.

FACTOR 3

IMENSIONALITY IONS PROJECT

FACTOR 2 {/<•• X /

REPORT

<*" ^^f

JTICAL SCICNCI

UNIVERSITY OF HAWAII LEB /

PER R«produc*d by

GRC NATIONAL TECHNICAL IN^RMATION StRVICE

Spnngfidd. Vl 21151

.GUA AFG,

JUL SO OTI

USEÜÜTaMJ

BEST AVAILABLE COPY

Dimensionality of Nations Project

Department of Political Science

University of Hawaii

RESEARCH REPORT NO. 12

The Patterns of Dyadic Foreign Conflict Behavior tor 1963

Dennis R. Hall and lO R. J. Pummel |UJ jUL 29 1971

WsEinns - B ^

June, 1963

D D C

Prepared in Connection with Research Supported by the National Science Foundation, Grant No. 1230, and the Advanced

Research Projects Agency, APPA Order No.1063, and Monitored by the

Office of Naval Research

Contract *NO0OlA-67-A-0387-00O3

This document has been approved for public release and sale; its distiibution is unlimited and reproduction in whole or in part Is permitted for any purpose of the United States Government.

• ' '^ y--']ic. rd-ase; L .-IK..! " .li^J

NOTICE TO USERS

Portions of this document have been judged by the NTIS to be of poor reproduction quality and not fully legible. However, in an effort to make as much information as possible available to the public, the NTIS sells this document with the understand- ing that if the user is not satisfied, the document may be re- turned for refund.

If you return this document, please include this notice to- gether with the IBM order card (label) to:

National Technical Information Service U.S. Department of Commerce Attn: 952.12 Springfield, Virginia 22151

Sj'iurilv CLissifii iilmn

DOCUMENT CONTROL DATA -R&D St ■ iirir»- t IH\ tiln ulion at tttiv, hndy «»/ uhutmrl und tntintint' .ifinof«fii.n nmst hf mntvrcd w /ti-n r/i»- nvt-r.itl rvftufl t ■» r lii\hiHcd)

i OMioiN»iiNO *C T i vi T v (C'orpiiMil« »mhorj

k Dimensionality of Nations Project University of Hawaii 2500 Citnpus Road, Honolulu, Hawaii 96822

ia.BtroRT lECURlTV < I »V'jl I if A 1 ION

Unclassified tb. OWOUP

J REPO« T TITLE

The Patterns of Dyadic Foreign Conflict Behavior for 196J

4 DESCRIPTIVE NOTES (Typ* ol rrpon and inrtuiivr tlmlrt)

Research Report No. 12 9 »Ii i MOBi»i (Fitmi name, middl» inlllml, lull nmm»)

Dennis R. Hall and R. J. Rummel

6 REPORT DATE

June 1968 Ta. TOTAL NO OF PACE»

39 7b NO OF RE !■»

10 •a. CONTRACT OH GRANT NO

N0001A-67-A-03G7-0003 (and NSF GS-1230 9a. ORIGINATOR'S REPORT NUMBERISI

Research Report No. 12 6. PROJEC T NO

*b OTHER REPORT NOIS) (Any olhrt ihurt thmt mmy 6« mmtlfnrd thit rmporl)

10 DISTRIBUTION STATEMENT

This document has been approved for public release and sale; its distribution is unlimited and reproduction in whole or in part is permitted for any purpose of the United States Government.

II. SUPPLEMENTARY NOTES

2500 Campus Road Honolulu, Hawaii 96822

12 SPONSORING MILI T «RT ACTIVITY

Advanced Research Projects Agency Washington, D. C.

I> ABSTRACT

Five patterns of dyadic foreign conflict behavior were delineated for 1963. The strongest of these patterns was negative communications, which was even stronger as a pattern than it had been for a 1955 dyadic foreign conflict study. The third pattern, violence intensity, and too. fourth pattern, warning and defensive acts, marked the general decrease in military activity from 1955 to 1963. The most militant conflict behavior in 1963 was that of China to Taiwan and Taiwan to China. There were a number of warning and defensive acts, most noteworthy being those of Indonesia to Malaysia end Malaysia to Indonesia. Negative sanctions and unofficial incidence of violence were the most stable patterns of foreign conflict behavior. The United States was involved in a number of sa.ictions directed against Cuba, the Dominican Republic, Russia, and South Vietnam, The unofficial incidents of violence occurred most frequently in the less developed, smaller nations, and w«re frequently directed against major world powers. The neg tive communication pattern characterized the primary behavior of the major world powers.

Data collection from The New York Timeg for the 1963 study assumed a typology of foreign conflict behavior subsuming the diverse reported events. The patterns of the 1963 study provided good evidence for the soundness of this typology - the five major categories of the typology emerged as the five basic, uncorrelated behavior patterns of dyadic foreign conflict behavior.

DD.FN0or.,1473 VN 0101.807.6001

(PAGE 1) UnrlflBBified

Sfcunlv (Irt'.sidi .limn

Unclassiiied Security Clansificalion

KEY wonoi KOLK WT

LINK C

NOLC

Foreign Conflict

Factor Analysis

Conflict Groups

DD.,r,..t473 «BACK, (PAGE 2)

Unclassified Sacwltf Classification

1

TABLE OF CONTENTS

1. Abstract 2

2. Conflict Patterns for Dyadic Behavior 3

3. Stability of Conflict Patterns from 1955 to 1963 ... 12

A. Types of Dyadic Conflict in 1963 10

TABLES

2.1 Correlation Matrix . « 1

2.2 Orthogonal Componeut Matrix 10a

2.3 Factor Correlation Matrix 10b

3.1 Lambda Transformation Matrix Normalized 1A

3.2 Difference Matrix, 1955 to 1963 15

FIGURES

2.1 Dyadic Foreign Conflict 1963 Variable List with Codes . . 5

2.2 Eigenvalues 6

2.3 Comparison of Rotated Factor Loadings 9

3.1 Variable List with Codes for Behavior Pattern Comparison 1955 to 1963 12

4.1 Dyadic Members of 1963 Foreign Conflict Behavior Groups . 19

4.2 Grouping Characteristics for Foreign Conflict Behavior in 1963 20

4.3 Group Profilec 21

REFERENCES

2

i. ABSTRACT

Five patterns of dyadic foreign conflict behavior were delineated

for 1963. The strongest of these patterns was negative communications,

which was even stronger as a pattern than it had been for a 1955 dyadic

foreign conflict study. The third pattern, violence intensity, and the

fourth pattern, warning, and defensive acts, marked the general decrease

in military activity from 1955 to 1963. The most militant conflict

behavior in 1963 was that of China to Taiwan aad Taiwan to China. There

were a number of warning and defensive acts, most noteworthy being those

of Indonesia to Malaysia and Malaysia to Indonesia. Negative sanctions

and unofficial incidence of violence were the most stable patterns of

foreign conflict behavior between 1955 and 1963, each accounting for

about ten percent of foreign conflict behavior. The United States was

involved in a number of sanctions directed against Cuba, the Dominican

Republic, Russia, and South Vietnam. The unofficial incidents of

violence occurred most frequently in the less developed, smaller nations,

and were frequently directed against major world powers. The negative

communication pattern characterized the primary behavior of the major

world powers.

Data collection from The New York Times for the 1963 study assumed

a typology of foreign conflict, behavior subsuming the diverse reported

events. The patterns of the 1961 study provided good evidence for the

soundness of this typology - the five major categories of the typology

emerged as the five basic, uncorrelated behavior patterns of dyadic

foreign conflict behavior.

2. CONFLICT PATTERNS FOR DYADIC BEHAVIOR

Systematic research may ultimately allow us to predict to

behavior in the international system. Long before this becomes

possible, however, it will be necessary to derive and test an empirical

typology of international behavior and to reduce behavior to its primary

patterns. One unit of behavioral analysis in this task is the dyad, a

pair of nations coupled by the behavior of one towards the other. One

type of such behavior is hostile, reflecting the existence of a conflict

situation between nations. Behavior of a hostile sort, or conflict

behavior, has been our concern in the research reported here. This is

not to ignore cooperative behavior, but rather to focus on conflict

behavior sufficiently to enable, in a larger study, the correlation of

those patterns with those involving cooperation.

Hostile acts between nations occurring in 1963 and reported in

daily issues of The New York Tiroes were collected for analysis using a

foreign conflict code sheet (Rummel, 1966). Each hostile act was

recorded as to actor, object, date, type of action involved, and des-

criptive information of the act. The code sheet employs an empirically

derived typology differentiating six primary categories of foreign

conflict behavior: warning and defensive acts, violent acts, negative

behavior acts, negative communications, urofficial violence, and non-

violent demonstrations. These primary categories are further divided

into approximately one hundred sub-categories.

All hostile acts for 1963 were recorded on code sheets, one

sheet to an act. The code sheets for a specific dyad were then aggregated

for 1963 and frequencies were calculated for each type of conflict act,

4

such as threat, mobilization, diplomatic snubs, accusations, anti-foreign

demonstrations, and clashes. Those acts that occurred with sufficient

frequency for analysis were selected as variables for this study.

Figure 2.1 lists the conflict variables for 1963 formed from the codings

of The New York Times.

Almost three-thousand conflict acts involving 275 dyads were

recorded for 1963. A listing of the raw data and descriptive statistics

on the variables are given in Appendix I.

The product-moment correlation matrix between the raw data on

the 24 conflict variables for 27: dyads is shown in Table 2.1. These

correlations were factor analyzed (component analysis) to delineate the

patterns of dyadic foreign conflict behavior. Since the component factor

model was used, the patterns dv.-. -«ated describe the total (common and

unique) variation among the conflict acts.

The question of when to stop factoring - how many patterns to

delineate - has been of central concern in applying factor analysis.

There is no pat answer; a decision has to be based on the aims of the

analysis. Component analysis of total variance typically delineates as

many dimensions as variables and this study is no exception.

The eigenvalues Indicate the amount of variance accounted for by

the dimensions. The squares of the eigenvalues divided by the number of

variables indicates the proportion of variance accounted for by the

dimensions. To rotate factors beyond the sixth dimension would involve

such a small proportion of variance in the data that Inclusion can be

justified only when theoretic interest in the total pattern of a dimension

outweighs the demand for parsimony.

