A STUDY 0F ENTERNATiONAL THREA?

Thesis for the flame or! M. .43..
MECHEGAN STA‘E‘E UNE‘4’ERSEW
J'a‘é‘an (10%, fir.

W635

IHEIIB

LIBRARY

Michigan Stave
University

 

ABSTRACT
A STUWY OF INTERNATIONAL THREAT

by John Cobb, Jr.

This is a study of sources of international threat as perceived
by people of different nationalities. It attempts to establish certain
correlates of international threat in terms of country characteristics
which include ideology, size of pOpulation and per capita wealth.

It is assumed here that nations may be defined as organizations,
i.e., that the citizens of any nation are interconnected by channels of
communication and that they exhibit certain regularities of behavior. One
of these regularities is assumed to be a high degree of consensus cone
cerning which other nations are threatening or dangerous. Perceived in-
ternational threat is defined here as a mechanism which controls the
growth of an international community of nations.

A sample of 24 countries was selected which was stratified by
ideology (Communist and non-Communist), population (large and small),
and level of per capita energy consumption (rich, medium and poor). These
three country characteristics-~plus the nationality of the subject-~were
hypothesized to be related to how threatening a country is judged to be.

The sample of subjects included peOple of five different
nationalities. Except for a special subsample of experts in international
relations, the sample was made up of graduate students at Michigan State
University. Nationality subsamples were made up of Four persons each,

and the entire sample consisted of only 2n peOple. With such extremely

John Cobb, Jr.

small and non-random subsamples, no claim can be made that findings are
representative of the five nationalities studied (Brazilians, Egyptians,
Indians, Formosan Chinese and Americans).

To operationalize international threat subjects were asked to
reSpond to a 21-point scale marked "safe/not threatening" at one end
and "dangerous/ very threatening" at the other. Subjects made an X on
this scale for each of the 2n countries to indicate how threatening they
judged it to be. These responses were coded as scores of 0—20 with 20
representing maximum threat.

These threat scores were analyzed in various ways. A Q factor
analysis of these sets of reSponses produced a five-factor solution in
which American, Indian and Brazilian "types" of responses were discernable,
plus two Egyptian "types." This offers an indication that the nationality
of the respondent was of importance--(83% of the variance was accounted
for)--in defining which countries are judged as threatening and not
threatening. It also suggests that despite the lack of sampling pro-
cedures, the fact that three of the nationality subsamples are indicated
to be different "types" of people lends some substance to the hope that
nationality differences and similarities have been tapped.

The analytic technique used to test most hypotheses was a factorial
analysis of variance. For this sample-owhich included subjects from
"neutralist" countries such as Egypt and India--Communist countries were
seen as more threatening than non-Communist countries, and large countries
were perceived to be more threatening than small countries, regardless of
ideology. The nationality of the subject again proved to be a strong effect

influencing judgements of international threat.

John Cobb, Jr.

The per capita wealth or energy consumption of a country was also
found to influence judgements about threat, but its effects were far more
complex than those hypothesized. This variable interacted with all of
the others in the analysis, and severely restricted the statistical in-
terpretation of the analysis of variance of threat scores. In general,
poor countries were seen as most threatening, rich countries as next
most threatening and countries of intermediate wealth as least threatening.

Among the three country characteristics examined, ideology was
found to have the strongest influence upon threat judgements. But the

nationality of the subject was an even stronger influence.

A STUDY OF INTERNATIONAL THREAT

by

John Cobb, Jr.

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF ARTS

Department of Communication

1964

ACKNOWLEDGEMENTS

I wish to express my appreciation to my advisor, Dr. Malcolm S.
MacLean, Jr. for guidance and assistance with this study from planning
'to presentation; to Dr. Eugene Jacobson, Dr. Randall P. Harrison and
Dr. John T. McNelly for constructive criticism and encouragementt to
Albert Talbott for assistance with the analysis of variance design and

for supervision of data processing; and to Mrs. Shirley Sherman for typing

this thesis.

ii

CHAPTER

I

II

III

IV

TABLE OF CONTENTS

INTRODUCTION: A Theoretical Exploration of the
Concept Of Threat e e e e e e e e e e e e e e e 0

Individual Organization . . . .

Perception of Threat

Stress and Strain e e e e e e e e

Perceived Threat in an International Context

VARIABLES AND HYPOTHESES

Threat, Size and Wealth
Myths and Realities . .

Nationality as a Variable

Hypotheses

O O O O O 0

Threat and Nationality of

O

O

O O O 0

Respondent

O
O
O

O O O

0

"Myths and Realities" Operationalized
Threat and Perceived Similarity . . .
Threat and world Government . . .

PROCEDURES AND METHODS OF

IdeolOgy
POPUlatiOn e e e e 0
Health and Energy . .
The Sample of People

Analytical Procedures
Othar Analyses e e e
Criteria Data . . . .
World Government . .

O O O O O 0

O O O O O O O O

RESULm O 0 O O O O O O 0

Analysis of Variance of
Ideology
POPUlfltiOn e e e e e e
IdeolOgy and Population
Energy 0 e e e e e e e
Ideology and Energy . .
Population and Energy .

ANALYSIS

0 O O O O O O O

Threat Scores

0 O O O O

O
O
o
e
O
O

Ideology, Population and Energy

Nationality Subsamples

Energy and Nationality

Ideology, Energy and Nationality
Population, Energy and Nationality

Some Additional Findings

iii

0 O O O O O

O O O O O 0

Ideology of Country and Nationality
Population of Country and Nationality

Multiple Correlations . . . . . . .

O
O
O
0
O
O
O
O

of Subject

0 0 O
O O O
O O O
O O O O
I O O O

O O C O O O O O
O O O O O O O 0
O O O O O O O O
O O O O O O O O
O O O O O O O O
O O O O O O O O

O O O O C O O O
O O O O O O O O
O O O O O O O O
O O O O O O O O
O O C O O O O O
O O O O O C O O

O O O O O O O C O
C O O O O O O O O
O O O O O O O O O

O O O O O O O O
O O O O O O O O O

of Subject

0 O O O O
O O O O O

O
O
O
0

so...

0 O O O O O O O O O O O

O O O O O O O O O O O O O O O O

O O O O O O O O O O O O

O

O O O O O O O O

O O O O O O O O O O O O O O O 0

Page

10
13
16

19

22
23
25
27
29
29
31
32

33

3“
35
36
39
“2
an
“5
“6

#7

52
5k
55
56
57
60
61
63
66
68
7O
73
75
77
83
85

CHAPTER

TABLE OF CONTENTS

QANCIYSISeeeeeeee
Similarity e e e e a e e 0
Correlations with Criteria
World Government . . . . .

V DISCUSSION . . . . . . . . 0

REFERENCES .

BIBLIOGRAPHY O O O O 0 Q 0 Q 5 5 .

APPENDIX A:

Ideology . .

POPUlation e e a e e e e 0
Energy a a e e e e e e e
Nationality a e e e e e e

Questionnaire e a e e e 0

iv

(continued)

0 O O O
O O O O
O O O O
O O O O
O O O O
O O O O
O O O O
O O O O

O O O O
O O O O
O O O O
O O O O
O 0 O O
O O O O
O O O O
O O O O

O O O O

O O O O

O O O O

O O O O

O O O O

Page

86
90
91
93

96
97
100
102
106

115

117

121

TABLE

8a

8b

9a

9b
10a
10b

lla

11b

12a

12b

13.

lua
lab

15

LIST OF TABLES

The Country Sample with Criteria Data. . . . . . .
Age and Field of Study of Subjects . . . . . . . .
Average Threat Scores for Subjects and Countries .

Per Capita Energy Consumption Averages for
SUbjCCt. .nd CCUDtPICQ o o o o o o o o s o o o o 0

Average Population Estimates for Subjects and
Countries o o o o e o o a o a o s o a s s s o o o

Average Ideology Estimates for Countries . . . . .
Summary of Four Analyses of Variance . . . . . . .

Ideology: Average Throat Scores for Communist and
non-Communist Countries o o o o o s o o s o o o o

“‘1“ Ideology EffCCt o o o e o o a o o s o s o 0

Population: Average Throat Scores for Large and
5m.11 countrIO‘ o o o e o o o o o o o o o o o o o

M‘In POPUl‘tIUD Bff‘Ct o e s o o e o o o o o s o o
Ideology and Population: Average Throat Scores. .
Ideology-Population Interaction . . . . . . . . .

Energy: Average Throat Scores for Exporimontor-
Classified High, Medium and Low Energy Countries .

Main Enorgy EffOCt o s s o o o o o o o o o o o o o

Enorgy: Average Throat Scores for Subject-Classified

High, Hedium and Low Energy Countries- . . . . . .

Main Energy Effect, Subject-Classified Analysis .

Experimenter Classification of Countiros by Per
Capita EUCng CODIUIPti.n e s e o e o o s o s o s

Ideology and Energy: Average Threat Scores . . . .
Ideology-Energy Interaction . . . . . . . . . . . .

Ideology and Energy: Average Threat Scores, Subject-
Clasaified An‘lysis o s o o o o o o s o o o s o o a

V

Page
38

“0

Q7

49

SO
51

53

SH

55

55
S6
56

56

57

S7

58

58

59
60

60

61

LIST OF TABLES (continued)

TABLE Page
16a Papulation and Energy: Average Threat Scores . . . . . 61
16b Population-Energy Interaction . . . . . . . . . . . . 61
17 Countries Experimenter Classified by Energy Level

and POpulatiOn o o e e e e e e e e e e o e o e e e e o 62
18 Population and Energy: Average Threat Scores,

Subject-Classified Analysis . . . . . . . . . . . . . 62
19a Ideology, Population and Energy: Average Throat

Scores e e e e e e e e e e o e e e e e o e e e e e e o 63
19b Ideology-Population-Energy Interaction . . . . . . . . 63
20 Countries Experimenter Classified by Ideology,

Population and Energy e e e e e o e o e e e e e e e e 6“
21 LOVGIS Of Perceived Threat e e e e e e e e e e o e e o 65
22a Ideology, Population and Energy: Average Threat

Scores, Subject-Classified Analysis . . . . . . . . . 66
22b Ideology-Population-Energy Interaction, Subject-

CIESBifin Analysis e e e o e o o e e e e o e e e e e 66
23a Nationality: Average Threat Scores for Six Subsanples. 67
23b “310 Nationality EffOCt e e e e e e e e e e o o e o e 67
2ua Ideology and Nationality: Average Threat Scores . . . 68
2% Ideology-Nationality Interaction . . . . . . . . . . . 68
25a Ideology and Nationality: Average Threat Scores,

Subject-Classified Analysis e e e e e e e o e e o e e 69
25b Ideology-Nationality Interaction, Subject-

C13831fied AhllYSia e e e e e e e e e e o e e o e e e 69
26a Population and Nationality: Average Threat Scores . . 70
26b Population-Nationality Interaction . . . . . . . . . . 70
27a Ideology, Population and Nationality: Average

Threat Scores e e e e o o e e o e e e e e e e e e e e 72

27b Ideology-PopulationcNationality Interaction . . . . . 72

vi

TABLE
28a
28b

29a

29b

30a
30b

31a

31b
32a
32b
32c
33

3k

35
36

37

38

39

NO

1:1

“2

LIST OF TABLES (continued)

Energy and Nationality: Average Threat Scores . . . .
Energy-Nationality Interaction . . . . . . . . . . .

Energy and Nationality: Average Threat Scores,
SUbjGCt'CICSSified Analy313 e e e e e e e e e e o e e

Energy-Nationality Interaction, Subject-
ClflS'ifiCd AnalySis o e o e e e e e e e o e e e o e e

Ideology, Energy and Nationality: Average Threat Scores

Ideology-Energy-Nationality Interaction . . . . . . .

Population, Energy and Nationality: Average Threat

Scores . . . . . . . o . . . . . . . . . . . . . e e
Population-Energy-Nationality Interaction . . . . . .
Pour-factor Interaction . . . . . . . . . . . . . .
Pmfihamefihm:MwnemmuS«ms....
Summary, Experimenter-Classified Analysis . . . . . .
Differences Between Two Analyses . . . . . . . . . .
Intro-Individual Correlations Between Sets of Threat
Scores and Sets of Ideology, Population and Energy

Estimates . . . . . . . . . . . . . . . . . . . . e
Multiple Correlations: Three Predictors of Threat . .

Q Analysis: Rotated Factor Loading: . . . . . . . . .

More and Less Threatening Countries for Nationality

TYPO: e a a e e e e o e e e e e e o e e e e e e e o o

Correlations of Similarity Estimates with Threat and
Ideology Estimates e e e o e e e e e e e e e e e o e

Correlations of Subjects' Estimates and U.N. Criteria

Data e e e e o e e o e e o e e e e e e o e e e e e o

Classification Scores by Country Characteristics . .

0

Mean Threat Scores for Subjects Classified by Responses

to World Government Statements . . . . . . . . . . .

Subsample Means for world Government and Threat Scores

vii

Page
73

7“

7a

75
76

76

77
77
78
80
81

82

8k
85

87

90

92

93

94
9M

CHAPTER I

INTRODUCTION

A Theoretical Exploration of the Concept of Threat

Note: This chapter examines various sources of theory and

research in an effort to provide a broad theoretical field

in which perceived international threat may be defined and

studied. A communication model derived from cybernetic theory

is proposed.

The variables and hypotheses of immediate theoretical

concern in this study are discussed in Chapter II.

The purpose of this study is to investigate the relationship be—
tween perceived threat and danger represented by a given country or class
of countries and the nationality of the perceiver. There will also be an
effort to ascertain the relationship between degree of perceived threat
and certain country characteristics such as national population, per capita
wealth, and political ideology.

In this study, threat is the dependent variable of major concern.
How is threat to be dealt with in theoretical terms? Efforts to study
threat have been made at various levels of generality. Several of these
levels will be examined here.

The most abstract efforts to include threat in scientific theory
have been in the field of cybernetics. Here efforts have been made to
formulate theory which deals with communication and control in a wide
diversity of systems--astronomical and chemical as well as biological

and social. The notion of threat is used in this body of theory in terms

of defining the elements of an on-going system or organization. This

study is concerned with three levels of organizations: (1) individuals,
(2) nations, and (3) groups or blocs of nations, so this theoretical
approach should be of some value here. A

In dealing with organisations in general, cybernetic theorists
make several axiomatic assumptions. In The_Human Use of Human Dermal,
Norbert Wiener says, "...certain organisms, such as man, tend for a time
to maintain and often even to increase the level of their organization...
The process by which we living beings resist the general stream of
corruption and decay is known as homeostasis... The oxygen and carbon
dioxide and salt in our blood, the hormones flowing from our ductless
glands, all are regulated by mechanisms which tend to resist any untoward
changes in their levels. These mechanisms constitute what is known as
homeostasis, and are negative feedback mechanisms." (1)

Neiner goes on to say, "It is the pattern maintained by this
homeostasis which is the touchstone of our personal identity... We are but
whirlpools in a river of overflowing water. No are not stuff that abides,
but patterns that perpetuate themselves." (2) In this approach, both
personality and culture may be conceptualized as patterns in the process
of transmission which change as they are constrained by communication
with other patterns.

H. Ross Ashby, another cybernetic theorist, suggests that external
threats and disturbances are not necessary for defining organization,
p=£_22, but that they are necessary in defining a viable organization
or an organism which acts so as to further it's own survival. Ashby says
that "there is not any property of an organization that is 'good' in any
absolute sense; all are relative to some given environment, or to some

given set of threats or disturbances, or to some given set of problems.” (3)

' He also says that "the organization (whether of a cat or an automatic
pilot or an oil refinery) is judged 'good' if and only if it acts so as to
keep an assigned set of variables, the 'essential' variables, within assigned
limits." (fl)

Here an effort will be made to develop a communication model of the
international situation in terms of Ashby's cybernetic theory. It will
be conceptualized as a situation in which previously isolated, independent
systems become interconnected in a larger system. ”An example," of such
a situation, says Ashby, "is the embryo nervous system, which starts
with cells having little or no effect on one another, and changes, by
the growth of dendrytes and formation of synapses, to one in which each
part's behavior is very much affected by the other parts." (5)

The proposed communication model would be one which starts with
nation-states having little effect upon one another, and changes, by the
growth of communication to one in which each country's behavior is very
much affected by that of the other countries.

In this view it is the establishment of increasingly large communi-
cation channels and flow of messages between countries which has replaced
relative independence with conditionality and imposed new constraints
upon the behaviors of countries.

Here it is assumed that to the extent that channels of international
communication exist, to that extent international organization exists
in that constraints have been imposed upon what were relatively in-
dependent parts. (Of course, it is not the formal, conscious, planned
sort of international organization that is being discussed here.) In

these terms, organization need only be defined as conditionality between

parts and regularity of behavior. A further implicit assumption is the
one already implied in the definition of homeostasis: systems or or—
ganizations tend toward states of equilibrium.

In this theoretical context, threat has no intrinsic qualities,
but, rather, expresses a relationship. First, an organization must be
defined as a number of interdependent parts: the behavior of each part
is constrained by the behavior of other parts. Each part has a number
of psosible states: each variable has a number of values. The laws or
regularities of the system are such that certain values of states are
preferred over others. A threat, then, is any element or event which
tends to drive the outcome away from the preferred values of the system's
variables.

For a mammal, body temperature must be maintained within certain
limits. Anything which tends to push body temperature too high or too
low represents a threat to the organism. For a business organization,
income and expenses must be kept in balance. Anything which pushes income
below expenses represents a threat to the organization.

To describe the proposed model of international relations further,
two more concepts must be introduced: boundary maintenance and systemic
linkage. The tendency of a system to maintain its independence and
relative isolation from other systems will be called boundary maintenance.
The tendency for a system to increase its communication and thus its
interdependence with other systems will be called systemic linkage.

An extreme example of boundary maintenance in the international
sphere is the present 0.5. policy toward Cuba. An example of systemic
linkage is the change of status of Hawaii and Alaska from territories to

states. Other examples of systemic linkage are the installation of the

”hot line" between Washington and Moscow, and the proposed commercial
flights between New York and Moscow.

In general, an increase in communication and trade indicates
systemic linkage, and an absence, or a reduction in these things indicates
boundary maintenance. Here it will be assumed that if a country is
perceived as threatening or dangerous, a policy of boundary maintenance
will be carried out. If a country is not seen as threatening or dangerous,
the potentiality of increased communication and interdependence exists.

In this view threat is a negative feedback mechanism which results
in the reduction of systemic linkage.

Such an idea is neither new nor novel. In a book published in
lens, two English anthropoIOgists discussed social change in terms of
equilibrium and disequilibrium:

"The positive or cultural forces of social change are new
ideals, ideas and intuitions of beauty, and the discovery of

new uses for material resources, of new facts and of new

techniques...

"The negative or structural forces are intollerable

opposition between groups and categories of peOple that

threaten, unless some change is quickly made, to destroy

law, logic and convention. These are the forces of dis-

equilibrium." (6)

Godfry and Monica Wilson go on to say that, "So long as foreigners
are regarded as dangerous enemies to whom one has no obligations, trade
or other cooperation with them is impossible." (7)

Some insight into the nature of international isolation and
independence as contrasted with the interdependence created by communi-

cation with other countries is offered by the following discussion of

Tibet by the current Dalai Lama:

"Perhaps the best known quality of Tibet in the recent

past was its deliberate isolation. In the world outside,
Lhasa was often called the Forbidden City. There were two
reasons for this withdrawal from the world. The first, of
course, was that the country is naturally isolated. Until
the last decade, the route from the border of India or
Nepal to Lhasa was a journey of two months across high
Himalayan passes which were blocked for a large part of the
year. From my birthplace in the borderland between Tibet
and China the journey to Lhasa was even longer, as I have
already told-~and that borderland itself was over a thousand
miles from the sea coast and ports of China.

Isolation was therefore in our blood. we increased our

natural isolation by allowing the fewest possible foreigners

into our country, simply because we had had experience of

strife, especially with China, and had no ambition whatever

except to live in peace and pursue our own culture and religion,

and we thought to hold ourselves entirely aloof from the world
was the best way of ensuring peace. I must say at once that

I think this policy was always a mistake, and my hope and

intention is that in the future the gates of Tibet will be

kept wide open to welcome visitors from every part of the

world." (8)

Here a change from a policy of isolation to a policy of increasing
external communication is suggested. Visitors from fereign countries
were judged as dangerous: now they are considered safe. A policy of
boundary maintenance has been abandoned in favor of a policy of systemic
linkage.

These opposite sorts of behavior are discernable in many kinds of
systems. In molecular systems with high mutual attraction like
molecules stick together and expose the least possible surface areas to
foreign contact. An example is a ball of mercury. In molecular systems
of high capillarity, there is a strong attraction for alien molecules:
water spreads out and clings to glass.

Nations in this century rarely assume that the total external

environment is threatening and maintain the extreme policy of isolation

pursued by Tibet. More often such a social system discriminates some

' elements in-its environment which are relatively safe or benign from
other elements which are relatively threatening or dangerous. And it
usually has in its repertory of responses both adaptive and defensive
reactions to unfamiliar or changed aspects of the environment.

From the theoretical viewpoint which has been developed here,
threat or danger posed by other countries provides negative feedback
which restrains and controls the growth of a global international system
made up of nation states.

It is also possible with such a model to use another unit of
analysis, and consider the "Communist bloc" and the ”non-Communist world"
as organizations or systems. There is certainly more communication and
commerce within these groups of countries than between them.

In terms of the model developed here, the degree of organization
existing in the non-Communist world should be indicated in consensus
among individuals in non-Communist countries that Communist countries are
threatening. However, not as much organization would be expected at
this level of analysis as would be expected when nations are the units
of analysis.

