ADV EXPERIMENTAL SOCIAL PSYCHOLOGY,VOL 3

Visit the link below to download the full version of this book:

https://medidownload.com/product/adv-experimental-social-psychologyvol-3/

Click Download Now

 CONTRIBUTORS

Numbers in parentheses indicate the pages on which the authors’ contributions begin.

ROBERTP. ABELSON,Depnrfnienf of Psychologi.?:Yole University, New

Haven, Connecticut ( I )

MICHAELARGYLE,Institute of Experimental Psychology, Oxford University,

Oxford, England ( 5 5 )

N. T. FEATHER,Department of Psychology, University of New England,

New South Wales, Australia (99)

IRVING L. JANIS, Department of Psychology, Yale University, New Haven,

Connecticut ( 166)

ADAM KENDON,’Institute of Experimental Psychology, Oxford University,

Oxford, England ( 5 5 )

SERGE MOSCOVICI,Ecole Pratique des Hautes Etudes, Universite‘ de

Paris, Paris, France (225)

PERCYH. TANNENBAUM, Mass Communications Research Center, University

of Wisconsin, Madison, Wisconsin (27 I )

‘Present address: Department of Psychology, Morrill Hall, Cornell University,

Ithaca, New York.
V
This Page Intentionally Left Blank
 PREFACE

This volume will report much that is new, and perhaps even surprising,
to many of its readers. For one thing, although social psychology is often
regarded as an exclusively American science, especially by Americans,
four of the six chapters were written by citizens of other nations: Argyle
and Kendon, Feather, Moscovici, and Tannenbaum (although both Feather
and Tannenbaum received their doctorates in this country and Tannenbaum,
fortunately for us, has remained in the United States). American social
psychologists typically confine their professional reading to American
journals to the neglect of the important non-American publications in
their field. This book, hopefully, will contribute to a greater, and multi-
directional, exchange of social-psychological information across national
boundaries.
Then too, if we are to judge from the problems receiving the greatest atten-
tion we might think social psychology dealt primarily with cognitively
mediated defensive or autistic processes. The chapters by Argyle and
Kendon and by Moscovici can help us to broaden our professional outlook.
Analyzing focused social interactions in terms of the performance of social
skills, Argyle and Kendon show how the individual’s actions toward other
people are regulated by a wide variety of determinants, including such
factors as the physical distance between the people, the reason they have
come together and the setting they are in, and the speech characteristics of
the other persons. Moscovici’s paper concentrates on the communicative
aspects of social interaction. Demonstrating a truly impressive cross-
national scholarship (as do Argyle and Kendon), Moscovici calls on social
psychologists to give greater attention to the interplay between social
relationships and language behavior. His chapter introduces a relatively new
(for many of our readers) and rapidly developing field of study and, more
than this, reports a number of important investigations conducted both in
Moscovici’s own laboratory and elsewhere.
To say there is more to social psychology than self-defense and cognitive
consistency, however, is not to deny the role of cognitive and defensive
processes in human behavior. Three papers in this volume are concerned
with such processes, and testify to the continuing interest in them. Like the
Argyle and Kendon and Moscovici chapters, they deal with communic-
ations, but this time are addressed to the person’s reactions to information
received from other people. Feather presents a general model derived from
balance-theory considerations for analyzing these communication effects.
vii
 ...
Vlll PREFACE

In addition to summarizing his own research, he provides us with a very

comprehensive survey of the relevant literature in this area. Tannenbaum’s
chapter is focused on the problem of inculcating resistance to persuasive
communications. Developing his ideas from the congruity version of the
cognitive consistency formulations, he reports the results of an extensive
research program and also points out the connection between his findings
and those obtained by McGuire (as given in his chapter in Volume 1 of this
serial publication). Janis also discusses resistance to persuasion attempts:
the resistance sometimes created by fear arousal. As many of our readers
know, Janis and Feshbach had found, in one of the best-known studies
conducted by the Yale Attitude Change Project, that fear-arousing com-
munications can produce a “defensive-avoidanceyyof the communicator’s
arguments. Later investigations, however, have not always yielded the
same results, and Janis advances a theoretical model designed to reconcile
these apparent contradictions.
Now that I have referred several times to general theoretical schemes,
let me repeat a comment I made in the Preface to Volume 1. The papers in
this serial publication will typically try to provide a theoretical integration
of findings in a particular research area rather than an exhaustive literature
survey. These theoretical generalizations give meaning to the facts accumul-
ated by investigators. Equally important, they can also stimulate the further
research endeavors which will lead, inevitably, to their demise and replace-
ment by better formulations. Most of the theoretical models with which we
are familiar in social psychology consist entirely of verbal statements.
As our science matures, however, theoretical generalizations will increas-
ingly be couched in mathematical terms. Abelson’s chapter in this volume
discusses the present status of mathematical models in social psychology
and can serve as a very helpful introduction to subsequent developments
along these lines.