Figure 2.1

DYADIC FOREIGN CONFLICT 1963 VARIABLE LIST WITH CODES*

Primary Hategory Variable No. Code Variable

warning and defensive acts WARNDF ALRTMB

Warning and Defensive Acts Alerts and Mobility

violent acts

3 PLNVIL - Planned Violent Acts A WARACT - Overt Violence 5 DISCMA - Discrete Military Actions 6 DAYVIL - Days of Violence

negative behavior acts

7 NEGACT 8 UNCNEG 9 SEVDPR

10 EXPREC 11 BCOTEM 12 AIDREB 13 NEGCOM 14 WRTCOM 15 ORLCOM 16 WRTORL 17 ACCUSN 18 PROTST 19 MINTHM

Negative Behavior Acts Unclassified Negative Acts Severence of Diplomatic Relations Expulsion or Recall Boycott or Embargo Aid to Rebels

negative communication

Negative Communication Written Negative Communication Oral Negative Communication Written or Oral Negative Communication Accusations Protests Minor Themes

20 UNOFVL - Unofficial Violence 21 ATKEMB - Attacks on Embaosy

unofficial violence 22 ATKPER - Attacks on Persons 23 ATKFLG - Attacks on Flag

non-violent demonstrations 24 NVIOLB - Non-Violent Behavior

*Primary code sheet categories will be separated by solid lines throughout the rest of this report. Variables 1-19 are Official Acts; Variables 20-24 are Unofficial Acts.

., .• ,. • . .

5.C

4.0

3.0

2.0 1

1.0

Figure 2.2

EIGENVALUES*

0.0 l 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 lb 19 20 21 22,23 24

Variables

*24 unrotated component factors (principle axes)

(SI

u

X c iH i-l JO U « n

M

O u

«IäI

a o

H 3 PJ

-< • O «M O vD

PM PNI

H

O 00 00 O <N —<

c •H Pvl

rH

C> \r\ oo ir\ C rs r^ r^

N o CM

rH

g fi

m <n »» p» o o o o

A 3 rH

i i i i

o •* o o

in v*; <f iri o O o O

n oo rH

1-i I i i I 1

C Ov 00 O <M l^-

IT» P"> m \C o o o o

n 3 rH

rH 1 1 t 1 1

O rH VO O C r- -» r»

so o r^ co o o o o

N 3

«O rH

rH 1 1 1 1 1

C CM VO VO <t O >* vO tM vD

IT» PI >» »ri o o c e

n o

m rH

rH 1 1 1 1 i

c •& m a «oi^jcocc- >* o

sr rH

rH 1 1 1 1 1

O VO 00 ^ C O" 00 o oo «o r^ 9* m f»

so pi p-» oo o o o o § rH

rH 1 1 1 1 1

q c

rH e» »r» «n r> CM «r> o o o o o o o

o- <■ ^ <■ o o o o p rH

r- 1 1 1 1 1 1 1 1

o o \o i» -* c* m o o c o o IH e

in «r» «* »n D O O O fe

rH rH

f-t i i i i 1 1 1 1

O evj O O ^-t o

O rn O <■» fx oo <"■> O O rH O O O O

■3- PO »O rH r> o o o O

o rH

l-l 1 1 1 1 i i i i 1

C vi C rH c c o o

rH C CO CO 00 C r» •H rH O O O rH O

c <t -» o« D O O O

o o o<

•Hill 1 1 1 1 1 1 1 (III

O C <-i >» (N c o c o o

rH tr\ VC «C O rH >0 rH O O rH O P« O

■O rH «M rH 3 C O O

in oo rH 1 1 1 I

§ <r tH sr r^ oo \C ~» »J f^l f-1

ON O rH d «r» O f» O rH O rH O M O

ff\ »n p»> p<. D O O O

n 3 rv

rH 1 1 1 1 1

o o 0

vr <r vT m CT> O O O O •-(

.* ri o m <r <i- PI c o o o o c o

•> PO PH n D O O O

»4 3 s£

r-t 1 1 1 1 i i i i i i i till 1

§g o ■^ vr sr o a O O C? O rH

PO m o «^ P^ »y rH o o o o c o o

^ PH N P"* D O O O 3 •

1

•n

iH rH 1 i i i i 1 ! 1 1 1 1 1 till

O tH VO O «» O o

fg CM r-l r-> «J O O rH C ^

,C rH »/■> Pt rH 0C 0> D O O rH rH O H

H rv in <• s o o c

»N

« -»

iH 1 1 1 1 1 1 1

O «^1 •-» vD O O W r-H o

O rH r- m o c a c o o

/^ PM m p^ o >f' vo S O O rH rH O rH

:. oo »o rH 3 o o o pn

t-l 1 < 1 • 1 1 1 1 i

O c

KO vC >-4 CM C i-t O O

o «* vo «* r>4 »H rH O O O

» rH IT» pg «» rH IT» -H O rH m rH O PM ^SS H

H •

rsi

c-t 1 1 1 1 1 1

O CO o o>

-* t-< O «^1 r-» <-J O O

o m vc •» t-J -H iH O O O

«- rH vO <H <T O »O H O rH P^ iH O PJ

T oo CM m > 5 rH © O " •

rH

iH 1 1 1 1 ' t

11 n ■J H < -J M U X H *■» «* U P Vi K W' >; ►J «aj >••' < O. ^. O C

<

5.

ü et r-> vn w m »- u: w tq fc^ ca r.i H oc Ö ?• r1. O O » W >• U M b V) u' (-I <

s r: s »J ^ f-1 £ 5 o u u; v) v. a _1 u o o » H P _-> (-: J JH t.) b t:

z; 5 O » 5 a, S

.1 pg ti Q e > S W H r u Q es C c ? S2 52 u: r - H H H ; ^ < < <: 5

e J -> -* >

u

u a. 41 »4 m

s o

m u u o Ü

Q >n CM

■*w*»^-r>-

To determine whether rotating only six factors would affect the

factor structure, we carried out additional rotations for different

numbers of factors. Figure 2.3 displays the different orthogonally rotated

component matrices. The best matching component patterns from each of four

rotations are shown together. Each of these groups of factors are named,

as indicated across the top of the display. The groups of patterns named

negative communications (NEGACO), unofficial violence (INCVIO), violence

Intensity (VIOINT), and warning and defensive acts (WRDEAC) are remarkably

stable on high loadings through the four rotations. The fifth group of

dimensions, negative sanctions (NEGSAN), indicate that some differences do

show up when rotations are based on different numbers of factors. The

negative sanctions pattern accounts for about nine percent of total

variance in the data and is retained for its theoretic interest.

The third (VIOINT) and sixth (MILACT) groups of dimensions have

variables loading on them that measure behavior in the violent acts code

sheet category. Since variables number 3 (PLNVIL) and A (WARACT) load

moderately on the third dimension (VIOINT), the six group of dimensions

(MILACT) can be dropped from the analysis in the interests of parsimony.

The fifth column in each group of dimensions in Figure 2,3 shows

the result of orthogonally rotating five image factors (Harman, 1967,

Chapter 8 - Rummcl, 1968, Chapter 5). Image analysis is a type of common

factor analysis in which the unique variance for each variable is removed

from the analysis. Image analysis delineates factor patterns from that

portion of variance in any one variable which can be predicted from the

remaining variables in a least squares sense. Appendix II gives the

covarlance matrix which was image factor analyzed along with the image.1

H U

1/3 IT» o

M

u

M «n O

M >

o

U

CM u

<N

CM O

il

K # • • • •

• ♦

• •

-J H < J M U Z M

a. 3 (A ^ M < a a IT» *

o in

o

I I

I I

•••••••

£••••••

a da a.

JSgg (Mi

u a; « S g s S§3

r-> 30 0\ O •-»CM

55 H » H oi vi V) O 3 H H U O

(*) ■« in >0 r^ ao 9i •H •-• -H r-i ^H iH iH

o < < <

o »H CM m fM CM CM CM

N O

u «• r> 00 »H «M CM

«M O •• CM 9« r> • • • • •

■ ■ ■ • • %* u u u u u o o o o o o W «J w w w

« u u u u u 9 « « « « « 3 ^ ^ •** •" So ■« «o m <n

CM •-( «I w ww ws^ M

<*4 • « • •

■ ■ ■ • >« fe »s >>

rH H r-l ^ • llll! -i ■ u u u u m t* o o o o 9 a M «i ^ ^ < «4 u u u u •e m m m v u n b. «b h lb O O

<J W W 4« U 8 a o n « «

« « tt b. w S 6 S 6 3 0 o o o « ^ fiiili u o u u •? « 1 I I ■ ■

Si-t CM «n •« u u u u M

10

orthogonally rotated five factor matrix. It should be clear from Figure

2.3 that Image analysis does not affect the pattern delineation sig-

nificantly. In other words, our results are invariant of choice of

common or component models.

The orthogonally rotated five component matrix for 24 dyadic

coniiict variables is shown in Table 2.2. This table corresponds to the

five factor results (C^) shown in Figure 2.3. The five rotated dimensions

account for 59.8 percent of total variance in our foreign conflict dati.

Dimension one - negative communication (NECACO) - alone accounts for 20.2

percent of the variance. The last column of figures in Table 2.2 are

the communallties (h2) of the variables, that is, the proportion of

variance in each of the variables accounted for by the five dimensions.

The five dimensions of Table 2.2 are mutually uncorrelated. Accord-

ingly, an oblique rotation (biquartimin technique) of the components was

computed and compared with the orthogonal rotation. Table 2.3 shows the

correlations between the five primary factors defining the oblique patterns.

The off-diagonal correlations are low, Indicating that the diiiensions are

indeed orthogonal for all practical purposes. Therefore, the obliquely

rotated results will be omitted from further consideration.

3. STABILITY OF CONFLICT PATTERNS FROM 1955 TO 1963

The delineation of five dyadic conflict patterns for 1963 gives

information only for one cross-section of time. Many other cross-sections

must be added to the analysis before conflict dynamics can be isolated.

A study of dyadic foreign conflict behavior in 1955 has been published

(Rummel, 1967) and will at least give a second cross-section to contrast

with the patterns delineation for 1963.