Stated in general terms, the model deveIOped here assumes that
personalities, national cultures and larger patterns of culture may all
be conceptualized as patterns in the process of transmission which maintain
stability by homeostasis. They change and modify themselves if, and
only if, they are constrained by communication with other patterns. Thus
both stability and change are taken into account in this communication
model at various levels of organization.

(It is interesting to note that in describing A ngss-Polity Survgy,

Banks and Textor state that the basic concept used is "pattern.") (9)

As communication imposes conditionalities or constraints upon the
behavior of each individual, he may be considered as an interdependent part
of a system. Thus the clan, tribe, feudal community and nation may be
considered systems to the extent that the communication net imposes
constraints upon the behavior of its members.

The socialized individual finds that to maintain his essential
biological variables in equilibrium, it is necessary to maintain
equilibrium within one or more larger social systems of which the in-
dividual is a constrained or interdependent part.

The present age of nationalism may be defined in terms of the
formalized boundary maintenance which takes place at national borderS.
This suggests that national communication networks impose constraints
upon the attitudes of their members concerning external threats. In
the terms used here, one index of national organization would be consensus

concerning a pattern of external threats.

Individual Organization

It is assumed here that individual and social change is always
mediated by communication. How does this operate at the level of
individuals?

Theodore Newcomb sees interpersonal communication as an ABX system
in which each message modifies the state of the two-person system in
the direction of symetry. In terms of what has been said previously,
this "process of co-orientation" can be viewed as change in a two-person

social system which is mediated by communication.

Just what happens when a new and relatively novel message is
received? The following description is in the nature of a hypothesis
which owes much to the work of D. E. Berlyne. In most general terms the
problem will be case as the introduction of a new element of information
into a system, and the consequent reorganization of the system.

A healthy organism devotes part of its time and energy to environ-
ment scanning or information seeking behavior. (Natural selection has
provided some assurance that the sensory apparatus is appropriate for
attending to external elements which are potential problems, disturbances
or threats to survival or well being.) The sense data or new information
which filters into the brain is checked against old information which
has been gathered during the past and stored in the memory. The result
of the check between sense data and memory data is a same/different
reaction.

When the check between new and old infermation yields a ”different”
response, the healthy individual focuses the senses on the new or un-
expected or changed element in the perceptual field. Since the individual
has no learned responses or attitudes concerning the new element, there
are conflicting response tendencies. The conflict brought about in the
self by new information which is contrary to expectations about the
environment may have a number of outcomes.

An "exploratory drive" may impel the individual to orient itself
toward the uncertain stimulus for maximum stimulation from it, to approach
it, and examine it. There is some evidence that unfamiliar objects or
events bring about exploratory behavior on the part of healthy organisms.

(This is something like the capillarity of water mentioned earlier in

10

which there is a strong attraction for alien molecules.)

The first thing that happens during the ”orientation reaction"
is a scanning-rescanning by the appropriate senses in an effort to increase
information and reduce uncertainty. Pavlov called this reaction to a
perceived change in the environment the "what is it?" response. Berlyne
calls it the orientation reaction. He points out that the same external
conditions may bring about a reaction of curiosity, fear, adaptive

behavior, or even laughter.

Perception of Threat

The next step in the process is an evaluative reaction of safe]
dangerous. If the evaluative reaction is "dangerous," than a fight/flight
or attack/retreat decision is made and appropriate behavior is carried
out to maintain the physical and psychological integrity of the in-
dividual.

If the evaluative reaction is "safe," than a number of choices are
opened including more information seeking and another evaluation. It
would seem that a likely sequence of action in response to new information
or a novel stimulus is information seeking, tentative "safe" evaluation,
more information seeking, another safe/dangerous evaluation, etc. This
sort of behavior is exemplified in the gingerly approach to an uncertain
stimulus. As long as the safe/dangerous evaluation remains tentative,
approach/avoid/approach behavior is likely. From the subjective view-
point of the individual, an attitude, orientation or policy has not yet
been decided upon.

An evaluation of "dangerous" brings about a defensive reaction,

while an evaluation of "safe" brings about an adaptive reaction. Both

11

adaptive and defensive reactions are generalized and pervasive
physiological states, and both act to counteract stimulation or in-
formation gathering, and reallocate available energy.

The defensive reaction has been thoroughly explored and fitted
out with terminology by Freudian psychologists. Ego-defensive behavior
includes denial, identification with the agreesor, repression, pro-
jection, etc. In Irving Sarnoff's definition, ego defense is a response
to intolerable fear to minimize incapacitating the ego's perceptual function.
He says that threatening events exert the greatest strain on the ego's
perceptual function. (10)

In Berlyne's terms the defensive reaction counteracts stimulation:
energy is diverted from the sense organs in favor of the musculature.

(In terms of a national system, this would be equivalent to allocating
less money to the diplomats and more to the military.)

One limitation of the intra-individual approach outlined above is
that it takes little note of the importance of other individuals in the
environment. Theodore Newcomb, in writing about the acquaintance process
as a prototype of human interaction suggests that between two given
people the process goes on indefinitely: "Participants, however familiar
with one another they may already be, acquire information about each
other, assess one another's attitudes, and either reinforce existing
states of orientation toward each other and toward the common world, or
change them, or develop new ones." (11)

This implies that perhaps the other persons and groups making up
the interpersonal environment of the individual should constantly be

considered as relatively novel stimuli in Berlyne's schema. Each new

12

encounter offers an opportunity for increasing information and reducing
uncertainty and for the evaluation of any information which is
classified as different from that received during previous encounters.

Newcomb sees the individual as confronted with the constant problem
of maintaining three kinds of equilibrium: (1) intrapersonal, (2) with
other individuals and groups, and (3) with the world that we have in
common with these persons and groups.

The communication input which provides the constraints within
and among these systems consists of (1) direct sensory experience, and
(2) the testimony of other people. As long as this communication input
is internally consistent, no question arises about the ”real" nature of
things. But when there is conflicting evidence perceived in communi-
cation input, then disturbing questions arise about "myths" and
"realities,” and conflicting response tendencies are activated.

"Thus," says Newcomb, "we become sensitive to the acquired, drive-
like state of strain" which, like other states of drive, may be tolerated
for a time at least, but whether endured or appeased, influences
behavior." (12)

Two kinds of tendencies in balance-promoting changes in orien-
tations discussed by Newcomb are the autistic and the realistic.

”Insofar as they are characterized by distortion of the testimony of
others in balance-maintaining fashion-~30 that one's orientations neither
toward those others nor toward the common world need be changed-owe

have referred to them as autistic. And insofar as balance is achieved

by changing one's own orientations (either attitudes or attractions)

rather than by distortions of others orientations, the adoptions, may,

l3
contrarywise, be regarded as realistic." (13)

Stress and Strain

This pressure toward equilibrium in interpersonal systems Newcomb
calls ”strain." "Human socialization necessarily includes the acquiring
of many attitudes that are consensual with others' (especially adults')
attitudes. To survive requires the acquiring of unfavorable attitudes
(like those of adults) toward possibly drowning water, or devouring
beasts, or macerating machinery. To adapt comfortably to a stabilized
family, community, or social order one must acquire the right attitudes
(attitudes like those of one's associates who have already adapted com-
forably to family, community, etc.) toward cabbages and kings and all
else that is culturally approved or disapproved. Socialization, in
short, includes the building of danger signals when attitudinal dis-
crepancies with trusted others are perceived." (13) These danger signals
provide negative feedback necessary for the maintenance of equilibrium
between an individual and a social system.

Richard S. Lazarus has studied a variable which he calls "stress."
"This anticipation of potential harm or motive thwarting is the key to
the concept of threat. I regard threat as the central intervening
variable in psychological stress." (1a)

In discussing the role of evaluation or "discriminating between
dangerous or threatening conditions and benign ones," Lazarus says,
"The process of appraising which circumstances are harmful and which are
benign is crucial to the production of stress reactions... It seems
likely that beliefs or expectations about events based both upon past

experience and the present stimulus configuration determine whether a

ll)

stimulus will be appraised as threatening." (15)

What has been labeled a safe/dangerous evaluation in previous
pages is called "cognitive appraisal" by Lazarus. "The concept of
cognitive appraisal implies that the same stimulus can be threatening or
not, depending upon the interpretation the person makes concerning its
fUture personal significance... The threat is not simply an attribute of
the stimulus; rather it depends for its threat value on this appraisal,
which in turn depends upon the person's appraisal of the meaning of the
stimulus fer the thwarting of motives important to him." (16)

In an experimental situation in which stress reaction was indexed
by skin conductance, Lazarus demonstrated that the stress reaction to
the same visual event could be significantly modified by varying
simultaneously presented verbal commentary (i.e., cognitive appraisal).
His conclusion was that the defensive verbal messages--constructed along
lines specified by the theory of ego-defense--significantly reduced the
threatening impact of a visual event (a motion picture).

Lazarus has demonstrated that "the same stimulus may be threatening
or not, depending upon the manner in which it is interpreted." (17)
Perhaps some individuals habitually interpret their environment as being
more threatening. Or in Berlyne's terminology, they may habitually make
use of a defensive reaction.

John A. Hammes has found that highly anxious individuals rate en-
vironmental stimuli higher on a "danger-egression-threat" continuum than
less anxious individuals. He operationalized high and low anxiety by
Taylor Manafest Anxiety Scale items in multiple choice form on the

Heineman Forced Choice Anxiety Scale. Hammes concluded that his findings

15

support the view that ”manifest anxiety" may be considered as a per-
sonality characteristic. (l8)

Authoritarianism is a well-known concept related to dichotomizing
of complex issues, acceptance of authority figures and maintaining a
view of the world as hostile and threatening. Non-authoritarianismo-as
indexed by a low score on the F-scale--is associated with perceptiveness
of others. Low scorers tend to be more sensitive to interpersonal re-
lations among their peers. In Berlyne's terms they use their senses more
and their muscles less.

Newcomb and others have found that nonauthoritarians are more
accurate in estimating the attitudes of people with whom they interact.
And 0. J. Harvey has found that "higher authoritarianism was reflected
in tendencies ‘to form judgmental norms or concepts of novel stimuli
more quickly, as manifested in faster reaction time and less variability
in the judgments entering into the formation of a concept." (19) In
terms of the hypothetical cognitive circuitry described earlier, it
would seem that high F-scale scorers move from orientation to evaluation
more quickly. And the faster reaction time suggests that they are able
to allocate energy from the sensory apparatus to the musculature more
quicklyo-a change associated with the defensive reaction.

Harvey found that high F—scale scorers demonstrated shorter delay
in perception of movements and that "...authoritarianism disposes toward
faster and more rigid structuring of novel stimuli." He also found
that "authoritarianism disposes the individual toward increased closed-
ness of his conceptual system and hence toward warding off events that

deviate very far from his simple and narrow-banded interpretative schema."

1.6

This suggests that authoritarianism is related to autisim, and
this was validated by Newcomb. He found that in an interpersonal re-
lationship with peers, "the relationship between authoritarianism and
autisim increases with acquaintance, and may be reversed at early stages
of acquaintance if objects of orientation are important enough to induce

very strong forces toward balance." (20)

Perceived Threat in an International Context

The foregoing pages have been an effort to make a case for the
position that judgments of "safe/dangerous" or "threatening/not threatening"
about aspects of the social reality of an individual are not a randomly
selected attitudinal dimension but rather a fundamental evaluative
operation which takes place as a system, organization, organism or person
maintains equilibrium in a world which includes external threats,
disturbances and problems.

It has further been suggested how this conceptual scheme can be
fitted into a communication model which defines individuals, nations,
and blocs of nations as systems whose degree of independence of other
systems is defined by degree of communication with other systems.

Minimum communication with other systems equals maximum independence
of the system. Maximum communication between systems imposes maximum
constraints and conditionalities upon system behavior. This suggests
that growth is taking place and that the systems are becoming sub-systems
of a larger system as communication increases. It is thus suggested
that the current international situation can be defined in terms of

"levels of interdependence." (21)

l7

~Within the highly integrated nation-states of the mid-twentieth
century, the integration of the system is expressed in terms of many
constraints upon the behavior of the individuals, and among these
constraints are those which impose consensus concerning threats or dis-
turbances external to the national system.

Ross Stagner has taken a similar position in suggesting that
"man come to value his nation, or other social group, as an essential part
of his environment, and mobilizes energy to protect it. Further, as
a part of this process, he distorts the input of information in such a
fashion as to protect valued aspects of his social environment, and
these distortions contribute in no small degree to the intensity and
bitterness of social conflicts." (22)

Stagner says that ”for both theoretical and practical reasons,
we should focus on how members of groups perceive other groups, and
their goals and tactics," and that "social conflicts are rational if we
grant the accuracy of the way in which the participants perceive the
issues." (23)

In detailing the role of perception he says that "it can magnify
certain information inputs, giving them greater weight; and secondly,
the obvious collary, it can diminish the importance of other cues. Finally,
actual distortions may occur in quality and magnitude." (24)

In this study, Stagner's admonition to focus on how members of
groups perceive other groups will be heeded by utilizing a sample made
up of peOple of different nationalities. Although primary interest is in
perceived threat in terms of nations and groups of nations, the ex-

tremely small sample of 2“ people certainly precludes generalizing about

18

nationalities. Findings must be interpreted not as generalizations
but rather as explorations which indicate that further research may be

fruitful.

CHAPTER II

VARIABLES AND HYPOTHESES

The myth is that every Communist state is an un-
mitigated evil and a relentless enemy of the free
world; the reality is that some Communist regimes
pose a threat to the free world while others pose

little or none...

J. Hilliam Fulbright
"...the peOple whose decisions determine the
policies and actions of nations do not respond to

the 'objective' facts of the situation, whatever

that may mean, but to their 'image' of the situation."

Kenneth Boulding

"National Images and International Systems"

19

20

On Wednesday, March 25, 196“; Senator J. William Fulbright,
Chairman of the Senate Foreign Relations Committee, made a speech
before the Senate in which he said:

"It is not Communist dogma as espoused within Russia but
Communist imperalism that threatens us and other peOples of the non-
Communist world. Insofar as a great nation mobilizes its power and
resources for agressive purposes, that nation, regardless of ideology,
makes itself our enemy. Insofar as a nation is content to practice
its doctrines within its own frontiers, that nation, however repugnant
its ideology, is one with which we have no proper quarrel."

The study to be described here is concerned with the countries,
and the kinds of countries, which are perceived as threatening "us
and other peoples of the non-Communist world," and how this is related
to such things as power, resources and ideology.

Senator Fulbright believes that simply because a country has a
different ideology should not be sufficient grounds for judging it to
be threatening or hostile. In this study people of different nationalities
were asked to say to what degree 2n countries are dangerous or threatening
to the security of themselves and their families. They were also asked
to say which of the countries have governments which are Communist, and
which, non-Communist.

The correlation of the responses concerning threat and ideology
makes it possible to say to what degree each of the people questioned
agree with the view that the threat represented by a country is independent
of its ideology. Senator Fulbright does not argue that these two things
are independent, but rather that there is not a one-to-one relation

between the two. In the same speech he said:

21

"I believe that the Communist world is indeed hostile to the free
world in its general and long-term intentions but that the existence of
this animosity in principle is far less important for our foreign policy
than the great variations in its intensity and character both in time
and among the individual members of the Communist bloc. Only if we
recognize these variations, ranging from China, which poses immediate
threats to the free world, to Poland and Yugoslavia, which pose none,
can we hope to act effectively upon the bloc and to turn its internal
differences to our own advantage and to the advantage of those bloc
countries which wish to maximize their independence.”

Thus Senator Fulbright argues that it is important to discriminate
among Communist countries which represent different degrees of threat.
He then states as pertinent examples that China poses immediate threats
to the free world while Poland and Yugoslavia pose none.

In this study the respondents were asked to express just this kind
of judgments about a sample of 2a countires which includes China, Poland
Yugoslavia, and also Cuba, a country which Senator Fulbright mentioned
in the following terms:

"I think we are bound to conclude that Castro is a nuisance but
not a grave threat to the United States... Cuban communism does pose
a grave threat to other Latin American countries..."

The data from this study yields information about the degree
to which a subsample of Americans see Cuba as a threat and comparisons
may be made with the responses of a subsample of Brazilians. It will
then be possible to say to what degree the Americans and Brazilians

questioned agree with Senator Fulbright's view of the threat posed by

22

Cuba to Americans and Latin Americans, and further to compare these

responses with those of Egyptians, Chinese and Indians questioned. While

the samples are far too small to make generalizations about national
populations, the data will yield some indications of whether members of

a nationality make similar judgments about threat, ideology, etc. and whether
the variation in reSponses within a nationality subsample are as great

as the variation of the whole sample of five nationalities.

Threat, Size and Health

In the March 25 speech, Senator Fulbright did not go into great
detail about the relationship between the theat a country poses and its
power and resources. However, in discussing the situation in Panama
he did cite the per capita income of Panama ($u29) and the Canal Zone
($u,228) as evidence of "the profound social and economic alienation
between Panama and the Canal Zone, and its impact on the national
feeling of the Panamanians, that underlines the current crisis."

In this investigation, one hypothesis is that the wealth of a
country is related to how threatening it is judged to be. That is,
rich Communist countries will be seen as more threatening than poor
Communist countries. Among non-Communist countries it is eXpected that
the opposite relationship will prevail: rich countries will be judged to be
less threatening than poor countries.

One further point should be made about Senator Fulbright's dis-
cussion of the crisis over the Panama Canal. He described Panama as
"a small nation with a weak economy...” and stated, "I am unable to
understand how a controversy with a small and poor country, with

virtually no military capacity can possibly be regarded as a test of our

23

bravery and will to defend our interests."

The Senator evidently does not see Panama as much of a threat,
and part of the reason for this is that it is both small and poor. Here
it will be hypothesized that a country's size (i.e., national population)
as well as its wealth help to determine whether it is seen as posing a

threat to the welfare and security of the person responding.

Myths and Realities

The central theme of Senator Fulbright's Speech was the divergence
between "cherished myths" and "objective facts." He said:

"As long as our perceptions are reasonably close to objective
reality, it is possible for us to act upon our problems in a rational
and appropriate manner. But when our perceptions fail to keep pace with
events, when we refuse to believe something because it displeases or
frightens us, or because it is simply startingly unfamiliar, then the
gap between fact and perception becomes a chasm, and action becomes
irrelevant and irrational.

"There has always--and inevitably-obeen some divergence between
the realities of foreign policy and our ideas about it. This divergence
has in certain resPects been growing, rather than narrowing; and we are
handicapped, accordingly, by policies based on old myths rather than
current realities...

"We are confronted with a complex and fluid world situation and
we are not adapting ourselves to it. We are clinging to old myths in the
face of new realities, and we are seeking to escape the contradictions by
narrowing the permissable bounds of public discussion, by relegating an

increasing number of ideas and viewpoints to a growing category of

2N»

'unthinkable thoughts.'"

The issue raised by Senator Fulbright is a broad one, and only
certain apsects of its can be examined within the confines of this
study. However, some of the divergences between "perceptions" and
"realities" will be examined here.

For example, it is possible to ask a person to make estimates of
the pOpulations of a sample of countries and to compare these estimates
with data in official United Nations publications. A correlation
coefficient will tell how closely these two sets of data agree, and will
give an indication of whether the respondent rank ordered the countries
in the same way as the U.N. statisticians.

But there are difficulties in making such comparisons, and China
makes a good example. It is easy enough to ask someone to make an
estimate or guess of the population of China, and this was done in this
study. But if this is the "perception," then what is the "reality"?
China reports no figures to the United Nations, and any figure printed
will be another estimate-~a carefully considered one no doubt, but still
an estimate. There are other countries in the sample which have taken
no national census in recent years, and, here again, pepulation figures
will be estimates.

If there are difficulties in specifying the "reality" of which
countries are large and which countries are small, and which are rich
and which are poor, the difficulties increase when ideology is to be
discussed. Which countries have governments which are "very communist,"
"somewhat communist" or "non-communist"? There are no tables of numbers

to consult to discover whether a government is more or less communist.

25

The reSpondents in this study were asked to rate the governments
of 2% countries on a eight-point scale starting with "most non-communist"
and ending with "most communist," and they completed the task without
major difficulties.

Thus the "perception" is easy enough to elicit, but the "reality"
is harder to find. At this point, it seems best to abandon the reality-
perception distinction and only to specify whose perception or classifi-
cation is being discussed, and under what conditions and at what time
the estimate was made.

Each person participating in this study gave his own estimates
concerning the degree of threat represented by each of 2% countries and
also its population, per capita wealth (energy consumption) and ideology.
Since the responses of any one person do not agree 100% with any other
person in the sample, it might be argued that fragments of 2a different
perceptual worlds of 2a people are represented by the data. Or one
might say that 2a varying realities are represented.

At any rate, each of these 24 sets of responses may be compared

with a 25th set--data selected from current United Nations publications.

Nationality as a Variable
Two samples are under consideration in this study: a sample of
2a countries, and a sample of 2n people. Some characteristics of each
sample have already been mentioned.
The People were selected primarily on the basis of their
nationality. There are five nationality subsamples each made up of four
persons. (The sixth subsample of people was made up of international

relations experts.)

26

All respondents were adult male college graduates. Each person
in the sample had lived in the country where he was born until he was
at least 21 years old with no extended periods of residence in foreign
countries, and in all cases the parents of the respondent were of the
nationality of the country where the respondent was born. In other
words, the persons who participated in the study were all relatively pure
national types in that they had been born and raised in the country
where their parents had been born and lived.

