.NH ¢.w ¢ AN mN
amm.v m.Nm mm mm VN
mA.m o.vom omH vmm MN
«NA.vm m.Am mm mH NN
AH.N o.mA om mm HN
«mm.NH m.mw Hv Hm NH
«NH.@ o o m ma
mm.H o.m m ma MH
mv.N m.m m m NH
co. c o N HH
mumsvmlHnO mmHocmcvmum mmHocmsvam meocmsvam mmHHommumu
. AS.HV cmumsnc¢ mmmHU Hm3oq mmmHU momma opoo

 

 

.mmMHo An mmHocwnvmum o©001|.m wHQMB

.mc H nuH3 Hm>oH mo. map #0 ucmoHMHcmHm u vm.m«

 

67

 

«No.0 o.NHH 0A mA oo
«mm.m o o A om
fiHO-h 00H H HH m¢
mm. m.NH m m cm
«mm.vH N.m N mN mm
SS.H o o m AS
MCoH 00H H V 0mm
swm.MH N.AN AH No Hm
oo. m.v m m mA
00. m.H H 0 Am
oo. o o H SS
Hv.m N.HH A v mm
*HM.OH o.m m AN mm
soa.m m.H H MH Am
mumsvmlHnO mmHocmsvam mmHocmsUmum mmHocmsvam mmHHommumo
HS.HV pmumsnpfi mmmHU Hm30H mmMHU momma moou

 

 

.SmsaHucoouu.S mHnme

68

said, though not agreement, disagreement, extension, topic
change, or disconfirmation) than did upper class dyads.

In fact, there is a significant difference overall between
lower and upper class question-asking with lower class
subjects doing more of it (a ratio of 83.2 to 59).

While there is pdeifference between classes on
assertion-agreements, there i§|a very large difference in
assertion-digagreements. In this sample, lower class sub-
jects disagreed with their Spouses almost five-and-one-half
times as much as upper class spouses. Perhaps there is
some support here for the belief that there are more argu-
ments in lower class homes, or for the argument George Bach
raises that increased education makes us think that fight-
ing is bad resulting in fewer upper class disputes.l
Additional support for these statements may be indicated
by the near Significant level of assertion-extensions
(code 23) among upper class participants, using the ad-
justed figures. It is a clearly significant chi-square
without adjustment.

Though the upper class subjects do not disagree as
often, and do extend more often, than lower-class subjects,
they have a significantly greater (four times as great)

rate of assertion-disconfirmations (code 25) than do

 

1G. Bach and P. Wyden, The Intimate Enemy. William
Morrow and Co., N.Y., 1969.

69

lower class pairs. This finding dovetails with the fact
that lower class subjects tend to give assertion-answers
(code 24) significantly more often than do upper class
subjects. AS disconfirmations indicate one person sending
a relational message to his or her partner that they "do
not exist" given the preceding definition of the relation-
ship, this is perhaps more powerful than mere rejection in
displaying disagreement. In fact, if done habitually,
some therapists believe it can lead to schizophrenia or
other abnormal behaviors.

Other Significant findings indicate that the lower
class does more of the following: (1) talking-over in
disagreement (code 52); (2) assertions followed by ques-
tions from the same individual (code 63); and (3) silence.
On the other hand, upper class dyads display more of the
following behaviors: (1) assertions which first agree and
then extend (code 27); (2) topic-changing orders; (3)
talking-over of a variety of kinds including (a) extensions
(code 53), (b) disconfirmations (code 55), (c) agreement
followed by extension (57), and (d) mumbles or filled
hesitation pauses (code 59); (4) statements offagreement
that are neither talking-over nor assertions (i.e.,
"right," "sure," "yup," etc.) (code 81); (5) unintelligible
utterances (code 89); and (6) laughter of both the self-

induced and both laughing kinds (codes 99 and 90);

70

All this seems to indicate that members of these
two social classes do pragmatically different things when
they communicate with their spouses. There appears to be
more overt disagreement in lower class dyadic discussions
and more competition to speak, more development of an
idea by extending what another has said, and more covert
disagreement in upper class dyadic interactions. Though
one might speculate as to which aspect of social class--
education, income or job type--contributes most to these
differences, the point is that that combination of ele-
ments which distinguishes the two broad socio-economic
classes can also be distinguished on a socio-interaction

basis.

Sex by Class Findings

At this point, frequencies are distinguished and
compared by both sex and class. Once again.our total N
is the number of exchanges, thus necessitating adjusting
the lower class figures by 1.6. AS a check, I ran both
adjusted and unadjusted Chi-squares. As a result of using ,
adjusted figures, no unadjusted significant relationships
are lost and three adjusted significant relationships, all
female, are gained (codes 219, 222, and 224). Similarly,
for the males only one unadjusted significant relationship
is lost (code 123) and one significant adjusted relation-
ship is acquired (code 152). Silence (code 00) also be-

comes Significant when adjusted frequencies are used.

71

As can be seen from Table 7, the significant sex
by class findings for males indicate that lower class
males: (1) tend to ask more questions (code 119); (2)
assert in disagreement more often (code 122); (3) disagree
and extend on that disagreement more often (code 128); and
(4) talk-over in disagreement (code 152) more frequently
than their upper class counterparts. In addition, upper
class males significantly more often than lower class
males: (1) make disconfirming assertions (code 125); (2)
topic-changing assertions (code 126); (3) assertions which
agree with and extend what came before (code 127); (4)
extend while talking-over (code 153); (5) disconfirm while
talking-over (code 155); (6) agree and extend while talk-
ing-over (code 157); (7) mumble or say "well, but," "and,
maybe," etc., during a talking-over speech (code 159);

(8) agree, but little else, to fill the spaces (code 181);
(9) issue more unintelligible comments (code 189); and
(10) laugh more at their own statements (code 199) (either
from the content, as a tension release, and/or a dis-
qualification).

AS for the lower class females in this sample,
they (1) asked Significantly more questions (code 219);

(2) used more assertion agreements and assertion disagree-
ments (codes 221 and 222); (3) answered more questions
(code 224); and (4) disagreed while talking over (code

252) more frequently than their upper class female

72

Table 7.--Significant sex by class findings.

 

 

Male Adjusted

 

 

 

 

 

 

Lower Upper .
Code (1.56) .
Categories Class Lower Class Class Adj. x2 (U23?3'
119 21 32.76 15 3.30' (1.00)
122 28 43.68 8 24.63 (11.11)
123 113 176.28 183 .125 (16.55)
125 1 1.56 11 6.00 (6.75)
126 5 7.8 21 6.05 (8.65)
127 5 7.8 19 4.68 (7.04)
128 6 9.36 3 3.27“ (.44)
152 11 17.16 6 5.38 (.94)
153 4 6.24 23 9.61 (12.0)
155 0 O 9 7.11
157 0 0 11 9.09
159 3 4.68 15 5.45 (6.72)
181 10 15.6 31 5.09 (10.75)
189 2 3.12 13 6.06 (6.66)
199 0 O 7 5.14
000 70 109.2 78 5.20 (0.43)
Female Lower Adjusted Upper
Code Class (1'62) Class 2 (Unad'
Categories Lower Class Adj. x X2)J.
219 20 32.40 16 5.56 (.44)
221 34 55.08 32 6.12 (0.00)
222 37 59.94 10 35.66 (15.51)
223 77 124.74 171 7.24 (35.63)
224 16 25.92 13 4.29 (0.31)
225 3 4.86 16 5.95 (7.58)
227 4 6.48 20 6.90 (9.37)
252 13 21.06 3 13.56 (5.06)
253 12 19.44 23 0.30 (3.45)
255 0 0 8 6.12
259 2 3.24 12 5.04 (5.78)
281 7 11.34 31 15.04 (13.92)
289 0 0 10 8.10
3.84 = significance at the .05 level with 1 df.

