ficwaA

 

.
D
.
2.

 

 

llllljlllllIjllljflljllllllflllflllllllfll‘fllll

THESLS

   
   

   
 
 

L I B R A R Y
Michigan Stan
University

 

This is to certify that the
thesis entitled
A PARADIGM FOR THE STUDY OF COMMUNICATIVE STYLE:
THEORY, MEASUREMENT, AND A RESEARCH APPLICATION

presented by
David Randall Brandt
has been accepted towards fulfillment

of the requirements for

Ph . D . degree in Commun i'ca t ion

Major professor

Date February 13, 1981

0-7639

 

 

-‘3Ub?22’f°0‘i

 

    
 
    

 
 

‘k-ifjfi 7"" ‘ I .'
c: "at!!!” {1

..‘.

 

 

 

 

OVERDUE FINES:

25¢ per day per item

RETURNING LIBRARY MATERIALS:

Place in book return to rem0\
charge from circulation recor

 

 

 

A PARADIGM FOR THE STUDY OF COMMUNICATIVE STYLE: THEORY,

MEASUREMENT, AND A RESEARCH APPLICATION

BY

David Randall Brandt

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Communication

1980

ABSTRACT

A PARADIGM FOR THE STUDY OF COMMUNICATIVE STYLE:
THEORY, MEASUREMENT, AND A RESEARCH APPLICATION

BY

David Randall Brandt

The study of communication content and its structure
has a counterpart in the investigation of communicative
style. Research on communicative style focuseS'on hgw_actors
use language and gesture in social contexts; it focuses less
on what an individual says, per se, and more on how s/he says
it.

This thesis attempts to develop a paradigm for the
study of communicative style which builds upon the work of
scholars and students from psychotherapy, social psychology,
linguistics, semiotics, and communication, to name a few
disciplines. Researcn findings, conceptual problems, and
methodological problems characteristic of many previous
approaches to stylistic inquiry are considered. Subsequently,
an attempt is made to develop and demonstrate the scientifc
utility of an observational framework for stylistic inquiry,
particularly as it pertains to communication in face-to-face
settings. Meta-methodologically, the importance of seeking
a balance between theoretic precision and social intelligi-
bility is stressed. It is argued that, to achieve social
generalizability, the framework must take into account the
normative and behavioral referents of the members of a social

system under study. It is also suggested, however, that

David Randall Brandt
the approach must incorporate measurement procedures which
satisfy the requirements of objectivization, relativization,
empirical verifiability, reliability, and discriminatory
sensitivity, in order to be useful for scientific inquiry.

In attempting to develop an observational framework
which satisfies the preceding requirements, a series of ini-
tial interactions between strangers were videotaped and
served as stimuli to be observed and scaled. Observers were
given very general definitions of a set of stylistic attri—
butes, and were asked to collaborate and consensually generate
a set of empirical referents for each attribute. Once a
common conceptual and empirical frame of reference was esta—
blished, observers underwent initial practice sessions in
which they viewed a sample of interactants and made ordinal
judgements of each attribute, as they perceived it to be
manifested by the various interactants observed. This stage
of preparation was followed by a second series of practice
sessions in which observers made direct magnitude estimates
of displayed levels of each attribute, relative to a numeri-
cal standard of 100, which corresponded to the "average level
of (a given stylistic attribute) displayed by most persons in
an initial interaction with a stranger." Subsequently, the
rating of actual stimulus tapes was conducted, and the data
obtained from this stage of the research were analyzed to
determine reliability, precision, and empirical verifiability.
The results of the initial research indicated that raters'

estimates were generally reliable, precise and empirically

David Randall Brandt
verifiable. These results paved the way for an application
of the observational procedures in a study of the relation
between communicative style and perceived interpersonal
attractiveness and effectiveness.

The results of the research application indicate that
there are at least two distinct communicative styles, one of
which is related primarily to percieved social attractiveness
and communicative effectiveness, and the other of which is
primarily related to perceived task attractiveness.

The final chapter of this thesis provides a discussion
and summary of the entire research. Specific sections are
devoted to (l) the available evidence pertaining to the
scientific utility of the observational framework, (2) the
results of the research application, (3) the implications of
the findings for future stylistic inquiry, and (4) limita—

tions and reservations concerning this research.

Accepted by the faculty of the Department of
Communication, College of Communication Arts, Michigan
State University, in partial fulfillment of the re-

quirements for the Doctor of Philosophy degree.

5?.
’ é

Director of Dissertation

Guidance Committee: flflf/(Qiifi/Zg/fi/LL , Chairman

AW Wu

U 1

W/LHM

 

 

This dissertation is dedicated to a most remarkable

and special lady, my grandmother, Mrs. Mary Louise Pfister.

ii

ACKNOWLEDGMENTS

I would like to express my gratitude to the persons
who provided support and inspiration, of various sorts,
throughout the completion of this project.

I am deeply indebted, both as a student and as a
friend, to Dr. Gerald R. Miller. His persistence in de-
manding my very best efforts, his patience during my growth
‘to personal and scholarly maturity (and periodic lapses
therefrom), and his enduring confidence in my potential as
a teacher and researcher deserve much greater recognition
than I can give here, so I'll just say "thanks G. R.," and
hope my intentions are understood.

I would also like to thank the members of my guidance
committee. Dr. Joseph Woelfel stimulated my thinking through-
out my graduate study at Michigan State, and offered patient
but thorough criticism of the early drafts of this disser-
tation, most of which were prepared in_absentia. Dr. William
A. Donohue tutored me in an alternative approach to communica—
tion inquiry (hermeneutics), and subsequently, forced me to
reconsider and defend my own perspective in its light. Dr.
William D. Crano was generous with his time and knowledge,
both of which, I hope, are reflected in this dissertation.

Several additional persons deserve acknowledgment here.
To Kim R. Kanaga, F. Joseph Boster, James DinKelacker, Elyse
Kaplan, George A. Barnett, and (of course) N. J. Stoyanoff

iii

go my enduring thanks for their personal and intellectual

support. These people have given me a new understanding

of the lyric "I'll get by with a little help from my friends."
Finally, I wish to express my thanks to my mother and

father, and other members of my family, who did not always

understand what I was doing in graduate school, but who

never failed to tell others how proud they were that I was

doing it!

iv

TABLE OF CONTENTS

Chapter

I INTRODUCTION AND OVERVIEW. . . . .
The Problem . . . . . . .

Overview. . . . . . . .

Summary . . . . . . . .

II LITERATURE REVIEW AND CRITIQUE . . .
Previous Research . . . . .

A Psychological Approach . .

A Linguistic Approach . . .

A Sociolinguistic Approach . .
Norton's Communicator Style Construct
Domain of the Communicator Style

Construct . . . . .

The Communicator Style Measure .

Previous Research Utilizing the

CSM. O O O O O O 0

Summary . . . . . . . .

III THE MEASUREMENT OF COMMUNICATIVE STYLE:

THEORY AND DEVELOPMENT . . . . .
Theoretic Criteria . . . . .

Social Referents Must Be Con—
sidered. . . .
Verbal and Nonverbal Behaviors
Must Be Considered as Comple-
mentary. . .
The System Must Be Quantifiable.

Methodological Criteria . . . .

Objectivization. . . . .

Page

11
ll
12
l6
17
22
23
27
27

35

38

39

4O

43
45
47

48

Chapter
III (continued)
Relativization . . . .

Empirical Verifiability . .
Reliability .- . . . .

Precision . . . . . .
The Observational Framework . . .
Overview. . . . . . .
Interactants . . .
Videotaping the Interactions
Raters . . . . .
Training the Raters. .
Summary and Conclusion . . .
IV THE MEASUREMENT OF COMMUNICATIVE STYLE:

RELIABILITY, PRECISION, AND EMPIRICAL VERI-
FIABILITY . .: . . . . . .

Reliability and Precision.

Empirical Verifiability . .
Procedures . . . . . .
Coders . . .
Behavioral Referents . .
Results .
V A RESEARCH APPLICATION . . . . .

Communicative Competence: Characteris-
tics of the Effective and Attractive

Interactant . . . . . . .
Die thOd o o o o o o o
Competence Operationalized
Observers . . . . .
Procedure . .
Results . . . . .

Reliability of Competence
Measures . .

vi

Page

48
49
49
50

51
51
53

54
56

56
67

71

72
77

78

78
78

79

83

84
88
88
89
89

9O

9O

Chapter Page
V (continued)

Intercorrelations Among Style,
Effectiveness, and Attractiveness

Variables . . . . . . 92
VI DISCUSSION AND CONCLUSION . . . . . 98
Scientific Utility . . . . ' 98
Communicative Style and Interpersonal
Competence . . . . . 105
Implications for Future Research . . 106
limitations of this Research . . . 111
Summary and Conclusions. . . . . 112
REFERENCES . . . . . . . . . . . . 116

vii

LI ST OF TABLES

Table ’. Page
1. Norton's (1978) Communicator Style Measure Items 28
2. Conceptual Definitions and Empirical Indicators

of Communicative Style . . . . . . . 58‘

3. Reliability Estimates for All Stylistic
Attributes . . . . . . . . . . 73

4. Selected Summary Statistics for Stylistic
Attributes . . . . . . . . . . 75

5. Estimated Information-Carrying Capacity of Scale
Values Based on Actual Ranges . . . . . 76

6. Summary of Results Pertaining to Empirical
Verifiability of Dominance . . . . . . 8O

7. Intercorrelations Among Style, Effectiveness,
and Attractiveness . . . . . . . . 91

8. Canonical Analysis Summary Table . . . . 96

viii

LI ST OF FIGURES

Figure Page

1. Giles and Powesland's Accommodation Model of
Speech Style. . . . . . . . . 21

ix

CHAPTER ONE

INTRODUCTION AND OVERVIEW

The study of communication content and its structure
has a counterpart in the investigation of communicative
style. Research on communicative style focuses on hgw.
actors use language and gesture in social contexts; it
focuses less on what an individual says, pg£_§§, and more
on how s/he says it. Even a cursory review of the social
science literature reveals that a concern for the nature
and social consequences of interpersonal communicative
styles is shared by scholars from a number of disciplines,
including psychiatry, psychology, sociolinguistics, and
communication.

That stylistic inquiry is concerned with h9w_indi—
viduals use language and gesture implies that both verbal
and nonverbal behaviors are relevant in designing and
executing empirical studies of communicative style, par-
ticularly within the context of human, face—to-face inter-
action. Further, such behaviors are probably best seen as
complementary in supplying information to interactants.
Given that the conceptualization and measurement of phe-
nomena are inextricably connected, any conception of com-
municative style which views verbal and nonverbal behaviors
as complementary requires that its corresponding method of

l

2
measurement reflect this complementarity. Thus, com-
municative style is conceptualized here as a set of.

anchored attributes, in terms g£_which the verbal and

 

 

nonverbal behaviors g; interactants may be typified and

 

 

 

differentiated. This thesis attempts to develop a para-
:digm for the study of communicative style which is con-

sistent with the preceding conceptualization.l

The Problem

 

Considerable attention has been devoted recently to the
proliferation of meta-theoretic and meta-methodological
debates in communication and other social science disciplines.

In fact, a recent issue of Communication Quarterly (Winter,

 

1977) is exclusively devoted to a symposium on alternative
models for communication theory and research. Although
methodological and theoretic pluralism should probably not
be discouraged in a discipline as pre-paradigmatic (Kuhn,
1970) as the field of communication, we must keep in mind
that applying different conceptions and methods of obser-
vation to the "same" phenomenon will likely yield different
outcomes, thus creating confusion as to the "objective"
character of the phenomenon (Woelfel, 1978). As such con-
fusion persists, consensus regarding a conceptual and
epistemic frame of reference will continue to elude com-
munication scholars. This, in turn, will likely inhibit the
inter-subjective verification of observations of communi-

cative events, and hence, theory-building in our discipline.

3

The value of considering a scientific discipline in
terms of its success in developing a paradigm for inquiry
has been discussed by Rossiter (1977):

To the extent that a community is paradigmatic,

it has achieved a level of maturity, which is

related to social organization and scientific

productivity. Scientific communities general-

ly do move from immaturity to maturity, dis-

organization to organization, and from being

aparadigmatic to paradigmatic. Increased
organization and maturity tend to accelerate

the rate of scientific production by the com—

munity (p.71).

The development of a methodological framework for the
precise description of phenomena is an important step
toward building a paradigm in any discipline. Hawes
(1977) argues that most "mature" sciences began with ex-
tensive descriptions. The importance of this process of
description has been underscored by Dubin (1969), who
notes:

In every discipline, but particularly in its

early stages of development, purely descriptive

research is indispensible. Descriptive research

is the stuff of which the mind of man, the theorist,

develops the units that compose his theories.

The more adequate the description, the greater is

the likelihood that the units derived from the

description will be useful in subsequent theory

building (p.85).

When one traces the growth of the physical sciences,
s/he notes that a crucial step in that growth process
was the establishment of a framework for the description

of phenomena. Born (1962) identifies three principles

associated with this framework. The first he calls

4

objectivization, which aims at "making observations as

 

independent of the individual observer as possible" (p.2).

The second he calls relativization, which suggests that

 

the meaning or definition of a concept can only be deter—
mined through its relation to some standard, or "relative

to the standpoint of the observer" (p.2). Empirical

 

verifiability is the third principle, which demands the

 

elimination of the unobservable from "that which is to be
called scientific knowledge" (p.3).

Development and adherence to the preceding principles
has expedited the establishment of a common epistemic and
conceptual frame of reference among physical scientists,
as well as consensus regarding which concepts, structures,
and observations should be considered "real." This is not
to suggest that the physical sciences are free from contro-
versy. The development of paradigmatic principles has,
however, led to increased research productivity and
scientific achievement.

This thesis is predicated on the assumption that the
deve10pment of a common framework for observing and record-
ing observations is a necessary step in the growth and
maturity of a science of human communication. This further
assumes that a science of human communication is possible,
and that communication scientists can benefit from the
example set by their counterparts in the physical sciences.
There are scholars who claim that imitation of the physical

sciences has dominated social science research, and that

’\

‘\

5
it has, by and large, proven unfruitful. With few ex-
ceptions, however, it is difficult to demonstrate that the
"imitations" actually resemble the imitated (Woelfel,
1978). In particular, one is hard-pressed to identify
measurement systems in communication, psychology, and
sociology which adhere to the principles of objectiviza-
tion, relativization, and empirical verifiability outlined
above. This lack of similitude will become more apparent
when we examine some previous approaches to the measurement.
of communicative style. The point is that, in the absence
of evidence that social scientists have successfully prac-
ticed the principles of scientific description provided by
physical scientists, it is difficult to claim that the
method of science has been adequately tested in relation to
social phenomena; the question of whether human communica-
tion is beyond the scope of scientific inquiry remains an
open one.

It has already been suggested that the study of what
people say has its complement in the study of how they say
it. This assertion alone does not warrant the claim that
communicative style should be explicated and developed as
a cardinal variable in communication research. Consider,
however, that, particularly when it is inconsistent with
verbal content, the nonverbal component of communicative
behavior often overrides the former as a basis of meaning
or interpretation (Mehrabian, 1971; Hymes, 1974). Con-

sider, also, that an entire discipline, sociolinguistics,

 

6
has been built around exploring the interrelations among
speech performance, social structure, and social differenti-
ation (Grimshaw, 1973). Finally, consider the growing
corpus of literature which focuses specifically on inter—
active or communicative style (e.g., Ring & Wallston, 1968;
Ring, Braginsky & Braginsky, 1966; Giles & Powesland, 1975;
Norton, 1977, 1978; Norton & Pettegrew, 1977; Norton &
Warnick, 1976; Norton & Miller, 1975; Miller, 1977; Pettegrew,
1977, 1978; Brandt, in press). Taken together, these con-
siderations suggest that communicative style is a signifi—
cant and pervasive component of social interaction. Further,
to conceptualize communicative style as a set of attributes
in terms of which the verbal and nonverbal behaviors of
individuals may be typified is consistent with Berger and
Luckmann's (1967) assertion that the "social reality of
everyday life is apprehended in a continuum of typifications,"
and that "social structure is the sum total of these typifi-
cations" (p.33). Such a conceptualization also lays the
foundation for the development of a precise method of de-
scribing individual communicative styles through the quanti-
fication of stylistic attributes. Given the preceding,
the purpose of this thesis can now be stated more formally:
This thesis proposes a paradigm or model for the study of
communicative style in face-to-face interaction; an attempt
is made to outline, develop, and apply a precise method for
observing and analyzing style in relation to its social

contexts.

7

Overview

This chapter introduces the reader to the area of in-
quiry, provides a rationale for such inquiry, and provides
a general overview of the contents of the thesis.

Chapter Two reviews the literature on communicative
style. Special attention is devoted to the conceptuali-
zation of style presented by Norton (1978), as it appears
to represent the most ambitious attempt yet to develop a
holistic communicator style construct. Research findings,
conceptual problems, and methodological problems pertain-
ing to Norton's style construct are subsequently discussed.
The chapter concludes with a discussion of steps toward
refining the conceptualization and measurement of communica-
tive style undertaken in this research.

Chapter Three attempts to develop an exemplar for the
empirical assessment of interactive or communicative style.
A theory and principles of measurement are phesented. Pro-
ceeding from a number of extant conceptions of social be-
havior, including attribution theory, symbolic interactionism,
learning theory, linguistics, and sociolinguistics, it is
argued that the average level of a given stylistic attribute
(e.g., attentiveness) displayed by "most persons" in a given
situation, serves as a benchmark in relation to which the
communicative conduct of an individual is assessed. It is
further argued that transforming this social system of com-
parison from a vernacular to a quantitative one should render

it useful for scientific description, provided the

8
requirements of precision, reliability, objectivization,
relativization, and empirical verifiability are satisfied.
A direct magnitude estimation procedure (Torgerson, 1958;
Hamblin, 1974; Shinn, 1974) by which the stylistic attri-
butes were scaled is described. Also, procedures by which
observers were prepared to apply the proposed method to
the precise typification and differentiation of individual
styles are discussed.

ChapteryFour presents data pertaining to the reliabia-
Bility, precision, and empirical verifiability of the pro-
cedures developed herein.

Chapter Five addresses a specific question to which
the model developed here was applied; a study of the re-
lation between communicative style and social perceptions
of interpersonal attractiveness and communicative effective-
ness is described. Some relevant literature on communica-
tive competence is reviewed. An operationalization of per-
ceived competence is offered. Subsequently, an experiment
designed to investigate the relation between communicative
style and communicative competence is outlined. Finally,

the results of the experiment are presented.