One of the primary purposes of the study is to find out whether
people of different nationalities, when asked the same questions in
approximately the same time and place reSpond in significantly different

ways to questions about international threat.

In the preceeding pages, a discussion of the concept of threat
in systematic, international, interpersonal and intrapersonal terms has
drawn on a large body of theoretical and practical material. How can
this be utilized in a small-scale exploratory study? Heinz Eulau
suggests that:

"The most feasable alternative is to deal with modest
propositions that require simultaneous manipulation of only

a few variables, but to do so in a larger conceptual system

that, although it cannot be tested directly, serves the very

useful purpose of guiding an investigation and giving it

theoretical significance." (25)

If threat is to be the dependent variable of major concern, what
are the independent variables which will help to account for variation
in threat? Rosecrance has examined international politics from a

"systems" viewpoint derived from Ashby, and he provides a statement which

may assist in developing a few "modest propositions":

27

"The variety of the disturbance is in turn dependent

upon certain intra-actor variables. The number of Options

which an elite in control of an actor will exercise in the

international system is a function of its particular ethos

(that is, of the direction in which control is to be

exercised), its control of disposable resources, and the

quantity of disposable resources available." (26)

This statement will be freely interpreted along with the remarks
of Senator Fulbright to develop hypotheses concerning international
threat. Thus an effort will be made to take into account viewpoints
from both the practical and theoretical worlds of international politics.

Some hypotheses have already been suggested. They will now be

restated along with further hypotheses.

Hypotheses

The first two hypotheses suggest that there is a relationship
between perceived threat and the ideolOgy of the government of a country
and its total national pOpulation.

In a sample of peOple made up of Egyptians, Indians, Brazilians,
Formosan Chinese and Americans, (1) Communist countries will be perceived
as more threatening than non-Communist countries, and (2) large countries--
regardless of ideology-—wi11 be perceived as more threatening than small
countries.

The first hypothesis may be interpreted as expressing a relation-
ship between the ideOIOgy of the stimulus country and the ideology of
the respondents which are assumed to be non-Communist in varying degrees.

The second hypothesis is more directly related to a country
characteristic, size of population. Although it is also possible to
think of this as a relationship between the pOpulation size of the

country being judged and the pOpulation of the country of the person

28

responding, this is not implied in the hypothesis. Regardless of

the pepulation of the home country of the reSpondent it is hypothesized
that large countries will be seen as more threatening than small
countries.

The third hypothesis relates international threat to wealth or
what Rosecrance calls "the quantity of disposable resources available."
In this study this will be considered as an energy variable, and it will
be indexed as per capita energy consumption. It is assumed that such
an index taps the same sort of underlying characteristic that is indexed
by per capita gross national product. Edward S. Mason of Harvard
University made an international study of per capita energy consumption and
per capita gross national product, and concluded that "no country at
this state of history can enjoy a high per capita income without being
an extensive consumer of energy." (27)

The sample of countries utilized in this study will be divided
into equal subsamples of countries of high, medium and low per capita
energy consumption. The hypothesis is that there will be no significant
differences between the degree of threat represented by countries of
different energy levels, but that there will be an interaction between
energy level and ideology. Specifically, it is predicted that rich
Communist countries will be seen as more threatening than poor Communist
countries, but that rich non-Communist countries will be seen as 3555:
threatening than poor non-Communist countries.

No further hypotheses will be made about possible interactions
between ideolOgy, population and energy, but any significant interactions

among these variables will be reported and examined.

29

Threat and Nationality of ReSpondent

Although it is expected that there will be significant differences
in mean threat responses of nationality subsamples, there are no detailed
hypotheses. This aspect of the study is exploratory, and the general
prediction is that there will be significant differences in the mean
threat reSponses of nationality subsamples when countries are divided
by ideology, pepulation or energy level.

In line with the view of a country as an organization which
constrains the behavior of its citizens, it is also hypothesized that
each nationality subsample will produce a pattern of threat reSponses
for the country sample which will be different from that of the other
nationality subsamples. In quantitative terms, this is to predict that
inter-subject correlations of sets of threat responses will be significantly
higher within nationality subsamples than between them.

To summarize, it is predicted that perceived international threat
is related to certain country characteristics (ideology, pepulation

and energy level) as well as to the nationality of the respondent.

"Myths and Realities" Operationalized

One difficulty with the hypotheses suggested above is that a
country which, according to United Nations statistics, has a small
population may be considered by one or more subjects as having a large
population. For example, when the countries of the world are rank
ordered on the basis of their populations, Australia is clearly one of
the smaller countries; yet many peeple might judge it to be a country

with a large population.

30

This is the problem of "myths and realities" raised by Senator
Fulbright. In this study, this problem will be dealt with by using two
different criteria for making judgements about the characteristics of a
country of interest here. In one case, the experimenter will classify
the countries in the sample as to ideology, population and energy
level--largely on the basis of the latest available U.N. figures. In
the other case the subjects will classify the countries in the sample
by giving their own estimates or judgements about the ideolOgy, population
and energy levels of the countries.

Thus the same variables of classification will serve as the basis
for two classifications of threat reSponses, and two analyses. The
experimenter-classified analysis will provide a basis for generalizations
concerning specific countries with characteristics Specified by U.N.
statistics. The subject-classified analysis will provide a basis for
making statements, for example, about small, poor, non-Communist countries,
although subjects may not always agree which countries in the sample
have these characteristics. In both cases the dependent variable will
be the threat responses of each subject to each country in the sample.

The estimates or judgements of each subject about the ideology,
population and energy level of each country makes it possible to use
this limited view of the perceptual world of each subject to make intra-
subject examinations of the relation of perceived international threat
to perceived country characteristics. In this way it is also possible
to test the assumption that these variables are significantly related on

a subject by subject basis.

31

It is also possible to ask whether the relationship between
subject estimates of country characteristics and U.N. figures on country
characteristics help to explain the threat reSponses of an individual.
Perceived populations and energy levels of countries for each subject
will be correlated with U.N. data, and these data will be examined in
terms of the "mean threat score" of each individual. This should give
some idea of whether lack of information or uncertainty is related to

how threatening the extra-national world is judged to be.

Threat and Perceived Similarity

It has already been hypothesized that non-Communist subjects will
perceive Communist countries as threatening. It is possible that other
kinds of perceived differences make a country or a nationality seem
threatening. In order to investigate this possibility, each subject
will be asked to say which nationalities (from the country sample) are
most similar to his own and most different to his own. Here the subject
will supply his own criteria for similarity and difference. The
hypothesis will be that these "difference scores" will be positively
correlated with the threat scores of an individual.

These difference scores will also be correlated with ideology
scores to provide a crude projective technique for assessing the
ideology of each subject on a Communist to non-Communist scale. If
Communist nationalities are regarded as "different" and non-Communist
nationalities are regarded as similar, the correlation will be highly
positive. It is assumed that the multinational sample will provide some

variation in correlations, although they are all expected to be positive.

32

Threat and World Government

Threat has been defined here as a mechanism which increases
national boundary maintenance and reduces the possibility of systemic
linkage. The threat responses discussed here have been thought of
primarily as bilaterial with an individual country as a stimulus and
a person of a given nationality as the respondent. However, it has
also been suggested that summing and averaging the threat scores of an
individual should give an indication of how dangerous he judges his
international environment to be.

If this is the case, then a person with a high mean threat score
should be against supra-national arrangements which links his country
with other countries in an international system. To investigate this,
subjects were asked to indicate one of five degrees of approval or
disapproval with the following statement:

I think that my country should give up some of its

soverignty to become part of a world government-~in the

same way that a state or privincial government gives up

some of its power to a national government.

The hypothesis is that when subjects are cast into five categories
according to the five possible responses to the statement, subjects who
most strongly disapprove of world government will have highest mean

threat scores, and subjects who most strongly approve of world government

will have lowest mean threat scores.

CHAPTER III

PROCEDURES AND METHODS 0? ANALYSIS

Two samples were utilized in this study: (1) a sample of 2a
countries used in the questionnaire, and (2) a sample of 24 people who
reSponded to the questionnaire.

The country sample was constructed by the experimenter by
selecting countries on the basis of (l) ideology, (2) population size,
and (3) energy consumption per capita.

The initial criterion for selection of countries was that they
appear in the United Nations Statistical Yearbook for 1962. An effort
was also made to limit the sample to countries which would be familiar
to college graduates whose professional interests were not in inter-
national relations. Hithin these limitations a country sample of
maximum geographic diversity was chosen.

Further details concerning the country sample will be described
on the following pages. But first, this will be prefaced by a general
description of the measuring instrument. It consisted of a brief
introduction and six sections. (Appendix A is a cepy of the questionnaire.)
The first five sections each asked for reaponses concerning the same
2H countries, and the sixth section included a world government question
and blanks for information about the subject's sex, age, occupation or
field of study, educational background, country of birth, nationality,
parent's nationality, and foreign countries lived in for at least one
year before the age of 21.

33

3a

The order of presentation of the sections was as follows:
(1) threat represented by countries (2) similarity of nationalities
(3) energy consumption per person by countries (4) national populations
and (S) ideologies of governments (Communist to non-Communist).

In each section the order of presentation of countries was
alphabetical, except for the section on national population in which
this order was reversed. Thus if the order of presentation has affected

the variability of reSponses, it is the same for all subjects.

Ideology

An initial selection of some 50 countries was made from the latest
available United Nations publications--principally the Ugh. Statistical
Yearbook and the U.N. Demggraphic Yearbook. These were divided by the
experimenter into Communist and non-Communist countries.

The increasing number of "neutralist" countries posed problems
which were resolved by the experimenter by casting some of them into
each ideological category. Neutralist nations are here defined as
20 countries whose representatives (1) attended a 1961 conference in
Belgrade, Yugoslavia, and (2) signed a letter to the Secretary General
of the United Nations asking that the 196k General Assembly be postponed
until after a meeting of heads of state of non-alligned nations
scheduled to meet in Cairo, Egypt, in October, 1964. (28)

Five of these countries appear in the country sample. Cuba and
Yugoslavia are experimenter classified as Communist countries while
Algeria, Indonesia and the Congo are classified as non-Communist
countries. (Banks and Textor classified Cuba as quasi-Communist rather

than Communist. Otherwise, the countries classified as Communist here

35

are identically classified by Banks and Textor.)

The subjects in this investigation were also asked to classify
the governments of each country in the sample on the basis of ideology.
They classified the countries according to degree of Communist or non-
Communist government orientation on a eight-point scale. Distribution
was forced normal so that only one country was allowed to be scored
"most non-Communist" and one, "most Communist" in the sample, etc.

This allowed for the countries to be dichotomized according to ideological
orientation on the basis of the judgements of the subjects. Countries
scored 1 through u were called "subject classified as non-Communist"

while those scored 5-8 were called "subject classified as Communist."

These two bases of classification were used with each of the
three major independent variables in this study. This allowed for the
judgements concerning threat to be cast into two matrices: one on the
basis of experimenter classifications, and the other on the basis of
subject classification. The same type of analysis was carried out on
the two matrices.

Experimenter classifications were made primarily on the basis of
United Nations published data, and the divergences between the two
matrices will yield some idea of how the perceptions and judgements of

the subjects vary from this criterion.

Pepulation
National population figures for mid-1962 were the latest available
in the United Nations Demoggaphic Yearbook for the countries under
consideration, and these were used for experimenter classification of

countries according to pepulation size. A dichotomy was also wanted on

36

this variable, and countries were classified as "large" if their
populations were reported as greater than 16 million in mid-1962, and
"small" of their pOpulations were 16 million or less.

Subjects were asked to estimate the populations of each of the
countries in the sample to the nearest million, and this provided the
basis of subject classifications on this variable. In this case, the
mid-point which discriminated "large" from "small" countries was
determined by the subjects' estimates. That is, the 12 countries which
were considered largest by each subject were classified as "large," and
the rest, as "small." If two or more countries at the subject's
pOpulation midpoint were estimated as having the same pOpulation, they

were randomly cast into the two categories by the experimenter.

Wealth and Energy

The initial decision was to use per capita wealth as one of the
three major independent variables of the investigation. However, an
examination of the U.N. Yearbook of National Account Statistics,¢1962,
indicated that the latest year that a reasonably complete set of
figures was available referred to the year 1958. Also, there were no
figures for most Communist countries.

An investigation of other possible indices indicated that per
capita figures for consumption of commercial energy were available for
a large number of countries for the year 1961, and that Communist
countries were included. Also, it had been demonstrated that for data
from 62 countries, there was a high positive correlation between per

capita gross national product and per capita energy consumption. (29)

37

The pre-selection of countries was classified into three levels

of energy consumption: high, medium and low. In the final sample of

u countries there are eight countries in each of these categories.
But because this variable is nested within the two mentioned previously,
the greatest variability within each level is present in the wealth or
energy variable.

Subject classifications of countries on this variable were based
on individual estimates of the wealth or energy consumption per capita
on a 15-point scale. Part of the instructions on this part of the
questionnaire said:

Now you are to make the best estimate that you can about

the wealth or poverty of the same list of countries. In

order to avoid the problem of equivalency of different

currencies, you are asked to judge each country in terms

of the amount of coal, hydroelectric power and other

sources of commercial energy available to the average

person in each country...

In general, you should consider the scale below and the

one on the next page, as asking you to make distinctions

between richer and poorer countries. The richer that

you think a country is--that is, the greater the income

of the average person--the farther to the right you should

make your "X". However, in case of doubt, you should

consider the commercial energy consumption per person

as the indicator of wealth...

Note that the scale values (0 to 8000) increase more

rapidly as the scale moves from the low values on the

left to the high values on the right.

Thus an effort was made to elicit distinctions between countries
perceived as rich and poor, and at the same time introduce the general
notion of the criterion index for this variable which was per capita
energy consumption.

In summary, a sample of 29 countries was constructed by the

experimenter in which 12 were classified as Communist and 12 as

38

non-Communist. The sample was so selected that there were 12 "large"
countries and 12 "small" countries. Six of the Communist countries

were "large," and six were "small"; six of the non-Communist countries
were "large," and six were "small." Within each of these four groups

of six countries, there were two relatively poor countries, two countries
of intermediate wealth, and two relatively rich countries.

The final result is a 2x2x3 design with two countries (one
replication) in each cell. The direction of the nesting assumes that
ideolOgy is most important in judgements concerning international threat,
and that population size and wealth are successively less important.

This design assumes that all of the cells can be filled with
countries of apprOpriate characteristics as described by U.N. data.

How well this was achieved is described by the following table:

TABLE 1
The Country Sample With Criteria Data
Energy
Population Consumption Index
(millions) (per capita)

Non-Communist Small Rich: New Zealand 2.9 2033
Switzerland 5.6 1951

Medium: Formosa 11.3 529
Alberia 11.6 25h

Poor: Congo lu.8 6H

South Viet-Nam 14.9 60

Large Rich: United States 186.6 80u2
France ”6.9 251k

Medium: Mexico 37.2 959
Japan 9n.9 1298

Poor: Pakistan 96.6 68
Indonesia 97.7 140

Communist Small Poor: Albania 1.7 337
North Korea 8.“ 528

Medium: Cuba 7.7 866
Hungary 10.1 2996

Rich: Czechoslovakia 13.9 6125

East Germany 16.0 u9u2

39

TABLE 1 (Continued)

Large Poor: North Viet-Nam 16.7 528
China (mainland) 686.U 528

Medium: Yugoslavia 18.8 sou
Roumania 18.7 1433

Rich: Poland 30.3 3182
Soviet Union 221.5 2921

The rich-medium-poor order is reversed in the Communist countries
so that this listing of pairs of countries represents hypothesized rank
order of threat. The prediction is that countries at the top will be
least threatening and countries at the bottom will be most threatening.

The cells are certainly not as homogeneous as might be desired.
This is eSpecially true in the Communist countries, but very little
selection was possible in this case. Bulgaria and Mongolia are the
only Communist countries in the U.N. publications which are left out

of the sample.

The Sample of PeOple

The main sample of subjects consisted of 20 male college graduates
who were all pursuing graduate studies at the same university. To this
was appended a special subsample consisting of four university professors
with professional interests in international relations.

An effort was made to introduce nationality as a variable in the
investigation by breaking the sample into five subsamples of four persons
each. The subsamples consisted of Brazilians, Indians, Chinese (from
Formosa), Egyptians and Americans. In all cases the nationality of the
subject was that of the country of birth, and none of the subjects had
been outside his home country for as much as a year before reaching 21

years of age. And in all cases the parents were both of the same

HO

nationality as the reapondent.

(Brazil and Egypt have similar levels of energy consumption and
per capita wealth while India is somewhat poorer. India and Egypt
are both "neutralist nations" in the terms mentioned earlier, and Brazil
had pursued a similar course quite recently.)

Except for the Americans, the sample of people was drawn from
a list of foreign students attending Michigan State University. In an
effort to obtain a stronger nationality effect in this small sample,
subjects were selected from students who had come to the United States
to do graduate work, and no undergraduates were included. Selection
of subjects was not random but rather was based on the willingness of
the subjects to respond to the questionnaire.

No effort was made to control for age or field of study. Data

from the questionnaires yields the following information:

 

TABLE 2
Age and Field of Study of Subjects
Brazilians Questionnaire
A e Field Date

Subject #11 3O Businesministration W
Subject #12 26 Business Administration 3/23/6“
Subject #13 #3 Sociology u/u/su
Subject #1u ( no) Business Administration 3/23/6u

E tians UAR
Subject £31 30 Education 3/25/69
Subject #22 30 Biochemistry (none)
Subject #23 23 Education 3/27/6u
Subject #24 35 Social anthropology 3/26/6u

Chinese (Formosa)
Subject §3I 27 Agricultural Economics 3/24/6u
Subject #32 27 Communication 3/23/6u
Subject #33 Chemistry u/u/su
Subject #3u 32' Agricultural Engineering M/S/6u
, Indians
Subject FBI 25 Agricultural Engineering 3/29/6”
Subject ##2 25 Electrical Engineering 3/26/6u
Subject #93 28 Biochemistry 3/2u/6u
Subject #uu 27 Plant Pathology 3/2u/6u

ul

TABLE 2 (continued)

Americans
Subject-m— 35 Communication 3/27/61;
Subject #52 29 Communication 3/2u/6u
Subject #53 Communication (none)
Subject #su '-' Communication (none)
U.S. Ex erts—
Subject #6! 5H International Relations 3/3/6u
Subject #62 #9 Agricultural Economics u/20/6u
Subject #63 ___ Economics (none)
Subject #SH no Journalism 3/27/6u

Here it Will be assumed that nationality is a much stronger
determinant of international attitudes than field of study, and while
the selection of the sample of peOple does not allow for a rigorous check
of this assumption, the duplication of departments in different
nationality subsamples will allow for an informal examination of this
assumption.

A factor analysis of the subjects in terms of their 2h international
threat reSponses will provide one means of checking this assumption.

An examination of the resulting factors should give some indication of
whether, say, biochemists and agricultural engineers are on the same
factor or on different factors according to their nationalities.

The same analysis should give an indication of whether the two
sub-samples of Americans (graduate students and faculty international
relations Specialists) are similar in their responses concerning inter-
national threat.

Host of the responses to the questionnaires were made during the
last week of March, 196“, and the remaining ones were made during April.

Three of the 0.8. experts made their responses some three weeks later

u2

than the rest of the sample. However, it is assumed that this time
differential was not significant enough to introduce variability in

reSponses.

Analytical_§rocedures

The basic analytical tool will be a factorial analysis of
variance in which the factors are ideology (two values), population
(two values), wealth (three values) and nationality (six values). The
alpha level will be set at the conventional .05 level, and greater
levels of significance will be mentioned.

Subjects' judgements about the threats of 2H countries will be
cast into two matrices (1) experimenter classified, and (2) subject
classified, and identical analyses of variance will be carried out on
each one. No hypotheses will be made about differences in the two sets
of results, but the differences will be examined and discussed.

The eXperimenter classified analysis is apprOpriate for making
statements about specific pairs of countries and groups of countries.
Here summation will be across subjects on reSponses about one or more
pairs of concrete countries.

In the subject classified analysis, it will no longer be possible
to make generalizations about specific countries, since these will vary
from one individual to another. Here the common quality of the rows
and columns being summed will be values of the variables of classification:
ideology, population and wealth. And here generalizations will refer to
these kinds of country characteristics.

The analysis of variance table for this design will consist of

fifteen possible sources of variance: four main effects, six two-factor

1+3

interactions, four three-factor interactions, and one four-factor
interaction.

Error terms will be drawn from six other analysis of variance
tables--one for each of the six subsamples of peOple. The sum of
squares for individual subjects (within subsamples) will be pooled for
the six subsamples and divided by the correSponding degrees of freedom
to produce a mean square error apprOpriate for testing the significance
of a main "nationality" effect.

Similarly, sums of squares for subject (within nationality)-
ideolOgy interaction for the six subsamples will be pooled and divided
by the correSponding degrees of freedom to produce an appropriate
denominator for the F-ratio testing the main ideology effect as well as
the ideology-nationality interaction. The six other error terms to
be used in this analysis will be computed in the same manner.

This computation of a series of different error terms should add
more precision to the analysis than is possible when a residual error
term is used in a fixed effects model. Individual subjects will be
considered as "replicates" of a given nationality.

Also, the set of error terms used here allows the use of analysis
of variance procedures with subsamples of subject reaponses in which
the variance is not homogeneous across subsamples.

Simple effects will be tested by t-tests in which the numerator
is the difference between a pair of cell means and the denominator will
be the mean square error term for the source of variance under con-

sideration. (30)

141+

Other Analyses

In order to examine the assumptions concerning the relationships
between the four main variables being examined, intra-individual
correlation coefficients will be computed for pairs of the sets of
responses for each subject. This will determine the direction and the
strength of the relationship between threat and ideology, threat and
population, and threat and energy. Here, individual differences will
be visible which were examined only indirectly--in mean square error
terms--in the analysis of variance.