73

counterparts. The upper class women utilized the following
codes significantly more often: (1) assertion-extension
(code 223); (2) assertion-disconfirmation (code 225); (3)
assertion in agreement and extension (code 227); (4) dis-
confirmation while talking-over (code 255); (5) mumbling
while talking-over (code 259); (6) agreeing but nothing
more, though not talking-over (code 281); and (7) more
mumbling (code 289).

A treatment by levels analysis of variance1 test
of significance was conducted to compare males and females
across class on codes 19,22,52,25,27,55,59,81,89,and 26.
Several of these indicated main effects but no interaction
effects. These indicated that both lower class couples
display more question-asking, assertion-disagreement and
talking-over disagreement than upper class couples.
Similarly, upper class couples tend to display more
disconfirming assertions, assertions in agreement and ex-
tension, talking-over disconfirmations, talking-over mum-
bling, flat statements of agreement and general mumbling
than lower class couples.

The only significant interaction effect was for
code 26--assertion-topic-changes. Here the finding was

that more upper class males exerted themselves in this

 

1W. L. Hays, Statistics for Psychologists, Holt,
Rinehart and Winston, N.Y., 1963.

 

74

fashion, while a Similar pattern was evident among lower

class females (see Table 8).

Table 8.--ANOVA Table for Code 26.

 

 

F
Critical
SS df MS F F Sig. Value Significant

 

Treatments 3.75 1 3.75 6.70 .025 5.29 Yes
Levels .42 1 .42 .75 .05 4.00 N.S.
Interaction 4.83 1 4.83 8.63 .01 7.08 Yes
Within 31.19 56 .56

 

Successive Interactions

To this point, we have been concerned with the fre-
quency of occurrence of individual code categories. These
are partially relational in the sense that the third digit in
the code refers to a meta-communicational aspect (agreement,

disagreement, etc.) of the interaction. As indicated in

Chapter III, we must look at least at pairs of statements
--exchanges--for a relational index. A frequency count of
pairs of statements, adjusted to account for the social
class time differential was taken. In all, there were 391
different pairs of codes which occurred at least once in
one of the classes, out of an 89 by 89 matrix. Of these

391 paired categories, twenty-two yielded significant

75

Chi-squares prior to adjusting the frequencies by a factor
of 1.6. After adjusting them, only fourteen of these
pairs remained significant, but four new pairs were added,
resulting in eighteen significant and three near signif-
icant chi-squares (see Table 9).

Using the adjusted frequencies, there is an indi-
cation that lower class husband-wife dyads, Significantly
more often than their upper class counterparts, tend to:
(1) follow a two second Silence with silence (have extended
periods of silence (code pair 00-00)); (2) answer more
general-type questions (code pair 19-24); (3) follow as-
sertion-disagreements with assertion-disagreements, or
with silence (code pairs 22-22 and 22-00); (4) follow
assertion-extensions with an assertion-disagreement, or a
talking-over disagreement (code pairs 23-22, and 23-52);
(5) follow discussion-Opening assertions with an assertion
agreeing with what was said (code pair 29-21); (6) follow
talking-over disagreements with assertion extensions (code
pair 52-23); and (7) have joint dyadic talking-over in
agreement when it occurs (code pair 51-51). This gives us
a picture of the over-riding tendencies of lower class
discussions, though admittedly over-simplified. One could
characterize them as assertions followed by a set of dis-
agreements and further assertions interspersed with si-
lence, with "rapid-fire" talking-over periods of both

disagreement and agreement.

76

.SS H cuHs Hm>mH Sc. cc» um ucconHcmHm

vm.m*

 

 

«SH.¢ SSH.¢ S o o mNImm
+AN.m «Nm.m SN SH SH MNIHS
«mH.m «MH.HH oN m.v m mNImS
SSH.S «SH.S SH o o MNIAS
«SH.S «SH.S A o o MNImm
Am.N «HA.A om N.mH NH MNImm
«NS.NH som.v v v.NN SH MNINS
4mm.S NH.m H N.HH A Hmle
SN. smm.S Hm S.vv SN MNIHS
«oo.v om.H H m m HNImN
«SH.S SSH.S A o o MNIAN
mm.N «HH.A SN SH SH HSIMN
«Am.m «SS.m SH S.H H mmIMN
soH.m SSH.S SH 0 o AmlmN
so.v som.A MN N.HH A mmlmN
«SS.HH «So.m m N.SN mH NSIMN
SA. «mm.v SN S.AH HH HmlmN
+AA.m «So.m MH m.v m ANINN
SH. vo.v SS v.oA vv MNIMN
«SN.S vo.N m S.mN SH NNImN
mm. «Sm.m Hm ¢N SH SclmN
«mm.SN «mo.SH o S.SN SH NNINN
«mv.m soo.v H S.NH m SSINN
«SH.m «SH.S A o o mmIHN
«wm.m SH.H SH ow SN @NImH
«Sm.oH mm.N m ¢.om mH octoo
mnmnvmlHno mumsv H mmHocmcvmum mmHocmcvam mmHocmdvmum mmHHommpmu

Umumcncd m|.£O mmMHO HmmmD mmMHO HS3OH cwumsmcfi mmMHO umBOH mpou

 

 

.HAHco unmoHMHcmHmv mcoHuomuwHSH cmHHmm mo mmHocmsvamul.m mHnt

77

In the other direction, the upper class dyads
studied, aswa whole, tended to have Significantly more:
(1) assertion-agreements followed by talking-over exten-
sions (code pair 21-53); (2) assertion-extensions followed
by both talking-over in agreement and extending, and non-
substantive talking-over (code pairs 23-27 and 23-59); and
(3) assertion-extensions following assertions which agree
and extend, disconfirming unsuccessful interruptions,
talking-over in agreement and then extending, talking-over
but saying nothing substantive, and saying nothing sub-
stantive though not talking-over (code pairs 27-23, 55-23,
57—23, 59-23, and 89-23, respectively). Using these ad-

justed statistics, we might characterize upper class
dyadic interactions as supportive and extending. But with

the disconfirmations that exist, there is strong convert

control.

Prior to adjusting the frequencies, upper class
couples could have also been portrayed as having a greater
tendency to: (1) follow assertion-extensions by silence
(code pair 23-00); (2) follow assertion-extensions with
assertions in agreement and extending what preceded it
(code pair 23-27--this pair has a near significant Chi-
square of 3.77 after adjustment); (3) follow assertion-

extensions by talking-over in agreement or general

78

agreement (code pairs 23-51 and 23-81); (4) follow both talking-
over in agreement and talking-over extensions with assertion-
extensions (code pairs 51-23 and 53-23); and (5) follow general
agreements with an assertion-extension (code pair 81-23--this
pair also has a near Significant chi-square (3.27) after adjust-
ment).

Based on these data, it is possible to begin to per-
ceive different discussion styles in the two social classes.

The differences in what is said in response to some other state-
ment as the stimulus, indicate that the lower class discussions
had more silence and more disagreements aired while upper class
dyads had more assertion-extensions perceded or followed by a
variety of unsuccessful interruptions or talking-over.

One of the questions arising from this study was whether
or not upper class dyads had a greater repertoire of responses or
speech possibilities than the lower class dyads. To determine
this, a distribution of the occurrence of code frequencies was
made. Starting with an 89 by 89 matrix in which each cell repre-
sents a pair of codes, the number of incidences of those rangeS‘
from zero to sixty-five. Zeroes were only scored when a frequency
existed for that code pair for one of the two classes. The fre-
quency of code pairs with these cell values in the lower class,
were anywhere from zero to two hundred and ten times and in the
upper class codes were used anywhere from zero to one hundred and
seventy-four. Variances were calculated for both of these dis-
tributions. The lower class variance was calculated to be
1604.3352, while the upper class variance was 1141.9450.

The resulting comparison yielded and F of

79

1.4049, exceeding the .05 level of significance (F=l.35).
This indicates that the upper class dyads displayed the use
of a greater repertoire of speech pairings, or in other
words, had more variability in paired patterns than the
lower class participants, even though the lower class vari-
ance is higher (a result of two hundred and ten instances of
zero as compared to seventy-four instances of zero, for a
code pair, for the upper class respondents. (See Figure l

for a visual comparison of the two classes.)

