U.S.

Copyright Law
(title 17 of U.S. code)
governs the reproduction
and redistribution of
copyrighted material.

Downloading this
document for the
purpose of
redistribution is
prohibited.
HANDBOOKOF
COMMUNICATION
SCIENCE
EDITORS
Charles R. Berger &• Steven H. Chaffee

SAGE PUBLICATIONS
The Publishers of Professional Social Science
Newbury Park Beverly Hills London New Delhi
Copyright © 1987 by Sage Publications, Inc.

All rights reserved. No part of this book may be reproduced or utilized

in any form or by any means, electronic or mechanical, including
photocopying, recording, or by any information storage and retrieval
system, without permission in writing from the publisher.

For information address:

SAGE Publications, Inc.

2111 West Hillcrest Drive
Newbury Park, California 91320
SAGE Publications Inc. SAGE Publications Ltd.
275 South Beverly Drive 28 Banner Street
Beverly Hills London EC1Y 8QE
California 90212 England

SAGE PUBLICATIONS India Pvt. Ltd.

M-32 Market
Greater Kailash I
New Delhi 110 048 India

Printed in the United States of America

Library of Congress Cataloging-in-Publication Data

Main entry under title:

Handbook of communication science.

Includes index.
1. Communication. I. Berger, Charles R.
II. Chaffee, Steven H.
P90.H294 1987 001.51 87-9783
ISBN 0-8039-2199-3
3 What Communication
Scientists Do
STEVEN H. CHAFFEE
Stanford, University

CHARLES R. BERGER
Nortdruestern University

‫حتم!آ^آل‬٦‫جط‬£‫ك‬،،!٢0‫ءآلأل‬٠‫ي!!عألد‬٠‫ئ‬
being without developing some expectations—we could even call them
implicit theories—about how we and others communicate. Our answer
lies in the way communication scientists think, more than anything else;
so we will review here some key feahires of thought that are common
among communication scientists but rare in everyday life. Many of the
principles we will discuss are shared with other social and behavioral
sciences, but some are peculiar to communication science and are points
of contention between it and other forms of communication study.

BACKGROUND ASSUMPTIONS

The concept of a science of human communication rests upon the opti­

mistic assumption that behavior can be both understood and improved
through systematic study. Further, it assumes that improvement must be
based upon understanding, which is the primary goal of a science. Like
all assumptions, these beliefs are not tested within the science itself.
Instead they provide a working point of view, which derives its validity
in the long run from the value of the knowledge produced. If we did not
believe that we could understand human communication, and so per­
haps improve upon it, we would not undertake scientific study with
those goals in mind. Still, we know that there are limits beyond which
communication is not subject to further understanding or improvement
through scientific research.
These limits do not concern us when we are acting in the role of com­
munication scientist. The working assumptions of a scientist are called
into question only after the research is done. If it fails to produce the

99
100 OVERVIEWS

hoped-for understanding or improvement, we might consider approach­

ing the communication problem through methods other than science (see
Farrell, Chapter 4 in this volume).
Besides explanation, basic goals of science include prediction and
control. By control we do not mean that communication scientists want
to manipulate individuals or institutions for their own purposes; scien­
tific control means that we would know how to produce a phenomenon
in which we are interested by setting up the conditions that account for it.
Understanding those conditions is the essence of prediction.
Whether the phenomenon we produce has a positive or a negative
impact is a question that lies beyond scientific discourse. For example,
the ability of physicists to produce atomic reactions has led to both nuclear
weapons and nuclear medicine. Theories of communication are used
both to sell people cigarettes and to persuade them not to smoke. As
citizens we can certainly oppose nuclear weapons or cigarette smoking.
That has little to do with our work as scientists, except that our selection
of research topics might reflect our personal values. Whether knowl­
edge about nuclear fission or how to market cigarettes should be sup­
pressed “for the good of society” is not a question to take up here. But
communication scientists know that our explanations and predictions
could be used to produce effects we would not want.

THEORY

Communication scientists think and talk about theory a lot. They work
toward development of theory, and they bemoan the fact that there is not
more good theory in the field. This raises two basic questions: What is
“theory”? And what makes a theory “good”?
We will spend some portion of this chapter describing what theories
are and how they can be evaluated. We will also present a way of thinking
about communication at four levels of analysis, which we believe can
promote more useful theoretical work. This effort is not itself a theory,
but a way to integrate many disparate domains of communication inquiry.
The world offers an incredible variety of communication phenomena
begging for explanation, so there is ample room for the development of
communication theories.

What Is a Communication Theory?

This entire book is filled with examples of communication theory.
Behind them lies an ongoing debate over the proper approach. We can­
not possibly deal with the full variety of positions here. Some advocate
rule-governed approaches, in which communication acts would be inter­
Chaffee, Berger I What Communication Scientists Do 101

preted as the result of people following certain rules for behavior;

research would seek to find out what those rules are (Cushman, 1977).
Others hold to a classically scientific law assumption; that is, that there
are general laws of human behavior as there are in, say, biology or
chemistry, and that research should test predictions from general propo­
sitions to see if they hold up (Berger, 1977; Hempel, 1965). Another
position is to assume that a general form of theory cuts across all sci­
ences and will be equally applicable to communication, such as the
general systems approach (Monge, 1973, 1977). A decade ago this
kind of debate occupied an entire issue of Communication Quarterly
(1977, Vol. 25). These meta-theoretical issues are important because
they influence the directions researchers take. But overemphasis on
them can discourage researchers from developing substantive theories
to explain communication.
A useful working definition of a theory is a set of constructs that are
linked together by relational statements that are internally consistent
with each other. Constructs are concepts that are formed inductively by
generalizing from particulars. An example is “interpersonal influence,”
which is a name that has gradually developed to refer to a variety of
observations of human activity. Constructs are abstractions; they are
given meaning through theoretical definitions. Interpersonal influence
may be defined conceptually as a change in one individual that is pro­
duced by another individual or a group.
Constructs vary in levels of abstraction. The more abstract they are,
the more of a particular domain they cover. Interpersonal influence is
more abstract than “salesmanship” but less abstract than “persuasion.”
Constructs may also be thought of as theoretical variables, which
means they may take on different values. The relational statements in
theories of communication are about two or more variables. For exam­
ple, we may theorize that there is more interpersonal influence between
people of great “empathy.” This is the most common kind of theoretical
statement: a prediction that two constructs will covary together.
Not all relationships are so simple; in human communication few are.
There are a number of types other than the positive relationship exem­
plified by empathy and interpersonal influence. Relationships can also
be inverse or negative, which means that increases in one variable are
associated with decreases in another. For example, increases in self­
disclosure produce decreases in one’s attractiveness. Other, more com­
plex relational statements can be used to link constructs. Nonlinear
relationships are common in communication. For example, it may be
that increases in empathy heighten interpersonal influence up to a cer­
tain level, but beyond that make no further improvement; this is called
an asymptotic function. Or perhaps very high levels of empathy are
associated with a decline in interpersonal influence, in which case the
102 OVERVIEWS