To examine the inter-relationships between ideology, population
and energy, and their effectiveness in predicting the threat scores of
an individual, multiple correlation coefficients will be computed for
each individual. These will give an indication of how much of the
variation in threat scores can be accounted for by ideology, population
and nationality together, and in the three possible pairings.

To examine the relationship between threat scores and subsamples
of subjects, a Q factor analysis will be computed. The hypothesis will
be that each nationality will have a different pattern of threat scores,
and that these differences will produce different factors for each
nationality.

Responses on the 21-point threat scale will be treated as
equivalent to a Q-sort of countries on a threat dimension. Each subject's
set of threat reSponses will be correlated with that of every other
subject, and the resulting matrix of intercorrelatinns will be submitted
to factor analysis. A principal axis solution will be obtained and

submitted to varimax rotation to produce orthOgonal simple structure.

The factor loadings for the individuals loading highest on each
factor will be weighted apprOpriately to produce a typical set of
threat responses for each factor, and these weighted responSes will be
converted to z-sccres. Similarly, sets of responses for each factor

concerning

0

ideology, population, energy, and similarity of nationality
will be produced. These will provide a check on hypothesized relation-
ships between threat and the independent variables in terms of
individual countries.

Reaponses concerning similarity of nationality have been mentioned
briefly in the description of the questionnaire and in the Q analysis.
These responses are similar to the ideology responses in that the in-
structions require that the nationalities of the 2“ countries be ranked
on an 8-point scale in a normal distribution pattern.

Correlations of similarity and threat scores for each individual
will give an indication of the direction and strength of the relation-
ship between these two sets of responses.

Correlations of similarity and ideology scores for each individual

should yield a rough index of the ideology of each subject.

Criteria Data
The United Nations data on national populations and per capita
energy consumption used to construct the country sample has already
been mentioned. Correlations of these data with sets of individual
responses will be used to give an indication of how closely each subject's
reSponses correSpond to the criteria.
Also, a composite index will be constructed by noting in how

many cases each subject's responses on ideology, population, and energy

1&6

result in classifying a country in the same coll--of a total of twelve--
as that in which U.N. data classifies it. A perfect score here would

be 2n. Those data can be srrsysd by subjects, and also by countries.

World Govern-out
Finally, the subjects will be divided into five groups an the
bssis of their response to the world government ststssont on the last
psgs of the questionnaire. The hypothesis here is s corrolationsl one:
strong agreement with the world government statement will be associated
with low nosn threst scores, and strong dissgresnont will be associated

with high Issn thrust scores.

CHAPTER IV

RESULTS

The principal findings of this study consist of an analysis of
variance of threat scores. But before these results are detailed, a
few general descriptive findings will he reported.

Below are the average threat scores of individual subjects and

individual countries in the two samples:

TABLE 3

Average Threat Scores for Subject and Countries

 

“7

Sub ects Countgigg‘
Brazilian #11 1.1667 Albania 5.0u17
Brazilian #12 1.2917 Algeria 3.0583
Brazilian #13 8.5u17 China (mainland) 13.9167
Brazilian #1# .8750 Congo “.3333
Egyptian #21 5.7083 Cuba 6.8750
Egyptian #22 2.8333 Czechoslovakia 5.0u17
Egyptian #23 1.2083 East Germany 6.2917
Egyptian #2“ .5ul7 Formosa 3.6250
’Chinese #31 11.1667 France ' u.usas
Chinese #32 “.2500 Hungary 5.5000
Chinese #33 1.8333 Indonesia “.6667
Chinese #3“ 11.0“17 Japan 3.0000
Indian #ul 1.6250 Hexico 1.5833
Indian #02 2.7083 NOV Zealand .6667
Indian in: 2.7917 North Korea 6.1250
Indian #uu 1.6667 North Viet-Nam 6.9167
Anorican #51 11.3333 Pakistan 5.2500
American 652 9.7083 Poland “.7917
American #53 10.2500 Roumania “.6250
American #5“ 8.H583 South Viet-Nan 3.n167

48

TABLE 3 (continued)

U.S. Expert £61 .7917 Soviet Union 9.7917
0.8. Expert #62 5.5000 Switzerland .3333
0.3. Expnrt #63 10.5000 Unitod States 3.7083
0.3. Expert #su 1.6667 Yugoslavia n.0u17

The questionnaire obviously produced considerably variation in
response, both in terms of subjects and of countries. The 21-point scale
was scored from zero to twenty. with no further information than that,
the expected mean score for any person or country would be 10.

Switzerland and New Zealand have the lowest means and also the
lowest standard deviations (not shown). China, the Soviet Union and
North Viet-Nam all have the highest means and the highest standard
deviations. It seems that it is easier to get international consensus
concerning countries which are not threatening than for which countries
are threatening.

The countries are listed in the sane order as that which was
used in the questionnaire. (They were equally spaced in the questionnaire;
the groupings of four are arbitrary.) It is possible that the order
of presentation might have affected threat scores, but if there is such
an effect it seems to be a minor one.

Just as threat responses were averaged for subjects and for
countries, it is also possible to average other sets of responses from
the questionnaire. Per capita wealth or energy estieates were made on

a 15-point scale scored from zero to 70.

“9

TABLE 0

Per Capita Energy Consumption Averages for Subjects and Countries

Sub ects
Brazilian #11 18.7
Brazilian #12 15.0
Brazilian #13 7.1
Brazilian #1“ 15.0
Egyptian #21 18.9
Egyptian #22 5.7
Egyptian #28 8.0
Chinesa #31 8.5
Chinese #32 17.1
Chinese #33 21.0
Chinese 33% 9.8
Indian #91 13.0
Indian #uz 16.7
Indian #us 6.0
American #51 20.0
American #52 13.“
American #53 19.9
American #59 13.5
U.$. Expert #61 10.7
U.S. Expert #62 18.5
0.8. Expert #63 1u.u
UeSe Export #6“ So].

In terms of the scale and scoring procedure used here, the average

Countries

Albania 5 0
Algeria 3 2
China (mainland) 5.0
Congo 1 5

Cuba 6.2
Czechoslovakia 13.2
East Germany 13.9

Formosa 6.3
France 30.3
Hungary 10.6
Indonesia 2.9
JCPHD 2'4. 0
"Cu“ 7e 1
New Zealand 21.2
North Korea 3.0
North Viet~Nam 2.8
Pakistan 3.0
Poland 9.0
Roumania 6.6
South Viet-Nam 2.7

Soviet Union 27.8

Switzerland 38.9
United States 68.3
Yugoslavia 12.1

0.".

N
0050‘”
e e e e
000°

H0)
000010
es a
0000

HP”
”Pm".
e

e
0000

energy consumption per capita of the 2n countries in this sample is

17.0 when U.N. criteria data is used.

are above this figure, and 17 are be

variation in how this sample of subjects judges the average per capita

wealth in this sample of countries.

The average estimates of 7 subjects

low it. There is considerable

Data

50

Subjects were asked to estimate the population of each country

in the country sample to the nearest million, and these estimates may

be treated just as the threat scores and energy estimates were:

TABLE 5

Average Population Estimates for Subjects and Countries

Subjects

Brazilian #11
Brazilian #12
Brazilian #13
Brazilian #1u

Egyptian #21
Egyptian'#22
Egyptian #23
Egyptian #29

Chinese #31
Chinese #32
Chinese #33
Chinese #38

Indian #91
Indian #uz
Indian #ua
Indian #uu

American #51
American #52
American #53
American #Sn

UeSe EXPCrt #61
0.8. Expert #62
0.6. Expert #63
U.S. Expert #6:.

(in millions)

Countries

Albania
Algeria
China
Congo

Cuba
Czechoslovakia
East Germany
Formosa

France
Hungary
Indonesia
Japan

Mexico

New Zealand
North Korea
North Viet-Nam

Pakistan
Poland
Roumania

South Viet-Nam

Soviet Union
Switzerland
united States
Yugoslavia

78.8
26.0
21.0
28.0

232.8
15.2
186.0
31.2

U.N. Data
(for countries)

1.7
11.5
686.“

1“.8

The mean for the U.N. population data used as criteria was 69.7

million. The mean average estimates of 13 subjects are above this

criterion, and 10 are below it.

Taken at face value, this indicates

51

that some subjects think that these countries are much more highly
pepulated than do others.

The second and third columns of figures refer to populations of
countries in millions. When U.N. data are used as criteria, the
populations of 18 countries are overestimated, and 6 countries are under-
estimated.

The questionnaire also asked for estimates of the ideology of
each country on an 8-point non-Communist to Communist scale. Subjects
were asked to judge the present governments of the countries in the
sample. Responses were forced into a normal distribution, and as a
result all subjects have the some means. Therefore only country averages

will be presented in this case:

TABLE 6

Average Ideology Estimates for Countries

United States 1.29 Roumania 5.00
Switzerland 2.33 Poland 5.12
Formosa 2.87 Hungary 5.25
Japan 2.92 Yugoslavia 5.81
France 2.96 North Korea 5.50
New Zealand 2.96 North Viet-Nam 5.62
South Viet-Nam 3.82 Albania 5.79
Mexico 3.59 East Germany 5.79
Pakistan 3.83 Cuba 5.79
Congo 3.92 Czechoslovakia 5.83
Algeria ”.08 Soviet Union 7.25

Indonesia n.17 China 7.62

S2

A low score indicates a non-Communist government and a high
score indicates a Communist government. The fact that the twelve
countries in the left hand column of Table 6 were all experimenter
classified as non-Communist and the other twelve countries as Communist
indicates general consensus between the subjects and the experimenter
with regard to this variable.

Standard deviations are not shown, but the highest ones were for

Albania and Formosa. All other standard deviations were below 1.0.

Analysis of Variance of Threat Scores

The findings to be reported in the following pages are the
results of four identical factorial analyses of variance of threat
responses. These responses or scores have been cast into two different
matrices, one in which countries are experimenter classified as to
ideology, papulation, and wealth, and another in which the countries
are subject classified with regard to these variables.

In each of these two approaches to classification, the threat
responses are quantified in two different ways (1) as raw scores, and
(2) as standardized scores. This produces the data for the four analyses.

Threat responses for each country were marked on a 2l-point scale
and are scored from zero to 20. In this analysis the 2n countries in
the sample are treated as 12 pairs of countries which are similar in
ideology, population and wealth or energy level. The raw scores of
pairs of countries are summed so that the sum of any two "replicate"
countries may vary from zero to no.

Standardized scores were produced by computing means and standard

deviations for each subject, and translating the raw scores into

53

standard scores with a mean of 50 and a standard deviation of 10.

Raw scores then, refer to absolute scale values while standard

scores refer to relative threat for a given individual.

In the following table the results of the four analyses are

summarized.

TABLE 7.

Summary of Four Analyses of Variance

Ideology (Communist/Non-Communist)
Population (Large/Small)
Ideology-Population

Energy (Rich/Hedium/Poor)
Ideology-Energy
Population-Energy
Ideology-Population-Energy
Nationality (six subsamples)
Ideology-Nationality
Population-Nationality
Ideology-Population-Nationality
Energy-Nationality
Ideology-Energy-Nationality
Population-Energy-Nationality

Ideology-Population-Energy-Nationality

Threat Scores

Experimenter
Classified

Raw Std.
.0005 .0005

0001 .005

.0005 .0005
.01 .025
.0005 .0005

.005 .0005

.05 Contr.
.005
.001 .025
.01 .01
.001 .005

Subject
Classified
Raw Std.
.0005 .0005
.025 .005

.01
.025 Controlled
.05 .005
.05
.01 .05

58

These findings will be discussed in detail in the following
pages. The discussion will center upon the analysis of the experimenter-
classified raw scores, and the other analyses will not be discussed in
detail except in cases where they differ significantly from it. For
example, the main ideology effect was highly significant in all four
analyses, and no purpose would be served by presenting details from
the four analyses.

Only significant results are noted in the table above, and blanks
indicate a lack of significance at the .05 level.

The order of presentation will be that of the analysis of
variance tables: (1) ideology, (2) population, (3) energy, and (8)
nationality. That is, the three nested "treatments" will be presented

first, and the "levels" last.

Ideology
The hypothesis is that Communist countries will be perceived as
more threatening than non-Communist countries.
The results are as fellows:
TABLE 8a

Ideology: Average Threat Scores for Communist
and non-Communist Countries

Threat Scores

N Mean S.D.
(observations)
Experimenter—classified non-Communist countries 188 6.89 8.88
Experimenter-classified Communist countries 188 13.18 12.28

The AOV table shows the following results as the contribution

of ideology to the variance in threat responses:

55

TABLE 8b

Main Ideology Effect

Source of Sum of Mean Observed
Variance Squares ‘2§_ Square Error F F.95
Ideology 3186.6806 1 3186.6806 132.7963 23.9968 8.81 with 1/18

df

The very large F values observed--significant beyond the .0005
level in all four cases-osuggest that no subtle distinction is being
examined here. Even though 12 of the 28 subjects are from what are
sometimes considered neutralist countries, the circumstances surrounding
the data gathering (i.e., on the campus of a 0.5. university) no doubt
contributed to this resounding rejection of a null hypothesis which

seems even less tenable after the fact than it did before.

Population
The theoretical hypothesis here is that, ideology aside, countries
with large populations will be seen as more threatening than those with
small populations.
Countries were again dichotomized with equal numbers of Communist

and non-Communist countries in each category. The results were:

TABLE 9a

Population: Average Threat Scores for Large and Small Countries

N
(Observations) Mean S.D.
Experimenter-classified small countries 188 8.38 10.62

Experimenter-classified large countries 188 11.28 11.32

56

TABLE 91)

Main Population Effect

Source of Sum of Mean Observed
Variance Squares 2£_ Sguare Error F F.95
Population 589.3889 1 589.3889 35.8188 16.8566 8.81 with 1.18

df

The conclusion is that whether countries are classified as large
and small by United Nations data by the experimenter or by the estimates
of each subject, it is possible to reject the null at the .05 level of
confidence and state that large countries are perceived as more threatening

than small countries by subjects of five different nationalities.

Ideology and Population
Are these two factors independent in their effects upon threat

scores, or do they interact? Here are the results of the experimenter

classified analysis:

TABLE 10a

Ideology and Papulation: Average Threat Scores for
non-Communist and Communist, Large and Small Countries

Small Large
Non-Communist 5.2778 7.6988
Communist 11.8861 18.7917

Here there are 72 observations in each cell. The AOV table reads:
TABLE 10b

Ideology..Popu1ation Interaction

Source of Sum of Mean Observed
Variance Squares ‘25_ Square Error F F.95

Idso.-Pop. 18.2222 1 18.2222 10.1667 1.3989 8.81 with 1/18 df

57

The interaction is not significant, and the effects of ideology
and p0pu1ation of a country upon its perceived threat are independent

for this sample of people.

Energy
In contrast with the first two dichotomous factors, this variable
is three-valued. Countries were clasified as high, medium or low in per
capita energy. The hypothesis is that there will be no significant
difference between the three cells when threat scores are classified in

this manner. The results of the experimenter classified analysis are:

TABLE 11a

Energy: Average Threat Scores for Experimenter-Classified
High, Medium and Low Energy Countries

N Mean Threat Scores Std. Dev.
(observations)
High energy (rich countries 96 8.9879 10.8808
Medium energy countries 96 8.0729 9.9792
Low energy (poor) countries 96 12.8167 12.1877
TABLE 11b

Main Energy Effect

 

Source Sum of Mean Observed
of Variance Squares 2: Square Error ' F F. 95
Energy 1013.3125 2 506.6563 20.8977 28.7177 3.26 with 2/36 df

The very large observed P value suggests that there is a highly

significant main energy effect, but the direction in the differences

58

between the means makes it difficult to interpret from the data
presented here. The following pages will show that the energy or wealth
factor interacts significantly with other factors, and that generalizations
about wealth or energy must be limited by these other factors.

When subjects classify the countries in the sample on per capita

energy consumption, the results are quite different:

TABLE 12a

Average Threat Scores for Subject-Classified
High, Medium and Low Energy Countries

N Threat Scores
(observations) Mean Std. Dev.
High energy (rich) countries 96 9.8687 10.7233
Medium energy countries 96 9.3858 11.0328
Low energy (poor) countries 96 10.5833 11.2331
TABLE 12b

Main Energy Effect, Subject-Classified Analysis

Source of Sum of Mean Observed
Variance Squares 2E Sguare Error F F. 95
Energy 85.8958 2 82.9879 50.8388 .8888 3.26 with 2/36 df

In this case there is no significant difference between these
three sets of observations. This is in marked contrast to the findings
in the experimenter-classified analysis.

Again, part of the explanation is that energy interacts with other

factors in the analysis, and this will be detailed on the following pages.

59

However, it should be noted that the relative order of the means
remains stable across the two analyses: the countries of intermediate
wealth are seen as least threatening, the rich countries next most
threatening, and the poor countries as most threatening.

Because of the contrast in the results, it seems worthwhile to
return for a moment to the experimenter-classified matrix, and to specify
the three groups of countries which resulted in highly significant

differences in threat scores:

TABLE 13

Experimenter Classification of Countries by per capita Energy Consumption

 

 

High per capita Medium per capita Low per capita
Energy Consumption Energy Consumption Energy Consumption
New Zealand 2033 Formosa 529 Congo 68
Switzerland 1951 Algeria 258 South Viet-Nam 60
United States 8082 Mexico 959 Pakistan 68
France 2518 Japan 1298 Indonesia 180
Czechoslovakia 5125 Cuba 866 Albania 337
East Germany 8982 Hungary 2896 North Korea 528
Poland 3182 Yugoslavia 908 North Viet-Nam 528
Soviet Union 2921 Romania 1833 China (mainland) 528
Mean 3,838 Mean 1,092 Mean 282
3.0. 2,088 3.13. 682 3.6. 223

It is the threat scores of these specific countries which are
used in the energy consumption cells in the experimenter-classified

matrix. It is obvious that the subject classified matrix cast these

60

some countries into these three categories in patterns which were

different enough to alter the significance of the results.

Ideology and Energy
In the experimenter-classified analysis, these two factors

interacted as follows:

TABLE 18a
Ideology and Energy: Average Threat Scores

Mean Threat Scores

Rich Medium Poor
Non-communist countries 8.7917 5.8333 8.8333
Communist countries 13.1082 10.3125 16.0000

Here there are 88 observations per cell. The AOV table shows:

TABLE 18b

Ideology: Energy Interaction

Source of Sum of Mean Observed
Variance Sguares ‘25. Sguare Error F F.95

Ideo.-Energy 185.8803 2 92.9201 17.6019 5.2790 3.26 with 2/36 df

The interaction of ideology with energy is significant at the .01
level of confidence. (Mean differences of 1.5 are significant.)

When ideology is controlled, poor countries are most threatening
at both levels, but there are significant differences between rich and
medium Communist countries. This difference is not significant in the
case of the non-Communist countries. When energy level is controlled,

Communist countries are more threatening at all three levels.

61

Results in the subject classified analysis were as follows:

TABLE 15

Ideology and Energy: Average Threat Scores, Subject-Classified Analysis

Rich Medium Poor
Non-Communist countries 5.3750 6.8125 7.35u1
Communist countries 13.5625 11.9583 13.8125

Here the relationships between the cell means are the same as
in the experimenter-classified analysis, but the pattern has been

attenuated enough to reduce it to statistical insignificance.

Papulation and Energy
Cell means for the experimenter classified analysis are as

follows:

TABLE 16a

Population and Energy: Average Threat Scores

Rich Medium Poor
Small countries 6.1667 9.5208 9.u583
Large countries 11.7292 6.6250 15.3750

Again in the 2x3 table, each cell mean is based on “8 observations.
Results from the AOV table are:
TABLE 16b

Population-Energy Interaction

Source of Sum of Mean Observed
Variance Sguares ‘RE_ Ssuare Error F F.95

Pop.-Bnergy 119u.6319 2 597.3160 22.9“68 26.0305 3.26 with 2/36 df

62

This is the largest P value observed in the eXperimenter-
classified analysis, and it seems worthwhile to examine the simple
effects in detail. Each of the six means refers to threat scores for

four specific countries:

TABLE 17

Countries Experimenter Classified by Energy Level and Papulation

Rich Medium Poor

New Zealand Formosa Congo

Small Switzerland Algeria South Viet-Nam
Czechoslovakia Cuba Albania
East Germany Hungary North Korea
united States Mexico Pakistan

Large France Japan Indonesia
Poland Yugoslavia North Viet-Nam
Soviet Union Rumania China (mainland)

Differences between each pair of means--exc1uding diagonals--
were tested with t-tests. All differences were significant with the
single exception of the difference between small, poor countries and
small medium (energy) countries. In contrast, the greatest difference
was between large, poor and large, medium (energy consumption) countries.

In the subject-classified analysis, the corresponding cell means

were:

TABLE 18
Population and Energy: Average Threat Scores, Subject-Classified Analysis
Rich Medium Poor
Small countries 7.2292 9.1667 9.9167

Large countries 11.7083 9.6081 11.7500

63

Tests of difference between means revealed only three which were
significant. The results may be summarized as follows: When wealth is
controlled, large countries are seen as more threatening than small
countries at the rich and poor level, but not at the intermediate level.
When population is controlled, there is only one significant difference
in the table: large, poor countries are more threatening than large

countries of medium wealth or energy consumption.