Chapter Six provides a discussion and summary of the

..._

work presented in the thesis. Specific sections are de-
voted to (1) the available evidence pertaining to the utility
of the measurement system developed here, (2) the results
pertaining to the relation between communicative style and

perceived attractiveness and communicative effectiveness,

9
(3) the implications of the findings for future com-
munication inquiry, and (4) problems and limitations of
this research. The final chapter concludes with a summary

of the entire thesis.

Summary

This chapter was designed to introduce the reader to
the problem being investigated, to provide a rationale for
such investigation, and to provide an overview of the
contents of this thesis. Communicative style was conceptual-
ly defined as a set of anchored attributes in terms of
which the verbal and nonverbal behaviors of individuals
might be typified and differentiated. The importance of
devising a descriptive framework for social constructs
such as communicative style was stressed, particularly with
regard to the development of a science of human communica-

tion. Finally, a brief description of the purpose and scope

of each chapter of the thesis was presented.

10

FOOTNOTES

 

1
Masterman (1970) found that Kuhn (1970) uses the

term "paradigm" in at least twenty-two different senses,
ranging from "a concrete scientific achievement" to a
"characteristic set of beliefs and preconceptions" (pp.
39-42). Rossiter (1977) furnishes the following precis
of Kuhn's conception of a paradigm: "A paradigm is a
world View about how theoretic work should be done in a
particular subject area which is shared by those who
actually do theoretic work in that area. It includes
agreements about: assumptions about the nature of the
phenomenon about which the theory is being built; variables
which are most important for study to understand the
phenomenon about which the theory is being built; and
acceptable methods for supporting assertions about the
phenomenon about which the theory is being built" (p.70).

The term "paradigm" is used here to denote a con-
ceptual and epistemic model for empirical inquiry, whose
principles are explicated in such a way that other scientists
and scholars can readily identify and apply or replicate
them. As such, this conception of a paradigm is strongly
rooted in operational philosophy (Rapoport, 1953).

CHAPTER TWO

LITERATURE REVIEW AND CRITIQUE

The purpose of this chapter is to provide an overview
of the literature on communicative style. A number of
linguistic, sociolinguistic, and psychological approaches
to style are reviewed and critiqued. Special attention is
devoted to the work of Norton and his assdbiates toward the
development of a holistic communicator style construct.

The conceptual domain of that construct, its origins, pre-
vious empirical studies, and conceptual and methodological
problems are discussed. Finally, the chapter outlines some

steps undertaken here to refine the conceptual and methodo-

logical framework for stylistic inquiry.

Previous Research

 

In one sense, the study of interactive or communica-
tive style is not entirely novel. Previous studies have
explored means by which interactive behavior might be
characterized (Bales, 1970) and/or very general styles
identified (Joos, 1959; Ring, Braginsky & Braginsky, 1966).
Research has also examined the effects of various indi-
vidual verbal and nonverbal cues on perceived communica-
tive effectiveness, interpersonal attraction, and impres-
sion formation (Mehrabian, 1967, 1969; Argyle & Kendon,
1967; Lowe & Goldstein, 1970; Holstein, Goldstein & Bem,

ll

12
1971; Leginsky & Izzett, 1973; Pacanowsky & Fink, 1976).
Typical findings indicate that some individual eXpressive
behaviors are significantly correlated with certain per-
ceptions of and/or liking for a communicator. Most of
these studies, however, tend to focus on isolated behaviors,
neglecting a holistic consideration of verbal and nonr
verbal behaviors in relation to a general communicative
style construct. As for the studies which have attempted
to identify and explore general communicative styles, a
number of conceptual and methodological problems can be
cited. Careful examination of some of these studies should
highlight some of the problems. Therefore, let us turn to
some previous approaches to the study of communicative style

in psychology, linguistics, and sociolinguistics.
A Psychological Approach

An early attempt to identify general communicative
performance styles is reflected in the work of Ring and his
associates. Ring, Braginsky, and Braginsky (1966), focusing
on actor-related sources of variance in interpersonal beha-

vior, defined the concept performance style as "an indivi—

 

dual's characteristic mode of interaction with others"
(p.205). These researchers proposed a typology of three
performance styles based on three specific actor attributes:
motivation, knowledge, and skill. The three performance
styles, labelled p, r, and g, were presumed to represent
"fundamental and qualitatively different styles of inter-

action with others" (Ring & Wallston, 1968; 147). According

13
to this typology, the performance style p is characterized
by a preference to avoid interpersonal contexts which call
for role enactment, a general lack of social agility re-
quired for successful performance-giving, and a deficiency
in knowledge of the role-demands appropriate to a variety of
social settings (Ring, 32 a1., 1966; 208). The performance
style a is characterized by skill in interpersonal rela-
tions, a somewhat Machiavellian motivational orientation,
and a knowledge of how to exploit social situations and
others for personal gain (p.212). The performance style 3
is typified by a high degree of behavior adaptability to
situational cues, a strong need or motivation for social
approval, and an increased sensitivity to the conventional
script demands of a gamut of social situations (p.215).
The performance styles E and g are very similar, except for
the motivation of the actor to give good social performances;
the r is motivated by perceived personal gain, whereas the
g is motivated by a need for social approval or confirmation
(p.215). .

To measure these performance styles, a 55-item, "true-
false" test, called the Performance Style Test (PST) was
constructed (Ring & Wallston, 1968). Typical items include
statements like "I can fit pretty easily with any group,"

"I like to conform to custom and to avoid doing things that
people I respect might consider unconventional," and "I can
usually get people to do what I want“ (Ring & Wallston, 1968;
150-151).

Typifications of individuals as "social chameleons,"

14
"social buffoons," or "Machiavellians" are generally pre—
alent in the vernacular of everyday social intercourse.
It was noted in Chapter One, however, that a scientifically
useful framework for the description of individual com-
municative styles should fulfill the requirements of ob-
jectivization, relativization, empirical verifiability,
reliability, and precision. It is possible to question
the utility of the preceding approach for failure to meet
a number of these requirements. First, the PST fails to
meet the requirement of objectivization, since measurement
is subject-centered (Torgerson, 1958; 49) and only an
individual's self-assessment is taken into account; assess- .
ments by and attempts at subjective verification among
several observers are not taken into account.

It is also difficult to empirically verify the styles
identified or "measured" by the PST, since no epistemic
frame of reference is explicated. Thus, the question of
hgw_a communicator behaviorally signals to others that s/he
"likes to conform to custom" cannot be addressed empirically.

Finally, both conceptually and methodologically, the
approach to communicative style developed by Ring and his
associates lacks precision. The three styles which com-
prise the typology are so vague and few in number as to
provide little information about the individuals which are
classified. This shortcoming can be attributed, in part,
to the fact that Ring, et al.'s approach is a classifica-

tory one. As Carnap (1966) has noted, however, even though

15’

classificatory information is inherently less precise than
comparative or quantitative information, the use of in-
creasingly precise classificatory systems facilitates in-
creased precision in the placement and differentiation of
objects and events, and thus, yields more information about
them. This also applies to the precision of the symbol set
or scaling procedure used to report comparisons of objects
or events which are placed in a given class (p.51). While
the use of a dichotomous scale (e.g., "true-false" scale)
may be apprOpriate for some judgment or measurement tasks
(e.g., "measuring" verdict in an empirical investigation
of juridic decision-making), its use in attempting to dif-
ferentiate a potentially vast range of communicators and
their respective styles is, at best, highly questionable.

Beyond the problems previously cited, it should be
noted that the approach of Ring and his associates is
admittedly aimed at identifying interpersonal and motiva-
tional orientations of individuals as a basis for typifying
"characteristic modes of interaction with others." As such,
this approach focuses less on an individual's verbal and
nonverbal behaviors, and more on his/her psychological
orientation to interaction. Since this thesis is aimed at
developing a framework for stylistic inquiry by which be:
haviors may be typified, and given the preceding criticisms,
the approach of Ring and his associates must be rejected as
inadequate for our purposes.

An alternative approach to stylistic inquiry is

l6
reflected in the work of a number of linguists. In the
next section, we will examine a linguistic approach to the

study of communicative style.
A Linguistic Approach

An attempt will not be made here to outline the theory
and method of stylistic analysis in linguistics; others
(e.g., Hymes, 1974) have addressed the task with considerably
more energy and expertise than can be mustered by this wri-
ter. Suffice it to say that linguistic studies of com-
municative style tend to focus primarily on linguistic
structure or the grammatical and phraseological elements of
language behavior in relation to social class, topical, situ-
ational, and relational variables.

'An example of a linguistic approach to the study of
communicative style is reflected in the work of Joos (1959).
Joos has pointed out that, for different types of encounters,
persons adopt different linguistic styles or modes of lan-
guage choice and sequencing. He distinguishes between

intimate, casual-personal, social—consultative, formal and

 

 

frozen linguistic styles, and provides some description of
their differences. For example, casual style is contrasted
with consultative style, in that the former is characterized
by ellipses in grammatical construction and the use of slang,
and no attempt is made (as in consultative style) to pro-
vide background information pertaining to the topic(s) of
conversation. Both casual and consultative styles differ

from intimate style, where there is no reference to public

17
information, but where speakers use language to define and
regulate their relationship (Joos, 1959; 109-112).

The work of Joos (1959) is significant in that it
points to the influence of topical, situational, and re-
lational variables on an actor's choice and sequencing of
language. The absence of an attempt to integrate such verbal
choices with nonverbal behaviors is a primary weakness in
his approach.' Indeed, Joos (1959) makes it the specific
condition of stylistic analysis that it crosscut the usual
grammatical comparmentalizations (pp.112-113). As Hymes
(1974) has noted, however, "Joos did not follow up his
conception with any indication of an empirical approach to
the identification and analysis of styles. It has been left
for scholars working within a sociolinguistic approach
(Gumperz and Ervin-Tripp) to do this. Much more remains to
be done" (p.178).

The typology of styles proposed by Joos (1959), there-
fore, is inadequate for the purpose of developing a model
for stylistic inquiry which is consistent with our con-
ceptualization of communicative style. As verbal and non-
verbal behaviors are viewed as complementary and inextricably
linked, they warrant simultaneous and integrated empirical

consideration.
A Sociolinguistic Approach

Sociolinguistics is said to be concerned with both the

structure of language and its performance by users of the

18
language, in relation to social structure and social dif-
ferentiation (Grimshaw, 1973). Recently, Giles and
Powesland (1975) have proposed what they term a "social
psychological model of speech diversity; they propose and
attempt to develop an accomodation model of speech style
based largely on the work of attribution theorists such as
Kelley (1973) and Jones and Davis (1965). In introducing
the reader to their purpose, Giles and Powesland state
that:

General appearance, facial expressions, and gestures

provide useful clues as to what sort of person we

are dealing with. What he says to us, whether it is

directly biographical or not, and whether we accept

it at face value or not, tells us a great deal about

him. The manner in which he speaks can also be used

as a basis of judgment, and it is with this source of

information that this book is mainly concerned (p.1).
That they are concerned with the manner in which persons
use language, in relation to social evaluation and response,
suggests that Giles and Powesland's focus is probably best
characterized as sociolinguistic.

The accomodation model proposed by Giles and Powesland
rests on the assumption that a person can elicit more
favorable interpersonal evaluations from others by reducing
dissimilarities between them. To accomplish this, speakers
often "adapt or accomodate their speech towards that of their
interlocutors...at least one member of an interactive dyad
tends to adopt the speech patterns of the person to whom he

is talking" (Giles & Powesland, 1975; 155-156). Giles and

Powesland term such accommodative changes in interactive

l9

behavior "convergent" behavior, whereas attempts to main-
tain or modify one's speech behavior to increase inter-
personal dissimilarity is termed "divergent" behavior.
According to Giles and Powesland, "speech convergence is a
strategy of identification with the speech patterns of an
individual internal to the social interaction, whereas
speech divergence may be regarded as a strategy of identi-
fication with regard to the linguistic norms of some
reference group external to the immediate situation" (p.156,
italics in the original). In attempting to elicit favorable
impressions or evaluations from others, a speech "accommo-
dator" undergoes the following process:

There is a dyad consisting of speakers A and B.

Assume that A wishes to gain B's approval. A then:

(1) samples B's speech and (i) draws inferences as

the personality characteristics of B, (ii) assumes

that B values and approves of such characteristics,

(iii) assumes that B will approve of him (A) to the

extent that he (A) displays similar characteristics,

' and (2) chooses from his speech repertoire patterns

of speech which project characteristics of which E

is assumed to approve (p.158).

Figure 1 displays the accommodation model of speech
style proposed by Giles and Powesland (1975). Based on
the implications of attribution theory, the model suggests
that the accommodation is likely to foster favorable speaker
evaluation and induce reciprocal accommodation, while non-
accommodation is likely to foster a relatively unfavorable

speaker evaluation and induce the listener to maintain his

regular speech style. However, when this behavior is

20
attributed to a lack of language ability or to external
pressures, rather than to a lack of effort on the part of
the speaker, then qualifications have to be made by the
listener and impressions formed are not so negative (pp.
164-165).

The model proposed by Giles and Powesland is both
provocative and empirically testable. Indeed, the authors
provide considerable evidence that accommodative changes
take place in language choice and speech accent (i.e.,
in verbal and paralinguistic behaviors) of individual
interlocutors. At least two weaknesses in their approach
may, however, be identified. First, while they claim that
"facial expressions and gestures provide useful clues as to
what sort of person we are dealing with" (p.l), Giles and
Powesland choose to focus on the purely verbal and para—
linguistic features of communicative style, and fail to
account for other nonverbal or coverbal aspects of behavior.
Thus, while they offer a descriptive framework for stylistic
inquiry which is more in accordance with the kind developed
here, the approach of Giles and Powesland falls short of
the mark for failure to capture some of the more pervasive
features of nonverbal interactive behavior. This may be
attributable to their concern for speech rather than 99m:
municative style. Regardless, the inclusion of additional
kinesic and facial cues seems required in order to develop
a holistic and descriptively robust framework for the empirical

assessment of communicative style.

 

 

,’\\\

21

eea.d .lmemav eceammzom a mmHnw some
mm>fluoa m.memmmm
ou mm Susannumocs mo
mmsmomn coaumcoesooom mo
muflaflnwmmom bum coaummo
Inmm m>eummcc beamscmuudhl. whammcnm Hmcucpxm

 

nomwmm m>wumcoEEooom mcoo
Imp on huflaenm u.cmms Hm
Ixmmmm mmsmomn waco nommmm
m>AHMbOEEooom «coflummo

 

 

lumm m>fiummmc cmpmscmuuénw Anoomv muflaflna.Tllllscoflumboaeooomcoz
coHumMOEEooom
wo coaumoonmeomn on
cam coaummohmm m>Hummmznl Amo xomHV uuommm
memcmm
mm>HUOE

m.umxmmmm mo mucflmunmocs

on mcflzo bmumscmuum usn

COHDMUOEEOUUm Hmooumflommn! whammmum accumuxm.TllllllcoHUMUOEEooom
can coflumwoumm manmno>mm

 

manfimmom ma coflumcoe
IEooom mo coaumooumfiomn
can coaummoumm manmno>mmnw uH0mmm

 

 

 

 

mmcommmm Hmcmumflq cofiummoumm Hmcmumfiq pod nommmm

camum sommmm mo Hmboz COHDMbOEEooom m.ccmHmmBOm can mmafiw

H WMDUHE

22

A second weakness in the approach of Giles and
Powesland, and one which it shares with some of the ap-
proaches previously examined, is a general failure to spe-
cify an epistemic frame of reference for determining
accommodative versus nonaccommodative shifts. How is con-
vergence of speech accent detected? How is divergence
likewise detected? This information is largely missing
from Giles and Powesland's discussion, and the need for its
explication, so that other scientists and scholars may em-
pirically test the predictions of the accommodation model,
seems sufficiently clear.

The preceding review is by no means exhaustive, but it
is generally representative of approaches to the study of
communicative style from various disciplinary perspectives.
More recently, an ambitious attempt to develop and opera-
tionalize a holistic and comprehensive communicator style
construct has been undertaken by Norton and his associates,
and considerable attention shall be devoted to this work in

the next section.

 

Norton's Communicator Style Construct

The foundations of the communicator style construct are
presented in Norton (1978), though research utilizing the
construct and its measurement, the Communicator Style Measure
(CSM), has been going on since 1974. Briefly, Norton con-
ceptualizes a communicator style construct consisting of

tend predictor variables (dominant, animated, attentive,

23
dramatic, open, impression leaving, relaxed, voice, friendly,
and contentious) and one dependent variable (communicator
image), which corresponds to an individual's perceived
effectiveness and attractiveness as a communicator. It
should be fruitful, at this point, to furnish an adumbration
of each of the variables which, together define the domain

of Norton's Communicator Style Construct.

Domain of the Communicator Style Constructl

Dominance

 

Dominance is one of the components of communicative
style identified by Norton (1978). According to Norton,
the dominant communicator "tends to take charge of social
interactions" (p.99). The dominant communicator tends to
speak frequently during social interaction and to control
the flow of information during conversation. Norton claims
that, as a style variable, dominance pervades the literature
(Bales, 1970; Leary, 1957; Mann, Gibbard & Hartman, 1967;
Schutz, 1958; Ideberman, Yalom & Miles, 1973), encompassing
a wide range of semantic and operational meanings. Norton
indicates that the literature tends to focus on physical mani-
festations of dominance, psychological correlates of domi-
nance, and dominance as a predictor of behaviors, attitudes,

and perceptions (p.99).

Dramatic

A dramatic communicative style is exemplified by the use

24
of exaggerations, fantasies, stories, metaphors, rhythm,-
voice, and other expressive devices to highlight or under-
score verbal content (Norton, 1978; 100). As a style
variable, dramatizing covaries with a number of communica-
tive phenomena, including self-image, anxiety, status,

ambiguity tolerance, and critical group functions (p.100).

Contentious

 

A communicator who is contentious has an aggressive and
argumentative style. Norton contends that, while few data
are available in the pertinent literature, contentiousness
emerged as a covariate of dominance in his pilot studies,
and it tended to entail negative attributes of communicative
style (p.100). Norton includes contentiousness in his style
construct because "it was thought that it would provide a

greater understanding of the dominance component" (p.100).