relationship is curvilinear. It is very important to know about asymp­

totic or curvilinear effects in communication so that extensive effort is
not put into activity that is useless or counterproductive. For example, in
political advertising a candidate can gain support by simply having his
name repeated several times—but lose support if it is mentioned so often
that people get tired of hearing it (Becker & Doolittle, 1975). It is rarely
the case that the more communication, the better.
Most theories have more than two constructs in them, and relation­
ships can become complex. They should, because human communica­
tion is quite complex. A three-variable theory may involve an interaction,
where the relationship between two variables depends upon values of
other variables. For example, self-disclosure might be positively related
to attractiveness for females, but for males these two variables might be
negatively related. In this example, sex, the third variable, determines
the relationship between the other two. Again, interactions are the rule
rather than the exception in communication. The most defensible
answer to many questions about relationships between communication
constructs is “It depends.”
Theories also contain boundary conditions. Boundary conditions
specify the domain of events the theory explains, and what lies out­
side it. A theory constructed to explain communication in work groups
might not apply to family groups. Or a theory about media use based on
the middle class might not successfully predict what very wealthy or
very poor people do. Boundary conditions can be introduced a priori as
part of the theory, but more often they grow out of empirical research
findings when it turns out that the theory does not work under some
conditions.
Scientific theories are useful to the extent that they can be tested to
assess their validity. To test theoretical relationships, it is necessary to
measure theoretical variables. The move from the theoretical level to
that of empirical research is accomplished by operationally defining
theoretical constructs. Operational definitions specify how constructs
are measured. For example, self-disclosure might be measured by a set
of questions asking the person how much she provides others with infor­
mation about herself. Or alternatively, the researcher might observe
people in interactions and keep track of how often they say things about
themselves. Finally, one might manipulate self-disclosure by having
people interact with a confederate who systematically varies the amount
of disclosure. In fact, all three of these ways of operationalizing self­
disclosure have been used.
Several important points should be made about operational defini­
tions. First, no one operational definition can possibly capture the full
meaning of a theoretical construct. Second, it is difficult to create an
operational definition that represents only the construct intended. There
Chaffee, Berger / What Communication Scientists Do 103

is always some slippage between constructs and what is done in a spe­

cific study. Therefore, multiple operations of a construct are preferable
to singular operations. If we operationalize a construct several different
ways, and get approximately the same result each way, we have much
more confidence that the research represents the construct intended.
There have been strong arguments for a multiple operations procedure
(Blalock, 1984; Webb, Campbell, Schwartz, & Sechrest, 1966) but it is
rarely used in studies of communication. Somewhat the same purpose is
served by replication of a study, when the second study uses different
operational definitions. This presumes, however, that the comparison of
the two studies will be made by someone in a review of the literature;
examples can be found in many chapters of this book, frequently point­
ing up discrepancies between studies that lead to reformulated theories.
While many theories contain constructs that are not amenable to opera­
tionalization (“nominal” constructs), a theory with too many is hard to
test. Psychoanalytic theory is an example. A theory that cannot be tested
may contain valuable insights, but it is difficult to evaluate. Often in
communication research we use operational measures to represent dis­
positional constructs that may not exist. For example, the notion of
“attitude” originated from observing consistencies in behavior, but no
one has ever seen an attitude. The term is deeply entrenched in our
literature, but the validity of the measures rests upon the success with
which they enter into relationships specified in theories.
Once theoretical constructs are operationally defined, we can test
specific hypotheses deduced from the more general theory. Since opera­
tional definitions never fully represent their theoretical constructs, there
is always some degree of doubt about the theoretical significance of a
particular hypothesis test. Even when specific hypotheses are supported
by research findings, one cannot be confident that the general theo­
retical proposition tested is fully supported. Replications with different
operational definitions under varying conditions are needed. For exam­
ple, if we postulated a positive relationship between exposure to violent
media content and propensity to act aggressively, we would want to test
this proposition using a number of measures of the two constructs, as
well as a number of people and situations. Corroboration of each spe­
cific prediction would increase our confidence in the general propo­
sition. This is not often the case, however. For example, the predicted
positive relationship between media violence and aggressiveness is
usually found when the operational measure concerns what is actually
watched, but not when it concerns the persons’s “favorite programs”; it
is also more often found in junior high school samples than at other ages
(Chaffee, 1971). Communication scientists spend a great deal of effort
trying to adjudicate such conflicts between studies. Are they due to
faulty operations, or to poorly specified boundary conditions, or is the
104 OVERVIEWS

theory simply inadequate? These possibilities are discussed in detail by

Blalock (1984).
Theories typically contain a number of theoretical relationships, and
it is not unusual for some propositions to be better supported than oth­
ers. Assuming that the lack of empirical support for a proposition is not
due to faulty operationalizations or an inappropriate test domain, the
theorist may be forced to abandon a theoretical proposition. Because
theories are internally consistent systems of propositions, the deletion of
one proposition will most probably entail changes in those that remain.
Such modifications are commonplace in communication, as in any sci­
ence. The theories of many of the most venerated names—Galileo, New­
ton, Einstein—contained errors.

Evaluating Theories
In discussing how communication scientists think about and develop
theories, we have also used many of the criteria for evaluating theories.
It is time to specify these. The following list of attributes of a good
theory is similar to those most communication scientists—who are typi­
cally professors—outline for their beginning graduate students. We
summarize our own list here to help theory builders (and borrowers) be
critical of theory construction efforts they encounter or undertake.

(1) Explanatory power: Here we are concerned with the theory’s ability to
provide plausible explanations for the phenomena it was constructed to
explain. Also considered here is the range of phenomena that the theory
explains; the greater the range, the more powerful the theory.
(2) Predictive power: This criterion assesses theoretical adequacy by mea­
suring the theory’s ability to predict events. It is, however, possible for
theories to predict but not be able to provide plausible explanations.
(3) Parsimvny: Simple theories are preferred to more complex ones, assum­
ing that both predict and explain equally well. The complexity of a
theory is directly related to the complexity of the reality it seeks to
explain.
(4) Falsifiability: Theories should be capable of being proved false. If a
theory is not, it cannot be said to have survived a test even if research is
consistent with it. Popper (1963) suggests that theoretical propositions
be attacked from a variety of angles to see whether they can withstand
efforts to disprove them. If there are numerous nominal constructs in a
theory, it may be difficult to falsify since negative evidence can be
explained away.
(5) Internal consistency: The internal logic of a theory can be assessed inde­
pendently of empirical tests. Theoretical propositions should be consis­
tent with each other. If they are not, empirical findings may be difficult
to interpret within the theory.
(6) Heuristic provocativeness: Good theories generate new hypotheses,
which expand the range of potential knowledge.
Chaffee. Berger I What Communication Scientists Do 105

(7) Organizing power: Useful theories not only generate new knowledge,
they are able to organize extant knowledge.