LEONARD BERKOWITZ
September, 1967
 CONTENTS
CONTRIBUTORS.................................................................. v
PREFACE.......................................................................... vii
CONTENTSOF OTHERVOLUMES ................................................ xi

Mathematical Models in Social Psychology

Robert P. Abelson
I . The Nature and Role of Mathematical Models ........................... 1
I1. Mathematical Models of Social Structure ................................. 7
111. Mathematical Models of Social Behavior ................................. 25
References .................................................................. 49

The Experimental Analysis of Social Performance

Michael Argyle and Adam Kendon
I . Introduction ................................................................ 55
11. Theoretical Framework .................................................... 56
I11. Structure of the Social Performance ...................................... 62
IV. Self-presentation and Competence ........................................ 80
References .................................................................. 91

A Structural Balance Approach to the

Analysis of Communication Effects
N . T. Feather
I. Introduction ............................................................. 100
I1. Balance Theory .......................................................... 101
111. The Communication Model ............................................ 107
IV . Studies Based upon the Communication Model ....................... 109
V. Other Relevant Evidence ............................................... 134
VI . Some Critical Problems ................................................. 146
VII . Concluding Remarks .................................................... 158
References ............................................................. 159
ix
X CONTENTS

Effects of Fear Arousal on Attitude Change:

Recent Developments in Theory and Experimental Research
Irving L. Janis
I . Introduction ............................................................... 167
I1. Theoretical Framework ................................................... 169
111. Experimental Findings .................................................... 187
IV . Theoretical Model for Analyzing Determinants of the Optimal
Level of Arousal .......................................................... 203
V . Discussion of the Model .................................................. 210
References ................................................................. 222

Communication Processes and the Properties of Language

Serge Moscovici
I . Language as a Field of Social Psychology ............................... 226
I1 . Theoretical and Experimental Observations of Linguistic
Behavior ................................................................... 237
111. Some Open Questions .................................................... 262
References ................................................................. 265

The Congruity Principle Revisited: Studies in the Reduction.

Induction. and Generalization of Persuasion
Percy H . Tannenbaum
I. Introduction .............................................................. 272
I1 . Strategies for the Reduction of Persuasion ............................. 275
111. Combinations of Congruity Strategies .................................. 283
IV . Mechanisms of the Refutation Treatment .............................. 291
V. Persuasion via Congruity and Information Processes .................. 304
VI . Generalization of Attitude Change ...................................... 310
VII . Concluding Remarks ..................................................... 316
References ................................................................ 318
AUTHORINDEX .......................................................... 321
SUBJECTINDEX .......................................................... 330
 CONTENTS OF OTHER VOLUMES

Volume 1
Cultural Influences upon Cognitive Processes
Harry C.Triandis
The Interaction of Cognitive and Physiological Determinants of
Emotional State
Stanley Schachter
Experimental Studies of Coalition Formation
William A . Gamson
Communication Networks
Marvin E. Shaw
A Contingency Model of Leadership Effectiveness
Fred E. Fiedler
Inducing Resistance to Persuasion: Some Contemporary Approaches
William J. McGuire
Social Motivation, Dependency, and Susceptibility to Social Influence
Richard H. Walters and Ross D. Parke
Sociability and Social Organization in Monkeys and Apes
William A . Mason
AUTHOR I N D E X S U B J E C T INDEX

Volume 2
Vicarious Processes: A Case of No-Trial Learning
A lbert Bandura
Selective Exposure
Jonathan L . Freedman and David 0. Sears
Group Problem Solving
L. Richard Hoffman
Situational Factors in Conformity
Vernon L. Allen
Social Power
John Schopler
From Acts to Dispositions: The Attribution Process in Person Perception
Edward E. Jones and Keith E. Davis
Inequity in Social Exchange
J. Stacy Adams
The Concept of Aggressive Drive: Some Additional Considerations
Leonard Berkowitz
AUTHOR I N D E X S U B J E C T INDEX
xi
This Page Intentionally Left Blank
 ADVANCES IN