Table 2.2

Orthogonally Rotated Matrix*

10a

Variable Dimensions

No. Name 1 2 3 4 5 h2

1 WARNDF .17 .04 + .07 (.92) .05 .88 2 ALRTMB .17 .02 +.04 (.90) .05 .84 3 PLNVIL .08 .06 (+.51) .22 -.14 .34 4 WARACT .08 .06 (+.70) .28 -.14 .59 5 DISCMA -.05 -.03 (+.91) -.13 -.03 .86 6 DAYVIL -.06 -.03 (+.89) -.15 -.02 .82 7 NEGACT .06 -.04 +.04 .12 (.96) .94 8 UNCNEG .10 .06 +.01 .10 (.65) .45 9 SEVDPR -.17 -.12 -.02 .31 .31 .24

10 EXPREC -.00 -.03 -.05 -.14 .49 .26 11 BCOTEM .00 -.05 -.03 -.08 .45 .21 12 AIDREB .00 -.06 +.33 -.05 .16 .14 13 NEGCOM ,(.98) -.04 +.00 .01 .04 .97 14 WETCOM (.85) -.02 -.02 -.18 .07 .76 15 0RLC0H (.72) -.03 +.01 .06 -.13 .54 16 WRTORL (.80) -.03 +.03 .22 .10 .69 17 ACCUSN .(.92) -.03 +.03 .02 -.08 .85 18 PROTST (.59) -.02 -.06 -.18 .30 .48 19 MIHTHM (.84) -.03 + .08 .13 -.01 .75 20 UNOFVL -.03 (.92) -.03 -.02 -.04 .86 21 ATKEMB -.04 .40 -.00 .25 -.07 .23 22 ATKPER -.04 (.88) +,ai «00 -.00 .78 23 AT1CFLG -.04 <.85) -.02 -.09 -.02 .74 24 NVIOLB -.06 .03 -.03 .39 -.11 .17

Percent of total Variance 20.21 10.69 10.41 9 75

Percent of Common Variance 33.80 17.88 17.41 14.61

8.73 59.79

14.61

^Component analysis(employing, ones in principal diagonal of product- moment correlation matrix) and principle axes technique. Cutoff eigenvalue is 1.92. Rotation is Varimax. From factor analysis of Table 2.1. Loadings greater or equal to an absolute value of .50 are shown in parentheses.

10b

Table 2.3

CORRELATIONS BETWEEN PRIMARY FACTORS*

NEGACO INCVIO VIOIKT WRDEAC NEGSAN

NEGACO 1.00 INCVIO 0.00 1.00 VIOINT 0.02 -0.03 1.00 WRDEAC 0.01 -0.07 0.04 1.00 NEGSAN 0.01 -0.09 0.00 0.03 1.00

*Blquartlmln rotation technique: 30 major cycles, 6938 Iterations.

11

For a number of reasons a direct visual comparison of the factor

loadings for the two periods cannot precisely define any shift in

patterns. First, the 1963 study employed 24 variables and the 1955

study employed 16. Second, the 1963 study delineated five primary

patterns of behavior, while the 1955 study delineated only four. Third,

independent rotation of the two studies msy have delineated different

appearing patterns, although still linear transformations of each

other. Fourth, the factor matrices are simply too large to compare

visually.

We have employed another approach requiring some modifications

in the original analysis. The technique, called transformation analysis,

rotates the patterns of one study to a best (least squares) fit with

those of the second study, removing the influences of the separate

rotations specific to each study (Ahmavaara, 1957; the technique is

discussed in Rummel, 1968, chapter 20). The technique requires that

both factor matrices involved in the transformation analysis have the

sane variables.

Although factor matrices for 1955 and 1963 with common variables

can be formed by omitting variables from each matrix uncommon to both,

we approached the problem differently. We refactor analyzed the 1955

and 1963 data matrices after removing variables uncommon to both. By

analyzing the same variables we were able to generate factor scores for

a research design to be reported in Hall and Rummel (1968).

12

Comparison of the same variables led to considerable complication

of the research design. The code sheet categories and subcategorles

which had occurred with enough frequency to be analyzed for the 1963

study (see Figure 2.1) were substantially different from the categories

and subcategorles occurring with sufficient frequency in 1955 (Rummel»

1967). Consequently, only sixteen of these categories and subcategorles

are common to the dyadic conflict variable lists for both years. Since

the sixteen conflict variables were fewer than those employed in the

1955 study, both the 1955 and 1963 studies had to be reanalyzed for

these alone.

Figure 3.1

VARIABLE LIST WITH CODES FOR BEHAVIOR PATTERN COMPARISON*

warning and defensive acts 1 WARHDF . 2 ALRTMB

Warning and Defensive Acts Alerts and Mobility

violent acts

3 PLNVIL - Planned Violent Acts 4 WARACT - Overt Violence 5 DISCMA - Discrete Military Actions 6 DAYVIL - bays of Violence

negative behavior acts 7 NEGACT - Negative Behavior Acts 8 BCOTEM - Boycott or Embargo

negative communication

9 NEGCOM - Negative Communication 10 WRTCOM - Written Negative Communication 11 0RLC0M - Oral Negative Communication 12 WRTORL - Written or Oral Negative

Communication 13 ACCUSA - Accusations 14 PROTST - protests

unofficial violence 15 ATKEMB - Attacks on Embassy non-violent demonstrations 16 KVIOLB - non-violent Behavior

*These variables are common to both the 1955 and the 1963 studies. Variab1es 1-14 are official acts; variables 15-16 are Unofficial acts.

13

The factor structure for the 1963 study remained largely unchanged

with the reduction in the number of variables and again we rotated five

component factors. These five patterns, accounting for 72 percent of

total variance in the sixteen variables, are named to correspond to the

patterns of the twenty-four variable analysis. To avoid any differences

between studies due to different numbers of factors, five patterns were

also delineated for the 1955 study. The 1955 patterns account for

78 percent of variation in the sixteen variables.

Transformation analysis follows the basic form of regression

analysis and, in matrix terms, is.

Lambda - (FJ^'FJ^)"1?^,

where Lambda is the matrix of transformation (regression coefficients)

of Fj^ matrix of factors to matrix F2, F^ Is the 1955 sixteen variable

factor matrix, and F2 is the 1963 sixteen variable factor matrix.

Taking the 1955 factor matrix as the independent study and the

1963 factor matrix as the dependent study (as in regression analysis)

we calculate Lambda. The Lambda transformation matrix is then used to

A

pre-multiply the 1955 study, Fj, to get a least squares estimate. F2,

of the 1963 patterns,

F2 = Fl LaiIlbda

The overall correspondence of F2 to ^2» ^e 1963 patterns estimated

from the 1955 factor matrix, can be given by an intraclass correlation

coefficient between all the loadlngr of F« and F9, that is, between the

transformed 1955 patterns and those for 1963. For our 1955 and 1963

patterns the intraclass correlation coefficient is .71, meaning that

there is a fifty percent congruence between the patterns of the two

studies.

14

The finding that the correlation is .71 between 1963 factors and

the best fit to these from 1955 suggests that the patterns of conflict

behavior for nation dyads are fairly stable from 1955 to 19C3. By

knowing the 1955 conflict patterns alone, we can predict the conflict

patterns for 1963 with about fifty percent reliability. The implica-

tion of this is that conflict behavior not only has regularity across

nations, a finding of importance by itself, but also across time.

The regularity that we have discussed so far refers to the

structure of conflict behavior defined by the five dimensions. How do

these dimensions stand up individually when 1955 factors are transformed

to a best fit with those of 1963? The Lambda transformation natrix when

normalized by rows gives the cosines (correlations) between the factors

of Fi (1955 factor matrix) and F2 (1963 factor matrix. F^ are the

rows in Table 3.1 and F2 are the columns.

Table 3,1

LAMBDA TRANSFORMATION MATRIX NORMALIZED*

NEGACO VIOINT URDEAC NEGSAN NVIOLB NLCACO (.96) -.14 -.12 -.16 -.14

1955 dimensions VIOINT -.04 (.73) .38 .08 (.56) of foreign WRDEAC .08 -.32 (.89) -.07 -.31 conflict NEGSAN .19 .03 .21 (.95) -.13

NVIOLB -.45 .10 (.63) .16 (.61)

*The sum of squares for the rows will equal 1.00. Cosines greater or equal to an absolute value of .50 are shown In parantheses.

The correlations of Lambda normalized are similar to the coefficient

of congruence (Rummel, 1968, Chapter 20) and measure the similarity

between factors that are not independently rotated. The highest conflict

pattern congruence is between Negative Communication in 1955 and 1963 (.96).

15

A difference matrix,

D = F2 - F2,

is calculated to determine the difference between the actual 1963

patterns and those that are the least squares estimates of 1963. The

elements in the body of the difference matrix (Table 3.2) can be

interpreted as the loading shifts from the 1955 patterns to those for

1963. Thus, variable 1, warning and defensive acts, loads .57 higher

onto the 1963 pattern Warning and Defensive Acts (WRDEAC), then it does

on the corresponding 1955 pattern.

Table 3.2 DIFFERENCE MATRIX, 1955 to 1963*

Sum Sq.

1 WARNDF .08 (-.38) (.57) .16 -.18 .53 2 ALRTMB ».03 -.03 (.44) -.03 .06 .20 3 PLNVIL -.10 (-.45) -.22 -.04 (.53) .55 4 WARACT .04 .03 .14 -.09 (.44) .23 5 DISCMA -.18 ( .50) -.25 -.07 -.20 .39 6 DAYVIL -.13 ( .49) -.27 .00 -.27 .40l

7 NEGACT (-.52) .02 .07 (.60) -.04 ■.64 8 BC0TEM .05 -.01 -.02 .00 -.12 .02 9 NEGCOM .17 -.03 .05 -.10 -.00 .04

10 WRTCOM .02 ,11 -.01. -.06 .04 .02 11 ORLCOM -.01 -.04 .05 (-.30) -.06 .10 12 WRT0RL ( .48) (-.30) .00 .05 -.10 •,34" 13 ACCUSA .10 -.03 .09 -.24 ,10 .09 14 PROTST -.02 .05 (-.31) ( .30) -.10 .19 15 ATKEMB -.04 .12 (-.32) -.20 ( .31) .26 16 NVI0LB -.01 .05 -.03 .04 -.27 .08

Sum Sq 0.61 0.97 0.93 0.70 0.85

♦Ahmavaara's transformation technique where the difference matrix, D ■ F2 - ^2« F2 isAthe matrix of 1963 factor loadings (component, varimax rotation), F2 is the linear transformation of Fj factor loading matrix (component, varimax rotation) to a least squares fit with F2. Differences greater or equal to an absolute value of .30 are shown in paranthesis.