Theory at Work: Agenda-Setting

A good example of how communication scientists use theory can be
found in the study of “agenda setting” by the news media. Cohen (1963)
wrote that the media may not be successful much of the time in telling
people what to think, but they are “stunningly successfol in telling peo-
pie what to think about:' McCombs translated this comment into a pre-
diction that topics emphasized in the press would be the topics people
think are important. This was first tested in the 1968 election campaign;
the researchers found that coverage of issues by a newspaper corre-
sponded fairly strongly with the issues voters used in deciding how to
vote (McCombs & Shaw, 1972).
The hypothesis has not always been suppported in subsequent re-
search. Weaver (1977) showed that it applies only to people with a high
“need for orientation,” and other researchers found the results hard to
replicate in a two-newspaper city (McLeod, Becker, & Byrnes, 1974)
or for local television (Palmgreen & Clarke, 1977). But the theory held
up well in a field experiment where people watched TV newscasts that
had been purposely altered to emphasize different political issues (Iyen-
gar, Peters, & Kinder, 1982). Their preferences among candidates
were strongly affected by the issues stressed in the experimental news
programs.
Agenda setting meets some of the tests we have outlined better than
others. As a simple two-construct prediction it is certainly parsimoni-
ous, and in its matching of the orders of two sets of issues it is internally
quite consistent. Most important to researchers, it has been heuristically
provocative; many studies have been organized around the idea. So it
has fairly strong predictive power. It is falsifiable, in that several studies
have searched for agenda-setting effects and not found them, although
there is some tendency for the term “agenda-setting fiinction” to survive
data that seem to falsify its prediction. On the other hand, it is a rela-
tively narrow theory, not very usefol for organizing knowledge beyond
the studies specifically directed at it. Its explanatory power is limited,
although it was improved by adding the “need for orientation” proviso.
We know how to look operationally for agenda-setting effects, and we
can recognize them clearly when they occur; but we do not know much
yet about how or why they occur-or why they often do not.
This brief example illustrates how theory relates to research in com-
munication science. There is an ongoing dialectic between the two. Ini-
tially, one scholar thought of a way of operationalizing the constructs,
tested the theory, and found encouraging results. Then, as this scholar
106 OVERVIEWS

and others tried to expand on the original findings, they discovered some
of its limitations and trimmed back the theory accordingly. In the absence
of a theory, this research would have had little point. Even when it turns
out to be incorrect in some ways, a theory is indispensable to a commu­
nication scientist trying to formulate new directions for study.

Theories of the Middle Range:

Limited Effects
Scholars sometimes lament the absence of a grand theory, such as
the atomic table or evolution, in communication science. We have many
specific hypotheses; but typically when many implications of a general
postulate are tested, the results include quite a few falsifications. Lazars-
feld’s pioneering group at Columbia University (see Delia, Chapter 2)
suggested that we should be aiming instead to develop “theories of the
middle range.” Their idea was that specific findings from replicated
studies would cumulate into empirical generalizations.
Unfortunately, communication scientists can be an impatient lot.
Many broad generalizations were drawn without the necessary replica­
tions, on the basis of one study or perhaps a couple of studies that used
very similar operationalizations. McLeod and Blunder (Chapter 9) dis­
cuss some premature conclusions about limited media effects that were
advanced by the Columbia group in the 1940s and 1950s. These propo­
sitions have not stood the tests of time and replication under different
operational conditions.
These theories of the middle range have not held up well partly
because they were not sufficiently modest in the scope claimed for them—
they might better have been labeled “upper-middle-range” theories—
but perhaps more because they were seized upon for their organizing
power rather than for predictive power or falsifiability. That is, the
limited-effects view of media was advanced as general knowledge about
communication, written into books (Berelson & Steiner, 1964; Klapper,
1960) and soon after into introductory textbooks. That is not the main
purpose of theory for a communication scientist, who is instead looking
for ideas to guide research. The precepts of the limited-effects model
were later used to organize research presentations demonstrating their
falsity. For example, Chaffee (1978) derived a series of predictions
about the effects of the presidential debates of 1976 from the limited-
effects model, and then showed how they had mostly been falsified by
various studies. The heuristic value of this middle-range theory was
considerable, even as its organizing power was being reduced.
This example illustrates the important point that the criteria for evalu­
ating a theory are not all consistent with one another. A communication
scientist must make some choices, emphasizing one criterion or another.
Chaffee, Berger I What Communication Scientists Do 107

Early proponents of middle-range theory were anxious to demonstrate

that their research was producing knowledge, and they built theories to
organize what had been found to that time. When a theory is framed for
that purpose, it is not being designed with falsifiability and further
empirical testing in mind.

LEVELS OF COMMUNICATION ANALYSIS

As we pointed out at the beginning of this chapter, the paucity of theory

in communication science is often bemoaned by its practitioners. In this
section we present a pretheoretical conceptual scheme meant to help
promote interrelated theory-building efforts. Communication scientists
assume they can find important patterns in social behavior through
observations of many similar actions. No one instance of communica-
tion, in this view, is especially important, although sometimes a single
event produces significant consequences. People are rarely conscious
that they are communicating and often have difficulty recalling details of
their communication after the event. Everyday activities such as alking,
reading road signs, or listening to the radio are not important enough for
us to mark carefully in time or space so that we can later reconstruct
them. The researcher faces a difficult task in imposing order upon com-
munication because it is taken so casually by the people involved in it.
One valuable approach to ordering the study of communication is to
think of several levels ofanalysis of communication events (Cushman &
Craig, 1976; Wright, 1959). Part II of this book is broken down by the
levels at which research is conceived and conducted. There is not total
agreement within the field on the precise definitions of levels, but most
of our colleagues are comfortable with a four-level breakdown.
The four levels of analysis we employ here are (1) the intraindividual
level, of processes that occur within the person in relation to communi-
cation activities; (2) the interpersonal level, where communicatory rela-
tionships involving two or a slightly larger number of persons are studied;
(3) the network or organizational level, where larger sets of persons are
studied in the context of a set of ongoing relationships; and (4) the mac-
roscopic societal level, wliere the communication properties and activi-
ties of large social systems are studied, often without immediate reference
to the people in those systems.
There is more than a heuristic value in organizing our presentation of
communication science in this way. These different levels tend to involve
different researchers, in pursuit of different goals. Research methods
vary widely across levels, as do the kinds of theories tested.
Levels of analysis should not be confused with topics of communi-
cation. Many communication scientists are drawn to their work by a
108 OVERVIEWS

concern for a specific communication function or context, such as the

socialization of new members to society or the conduct of political com­
munication. These substantive topics may be studied at several levels—
and often at all four. For example, in socialization there are individual
processes (learning), interpersonal (parent-child interaction), network
and organizational influences (peer groups, schools, churches), and mac­
rostructural factors (cultural influences of mass media). Each level needs
to be considered for a full picture of communication in the overall pro­
cess of socialization. Similarly, in political communication we find
studies of opinion formation (intraindividual), family influence (inter­
personal), reference groups and ethnic groups (network/organizational),
and campaign broadcasting policies (macrosystemic).
Just as a research topic cuts across levels, there are generic issues and
questions about communication at each level. These questions may
receive different answers in the hands of researchers faced with different
theoretical problems. But they must be resolved in some fashion by each
researcher.