Relational Analysis

Utilizing the recoding rules as outlined in Chapter
III each of the ninety dialogues was recoded. Frequencies
were obtained for males and females in both classes over
the nine relational categories (++, ++, +8, ++, ++, +8,
8+, 8+, 88). Table 10 gives the resulting frequencies.

Using this data chi-square tests of significance L”
were made (using Yate's Correction when necessary) to
determine if sex, salience or class made a difference in
relational communication. By comparing males to females
across tOpic salience (taking each of the relational codes
separately), there is only one significant relational
category. This one relational category (++) indicates
that in the upper class, males tended to contribute more
++ to the significant chi-square, primarily in the high
salience topic (see Table 11).

These findings begin to indicate support for the
expectation that at a relational level couples are con-

sistent over tOpics varying in salience. Further support

80

mo3oH ".IIII

momma n IIII

monomouoo opoo comHom mo moHococvomm

MuQAMMNMHMcmoNSNANSNMNAmMN NNHNSNE Q A\ maxim. «x t 2 S A S h 1..

 

 

I. 411*)? a
an S»..-
SN.H u m mums;
Hm>mH So. as
accoHSHcmHS s
. I mo> I
.Sscs.H u SSSS HHHH u H u m

 

Nmmm.v0SH mo>

 

h N H c

11....-A1A411

HWII g~épgASAuS$nmx

 

 

... 8
H N
n . A
. . 0%
A H mm moHocosv
.. .. hm ImHM
I . oo\ omonu
A . cHHs
m m\mmHmm «o
m" M\ monEdc
.1“ SHAH H38.
oom

MNN

Figure 1.--Comparison of class variances (an approximation).

81

 

 

SS HN SN SS.SS NN NH SH mm
S S S S S S S +m
H S H S S S S +m
S S S S S S S m+
SS SN SH NN.NN HH S S ++
HA SS NS SA.SA SS SH SN ++
S S S S S S S m+
SS SH SH SA.SA AS AH SN ++
SSH AS SS NS.NS HS HN SN ++

mouomm n NS.N oUSMHHmm 30H
SS SH SN SN.SN AH A SH mm
S S S S S S S +S
S S S S S S S +m
S S S S S S S m+
AH S S NS.HN SH S S ++
SA AS SS SS.SA AS SS SN ++
S S S S S S S S+
SS SH SH SS.AS SS SH NN ++
SS SS SS SS.SS SS SS SS ++

mouoom n oocoHHmm oHpSHz
SS SH SN SS SN NH SH mm
H S H S S S S +m
S S S S S S S +m
N N S S S S S m+
NN SH A SS.SN AH A SH ++
SA SN AS SA.AA SS SS HN ++
S S S S S S S m+
SS SH HH SS.HS NN SH NH ++
SS SS SS SH.NS SS SS SS ++

mouomm n oocoHHmm cmHa

Hmuoe monEom monz Hapoa moHoEom moHoz

 

mmoHU momma

Hmuos cmumsAcS

 

mmmHO mo30H

opoo HSSOHumHom

 

 

.xom paw mmMHo An moHocoawomM opoo HoGOHHMHomII.SH oHnt

82

Table 11.--Ma1e-Fema1e Comparison over Topic Salience, by

 

 

 

Class.
Relational
Codes Chi-square
++ .636
++ .666
Lower Class ++ 1.790
++ .140
SS .280
++ 1.210
++ 1.710
Upper Class ++ 7.486
++ .473
SS .673

 

5.99 = .05 level of Significance with 2 df.

is given this expectation with the results of a whole pat-
tern within class comparison across topic salience. Using
the totals for the major five relational categories (the
other four categories occurring too infrequently to be
included), a three-by-five matrix was tested against the
Chi-square distribution for each class. It required a
Chi-square of 15.51 or better for it to achieve the .05
level of significance. Neither lower (12.95) nor upper
(6.32) class achieved significance. Therefore, we may
draw the conclusion that salience apparently does not
affect dyadic relational communication. Though salience
and therefore content may change, in the aggregate, there
is relational consistency within each of the two classes

of people studied. This offers support for a major

83

assumption in therapy, expressed by Watzlawick and others,
that there is relational stability in groups with-a-history.
(Individual consistency will be discussed in the next
section.)

Using the adjusted total frequencies for the lower
class dyads, a class comparison of relational patterns was
also conducted. The results indicate that upper and lower
class participants discussed the high salient topic in
about the same relational manner, the middle salient tOpic
somewhat differently though not significantly so, and the
low salient topic significantly different. Requiring a
Chi-square of 9.49 for significance with four degrees of
freedom at the .05 level, the class-pattern comparison
across topics yielded Chi-squares of .58, 7.25, and 21.02,
respectively.

As the above results came from a whole pattern
analysis, individual relational code analyses were made
across claSS to determine what contributed to this dif-
ference. The major contributions to this increasing Chi-
square came from the 88 category in the middle salient
topic (4.41=x2; 3.84 needed for significance), with upper
class dyads exhibiting more of this symmetrical behavior,
and the ++ category in the low salient topic (15.26=x2;
3.84 needed for Significance), with lower class dyads

demonstrating more of this competition toward one-up

behavior.

84

It is interesting to note that some of those who
have studied these relational patterns would suggest that
the ++ relationship, because it is unstable will eventually
resolve itself into a stable complementary relationship
with one party being in meta-complementary control--i.e.,
"I'll leg you control me." Given the emergence of this
difference in the way classes relate under the conditions
of this experiment (one contact and one interaction situa-
tion type), we may want to pursue the matter of inter-class
relational differences in future research. AS of now, we
can conclude that there is consistency within a class
across both sex and topic salience, a difference in con-
tent-code frequencies across class and there may be a
relational difference in the way the two classes communi-

cate.

Individual Consistency over Topic Salience

To determine the parameters of "normal" family
communication in the dyad, it was felt that both aggregate
(above) and individual data were necessary to demonstrate
consistency of relational communication over tOpic salience.
This would be necessary for us to predict not only to a
mass tendency in the general population, but also to indi-
vidual dyads within it. To accomplish this each of the
thirty males and females were checked as to the consistency

of their relational communication over the three tOpics.

85

Each relational code was compared across t0pic salience
using a chi-square test with Yate's correction because of
generally low cell size. The chi-square tests were run
both by individual relational categories across sex (a
series of two-by-three matrixes) and by using total rela-
tional scores, the whole pattern for each couple across
tOpic salience (thirty matrixes varying in size from two-
by-three to five-by-three).

Taking the individual relational code comparisons
first, very few significant relationships were found--
among lower class males and females, only seven Significant
or near significant relationships out of one hundred and
five chi-squares and only twelve significant or near
significant differences out of one hundred and seventeen
chi-squares in the upper class dyads. This is about what
one would expect to find by change alone. Taking the
whole dyadic pattern for each couple and checking its
stability over the discussion tOpics resulted in only
one significant relationship among the thirty couples
tested. Both of these results reinforce the belief that
there is stability in relational communication in groups-
with-a-history. It is interesting to note that the one
couple whose relational pattern analysis attained signif-
icance (19.1585=x2; 15.51 needed for significance with 8df)
also contributed the largest number of individual rela-
tional code significant chi-squares (4). This same couple

will later be discribed as having a "fluid" relationship--a

86

pattern which Sluzki and Beavinl describe as the ideal

state for normals.

Individual Cases

One of the concerns in developing a system capable
of coding relational interactions, is the ability to dis-
cern emerging patterns. Though the data indicate that all
but a very small percentage of the interactions of thirty
dyads studied were consistent over topics, it is yet
another problem to begin to label them.

Sluzki and Beavin2 refer to seven theoretical
patterns, while Watzlawick et_21,3 refer to two relational
and two meta-patterns. Identifying a pattern is somewhat
analOgous to understanding music theory. In both cases
one is concerned with variations on a theme, rhythm, change
of key and endings. In relational communication we begin
with a pair of statements, or an exchange, and watch for
repetition or variation around that theme. In a more
therapeutic setting, one might also watch for more unique
orchestrations such as patterned disqualifications, habitual
disconfirmations, or some combination of these. I believe

that as our understanding of relational communication

 

lSluzki and Beavin, Op. Cit.