■■■■, ■ . ■ ■■ ■-., ........

16

The sum of squared loadings in the rows of Table 3.1 (SUMSQ)

Indicate the magnitude of shift over all patterns for the individual

variables from 1955 to 1963. The sum square of rows is thus an Index

of variable stability for the patterns between the two years. The

sum square of columns in Table 3.1 indicate the magnitude of shift over

all variables for individual patterns from 1955 to 1963, and is thus an

index of pattern stability for the variables between the two years.

4. TYPES OF DYADIC CONFLICT IN 1963

With the five dimensions of 1963 conflict behavior delineated on

24 conflict variables (section 1 of this report), we can locate in this

conflict apace each of the 275 dyads that had some conflict in 1963 in

terms of their factor scores. The factor scores for dyads on these

five dimensions give us a profile of the conflict behavior of each dyad.

With these profiles we can then determine what dyads have similar profiles ■

similar conflict behavior - and then group dyads in terms of this

similarity. These groups will define taxonomy of dyadic conflict behavior

for 1963 (see discussion in Rummel, et al 1969, Chapter 11).

A problem in developing such a taxonomy is the large number of dyads.

To pare down the 275 dyads to a manageable number for the capacity of

our computer programs, and recognizing that dyads with little conflict

behavior on any dimension would group together, we eliminated from our

factor score matrix all dyads with standardized factor scores less than

an absolute value of 1.5 standard deviations on any conflict dimension.

This left 61 dyads on which to build a taxonomy.

17

The five patterns of foreign conflict behavior In 1963 are

orthogonal and can be assumed to form a Cartesian coordinate system

within which the dyadic factor pattern scores can be plotted. Each of

the 61 dyads has a unique place In the space of the five patterns and

each point Is a unique Euclidean distance from all other points in the

space. Dyads which are .near one another In this space have similar

profiles of behavior. Thus, the Euclidean distance can be employed as

our measure of profile similarity - of conflict behavior similarity -

for grouping dyads.

Two methods were employed to group the 61 dyads on their profile

similarity — the hierarchical clustering scheme (Johnson, 1967), and

factoring distances (Rummel, 1968, Chapter 22). The method of factoring

distances gave us six factor groups, for which we calculated and plotted

group profiles to find the distinguishing similarities of the different

groups (Hall, 1968).

Figure 4.1 lists the group members along with their factor loadings

to indicate the degree to which each member approximates the group's

center. The dyads are listed in order of descending centrality for

each of the groups. Figure A..? presents a tabulation of the similarity

characteristics for each group.

The group mean plots, with one standard deviation confidence

intervals, are shown in Figure A.3. The plots show the group profiles

across the means of the population of foreign conflict study dyads on

the five conflict patterns for 1963. The confidence Interval for a

group which is entirely above or below the population mean on a pattern

is taken as a characteristic of the group. Thus, group III in Figure

4.3 is characterized by extensive negative communication, low Incidence

of violence and moderately low violence intensity.

18

Appendix III presents a display showing the correspondence

between factor groups (Figures 4.1-4.3) and connectedness method groups

from the hierarchical clustering scheme.

19

Figure 4.1 DYADIC MEMBERS OF 1963 FOREIGN CONFLICT BEHAVIOR GROUPS*

GROUP I (20 members)

.85 Senegal - Portugal

.84 USA - Dominican Republic

.82 Pakistan - India

.82 Cambodia - USA

.80 Netherlands - France

.78 Belgium - France

.77 Morocco - Egypt

.76 India - Union of South Africa

.76 USA - Cuba

.75 Congo (Leopoldvllle) - Russia

.75 United Kingdom - Indonesia

.74 Guatemala - United Kingdom

.71 USA - Haiti

.71 USA - Russia

.71 Venezuela - Haiti

.70 Haiti - USA

.69 USA - South Vietnam

.65 United Kingdom - Somalia

.64 India - Pakistan

.60 France - Russia

GROUP II (4 members)

.89 Yugoslavia - Albania

.88 Albania - Yugoslavia

.77 China - Taiwan

.74 Taiwan - China

GROUP III (6 members)

.87 Russia - USA

.84 China - USA

.79 China - Russia

.76 India - China

.76 Cuba - USA

.61 Russia - China

GROUP IV (4 members)

.84 Malaysia - Indonesia

.83 Dominican Republic - Haiti

.66 Indonesia - Malaysia

.60 France - Brazil

GROUP VI (20 members)

.75 United Kingdom - Yemen

.75 South Vietnam - Cambodia

.73 Israel - Jordan

.72 Egypt - Saudi Arabia

.72 Syria - Israel

.70 Japan - USA

.69 North Korea - USA

.69 Israel - Syria

.68 Brazil - France

.66 Egypt - Israel

.64 Jordan - Israel

.62 Iraq - Israel

.60 South Vietnam - USA

.60 Bulgaria - USA

.59 Ethiopia - Somalia

.59 Somalia - United Kingdom

.58 Malaysia - Philippines

.57 United Kingdom - Russia

.55 Yemen - United Kingdom

.50 Lebanon - Syria

♦Behavior for grouping is from the first nation toward the second. Factor analysiu (principle axis, varimax rotation) of similarities matrix for 61 dyads. Group members have loadings greater than or equal to .5 on the factor, or highest loading on a single factor. Member loading indicates degree to which member approximates group modal behavior.

GROUP V (7 members)

.79 Venezuela - USA

.76 Indonesia - United Kingdom

.75 Columbia - USA

.68 Venezuela - United Kingdom

.63 Iran - USA

.63 Ecuador - USA

.56 Taiwan - Japan

20

Figure 4,2

GROUPING CHARACTERISTICS FOR FOREIGN CONFLICT BEHAVIOR IN 1963

GROUPS

Patterns I II III IV V VI

low 1 NEGCOM

2 1NCVIO

3 VIOINT

4 NBIOLB

5 NEGSAN

very Wigh

low low low

low

high

Patterns are

very high low

low

low

IV V

• • low

• •

• •

quite hiRh moderately low

very high

moderately low

• • moderately low

low

NEGCOM - negative communication INCVIO - Incidence of violence VIOINT - violence intensity NVIOLB - non-violent behavior NEGSAN - negative sanctions

PERCENT OF PRüFILE VARIANCE ACCOUNTED FOR BY THE SIX BEHAVIOR GROUPS*

GROUP I 27.67 percent

GROUP II 8.02 percent

CROUP III 11.07 percent

GROUP IV 8.56 percent

GROUP V 9.41 percent

GROUP VI 23.18 percent

*The percent of profile variance figures measure that amount of total variation among all dyads on their conflict profiles accounted for by the group. The percentages are derived by summing the factor loadings for a group and dividing by 61 the number of dyads in the study.

fMCl.l 0

'JOZfjT

SJi'IOLß

Figure 4,3 Mean ProflleB for Conflict Croups

21

.//• r i

t ■> ' •»

-U.

ZtajpEL

* 'h - u.

*—- J : ? i -— i -* 2 * vi r* fi

'jt^rCOM i

^^ rrw"

r/\ Cl/t-O «*•

>Zc Tf\. T /T* i

< *«I4BH ^

22

"SAiCilÖ

Figure 4.3 (cont.)

-s

' t'TCI MT

b/louPt

x JL

GttoupJO.

^4 -üiä. tSL ti

Appendix I

Par.e 1 DATA LISTING AND STATISTICS - 1963 DYADIC FOREIGN CONFLICT

Variable Format Code tor Data Listing 1963 Dyadic Foreign Conflict

nation code numbers to specify dyad* warning and defensive acts alerts and mobilizations planned violent acts overt violence discrete military actions days of violence negative behavior acts unclassified negative icts severence of diplomatic relations expulsion or recall boycott or embargo aid to rebels negative communications written negative communication oral negative communication written or oral negative communication accusations protests minor themes unofficial violence

col. 42 attacks on embassy col. 43 attacks on person col. 44 attacks on flag col. 45 non-violent behavior

*First three columns specify the I.D. number of the actor nation; col. 4-6 specify the I.D. number of the object nation.

col. 1-6 col. 7 col. 8 col. 9-10 col. 11 col. 12-13 col. 14-16 col. 17-18 col. 19-20 col. 21 col. 22 col. 23 col. 24 col. 25-27 col. 28-29 col. 30-31 col. 32-33 col. 34-35 col. 36-37 col. 38-39 col. 40-41

Appendix 1 page 2

PROBLEM 124 RUN 4 CORRECTED DYADIC FOREIGN CONFLICT DATA IMAGE

VARIABLE NO. NAME MEAN SE ST DEV SF, SKEW KURTOSIS

1 WARNDF 0.076 0.023 0.388 0.087 6.736** 53.513** 2 ALRTMM 0.069 0.022 0.370 0.090 7.30C * 63.060** 3 PLNDYL 0.127 0.034 0.561 0.095 5.258** 29.451 4 WARACT 0.101 0.018 0.302 0.024 2.655** 5.083** 5 DISCMA 1.210 0.528 8.770 2.430 8.942** 82.796** 6 VIOLAC 10.409 5.251 87.234 24.459 9.071** 84.791** 7 NEGBEH 0.547 0.047 0.787 0.068 2.074** 6.280** 8 UNCNEG 0.174 0.029 0.481 0.057 3.399** 13.638« 9 SEVDPR 0.127 0.020 0.333 0.023 2.255** 3.109**