ANALYTIC AND SYNTHETIC ISSUES

FOR RESEARCH

An important function of our levels scheme is to demonstrate how

general questions about communication can be asked at each level, and
how questions can foster integration of levels. Communication scien­
tists typically confine a research effort to one level. This means that, for
example, those who study informal interactions could overlook influ­
ences of individual cognitive mechanisms or wider social networks.
Their theories in turn would have only limited explanatory power.

Analytic Issues Within Levels

In this section we consider questions and issues to be dealt with at
each level. These questions are general, and responses to them are con­
sidered in each level chapter of Part II.
Structural issues. “Structure” refers to ways in which units of a com­
munication system are linked. First, what units make up the communi­
cation structure? At the individual level, the unit of analysis is often a
person’s behavior, belief, or cognition; at the macro level, the units
might be communication institutions. The problem of specifying units
at any level can be difficult. For example, some interpersonal communi­
cation researchers argue that interactions are best understood from an
intraindividual cognitive perspective (Planalp & Hewes, 1982); others
(Millar & Rogers, 1976), however, insist that interpersonal communica­
Chaffee, Berger / What Communication Scientists Do 109

tion needs to be examined at the level of the “interact” rather than the
individual act. Our point is that only after an investigator has settled
upon a unit of analysis is it possible to specify what linkages exist among
units.
Early theories of social cognition (Festinger, 1957; Harvey, Hunt, &
Schroder, 1961; Rokeach, 1960; Schroder, Driver, & Streufert, 1967)
were concerned with the ways cognitions are linked. Models of semantic
memory emphasized linkages among memory nodes and how these
linkages influence recall, as in Anderson and Bower’s (1973) and Ander­
son’s (1976) model of human associative memory (HAM). Models of
discourse comprehension and processing also concern themselves with
linkage issues (Abelson, 1981; Schank& Abelson, 1977; vanDijk, 1980;
van Dijk & Kintsch, 1983).
Similar questions can be raised at the other three levels. In dyadic com­
munication systems we can look at linkages between persons, which can
have attributes such as reciprocity. Analysis of different linkage patterns
is an active area of organizational communication research (Farace,
Monge, & Russell, 1977; Rogers & Agarwala-Rogers, 1976). Institu­
tions can be linked too. The broadcast networks in the United States are
linked by competition with one another in the marketplace, but in mutu­
ally supportive ways in their collective opposition to public regulation.
Some mass media institutions are controlled by corporate conglomer­
ates, making it difficult for the media to be critical of capitalist institu­
tions (Golding & Murdock, 1977).
Once linkages are specified, we can ask how active they are in rela­
tionship to each other. Network analysts of formal organizations focus
on this variable, asking, “What units of the system are highly active
information processors?” or, “What units are relative isolates?” Direc­
tional studies distinguish information “receivers” from “givers.” A
related issue is complexity. A communication system becomes more
complex with more links. How does complexity influence communica­
tion in the system? What does increasing complexity do to the units?
Can a system be “too complex”?
We can also consider the organization of a system. Are the units linked
as a hierarchy? Or in a linear fashion? Many models of individual cogni­
tion and of formal organization favor a hierarchical structure. Before the
1940 voting study (Lazarsfeld, Berelson, & Gaudet, 1948), academics
assumed that influential people were of higher social status than their
followers. On many topics, though, “opinion leaders” are a lot like the
people they influence (Katz & Lazarsfeld, 1955).
Other kinds of organization are possible too. Some cognitive psycholo­
gists argue that sometimes people process information in a relatively
linear fashion, by “scripts” (Abelson, 1976, 1981; Schank & Abelson,
1977). Scripts are expectations for event sequences that people encoun­
110 OVERVIEWS

ter repeatedly. In a restaurant, for example, we expect a series of events

to unfold because the restaurant script is structured in about the same
order every time. A script represents a linear-order organization princi­
ple that can occur at many levels of abstraction (Abelson, 1976). In
social or institutional relations, nonhierarchical forms of organization
can develop. Nations maintain elaborate formal protocols for diplomacy
(and even war).
For the communication scientist, the structural organization of com­
munication systems affects such variables as communication frequency
and time, direction of influence, message content, communicator style,
and who interacts with whom. These parameters may be related to the
evolution of structural organization. For example, forming a task-oriented
group whose members vary in communicative style can affect the power
structure that evolves. Once formed, that structure influences people
who later join the group. Similarly, at the individual level the structure
of prior knowledge influences processing of new information; current
information may in turn alter knowledge structures.
Communication goals. In the 1970s some social psychologists
(especially Harré & Secord, 1973) raised philosophical issues about
dominant theories and methodologies. One criticism was that social
psychology is based on a mechanistic conception of behavior, as if
humanity is at the mercy of powerful environmental influences that shape
behavior through stimuli, reinforcements, and punishments. Harré and
Secord (1973) charged that by searching for causes of behavior in labo­
ratory experiments, social psychologists make the same assumptions
about human actions as do radical behaviorists. They proposed instead
that social behavior should be viewed as emanating from the person, as
an actor capable of making choices. They contended that people specify
goals in social situations and take action to achieve those goals.
Current theory in cognitive science and artificial intelligence allows
for both kinds of explanations. System units, either persons or groups,
do formulate and pursue goals. In some contexts, such as family com­
munication, individual members may not have well-articulated goals,
and yet the family as a unit may have goals. A number of questions can
be asked within each level concerning these goals.
The “uses and gratifications” approach to mass media suggests that
individuals may consume media content to gratify such needs as passing
time, acquiring information, arousal, and companionship (Blunder &
Katz, 1974; Greenberg, 1974). A person might also generate symbols in
order to persuade, inform, or entertain an audience. At the level of social
and group interaction, Bales (1950) suggested that communication can
serve either task or socioemotional group goals. Parsons (1955) distin­
guished between instrumental and socioemotional goals in families. It is
questionable whether persons, groups, or institutions can give adequate
Chaffee, Berger I What Communication Scientists Do Ill