Ideology, Population and Energy
When these three factors are taken into consideration at the same
time, as variables of classification, the 288 observations which form the
basis for this analysis are cast into twelve cells with 2n observations

in each cell. The mean threat scores in each cell refer to two countries.

TABLE 19a

Ideology, Population and Energy: Average Threat Scores

Non-Communist Countries Communist Countries
Small Large Small Large
Rich 1.0000 8.5833 11.3333 19.8750 Rich
Medium 7.0833 “.5833 11.9583 8.6667 Medium
Poor 7.7500 9.9167 11.1667 20.8333 Poor

The figures above are from the experimenter-classified analysis.
The AOV table data are:
TABLE 19b

Ideology-Population-Energy Interaction

Source of Sum of Mean Observed
Variance Sauares 2:. Sguare Error F P.95

Id.-Pop.-Bn. 825.0986 2 313.M253 18.69““ 11.3683 3.26 with 2/36 df

614

The observed F is significant at the .01 level, and therefore it

seems useful to examine the countries in each cell:

TABLE 20.

Countries Experimenter Classified by Ideology, Population and Energy

Non-Communist Countries Communist Countries
Small Large Small Large
Rich New Zealand United States Czechoslovakia Poland
Switzerland France East Germany .Soviet Union
Medium Formosa Mexico Cuba Yugoslavia
Algeria Japan Hungary Roumania
Poor Congo Pakistan Albania North Viet-Nam
South Viet-Nam Indonesia North Korea China (mainland)

The means in the last table are difficult to summarize. Ex-
ceptions are evident for the main effects which have been reported.
(For examining this table, a difference of 3.1625 between two means
indicates a significant difference.)

Communist countries have been found more threatening than non-
Communist countries, but Yugoslavia and Roumania are seen as no more
threatening than four of the six non—Communist cells. However, with
population and energy controlled, there are significant ideology
differences in each pair of six cells.

Large countries have been found more threatening than small
countries, but when ideology and energy consumption are controlled, the
Opposite is found at the medium energy consumption level: Formosa and
Algeria are seen as more threatening than Mexico and Japan. And Cuba

and Hungary are seen as more threatening than Yugoslavia and Roumania

65

by this sample. The large-small differences are not significant in
the case of poor non-Communist countries, and are greatest in the case
of poor Communist countries.

Poor countries are more threatening than rich countries only in
the small, non-Communist cells and the large, Communist cells. The
generalization that poor countries are more threatening than medium energy
consumption countries is sustained for the large countries, but rejected
for the small countries in this sample.

It is difficult to summarize these findings, but it is possible
to note six levels of threat which are attributed to the countries in these

cells, plus a seventh level which overlaps two levels:

TABLE 21

Levels of Perceived Threat

 

Least threatening Most threatening
New Zealand Mexico Congo Albania Soviet Union China(mainland)
Switzerland Japan South Viet-Nam North Korea Poland North Viet-Nam
United States Czechoslovakia
France East Germany
Yugoslavia Cuba
Roumania Hungary
Pakistan
Indonesia
Formosa
Algeria

In the subject-classified analysis, it will not be possible to

specify the countries in each cell. Here are the cell means:

66

TABLE 22a

Ideology, Population and Energy: Average Threat Scores,
Subject-Classified Analysis

Non-Communist Countries Communist Countries
Small Large Small Large
Rich 3.8583 7.2917 11.0000 16.1250
Medium 7.0833 6.5ul7 11.2500 12.6667
Poor 7.0833 7.6250 11.7500 15.8750
TABLE 22b

Ideology-Population-Energy Interaction,
Subject-Classified Analysis

Source of Sum of Mean Observed
Variance Sguares 2£_ Sauare Error F F.95

Id.-Pop.-En. 11.1319 2 5.5660 3u.2766 0.1623 3.26 with 2/36 df

Here the ordinal relationships between the means in most cases
remains the same as in the experimenter-classified analysis, but the
magnitude of the differences has decreased, and the error term has in-
creased. The result is that when ideology is controlled there are no
significant differences between population or energy levels. However,
when population and energy are controlled, there are significant

differences between the two values of ideology in every case.

Nationality Subsamples
In previous pages threat scores for the entire sample of people
have been cast into cells on the basis of three "treatment" variables.

Now, "levels" or subsamples of subjects will be brought into the analysis.

67

To put it another way, the focus of interest changes from characteristics
of the stimulus country to characteristics of the respondent.

The first question: Is there a main nationality effect? That
is, are there significant differences between mean threat scores of sub-

samples fOr the entire sample of countries?

TABLE 23a
Nationality: Average Threat Scores for Six Subsamples

N Threat Scores

(observations) Mean S.D.
Brazilians ”8 6.0833 10.01u5
Egyptians 48 5.1858 7.3802
Chinese (Formosan) us 18.1u58 12.887u
Indians #8 4.3958 6.7026
Americans #8 19.8750 9.u3u3
0.3. Experts #8 9.2292 9.5910

Since there is no breakdown of countries, the data here is the

same for the experimenter-classified and the subject-classified analysis:

TABLE 23b

Main Nationality Effect

Source of Sum of Mean Observed
Variance Sguares ‘25_ Sauare Error F P.95

Nationality 8899.091? 5 1779.8083 503.0u63 3.5381 2.77 with 5/18 df

The main nationality effect is significant at the .05 level, but

it is limited by interactions to be reported on the following pages.

68

Ideology of Country and Nationality of Subject
When threat scores are divided into those of Communist and non-
Communist countries, are there significant differences across the
nationality subsamples? Here are the results in the experimenter-

classified analysis:

TABLE 24a

Ideology and Nationality: Average Threat Scores

Non-Communist Communist Difference
Brazilians .6667 11.5000 10.8333
Egyptians n.5u17 5.7500 1.2083
Chinese 9.8333 18.8583 8.6250
Indians u.oooo “.7917 .7917
Americans l3.u167 26.3333 12.9166
U.S. Experts 6.US83 12.0000 5.5u17
TABLE 2ub

Ideology-Nationality Interaction

Source of Sum of Mean Observed
Variance Sguares 2P; Sauare Error F F. 95

Nationality- 1509.9861 5 301.9972 132.7963 2.27M1 2.77 with 5/18 df
Ideology

Here the interaction approaches but does not reach significance
at the .05 level. The differences between Communist and non-Communist
means are presented for comparison with the subject classified analysis

which follows:

69

TABLE 258

Ideology and Nationality: Average Threat Scores,
Subject-Classified Analysis

 

Non-Communist Communist Difference
Brazilians 1.2083 10.9583 9.7500
Egyptians “.5“17 5.7500 1.2083
Chinese 9.7916 18.5000 8.708“
Indians 3.9167 “.8750 .9583
Americans 13.1250 26.6250 13.5000
U.S. Experts 6.5000 11.9583 5.“583
TABLE 25b

Ideology-Nationality Interaction

Source of Sum of Mean Observed
Variance Sauares 9}; Sguare Error F F.95

Nationality- 1““5.2813 5 289.0562 101.1030 2.8590 2.77 with 5/18 df
Ideology

Here the interaction between ideology and nationality is signifi-
cant. An examination of pairs of means with t-tests indicates that when
subject nationality is controlled, there are no significant differences
between Communist and non-Communist countries. Neither are there
significant differences across nationalities in the case of non-Communist
countries. However, in the case of Communist countries, there are
significant differences between the Americans and the Egyptians in the
sample, and also between the Americans and the Indians. Here nationality

of subject is a somewhat stronger effect than ideology of country.

70

Population of Country and Nationality of Subject
Does the perceived threat of large, as opposed to small, countries
vary significantly across nationality subsamples? Here are the mean cell

values and P-ratio data for the experimenter-classified analysis:

TABLE 26a

Population and Nationality: Average Threat Scores

 

Small Large Difference
Brazilians 5.5833 6.5833 1.0000
Egyptians 3.6667 6.6250 2.9583
Chinese 11.7083 16.5833 “.8750
Indians 1.5000 7.2917 5.7917
Americans 19.1667 20.5833 1.“166
U.S. Experts 8.6667 9.7917 1.1250
TABLE 26b

Population-Nationality Interaction

Source of Sum of Mean Observed
Variance Sguares 2E. Sauare Error F F.95

Pop.-Nat'1. 25“.6111 5 50.9222 35.81“8 1.“218 2.77 with 5/18 df

Here, there is no significant interaction between the nationality
of the subject and the population size of the country. In other words,
judgements of the degree of threat associated with large in contrast
to small population countries is relatively stable across nationality

sub-groups in this sample.

71

In the table, a significant difference between a pair of means
must be greater than 6.27. When nationality is controlled, there is no
significant difference between the threat scores of large and small
countries, although the means for large countries are higher in every
case.

When population is controlled, there are significant differences
between nationalities. For small countries the mean for the American
subsample is greater than for any other subgroup; fer large countries
the American mean is greater than the rest with the exception of the
Chinese.

For large countries, the Chinese mean is greater than that of the
Egyptians, Indians, Brazilians and U.S. experts. For small countries,
it is greater than that of Egyptians and Indians. And for small
countries only, the mean of the U.S. experts is greater than that of the
Indian subsample.

In this table, nationality of the subject is a stronger effect
than the population of the country being judged with regard to threat.
Chinese and Indian subjects make the greatest distinction between large

and small countries in the experimenter-classified analysis.

Ideology, Population and Nationality
In the following table of means, control for ideology has been

added to the two factors just previously discussed:

72

TABLE 27a

Ideology, Population and Nationality: Average Threat Scores

Non-Communist Communist

Small Large Small Large
Brazilians .1667 1.1667 11.0000 12.0000
Egyptians 2.7500 6.3333 “.5833 6.9167
Chinese 7.8333 11.8333 15.5833 21.3333
Indians 1.“167 6.5833 1.5833 8.0000
Americans 13.5833 13.2500 2“.7500 27.9167
U.S. Experts 5.9167 7.0000 11.“167 12.5833

TABLE 27b
Ideology-Population-Nationality Interaction

Source of Sum of Mean Observed
Variance Sauares 2f; Sguare m F £1.23
Id.-Pop.- “1.1111 5 8.2222 10.1667 0.8087 2.77 with 5/18 df
Nationality

This triple interaction is not significant. However, it seems
useful to examine the means in this table with t-tests. Here any
difference greater than 3.5600 is significant. When papulation and
nationality are controlled, there are not always significant differences
between Communist and non-Communist countries. They exist for the
Brazilian, Chinese, American and U.S. expert subsamples, but not for
Egyptians and Indians. These differences are stable across large and

small countries.

73

When ideology and nationality are controlled, small-large
differences appear in the case of Communist countries for Chinese and
Indians, and in the case of non-Communist countries for Chinese, Indians
and Egyptians.

When ideology and population are controlled, nationality differences
are evident in all four columns. These will not all be detailed, but
it can be seen that the similarity between the Indian and Egyptian sub-
samples is stable across the four columns. It is also interesting to
note that Americans can not only be clearly discriminated from the other

nationality subsamples, but from the U.S. experts as well.

Energy and Nationalitv
Are there significant differences in threat scores across nationality
subsamples for countries of three different energy consumption levels?

Here are the findings from the experimenter-classified analysis:

TABLE 28a

Energy and Nationality: Average Threat Scores

Rich Medium Poor
Brazilians 6.7500 5.5625 5.9375
Egyptians 7.6875 2.6875 5.0625
Chinese 12.5625 12.5625 17.3125
Indians 1.7500 1.5625 9.8750
Americans 16.9375 18.0000 2“.6875
U.S. Experts 8.0000 8.0625 11.6250

7“

TABLE 28]:

Energy-Nationality Interaction

Source of Sum of Mean Observed
Variance Sauares 2E; Sauare Error F F. 95
En.-Nat. 862.7708 10 86.2771 20.“977 “.2091 2.10 with 10/36 df

The interaction is significant at the .01 level. When nationality
is controlled the only significant differences between rich, medium and
poor countries are those between medium and poor countries for Indians
and Americans. (A significant difference here must be greater than
5.3806). When energy level is controlled, a number of differences be-
tween nationality subsamples are visible, but not all are stable across
the three columns.

The corresponding figures for the subject-classified analysis are:

TABLE 29a

Energy and Nationality: Average Threat Scores,
Subject-Classified Analysis

Rich Medium Poor
Brazilians 6.6250 6.1250 5.5000
Egyptians 8.5000 3.000 3.9375
Chinese 12.6875 1“.1875 15.5625
Indians 5.1875 2.6875 5.3125
Americans 16.8125 20.6875 22.1250
U.S. Experts 7.0000 9.6250 11.0625

75

TABLE 29b

Energy-Nationality Interaction

Source of Sum of Mean Observed
Variance Sguares '2: Ssuare Error F F.95
En.-Nat 715.0625 10 71.5062 50.83““ 1.0667 2.10 with 10/36 df

In contrast with the previous data, the interaction of energy and
nationality is not significant in the case of the subject-classified
analysis. The error term has increased considerably, and here a
significant difference between a pair of means must be greater than
13.3039.

When nationality is controlled, all differences between threat
scores for countries of different energy levels disappear. When energy
level of country is controlled, some nationality differences remain.
Americans can be discriminated from Brazilians, Egyptians and Indians
for medium and poor countries, but not for rich countries. Here, again,

nationality is the stronger effect.

Ideology, Energy and Nationality
What happens to the findings when ideology is added to the two
factors examined above? Here are the results of the experimenter.

classified analysis:

TABLE

76

30a

Ideology, Energy and Nationality: Average Threat Scores

Non-Communist Communist
Rich Medium Poor Rich Medium Poor
Brazilians 1.625 0.250 0.125 11.875 10.875 11.750
Egyptians 8.750 2.000 2.875 6.625 3.375 7.250
Chinese 5.750 11.625 12.126 19.375 13.500 22.500
Indians 1.500 1.500 9.000 2.000 1.625 10.750
Americans 7.500 12.375 20.375 26.375 23.625 29.000
U.S. Experts 3.625 7.250 8.500 12.375 8.875 1“.7SO
TABLE 30b

Ideology-Energy-Nationality Interaction

Source of Sum of Mean Observed

Variance Sguares 2E. Sguare Error F F.95

Id.-En.- 53“.“931 10 S3.““93 17.6019 3.0366 2.10 with 10/36 df
Nat'lty.

In this table a significant difference between a pair of means
must be greater than 9.2“1. This triple interaction is significant at
the .01 level. When energy level and nationality are controlled, there
are differences in threat scores between Communist and non-Communist
countries for Brazilians at three energy levels, for Chinese at two energy
levels (rich and poor), and for Americans at two levels (rich and medium).

When ideology and nationality are controlled, the only significant

difference for energy levels of countries is the discriminations that the

American subjects make between rich and poor non-Communist countries.

77

And when ideology and energy are controlled, there are nationality
differences in every column except the one for rich, non-Communist countries.
In no case are there significant differences between Egyptians and
The similarity of Egyptians and Indians is also stable across

Indians.

the two analyses.

Population, Energy and Nationality
The final triple interaction to be examined is similar to the
last section, but ideology has been replaced by pOpulation. The data

from the experimenter-classified analysis follows:

TABLE 31a
Population, Energy and Nationality: Average Threat Scores

Small Countries Large Countries

Rich Medium Poor Rich Medium Poor
Brazilians 5.250 7.500 “.000 8.250 3.625 7.875
Egyptians 2.875 2.500 5.625 12.500 2.875 “.500
Chinese 8.375 13.875 12.875 16.750 11.250 21.750
Indians 1.500 1.375 1.625 2.000 1.750 18.125
Americans 12.625 21.500 23.375 21.250 1“.500 26.000
U.S. Experts 6.375 10.375 9.250 9.625 5.750 1“.000

TABLE 31b

Population-Energy-Nationality Interaction

Source of Sum of Mean Observed
Variance Sauares 2E; Ssuare Error F
Pop.-En.- 9“6.618l 10 9“.6618 22.9“68 “.1253 2.10 with 10/36 df

“at. ltye

78

The interaction of these three factors is significant at the .01
level. In this table a difference of 12.0“7 between a pair of means is
the minimum for significance.

When energy and nationality are controlled, in 13 of the 18 cases
here the mean threat score for large countries is greater than for small
countries, but in only one case does the difference attain significance:
the judgements of Indians concerning poor countries.

When pOpulation and nationality are controlled, there are no
energy effects for small countries, and only one for large countries
which reaches significance: in the judgement of the Indian subsample,
large, poor countries are more threatening than large rich countries.

When pOpulation and energy are controlled, there are significant
nationality effects in every column except the one for small, rich
countries.

On the next page are displayed both the experimenter-classified
and subject-classified tables of means subdivided by all four factors.

Each of the 72 cells consists of four observations. The AOV table data

are:
TABLE 32a

Four-factor Interaction
Source of Sum of Mean Observed
Variance Sauares EF- 5 uare m F 32.9.:
Experimenter-Classified
Id.-Pop.- 218.6181 10 21.8618 18.69““ 1.169“ 2.10 with 10/36
En.-Nat1. df
Subject-Classified
Id.-Pop.- “35.6180 10 “3.5618 33.3715 1.3053 2.10 with 10/36

En.-Natl. df

muneaxm .m.0
mceownee<
snowman
omecwzo
mcmuummwu

mcewawnenm

.munonxm .m.0
q. mceofinoa<
7
mcemvcH
onecanu
mceuua>wm

mamwawumnm

mb.md
00.0w
00.HH
mm.:m
00.5

mm.m

noom Eran»:

mN.HA
0m.0m
00.N
om.0m
mh.w

00.HH

mb.NH
0m.bn

mw.b

mm.mn
om.md
0m.:H

scam

momnpcsou owned

00.0H
00.30
mb.md
mw.mm
0m.m

0m.mH

mm.0
om.HN
mh.H
mb.NH
0m.w

mm.b

Loom suave:
newpucsoo swung

0m.NH
mu.0m
0m.m
00.nm
mb.m
mm.na

some

00.0d
mn.mw

mN.H
00.5H
0m.:
mb.0

Loom

mN.HH
00.0w
mh.H
mb.0H
mm.:

00.0H

sane»:

mh.h
0m.:m
00.0
0m.mH
0m.m
mb.m

scam

newnunooo Hamew

”OHEHCflPOU u. a.“ GSEEOO

0m.0H
00.3“
mb.H
mb.0H
00.0
00.0

noon

0m.0
0m.NH
0m.h
mh.:H
0m.
00.:

noon

00.0
mN.HH
0m.H
mm.mH
mN.H

00.0

ashes:

00.5
mN.:H
mm.d
00.m
00.na
mm.m

scam

mofinucsoo omnmq

eoemmmmaao-poomnsm

0m.HH
mb.mm
0m.H
mm.:a
mm.:

0m.:H

ease»:

mm.wa
0m.:m
0m.H
mh.mH
0m.m
0m.0a

goes

mewsucsoo HHmEm

moanucnoo umficneeoo

vofiwwmmeaounoucoEmLomxm

00.0

00.0H

0m.0H

mw.mH

0m.

mm.

Loom

newnucsoo umwcseeooucoz

mm.m
0m.b
mh.H
mh.m
mm.m

00.0

Esfivmz

mb.0
mm.:H
0m.H
0m.0a
mm.mH
mm.m

coax

newnpcsoo ownmq

mofipucooo unficseeooncoz

monoow peonch ommno>< “czowxmosm nouommnssom

nan mqm<k

00.0
mh.mw
0m.H
0m.m
mb.m
00.0

noon

00.b
00.5H
0m.m
mm.m
mb.m

00.0

scams:

0m.

00.H
mm.0
00.0
00.m
00.0

nowx

mvmnucsoo HHme

00.0

mh.wm

0m.H

00.0

mw.m

00.0

9000

mm.0
mw.hH
mN.H
0m.0H
mm.

0m.

sense:

om.
mu.
0m.H
00.H
mm.“
00.0

noes

mmHnucsoo HHmEm

.me
OE<

.UcH

e§W

Ohm

.me
.E<
.00H
.50
.%MM

Ohm

80

In neither case is the interaction significant.

With only four observations in each cell, there are very few
significant differences in the tables. Here t—tests will be abandoned
in favor of a search for an examination of the directions of increases
and decreases in threat scores.

When pOpulation, energy and nationality are controlled, in 32 of
the 36 cases there are increases from non-Communist to corresponding
Communist cells. In the cases of small, rich countries and large
medium energy countries, there are no differences for Indians. And in
the case of large, rich and large, medium countries, there are larger
scores for non-Communist countries for Egyptians.

When the other three factors are controlled, the effects of the
total population of a country interact strongly with nationality at the
medium energy level, but interact very little with nationality in the
case of rich and poor Communist countries. This pattern is stable across
both tables. For all six subsamples, the means for large Communist
countries--both rich and poor--is greater than the means for small
Communist countries.

It is even more difficult to generalize about the effects of
energy level upon threat scores in these tables. The Chinese, Indian
and American subsamples tend to see large, poor countries as more
threatening than small, poor countries. And in the case of the American
subsample, poor countries are almost always seen as more threatening
than rich countries. This tendency is also visible in the case of the

U.S. experts.

81

Little can be added about nationality effects which has not been
visible in previous tables: There is a strong tendency for the American
subsample to have the highest threat score in every category. And in
every category the scores of U.S. experts are lower than those of the
American subsample. Egyptians and Indians tend to have relative low
scores in all categories.