Animatedness

 

Nonverbal, particularly kinesic cues, define the domain
of the animated component of style. According to Norton,
an animated style is characterized by frequent and sustained
gestures, frequent and sustained eye contact, and the use of
a variety of facial expressions (p.100). Relevant literature
indeed suggests that the active use of gestures, postures,
body movements, eye contact, and facial expressions charac-
terize the nonverbally animated individual (Dittman, 1962;

Goffman, 1961; Rosenfeld, 1966; Schefflen, 1965).

25

Impression Leaving

 

The concept of impression leaving centers around
whether a communicator is remembered because of his/her
expressive behavior. This variable seems best conceived as
a function of rather than a component of an individual's
communicative style. Nonetheless, Norton includes impres-
sion leaving as a style component, and suggests that "a
person who leaves an impression should manifest a visible or

memorable style of communicating" (p.100).

Relaxed
Norton notes that Sullivan (1953), who defined psy-
chiatry as the study of interpersonal relations, points

to the anxious-not-anxious, relaxed-not relaxed, and tense-

 

 

not tense dimensions of behavior as a key to personality.

 

Similar dimensions are included in Ruesch (1957), Bales (1970),
and Mann‘s (1967) systems for analyzing interpersonal pro-
cesses. Thus, Norton includes relaxedness as a stylistic
variable in his system, claiming that it "opens the door to

rich and complex analyses" (p.100).

Attentive

 

The attentive communicator makes sure that the other
person (i.e., interlocutor) knows that s/he is being listened
to. Norton points out that their is little empirical re-
search describing attentiveness pg£_§e_as a style variable.
He notes that attentiveness is generally embedded in the
interpersonal and theraputic literature under the label

"empathy" or "listening" (Rogers, 1951). In addition,

.\

26
under the heading of "attentiveness" a number of nonverbal
covariates of conversational behavior have been researched

(Duncan, 1972; Naiman & Breed, 1974).

Open

Behavior associated with the Open component of style
includes communicative activity which is "conversational,
expansive, affable, convivial, gregarious, unreserved, un-
secretive, frank, possibly outspoken, definitely extroverted,
and obviously approachable" (p.101). Stylistically, the
open communicator readily reveals personal information about
the self during social interaction. Norton contends that
research on openness is abundant, and that typical findings
indicate that an open communicator is perceived as attractive
and trustworthy" (p.101). He further maintains that open—
ness is a pertinent style variable because it relates to

-trust, reciprocity, nonverbal behavior, and liking (p.101).

Friendly

 

Stylistically, friendliness serves a sort of "stroking"
function in social intercourse. A friendly communicative
style is typified by behavior ranging from simply being un-
hostile to encouraging more intimate behavior. Norton
suggests that friendliness has been treated by previous
writers in terms of confirmation/d1sconfirmation (Watzlawick,
Beavin & Jackson, 1967), and supportiveness or stroking
(Steiner, 1974). In general, a friendly communicator en-
courages and acknowledges input from others during inter—

action, and attempts to be supportive and non-threatening (p.101).

27

Communicator Image

 

The communicator image variable corresponds to an
individual's perceived attractiveness and effectiveness
as a communicator. The items in the CSM which are
designed to measure it basically attempt to tap a person's
image of his/her communicative ability. It is assumed
that a person who has a "good" communicator image finds
it easy to interact with others, whether they are intimates,

friends, acquaintances, or strangers (p.101).
The Communicator Style Measure

Communicator style is operationalized by Norton in
the form of the Communicator Style Measure (CSM). The
CSM consists of a series of items, each presumed to ex-
press the essence of or otherwise "tap" a given style
variable, for each of the variables included in the style
construct. Table 1 displays the final set of items chosen
from an initial pool of 101 by Norton (1978). The items
are typically scaled using four or five point, Likert-type
scales, and data are obtained by having respondents rate
either themselves, or some other communicator, in terms of
the CSM. Items are summed within each variable, but not
across variables, since cumulativity across variables is

not assumed (Norton, 1978).
Previous Research Utilizing the CSM

The CSM, or modifications thereof, have been used in

28

TABLE 1

Norton's (1978) Communicator Style Measure Items*

 

 

 

 

 

 

Dominant
(1) In most social situations, I generally speak very
frequently.
(2) In most social situations I tend to come on strong.
(3) I have a tendency to dominate informal conversa-
tions with other people.
(4) I try to take charge of things when I am with people.
(5) I am dominant in social situations.
Dramatic
(6) My speech tends to be very picturesque.
(7) I very frequently verbally exaggerate to emphasize
a point.
(8) Often I physically and vocally act out what I want
to communicate.
(9) Regularly I tell jokes, anecdotes, and stories when
I communicate.
(10) I dramatize a lot.
Contentious
(11) Once I get wound up in a heated discussion I have
a hard time stopping myself.
(12) Very often I insist that other people document or
present some kind of proof for what they are arguing.
(13) In arguments I insist upon very precise definitions.
(14) When I disagree with somebody I am very quick to
challenge them.
(15) I am very argumentative.
Animated
(16) I actively use facial expressions when I communicate.
(17) I am very expressive nonverbally in social situa-
tions.
(18) I tend to constantly gesture when I communicate.
(19) People generally know my emotional state, even if
I do not say anything.
(20) My eyes tend to reflect to a very great degree

exactly what I am feeling when I communicate.

(continued)

29

TABLE 1 (continued)

Impression Leaving

(21)
(22)
(23)
(24)
(25)
Relaxed

**(25)
(27)

(28)
**(29)

(30)

Attentive

 

(31)

(32)
(33)
(34)
(35)

What I say usually leaves an impression on
people.

I leave people with an impression of me they
tend to remember.

The first impression I make on people causes
them to react to me.

The pay I say something usually leaves an im-
pression on people.

I leave a definite impression on people.

I am conscious of nervous mannerisms in my speech.
As a rule, I am very calm and collected when I
talk.

Under pressure I come across as a relaxed speaker.
The rhythm or flow of my speech is affected by my
nervousness.

I am a very relaxed communicator.

I can always repeat back to a person exactly what
was said.

I always show that I am very empathic with people.
I am an extremely attentive communicator.

I really like to listen vegy carefully to people.
I deliberately react in such a way that people
know that I am listening to them.

 

 

I readily reveal personal things about myself.
I am an extremely open communicator.

Usually I do not tell people very much about
myself until I get to know them uite well.

As a rule, I openly express my feelings or emo-
tions.

I would rather be open and honest with a person
than closed or dishonest, even if ip_£§_painful
for that person.

 

 

(continued)

 

Friendly

(41)
(42)

(43)

(44)
(45)

30

TABLE 1 (continued)

I always prefer to be tactful.

Most of the time I tend to be ver ry encouraging
to people.

Often I express admiration to a person even if
I do not strongly feel it.

I am an extremely friendly communicator.

I habitually acknowledge verbally other's con-
tributions.

Communicator Image

(46)

(47)
(48)

(49)

(50)

The 321.1 communicate influences my life both
positively and dramatically.

I am a very good communicator.

I find it very easy to communicate on a one-to-
one basis with strangers.

In a small group of strangers I am a very gOOd
communicator.

Out of a random group of five people, including
myself, I would probably have a better communi-
cator style than 1, 2, 3, or 4 of them.

 

 

 

* From Norton (1978), p.103.
** Reverse the scoring of this item.

31

a number of studies to date. Norton (1978) reports that
two clusters of style variables, each interpreted as an
active communicative style in its own right, best predict
a "good" communicator image: for the first, a kind of
active listening seems to be the defining quality, whereas
for the second, an active sending of messages seems to be
the defining quality (pp.105-106). In a related study,
Norton and Pettegrew (1977) report that a dominant and
open communicative style (i.e., a style characterized by
the active sending of messages) was judged significantly
more attractive than a dominant and not open, not dominant
and not relaxed, or a not dominant and not relaxed style.

In the area of therapeutic communication, Pettegrew
(1977) and Pettegrew and Thomas (1978) have demonstrated
that persons manifest a sort of "therapeutic" communicator
style, particularly when approached by a person undergoing
emotional uncomfortableness, or stress, and that even this
"specialized" communicator style varies depending on whether
the situation is defined as formal or informal. Specifi-
cally, in formal therapeutic relationships, an attentive,
animated, and impression leaving style best predicts a
positive therapeutic climate, whereas in informal relation-
ships, a friendly, attentive, dominant, and impression
leaving style best predicts such a climate.

In the area of instructional communication, Norton
(1977) found that an attentive, relaxed, friendly, and

impression leaving communicative style was positively

 

32
related to teaching effectiveness.

Bradley and Baird (1977) investigated the relationship
between several approaches to management or leadership in
organizations and communicative style. Their findings
indicate that democratic management was characterized by
a relaxed, animated, attentive, and friendly communicative
style. Laissez-faire leadership was similar, except for a
lack of animatedness in communicative behavior. The auto-
cratic leader, on the other hand, was characterized by a
primarily dominant, and to a lesser extent, relaxed com-
municative style.

In other studies, the CSM has been employed to examine
interactive assertiveness (Norton & Warnick, 1976), con-
versational dominance in dyads (Norton & Miller, 1975;
Miller, 1977) and perceived communication ability and
affiliation in triads (Norton, Schroeder & Webb, 1975).
Typical findings indicate that communicative style, par-
ticularly dominant behavior, covaries with perceived com-
municative effectiveness.

The initial efforts of Norton and his associates have
been of considerable value in terms of the development of
a holistic communicator style construct. Further, the
framework proposed by Norton, pp 31. seems generally suitable
for the purpose of the present inquiry. First, it provides
a set of attributes by which the communicative behaviors
of individual interlocutors may be typified. Secondly, the

attributes seem appropriate for the description Of both

33
verbal and nonverbal behaviors. Finally, the approach
has been used in a number of investigations of communica-
tive style to date, so that a data base exists in relation
to which findings from studies such as the present can be
compared.

As with any new construct and/or set of procedures
designed to empirically assess it, careful scrutiny and
attempts at refinement are not only warranted, but ought
to be encouraged among members of a scientific community.
Further, it was earlier argued that a scientifically use-
ful system of observation and description should meet the
requirements of objectivization, relativization, empirical
verifiability, reliability, and precision. While Norton's
communicative style framework conceptually approaches what
we are seeking, the CSM fails to fulfill any of the above
requirements. First, the requirement of objectivization
is not satisfied, since measurement is subject-centered,
and only an individual's self-assessment is taken into
account. With few exceptions, observers' assessments of
an individual's communicative style are ignored.

Second, the criterion of relativization is not satis-
fied. Only in the case of item 50 under the variable "com-
municator image" (see Table 1) is any referent for com—
parison specified. In effect, communicators are asked to
assess their own styles in a sort of social vacuum. Thus,
it is not surprising that respondents in many previous

studies report that they are good communicators across a

34

widely diverse set of communicative styles (Norton, 1978).

Third, the requirement of empirical verifiability is
not satisfied by the CSM. Regardless of whether a person
rates himself/herself or is rated by others, to characterize
that person's style as attentive or dominant implies that
there are behaviors associated with attentiveness and domi-
nance which are observable and socially intelligible. Yet,
no such set of behaviors is specified, and thus, no empiri-'
cal basis for the characterizations can be offered.

A fourth flaw in the CSM is its failure to satisfy
the scaling requirement of precision. The final decision
concerning how to scale a theoretic construct or variable
should be solidly based on assumptions about the phenomenon
being measured, as well as the purpose of the measurement
(Nunnally, 1967). Recall that one of the primary aims
of the CSM is to provide observers with means by which
various communicators may be differentiated with regard to
displayed levels (or perceived levels) of stylistic attri-
butes. Given the potentially vast range of communicators
(and their respective styles) which an observer may en-
counter over time, it is questionable whether the scaling
method used in previous studies (four or five point, Likert-
type scales) affords sufficient precision to allow observers
to report perceived differences between communicators,
particularly if such differences are fine ones.

Finally, there is evidence that the CSM does not fulfill

the scaling requirement of reliability. Norton (1978) reports

35
internal reliabilities2 as low as .37 (friendly), using
only a four point scale. Only one variable achieved a
reliability greater than .80 (dominant). Further, Brandt
(1976) reports even lower internal reliability (alpha)
coefficients when the range of the scale is increased to
a lOO-point continuum (.27 to .63). In the case of both
data sets, the average reliability estimates were approxi-
mately .60 to .65. This is not, by any standard of which
the author is aware, an optimum level of reliability of
measurement.

In short, while Norton's approach to the study of
communicative or interactive style is conceptually closer
to the type of system we seek to develop in this thesis,
it suffers from at least five methodological shortcomings.
It is on the reconciliation of these methodological short-
comings that the next chapter, in part, focuses. A theory
and principles of operationalizing the system of stylistic
attributes identified by Norton are outlined. An attempt
is made to show how the system fulfills each of the require-
ments of a framework for the scientific description of
phenomena. Also, procedures by which observers were pre-
pared to apply the observational system to the precise
typification and differentiation of individuals' communica-

tive styles are reviewed.

Summary

This chapter was designed to provide an overview of

36
some of the literature on communicative style. A number
of psychological, linguistic, sociolinguistic, and com-
munication approaches to the study of style were reviewed
and critiqued. Each approach was shown to fail to fulfill
at least one or more of the requirements of objectivization,
relativization, empirical verifiability, reliability, or
precision. The approach of Norton and his associates was
found to be most compatible with the goals of the present
inquiry. Methodological shortcomings of Norton's approach
were outlined and a proposal for their reconciliation was

offered.

37

FOOTNOTES

 

1
This sectidn largely paraphrases Norton's (1978)
description of the origins and domain of a communicator
style construct. For a more detailed discussion, see
Norton (1978).

2
Norton (1978) does not report the method used to
determine internal reliability.

CHAPTER THREE
THE MEASUREMENT OF COMMUNICATIVE STYLE:

THEORY AND DEVELOPMENT

Earlier in this thesis, communicative style was defined
as a set of anchored attributes through which the verbal
and nonverbal behaviors of individuals may be typified and
differentiated. It was further asserted that verbal and non-
verbal behaviors are probably best viewed as complementary
in supplying the informational cues which are produced and
processed by communicators. It was also argued that con-
ceptualizing communicative style in such a manner requires
that it be measured accordingly, and that the measurement
system should meet the requirements of objectivization, rela-
tivization, empirical verifiability, reliability, and precision.
In the preceding chapter, it was shown that while the set of
stylistic attributes proposed by Norton and his associates
are suitable for our purposes, the manner in which they have
been operationalized in previous research fails to fulfill
any of the requirements listed above.

The purpose of this chapter is to outline procedures
for the empirical assessment of communicative style using
the set of attributes proposed by Norton, Theoretic and
methodological criteria for a precise measurement system are

described. Procedures for developing such a measurement

38

 

39
system are also described. Finally, an attempt is made to
show how the preceding theoretic and methodological criteria

were fulfilled by this measurement system.

Theoretic Criteria

 

A system for the measurement of communicative style,
if it is to be useful in the development of a science of

'communication, must fulfill the following requirements:

1. The system must take into account the norms,
standards, concepts, and other referents which
determine or constitute the "social reality"
of the members of the social system(s) of

interest.

2. The system must consider verbal and nonverbal
behaviors as they function together systemically;

it must treat them as complementary.
3. The system must be quantifiable.

Each of these requirements is derived from a number of
philosophical, social scientific, and communication sources,
as well as the author's conceptualization of the relation
between science and the study of face-to-face human interaction.
It should prove fruitful, at this point, to discuss each

criterion separately.

 

40

Social Referents Must Be Considered

Winch's The Idea pf a Social Science, published in 1958,

 

promotes an "understanding" methodology in the tradition of
weber and Wittgenstein. Winch argues that the social
scientist must understand the "meaning" of the behavioral
data s/he gathers in order to establish their correspondence
to "social reality." According to Winch, the social scientist
achieves this understanding by describing the data in terms
of the concepts and rules which constitute the social reality
of the members of the social system s/he studies. Thus, the
description and explanation of social behavior must employ
the same conceptual and epistemic framework used by social
actors and observers themselves. This position is reiterated
by Schutz (1970), who claims that:

each term in a scientific model of human action must

be constructed in such a way that a human act performed

...by an individual actor in the way indicated by the

typical construct would be understandable for the actor

himself as well as his fellowmen in terms of common-

sense interpretations of everyday life (p. 279).

In a more recent publication, Hewes (1978) has distin-
guished between the above position, termed the "induced per-
spective" and the "theoretically imposed perspective" as
alternative approaches to the scientific study of interpersonal

communication. Hewes specifies three broadly conceived per—

spectives in the process of making the preceding distinction:

41

Social scientists hold differing, broad perspectives

which dictate the way in which they approach research.

For example, one may impose a theoretical structure

on the world to gain explanatory power...one could

induce an explanation from actors in everyday life...

or one could combine these strategies in some way

(p. 156).
The essence of debate between the imposed versus induced
theoretic camps seems to turn on the question of whether
to emphasize the subject (social actor) or theorist (researcher)
as the primary source of variables for inclusion in the
theoretic and research models. As Hewes (1978) characterizes

it, the strength of theoretically imposed perspective is

its precision and control in examining relationships among

 

phenomena and constructs imposed on the phenomena. The major
weakness of this perspective lies in its frequent sacrifice

of ecological validity (Brunswick, 1947), or generalizability

 

of findings to the "real world" of social agents. The induced
perspective, on the other hand, dictates that both the
theoretic constructs and their interrelationships must be

cast in terms comprehensible to social actors in the "language

 

of everyday life." Hence, Hewes (1978) notes that:

As a consequence, such rules (contructs and interrela-
tionships) gain comprehensibility at the cost of pre-
cision. The predictive power of a theory is a direct
function of its constructs gpd_the precision of the re-
lationships among its constructs. The language of every-
day life is not notoriously precise (p. 164).

A third alternative mentioned, but not developed by

Hewes (1978), is an approach which combines the above strategies.

42

Hewes does, however, offer the following observation:

To the extent that one can choose the formal represen—

tational system so that it does minimal violence to

subject-based intuitions, to that extent one may be

able to attain the twin goals of understanding and

prediction. The problem, of course, is how one makes

the correct choice of a methodology which leaves sub-

jects' intuitions unbloodied or at least unbowed (p. 164).
Indeed, how one goes about the business of selecting or
developing such a system is, by and large, open to discussion.
Recalling that ours is a concern for the development of an
observational framework for the precise typification of indi-
vidual communicative styles, at least two steps toward
addressing the above issue may be identified. First, it may
be useful to have members of the social system under investiga-
tion make and record observations, rather than the researcher
or theorist. In this manner, the observational data obtained
should reflect greater input on the part of that social system.
Second, by having system members consensually generate the
empirical referents (verbal and nonverbal cues) which consti-

.tute attentive, open, dominant, and other stylistic behaviors,

within the social system under study, problems of theoretical

 

imposition of meaning on the system should be reduced.