EXPERIMENTAL
SOCIAL PSYCHOLOGY

VOLUME 3
This Page Intentionally Left Blank
 MATHEMATICAL MODELS IN SOCIAL
PSYCHOLOGY

Robert P.Abelson
DEPARTMENT OF PSYCHOLOGY
YALE UNIVERSITY
NEW HAVEN, CONNECTICUT

I. The Nature and Role of Mathematical Models . . . . . . . . . . . . . . . 1

A. Introductory Considerations ..................... 2
B. The Measurement Problem . . . . . . . . . . . . . . . . . . . . . . 3
C. Historical Background ........................ 4
D. Applications in Social Psychology . . . . . . . . . . . . . . . . . . . 5
11. Mathematical Models of Social Structure . . . . . . . . . . . . . . . . . 7
A. Description of Properties of Sociograms . . . . . . . . . . . . . . . . 7
B. The Balance Principle . . . . . . . . . . .. . . . . . . . . . . . . 13
C. Large Social Structures ....................... 19
TIT. Mathematical Models of Social Behavior . . . . . . . . . . . . . . . . . . 25
A. Social Modification of Opinions . . . . . . . . . . . . . . . . . . . . 25
B. Social Influence on Decisions, Choices, or Actions . . . . . . . . . . . 39
References ................................ 49

I. The Nature and Role of Mathematical Models

The phrase “mathematical model” strongly connotes the ultimate in
precise scientific analysis. It has sometimes been argued that social psycho-
logy has not sufficiently advanced theoretically to reap the benefits of precise
mathematical analysis, yet at the same time it is widely granted that theore-
tical progress is critically dependent on the organizing power that mathe-
matical models can provide. This chapter is intended as a contribution to
the continuing bootstrap operation by which the level of mathematization
of social psychology is gradually being raised.
‘This chapter was written while the author was a beneficiary of the atmosphere and
facilities of the Center for Advanced Study in the Behavioral Sciences.

1
2 ROBERT P. ABELSON

A. INTRODUCTORYCONSIDERATIONS
Mathematical models are used in the physical sciences as a technique for
forming abstractions from reality, manipulating the abstractions, and
applying the results again to reality to encompass other consequences. It is
necessary for the abstracting principle to be precisely specified and to be
coordinated with essential features of the reality under consideration. The
power of the technique depends on the flexibility with which the abstraction
can be manipulated in concert with other relevant abstractions.
Newton’s law of gravitation provides a simple (if overworked) example.
The property of mass is the essential feature of physical bodies appropriate
to gravitational force. Newton’s law specified the size of the gravitational
force between two bodies solely in terms of their masses and the distance
between them. Together with the mathematical tools of the differential
calculus and another abstract law relating force to acceleration, a complete
mathematical model is formed enabling the calculation of the orbits of
natural and artificial bodies. For present purposes, one thing is especially
noteworthy about this example.
A great many properties of physical bodies (such as color, volume,
hardness, and temperature) do not appear in the mathematical model. The
property of mass is an “abstraction” in the sense not only of being mathe-
matically symbolizable but also by virtue of all the other properties of the
real objects that are ignored. It is sometimes supposed that mathematical
models cannot usefully be constructed in the social sciences because human
behavior is too complex to be represented mathematically. This argument
seriously underestimates the very high level of abstraction necessary to the
construction of mathematical models, even in the physical sciences. One
never tries to represent the objects of study literally, only very limited aspects
of them. Of course, the difficulty lies in the selection of essential aspects.
Newton would not have got very far with a gravitational law dependent on
the color rather than the mass of physical bodies.
There are two quite distinct reasons for the potential scientific super-
iority of mathematical over purely verbal descriptions: the greater precision
of mathematical descriptions, and their greater manipulability. Precision is
clearly a crucial feature of models applied to physical situations as, for
example, in the calculation of orbits. In applications to social situations,
however, precision is less important whenever present difficulties in mea-
surement and in the control of extraneous variables render imprecise the
empirical testing of models. But even if models are used only in the service
of generating very approximate data fits, the greater manipulability of ma-
thematical models is a very powerful advantage. Sometimes this advantage is
difficult to achieve because mathematical derivations run into complica-
 MATHEMATICAL MODELS IN SOCIAL PSYCHOLOGY 3