10 EXPREC 0.116 0.022 0.363 0.049 3.740** 17.951** 11 BCTEMB 0.058 0.016 0.263 0.042 4.939** 26.252** 12 AIDREB 0.033 0.011 0.178 0.028 5.292** 26.194** 13 NEGCOM 1.489 0.200 3.324 0.554 4.971** 28.682** 14 WRTCOM 0.504 0.106 1.757 0.320 5.561** 34.551** 15 ORLCOM 0.362 0.055 0.922 0.204 5.976** 52.225** 16 WRTORL 0.591 0.083 1.381 0.243 4.867** 32.109** 17 ACCUSN 0.757 0.141 2.338 0.480 6.286** 44.579** 18 PROTST 0.337 0.064 1.058 0.198 5.421** 36.795** 19 MINTHM 0.207 0.042 0.701 0.117 4.852** 28.853** 20 UNOFPV 0.214 0.058 0.958 0.364 11.333** 157.582** 21 ATKEMB 0.058 0.015 0.249 0.036 4.494** 21.318** 22 ATKOFP 0.062 0.019 0.319 0.063 6.078** 41.252** ?.3 ATKFLB 0.072 0.019 0.311 0.058 5.387** 36.187** 24 NVIOLB 0.116 0.031 0.520 0.134 7.506** 71.449**

SE - SKEW - 0.147 SE - KURTOSIS « 0.292

* Significant at .05 level ** Significant at .01 level

DIMENSIONALITY OF ' NATIONS PROJECT Appenc llx I Page ! 3

Alphabetical Listing of Nations Code Code

I.D. Political Unit Abbreviation I.D. Political Unit Abbreviation 1. Afghanistan AFG 45. Korea (Dem. Rep.) KON 2. Albania ALB 46. Korea (Rep. of) K0S 3. Argentina ARG 80. Laos LAO 4. Augtralia AUL 47. Lebanon LEB 5. Austria AUS 48. Liberia LBR 6. Belgium BEL 49. Libya LBY 7. Bolivia BOL 93. Madagascar MAD 8. Brazil BRA 94. Malaysia MAL 9. Bulgaria BUL 95. Mali MLI

10. Burma BUR 96. Mauritania MAT 11. Cambodia CAM 50. Mexico MEX 83. Cameroon CAO 97. Morocco MOR 12. Canada CAN 51. Nepal NEP 84. Central African Rep. CEN 52. Netherlands NTH 13. Ceylon CEY 53. New Zealand NEW 85. Chad CHA 54. Nicaragua NIC 14. Chile CHL 98. Niger NIR 15. China CHN 99. Nigeria NIG 16. China (Rep. of) CHT 55. Norway NOR 17. Colombia COL 56. Outer Mongolia OUT 87. Congo (Brazzaville) CON 57. Pakistan PAK 86. Congo (Leopoldville) COP 58. Panama PAN 18. Costa Rica COS 59. Paraguay PAR 19. Cuba CUB 60. Peru PER 20. Czechoslovakia CZE 61. Philippines PHI 88. Dahomey DAH 62. Poland POL 21. Denmark DEN 63. Portugal POR 22. Dominican Republic DOM 64. Rumania RUM 23. Ecuador ECU 65. Saudi Arabia SAU 24. Egypt (ÜAR) EGP 100. Senegal SEN 25. £1 Salvador ELS 204. Sierra Leone SIE 26. Ethiopia ETH 101. Somalia S0M 27. Finland FIN 66. Spain SPN 28. France FRN 102. Sudan SUD 89. Gabon GAB 67. Sweden SED 29. Germany (D.D.R.) GME 68. Switzerland SWZ 30. Germany (Fed. Rep.) GMW 69. Syria SYR 90. Ghana GHA 213. Tanganyika TAN 31. Greece GRC 70. Thailand TAI 32. Guatemala GUA 103. Togo TOG 91. Guinea GUN 104. Tunisia TUN 33. Haiti HAI 71. Turkey TUR 34. Honduras HON 72. Union of S.Africa UNS 35. Hungary HUN 73. USSR USR 36. India INE 74. United Kingdom UNK 37. Indonesia INS 75. USA USA 38. Iran 1RN 105. Upper Volta UPP 39. Iraq IRQ 76. Uruguay URA 40. Ireland IRE 77. Venezuela VEN 41. Israel ISR 81. Vietnam (N) VTN 42. Italy ITA 82. Vietnam (S) VTS 92. Ivory Coast IVO 78. Yemen YEM 43. Japan JAP 79. Yugoslavia YUG 44. Jordan J0R

. . ■

Appendix II Page 1

IMAGli: ANALYSIS TABLE OF FACTOR LOADINGS AND -.VARIANCE MATRIX

Image Orthogonally Rotated Factor Matrix*

Fact :or Number 1 2 3 4 5 Sum Squares Over Variables 4.700 2.396 2.C48 2.177 1.925

Variable Communality No. Name 5 Factors

1 WARNDF 0.899 .0.157 0.075 0.092 -0.927 0.033 2 ALRTMM 0.875 0.161 0.051 0.065 -0.918 0.017 3 PLNDYL 0.220 0.079 0.403; -0.093 -0.202 0.048 4 WARACT 0.447 0.075 0.603 -0.262 0.038 5 DISCMA 0.924 -0.049 ( 0.952) 0.017 0.118 -0.040 6 VIOLAC 0.904 -0.058 0.938 0.023 0.140 -0.043 7 NEGBEH 0.687 0.092 -0.036 (0.818) -0.083 -0.038 8 UNCNEG 0.476 0.107 -0.029 (0.674) -0.068 0.071 9 SEVDPR 0.270 -0.156 -0.020 0.431 -0.221 -0.102

10 EXPREC 0.311 -0.005 -0.060 (0.535) 0.143 -0.032

11 BCTEMB 0.255 -0.003 -0.037 0.493 0.090 -0.050 12 AIDREB 0.142 -0.010 0.236 0.288 0.035 -0.045 13 NEGCOM .0.956 (0.976) 0.005 0.022 -0.011 -0.041 14 WRTCOM C0..772:. (0.858) -0.021 O.OS? 0.179 -0.P2'5 15 ORLCOM 0.491 (0.687) 0.030 -0.102 -0.079 -0.038 16 WRTORL 0.687 (0.793) 0.028 0.071 -0.224 -0.037

17.. ORLCOM 0.833 (0.^08^ 0.036 -0.068 -0.020 -0.035 18 PROTST 0.424 (0.571) -0.073 0.245 0.181 0.017

19 MINTHM 0.679 (0.799) 0.070 -0.062 -0.177 -0.033 20 UNOFPV 0.664 -0.C35 -0.003 -0.056 -0.012 (0.812)

21 ATKEMB 0.105 -0.028 -0.000 -0.046 -0.165 0.274

22 ATKOFP 0.599 -0.036 0.012 0.006 0.001 (0.773)

23 ATKFLG 0.562 -0.044 -0.014 -0.050 0.059 (0.744)

2A NVIOLB 0.063 -0.043 0.010 -0.045 -0.241 0.025

*Varlmax Technique, Loadings greater or equal to an absolute value of .50 shown In parantheses.

M OS

M

I 8 W

i r' v r * A* M' f '

M o a^ ii rl Ft fi lo

•* Ct O P^ Kt r» CN vo m o

r- •j

■ i • i

o es

1

m •

m en XJ C1

lin f-i in vo p O o C i • • • • till

Lj ^ sO vC p o o o

g |in p

ON »i

■-I

m vo so o> fO l^-

■ • •

CM 00 O CO ON SO <} >£

• • • V

f^. »H C4 0> O oo m >o r-< so

IIII

j<* CM NO vO o o o o 1 • « • * IIII

Ko CM NO rs p o o o III 1

|»n <j- en in o O O c

i i

len lc | «

1

H O •

1

tn o

-O iH

in ■-4

vo .-• in r». r» ON o^ in in r^ in ir,

i I I r in m -j m p c o cr

i

o *-*

t>. •* •» «C O ON 00 0'> 00 vO »"» ON m r~.

i i > I ko m r- oo o o o o

1

o en r-4

IIII 1

CO

m 00 SO tH

!»-< oo en .* ^r o m jc o c o o O O i • • ■ • • • «

1 i I i i o »-« «n m en f* »n O O O O O H O

f>7 <n -H m o o c o • • • • IIII

m n i-H NO o o o o

o o •

1

o o

CM i-l

i-t r-l

r- .-H O"- r» r^ p-4

1 1 11 1 lo en oo •>» "-i co en |o o o o o o o

IIII

*» «» en en O O o O

1

vO o

O iH

00 <H <^ 00 r~ r-l fH «H

11 i r r rH ON NO 00 r>> O CNI

U-* O O O O rH J

i r r r NO •* CM O O O O i-l

1

NO o

ON

r«. n m c^ f^ 00 t-t r-t r-< f-l

1 1 1 1 1 1 l] LH NO r* vO r-l <M M r-t C O »-1 O <M O

r r r i r". O -^ r-4 o o o o|

in o

00

•C vO CM »T »D r^ «j\ m co r« («4 o

ON O CM fM CM O O O <-* O "^ O es fH

1 " 1* ON in m NO o o o o 8 r»

• 1 i i r i 1

00

00

1.0

0

CO "ij" LO «vj- fO ^ O O O O O i-t

■a- en IH m >T -T enl o o o o o o o

CM en es «nT o o o ok

N 3

«o

1 1 t 1 1

•» *T «» «.t rn Cv O O O C O f-l

1 1 1 1 1 1

en en o en «M •* CM! o o o o o o o

i i r r rn CM es «n K o o o ok 3

in

<N o r« r~ •* m

1 1 1 1 1

n en r- r-t en CM O O C \-< O -I

ii iiii

^c CM m <n o oo OH o O O »H M C «Mj

i i' i i •M >» en -^|f D o C Ok

1

3 •*

<r c o «^ U^ ^ (N rM

II il

m iH rH 05 «A O D O O O O O

i i I NO CN m m o NO eol1

OOO-H-^OML

1 1 ;> m -» .HIO 5 O o op

m

1 1 1 1 1 l 'if . . . . .