2Sluzki and Beavin, op. cit.

3Watzlawick, Jackson, and Beavin, op. cit.

87

grows, this dissertation being but a small step along the
way, that this analogy will become more interesting. Some
of the more basic patterns are discernable from the data
gathered for this study.

TO take the most obvious of patterns first--stab1e
complementarity--we would find, for example, an exchange
in which participant A was both first to Speak and rated
one-up (+), that participant B was one down (+). In a
stable complementary relationship this will be true every
time participant A is coded first, and the Opposite will
hold true whenever participant B is taken first in the
exchange (++).

An example of this type of relationship comes from
one of the lower class dyads. In their discussion Of the
high salient topic the female was scored ++ five times
while the male was scored ++ four times. This would indi-
cate a perfect complementary relationship with the male in
one one-down position. An example from the dialogue in
this tOpic is illustrative of this kind of interaction:

119--H: But then I don't remember what'd I say to him?

224--W: You eXplained the wrong ways and the right
ways and childbirth and other things.

121--H: Oh yah, I remember that . . . let's see . . .

223--W: Over in E. Lansing they already have a program
in reproduction SO he really already knew the
right way.

For this couple, though there was some one-up/one-up com-

petition during the middle salient tOpic, all their

88

conversations were predominantly one-up/one-down or re-
solved into this pattern following these relationally
competitive bursts.

One of the upper class couples illustrates rather
clearly another of these patterns of interaction--the
stable symmetrical relationship. This doesn't mean that
they have only symmetrical exchanges, rather it means that
over time they have an equal number Of similar exchanges.
In this case this was true of each of the topics. The

following example comes from the middle salient topic.

Relational Scores

Husband Wife
++=6 ++=5
++=2 ++=1
++=3 ++=5
++ = 1 ++ = 1
53:4 ss=4

A typical coded sequence from that dialogue is as follows:

Start

assertion-extension = 223 F = 88
assertion-extension = 123 M = ++
talking-over-in-mumbles = 259 F = ++
assertion-extension = 123

. . _ M=SS
assertion-exten51on — 223 F = ++
talking-over-agreement = 151 M = ++
assertion-extension = 223 F = ++
talking-over-agreement = 151 M = ++
assertion-extension = 223 F = 88
assertion-extension = 123

In this example, the couple begins with a symmetrical ex-

change Of assertion-extensions. The female then is

89

one-down via an unsuccessful attempt to interrupt. This is
followed by another symmetrical exchange scored for the husband
this time. Then there are a set of exchanges in which.the hus-
band is talking-over in agreement with What his wife is saying.
The pattern concludes with another symmetrical exchange.
This type of chain repeats itself, equalizing out over
time and alternating positions of control for short periods
of time prior to reasserting the symmetrical base.

Sluzki and Beavin refer to one of their patterns
as Assymmetrical Competition toward One-down and Symmetry.1
An example of this pattern was exhibited by another lower
class couple. It is perhaps most strikingly found in
their discussion of the middle salient topic, though,
again, they are consistent over topics. Their relational

score is as follows:

Husband Wife
++ = 8 ++ = 8
++ = 1 ++ = 2
SS=4 SS=4

One of the patterns in that dialogue which demonstrates

this is:

assertion-extension = 223 F‘= 88
assertion-extension = 123 M = ++
talking-over-disagreement = 252 F = ++
assertion-extension = 123 M = 88
assertion-extension = 223 F = 88
assertion-extension = 223 M = ++
assertion-agreement = 221 F = ++
assertion-extension = 123 M = 88
assertion-extension = 223

 

1Sluzki and Beavin, Op. cit.

90

This pattern is typical for this dyad and demonstrates
almost a dance which this couple is engaged in. Every
time they have been symmetrical for awhile, she defines
herself as one-down only to later become symmetrical again
and then, again, one-down, etc.

A pattern which Watzlawickl and Bateson2 discuss
is escalating symmetry. Here, individuals try to top each
other culminating usually in one person's losing the point
(being one-down) followed by a symmetrical exchange to
redefine the relationship. An example of this comes from

another lower class dyadwhich had the following interaction:

assertion-disagreement = 122 _
o 0 o M - ++
quest1on-1n-d1sagreement = 212 F = ++
assertion-disagreement = 122 _
I o o o M — ++
instruction-1n-disagreement = 232 _

. . _ F - ++
question-Of-a-typ1cal-nature — 119 M = ++
assertion-disagreement = 222 _

O I F - ++
talk1ng-over-extens1on = 153 M = 88
talking-over-extension = 253

In another of their discussions, they had the following
exchange:

assertion-extension 223

assertion-extension = 123 a : if
assertion-disagreement = 222 _

. . F - ++
assertion-d1sagreement = 122 M = ++
question-of-a-typical-nature = 219 F = ++
talking-over-mumbling = 159 M = 88
talking-over-mumbling = 259

 

lWatzlawick, Jackson and Beavin, Op. cit.

2G. Bateson, Op. cit.

91

Throughout their discussion, a similar trend is seen.
Competition to outdo the other partner is followed by a
complementary exchange and then a symmetrical exchange as
the demarcation of the end Of that escalation. A completed
competition is generally followed, for this couple, by
either silence or laughter, both individual and mutual.
Over time, the number of competition wins and losses equal-
izes out.

There are other patterns which. Sluzki and.Beavin
discuss but which either didn't occur in this sample or
which this author cannot yet consider himself qualified to
identify positively. The only other pattern identifiable
is that non-pattern which. Sluzki and Beavin labeled

"Fluid."l

This they believe to be the pattern utilized by
true "normals." One upper class couple displayed this
tendency more Significantly than any of the others. It is
best shown by the high chi-squares on the sex by salience
dimension of their relational communication, indicating
lack of consistency over tOpics. They contribute four of

the twelve Significant or near significant chi-squares and

the only significant pattern chi-square, discussed earlier.

Husband Wife
Chi-squares Chi-squares
++ = 4.79* ++ = 2.18
++ = .55 ++ = .37
++ = 4.24* ' ++ = 5.45*
++ = .37 ++ = .37
88 = 1.34 88 = 4.16*

 

lSluzki and Beavin, Op. cit.

92

It would be precisely this lack Of consistency, in the

long run that would, perhaps, identify the truly normal
dyad. It would indicate that their relational rule was to .
have a flexible rule rather than a rule placing them con-
sistently in a complementary, symmetrical or some competing

pattern.

Summary

In summation, then, based on the data gathered in
this exploratory experiment, we can say that by examining
ninety dialogues from thirty couples divided into two
social classes, the Significant findings indicate: (1)
Upper and lower class dyads discuss topics varying in
salience significantly different lengths of time but~
within the social class there is no Significant difference
in the length of time they discuss these topics; (2) Upper
and lower class pairs use significantly different approaches
to discussion with the lower class displaying more open
disagreement and Silence and the upper class dyads dis-
playing more covert disagreement and more symmetrical
kinds of behaviors, such as assertion-extensions and
assertion-agreement-extensions; (3) There is consistency
of both content codes and relational codes over topic
salience by both sex and class, and by both aggregated
and individual data indicating strong support for the long

standing assumption that though people may say different

93

things or use different words, people with-a-history-
together are consistent in the way they relate to each
other; (4) There is a tendency for the two classes studied
to relate differently to their spouses; and (5) it is pos-
sible to identify individual patterns as being stereotypic,

consistent, or stable.

CHAPTER V
SUMMARY AND CONCLUSIONS

"I am but an egg about to be hatched."