In short, to establish correspondence between our
observational framework and the domain of social phenomena
for which it is designed, the conceptual and epistemic refer-
ents of members of the social system under study should be
taken into account. An attempt was made in this study to do

so by incorporating the two steps mentioned above.

 

43
Verbal and Nonverbal Behaviors Must Be

Considered as Complementary

In reviewing research on the study of verbal and non-
verbal behavior in face-to-face human interaétion, it is
interesting to note that the majority of relevant studies
has focused on one or more isolated cues, while generally
neglecting a holistic consideration of the same. For com-
munication researchers, such a research strategy is question-
able. In fact, a number of scholars have commented to this
effect. For example, Keiser and Altman (1976) argue that
"a single nonverbal behavior does not occur in a vacuum, but
blends together with other behaviors to form complex patterns"
(p. 147). Similarly, Leathers (1976) observes that "we rarely
communicate solely by facial expression or gesture, or pos—
ture, or by our use of space and clothing, or by vocal means.
All of the subsystems of nonverbal communication function
together in a high proportion of instances" (p. 214).

In speaking of the general acceptance and use of measure-
ment systems by a consensus of scholars, Mehrabian (1972)
writes:

notation and category systems generally have failed

to relate meaningfully to communicator states, feelings,

emotions, communicator characteristics, relations among

communicators, and other communication behaviors. There-
fore, the systems are seldom used by other investigators.

In contrast, the categories that have been elaborated with

a View to their significance in the communication process

...have been far more productive in generating research
and empirical findings for the communication process (p. 6).

44
Knapp (1978) has reflected that "after many hours

of viewing visual materials, you sometimes get the uneasy
feeling that you are observing and recording minute behaviors
which may have relatively little real—life impact -- if any
at all. You wonder whether the interacting parties are cog-
nizant of fleeting movements which, only with the advantage
of videotaped replays, you are able to observe" (p. 148).

Knapp continues:

Some may question the relevance of looking at micro-
momentary facial expressions, eyebrow flashes, and
pupil dialation by asking the question: "Are such
behaviors perceived in everyday interaction?" Even

if the answer is no, this does not suggest that such
research is unimportant or even irrelevant, but it
raises the question of observational priorities far
those concerned with understanding human communication.
And, it reiterates the need to establish observational
categories which are meaningful to human interaction
(p. 148; italics mine).

 

 

 

Finally, Ellis (1977) issues a "plea for the develop-
ment of meaningful molar category systems for the study of

human interaction. However, Ellis cautions:

A diversity of behaviors which can be classified into

a single category is important and desirable if the

link is sound...the crucial issue not the molarity of

the category (per se), but the correspondence between

the category and behavioral instances...the problem

with most observational analysis in communication
research is that few, if any, researchers devote enough
time to their category system...I could find no instances
where researchers provided explicity empirical referents
for their categories of interaction (pp. 6—7; italics mine).

The above arguments suggest the need to develop an

observational framework which acknowledges the functional

45
complementarity of verbal and nonverbal behaviors in face-
to-face interaction. This thesis, in part, is aimed at
developing such a system. Further, an attempt is made to
avoid the correspondence problems noted by Ellis (1977)
by (a) having observers generate a set of empirical referents
(verbal and nonverbal) for each stylistic attribute; (b)
having the observers rely on these empirical referents in
making holistic appraisals of the communicative styles of
communicators whom they observe; and (c) statistically
assessing the correspondence between the holistic estimates

and the individual empirical referents.

The System Must Be Quantifiable
Put simply, the language of science is mathematics
(Kramer, 1970). Numbers are the simplest and the most
universal linguistic invariants. What one may estimate in
a given vernacular as "many," another may estimate as
"few." The meaning of "14," however, is relatively
unambiguous, both in its formal sense and with reference to

aggregates of objects. Because we are interested here in

 

developing a precise system for the scientific description
of the communicative styles of individuals engaged in face-
to-face interaction, such a system must, of consequence, be
quantifiable.

The simplest known quantification is counting. Most
studies of verbal and nonverbal behavior in face-to-face

settings rely either "counting" methods of assessing such

46

behaviors (e.g., frequency of eyebrow flashes, head nods,
forward leans, word usage, etc.) or "timing" methods of
assessing these behaviors (e.g., duration of eye contact
or of a speaking turn). Both counting and timing yield
information about the frequency and duration of verbal and
nonverbal behaviors, respectively. They do not, however,
exhaust the measurement possibilities. Measurement involves
the establishment of some standard or unit, hence the recog-
nition of some invariant and an agreement to speak in terms
of that invariant. Part of the problem addressed in this
thesis is the development of a standard or unit for scaling
(measuring) the molar or qualitative attributes of communica-
tive style.

imadigressfor a moment, it is possible to raise the
question of whether such molar or qualitative attributes as
attentiveness or openness can be translated into quantitative
terms. There are some students of communication who have
serious doubts, while others are sincerely at a loss to
imagine how numbers can be applied meaningfully to the
"qualities" of communication. It is probable that these same
students would, however, agree that some individuals display
more openness in communicating with others than do other
individuals. Similarly, they might even confess to having
heard and/or made the remark "s/he wasn't paying any attention
to me" (i.e. level or degree of attentiveness = O in this
instance). Interestingly, if one looks at the history of the

more "exact" sciences such as physics, one discovers that it is

47
essentially a history of quantified qualities. As Rapport
(1953) has noted:

Temperature, elasticity, viscosity, luminosity,

color, energy are all examples of quantified quali-

ties. The history of chemistry is not too different.

The backbone of chemistry is the atomic theory,

which is essentially a quantification of quality

based on the discovery that innumerable "qualitatively

different" substances can be described as combinations

in strict quantitative proportions of comparatively
few (less than 100) basic substances, the elements

(p. 158).

The point is that in a scientific approach to the study
of human communicative behavior, specifically face-to-face
interaction, precision of description is essential. The
development and use of a precise quantitative measurement
system affords such precision, provided that such a system
can be "meaningfully" applied by observers to the social
phenomena of interest. Whether this provision is met depends,
by and large, on the degree of rigor with which we specify
rules for definition and application of the quantitative

system. This brings us to methodological requirements of the

system.

Methodological Criteria

 

Methodologically, it is proposed that a system for the
measurement of communicative style in face-to-face interaction,
to be useful in the development of a science of communication,
must fulfill the requirements of objectivization, relativiza-
tion, empirical verifiability, precision, and reliability.

These criteria were discussed earlier in Chapter One, but will

48

be briefly listed and once more outlined.
ijectivization

The criterion of objectivization aims at making observa-
tions as independent of an individual observer as possible.
Stated differently, there is bound to be an element of
subjectivity in any human observation -- what one observer
sees as "dominant" behavior may not be so typified by another
observer. The criterion of objectivization attempts to
establish criteria for identifying some event or act as an
instance of some class or variable or attribute in such a way
that the criteria are minimally dependent on the position or

perspective of any single observer.
Relativization

The criterion of relativization requires that the meaning
or value of an event or act can only be determined through its
relation to some standard unit or referent. To say that a
person was "totally inattentive" or "very attentive" can only
be rendered meaningful given some standard for comparison.

If, on the other hand, we provide a standard or anchor such as
"the average level of attentiveness displayed by most people
in, say, an initial interaction," then we can assess the
attentiveness of any particular individual engaged in a similar
interaction relative to the aforementioned standard. Further,
if we arbitrarily assign this standard of "the average level

of attentiveness displayed by most people in an initial inter-

49
action" some numerical value such as 100, then we can make

the aforementioned comparison and judgment quantitatively.

 

Empirical Verifiability

The criterion of empirical verifiability demands that
empirical referents for observations and measurements be
specified. It asks the question ”Unwhat behavioral events
or phenomena does scale value x for a given attribute or
variable correspond?" In the present research, it is important
that we demonstrate empirical verifiability by determining
correspondence between observers' holistic quantitative
appraisals of the level of a given stylistic attribute dis-
played by a communicator, and the instance(s) of verbal and
nonverbal referents upon which such appraisals were supposedly

based.
Reliability

Reliability of measurement concerns "the extent to
which measurements are repeatable by different persons using
(making) the same measure of an attribute (Nunnally, 1967;
p. 172). The critical issues in reliability are agreement
among observers and the consistency with which a measurement
procedure is applied in the observation of phenomena. The
system developed here, to be scientifically useful, must ex-

hibit high reliability.

50

Precision

Precision refers here to the degree to which a scaling
procedure incorporated within some observational framework
allows observers to make perceived discriminations between
stimuli regarding to what degree each manifests some attribute
of interest, while minimizing the standard error of measurement.
In Chapter Two it was noted that the purpose of the measurement
and assumptions regarding the phenomena being measured are
important criteria for deciding how to scale the construct or
attribute of interest. In addition, it can be argued that the
scale selected for measurement should provide observers with
at least as many symbols as the number of discriminations they
are capable of making. In an information-theoretic sense, the
scale must be capable of carrying maximum information about
the phenomena being measured. The information carrying capa—

city of a scaling system is given by the following formula:

log x
H = logzx = n

where H = information content in bits and x = the number of
possible different values a scale can take. From this formula,
it can be shown that the information carrying capacity of the
typical five-point Iikert scale is 2.32 bits, which is quite
small. By comparison a five-digit Arabic number (where the

set of real numbers comprises the scale) can carry 16.61 bits
(Woelfel, 1978). Whether observers will report magnitudes or

values which maximally exploit the information carrying

51

capacity of a scaling system is an empirical question. It
should be emphasized, however, that failure a priori to
provide observers with a sufficiently precise scale is
methodologically analogous to problems of theoretical imposi-
tion of a conceptual structure on the members of a social
system discussed earlier in this chapter. In short, our
observational framework must feature precise scaling.

Having outlined a set of theoretic and methodological
criteria, it is possible to proceed to a discussion of the

procedures by which their fulfillment was attempted.

The Observational Framework

 

 

‘Overview

A series of initial interactions between strangers in
a laboratory setting were videotaped and served as the
stimuli to be scaled. Some of the interactions were randomly
selected for the purpose of training raters to use the observa-
tional framework developed here. Three male and three female
raters were given conceptual definitions of each stylistic
attribute and were subsequently asked to collaborate and
consensually generate a set of empirical referents (verbal and
nonverbal behaviors) corresponding to each. After all raters
were familiar with the conceptual and epistemic framework,
they viewed the training videotaped segments and practices
making ordinal judgments of the degree of attentiveness, open-

ness, friendliness, dominance, relaxedness, preciseness,

animatedness, contentiousness, and impression leaving manifest

52
by each interactant observed.
The next stage of training required the raters to once
again observe the training segments and to make direct

magnitude estimates of the degree to which each interactant

 

displayed attentiveness, openness, friendliness, relaxedness,
animatedness, dominance, preciseness, contentiousness, and
impression leaving given a numerical standard of 100, which
corresponded to "the average amount of (a given style attri-
bute) displayed by most persons in an initial interaction
with a stranger." This stage of training was continued until
sufficient inter-rater reliability was established (alpha =
.90 or better).

When the coding of the actual experimental data began,
the six raters were randomly assigned to one of two threesomes:
one trio rated all persons A in the dyads and the other trio
rated all persons B. During this stage, each rater would
come in individually, at a different time of the day, three
days per week. S/he would View a given videotaped interaction,
stop the VTR equipment, record his/her estimates, and then
continue until all interactants scheduled for that session had
been rated. Upon completion of a given viewing/rating session,
the rater turned-in his/her rating forms and was finished
until the next session. When all data had been collected for
all videotaped interactions, the raters were gathered together

and completely informed of the nature of the entire research.

53

Interactants

 

Approximately two-hundred undergraduates enrolled in
various introductory and advanced courses in communication
at Michigan State University were solicited for participation
in a study of "dyadic interaction and problem-solving."
Participation was strictly on a volunteer basis. As induce-
ment to participate, the potential participants were given
'the following preview of the research:

We are interested in the way people interact when

they meet someone for the first time. Further,

we are interested in determining whether having

an opportunity to interact with this person, to

become acquainted with him/her, makes a difference

in how successfully you can work together in a

subsequent problem—solving situation.

For your participation in this study, you should get

at least three things in return: (1) you will get

a chance to meet someone new; (2) you will have an

opportunity to find out more about how communication

researchers study communication; and, of course, (3)

you will receive extra-credit toward your final grade

in the communication course in which you are currently

enrolled.
Following this preview, a "sign-up" sheet was circulated and
later collected by the experimenter, who thanked the students
and their course instructor for the time to describe this
research and to request volunteers.

In response to this solicitation procedure, a total of
92 students volunteered to participate. These participants
were randomly assigned to one of 46 dyads, and were only

reassigned to another dyad if they were acquainted with the

person in the dyad to which they ahd been originally assigned

(reassignment was necessary in the case of only one dyad).

54

The purpose of this procedure was to ensure that all inter-

actants were engaged in an initial interaction with a stranger.

Videotaping the Interactions

 

On the particular evening its members were scheduled to
participate in the study, each dyad was brought into a room
equipped with video cameras, and two swivelling, "highback"
chairs which faced each other and were separated by a small
coffee table. The cameras were positioned such that a full-
screen frontal View of each interactant was obtained.

Prior to their interaction, each dyad was greeted by the
experimenter and its members were introduced. The members
were told that since they were unacquainted, and since they
would be required to work together on a problem-solving task,
they could spend the first five or ten minutes of the session
just "getting acquainted." It was further suggested that in
order to facilitate the initiation of conversation, to "break
the ice," the members might begin by discussing their views on
"changing sex roles in contemporary society." This topic was
selected since a pilot study indicated that the majority of
undergraduates in the experimental population have knowledge
of and opinions regarding it. Not only did supplying the
interactants with a topic facilitate ease in initiating
conversation, but it also served to hold communication content
constant for at least a portion of all interactions recorded
(though most dyads managed to address topics other than this

one during their conversations). It was assumed that holding

55
content constant would provide raters with a basis for dif-
ferentiating interactants according to how information was

presented, as well as what information was presented.

 

Following the experimenter's introduction and instruc-
tions, he left the room, the audio and video recording equip-
ment was activated, and the participants were left to begin.
During the interactions the experimenter waited in another
room where he could monitor the conversations and keep track
of their duration. Approximately 8-10 minutes after the
initiation of a given conversation, the experimenter returned
and informed the interactants that it was now time to begin
the problem-solving stage of the research. At this time, the
members of the dyad were led by a second experimenter (who
also was responsible for activating the audio and video record-
ing equipment in this study) to another location in the building
in which the interactions took place. Upon arrival at that
location, the participants were told that there would be no
problem-solving phase of the research, were informed of the
entire nature of the research, and asked if they objected to
having the videotaped recording of their conversation shown to
other students. No participants refused permission to use the
tapes. The second experimenter attempted to answer any questions
pertaining to the research, and upon completion, thanked the
participants and told them they could leave, but to please not
discuss the study with anyone else who was scheduled to particie
pate in the study.

This procedure was repeated five evenings per week, for

56
approximately three weeks, until all 46 dyads had been

recorded (average of 3-4 dyads per evening).

Raters

The raters were six undergraduate students (three males
and three females) enrolled for independent study credits with
the experimenter. The students were informed prior to
enrollment that the independent project would be a practicum in
the analysis of behavior in face-to-face interaction, and that
they would be required to undergo a considerable training
period, followed by several weeks of viewing and coding video-
taped interactions. In return for their efforts, the students

were offered 3 credit-hours of independent study credit.

Training the Raters

 

 

The raters were trained for about eight hours per week over
a period of six weeks. Ten dyadic interactions, randomly
selected from the original 46, were used for training and
practice purposes. At first, the raters were given the
conceptual definition of a stylistic attribute as described by
Norton (1978). Based on the abstract conceptual definition of a
given style attribute, the raters conferred and consensually
generated a common set of observable behaviors which served as
empirical referents corresponding to that attribute. Thus,
for example, attentiveness was conceptually defined as "a
tendency to listen, to show interest in what the other is saying,

and to deliberately react in such a way that the other knows

s/he is being listened to." In response to this conceptual

57

definition of attentiveness, the raters consensually generated

the following verbal and nonverbal behavioral referents:

(l)
(2)

(3)

(5)

frequency and duration of eye contact

frequency of affirmative head nods, "uh-huhs"
and other "backchanneling" behaviors

frequency with which a communicator would repeat,
rephrase, or paraphrase the other's statements
back to him/her

frequency of requests for clarification or elabora-
tion of the other interactant's previous statement

body-postural orientation (whether directed toward
or away from the other interactant).

This process was replicated for each of the nine stylistic

attributes included in the observational framework. The pur—

pose of this stage of training was twofold: (l) to allow

raters to compare information pertaining to both verbal and

nonverbal behaviors considered by each as indicators of a

given attribute; and (2) from this comparison, to generate

a common conceptual and epistemic frame of reference to be

used by raters in making and reporting their observations.

A detailed description of the conceptual definitions

and empirical indicators upon which raters' judgments were

based appears in Table 2.

58

TABLE 2
Conceptual Definitions and Empirical Indicators

of Communicative Style

 

ATTENTIVENESS

Definition - a tendency to listen, to show interest

in what the other is saying, and to deliberately react
in such a way that the other knows s/he is being
listened to.

 

Indicators - (1) frequency and duration of eye contact;
(2) frequency of affirmative head nods, "uh-huhs" and
other "backchanneling" behaviors; (3) frequency with
which a communicator would repeat, rephrase, or para—
phrase the other's statements back to him/her;

(4) frequency of requests for clarification or elabor-
ation of the other interactant's previous statement;
(5) body-postural orientation (whether directed toward
or away from the other interactant).

 

DOMINANCE

Definition - a tendency to "take charge" of the inter-
action and/or to attempt to lead or control the
behaviors of others in it.

 

Indicators - (1) frequency and duration of speaking;

(2) direction of topic(s) of conversation; (3) frequency
of interrupting behavior; (4) number of direct glances
at the other interactant.

 

OPENNESS

Definition - a tendency to reveal personal things about

the self, to easily express feelings and emotions, and
to be frank and sincere.

 

Indicators - (1) frequency of self-reference statements

of personal opinion or experience; (2) frequency of "high-
risk" self-disclosive statements; (3) frequency of
requests for other to express his/her feelings or opinions.

 

59

TABLE 2 (continued)

ANIMATEDNESS

Definition - a tendency to provide frequency eye contact,
to use facial expressions, and to gesture often.