To summarize the results which have been presented, and to display
the relative differences between mean squares, error terms, and observed

F values, these elements in the two AOV tables will be presented to-

gether here:
TABLE 32c

Experimenter-Classified Analysis

 

 

Source__of Variance Mean Square M f; if; £315-
Ideology 3186.6806 132.7963 23.9968 1/18 “.“1
Population 589.3888 35.81“8 16.“566 1/18 “.“1
Ideology—Population l“.2222 10.1666 1.3989 1/18 “.“1
Energy 506.6563 20.“977 2“.7177 2/36 3.26
Ideology-Energy 92.9201 17.6019 5.2790 2/36 3.26
Population-Energy 597.3160 22.9“68 26.0305 2/36 3.26
Ideology-Population-Energy 212.52“3 18.69““ 11.3683 2/36 3.26
Nationality 1779.8083 503.0“63 3.5381 5/18 2.77
Ideology-Nationality 301.9972 132.7963 2.27“1 5/18 2.77
Population-Nationality 50.9222 35.81“8 1.“218 5/18 2.77
Ideology-Pop.-Nationality 8.2222 10.1666 .8087 5/18 2.77
Energy—Nationality 86.2771 20.“977 “.2091 10/36 2.10
Ideology-Energy-Nationality sa.uu93 17.6091 3.0366 10/36 2.10
Population-Energy-Nationality 9“.6618 22.9“68 “.1253 10/36 2.10
Ideol.-Pop.—Energy-Nationality 21.8618 18.69““ 1.169“ 10/36 2.10

82

Of the eight error terms used in this analysis, it can be seen that
the one for nationality is, by far, the largest. And despite this, the
main nationality effect is significant. (In the following pages, the
results of a Q factor analysis show that three of the nationality sub-
samples load highest on separate factors, while a fourth is Split into
two factors.)

Of the three very large observed F values, two are for main
effects and the third is for population-energy interaction.

when each of the 288 observations in the experimenter-classified
analysis is correlated with the corresponding observation in the
subject-classified analysis, the product moment correlation between the
two sets of observations is .87. Despite this high correlation, there
are five cases among the fifteen sources of variance in which one raw
score analysis produces a significant effect and the other does not.

Only in the case of ideology-nationality interaction are the
results significant in the subject-classified analysis and not sig-

nificant in the eXperimenter-classified analysis.

TABLE 33

Differences Between Two Analyses

 

 

Leve1§_of Significance Difference
Experimenter Subject
Classified Classified
IdeolOgy .01 .01
Population .01 .05
Ideology-Population n.s. n.s.

Energy .01 n.s. E.C. greater

83

TABLE 33 (continued)

 

Ideology-Energy .01 n.s. E.C. greater
Population—Energy .01 n.s. E. C. greater
Ideology-Population-Energy .01 n.s. E.C. greater
Nationality .05 .05
Ideology-Nationality n.s. .05 S.C. greatgg,
Population-Nationality n.s. n.s.
Ideology-Population-Nationality n.s. n.s.
Energy-Nationality .01 n.s. E.C. greater
Ideology-Energy-Nationality .01 .05
Population-Energy-Nationality .01 .01
Ideol.-Pop.-Energy-Nationality n.s. n.s.

Some Additional Findings

In the analyses of variance, subsamples of people and pairs of
countries have been treated as units of analysis. On this and the
following pages additional data will be presented in terms of individual
people and countries.

One assumption of this investigation is that the threat repre-
sented by a country is related to the ideology, pOpulation and energy
level of the country as well as the nationality of the respondent. Intra-
individual correlation coefficients should yield an indication of whether

these relationships hold for this sample of people and countries:

8“

TABLE 3“

Intra-Individual Correlations Between Sets of Threat Scores
and Sets of Ideology, Population and Energy Estimates for 2“ Countries

High Threat- High Threat- High Threat-High
Communist IdeolOgy Large POpulation Energy Consumption
Brazilian #11 .70 .“6 -.10
Brazilian #l2 .“8 .57 -.1“
Brazilian #13 .7“ .31 .06
Brazilian #1“ .53 .63 -.16
Egyptian #21 .22 -.18 -.““
Egyptian #22 .21 .73 .68
Egyptian #2“ -.25 .39 .76
Egyptian #2“ .09 .07 .02
Chinese #31 .68 .33 -.“6
Chinese #32 .29 .7“ .38
Chinese #33 .53 .82 -.2“
Chinese #3“ .68 .22 -.“2
Indian #“1 .26 .67 -.19
Indian #“2 .“8 .9“ -.15
Indian #“3 .17 .62 -.07
Indian #““ .18 .““ -.02
American #“1 .61 .25 -.33
American #52 .87 .27 -.“9
American #53 .75 .“6 -.“1
American #5“ .73 .36_ -.““
U.S. Expert #61 -.OS .35 .52
U.S. Expert #62 .63 .62 -.23
0.8. Expert #63 .75 .2“ -.69
U.S. Expert #6“ .55 .7“ -.16

When N=2“, r must be greater than .3““ to be significantly
different from zero. 0f the 72 correlation coefficients presented,
only 60% are significant. Moreover, there seems to be some patterning
by nationalities. In no case is the assumption of a relationship between
threat and ideology validated for the Egyptian subsample, and in only
one case is it validated for the Indian subsample. In neither the

Indian nor the Brazilian subsamples is the threat-energy relationship

85

validated. However, in some cases the low linear r may be due to
curvalinearity indicated in earlier analyses.

0n the positive side, the strongest relationship between threat
and population is in the Indian subsample, and the strongest relation—
ship between threat and ideology is in the American subsample. There
is only one case (Subject #2“) in which none of the three independent

variables is related to threat.

Hultiple Correlations
When the three independent variables are combined or paired, how
well do they predict the threat scores of an individual? In the following

table a multiple r for the three is followed by the corresponding r2,

2
and then by P for the three possible pairs of independent variables:

TABLE 35
Multiple Correlations: Three Predictors of Threat
M 1 1 1 v 1 ‘ 1 ”2 ”2 ”2
u t P e .11 tfip 9 Delete Delete 61 te
r r Energy Pop. I eo ogy

Brazilian #11 .77 .59 .57 .58 .3“
Brazilian #12 .68 .“6 .“3 .23 .“0
Brazilian #13 .87 .75 .57 .75 .10
Brazilian #1“ .73 .53 .5“ .29 .“2
Egyptian #21 .“7 .22 .12 .22 .20
Egyptian #22 087 975 .5“ eug e75
Egyptian #2“ all 001 001 001 .00
Chinese #31 e71 051 a“? 0'49 036
Chinese #32 .82 .68 .55 .38 .65
Chinese #33 .93 .86 081 028 081
Chinese #3“ .69 .“8 .“7 .“7 .26
Indian #“1 .73 .5“ .“6 .07 .50
Indian #“2 .98 .95 .89 .23 .95
Indian #“3 .69 .“7 .“0 .03 .““

Indian #14“ 0‘26 e21 020 .03 021

86

TABLE 35 (continued)

American #51 .61 .37 .37 .37 .19
American #52 .87 .77 .76 .76 .3“
American #53 .79 .62 .62 .57 .“1
American #5“ .76 .57 .5“ .55 .35
U.S. Expert #61 .61 .37 .15 .29 .37
U.S. Expert #62 .75 .57 .57 .“2 .““
U.S. EXpert #63 .82 .67 .57 .67 .52
0.8. Expert #6“ .80 .6“ .6“ .32 .57

The multiple correlation coefficients in the first column are
significantly different from zero in 21 of the 2“ cases. The exceptions
are subjects #21, #2“, and #““. For 16 of the 2“ subjects, the three
predictor variables can account for over half of the variation in
threat scores.

Concerning the last three columns, there is no clearly discernable
pair of independent variables which is better than the other two. Only
in the case of the Indian subsample does there seem to be a notable

nationality effect here.

Q Analysis

The previous section has examined the relationship between ideology,
population, and energy and the dependent variable, threat. 'In order to
examine the relationship between threat and the nationality of the respon-
dents, a Q factor analysis of threat scores was used. In Table 36, a
factor represents a grouping of subjects around a common pattern of threat
scores. (The highest factor loading for each subject is underlined).

The Kiel-Wrigley criterion indicated a four-factor solution, but

the five-factor solution was selected since it more clearly discriminates

87

four of the five nationalities in the sample of people. Here are the

rotated factor loadings:

TABLE 36

Analysis: Rotated Factor Loadings
Factor Commu—
Pactor III Factor IV V(Egyp- na ities
(Egyptian I) (Brazilian)tion II)(h )

Factor I Factor II
(American) (Indian)

Brazilian #ll .39 .30 .17 .80 .00 .91
Brazilian #12 .22 .2'4 .90 .17 .16 .89
Brazilian #13 .52 .06 .22 .“8 .35 .67
Brazilian #19 .76 .15 .35 .23 -.11 .79
Egyptian #21 .29 .26 .22 .03 ._3 .57
Egyptian #22 .13 .23 .89 .28 .01 .9u
Egyptian #23 .ou. -.10 .9"! -.09 -.oo .87
Egyptian #214 .00 .07 .TO' -.33 .31: .67
Chinese #31 .7“ .31 .21 -.20 -.09 .7“
Chinese #32 [T8 .28 .69 .“l -.l“ .77
Chinese #33 .u2 .51 .35 .59 -.16 .82
Chinese #39 .73 .12 -.05 .73” .29 .68
Indian #91 .06 .97 -.02 .15 -.02 .96
Indian #92 .25 .‘7‘6' .10 .50 .01 .90
Indian Ma .05 .2 .01 .15 -.03 .97
Indian #““ .06 .22_ ~.Ol .16 -.02 .97
American #51 .22_ -.ll .23 .33 .18 .82
American #52 .§Z_ .01 -.O9 .3“ .10 .89
American #53 .83” .28 .ll .22 -.0“ .85
American #5“ .§2_ .16 .0“ .3“ .35 .73
U.S. Expert #61 .0“ —.12 '§2. .21 .08 .85
U.S. Expert #62 .g .08 .38 .§_p_ .oo .86
U.S. Expert #63 .§_‘{_ -.00 -.1u .33 -.18 .91
U.S. Expert #6“ .“2 .38 .21 '22. -.Ol .88
% of variance 2“% 18% 15% 19% 6%
Total = 83%

This five-factor solution accounts for 83% of the variation in
threat scores. It is relatively easy to interpret in terms of nationality

subsamples. However, the Chinese and the 0.8. expert subsamples do not

88

produce factors of their own.

By weighting the highest loading subjects on each factor appro-

priately, it is possible to produce a set of threat scores for each type

of subject. These findings will not be reported in detail, but rather

will concentrate on the countries that each type sees as more or less

threatening than does any of the other types.

(Eight countries which

were consensus items for the five types will be eliminated from the

discriminating items.)

TABLE 37.

More and Less Threatening Countires for Nationality Types

More Threatening,

North Viet-Nam
North Korea
Indonesia
Albania

Pakistan
Switzerland

China
Cuba
East Germany

Soviet Union
United States
France

Hungary
Congo
Mexico

New Zealand

Factor

I (American)

II (Indian)

IV (Brazilian)

III (Egyptian I)

V (Egyptian II)

Less Threatenigg’

Switzerland
New Zealand
United States
Mexico

Cuba
North Viet-Nam
Soviet Union

Congo
France

Albania
Hungary
North Korea

Indonesia
Pakistan
East Germany
China

The eight countries for which threat scores were relatively stable

across the five factors were Algeria, Czechoslovakia, Formosa, Japan,

89

Poland, Roumania, South Viet-Nam and Yugoslavia. Weighted threat scores
for these countries tended to be in the middle of the threatening-not
threatening scale, and slightly to the not threatening side. Five of
the eight were experimenter-classified as medium energy consumption
countries.

By weighting the standardized scores for other variables in the
same way that the threat reaponses were weighted for each of the five
factors, it is possible to make further generalizations concerning
relationships between variables and how these relationships vary from
one nationality to another. Some findings concerning Factor I (American)
responses will be reported as an example:

The American "type" here sees North Viet-Nam, Indonesia and
Yugoslavia not only as more threatening than do the other "types" but
also as more Communist than they do. North Korea is seen as more
threatening, more Communist and poorer. Algeria is seen by the American
"type" as more threatening, more Communist, poorer and more different,
relative to the responses of the other nationality types.

On the other hand, Switzerland is seen as less threatening and more
similar. France is seen as larger and richer. New Zealand is seen as
less threatening, more non-Communist, richer and more similar than it
is seen by other types. And Hungary is seen as less Communist and more
similar in nationality.

However, not all of the reSponses fall together so neatly. For
the American type of response, Poland is seen as more threatening and
larger, but also as more similar in nationality. Cuba is seen as larger

and more different. Japan is seen as less threatening and smaller while

90

Mexico is seen as less threatening and larger. Mexico is also seen as

richer and more different.

Similarlity
There is one set of responses for which results have not yet
been reported. Subjects were asked to say which of the 2“ nationalities
of the country sample were most similar and most different from their
own in terms of an eight-point scale. The correlations of these reaponses

with threat scores and ideology estimates were as follows:

TABLE 38

Correlations of Similarity Estimates
with Threat and Ideology Estimates

Low Similarity- Low Similarity-
High Threat Communist IdeOIOgy
Brazilian #ll .52 .67
Brazilian #12 .02 .31
Brazilian #13 .23 .3“
Brazilian #1“ -.35 .0“
Egyptian #21 .53 .29
Egyptian #22 .oo .01
Egyptian #23 .“0 .00
Egyptian #2“ .0“ .2”
Chinese #31 .01 .29
Chinese #32 -.3“ .16
Chinese #33 -.30 .18
Chinese #3“ .29 .55
Indian #“1 -.“5 -.06
Indian #“2 .“1 .“1
Indian #“3 -.50 .09
Indian #““ -.17 -.10
American #51 .“9 .56
American #52 .“1 .“3
American #53 .50 .56
American #5“ .“5 .27
U.S. Expert #61 -.35 -001
U.S. Expert #62 .36 .57
U.S. Expert #63 .63 .3u

U.S. Expert #6“ .19 .u7

91

A low similarity score indicates that a nationality is very
similar to that of the subject. Thus the American subjects are most
likely to score a less similar nationality as belonging to a more
threatening country. The opposite tendency can be noted in three of the
Indian subjects.

The similarity-ideology correlations should give a rough indication
of the ideology of the subject. Here a high correlation indicates that
the same countries have been judged as non-Communist and "similar" in

nationality.

Correlations with Criteria

In the construction of the country sample, United Nations data
on national populations and per capita energy consumption were utilized.
As an indication of how much the estimates of each subject varied from
the U.N. figures, correlation coefficients were computed for the
estimates of each subject and the criteria.

The third column below consists of a similar indicator: the
number of cases in which the ideology, population and energy responses
of the subject resulted in classifying a country in the same way that
it was classified in the country sample constructed by the experimenter
on the basis of U.N. data.

Estimates concerning populations were obviously better than
estimates concerning per capita wealth or energy consumption. Ideology
estimates of each subject were correlated with every other subject,
and produced correlation coefficients above .70, except for Subject #““.
A similar correlation matrix for energy consumption estimates showed

less consensus in this sample for this variable than for ideology.

92

TABLE 39

Correlations of Subjects' Estimates and U.N. Criteria Data

Population Energy Classification
Correlation Correlation Score
Brazilian #11 .7“ .“6 9
Brazilian #12 .97 .62 12
Brazilian #13 1.00 .73 16
Brazilian #1“ 1.00 .86 15
Egyptian #21 .79 .69 13
Egyptian #22 .99 .62 10
Egyptian #23 .90 .62 9
Egyptian #2“ .95 .52 9
Chinese #31 .97 .66 12
Chinese #32 .99 .5“ 16
Chinese #33 .97 .62 13
Chinese #3“ .97 .63 7
Indian #“1 .99 .56 15
Indian #“2 .99 .60 10
Indian #“3 .97 .66 9
Indian #““ .7“ .6“ 6
American #51 .92 .69 16
American #52 1.00 .63 13
American #53 .99 .62 11
American #5“ .97 .61 ll
U.S. Expert #61 1.00 .81 13
U.Se Expert #62 099 070 12
U.S. Expert #63 .99 .72 9
U.S. Expert #6“ 1.00 .7“ 18

From the results in the previous table, it would seem that the
U.S. experts live up to their title, although the Brazilians fare almost
as well. Although the variability within subsamples may be as great
as that between subsamples, the Egyptians and the Indians seem to be
the lowest scoring subsamples for this table.

It is also possible to array the classification scores in another

way to show which kind (subject-classified) of countries are classified

93

in accord with U.N. data:

TABLE “0

Classification Scores by Country Characteristics

Non-Communist, small, rich 3“ Communist small, poor 17
medium 17 medium 16

poor 31 rich 1?

large, rich “1 large, poor 23

medium 2“ medium 11

poor “0 rich 23

There is more subject-U.1. data agreement in the non-Communist
than in the communist countries, and more agreement concerning large
countries than concerning small countries. It is also interesting to.
note that in each group of three countries, the medium energy country
has the lowest score. This tendency is stronger in small, non-Communist

and large, Communist countries.

World Government
On the final page of the questionnaire, subjects were asked to
indicate their degree of agreement with a statement that their country
should give up some of its sovereignty to a world government. The
hypothesis was that low average threat scores would be associated with
strong agreement with the world government statement. The results

were 3

9“

TABLE “1

Mean Threat Scores for Subjects Classified by Responses to
World Government Statement

Mean Threat

N Score
Strongly agree 7 13.58
Agree 8 9.8“
Uncertain “ 7.71
Disagree 2 9.00
Strongly disagree 3 “.28

The results are in the opposite direction of what was hypothesized:
strong agreement with the world government statement is associated with
high threat scores. This is partially explained by the differences in
mean threat scores of subsamples which have already been examined.

When "strongly agree" is given a score of one and "strongly disagree"

is given a score of five, the results by subsamples are as follows:

TABLE “2

Subsample Means for World Government and Threat Scores

World Government Threat
Mean Mean
Brazilians 2.25 6.08
Egyptians 3. 25 s . 15
Chinese 2.25 l“.15
Indians 3.00 “.“0
Americans 2.25 19.87

U.S. Experts 1.50 9.23

95

The greater disagreement of Egyptians and Indians with the
world government statement, and their relatively low threat scores is

a partial explanation for the previous table.

CHAPTER V

DISCUSSION

We are here, as on a darkling plain
Swept with confused alarms of struggle and flight

Where ignorant armies clash by night.

Matthew Arnold

This study has been an effort to examine the nature of threat
in an international context in terms of its relationships to several
other variables. Organization theory suggests that threat may be defined
as a relationship between an organization and external elements or
events which tend to drive certain variables of the organization away
from their preferred values.

A communication model of the international situation was suggested

n which the function of threat was to increase boundary maintenance

He

or to decrease systemic linkage between countries. It was also suggested
that groups of countries might be examined as organizations, and that
one check on their decree of organization would be that there be consensus
across nationalities concerning countries which are threatening.

To test this hypothesis-~83 well as others to be restated--a
sample of subjects made up of equal subsamples of graduate students of
different nationalities was selected. Deepite the fact that the

nationality of the subject was of primary interest, such a small-scale

96

97

study can make no claim to generalize findings to university graduates
of different nationalities. The investigation must be considered an
exploratory one, and findings only as suggestive of directions for
further research.

The sample of people was necessarily made up of non-Communist
nationalities, and this was capitalized upon in the case of the hypothesis
suggested above. One indication of organization in the non-Communist
world would be consensus across several nationalities that Communist
countries are threatening.

In order not to hypothesize an obvious relationship, five nationality
subsamples were selected not only on the basis of geOgraphic diversity
but also to include nationalities which might be considered as neutral
in terms of a Communist/non-Communist dichotomy. The nationalities in
the sample were Brazilians, Egyptians, Indians, (Formosan) Chinese, and
Americans.

In selecting a sample of countries which would become items in a
questionnaire, two other criteria were added to ideology. These were
the size of the national pepulation, and the per capita energy consumption
of the country. It was hypothesized that, in this multinational sample,
Communist countries would be seen as more threatening than non-Communist
countries and that large countries--regardless of ideology-~wou1d be seen
as more threatening than small countries. (The more complex factor of

per capita energy consumption will be discussed later.)

IdeolOgy
As expected, it was found that the ideology of a country had a

very strong effect upon how threatening a country was judged to be.

98

Communist countries were seen §s_cgnsiderably_more threatgning than
non-Communist countries. Moreover, when the countries were Split into
Communist and non-Communist groups on the basis of the judgments of each
subject, the effect was slightly stronger.

This is to say that there was not complete agreement in the sample
of people concerning which countries should be classified as Communist,
but once a country was perceived as Communist it was very likely to be
judged as relatively threatening.

When population and energy consumption are controlled, the effect
of ideology on threat scores is shown to be influenced by these factors.
If a Communist country is large it is seen as somewhat more threatening
than if it is small, but this must be modified by the finding that both
rich and poor large Communist countries are seen as more threatening
than those of intermediate wealth. This can be partially explained by
the high threat generally attributed to both the Soviet Union (relatively
rich) and China (relatively poor).

When the nationality of the person reSponding is brought into the
analysis, this has a definite influence on the importance of ideology
in determining threat scores. For the Indian and Egyptian subsamples,
ideolOgy doesn't make much difference. For the Chinese and Brazilian
subsamples, it has a greater effect; and for the American subsample it
has the greatest effect.

These results are in line with the findings of Lazarus that it is
not the stimulus characteristics alone which define threat, but rather
the way in which they are interpreted. These results also suggest that

a case for non-Communist "organization" can be made for subjects in the

99

American, Brazilian and Chinese subsamples, but the case is much weaker
for subjects in the Indian and Egyptian subsamples.

The similarity-ideology correlationships (Table 38) are fairly
reliable indicators of the importance of ideology in influencing judgements
about threat. (Compare Tables 2“~25 with Table 38). The correlations
of the American subsample are highest, followed by those of the Brazilian
and Chinese subsamples. Those of the Egyptian and Indian subsamples
are lower; two Indian subjects have slightly negative correlations.