A. Heinlein
Stranger in a Strange Land

Review

The purpose of this study was to explore and ex-
plicate the communication patterns of "normal" husband-
wife dyads. Following a review of the literature,
this author decided that the principle progress in family
communication had come from the works of a group of thera-
pists known as "system purists." This research began with
Gregory Bateson and has developed through to Sluzki and
Beavin's attempt to operationalize his variables. Taking
this prOgress one step further, an interaction coding
scheme was develOped which codes (a) speaker, (b) speech,
and (c) that Speech as a response to the preceding stimulus,
or meta-speech. A set of rules was created to reduce the
resultant eighty-nine categories down to nine relational
codes, whose relational frequencies were derived from

pairs of category codes or dyadic exchanges.

94

95

It is this relational aspect and not a specific
content analysis which is important. Relational communi-
cation serves as a constant guage on the state of the
system for components in that system. It also serves
the researcher and therapist as an indication of the
strengths and weaknesses of that system.

Given that every message bears both content and
relational elements and that it serves to define, redefine,
and reinforce relational rules, these rules ought to be
quantifiable by the social scientist. Although in theory
it should be possible to make any kind Of a response to
any kind of a stimulus statement, in fact, responses are
patterned, rather than random in groups-with-a-history.

The vast majority of the research to date has been
conducted with abnormal families as the subjects. The
data gathered from these families seem to indicate that
their patterns of communication are both rigid and de-
structive. They have greater periods of silence, take
longer to arrive at decisions, disqualify disconfirm, and
are non-supportive in a very consistent fashion. From
these data family communication researchers tend to eXpect
normal families to be more flexible in the sense that
they have more than one pattern at their disposal, but

are consistent relationally over time and tOpics.

96

Given these and other assumptions, it was felt
that an exploratory study examining so-called "normal"
husband-wife dyads was necessary. The principles of a
good family communication study as outlined by Jay Haley
were essentially followed. TOpics were developed which
would be relevant to the target audience--upper and lower
class couples with children under ten years of age. These
three tOpics, rank ordered by salience by each couple,
required them to discuss practical alternatives for them
and arrive at some decision regarding what to do if (1)
their child was overheard discussing sex with a friend
with misinformation; (2) night-time television program-
ming began to Offer programs for which their children
wanted to stay up late; and (3) what they would do with
their family in case of a Civil Defense emergency.

After oftaining audio-tapes from fifteen lower and
fifteen upper class couples, these ninety dialogues were
analyzed for time, salience, class, sex, and pattern dif-
ferences on both aggregated and individual data. Given
the restrictions of a non-random sample, one contact with
each couple, and taping of three similar tasks, it was
found that couples tend to possess patterns or rules of
communication and that these patterns are not only dis-
cernable in the normal pOpulation but consistent for these
dyads over tOpics varying in salience. Further, it was
found that these patterns may not be randomly distributed
over the general pOpulation, but rather, may be differen-

tiated at least by class. In fact, strong evidence is

97

presented indicating relational consistency in both aggre-
gate and individual cases over tOpics.

Content-structural distinctions between upper and lower
class couples were also discovered which demonstrate that
they do in fact talk differently, though on a relational
level this differentiation is not quite as dramatic. More
specifically, lower class subjects had more Open disagree-
ments while the upper class subjects used the subtler, but
perhaps more powerful method of disconfirmation more fre-
quently. Also, there was a greater incidence Of supportive
or symmetrical behavior among upper class dyads than among
lower class dyads.

There appears to be considerable Similarity across
sex within a social class, such that both males and females
in the lower class tend to ask more questions and disagree
in a variety of ways, than do their upper class counter-
parts. Similarly, both upper class men and women tend to
use assertion-extensions, talking-over-extensions, asser-
tion-agreement-extensions and disconfirmations more fre-
quently than lower class dyads. There is also an interac-
tion effect indicating a greater tendency for upper class
males and lower class females to make assertion-tOpic-
changes more frequently than their Spouses.

Using paired exchanges as the basic relational
unit it was possible to demonstrate the greater variability

of upper class responses, the consistency within class and

98

for individual dyads over topics varying in salience, and
the relational distinctions between classes. There appears
to be a greater tendency for symmetrical behaviors to

occur among upper class pairs and for more complementarity
and/or behaviors that should resolve in complementary or
meta-complementary patterns among lower class dyads.
Finally, it was also possible, using this relational anal-
ysis system, to identify several major interaction patterns
occurring in the normal population. These included, stable
complementary, stable symmetrical, asymmetrical competition
toward one-down and symmetry, escalating symmetry (which
Sluzki and Beavin might label asymmetrical competition
toward one-up and symmetry), and fluid.

An interesting point to note here is that no "eco-
logical" problems were indicated. Data was analyzed for
both aggregate and individual consistency across topics
and similar results were found in both instances. This
allows for generalization to both its occurrence in a
'class of people and to expectations about particular dyads
within that population. The major restriction on this
study comes from the fact that data was gathered as a
result Of only one meeting with each of these dyads, and
they discussed topics similar in organization. This was
consciously done to reduce variability in this initial
exploratory study. Future research, to be discussed later,

should explore these other possibilities as well.

99

On the whole nonparametric statistical tests were
used except where the level of measurement approached the
interval level, as in the question of time talked, or when
it was desirous to determine if there was an interaction
effect. These nonparametric tests were used primarily
because of low cell frequency size, nominal or at best
ordinal data, and the inability to fully achieve the as-r/
sumption of a fully random sample. In using nonparametric
statistics there is a loss Of power-efficiency. Yates cor-
rection was used with small cell Size and causes no prob-
lems when either the degrees Of freedom are greater than'

D O . I . 1
one, or when cell Size is greater than five.

Other Statistical Analysis

 

Given the sequential data it should be possible in
the future to do some further analysis. Specifically,
some form Of cluster analysis and/or sequence analysis may
be called for. This type Of analysis would become par-
ticularly relevant once similar data.havebeen gathered
from a matched group of abnormal dyads. (To be discussed

later.)

 

1 - ... . . .
~Op. C1t., Quinn McNeman, and S. Siegel, Non-
parametric Statistics, McGraw—Hill, N.Y., 1956.

...J

100

Cluster analysis depends "upon the identification
of clusters and presumed factors by searching for inter-
related groups Of correlation coefficients or other meas-
ures of relation."1 Perhaps the best form of cluster
analysis for this type of data would be McQuitty's linkage
analysis system. This is a method of clustering either
people, items or any other objects, which have distinctive
cluster-characteristics.2 Elementary linkage analysis
defines the linkage Of one variable to another as the
largest index of association which a variable has with any
or all of the other variables. The data determine which
variables are assigned to which clusters by their highest
index Of association.

For this interactional data, a cluster analysis
would take the most frequently paired categories and
cluster around them other pairs, from highest to lowest
frequencies, which are related to the initial pair. For
example, this might tell us that the incidence Of a 23-23
is clustered with 53-23, 23-22, 23-27, but not with 25-00,
or 90-51, etc. Given that cluster analysis is closely

related to factor analysis, although it determines a typal

 

1F. N. Kerlinger, Foundationsgof Behavioral Re-
search, Holt, Rinehart and Winston, Inc., N.Y., 1964, p.
659.

 

2L. L. McQuitty, "Elementary Linkage Analysis for
Isolating Orthogonal and Oblique types and Typal Relevan-
cies," Educational and Psychological Measurement, Vol. 7,
1957, pp. 207-229.

101

structure rather than a simple structure, it ought to be
possible to label these clusters. For example, there
might be supportive, non-supportive, argumentative, com-
petitive, disconfirming, complementary or symmetrical
clusters emerging from this data. TO be able to demon-
strate a greater frequency Of occurrence Of one cluster
type in the abnormal pOpulation surely warrants further
investigation.

Another method of analysis potentially usable in
the future is sequence analysis. Perhaps the most widely
used is the Markov chain analysis. Here, as in relational
theory, each act can be looked at as stimulus, response,
and reinforcement. Suppes and Atkinson discuss the possi-
bility that a sequence of apparently random variables C1,
C2, C3 . . ., Cn . . ., is a Markov chain. They conclude
that for it to be a Markov chain:

. . . g i v e n. a knowledge Of the conditioning on
trial n, the conditional probability is unchanged by
knowledge of the conditioning on any trials preceding
n. This fact is characteristic of Markov processes.
The process is a Markov chain when the transition pro—
babilities are independent of n, that is, constant
over trials.