tions, but in that case recourse may often be had to computer-aided numer-
ical methods or to the technique of computer simulation.
Computer simulation models rcpresent an interesting alternative to
mathematical models. They also involve rigorous mathematical specifica-
tions of processes, but in place of analytic derivation of the consequences
of these specifications, the model is “set in motion” on the computer, supply-
ing a symbolic imitation of the processes being modeled. This alternative
has the advantage of broadening the range of mathematical and empirical
possibilities that can be handled; on the other hand, it requires consider-
able effort and the outcomes do not have the neat, compact generality of
analytic mathematical solutions. Computer simulation models in social
psychology are reviewed in great detail by Abelson (1967).
B. THE MEASUREMENT PROBLEM
It is a difficult task to abstract essential theoretical variables for ex-
plaining given aspects of human behavior, but the problem of measuring
these variables often seems even more forbidding. There are two broad
classes of potential measures, based on counting and scaling, respectively,
and the problems associated with the two are quite different.
The operation of counting whether and when particular qualitative
responses have occurred offers no problems of philosophical principle. In
practice, however, counting often leads to probability models that are tricky
to analyze mathematically and test empirically. Common examples of qua-
litative responses in social psychological models are sociometric choices,
occurrences of communicative acts, and choices of alternative responses in
two-person games. Coleman (l964a), in his excellent treatment of mathe-
matical models in sociology, recommends more extensive use and develop-
ment of models based on counting, especially in view of the increasing
availability of data gathered from large survey samples.
If the mathematical modeler wishes to use quantitative quasi-continu-
ous measures appropriate for sensitive mathematical treatment, such as by
differential equations, he must be prepared to deal with the issue of the
nature of measurement in the social sciences. The question usually raised
(cf. Stevens, 1951) concerns the justification for applying the properties of
the real-number system to the numbers associated with “arbitrary” em-
pirical procedures, such as the assignment of scores on an attitude scale.
One constructive stance on this question is pragmatic: the appropriat-
eness of any given assignment procedure is judged in terms of the neatness
and usefulness of its theoretical and empirical consequences rather than by
a priori considerations. Fortunately, there is recent mathematical evidence
(Abelson and Tukey, 1963; Shepard and Carroll, 1966; Guttman, 1967)that
a number of seemingly arbitrary scaling procedures are much more robust
4 ROBERT P. ABELSON

in practice than might have been supposed when Stevens and others origi-
nally advanced their warnings.
What measurement procedures in the social sciences have lacked has
been a notion of fundamentality as powerful as N. R. Campbell’s (1928)
concept of “fundamental measurement” in the physical sciences. A recent
breakthrough has occurred however, foreshadowed by suggestions of
Adams and Fagot (1959) and culminating in the publication of the Luce and
Tukey (1964) paper on “conjoint measurement.” Those who are not satisfied
with the rough-and-ready pragmatic approach to measurement problems
will now find available much firmer axiomatic foundations on which to
build. Space does not permit an extended treatment of the matter here but,
in brief, the basic ideas are as follows.
The axiomatic system for Campbellian measurement rests on the
operation of “concatenation” of properties of distinct objects, as when two
sticks are placed end to end and the length of the concatenate is found to
equal the sum of the lengths of the individual sticks. It has been extraordi-
narily difficult to imagine how one might concatenate psychological pro-
perties, however. What Luce and Tukey propose is an alternative operation
intrinsic to a different, but no less fundamental, axiomatic system: the oper-
ation of “conjoining” the levels of two separate independent variables. To
conjoin level i of variable A with leveljof variable B requires the application
of these variable levels simultaneously to an experimental unit on which
some effect X r j is assessed. This is exactly the operation typically employed
in two-way factorial design. Conjoint measurement on X (and on A and B)
is achieved if it turns out that additivity of the independent variables obtains;
that is, if over all levels iandjthere are no interaction effects ofA and B on X .
The detailed relationship of this conceptual system to specific empirical
cases has not been worked through, but there has been a burst of recent work
on several aspects of conjoint measurement (Luce, 1966; Tversky and
Zivian, 1966; Krantz, 1964; Roskies, 1965; Tversky, 1967; Goode, 1964;
Kruskal, 1965). Against this background of axiomatic ferment the practical
man, of course, continues to construct his models, even as in the physical
sciences the mathematical applications ran far ahead of Campbell’s attempt
to rationalize the measurement process.
C. HISTORICAL
BACKGROUND
The early applications of mathematical models to psychology were in-
dividual pioneering efforts (in this country primarily associated with the
name of L. L. Thurstone) that did not meet with ready acceptance by the
discipline at large. Miller (1964) has provided a good summary of some of
these early attempts, which fell mainly in the areas of psychophysics and
mental test theory. In the 1950’s mathematical learning theory was launched