If the similarity-ideology correlations are interpreted as in-
dicators of the ide010gy of the re5pondent along a Communist to non-
Communist scale, then the influence of ideology as a stimulus characteristic
must be reinterpreted as a relationship between the ideology of the
respondent and the perceived ide010gy of the country.

In these terms the subjects in this study with similarity-

Po

ideology correlations near zero can be defined as not 3 gnificantly
linked to the Communist nor the non-Communist "world"--or organization
in the sense that it has been used here. That is to say that their
behavior is not significantly constrained by communication with either
of these organizations.

For example, two of the Egyptian subjects and three of the Indians
subjects have ideology-similarity correlations which are quite close to
zero. The highest threat-ideology correlation among these five subjects
is .26. In contrast, the lowest threat-ideology correlation in the
American subsample is .61.

In terms of the theory which has been cited, these data can be

interpreted as indicators that the American subjects are highly con-

strained parts of a non-Communist system or organization while the Indian

100

and Egyptian subjects tend to be on the periphery of this system.

One difference between the effect of ideology and the effect of
population and energy level should be noted. Population and energy effects
tended to be attenuated in the subject-classified analysis, while
ideology effects sometimes increased slightly. This was observable in
the Chinese, the American and the Indian subsamples. The Opposite
occurred in the case of the Brazilian subsample.

One unexpected finding about ideology was that there was more
consensus in this multinational sample concerning ideolOgy than there was
concerning per capita wealth or energy consumption. Subjects rated the
governments of 24 countries on an 8-point scale from most Communist to
most non-Communist in a forced normal distribution. In an intercorrelation
matrix, all correlations were above .70 except those for Subject #uu--a
striking deviant in this respect.

(A similar matrix for per capita energy consumption produced
notably lower correlations, but they were based on a lS-point scale in
which distribution was not forced.)

There are no widely accepted criteria for judging which govern-
ments are most to least Communist and most to least non-Communist, but
there seems to be a fair degree of consensus across nationalities in

this sample about this.

Population
Generally Speaking, large countries were perceived as more
threatening_than small countries but this must be qualified. The most
striking exception is that threat scores for Mexico, Japan, Yugoslavia

and Roumania--classified as relatively large-~were lower than those for

101

Formosa, Algeria, Cuba and Hungary--classified as small. All eight of
these countries were experimenter-classified as medium per capita energy
countries-~neither rich nor poor.

In the correSponding case in the subject-classified analysis, the
unexpected reversal in the direction of threat scores disappears, and
there is no significant difference in the threat scores for large and
small countries in the medium energy category.

In interpreting these findings, it is helpful to examine the
scores which were developed to indicate that a subject's classification
of a country was in agreement with criteria data. Scores for medium
energy countries are relatively low. This helps to explain the differences
in the two analyses mentioned above.

When the eight countries mentioned above are segmented into
Communist and non-Communist pairs, the small countries are still con-
sidered the more threatening by the subjects in this sample.

In general, the large-small breakdown in countries makes a
contribution to the variation in threat scores which is independent of
the ideology of the country and the nationality of the respondent.
However size of population is strongly affected by the per capita energy
consumption of the country.

The kind of subject-country relationship noted in the ideology
results is less visible in the case of population. Across the two
analyses it is the Chinese and Indian subsamples that make the greatest
small-large distinctions in attributing threat to certain countries.
That they both agree on the strong threat posed by mainland China is only
a partial explanation since there are 12 countries in each category when

scores are broken down by nationality and pOpulation.

102

An examination of the threat—population correlations (Table 3a)
shows the relationship between perceived threat and estimated population
of all 2n countries for each individual. There is a great deal of
variation within sub-samples, but correlations for Indians are .uu and
higher while correlations for Americans are .u6 and lower. But with
the exception of one Egyptian subject, all threat-population correlations
are positive.

In the case of ideology, it seemed to be the most non-Comnunist
subjects who saw Communist countries as most threatening. At least the
data from this study has been interpreted in these terms. But in the
case of population, it is the subjects from a very large country who
tend to see large coun rics as threatening. The significance of this
finding, however, may be nation-specific: Indians tend to see China and

Pakistan as threatening, and they are both large countries.

Energy

The hypothesis here was that there would be no main energy effect
on threat scores, but an interaction with ideology. In the case of
the non-Communist countries, it was expected that threat would increase
from rich to medium to poor countries, and in the case of Communist
countries it was hypothesized that threat would increase from poor to
medium to rich countries. This did not prove to be the case. In this
study energy was much more complex in its effects.upon threat scores
than either ideology or population.

In the first place, several significant differences between the
experimenter-classified and subject-classified analyses involve the

energy factor. In five sources of variance in which the experimenter-

103

classified analysis produced significant differences while the subject-
classified analysis did not, the energy factor was involved.

The experimenter-classified analysis produced a significant main
energy effect, a significant interaction of energy with ideology,
population and nationality plus a triple interaction of ideology-pOpulation-
energy. In the subject-classified analysis, none of these effects were
significant.

Perhaps the most striking difference between the two analyses is
that the main energy effect in the experimenter-classified analysis was
significant at the .0005 level while in the subject-classified analysis
it was insignificant, i.e., there were no significant differences in
threat scores when subjects divided the 2n countries into three categories:
rich, medium, and poor in terms of per capita energy.

Generally, poor countries were segp_as more_thgeatening_thanggich
countries,_and these as more threatening than medium eneggy countries.

The direction of these differences was stable across the two
analyses. When controls are added, this direction holds for Communist
countries and for large countries but not For non-Communist nor for
small countries. When nationality of the subject is added as a control,
the Specified direction hold for all six subsamples in the case of
Communist countries, but only in the experimenter-classified analysis.
It holds for four nationalities in the case of large countries.

In the light of these findings, the value of the threat-energy
correlations for each individual (Table 34) must be questioned, since
the assumption of linearity is not always fulfilled. However the fact

that 18 of the 2H are negative validates the analysis of variance finding

101‘

that poor countries are generally seen as nost threatening.

The consensus its-s across the five typos of threat responses
found in the Q analysis can also be interpreted in terms of on energy
offoct. The eight countries for*which throat scores were relatively
stable across five typos of responses included five countries which are
experimenter-classified as sodiul energy consumption countries: For-cos,
Algeria, Japan, Yugoslavia and Ron-unis. The general tendency was to
consider then as neither completely non-threatening nor very threatening.

One further interpretation is possible fro-utbo analysis of throat
scores in terns of energy levels with the nationality of the subject
as a control. The Anoricsn subsample and the U.S. exports seo rich
countries as least threatening, sodiun energy countries as more threatening
and poor countries as most threatening. The Chinese subjects also tend
to saintain this order. In contrast, the Brazilian and Egyptian subjects
soo rich countries as most threatening. The Indian subsample is like
the American subjects in seeing poor countires as lost threatening,
but they rank rich countries second.

This complex interaction of energy level with nationality is difficult
to interpret, but there is an indication that subjects from a rich country
are less likely to see rich countries as threatening than subjects from
relatively poor countries. The Chinese agreement with the American
pattern is indicated also by the results of the Q factor analysis: two
of the Chinese subjects load very high on the ”American" factor, on in-
dication that they nay have been more "A-oricanizod," at least in terms
of their threat responses.

Perhaps the effect of energy level of country upon perceived

throsto-ss indicated in those findings-csn best be generalized in tor-s
which take both stimulus country and subject nationality into account.

lCS

An energy hypothesis based on these findings would predict that
the perceived threat of a country increases as per capita energy con-
sumption increases above a certain point and also increases as per capita
energy consumption decreases below this certain point. The influence of
the reopondent would be a factor in defining the point of minimum threat.
For peeple from rich countries the point would be higher and for people
from poor countries it would be lower.

The rationale for such a hypothesis would be that in terms of
calories,_clothing and shelter, human beings require a certain amount of
energy per year in order to survive. The average person in a highly
industrialized country consumes perhaps a thousand times as much energy
per year as such a person in a very poor country. Countries which are
poor are seen as threatening to other countries because they want to
increase their consumption of energy, and may be perceived as wanting to
increase it at the expense of other countries. Countries which are, say,
above the world median in per capita energy consumption are seen as
threatening because of the surplus energy at their diSposal which may be
used for aggressive purposes.

It has been found that other factors interact with energy, so
such a hypothesis must be qualified. It will have a stronger effect
for large countries than for small countries, and it will have a stronger
effect for countries perceived as having a different ideology than for
countries having the same ideology.

An alternative hypothesis would eliminate the per cap:ta aSpect of
energy consumption and use a total for each country. In terms of the

responses elicited in this study, such a figure for each country for

106

each subject could be obtained simply by multiplying the estimated
per capita energy consumption by the estimated population of a country.

If power is defined as energy release per time unit, and energy
consumption is equated with energy eXpenditure, then this variable might
better be defined as national power. In retrOSpect, this variable
should be a more appropriate kind of measure to relate to threat and
ideology (and similarity) which treat a country as an undifferentiated
unit of analysis.

A disadvantage of this procedure would be that, according to
findings for this sample, there is little variability in pOpulation estimates
of countries across subjects while there is greater variability in
estimates concerning wealth. A power factor would obscure the contri-
bution that wealth estimates make in increasing the variability, "error,"

and uncertainty in a given sample of subjects.

Nationality

There are a number of factors involved in this study which might
conceivably reduce the effect of the different nationalities oF the
subjects in this study. Except for one subsample, subjects were all
graduate students in an American university. The questionnaire was in
English. The experimenter was an American, and an "aquiesence effect"
might have been produced. American mass media were the most available
source of information about international events. The directions for
the scale on sources of international threat and danger did not ask
about threat to the subject's country but rather in terms of "you and
your family."

DeSpite all of these things, nationality proved to be the strongest

factor in the analysis.

107

An unexpected result was a main nationality effect. This has
various possible interpretations. One is that the differences in mean
threat scores across nationalities indicate reliable differences in
terns of perceived international threat.

If the responses are taken at face value, the relatively high
threat responses of the American subjects night be taken as an indication
that they live in a more threatening international environment than
the other subjects. An assumption here is that the country sample
is indicative of the international environment in general.

In this view the Indian and Egyptian subjects perceive their
international environ-ant as relatively non-threatening; the Chinese
responses were sore similar to those of the Anericans, and the Brasilian
subsample was in between these two pairs.

It is interesting to note that the subsample of U.S. experts
was closest to the Brazilians in terns of both lean and standard
deviation of threat scores, and significantly lower than the Marina
subsample.

It is difficult to say whether this was due to expertise or to
other factors. The four Michigan State faculty nenbers which made up
this subsample were some 10 to 15 years older than the graduate students
in the Aserican subsample, and this is a possible source of difference.
The relatively high accuracy of Brasilian and 0.5. expert responses in

terns of U.N. criteria data lends sole support to the interpretation that

108

this intermediate level of perceived international threat is associated
w;th greater knowledge or the international situation.

if the threat scale used here may be equated with Hammes' "danger-
aggression-threat" scale, then subjects in the American subsample must
be considered as manifesting higher anxiety than any other group of
subjects, including the 0.8. experts. It seems entirely possible that
Ph.D. candidates in the American subsample might exhibit greater manifest
anxiety than the faculty members who reaponded to the questionnaire.

An entirely different interpretation would be to assume that,
subject matter aside, persons of different nationalities react to such a
questionnaire in different ways. Such subject-measuring instrument
interaction might lead Indian subjects to tend to make X's on the left-
hand side of the page and American subjects to check near the middle.

It is certainly possible that such an effect contributes to the
differences in scoring patterns across nationalities. However, this makes
the differences between the two subsamples of Americans more difficult

to explain.

If differences in mean threat scores of individuals are to be
interpreted as a hint of national differences in anxiety levels, then
the reSponses to the world gOVernment statement add difficulties to this
interpretation.

It was hypothesized that individuals with low mean threat scores
would tend to agree with the world government statement, and that in-
dividuals with high mean threat scores would tend to disagree. The results
of this study show that the two sets of reSponses are related, but the

tendency is in the opposite direction of the one hypothesized: strong

109

agreement with the world government statement tends to be associated
with relatively high threat scores.

There were two reasons for the use of the world government state-
ment in this study. One was that negative responses to a similar statement
had been shown to be regularly correlated with high t‘-scale scores across
six EurOpean nationalities. The other was that agreement could be in-
terpreted as favoring "systemic linkage" with Foreign countries in
general.

Conversely, the threat scale responses may be interpreted as in-
dicating the degree of "boundary maintenance" appropriate to individual
foreign countries. In these terms, the hypothesis is that a person who,
in general, sees a low need for boundary maintenance with the countries
in this sample will favor systemic linkage with foreign countries in
general.

This did not prove to be the case, and the findings suggest that
other variables must be added to any effort at explanation. Nationality
of the reSpondent is probably one such variable. Egyptian and Indian
subjects tend to have low mean threat scores and also tend to disagree
with the world government statement. American and Chinese subjects
tend to have high mean threat scores and tend to agree with the world
government statement.

Brazilians and 0.8. experts have intermediate levels of threat
scores and tend to agree with the world government statement. It is the
U.S. experts group which shows the strongest agreement of the six sub-

samples.

110

The extremely small number of subjects in this sample make any
generalizations questionable. And, in the case of the world government
question, the use of one response per subject must be considered as a
very crude index of a general attitude concerning foreign countries.
Still, the direction of the findings at least allows some basis for the
formation of new hypotheses.

It is perhaps better to interpret the mean threat scores of in-
dividuals as indicative of an arousal level concerning international
affairs. If this is quite low, as in the case of the Indian and
Egyptian subjects here, it may indicate that there is little attention
to or interest in events outside the home country. The national system
is perceived as relatively independent of other national systems, and
most attention is on events within the national system. External events
are noticed in gross, black and white terms. A few countries are seen
as very threatening, and the rest are hardly noticed.

The independence of the national system is valued, and there is
little desire for systemic linkage with other countries. There is
relatively little information about the external world.

On the other hand, arousal level concerning events external to
the national system may rise so high that anxiety is a chronic condition,
and many foreign countries are then seen as quite threatening. This is
more likely in an industrialized country with a great deal of regular
communication with other countries. Here there will be greater knowledge
of foreign countries and international events, and there will be finer
discriminations concerning the threat associated with individual Foreign

countries. In this stud= the American and Chinese subsam les anoroximate
O P Ht

this pattern of behavior.

Between these two possibilities is a third which is represented
by the U.S. expert and Brazilian subsamples in this study. Here, the
national System is also perceived as having a great deal of communicatio.
and thus interd,pendence with other national systems. But this position
is characterized by more information and less uncertainty about ex-
ternal events: international communication is perceived as having
consensual as well as conflictual elements. The result is less anxiety
about the international situation; or, in terms of this study, lower
threat scores.

These three types of response complexes are suggestive rather than
exhaustive. They are limited, for example, to cases in which a national
communication network influences the behavior of the individuals within
a country. Also, it is assumed that education has been held fairly
constant in five of the subsamples here, and this is certainly a variable
both within and between countries.

(P.A. Smith has Found that in the United States people with low
education and low prestige tend to react to crisis events and prefer
an aggressive foreign policy while better educated people with higher
prestige see international relations as a flow of events and are generally
in favor of a more moderate policy.)

The results of the Q factor analysis substantiate the important
role of nationality in the perception of international threat. The hy-
pothesis was that the threat reSponses of nationality subsamples would
produce different factors. This was best demonstrated in the case of

the American and the Indians subsamples, and to a lesser degree in the

112

case of the Brazilian subsample. The four Egyptian subjects produced
two factors.

Neither the Chinese nor the U.S. eXperts produced factors of
their own, but were distributed over the American, the Brazilian, and
one of the Egyptian factors (Factor III).

The fact that only one 0.8. eXpert had his highest loading on the
American factor might be interpreted as suggesting that knowledge at
the eXpert level reduces the constraint that is imposed by nationality
in evaluating sources of international threat.

The fact that two of the Chinese subjects have higher loadings on
the American Factor than one of the American subjects raises a question
about the "Chineseness" of these subjects. It would be interesting to
know how much time these subjects have spent in the United States in
comparison with other foreign subjects in the sample, but such information
was not requested in the questionnaire.

The 1 analysis of responses concerning international threat not
only accounted for 83% of the variation in threat scores, but also defined
the reaponses of each individual in terms of each of the five factors.
Although it is outside the bounds of this study, such a technique may
offer a quantitative approach to indexing acculturation. It does offer a
rather simple tool for quantifying international similarities and
differences in common terms, i.e., the names of countries. If nationality
of subject constrains threat responses as much as is suggested here, this
might be done with relatively small national samples.

Along with the clear results obtained from the Q factor analysis

(Table 36) in explaining variation in threat reSponses by the nationality

113

of the person reSponding, the multiple correlation coefficients (Table 35)
suggest that perceived country characteristics can also be used with
some success to predict threat responses.

The effectiveness of ideology, population and per capita energy
responses as predictors of threat scores seems to vary as much within as
between subsamples. There is at least one subject in each subsample
in which less than no% of the variation in threat scores is accounted for
by these three predictor variables.

And, except for the 0.3. experts, there is one person in each
nationality subsample in which the three selected variables account for
at least 75% of the variation in threat scores. ?erhaps the U.S.
eXperts take more things into account in judging how threatening a country
is.

Among the country characteristics which have been taken into
consideration here, ideoIOgy has been found to have the greatest influence
upon perceived international threat. Ideology is a particularly strong
predictor when it is interpreted as a relationship between the person
responding and the country being judged.

Per capita energy consumption has also been shown to be related to
the perceived threat of a country, but its effect is not a simple one.

It effect seems to increase as threat increases. Uncertainty or lack of
knowledge about the per capita wealth of a country seems to cause a
regression toward a mean or average t reat value for countries in general.

There is an indication that perhaps energy should also be in-

LJ.

terpreted as a relationship: sub ects from rich countries are less likely

to see rich countries as threatening than subjects from poor countries.

llu

The third kind 0: country characteristic examined here was
national population. This sample of people made Fairly accurate estimates
of country populations, and this variable is stable across nationalities
and regardless o? ideology. Here there seems to be the strongest case

For treating a variable as a stimulus ctaracter'stic which does have an

..

0
eFFect on perce;vei threat.

0' all the variables considered here, the n tionality or the
resyondent has the strongest eifect Upon the perceived threat attributed

to a sample 0? countries. finthropCIOgists would be pleased with the

Po

nternational threat Seem to be, above all, in

peoples.

REFERENCES

Quotations from Senator Fulbright's speech are taken from Congfiassional

 

 

 
 

 

 

 

 

 

Record, Vol. 110, No. 56, Washington, Wednesday, March 25, .

l. Hienor, N., Thg_flumagugso ofgguman Boingg, Boston: Houghton Hifflin,
195“. p. 95-960

2. Ibid., p. 96.

3. Ashby, H.R., ”Principles of tbo Self-Organizing System," Principles
Of SOIf’O’ aniz.ti.n. Ho VOD F°.r8t.r ‘nd Go we Zopf, JP.. . .0.
Now—York: Porgamon Press, 1962, p. 266.

u. Ibid., p. 263.

5. Ibid., p. 266.

6. Wilson G. and H. Hilson, The Anal sis of Social Chan o, Cambridge,
England: Cambridge Univsrs ty ross, , p. - 3 .

7. Ibid., p. 158-159.

8. Dalai Lama, The, !y_Land and My People, New York: McGraw-Hill, 1962,
Do 590

9. Banks, A.S. and R.B. Textor, A Cross-Polity Surv , Cambridge,
Massachusetts: The Massachusetts institute of Toghnology Press,
1963. p. 30

10. Sarnoff, I., "Psychoanalytic Theory and Social Attitudes," Public
Qpinion Quarterly, 1960, 2a, p. 255.

ll. Newcomb, T., The Ac uaintanco Process, New York: Holt, Rinohart
and Winston, I56I, p. 555.

12. Ibid., p. 260.

13. Ibid., p. 260.

1a. Lazarus, R.S., ”A laboratory approach to the dynamics of psychologidhl
stress," Administrative Scisncegguarterly, 1963, 8, p. 199-200.

15. Ibid., p. 200.

16. Ibid., p. 200.

17. Ibid., p. 212.

18. Hammes, J.A., "Manifest anxiety and perception of environmental

threat," Journal of Clinical Psychology, 1961, 17, p. 25-26.

115

19.

20.

21.

22.

23.
2“.

25.

26.

27.

28.

29.

30.

, 'nd “inaton. I535. po I650

116

Harvey, 0.J., ”Authoritarianism and conceptual functioning in
varied conditions," Journal 9: Porsgnality, 1963, 31, p. H70.

Newcomb, 7., op. cit., p. 139.

Berlo, D.K., The Process of Communication, New York: Holt, Rinehart

Stagner, R., "Personality dynamics and social conflict," Personality
and Social Systems, Now York: wiloy and Sons, 1963, p. use.

Ibid., p. «91.

Ibid., p. “91.

Eulau, H., The Behaviorial Persuasion in Politics, Now York: Random
Houso, 1965, p. 50.

Rosecrance, R.N., Action and Reaction in_!grld Politics, Boston:
Little, Brown and 50., 1965, p. 225.

Mason, 8.8., as quoted by Schurr, S.H. in "Enorgy," Tochnologz and
Economic Developmont, Now York: Knopf, 1963, p. 72.

Toltsch, K., "20 oountrios ask delay of U.N. assembly sossion,"
The New York Times, May 2, 1963, p. 7.

Schurr, S.H., Op. Cite. (27). Po 72o

 
 

Lindquist, B.P., Dosi and Anal sis of B orimonts in Pa cholo
ang_Education, Boston, on; ton n, , p.