verbal descriptions of their goals. Norman’s (1981) activation-trigger-

schema (ATS) model of cognitive processes and human action argues
that people may be aware of their general goals but not of the many spe­
cific subgoals that must be reached to accomplish the main goal.
Another problem is the multiplicity of goals in many situations. When
individuals, groups, or institutions act, they may be pursuing several
goals with a single action. Researchers on natural language processing
recognize this problem in their attempts to develop computer programs
that understand and produce natural language (Wilensky, 1983). The
utterance, “I like you very much,” might be an attempt to induce the
recipient to like the person making the statement (ingratiation), or to
make the recipient more receptive to future requests for favors, or both.
It is difficult to design computer programs to understand such statements.
Individuals, groups, and institutions may intentionally mislead their
constituents. Statements about one’s communicative goals may in turn
mislead the researcher. Goffman (1969) analyzed the moves and coun­
termoves individuals and governments make to mislead observers about
their true intentions. There is a considerable literature on deception in
interpersonal communication (Knapp, Hart, & Dennis, 1974; Miller,
de Turck, & Kalbfleisch, 1983; also see Knapp, Cody, & Reardon,
Chapter 13), but the practice is certainly not limited to this level. There
is evidence, for example, that television advertisements are regarded as
deceptive by many people, including grade school children (Ward, Wack-
man, & Wartella, 1977).
Unintended effects. While communicative goals are being pursued,
unanticipated effects may be produced. Such effects have been observed
in public information campaigns and diffusion studies, but are not con­
fined to these settings. A person may produce messages for one purpose
only to find that they have also produced outcomes that were both unan­
ticipated and undesired. Unintended effects can also occur at the group
and organizational levels. Communication scientists are especially atten­
tive to “side-effects” or “latent functions” of communication.
Unintended effects can occur simply because the creator of a message
is unaware of the usual effects it might have. An example is television
violence, which has the primary purpose of attracting large audiences
but can also have the unintended effect of stimulating aggressive behav­
ior (National Institute of Mental Health, 1982). Or, a non-native speaker
of English might utter a statement in a way that produces a humorous or
angry response. The receiver’s noncompliance may stem from the for­
eign speaker’s lack of knowledge about appropriate communication.
For example, an employee does not “give orders” to the boss.
A second source of unintended effects is unanticipated contingencies
in the communication episode. During a conversation, a person may
disclose information related to other issues so that the emotional tone of
112 OVERVIEWS

the conversation changes quickly. For example, two interactants may

intend an amiable conversation, but end up in an argument due to sensi­
tive comments. Information from mass media is especially prone to
unintended effects; once a message is sent via the media, it is difficult to
“take it back.” In face-to-face encounters, retractions can often prevent
major damage.
Predispositions of message receivers can also cause unintended
effects. While a given communicator may know what is appropriate to
say in a situation, individual differences in the audience may interact
with incoming messages in unforeseen ways. In persuasion, influence
agents may overgeneralize about their audiences, imagining for example
that they are more (or less) hostile than they actually are. Often people in
everyday interactions are surprised to find that others disagree (or
agree) with them on an issue.
Unintended effects are not always “bad,” of course. Sometimes peo­
ple who initially interact to solve a problem become close friends. Influ­
ence agents might not only persuade their audiences, they may inform
and entertain them as well. However, unintended effects are probably
negative as often as not. People who influence others may also induce
those others to dislike them, for example.
Processing capacity. Miller (1956) noted that humans are limited to
the perception of about seven (“plus or minus two”) items at one time,
which he took to be an individual’s maximum information processing
capacity. This work spawned an enduring interest among experimental
psychologists in information processing. Individuals can process only a
fraction of the information available in a given situation; they rely on
internal knowledge structures to help “fill in the gaps” due to these limi­
tations. Taylor and Fiske (1978) demonstrated how judgments can be
inordinately affected by a salient stimulus in the environment. Target
persons who are made more salient than other group members are judged
by observers to be most influential in the group, even when their con­
tributions do not exceed those of the other members. People erroneously
“recall” actions from stereotyped sequences they have observed,
although the recalled action did not actually occur. What is being
accessed is the scripted expectation that the action should have
occurred, not the action itself (Bower, Black, & Turner, 1979). Nisbett
and Ross (1980) and Kahneman, Slovic, and Tversky (1982) have docu­
mented several biases among people making judgments about others.
These biases arise, in part, because of various “heuristics” we use to
overcome our inability to process all relevant information.
Communication scientists have developed some evidence on limita­
tions of social systems. Allport and Postman’s (1947) classic studies of
rumor transmission concluded that messages sent in sequence through a
communication system become distorted. Bartlett’s (1932) demonstra­
Chaffee, Berger / What Communication Scientists Do 113

tions of memory distortion in serial transmission suggests memory limi­

tations are a primary cause. But later authors (e.g., Shibutani, 1966)
questioned the serial nature of rumor transmission, asserting that rumors
are frequently generated by groups trying to construct an explanation for
events they cannot otherwise explain. He argues that when people do not
get adequate information from the mass media about causes of various
events, they will concoct a plausible explanation. Rumors arise, in this
view, not because an original story gets distorted as it passes from per­
son to person, but because people combine their individual explanations
to form an overall story. Coleman (1957) notes the limits of the media,
when rumor and slander circulate rapidly during a community contro­
versy. The information-processing limitations of organizations and
institutions have also been studied (March & Simon, 1958).
technologies. Communication scientists are highly alert to the
introduction of new communication technologies, which have major
impact on the capacities of communication systems to process large
amounts of information. Widespread use of computers in institutional,
organizational, and individual settings expands the abilities of these sys­
tems to process, store, and retrieve messages. Current emphasis in dis­
cussions of computer technologies is on the quantity of information that
can be handled. Less attention has been paid to the issue of information
quality (Lester, 1981). Similarly, promoters of cable television stress
the number of channels available rather than programming quality.
Some communication scientists become so enamored with the quan­
titative aspects of new technologies that they overlook more important
questions connected with their widespread adoption. There is not much
healthy skepticism concerning the future impact of these technologies.
For example, advertising for home computers emphasizes that tasks
such as balancing a checkbook and finding recipes and telephone num­
bers can be facilitated by home computers. But these mundane tasks can
usually be handled faster, more easily, and more economically using
what might in reaction be called “low technology” (e.g., pocket calcu­
lators, index cards). The home computer may still be a technology in
search of uses (and gratifications) for most consumers, despite the
enthusiasm of some academics.
The expanding memory capacity of computers raises issues of indi­
vidual and institutional privacy. With the increased use of computers by
government and commercial agencies, many groups have become con­
cerned about the kinds of information the government collects, security
of stored information, and possible invasive uses. Instances of home
computers being used to “break in” to data storage systems raise issues
about the status of information stored in government, hospital, and busi­
ness computers. As personal use of computers expands beyond playing
computer games, which was the main early use of home computers
114 OVERVIEWS

(Rogers, 1986), problems of social management proliferate. So do

negative societal side-effects that stem from socioeconomic inequalities.
System constraints. At each level of analysis there are other con­
straints on communication systems that limit their capacities to achieve
certain goals. At the individual level, the way human sense organs are
arrayed determines how well people can adapt to changes in their envi­
ronments. Because our eyes are set the way they are and because we do
not have a very well-developed ability to localize sound (i.e., to deter­
mine its direction), it is difficult for us to respond in a discriminating
manner to events that occur behind us. People can make considerably
finer adjustments in their actions when stimuli are within their view and
when both the eyes and the ears can help to locate things.
At the dyadic and social network levels, there are additional con­
straints too. People meeting for the first time are expected to converse in
pleasant ways rather than unpleasant ones. In dating relationships, indi­
viduals may be especially careful to observe norms of politeness and
social appropriateness in order to make themselves attractive. Confor­
mity to these conventions can prevent people from knowing one another
as individuals (Jones & Davis, 1965). One of the few things conformity
tells an observer is that the person is like many others; it thus prevents
leakage of information that might give one insight into the other’s per­
sonality. Social norms and rules may make interactions in public places
more congenial, but these interactions may be vacuous when it comes to
learning about the people in them.
A similar problem occurs in the context of interactions in formal
organizations. When people perform organizational roles, those roles
may be functional for task accomplishment but dysfunctional for emo­
tional satisfaction or the development of friendships. This is a problem if
people expect work not only to provide them with material rewards but
with socioemotional rewards as well. The same difficulties occur in
interactions between parents and children; parental role demands (e.g.,
supervision, training) may preclude the parents from some behaviors
(e.g., play) they might want to share.
At the level of social institutions there is at least one rather glaring
institutional constraint that makes achievement of goals difficult. The
American public has demanded progressively more services from gov­
ernment at all levels. Various sectors of the public have looked to gov­
ernment to solve social problems concerned with education, health
care, poverty, and the like. But even with massive infusions of tax
money, government cannot ameliorate many of these problems. Most
likely, the government could not solve these problems no matter how
much money it has. It is not so much a matter of size (“big govern­
ment”) as it is the organizational structure of government agencies and
the “top-down” problem-solving approach it fosters. Perhaps a struc­
Chaffee, Berger / What Communication Scientists Do 115