 

Indicators - (l) vocal variation in terms of (a) pitch,
(b) loudness, and (c) range, (2) frequency of gesturing
and body movement while talking, (3) frequency of
facial expressions, (4) intensity of facial expressions,
(5) eye movements.

 

RELAXEDNESS

Definition- a tendency to be calm and collected, not
nervous under pressure, and to not show nervous

 

 

mannerisms.
Indicators - (1) leaning back in one‘s chair, (2) degree
of "steadiness" in the voice, (3) frequency of stuttering

and other verbal nonfluencies (inversely related to
relaxedness), (4) frequency of mentioning apprehension
or nervousness concerning the interaction (inversely
related to relaxedness), (5) postural rigidity.

FRIENDLINESS

Definition - a tendency to be encouraging to others, to
acknowledge others' contributions to the interaction, and
to openly express admiration and supportiveness.

 

Indicators - (1) frequency of agreement with the other's
previous statement or point of View, (2) frequency of
acknowledgment of the worth of the other's statement,

(3) frequency of smiles, (4) duration of smiles,

(5) frequency of forward leans, (6) duration of forward
leans, (7) frequency of statements reflecting interest
in the other's statements, (8) frequency of statements
of personal approval or liking for the other.

 

IMPRESSION LEAVING

Definition - a tendency to be remembered because of what
one says and/or the way one says it.

 

Indicators - frequency of unique nonverbal mannerisms
or gestures, (2) frequency of unique verbal expressions

 

61
The next step in the development of the observational
framework consisted of identifying and incorporating a scaling
procedure by which raters could quantify their observations
with maximum precision. The procedure employed here was

direct magnitude estimation (Stevens, 1956; Torgerson, 1958;

 

 

 

Hamblin, 1974; Shinn, 1974). When making direct magnitude
estimations, observers attempt to "match" a numerical
magnitude of a given attribute being displayed by some stimulus
being scaled, given some standard magnitude for comparison.
The choice of standard is important in the development
of any scaling procedure. Woelfel (1974) has argued that
"choosing as Rod S (standard or anchor) some ordinary language
symbol whose relation to other such symbols is stable over
time might make the results of the measurement more clearly
interpretable in terms of the ordinary language system than
would a Rod S defined by a symbol whose meaning fluctuates
in the vernacular" (p. 4). With specific reference to direct
magnitude estimation, Hamblin (1974) recommends the following
guidelines for the choice of standard and use of the scaling
procedure:
1. Use a standard whose level does not impress the
observer as being extreme (i.e., use a standard in

the middle of the stimulus range.

2. Present variable stimuli that are both above and
below the standard.

3. Call the standard by a number which is easily
multiplied and divided.

4. Assign a number to the standard only, and leave the
observer completely free to decide what he will call
the variable. If the experimenter assigns numbers

62
to more than one stimulus, he introduces constraints
of the sort that forces the observer to make
categorical rather than magnitude judgments.

5. Use only one standard in any given experiment.

6. Randomize the order of presentation of stimuli to
be scaled.

7. Let the observer present the stimuli to himself/
herself. S/he can then work at his/her own pace
and will be more apt to be attending properly when
the stimulus comes on. (pp. 64—65)

The above recommendations were incorporated in the present
research. Special attention, however, should now be devoted
to the specific standard chosen for this study.

The choice of scaling standard used by raters in this

study was largely an outgrowth of the author's reading of

Mead (1934), Sherif (1936), Sherif and Sherif (1964), Jones
and Davis (1965), Labov (1968), and Miller and Steinberg
(1975). Mead coined the term "generalized other" to refer
to the general belief, attitudinal, and behavioral patterns
of "most persons" in a given social system, as seen from the
perspective of an individual social actor. The individual,
according to Mead, refers to this "generalized other" in
responding to the conduct of others, as well as in attempting
to align his/her action or conduct with others in the social
system. One interpretation of the generalized other is that
it serves as a generally stable referent in relation to
which individuals align and interpret their own and others'
social conduct.

Similarly, Sherif and Sherif (1964) suggest that members

63
of a social system arrive at common definitions of what is
acceptable and unacceptable with regard to the social prac—
tices of members. Sherif and Sherif refer to these standard—
ized practices and evaluations as norms, and suggest that
"the standardized practices and evaluations which are called
norms need not be identical with typical or statistically
average behavior of a group (though they may be). They often
embody conceptions of expected behavior, of the goals or ideals
of a 'good member'" (p. 166). Further, Sherif and Sherif

suggest that the norms of the social system are the standards

 

by_which behavior is appraised (p. 166). It is interesting to

 

 

note that Iabov (1968) makes a similar argument vis a vis
language behavior and social evaluation in speech communities.
In terms of actual perceptual patterning and information
processing, Miller and Steinberg (1975) identify two basic
processes which persons use in evaluating and making pre-

dictions or assessments about others. Stimulus generalization

 

refers to the process of reacting to others” behaviors of
others one has encountered in similar situations, over time.

Stimulus discrimination, on the other hand, is a process by

 

which one reacts to and/or evaluates another's behavior based
on how it differs from those of others one has encountered

in similar situations, over time. Impressions formed and
evaluations of observed others are said to be a function of
the degree to which such persons are generalized or discrim-

inated in relation to socially shared norms, classes, or

stereotypes.

64

Each of the preceding conceptualizations shares the
nation that members of a given social system assess the social
conduct of individuals via comparison with others, the latter
often providing standards or exemplars. Perhaps the best
example of the incorporation of this notion into social
theory is the theory of correspondent inferences developed
by Jones and Davis (1965). The theory of correspondent
inferences attempts to explain a perceiver's inferences about
the intentions or dispositions behind another‘s actions. It
is assumed that the observer or perceiver makes initial deci-
sions concerning (a) whether another can forsee the social,
physical, or economic consequences of his/her actions and (b)
whether another has the ability to produce these and/or other
consequences. Assuming both a and b are affirmed, the perceiver
is then said to infer intentions behind another's actions
which, in turn, are used to infer stable personal attributes
or "dispositions" on the part of him/her. Formally stated,
"given an attribute—effect linkage which is offered to explain
why an act occurred, correspondence increases as the judged

value of the attribute departs from the judge's conception of

 

 

 

the average person's standing on_that attribute" (Jones &

 

 

Davis, 1965; 224, italics mine). In other words, the
attribution of a trait or disposition on the basis of a given
action is determined by the degree to which the action departs
from normative expectations.

Following from the above lines of thought, it is possible

to conceive of a sort of extant system of social differentiation

65
wherein the communicative conduct of "most persons" serves
as a standard or norm in relation to which one's own and
others' communicative conduct is appraised and/or evaluated.
Beyond the sources previously discussed, the use of such
normative standards in aligning and evaluating communicative
behavior is consistent with Stokes and Hewitt's (1976)
discussion of the connection between culture and personal
conduct, as well as Snyder's (1974) conceptualization of the
"selfrmonitoring" individual. In terms of communicative style,
a certain level of attentiveness, relaxedness, openness,
dominance, etc., is probably normatively expected in initial
interactions. "Most persons" display such levels, at least
within a certain range, and individuals who deviate from
these levels (i.e., who are unusually open or excessively
inattentive, etc.) may be readily differentiated (i.e., as-
signed a value for the style attribute in question) via
reference to the normative standard. In effect, we have the
potential for an observational framework in which the measure-
ment standard is solidly lodged in the social system of
interest. Given that this standard exists, and that persons
are at least qualitatively differentiated according to it,
it is neither unreasonable nor necessarily difficult to
transform this system of "measurement" from a qualitative
to a quantitative one. Provided that the requirements of
reliability, precision, and empirical verifiability can be
satisfied, the system would thus be rendered useful for

scientific inquiry.

66
Such was the rationale underlying the choice of standard
and method of measurement employed in this study. Raters
were instructed to interpret a numerical standard of 100

has equal to "the average level of (a given

 

style attribute) displayed by most people in an initial
interaction situation." Thus, for each interactant observed,
magnitude estimates were made by having raters compare his/
her behavior with the above standard for each of the nine
stylistic attributes. If an interactant's behavior was seen
as average for a given style attribute, s/he was assigned a
value of 100 for that attribute. If the behavior was seen as
exceeding the average, it was assigned an appropriate non-
negative integer greater than 100. If the behavior was seen
as below the average, it was assigned an appropriate non-
negative integer less than 100.

Using the ten sample dyads, the raters practiced making
magnitude estimates, referring the above standard, for about
four weeks. During the practice sessions, an interactant
was observed, raters made independent judgments, and then
compared their results. This was done in order to facilitate
learning of comparable meanings of individual magnitude
estimates among all raters, as well as to establish acceptable
inter-rater reliability.

When the coding of the actual experimental data (36
videotaped dyads) began, the six raters were randomly

assigned to one of two threesomes: one trio rated all persons

A in the dyads, while the other trio rated all persons B.

67
This was done in order to circumvent problems of cross-
contamination and dependence encountered when the same trio
rated both interactants in the same dyad.

Coding of the 72 interactants took approximately four
weeks. The standard procedure for coding was as follows:
Each rater came in individually, at a different time of the
day, three days per week. The rater viewed a given videotaped
interaction, stopped the VTR equipment, recorded his/her
estimates, and continued to the next interactant. The order
in which style attributes were estimated was randomized for
each interactant to avoid possible order effects. Similarly,
the order in which interactants were presented to raters
was also randomized. This procedure was followed for all
interactants scheduled to be viewed on a given day. Upon
completion of a given viewing/coding session, the rater
reported to the experimenter's office, turned in his/her
rating forms, and was finished for the day. ‘When all inter-
actants had been coded, the raters were gathered together

and the nature of the entire project was revealed.

Summary and Conclusion

 

The purpose of this chapter was to outline procedures
for the empirical assessment of communicative style in face-
to-face settings, using the set of attributes proposed by
Norton (1978). Three theoretic criteria were proposed:
(1) that the system must take into account the norms, standards,

concepts, and other referents which determine or constitute

68
the "social reality" of the members of the social system
of interest; (2) that the system must consider verbal and
nonverbal behaviors as complementary; and (3) that the
system must be quantifiable. Methodologically, it was
proposed that the measurement system must meet five require-
ments: (1) objectivization; (2) relativization; (3) reli-
ability; (4) precision; and (5) empirical verifiability.

The system proposed herein seems to meet the theoretic.
requirements outlined above. Considerable effort is expended
:to solicit and incorporate the input of members of the social
system under study, particularly by having these persons
specify the verbal and nonverbal referents used in assessing
displayed levels of the various stylistic attributes, and by
using a standard of measurement which theoretically is
derived from, rather than imposed upon, the observers. Having
observers appraise molar attributes like openness, attentive-
ness, dominance, and the like requires them to consider
individual verbal and nonverbal cues as they cluster or
function systematically, thus fulfilling the second require-
ment. Finally, by having observers use the direct magnitude
estimation technique, the measurement system is thus quantified.

Methodologically, it is possible to argue that both
the criterion of objectivization and the criterion of relativi-
zation were satisfied. Measurement is stimulus-centered and
is based on an epistemic framework which is the product of
a consensus of observers, rather than any single observer's

perspective or vantage point. Also, by definition, the direct

69
magnitude estimation procedure requires the inclusion of a
standard in relation to which other stimuli are compared and
assigned values for a given attribute. Thus, the criterion
of relativization is satisfied.

As for the criteria of reliability, precision, and
empirical verifiability, determining whether they are satis—
fied in the approach developed here requires that the data
gathered by raters (which was described earlier) be subjected
to a series of statistical analyses. Details concerning
these analyses, as well as the results pertaining to the
reliability, precision, and empirical verifiability of the
observational framework developed here, are presented in

Chapter Four.

70

FOOTNOTES

1 Post-interaction interviews with interactants
indicated that none were particularly apprehensive about
or inhibited by the presence of the video equipment. In
fact, several interactants cited previous classroom and/or
research experience as a basis for their feelings during
the interaction sessions. Thus, unless the interactants
were lying, the effect of knowing that they were being
videotaped should have had minimal impact on their behavior.

CHAPTER FOUR
THE MEASUREMENT OF COMMUNICATIVE STYLE: RELIABILITY,

PRECISION, AND EMPIRICAL VERIFIABILITY

In Chapter Three, theoretic and methodological criteria
for the development and evaluation of an observational
framework for the empirical assessment of communicative style
were presented. Of the five methodological criteria, it was
proposed that reliability, precision, and empirical verifia-
bility must be satisfied based on the collection and analy-
sis of relevant data. The critical issues related to preci-
sion and reliability are threefold: (1) Are multiple indica—
tors of the same latent variable consistent?; (2) Is there
sufficient inter-observer agreement, such that error of
measurement is minimized?: and (3) Does the scaling procedure
incorporated within the framework allow observers to report

perceived differences between stimuli which are being scaled

 

with regard to some attribute? The critical issue in demon-
strating empirical verifiability centers on whether behavioral
referents for reported values of a molar attribute can be
specified, measured, and shown to co-vary with those reported
values.

This chapter presents the results of a series of data
analyses pertaining to the reliability, precision and em-
pirical verifiability of the measurements obtained from raters

in the research described in Chapter Three.

71

72

Reliability and Precision

 

 

Estimates of inter-observer consistency and agreement
were obtained by computing the intraclass correlation co-

efficientl

for each of the stylistic attributes. This
technique is appropriate since it provides an estimate of
both reliable variance in observers' ratings, while at the
same time accounting for inter-observer agreement by in-
cluding the "between raters" variance in the error term in
the analysis of variance from which the coefficient is
ultimately derived.

The results of this analysis are summarized in Table
3. The results indicate generally high reliability and
agreement among observer estimates, with coefficients rang-
ing from .714 (friendly) to .949 (dominant). Since the use
of a scaling procedure affording a theoretically infinite
number of symbols by which observers could report their ob-
servations correspondingly increases the possibility of
obtaining greater “between rater" variance, these results
can be regarded as especially encouraging; they suggest that
with sufficient practice, observers can reliably estimate
magnitudes of stylistic attributes using a relatively complex
scale.

Regarding the general discriminatory power or precision
of the scaling procedures incorporated within this observa-
tional framework, the data summarized in Tables 4 and 5

are illustrative. Table 4 provides summary statistics for

73

TABLE 3

Reliability Estimates for All Stylistic Attributes

 

 

Attribute Intraclass Coefficient
Impression Leaving .835
Open .891
Attentive .914
Animated .867
Relaxed .923
Friendly _ .714
Precise .865
Dominant .949

Contentious .753

 

74

all stylistic attributes: the overall mean rating, mini-
mum mean rating, maximum mean rating, range and standard
error of measurement for eacH of the attributes are pre-
sented. Initially, it is apparent that the scaling proce-
dure-permitted raters to report a wider range of perceived
differences between stimulus interactants, for each of the
stylistic attributes, than would have been possible using
less precise scaling methods (e.g., five-point, Likert-type
scales). In this sense, the procedure demonstrates greater
precision in terms of discriminatory power.

A second means of evaluating precision pertains to
the information potential of the procedures. Recalling the
formula presented in Chapter Three for computing the in—
formation-carrying capacity of a scaling system, and apply-
ing it to the actual range of scale values for each stylistic
attribute, it is possible to obtain an estimate of the in-
formation-carrying capacity of the magnitude estimates for
each attribute. Table 5 summarizes these estimates. The
results indicate that, even for the lowest estimate (6.64 bits
for animated) the information-carrying capacity of the mag-
nitude estimation procedure was almost three times as great
as that of a five-point, Likert-type scale. Further, the
estimates summarized in Table 5 can be regarded as conserva-
tive since (a) they are based on the range of mean scale
values, rather than on the range reported by any single ob-
server, and (b) they are based only on values reported in

this research; the potential still exists for a wider range

75

TABLE 4

Selected Summary Statistics for Stylistic Attributes

 

 

Attribute Mean Min. Max. Range Gmeas
Impression- 108.95 83.33 190.00 106.67 4.021
Leaving
Open 118.38 76.67 162.33 85.66 3.581
Attentive 115.09 77.00 173.33 96.33 2.700
Animated 111.24 95.67 165.33 69.66 4.365
Relaxed 113.90 84.67 179.67 95.00 3.967
Friendly 111.50 96.67 170.00 73.33 3.322
Precise 111.61 50.00 142.67 92.67 2.981
Dominant 106.77 57.67 168.33 110.66 3.220
Contentious 96.27 40.67 122.67 82.00 2.581

 

76

TABLE 5

Estimated Information-Carrying Capacity of Scale

Values Based on Actual Ranges

 

 

Attribute H
Impression Leaving 7 , 7.31
Open 6.97
Attentive 7.15
Animated 6.64
Relaxed 7.13
Friendly 6.72
Precise 7.10
Dominant 7.37
Contentious 6.90

 

i:
H designates information—carrying capacity in bits.

77

of values to be reported by alternative groups of observers
in alternative research settings.

The preceding data seem to indicate that the procedures
demonstrate reliability and precision, well within conven-
tional ranges of acceptability. Let us now turn our atten-

tion to the issue of empirical verifiability.

Empirical Verifiability

 

The critical issue in demonstrating empirical verifi-
ability centers on whether independently obtained measure-
ments of the empirical or behavioral referents generated
by the obServers in this Study can be shown to co-vary with
the direct magnitude estimates the observers reported.

It was not possible to obtain independent measures of
each set of indicators for each stylistic attribute. Neither
human nor financial resources were available for this purpose.
Therefore, the decision was made to take one of the attri-
butes, dominance, and make as complete a-check of its em-
pirical verifiability as possible. Since all stylistic
attributes were scaled similarly, and behavioral referents
were generated by the same observers for each attribute,
should the appraisals of dominance and the individual be-
havioral referents be substantially correlated, then we may
tentatively place confidence in the empirical verifiability

of the measurement system.

78
Procedures

Coders

Three undergraduate students enrolled in a course in
nonverbal communication at Michigan State University volun-
teered to observe the sample of interactants and to code
each stimulus interactant with regard to displayed verbal
and nonverbal cues reflective of dominant behavior. Coding
sessions took place twice a week for the full nine weeks of
the quarter: the first two sessions‘were training sessions
in which the individual behaviors were defined, and measure-
ment procedures outlined and practiced; the remaining ses-

sions, which normally lasted from two and one-half to three

hours each, were devoted to actual coding.