 
 

B 13 LI OGRAPHY

Note: This list does not include references which have been cited.

Adler, K. and D. Bobrow, "Interest and influence in foreign affairs,"
Public Qpinign Quarterly, 1956, 20, 89-101.

Almond, G. and J. S. Coleman, et. al., The Politics of Developingm
Princeton University Press, 1960.

Ashby, W.R., An Introduction toUCybernetics, New York: Wiley, 1958.
Bernard, J., et. a1., The Nature of_Conflict, Paris, France, UNESCO, 1957.
Boulding, K.E., Conflict and Defense, New York: Harpers, 1962.

Cantril, H., ed., Tensions that Cause wars, Urbana, University of Illinois
Press, 1950.

Cantril, H. and N. Buchanan, How Nations See BachyOther, Urbana, University
of Illinois Press, 1955;

Cantril, H. and L.A. Free, "Hopes and fears for self and country," The
American Behaviorial ScientiSt, 1962, 6, (2, Supplement).

Carr, L.J., "A situational approach to conflict and war," Social Forces,
1906, 20, 300-303.

Cattell, R., H. Breul and H.P. Hartman, "An attempt at more refined
definition of the cultural dimensions of syntality in modern

nations," American Sociological Review, 1951, 17, 008-021.
Christiansen, B., "Attitudes toward forei affairs as a function of
personality," BsIo, Norway: 5510 Ufigversity Press, 1§§§.

Deutsch, K. H., "0n communications models in the social sciences," Public
Opinion Quarterly, 1952,16 (3).

Deutsch, K.H., Nationalism and Social Communication, New York: The Technology
Press, 1935.

Deutsch, K.N., Political Communisy at the International Level: Prdblems of
Definition an easurement, ew or : e ay, .

Deutsch, K.W., "Shifts in the balance of communication flows: a problem
of measurement in international relations," Public gainion Quarterly,
1956, 20, 103-150.

 

Deutsch, K.W., "Toward an inventory of basic trends and patterns in com-
parative and international politics," American Political Science
Review, 1960, SN, 30-57.

117

118
Dodd, S.C., "A social distance test in the Near East," American Journal of
Sociology, 1935, “1, 19“-20“.
Dunn, F.S., War and the Minds of Men, New York: Harper, 1950.
Free, L.S., Six Allies and a Neutral, Glencoe, 111.: The Free Press, 1959.

Gladstone, A., "The conception of the enemy," Journal of Conflict Resolution,
1959, 3, 132-137.

Grace, H.A. and G. L. Grace, ”Hostility, communication and international
tensions, III, The hostility factors," Journal of Educational
Pszchologz, 1951, “2, 293-300.

Grace, H.A. and J.0. Neuhaus, "Information and social distance as pre-
dictors of hostility toward nations," Journal of Abnormal and Social

Psychology.

Haas, 5.8. and A.S. Whiting, Dynamics of international Relations, New York:
McGraw-Hill, 1956.

Herskovits, N.J., AcculturationI The Study of Culture Contact, New York:
J.J. Augustin, l .

Hyman, H.H., Political Socialization, Glencoe, 111.: The Free Press, 1959.
Isaacs, 3., Scratches on Our Hinds, New York: John Day, l9“8.

Jansson, J.H., "The role of political ideologies in politics," Inter-
national Relations, 1959, 1.

Jonassen, C.T. and S.H. Peres, Interrelationshi s of Dimensions of
Community_Systems, ColomBus; 5510 State Efiiversity Press, 1960.

Keenan, G.F., "The illusion of security," Atlantic Monthly, August,
195“, 31-3“.

Kennan, G.P., "Polycentrism and western policy,” Porei Affairs, 196“,
“2, 171-183. (Quoted in Senator Fulbright's speech 0? “arch 25, 196“).

Kendell, P., "The ambivalent character of natio-alism among Egyptians
professionals," Public Opinion Quarterly, 1956, 20, 277-292.

Klineberg, 0., Tensions Affectin International Understanding, New York:
Social Science Research Council, I555.

Lasswell, H.D., florid Politics and Personal Insecurity, New York: McGraw-
Hill, 1935.

 

Locke, A. and B.J. Stern, When Peo les Meet: A Study in Race and Culture
Contacts, Barnes 8 Noble, I950.

119
Liska, 6., International Equilibrium, Cambridge: Harvard University Press,
1957.
Mack, R.W. and R.C. Snyder, "The analysis of social conflict: toward an

overview and synthesis," Journal of Conflict Resolution, 1957, l,
212'2u80

Morganthau, H., Politics Amon Nations: The Stru 1e for Power and Peace,
New York: Knopf, Secons edition 1955.

Nehnevasjsa, J., "Anticipating ideological conflict," The American
Behavioral Scientist, 1961, “, 12-1“.

 

Osgood, C.E., "An analysis of the cold war mentality," Journal of Social
IssuesI 1961, 17, 1-15.

Parsons, T., "Certain primary sources and patterns of aggression in the

western world," in Bryson, L., et. al., Conflicts of Power in
Modern Culture, 19“7, 29-“8.

Pear, T.H., Ps cholo ical Factors in Peace and War, New York:
PhilospEicaI Eibrary, 1955.

Pool, I. de 3., S 013 of Internationalism, Stanford: Stanford University
Press, 19 .

RapOport, A., Fights, Games and Debates, Ann Arbor: The University of
Michigan Press, .

Rogers, C.R., "Communication: its blocking and its facilitation," Btc.:
A Review of General Semantics, 1952, 9, 83-88.

 

Rostein-Rodan, P.N., "International aid for underdeveloped countries,"
The Review of Economics and Statistics, 1961.

Rothstein, J., Communication 0r anization and Science, Indian Hills,
Colorado: The Falcon wing.Press, I558.

Rummel, R.J., Dimensions of conflict behavior within and between nations,"
General Systems Yearbook, 1963, 8.

 

Saul, L.J., et. al., "Hostility, a bibliography in the psychological
sciences with supplementary references from other social sciences,"
The Americ§n_8ehavioral Scientist, 1961, “, Supplement.

Schelling, T., The Strategy of Conflict, Cambridge: Harvard University
Press, 1960.

Scott, N., "Correlates of international attitudes," Public Qpinion Quarterly,
1958-59, 22, “, “6“-“72.

Scott, W., "Rationality and non-rationality of international attitudes,"
Journal of Conflict Resolution, 1958, 2, 8-16.

120

Sherif, M., et. al., Inter-Grou Conflict and Coo eration, Norman:
The University of aklahoma, Institute of Group Relations, 1961.
Smith, P.A., "Opinions, publics and world affairs in the.United States,"

Western Political Quarterly, 1961, 1“, 698-71“.

Smith, 3. L. and C. M. Smith, International Communication and Political
inion, A ide to theflliterature, Princeton: Princeton University
Press, 19 6.

Spaulding, I. A., "Of human values," Sociology and Social Research, 1963,
,(2).

Stevenson, N., The Study of Behavior, Chicago: The University of Chicago
Press, 1956.

Strachey, A., The Unconscious Motives of War: A Psycho-analytical
contribution, Condon: Allen I Unwin, 1957.

Thurstone, L.L., "An experimental study of nationality preferences,”
Journal ongeneral Psychology, 1928, l, “OS-“25.

Weiss, P. and J.L. Kavinau, "A model of growth and growth control in
mathmatical terms," Journal of General Physiology, 1957, “1, 1-9.

Zinnes, D.A., "Hostility and international decision-making," The
Journal of Conflict Resolution, 1962, 6, 236-2“3.

Zipf, G.K., National Unity and Disunity, Bloomington, Indiana: Principia
Press, 19“1.

APPENDIX A

QUESTIONNAIRE

Project No.
INTERNATIONAL SURVEY Phase No.
32“ Union Building Subject No.

 

 

This is a study about the general international situation as it appears
in the early months of 196“. Your cooperation is requested in answering
the various questions on the following pages. Your responses should be
your best estimates based on your previous knowledge and careful judgement.

For each question please make the best estimate or guess that you
can at this time. It is important that you respond to all questions.

You may take as much time as you need.
Below are the instructions for the topic on the next page.

Everyone is interested in maintaining and protecting their own and
their family's peace and well being. In the light of the current inter-
national situation, which of the countries listed represents the greatest
threat or danger to your own and your family's well being and security?

If you think a certain country is extremely threatening and dangerous
to the security of you and your family, place an "X" in one of the spaces
near the extreme right of the page on the appropriate line. If you think
that a country is not at all dangerous nor threatening, then place an
”X" near the,far left of the page on the appropriate line.

The more threatening and dangerous that you think that a country is,
the farther to the right you should place your "X" for that country. You
should think of the horizontal line as representing maximum threat and
danger on the right and steadily decreasing to the complete absence of
threat and danger on the left.

Below are samples to illustrate the topic on the next page. In the
following example, Bulgaria is estimated to be a fairly threatening and
dangerous country, and the "X" has been placed on the right-hand side of
the page:

PERFECTLY SAFE VERY DANGEROUS

Not threatening Threatening

In the next example, Bulgaria is judged to be fairly safe, and to
represent very little threat:

PERFECTLY SAFE VERY DANGEROUS
Not threatening Threatening
Bulgaria:_:_: : :A: ° ° ° ' ' ° ° ' : : ° ° ° :__:_:Bu1garia

After you understand the instructions, turn the page.

2

In the light of the current international situation, which of the countries
listed below represent the greatest threat or danger to your own and your
family's well being, safety and security?

The more threatening and dangerous that you think a certain country is,
the farther to the right you should place your "X" for that country. Check
back to the examples on the previous page if you do not clearly understand
the instructions.

Read through the whole list of countries on this page before you begin
to answer. Be sure to use as many of the different horizontal spaces as
you need to indicate the different degrees of threat and danger represented
by the countries listed on this page.

PERFECTLY SAFE

Not threatening

Albania:__n__r__.__r__.
Algeria :_:_: ° '-
ainland)China:_:

:Albania

— -.—.‘ — - _ _._ - - —.- — — _ — —::: Algeri‘
__ _:_:_:_:_:__:_:_:_:_._:_ __:__.__:_._:__: (mainland)Chir.

Cuba:_:':'°~:':~::"::°:°':Cuba

O O O O I O O O O O O O O
-_--_‘--‘——-—----_-—

.0 O.
O. O.
O. O.

O.

C. O.
O. O.
O. O.
O. O.
O. O.
O. O.
O. O.

O.
O.

ze chos lovak i a :_:_:_;_:__:_:_:_:_:_:_:_:_:_:_:_:_:_:_:_:_:_: Cze chos lovak i 3
East Germany :__:__:___:_:_:_:___:__:__:__:__:__:_:_:_:_:_:__:_:_:__:East Germany
aiwan ) Formosa :__:__:_:_:___:__:_:__:_:_:_:_:_:__:__:_:_:_:__:__:_: Formosa( Taiwan
France :__:_:__ :_:___:__:_:__:__:__:_:_:__:_:_:_:_:_:_:__:__: France
Hungary :_:___:__:__:_:__:__:_:__:__:_:_:__:_: d:__:__:__:_:__:_:Hungary
Indones i a z_:_:_:__:_:_:_:_:_:_:_:_:_:__:_:__:__:_:__:_:__: Indonesi a
Japan:__:-°°-':-:°°--'----:°::Japan

Mexico:_::°°'::°-°°°°'-::°::__:Nexico

NewZealand:__: °° ° ° ° ° ° ° ° ' ' -°':° ' :_:NewZea1and

North Korea:___:__:_:_:__:_:__:__:_:_:_:_:__:__:_:_:_:__:_:_:__:North Korea
forth Viet-Nam :__:__:_:__:__:__:_:_:_:_:__:_:_:_:_:_:_:__:_:__:__: North Viet-Nan
Pakistan :_’_‘_’_’_:_‘_‘_:_'_’_:_‘_’_3_’_’_’_.’_’_’_: Pakis tan
Poland:_::::::::°:':::°::':_:_:Poland

Roumania::::":::::'°°°::‘°:::Roumania
— -~

:_: South Viet-Nan

'»outh Viet-Nam:_:__:__:__:__:_:__:_:_:_:_:__:__:__:__:_:__:_:_
Soviet Union :_:__:__:__:___:__:_:_:_:_:_:_:__:_:__:._:_:__:__:_:_:Soviet Uhion
Sw it zerland :__:___:__:__:_:__:_:_:__:_:__:__:__:_:__:_:_:_:_:_:_: Swi t zerland
United States :__:_:__:__:__:___:__:___:__:__:_:_:_:___:_:_:__:__:_:__:_: United States

Yugoslavia:_:__° ° ' ° ° ° : ' ° ° ' ° ° ° ° : : ' :_:Yugoslavia

Make sure you have made an "X" in a Space on each line to represent the
degree of threat or danger represented by each country. Then turn the page.

The information requested below is your best estimate of the
similarities and differences between people of different nationalities.
Of course, not all people of one nationality have the same characteristics,
and in responding to the questions below, you should have in mind the
average person--neither the best nor the worst of a country's people.
1. Look at the names of all the nationalities listed below. Which
nationality is most similar to your own? Write a "1" in front of that
nationality. (If your own nationality is listed, write a "l” in front
Of ite)
2. Which nationality of those listed is most different from your own?
Write an "8" in front of the nationality which is most different from
your own.
3. From the nationalities which you have not marked yet, select two that
are most similar to the peOple of your country. Place a "2" in £535%
of each of them.
“. Choose two of the remaining nationalities which are most different
from your 053:.and place a "7" in front of each of these two nationalities.
5. Now choose four other nationalities which--among those which remain--
are most similar to your own, and place a "3" in front of each of them.
6. Now choose four more nationalities among those remaining which are
most different from your own, and place a ”6" in front of each of these
four.

If you have followed the instructions above, there should be ten
nationalities remaining which you have not yet numbered.
7. Choose five of the nationalities which are left which are more similar
to your own, and place a "“" in front of each of them.
8. Place a "5" in front of the five nationalities which are left.

Albanians _____Mexicans

I Algerians _____New Zelanders
(mainland) Chinese _____North Koreans
Congalese _____North Vietnamese
Cubans _____Pakistani
Czechoslovakians ‘____;Poles
East Germans ____Roumanians
Formosan Chinese _____South Vietnamese
Frenchmen _____fiussians ‘
Hungarians _____Swiss
Indones ians wAmeri cans ( USA)
Japanese ‘____;Yugoslavians

After you have placed a number in front of all of the nationalities
listed above, according to the instructions, turfi'fhe page.

u

Now you are to make the best estimate that you can about the
wealth or poverty of the same list of countries. In order to avoid the
problem of equivalency of different currencies, you are asked to judge
each country in terms of the amount of coal, hydroelectric power and
other sources of commercial energy available to the average person in
each country. a

Before making any X's, you should look at all of the countries--
especially the list on the next page--and decide whether the ones listed
are generally rich or poor, and whether they vary a little or a great
deal in wealth. Be as accurate as possible in indicating the differences
of wealth between countries.

For example, the average person in a very highly industrialized
country consumes 8,000 units of commercial energy a year while the average
person in a poor, underdevelOped and unindustrialized country, such as
Yemem, consumes far less.

In general, you should consider the scale below--and the one on the
next page, as asking you to make distinctions between richer countries
and poorer countries. The richer that you think a country is--that is,
the greater the income of the average person--the farther right you
should make you "X". However, in case of doubt, you should consider the
commercial energy consumption per person as the indicator of wealth.

Note that the scale values (0 to 8000) increase more rapidly as the
scale moves from the low values on the left to the high values on the
right.

05 q c» 0, ca 6?

o, 0: 0) 0:

~90? ssdsesisi’sf
£e7th76§IVIPId
, (a 6? c! c; c> Q (5 I 6? c> é? c> <3 0
zuczwémflW/‘b 9’ to Q ‘0

 

O
O

Argentina_:_:_:_:__:__.__.__:_:L:_:_:_:_:_Argentina

Bulgaria_:__° ° : ' : ' ° ° :X:_° °__:__Bulgaria

Portugal__° ° ° ° : ' ° ' ° : ° ° ° :__Portugal

0 O O O O O O O C C 0
.-~.-._*----*

In the example above, one person has estimated that in Argentina
the average energy consumption per person is 1000-1500 units in Argentina,
and 1500-2000 units in Bulgaria.

For practice, before turning the page, make an estimate of the
wealth per person of Portugal. Then turn the page.

wealth of each country.

(mainland)China: :

lllllll

Make an "X" on each of the lines below to indicate the per ca ita

Energy ConsumptiggnflgPer Person

Albania :__:__° °

0
O
O
C
O
0

Algeria :__:_: ° '

O.

0 O 0 O O 0

Congo:_: ~ . .

O O 0 O O O O O O 0
-—_‘--“-‘—~—

Cuba :__: :

Czechoslovakia-g . . .

e e e e O I O O
-—_-——-_--

East Germany:__: ° : 0

(Taiwan) Formosa ;-:_. . .

O I O O O O O O O O O O
’-—--—_----—

France :___: : ° ~
Hungary :_. . . .
Indones ia :__: ° ° °
Japan :__: ' ° '
Mexico :__: ° ° °

New Zealand: : ' ' '
North Korea ::: ' ' °

North Viet-Nam: : : ' °

Pakistan :__:___° ° °

Poland:_:
Roumania:_: ° ' °

South Viet-Nam: : - - -

Soviet Union:_:_° ° ‘

Switzerland:_: . . .

United States:__:__° : '

YugoslaVia :__:_' ° ‘

0

O

O

O

O O O O O

O O O O 0 O

O O O O 0

O O O

O O O O O O

O O O O O O

O 0 O O O O

O 0 O O O

O O O O O O

O

O

2

O O

O

O O O O O O O O O O O O
—-—_‘__—*----

O

O
O

O
O

2

.__Albania

Make sure you have made an "X" on each line to indicate the wealth

per person of each country.

Then turn the page.

For detailed instructions, see the prev one page.

 

.__Algeria ____
_China( mainland )_
_Congo _-
__Cuba .........
_Czechoslovakia
East Germany
:Formosa( Taiwan:
_France .._._....
_Hungary
_Indonesia
Japan
-Hexico
-New Zealand
—North Korea
—North Viet-Nar—
:Pakistan “
Poland
:Roumania
_South Viet-Nam____
._Soviet Union
_Switzerland
_United States
_Yugoslavia

In the space provided below, please make your best estimate of
the total national population of each of these countries to the nearest
million. For example, if yourbest guess of the total population of Italy
is fifty million, write a "50" in the space by the name of the country:

0’0 Italy

Now make your best estimate of the populations of the countries

listed below:

Yugoslavia Japan

United States Indonesia

Switzerland Hungary

Soviet Union France

South Viet-Nam Formosa (Taiwan)

Roumania East Germany ‘______
Poland Czechoslovakia

Pakistan Cuba

North Viet-Nam Congo

North Korea China (mainland) ______
New Zealand Algeria

Mexico Albania

Before turning the page, please make sure that you have made at
least a rough guess about the pOpulation of each country to the nearest

mi 11101).

Below are listed the same countries you have seen on previous pages.
This time you are to examine them in terms of the degree to which each
government is communist or non-communist.

1. Look at the names of all the countries listed below. Which country
has a government which is the most non-communist, that is, which government
has the least communist influence of-Ehy on the list? Hark a "l" in front
of this country.

2. Which country has a government which is the most completely
communist one on the list? Place an "8" in front of this country.

3. From the countries you have not yet marked, select two countries
which have the most non- communist governments. Hark a "2“-innfront of
each of these 512 comries.

“. Choose two of the remaining countries which have the most communist
governments. MEFF'a "7" in front of each of these two countries.

5. Now choose four of the countries you have not yet marked which
have the most Eggrcommunist governments, and place a "3" in front of
these four countries.

6. Choose four of the countries you have not yet marked which have the
most communist governments, and place a "6" in front of these four countries.

7. If you have followed carefully the instructions above, there should
now be ten countries on the list below that you have not yet marked.

Select five of the ten which have the most Egg-communist governments, and
place a "“" in front of these five countries.

8. Place a "5" in front of the five countries which are left.

 

 

‘ Albania Mexico
Algeria New Zealand
China (mainland) North Korea
Congo North Viet-Nam Summary_of_lnstructions
Cuba Pakistan 1 one (most non-
Communist)
Czechoslovakia Poland 2 twos
“ threes
East Germany Roumania 5 fours
"' ' 5 fives
Formosa (Taiwan) South Viet-Nam “ sixes
2 sevens
France Soviet Union 1 eight
"'"'""' """' (most
Hungary Switzerland Communist)
Indonesia United States
Japan Yugoslavia

After you have placed a number in front of each of the countries according
to the instructions above, turn the page.

Please indicate your disapproval or approval of the following statement:
I think that my country should give up some of its soverignty to

become part of a united world government--in the same way that a state

or provincial government gives up some of its power to a national government.
Strongly disapprove Approve

Uncertain
Disapprove Strongly approve

Thank you for your cooperation. Your responses will be treated in
a completely confidential manner, and no names of individuals will be
used in any report of this study. Please fill in the biographical in-
formation requested:
Name__ _(name not necessary)
Sex: ____Male ____Female
Age: ‘4‘
Occupation or field of study:
College or university graduate Yes No
In what country were you born?
What is your nationality? ___
Is either your mother or your father of a nationality different from
that of the country in which you were born? ____Yes ___No
Name any country--or countries--other than your own that you lived in
for more than one year before the age of 21:_¢ ‘4

Date
THE END

COMMENTS:

   

M'TITI'I‘HLMTILEJ/fllflzfi!Lifliujfyifllflflfljfiflmfl'Es