o- "■ \D t-4 (sj -.; r-« c o

H ON en r» in iH -1 o »-• o o o

oo »H m r-i vr rH vole •-4 O r-i «n I-H c evik

s r* o m ICM 3 -< O Oks «s

i N f' O CS -• f-l o o

1 1 1 N ri \u "t ft f >\ -« r-l fH O O O «HO r-ICOr-IOtMlC

* B ' " 1 *

t i

n ND <M en ICM ^ r-l O C (cs

i-4

55

SS n 1

►J H < J M 5 s ^ ►J 3 ^ <

<*■» ^ »n \JI

1 1 |

H u « o ^ pal b u- cv, UJ Ui wi •< IT: l-i cj H a: u ü £• <"- c a w ä w ,•; u MI a: 5 '/J ui c3 •<

fo ft, W g fj Csj fi •* m NO r^ « c>|j

1

J « OS ü .-

■} < < <h

5 r-t CM <n L vl ^J rVi fvl W<

r

u X

at u « (A u C

I s o

a

«n

CM

«

, ..■ •.■.,:•■..• ■•■ ■■-..:■■

CORRKCTED DYAÜIC FOREIGN CONFLICT DATA: RAW DATA LISTING

Appendix It

OLUJl ■;".: rHOOOCuOOiüOOüOÜ004030l000200QO.OOOOOO -..'. ■■7-y:i.r ii'iovcODnooo.nooioooiooooooooooocoo .. ' '... I ■)■- „ u : oO'" . C 0 1 00 1 JOnOüOOOOOOCüOOOOOOOOOOG J.. > ■!> .v''OvODCOCO "'OOOvüOOlOOOluOOOöOOOO.lOOlO Jj^u 5/'. .OOGOCOOOOOOOuüOOOlOOOOülOiOOOCOOOOOö 0M37.ir0C.^ JOr.OOOlCOOlOOOOOOOOOüOOOOOOCOOOOOO C. 5f'^,' "uo. C JOOOOOOOOOÜOGOOIOOOOOIOIOOOOOOOOOO ■J '■ ?; "_■,.. .'.ü"'v002oiooiooaidooooiöiooocoooooo J. '^bi- :.(.■ . , 0:000 rrOOüCOOGOOüOOüOCOOOOCOOOOGl :'./'.. 1^L..U , i-jj üOOlOlQOOüO'JOOOOOOOOOOOOCOOOftOQ OJ70 IOCOOOOQOOC000COOÜOüOülüOOOO1000100000000 Ou-m^OOv'OOOüOüOOOOOOOOOOOOOOOOOOOOOOOOUOOO 0>02i>ll(\ .VJO00C0OGO00C0ÜO1OO0O01OOÜO0C0OOOO0 OL 'i.)7 v.v- -: XH 0030^00000000000000000000010010 0; •7-..-. • . J jcr ocoiooooooooiooroüionoooooioioü v. ' i •:: JO( o )0iüa>?jO0Gocoü200C2üO02üO0loooooo OO'OT^rüOüOjOOOOOOOOOOOCOOOOOOOOOOOOOOCOlllOO ono^b'joeooooooooiaioooooocoooooooooooooooooo Ü HO7CC0000ÜO0CO0OCO0O0O0010001000ICOO.CODOO.OQ .ir 7^. d.luO'.eonioioooooocooooooooooocoooooo Ji ir7 • :. r, ,. J...OO1 00 1)30000003000102020000000000 JliJü,-'!.! u 100CvGO 1001OOOOOQOD00^0000000000001 OUO 7b COCGOC 00000000000000 3000003020000000000 :iio7):coooooooooooooooooooooooooooooooooooüi Ji-fGOJ^ÜÜHCjOJQOOOOüQOOOOOOOOOOOOCOOOOOOHOOO Jl'iJJ'JvCL JJXOOl^lOuOOOOOOOOüOOOüOOOCOOOOOO 31'tOl ^iOunjcnCQOiroOlOC'XlOOOOOOOOOOOOOOQOOOOO Olv'V- r^ u.Ou^vOOlOlOOOOOGOQOOOOOOOOOOOOOOO.O-C 0 1 '>;) 1 ? "i n2 1 ^ 1 SOloocooooooooooooooooooonoooooo Jl-y 2- "LCÜOOOJOOOOOOOOOOOOIOÜOQOIOOOIOOOOOOOO Ji iO ■i» kJLwCOOOOOO?OOOnoOO503O2OOO2O2OCOOOOOC )i . ' HCuOtJCOOOCOOOOOOOOOOlOlOOOOOOOlOOOOOOOO . I > 7 , -OCvJD-OOOI00000001915020217020lOCOOOG J1-.>J7V\,CV0UJ:0000 00000000 10 lOOCOO'JO IOÜOOOOCG 0 I'r> 75 "L;OC 0000 :O0O0OÜOC002211020920000AOOOOOC OiW ' 'CCO0a00G0000POC0OOO3010200O20O00OO0OO0 01->; 3c CtotOOOOL'COO 0000000010000010000000000J3D D 1 o.' 1 >' 00 11 J 1 s 100 DOOOOOOOOOOCOOOOOOOOCOOOOOO jioj-t i~' ^wOuo^rooiJioonoooiooocGioooioooioiio .1/ '1 ' .vCfJJJ V.G010O1C0OOO0OOOOOO00OOOCOOOÜOO 01 M 7 J^CC 0000-'0000 0000000000000000000000402 10 01 7f« 7 70000«: 00 3G0 30000000000000000000000010010 }1'101901 ouUJO0000l00l.0O0O0OQ0000000000000000.0_ .•i< )2.- M - t o..v,',}onoi inoiocoooooooooocoocooooooo u-«. i v'or- , iio.rooi 00 10000000ÜC ooooooo ocoooooo . i ■ •:• r\ .«..r-j jnrOGOOOOOOOOlOOGlOOÜOOOOCOQOOOO Ji901 rCGOCLviOOOOOOOOOOOOOOlOOOlOOOOOOOOOboOOC oi ^J200ucjüoooooooooooqooioooiooooooorooooüo jHo>c0CCoOwO00O00n0Q0OOO020200OO00020C00O.Q0.0 jk'iCiy. < v 0 '..COv GOO0O0Ü0000100010O0OGOOGO0O0C0 Jl »0 r- I 1011 J?X/20Io0000002606071320C2ü6000000 jl /.m\00nujOwuw00O000OOC20C02Ü00iOO0000OD.0-C ü^JOr..' '.,000030000 I C0010000000000000000COOOOOO 32^1.) ^Cuu- uJO'jOOOinOOlOOOOCOOOUOOOOOOOOOOOOOO

Append jr. II

0«?>!Ol4COOCÜüOOOp0100lOOOOOOOOOOOOGGCOOOOOOOOO 02^0 3 3^00000000000000000080002060400020110010 022O340OOt'O0O00O000O000000OO000000000OC0I00 1 0 022u7bCeOOOoOOOOOOOOOOOüOO100000101000COluOlC 0220 77000000000000000000000000000000000010010 023019000000000001 oo locfooooooooooooooocooo 000 02303300000000000100lOGOOOOOOOOOOOOOOGOOOOOOO 02 30 ^000o0ÜOCO000C000COOOO00OO0O0O00OCO2üil 10 024039000000000001roooc0001000100010000000000 0240411 IC110100100000000003010101020001000000 024063000000000001001000000000000000000000 000 02406i»00041020020000000000100000 lOiOOOCOOOOOO O24O69CU00O0OOO0O1000000005030101030001000000 02^0 72 0000000000010000100000000000000 00000 0 00 024097GÜ00ÜÜ00000000000000101OOuOOOO100000000 02^019000000000001001000000000000000000000000 026063000000000001001000000000000000000000000 026072000000000000000000001000100000000000000 0260 75000000000000000000001000001000100000000 02610100000000000101000000 30002010 3000 3000000 02 800822CC00000OO0000000000000000000000000000 02ÖO4J0ÜOO00O0OP000OO00OOO100000100C100000O00 028066000000000000000000000000000000000000001 028073C0000000O0O1000000006050001000600000000 028074000000000001010000000000000000000000000 029002CuGüOüOOÖOOO000000001000100010000000000 U2 9015OOOuCOOOOOOO000000002000200020000000000 0290 28000000000000000000001010000000lOOOOOOCG 02907300000000000101000000000000000000coooooc 02907400000000000000000000lOlOOOOOOOiOCOOOOOO 029075O00000000Q000O00OOO0302OP0100020100000G 029U66000000000000 000000001000001000100000000 03C190CÜCC0000000ICOIOOOOOOOOOOOOOOOOOCOOOOOO 0302SGOOGOOGOOCpOl00001C00OOOOO00000O0C000000 03O74Ü0C00000OÖ000OO"000OOOO0000O00ü0O00Ol) 000 031073000000000000000000001000001000100000000 032019000000000001001000000000000000000000000 032074000^00000002011060000000000000000000000 0J20750Ü0000000000000000000000000000000011001 0 3 30 1^9000000000001 COlüOCOOOOOOOOOOüCOOCOCG0 00 03 3022 000000000001001000004000004040001000000 0 3 30 52C0000C000O010001O000OO00OO00Ü000CO0OO00 03 3O75C00OO000000_20O0200O01QO0lO0O1000C01001O Ü34019^000 0 00000010010000000000000000000000 00 0 3 bor>000 0 0000000000000000 1000100010000000000 0 3'3042 OOCOOCOOOOO1000 100000000000000000000000 036015000000000001010000017100106080703000000 036057000000000001000100005010004010301000000 0360 72 000000000002001010000000000000000000 000 03 70 740001 10100101010000POOCOOOOOOÜOOOC022-.,21 0 3707JCOOGOOOOOOOO00000000100000101C00C000CC? 03 70 94210310 30030 3011000002000 2000)00000 2110 1 0380 75C00000000000000000000000000000000020 200 0 390090000000000010001oooooooooooooooooocooco

Appendix II Page 5

Oi ^y^OOuuüODOOOlOOOlOOOOOüOOOOOüÜOOOCOOOOOO U3 K.',\ lir, .;i.'0 "ivOOl'-OOOOlOOOüOOOOOOOOOOCOOOOCO .»■1 * 'r>t. .„" ;. j^O^uOl OlOOGOOnOOOJoOOOÜOOCOOOOno L v. 7 ^^ . :. OJJOOJ-OüC^LJO^OOO?U10 3000 100000C

0 iSüT^JOv/JO J030OO0000C000G1000100010OO00OOO00 0 i JuT'j^OO JGUOOOOOOCOOOOOOOIOOOIOOOIOOOOOOOOOO

Oi l.^'. JoCo JCODOOQOC'OOüOOUOOOOGOOOOOOOOOOOOOOI.