Behavioral Referents

 

The behavioral indicators of dominance were as follows:

(1) total number pf speaking turns (where a speaking turn was

 

 

defined as a period of talk by one interactant accompanied
by the other interactant's assumption of the "listener's role");

(2) direction of topic 9: conversation (frequency with which a

 

 

change in the tOpic of conversation was initiated by the

interactant being coded); (3) frequency of interrupting PET

 

 

havior; (4) duration 9f direct glances at the conversational

 

 

partner; (5) total duration 9: speaking. The first three in-

 

dicators were measured by having coders observe and record
each occurrence of the behavior on a "check-list"; with the

occurrence of each perceived instance of a behavior, the coder

 

79

made a check-mark on the line(s) adjacent to that behavior.
The fourth and fifth indicators were monitored by having
coders use a stop watch.

Five separate passes through the 30 videotaped inter-
actions were required in order to concentrate on a single
behavior at a time. As was the case with raters who made
direct magnitude estimates of style, the coders of indivi-
dual behavioral cues saw only one interactant (full-screen,

frontal view) in a given dyad.

Results

Table 6 summarizes the results of the tests of the
empirical verifiability of dominance. With the exception
of duration of direct glances, all other correlations are
both significant and substantial (.51 to .83). Note also
that the intraclass correlation coefficients (i.e., re-
liability estimates) are generally high (.67 to .92). In
general, it appears that the relationship between the
direct magnitude estimates and the individual indicators
is substantial, which suggests that, at least with regard
to dominance, the measurement system is empirically verifi-
able. Though no test of empirical verifiability was made
for any other attribute, for reasons mentioned earlier in
this chapter, we may tentatively suggest that it is reason-
able (albeit empirically questionable) to expect similar
findings regarding these attributes. More will be said about

this in Chapter Six.

80
TABLE 6

Summary of Results Pertaining to Empirical

Verifiability of Dominance

 

 

Behavior 5* £**
Speaking Turns .65 .85
Direction of Topic .51 .67
Interruptions .83 .72
Duration of Glances .17a .80
Duration of Speaking .79 .92

 

Unless indicated by the subscript a, all Pearson
correlations are significant at or beyond the .05
level of confidence with 70 degrees of freedom
(two-tailed).

'3':
Pearson product—moment correlation coeff1c1ent

** . . .
Intra-class correlation coeff1c1ent

81

The purpose of this chapter was to describe a series
of analyses pertaining to the reliability, precision, and
empirical verifiability of the measurement approach de-
veloped in this research. Based on these results, there
is evidence that the system is precise, reliable, and cate-
gorically empirically verifiable.

The data presented thus far augur well for the scien-
tific utility of this measurement system in the study of
communicative style in face-to-face interaction. The next
logical step consists of actually applying the procedures
in a research setting.

Chapter Five addresses a specific question to which
the measurement approach developed here was applied. A
study of the relation between interactive style and social
perceptions of interpersonal attractiveness and communica-
tive effectiveness is described. Some relevant literature
on communicative effectiveness and competence is reviewed.
Subsequently, an experiment designed to investigate the
relation between style and competence is outlined. Finally,

the results are presented and discussed.

82

FOOTNOTES

1 This approach was developed by Ebel, R. L. Estima-
tion of the reliability of ratings. Psychometrika, 1951,
16, 407—424.

 

CHAPTER FIVE

A RESEARCH APPLICATION

Berger (1977) has characterized most empirical studies
of communicative or interpersonal style as follows:

The general thrust of research on communicator

style is to isolate various facets of self-pre-

sentation which influence such factors as at-

tractiveness and effectiveness in interaction

(pp. 219-220).

Among the commonly cited bases of interpersonal attrac—
tion are homophily, physical appearance, spatial propinquity,
and behavioral reciprocity (Byrne, 1971). In addition to '
these, it seems reasonable to expect that the manner in
which persons communicatively present themselves and/or
respond to the communicative self-presentation of others
should have a significant impact on attributions made about
their interpersonal attractiveness and social competence.
That is, reactions to relative strangers may largely be de-
termined by their communicative or interpersonal style.
Indeed, Berger's characterization of research on style above
suggests that this hunch has guided researchers to a great
degree.

This chapter presents a study of the relation between
communicative style and perceived social competence to which

the measurement model developed here was applied. Specifically,

83

84

a study of the relations between style and social perceptions
of interpersonal attractiveness and communicative effective-
ness is described. Some relevant literature on communica-i
tive effectiveness and competence is reviewed. An opera-
tionalization of perceived competence is offered. Sub—
sequently, an experiment designed to investigate the relation
between style and.competence is outlined. Finally, the

results are presented and interpreted.

Communicative Competence: Characteristics of the

 

 

Effective and Attractive Interactant

An area of research which has received increased atten-

tion from communication scholars recently pertains to com-

municative competence. Wiemann (1977) defines communicative

 

competence as:

the ability of an interactant to choose among

available communicative behaviors in order that

he may successfully accomplish his own inter-

personal goals during an encounter while main-

taining the face and line of his fellow interac-
tants, within the constraints of the situation

(p. 198).

In justifying the importance of inquiry into the ef-
ficacy of communicative behavior, proponents of the com—
municative competence position maintain that in order for
persons to achieve personal and social goals through com-
munication, they must learn to enact behavioral routines

which are deemed appropriate to the particular individuals

and social situations defining the interaction.

85

While virtually all scholars who have written about
communicative competence adopt a performance-based per-
spective (i.e.,_emphasis is placed on the necessity of
both the knowledge of efficacious and appropriate modes
of interaction and the ability to perform the same), dif-
ferent writers seem to emphasize different features of
competence.

Some writers focus on the ability of a social actor
to achieve his/her goals through communication. For instance,
Parks (1976) conceives of the competent communicator as "one
who maximizes his or her goal-achievement through communica—
tion" (p. 1). Parks continues to suggest that such a
formulation focuses on the instrumental function of indivi-
dual communicative conduct, by choice, because of the purpo-
sive nature of communication. This perspective is con—
sistent With the view of Miller and Steinberg (1975) that
"the basic function of communication is to control the
environment so as to realize certain physical, economic, and
social rewards from it" (p. 62). Such a perspective of
communication and communicative competence does not neces-
sarily exclude or ignore the importance of others' evalua—
tions of and impressions concerning an individual's goal-
seeking social conduct. In its strictest interpretation,
however, such a perspective does not require that others'
perceptions or evaluations be taken into account in assessing

an individual's communicative competence.

86

Another perspective of communicative competence focuses
on the ability of a social actor to manifest coordinative
communication skills. Cushman and Craig (1976), for instance,
have suggested that competence rests on the degree to which
an individual develops and is able to perform three classes

of communication skills: listening, cueing, and negotiation:

 

 

Listening skills turn on our ability to recognize
differing types of statements and the respective
self-object relationships which they designate.
Cueing skills turn on our ability to translate our
own relationships to objects into the vocabularies
of diverse others. Negotiation skills turn on our
ability to recognize the positions of others, to
cue others as to our position, and to develop the
appropriate strategies for reconciling differences
in our expectations toward situations (p. 55).

A third perspective, which focuses on goal-directedness,
coordinative skills, and others' evaluations of an actor's
communicative conduct is reflected by writers like Bochner
and Kelly (1974), Pearce (1976), and Wiemann (1977).

Bochner and Kelly (1974), for example, conceptualize communi-
cative competence as follows:

competence can be judged by the following criteria:

(1) ability to formulate and achieve objectives;

(2) ability to collaborate effectively with others;

i.e., to be interdependent; and (3) ability to adapt

appropriately to situational and environmental varia-

tions (p. 288; italics in the original).
The emphasis of Bochner and Kelly on the "ability to colla-
borate effectively with others" is closely paralleled by

Wiemann's (1977) definitional emphasis on "maintaining the

face and line of fellow interactants" (p. 198) in that both

87
stress an "other-orientedness" dimension of communicative
competence. Taken together, the conceptualizations of
Bochner and Kelly (1974) and Wiemann (1977) suggest a
perspective of competence in which the importance of goal—
achievement, interactive skills, 229 sensitivity to
both situations and other persons are collectively stressed.
Such a perspective highlights the social rather than indi-
vidual nature of interpersonal communication; it suggests
that goal-attainment in the absence of an awareness of inter-
personal interdependence and social evaluation cannot describe
a "competent" communicator from other than a socially myopic
perspective. This perspective also enjoins researchers to
consider both individual interactive conduct and social reactions
to such conduct in order to best understand and explain inter-
personal competence. This perspective is also embodied in the
present thesis.

While the previous view of interpersonal competence may
be conceptually suggestive, few data are available which allow
an assessment of its empirical utility, particularly with
regard to observer's reactions to or evaluations of an actor's
communication behavior. Cne exception is the work of Wiemann
(1977) who, after providing a comprehensive overview of perti-
nent research, argued that individual verbal and nonverbal

behaviors function to create impressions of affiliation/support,

 

social relaxation, empathy, behavioral flexibility, and inter-

 

 

action management skills. These, in turn, lead to judgments of

 

relative competence. Wiemann (1977) had observers watch

88

videotapes of individuals who displayed varying amounts of these
five classes of behavior, and then had the observers rate the
individuals in terms of their communicative competence. The
results generally supported his argument, with communicators
who displayed increasing amounts of affiliation/support,
social relaxation, empathy, behavioral flexibility, and inter-
action management skills being perceived as increasing competent.
While the findings of Wiemann (1977) are enlightening, it
is possible to question whether the five classes of behaviors
examined are sufficiently precise or inclusive to allow ob-
servers to discriminate between individuals and their respective

styles. If, as Wiemann and Knapp (1975) have asserted, we
usually make judgments about people based on the way they inter-
act" (p. 75), then it may be necessary to focus more clearly
on the relation between communicative style and social eval~
uations of interpersonal competence.

The research question addressed in the study reported in
this chapter may be stated as follows:

Are there identifiable patterns of relations between
communicative style and interpersonal competence?

Method

Competence Operationalized

 

Rather than employing a competence subscale, as did
Wiemann (1977), the social evaluative component of competence
was operationalized in terms of three variables: (1) communi-

cative effectiveness; (2) social attractiveness; and (3) task

 

 

 

89

attractiveness. Since research by McCroskey, Hamilton, and

 

Weiner (1974) indicates that interpersonal attractiveness is
not unidimensional, scales developed by those researchers

for the measurement of social and task attractiveness were
used in this research. Further, the three most reliable items
from Norton's (1978) communicator image subscale and Wiemann's
(1977) competence subscale were employed to measure communica-

tive effectiveness.

Observers

 

Observers, from whom measures of perceived communicative
effectiveness, social attractiveness, and task attractiveness
were obtained, consisted of 180 undergraduates, enrolled in
various introductory and advanced courses in communication,
who volunteered to participate in a study of "perceptions

of behavior in dyadic interactions."

Procedure

 

The observers were divided into two groups (90 persons
per group), one of which viewed all persons A in the dyadic
interactions recorded earlier (see Chapter Three), and the
other of which viewed all persons B. Further, members of
each group were randomly assigned to one of three subgroups.
The purpose of such assignment was to pair the mean attrac-
tiveness ratings and effectiveness ratings from each subgroup
with the style estimates of one of the raters (through ran-
domization) for all interactants observed.1 This procedure

yielded 216 such pairs for data analysis.

90

The basic procedure consisted of presenting the
videotaped interactions to the observers, allowing them to
observe each for approximately five minutes, and then having
the observers evaluate the interactant in terms of the three
communicative effectiveness items, three social attractiveness
items, and three task attractiveness items.

Data were collected over a period of three evenings
(12 interactants observed per evening) for one week. Separate
viewing rooms were used for each group of observers. Upon
completion of this stage of data collection, observers were
completely informed of the nature of the research, were
thanked for their participation, and told they could leave.
A11 observers received credit for participating in a communi-
cation research project which was applied to the final grade
received by each in the particular course in which s/he was
enrolled. This was done in accordance with departmental and

university policy concerning research involving human subjects.
a

Results

Reliability of Competence Measures

 

 

Alpha analysis was used to determine internal reliability
of the items measuring perceived communicative effectiveness,
social attractiveness, and task attractiveness. Coefficients
of .976, .906, and .969 were obtained for social attractive-
ness, task attractiveness, and communicative effectiveness,
respectively. The high reliability of observer ratings was

expected as a function of (a) the large number of observers

91

v5.

HH

.Awam n mpv mocwpflmcoo mo Hm>mH mo. may
pso>mn no um uCGOHMHcmHm mum mcoflumamwuoo Ham Axv Mmflumpmo as %Q pmpmoflpsfl mmchD

am. am.: mm. mm. >4. mm. as. am. me. am. pommmm m>flumOHcsssoo
me. wao.u ea. Hm. ma. mm. mm. as. mm. mm. mmmcm>nuomunu¢ xmme
«40.: mm. mm. Hm. ov. as. mm. mm. om. m>fluomupu¢ Hmfloom
om. m4. 44o.n 4mo.a 4mo.u ma.n ea.u om.u msoflucmucoo
em. me. «no.1 mm. ma. as. mm. ucmcflsoo
ma. mm. mm. we. mm. as. mmfiomnm
mm. me. am. H4. H4. saccmnnm
am. me. mm. mm. emxmamm
mm. mm. as. emumeflaa
om. mm. m>auamupm
He. ammo
msfl>wmq :OfimmmumEH
OH m m a m m a m N H

mmmcm>flpomuup< paw .mmmcm>fluommwm .mawpm ms084 mcoflpmHmeOOQmucH

b mqmdfi

Away
AHHV
Aoav
Amv
Awe
Anv
Amv
“my
Aev
Amy
Amy
Add

92
used, (b) the fact that means from subgroups were entered
into the analysis rather than individual observer scores,
since our interest was in determining social (i.e., aggregate)
rather than individual perceptions of effectiveness and

attractiveness.

Intercorrelations Among_Style, Effectiveness, and

 

Attractiveness Variables

 

Table 7 displays the zero-order correlations among all
variables examined in this study. As is apparent, nearly
all variables are significantly correlated. Since statistical
significance is largely contingent on sample size, as well as
the "true" relationships between variables (Hayes, 1973),
and since the sample of observations in this study was somewhat
large (N = 216), it may be of less value to focus on the
statistical significance of these correlations, and instead
turn our attention to their interpretability xii a_zi§ magnitude
and comparability to previous studies. More will be said about
this in the final chapter of this thesis. For now, it may be
useful to note the following: (1) as was found by Norton
(1978), impression leaving correlated strongly with communi-
cative effectiveness, social attractiveness and task attract-
iveness (.81, .80, and .73, respectively), the latter variables
being roughly analogous to the "communicator image" component
of Norton's research; (2) impression leaving, openness,
attentiveness, animatedness, and relaxedness were at least
moderately (.35) or more strongly (.49 to .81) correlated

with all effectiveness and attractiveness variables; and

93

(3) communicative effectiveness, social attractiveness, and
task attractiveness, which were conceptualized as comple-
mentary components or reflectors of interpersonal competence,
were in fact highly intercorrelated. In addition, while it
was not strongly related to task attractiveness, friendliness
is moderately correlated with social attractiveness and

communicative effectiveness (.51 and .47, respectively).

Canonical Correlation Analysis

 

The purpose of this study was to identify relationships,
if any, between a set of communicative style variables and a
set of interpersonal competence variables. The method of
canonical analysis was selected because it permits us to assess
whether linear combinations of certain style variables are
significantly correlated with linear combinations of communicative
effectiveness and attractiveness variables (Van de Geer, 1971;
Harris, 1975; Finn, 1974). Table 8 summarizes the results of
the canonical analysis. The results indicate two significant
canonical correlations between the set of communicative style
variables and the set of effectiveness and attractiveness vari-
ables. Both canonical variate coefficients (i.e., the coeffi—
cients in the linear regression equations which generates the
canonical variate from the original variables) and the canonical
loadings (i.e., standardized coefficients expressing the cor—
relations between the original variables and the canonical
variate) are reported.2 For the first canonical variate, the

canonical correlation (RC) between the first and second set of

94
variables is .898 (x2 = 604.25; df = 27; p< .001).
An examination of the canonical loadings indicates that, in
the communicative style set (first set) impression leaving,
open, animated, and relaxed are strongly related, while domi—
nant, attentive, and friendly are moderately related to the
first canonical variate. From the second variable set,
communicative effectiveness, social attractiveness, and task
attractiveness are all strongly related to the first canonical
variate.

For the second canonical variate, the canonical correlation
between the first and second set of variables is .768
(x2 = 258.38; df = 16; p <.001). From the communicative style
set, attentiveness, friendliness, and preciseness are strongly
related, and relaxedness is moderately related with the second
canonical variate. No other significant canonical correlations
were found among these data.

The results seem to indicate that there are at least two
subsets of the communicative style variables which are highly
and linearly related to two corresponding subsets of the communi-
cative effectiveness and attractiveness variables (interpersonal
competence set). For purposes of interpretation, canonical
loadings will be utilized in accordance with the recommendations
of Thorndike and Weiss (1973) and Tucker and Chase (1976).

For purposes of exposition, let us refer to the subsets of the
first variable set as "communicative styles" in their own right,
and to the subsets of the second variable set as types or

dimensions of perceived interpersonal competence. Accordingly,

95

TABLE 8

Canonical Analysis Summary Table

 

Original First Canonical Second Canonical
Variable Variate Variate

I II I II
FIRST SET
Impression Leaving 1.07 .933 .06 ‘ .033
Open .18 .752 .06 .206
Attentive - .56 .662 .31 .627
Animated .45 .783 - .79 - .137
Relaxed - .27 .792 - .54 .485
Friendly .15 .520 .69 .518
Precise .20 .415 .08 .633
Dominant - .32 .522 .34 .210
Contentious .11 .087 .42 - .292
SECOND SET
Social Attractiveness .66 .980 - .29 .017
Task Attractiveness .15 .816 -1.39 .576
Communicative Effective- .25 .935 1.52 .239
ness

 

I = Canonical Variate Coefficients
II = Canonical Component Loadings

First Canonical Variate: RC = .898; Eigenvalue = .808;

2

x 604.25; df = 27; p < .001

Second Canonical Variate: RC .768; Eigenvalue = .591;

X2

258.38; df = 16; p < .001

96

the results pertaining to the first canonical variate suggest
that an impression leaving, open, attentive, animated, relaxed,
and to a lesser extent, friendly and dominant style will elicit
perceptions of both social and task attractiveness (i.e., persons
will perceive the interactant as potentially likeable and easy
to work with), and perceived communicative effectiveness (i.e.,
the interactant's style will be seen as skilled and pro-social).
In terms of the second canonical variate, the results suggest
that an attentive, friendly, precise, and to a lesser extent,
relaxed communicative style will elicit perceptions of task
attractiveness (i.e., the interactant will be seen as a potentially
desirable working partner or associate).