ture that encourages a “bottom-up” approach to problem solving would

be more satisfying. It has been noted that government authorities who
live in the environs of Washington, D.C., have little awareness of issues
confronting other social classes and regions of the country. Some com­
munication scientists think this isolation might be remedied by strategic
use of new communication technologies, which could enable more peo­
ple to have direct access to political elites (e.g., Siune & Kline, 1975).
System change. Natural communication systems evolve over time, but
some changes are more rapid and dramatic than others. Human anatomy
might very gradually evolve so that the eyes would be repositioned to
increase the width of the visual field, or the internal structure of the eye
might be altered. But technologies could much more quickly be devel­
oped to compensate people for evolutionary shortcomings. Examples
include optical instruments that allow people to see for long distances in
the dark, or devices to project voices far away (amplifiers).
In contrast to the slow pace of human anatomy and physiological evo­
lution are the sometimes rather rapid changes in people’s beliefs, atti­
tudes, and behaviors. One of the most extensively researched areas in
communication has been persuasion. The early genres of persuasion
research were cast at the individual level; almost all the studies viewed
influence as a one-way process in which a source induced change in indi­
vidual audience members. It has become apparent to communication
scientists that persuasion can be more fruitfully studied in the context of
interpersonal influence. Most probably, a majority of persuasive attempts
occur in situations where the person being influenced can resist the
influence and try to exert counterinfluence. While the early persuasion
research paradigms concerned themselves with resistance issues, little
attention was paid to counterinfluence. Investigators of communication
at the relational level (Ericson & Rogers, 1973; Millar & Rogers, 1976)
have explicitly incorporated counterinfluence into their coding schemes.
Despite considerable interest in effects of organizational and institu­
tional changes on productivity and satisfaction, little attention has been
paid to organizational and institutional change itself. Like individuals,
large formal social entities undergo both evolutionary and revolutionary
changes. Some are purposely induced and some are unintentional. Com­
munication scientists seek to find conditions that produce these changes.
That is, what variables are responsible for producing change and what
conditions trigger these causal variables? We also need to understand
how and why changes are resisted by individuals, groups, organizations
and institutions. System goal conflicts are important to change and resis­
tance, but information-processing capacity, constraints, and structure
are also involved.
Outcome evaluation issues. A critical issue in public communication
is “communication effectiveness.” Whether one is concerned with the
116 OVERVIEWS

individual or the institutional level, communication practitioners are

deeply concerned with promoting effective communication. This notion
is variously offered as a panacea for marital problems, voter apathy, and
corporate image difficulties. Criteria of effectiveness obviously vary
with these different situations. Effects sought in communication may be
to inform, to persuade, to be liked, to resolve a dispute, and so forth. A
journalist or teacher may be “effective” if their efforts are informative,
while a salesperson must sell, a mediator must resolve disputes, and a
public health campaign must reduce morbidity and mortality. Most
organized communication programs are undertaken to achieve many
proximal, intervening, and ultimate mission-determined goals (see Rog­
ers & Storey, Chapter 26). Evaluation research, the measurement of
communication effectiveness, usually involves matching the criterion
measures to this complex set of pragmatic goals. Communication scien­
tists often shy away from these “mundane” questions. But this attitude
may be doing their research enterprise a disservice in the name of intel­
lectual purity. There are many examples of healthy collaborations
between “pure” physical sciences and engineering. Communication
science has yet to develop an identifiable cadre of “communication
engineers,” but mission-oriented agencies see communication problems
at the root of many social evils and seek solutions.
Who should be the judge of communication effectiveness? Some argue
that judgments of communication experts or critics should be employed,
rather than the impact a communication has on its audience. In this view,
the message and its presentation would be judged effective or not on its
own merits. The alternative view, which is shared by most communica­
tion scientists, is that communication should be evaluated in terms of
goal achievement. Messages that accomplish their goals are effective;
those that do not are ineffective. Most people appear to understand the
importance of effective communication; perhaps what the public does
not understand are the conceptual complexities that arise when effec­
tiveness indicators are examined critically. It is a responsibility of com­
munication scientists to analyze these complexities and develop theories
and research programs that examine them in detail.
Designing communication systems. Given that communication sys­
tems have goals, can systems be designed to assure that their goals will
be realized in an efficient manner? We noted earlier individuals often
process information in ways that produce judgmental biases. Nisbett
and Ross (1980) suggest that one way to ameliorate these information
processing deficits is to train people in statistical inference. (Unfortu­
nately, elsewhere in their book they cite evidence that persons with
extensive training in statistics are almost as likely as naive subjects to
make inferential errors.)
Chaffee, Berger / What Communication Scientists Do 117

At the interpersonal level, there is some interest in redesigning vari­

ous kinds of relational arrangements. The 1970s saw considerable atten­
tion to redefinition of marriage arrangements (O’Neill & O’Neill, 1972)
and gender relationships. Kramarae (1981) argues that because males
were primarily responsible for development of language, most languages
contain biases that prevent females from expressing themselves as they
wish. Some minor linguistic changes have been effected, but languages
are notoriously resistant to significant change.
Organizations have a longstanding concern for design of “optimal”
communication systems to achieve goals. In businesses, the profit motive
has a lot to do with this interest, but it is also true that effectiveness
criteria are less ambiguous in organizations than at the individual or
interpersonal levels. McClelland (1961), in his work on achievement
motivation, pointed out that money earned is a rather unambiguous
index of success, whereas emotional outcomes of interpersonal relation­
ships are considerably more difficult to evaluate.
There has been some interest in redesigning certain institutions to
optimize their performance. Proposals for reorganization of the federal
government are often made, and there has been significant change in
religious institutions because of decreases in church membership across
many denominations. But threat of extinction may be necessary before
some institutions will initiate internal modification. In any event, com­
munication scientists are often involved in redesign of communication
systems at all levels.