Ji 1 •U'i'i.L 11 : *0J3OUt.öUu0u001ü000O1010O0C00ÜO00 'j'( I ..orJ ^ vu^io^ ^bO1' GoCOuOOAOCUl03C JOOO^OOGOCC ui r^ifHiGC . ^:^:30001r'?001Co0000o0ü000C00C000000 0<*<>M JC00000 00000IQCOIÖCOOOOGOOOOOOOOOCOOOOOO 0^ iOU.OOCOOJOOOOOlOOOlbooOOOOOOOOOOOOOOOOOOOO 0<t i »7 jOuOwüüOOOOOOOOOCDOOOOOOOOOOOOOOOCOOOOOb us'. i?i : wOoouO JOOOO'. OOOnOOOlOOOOÜlOüOOOOOOOOOO V.s ^1 rcGvlOiOOiOiu'COOOlOOlOlOOOOC lOOOlOOOOOO o'o'io 7'>COOuOCOOCü000000O000lO001O00lC0O0000OOO 04'30 7-JrjO^ I 3^00^0000000000 30100020 3000 1000000 Ü4MJ4 iOÜÜCOOOOOCOOOOOOOOOOOOOOOOOOOOOOCOlOOOO OW06»1002 10 30020000000000?OOC00202Ü002000000 ^'iWj7- Oi C. J JOMOOOIOOOOIOOOOOOCOOOOOOOOGOOCOOC J v V/ ^JL •: TOOOOi )001COOOOOCOOJOOOCOOCOOOOOO j J.;. 7 .GwOiiUGCjOCOC }OOOOCOOOOOCOOCüCüOOC0100l G O'J 10 rj^üCCOjOOOOOJGOOOOOOO 1000001010000000000 0^10^2:00000000000000000000000000000000000001 0,3^02ü J^.\/OO000ü02:0ü020O000CO0OOO0OO0000QQOC 0 J ^ 2 iCGO'jcJOoOC 00000000001000001OOC lOCOOOOOO O^^W^CC JOGCOGOOICOIOOOOOOOOOOOOOOOOOCOOOOOC 0-J70 3O >..cOOüJ0003C0031000401C00302010200000C 0,>bOl'-)C:J000ü00000in010000000000w0000000000000 01JOJ7 ^LOCOGOOOCOOCOOOOOOO1010000000100000000 Ob^Jl ^vOOoOOOOOOluOlOOOOOOOOOOOOOOOOOOOOOOOC 'Jo,..l -"X'-M'JOO^OC OKOIOOOOOIOOOOOIOIOOOCOOOOOO 0('OJ 7 iOOOO JuOOOOOOuOOOOOOOlOOOOOlOlOOOOOOOOOO joi: ^♦OJOOOCOOCOOOOOOOOOOOIOOOIOOOIOOOCOOOOOO

0ül09Hno0030ü00001010000000000000000000000000 0o201JCJOUOOOOOOOOOOOOOOOOIOOOIOOOIOOOOOOOOOO

0o/:7'iC0000000C0C1000ll,00000CüO0000ü00000000Q 3oi07iC-CGOOOOOCOOOOOOOOOO1000 100010000000000 (■(. i'W ^GOtuOOOOOOOOOOOaOOOlOOOlOOOlOOOCOOOOOO j<. Vjti/''OOOroOO0OC0 ?OOOCg00100G00100010COOOOOO. 06310OOÜCC tJOOOOCOOOOOOOOOlOOOOOlOlOOOOOOOOOO U(.^J2VXJO 1020020000000000300010202000 1000000 0N(.0 3 000t.UOOCOCO'.OOüOOOOOOOOOOOUOCOOOO 11000 QüOJ7 .' ^Ou '0 T0CO0Dü0ü0U0OO0000O00O00O0CO10102 ji.hOV/ 'GO^OjOOOoCOOOünOOOOOOOOOOOOOOOOOOUOOO Oo ?.: 7 iOv'. u^OOOCOOl "lOGiOOOOlOlOOOOOOOlOOOOOOOO {)O/0 7-.000000000000000000001000001010000000000 Of)-)0 2sruOOOüC000000000000020001010?OOOCOOOOOO 0^041 iOOllOb00b00C0O00OOO300O0O3O2O00l0000OO w n7l ^,'i iCOOCOOOOOOCOOOOOOOCOOOuCOOOCOllOOO Jo t ' 7^ \;CL .'.moOOOCOOCOOOOlOOOOClOlOOOOOOOOOO

Ou »0 7J^ jv. )vCOOCOOuOOÜ )000100GOC1010000ÜOOOOO jl. J11:OCuOOü00000000000004000202030000000000 0 71064000000000001000100000000000000000000000

Appendix 11 Pag« 6

OTiJtl .:i L'iCJOOOOOOt 00000)020001 0 1 OuOOO 00000 00 o f. v n'-.v: jco: ^ ooo JOOOOCOO iooo i ocfcco JCOOOCOO j ^ J'* j n ( ^ joonvoo ):OüOOOOOIüOOi OCOUOODCOOOOOO jfP'n'j ^».i ■" JOcanOQ 100000002000 UlOuC 00 coooooc 0 7206 7COOt )C0OC000nOOC00ü0?O0OlO10OO0ü0OOü0O0 UZi'OT^roOOODOOOOOOCOOuOOOO-JOOOPOlOlOüuOOOOOOO 0 7^0 7rjOuOui.JOOOOOOCOOOOOCO 300020 101000COOOOOO J7J.J •; '" .' oO.'OTCOOOCOüCOCOOlÜOClOOOlOüOCOOOOOC J7.. Mcf'-CO i.r ."»ODOJ-'-OOOOOOOl0100000ICC^3000000 . 7 v isrjC UMJO lOC02'..001Ul01 H IOÜ020902C 2000000 17 J02.r?'."0n0OÜCC0000000O00002u2Ü000C00100000000 07 s'. Mi''.:Güu000000100000000202000uCOC2COOOOOOC 07 JOiti.'.f.c 111,102100000000001010000000000000000 J7 < ; w.\ ..CULO n 0C0O0O0O0006C300O302O100011000 )7 n'*3::c :JL; o^jooonooo^ooopoiooüiojoiücoiöooc >7 >'(,'<'.. juOJ'i'M.OOO: OÜ0O000202CO0O0 10000000C0G 0 7 ^.■'71',uC"Jj:OJOCOO':Oö000001010000000CüCOOOÜOt J7 10 7««':CO JOCO'^OOOO^OOOOC^O 1010300000200000000 >7iJ 7t.rouoOOOC00010lOOOOO?&121003l40605000000 J7.-. "c \ ;. vJ''^;00.i OOCOCOO JG2G100U003O100O000 ..'7'-MM ," .lüOOuOOJtOOOOCOOlCOCJOlOOOlOCOOOOOO J7'»'J2VOIJCGJJ0üUOO1O0C1OC60OOOÜOO0OÜüO0OO00O0C

O7'»J2<»'")OÜOC00OO00üC0Ü0O00O 10001000 lOOOCOOOOOO O7<.O2eCüOCuO00OO0lCO0000004C0020204C000000000 07<.5ii:o?iCüOOOOOOOOOOOOOOOOOü00000000002000l 'j{<<: > / 1 i:( *C0 ■0003.,30oO000?u0C0C20001C coooooo j7't; 4;i\ CC^JOuOOOOCOOCOOCOlCOCOO 1000000000000 O7'iOti,y"-rijCJ0Ci 00 ü }COC 00 :C 1000001 OOOOCIOOOOOO 0 7^.0 72 ^LcOOOCOOO3000000boooooooooooooooooooi 3 7^0 7.VJJ» uOJO300030O0CO0006O50001010500000000 0 7^:7 'jC.bZ . C „ 00 JOO O COOOOOC O 30101C l O i Cl 0 2000001 j 7'i" r- " i, t'l :'2CL2OOCOüOOOO0OO000OO0ü0OOCOOOO00

O/«, n<{ ^..M O^CC :0030000000010ÜC00100000COOOOOO 07^101r 0010100102010001000000000000000000000 0/ jOO^ruCOOOOOOOOOOOOOOOOlOOClOOOOOOOOCOOOOC 0 7b0 0S',CCu0oOGOO00.")O00O0p0O0000000O00C COOOOOl 0 7t>3Ü7COC 00300000000000000101000000ÜPOCOOOOOC .TT'JOOV' uuOOO^OOi: 100000010 lOOCOOOt'100000000 )7l>jl 1 ,', L^OO^OOOlCOClOOOOIOOCOOlOüCOOCOOOOOO • 7 j.l.- :0I..J:^0.I0O00COOO0O0C200C3O?01C1O00000O0

J7'>ül i^;'i.0 3C0CO01COOOCOC0 3ClOJO2OG01O0OOOCOC ■j7')',irj',Oe..OjOGOOOlOOOOOiOOOCOOOOOOOOOOCOOOOOO 0 7t>'jiy',LüO' CÜ00O02OOOO2C0JJO1010101020 2000001 jl > S.l'j.rjhj)' OCOiriOO'iOOJOCOOOOCOOOOOCOOOOOO J/->' ^<"^....'C00ü3000'>"211','ü00O; OGOOOÜCOOOCOOOOOO

■ 7'.''J v^Cl i.OOO^CCOOGOOOOOOOlOOClOOCOOlOOOOOOOO J7 j'.V<.nuOÜO jOGOOOOOOOOOOpOl 00010000000 1000000 0 7 tO 2LCOCC0O0OO000COO000OO1010000000000000000 0 7'333il IGcOüOCOOOJC 11000032010001000200000000 j t'jj \^r s. OCUOJOJOOICIOOOOOOOOOOOOOOOOOOCOOOOOO

;7^0 i) Vo^ .• JODüOO 1 10000000000000000000000000 . I >o ,1 *yj(<K 3003». 001 :iocoocoioooioooooooiocoooo 07>0AK0Cv0C0i.0000CO0CG000l00Ol000C000000O001 u7bOA^OtOO102032000000000010100000lOOGOOOOOOC