The results obtained here are generally consistent with
those obtained by Norton (1978), Norton and Pettegrew (1976),
and Wiemann (1977). In the next and concluding chapter of this
thesis, these findings will be considered in detail. Specifically,
Chapter Six provides a discussion and summary of all the work
presented in this thesis. Individual sections are devoted to
(1) the available evidence pertaining to the utility of the
measurement system developed here, (2) the results pertaining
to the relation between communicative style and perceived inter-
personal competence, (3) the implications of the findings for
future communication inquiry, and (4) problems and limitations
of the research. The final chapter concludes with a summary

of the entire thesis.

97

FOOTNOTES

1
See Chapter Three for a detailed discussion of
the manner in which ratings of the videotaped interactants'
communicative styles were obtained.

2 While the canonical loadings are generally more
readily interpretable, and while they are informative as
to the degree to which the original variables are core
related with the canonical variates, the canonical variate
coefficients are informative with regard to the relative
weight of each of the original variables in the linear
regression equation which generates the canonical variate(s).
Thus, both items of information are presented for the reader's
consideration.

CHAPTER SIX

DISCUSSION AND CONCLUSION

The purpose of this chapter is to review and critique
the work presented in this thesis. Specific sections are
devoted each to (l) the available evidence pertaining to
the utility of the measurement system developed here, (2)
the results pertaining to the relation between interpersonal
competence and communicative style, (3) the implications of
the findings for future communication inquiry, and (4)
problems and limitations of this study. The chapter con—

cludes with a summary of the entire thesis.

Scientific Utility

 

In Chapter Three, a set of theoretic and methodological
criteria were specified by which the scientific utility of
a measurement system developed for the study of face-to-face
interaction, in general, and communicative style, in parti-
cular, might be assessed. The theoretic criteria are:
(l) the system must take into account the norms, standard,
concepts, and other referents which constitute the "social
reality" of the members of the social system of interest;
(2) the system must consider verbal and nonverbal behaviors
as functionally complementary; and (3) the system must be

quantifiable. Methodologically, it was proposed that the

98

99

measurement system must meet the requirements of (1) ob—
jectivization, (2) relativization, (3) reliability, (4)
precision, and (5) empirical verifiability. To the degree
that a given measurement system satisfies these theoretic
and methodological criteria, its potential scientific
utility is heightened. Mindful of the above, let us assess
the scientific utility of the system developed in this study.

It is my belief that the system developed herein
meets the theoretic requirements outlined above. Considerable
effort was expended to solicit and incorporate the input of
members of the social system in which observations were re-
corded. In particular, two steps were taken to fulfill the
first theoretic requirement: (1) observers consensually
generated the verbal and nonverbal referents used in assessing
displayed levels of the various stylistic attributes, rather
than having empirical referents supplied to them by the-
researcher; and (2) a standard of measurement was employed
which was derived from observer's experiences, rather than an
arbitrarily imposed one. Such an approach is consistent
with Hewes' (1978) call for a "combinational" approach that
facilitates both precision and comprehensibility in communi-
cation research.

The second theoretic requirement seems to have been
fulfilled as a result of having observers offer holistic
appraisals of molar attributes like openness, attentiveness,
dominance, and so forth, requiring them to consider individual

verbal and nonverbal behaviors as they cluster or function

 

100
systematically in human communicative exchanges.

Finally, by having observers use the direct magnitude
estimation procedure, the requirement that the system be
quantified was satisfied.

Methodologically, the criterion of objectivization
appears to have been satisfied. Measurement was stimulus-
centered and was based on an epistemic framework which was
the product of a consensus of observers, rather than that of
any single observer's perspective or vantage point.

Furthermore, the methodological criterion of relativi-
zation was also apparently satisfied. By definition, the
direct magnitude estimation procedure requires the inclusion
of a benchmark in relation to which all other stimuli being
scaled are compared and assigned scale values.

The results pertaining to reliability and precision
offer further evidence of the potential scientific utility
of the measurement system. Intra-class correlation coeffi-
cients were generally high for all stylistic attributes,
ranging from .714 (friendly) to .949 (dominant). The range
of ratings reported by raters was clearly greater for all
stylistic attributes than would have been possible using
many conventional scaling methods. By the same token, the
percentage of reliable variance produced by rater estimates
was generally high for all attributes, while the standard
error of measurement was in no case greater than 4.365. This
is a relatively small error rate when one considers the com-

plexity of the scale used (i.e., theoretically, all positive

101

integers comprise the scale). Thus, the criterion of
precision is satisfied since the scaling procedure incorporated
within the observational framework developed here is sensitive
enough to allow observers to report perceived differences
between stimuli being scaled, while minimizing error of
measurement.

The critical issue in demonstrating empirical verifia-
bility centers on demonstrating correspondence between raters'
holistic appraisals of a given stylistic attribute, and the
empirical or behavioral referents claimed by those raters
to provide an objective epistemic framework upon which the
appraisals are based. As was noted in Chapter Four, neither
human nor financial resources were available for the task
of checking the empirical verifiability of the ratings for
all attributes. Dominance was arbitrarily chosen and tested
for empirical verifiability. With the exception of duration

of direct glances, all other correlations between individual

1
)

referents and the raters' magnitude estimates were signifi-
cant and substantial (.51 to .83). While research by Exline
(1972) has indicated that duration of eye contact is posie
tively related to control or power in dyadic interactions,
it should be noted that the status of the interactants was
independently manipulated prior to the interaction, so that
dominant role-bound or submissive role-bound behavior may
very well have been a function of factors exogenous to the
interaction, rather than emergent during the interaction.

No such independent manipulation of power or status was

102

made in this research. Rather, the interactive setting was
that of an initial interaction between strangers whose
statuses may be presumed to be fairly equal going into such
interaction (Goffman, 1967). Also, the duration of eye
contact was also listed as a nonverbal indicant of attentive-
ness. An independent check of the relation between appraisals
of attentiveness and duration of eye contact produced a
Pearson g of .912 (df = 28, p < .001). Thus, it would appear
that, though they listed it as an indicator of dominance, the
raters associated duration of eye contact with attentive
behavior. This suggests that situational and relational
factors may influence how an individual verbal and/or
nonverbal cue is processed and interpreted by observers of
face-to-face interaction, and that perhaps such factors
should receive greater attention in preparing observers to
utilize these procedures. Nevertheless, the results generally
indicate that, at least in terms of dominance, the measure?
ment system is empirically verifiable. We can tentatively
expect to obtain similar findings regarding the other attrir
butes, since procedures for preparing and making measurements
of those attributes were uniform throughout the research.

The question of validity frequently arises with regard
to the development of new measurement procedures. It is
difficult to assess the validity of the system developed
here based on one study. Cm one hand, an argument for
predictive validity might be made since, as will be seen

shortly, the results pertaining to the relation between style

103
and perceived attractiveness are generally consistent with
previous studies (c.f., Norton & Pettegrew, 1976; Norton,
1978; Wiemann, 1977). Determining the validity of this
set of procedures will, however, probably hinge on at least
two additional factors, both of which can be said to relate
to content validity: (l) the degree to which the epistemic
framework generated by observers in this study is comparable
to those generated by observers in future studies; and (2)
the degree to which the same are comparable across the
situations in which observations are made.

In sum, the available evidence pertaining to the satis-
faction of theoretic and methodological criteria augur well
for the scientific utility of the measurement system developed
here for the study of communicative style in face-to-face
settings. At least one application of the system to a
specific research question has been made and was described
in Chapter Five. Let us now turn our attention to discussing

the results of that application.

Communicative Style and Interpersonal Competence

 

 

The results of the canonical analysis indicate that
there are at least two subsets of communicative style
variables which are highly and linearly related to two
corresponding subsets of perceived communicative effective-
ness and attractiveness variables. For purposes of expo-
sition, let us refer to these subsets of the first variable
(style) set as "communicative styles" in their own right,

and to the subsets in the second variable (competence)

 

104

set as "types" of perceived interpersonal competence in

terms of communicative effectiveness and attractiveness.
Accordingly, the first canonical correlation suggests that

an impression leaving, Open, attentive, animated, relaxed,

and to a lesser extent, friendly and dominant style, will
elicit perceptions of both social and task attractiveness
(i.e., the interactant's style will be seen as both potentially
likeable and easy to work with) and perceived communicative
effectiveness (i.e., the interactant's style will be seen as
skilled and pro-social).

These results are generally consistent with those obtained
by Norton and Pettegrew (1976), Norton (1978), and Wiemann
(1977). For example, Norton (1978; Study 1) found that
impression leaving, attentive, relaxed, open, animated, and
friendly all related to a "good" communicator image on the
first dimension resulting from a smallest space analysis.
Norton and Pettegrew (1976) found the relaxed, friendly, and
dominant style components to comprise an interpersonally
attractive style. Generally speaking, the results of these
and the present study suggest that a kind of communicative
or interactive style which is characterized by active
listening, supportiveness of the conversational partner(s),
openness and candor, and the ability to facilitate inter—
action through the control of information flow and mainte-
nance of a relaxed atmosphere tends to elicit attributions
of interpersonal attractiveness and communicative effective—
ness. As such, these findings are highly comparable to

' Wiemann's (1977) findings that supportiveness, social

105

relaxation, empathy, behavioral flexibility, and interaction-
management skills promote perceptions of interpersonal
competence. In short, a fairly stable pattern of association
between style and competence components has been identified
across a number of observational settings.

In terms of the second canonical variate, the results
suggest that an attentive, friendly, precise, and to a
lesser extent, relaxed style will elicit perceptions of task
attractiveness (i.e., the interactant‘s style will be
viewed as facilitative of a positive working atmosphere).
While little research has been conducted directly relating
communicative style to interpersonal competence in a work
setting, it should be noted that Bradley and Baird (1977)
found that managers who were perceived as democratic and
likeable displayed an attentive and friendly style, which
is generally consistent with the results obtained in this
study. Both investigations seem to suggest that a person
who is relaxed (i.e., does not breed tension as a result of
his/her interactive behavior), precise (i.e., facilitates
productivity and efficiency through his/her choice of words
and gestures), friendly, and attentive makes an attractive
working associate. This conclusion, however, must remain
tentative. Researchers in the area of organizational and
industrial communication may wish to pursue studies of the
relation between communicative style and perceived task
attractiveness, in order to better understand the nature of

interpersonal competence in work settings.

106

Implications for Future Research

 

 

It has already been suggested that the measurement
procedures developed and employed here should prove useful
in the scientific study of human, face-to—face interaction.
The study reported in Chapter Five represents an initial
research application of the procedures. Specifically,
this study attempted to increase our understanding of the
link between social performance and perceived social
competence by identifying some relations between communi-
cative style and perceived interpersonal competence. Depend-
ing on the settings and situations in which style and
competence are investigated, it is likely that patterns
of association other than those identified here will be
discovered. This prospect will, hopefully, encourage
others to pursue such investigations. Beyond this prospect,
a number of implications and additional suggestions may be
considered.

The components of communicator style identified by
Norton (1978) are clearly descriptive of certain molar
classes of interactive behaviors, both verbal and nonverbal,
which are regularly displayed by actors in social settings.
In fact, they were selected for scaling in the present
study for three reasons: (1) they offer a set of attributes
in terms of which the communicative behaviors of individuals
may be typified; (2) the attributes seem appropriate for the
description of co—occurring verbal and nonverbal behaviors;

and (3) their selection permitted a more direct comparison

 

107

of the results obtained here with those obtained in
previous studies. These components are, however, by no
means exhaustive. In particular, the components included
in the style construct, with the exception of contentious,
describe primarily positive or pro—social behaviors. It is
entirely possible to conceive of additional, less positive
or pro-social attributes which characterize and even prepon—
derate the communicative styles of some individuals. Future
studies should address this possibility.

It may be of some theoretical and practical utility

to distinguish communicative style components which are

 

basically expressionistic in nature (i.e., those which reflect
observable expressive behaviors of an interactant) from those

which are basically impressionistic in nature (i.e., those

 

which reflect impressions formed by interactants and/or
observers regarding an individual's behavioral conduct). For
example, attentive, animated, dominant, and precise seem to
fall into the former category, while friendly and impression
leaving seem better suited to the latter category. This
distinction may be especially important with regard to the
measurement procedures developed in this research. As an
attribute reflects increasingly observable or expressionistic
behaviors, it should be increasingly easy for observers to
generate empirical referents for it. The theoretic require—
ments of the measurement system developed here dictate that
such referents be specified. To the degree that the subject—

ive or impressionistic nature of an attribute inhibits its

108

specification, the precision and accuracy of the system, and
thus, its scientific utility, may be questionable. This
implies that we may not only want to distinguish expression-
istic from impressionistic style components, but may also
seek to minimize the inclusion of the latter in an observa-
tional framework such as that proposed in this thesis.
Mention of the problem of selecting attributes or
components for inclusion in our observational framework
raises a related issue. The work presented in this thesis

may be criticized in part for imposing a set of attributes

 

on observers, rather than having observers identify such
attributes. The issue, of course, relates to the distinction
between an "induced" versus a "theoretically imposed"
perspective of theory construction and empirical inquiry in
communication, which was discussed in Chapter Three. The
approach taken here falls somewhere between the two extremes.
That is, while the concepts scaled were imposed on observers
by the researcher, the impirical referents which largely
constituted their operationalization were supplied by or
induced from the observers. Such an approach is similar

to Denzin's (1970) description of the use of sensitizing

 

concepts as endemic to a symbolic interactionistic perspective

 

of social behavior:

By sensitizing concepts, we refer to concepts that

are not immediately transformed into rigid, operational
definitions via an attitude scale or check-list.
Rather, sensitizing concepts are deliberately left
nonoperationalized until the investigator enters the
field and learns the specific meanings attached to the
processes represented by his concept...This process of

109

sensitizing-aficoncept permits the sociologist to
discover what 18 unique about each empirical instance
of the concept while he uncovers what it displays

in common across many different settings. Such a
conception forces (in fact allows) the sociologist

to pursue his interactionist view of reality to the
empirical extreme (pp. 455-456).

 

One may extend the above line of reasoning to suggest
that, if a researcher is interested in typifying the com-
municative styles of individuals in terms of a set of
attributes which constitute the social reality of those
individuals, the attributes themselves should be culled
from a sample of members of the social system under study.
One rather interesting question implied by such an approach
concerns whether the sets of attributes supplied by samples
across studies (a) reveal a central or core set of stylistic
components, and (b) yield components comparable and basic-
ally similar to those identified and imposed on observers
in studies like this one. Researchers cannot ignore this
issue in future explorations of communicative style in face-
to—face interaction.

It seems likely-that communicators choose to display
and/or mask certain elements of their respective communicative
styles as a function of their individual goals, the presence
and identities of others with whom they interact, as well
as the social situations in which such interactions take
place. Typifications of an individual's communicative style
and perceptions of his/her interpersonal competence may vary
accordingly. This suggests that researchers ought to address

the possibility of both context-specific and cross-contextual

110

relations between components of communicative style and other
variables, in planning and executing their research efforts.

In Chapter Two the "accomodation model" of Giles and
Powesland (1975) was discussed. Recall that this model is
predicated on the assumption that a person can elicit more
favorable interpersonal evaluations from others by reducing
dissimilarities between himself/herself and these other persons.
To accomplish this convergence or dissimilarity reduction,
interactants may adapt their speech patterns toward the speech
patterns of fellow interactants. Mindful of the preceding,
the use of the procedures developed in this thesis may be
useful in studies of "style accomodation," particularly
those examining convergence behavior and relational develop-
ment longtitudinally.

Finally, while the development of the measurement
procedures was, in part, predicated on the notion that
communicative style is a multiple-attribute construct, the
use of these procedures in the study of individual attributes
may prove fruitful. Examples of such studies might include
the study of dominance in face-to-face interaction, studies
of selfedisclosure in face—to-face interaction, or studies
of attentiveness in face-to-face interaction. Also, the
general set of procedures might also be applied to attributes
of interest to researchers which were not eXplored in this
research. Interested researchers could employ the procedures
to obtain precise, quantitative appraisals of any number
of communication or behavioral attributes, provided empirical

referents for these attributes could be specified.

 

111

Limitations g£_this Research

 

 

This study is limited by at least four factors: (1)
possible problems arising from the imposition of stylistic
attributes on observers; (2) the absence Of evidence per-
taining to the empirical verifiability of stylistic attributes
other than dominance; (3) the nature of the interactive set—
ting; and (4) the role played by observers in this study.

The first two factors have already been discussed in some
detail, and the reader is merely reminded of them here. The
latter two warrant further consideration.

The situation in which interactants were observed,
although pervasive in everyday social intercourse, is compara-
tively unique. That is, for any given human relationship,
the participants have only one initial interaction. As Wiemann
(1977) has noted, "criteria used to evaluate others probably
change over time; people may be more tolerant of some types
of behavior in initital interactions because they a£e_initial
interactions" (p. 211). While it may be argued that a basic
or elementary set of communicative style and interpersonal
competence attributes remains the foundation for social typi—
fication and evaluation of an interactant's conduct, more ideo—
graphic criteria may evolve as an interpersonal relationship
develops. For example, Suttles (1970) suggests that one
indicator of friendship is the development of relational norms
between friends which are often at odds with more general
cultural norms regarding similar behaviors. It is the agree-

ment, implicit or explicit, to break cultural norms which

112
marks the relationship. The specific ways in which ideo-
graphic factors such as the development of relational norms
might impinge upon the process of typifying communicative
style and interpersonal competence remain, by and large, Open
to empirical inquiry. Thus, as was earlier suggested,
researchers ought to explore both situation-specific and cross-
situational contingencies influencing relations between
communicative style and perceived interpersonal competence.
In the meantime, any attempt to generalize the findings of
this research beyond‘an initial interaction setting must be
made cautiously until pertinent data are available.

By assuming the role of non—participating observers of
face-to-face interaction, the raters in this research assumed
a different perspective of the behavior of interactants than
the perspectives most likely assumed by the interactants them—
selves. This is not a weakness from the standpoint of the
scientific enterprise, but it may be interpreted as a potential
limitation of the present study by those interested in the
perceptions of actors engaged in interpersonal processes.

It should be noted that interactants in this study were asked
to complete a version of Norton's (1978) Communicator Style
Measure, for the purpose of assessing the communicative styles
of their respective dyadic partner. These data, while not
directly related to the goal of the present thesis, will be

analyzed at a later time.