To apply the Markov process to this data would

necessitate the identification Of conditional probabilities

attached to the wide variety of alternative response

 

1P. Suppes and R. C. Atkinson, Markov Learning
Mgdels for Multiperson Interactions, StanfOrd University
Press, Stanford, Calif., 1960.

102

categories to some given stimulus category. In other
words, the development of a matrix of probabilities for
response choices Open to an individual as a function Of
the state in which that individual finds himself at each
stage of the interaction. The frequency of paired code
responses is a start in this direction but will require
further investigation before its immediate payoff for this

data can be assessed.

Future Research Suggestions

 

Given the success of this preliminary study of
"normal" family communication in the dyad and the develop-
ment Of an analysis scheme, the concern must now be with
how to further the systematized experimentation into this
area. This future research falls primarily into three
categories--short-term follow-up studies, long-term devel-

opmental studies and greater than dyadic studies.

Short-term Studies

Ideally, an identical study of "abnormal" husband-wife
pairs should be conducted. It would ask upper and lower class
couples in which at least one member of a pair was undergoing
therapeutic assistance, to discuss identical tOpics. This
type of study should reveal the significant relational
differences between normal and abnormal dyads with-a-

history. It would be possible to again note the

103

similarities across class and consistency over topic sal-
ience. A relational study such as this might also go a
long way toward discovering a more operationally valid and
realistic way of determining abnormal interaction behavior.

A similar short-term study dealing solely with
parents of diagnosed schizophrenics might also be con-
ducted. Here, too, both content and relational codes may
uncover the stereotypic pattern of interaction utilized by
these couples which not only distinguishes them from "nor-
mal" couples, but also results in abnormal offspring.

Also, other short-term studies of normal couples
are called for. Besides the discussion of topics varying
in salience to determine consistency, conflict resolution
studies might be tried to discover homeostatic functions
after problem arousal. Real conflict which the couple is
currently coping with would be the best for this type of
study, but revealed-differences technique or other conflict
arousing stimuli would probably reveal some of these
patterns.

Another potential study would replicate this one
but allow for variation in length of association of hus-
band-wife dyads. Obtaining data from newlyweds to "golden-
agers" as verification of this data appears to be another
logical "follow-up" study. Throughout these studies, emphasis
should be placed on the development Of criterion measures to
evaluate relational satisfaction. This could result in an
Operationalization of abnormal communication and/Or an index

of relationships with destructive tendencies.

104

Long-term Developmental Investigations

The purpose of such long-term developmental studies
would be to demonstrate the evolution of interaction pat-
terns from exploratory to unstable to stable. Also, it
would be worthwhile to investigate the conditions leading
to changes in tactics and to their stability over time.

It is at this point that some non-verbal categories Should
be added so that the transition of simple matter-energy
kinesics to their utilization as information frames can

be traced.

Such long-term studies could be conducted with
roommate pairs, engaged couples or friendship pairs in
grade school. The goal would be to obtain these pairs of
subjects very close to their "first meeting"--for roommate
pairs, their first day on campus, for engaged couples,
shortly after the engagement, etc. Then, reexamine these
dyads periodically noting changes in their relational
communication.

Another developmental study certainly worth doing,
though rather difficult to gather subjects, would be to
add data on the parents of the above mentioned dyads.

This could give the first verification of transference of
relational behaviors from one system to another.

In addition, a study Similar to this one might be

conducted. The only addition would be to take the tapes

105

on three or more different evenings. This might offset

any problems incurred by a one-time exposure.

Larger than Dyadic Groups

The method of analysis developed for this paper
ought to be equally applicable to other groups-with-a-
history with a larger group N. Initially such studies
should concentrate on the study of triads and their
coalition formation.

Georg Simmel began the work in this field by ex-
ploring the triad in the 1890's.1 Caplow gave perhaps the
first truly systematic treatment of coalitions.2 His
belief that the most desirable coalition partner is the
weakest partner who is strong enough to produce a winning
coalition, however, leaves the subtleties of long-term
relationships out.

The importance of the coalition concept contribu-
tions to family and other communication system analyses
lies in its ability to better conceptualize some Of the
dynamic interaction processes--both verbal and non-verbal
--that occur within the system. Gamson apparently grasped
the volatility of coalitions in an interacting system-and

defined coalitions as temporary events:

 

lT. Caplow, Two A ainst One, Prentice-Hall Sociology
series, N.Y., 1968' pp. 1 -150

2T. Caplow, "A Theory of Coalition Formation,"
American Sociological Review, 21, 4, 1956, pp. 489-493.

 

 

106

There is generally little value consensus in a coali-
tion and its stability requires tacit neutrality of
the coalition on matters which go beyond the immediate
prerogatives. The coalition is an instrument for
achievement of widely ranging and even incompatible
goals. Two members may realize their mutual goal
antagonisms but such decisions lie in the future and
the present alliance may make both better able to
achieve a wide ringe of goals not all of which will

be incompatible.

Given this orientation, perhaps we can better
understand Watzlawick's statement that:

We believe that any future taxonomy of families will
rely heavily on the identification and classification
of typical coalition patterns. At present, for in-
stance, recurrent data suggest that one of the schizo-
phrenic patient's main concerns is manipulating coali-
tions so as to keep his parents at an optimal distance
from each other. It would appear that, far more than

a normal child, he feels a threat to his survival when
his parents either come too close together or drift too
far apart.

In addition to this, both Wynne and Haley have suggested
that coalitions, splits and, perhaps even meta-coalitions,

are the very essence of maintaining family equilibrium.3

 

1W. A. Gamson, "A Theory of Coalition Formation,"

American Sociological Review, 26, 1961, pp. 373-382, and

W. A. Gamson, "An Experimental Test of a Theory of Coali-
tion Formation," American Sociological Review, 26, 1961,

pp. 565-573.

2P. Watzlawick, op. cit., Anthology.

3L. C. Wynne. "The Study of Intrafamilial Align-
ments and Splits in Exploratory Family Therapy," in Ex:
ploring the Basefog Family Therapy, N. W. Ackerman, F. L.
Beatman, and S. . Shérman (eds.)7 Fam. Serv. Asc. of
Amer., N.Y., pp. 95-115, and J. Haley, "Family Experiments:
A New Type of Experimentation," Family Process, 1, 1962,
pp. 265-293.

 

107

Using this scheme of relational analysis it should
be possible to study coalition formation, and change, as
well as other tactics of control in any group-with-a-his-
tory. One would assume that those members of the group in
a coalition together would exhibit symmetrical relational
behaviors, whereas they would exhibit a somewhat more com-
plementary or competitive relational style with members of
the group not in the coalition. Changes in coalitions and

competition to form coalitions could also be mapped.

Conclusions

 

To engage in a study of interactions based on the
participants communication behavior we must first fully
understand the dynamics of the small group and the subtle-
ties of communication existing within that group when it
possesses a history. A systems conceptualization of the
family must draw upon the findings of both small-group
research and family communication research to understand a
process conceptualization of communication behavior in
this relationship.

The coding scheme which emerged from this process
conceptualization is apparently effective in analyzing rela-l
tional communication. This study has shown the existence of
significant class discussion differences,as well as Offering

statistical support to the belief that the relational patterns

108

peOple use in their interactions in dyads with—a-history
are consistent over topics and are discernable.

Some of the communication patterns of "normal"
dyads have been revealed and described and suggestions for
further research in this area have been offered. Beyond
this, further exploration into abnormal and normal inter-
action differences in relational rules, homeostatic mech-
anisms, redundant kinesic behaviors, and tactics of con-
trol, is called for. Only additional large-scale research
can uncover the elusive nature of man's communication

behavior.

B IBLIOGRAPHY

BIBLIOGRAPHY

Bach, G. and P. Wyden. The Intimate Enemy. William Morrow
and Co., N. Y.,il969.