Synthetic Issues Among Levels

Our previous discussion has dealt with questions within each level of
analysis; we now consider relationships between and among levels. We
are not advocating a reductionist position that argues for the explanatory
primacy of any one level of analysis. For example, we would reject the
position that all other levels can ultimately be explained by individual
cognitive processes or, at the other extreme, by social structure. But in
some instances one level may offer a better explanation of a communica­
tion phenomenon than does another level. Research at one level is some­
times more satisfactorily explained by theory at another level.
Interlevel effects. A number of questions can be raised regarding inter­
level effects. For example, we can ask how interactions between indi­
viduals and institutions influence individuals’ cognitive development, as
in Newcomb’s (1947) classic Bennington College study. Research has
demonstrated how individual attitudes and cognition can be changed
by different organizational milieux (Lieberman, 1956). Large formal
organizations consist of interrelated smaller work groups, and these
118 OVERVIEWS

smaller units have an impact upon the entire organization; in turn, organi­
zationwide actions influence each group.
For many of the phenomena communication scientists study, events at
one level of analysis may be at least partially explained by recourse to
events at another level of analysis. But at this point in the history of com­
munication science it is difficult to say that any one level is most likely to
provide satisfying explanations. Our more modest hope is that com­
munication scientists will become sensitive to the influences that phe­
nomena at each level can exert.
Transcendent principles. Finally, let us consider the search for prin­
ciples that operate in a similar fashion across all levels. This kind of
“horizontal” integration of levels is akin to the basic objective of general
systems theory (GST; von Bertalanffy, 1968). While the vocabulary of
systems has made a strong impact upon communication science (e.g.,
Monge, 1973,1977), actual implementations of systems approaches are
rare. The systems approach has been at most a stimulating metaphor.
This does not, however, mean that the global objective of GST is not a
reasonable one to pursue in communication science. We could well
explore such questions as, Are there similarities between the ways indi­
vidual cognitive systems are organized and the structure of formal
organizations? Both models of human cognition and models of formal
organization stress that superordinate units exert influence over subordi­
nate units. Given this isomorphism, do the systems operate in any simi­
lar ways? Rokeach (1960) discusses the organization of belief systems
that are open to discrepant information from the environment and those
which are less permeable. We should expect to find structural parallels
between the openness of belief systems and the openness of groups,
organizations, and institutions to outside information.
It is most unlikely that we could find a communication scientist who
would undertake to test such a relationship across all four levels. More­
over, given the current structure of communication science, even if indi­
vidual researchers were investigating our hypothesized relationship at
each level, they would probably be unaware of one another’s research
efforts, or would not recognize the similarities. We have in the later sec­
tions of this chapter shifted somewhat from describing what communi­
cation scientists do to discussing what we feel they ought to do in order to
advance the discipline. We have continually stressed the different levels
because few communication scientists are used to thinking across levels.
There may indeed be principles that transcend our analytic levels, and
which have already been verified by research. But the relative isolation
of research literatures may prevent us from discerning these transcen­
dent generalizations.
Chaffee, Berger I What Communication Scientists Do 119

CONCLUSION

We have addressed a number of issues surrounding the conduct of

communication research and the building of communication theory. Not
all communication scientists will subscribe to the assumptions we have
presented, the issues we have raised, or our blandishments to consider
multiple levels of analysis. The concept of communication science is not
yet an established, operating reality. We have noted many more disconti­
nuities between levels of analysis than commonalities. Still, we have
been able to locate and describe the heart of what appears to be an emer­
gent, unified discipline. In addition to a central core of assumptions and
issues, a field of study must also have its outer boundaries. Not all com­
munication phenomena are studied by scientific methods. Our purpose
here has been to outline the work of those who are within the scientific
tradition.

REFERENCES

Abelson, R. P. (1976). Script processing in attitude formation and decision making. In

J. S. Carroll & J. W. Payne (Eds.), Cognition and social behavior. Hillsdale, NJ:
Lawrence Erlbaum.
Abelson, R. P. (1981). Psychological status of the script concept. American Psycholo­
gist, 36, 715-729.
Allport, G. W., & Postman, L. J. (1947). The psychology of rumor. New York: Henry
Holt.
Anderson, J. R. (1976). Language, memory, and thought. Hillsdale, NJ: Lawrence
Erlbaum.
Anderson, J. R., & Bower, G. H. (1973). Human associative memory. Washington,
DC: V. H. Winston.
Bales, R. F. (1950). Interaction process analysis. Reading, MA: Addison-Wesley.
Bartlett, E C. (1932). Remembering: A study in experimental and social psychology.
Cambridge: Cambridge University Press.
Becker, L., & Doolittle, J. (1975). How repetition affects evaluations of and informa­
tion seeking about candidates. Journalism Quarterly, 52, 611-617.
Berger, C. R. (1977). The covering law perspective as a theoretical basis for the study
of human communication. Communication Quarterly, 25, 7-18.
Berelson, B. R., & Steiner, G. A. (1964). Human behavior. Orlando, FL: Harcourt
Brace Jovanovich.
von Bertalanffy, L. (1968). General systems theory. New York: Braziller.
‫ة'آلأ‬١0‫ة‬,‫آل‬١. M. ‫اع‬9‫ه<؟‬١. Conceptualization and measurement in the social sciences.
Newbury Park, CA: Sage.
Blumler, J. G., & Katz, E. (1974). The uses ofmass communications. Newbury Park,
CA: Sage.
Bower, G. H., Black, J. B., & Turner, T. J. (1979). Scripts in memory for text. Cogni-
tive Psychology, 11, ١٦٦-1.
Chaffee, s. H. (1971). Television and adolescent aggressiveness (overview). In
G. Conk&.١٩n٠٢i ٠٠٠١, Television and social behavior: Vol. Ill,Tele­
120 OVERVIEWS