Summary and Conclusions

 

The purpose of this thesis was to propose a paradigm

 

 

113

for the study of interactive or communicative style in human
face—to-face interaction. Communicative style was conceptual-
ized and operationalized as a set of anchored attributes in
terms of which the verbal and nonverbal behaviors of individual
interactants may be differentiated and typified.

Chapter One introduced the reader to the area of inquiry,

"I... .-———.-

and pvaided an overview of the contents of the thesis.

Chapter Two presented a review of previous research on

a...
"M

- -_....._-n.~_,.. -...—.-I

communicative style. Special attention was devoted to the
work of Norton (1978) and his asSociates. An attempt to
specify conceptual and methodological problems with previous
studies was accompanied by an attempt to identify steps toward
resolving these problems.

...nI-u’” '

In Chapter Three, an attempt to develop a paradigm for
the empirical assessment of style was made. A series of
conceptual and methodological requirements were specified in
terms of which the procedures might be assessed. Proceeding
from a number of extant conceptions of social behavior,
including attribution theory, learning theory, symbolic inter-
actionism, linguistics, sociolinguistics, and the psychology
of social norms, it was argued that the average level of a
given stylistic attribute displayed by "most persons" in a
given situation serves as a benchmark in relation to which the
communicative conduct of an individual interactant is assessed.
It was further argued that transforming this social system of

typification from a vernacular to a quantitative one would

render it useful for scientific description, provided the

114

requirements of reliability and precision, objectivization,
relativization, and empirical verifiability were satisfied.
Subsequently, the procedures by which this transformation
was attempted and applied here were described.

Chapteero r presented data pertaining to the reliability
and precision, and empirical verifiability of the procedures
developed herein. The results indicated that measurements
obtained utilizing these procedures were both reliable and
precise, and that in terms of at least one attribute
(dominance), the measurements were empirically verifiable.

In Chapter Five, a study of the relation between
communicative style and perceived interpersonal competence
was described. Relevant literature pertaining to communica-
tive competence was briefly reviewed and an operationalization
of competence offered. Subsequently, the procedures by which
the study was conducted were described. The results of the
study indipated at least two different communicative styles
are related to two different types of perceived interpersonal
competence.

This chapter concludes the thesis with a discussion of
the work presented in preceding chapters.

In concluding this discussion, two points are worth
noting. First if, as Hewes (1978) has suggested; any test
of a theory is "really a test of the assumptions of that
theory and the assumptions of the methodology used in the

study" (p. 165), and communicative style is conceptualized

as a multi-behavioral construct, then the use of measurement

115
procedures such as those proposed in this thesis is recom-
mended, particularly if the multiple behaviors which constitute
the various components of style are viewed as functionally
complementary. To focus on discrete, isolated behaviors
at the empirical level would be inconsistent with assumptions
about the phenomena under study.

Second, if Hewes (1978) is correct in asserting that.
"researchers should have at their disposal the widest possible
range of methodological tools" (p. 165), then the addition
of these procedures to the methodological repertoire of
communication scholars is desirable. This is not an argument
in favor of the proliferation of observational methods for
the sake of proliferation, since we have already noted the
potential confusion which may result from applying different
conceptions and methods of observation to the "same"
phenomenon. Rather, it is a suggestion that access to and the
use of alternative observational procedures should, for the
present, provide multiple bases for establishing the empirical
stability of relations between communicative style and other

variables of interest.

REFERENCES

Argyle, M. & Kendon, A. The experimental analysis of social
performance. In In Berkowitz (Ed.), Advances £3 Experi-
mental Social Psychology, Vol. 3. New York: Academic
Press, 1967, 55-98.

 

 

Bales, R. Personality and Interpersonal Behavior. New York:
Holt, Rinehart & Winston, 1970. '

 

 

Berger, C. R. Interpersonal communication theory and research:
An overview. In B. D. Ruben (Ed.), Communication 1
Yearbook I. New Brunswick: Transaction Books, 1977; L

217-228. -

 

Berger, P. L. & Luckmann, T. The Social Construction pf
Reality. New York: Anchor Books, 1967.

 

Blalock, H. M. Social Statistics. New York: McGraw—Hill,
1972.

 

 

Bochner, A. & Kelly, C. Interpersonal competence: Rationale,
philosophy, and implementation of a conceptual frame-
work. Speech Teacher, 1974, 23, 279-301.

 

Born, M. Einstein's Theopypf Relativity. New York: Dover,
1962.

 

 

Bradley, P. H. & Baird, J. E. Management and communicator
style: A correlational analysis. Central States Speech
Journal, 1977, 28, 194—203.

 

Brandt, D. R. Communicator style and interpersonal attraction
in dyads. Unpublished manuscript, Department of
Communication, Michigan State University, 1976.

Brandt, D. R. On linking social performance with social
competence: Some relations between communicative style
and attributions of interpersonal attractiveness and
effectiveness. Human Communication Research, in press.

 

 

Brunswick, R. Systematic and Representative Design pf
Psychological Experiments with Results 13 Physical and
Social Percgption. Berkeley: University of California
Press, 1947.

 

 

 

116

117

Byrne, D. The Attraction Paradigm. New York: Academic
Press, 1971.

 

Carnap, R. Ag Introduction Eg_the Philosophy pg Science.
New York: Basic Books, 1966.

 

 

Cronback, L. J. Coefficient alpha and the internal structure
of tests. Psychometrika, 1951, 16, 297-334.

 

CUshman, D. P. & Craig, R. T. Communication systems: Inter-
personal implications. In G. R. Miller (Ed.),
Explorations in Interpersonal Communication. Beverly
Hills: Sage PEblications, 1976, 37—58.

 

 

Danes, J. & Woelfel, J. An alternative to the "traditional"
scaling paradigm in mass communication research:
Multi—dimensional reduction of ratio judgments of
separation. Paper presented at the annual meeting of
the International Communication Association, Chicago,
Illinois, 1975.

Denzin, N. K. The methodologies of symbolic interaction:
A critical review of research techniques. In G. P.
Stone and H. A. Farberman (Eds.), Social Psychology
Through Symbolic Interaction. Waltham: Xerox Press,
1970; 455-456.

 

 

Dittman, A. The relationship between body movements and

moods in interviews. Journal pf Consulting Psychology,
1972, 23, 283-292.

 

Dubin, R. Theopy Building. New York: Free Press, 1969.

 

Duncan, S. D. Some signals and rules for taking speaking
turns in conversations. Journal pf Personality and
Social Psychology, 1972, 23, 283—292.

 

 

Ellis, D. G. Issues in analyzing sequential interaction
data: A plea for rigor in matters of observation.
Paper presented at the annual convention of the Speech
Communication Association, Washington, D. C. 1977.

Exline, R. V. Visual interaction: The glances of power and
preference. In J. K. Cole (Ed.), Nebraska Symposium pp
Motivation. Lincoln: University of Nebraska Press,
1972; 162-205.

 

 

Finn, J. D. A General Model for Multivariate Analysis.
New York: Holt, Rinehart.&Wihston, 1974}

 

Giles, H. & Powesland, P. F. Speech St 1e and Social Evalua-
tion. New York: Academic Press, 1 75.

 

118

 

Goffman, E. Encounters: Two Studies ip_the Sociology pf
Interaction. Indianapolis: Bobbs-Merri , .

 

Goffman, E. Interaction Ritual. New York: Anchor Books,
1967.

 

Grimshaw, A. D. Sociolinguistics. In I. de Sola Pool,
G. W. Frey, W. Schramm, N. Maccoby, & E. B. Parker,
(Eds.), Handbook of Communication. Chicago: Rand
McNally, 1973; 49-92.

 

Hamblin, R. L. Social attitudes: Magnitude measurement
-and theory. In H. M. Blalock (Ed.), Measurement $3 the
Social Sciences: Theories and Strategies. Chicago:

Aldine, 1974; 61-120.

 

 

Harris, R. J. A Primer 2f Multivariate Statistics. New York:
Academic Press, 1975.

 

Hawes, L. C. Alternative theoretical bases: Toward a pre-
suppositional critique. Communication Quarterly, 1977,
25, 63-68.

 

Hays, W. L. Statistics for the Social Sciences. New York:
Holt, Rinehart & Winston, 1973.

 

Hewes, D. E. Interpersonal communication theory and research:
A meta-methodological overview. In B. D. Ruben (Ed.),
Communication Yearbook II. New Brunswick: Transaction

Books, 1978; 155-170.

 

Holstein, C. M., Goldstein, J. M. & Bem, D. J. The importance
of expressive behaviors, involvement, sex, and need-
approval in induced liking. Journal of Experimental
Social Psychology, 1971, Z, 5 4-54 . _—

 

 

Hymes, D. Foundations g: SociOlinguistics. New York: Holt,
Rinehart & Winston, 1974.

 

 

Jones, E. E. & Davis, K. E. From acts to dispositions: The
attribution process in person perception. In L. Berkowitz
(Ed.), Advances in Experimental Social Psychology. New
York: Academic Piess, 1965; 219-266.

 

 

Keiser, G. J. & Altman, I. Relationship of nonverbal behavior
to the social penetration process. Human Communication

 

 

Kelly, H. H. Attribution theory in social psychology. In
D. Levine (Ed.), Nebraska Symppsium gp_Motivation.
Lincoln: University of Nebraska Press, 1967; 192-238.

 

 

119

Knapp, M. L. Nonverbal Communication Ip_Human Interaction,
2nd ed. New York: Holt, Rinehart & Winston, 1978.

 

 

Kramer, E. E. The Nature and Growth pf Modern Mathematics,
Vol. 1. Greenwich, Connecticut: Fawcett Premier BoOks,

1970.

 

 

Kirk, R. E. Introductory Statistics. Monterey, California:
Brooks/Cole Publishing Company, 1978.

 

Kuhn, T. S. The Structure 9f Scientific Revolutions, 2nd ed.
“ Chicago: University of Chicago Press, 1970.

 

 

Labov, W. Reflections of social processes in linguistic
structures. In J. A. Fishman (Ed.), Readings i3 the
Sociology pf Language. The Hague: Mouton, 1968; 240-
251.

 

 

 

Leary, T. Interpersonal Diagnosis of Personality: A Functional

 

Theory gf_Methodology for Personality Evaluatiofi. New
York: Ronald Press, 1957.

 

Leathers, D. G. Nonverbal Communication Systems. Boston:
Allyn & Bacon, 1976.

 

Leginsky, W. & Izzett, R. Linguistic styles as indices of
interpersonal distance. Journal 9: Social Psychology,
1973, 91, 291-304.

 

Lieberman, M., Yalom, I. & Miles, M. Encounter Groups: First
Facts. New York: Basic Books, 1973.

 

Iowe, C. A. & Goldstein, J. W. Reciprocal liking and attribu-
tions of ability: Mediating effects of perceived intent
and personal involvement. Journal of Personality and
Social Psychology, 1970, 16, 29I-29TT

 

 

Mann, R., Gibbard, G. & Hartman, J. Interpersonal Styles and
Group Development. New York: Wiley, 1967.

 

 

 

Masterman, M. The nature of a paradigm. In I. Lakotos &
A. Musgrave (Eds.), Criticism and the Growth pf Knowledge.
Cambridge: Cambridge University Press, 1970; 33-42.

 

 

McCroskey, J. C., Hamilton, R. R. & Weiner, A. N. The effect
of interaction behavior on source credibility, homophily,
and interpersonal attraction. Human Communication
Research, 1974, 1, 42-52.

 

 

Mead, G. H. Mind, Self, and Society. Chicago: University of
Chicago Press, 1934.

 

 

 

120

Mehrabian A. Inference of attitudes from the posture,
orientation, and distance of a communicator. Journal 9:
Consulting and Clinical Psychology, 1968, 32, 296-308.

Mehrabian, A. Significance of posture and position in the
communication of attitudes and status relationships.
Psychological Bulletin, 1969, 11, 359-372.

Mehrabian, A. Nonverbal Communication. Chicago: Aldine,
1972.

Meyer, S. L. Data Analysis for Scientists and Engineers.
New York: Wiley, 1975.

 

Miller, G. R. & Steinberg, M. Between People: A New
Analysis 9: Interpersonal Communication. Chicago:
Science Research Associates, 1975.

 

Miller, L. D. Dyadic perceptions of communicator style:
Replication and confirmation. Communication Research,
1977, g, 87-121.

 

Norton, R. W. Teacher effectiveness as a function of com-

 

 

municator style. In B. D. Ruben (Ed.), Communication
Yearbook I. New Brunswick: Transaction Books, 1977;
525—542.

Norton, R. W. Foundations of a communicator style construct.
Human Communication Research, 1978, 4, 99-112.

 

Norton, R. W. & Miller, L. D. Dyadic perceptions of communi-
cator style. Communication Research, 1975, 2, 50-76.

 

Norton, R. W., Schroeder, A. & Webb, J. Communicator style
as a function of effectiveness and attractiveness in
triads. Unpublished manuscript, Department of Speech
and Theatre, University of Michigan, 1975.

Norton, R. W. & Pettegrew, L. S. Communicator style as an
effect determinant of attraction. Paper presented at
the annual convention of the International Communication
Association, Portland, 1976.

Norton, R. W. & Warnick, B. Assertiveness as an interpersonal
construct. Human Communication Research, 1976, 3, 62-66.

 

Naiman, T. & Breed, G. Gaze duration as a cue for judging
conversational tone. Representative Research 13 Social
Psychology, 1974, 5, 115-122.

 

 

Nunally, J. C. Psychometric Theory. New York: McGraw-Hill,
1967.

 

 

121

Pettegrew, L. S. An investigation of therapeutic communicator
style. In B. D. Ruben (Ed.), Communication Yearbook
I. New Brunswick: Transaction Books, 1977? 593—604.

 

Pettegrew, L. S. & Thomas, R. C. Communicator style differ-
ences in formal vs. informal therapeutic relationships.
In B. D. Ruben (Ed.), Communication Yearbook II. New
Brunswick: Transaction Books, I978;‘522:537.——

 

Parks, M. Communication competency. Paper presented at the
annual convention of the Speech Communication
Association, San Fransisco, 1976.

Pearce, W. B. The coordinated management of meaning: A
rules-based theory of interpersonal communication. In
G. R. Miller (Ed.), Explorations in Interpersonal
Communication. Beverly Hills: sage, I976; I7-35.

 

 

Pacanowsky, M. E. & Fink, E. L. The effects of perceived
potential reward on individuation. Paper presented at
the annual convention of the International Communication
Association, Portland, 1976.

Rapoport, A. Qperational Philosophy. New York: Harper &
Row, 1953.

 

Ring, K., Braginsky, D. & Braginsky, B. Performance styles
in interpersonal relations: A typology. Psychological
Reports, 1966, 18, 203-220.

 

Ring, K. & Wallston, K. A test to measure performance styles
in interpersonal relations. Psychological Reports, 1968,
EE, 147-154.

 

Rogers, C. Client-Centered Therapy. Boston: Houghton
Mifflin, 1951.

 

Rosenfeld, H. Instrumental affiliative functions of facial
and gestural expressions. Journal g£_Personality and
Social Psychology, 1966, 4, 65-72.

 

 

Rossiter, C. M. Models of paradigmatic choice. Communication

 

Quarterly, 1977, 25, 69-73.

 

Scheflen, A. Quasi—courtship behavior in psychotherapy.
Psychiatry, 1965, 28, 245-257.

 

Sherif, M. The Psychology of Social Norms. New York:
Harper & Row, 1936. ‘—

 

 

 

Sherif, M. & Sherif, C. W. Reference Groups. New York:
Harper & Row, 1964. ‘

 

122

Shinn, A. M. Relations between scales. In H. M. Blalock
(Ed.), Measurement ip_the Social Sciences: Theories and
Strategies. Chicago: Aldine, 1974; 121—158.

 

 

 

 

Schutz, W. FIRO: A_Three-Dimensional Theory of Interpersonal
Behavior. New York: Holt, Rinehart & Wifigton, 1958.

 

 

 

Steiner, C. Scripts People Live: Transactional Analysis pf
Life Scripts. New York: Grove Press, 1974.

 

 

 

Stevens, S. S. The direct estimation of sensory magnitudes:
Loudness. American Journal 9: Psychology, 1956, 33,
1—25.

 

 

Stokes, R. & Hewitt, J. P. Aligning actions. American
Sociological Review, 1976, £3, 838—849.

 

 

Snyder, M. The self-monitoring of expressive behavior.4 -
Journal 2: Personality and Social Psychology, 1974, 33,
526-547.

 

Sullivan, H. The Interpersonal Theory pf Psychiatpy. New
York: W. W. Norton & Company, 1953. '

 

 

Suttles, G. D. Friendship as a social institution. In G. J.
McCall, M. M. McCall, N. K. Denzin, G. D. Suttles &
S. B. Kurth, Social Relationships. Chicago: Aldine,
1970; 95-135.

 

Thorndike, R. M. & Weiss, D. J. A study of the stability
of canonical correlations and canonical components.
Educational and Psychological Measurement, 1973, 33,

I233I347—_—

 

Torgerson, W. S. Theory and Methods 9: Scaling. New York:
Wiley, 1958.

 

Tucker, R. K. & Chase, L. J. Canonical correlation in human
communication research. Human Communication Research,
1976, 3, 86-96.

 

Van de Geer, J. P. Introduction pg Multivariate Analysis
for the Social Sciences. San Fransisco: W. H. Freeman,
1971.

 

 

 

Watzlawick, P., Beavin, J. & Jackson, D. Pragmatics 9:
Human Communication: A Study pf Interactional Patterns,
Pathologies, and Paradoxes. New York: W. W. Norton &
Company, 1967.

 

 

 

 

Wiemann, J. M. & Knapp: M. L. Turn-taking in conversations.
Journal 2: Communication, 1975, 33, 75-92.

 

123

Wiemann, J. M. Explication and test of a model of communicative
competence. Human Communication Research, 1977, 3,
195—213. ‘

 

Winch, P. The Idea of a Social Science and Its Relation

pp Philosqphy. ILOHdon: Routledge and Kegan Paul,
958.

 

 

 

Woefel, J. Metric measurement of cultural processes. Paper
presented at the annual convention of the Speech
Communication Association, Chicago, 1974.

Woelfel, J. Science and human phenomena. Unpublished
manuscript, Department of Rhetoric and Communication,
State University of New York at Albany, 1978.

 

 

RIES

1111311111

E UN

w

3

HIGH

\nmllull

312931

 

613