Bales, R. F. Interaction Process Analysis. Addison-
Wesley, Cambridge, Mass., 1950.

Bateson, G. "Minimal Requirements for a Theory of Schiz-
ophrenia," Archives of General Psychiatry, 2, 1960.

Bateson, and D. D. Jackson. "Some Varieties of Pathogenic
Organization," in Disorders of Communication.
David Mck. Rioch (ed.), Research Publ. A.R.N.M.D.
42: 1964.

ll/

 

Beels, C. C. and A. Ferber. "Family Therapy: A View,"
Family Process, 8, March, 1969.

 

Becker, E. The Birth and Death of Meaning. Free Press of
Glencoe, 1962.

Bergel, E. Social Stratification. McGraw-Hill Book Co.,
N.Y., 1962.

 

Berlo, D. K. "The Context for Communication Study," a
recent essay, Michigan State University, 1969.

Bodin, A. M. "Family Interaction: A Social-Clinical Study
of Synthetic, Normal and Problem Family Triads," a
paper read at the Western Psychological Assoc.,
1966.

Bodin, A. M. "Conjoint Family Assessment," in Advances in
Psychological Assessment. P. McReynolds (ed.).
Science and Behavior Books, Palo Alto, California,
1968.

Boszormenyi-Nagy, I. and J. L. Framo (eds.), Intensive
Family Therapy, Harper and Row, Pubs., New York,
1965.
a. I. Boszormenyi-Nagy, "A Theory of Relationships:
Experience and Transaction;
b. J. L. Framo, "Systematic Research on Family
Dynamics."

 

 

109

110

Buber, M. I and Thou. Charles Scribner's Sons, N. Y.,
1958.

 

Caplow, T. "A Theory of Coalition Formation," American
Sociological Review, 21, 4, 1956.

 

Caplow, T. Two Against One. Prentice-Hall Sociology
Series, N.Y., 1968.

 

Cartwright, D. and A. Zander. Group Dynamics: Research
and Theory, Row Peterson, Evanston, Ill., 1962.

 

Cohen, A. M. "Changing Small Group Communication Net-
works," Administrative Science Quarterly, 6, 1962.

Cooley, C. H. Human Nature and the Social Order. Free
Press of Glencoe, 1956.

Ferber, A. and M. Mendelsohn. "Training for Family Ther-
apy," Family Process, 8, March, 1969.

 

Ferreira, A. J. "Decision-Making in Normal and Patholog-
ical Families," Archives of General Psychiatry,
8, 1963.

Ferreira, A. J. and W. D. Winter. "Family Interaction
and Decision-Making," Archives of General Psy-

chiatry, 13, 1965.

Ferreira, A. J., W. D. Winter and E. J. Poindexter. "Some
Interaction Variables in Normal and Abnormal Fam-
ilies," Family Process, 5, 1966.

 

Flanders, R. "Interactional Analysis," in Interaction
Analysis, Amidon and Hough (eds.), MacMillan Publ.
Co., N.Y., 1963.

 

Frank, G. H. "The Role of the Family in the Development
of Psychopathology," Psychological Bulletin, 64,
1965.

 

Franch, Jr., J. R. P. and B. H. Raven. "The Bases of
Social Power," in Studies in Social Power, D.
Cartwright (ed.). Ann Arbor, Mich., Univ. of
Michigan Press, 1959.

 

Gamson, W. A. "A Theory of Coalition Formation," American
Sociological Review, 26, 1961, and W. A. Gamson,
1'An Experimental Test of a Theory of Coalition
Formation," American Sociological Review, 26, 1961.

 

 

111

 

Gross, E. "Symbiosis and Consensus as Integrative Factors
in Small Groups," American Sociological Review, 21,
1956.

Haley, J. "Family Experiments: A New Type of Experimenta-
tion," Family Process, 1, 1962.

 

Haley, J. Strategies of Psychotherapy, Grune and Stratton,
Inc., N.Y., 1963.

 

Haley, J. "Research on Family Patterns: An Instrument
Measurement," Family Process, 3, 1964.

Hall, A. D. and R. E. Fagen. "Definition of System."
General Systems Yearbook, l, 1956.

 

Hare, A. P. Handbook of Small-Group Research, The Free
Press, Glencoe, N.Y., 1962.

 

Hays, W. L. Statistics for Psychologists, Holt, Rinehart
and Winston, N.Y., 1963?

 

"Home Fallout Protection Survey for the State of Michigan
—-1969," A study conducted for the Office of Civil
Defense by the Department of Communication, publ.
April, 1970.

Jackson, D. "Family Interaction, Family Homeostasis, and
Some Implications for Conjoint Family Psychother-
apy," in Science and Psychoanalysis. J. H. Masser-
man (ed.), Vol: 5, Grune and Stratton, N.Y., 1959.

 

Jackson, D. "Family Rules: The Marital Quid Pro Quo,"
Archives of General Psychiatry, 12, 1965.

 

Jackson, D. and P. Watzlawick. "Development of a Struc-
tured Family Interview," final Progress Report to
the National Assoc. for Mental Health, Jan., 1966.

Jackson, D.D. (ed.). Therapy, Communication and Change,
Vols. 1 & 2, Science and Behavior Books, Inc.,
Palo Alto, California, 1968.

 

Jourard, S. M. The Transparent Self. D. Van Nostrant Co.,
Princeton, N. J., 1964.

 

Kelley, H. H. "Communication in Experimentally Created
Hierarchies," Human Relations, 4, 1951, pp. 39-66;
and J. Ruesh, J. BIOOk and L. Bennett, "The Assess-
ment of Communication," Journal of Psychology, 35,
1953.

 

 

112

Kelman, H. "Compliance, Identification and Internalization:
Three Processes of Attitude Change," Journal of Con-
flict Resolution, 2, 1958.

 

Kerlinger, F. N. Foundations of Behavioral Research, Holt,
Rinehart aid Winston, Inc., N.Y., 1964:

Krippendorff, K. "On Generating Data in Communication Re-
search," presented at the National Society for the
Study of Communication, February, 1969.

Laing, R. D., H. Phillipson and A. R. Lee. Interpersonal
Perceptions, Springer Publ. Co., N.Y., I966.

 

Leary, Timothy. Multi-level Measurement of Interpersonal
Behavior, Berkéley, California: Psychological Con-
sultation Service, 1956.

Leavitt, H. J. "Some Effects of Certain Communication
Patterns on Group Performance," Journal of Abnor-
mal and Social Psychology, 66, 1951.

Leik, R. "Instrumentality and Emotionality in Family Inter-
action," Sociometry, 26, 1963.

 

Lennard, H.L. and A. Bernstein. The Anatomy of Psyghother-
apy. Columbia University Press, N.Y., 1960.

Lerner, P. "Resolution of Intrafamilial Role Conflict in
Families of SchiZOphrenic Patients," Journal of
Nervous Mental Disorders, 3, 1965.

Locke, H. J. Predicting Adjustment in Marriage, Henry
Holt and Co., N.Y., 1951.

 

Lyle, J. "Communication, Group Atmosphere, Productivity
and Morale in Small Task Groups," Human Relations,
14, 1961.

McLeod, J.M., S. H. Chaffee and D. B. Wackman. "Family
Communication: An Updated Report," 1967, a paper
presented to the Theory and Methodology Division
Assoc. for Ed. in Journalism.

McNemar, Quinn. Psychological Statistics, J. Wiley & Sons,
InCO' N.Y., 1962. ‘

McQuitty, L. L. "Elementary Linkage Analysis for Isolating
Orthogonal and Oblique types and Typal Relevancies,"
Educational and Psychological Measurement, Vol. 7,
1957.

113

Mead, G. H. Mind, Self, and Society. Univ. of Chicago
Press, Chicago,I1I., 1934.

Miller, J. G. "Living Systems: Cross-level Hypotheses,"
Behavioral Science, Oct. 10, 1965.