vision and adolescent aggressiveness (pp. 1-34). Washington, DC: Government

Printing Office.
Chaffee, S. H. (1978). Presidential debates—are they helpful to voters? Communica-
tion Monographs, 49, 330-346.
Cohen, B. (1963). The press and foreign policy. Princeton, NJ: Princeton University
Press.
Coleman, J. (1957). Community conflict. New York: Free Press.
Cushman, D. p. (1977). The mles perspective as a theoretical basis for the study of
human connmunication. Communication Quarterly, 25, 30-45.
Cushman, D. p. & Craig, R. T. (1976). Communication systems: Interpersonal impli-
cations. In G. R. Miller (Ed.), Explorations in interpersonal communication. New-
bury Park, CA: Sage.
Ericson, p. M., *Rogers, L. E. (1973). New procedures for analyzing relational com-
munication. Family Process, 12, 245-267.
Farace, R. V., Monge, p. R., & Russell, H. M. (1977). Communicating and organiz-
ing. Reading, MA: Addison-Wesley.
Festinger, L. (1957). A theory ofcognitive dissonance. Stanford, CA: Stanford Univer-
sity Press.
Goffman, E. (1969). Strategic interaction. Philadelphia: University of Pennsylvania
Press.
Golding, R, & Murdock, G. (1977). Theories of communication and theories of soci-
ety. Communication Research, 5, 339-356.
Greenberg, B. s. (1974). Gratifications of television viewing and their correlates for
British children. In J. G. Blumler & E. Katz (Eds.), The uses ofmass communica-
tions. Newbury Park, CA: Sage.
Harre, R., & Secord, p. F. (1973). The explanation of social behavior. Totowa, NJ:
Littlefield, Adams & Co.
Harvey, o. J., Hunt, D. E., & Schroder, H. M. (1961). Conceptual systems and per-
sonality organization. New York: John Wifey.
Hempel, c. G. (1965). Aspects ofscientific explanation. New York: Free Press.
Iyengar, s., Peters, M., *Kinder, D. (1982). Experimental demonstrations of the "not
so minimal” consequences of television news programs. American Political Science
Review, 76, 848-858.
Jones, E. E., *Davis, K. E. (1965). From acts to dispositions: The attribution process
in person perception. In L. Berkowitz (Ed.), Advances in experimental socialpsy-
chology (Vol. 2, pp. 219-266). New York: Academic.
Kahneman, D., Slovic, p., *Tversky, A. ( 1982). Judgment under uncertainty: Heuris-
tics and biases. Cambridge: Cambridge University Press.
Kaz, E., & Iazarsfeld, p. F. (1955). Personal influence. New York: Free Press.
Klapper, J. T. (1960). The effects ofmass communication. New York: Free Press.
Knapp, M. L., Hart, R. p, & Dennis, H. s. (1974). An exploration of deception as a
communication construct. Human Communication Research, 7, 15-29.
Kramarae, c. (1981). Women and men speaking. Rowley, MA: Newbury House.
Iazarsfeld, p. F., Berelson, B., & Gaudet, H. (1948). The people’s choice. New York:
Columbia University Press.
‫ةعا‬١‫ًاج‬,‫لمأ‬.‫ذأ‬. )١1, ‫ج؟‬١‫ه‬٢‫ةئ‬٢ ١ . Embedding network constructs in Q theoreticalframe-
work: A preliminary formulation of a model of intraorganizational communication
behavior. Paper presented at the Western Speech Communication Association, San
Jose, CA.
Lieberman, s. (1956). The effects of changes in roles on the attitudes of role occupants.
Human Relations, 9, 385-402.
March, J. G., & Simon, H. A. (1958). Organizations. New York, Wiley.
McClelland, D. c. (1961). The achieving society. New York: Van Nostrand.
Chaffee, Bergei I What Communication Scientists Do 121

McCombs, M., & Shaw, D. (1972). The agenda-setting function of mass media. Public
Opinion Quarterly, 36, 176-187.
McLeod, J., Becker, L., & Byrnes, J. (1974). Another look at the agenda-setting func­
tion of the press. Communication Research, 1, 131-166.
Millar, F. E., & Rogers, E. (1976). A relational approach to interpersonal communica­
tion. In G.R. Miller (Ed.), Explorations in interpersonal communication. Newbury
Park, CA: Sage.
Miller, G. A. (1956). The magical number seven plus or minus two. Psychological
Review, 63, 81-97.
Miller, G. R., deTurck, M. A., &Kalbfleisch, P. J. (1983). Self-monitoring, rehearsal,
and deceptive communication. Human Communication Research, 10, 97-117.
Monge, P. R. (1973). Theory construction in the study of communication: The system
paradigm. Journal of Communication, 23, 5-16.
Monge, P. R. (1977). The systems perspective as a theoretical basis for the study of
human communication. Communication Quarterly, 25, 19-29.
National Institute of Mental Health. (1982). Television and behavior: Ten years ofsci­
entificprogress and implicationsfor the Eighties. Rockville, MD: National Institute
of Mental Health.
Newcomb, T. M. (1947). Some patterned consequences of membership in a college
community. In T. M. Newcomb, E. L. Hartley, et al. (Eds.), Readings in social
psychology (pp. 345-357). New York: Henry Holt.
‫دلحجدا‬., &loss,i. (1‫ا‬١. Human inferences: Strategies and shortcomings of
socialjudgment. Englewood Cliffs, NJ: Prentice-Hall.
Norman, D. A. (1981). Categorization of action slips. Psychological Review, 88, 1-15.
O’Neill, N., & O'Neill, G. (1972). Open marriage: A new life stylefor couples. Phila-
delphia: Lippincott.
Palmgreen, p., & Clarke, p. (1977). Agenda-setting with local and national issues.
Communication Research, 4, 4١S-4Sl.
Parsons, T. (1955). Family structure and the socialization of the child. In T. Parsons &
R. F. Bales, Family, socialization and interactionprocesses. New York: Free Press.
Planalp, s., & Hewes, D. (1982). A cognitive approach to communication theory:
Cogito ergo dico? In M. Burgoon (Ed.), Communication yearbook 5. New Bruns-
wick, NJ: Transaction Press.
Yo۴‫؟‬,Y..K. (١96٦١. Conjectures and refutations: The growth ofscientific knowledge.
Idon: Routledge & Kegan Paul.
Rokeach, M. (1960). The open and closed mind. New York: Basic Books.
Rogers, R. 1 ‫ا‬. )‫ل‬١. Communication technology: The new media in society. ‫جال‬١‫يل‬
York: Free Press.
Rogers, E. M., & Agarwala-Rogers, R. (1976). Communication in organizations.
New York: Free Press.
Schank, R. C., & Abelson, R. P. (1977). Scripts, plans, goals and understanding.
Hillsdale, NJ: Lawrence Erlbaum.
Schroder, H. M., Driver, M. J., & Streufert, S. (1967). Human information process­
ing. New York: Holt, Rinehart & Winston.
Shibutani, T. (1966). Improvised news: A sociological study of rumor. New York:
Bobbs-Merrill.
Siune, K., & Kline, F. G. (1975). Communication, mass political behavior, and mass
society. In S. Chaffee (Ed.), Political communication (pp. 65-84). Newbury Park,
CA: Sage.
Taylor, S. E., & Fiske, S. T. (1978). Salience, attention, and attribution: Top of the
head phenomena. In L. Berkowitz (Ed.), Advances in experimental social psy­
chology (Vol. 11, pp. 249-288). New York: Academic Press.
122 OVERVIEWS

van Dijk, T. A. (1980). Macrostructures. Hillsdale, NJ: Lawrence Erlbaum.

van Dijk, T. A., & Kintsch, W. (1983). Strategies of discourse comprehension. New
York: Academic Press.
Ward, S., Wackman, D., & Wartella, E. (1977). How children learn to buy. Newbury
Park, CA: Sage.
Weaver, D. (1977). Political issues and voters’ need for orientation. In D. Shaw &
M. McCombs (Eds.), The emergence ofAmerican political issues. St. Paul, MN:
West.
Webb, E. J., Campbell, D. T., Schwartz, R. D., & Sechrest, L. (1966). Unobtrusive
measures: Nonreactive research in the social sciences. Chicago, IL: Rand McNally.
Wilensky, R. (1983). Planning and understanding: A computational approach to
human reasoning. Reading, MA: Addison-Wesley.
Wright, C. (1959). Mass communication: A sociological perspective. New York: Ran­
dom House.