Appendix II Page 7

■v/v. w '•'.''■ 00 J 10001C10ÜOO'')Ü100000 IOÜüOOOOOOOüO

>/' ",c ■. .r-o. uooii loooc ■'OOOOOOUOCCOOCOOüOCC j/ (»«.^r >.■.•. :■ ^Cii • i jooojüococ^ocoocOv^roooocv; O/'J.'O'.:. . : ^JOüöC'OiOuüCOOOiCOCOJlÜCuOuCOOOOCO ) 7 r>v(.M-,.,,• v. OOOJOOO JOOCOOOUIOOOOOIOCOOOIOOOOOO jr.- ?/ L. ;.!LJ -cooIC IOCOOOOOOOCOOOOOOCOCDOOOOO <jl .. 7 i'ci.t J^OOCC-O^^aoHOC 1305000801lOOCOCOOOC . r. ^.'. . .„' uOOO'COOCOOClOOOOOlOUClOCnOOOOO .^ fj ■' ,.d0 U COO'.COO^OOOCOOOOCOJOOOOCOOOOOl )/. •'! ■ . .■ OJ-^OOO'-OOCOOnOlOOCDOlClCOCCOOOOOC

J7'JCH/ 1 l( ■JOJOOCOO? '30000 )08ül0 0501020 100000C J/MjMt.-CC'cO.'Oi'OCOlOlOCOOOOOüOOüOOOOOOOCOOOOOO J/r,u7'?'Cco03nncOOQCOOOOOnOOOOC'OOCOÖOOOCOllOCO j7 I \':\y' ■jO\Uülir0l0,,C0ü<»ÜOC30lülOO0C0000O0 .)f f :< . .v.v.vwOOOOOl 000 lOOOCOOOOOuOOOOOCf 000000 J { ( > i •. *.. .■ 0 JUOOOO^bOlLOlOuoOOCOOOOüOOOCOOOOOC 07707H's0Ot00O5C0OOC000OO0OO0O0O0O0O000C02lllC 0 7 70 7f)C:0uJüO0CO0JCO000O0O0O0OD0O0000öCUO330 C7 /^t.fJLcuJOruOOlClOOOOOOüOOOOOOOOCOOrOOOOOO :/■.'*■' •. .'.: 0 OOOJ'OOOOCO'IKOCOOIOOC^OCOOOOOC

j/>i"u'> i.i^' .ÜOjOOOOCOOOlÜOOOOiOOOOOCOOOOOO jl.<: 7<i>\, H.101 '0201^10000000000000000000010010 //;. I'-i-u'iuOOOOOOOOOOOOOOOO 100000101 ooocoooooc OHLJ 7^0.0COO00000C',ü00O0OO30000O30200U0000000 OH K) 7'>:0Cui>000COOO0O0OOCOO^O101OOO20O0CO0O00O J*. i i 7 >' jOu-. JOOuOOO^OOOOOOOiOlCOOOOlOOOCOOOOOC 0.^ I . b"'). i u .cD ^OUOl "OOOOlOOGOOCOOOOOOOOOOOOOCC jf.^'ii;. ■ ^i.^..o?o.",oocoooojr,o.i;ioccxoocoooccc OJJ^O^Ö^OOLOJOOOOOOOOOOOOOOOOOCOOOOOOOOCOOOOOI OH-.'O rjoioeaooocooooooooc 00^020101020100010010 DLi JOb i'OCOOCO 'OCOl J010O(.000000000Ü0C00C000000 J--L )-^i\:CK \ ,) y COIOOOIOOOOOOOOOOOOOOOOCOOOOOC Ot .ct. ■ (v H OJCOOO-'OOtnoOlOOlOOOOOlOOOCOOOOOO OfJoi. 7.. ' ;. lvJO:oO92',2OCOOOOOO0O0OOOOCOO0OOOOOC 0Üo ) 7'. "ro: 000 30001000100000000000000000011000 oi Mo lOuoiooooooooooooonoo1000100010001000000 0 )jlO iC.cCJO00OO0QC000OOOO2O10001Ü100000.0O.0OO O^H U l j .'; 1,? K;101'*01ÜCOÜ0^01C003Ü20CG?00000^ ■J »'V 'fii! 1 • ;.>.,( ;oro 1 cou0000000cooco:000coooo01 J ' . 7 . i.'Ct .}0'.ICO00J0000üOOIO0COO1010O0C000000

0 ^.W/ ^CL'.U JJOOOOIOOOOIOOOOOOOOOOCOCOOCOOOOOO

j r/'.'i )r .COCOOOCCOI'JOIOOOOOOOOOOOOOOOOOOOOOOOO 0W % \\ COOO^OOOJOIOIOCOOOCOGOOOOOOOOOOOOOOO OW, .' ■ i c .0 UOOO'-.OOOOOOOIOOOOOIOOOIOCOOOOOO

' / ". J ; ' ... LOJOOOIOOIOOOOOOOOCOOOOOOOOOOOOOOO

'.: 1 r: 0 < '.v coco )C00i :OOIO:OOOOOCOOOüOUOüCOOUOOO 01.') 7/rv,i-'00Ci0r0001 JOOGIOOOOOOOOOOOOOOOCOOÜOOO I., j^h JOGOC ^OOOOOO^Ol 1110000000000000000000000 1. 1026'- .OL'J000000000000^002010100020101010100 1 .1 r-11'liJlO HOl'OUOOOuluCOlOOOlCOOOOOOOOO L ji A7'>' .i:0'".O'?(,'L'0OnO00OCO01OO01O00iCO0CO00OC0 1..".) 7.-", <;. jjf. JOOOI .lOOOlüOOOOOCOOOOOOOOCOOOOOO 21 iOu ^t.CuOOOJOOüOOOOOOGOOOOOüOOOüOOOOOOlOOlO 2 1 J J 7:' OC 000000000 1 nOOÜOOOOüOOOOOOOOüOCOOÜOOO

Appendix II ?age 8

JT'AIC.: 0c^üi'<C9o0OOU000Oüü00ü0üOU0üü0OU00000l;0 HI v. "ct. IM )üioo ,ooooopöocoonooooooor:oooc oo Jf>'M'« f ' .TL 1 j jn .?o J'iOOOOOOOOOOOOÜCuC'UOt.'rOOüOUO 0J4J7 i' LwlOlOvlOJOOOÜOOOOOüOOOOOOOOCCCOCOOCC ObijR^'.vu^jüOJoooicooooicooüOcoooooooocoiocoo ÜOCJJGDjOOOOOOOOOOCOOOOOOOOüOOOOOÜÜOOOioooooo

Appendix III

Page 1

Part 4 of the report presented six groups of dyadic conflict

behavior for 1963 and Indicated that we had used several techniques for

grouping the 61 dyads on their Euclidean distances. The factor groups

were selected for presentation and discussion in the body of the report

because they seemed most reasonable and straightforward. We are not

completely satisfied with any of our techniques for grouping and are at

present working on a research report to explicate some of the dif-

ficulties encountered (Phillips, not in references, forthcoming). We

had thought that hierarchical clustering scheme connectedness method

groups would be relatively easy to interpret, but find that it is not

easy to choose groups from among the sevc/.y or so given us by the method.

The hierarchical method builds groupe by relaxing the requirements for

grouping similarity as the technique works from the individual dyad to

a grouping of all dyads. Three levels of similarity were selected and

plotted to give a visual comparison between th3 factor groups and the

hierarchical connectedness groups.

Factor Croup Dyads f INS-MAL • AUSA-VTS«

MUNK-INSr /,«USA-HAIt

/^UNK-SOM» ; AVEN-HAI< ,

y /^GUA-UNK»^

Hierarchical Clustering Appendix III

/ ///

Factor I

Factor VI

Factor V

"actor IV

Factor III*—«^

Factor II*

IND-UNS♦ MOR^EOPK

^BBL-FRN^ ^»NTH-FRN«

^•COP-USR» -•USA-0ÜB»

-J» CAM-USA» _ f iro-PAK^ />• FRN-USR»

UNK-USR- ETH-SOM^

>LEB-SYBr ^EGP-ISR^ ,*ISR-SYR-

yKON-USA«', > VT3-'JAM*

UriK-EEM*7

ISR-JOR//- ÜOP-SAIJ// / sra-isRA' JOR-ISR/

-^IRQ^ISR« SCM-UNK« JAP-ÜSA« BRA-FRN« MAL-PHI- HAI-USA« VTS-USA« BUL-USA.

^•ECU-USA« *IRN-USA« •VEN-ÜNK. * COL-USA- *CHT-JAP« •PAK-IND«

PRN-BRA« YEM-UNKv SEN-POR« USA-DOM' USR-CHN»

^IND-CHN v—^CHN-USR

-v--«CHN-ÜSA ■~V USR-USA

\*INS-UNK * USA-USR - CHT-CHN ^ CHN-CHT - YUG-ALB

V-- -ALB-rUG > CUB-USA \VV DOM-HAI

MAI-INS VEN-USA*-

\ v. W;

/// •;//

/ /

/

i m.

'

REFERENCES

1. Ahmavaara, Yrjo, 1957, "On the Unified Factor Theory of Mind," Annales Akademiae Scieniarum Fennicae. Serial B, 106, Helsinki.

2. Hall, Dennis, 1968, "D.O.N. Computer Program Profile," Dimensionality of Nations Project, University of Hawaii, Research Report No. 14.

3. Hall, Dennis and Rudolph Rummel, 1968, "The Dynamics of Dyadic Foreign Conflict, Behavior - 1955 to 1963," Research Report No. 14, Dimensiona- lity of Nations Project, University of Hawaii.

4. Harman, Harry, 1960,"Modern Factor Analysis," Chicago: University of Chicago Press.

5. Johnson, Stephen C, 1967, "Hierarchical Clustering Schemes," Psychometrika. Vol. 32, (September), pp. 241-254.

6. Philipps, Warren, 1968, forthcoming, Dimensionality of Nations Project, University of Hawaii, mimeo.

7. Rummel, R. J., 1965. "A Foreign Conflict Code Sheet," World Politics. Vol. XVIII, January 1966, No. 2.

8. , 1967, "Some Attributes and Behavioral Patterns of Nations," Journal of Peace Research, No. 2, 1967, pp. 196-206.

9. , 1968, "Applied Factor Analysis," Evanston: Northwestern University Press, forthcoming.

10. , Jack Sawyer, Harold Guctzkow, and Raymond Tanter, 1969, The Dimensionality of Nations, forthcoming.