 

Murrell, S. A. and J. G. Stachowiak. "Consistency, Rigid-
ity and Power in the Interaction of Clinic and
Non-clinic Families," Journal of Abnormal Psychol-
ogy, 72, 1967.

O'Rourke, J. F. "Field and Laboratory: The Decision-
Making Behavior of Family Groups in Two Experimen-
tal Conditions," Sociometry, 26, 1963.

 

Parsons, T. and R. F. Bales. Family, Socialization, and
Interaction Process, Free Press, Glencoe, Ill.,
1955.

 

Ruesch, Jurgen and Gregory Bateson. Communication, The
Social Matiny of Psychiatry, New York, Norton,
1951.

 

 

Ryder, R. and D. Goodrich. "Married Couples Responses to
Disagreement," Family Process, 5, 1966.

 

Satir, V. Conjoint Family Therapy, Palo Alto, Science
and Behavior BoOks, Inc., 1968.

Scheflen, A. E. "Stream and Structure of Communicational
Behavior," Behavioral Studies Monograph #1, E.
Penn. Psych. Inst., 1965.

Siegel, S. Non-parametric Statistics, McGraw-Hill, New
York, 1956.

Sluzki, G. E. and J. H. Beavin. "Simetria y Complemen-
‘ taridad: Una Definicion Operacional y una Tip-
ologia de Parejas" (Symmetry and Complementarity:
An Operational Definition and a Typology of Dyads)
Acta Psiquidtrica y Psicologica de America Latina,
11, 1965.

Sluzki, G. E., J. H. Beavin, A. Tarnopolsky and E. Veron.
"Transactional Disqualification," Archives of
General Psychiatry, 18, 1967.

 

 

Strodtbeck, F. L. "Husband-Wife Interaction Over Revealed
Difference," American Sociological Review, 16, 1951.

 

114

Sullivan, H. S. Conceptions of Modern Psychiatry, reprinted
from Psychiatry: Journal of the Biology and Pathol-
ogy of Interpersonal Relations, 3, 8, Feb. 1940 and
May, 1945.

Suppes, P. and R. C. Atkinson, Markov Learning Models for
Multiperson Interactions, Stanford University Press,
Stanford, Calif., 1960.

Szasz, T. S. The Myth of Mental Illness, Paul B. Hoeber,
Inc., New York, 1961.

Thibaut, J. W. and H. H. Kelley. The Social Psychology of
Groups, John Wiley and Sons, Inc., N.Y., 1959.

Trager, G. L. "Paralanguage: A First Approximation," in
Studies in Linguistics, 13, 1958.

Turner, R. "The Ethnography of Experiment," American Behav-
ioral Scientist, April, 1967, Sage Publ., Inc.

 

Verwey, N. E. "Relationship of Adaptability to Interaction-
al Contingency and Interpersonal Prediction," un-
published dissertation, University of Washington,
1962.

Watzlawick, P. An Anthology of Human Communication. Science
and Behavior Books, Inc., Palo Alto, California,
1964.

Watzlawick, P. "Development of a Structured Family Inter-
view," January, 1966, a final report to the Nation-
al Assoc. for Mental Health.

Watzlawick, P., J. H. Beavin and D. Jackson. Pragmatics
of Human Communication, W. W. Norton and Co., N.Y.,

 

 

1967.

Winter, W. D. and A. J. Ferreira. "Interaction Process
Analysis of Family Decision-Making," Family Process,
6, 1967.

Wynne, L. C., I. M. Ryckoff, J. Day and S. I. Hirsch.
"Pseudo-Mutuality in the Family Relations of Schiz-
ophrenics," Psychiatry, 21, 1958.

 

Wynne, L. C. "The Study of Intrafamilial Alignments and
Splits in Exploratory Family Therapy," in Explor-
ing the Base for Family Therapy, N. W. Ackerman,
FTrt. Beatman, and S. N. Sherman (eds.), Fam. Serv.
Asc. of Amer., N.Y.

APPENDIX

MICHIGAN STATE UNIVERSITY East Lansing, Michigan 48823

 

College of Communication Arts Department of Communication
Cable: Commdept
Dear Lansing Area Resident:

A Public Opinion Survey is currently being con-
ducted in this area by a Doctoral candidate in the Depart-
ment of Communication at Michigan State University. We
are interested in what many of the residents of the Lansing
community think about several interesting subjects.

We are interested in having you and your Spouse
participate in this survey, if you have children under ten
(10) years old. This is necessary as the topics in the
survey are of interest to families with young children.

This Public Opinion Survey will remain totally
confidential and you will never be identified by name, or
singled out as an example. We request your permission to
tape record your discussion, as this will help us analyze
our data.

You are part of a scientifically selected sample
and your participation is of considerable importance to
us. If you have any questions, please feel free to ask
the person bearing this letter.

Sincerely,
Robert A. Mark, Instr.

Department of Communication
Michigan State University

115

MICHIGAN STATE UNIVERSITY East Lansing, Michigan 48823

 

College of Communication Arts Department of Communication

PUBLIC OPINION SURVEY - 1970

Instructions:

You will be presented with a series of situations
which couples encounter from time to time. These situa-
tions can be handled in a variety of ways. We are inter-
ested in your opinions on these matters and would like for
you to discuss as many of the alternatives that seem im-
portant to you. From these different views, select an
alternative which you both agree on.

There are no right or wrong solutions . . . only
what both of you agree on as mo§t_appropriate answer for
the situation. Do not worry about the time, as you may
take as much or as little time on each of the following

three situations as you need. Do you have any questions?

Turn the page and begin on the first topic. When you have

finished that one, proceed to the next, etc.

116

117

Imagine that you, your spouse and child (children)
are at home together on a Sunday afternoon, and you are
watching television or listening to the radio when a Civil
Defense alert is broadcast. This broadcast informs you
that you have only 15 minutes to prepare and take shelter
prior to disaster. How would you, or you and your spouse
handle this situation? Please discuss the possible alter-

natives and decide on a course of action.

(When you have completed discussing this topic, and reached
a decision, please go on to the next situation.)

118

Imagine that your child (children) are about 12
years Old and have been receiving some information about
human reproduction in the classroom, as well as from other
children. You accidentally overhear them discussing it
with a friend and you realize that their information is

quite incorrect and misleading. They are not aware that

 

you overheard them. How would you, or you and your spouse
handle this situation? Please discuss the possible alter-

natives and decide on a course of action.

(When you have completed discussing this topic, and reached
a decision, please go on to the next Situation.)

119

Imagine that it is now September, 1970. Your
child (children) are just beginning a new school year.
The major television networks (NBC, CBS, and ABC) have
decided to run more "educational" programming in the
evening hours. In addition to this, there are several
new series on television that are popular among your
child's (children's) friends, also scheduled for late
night viewing. Now, your child (children) have asked to
be allowed to stay up later than they do now to view these
programs. How would you, or you and your Spouse handle
this situation? Please discuss the possible alternatives

and decide on a course of action.

(When you have completed discussing this topic, and reached
a decision, please go on to the next situation.)

120

Will you and your spouse please come to a joint
agreement as to which of the three topics you discussed--
sex education, television viewing attitudes, and Civil
Defense p1anning--was mogp salient or important to you,
and which of these topics was leogp salient or important

to you. Write your decision below:

Most Important/salient:

 

Least Important/salient:

 

121

You have just completed the major portion of this
Public Opinion Survey. We do, however, need some addi-
tional information, which, asis true of this entire

survey, will be kept totally confidential.

1. Please list the ages and sex of your children below:

age sex age sex

2. What was the last year of school you completed?
(Circle the last year)

Grade School: 1 2 3 4 5 6 7 8

High School: 1 2 3 4

Vocational Training: 1 2 3 4 5

University or College: 1 2 3 4 5 6 7 8

3. To be answered only by the Spouse(s) who has (have)
been earning an income during the past year: What is
the title of the position you currently hold at your

place Of employment:

 

In this position, what is it that you do_at work?
(Please give a Short explanation of your job responsi-

bilities.)

 

 

   

M71111111071111)(([Jflfiflijwflfflfimfl‘fi